Principles and Practice of Clinical Research [4 ed.] 9780128499054

Citation preview

PRINCIPLES AND PRACTICE OF CLINICAL RESEARCH FOURTH EDITION Edited by

JOHN I. GALLIN FREDERICK P. OGNIBENE LAURA LEE JOHNSON

Academic Press is an imprint of Elsevier 125 London Wall, London EC2Y 5AS, United Kingdom 525 B Street, Suite 1800, San Diego, CA 92101-4495, United States 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom Copyright © 2018 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-849905-4 For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Mica Haley Senior Content Strategist: Kristine Jones Editorial Project Manager: Fenton Coulthurst Production Project Manager: Kiruthika Govindaraju Designer: Matthew Limbert Typeset by TNQ Books and Journals

Contents List of Contributors Acknowledgments Preface

3. Integrity in Research: Principles for the Conduct of Research

xiii xv xvii

MELISSA C. COLBERT, ROBERT B. NUSSENBLATT, MICHAEL M. GOTTESMAN

1. A Historical Perspective on Clinical Research

Guidelines and Principles for the Conduct of Research Scientific Integrity and Research Misconduct Responsibilities of Research Supervisors and Trainees Data Management, Archiving, and Sharing Research Involving Human and Animal Subjects Collaborative and Team Science Conflict of Interest and Commitment Peer Review Publication Practices, Responsible Authorship, and Results Reproducibility Study Questions Acknowledgments References Further Reading

JOHN I. GALLIN

The Earliest Clinical Research Greek and Roman Influence Middle Ages and Renaissance Seventeenth Century Eighteenth Century Nineteenth Century Twentieth Century and Beyond Summary Questions References

1 2 2 3 4 7 11 14 14

I

36 36 38 39 40 41 42 45 45 45 46

4. Institutional Review Boards

ETHICAL, REGULATORY AND LEGAL ISSUES

JULIA SLUTSMAN, LYNNETTE NIEMAN

Historical, Ethical, and Regulatory Foundations of Current Requirements for Research Involving Human Subjects Institutional Review Boards Clinical Researchers and Institutional Review Boards Evaluation and Evolution of the Current System of Research Oversight and Institutional Review Boards Conclusion Summary Questions References

2. Ethical Principles in Clinical Research CHRISTINE GRADY

Distinguishing Clinical Research From Clinical Practice Ethics and Clinical Research History of Ethical Attention to Clinical Research Codes of Research Ethics and Regulations Research on Bioethical Questions Ethical Framework for Clinical Research Ethical Considerations in Randomized Controlled Trials Conclusion Summary Questions References

33 34

19 20 20 22 23

47 50 57 57 59 59 59

5. Accreditation of Human Research Protection Programs

23 27 29 29 30

ELYSE I. SUMMERS, MICHELLE FEIGE

A Brief History Principles of Accreditation

v

63 64

vi Human Research Protection Programs: The Shift to Shared Responsibility The Accreditation Standards Steps to Accreditation Value of Accreditation Summary Questions References

CONTENTS

65 66 70 70 72 72

6. The Regulation of Drugs and Biological Products by the Food and Drug Administration MOLLY M. FLANNERY, AMY E. McKEE, DIANE M. MALONEY, JONATHAN P. JAROW

Background Mission and Terminology Drug and Biological Product Life Cycle Compliance Summary Summary Questions

73 74 76 84 84 84

7. International Regulation of Drugs and Biological Products

93 95 98 98

CHRISTOPHER O. OLOPADE, MICHELLE TAGLE, OLUFUNMILAYO I. OLOPADE

99 100 103 106 106 107

9. The Role and Importance of Clinical Trial Registries and Results Databases

PAUL G. WAKIM, PAMELA A. SHAW

Why Monitor? Who Monitors? What to Monitor? When and How Often to Monitor? Special Topics Summary Summary Questions Acknowledgments References

127 128 130 136 137 138 139 139 139

The Reasons The Drug The Target The Trials Cassandra Revealed Extended Studies Fialuridine Toxicity Reassessing the Preclinical Studies Research Oversight The Investigations Begin Scientific Misconduct The Food and Drug Administration The National Institutes of Health The Institute of Medicine The Media The Congress The Law Epilogue Acknowledgments References Further Reading

143 144 145 145 147 147 147 149 149 150 150 151 152 152 153 154 154 155 157 157 158

12. Legal Issues in Clinical Research

161

VALERIE H. BONHAM

DEBORAH A. ZARIN, REBECCA J. WILLIAMS, TONY TSE, NICHOLAS C. IDE

Introduction Background Current Policies

10. Data and Safety Monitoring

STEPHEN E. STRAUS

87 88

8. Clinical Research in International Settings: Opportunities, Challenges, and Recommendations

Introduction Challenges Recommendations Conclusion Summary Questions References

116 118 120 121 123 123 123

11. Unanticipated Risk in Clinical Research

THERESA MULLIN

Introduction Background Overview of the International Council on Harmonisation Technical Harmonization Process International Council on Harmonisation Guidelines Most Relevant to Clinical Research Future Work in Regulatory Harmonization References

Registering Clinical Trials at ClinicalTrials.gov Reporting Results to ClinicalTrials.gov Using ClinicalTrials.gov Data Looking Forward Conclusion Summary/Discussion Questions References

111 112 115

Introduction Protecting Individual Participant Interests Special Protections for Fetal Tissue, Human Embryos, and Human Embryonic Stem Cells Conflict of Interest and Financial Disclosure

161 162 166 167

vii

CONTENTS

Public Transparency: Registration and Results Reporting Recordkeeping and Privacy Protection Data Sharing and Individual Consent Conclusion Summary/Discussion Questions References

168 168 171 173 173 173

13. National Institutes of Health Policy on the Inclusion of Women and Minorities as Subjects in Clinical Research JANINE CLAYTON, JULIANA BLOME

National Institutes of Health Policy Scientific Considerations and Peer Review Role of the Institutional Review Board Challenges to Enrolling Volunteers Women of Childbearing Potential, Pregnant Women, and Children Demographic Trends in Clinical Trial Participation What Have We Learned? Conclusion Summary Questions References Further Reading

177 179 180 181 182 183 184 186 186 186 187

14. Clinical Research: A Patient Perspective JERRY SACHS

The PatienteScientist Partnership Walking Away: Why Patients Refuse to Participate in Clinical Trials The Trial Begins: Understanding the Patient Experience Understanding the Caregiver The Role of the Patient Representative The Role of Palliative Care Managing Difficult News Effective Patient Communications: Recommendations and Considerations The Assertive Patient: Ally in Scientific Research Conclusion Further Reading

190 191 193 195 195 196 196 197 198 199 199

II

204 213 213 216 217 217 217 217 217

16. Writing a Protocol ELIZABETH A. BARTRUM, BARBARA I. KARP

Introduction Regulatory Oversight Writing a Protocol Elements of a Protocol Summary Acknowledgments References

219 219 220 220 229 229 229

17. Design of Observational Studies LAURA LEE JOHNSON

Introduction Ecological (Correlational) Studies Case Reports and Case Series Single Time Point Studies: Cross-sectional Studies, Prevalence Surveys, and Incidence Studies Case-Control Studies Cohort Studies: Retrospective, Prospective, and Studies Nested Within a Cohort Odds Ratios, Risk Ratios, Relative Risks, and Attributable Risk Mistakes, Misconceptions, and Misinterpretations Conclusions Questions Acknowledgments Disclosures References

231 233 233 234 236 239 242 243 247 247 247 247 247

18. Design of Clinical Trials and Studies CATHERINE M. STONEY, LAURA LEE JOHNSON

STUDY DESIGN AND BIOSTATISTICS 15. Development and Conduct of Studies CATHERINE M. STONEY, LAURA LEE JOHNSON

Development and Conduct of Studies How to Choose a Study Design

Development and Importance of a Study Protocol Equipoise Manual of Operating Procedures Reporting the Results Conclusions Summary Questions Acknowledgments Disclosures References

203 204

Design of Clinical Trials The Purpose of Clinical Trials and Clinical Studies Understanding the Spectrum of the Research Continuum Clinical Trial Designs Critical Issues in Clinical Study Design Control Groups

250 250 251 255 258 258

viii Placebo Responses Mistakes and Misconceptions Conclusions Summary Questions Acknowledgments Disclosures References Further Reading

CONTENTS

261 262 266 266 266 266 266 268

19. The Role of Comparative Effectiveness Research JOE V. SELBY, EVELYN P. WHITLOCK, KELLY S. SHERMAN, JEAN R. SLUTSKY

Introduction A History of Comparative Clinical Effectiveness Research The Patient-Centered Outcomes Research Institute The Role of Comparative Clinical Effectiveness Research in the Nation’s Medical Research Enterprise The Methods of Comparative Clinical Effectiveness Research Study Designs for CER Studies Evidence Synthesis in CER Building a National Infrastructure for the Conduct of Comparative Effectiveness Research Conclusions References

NAOMI L. GERBER, JILLIAN K. PRICE

Introduction to Patient-Reported Outcomes, Measures of Function, and Health-Related Quality of Life Systematic Reviews Outcomes: Functional Measures and PatientReported Outcomes Summary Questions References Further Reading

269

22. Meta-analysis of Clinical Trials

270

JUNFENG SUN, BRADLEY D. FREEMAN, CHARLES NATANSON

271 273 275 278 285 287 290 290

20. Using Large Data Sets for Population-Based Health Research LEIGHTON CHAN, PATRICK McGAREY, JOSEPH A. SCLAFANI

Introduction What Are the Original Sources for These Data? Uses of Secondary Data in Health Research Strengths Limitations (and Solutions) Surveys Linking Data Sets Ethical Considerations Future Directions and Conclusions Summary Questions References

21. Measures of Function and Health-Related Quality of Life

293 294 294 296 297 298 298 299 300 300 300

Techniques of Meta-analysis Meta-analysis of Clinical Trials of Antiinflammatory Agents in Sepsis Conclusions Summary Questions References

303 305 306 313 313 315

318 321 323 323 324

23. Issues in Randomization PAMELA A. SHAW, LAURA LEE JOHNSON, CRAIG B. BORKOWF

What Is Randomization? Importance of Randomization History of the Randomized Trial Randomization Methods Issues in Implementation Special Considerations Conclusion Summary Questions Acknowledgments Disclosures References

329 330 330 331 333 335 337 338 338 338 338

24. Hypothesis Testing LAURA LEE JOHNSON, CRAIG B. BORKOWF, PAMELA A. SHAW

Introduction Basic Concepts in Hypothesis Testing Formulation of Statistical Hypotheses in the Motivating Examples One-Sample Hypothesis Tests With Applications to Clinical Research

342 343 345 346

ix

CONTENTS

Two-Sample Hypothesis Tests With Applications to Clinical Research Hypothesis Tests for the Motivating Examples Common Mistakes in Hypothesis Testing Misstatements and Misconceptions Special Considerations Conclusion Summary Questions Acknowledgments Disclaimers References

349 351 353 353 354 356 356 357 357 357

25. Power and Sample Size Calculations CRAIG B. BORKOWF, LAURA LEE JOHNSON, PAUL S. ALBERT

Introduction Sample Size Calculations for Precision in Confidence Interval Construction Sample Size Calculations for Hypothesis Tests: One Sample of Data Sample Size Calculations for Hypothesis Tests: Paired Data Sample Size Calculations for Hypothesis Tests: Two Independent Samples Advanced Methods and Other Topics Conclusion Exercises Acknowledgments Disclaimers References

359 361 362 364 366 368 369 370 371 371 371

26. An Introduction to Survival Analysis LAURA LEE JOHNSON

Introduction Features of Survival Data Survival Function Common Mistakes Conclusion Questions Acknowledgments Disclaimers References

373 374 375 380 380 381 381 381 381

403 405 406 406 407 407 407

28. Large Clinical Trials and RegistriesdClinical Research Institutes ROBERT M. CALIFF

Introduction History Phases of Evaluation of Therapies Critical General Concepts Expressing Clinical Trial Results Concepts Underlying Trial Design General Design Considerations Legal and Ethical Issues Hypothesis Formulation Publication Bias Statistical Considerations Meta-analysis and Systematic Reviews Understanding Covariates and Subgroups Therapeutic Truisms Operational Organization for Large-Scale Clinical Research Integration Into Practice Controversies and Personal Perspective The Future Summary Questions References

412 412 413 414 415 417 420 422 427 427 428 430 431 432 433 437 437 439 440 440

III TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH 29. Intellectual Property and Technology Transfer BRUCE GOLDSTEIN

27. Intermediate Topics in Biostatistics PAMELA A. SHAW, LAURA LEE JOHNSON, MICHAEL A. PROSCHAN

Special Topics in Trial Design Special Considerations in Data Analysis Regression to the Mean Diagnostic Testing Special Considerations in Survival Analysis

Missing Data Causal Inference in Observational Studies Concluding Remarks Summary Questions Acknowledgments Disclaimers References

384 392 394 396 400

Introduction Part One: Intellectual Property Generally Part Two: Patents and Technology Transfer Part Three: Technology Transfer Agreements Conclusion Brief Glossary of Critical Terms in Patenting Review Questions References

448 448 487 503 518 519 519 520

x

CONTENTS

30. Data Management in Clinical Trials

33. Evaluating a Protocol Budget

DIANE C. ST GERMAIN, MARJORIE J. GOOD

PHYLLIS KLEIN

The Research Team Data Management Auditing Unanticipated Problems and Adverse Event Monitoring and Reporting Legal and Regulatory Issues Related to Data Reporting Follow-Up and Analysis Record Retention Conclusion Summary Questions References

531 533 538 540 542 543 543 544 544 544

31. Clinical Research Data: Characteristics, Representation, Storage, and Retrieval JAMES J. CIMINO

Introduction Data as Surrogates Types of Data Data Standards Data Capture, Storage, and Retrieval Responsible Stewardship of Data Cooperative Sharing Efforts Summary Summary Questions References

547 547 550 550 551 553 555 556 556 557

32. Management of Patient Samples KAREN E. BERLINER, AMY P.N. SKUBITZ

Introduction Successful Research Rests on a Foundation of Careful Planning The Role of Pre-analytic Variables in Research Using Patient Specimens Training and Accreditation The Importance of Good Record Keeping Specimen Tracking Specimen Collection Specimen Handling Specimen Transit Specimen Storage Access to Patient Samples Specimen Culling, Transfer of Collections, and Repository Closings Summary Questions References

559 560 560 561 562 562 563 565 565 566 567 567 567 568

Overview Institutional Review Board Fees Overhead or Indirect Cost Determining the Hourly Rate The “Per Patient” Budget Start-Up Cost and Invoiced Items Submitting Your Budget to the Sponsor for Approval Areas of Concern Walking Away Wrapping Up

571 572 572 572 573 577 582 585 586 586

34. Getting the Funding You Need to Support Your Research: Navigating the National Institutes of Health Peer Review Process VALERIE L. PRENGER

Overview of National Institutes of Health The National Institutes of Health Peer Review Process for Grants Hints for Preparing Better Grant Applications Recent Changes to Application Procedures for National Institutes of HealtheFunded Clinical Trialsd More to Come Revising Unsuccessful Applications National Institutes of Health Grant Programs for Clinical Researchers at Various Stages in Their Careers How to Stay Informed About National Institutes of Health Peer Review

590 594 600

605 605

607 609

35. Philanthropy’s Role in Advancing Biomedical Research ELAINE K. GALLIN, MARYROSE FRANKO, ENRIQUETA BOND

Introduction Organization of the Philanthropic Sector and Terminology History of the Philanthropic Sector Areas of Contribution Conclusions and Future Directions Summary Questions References

611 613 615 617 628 629 629

xi

CONTENTS

Additional Educational Approaches and Support for Training Conclusions Summary/Discussion Questions References

IV CLINICAL RESEARCH INFRASTRUCTURE 36. Identifying, Understanding, and Managing Patient Safety and Clinical Risks in the Clinical Research Environment LAURA M. LEE, DAVID K. HENDERSON

Identifying and Managing Clinical Risk in the Clinical Research Environment Building a Road map to Safe and High-Quality Care and Research Support: Applying the Principles of High Reliability in the Clinical Research Environment Leveraging Patient Safety and Quality Improvement Techniques in the Conduct of Clinical Research Proactively Assessing Clinical and Operational Risk Electronic Surveillance for Errors and System Failures Patient Safety and Clinical Quality Measures Assessing Clinical Research Participants’ Perceptions of the Clinical Research Experience Conclusion Summary Questions References

633

635 635 638 641 641 642 642 643 643

37. Clinical Pharmacology and Its Role in Pharmaceutical Development SUE CHENG, KONSTANTINA M. VANEVSKI, JUAN J.L. LERTORA

Clinical Pharmacology as a Translational Discipline Overview of Drug Development Current State of Affairs in Drug Development Contribution of Clinical Pharmacology The Role of the Regulatory Agency Summary Questions References

645 646 647 649 654 656 656

39. Clinical Research Nursing: A New Domain of Practice GWENYTH R. WALLEN, CHERYL A. FISHER

Introduction Clinical Research Nursing: An Evolving Practice Specialty Defining and Documenting the Specialty of Clinical Research Nursing Legal Scope of Practice Issues Tools to Assist a Principal Investigator in Staffing a Study Future Considerations Summary/Discussion Questions Acknowledgments References

671 672 674 679 680 682 684 684 684

40. The Importance and Use of Electronic Health Records in Clinical Research JON W. McKEEBY, PATRICIA S. COFFEY

Electronic Medical Record Electronic Health Record Electronic Health Record Architecture Clinical Research Information Systems Using an Electronic Health Record in Clinical Research Secondary Use of the Electronic Health Record for Clinical Research Legislation and the Electronic Health Record Summary Summary Questions Terms References Further Reading

687 688 688 688 692 698 698 699 699 701 701 702

41. The Clinical Researcher and the Media JOHN T. BURKLOW

38. Career Paths in Clinical Research FREDERICK P. OGNIBENE

Background Student and Resident Training in Clinical Research PhysicianeScientist Workforce Clinical Research Curriculum and Training NIH Clinical Center Core Curriculum

668 669 669 669

661 662 664 665 666

What Makes News in Science and Medicine? Published SciencedThe Media’s Bread and Butter Novelty The Unexpected Celebrity Controversy Impact

704 704 704 705 705 705 706

xii Why Talk to Reporters? Why Reporters Want to Talk to You Why You Should Talk to Reporters Social Media: What to Keep in Mind Engaging the MediadThe Process A Word About Email, the Web, and Social Media The Interview What if You Are Misquoted? What the Public Does Not Know About Science Unexpected Questions When the News Is Not Good A Word About Investigative Reporters The Freedom of Information Act Embargoes When to Contact Your Communications Office Conclusion Summary Questions

CONTENTS

706 706 706 707 707 707 708 710 710 710 710 710 711 711 712 712 712

42. Information Resources for the Clinical Researcher JOSH A. DUBERMAN, PAMELA C. SIEVING

Introduction Organization and Features of Information Resources Origin

714 714 715

Content and Structure Search Capabilities Citation Searching Access and Business Models Familiarity and Currency Biomedical Databases Bioinformatics Resources Data Management Data Integration and Precision Medicine Bibliometrics Bibliographic Managers Resource Selection and Search Strategy Educational Resources Final Notes Acknowledgments Disclosure References

Appendix 1: Answer Key to Summary Questions Appendix 2: Acronyms Index

715 717 721 721 723 724 744 746 746 747 748 748 749 749 750 750 750

753 761 775

List of Contributors Paul S. Albert National Institutes of Health, Rockville, MD, United States

Maryrose Franko Health Research Alliance, Research Triangle Park, NC, United States

Elizabeth A. Bartrum National Institutes of Health, Bethesda, MD, United States

Bradley D. Freeman Washington University School of Medicine, St. Louis, MO, United States

Karen E. Berliner National Institutes of Health, Bethesda, MD, United States

Elaine K. Gallin QE Philanthropic Advisors, Potomac, MD, United States

Juliana Blome National Institutes of Health, Bethesda, MD, United States

John I. Gallin National Institutes of Health, Bethesda, MD, United States

Enriqueta Bond United States

Naomi L. Gerber George Mason University, Fairfax, VA, United States; Inova Health System, Falls Church, VA, United States

QE Philanthropic Advisors, Warrenton, VA,

Valerie H. Bonham National Institutes of Health, Bethesda, MD, United States

Bruce Goldstein a National Institutes of Health, Rockville, MD, United States

Craig B. Borkowf Centers for Disease Control and Prevention, Atlanta, GA, United States

Marjorie J. Good National Cancer Institute, National Institutes of Health, Rockville, MD, United States

John T. Burklow National Institutes of Health, Bethesda, MD, United States Robert M. Califf Duke University School of Medicine, Durham, NC, United States; Verily Life Sciences (Alphabet), South San Francisco, CA, United States; Stanford University Department of Medicine, Stanford, CA, United States Leighton Chan National Institutes of Health, Bethesda, MD, United States Sue Cheng Bayer HealthCare Pharmaceuticals, Inc., Whippany, NJ, United States James J. Cimino University of Alabama School of Medicine, Birmingham, AL, United States Janine Clayton National Institutes of Health, Bethesda, MD, United States Patricia S. Coffey National Institutes of Health, Bethesda, MD, United States

David K. Henderson National Institutes of Health, Bethesda, MD, United States Nicholas C. Ide National Institutes of Health, Bethesda, MD, United States Jonathan P. Jarow U.S. Food and Drug Administration, Silver Spring, MD, United States Laura Lee Johnson U.S. Food and Drug Administration, Silver Spring, MD, United States Barbara I. Karp National Institutes of Health, Bethesda, MD, United States

Laura M. Lee National Institutes of Health, Bethesda, MD, United States

Josh A. Duberman National Institutes of Health, Bethesda, MD, United States Michelle Feige Association for the Accreditation of Human Research Protection Programs, Inc., Washington, DC, United States

Molly M. Flannery U.S. Food and Drug Administration, Silver Spring, MD, United States

Christine Grady National Institutes of Health, Bethesda, MD, United States

Phyllis Klein Washington University, St. Louis, MO, United States

Melissa C. Colbert National Institutes of Health, Bethesda, MD, United States

Cheryl A. Fisher National Institutes of Health, Bethesda, MD, United States

Michael M. Gottesman National Institutes of Health, Bethesda, MD, United States

Juan J.L. Lertora Duke University School of Medicine, Durham, NC, United States Diane M. Maloney U.S. Food and Drug Administration, Silver Spring, MD, United States Patrick McGarey National Institutes of Health, Bethesda, MD, United States Amy E. McKee U.S. Food and Drug Administration, Silver Spring, MD, United States

a

Mr. Goldstein is a patent attorney serving as the Assistant Director for Monitoring & Enforcement Unit in the NIH Office of Technology Transfer. This chapter reflects the personal views of Mr. Goldstein, not of his employer. No official support or endorsement by the National Institutes of Health or the United States Government is intended or should be inferred.

xiii

xiv

LIST OF CONTRIBUTORS

Jon W. McKeeby National Institutes of Health, Bethesda, MD, United States

Amy P.N. Skubitz University of Minnesota, Minneapolis, MN, United States

Theresa Mullin U.S. Food and Drug Administration, Silver Spring, MD, United States

Jean R. Slutsky Patient-Centered Outcomes Research Institute (PCORI), Washington, DC, United States

Charles Natanson National Institutes of Health, Bethesda, MD, United States

Julia Slutsman National Institutes of Health, Washington, DC, United States

Lynnette Nieman National Institutes of Health, Bethesda, MD, United States

Diane C. St Germain National Cancer Institute, National Institutes of Health, Rockville, MD, United States

Robert B. Nussenblatt y National Institutes of Health, Bethesda, MD, United States

Catherine M. Stoney National Institutes of Health, Bethesda, MD, United States

Frederick P. Ognibene National Institutes of Health, Bethesda, MD, United States

Stephen E. Straus y National Institutes of Health, Bethesda, MD, United States

Christopher O. Olopade The University of Chicago, Chicago, IL, United States

Elyse I. Summers Association for the Accreditation of Human Research Protection Programs, Inc., Washington, DC, United States

Olufunmilayo I. Olopade The University of Chicago, Chicago, IL, United States Valerie L. Prenger National Institutes of Health, Bethesda, MD, United States Jillian K. Price Inova Health System, Falls Church, VA, United States Michael A. Proschan National Institutes of Health, Bethesda, MD, United States Jerry Sachs National Institutes of Health, Bethesda, MD, United States Joseph A. Sclafani National Institutes of Health, Bethesda, MD, United States; Medstar Georgetown University/ National Rehabilitation Network, Washington, DC, United States Joe V. Selby Patient-Centered Outcomes Research Institute (PCORI), Washington, DC, United States Pamela A. Shaw University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, United States Kelly S. Sherman Patient-Centered Outcomes Research Institute (PCORI), Washington, DC, United States Pamela C. Sieving Sieving Information Solutions, Bethesda, MD, United States

y

Deceased.

Junfeng Sun National Institutes of Health, Bethesda, MD, United States Michelle Tagle The University of Chicago, Chicago, IL, United States Tony Tse National Institutes of Health, Bethesda, MD, United States Konstantina M. Vanevski Bayer HealthCare Pharmaceuticals, Inc., Basel, Switzerland Paul G. Wakim National Institutes of Health, Bethesda, MD, United States Gwenyth R. Wallen National Institutes of Health, Bethesda, MD, United States Evelyn P. Whitlock Patient-Centered Outcomes Research Institute (PCORI), Washington, DC, United States Rebecca J. Williams National Institutes of Health, Bethesda, MD, United States Deborah A. Zarin National Institutes of Health, Bethesda, MD, United States

Acknowledgments

The editors extend special thanks to Ms. Jennifer Simmons for her energetic administrative support in coordinating the many activities associated with the development of the fourth edition of this textbook, Ms. Rona Buchbinder for her dedicated and excellent editorial assistance, and Ms. Kristine Jones, Ms. Molly McLaughlin, and Mr. Fenton Coulthurst at Elsevier for their patience and perseverance in bringing this huge undertaking to fruition. Very special thanks to all of the authors who contributed outstanding, up-to-date chapters to this fourth edition and the numerous patients, study participants, and course participants over the years who inspired them.

xv

Preface

The positive reactions and feedback to the first three editions of Principles and Practice of Clinical Research have been appreciated and reinforced the importance of this textbook to the discipline of clinical research. In each edition the content of the textbook has been updated and new information added. The textbook nearly doubled in size from the second to the third editions as an expanded and comprehensive section on biostatistics was included. The critical importance of study design and biostatistics, coupled with enhanced research regulatory requirements prompted the addition of a new editor, Laura Lee Johnson, PhD. Dr. Johnson has been a colleague for years, having been a faculty member and currently codirector of the National Institutes of Health (NIH) Clinical Center’s “Introduction to the Principles and Practice of Clinical Research” (IPPCR) course. After many years at the NIH, she is now Acting Director of the Division of Biometrics III in the Center for Drug Evaluation and Research at the U.S. Food and Drug Administration (FDA). She is an extremely welcome addition as the third editor of this textbook. IPPCR started at the NIH Clinical Center in 1995 and was the impetus for the first edition of this textbook. Currently IPPCR is a web-based course using recorded lectures by many of the textbook’s authors, online bulletin boards for each lecture, and local study groups hosted by volunteer institutions around the world. In the 2016e17 academic year we had over 8,800 registrants at 270 sites around the world. Since its inception the course has had nearly 38,000 participants formally enroll and an even wider audience informally taking the course or watching lectures via YouTube. In addition, the textbook has been translated into Chinese, Japanese, and Russian and has been used for live intensive IPPCR courses taught in China, Nigeria, Russia, India, Brazil, and South Africa. Based on broad international needs and interest in enhancing clinical research infrastructure around the world, this fourth edition includes an expanded chapter on clinical research in international settings as well as a new chapter focusing on international regulation of drugs and biologics. It also includes updated content on large clinical trials and registries as well as a new chapter focusing on the emergence of the important role of comparative effectiveness research. Since clinical research has become more complex and thus, potentially, more risky, there is a new chapter devoted to identifying clinical risks and managing patient safety in a clinical research setting. There also is new content on the use of electronic health records in clinical research and a very detailed presentation of the broad utility and application of informational resources in clinical research. With the growth of the clinical research enterprise and the need to ensure that the highest standards are maintained, chapters about accreditation of human research protection programs, regulatory sciences, and research integrity have been enhanced. We hope that this book provides its audience with a deeper understanding of the broadening scope of the global clinical research enterprise. The textbook provides not only details about clinical research mechanics and practical information but also introduces the reader to the complexities and intricacies of ensuring safe, ethically sound, and scientifically rigorous clinical research. All clinical investigators must consider the safety of research subjects enrolled in their investigational protocols while navigating the research pathways from the bedside to the bench and back. We are proud of this book on so many levels and hope that the passion, expertise, and dynamic quality of our contributors and their content are appreciated by you, the readers. John I. Gallin, MD Frederick P. Ognibene, MD Laura Lee Johnson, PhD

xvii

C H A P T E R

1 A Historical Perspective on Clinical Research John I. Gallin National Institutes of Health, Bethesda, MD, United States

O U T L I N E The Earliest Clinical Research

1

Greek and Roman Influence

2

Middle Ages and Renaissance

2

Seventeenth Century

3

Eighteenth Century

4

Nineteenth Century Twentieth Century and Beyond

11

Summary Questions

14

References

14

health services research; epidemiology; and community-based and managed care-based research.

If I have seen a little further it is by standing on the shoulders of giants. Sir Isaac Newton (1676).

The successful translation of a basic or clinical observation into a new treatment of disease is rare in an investigator’s professional life, but when it occurs, the personal thrill is exhilarating, and the impact on society may be substantial. The following historical highlights provide a perspective of the continuum of the clinical research endeavor. These events also emphasize the contributions that clinical research has made to advances in medicine and public health. In this chapter, and throughout the book, a broad definition of clinical research from the Association of AmericanMedicalCollegesTaskForceonClinicalResearch is used.1 This task force defined clinical research as

THE EARLIEST CLINICAL RESEARCH Medical practice and clinical research are grounded in the beginnings of civilization. Egyptian medicine was dominant from approximately 2850 BC to 525 BC. The Egyptian Imhotep, whose name means “he who gives contentment,” lived slightly after 3000 BC and was the first physician figure to rise out of antiquity.2 Imhotep was a known scribe, priest, architect, astronomer, and magician (medicine and magic were used together); he performed surgery, practiced some dentistry, extracted medicine from plants, and knew the position and function of the vital organs. Imhotep likely provided the first description of cancer in one of his 48 clinical case reports. In case 45, he reported, “If you examine (a case) having bulging masses on (the breast) and you find that they have spread over his breast; if you place your hand upon (the) breast (and) find them to be cool, there being no fever at all therein when your hand feels him; they have no granulations,

a component of medical and health research intended to produce knowledge essential for understanding human disease, preventing and treating illness, and promoting health. Clinical research embraces a continuum of studies involving interaction with patients, diagnostic clinical materials or data, or populations, in any of these categories: disease mechanisms; translational research; clinical knowledge; detection; diagnosis and natural history of disease; therapeutic interventions including clinical trials; prevention and health promotion; behavioral research;

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00001-0

7

1

Copyright © 2018. Published by Elsevier Inc.

2

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

contain no fluid, give rise to no liquid discharge, yet they feel protuberant to your touch you say concerning him: ‘This is a case of bulging masses I have to contend with.’ Bulging tumors of the breast mean the existence of swellings on the breast, large, spreading, and hard; touching them is like touching a ball of wrappings, or they may be compared with unripe hemat fruit, which is hard and cool to the touch.”3 Evidence also shows that ancient Chinese medicine included clinical studies. For example, in 2737 BC, Shen Nung, the putative father of Chinese medicine, experimented with poisons and classified medical plants,4 and I. Yin (1176e1123 BC), a famous prime minister of the Shang dynasty, described the extraction of medicines from boiling plants.5 Documents from early Judeo-Christian and Eastern civilizations provide examples of a scientific approach to medicine and the origin of clinical research. In the Old Testament, written from the 15th century BC to approximately the 4th century BC,6 a passage in the first chapter of the Book of Daniel describes a comparative “protocol” of diet and health. In the setting of Babylon where Israelites defiled the sin of eating rich food, Daniel described the preferred diet of legumes and water made for healthier youths compared with the king’s rich food and wine: Then Daniel said to the steward. “Test your servants for ten days; let us be given vegetables to eat and water to drink. Then let your appearance and the appearance of the youths who eat the king’s rich food be observed by you, and according to what you see deal with your servants. So he harkened to them in this matter; and tested them for ten days. At the end of ten days it was seen that they were better in appearance and fatter in flesh than all the youths who ate the king’s rich food. So the steward took away their rich food and the wine they were to drink, and gave them vegetables.” Daniel 1:11e16

The ancient Hindus excelled in early medicine, especially in surgery. Sushruta, the father of Indian surgery, resided in the court of the Gupta kings in about 600 BC and wrote medical texts about surgery, the most famous being Sushruta Samhita, an encyclopedia of medical learning. In addition, there is evidence of Indian hospitals in Ceylon in 437 BC and 137 BC.7

GREEK AND ROMAN INFLUENCE Although early examples of clinical research predate the Greeks, Hippocrates (460e370 BC) is considered the father of modern medicine, and he exhibited the strict discipline required of a clinical investigator. His emphasis on the art of clinical inspection, observation, and documentation established the science of

medicine. In addition, as graduating physicians are reminded when they take the Hippocratic oath, he provided physicians with high moral standards. Hippocrates’ meticulous clinical records were maintained in 42 case records.8 These case studies describe, among other maladies, malarial fevers, diarrhea, dysentery, melancholia, mania, and pulmonary edema with remarkable clinical acumen. On pulmonary edema, he wrote the following: Water accumulates; the patient has fever and cough; the respiration is fast; the feet become edematous; the nails appear curved and the patient suffers as if he has pus inside, only less severe and more protracted. One can recognize that it is not pus but water.if you put your ear against the chest you can hear it seethe inside like sour wine.9

Hippocrates also described the importance of cleanliness in the management of wounds. He wrote, “If water was used for irrigation, it had to be very pure or boiled, and the hands and nails of the operator were to be cleansed.”10 Hippocrates used the Greek word for “crab,” karkinos, to describe cancer. The tumor, with its clutch of swollen blood vessels around it, reminded Hippocrates of a crab dug in the sand with its legs spread in a circle.11 Hippocrates’ teachings remained dominant and unchallenged until the time of Claudius Galen of Pergamum (c.130e200 AD), the physician to the Roman Emperor Marcus Aurelius.12 Galen was one of the first individuals to utilize animal studies to understand human disease. By experimenting on animals, he was able to describe the effects of transection of the spinal cord at different levels. According to Galen, health and disease reflected the balance of four humors (blood, phlegm, black bile, and yellow bile), and veins contained blood and the humors, together with some spirit.12 Inflammation, described by Galen as a red, hot, and painful distention, was attributed to excessive blood. Tubercles, pustules, catarrh, and nodules of lymph, all cool, boggy and white, were attributed to excesses of lymph. Jaundice was an overflow of yellow bile. Cancer was attributed to black bile as was melancholia, the medieval term for depression. Thus cancer and depression were closely intertwined.13

MIDDLE AGES AND RENAISSANCE In the Middle Ages, improvements in medicine became evident, and the infrastructure for clinical research began to develop. Hospitals and nursing, with origins in the teachings of Christ,14 became defined institutions (although the beginnings of hospitals can be traced to the ancient Babylonian custom of bringing the sick into the marketplace for consultation, and the Greeks and Romans had military hospitals).

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

3

SEVENTEENTH CENTURY

The Persian al-Razi (865e925) discovered the use of alcohol as an antiseptic and wrote the first treatise on pediatrics, as well as more than 180 books and articles.15 Persian scientists emphasized the importance of methodology, and Ibn al-Haytham (Alhazen) wrote his Book of Optics, for which he is regarded as the father of optics.16 The surgical needle was invented and described by Abu al-Qasim al-Zahrawi in his Al-Tasrif in the year 1000.17 The Iraqi surgeon Ammar ibn Ali al-Mawsili invented the first injection syringe in the 9th century using a hollow glass tube and suction to extract and remove cataracts from patients’ eyes.18 By the 1100s and 1200s, hospitals were being built in England, Scotland, France, and Germany. Early progress in pharmacology can be linked to the Crusades and the development of commerce. Drug trade became enormously profitable during the Middle Ages. Drugs were recognized as the lightest, most compact, and most lucrative of all cargoes. Records of the customhouse at the port of Acre (1191e1291) show a lively traffic in aloes, benzoin, camphor, nutmegs, and opium.19 Influences of Arabic pharmacy and contact of the Crusaders with their Muslim foes spread the knowledge of Arabic pharmaceuticals and greatly enhanced the value of drugs from the Far East. The Persian Ibn Sina-Avicenna (980e1037), a leader in pharmacy, philosophy, medicine, and pharmacology, wrote The Canon of Medicine, which describes seven conditions for “the recognition of the strengths of the characteristics of medicines through experimentation”: ensuring the use of pure drugs, testing the drug for only one disease, use of control groups, use of dose escalation, requirement of long-term observation, requirement of reproducible results, and requirement of human over animal testing.20 Documentation through case records is an essential feature of clinical research. Pre-Renaissance medicine of the 14th and 15th centuries saw the birth of “Consilia” or medical case books, consisting of clinical records from the practice of well-known physicians.21 Hippocrates’ approach of case studies developed 1700 years earlier was reborn, particularly in the Bolognese and Paduan regions of Italy. Universities became important places of medicine in Paris, Bologna, and Padua. Clinical research remained mostly descriptive, resembling today’s natural history and disease pathogenesis protocols. In 1348, Gentile da Foligno, a Paduan professor, described gallstones.21 Bartolomeo Montagna (1470), an anatomist, described strangulated hernia, operated on lachrymal fistula, and extracted decayed teeth.21 The Renaissance (1453e1600) represented the revival of learning and the transition from medieval to modern conditions; many great clinicians and scientists prospered. At this time, many of the ancient Greek dictums of medicine, such as Galen’s four humors, were discarded. Perhaps the most important anatomist of

FIGURE 1.1 Leonardo da Vinci self-portrait (red chalk); Turin, Royal Library. From Da Vinci L. Copyright in Italy by the Institute Geografic DeAgostini S.p.A. e Novara. New York: Reymal & Company; 1956, Fig. 1 [Wikipedia].

this period was Leonardo da Vinci (1453e1519) (Fig. 1.1).22 Da Vinci created more than 750 detailed anatomic drawings (Fig. 1.2). In 1533, Andreas Vesalius, at age 19, was beginning his incredible career as an anatomist. His dissections and recordings of the human anatomy recorded in detailed plates and drawings of patients with cancer failed to note black bile in any cancer, regardless of the organ involved, and provided the basis for dismissing Galen’s theory of the role of black bile in cancer.23

SEVENTEENTH CENTURY Studies of blood began in the 17th century. William Harvey (1578e1657) convincingly described the circulation of blood from the heart through the lungs and back to the heart and then into the arteries and back through

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

4

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

FIGURE 1.3 Christopher Wren’s drawing of the brain shows blood vessels discovered by Thomas Willis. From Knoeff R. Book review of soul made flash: discovery of the brain and how it changed the world by C. Zimmer. Nature 2004;427:585.

FIGURE 1.2 Example of anatomic drawing by Leonardo da Vinci. Trunk of female human body, with internal organs seen as though ventral side were transparent. From Da Vinci L. Copyright in Italy by the Institute Geografic DeAgostini S.p.A. e Novara. New York: Reymal & Company; 1956. p. 369 [Wikipedia].

the veins.24 Harvey emphasized that the arteries and veins carried only one substance, the blood, ending Galen’s proposal that veins carried a blend of multiple humors. (Of course, today we know that blood contains multiple cellular and humoral elements, so to some extent Galen was correct.) The famous architect Sir Christopher Wren (1632e1723), originally known as an astronomer and anatomist (Fig. 1.3), in 1656 assembled quills and silver tubes as cannulas and used animal bladders to inject opium into the veins of dogs.25 The first well-documented transfusions of blood were done in animals (dogs) in 1667 by Richard Lower and Edmund King in London26 and were mentioned in Pepys’ diary.27 The first transfusions into humans are attributed to the French physician Jean-Baptiste Denys who in June 1667 transfused sheep blood into a 15year-old boy who survived. James Blundell performed the first modern transfusions in humans in 1818, some of whom survived.28 Transfusions did not become an accepted approach until Landsteiner discovered the major A, B, AB, and O blood groups in 1900 and 1901.29

The 17th century also brought the first vital statistics, which were presented in Graunt’s book Natural and Political Observations Mentioned in a Following Index, and Made Upon the Bills of Mortality.30 In this book of comparative statistics, population and mortality statistics were compared for different countries, ages, and sexes in rural and urban areas. Use of data on mortality among groups would have major importance in future clinical studies.

EIGHTEENTH CENTURY The 18th century brought extraordinary advances in the biological sciences and medicine. At the end of the 17th century, Antony van Leeuwenhoek of Delft (1632e1723) invented the microscope. Although he is best known for using his microscope to provide the first descriptions of protozoa and bacteria, Leeuwenhoek also provided the first description of striated voluntary muscle, the crystalline structure of the lens, red blood cells, and spermatozoa (Figs. 1.4 and 1.5).31 Modern clinical trials can be recognized in the 1700s. Scurvy was a major health problem for the British Navy. William Harvey earlier had recommended lemons to treat scurvy but argued that the therapeutic effect was a result of the acid in the fruit. James Lind (Fig. 1.6), a native of Scotland and a Royal Navy surgeon, conducted a clinical trial in 1747 to assess this hypothesis by

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

5

EIGHTEENTH CENTURY

0*0

PHILOSOPHICAL TRANSACTIONS.

For the Months of Jng*Ji and Stftemttr. Stftemt. 21. 1674, The CONTENTS. llxnfatictl Qtftrtfttutt firna «#r. Leeuwcnboeck, «fotf Blood, Milk, Bones, tkt Brain, Spitle, Coticula, Sweat, Fact, Teares 5 ummtmutei t» t** Letttrs to the PaUfitr. An AtuHit tf* ntttkU Ctfe if* Dropfy, mfoktn fir Grtvibtk* at+yt**g W*mt* ^ imftrttAh* Lttnttd Ptyfoi* MI* Holland. jfa4tt*at,f three E»hr LDE SECRET IQKE JN-IM.ALI Ctgtutt, 4»tb. Gall. Co'e, At D. IF. Ertfmi BtrthiM SELECT A GEO eX£ fR \CA. III. LOG ICJ, fat An CritoUi; f * Gtllit* i» iMnnu* Strmutem vtrfc SUM A*t*u&o*rfait *p* tke Latin Verfon, mult by C. S. of the Pbil.TranUflions*/' J. 1665.1666.1667. l OtftnnttoKfr** e^C Leeowrahoeck, u#tr»*7 Blood, Milk, Bones, /fc Bnio,SpitIe,36^ Cuticula,^. MmmmtiUAk tkt JUt Obfervtr i» tin fMjbtr in * Letttr, lUteljUK r. 1674, Sir,

Y

Ours of 14* of Mrit\a& was my welcome to roe j Wheocelnnd«ftoodwitfj great cooteotfneot, that my Microfcopical Cooxnookatiom badnot been nnaocepoble co yoaaod yoor Philofophkal Frkndr; wUcb hath encouraged R

**

van leeuwenhoek and his little animals. New York, Dover: A Collection of Writings by the Father of Protozoology and Bacteriology; 1960 [Original work published in 1932].

FIGURE 1.5 Title page from Leeuwenhoek’s paper, “Microscopical Observations.” From Dobell C. Antony van leeuwenhoek and his little animals. New York, Dover: A Collection of Writings by the Father of Protozoology and Bacteriology; 1960 [Original work published in 1932].

comparing three therapies for scurvy (Table 1.1).32 Twelve sailors with classic scurvy were divided into six groups of two each, all given identical diets; the various groups were supplemented with vinegar, dilute sulfuric acid, cider, seawater, and a nutmeg, garlic, and horseradish mixture, along with two oranges and one lemon daily. Sulfuric acid, vinegar, seawater, cider, and the physician’s remedy had no benefit. Two sailors receiving citrus fruit avoided scurvy. Although not significant because of sample size, this early clinical study formed the basis for successful avoidance of scurvy with citrus fruit. Studies with sulfuric acid, vinegar, and cider excluded acid as a likely explanation for the beneficial effect of citrus fruit. The 18th century saw great progress in the area of surgery. A remarkable succession of teachers and their students led these studies. Percival Pott of St. Bartholomew’s Hospital described tuberculosis of the spine, or Pott’s disease.33 John Hunter, Pott’s pupil, was the founder of experimental and surgical pathology and was a pioneer in comparative physiology and experimental morphology. Hunter described shock, phlebitis, pyremia, and intussusception and reported major findings of inflammation, gunshot wounds, and surgical

diseases of the vascular system.33 Hunter’s student Edward Jenner (1749e1823)33 introduced vaccination as a tool to prevent infectious diseases (Fig. 1.7).34 Jenner was aware that dairymaids who had contracted cowpox through milking did not get smallpox. In 1798, Jenner conceived of applying this observation on a grand scale to prevent smallpox.35 Jenner was not the first to conceive of the idea of inoculation for smallpox. The Chinese had thought of this earlier, and Sir Hans Sloane had done small studies in 1717 using variolation (inoculating healthy people with pus from blisters obtained from patients with smallpox).36 In 1718, after providing variolation vaccination of her 3-year-old son in Turkey and, 3 years later, her 5-year-old daughter in England, Lady Mary Worley Montagu introduced the Ottoman practice of variolation to the West.36 In addition, James Jurin published several articles between 1723 and 1727 comparing death from natural smallpox in people who had not been inoculated versus those who had been inoculated. Jurin showed that death occurred in 5 of 6 subjects in the first group compared with 1 in 60 in the latter,37 providing one of the first studies using mortality as a critical clinical end point. In 1734, Voltaire wrote, “The Cirassians [a Middle Eastern people] perceived that of a thousand

FIGURE 1.4 Antony van Leeuwenhoek. From Dobell C. Antony

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

6

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

FIGURE 1.6

TABLE 1.1

FIGURE 1.7 Edward Jenner (painting by Sir Thomas Lawrence). From Garrison FH. History of medicine. Philadelphia: Saunders; 1917. Reprinted 1963.

James Lind.

Treatment of Scurvy by James Lind

Treatment Arm

Cured

p-valuea

Sulfuric acid

0/2

>0.05

Vinegar

0/2

>0.05

Seawater

0/2

>0.05

Cider

0/2

>0.05

Physicians

0/2

>0.05

Citrus fruit

2/2

>0.05

a

Compared to patients in the five areas of the trial; no placebo group.

persons hardly one was attacked twice by full blown smallpox; that in truth one sees three or four mild cases but never two that are serious and dangerous; that in a word one never truly has that illness twice in life.”38 Thus, Voltaire recognized natural immunity to smallpox, which was an important concept for future vaccinology. In 1721, Cotton Mather demonstrated that variolation protected citizens of the American colonies in Massachusetts,39 and, in 1777, George Washington used variolation against smallpox to inoculate the

Continental Armydthe first massive immunization of a military group.40 In 1774 Benjamin Jesty, a cattle breeder in Dorset, England, inoculated his wife and two sons with cowpox to protect them during an outbreak of smallpox. Jenner, based on his clinical observation that persons who had cowpox were protected from smallpox and his subsequent work showing that people inoculated with cowpox were protected when challenged with smallpox, was the first to try vaccination on a large scale using scabs from cowpox to protect against human smallpox. Jenner was the first to use experimental approaches to establish the scientific basis for vaccination and he transformed a local country tradition into a viable prophylactic principle. Jenner’s vaccine was adopted quickly in Germany and then in Holland, Denmark, the rest of Europe, and the United States. The 1700s were also the time when the first known blinded clinical studies were performed. In 1784 a commission of inquiry was appointed by King Louis XVI of France to investigate medical claims of “animal magnetism” or “mesmerism.” The commission, headed by Benjamin Franklin and consisting of such

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

NINETEENTH CENTURY

distinguished members as Antoine Lavoisier, JeanSylvain Bailly, and Joseph-Ignace Guillotin, had as a goal to assess whether reported effects of this new healing method were due to “real” force or to “illness of the mind.” Among the many tests performed, blindfolded people were told that they were either receiving or not receiving magnetism when in fact, at times, the reverse was happening. Results showed that study subjects felt effects of magnetism only when they were told that they received magnetism and felt no effects when they were not told this, whether or not they were receiving treatment.41 This was the beginning of the use of blinded studies in clinical research. In addition to the first blinded or masked studies, Franklin, for the first time, also pointed out the importance of the placebo effect. The 18th century provided the first legal example that physicians must obtain informed consent from patients before performing a procedure. In an English lawsuit, Slater vs. Baker & Stapleton, two surgeons were found liable for disuniting a partially healed fracture without the patient’s consent.42 This case set an important precedent described by the court: “Indeed it is reasonable that a patient should be told what is about to be done to him that he may take courage and put himself in such a situation as to enable him to undergo the operation.”

7

of the great Justice Holmes, read his article, “On the Contagiousness of Puerperal Fever,”46 to the Boston Society for Medical Improvement (Fig. 1.8). Holmes stated that women in childbed should never be attended by physicians who have been conducting postmortem sections on cases of puerperal fever; that the disease may be conveyed in this manner from patient to patient, even from a case of erysipelas; and that washing the hands in calcium chloride and changing the clothes after leaving a puerperal fever case was likely to be a preventive measure. Holmes’ essay stirred up violent opposition by obstetricians. However, he continued to reiterate his views, and in 1855 in a monograph, Puerperal Fever as a Private Pestilence, Holmes noted that Semmelweis, working in Vienna and Budapest, had lessened the mortality of puerperal fever by disinfecting the hands with chloride of lime and the nail brush.47 Ignaz Philipp Semmelweis (1818e65) performed the most sophisticated preventive clinical trial of the 19th century, which established the importance of hand washing to prevent the spread of infection (Fig. 1.9).48

NINETEENTH CENTURY In the first days of the 19th century, Benjamin Waterhouse, a Harvard professor of medicine, brought Jenner’s vaccine to the United States, and by 1802 the first vaccine institute was established by James Smith in Baltimore, Maryland. In 1813 this led to the establishment of a national vaccine agency by the Congress of the United States under the direction of James Smith.43 Jenner’s vaccination for smallpox was followed by other historic studies in the pathogenesis of infectious diseases. In the mid-1800s, John Snow (1813e58), an anesthesiologist by training, performed the classic studies that determined how cholera was spread in contaminated water. Snow’s studies, which included the first use of statistical mapping, identified contaminated water as the source of cholera. For his work, John Snow is widely considered to be the father of modern epidemiology.44 The French physician Pierre Charles Alexandre Louis (1787e1872) realized that clinical observations on large numbers of patients were essential for meaningful clinical research. He published classical studies on typhoid fever and tuberculosis, and his research in 1835 on the effects of bloodletting demonstrated that the benefits claimed for this popular mode of treatment were unsubstantiated.45 On February 13, 1843, one of Louis’ students, Oliver Wendell Holmes (1809e94), the father

FIGURE 1.8 Oliver Wendell Holmes. From Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 435. Reprinted 1963.

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

8

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

FIGURE 1.9 Ignaz Philipp Semmelweis. From Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 436. Reprinted 1963.

Semmelweis, a Hungarian pupil, became an assistant in the first obstetric ward of the Allgemeines Krankenhaus in Vienna in 1846. Semmelweis was troubled by the death rate associated with puerperal or “childbed” fever. From 1841 to 1846, the maternal death rate from puerperal sepsis averaged approximately 10%, and in some periods was as high as 50%, in the First Maternity Division of the Vienna General Hospital. In contrast, the rate was only 2% or 3% in the Second Division, which was attended by midwives rather than physicians. The public knew the disparity, and women feared being assigned to the First Division. Semmelweis became frustrated by this mystery and began to study cadavers of fever victims. In 1847, his friend and fellow physician Jakob Kolletschka died after receiving a small cut on the finger during an autopsy. The risk of minor cuts during autopsies was well known, but Semmelweis made the further observation that Kolletschka’s death was characteristic of death from puerperal fever. He reasoned that puerperal fever was “caused by

conveyance to the pregnant women of putrid particles derived from living organisms, through the agency of the examining fingers.” In particular, he identified cadaveric matter from the autopsy room, with which midwives had no contact, as the source of the infection. In 1847 Semmelweis insisted that all students and physicians scrub their hands with chlorinated lime before entering the maternity ward, and during 1848, the mortality rate on his division dropped from 9.92% to 1.27%. Despite his convincing data, colleagues rejected Semmelweis’ findings and accused him of insubordination. The dominant medical thinking at the time was that high mortality in the charity hospital was related to the poor health of impoverished women, despite the differences between control (no chlorinated lime hand washing) and experimental (washing with chlorinated lime) divisions. Without any opportunity for advancement in Vienna, Semmelweis returned to his home in Budapest and repeated his studies with the same results. In 1861, he finally published The Etiology, Concept, and Prophylaxis of Childhood Fever.48 Although Holmes’ work antedated Semmelweis by 5 years, the superiority of Semmelweis’ observation lies not only in his experimental data but also in his recognition that puerperal fever was a blood poisoning. The observations of Holmes and Semmelweis represent a critical step for medicine and surgery. In addition to the discovery of the importance of hand washing, the first well-documented use of ether for surgery (1846) by William Thomas Green Morton, a Boston dentist, with Dr. John Collins Warren as the surgeon at the Massachusetts General Hospital, occurred during the 19th century.49 The discovery of anesthesia led to the dissociation of pain from surgery, allowing surgeons to perform prolonged operations. Oliver Wendell Holmes is credited with proposing the words anesthetic and anesthesia.49 Recognition of the importance of hand washing and the discovery of anesthetics were essential findings of the 19th century that were critical in the development of modern surgery. In 1865, a Scottish surgeon named Joseph Lister recognized the importance of keeping surgical wounds clean and wrote “.that the decomposition in the injured part might be avoided.by applying as a dressing some material capable of destroying the life of the floating particles.” Based on the observation that carbolic acid cleansed raw storage, Lister began to apply carbolic acid to wounds with great success, establishing the importance of antisepsis in the operating room.50 The work of Holmes and Semmelweis on the importance of hand washing opened the door for Pasteur’s work on the germ basis of infectious diseases. Louis Pasteur (1822e95) was perhaps the most outstanding clinical investigator of the 19th century (Fig. 1.10). He was trained in chemistry. His fundamental work in

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

NINETEENTH CENTURY

9

bacterial culture and identification easy and widely available. Koch cultured the tubercle bacillus and identified the causative agent for anthrax, which was later used by Pasteur to develop a vaccine, and he established Koch’s postulates to prove that an infectious agent causes disease (Fig. 1.11).51 The studies of Pasteur and Koch were performed during the same period as the work of the Norwegian Gerhard Armauer Hansen (1841e1912). In 1874, based on epidemiologic studies in Norway, Hansen concluded that Mycobacterium leprae was the microorganism responsible for leprosy. Hansen’s claim was not well received, and in 1880, in an attempt to prove his point, he inoculated live leprosy bacilli into humans, including nurses and patients, without first obtaining permission. One of the patients brought legal action against Hansen.

FIGURE 1.10 Louis Pasteur. One of the remarkable facts about Pasteur was his triumph over a great physical handicap. In 1868 at age 46, just after completing his studies on wine, he had a cerebral hemorrhage. Although his mind was not affected, he was left with partial paralysis of his left side, which persisted for the remainder of his life. This photograph, taken after he was awarded the Grand Cross of the Legion of Honor in 1881, gives no hint of his infirmity. From Haagensen CD, Lloyd EB. A hundred years of medicine. New York: Sheridan House; 1943. p. 116.

chemistry led to the discovery of levo and dextro isomers. He then studied the ferments of microorganisms, which eventually led him to study the detrimental causes of three major industries in France: wine, silk, and wool. Pasteur discovered the germ basis of fermentation, which formed the basis of the germ theory of disease.51 He discovered Staphylococcus pyogenes as a cause of boils and the role of Streptococcus pyogenes in puerperal septicemia. In other studies, he carried forward Jenner’s work on vaccination and developed approaches to vaccine development using attenuation of a virus for hydrophobia (rabies) and inactivation of a bacterium for anthrax. The work of Pasteur was complemented by the studies of Robert Koch (1843e1910), who made critical technical advances in bacteriology. Koch was the first to use agar as a culture medium, and he introduced the Petri dish, pour plates, and blood agar to make

FIGURE 1.11 Robert Koch. His career in research began in 1872, when his wife gave him a microscope as a birthday present. He was then 28 years old, performing general practice in a small town in Silesia. This was an agricultural region where anthrax was common among sheep and cattle, and it was in the microscopic study of this disease in rabbits that Koch made his first great discovery of the role of anthrax bacilli in disease. From Haagensen CD, Lloyd EB. A hundred years of medicine. New York: Sheridan House; 1943. p. 132.

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

10

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

The court, in one of the early cases demonstrating the importance of informed consent in clinical research, removed Hansen from his position as director of Leprosarium No. 1, where the experiments had taken place. However, Hansen retained his position as chief medical officer for leprosy52 and later in his life received worldwide recognition for his life’s work on leprosy. In the same era, Emil von Behring (1854e1917) demonstrated in 1890 that inoculation with attenuated diphtheria toxins in one animal resulted in production of a therapeutic serum factor (antitoxin) that could be delivered to another, thus discovering antibodies and establishing a role for passive immunization. On Christmas Eve of 1891, the first successful clinical use of diphtheria antitoxin occurred.51 By 1894, diphtheria antiserum became commercially available as a result of Paul Ehrlich’s work establishing methods of producing high-titer antisera. Behring’s discovery of antitoxin was the beginning of humoral immunity, and in 1901 Behring received the first Nobel Prize. Koch received the Prize in 1905 (Fig. 1.12). The Russian scientist Elie Metchnikoff (1845e1916) discovered the importance of phagocytosis in hostdefense against infection and emphasized the

FIGURE 1.12 Emil von Behring. From Hirsch JG. Host resistance to infectious diseases: a centennial. In: Gallin JI, Fauci AS, editors. Advances in host defense mechanisms: vol. 1. Phagocytic cells. New York: Raven Press; 1982. p. 7.

importance of cellular components of host defense against infection.53 Paul Ehrlich (1854e1915) discovered the complement system and asserted the importance of the humoral components of host defense. In 1908, Metchnikoff and Ehrlich shared the Nobel Prize (Figs. 1.13 and 1.14). At the end of the 19th century, studies of yellow fever increased awareness of the importance of the informed consent process in clinical research. In 1897, Italian bacteriologist Giuseppe Sanarelli announced that he had discovered the bacillus for yellow fever by injecting the organism into five people. William Osler was present at an 1898 meeting at which the work by Sanarelli was discussed, and Osler said, “To deliberately inject a poison of known high degree of virulency into a human being, unless you obtain that man’s sanction.is criminal.”54 This commentary by Osler had substantial influence on Walter Reed, who demonstrated in human volunteers that the mosquito is the vector for yellow fever. Reed adopted written agreements (contracts) with all his yellow fever subjects. In addition to obtaining signed permission from all his volunteers, Reed

FIGURE 1.13 Elie Metchnikoff in his 40s. From Tauber AI, Chernyak L. Metchnikoff and the origins of immunology. New York: Oxford University Press; 1991, Fig. 5 [Wikipedia].

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

TWENTIETH CENTURY AND BEYOND

11

TWENTIETH CENTURY AND BEYOND

FIGURE 1.14 Paul Ehrlich. From Hirsch JG. Host resistance to infectious diseases: a centennial. In: Gallin JI, Fauci AS, editors. Advances in host defense mechanisms: vol. 1. Phagocytic cells. New York: Raven Press; 1982. p. 9.

made certain that all published reports of yellow fever cases included the phrase “with his full consent.”54 On November 8, 1895, Wilhelm Ro¨ntgen (1845e1923), a German physicist, produced and detected electromagnetic radiation, and on December 22, 1895, he took the first X-ray of his wife’s hand. For this achievement, Ro¨ntgen won the first Nobel Prize in Physics in 1901 (Fig. 1.15A and B). Toward the end of the 19th century, women began to play important roles in clinical research. Marie Curie (1867e1934) and her husband Pierre won the Nobel Prize in Physics in 1903 for their work on spontaneous radiation; in 1911 Marie Curie won a second Nobel Prize (in chemistry) for her studies on the separation of radium and description of its therapeutic properties. Marie Curie and her daughter Irene Joliot-Curie (who won a Nobel Prize in Chemistry in 1935 for her work synthesizing new radioactive elements leading to the discovery of uranium fission) promoted the therapeutic use of radium during World War I (Fig. 1.16).55 Florence Nightingale (1820e1910), in addition to her famous work in nursing, was an accomplished mathematician who applied her mathematical expertise to dramatize the needless deaths caused by unsanitary conditions in hospitals and the need for reform (Fig. 1.17).56

The spectacular advances in medicine that occurred during the 20th century would never have happened without centuries of earlier progress. In the 20th century, medical colleges became well established in Europe and the United States. The great contributions of the United States to medicine in the 20th century are linked to an early commitment to strong medical education. The importance of clinical research as a component of the teaching of medicine was recognized in 1925 by the American medical educator Abraham Flexner, who wrote, “Research can no more be divorced from medical education than can medical education be divorced from research.”57 Two other dominant drivers of progress in medicine through clinical research were government investment in biomedical research and private investment in the pharmaceutical industry. These investments, closely linked with academia, resulted in enhanced translation of basic observations to the bedside. Paul Ehrlich coined the term “chemotherapy” and popularized the concept of a “magic bullet” (chemicals injected into the blood to fight various diseases, particularly those caused by parasites). Ehrlich, in 1910, working with his assistant Sahachiro Hata, developed Salvarsan (arsphenamine), a trivalent arsenic-based chemotherapeutic to cure syphilis. Salvarsan and later Neosalvarsan were commercialized by Hoechst AG as one of the first pharmaceuticals (antibiotics). Sir Alexander Fleming’s discovery of penicillin in 1928 in Scotland spawned expansion of the pharmaceutical industry through the development of antibiotics, antiviral agents, and new vaccines. The Canadian physician Frederick Banting and medical student Charles Best’s discovery of insulin in 1921 was followed by their collaboration with the Canadian chemist James B. Collip who assisted with purification of insulin from cows for use in humans and then the Scottish physiologist J.J.R. Macleod’s confirmatory studies on use of insulin in humans provided lifesaving long-standing management of diabetes for which Banting and Macleod won the Noble Prize in Physiology and Medicine in 1923. The discovery of insulin was followed by the discovery of multiple hormones to save lives. In the 1920s and 1930s, Sir Ronald Aylmer Fisher (1890e1962), from the United Kingdom, introduced the application of statistics and experimental design.58 Fisher worked with farming and plant fertility to introduce the concepts of randomization and analysis of variancedprocedures used today throughout the world. In 1930, Torald Sollmann emphasized the importance to a study of controlled experiments with placebo and blind limbsda rebirth of the “blinded” or “masked”

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

12

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

FIGURE 1.15

FIGURE 1.16

(A) Wilhelm Conrad Ro¨ntgen. (B) Print of Wilhelm Ro¨ntgen’s first X-ray of his wife’s hand.

Marie Curie (1867e1934).

FIGURE 1.17 Florence Nightingale (1820e1910).

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

TWENTIETH CENTURY AND BEYOND

studies originated by Benjamin Franklin in 1784. Sollmann wrote, “Apparent results must be checked by the ‘blind test,’ i.e., another remedy or a placebo, without the knowledge of the observer, if possible” (Fig. 1.18). He said “Observations without adequate controls and checks are practically useless.”59 Through these approaches, many new drugs for treatment of hypertension, cardiovascular disease, manic depression, and epilepsy, to name a few, were developed. The spectacular advances of the 20th century were associated with troubling events in clinical research that heightened public attention and formalized the field of clinical bioethics. Nazi human experimentation led to the Nuremberg Code in 1947, which was designed to protect human subjects by ensuring voluntary consent of the human subject and by asserting that the anticipated result of research must justify its performance. The Tuskegee syphilis experiments initiated in the 1930s and continued until 1972 in African American men and the Willowbrook hepatitis studies of the mid-1950s in children with Down syndrome highlighted the need to establish strict rules to protect research patients. In 1953 the US National Institutes of Health (NIH) issued “Guiding Principles in Medical Research Involving Humans,” which required prior review by a medical committee of all human research to be conducted at the newly opened NIH Clinical Center. In 1962, the Kefauver-Harris amendment to the 1938 US Federal Food, Drug, and Cosmetics Act stipulated that subjects must be told whether a drug is being used for investigational purposes and that subject consent must be obtained. In 1964, the World Medical Assembly adopted

FIGURE 1.18 Testing puddings and gelatins at Consumers Union. Copyright 1945 by Consumers Union of U.S., Inc., Yonkers, NY. Reprinted with permission from the April 1945 issue of Consumer Reports.

13

the Declaration of Helsinki, stressing the importance of assessing risks and determining that risks are outweighed by potential benefits of research. In 1966, Henry Beecher pointed out major ethical issues in clinical research.60 During the same year, the US Surgeon General issued a memo to the heads of institutions conducting research with Public Health Service grants requiring prior review of all clinical research. The purposes of this review were to ensure protection of research subjects, to assess the appropriateness of methods employed, to obtain informed consent, and to review risks and benefits of the research; thus institutional review boards were established. In 1967, the U.S. Food and Drug Administration added the requirement that all new drug sponsors must obtain informed consent for use of investigational drugs in humans. Over the past 50 years, clinical research has become big business. The pharmaceutical and biotechnology industries have engaged university-based clinical investigators in the business of clinical research. For example, in the United States interaction between federal investigators and industry, encouraged by the US Congress when it passed the Federal Technology Transfer Act in 1986, has successfully increased the translation of basic research to the bedside by US government scientists. At the same time, however, the relationship between industry and academia has grown closer, and new ethical, legal, and social issues have evolved worldwide. Clinical investigators have become increasingly associated with real and perceived conflicts. Examples of these issues include promoting an investigator’s financial or career goals while protecting the patient, protecting “unborn children” while pursuing the potential use of embryonic stem cells to rebuild damaged organs, and protecting patient confidentiality as a result of gene sequencing. As a consequence of these issues, the public has engaged in debate about the well-being of current and future generations of patients who volunteer to partner with a clinical investigator on protocols. The 20th century saw incredible advances in genomics, including the Nobel prizes to Watson and Crick for the description of the double helix model of DNA61 and to Barbara McClintock for her work in the 1940s for studies of corn indicating an organism’s genome is not a stationary entity, but rather is subject to alteration and rearrangement through transposable elements or jumping genes.62 In the 1970s Janet Rowley discovered that translocation between chromosome 8 and 21 caused acute myelogenous leukemia and between chromosome 15 and 17 caused promyelocytic leukemia.63 These and other genomic and molecular discoveries have provided opportunities for conducting clinical research in the 21st century that are greater than ever. A new urgency to move clinical research findings from the laboratory to the patient and into the community has prioritized

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

14

1. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

translational research globally. Today, understanding and meeting public concerns is as important for the clinical investigator as performing the clinical study. Principles for conducting clinical research have evolved from centuries of experience. As the science moves forward, ethical, legal, and social issues pose special challenges for the clinical investigator. These challenges are the focus of the following chapters of this book.

SUMMARY QUESTIONS 1. The definition of clinical research embraces a continuum of studies in which of these following categories? (More than one item can be selected.) a. Behavioral research b. Health services research c. Epidemiology d. Disease mechanisms e. Translational research f. Diagnosis and natural history of disease g. Therapeutic interventions including clinical trials h. Prevention and health promotion i. Community-based and managed care-based research j. All of the above 2. True or False: Although early examples of clinical research predate the Greeks, Galen (AD 129 216) is considered the father of modern medicine, and he exhibited the strict discipline required of a clinical investigator. 3. Ignaz Semmelweis performed the most sophisticated preventive clinical trial of the 19th century, which established the importance of hand washing. Circle all of the following statements related to Semmelweis’ work that are true: (More than one item can be selected.) a. Semmelweis started his career as a student on an obstetric ward b. The death rate from puerperal sepsis reached 90% in select maternity divisions c. The second division used midwives and the death rate was only 2%e3% d. Semmelweis started his work by studying cadavers e. Semmelweis introduced hand washing with chlorinated lime to decrease mortality rates f. Despite convincing data, Semmelweis’ work was condemned by colleagues 4. The first blinded clinical study was done by which of the following? a. Hippocrates b. Galen c. James Lind d. Benjamin Franklin e. Louis Pasteur

References 1. Association of American Medical Colleges Task Force on Clinical Research 2000, vol. 1. Washington, DC: Association of American Medical Colleges; 1999. p. 3. 2. Thorwald J. Science and secrets of early medicine. New York: Harcourt, Brace and World; 1962. 3. Mukherjee S. The emperor of all maladies. New York: Scribner; 2010. p. 40. 4. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 73e4. Reprinted 1963. 5. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 70. Reprinted 1963. 6. Lane B. In: Burgland L, editor. Reading the bible with understanding. St. Louis, MO: Concordia: How We Got the Bible; 1999. 7. Saraf S, Parihar RS. Sushruta: the first plastic surgeon in 600 B.C. Intern J Plast Surg 2007;4(2). 8. Adams F. The genuine works of Hippocrates. New York: William Wood; 1886. 9. Lyons AS, Petrucelli RJ. Medicine, an illustrated history. New York: Abradale Press; 1987. p. 216. 10. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 98. Reprinted 1963. 11. Mukherjee S. The emperor of all maladies. New York: Scribner; 2010. p. 47. 12. Logic Nutton V. Learning, and experimental medicine. Science 2002;295:800e1. 13. Mukherjee S. The emperor of all maladies. New York: Scribner; 2010. p. 48. 14. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 176. Reprinted 1963. 15. Ligon BL. Rhazes: his career and his writings. Semin Pediatr Infect Dis 2001;12(3):266e72. 16. Hehmeyer I, Khan A. Islam’s forgotten contributions to medical science. Can Med Assoc J 2007;176(10):1467e8. 17. Ahmad Z. Al-Zahrawidthe father of surgery. ANZ J Surg 2007; 77(Suppl. 1):A83. 18. Finger S. Origins of neuroscience: a history of explorations into brain function. New York: Oxford University Press; 1994. p. 70. 19. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 180. Reprinted 1963. 20. Sajadi M, et al. Ibn Sina and the clinical trial. Ann Intern Med 2009; 150:640e3. 21. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 166e7. Reprinted 1963. 22. Da Vinci L. Copyright in Italy by the Istituto Geografico De Agostini S.p.A. e Novara. New York: Reymal & Company; 1956. 23. Mukherjee S. The emperor of all maladies. New York: Scribner; 2010. p. 53. 24. Wintrobe MM. Blood, pure and eloquent. New York: McGraw-Hill; 1980. 25. Wintrobe MM. Blood, pure and eloquent. New York: McGraw-Hill; 1980. p. 661e2. 26. Wintrobe MM. Blood, pure and eloquent. New York: McGraw-Hill; 1980. p. 663. 27. Nicolson MH. Pepys’ diary and the new science. Charlottesville: University Press of Virginia; 1965. p. 663. Quoted in reference 13. 28. Blundell J. Observations on the transfusion of blood. Lancet 1828; 2(2):321. 29. Landsteiner K. On the individual differences in human blood. In: Nobel lectures, physiology or medicine 1922e1941. Amsterdam: Elsevier Publishing Company; 1965 30. Graunt J. Natural and political observations mentioned in a following Index, and made upon the Bills of mortality. London, 1662. Reprinted by Johns Hopkins press, Baltimore. 1939 Quoted in Lilienfeld AM. Centeris Paribus: the evolution of the clinical trial. Bull Hist Med 1982;56:1e18.

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

REFERENCES

31. Dobell C. Antony van Leeuwenhoek and his little animals. New York, Dover: A Collection of Writings by the Father of Protozoology and Bacteriology; 1960 [Original work published in 1932]. 32. Lind J. A treatise of the scurvy. Edinburgh, UK, sands, Murray and Cochran, 1753, p. 191e193. Quoted in Lilienfeld AM. Centeris Paribus: the evolution of the clinical trial. Bull Hist Med 1982;56:1e18. 33. Haagensen CD, Lloyd EB. A hundred years of medicine. New York: Sheridan House; 1943. 34. Wood GB. Practice of medicine. Philadelphia: Collins; 1849. 35. Jenner E. An inquiry into the causes and effects of the variolae vaccinae. London: Sampson Low; 1798. 36. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 373. Reprinted 1963. 37. Miller G. The adoption of inoculation for smallpox in England and France. Philadelphia: university of Pennsylvania press; 1957, 114e118. Quoted in Lilienfeld AM. Centers Paribus: the evolution of the clinical trial. Bull Hist Med 1982;56:1e18. 38. Plotkin SA. Vaccines: past, present and future. Nat Med 2005;11: S5e11. 39. Harper DP. Angelical conjunction: religion, reason, and inoculation in Boston, 1721e1722. The Pharos Winter 2000:1e5. 40. Fenn EA. Pox Americana. The great small pox epidemic of 1775e82. New York: Hill and Wang; 2001. 41. Franklin B. Animal and other commissioners charged by the king of France. Animal magnetism. 1784. An historical outline of the “Science” made by the committee of the Royal Academy of Medicine in Philadelphia translated from the French. Philadelphia: H. Perkins; 1837. 42. Slater vs. Baker & Stapleton (1767) 95, Eng. Rep. 860. Quoted in Appelbaum PS. In: Lidz CW, Meisel A, editors. Informed consent. Legal theory and clinical practice. New York: Oxford University Press; 1987. 43. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 375. Reprinted 1963. 44. Hempel S. The strange case of the broad street pump: John Snow and the Mystery of Cholera. Berkley: University of California Press; 2007. 45. Morabia APCA. Louis and the birth of clinical epidemiology. J Clin Epidemiol 1996;49:1327e33.

15

46. Holmes OW. On the contagiousness of puerperal fever. N Engl J Med 1842e1843;1:503e30. Quoted in reference 3, 435. 47. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 435. Reprinted 1963. 48. Semmelweiss IP. Die Aetiologie, der Begriff und die Prophylaxis des Kindbettfiebers. C.A. Hartleben: Budapest and Vienna; 1861. p. 436. Quoted in reference 3. 49. Garrison FH. History of medicine. Philadelphia: Saunders; 1917. p. 506. Reprinted 1963. 50. Lister J. Classics in infections diseases. On the antiseptic principle of the practice of surgery. Rev Infect Dis 1987;9(2):421e6. 51. Hirsch JG. Host resistance to infectious diseases: a centennial. In: Gallin JI, Fauci AS, editors. Advances in host defense mechanisms: vol. 1. Phagocytic cells. New York: Raven Press; 1982. 52. Bendiner E. Gerhard Hansen: hunter of the leprosy bacillus. Hosp Pract December 15, 1989:145e70. 53. Tauber AI, Chernyak L. Metchnikoff and the origins of immunology. New York: Oxford University Press; 1991. 54. Lederer SE. Human experimentation in America before the Second World War. Baltimore: Johns Hopkins University Press; 1995. 55. Macklis RM. Scientist, technologist, proto-feminist, superstar. Science 2002;295:1647e8. 56. Cohen IB. Florence Nightingale. Sci Am 1984;250:128e37. 57. Flexner A. Medical education. A comparative study. New York: Macmillan; 1925. 58. Efron B. Fisher in the 21st century. Stat Sci 1998;13:95e122. 59. Sollmann T. The evaluation of therapeutic remedies in the hospital. J Am Med Assoc 1936;94:1280e300. 60. Beecher HK. Ethics and clinical research. N Engl J Med 1966;274: 1354e60. 61. Watson J, Crick F. Molecular structure of nucleic acids. A structure of deoxyribonucleic acid. Nature 1953;171:737e8. 62. Pray L, Zhaurovak K. Barbara McClintock and the discovery of jumping genes (transposons). Nat Education 2008;1:169. 63. Drucker BJ. Janet Rowley (1925e2013). Geneticist who discovered that broken chromosomes cause cancer. Nature 2014;505:784.

I. A HISTORICAL PERSPECTIVE ON CLINICAL RESEARCH

P A R T I

ETHICAL, REGULATORY AND LEGAL ISSUES

C H A P T E R

2 Ethical Principles in Clinical Research Christine Grady National Institutes of Health, Bethesda, MD, United States

O U T L I N E Distinguishing Clinical Research From Clinical Practice

19

Ethics and Clinical Research

20

History of Ethical Attention to Clinical Research Benefit to the Individual Benefit to Society Protection of Research Subjects Research as a Benefit Community Involvement in Research

20 20 20 21 21 21

Codes of Research Ethics and Regulations

Value and Validity Fair Subject Selection Favorable Risk/Benefit Ratio Independent Review Informed Consent Respect for Enrolled Subjects

23 24 25 25 26 27

Ethical Considerations in Randomized Controlled Trials

27

22

Conclusion

29

Research on Bioethical Questions

23

Summary Questions

29

Ethical Framework for Clinical Research

23

References

30

Clinical research has resulted in significant benefits for society, yet continues to pose profound ethical questions. This chapter briefly describes: five overlapping but distinct eras reflecting the history of clinical research ethics; codes of research ethics; and seven ethical principles that guide clinical research ethics and particular ethical challenges in randomized controlled trials (RCTs).

hypotheses and permits generalizable conclusions useful in understanding human health and illness, improving medical care or the public health, and developing safe and effective interventions to prevent, diagnose, and treat disease. As such, research serves the common or collective good; the individual subject participating in clinical research may or may not benefit from participation. Clinical research is distinct from clinical practice in that each has different, yet not mutually exclusive, purposes, goals, and methods.1 Clinical practice involves diagnosis, prevention, treatment, and care for a particular individual or group of individuals with the goal of meeting the health needs of and benefiting that individual(s). Clinical practice is based on evidence or experience, is designed to enhance the patient’s well-being, and has a reasonable expectation of success. Usual methods in clinical practice are evidence-based and

DISTINGUISHING CLINICAL RESEARCH FROM CLINICAL PRACTICE Clinical research involves the study of human beings in a systematic investigation of health and illness, designed to develop or contribute to generalizable knowledge. The goal of clinical research is to gather knowledge through a set of activities that tests

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00002-2

19

Copyright © 2018. Published by Elsevier Inc.

20

2. ETHICAL PRINCIPLES

guided by standard practice and experience. The risks of interventions or procedures employed in clinical practice are justified by the prospect of therapeutic benefit to the individual. In contrast, clinical research aims to generate useful knowledge and is not designed to meet the health needs of, nor necessarily to benefit, individual patient participants. Although an individual may receive quality patient care and treatment when participating in research, this is not the goal of research, and much research does not directly benefit individual participants. Further, frequently used research methodologies, such as randomization, blinding, dose escalation, placebo controls, and others are rarely found and might be considered unacceptable, in clinical practice. In clinical research, some risk is justified by the importance of the knowledge to be gained rather than benefit to the individual participant.

ETHICS AND CLINICAL RESEARCH Two fundamental ethical questions regarding clinical research are important to consider: (1) why should we do research with human beings and (2) how should it ethically be done? Two competing considerations frame these questions: clinical research is valuable in generating practical knowledge useful for advancing or improving medical care and health, yet respect for the rights, welfare, dignity, and freedom of choice of individual humans is indispensable. Research with human beings is essential to advancing or improving medical care and/or the public health and providing health professionals with the knowledge and evidence necessary to appropriately and safely care for patients. The pursuit of knowledge through research should be rigorous to inform effective and safe clinical practice, and progress would not be possible without rigorous clinical research. Conducting clinical research designed to enhance the understanding of human health and illness may be more than a social good; arguably it is a social imperative.2 Although progress in medical care and health is a societal good, some contend it is an optional good,3 and that other considerations, such as the primacy of the individual, should take precedence. Whether improvement in medical care or health through clinical research is an option or an imperative, limits are necessary. Human research participants are the means to securing practical knowledge, but because people should not be treated “merely [as] means to an end, but always as ends in themselves,”4 the need to respect and protect human research participants is paramount. The primary ethical tension in clinical research, therefore, is that a few individuals are asked to accept some research burden, risk, or inconvenience to benefit

others, including future persons and society. Ethical requirements aim to minimize the possibility of exploiting research participants by ensuring that they are treated with respect while contributing to the generation of knowledge, and their rights and welfare are protected throughout the process of research.

HISTORY OF ETHICAL ATTENTION TO CLINICAL RESEARCH Throughout history, perception and acceptance of the methods, goals, and scope of clinical research have evolved significantly, as have attention to and appreciation of what respecting and protecting research participants entails. A brief detour through the history of clinical research illustrates these changing perspectives.5

Benefit to the Individual Historically and for hundreds of years, there was little basis for a distinction between experimentation and therapy because most therapy was experimental, and systematic evidence of the effectiveness of medical interventions was rare. Experimental therapies were used in the hopes of benefiting ill patients, but such “therapy” frequently contributed to or caused morbidity or mortality. Systematic research was sporadic. Most researchers were medical practitioners, motivated to do what they thought best for their patients, and trusted to do the right thing. Fraud and abuse were minimized to some extent through peer censorship because no specific codes of ethics, laws, or regulations governed the conduct of research. Early regulations, such as the Pure Food and Drug Act of 1906 in the United States, prohibited unsubstantiated claims on medicine labels. Yet, research began to grow as an enterprise only after the development of early antibiotics like penicillin and the passage of the Food, Drug, and Cosmetic Act in 1938, which required evidence of safety before a product was marketed.6

Benefit to Society Around the time of World War II, there was a dramatic shift in clinical research with tremendous growth in the research enterprise. Pharmaceutical companies were established; large amounts of both public and private money were devoted to research; and research became increasingly centralized, coordinated, standardized in method, and valued. Human subjects research entered what has since been described as an “unashamedly utilitarian phase.”7 Individuals often were included in research because they were

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

HISTORY OF ETHICAL ATTENTION TO CLINICAL RESEARCH

available, captive, and marginalized, and they were seen as making a contribution to society. The federal government and the pharmaceutical industry supported intensive research efforts to develop vaccines and antibiotics for infectious diseases to help soldiers, as infectious diseases were a significant problem for the armed services. During this era, research was commonly conducted in prisons, orphanages, and homes for the emotionally or developmentally disturbed, as well as with other institutionalized groups. The distinction between research and therapy was fairly clear; subjects not necessarily in need of therapy were accepting a personal burden to make a contribution to society. A utilitarian justification served as the basis of claims that some individuals could be used for the greater common good. Revelations of Nazi medical experiments and war crimes, and the Nuremberg trial of Nazi doctors, raised public and professional concerns about the justification and scope of research with human subjects.8

Protection of Research Subjects In the late 1960s and early 1970s in the United States, shock and horror at stories of abuse of human subjects led to intense scientific and public scrutiny and reflection, and debate about the scope and limitations of research involving human subjects. A renowned Harvard anesthesiologist, Henry Beecher, published a landmark article in the New England Journal of Medicine in 19669 highlighting ethical problems in 22 research studies conducted in reputable US institutions. Exposition of studies such as the hepatitis B studies at Willowbrook, the U.S. Public Health Service Tuskegee syphilis studies, and others generated intense public attention and concern. Congressional hearings and action led to passage of the 1974 National Research Act (PL 93e348) and establishment of the US National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research.10 This extremely influential body authored multiple reports and recommendations about clinical research, including reports on research with children and on institutional review boards (IRBs). Included in its legacy is the Belmont Report, in which ethical principles underlying the conduct of human subject research and their application are explained.11 The Commission’s work emphasized the need to protect individuals participating in research from potential exploitation and harm, and provided the basis for subsequent federal regulations codified in 1981 in Title 45, US Code of Federal Regulations (USCFR), Part 46 (45CFR46), titled “Protection of Human Subjects,” and similar FDA regulations (21 CFR.50 and 56). In 1991, the Department of Health and Human

21

Services (DHHS) regulations became the currently operative Common Rule,12 which governs the conduct of human subjects research funded by 17 US federal agencies. The major thrust of these federal regulations and of many existing codes of research ethics continues to be protection of subjects from the burdens and harms of research.

Research as a Benefit Events in the late 1980s and 1990s altered some public perspectives on clinical research. Certain articulate and vocal activists asserted that research participation can offer an advantage that individuals want access to, rather than simply harm to be protected from.13 According to this perspective, as espoused by human immunodeficiency virus (HIV) and breast cancer activists and others, participation in research is a benefit, protectionism is discrimination, and exclusion from research can be unjust. Empirical studies have demonstrated that oncology patients, for example, who participate in clinical trials benefit through improved survival.14,15 Activism and changes in public attitudes about research led to substantive changes in the way research is done and drugs are approved. In addition to the possible benefits of participation for individuals, it was claimed that certain traditionally underrepresented groups were being denied the benefits of the application of knowledge gained through research.16 Since 1994, the US National Institutes of Health (NIH) has required that those who receive research funding must include previously underrepresented women and ethnic minorities.17 Since 1998, NIH guidelines have required the inclusion of children in research or justification for their exclusion.18

Community Involvement in Research In subsequent years, the growth of genetics research, research with stored biospecimens and data, and international collaborative research, in particular, have highlighted the value of greater public and community involvement in research. Clinical research does not occur in a vacuum but is a collaborative social activity that requires the support and investment of involved communities; and it also comes with inherent risks and potential benefits for communities and groups. As such, involvement of the community (1) in helping to set research priorities, (2) in planning and approving research, (3) in evaluating risks and benefits during and after a trial, and (4) in influencing particular aspects of recruitment, informed consent, and the realization of community benefits demonstrates respect for the community and can facilitate successful research.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

22

2. ETHICAL PRINCIPLES

CODES OF RESEARCH ETHICS AND REGULATIONS Throughout this history, several influential documents have helped to shape our sense of the contours of ethical research (Table 2.1). Most were written in response to specific crises or historical events, yet all have accepted an underlying assumption that research as a means to progress in medical care or health is a social good. The Nuremberg Code, a 10-point code on the ethics of human experimentation, was written as the concluding part of the judgment at the Nuremberg Trials (1949).19 Established in response to Nazi experimentation, the Nuremberg Code recognized the potential value of research knowledge to society but emphasized the absolute necessity of voluntary consent of the subject. The Nuremberg Code established that ethical research must prioritize the rights and welfare of the subject. Most subsequent codes and guidelines for the ethical conduct of research have maintained this emphasis and all have incorporated requirements for informed consent. The Declaration of Helsinki was developed by the World Medical Assembly (WMA) in 1964 as a guide to the world’s physicians involved in human subject research.20 The Declaration of Helsinki recognizes that some, but not all, medical research is combined with clinical care and emphasizes that patients’ participation in research should not put them at a disadvantage with respect to medical care. The Declaration of Helsinki also recognizes legitimate research with people who cannot give their own informed consent, such as children and the cognitively impaired, but for whom informed permission could be obtained from a legal guardian. The Declaration of Helsinki has had considerable influence on the formulation of international, regional, and national legislation and regulations governing clinical research. The Declaration of Helsinki has been revised multiple times by the WMA (1975, 1983, 1989, 1996, 2000, 2008, and 2013) TABLE 2.1

Selected Codes and US Regulations Guiding Clinical Research

• The Nuremberg Code (1949) • The World Medical Association Declaration of Helsinki (1964, 1975, 1983, 1989, 1996, 2000, 2008, and 2013) • The National Commission’s Belmont Report (1979) • CIOMS International Ethical Guidelines for Biomedical Research Involving Human Subjects (1982, 2002, 2015) • International Conference on Harmonization Guidelines for Good Clinical Practice (1996) • Title 45, USCFR, Part 46, “The Common Rule” • Title 21, USCFR, Part 50 (“Protection of Human subjects”) and 56 (“Institutional Review Boards”)

and is considered a living document. Certain provisions of the Helsinki Declaration, such as posttrial obligations and the use of placebo controls, have been topics of continued debate among international researchers. The Belmont Report, published by the US National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, describes three broad ethical principles that guide the conduct of research and form the “basis on which specific rules could be formulated, criticized, and interpreted.”11 These three principles are respect for persons, beneficence, and justice. Respect for persons requires respect for the autonomous decision-making of capable individuals as applied in the process of informed consent and also calls for protection of those with diminished autonomy. Beneficence requires protecting individuals from deliberate and unnecessary harm, as well as maximizing benefits and minimizing harms, and is applied to clinical research through careful risk/benefit evaluation. Justice demands a fair distribution of the benefits and burdens of research and is applied in the Belmont Report to fairness in the processes and outcomes of selecting research subjects. In 1982, the Council of International Organizations of Medical Sciences (CIOMS), in conjunction with the World Health Organization (WHO), issued International Ethical Guidelines for Biomedical Research Involving Human Subjects, which were revised in 1993, 2002, and 2015.21 The CIOMS guidelines acknowledge that background circumstances sometimes differ between low- and middle-income and high-income countries, and there may be differences in the primacy of focus on the individual and individual rights. CIOMS set out to apply the Helsinki principles to the “special circumstances of many technologically developing countries.” CIOMS adopted the three ethical principles spelled out in the US National Commission’s Belmont Report and maintains most of the tenets of Nuremberg and Helsinki but has provided additional and valuable guidance and commentary on externally sponsored research and research with vulnerable populations. The 2015 revision restructures and expands many previously existing guidelines and adds new guidelines on compensation for research-related injury, research with stored biospecimens and data, and implementation science, among others.21 Federal regulations found in Title 45, USCFR, Part 46 (45CFR46),12 were promulgated in 1981 for research funded by DHHS (formerly the Department of Health, Education, and Welfare), and at Title 21 USCFR, Part 50 and 56 for the U.S. Food and Drug Administration (FDA).22 FDA regulations are similar, but not identical, to those found in the Common Rule.23 Compliance with these and other FDA regulations is required for

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

ETHICAL FRAMEWORK FOR CLINICAL RESEARCH

research investigating FDA-regulated products, such as drugs, biologics, and medical devices. DHHS regulations were extended in 1991 as the Federal Common Rule, applicable to research funded by 17 US federal agencies (not including the FDA). Based on recommendations of the National Commission, the Common Rule stipulates both the membership and the function of IRBs, and the criteria that an IRB should apply when reviewing a research protocol to determine whether to approve it. The Common Rule also delineates the information that should be included in an informed consent document, how consent should be documented, and criteria for waiver or alteration of informed consent. Subparts B, C, and D of 45CFR46 describe additional protections for DHHS-funded research with fetuses and pregnant women, prisoners, and children, respectively. In 2017, a final revision to the Common Rule was published in the Federal Register, with the most extensive changes to the Common Rule since 1991.24 The International Conference on Harmonization (ICH) sought to harmonize regulatory guidelines for product registration trials for the United States, the European Union, and Japan. The ICH Good Clinical Practice (GCP) (E-6) Guidelines provide widely accepted guidance promoting the ethical conduct of research and reporting of accurate and reliable data.25 The World Health Organization produced good clinical research guidelines that incorporated ICH-GCP and also included types of clinical research beyond drugregistration trials.26 Good clinical practice guidelines are being adopted by countries around the world to guide the conduct of research.

RESEARCH ON BIOETHICAL QUESTIONS The historical evolution of clinical research ethics and the development of guidelines and regulations was largely in response to particular events or scandals. The Nuremberg Code, for example, was a response to atrocities performed by Nazi research doctors during World War II; the formation of the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research was in response to revelations of the U.S. Public Health Service syphilis studies in Tuskegee. Our systems for protection of human subjects, the focus of the ethics of clinical research, and the existing regulations grew out of these efforts. Another essential way to inform our thinking about the ethics of clinical research, and one that has gained traction in recent decades, is research on bioethical questions. Bioethics research is usually conducted using one or more of the following methodologies: historical inquiry, conceptual analysis, empirical studies, or policy analysis.27 For

23

example, bioethics research on voluntariness, an essential part of informed consent, could better our understanding of what voluntariness means and how to maximize it in the process of informed consent. Such research might include an analysis of the concept. Recognizing that all decisions and actions can be influenced by one’s understanding, previous experiences, religion and culture, and the influences of respected others, distinguishing what makes a choice sufficiently voluntary from a choice that is controlled is important. Conceptual bioethics research also might examine the concepts of coercion, undue influence, and manipulation, which are different possible controlling influences.28 Empirical research might seek to elucidate how people actually choose research participation, what sources of influence and pressure they identify, whether they perceive they could say no to participation and under what circumstances, experiences of manipulation or undue influence, and other phenomena. Requirements for rigorous and ethical research on topics in bioethics are similar to those for ethical clinical research.

ETHICAL FRAMEWORK FOR CLINICAL RESEARCH A systematic framework of principles for ethical clinical research was derived from guidance provided in various ethical codes, guidelines, literature, and bioethics research. This proposed framework of principles is meant to apply sequentially and universally to clinical research.29 According to this framework, ethical clinical research should satisfy the following requirements: social or scientific value, scientific validity, fair subject selection, favorable risk/benefit ratio, independent review, informed consent, and respect for enrolled subjects30 (Table 2.2). Each will be described briefly.

Value and Validity The first requirement of ethical research is that the research question must be worth askingdthat is, have potential social, scientific, or clinical value. The anticipated usefulness of knowledge to be gained in understanding or improving health or health care is the crux of determining valuednot whether study results are positive or negative. A study should have sufficient social value to justify asking individuals to assume risk or inconvenience in research and to justify the expenditure of resources.31 A valuable research question then ethically requires validity and rigor in research design and implementation to produce valid, reliable, interpretable, and generalizable results. Poorly designed researchdfor

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

24 TABLE 2.2

2. ETHICAL PRINCIPLES

Ethical Framework for Clinical Research

Principles of Ethical Clinical Research

Description

Value

Research poses a clinically, scientifically, or socially valuable question that will contribute to generalizable knowledge about health or will be useful in improving health. Research is responsive to health needs and priorities.

Validity

The study has an appropriate and feasible design and end points, rigorous methods, and a feasible strategy to ensure valid and interpretable data.

Fair subject selection

The process and outcomes of subject and site selection are fair and are based on scientific appropriateness, minimization of vulnerability and risk, and maximization of benefit.

Favorable risk/benefit ratio

Study risks are justified by potential benefits and the value of the knowledge. Risks are minimized and benefits are enhanced to the extent possible.

Independent review

Independent evaluation of adherence to ethical guidelines in the design, conduct, and analysis of research.

Informed consent

Clear processes for providing adequate information to and promoting the voluntary enrollment of subjects.

Respect for enrolled participants

Study shows respects for the rights and welfare of participants both during and at the conclusion of research.

example, with an inappropriate design, inadequate power, insufficient or sloppy data, or inappropriate or unfeasible methodsdis harmful because human and material resources are wasted and individuals are exposed to risk for no benefit.30

Fair Subject Selection Fair subject selection requires that subjects be chosen for participation in clinical research based first on the scientific question, balanced by considerations of risk, benefit, and vulnerability. As described in the Belmont Report, fairness in both the processes and the outcomes of subject selection prevents exploitation of vulnerable individuals and populations and promotes equitable distribution of research burdens and benefits. Fair

procedures means that investigators should identify groups or individuals who would be appropriate for scientific reasonsdthat is, for reasons related to the problem being studied and justified by the design and the particular questions being askeddnot because of their easy availability or manipulability, or because subjects are favored or disfavored.11 Extra care should be taken to justify the inclusion of vulnerable subjects, as well as to justify excluding those who stand to benefit from participation. Exclusion without adequate justification can be unfair; therefore, eligibility criteria should be as broad as possible, consistent with the scientific objectives and the anticipated risks of the research. Distributive justice is concerned with a fair distribution of benefits and burdens, thus expected benefit and burden in a particular study is an important consideration for subject selection. Scientifically appropriate individuals or groups may be fairly selected consistent with attention to equitably distributing benefits and burdens, as well as minimizing risks and maximizing benefits. Persons are considered vulnerable when their ability to protect or promote their own interests is compromised, often because of an impaired capacity to provide informed consent. Although disagreement remains about the meaning of vulnerability in research and who is actually vulnerable,32 there is support for the idea that among scientifically appropriate subjects, the less vulnerable should be selected first. For example, an early drug safety study should be conducted with adults before children, and with consenting adults before including those who cannot consent. Certain groups, such as pregnant women, fetuses, prisoners, and children, are further protected by specific regulations requiring additional safeguards in research. According to US regulations, determination of the permissibility of research with children depends on the level of research risk and the anticipated benefits. Accordingly, (1) research that poses minimal risk to children is acceptable, (2) research with more than minimal risk must be counterbalanced by a prospect of direct therapeutic benefit for the children in the study, (3) for research with small amounts of additional risk (minor increment over minimal), but without the prospect of direct therapeutic benefit for the children can sometimes be justified by the importance of the question for children with the disorder under study, or (4) research without a prospect of benefit that poses greater than minimal risk to participating children can only be conducted if approved by a special panel convened by the US Secretary of the DHHS.33 Enrolling children in research also requires permission from their parents or legal guardians, along with the child’s assent whenever possible. Fair subject selection also requires considering the outcomes of subject selection. For example, if women, minorities, or children are not included in studies of a

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

ETHICAL FRAMEWORK FOR CLINICAL RESEARCH

particular intervention, then study results may be difficult to apply to these groups in practice, and interventions could actually be harmful. Therefore, study populations recruited for research should be representative of the populations likely to use the strategies tested in the research.34 Similarly, it has been argued that justice requires subjects to be among the beneficiaries of research. This means that subjects should be selected as participants in research from which they or others like them can benefit and should not be asked to bear the burdens of research from which they can reap no benefits. This understanding of justice has raised important and challenging questions in the conduct of collaborative international research. Some have argued that if an experimental drug or vaccine is found effective in a certain tested population, there should be prior assurance that population will have access to the drug or vaccine.35 Alternatively, subjects or communities should be assured of and involved in negotiation about fair benefits derived from research that are not necessarily limited to the benefits of available products of research.36

Favorable Risk/Benefit Ratio The ratio of risks to benefits in research is favorable when risks are justified by benefits to participants or to society, and when research is designed in a way that minimizes risk and enhances benefit for participating subjects. The ethical principle of beneficence obliges that we (1) protect people from deliberate or unnecessary harm and (2) maximize possible benefits and minimize possible harms. A widely accepted principle states that one should not deliberately harm another individual regardless of the benefits that might be made available to others as a result. However, as the Belmont Report reminds us, offering benefit to people and avoiding harm requires learning what is of benefit and what is harmful, even if in the process some people may be exposed to the risk of harm. To a great extent, clinical research is an activity designed to learn about the benefits and harms of unproven methods of diagnosing, preventing, treating, and caring for human beings. The challenge for clinical investigators and review/oversight groups is to decide in advance when it is justifiable to seek certain benefits despite the research risks, what level of risk is acceptable, whether risks have been minimized to the extent possible, and when it is better to forego the possible benefits because of the risks. This is called a risk/benefit assessment. The calculation and weighing of risks and benefits in research can be complicated. When designing a study, investigators consider whether the inherent risks are

25

justified by the expected value of the information and any possible benefit to the participants. Studies should be designed so that risks to participants are minimized and benefits are enhanced. When reviewing a study, an IRB identifies possible risks and benefits and determines whether the relationship of risks to benefits is favorable enough that the proposed study should go forward or instead be modified or rejected. When reviewing studies with little or no expected benefit for individual subjects, the IRB determines whether the anticipated risks or burdens to study subjects are justified only by the potential value of the knowledge to be gained, a particularly challenging risk/benefit assessment. Prospective subjects make their own risk/benefit assessment of whether the risks of participating in a given study are acceptable to them and are worth their participation. A risk/benefit assessment can include consideration of many types of risks and benefits, including physical, psychological, social, economic, and legal. For example, in a genetics study, physical risks may be limited to a blood draw or a buccal swab, so assessment of potential psychological and social risks is more important. Investigators, reviewers, and potential subjects may not only have dissimilar perspectives about research but also are likely to assign different weights to risks and benefits. For example, IRBs consider only health-related benefits of the research in justifying risks, whereas subjects are likely to consider access to care and financial compensation as important benefits that may tip the balance in favor of participation. Acknowledging that risk/benefit assessment is not a straightforward or easy process does not in any way diminish its importance. An important step in evaluating the ethics of clinical research involves not only careful attention to potential benefits to individuals or society of a particular study in relation to its risks, but also consideration of the risks of not conducting the research.

Independent Review Independent review is a process that allows evaluation of the research for adherence to established ethical guidelines by individuals with varied expertise and no personal or business interests in the research. For most clinical research, this independent review is carried out by an IRB or research ethics committee (REC). Using criteria detailed in US federal regulations,12,22 IRBs evaluate the value of doing the study, the risks involved, the fairness of subject selection, whether the risks have been sufficiently minimized and are justified, and the plans for obtaining informed consent; they then decide whether to approve a study, with or without modifications, to table a proposal for major revisions or more

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

26

2. ETHICAL PRINCIPLES

information, or to disapprove a study as unacceptable (See also Chapter 4). Independent review of the risks of proposed research by someone other than the investigator has been described as a “central protection for research participants.”37 Nonetheless, there is concern that the current IRB system in the United States is outdated given the current profile of clinical research, and also is bureaucratic, beset with conflicts, and in need of reform.38 Both the 2017 revisions to the Common Rule and recent NIH policy require single IRB review for domestic multisite studies.24,39

TABLE 2.3 The Process of Informed Consent Informed Consent Elements Description

Considerations and Challenges

Disclosure of information

Information about the study based on a “reasonable” person standard is disclosed to prospective participants. Disclosure takes into account subjects’ language, education, familiarity with research, and cultural values. Both written information and discussion are usually provided.

There is a need to balance the goal of being comprehensive with attention to the amount and complexity of information, to give participants the information they need and facilitate understanding.

Understanding

Knowledge of study purpose, risks, benefits, alternatives, and requirements.

Empirical data show that participants often do not have a good understanding of the details of the research.

Voluntary decision- Free from coercion making and undue influence. Subject is free to choose not to enroll.

Many possible influences affect participants’ decisions about enrolling in research. Avoid controlling influences.

Informed Consent Once a proposal is deemed valuable, valid, with acceptable risks in relation to benefits and fair subject selection, individuals are recruited and are asked to give their informed consent. The process of informed consent shows respect for persons and their autonomy, giving prospective subjects the opportunity to make autonomous decisions about participating and remaining in research, and respecting their choices about participation. We show lack of respect when we do not provide the necessary information to make a considered judgment, pressure an individual to make a particular judgment or deny him or her the freedom to act on judgments. The process of informed consent involves: disclosure of study information, comprehension of the information, voluntariness with respect to the decision, and authorization40 (Table 2.3). Information provided to subjects about a research study should be adequate, according to a “reasonable volunteer” standard, balanced, and presented in an understandable manner. Information should be provided in the language of the subject, at an appropriate level of complexity given the subject’s age, educational level, and culture. US federal regulations detail the types of information that should be included in informed consent12,22; this is essentially information that a reasonable person needs to know to make an informed decision about initial or ongoing research participation. Ideally, individuals receive the necessary information, understand it, process it in the context of their own situation and life experiences, and make a “voluntary” choice free from coercion or undue influence. The process of initial research informed consent usually culminates with the signing of a consent form. However, respect for persons requires that subjects continue to be informed throughout a study and are free to modify or withdraw their consent at any time. Although widely accepted as central to the ethical conduct of research, achieving informed consent is challenging. Determining the appropriate amount and complexity of information for disclosure is not straightforward. Written consent documents have become long

Authorization

Usually given by a For some individuals or signature on a written communities, requiring consent document. a signature reflects lack of appreciation for their culture or literacy level.

and complex, and large amounts of information may actually hinder subject understanding. Scientific information is often complex; research methods are unfamiliar to many people; and subjects have varying levels of education, understanding of science, and knowledge about their diseases and treatments, and are dissimilar in their willingness to enter into dialogue. Besides the amount and detail of information, understanding may be influenced by who presents the information and the setting. In some cases, information may be more accessible to potential subjects if presented in group sessions or through print, video, or other media presentations. Determining whether a subject has the capacity to consent and understands the particular study information is challenging. Capacity to provide consent is study specific. Individuals who are challenged in some areas of decision-making may still be capable of consenting to a particular research study. Similarly, individuals may not have the capacity to consent to a particular study, even if generally able to function in other areas of their lives. Assessing capacity might take into account an individual’s educational level and familiarity with science

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

ETHICAL CONSIDERATIONS IN RANDOMIZED CONTROLLED TRIALS

and research, as well as evidence of cognitive or decisional impairment. In some but not all cases, mental illness, depression, sickness, desperation, or pain may interfere with a person’s capacity to understand or process information. Empirical research on informed consent shows that participants do not always have a good understanding of the purpose or potential risks of the research studies for which they gave their consent.41 Informed consent to participation in research should be voluntary, and free of controlling influences, coercion and undue influence.40 Terminal or chronic illness, exhaustion of other treatment options, and lack of health insurance may limit a participant’s options but do not necessarily render decisions involuntary. Payment and other incentives, trust in health care providers, dependence on the care of clinicians, family pressures, and other factors commonly influence decisions about research participation. Most of the time, these are acceptable influences, but some worry that under certain circumstances, they can become controlling. Given these multiple factors, it is important to ensure that prospective subjects have and perceive that they have the option to say no to research participation and to do so with impunity. Research has demonstrated that active and ongoing dialogue and discussion between the research team and subjects, opportunities to have questions answered, waiting periods between the presentation of information and the actual decision to participate, the opportunity to consult with family members and trusted others, a clear understanding of alternatives, and other strategies can serve to enhance the process of informed consent.42,43

Respect for Enrolled Subjects Research participants deserve continued respect after enrollment, throughout the duration of the study, and when the study ends. Respect for subjects is demonstrated through appropriate clinical monitoring and attention to participants’ well-being throughout the study. Adverse effects of research interventions and any research-related injuries should be treated. Private information collected about subjects should be handled confidentially, and participants informed about the limits of confidentiality. Research subjects should be reminded of their right to withdraw from the research at any time without penalty. A change in clinical status or life circumstances, as well as new information from the study or other studies, may be relevant to a person’s willingness to continue participation. Investigators should make plans regarding the end of the trial, including participants’ continued access to successful interventions when indicated and to study results after the study is finished.

27

In summary, ethical clinical research is conducted according to the seven principles delineated in Table 2.2. Application of these principles to specific cases will always involve judgment and specification on the part of investigators, sponsors, review boards, and others involved in clinical research.

ETHICAL CONSIDERATIONS IN RANDOMIZED CONTROLLED TRIALS RCTs remain the principal method and “gold standard” for demonstrating safety and efficacy in the development of new drugs and biologics, and other interventions. An RCT has several characteristic features: RCTs are controlled, randomized, and usually blinded, and the significance of the results is determined statistically according to a predetermined algorithm. An RCT typically involves comparison of two or more interventions (e.g., Drug A vs. Drug B) to demonstrate that they are similar or that one is superior in the treatment, diagnosis, or prevention of a specific disorder. RCTs present a spectrum of unique ethical problems (Table 2.4). “In considering the RCT, the average IRB member must be baffled by its complexity and by the manifold problems it represents.”44 The ethical justification to begin an RCT is usually described as that of “an honest null hypothesis,” also often referred to as equipoise or clinical equipoise.45 In an RCT comparing interventions A and B, clinical equipoise is satisfied if there is no convincing evidence about the relative merits of A and B (e.g., evidence that A is more effective than or less toxic than B). The goal of an RCT is to provide credible evidence about the relative value of each intervention. Equipoise rests on a therapeutic commitment that patients should not receive a treatment known in advance to be inferior, nor should they be denied effective treatment that is otherwise available. Doubt based on lack of evidence about which intervention is superior justifies giving subjects an equal chance to get either one and makes it ethically acceptable to assign half or some portion of subjects to different treatments provided in an RCT. There remains some disagreement about the meaning, justification, and application of equipoise in clinical research. Some argue that equipoise is based on a mistaken confluence of research with therapy and therefore should be abandoned.46 Another controversy in RCTs involves what should count as “convincing” evidence. Some worry that the common acceptance of statistical significance at the P ¼ .05 level potentially discounts clinically significant observations. Statisticians recently criticized overreliance and misuse of the p-value, reiterating that it cannot

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

28 TABLE 2.4

2. ETHICAL PRINCIPLES

Selected Ethical Considerations in Randomized Controlled Trials (RCTs)

Features of RCTs

Description

Considerations

Equipoise

No convincing evidence that one intervention is better (i.e., more effective or less toxic) than another.

How to factor in early evidence? Does a requirement for equipoise conflate research and therapy?

Choice of control

Appropriate choice of control is necessary for scientific validity and generalizability.

Choice of control is not simply a scientific decision. Placebos as controls require ethical justification.

Randomization

Random assignment decreases bias and controls for many factors.

Random assignment does not allow for autonomous preferences.

Blinding

Single or double blinding is often used to decreases bias.

Research participants consent to temporarily suspend knowledge of which intervention they are receiving. In rare cases, a blind may need to be broken to manage certain clinical problems.

Sharing preliminary information

As evidence accumulates information about risks and benefits may change, and equipoise may be disturbed.

Study monitors, independent data and safety monitoring committees monitor data to help determine when the study should be stopped or altered, or information should be shared with participants.

tell you the probability that results are true or due to random chance, but only the probability of seeing results given a particular hypothetical explanation.47 People also disagree about the extent to which preliminary data, data from previous studies, data from uncontrolled studies and pilot studies, and historical data do or should influence the balance of evidence. In some cases, the existence of these other types of data may make equipoise impossible. However, data from small, uncontrolled, or observational studies can lead to false or inconclusive impressions about safety or efficacy. RCTs are usually monitored by data and safety monitoring committees who see data at specified time points during the trial and can recommend altering or stopping a trial based on prespecified boundaries for safety, efficacy, or futility.48

Another important scientific and ethical consideration in RCTs is the selection of outcome variables by which the relative merits of an intervention will be determined. Different conclusions may be reached depending on whether the efficacy of an intervention is a measure of survival or of tumor shrinkage, symptoms, surrogate end points, quality of life, or some composite measure. The choice of end points in a clinical trial is never simply a scientific decision. In an RCT, subjects are assigned to treatment through a process of randomization, rather than on the basis of individual needs and characteristics. The goal of random assignment is to control for confounding variables by keeping two or more treatment arms similar in relevant and otherwise uncontrollable aspects. Also, RCTs are often single blind (subject does not know which intervention he or she is receiving) or double blind (both subject and investigator are blinded to the intervention). Random assignment and blinding are methods used in clinical trials to reduce bias and enhance study validity. Although compatible with the goals of an RCT, random assignment to treatment and blinding to treatment assignment may seem incompatible with the best interests or autonomy interests of the patient-subject. In some placebo-controlled blinded studies, both subjects and investigators can guess (often because of side effects) whether they are receiving active drug or placebo, potentially thwarting the goal of reducing bias.49 The necessity and adequacy of blinding and randomization should be assessed in the design and review of each proposed research protocol. When randomization and blinding are deemed useful and appropriate for a particular protocol, two ethical concerns remain: (1) preferences for an intervention and information about which intervention a subject is receiving may be relevant to autonomous decisions and (2) information about which intervention the subject is receiving may be important in managing an adverse event or a medical emergency. With respect to the first concern, subjects should be informed about the purpose of the research and should be asked to consent to random assignment and a temporary suspension of knowledge about which intervention they are receiving. To balance the need for scientific objectivity with respect for a research subject’s need for information to make autonomous decisions, investigators should provide subjects with adequate information about the purpose and methods of randomization and blinding. Subjects are asked to consent to a suspension of knowledge about their treatment assignment until completion of the protocol or some other predetermined point, at which time they should be informed about which intervention they received in the clinical trial. In some cases, knowledge of which medications a subject is receiving may be important in the treatment

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

29

SUMMARY QUESTIONS

of adverse events or other medical emergencies. To balance the need for scientific objectivity with concern for subject safety, investigators should consider in advance the conditions under which a blind may be broken to treat an adverse event. Specifically, the protocol should specify where the code will be located, the circumstances (if any) under which the code will be broken, who will break it, how the information will be handled (i.e., will the investigator, the subject, the IRB, and the treating physician be informed?), and how breaking of a blind will influence the analysis of data. Research subjects should always have information about whom they should contact in the event of an emergency. The IRB should be satisfied that these plans provide adequate protection for patient safety. Plans also should be made for what will happen at the end of a trial. Some argue that those who volunteer for RCTs, especially in externally sponsored international research, deserve assurance in advance about access to interventions proven to be beneficial in the RCT. Investigators should plan for whether and how subjects randomized to an intervention that is benefiting them will continue to receive that intervention, and how those randomized to the inferior intervention might be given an opportunity to receive the better one. Considerable disagreement remains regarding the extent of the obligation of researchers or sponsors to ensure posttrial access. A participant may be concerned about participating in an RCT if one of the potential treatment assignments is placebo. Some people perceive randomization to placebo in clinical trials as problematic because it potentially deprives the individual of treatment that he or she may need. On the other hand, without proof of the safety and efficacy of an experimental treatment, it is possible that those randomized to placebo are simply deprived of potentially toxic side effects or of a useless substance.50 Scientifically, comparison of an experimental drug to placebo can allow efficient and rigorous establishment of efficacy. The alternative is an RCT that compares the investigational drug to an already established therapy, if one exists, which can be designed to test superiority or noninferiority of the two agents (i.e., the experimental drug is similar to the standard therapy control within a noninferiority margin). Some authors suggest that both scientific design and possible risk to subjects should be determinants of the acceptability of placebo.51 Most accept that the use of a placebo control in research is justified when (1) there is no proven effective treatment for the condition under study; (2) withholding treatment poses negligible risks to participants; (3) there are compelling methodological reasons for using placebo, and withholding treatment does not pose a risk of serious harm to participants; and, more controversially, (4) there are compelling methodological

reasons for using placebo, the research is intended to develop interventions that can be implemented in the population from which trial participants are drawn, and the trial does not require participants to forgo treatment they would otherwise receive.52 Most agree, however, that if the outcome for the patient of no treatment or placebo treatment is death, disability, or serious morbidity, a placebo control should not be used.53

CONCLUSION Ethical principles and guidance related to the conduct of clinical research with human participants help to minimize the possibility of exploitation and promote respect for and protection of the rights and welfare of individuals who serve as human subjects of research. This chapter has reviewed the historical evolution of research ethics, a systematic ethical framework for the conduct of clinical research, and ethical considerations of some of the unique features of RCTs. In addition to adherence to principles, codes of ethics, and regulations, the ethical conduct of human clinical research depends on the thoughtfulness, integrity, and sagacity of all involved.

SUMMARY QUESTIONS 1. Scientific validity is important to evaluating the ethics of clinical research. Without rigorous scientific validity, the research outcomes are not reliable so persons are unnecessarily asked to accept risk and burden. Assessing scientific validity includes consideration of: a. Sample size and study design b. Costs and budget c. Informed consent d. Amount of compensation to participants 2. Disclosure of which of the following items is necessary for an informed consent document? a. A statement that the study involves research and the study’s purpose b. An explanation of the proposed treatment or intervention and procedures c. The foreseeable risks and benefits of study participation d. All of the above 3. Although research participants are often exposed to some risk in clinical research, there are limits to the amount of acceptable risk. In evaluating risk in clinical research, it is commonly accepted that: a. Only known risks to participants are permitted b. Risks should be minimized and justified by the benefits or value of the study

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

30

2. ETHICAL PRINCIPLES

c. To compensate for possible risks, individuals must receive therapeutic benefit d. Risks are acceptable only when competent adults consent 4. In the proposed ethical framework of seven principles for evaluating clinical research, the final principle “respect for enrolled subjects” is understood to include at least: a. Establishing a contract between the subject and the researcher b. Monitoring the subject’s welfare and protecting confidentiality of information c. Keeping the financial costs of participation reasonable d. Informing the subject of new information only after the study is published 5. Most ethical research calls for the voluntary informed consent of the research participant. In informed consent, the research participant’s decision to participate in research is considered “voluntary” if it is free from: a. Any outside influences b. Payment or other inducements c. Coercion or undue influence d. Misunderstanding or cognitive impairment 6. Multiple codes and regulations provide guidance about the ethical conduct of research. One influential “living” document written by the World Medical Association has had considerable influence in developing national and local guidance. That document is: a. The Bill of Rights b. The Declaration of Helsinki c. The Nuremberg Code d. The Declaration of Lisbon

References 1. Brody H, Miller FG. The clinician-investigator: unavoidable but manageable tension. Kennedy Inst Ethics J 2003;13(4):329e45. 2. Eisenberg L. The social imperatives of medical research. Science 1977;198:1105e10. 3. Jonas H. Philosophical reflections on experimenting with human subjects. In: Freund P, editor. Experimentation with human subjects. New York: Braziller; 1970. 4. Kant as quoted in Beauchamp T. In: Childress J, editor. Principles of biomedical ethics. 4th ed. New York: Oxford University Press; 1994. p. 351. 5. Emanuel E, Grady C. Four paradigms of clinical research and research oversight. In: Emanuel G, Crouch L, Miller W, editors. Oxford textbook of clinical research ethics. NY: Oxford University Press; 2008. p. 222e30 [Chapter 22]. 6. US FDA. Significant dates in U.S. food and drug law history. Available at: http://www.fda.gov/AboutFDA/WhatWeDo/ History/Milestones/ucm128305.htm. 7. Rothman D. Ethics and human experimentationdHenry Beecher revisited. N Engl J Med 1987;317:1195e9.

8. Wiendling P, The Nazi medical experiments, Annas G, Grodin M. The Nuremberg code [Chapters 2 and 12 respectively in Emanuel]. In: Grady, Crouch, Lie, Miller, Wendler, editors. Oxford textbook of clinical research ethics. NY: Oxford University Press; 2008. p. 18e30. 136e140. 9. Beecher HK. Ethics and clinical research. N Engl J Med 1966;274: 1354e60. 10. Porter J, Koski G. Regulations for the protection of humans in research in the United States. In: Emanuel, Grady, Crouch, Lie, Miller, Wendler, editors. Oxford textbook of clinical research ethics. NY: Oxford University Press; 2008. p. 156e67 [Chapter 15]. 11. National Commission for the Protection of Human Subjects of Biomedical, Behavioral Research. The Belmont report: ethical principles and guidelines for the protection of human subjects of research. Washington, DC: U.S. Government Printing Office; 1979. 12. U.S. code of federal regulations title 45, part 46. Available at: www. hhs.gov/ohrp/humansubjects/guidance/45cfr46.htm. 13. National Research Council. The social impact of AIDS in the United States. Washington, DC: National Academy Press; 1993. 14. Herbert-Croteau N, Brisson J, Lemaire J, Latreille J. The benefit of participating to clinical research. Breast Cancer Treat Res 2005; 91(3):279e81. 15. Bleyer A, Montello M, Budd T, Saxman S. National survival trends of young adults with sarcoma: lack of progress is associated with lack of clinical trial participation. Cancer 2005;103(9):1891e7. 16. Dresser R. Wanted: single, white male for medical research. Hastings Cent Rep 1992;22(1):21e9. 17. National Institutes of Health. Guidelines for the inclusion of women and minorities as subjects in clinical research. NIH guide for grants and contracts. Bethesda, MD: National Institutes of Health; March 18, 1994. 18. National Institutes of Health. NIH policy and guidelines on the inclusion of children as participants in research involving human subjects. NIH guide for grants and contracts. Bethesda, MD: National Institutes of Health; March 6, 1998. 19. The Nuremberg Code. Available at: www.hhs.gov/ohrp/references/ nurcode.htm. 20. World Medical Assembly. Declaration of Helsinki. Available at: www.wma.net/e/ethicsunit/helsinki.htm. 21. Council for International Organizations of Medical Sciences. International ethical guidelines for biomedical research involving human subjects. Geneva: CIOMS/WHO; 2002. Available at: www. cioms.ch. 22. U.S. code of federal regulations title 21, part 50 “protection of human subjects” and part 56. 23. US FDA. Comparison of FDA and DHHS humans subject protection regulations. Available at: http://www.fda.gov/ScienceResearch/ SpecialTopics/RunningClinicalTrials/EducationalMaterials/ ucm112910.htm. 24. Department of Homeland Security, et al. Final rule. Federal policy for the protection of human subjects Fed Regist 2017;82(12): 7149e274. Available at: https://www.gpo.gov/fdsys/pkg/FR2017-01-19/pdf/2017-01058.pdf. 25. International conference on harmonisation of technical requirements for registration of pharmaceuticals for human use: guidelines for good clinical practice 1996: E6(R1). Available at: http:// www.ich.org/products/guidelines/efficacy/article/efficacyguidelines.html. 26. World Health Organization (WHO) handbook for good clinical research practice (GCP): guidance for implementation. Available at: http://apps.who.int/medicinedocs/documents/s14084e/ s14084e.pdf. 27. Emanuel E. Researching a bioethical question. In: Gallin JI, Ognibene FP, editors. Principles and practice of clinical research. 3rd ed. London: Elsevier Inc.; 2012. p. 31e42 [Chapter 3].

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

28. Largent E, Grady C, Miller F, Wertheimer A. Misconceptions about coercion and undue influence: reflections on the views of IRB members. Bioethics 2013;27(9):500e7. 29. Emanuel E, Wendler D, Grady C. What makes clinical research ethical? J Am Med Assoc 2000;283(20):2701e11. 30. Emanuel E, Wendler D, Grady C. An ethical framework for biomedical research. In: Emanuel, Grady, Crouch, Lie, Miller, Wendler, editors. Oxford textbook of clinical research ethics. NY: Oxford University Press; 2008. p. 123e35 [Chapter 11]. 31. Freedman B. Scientific value and validity as ethical requirements for research: a proposed explanation. IRB Rev Hum Subjects Res 1987;9(5):7e10. 32. Levine C, Faden R, Grady C, Hammerschmidt D, Eckenwiler L, Sugarman J. Consortium to Examine Clinical Research Ethics. The limitations of “vulnerability” as a protection for human research participants. Am J Bioeth 2004;4(3):44e9. 33. U.S. code of federal regulations. Title 45, part 46. Subpart D. 34. Meltzer L, Childress J. Fair participant selection. In: Emanuel G, Crouch L, Miller W, editors. Oxford textbook of clinical research ethics. NY: Oxford University Press; 2008. p. 377e85 [Chapter 35]. 35. Emanuel E. Benefits to host countries. In: Emanuel, Grady, Crouch, Lie, Miller, Wendler, editors. Oxford textbook of clinical research ethics. NY: Oxford University Press; 2008. p. 719e28 [Chapter 65]. 36. Participants in the 2001 Conference on Ethical Aspects of Research in Developing Countries. Fair benefits for research in developing countries. Science 2002;298:2133e4. 37. National Bioethics Advisory Commission. Ethical and policy issues in research involving human participants: vol. 1. Report and recommendations. Available at: www.bioethics.gov/reports/past_ commissions/nbac_human_part.pdf. 38. Emanuel E, Wood A, Fleischman A, et al. Oversight of human participants research: identifying problems to evaluate reform proposals. Ann Intern Med 2004;141(4):282e91. 39. National Institutes of Health. Final NIH policy on the use of a single institutional review board for multi-site research; 2016. Available at: http://grants.nih.gov/grants/guide/notice-files/NOT-OD-16094.html.

31

40. Faden R, Beauchamp T. A history and theory of informed consent. New York: Oxford University Press; 1986. 41. Mandava A, Pace C, Campbell B, Emanuel E, Grady C. The quality of informed consent: mapping the landscape. A review of empirical data from developing and developed countries. J Med Ethics 2012;38:356e65. 42. Flory J, Emanuel E. Interventions to improve research participants’ understanding in informed consent for research: a systematic review. J Am Med Assoc 2004;292:1593e601. 43. Nishimura A, Carey J, Erwin PJ, Tilburt JC, Murad MH, McCormick JB. Improving understanding in the research informed consent process: a systematic review of 54 interventions tested in randomized control trials. BMC Med Ethics July 23, 2013:14e28. 44. Levine R. Ethics and regulation of clinical research. 2nd ed. Baltimore: Urban & Schwarzenberg; 1986. 45. Freedman B. Equipoise and the ethics of clinical research. N Engl J Med 1987;317(3):141e5. 46. Miller F, Brody H. A critique of clinical equipoise: therapeutic misconception in the context of clinical trials. Hastings Cent Rep 2003;33(3):20e8. 47. Wasserstein RL, Lazar NA. The ASA’s statement on p-values: context, process, and purpose, the American Statistician. To link to this article. 2016. http://dx.doi.org/10.1080/00031305.2016.1154108. 48. FDA. The Establishment and operation of clinical trial data monitoring committees for clinical trial sponsors. Available at: http:// www.fda.gov/regulatoryinformation/guidances/ucm127069.htm. 49. Fisher S, Greenberg R. How sound is the double-blind design for evaluating psychotropic drugs? Nerv Ment Dis 1993;181(6):345e50. 50. Levine R. The use of placebos in randomized clinical trials. IRB Rev Hum Subjects Res 1985;7(2):1e4. 51. Emanuel EJ, Miller FG. The ethics of placebo-controlled trialsda middle ground. N Engl J Med 2001;345(12):915e9. 52. Millum J, Grady C. The ethics of placebo-controlled trials: methodological justifications. Contemp Clin Trials September 12, 2013;36(2): 510e4. 53. Miller F, Brody H. What makes placebo-controlled trials unethical? Am J Bioeth 2002;2(2):3e9.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

3 Integrity in Research: Principles for the Conduct of Research Melissa C. Colbert, Robert B. Nussenblatty, Michael M. Gottesman National Institutes of Health, Bethesda, MD, United States

O U T L I N E Guidelines and Principles for the Conduct of Research33

Peer Review

41

Scientific Integrity and Research Misconduct

Publication Practices, Responsible Authorship, and Results Reproducibility Publication Practices Authorship Reproducibility

42 42 42 43

34

Responsibilities of Research Supervisors and Trainees 36 Data Management, Archiving, and Sharing Data Management Archiving Data Sharing

36 36 37 38

Study Questions

45

Research Involving Human and Animal Subjects

38

Acknowledgments

45

Collaborative and Team Science

39

References

45

Conflict of Interest and Commitment

40

Further Reading

46

GUIDELINES AND PRINCIPLES FOR THE CONDUCT OF RESEARCH

scientific freedom and creativity. The writers of the Guidelines tried to take into account the major differences in commonly accepted behaviors among different scientific disciplines. The initial version was issued in 1990, and it has subsequently been revised and reissued several times.1 In the latest version, important NIH policies have been added to what is now called the “Guidelines and Policies for the Conduct of Research in the Intramural Research Program at NIH” including how requirements pertaining to embryonic and fetal tissue research, creating a diverse and inclusive workforce and human subject research protections. The Guidelines serve as a framework for the education of NIH scientific staff in research conduct issues,

In the late 1980s, the leadership of the National Institutes of Health (NIH) Intramural Research Program (IRP) decided to develop a set of guidelines for the conduct of research at NIH that could be used as a basis of discussion, as well as education, of all scientific staff including those in training. The Guidelines for the Conduct of Research in the Intramural Research Program at NIH (referred to as the Guidelines) were “developed to promote the highest ethical standards in the conduct of research by intramural scientists at NIH.” The intent was to provide a framework for the ethical conduct of research without inhibiting

y

Deceased.

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00003-4

33

Copyright © 2018. Published by Elsevier Inc.

34

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

through discussion sessions and more formal courses, as well as a reference book. In 1995, the NIH Committee on Scientific Conduct and Ethics (CSCE) was established for the IRP to help set policies on these issues, as well as to set in place mechanisms for teaching the principles of scientific conduct and to establish mechanisms to resolve specific cases. This CSCE has been responsible for the last four versions of the Guidelines. They also created a computer-based Research Ethics course that all new scientific staff must complete, to ensure that everyone has the same basic understanding of the policies and regulations governing the responsible conduct of research, available to the public.2 Finally, the CSCE selects the topic, and interesting case studies, for yearly research ethics discussions in which all scientific staff participate.3 In addition to the Guidelines, NIH has other Guides, such as Sharing Research Resources, Standards for Clinical Research within the IRP, Human Biospecimen Storage and Tracking, Scientific Record Keeping, Training and Mentoring, Handling Research Misconduct Allegations and Food and Drug Administration (FDA) Amendments Act (FDAAA) Reporting of Research Results, which have been collected on a convenient public location: the NIH Sourcebook.4 Other institutions develop policies for the conduct of research for their investigators, which are specific for their needs. Books, textbooks, and symposia or colloquia proceedings5e7 that address scientific conduct and/or misconduct, as well as internet-based learning programs at many institutions,8e12 have increased. As a result of the mandate from the Office of Science and Technology Policy in the White House for the Office of Research Integrity (ORI), Department of Health and Human Services, to become primarily an educational office, ORI has been funding grants to support institutions in the development of research conduct materials and courses that can be made available widely to any institution interested in using them.13 The NIH Guidelines cover research integrity; mentore trainee relationships; data management, sharing and archiving; research involving human and animal subjects; collaborations and team science; conflict of interest and commitment; peer review and privileged information; publication practices; responsible authorship and results reproducibility; social responsibilities; and dual use research, among other issues and policies. These enumerated topics form the basis of the remainder of the chapter.

SCIENTIFIC INTEGRITY AND RESEARCH MISCONDUCT Scientists at the NIH, like scientists everywhere, should be committed to the responsible conduct of

research and the scientific method in seeking new knowledge. We expect that all research staff in the NIH IRP will maintain exemplary standards of intellectual honesty in designing, conducting, and presenting research as befits the leadership role of the NIH. The principles of the scientific method include formulation and testing of hypotheses, controlled observations or experiments, analysis and interpretation of data, and oral and written presentations of all of these components to scientific colleagues for discussion and further conclusions. The scientific community and the general public rightly expect adherence to exemplary standards of intellectual honesty in the formulation, conduct, and reporting of scientific research. Research integrity must form the foundation for the conduct of science, which underpins the reputation of the scientific community and supports confidence of the general public. Research misconduct undermines this basic foundation and erodes the public’s trust. The issue of research misconduct became one of interest to the public in the 1980s as a result of several cases involving high-profile scientists. In response to this, the Institute of Medicine (IOM) convened a committee, under the chairmanship of Dr. Arthur Rubenstein, to examine the issues. In 1989, this committee drafted and published “The Responsible Conduct of Research in the Health Sciences”.14 The IOM revisited the topic in 2001, again chaired by Dr. Rubenstein, resulting in a second report, “Integrity in Scientific Research, Creating an Environment that Promotes Responsible Conduct” the following year.15 These reports recognize the dangers of research misconduct and other egregious behaviors. They proposed that institutions as well as scientists develop standards for the ethical conduct of research to focus on promoting a research environment that values integrity and reproducibility in highquality research and diminish research misconduct. In 2015, the NIH made a marked change in its responsible conduct of research training requirements emphasizing that research ethics and integrity are the foundations for all good science and form a natural part of the culture and daily life at NIH. In addition to the required online training modules (available to the general public for many years), in-person training and an introduction to research ethics were extended to our junior scientists and students. Courses and interactive presentations for postdoctoral and visiting fellows now provide more options, and several different media modalities, including video vignettes and recorded workshops covering potentials and pitfalls in modern technologies for cell and structural biology as well as genome technology, emphasize reproducibility.16 In spite of genuine efforts and the good examples of investigators, cases of research misconduct persist.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

SCIENTIFIC INTEGRITY AND RESEARCH MISCONDUCT

Of concern, the trend has been increasing nationwide. The number of retraction of scientific papers, whether for findings of research misconduct or error, has risen sharply. Reports in 2012 suggest that research misconduct was often associated with greater than 60% of the retractions.17 This begs the question: has research misconduct actually increased or are we just more aware of the problem? In the past several years, the internet has become a platform for watchdogs, anonymous allegations, and widespread discussions of suspected research misconduct. Retraction Watch, a forum started in 2010 by Ivan Oransky and Adam Marcus18 makes an effort to widely publicize cases of research misconduct. Their work over the past 6 years has shown that many more retractions occur annually than was once thought. Although it identifies misconduct, the distinction between misconduct and research errors is not always made. A recent publication looking at retractions over several decades analyzed the effects on scholarly impact on authors and institutions from which the retractions came, and whether this translated to disciplines investigated. The results indicated negative outcome on author stature, particularly if the retraction was due to research misconduct, but only limited effect on the areas of scientific study.19 PubPeer, another web blog invites discussion and comment on recent publications.20 While a worthy undertaking, it appears to have devolved into what often seems an overzealous analysis of figures, particularly Western blots or other types of gel-based analyses, similar to the suspended website Science Fraud. Many of the postings are anonymous, highly critical and obliquely allege misconduct, as opposed to promoting thoughtful comment and critique. In the late 1990s and in early 2000, the federal policy defining research misconduct was released. This policy defined certain standards to be followed when reporting a finding of misconduct and describes a three-step process for assessing and then establishing whether research misconduct occurred (Table 3.1). The most recent NIH Intramural Policies and Procedures for Research Misconduct Proceeding is consistent with U.S. Public Health Service (PHS) regulations 42 CFR part 93. The PHS policy was the first universally applicable guidance for federally supported research. The IRP also provides an abridged Guide to the Handling of Research Misconduct Allegations as a pdf, available on the NIH Sourcebook.21 One of the most difficult and controversial aspects of finding research misconduct involves intent. Misconduct must be found to have been committed intentionally, knowingly, or recklessly. The ORI and other individuals and organizations are beginning to deal critically with how to define reckless behavior in

TABLE 3.1

35

Federal Definition of Scientific Misconduct, Standards, and Process by Which It Is Assessed

I. RESEARCH MISCONDUCT DEFINED Research misconduct is defined as fabrication, falsification, or plagiarism in proposing, performing, or reviewing research, or in reporting research results. Fabrication is making up data or results and recording or reporting them. Falsification is manipulating research materials, equipment, or processes, or changing or omitting data or results such that the research is not accurately represented in the research record. Plagiarism is the appropriation of another person’s ideas, processes, results, or words without giving appropriate credit. Research misconduct does not include honest error or differences of opinion. II. FINDINGS OF RESEARCH MISCONDUCT A finding of research misconduct requires that: There be a significant departure from accepted practices of the relevant research community; The misconduct be committed intentionally, or knowingly, or recklessly; and The allegation be proven by a preponderance of the evidence. III. PROCESS FOR ASSESSING OCCURRENCE OF RESEARCH MISCONDUCT Allegation Assessmentddetermination of whether allegations of misconduct, if true, would constitute misconduct and whether the information is sufficiently specific to warrant and enable an inquiry Inquirydthe process of gathering information and initial fact-finding to determine whether an allegation of misconduct warrants an investigation Investigationdthe formal examination and evaluation of all relevant facts to determine if scientific misconduct has occurred, and if so, to determine the person(s) who committed it and the seriousness of the misconduct.

scientific research and its role in research misconduct. “Recklessness” lies along a continuum of intent, moving from negligent to reckless to knowing and finally intentional. An important consideration is the standard of what a reasonable person would do under similar circumstances and what is acceptable to the relevant scientific community. Evaluation of context and of risk is at the heart of the process. ORI is planning discussions with legal experts and research integrity officers to establish guidance on interpretation of a standard for a decision of recklessness. Institutions may handle imposition of sanctions and appeal processes within certain guidelines. For the Department of Health and Human Services, the policy provides for ORI oversight of completed investigations.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

36

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

RESPONSIBILITIES OF RESEARCH SUPERVISORS AND TRAINEES NIH’s mission is to improve the health of the public through support of biomedical research as well as the training of biomedical scientists. The quality of research as well as the training provided students depends in large part on the relationship between the mentor and trainees in each NIH laboratory. The goals of a mentoretrainee relationship are to ensure that fellows receive the best possible training in how to conduct research and how to develop and achieve career goals. Mentoring and being mentored are essential life-long components of professional life. Research supervisors should always be mentors, but in addition trainees should be encouraged to seek out other mentors who may provide additional expertisedtogether they form the basis of a professional network. Characteristics of a good mentor include an interest in contributing to the career development of another scientist, research accomplishments, professional networking, accessibility, and past successes in cultivating the professional development of their fellows. The trainees themselves must be committed to the work of the laboratory and the institution, to the achievement of their research and career goals, and to actively participate in their training. It is the responsibility of the mentor to provide a rich research environment in which the trainee has the opportunity to acquire both conceptual and technical skills of the field. In this setting, the trainee should be provided with clear expectations and undertake a significant piece of research, usually chosen as the result of discussion between supervisor and trainee. Good communication is critical to success, including time spent by the mentor in reviewing primary data with the trainee. Mentors should consider the overall size of the research group and perhaps limit the number of trainees for whom they can provide an appropriate and productive training experience. The use of a “Welcome Letter” or Lab Compact is a useful tool to introduce trainees to the specific expectations and responsibilities of both trainees and mentors in the lab.22 Such a document is gaining acceptance in many institutions.23e25 Among the skills that trainees should acquire during their fellowship period are training in scientific investigationdhow to choose a first-rate research project, how to carry out the necessary experiments and analyses in an appropriate and rigorous way, how to incorporate knowledge of the research field and published literaturedwith the ultimate goal of developing increasing independence throughout the training period; training in communication skills, both written and oral; training in personal interactions, including negotiations,

persuasion and diplomatic skills, and in networking; and training in scientific responsibility, the legal and ethical aspects of carrying out research. In addition, fellows should be considering career pathways, in consultation with their mentors, being sure to survey the many options available to scientists these days. To ensure a rich and stimulating laboratory experience, mentors should strive to establish a diverse, talented research group. Attention should be given to assure that all trainees and employees (at NIH) are valued and included as respected members of the scientific community. While respecting cultural differences of the community, at the laboratory level this includes preparing records and conducting laboratory business in English, as a common language of science.

DATA MANAGEMENT, ARCHIVING, AND SHARING Data Management Scientific data may be divided into three categories: experimental protocols; primary data, which include instrument setup and output, raw and processed data, statistical calculations, photographic images, electronic files, and patient records; and procedures of reduction and analysis. Any individual involved in the design or execution of an experiment and subsequent data processing is responsible for the accuracy of the resultant scientific data and must be meticulous in the acquisition and maintenance of them. These individuals may include, in addition to the person responsible for actually carrying out the experiment, the principal investigator, postdoctoral fellows, students, research assistants, and other support staff such as research nurses. Research results should be recorded in a form that allows continuous access for analysis and review, whether via an annotated bound notebook or computerized records. All research data must be made available to the supervisor, as well as collaborators, for immediate review. Data management, including the decision to publish, is ultimately the responsibility of the principal investigator. Martinson et al.26 carried out a survey that asked respondents to report which, if any, questionable research practices they had engaged in over the previous 3 years. Among those who responded (46% of those surveyed), 27.5% reported that they had kept inadequate research records, suggesting that lack of appropriate record keeping is a serious problem. A follow-up study in 2012 confirmed the surprisingly high prevalence of questionable practices, such as poor record keeping and suggested, unfortunately, this may be becoming more the norm.27

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

DATA MANAGEMENT, ARCHIVING, AND SHARING

Good science requires keeping good records, both in the research laboratory and in clinical research. Good record keeping facilitates communication within the research team and preparation of data for publication or submission of intellectual property for patents. Properly maintained research records are required by many federal agencies, particularly those that regulate radioactivity, animal use, and FDA-regulated products. The importance of record keeping and data quality in clinical research, where protocols use drugs, devices, or biologics are frequently audited by federal agencies, cannot be overstated (see also Chapter 30 on Data Management in Clinical Trials). In any clinical setting, there are clear distinctions between patient care and clinical research. Patient care records are part of their medical record and usually accompany the patient wherever they go. In clinical research, the investigator must follow the clinical protocol and document the experience of the research participant. These records are kept at the research site, need to be well ordered, and are often organized in a research or regulatory binder. Documentation usually includes such things as records of institutional review board (IRB) actions, drug/device accountability as well as other materials and records, which demonstrate the site follows Good Clinical Practices. Clinical data should be retained as directed by federal regulations (Table 3.2).

TABLE 3.2 Scientific Record Keeping I. REASONS WHY GOOD RECORDS ARE IMPORTANT IN SCIENTIFIC RESEARCH 1. Good record keeping is necessary for data analysis, publication, collaboration, peer review, and other research activities. 2. Good record keeping is required by the NIH to meet the accepted policies and standards for the conduct of good science. 3. Good record keeping is necessary to support intellectual property claims. 4. Good record keeping can help defend you against false allegations of research misconduct. 5. Good record keeping is important in the care of human subjects. II. RESEARCH RECORDS SHOULD DESCRIBE OR EXPLAIN THE FOLLOWING 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Name of the person making the Record What was done When the record was madedmonth, day, year The purpose of the research The project associated with the research The methodology involved The materials used The findings Interpretation of the findings Future plans Continued

TABLE 3.2

37

Scientific Record Keepingdcont’d

III. CLINICAL RECORDS 1. Medical record of patient care documents a. Why the patient is heredhistory/diagnosis? b. What was donedtreatments, procedures, tests? c. When was the action performed? d. Who performed the action or activity? e. What was the outcome of caredresponse, prognosis? 2. Clinical Research Records for a Regulatory Binder include a. IRB approved documentation b. Participant information c. Technology agreements d. Personnel documents (FDA study) e. Site monitoring (FDA study) f. Laboratory information (FDA study) g. Pharmacy documentation (FDA study) h. Study documentation (FDA study) i. FDA regulatory Documentation IV. RECORDS RETENTION 1. Research Data: a. Records of basic researchd7 years b. Records that support patent or invention rightsd30 years after patent filed c. Records of historical significancedtransferred to the National Archives maintained permanently 2. Clinical Research Data: a. Data: Subject to FDAAA i. Results deposited in ClinicalTrials.gov within 12 months after primary completion date ii. 2 years following the date a marketing application is approved iii. 2 years after the investigation is discontinued and the FDA is notified that no application is to be filed b. Data subject to NIH regulations i. Results deposited in ClinicalTrials.gov within 12 months after primary completion date IRB, institutional review board; FDA, Food and Drug Administration; FDAAA, FDA Amendments Act; NIH, National Institutes of Health.

Archiving At the NIH, all data collected, as well as laboratory notebooks, research records, and other supporting materials such as unique reagents, belong to the government and must be retained for a period of time sufficient to allow for further analysis of the results as well as repetition by others of published material. Intramural research records are the property of the NIH and must be maintained for a period of time as dictated. All records must be maintained at least 7 years after completion of the project. Records supporting intellectual property rights (patents or inventions) must be maintained for 30 years after the patent is filed. Records of historical significance are transferred to the National Archives and are maintained permanently.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

38

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

Data Sharing Once publications have appeared, supporting materials must be made available to all responsible scientists seeking further information or planning additional experiments when possible; for example, aliquots of any monoclonal antibody that derives from a continuously available cell line must be provided, whereas the final aliquots of a polyclonal antibody, needed by the original lab to finish additional experiments, do not. The NIH IRP, in line with other research institutions, has required that transgenic or knockout mouse lines be made available, preferably through deposition in a commercial mouse facility. Requests for human samples require IRB review and approval prior to sharing to ensure that confidentiality issues are covered. All primary research data in the IRP are subject to the Freedom of Information Act (FOIA). In 2014, the NIH instituted a policy that requires all grants and contracts to include a plan to share genomic data.28 This policy covers large-scale human and nonhuman genomic data, including genome-wide association studies, single-nucleotide polymorphisms arrays, genome sequence, transcriptomic, metagenomic, epigenomic, and gene expression data. This facilitates the opportunity for further understanding of factors that influence health and disease. NIH established a dedicated website where information is stored on data repositories, NIH-funded databases and NIH database collaborators.29 The FDAAA of 2007 was designed to improve public access to information about clinical trials of FDAregulated products and devices. In 2015, a Notice of Proposed Rule Making expanded FDAAA rules regarding registration of trials and results reporting. Concurrently, NIH issued a draft policy to promote broad and responsible dissemination of information on clinical trials funded by the NIH through registration and submission of summary results information to ClinicalTrials.gov. This policy expands the scope of studies covered, identifies which studies require results reporting, but also included a more cogent definition of what constitutes a clinical trial covered by the policy. As of this writing, these policy changes are still under review (see also Chapter 9). Information is available at a number of websites.30,31

RESEARCH INVOLVING HUMAN AND ANIMAL SUBJECTS The use of humans and animals in research is essential to the NIH mission for improving human health but such research entails special ethical and legal considerations. Many chapters in this textbook address the

issues related to carrying out human subject research and readers may wish to consult the Office of Human Subjects Research Protections, Policies and Procedures.32 While research ethics and training have traditionally focused on bench scientists, increasing educational efforts have been targeted to clinical and translational scientists. Translational science involving human subjects is subject to unique ethical issues that differ substantially from basic research. Given that clinical research is highly regulated by both the Office of Human Research Protection and the FDA, it is of concern that the survey by Martinson et al.26 reported that 0.3% of those responding said they had ignored major aspects of human subject requirements while 7.6% circumvented certain minor aspects. Organizations such as Public Responsibility in Medicine and Research (PRIM&R) and Collaborative Institutional Training Initiatives (CITI) once primarily concerned with education in protecting human subjects have increasingly sponsored more workshops, webinars, and online courses in the Responsible Conduct of Research for clinical scientists. CITI created a discipline-specific public access course with a module for biomedical research.33 Other online resources available include a Research Networking for clinical researchers supported through the Clinical and Translational Research Awards program.34 Several institutions also maintain online case libraries of issues related to ethical research with examples taken from clinical scenarios.34 The collection of human biospecimens (Table 3.3), a valuable and unique resource, must be handled under the highest ethical and scientific standards. Under the Notice of Published Rule Making, The Common Rule (45 CFR 46) may soon require a broad consent for secondary use of collected biospecimens, in which persons give consent to future unspecified research uses. NIH developed specific Guidelines for Human Biospecimen Storage and Tracking within the NIH IRP, updated in 2013.35 All samples, irrespective of whether they were obtained during standard of care or for research and regardless of whether individuals are still TABLE 3.3 Human Biospecimens Biological Materials or Derivatives Thereof Include 1. 2. 3. 4. 5. 6. 7. 8.

DNA Cells or cell lines Tissue (bone, muscle, connective tissue, skin) Organs, (heart, liver, bladder, kidney, etc.) Blood Gametes (sperm, ova) Embryos and fetal tissue Waste (urine, feces, sweat, hair and nail clippings, shed epithelial cells, placenta)

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

COLLABORATIVE AND TEAM SCIENCE

living, are covered by a new tracking system. The guidelines cover legal and ethical considerations for the collection, storage, use, sharing, and disposal of all human materials. A major ethical and legal issue related to biospecimen banking and custodianship involves ownership and disposition of these materials. Several prominent lawsuits have been reviewed and adjudicated recently, which deal with privacy issues, tissues and DNA as property and as potential sources of income.36,37 Although often contradictory, Federal rulings generally have stated that once tissues are removed, they are no longer the property of the donor but belong to the research institute where the study originated.38 Human biospecimens obtained by NIH researchers are considered federal property and must remain in the custody of NIH, although materials are made available for use by specific written agreement. The use of laboratory animals often is essential in biomedical research, but in using animals, a number of important points must be kept in mind. Animals must always be cared for and used in a humane and effective way, with procedures conducted as specified in an approved protocol. The use of animals in research must be reviewed by an Animal Care and Use Committee (ACUC), in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) guidelines. ACUC committees perform the following functions: review and approve protocols for animal research; review the institute’s program for humane care and use of animals; inspect all of the institution’s animal facilities every 6 months; and review any concerns raised by individuals regarding the care and use of animals in the institute. The NIH phased out all laboratory experiments with chimpanzees in 2015.39 In the 2016 spending bill, Congress strongly urged specific review of nonhuman primates in NIH-funded biomedical research. A workshop is planned for the summer of 2016 to develop policies and procedures for research with nonhuman primates. An investigator’s responsibilities in using animals for research include humane treatment of animals; following all procedures that were specified in the approved protocol; following the general requirements for animal care and use at the institution; and reporting concerns related to the care and use of laboratory animals. The policies and regulations for the utilization and care of laboratory animals are primarily concerned with minimizing or alleviating the animal’s pain and utilizing appropriate alternatives to animal testing when possible. In recent years great emphasis has been placed on the three R’sdreduction, refinement, and replacement (Table 3.4). However, experiments with

TABLE 3.4

39

The Three R’s in Animal Research

Reduction: Reduction in the numbers of animals used to obtain information of a certain amount and precision. Refinement: Decrease in the incidence or severity of pain and distress in those animals that are used. Replacement: Use of other materials, such as cell lines or eggs, or substitution of a lower species, which might be less sensitive to pain and distress, for a higher species.

animals should always consider sex as a biological variable, with appropriate experimental design to evaluate this important variable.40

COLLABORATIVE AND TEAM SCIENCE Research collaborations facilitate progress and should be encouraged. As research methods become more specialized and resources diminish, team science is not only attractive but in some cases necessary. The ground rules for collaborations, including authorship issues, should be discussed openly among all participants from the beginning. Research data should be made available to all scientific collaborators on a project upon request. Although each research project has unique features, certain core issues are common to most of them. Successful collaborations are characterized by a strong sense of direction, a willingness to commit time and effort, an efficient communication strategy for discussion among the group members, a system for reevaluation as the project progresses, and a clear definition of roles and responsibilities. It is advisable that the ground rules for collaborations, including eventual authorship issues, be discussed openly among all participants from the beginning. The NIH Ombudsman Office has developed a useful set of criteria to consider establishing collaborations and a Field Guide for Team Science shown in Table 3.5.41 Whenever collaborations involve the exchange of biological materials, they are routinely formalized by written agreements. Material Transfer Agreements used for simple transfer of proprietary research material without collaboration, for example, if you request a reagent from, or give one to, a colleague outside of NIH. Collaborative Research and Development Agreement (CRADA): Agreements between one or more NIH laboratories and at least one nonfederal group (private sector, university, not-for-profit, nonfederal government). CRADAs provide a protected environment for long-term collaboration; they confer intellectual

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

40

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

TABLE 3.5

Questions for Scientific Collaborators

Although each research project has unique features, certain core issues are common to most of them and can be addressed by collaborators posing the following questions: OVERALL GOALS 1. What are the scientific issues, goals, and anticipated outcomes or products of the collaboration? 2. When is the project over? WHO WILL DO WHAT? 1. What are the expected contributions of each participant? 2. Who will write any progress reports and final reports? 3. How, and by whom, will personnel decisions be made? How, and by whom, will personnel be supervised? 4. How, and by whom, will data be managed? How will access to data be managed? How will long-term storage and access to data be handled after the project is complete? AUTHORSHIP AND CREDIT 1. What will be the criteria and the process for assigning authorship and credit? 2. How will credit be attributed to each collaborator’s institution for public presentations, abstracts, and written articles? 3. How, and by whom, will public presentations be made? 4. How, and by whom, will media inquiries be handled? 5. When and how will intellectual property and patent applications be handled? CONTINGENCIES AND COMMUNICATING 1. What will be the mechanism for routine communications among members of the research team (to ensure that all appropriate members of the team are kept fully informed of relevant issues)? 2. How will decisions about redirecting the research agenda as discoveries are made be reached? 3. How will the development of new collaborations and spin-off projects, if any, be negotiated? 4. Should one of the principals of the research team move to another institution or leave the project, how will data, specimens, lab books, and authorship and credit be handled?

property rights to NIH inventions and are handled by the technology Transfer Office in each Institute. The NIH Office of Technology Transfer developed a set of FAQs to help investigators determine which instrument is most appropriate.42

CONFLICT OF INTEREST AND COMMITMENT Conflict of interest is a legal term that encompasses a wide spectrum of behaviors or actions involving personal gain or financial interest. According to Frank Macrina, “a conflict of interest arises when a person exploits, or appears to exploit, his or her position for personal gain or for the profit of a member of his or her immediate family or household.”43 The existence

of a conflict of interest may adversely affect the ability to objectively carry out scientific studies and report their results. Potential conflicts of interest may not be recognized by others unless disclosed; disclosure should include all relevant financial relationships. Disclosure is made to the appropriate organization depending on the activity: to one’s research institution while carrying out the research; to the funding agency when involved in peer review of grants; to meeting organizers when giving an invited presentation; and to journal editors when asked to referee articles, or when submitting one’s own manuscripts for consideration. Three-tenths percent of respondents to the survey26 on inappropriate research behaviors reported “not properly disclosing involvement in firms whose products were based on their own research,” suggesting that this is an issue that needs to be further addressed. The personal integrity of the physician is a paramount concern of society that dates back to the beginning of written history. Because clinical research involves a somewhat different relationship between investigators (many of whom are not physicians) and patients, it has been necessary to develop a new set of community standards to assure the integrity of the clinical research process. One set of ethical standards relates to the need to protect human subjects involved in clinical research. Another concern relates to the way in which real or perceived conflicts of interest may affect the integrity of clinical research. A conflict of interest occurs when other interests that the physician may have undermined, or appears to undermine, his or her objectivity and conduct in meeting those goals. A chronic concern is the interaction between industry and clinical researchers in the handling of clinical trials, and the potential for conflict of interest. Given the enormous costs of clinical trials, combined with the desire of clinical investigators to try the latest drugs, which are often only available from drug companies, increasingly companies serve as the sponsors of clinical trials (70% of the funding for such trials) and as such, may seek control over the research protocol and publication of the results. Even the appearance of such a conflict, without intent on the part of the investigator, is corrosive to the integrity of clinical investigation. Although most clinical investigators will deny vehemently that their financial interests would affect their research and clinical activities, studies have shown that interactions with pharmaceutical firms can have an effect on decision-making by physicians.44,45 Those who were receiving remuneration of some kind from pharmaceutical firms were more likely to support the safety of the drugs of those companies, and we can presume that research activities would be similarly affected. In 2005 federal restrictions on relations with pharmaceutical companies and the

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

41

PEER REVIEW

biotechnology industry were enacted. Although not a total ban, collaborations and interactions remain strong. Agreements with industry, including CRADAs and Clinical Trials Agreements are reviewed thoroughly by the NIH Technology Transfer and Ethics offices, demonstrating that industry interactions can still flourish and conflicts avoided within ethical guidelines that strengthen public trust. Preventing conflicts of interest (COI) in clinical studies is of particular concern to NIH, where conflicts have the potential to skew enrollment, data acquisition and analysis, and outcomes. NIH historically has provided advice and guidance on COI for its employees for many years and updated its “Guide to Avoiding Financial and Non-Financial Conflicts or Perceived Conflicts of Interest in Clinical Research at NIH” (Guide) in 2012.46 The guide is intended to provide assistance to those engaged in clinical research, IRB, and Data and Safety Monitoring Board (DSMB) members in avoiding real or perceived financial and nonfinancial conflicts of interest. Another concerning conflict involves the propensity of academic scientists to establish start-up companies sponsored or supported by their institutions as a result of the Bayh-Dole Act (1980), which allows investigators to profit from their intellectual property. A classic example of such a conflict involved the patenting of the BRCA1 gene associated with hereditary susceptibility to breast cancer by Myriad Inc, in 1991. Myriad Inc. held seven patents that prohibited the use of the genetic sequence for testing of patients for the likely presence of this high-risk gene mutation. In a class action lawsuit filed by the American Civil Liberties Union on behalf of patients, the courts in 2010 overruled the ability of Myriad to commercialize genetic testing, by demonstrating that profit motivated the patent without concern for medical advancement in the detection and treatment of breast and ovarian cancer.47 While conflict of interest issues has been a main concern in public policy, conflict of commitment can be equally important. This refers to the idea that someone has agreed to do more things than possible, especially activities that have no direct bearing on their employment responsibilities. These could be compensated or uncompensated activities, such as work with professional or nonprofit organizations in off-duty hours, which take away from primary responsibilities. Examples could include excessive commitments of time for work on behalf of scientific societies, nongovernmental organizations, or participation in outside private clinical practice. Similarly, overcommitment can lead to ethical problems. For example, when someone takes on too many trainees, or oversees too many clinical trials as Principal investigator, they become incapable of giving their best effort to all of them. If an investigator cannot find the

time to meet with a fellow, to review data and results, critique the first draft of a manuscript within a few days or a week, or to personally supervise the running of a clinical trial, that is a strong sign of overcommitment. Failure to personally oversee (1) basic research projects, a common occurrence in many research misconduct cases, or (2) clinical research requirements such as adequate monitoring of FDA-regulated products is one of the most common findings cited in audit reports and FDA warning letters.

PEER REVIEW Peer review is defined as a critical evaluation, conducted by one or more experts in the relevant field, of either a scientific documentdsuch as a research article submitted for publication, a grant proposal, or a study protocoldor a research program. One requisite element for peer review is the need for reviewers to be experts in the relevant subject areas. At the same time, real or perceived conflict of interest arising as a result of a direct competitive, collaborative, or other close relationship with one of the authors of the material under review should be avoided. All evaluations should be thorough and objective, fair and timely, and based solely on the material under review: information not yet publicly available cannot be taken into consideration. The use of multiple reviewers mitigates to some extent one inappropriate review, but nevertheless reviewers should strive to provide constructive advice and avoid pejorative comments. Since reviews are usually conducted anonymously, it is incumbent on the reviewer to protect the privileged information to which he or she becomes privy. No reviewer should share any material with others unless permission has been requested and obtained from those managing the review process. One of the marks of a good mentor is someone who teaches trainees how to handle peer review by asking them to review a submitted manuscript, but it is incumbent on the mentor to notify the journal that he/she plans to do so and get explicit permission before doing so. Sadly, peer review fraud has begun to appear with alarming regularity. Between 2012 and 2015, several prominent publishers have reported being scammed by fraudulent reviewers, many times by the authors of the papers themselves. In 2014, a Nature article reported that over 110 papers were retracted owing to being implicated in peer review fraud.45 In 2015 Hindawi, which publishes 437 academic journals, retracted 32 articles; BioMedCentral, publisher of 277 journals, and Springer, which owns BioMedCentral, together retracted 74 of their articles when confronted with evidence of reviewer fraud by both authors and editors.48

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

42

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

These retractions cover a variety of disciplines and authors from several countries. The ease with which this took place was astounding and relates mostly to the common practice that permits authors to recommend reviewers. Often the author established a separate email account under an assumed name and put this individual forward as a reviewer. Journals seldom checked the legitimacy of the proposed reviewers, who were unknown and unaffiliated with an academic institution. Because of the avalanche of papers that inundate publishers and editors, more and more reliance is placed on the use of publishing software where reviews are distributed and processed. In the cases cited, evidence implicates the ease with which vulnerabilities such as passwords in publishing software, such as ScholarOne or Editorial Manager were hacked. A computer security expert from Harvard commented “As you make the system more technical and more automated, there are more ways to game it.”49 And as long as the currency of scientific achievement remains the number of papers published, new and more inventive ways to game the system seem to reappear with regularity.

PUBLICATION PRACTICES, RESPONSIBLE AUTHORSHIP, AND RESULTS REPRODUCIBILITY Publication Practices Publication of results fulfills a scientist’s responsibility to communicate research findings to the scientific community, a responsibility that derives from the fact that much research is funded by the federal government using taxpayers’ money. Publication of clinical studies also fulfills the responsibility to provide scientific benefit in return for putting human subjects at risk. Other than presentations at scientific meetings, publication in a scientific journal should normally be the mechanism for the first public disclosure of new findings. An exception may be appropriate when serious public health or safety issues are involved. Timely publication of new and significant results is important for the progress of science but each publication should make a substantial contribution to its field. Fragmentary publication of the results of a scientific investigation or multiple publications of the same or similar data are not appropriate. Publications share findings that benefit society and promote human health, but publications also establish scientific principles and precedence. Credit for a discovery belongs to the first to publish, and reputations and research funding are based on the number and impact of publications, as are improved opportunities for prestigious positions.

The recent proliferation of so-called Predatory journals should be a cautionary note when considering a journal for submission of results.50 Starting around 2008, concerns about predatory publishing has been growing, and the number of dubious article published in these journals in 2014 reached 420,000. These journals attract authors with false or misleading information and solicit articles from well-known authors to add prestige to their journals. Characteristics of a predatory journal may include publishing with little or no peer review (rapid turn-around time), aggressive solicitation of submissions, splashy websites, imitating journal names with subtle changes or miss-represented titles, misleading claims, or fake impact factors. If in doubt, best refer to an indexed list of journals, such as found in PubMed.

Authorship Authorship is the primary mechanism for determining the allocation of credit for scientific advances and is thus the primary basis for assessing a scientist’s contributions to developing new knowledge. As such, it not only conveys great benefit but also significant responsibility. Authorship involves the listing of names of participants in all communications (oral or written) concerning experimental results and their interpretation, as well as making decisions about who will be the first author, the senior author, and the corresponding author. Yet authorship on publications often generates some of the most difficult disputes among scientists, because of its importance for careers. The NIH established benchmarks for authorship credits, developed by the CSCE. Furthermore, they determined a procedure to adjudicate authorship disputes. Recommendations range from mediation through the NIH Ombudsman (note: authorship disputes constitute the single largest group of scientific complaints with which the NIH Ombuds office deals), to establishing a peer review panel empowered to make binding recommendations, to a final decision by the Scientific Director of the institute or the Deputy Director for Intramural Research.51 Authorship is justified by a significant contribution to the conceptualization, design, execution, and/or interpretation of the research study and a willingness to assume responsibility for the study. Other ways to establish credit for contributions besides authorship include acknowledgments and references. Acknowledgments provide recognition of individuals who have assisted the research by their encouragement and advice about the study, editorial assistance, technical support, or provision of space, financial support, reagents, or specimens. References acknowledge others’ discoveries, words, ideas, data, or analyses and must be cited in

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

PUBLICATION PRACTICES, RESPONSIBLE AUTHORSHIP, AND RESULTS REPRODUCIBILITY

a way that others can find the reference and see the contribution. According to results from the 2005 survey on questionable practices, 1.4% reported that they had used others’ ideas without obtaining permission or giving credit.26 When should authorship issues be discussed? Although there is no universal set of standards for authorship, each research group should freely discuss and resolve questions of authorship before and during the course of a study. Each author should fully review material that is to be presented in a public forum or submitted (originally or in revision) for publication. Each author should indicate a willingness to support the general conclusions of the study before its presentation or submission. Settling authorship issues should be finalized as early as possible to avoid any conflicts over credit for scientific work. With the recent increase in numbers of authors on publications, the problem has increased in magnitude. A number of studies in the late 1990s and early in 2000s, defined several categories of irresponsible authorship.52,53 These include honorary authorshipdan author who does not meet the criteria; ghost authorshipd failure to include as an author, someone who made substantial contributions to the article; refusal to accept responsibility for an article despite ready acceptance of credit; and duplicate and redundant publications. Rennie and colleagues52 carried out a study based on the following hypotheses: research articles in largecirculation prestigious medical journals would be more likely to have honorary authors while review articles in smaller-circulation journals that publish symposia proceedings would be more likely to have ghost authors. Despite disproving the hypotheses, the study showed a significant misuse of authorship in biomedical journals that ultimately led to a number of changes regarding authorship criteria. The International Committee of Medical Journal Editors (ICMJE) issued a set of Uniform Requirements for Manuscripts Submitted to Biomedical Journals54 to address standards for authorship and more recently established “ethical principles related to publication in biomedical journals.” They defined an author as someone “who has made substantive intellectual contributions to a published study” and provide a set of criteria for authorship as shown in Table 3.6. In a follow-up survey, Wisler et al. noted that although allover inappropriate authorship (e.g., ghost authorship) declined, there was little to no change in levels of honorary authorship.53,55 The Journal of the American Medical Association authorship policy more specifically states that all authors must describe their specific contributions as well as the contributions of those included in the acknowledgments.56 Authors must determine the distribution of authorship between who authors and who should simply be acknowledged. Authors should

TABLE 3.6

43

International Committee of Medical Journal EditorsdCriteria for Authorship

Authorship should be based on: • Substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data • Drafting the article or revising it critically for important intellectual content • Final approval of the version to be published Authors should meet all three conditions. Furthermore, all persons designated as authors should qualify for authorship, and all those who qualify should be listed.

be listed in order of actual degree of contribution, based on discussions among the authors. The NIH CSCE developed a visual graphic of general guidelines, based on a sliding scale developed by Evelyn Ralston at the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), to help authors work through these issues Table 3.7. The Annals of Internal Medicine, in concordance with the general principles illustrated in Tables 3.6 and 3.7, further notes that the following, by themselves, are not criteria for authorship: holding a position of administrative leadership; contributing patients or reagents; or collecting and assembling data. Adhering to these criteria will result in a significant change to the way authorship is determined for clinical studies and may require a culture change.57

Reproducibility Although each paper should contain sufficient information for the informed reader to assess its validity, the principal method of scientific verification is not simply a review of submitted or published papers, but the ability of others to replicate the results. Concerns are mounting about the current system for ensuring reproducibility in biomedical research.58 Poor training in experimental design and publications missing basic elements of design and increased pressure to publish in high-impact journals with space limitations all affect methodological completeness. A rush to publish for promotion and tenure decisions also contribute to lack of reproducibility. Likewise the dearth of publications on negative data or failure to correct flawed methodology in previously published work is at fault. NIH proposed several recommendations for enhancing reproducibility.58 Each paper should contain all the information necessary for other scientists to repeat the work. When considering reproducibility in research, two specific issues should be addressed: potential biases and rigorous experimental design (Table 3.8). Experimental bias in reporting results can be unconscious

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

44

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

TABLE 3.7

General Guidelines for Authorship Contributions

TABLE 3.8

Experimental Design and Reproducibility

Issues that affect experimental design and reproducibility of research results Bias: Prejudice in favor or against one idea, thing, person, or group compared with another usually in a way considered to be unfair. Bias in research may be unconscious or unintentional. Scientific Rigor: The strict application of the scientific method to ensure robust and unbiased experimental design, methodology, analysis, interpretation, and reporting of results. Links to NIH Training Videos focusing on issues critical for reproducibility https://oir.nih.gov/sourcebook/ethical-conduct/responsibleconduct-research-training/instruction-responsible-conduct-researchpostdoc-irta-crta-vf-research-0 Discussion documents are provided for each vignette. 1. 2. 3. 4.

Lack of transparency Blinding and randomization Biologic and technical replicates Sample size and exclusion criteria

and unintentional. Causes may include unknown or unavoidable differences between comparison groups, less than ideal experimental designs, systematic errors introduced between test groups, and poor methodology, analysis, interpretation, or reporting of results. Blinding can minimize bias through randomization or stratification, reporting all data (both positive and negative), and establishing criteria to identifying outliers. Rigor in scientific research includes consideration of experimental design, rationale for selecting endpoints or model systems, use of both positive and negative controls, consistent experimental conditions, sample size, power calculation, statistical methods used for analysis and interpretation of results, and maintaining rigorous, detailed laboratory records.59 Good experimental design is a critical factor for reliable reproducibility of research results. A clearly stated question is important, as are the choice of the appropriate tools that allow clear answers to the question.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

Working with a statistician in advance of starting experiments will help to determine the model of analysis and the number of data points needed to make a valid conclusion.60 Judicious publication of new and significant results is important for the progress of science, but each publication should make substantial contribution to its fields. Fragmentary publications of results of a scientific investigation or multiple publications of the same or similar data are not appropriate.

STUDY QUESTIONS 1. Which of the following is not a criterion for research misconduct? a. Making up data b. Appropriating ideas from someone else’s research application c. Failure to retain research records d. Changing results to match the hypothesis 2. A finding of research misconduct requires that a. The allegations must be proven by a preponderance of the evidence b. The misconduct is committed knowingly, recklessly, or intentionally c. There has been a significant departure from accepted practices of the relevant research community d. All of the above 3. Authorship for a publication should be based upon a. Substantial contributions to conception and design b. Financial interest of the sponsor of the study c. Personal relationship with the PI of the study d. Reputation of a colleague 4. Which factor(s) do/does not contribute(s) to lack of reproducibility? a. Increased emphasis on provocative statements b. Publication of negative data c. Pressure to publish in high-impact journals d. Promotion and tenure incentives for publications

Acknowledgments We are grateful to the many colleagues who have contributed to the ideas presented in this chapter. In particular, we thank the contributions of the members of the NIH Committee on Scientific Conduct and Ethics (CSCE) to the latest revision of the Guidelines and Policies for the Conduct of Research in the Intramural Research Program at NIH. Since the inception of the CSCE, its many members have contributed to the development and continued refinement of the NIH Responsible Conduct of Research Education program. And finally to Robert Nussenblatt, our esteemed and respected collaborator on issues related to research integrity. Bob passed away during the preparation of the revision of this chapter and we dedicate it to his memory.

45

References 1. Guidelines for the conduct of research in the intramural program at NIH. https://oir.nih.gov/sites/default/files/uploads/sourcebook/ documents/ethical_conduct/guidelines-conduct_research.pdf. 2. NIH responsible conduct of research. http://researchethics.od.nih. gov. 3. NIH ethics cases. https://oir.nih.gov/sourcebook/ethical-conduct/ responsible-conduct-research-training/annual-review-ethics-casestudies. 4. NIH sourcebook. https://oir.nih.gov/sourcebook. 5. National Academy of Sciences. On being a scientist. 3rd ed. Washington, DC: National Academy Press; 2009. http://www.nap. edu/read/12192/chapter/1. 6. Resnik DB, Rasmussen LM, Kissling GE. An international study of research misconduct policies. Account Res 2015;22:249e66. http:// dx.doi.org/10.1080/08989621.2014.958218. 7. Anderson E, Solomon S, Heitman E, DuBois JM, Fisher CB, Kost RG, Lawless ME, Ramsey C, Jones B, Ammerman A, Friedman- Ross L. Research ethics education for community-engaged research: a review and research agenda. J Empir Res Hum Res Ethics 2012;7:3e19. http://dx.doi.org/10.1525/jer.2012.7.2.3. 8. Epigeum, Imperial College in London. https://www.epigeum. com/epigeum/. 9. Center for research ethics and bioethics, Uppsala University, Sweden. http://www.crb.uu.se/education/ethics-training/. 10. European Network of Research Ethics Committees. http://www. eurecnet.org/materials/index.html-m. 11. Responsible Conduct of Research Education Consortium. http:// www.indiana.edu/wappe/rcrec.html. 12. Poynter Center for the Study of Ethics and American Institutions. http://poynter.indiana.edu/index.shtml. 13. Office of Research Integrity. https://ori.hhs.gov/rcr-casebookstories-about-researchers-worth-discussing. 14. Institute of Medicine. The responsible conduct of research in the health sciences. Washington DC: National Academy Press; 1989. 15. Institute of Medicine. Integrity in scientific research, creating an environment that promotes responsible conduct. Washington DC: National Academy Press,; 2002. 16. NIH Responsible Conduct of Research Training. https://oir.nih. gov/sourcebook/ethical-conduct/responsible-conduct-researchtraining. 17. Fang EC, Steen RG, Casadevail A. Misconduct accounts for the majority of retracted scientific publications. PNAS 2012: 17029e33. http://dx.doi.org/10.1073/pnas.1212247109. 18. Retraction watch. http://retractionwatch.com. 19. Shuai X, Moulinier I, Rollins J, Custis T, Schilder F, Edmunds M. A multi-dimensional investigation of the effects of publication retraction on scholarly impact. 2016. http://arxiv.org/abs/1602.0912320. 20. PubPeer. https://pubpeer.com. 21. National Institutes of Health Intramural Research Program Policies & Procedures for Research Misconduct Proceedings. https://oir. nih.gov/sites/default/files/uploads/sourcebook/documents/ ethical_conduct/policy-nih_irp_research_misconduct_proceedings.pdf. 22. Bennett LM, Marais R, Gadlin H. The ‘Welcome letter’: a useful tool for laboratories and teams. J Transl Med Epidemiol 2012;2:1035. 23. McMahon T. Aligning expectations. In: Pfund C, House S, Asquith P, Spencer K, Silet K, Sorkness C, editors. Mentor training for clinical and translational researchers. Basingstoke, England: W. H. Freeman; 2012. p. 43e7. https://mentoringresources.ictr.wisc. edu/sites/default/files/McMahon_UW_Compact_Example.pdf. 24. Ramsey N. Working with normal ramsey, a guide for research students. Tufts University; 2014. https://www.cs.tufts.edu/wnr/students/ guide.pdf.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

46

3. INTEGRITY IN RESEARCH: INDIVIDUAL AND INSTITUTIONAL RESPONSIBILITY

25. AAMC. https://www.aamc.org/initiatives/research/gradcompact. 26. Martinson BC, Anderson MS, de Vries R. Scientists behaving badly. Nature 2005;435:737e8. 27. John LK, Lowenstein G, Prelec D. Measuring the prevalence of questionable research practices with incentives for truth telling. Psychol Sci 2012;23:524e32. http://dx.doi.org/10.1177/0956797611430953. 28. Genomic data sharing. http://grants.nih.gov/grants/guide/ notice-files/NOT-OD-14-124.html. 29. NIH databases. https://gds.nih.gov/02dr2.html. 30. Clinical Trials website. https://clinicaltrials.gov. 31. NIH guide notice NOT-OD-15-019. https://grants.nih.gov/ grants/guide/notice-files/NOT-OD-15-019.html. 32. Office of Human Subjects Research Protections. http://ohsr.od. nih.gov/OHSR/pnppublic.php. 33. Responsible Conduct of Research for Clinical Scientists. https:// www.citiprogram.org/rcrpage.asp? language¼english&affiliation¼100. 34. Research Networking for Clinical and Translational Scientists Awards. https://ctsacentral.org/consortium/best-practices/researchnetworking/. 35. Guidelines for Human Biospecimen Storage and Tracking within the NIH Intramural Research Program. http://sourcebook.od. nih.gov/oversight/BiospecimenGuidelines.pdf. 36. Charo RSA. Body of research-ownership and use of human tissue. New Engl J Med 2006;355:1517e9. 37. Roche PA, Annas GJ. New genetic privacy concerns. GeneWatch 2007; 20(1). http://www.councilforresponsiblegenetics.org/GeneWatch/ GeneWatchPage.aspx?pageId¼196&archive¼yes. 38. Mascalzoni D, Dove ES, Rubinstein Y, Dawkins HJS, Kole A, McCormak P, Woods S, Riess O, Schaefer F, Lochmuller H, Knoppers BM, Hansson M. International charter of principles for sharing biospecimens and data. Euro J Hum Genet 2015;23: 712e28. http://dx.doi.org/10.1038/ejhg.2014.197. 39. Kaiser J. NIH to end all support for chimpanzee research. Science 2015. http://dx.doi.org/10.1126/science.aad7458. 40. Clayton JA, Collins FS. Policy: NIH to balance sex in cell and animal studies. Nature 2014;509:282e3. http://dx.doi.org/ 10.1038/509282a. 41. Bennett LM, Gadlin H, Levine-Finley S. Collaboration & team science: a field guide. 2010. http://TeamScience.nih.gov. 42. Technology transfer FAQs. http://www.ott.nih.gov/crada-mtafaqs. 43. Macrina FL. Scientific integrity: an introductory text with cases. Washington DC: ASM Press; 2000. 44. Stelfox HT, Chua G, O’Rourke K, Detsky AS. Conflict of interest in the debate over calcium-channel antagonists. N Engl J Med 1998; 338:101e6. 45. Blumenthal D, Campbell EG, Anderson MS, Causino N, Louis KS. Withholding research results in academic life science. Evidence from a national survey of faculty. JAMA 1997;277:1224e8.

46. A guide to avoiding financial and non-financial conflicts or perceived conflicts of interest in clinical research at NIH. https://oir.nih.gov/sites/ default/files/uploads/sourcebook/documents/ethical_conduct/ guide-avoiding_conflict_interest_clinical_research.pdf. 47. Siegal D. Myriad settles with quest in cancer gene test patent MDL, law 360. 2015. http://www.law360.com/articles/619520/myriadsettles-with-quest-in-cancer-gene-test-patent-mdl. 48. Ferguson C, Marcus A, Oransky I. Publishing: the peer-review scam. Nature 2014;515:480e2. 49. Bohannon J. Who’s afraid of peer review? Science 2013;342:60e5. http://dx.doi.org/10.1126/science.342.6154.60. 50. Beware of predatory publishers, NIH catalyst. JanuaryeFebruary 2015. 51. Procedure for authorship resolution. https://oir.nih.gov/sourcebook/ ethical-conduct/responsible-conduct-research-training/processesauthorship-dispute-resolution. 52. Flanagin A, Carey LA, Fontanarosa PB, Phillips SG, Pace BP, Lundberg GD, Rennie D. Prevalence of articles with honorary authors and ghost authors in peer-reviewed medical journals. JAMA 1998;280:222e4. 53. Baskin PK, Gross RA. Honorary and ghost authorship. BMJ 2011; 343:d6223. http://dx.doi.org/10.1136/bmj.d622331. 54. International Committee of Medical Journal Editors Uniform Requirements. http://www.ICMJE.org. 55. Wislar JS, Flanagin A, Fontanarosa PB, DeAngelis CD. Honorary and ghost authorship in high impact biomedical journals: a cross sectional survey. BMJ 2011;343:d6128. http://dx.doi.org/10.1136/ bmj.d6128. 56. JAMA. http://jama.jamanetwork.com/public/InstructionsForAuthors. aspx#dvTopNav; http://www.icmje.org/recommendations/. 57. Prinz F, Schlange T, Asadullah K. Nature rev. Drug Disc 2011;10: 712e3. 58. Collins FS, Tabak LA. Policy: NIH plans to enhance reproducibility. Nature 2014;505:612e3. http://dx.doi.org/10.1038/505612a. http://www.nature.com/news/policy-nih-plans-to-enhancereproducibility-1.14586. 59. Landis, et al. A call for transparent reporting to optimize the predictive value of preclinical research. Nature 2014;490:187e91. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3511845/. 60. Kilkenny C, Parsons N, Kadyszewski E, Fesitng MFW, Cuthill ID, Frey D, Hutton J, Altman DG. Survey of the quality of experimental design, statistical analysis and reporting of research using animals. PLoS One 2009;4:e7824. http://dx.doi.org/10.1371/ journal.pone.0007824.

Further Reading 1. Zinner DE, DesRoches CM, Bristol SJ, Clarridge B, Campbell EG. Tightening conflict-of-interest policies: the impact of 2005 ethics rules at the NIH. Acad Med 2010;85:1685e91.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

4 Institutional Review Boards 1

Julia Slutsman1, Lynnette Nieman2

National Institutes of Health, Washington, DC, United States; 2National Institutes of Health, Bethesda, MD, United States

O U T L I N E Continuing Review of Research

Historical, Ethical, and Regulatory Foundations of Current Requirements for Research Involving Human Subjects 47 Historical Foundations 47 Ethical Foundations 49 Regulatory Foundations 49 Institutional Review Boards Key Concepts and Definitions From the Common Rule Research Exempt Research Activities Minimal Risk and Expedited Review Procedures Institutional Review Board’s Review of Research Institutional Review Board Membership Criteria for Institutional Review Board Approval of Research

Clinical Researchers and Institutional Review Boards 57 Evaluation and Evolution of the Current System of Research Oversight and Institutional Review Boards 57 Proposed Changes to Current Oversight of Research With Human Subjects 57 Critique and Proposed Changes to Institutional Review Board Operations 58

50 50 51 51 51 51 51 52

Conclusion

59

Summary Questions

59

References

59

HISTORICAL, ETHICAL, AND REGULATORY FOUNDATIONS OF CURRENT REQUIREMENTS FOR RESEARCH INVOLVING HUMAN SUBJECTS

In the United States, the rights and welfare of human research subjects take precedence over the advance of scientific knowledge. Ethical guidelines, federal regulations, local institutional policies and procedures, and the knowledge and integrity of researchers and research staff all contribute to promoting the protection of human subjects. Our society has decided by law that objective, ongoing review of research activities by a group of diverse individuals is most likely to protect human subjects and promote ethically sound research. Prospective review of research by institutional review boards (IRBs) provides an important assurance that the rights and welfare of human subjects are given serious consideration. This chapter focuses on the development of US federal regulations concerning research involving human subjects and the roles and responsibili ties of IRBs. Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00004-6

56

Historical Foundations Concerns about the ethics of the practice of medicine have a long history, but until the mid-20th century, they were mostly centered on the practice of therapeutic medicine, not research medicine and are summarized in Chapter 2, but restated here in slightly different context for different and enhanced emphasis. In 1946, 23 Nazi physicians went on trial at Nuremberg for crimes committed against prisoners of war and inmates

47

Copyright © 2018. Published by Elsevier Inc.

48

4. INSTITUTIONAL REVIEW BOARDS

of concentration camps. These crimes included exposure of humans to extremes of temperature, performance of mutilating surgery, and deliberate infection with lethal pathogens. During the trial, fundamental ethical standards for the conduct of research involving humans were codified into the Nuremberg Code, which sets forth 10 conditions that must be met to justify research involving human subjects.1 Two important conditions are (1) the need for voluntary informed consent of subjects and (2) a scientifically valid research design that can produce fruitful results for the good of society. The Nuremberg Code was accepted in principle by 48 of 58 original signatory nations of the Charter of the United Nations as part of the Declaration of Human Rights. (Others abstained or did not vote.) However, in the United States, the existence of the Nuremberg Code was not widely appreciated. Researchers and physicians who were familiar with it generally believed that its requirements narrowly applied to research conducted by German researchers, and that it had little applicability or relevance to research conducted in the United States.2 In fact, full implementation of the first condition of the code in the United Statesdthe voluntary consent of subjects who are able to exercise free power of choicedwould have severely curtailed, if not eliminated, research involving prisoners, minors, and other individuals determined to lack capacity for providing informed consent. In the United States during the 1950s through the mid-1970s, many chemotherapeutic agents for cancer and other diseases/disorders were tested initially in healthy prisoners; in fact, some pharmaceutical companies had research buildings located on or near prisons to facilitate their research activities. Therefore, implementation of the code would have had major, dramatic effects on the conduct of research in the United States and in fact many significant changes to these practices would occur later on time. Most countries accepting the principles of the code, including the United States, had no mechanism for implementing its provisions. In 1953, the National Institutes of Health (NIH) opened the Clinical Center (CC), its major research hospital in Bethesda, Maryland, which subsequently developed the first US public policy for the protection of human subjects. The policy, which was applicable to intramural research at the NIH CC, required peer review of research protocols enrolling healthy subjects and was consistent with the Nuremberg Code in that it gave special emphasis to the protection of healthy, adult research volunteers who had little to gain directly from participation in research.3 The CC policy was innovative not only for its adoption but also for providing a mechanism for prospective review of research by individuals who had no direct involvement or intellectual investment in the research.

This was the beginning of the research review mechanismdthe IRBdthat is now fundamental to the current system of human subject protection throughout the United States. In fact, the first two research protocols submitted to the institutional review committee of the CC were disapproved because the committee judged that research-related risks to healthy volunteers were too high.4 However, the initial CC requirements for prospective review of research and obtaining subjects’ informed consent were applicable only to research involving healthy volunteers, not patients. In excluding research involving patients from these requirements, the policy was consistent with contemporaneous thinking of US physicians/researchers; most were reluctant to set forth explicit rules for the conduct of research involving patients, arguing that such rules would impede research and undermine trust in the physician.2 In the 1960s, increased pharmaceutical industry sponsored funding for clinical research expanded with the passage of the KefauvereHarris bill, which required manufacturers to perform research establishing safety and efficacy of drugs prior to marketing. The bill also required that subjects participating in research subject to the US Food and Drug Administration (FDA) purview must provide their informed consent prior to participation, becoming the first federal statute requiring protections for human subjects.3 Interest in the rights of research subjects grew not only because of a general increase in attention to human rights in the United States but also because of a number of highly publicized clinical research abuses. Indeed, the current laws that govern the conduct of federally funded research with human subjects were formulated in part in reaction to research abuses and scandals. In 1966, Henry Beecher, a highly respected physician/investigator from Harvard University, shocked the medical community when he reported that unethical and questionably ethical practices were common in the conduct of human subjects research at many of the premier research institutions in the United States.5 One of 22 examples of research misconduct that Beecher described involved investigators at the Jewish Chronic Disease Hospital in New York who injected elderly, indigent people with live cancer cells, without their consent, to learn more about the human immune system. Although no apparent harm to subjects were documented, the investigators were cited for fraud, deceit, and unprofessional conduct.6 In 1964 the World Health Organization recognized the need for research ethics guidelines that were broader in scope than the Nuremberg Code by adopting the “Declaration of Helsinki: Recommendations Guiding Medical Doctors in Biomedical Research Involving Human Subjects.”7 In the ensuing years these guidelines have been revised numerous times and are currently in use throughout the world.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

HISTORICAL, ETHICAL, AND REGULATORY FOUNDATIONS OF CURRENT REQUIREMENTS FOR RESEARCH INVOLVING HUMAN SUBJECTS

In 1966, the NIH, under the directorship of Dr. James Shannon, issued the first Public Health Service Policy on the Protection of Human Subjects.3 This policy, which applied to research conducted or supported by the then Department of Health, Education, and Welfare (HEW), including the NIH, required prospective review of human subjects research, taking into account the rights and welfare of involved subjects, the appropriateness of methods used to secure informed consent, and the risks and potential benefits of the research. The policy included the requirement that consent should be documented by the signatures of subjects or their representative(s). Several events in the early 1970s led to renewed and intense efforts in the United States to enhance protections for research participants by putting in place regulations governing human subjects research. Most notable was the revelation that, since the 1930s, 400 syphilitic and 200 healthy black men in Tuskegee, Alabama, had been involveddwithout their knowledged in the Tuskegee Syphilis Study, a US Public Health Service sponsored, decades-long study on the natural history of syphilis.8 These men were systematically denied penicillin even after its introduction as standard treatment for the disease. Beginning in 1971, The Health Subcommittee of the Senate Committee on Labor and Human Resources held hearings on this study and on other alleged health-care abuses of prisoners and children.3 Outcomes of these and additional hearings included (1) enactment of the National Research Act of 1974, requiring HEW to codify its policy for the protection of human subjects into federal regulations, which it did in 1974; (2) formation of the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research; and (3) imposition of a moratorium on research conducted or supported by HEW involving live human fetuses until the National Commission could study and make recommendations on this activity.4 The National Commission, which functioned from 1974 to 1978, evaluated existing HEW regulations, recommended improvements to the Secretary of HEW, and issued reports on research involving pregnant women, live human fetuses, prisoners, children, the mentally disabled, and the use of psychosurgery. The Commission issued The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research in 1979.5 This major advancement in the development of public policy provided guidance for distinguishing therapeutic medicine from research, identified an analytic framework of three fundamental ethical principles for the protection of human subjects, and illustrated how ethical principles should be applied to the conduct of human subjects research. In 1979, the HEW began to revise the 1974 regulations, and in January 1981 the (renamed) Department

49

of Health and Human Services (DHHS) gave final approval to the Title 45 section 46 Code of Federal Regulations (CFR) governing protection of human subjects (45 CFR 46).6 Initially, these regulations were applicable only when research was conducted or supported by DHHS, but in June of 1991, the core of the regulations (Subpart A)dreferred to as the Common Ruledwas adopted by 16 other federal department agencies.7

Ethical Foundations The ethical framework for US laws governing human subjects research protection is articulated in The Belmont Report. This document establishes three fundamental ethical principles that are relevant to all research involving human subjects: (1) respect for persons, (2) beneficence, and (3) justicedand demonstrates how they apply to the conduct of research involving human subjects.6 Respect for persons acknowledges the dignity and autonomy of individuals and requires that subjects give informed consent to participation in research. However, not all individuals are capable of selfdetermination, and The Belmont Report acknowledges that people with diminished autonomy are entitled to additional protections. For example, some individuals may need extensive protection, even to the point of excluding them from research activities that may harm them, whereas others require little protection beyond making sure that they undertake research freely, with awareness of possible adverse consequences.5 Beneficence requires that the benefits of research are maximized, possible harms minimized and that risks are assessed to be reasonable in relation to potential benefits. This principle underlies careful analysis by researchers and IRBs of the risks and benefits of research protocols.6 Justice requires fair selection and treatment of research subjects and a fair distribution of the risks and benefits of research. For example, subjects should be equitably chosen to ensure that certain individuals or classes of individuals are not systematically selected for or excluded from research, unless there are scientifically or ethically valid reasons for doing so. Also, unless there is careful justification for an exception, research should not involve people from groups that are unlikely to benefit directly or from subsequent applications of the research.6 These three principles are not mutually exclusive. Each principle carries strong moral force and difficult ethical questions arise with regard to balancing the principles when they come into conflict.

Regulatory Foundations Biomedical and behavioral human subjects research funded or supported by the DHHS, including the NIH,

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

50

4. INSTITUTIONAL REVIEW BOARDS

is under the purview of regulations for the protection of human subjects included in 45 CFR 46.7 These regulations embody the principles of The Belmont Report. Taken together, The Belmont Report and 45 CFR 46 articulate the minimal ethical standards and legal obligations of those who conduct, review, and oversee research. Also, regardless of the funding source, all clinical trials in the United States involving investigational drugs or devices are under the regulatory purview of the FDA, which endorses 45 CFR 46. Additional FDA regulations contained in Title 21 sections 50 and 56 CFR govern the development and approval of drugs, biologics, and devices, regardless of the funding source.8 FDA and DHHS regulations on the protection of human subjects and IRBs are generally consistent, although some differences have been noted.9 The regulatory apparatus of the DHHS for overseeing the protection of human subjects involved in the research that it funds consists of two major tiers of reviewdone at the federal level and the other at the institutional level. For example, as a condition for receipt of NIH research funds, institutions must assure in writing that personnel will abide by ethical principles of The Belmont Report and the requirements of 45 CFR 46. These attestations are referred to as federalwide assurances (FWAs) of compliance. FWAs are negotiated and approved by the Office for Human Research Protections (OHRP) on behalf of the Secretary of the DHHS. In March of 2011, OHRP held 8557 active assurances with entities in the United States and 2457 assurances with international entities (personal communication with OHRP). All assurances set forth the institution’s policies and procedures for review and monitoring of human subjects research activities, including IRB membership requirements and review and record-keeping procedures. A variety of administrative actions can be taken by OHRP for violation of the requirements of 45 CFR 46 or the terms and conditions of an institution’s assurance of compliance. OHRP has published analyses of compliance oversight investigations. Compliance oversight investigations conducted from 1990 through early-2000 resulted in restrictions of clinical research activities or corrective measures in 38 US research institutions. Corrective actions included temporary suspension of all DHHSfunded clinical research in some institutions, the requirement that some or all investigators conducting research in these institutions receive appropriate additional education concerning the protection of human subjects, and quarterly reports to the DHHS of the institution’s progress in correcting identified deficiencies.10 Compliance oversight investigations conducted from 2005 to the end of 2010 resulted in restrictions or suspension of research activities in three US research institutions. Actions included temporary suspension of all federally funded research in some institutions,

assignment of a new Authorized Institutional Official on the Assurance, and the requirement to rereview all DHHS or US. federally supported human subject protocols that had not undergone adequate continuing review.11 From 2011 to May of 2016, OHRP has sent 71 compliance oversight determination letters. OHRP’s website provides letters of compliance oversight determination organized by date as well as by the type of noncompliance action addressed (for more recent years).12 Some clinical research conducted in the United States does not fall under federal regulations because it is not funded by the federal government, or because it does not involve compounds under the jurisdiction of the FDA. This includes some research conducted in educational environments (such as colleges and universities) and other nonfederally funded research. The quantity of such research and the settings in which it is being conducted are not known. Efforts were made to bring all US clinical research under the purview of federal regulations, however such provisions were not added to the Final Rule published on January 19, 2017. When filing an FWA, individual US institutions can voluntarily indicate to DHHS that they will apply to the rule to all human subjects research that they conductdregardless of the funding source. It is important to note that in addition to the regulation at 45 CFR 46, many states and institutions have additional laws and requirements that apply to human subjects research.

INSTITUTIONAL REVIEW BOARDS Key Concepts and Definitions From the Common Rule Because clinical investigators seek generalizable knowledge applicable to persons other than their individual patients, the pursuit of this goal may not always promote the welfare of individual patients. Accordingly, DHHS and FDA regulations require most proposed clinical research to undergo prospective independent review by an IRB as a mechanism for promoting ethically sound research. Although the IRB system is not perfect, conscientious IRBs reassure the US public that people not directly involved in the research consider seriously the rights and welfare of human subjects before research may begin. It is through this process of research review and approval that investigators, research institutions, IRB members, and others are held publicly accountable for their decisions and actions. The federal regulations at 45 CFR 46 include definitions of terms, recordkeeping requirements, composition of IRBs, and responsibilities and requirements for IRB review and approval of research involving human subjects.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

INSTITUTIONAL REVIEW BOARDS

Research Research is any systematic investigation designed to develop or contribute to generalizable knowledge (45 CFR 46.102[d]). A human subject is a living individual about whom an investigator obtains (1) data through interaction or intervention with the individual or (2) identifiable private information (45 CFR 46.102[f]). For example, consider the situation in which a physician asks the hospital medical records department to make available for review the medical records of all patients with a diagnosis of human immunodeficiency virus (HIV) infection. The physician wants to learn about the medical management of these patients treated in the hospital and its clinics during the past 5 years and to publish an analysis of that management. According to the preceding definitions, if the physician reviews medical records of patients who are no longer living, he or she is conducting research, but it does not involve human subjects (defined as living individuals). However, if the physician reviews medical records of patients who are still living, he or she is conducting research involving human subjects.

Exempt Research Activities Not all research involving human subjects requires prospective IRB review and approval. Although they involve human subjects, six categories of research are exempt from the requirements of 45 CFR 46 for IRB review in the version of the Common Rule in effect as of the writing of this Chapter. Changes and additional exemption categories are described in the changes to the Common Rule published on January 17, 2017.12a The general rationale behind the six categories of exemptions is that although the research involves human subjects, it does not expose them to physical, social, psychological, or other risks beyond those encountered in daily life. One example of exempt research is the study of existing records (e.g., pathologic samples, medical records) if these sources are publicly available, or if the investigator records the information in such a way that subjects cannot be identified directly or through identifiers linked to the subjects. Therefore, in the previous example, in which the researcher wants to study existing medical records, the research may be exempt from the requirement for IRB review and approval if the researcher records information from the medical charts in an anonymous fashion (no links or codes identifying patients). However, many hospitals have more restrictive policies concerning the research use of medical records and pathologic samples, and researchers should be familiar with relevant institutional policies. Survey and questionnaire research may be exempt unless the information elicited, if disclosed outside the research, could reasonably place subjects at risk for criminal or civil liability or could be damaging to subjects’

51

financial standing, employability, or reputation. Therefore, a questionnaire or survey should not be exempted if, for example, it elicits information about illegal behaviors, such as drug use, child or spousal abuse, or other sensitive issues such as sexual and other private behaviors. Institutional procedures vary for making determinations about whether proposed research activities are exempt. Investigators are not authorized to make final determinations about whether their proposed research activities are exempt from the requirement for prospective IRB review and approval. In some institutions, the IRB makes these determinations; in others, an office for research regulation or its equivalent makes these determinations.13 Thus, researchers should be familiar with their institution’s procedures for requesting exemptions.

Minimal Risk and Expedited Review Procedures Minimal risk means that “the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or the performance of routine physical or psychological examinations or tests” (45 CFR 46.102(i)). Some minimal risk research activities are eligible for IRB review through expedited review procedures. This means that the IRB chair and/or other experienced IRB members designated by the chair may approve (but not disapprove) the research on behalf of the IRB. The expedited review process was put into place to streamline and facilitate IRB review of certain minimal risk research activities. The changes to the Common Rule published on January 17, 2017, known as the Final Rule, also include an additional category of IRB review, called limited IRB review, which can apply to certain exempt and nonexempt research.

Institutional Review Board’s Review of Research When a researcher proposes to do research that is not exempt, he or she submits a research protocol for review. If a decision cannot be expedited, it must be reviewed by a convened IRB. A protocol is the researcher’s written description of the planned research, as well as a discussion of issues related to the protection of subjects. The following sections provide some of the regulatory requirements for IRB composition and criteria for IRB review and approval of research involving human subjects. Institutional Review Board Membership Federal regulations set minimal IRB membership standards for review of research under the Common Rule. All IRBs must have at least five members (45 CR 46.107): at least one whose primary concerns are in scientific areas, one whose primary concerns are in nonscientific areas, and one who is not otherwise affiliated with the institution. Also, when in its judgment

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

52

4. INSTITUTIONAL REVIEW BOARDS

the IRB requires expertise beyond or in addition to that available through its members, it may invite individuals with competence in special areas to assist in its reviews. These requirements for membership are grounded in the belief that the protection of human subjects is promoted by an objective review of research activities by a group of diverse individuals who have no direct involvement in the research. Because IRBs are often situated at or near the site of the research, members are expected to have knowledge of and sensitivity to specific conditions affecting the conduct of the research and the protection of participants in geographic communities proximal to the research institution. For example, if a US-sponsored research protocol will enroll subjects in other countries, there, rather than, or in addition to, US IRBs may be important to help identify and resolve particular cultural, religious, or other matters related to the research. Criteria for Institutional Review Board Approval of Research To approve research, an IRB must determine that it meets minimal requirements. Table 4.1 includes the seven minimal regulatory criteria for IRB review and approval (with citations to their sources in 45 CFR 46.111) along with questions that IRBs in the Intramural Research Program (IRP) at the NIH often consider when reviewing research protocols. All clinical researchers, particularly principal investigators (PIs), must be familiar with these requirements and must understand how they apply to their research protocols. In addition, the human subjects protection regulation at 45 CFR 46, state laws, guidance documents from OHRP and published frameworks for ethical review of research provide a broad range of resources for investigators and IRBs. 1. The proposed research design is Scientifically Sound and Will Not Unnecessarily Expose Subjects to Risk (Table 4.1, #1). At a minimum, the IRB should determine that the hypothesis is clear, and that the study design is appropriate. If a research protocol is poorly designed and investigators are not likely to obtain meaningful information, it is not ethically justifiable to expose subjects to any risk, discomfort, or inconvenience.18 However, although IRBs have some members with scientific expertise, they are not constituted to act as primary scientific review committees. In many institutions, protocols undergo pre-IRB scientific review to ensure that protocols sent to the IRB are likely to yield scientifically meaningful results. This is a desirable approach because it allows the IRB to focus its major attention on the protection of subjects. In any event, an IRB should not approve a research protocol that it does not believe to be scientifically sound.

2. Risks are Reasonable in Relation to Anticipated Benefits (Table 4.1, #2). The IRB is required to determine the risks, discomforts, burdens, and benefits of participation in the protocol under consideration. Investigators should describe the informed consent form in the protocol and should describe the risks and benefits of the research in the protocol and informed consent form. Risk is the probability of harm or injury (types of risk include physical, psychological, social, and economic) occurring as a result of participation in a research study. Risk varies in magnitude, but only minimal risk is defined by federal regulations. IRBs are required to assess research-related benefits. The term benefit is not defined in the regulations. Generally, the benefits of research fall into two major categories: (1) direct benefits to individual subjects, for example, in the form of health improvements from the intervention being studied (cure or diminution of symptoms of a disease/ disorder), and (2) benefits to others (e.g., society at large and future patients) because of advancements of knowledge through research.19 To approve research, an IRB must determine that “risks to subjects are reasonable in relation to anticipated benefits, if any, to subjects and the importance of the knowledge that may reasonably be expected to result” (45 CFR 46.111(2)). If research subjects stand to benefit directly from participation in research, because they are receiving treatment or diagnostic procedures, higher risks and discomforts may be justifiable. On the other hand, in research for which there is no prospect of direct benefit to individual subjects, such as research involving healthy volunteers, the IRB must evaluate whether risks to subjects presented by research-related procedures/interventions solely to obtain generalizable knowledge are ethically acceptable. For example, in the IRP of the NIH, IRBs are expected to categorize research-related benefits and risks according to the criteria in Table 4.2. 3. Risks to Subjects are Minimized (Table 4.1, #6). Efforts of the IRB to minimize risks are closely related to, and most likely will be discussed along with, criteria #1 and #2, above. Even when research risks are justifiable and unavoidable, they often can be reduced or managed effectively. IRBs are responsible for assuring that risks are minimized to the extent possible. Ways to minimize risks may include, but are not limited to, assuring that (1) adequate safeguards are incorporated into the protocol that reduce the probability and/or severity of harm(s), (2) monitoring of participant safety and data integrity is appropriate, and (3) investigators are competent in the area(s) being studied. Data safety and monitoring by an

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

INSTITUTIONAL REVIEW BOARDS

53

TABLE 4.1 Institutional Review Board (IRB) Protocol Review Standards: Regulatory Requirements for IRB Review and Documentation in the Minutes Regulatory Review Requirements (46.111)

Possible Questions for IRB Discussion

1. The proposed research design is scientifically sound and will not unnecessarily expose subjects to risk. (45 CFR 46.111(a)(1)(i))

a. Is the hypothesis clear? Is it clearly stated? b. Is the study design appropriate? c. Will the research contribute to generalizable knowledge? Is it ethically permissible to expose subjects to risk?

2. Risks to subjects are reasonable in relation to anticipated benefits, if any, to subjects and the importance of knowledge that may reasonably be expected to result. (45 CFR 46.111(a)(2))

a. What does the IRB consider the level or risk to be? (See risk assessment in Table 5.2.) b. What does the principal investigator consider the level of risk/ discomfort/inconvenience to be? c. Is there prospect of direct benefit to subjects? (See benefit assessment in Table 5.2.)

3. Subject selection is equitable. (45 CFR 46.111(a)(3))

a. Who is to be enrolled? Men? Women? Ethnic minorities? Children (rationale for inclusion/exclusion addressed)? Seriously ill persons? Healthy volunteers? b. Are these subjects appropriate for the protocol?

4. Additional safeguards are required for subjects likely to be vulnerable to coercion or undue influence. (45 CFR 46.111(b))

a. Are appropriate protections in place for vulnerable subjects (e.g., pregnant women, fetuses, socially or economically disadvantaged, decisionally impaired)?

5. Informed consent is obtained from research subjects or their legally authorized representative(s). (45 CFR 46.111(a)(4)) Informed consent will be appropriately documented. (45 CFR 46.111(a)(5)

a. Does the informed consent document include the eight required elements? b. Is the consent document understandable to subjects? c. Who will obtain informed consent (principal investigator, nurse, or other)? In what setting? d. If appropriate, is there a children’s assent? e. Is the IRB requested to waive or alter any informed consent requirement?

6. Risks to subjects are minimized. (45 CFR 46.111(a)(1)) “When appropriate, the research plan makes adequate provision for monitoring the data collected to ensure the safety of subjects.” (45 CFR 46.111(a)(5)

a. Does the research design minimize risks to subjects? b. Would use of a data and safety monitoring board or other research oversight process enhance subject safety?

7. Subject privacy and data confidentiality are maximized. (45 CFR 46.111(a)(7))

a. Will personally identifiable research data be protected to the extent possible from access or use? b. Are any special privacy and confidentiality issues (e.g., use of genetic information) properly addressed?

This table describes regulatory requirements for IRB review of research studies and suggested discussion question for each requirement.

independent person or committee (DSMC) may be appropriate. It is important to assure that IRBs communicate effectively with DSMCs.19 4. Subject Selection is Equitable (Table 4.1, #3). The ethical principle of justice, which requires fair distribution of both the burdens and the benefits of research, underlies the requirement for equitable selection of research subjects. On the one hand, when the NIH funds research, it expects the findings to be of benefit to all persons at risk for the disease, disorder, or condition under study. NIH also requires participation of women and minorities (see Chapter 13), as these populations have historically been underrepresented in clinical research20 and because there are relevant physiological differences and differences in prevalence rates of conditions between the sexes and across ethnic groups “that can affect how disease and treatment manifest themselves.”21

On the other hand, IRBs are required to ensure that subjects (e.g., indigent persons, racial and ethnic minorities, persons confined to institutions, individuals who are socially or politically disenfranchised) are not being systematically selected merely because of their vulnerability, easy availability, their susceptibility to undue influence, their compromised position, or their manipulability, rather than for reasons directly related to the goals and questions of the research. When defining the appropriate group of subjects to be studied in a research protocol, researchers take into account: scientific design, susceptibility to risk of potential subjects (selecting subjects in a way that minimizes risk), the likelihood of direct benefits to subjects and society, and considerations of practicability and fairness. If vulnerable groups of subjects are included in the research, the IRB should determine that

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

54 TABLE 4.2

4. INSTITUTIONAL REVIEW BOARDS

Template Institutional Review Board (IRB) Assessment of Research-Related Risks and Benefits

RISK Definition of minimal risk: Minimal risk means that the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests [45 CRF 46.102(i)]. What is the appropriate risk category for the protocol under consideration? • The research involves no more than minimal risk to subjects. • The research involves more than minimal risk to subjects. • The risk(s) represents a minor increase over minimal risk. • The risk(s) represents more than a minor increases over minimal risk. BENEFIT Definition: A research benefit is considered to be something of healthrelated, psychosocial, or other value to an individual research subject, or something that will contribute to the acquisition of generalizable knowledge. Money or other compensation for participation in research is not considered to be a benefit but, rather, compensation for researchrelated inconveniences. The appropriate benefit category for the under considerations is • The research involves no prospect of direct benefit to individual but is likely to yield generalizable knowledge about the subject’s disorder or condition. • The research involves the prospect of direct benefit to individual subjects. This template provides a framework for IRBs to assess the risks and benefits associated with a given protocol.

protections for these groups are adequate; such safeguards could include consent monitoring, capacity assessment, and adequate data and safety monitoring. IRBs are expected to determine that subject selection as proposed by the researcher in his or her research protocol is scientifically justified and ethically appropriate. 5. Informed Consent is Provided by Research Subjects or Their Legally Authorized Representative(s) (Table 4.1, #5). Although the requirement to obtain informed consent has substantial foundations in law, it is essentially an ethical imperative. It is through informed consent that researchers make operational their duty to respect the rights of prospective subjects to be self-determining, for example to be left alone, to make free choices consistent with the values that are meaningful to them, and to have private information about them shared only in ways for which they give permission.22 However, in practical terms, signing the consent document is the only one element in a subject’s decision-making process about participating in a research protocol. The decision-making process of

prospective subjects can be influenced by various factors, including (1) the written consent document and discussion about the study with the investigator, (2) the knowledge and skills of professionals involved in the process and their relationship to potential subjects (e.g., researchers, nurses), (3) the prospective research subject (e.g., his or her medical and emotional state, primary language, ethnic/cultural background, financial considerations, other personal factors), and (4) the circumstances in which the process takes place (e.g., an emergency room, private practice setting, academic institution). IRBs spend considerable time reviewing the written informed consent document(s). The role of the IRB is to ensure that the consent document contains required elements of consent (Table 4.3) and that it is written at a reading level, and in a format, understandable to prospective subjects. The Final Rule adds additional requirements for informed consent. In addition to reviewing the consent document, IRBs can influence the informed consent process by ensuring that individuals obtaining consent are qualified to take on this important responsibility. For example, an IRB should take into consideration who will obtain informed consent to participate in the protocol and under what circumstances. Depending on the complexity and TABLE 4.3 General Requirements for Informed Consent (45 CFR 46.116) In seeking informed consent, the following information shall be provided to each subject: 1. A statement that the study involves research, and an explanation of the purposes of the research; the expected duration of the subject’s participation; a description of procedures to be followed; and identification of any procedures that are experimental. 2. A description of any foreseeable risks or discomforts to the subject. 3. A description of any benefits to subjects or to others that may reasonably be expected from the research. 4. A disclosure of appropriate alternative procedures or courses of treatment, if any, that might be advantageous to the subject. 5. A statement describing the extent, if any, to which confidentiality of records identifying the subject will be maintained. 6. For research involving greater than minimal risk, an explanation as to whether any compensation and an explanation as to whether any medical treatments are available if injury occurs and, if so, what hey consists of and where further information may be obtained. 7. An explanation of whom to contact for answers to pertinent questions about the research and research subjects’ rights, and whom to contact in the event of a research-related injury to the subjects. 8. A statement that participants is voluntary, refusal to participate will involve no penalty or loss of benefits to which the subject is otherwise entitled, and the subject may discontinue participation at any time without penalty or loss of benefits to which the subject is otherwise entitled. This table summarizes regulatory requirements for informed consent forms.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

INSTITUTIONAL REVIEW BOARDS

risks associated with a research study, an experienced senior researcher, rather than a junior person, may be required to obtain consent. Also, IRBs may exercise their authority to observe or request a third party to observe the consent process and the conduct of the research (45 CFR 46.109(e)). Obtaining informed consent for research participation is a complex process; therefore, it has been a topic of interest, discussion, and publication for many years. In 1966, Dr. Henry Beecher, refer to Chapters 2 and 3, wrote that the two most important elements in ethical research involving human subjects are informed consent (which he acknowledged in some cases was difficult, if not impossible, to obtain) and the “presence of an intelligent, informed, conscientious, compassionate, responsible investigator.”15 His ideas still ring true today. Even though the role of the IRB in promoting subjects’ informed consent is important, it is primarily the responsibility of the investigator who is obtaining the consent to ensure that it is, in fact, informed and valid. 6. Additional Safeguards are Required for Subjects Likely to be Vulnerable to Coercion or Undue Influence (Table 4.1, #4). Vulnerable research subjects are individuals who are relatively or absolutely incapable of protecting their interests. In other words, “they have insufficient power, prowess, intelligence, resources, strength, or needed attributes to protect their own interests through negotiations for informed consent.”23 Vulnerable subjects do not represent a homogeneous group but rather fall into heterogeneous groups whose participation in research may require additional protections. Table 4.4 is a noninclusive list of vulnerable or potentially vulnerable research subjects. It lists individuals who TABLE 4.4 Vulnerable (or Potentially Vulnerable) Research Subjects This is a noninclusive list of common categories of research subjects who have limitations to their capacity to provide informed consent and/or who may be susceptible to coercion or undue influence in decisions about research participation: Comatose people Critically ill people Mentally retarded people/people with dementias/some psychiatric diseases Children NoneEnglish speaking people Educationally/economically deprived people Prisoners Seriously/terminally ill people Paid research volunteers This table enumerates common categories of vulnerable subjects (e.g., subjects who have limitations to their capacity to provide informed consent and/or who may be susceptible to coercion).

55

have no, or limited, capacity to provide informed consent, as well as those who may be particularly susceptible to undue influence or coercion. Federal regulations direct IRBs to ensure that when some or all subjects are likely to be vulnerable to coercion or undue influence, such as children, prisoners, pregnant women, mentally disabled persons, or economically or educationally disadvantaged persons; additional safeguards have been included in the study to protect the rights and welfare of these subjects (45 CFR 46.111(b)). However, little additional practical guidance is provided, except when subjects of research are pregnant women (45 CFR 46, Subpart B), prisoners (Subpart C), and children (Subpart D). Recent guidance on research involving prisoners and children has been useful to IRBs in reviewing such research.24,25 Otherwise, IRBs, in consultation with investigators, are expected to determine when subjects are likely to be vulnerable to coercion or undue influence and therefore at increased risk of acting against their own best interests and to provide additional safeguards appropriate to the particular research protocol under consideration. For example, persons suffering from prolonged or serious illnesses that are refractory to standard therapies, or for which no standard therapies are available, should be considered vulnerable. Although these sick individuals may have the intellectual capacity to provide informed consent, attention must be paid to the validity of the consent. Because of limited choices, out of desperation they may be willing to take serious risks, even for a highly remote prospect of direct benefit. Although this is not necessarily inappropriate, researchers and IRBs need to give careful attention to the informed consent process in protocols studying terminally ill or very sick people. To evaluate the validity of the consent, the IRB may ask that an “uninterested” individual, such as a social worker, a physician not involved in the research, or a research subject advocate, discuss the research study and other clinical or research alternatives with prospective subjects.26 Attention has been given to additional protections for research involving individuals with mental disorders,27 research involving cognitively impaired subjects,28 and research conducted in emergency circumstances.29 7. Subject Privacy and Confidentiality Are Maximized (Table 4.1, #7). Confidentiality refers to the management of information that an individual has disclosed in a relationship of trust; an expectation is that it will not be divulged to others in ways that are inconsistent with an understanding of original disclosure without the person’s permission. Privacy is defined in terms of having control over the extent, timing, and

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

56

4. INSTITUTIONAL REVIEW BOARDS

circumstances of sharing information about oneself (physical, intellectual, or behavioral) with others. Biomedical and behavioral research may invade the privacy of individuals or may result in a breach of data confidentiality. In certain circumstances, a breach of confidentiality may present a risk of serious harm to subjects, for example, when a researcher obtains information about subjects that, if disclosed by the researcher, would jeopardize their jobs or lead to their prosecution for criminal behavior. In other circumstances, such as observation and recording of public behavior, the invasion of privacy may present little or no harm. However, the need for confidentiality exists in virtually all studies in which data are collected about identified subjects.30 In most research, ensuring confidentiality is a matter of following best practices for handling paper and electronic records and also biological samples. While it is not possible to absolutely protect privacy or maintain confidentiality, researchers should be aware of evolving best practices and implement up to date security measures to protect data confidentiality. Common practices include identifying private locations for conducting subject assessments and interviews, substituting codes for personal identifiers, properly disposing of computer sheets and other papers with confidential identified information, deidentifying data, limiting access to identifiable data in a manner consistent with applicable regulations and laws (such as the Privacy Act of 1972 and the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule), and/or storing research records in locked cabinets and in secure and encrypted electronic media. Most researchers are familiar with these routine precautions taken to maintain the confidentiality of data. At a minimum, IRBs should assure themselves that adequate protections will be taken to safeguard the confidentiality of research information to the extent possible. The types and stringency of measures depend on the type of information to be gathered in the study. In any case, guarantees of “absolute” confidentiality should be avoided; in fact, the limits of confidentiality should be clarified. For example, federal officials have the right to inspect research records, including informed consent documents and individual medical records, to ensure compliance with the rules and standards for their programs (e.g., FDA inspections of clinical trial records). More elaborate procedures may be needed in studies in which data are collected on sensitive matters such as sexual behavior, criminal activities, and genetic predilection to disease and studies where social media or Web-based tools are used for data collection. Other federal, state, or local laws address the confidentiality and maintenance of protected health information (PHI). For example, in the research context the

HIPAA Privacy Rule gives patients certain rights over their health information and sets rules and limits on who can look at and receive this information (such as the right to provide an authorization for sharing of data unless a waiver of authorization has been approved by a Privacy Board or IRB). HIPAA, also referred to as the “Privacy Rule,” was a federal legislative response to public concern over potential abuses of the privacy of health information in medical care. The Privacy Rule establishes a category of health information, referred to as PHI, which may be used or disclosed to others only under certain conditions. PHI includes what health-care professionals typically regard as a patient’s personal health information, such as information in a patient’s medical chart. The rule applies to identifiable health information about subjects of clinical research gathered by researchers who qualify as “covered health-care providers.” Therefore, researchers associated with entities covered under HIPAA must understand its requirements to protect the confidentiality of research subjects.31 Continuing Review of Research IRBs are required to conduct continuing review of approved research at least annually, or sooner if they determine that the research presents significant physical, social, or psychological risks to subjects (45 CFR 46.109(e)). Continuing review is required to assure IRBs, investigators, research subjects, and the public that ongoing assessment will protect the rights and welfare of subjects. Requirements regarding what information investigators must submit to an IRB at the time of its continuing review vary according to institutional requirements and whether any subjects continue to be seen. For example, in the IRP of the NIH, investigators are required to submit materials including a copy of the currently approved protocol consent document; a concise summary of the protocol’s progress to date; the reason(s) for continuing the study; the gender/ethnic breakdown of subjects recruited to date; and any scientific developments that bear on the protocol, especially those that deal with risk(s), burdens, or benefits to individual subjects. Also, at the time of continuing review, protocol investigators must report any new equity, consultative, or other relationships with non-NIH entities that might present a real or apparent conflict of interest in the conduct of the protocol. By contrast, if no subjects continue to be seen, a more abbreviated and expedited process is possible. At its continuing review, or at any other time, an IRB may suspend, modify, or terminate approval of research that has been associated with serious harm to subjects or is not being conducted in accord with federal regulatory requirements, ethical guidelines, and/or institutional policies. The Final Rule allows IRBs to not require continuing review for research that is no greater than minimal risk.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

EVALUATION AND EVOLUTION OF THE CURRENT SYSTEM OF RESEARCH OVERSIGHT AND INSTITUTIONAL REVIEW BOARDS

CLINICAL RESEARCHERS AND INSTITUTIONAL REVIEW BOARDS Successful clinical researchers know that strong ethical practices go hand in hand with high-quality scientifically valid research involving human subjects. These researchers understand the IRB’s mandate to protect human subjects and strive to work effectively with the IRB. Researchers’ knowledge of and expertise in the ethical dimensions of their research activities are important to IRBs for several reasons. First, clinical researchers can help educate IRBs about human subject protection issues related to their research protocols. It is helpful to IRBs in understanding and resolving human subject protection issues if PIs are knowledgeable about the IRB review standards and are expert in applying them to their protocols. For example, when writing a protocol to test an investigational drug in persons with Alzheimer’s disease, the researcher should provide clear scientific justification in the protocol for including individuals with cognitive impairment in research. The investigator should describe procedures for (1) assessing whether subjects have the capacity to provide consent, (2) selecting a legally authorized representative to give permission for subjects who cannot provide consent, and (3) any additional protections afforded to subjects. The PI may propose that a person otherwise not involved in the research should monitor the informed consent process to ensure that subjects and/or their representatives understand the investigational nature of the study and its risks. This approach assists the IRB greatly by providing it with a thorough overview of human subject protection issues specific to the protocol under review, along with a description of measures proposed by the PI to resolve these issues. Second, in the early phases of scientifically innovative research, ethical and human subject protection issues may be unique and/or unclear; researchers who are experts in the scientific and ethical aspects of their research can provide IRBs with invaluable guidance in areas of uncertainty. IRB decisions are matters of judgment, and when highly innovative research is reviewed, it is particularly important that such judgments take into account relevant ethical thinking and scientific knowledge. Increasingly, institutions conducting biomedical research have processes for research ethics consultation, similar to clinical ethics consultation to support stakeholders in research studies. The NIH IRP was among the first institutions in the country to establish a clinical research ethics consultation service and has published approaches and cases in such consultation.32

57

EVALUATION AND EVOLUTION OF THE CURRENT SYSTEM OF RESEARCH OVERSIGHT AND INSTITUTIONAL REVIEW BOARDS Proposed Changes to Current Oversight of Research With Human Subjects In the past 30 plus years, since the current research system was put into place, the landscape of human subjects research has changed considerably. Funding sources have changed. For example, the pharmaceutical industry’s share of total funding of biomedical research has increased from 32% in 1980 to 62% in the early 2000s, whereas the federal government’s share has fallen.33 There has been an increase in multisite studies as opposed to single-institution studies that were the norm when the research oversight system was conceptualized. Because of globalization and advances in technology, there is a greater diversity of research including more health services and comparative effectiveness research, international research, biospecimen research utilizing ever advancing analytic tools and “big data” research leveraging data from across studies and sources.34,35 Concerns about sufficient oversight and the potential for institutional and investigator conflict of interest were voiced in the wake of the deaths of a research subject in a gene therapy trial and a healthy volunteer in an asthma challenge study.36,37 The US Government Accountability Office (GAO) and the Institute of Medicine (IOM) conducted formal evaluations of the human subject protections oversight system, including the function of IRBs, and the efficiency and effectiveness of the current system have come under scrutiny and critique by academicians.32,38 Concerns raised included “(1) structural problems deriving from the organization of the system as established by federal regulations, (2) procedural problems stemming from the ways in which IRBs operate, and (3) performance assessment problems from the systematic assessment of current protections.”39 Recommendations made by the GAO and IOM were aimed at improving the education of researchers, IRB members, and institutional officials overseeing research involving human subjects; ensuring that IRBs have sufficient time and resources for adequate review; and strengthening federal oversight of research. Others have called for more significant changes, including establishment of a single regulatory office for all human subjects research conducted in the United States.39 In response to numerous proposals for reform of the system, OHRP promulgated an Advance Notice of

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

58

4. INSTITUTIONAL REVIEW BOARDS

Propose Rule Making (ANPRM), “Human Subjects Research Protections: Enhancing Protections for Research Subjects and Reducing Burden, Delay and Ambiguity for Investigators.”40 OHRP subsequently published a Notice of Public Rulemaking in 2014.41 The NPRM was revised based on the 2186 comments received and the Final Rule was published on January 19, 2017. Its provisions go into effect on January 19, 2018, except for the requirement for single IRB review of multisite studies which goes into effect one year later. The Final Rule proposes changes in eight broad areas: (1) adding additional requirements for informed consent forms and processes; (2) allowing the use of broad informed consent for secondary analyses of stored biospecimens; (3) creating a category of IRB review called limited IRB review, which can take multiple forms; (4) adding additional exemption categories and allowing exemption determinations to be made without administrative, institutional, or IRB review; (5) modifications to requirements for waiver or alteration of consent for biospecimen research to make such mechanisms less common; (6) mandate the use of single IRBs for review of domestic multisite studies (with provision for certain exemptions); (7) eliminate the requirement for continuing review for studies that are limited to data analysis or limited observational follow up activities where review is thought to not meaningfully contribute to the protection of human subjects; (8) additional requirements, designed to increase subject understanding, for informed consent forms.44 One of the most immediate and dramatic changes is the requirement for single IRB review of multisite studies. Review of multisite research studies by IRBs at each site has been cited as a common impediment to the efficient conduct of multisite research. Some have argued that this review is redundant, impedes the timely conduct of research, increases the complexity of informed consent forms and does not enhance the protections of human subjects.16,45e48 Another disadvantage of institutionally based IRBs is the potential for IRBs to focus on the interests of the institution as opposed to those of participants, thereby generating potential for conflict of interest.39 The regulations at 45 CFR 46.114 allow an institution to cede IRB review for a study conducted under its FWA to another institution, and reliance on a single or central IRB for multisite research is becoming increasingly common. NIH has published a policy requiring single IRB review for most NIH-funded multisite studies.49 There remains resistance among institutions, IRB members and investigators to use of a single IRB, especially one that is not local to the community being studied.50,51 There is some experience with multiple models of central IRB review, and while data are limited there is some preliminary evidence to suggest that efficiency in review time and potentially costs may be possible with mature, well-structured central IRB operations.52e55

Critique and Proposed Changes to Institutional Review Board Operations The function of the IRB system, distinct from the overall research oversight system, is currently under considerable criticism.56 Some claim “that IRBs should be radically overhauled.”57 The current IRB system deserves serious reevaluation; its strengths should be acknowledged and supported, and its weaknesses should be addressed.58e60 One challenge to broader reform of IRB operations is a lack of empirical studies and established performance measures for the quality of IRB review. Despite their long-standing central role for the protection of human subjects, relatively little published research has explored their deliberations.61,62 Most studies examined only IRB records and procedures and IRB members’ knowledge and attitudes; little published work has evaluated the protocol review activities of IRBs as conducted in their convened meetings. One approach to achieving consistency and potentially high-quality IRB review is to undergo review by an outside entity according to published standards. The IOM encouraged the development of such institutional accreditation standards that build on federal regulations and urged that accrediting organizations be nongovernmental entities.39 The Association for the Accreditation of Human Research Protection Programs (AAHRPP) is the main accrediting organization in the United States (see Chapter 5). Its standards indicate that IRBs are one of several important elements (or domains) in an institution’s overall Human Research Protection Program (HRPP). Other critical domains addressed by these standards are the roles and responsibilities of the institution, institutional/organizational officials, researchers, research staff, and research subjects. The process of accreditation includes organizational self-evaluation and site visits by independent human subject protection experts. As of June 2014, greater than 60% of US research intensive institutions and 65% of US medical schools, as well as pharmaceutical companies, independent IRBs, and international institutions have received or have begun application for full accreditation by AAHRPP.63 Interestingly, some of the strengths of the IRB system also contribute to its potential weaknesses. For example, having IRBs situated at the site of the research provides the advantage that research is reviewed by people most likely to be familiar with the researchers and with institutional and other local factors relevant to the protection of research participants. However, if IRB members are predominantly employees of the research institution, there is the potential for conflict of interest, particularly when reviewing research protocols involving large amounts of grant or other support money to the institution. Many organizations take steps to address and

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

59

REFERENCES

minimize these real or potential conflicts of interest of IRB members.38 One strength of IRBs situated in research institutions is that IRB can have an important educational role within the organization.32 For example, the NIH IRP has 12 IRBs consisting of approximately 200 members who provide a significant educational resource to the NIH research community. The ability of an IRB to fulfill its mandate is influenced by several factors, including the knowledge and experience of members and institutional resources and commitment. IRB decisions are matters of judgment and therefore depend on an understanding and wise application of ethical guidelines and regulatory requirements, as well as an appreciation of local influences such as cultural considerationsdthough this last function need not be performed by a local IRB. Efforts to improve the abilities and procedures of IRBs should promote independence of IRB review; identify which functions commonly performed by IRBs may be better situated elsewhere in an institution; promote a culture of awareness of conflict of interest among institutions and investigators and implement strong policies for identifying and managing conflict of interest; minimize duplicative IRB review; conduct research to develop measures of IRB consistency and thoroughness; and identify the costs associated with IRB review and other functions that IRBs are required to take on to ensure adequate funding.

CONCLUSION Research involving human subjects, even if they may benefit directly from participation, is a different kind of enterprise from the routine practice of medicine. The goals of research include not only the welfare of individual subjects but also the gathering of scientific data for future application. Our society has granted a conditional privilege to perform research with human subjects: the research must be scientifically sound and must be conducted in a manner that protects the rights and safeguards the welfare of participants. The IRB system is well developed but is ever evolving. The current US system for protecting human research subjects, including the role of IRBs, is undergoing serious evaluation and change. Successful evolution of the system depends on learning from the past, understanding current and future needs, and applying knowledge to implement meaningful improvements. Researchers, research participants and institutions, and others, particularly the American people who bear the burdens of research and to whom the benefits accrue, all have a stake in the process.

SUMMARY QUESTIONS 1. The mandate of IRBs provided in federal regulations (45 CFR 46) is to a. Conduct primary scientific review of research protocols to ensure that they are scientifically sound b. Protect the rights and safeguard the welfare of human research subjects by conducting independent ethical review c. Ensure that research protocols are consistent with contemporaneous national public policies d. Approve research protocols quickly to support acquisition of research funding 2. Which of the following statements is/are correct? a. Institutional officials may choose not to allow an IRB-approved research protocol to go forward but they are not permitted to override IRB disapproval b. IRBs are required to conduct continuing review of active research protocols at least once per year c. IRBs are expected to identify when potential researchers may be vulnerable to coercion or undue influence and to provide additional appropriate protections d. Vulnerable research subjects include, but are not limited to, children, prisoners, and pregnant women e. All of the above 3. Our society has decided that review of research involving human subjects should undergo review by IRBs because a. Clinical researchers have an inherent conflict of interest when balancing their roles as researchers and health-care professionals b. Review of research by persons who do not have a role in the research is one way to promote ethically sound research c. IRB review is one way research institutions, researchers, IRBs, and others are held publicly accountable for their decisions and actions regarding their clinical research activities d. IRBs are made up of diverse persons with varied backgrounds to promote a comprehensive approach to protecting the rights and safeguarding the welfare of research subjects e. All of the above

References 1. Levine J. The nuremberg code. Ethics and regulation of clinical research. New Haven, CT: Yale University Press; 1988. 2. Katz J. The Nuremberg code and the Nuremberg trial: a reappraisal. JAMA 1988;276:1662e6. 3. McCarthy C. The origins and policies that govern institutional review boards. In: Emanuel E, Grady C, Crouch R, Lie R, Miller F, Wendler D, editors. The oxford textbook of clinical research ethics. New York: Oxford University Press; 2008 [Chapter 50].

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

60

4. INSTITUTIONAL REVIEW BOARDS

4. U.S. Congress. National research act (PL 93-348). July 12, 1974. https://history.nih.gov/research/downloads/PL93-348.pdf. 5. National Commission for the Protection of Human Subjects of Biomedical & Behavioral Research. The Belmont report: ethical principles and guidelines for the protection of human subjects of research. Publication No. 887e809. Washington, DC: Government Printing Office; 1979. p. 4e8. 6. DHHS. Regulations for protection of human subjects, 45 CFR, Part 46. http://www.hhs.gov/ohrp/regulations-and-policy/regulations/ 45-cfr-46/index.html; 1981. 7. Porter J, Koski G. Regulations for the protection of humans in research in the United States. In: Emanuel E, Grady C, Crouch R, Lie R, Miller F, Wendler D, editors. The oxford textbook of clinical research ethics. New York: Oxford University Press; 2008. p. 156e67 [Chapter 15]. 8. Parts 50 FDA regulations for the protection of human subjects, 21 CFR. 1996. p. 56. 9. Food and Drug Administration. Guidance for institutional review boards and clinical researchers. Information sheet on “Significant Differences in FDA and HHS Regulations”. 2000. Retrieved from: http://www.fda.gov/ScienceResearch/SpecialTopics/ RunningClinicalTrials/EducationalMaterials/ucm112910.htm. 10. Borror K, Carome M, McNeilly P, et al. A review of OHRP compliance oversight letters. IRB: Ethics Hum Res SeptembereOctober 2003;25(5):1e4. 11. Weil C, Rooney L, McNeilly P, Cooper K, Borror K, Andreason P. OHRP compliance oversight letters: an update. IRB: Ethics Hum Res 2010;32:1e6. 12. Office of Human Research Protections. http://www.hhs.gov/ ohrp/compliance-and-reporting/index.html. 12a. HHS. https://www.gpo.gov/fdsys/pkg/FR-2017-01-19/html/ 2017-01058.htm. 13. Wichman A, Mills D, Sandier AL. Exempt research: procedures in the intramural research program of the National Institutes of Health. Rev Hum Subjects Res MarcheApril 1996:3e5. 14. Faden RR, Beauchamp TL. A history and theory of informed consent. New York: University Press; 1986. 15. Beecher HK. Ethics and clinical research. N Engl J Med 1996;274: 1354e66. 16. Pogorzelska M, Stone PW, Gross Cohn E, et al. Changes in the institutional review board submission process for multicenter research over 6 years. Nurs Outlook 2010;58:181e7. 17. Levine J. The nuremberg code. Ethics and regulation of clinical research. New Haven, CT: Yale University Press; 1988. p. 427e9. 18. Emanuel E, Wendler D, Grady C. An ethical framework for biomedical research. In: Emanuel E, Grady C, Crouch R, Lie R, Miller F, Wendler D, editors. The oxford textbook of clinical research ethics. New York: Oxford University Press; 2008. p. 123e40. 19. Taylor HA, Chaisson L, Sugarman J. Enhancing communication among data monitoring committees and institutional review boards. Clin Trials 2008;5:277. 20. NIH policy and guidelines on the inclusion of women and minorities as subjects in clinical research e amended. October 2001. Retrieved from: http://grants.nih.gov/grants/funding/women_min/ guidelines_amended_10_2001.htm. 21. Dresser R. Wanted: single, white male for medical research. Hastings Cent Rep 1992;1(22):24e9. 22. Levine J. The nuremberg code. Ethics and regulation of clinical research. New Haven, CT: Yale University Press; 1988. p. 96. 23. Levine J. The nuremberg code. Ethics and regulation of clinical research. New Haven, CT: Yale University Press; 1988. p. 72. 24. Institute of Medicine. Ethical considerations for research involving prisoners. Washington, DC: IOM; 2006. 25. Institute of Medicine. Ethical conduct of clinical research involving children. Washington, DC: IOM; 2004.

26. American Academy of Neurology. Position statement. Ethical issues in clinical research in neurology. Neurology 1998;50:592e5. 27. National Bioethics Advisory Commission. Research involving persons with mental disorders that may affect decision making capacity. Washington, DC: Government Printing Office; 1998. 28. Alzheimer’s Association. Research consent for cognitively impaired adults: recommendations for institutional review boards and investigators. Alzheimer Dis Assoc Disord 2004;18:171e5. 29. Food and Drug Administration. Protection of human subjects: exception from informed consent requirements for emergency research; Revised 2010. 21 CFR, Part 50.24. 30. Standards for privacy of individually identifiable health information (2000). 45 CFR 46, Parts 160, 165 Fed Reg 2000;65: 82462e82829:(See also HHS Office for Civil Rights, Retrieved from: http://www.hhs.gov/ocr/hipaa. For effect of the Privacy Rule on clinical research, see http://privacyruleandresearch.nih. gov/clin_research.asp.). 31. U.S. Government Accountability Office, Health, Education, and Human Services division. Report to the ranking minority member, Senate Commission on governmental affairs. 1996. Continued Vigilance, critical to protecting human subjects. Publication No. 96e72. 32. Danis M, Largent E, Wendler D, et al. Research ethics consultation: a casebook. New York: Oxford University Press; 2012. 33. Bekelman JE, Li Y, Gross CP. Scope and impact of financial conflicts of interests in biomedical research. JAMA 2003;289:454e65. 34. Emanuel E, Menikoff J. Reforming the regulations governing research with human subjects. NEJM 2011;365(12):1145e50. 35. Lo B, Barnes M. Federal research regulations for the 21st century. NEJM 2016;374:1205e7. 36. Savulescu J, Spriggs M. The hexamethonium asthma study and the death of a normal volunteer in research. J Med Ethics 2002;28:3e4. 37. Resnik D, Ariansen J, et al. Institutional conflict of interest policies at U.S. Academic Research Institutions. Acad Med 2016;91(2):242e6. 38. Institute of Medicine. Responsible research: a systems approach to protecting research participants. In: Federman D, Hanna K, Rodriguez L, editors. National Academies Press; 2002. Retrieved from: http://www.nap.edu/search/?term¼ResponsibleþResearch% 3AþAþSystemsþApproachþtoþProtectingþResearchþParticipants. 39. Emanuel E, Wood A, Fleischman A, et al. Oversight of human participants research: identifying problems to evaluate reform proposals. Ann Intern Med 2004;141:282e91. 40. Department of Health, Human Services. Human subjects research protections: enhancing protections for research subjects and reducing burden, delay, and ambiguity for investigators. Fed Regist 2011;76(143):44512e31. Retrieved from: http://www.gpo.gov/ fdsys/pkg/FR-2011-07-26/html/2011-18792.htm. 41. Department of Health and Human Services. Federal policy for the protection of human subjects. Fed Regist 2015;80(173):53933e4061. 42. Katz J, Capron A, Glass E. Experimentation with human beings: the authority of the investigator, subject, professions, and state in the human experimentation process. New York: Russell Sage Foundation; 1972. p. 9e44. 43. Brandt A. Racism and research: the case of the Tuskegee syphilis study. Hastings Cent Rep 1978;8(6):21e9. 44. Menikoff J, Kaneshiro J, Prtichard I. The Common Rule updated. NEJM January 19, 2017;376(3). 45. Green LA, Lowery JC, Kowalski CP, et al. Impact of institutional review board practice variation on observational health services research. Health Serv Res 2006:214e30. 46. Beardsmore CS, Westaway JA. The shifting sands of research ethics and governance: effect of research in paediatrics. Arch Child Dis 2007;92(1):80e1. 47. Greene SM, Geiger AM. A Review finds that multicenter studies face substantial challenges but strategies exist to achieve Institutional Review Board approval. J Clin Epidemiol 2006;59:784e90.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

48. McWilliams R, Hoover-Fond J, Hamosh A, et al. Problematic variation in local institutional review of a multicenter genetic epidemiology study. JAMA 2003;290:360e6. 49. National Institutes of Health. Final NIH policy on the use of a single “Institutional Review Board for Multisite Research”. 81 FR 40325. 2016. p. 40325e4033. Retrieved from: https://www.federalregister.gov/ articles/2016/06/21/2016-14513/final-nih-policy-on-the-use-ofa-singleinstitutional-review-board-for-multi-site-research. 50. Loh ED, Meyer RE. Medical schools’ attitudes and perceptions regarding the use of central institutional review boards. Acad Med 2004;79:644e51. 51. Klitzman R. How local IRBs view central IRBs in the US. BMC Med Ethics 2011;12:13. 52. Christian MC, Goldberg JL, Killen J, et al. A central institutional review board for multi-institutional trials. N Engl J Med 2002;346: 1405e8. 53. Wagner TH, Murray C, Goldberg J, et al. Cost and benefits of the national cancer Institute central institutional review board. J Clin Oncol 2007;28(4):662e6. 54. Kaufmann P, O’Rourke PP. Central institutional review board review for an academic trial network. Acad Med March 2015; 90(3):321e3. 55. Slutsman J, Hirschfeld S. A federated model of IRB review for multisite studies: a report on the national children’s study federated IRB initiaitve. IRB: Ethics Hum Res 2014;36(6):1e6.

61

56. Edgar H, Rothman D. The institutional review board and beyond: future challenges to the ethics of human experimentation. Millbank Q 1995;73:489e506. 57. Annas G. Research censorship on campus. Review of the censor’s hand: the misregulation of human-subject research, by Carl. E. Schneider. The new rambler review: an online review of books. Retrieved at: http://newramblerreview.com/component/content/article? id¼102:research-censorship-on-campus; 2015. 58. Chadwick GL, Dunn CM. Institutional review boards: changing with the times? J Public Health Manag Pract 2000;6:19e27. 59. Grady C. Do IRBs protect human research participants? JAMA 2010;304:1122e3. 60. Beh HG. The role of institutional review boards in protecting human subjects: are we really ready to fix a broken system? L Psychol Rev 2002;26:1e47. 61. Abbott L, Grady C. A systematic review of the empirical literature evaluating IRBs: what we know and what we still need to learn. J Empirical Res Hum Res Ethics 2011;6:3e20. 62. Candilis PJ, Lidz CW, Arnold RM. The need to understand IRB deliberations. IRB: Ethics Hum Res 2006;28:1e5. 63. Association for the Accreditation of Human Research Protection Programs (AAHRPP). News release: latest AAHRPP accreditations include NY state department of health and India’s national comprehensive cancer center. June 12, 2014. Retrieved from: http://www.aahrpp. org/learn/news-releases.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

5 Accreditation of Human Research Protection Programs Elyse I. Summers, Michelle Feige Association for the Accreditation of Human Research Protection Programs, Inc., Washington, DC, United States

O U T L I N E A Brief History

63

Principles of Accreditation What AAHRPP Expects From Organizations What Organizations Can Expect From AAHRPP

64 64 64

Domain II: Institutional Review Board or Ethics Committee Domain III: Researcher and Research Staff

Human Research Protection Programs: The Shift to Shared Responsibility 65 The Accreditation Standards Domain I: Organization

66 66

Steps to Accreditation

70

Value of Accreditation

70

Summary Questions

72

References

72

A BRIEF HISTORY

A global nonprofit organization, the Association for the Accreditation of Human Research Protection Programs, Inc. (AAHRPP), was founded in 2001 to accredit high-quality human research protection programs (HRPPs) in every sector of the human research enterprise. Accreditation-eligible organizations include academic institutions, contract research organizations, government agencies, hospitals, independent institutional review boards (IRBs), private entities, research institutes, and dedicated research sites. AAHRPP accreditation demonstratesdto the public, research participants, other research organizations, and government and industry sponsorsdthat an organization meets rigorous standards for protecting research participants and is committed to high-quality, ethical research. The AAHRPP-accreditation model is voluntary, peer-driven, and educationally based and has the complementary outcome of raising research participant protection to an organizational priority.

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00005-8

67 69

For the US research community, the late 1990s and early 2000s will long be remembered for a number of high-profile research protection deficiencies, followed by the temporary government shutdown of several prestigious research programs. One of the most serious failures resulted in the death on September 17, 1999 of Jesse Gelsinger, a student enrolled in a gene-transfer trial at the University of Pennsylvania. The Gelsinger case shined a spotlight on issues including deficiencies in the informed consent process, investigator conflict of interest, and the reporting of adverse events. The case also prompted congressional hearings on the safety of US clinical trials and contributed to calls for fundamental improvements to safeguard participants and restore public confidence in research. Two entities, the nonprofit Institute of Medicine (IOM)1 (now the National Academy of Medicine) and the National

63

Copyright © 2018. Published by Elsevier Inc.

64

5. ACCREDITATION OF HUMAN RESEARCH PROTECTION PROGRAMS

Bioethics Advisory Commission,2 issued reports acknowledging that accreditation offered promise as part of a multipronged solution. Seven highly respected organizationsdthe Association of American Universities (AAU), Association of American Medical Colleges, Association of Public and Land-grant Universities, Consortium of Social Science Associations, Federation of American Societies for Experimental Biology, National Health Council, and Public Responsibility in Medicine and Researchdled the charge to establish AAHRPP. These “Founding Members” incorporated AAHRPP in April 2001, the same month the IOM issued its report, Preserving Public Trust: Accreditation and Human Research Participant Protection Programs. Six months later, AAHRPP began developing its accreditation standards, which were released in February 2002. The first accreditations followed 14 months later. As of September 2017, 247 organizations were AAHRPP accredited, including 46 that are located outside the United States. AAHRPP has accredited organizations in 47 US states and in Belgium, Brazil, Canada, China, India, Mexico, Republic of Korea, Saudi Arabia, Singapore, South Africa, Taiwan, and Thailand. All major independent US IRBs have earned AAHRPP accreditation. In addition, 70% of US medical colleges and 84% of the top National Institutes of Health (NIH)-funded academic medical centers are either AAHRPP accredited or have begun the accreditation process. The intramural research program of the US NIH, the world’s largest public funder of research, has earned accreditation, as has Pfizer, Inc., the largest industry sponsor of clinical research. Although AAHRPP accreditation is substantively grounded in the US federal regulations, the AAHRPP accreditation standards and process also reflect the local requirements of any non-US jurisdiction where accreditation is sought. Perhaps most important, AAHRPP accreditation has proved to promote highquality research and to help organizations worldwide strengthen their HRPPs.

PRINCIPLES OF ACCREDITATION AAHRPP has adopted nine principles that serve as the foundation for accreditation and the AAHRPP accreditation standards.3 Together, these principles also set forth what AAHRPP expects from organizations and what organizations can expect from AAHRPP.

What AAHRPP Expects From Organizations 1. Protecting the rights and welfare of research participants must be an organization’s first priority. An

organization should promote a research environment where ethical, productive investigation is valued. 2. Protecting research participants is the responsibility of everyone within an organization and is not limited to the IRB or ethics committee (EC). Accreditation examines whether the policies and procedures of the organization as a whole result in a coherent, effective system to protect research participants and whether every role-player understands his or her responsibilities. 3. Striving to meet or exceed the federal requirements and to continually seek new safeguards for protecting research participants while advancing scientific progress must be integrated into an organization’s mission. Organizations can rely on AAHRPP standards to fill a void in instances where there is little or no regulatory guidance on research protections. Generally, where there is clear regulatory language and guidance that spells out a government agency’s view, AAHRPP standards should not be interpreted to require greater than what the regulations require.

What Organizations Can Expect From AAHRPP 4. The standards for protecting participants in human research will be clear, specific, and applicable to research across the full range of settings (e.g., university-based biomedical, behavioral and social science research; IRBs; hospitals; government agencies; and others). Standards will address any special concerns (e.g., the use of vulnerable populations or heightened risk to privacy and confidentiality) that may arise in each setting. 5. The standards will identify outcome measures that organizations can use to assess and demonstrate quality improvement over time. 6. The standards will be performance based, using objective criteria and measurable outcomes to evaluate whether an HRPP effectively implements the standards. The application for accreditation will be reviewed by the Council on Accreditation, which will determine whether standards are met and, where appropriate, will include commendations for “areas of distinction” for organizations that demonstrate exemplary or innovative practices. When organizations struggle to meet specific standards, the Council will recommend various methods of satisfying the unmet requirements. 7. The accreditation process will provide a clear, understandable pathway to help organizations achieve accreditation, and AAHRPP staff will be available with assistance throughout the process.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

HUMAN RESEARCH PROTECTION PROGRAMS

8. The accreditation process will be collegial, educational, and interactive and will include discussion with and constructive feedback from AAHRPP staff. The accreditation process will identify areas in which the HRPP does not yet meet the standards and give organizations the opportunity to discuss potential program improvements. 9. The accreditation process will be responsive to changes in federal regulations and to standards that will evolve based on the changing landscape of the research enterprise.

HUMAN RESEARCH PROTECTION PROGRAMS: THE SHIFT TO SHARED RESPONSIBILITY For decades, research oversight was considered the purview of IRBs and ECs. Calls for broader, organizational responsibility came after the same series of US research incidents that helped give rise to AAHRPP accreditation. In 2000, the year before AAHRPP’s founding, AAU issued a report4 recommending increased vigilance by senior university management and training for all personnel involved in research with human participants. In essence, AAU was making a case for broader responsibility for research protections beyond the IRB. The IOM took an even stronger position. In its April 17, 2001 report, Preserving Public Trust: Accreditation and Human Research Participant Programs, IOM argued for a more comprehensive approach toward human research protections. Specifically, the report called for “a broader human research participant protection system than just the IRB, with multiple functional elements that in total are referred to as human research participant protection programs, or HRPPPs.” Under this approach, as shown in Fig. 5.1, responsibilities

FIGURE 5.1

Human research protection program.

65

are shared among numerous players, including senior organization officials, conflict of interest committees, education programs, auditing and compliance oversight functions, offices of sponsored programs, pharmacy services, researchers and research staff, and the IRB or EC. As its name indicates, from the beginning AAHRPP recognized and advocated for a systematic, organization-wide approach to research protections. Although AAHRPP does not mandate that organizations label their protection programs as HRPPs, the accreditation standards do require a comprehensive, integrated program that affords protections for all research participants. Furthermore, the AAHRPP accreditation standards acknowledge the responsibilities of all those involved in research, including IRB members and IRB staff, researchers and research staff, sponsors, and others across the research enterprise. The AAHRPP requirement for shared responsibility and an integrated HRPP applies even for an organization that outsources all research review to external IRBs. Under these circumstances, the research organization still has an obligation to protect research participants and ensure the integrity of the research. To help organizations understand and meet the accreditation standards, in 2009 the AAHRPP Board of Directors identified the following characteristics of high-quality HRPPs. • The commitment to human research permeates the entire organization, starting with senior leadership who set the example by promoting ethical and productive human research. • Researchers, IRB professionals, and others involved in protecting research participants communicate and collaborate to achieve a shared vision that emphasizes the importance of ethical human research. • The organization sets as a high priority the protection of human research participants and safeguarding their well-being. • The organization advances discovery by publishing or sharing new knowledge. • The consent process is ethical, comprehensive, and informative. • The organization manages conflicts of interest to preserve the integrity of human research. • The organization recognizes and fulfills its responsibility to the public by promoting community outreach and education efforts that help build public trust and support for human research. • Researchers and IRBs view the HRPP as efficient and effective.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

66

5. ACCREDITATION OF HUMAN RESEARCH PROTECTION PROGRAMS

THE ACCREDITATION STANDARDS AAHRPP accreditation standards are divided into three domains: Organization, IRB or EC, and Researcher and Research Staff. These domains represent the three primary spheres of responsibility within an HRPP. Each domain contains practice-based standards; elements further specify what is required to meet a standard.

Domain I: Organization Domain I focuses on the overarching organization, including policies for financial disclosures, clinical trials, education and training in research ethics, scientific review, community engagement, and quality improvement. Standard I-1: The organization has a systematic and comprehensive HRPP that affords protections for all research participants. Individuals within the organization are knowledgeable about and follow the policies and procedures of the HRPP. Element I.1.A. The organization has and follows written policies and procedures for determining when activities are overseen by the HRPP. Element I.1.B. The organization delegates responsibility for the HRPP to an official with sufficient standing, authority, and independence to ensure implementation and maintenance of the program. Element I.1.C. The organization has and follows written policies and procedures that allow the IRB or EC to function independently of other organizational entities in protecting research participants. Element I.1.D. The organization has and follows written policies and procedures setting forth the ethical standards and practices of the HRPP. Relevant policies and procedures are made available to sponsors, researchers, research staff, research participants, and the IRB or EC, as appropriate. Element I.1.E. The organization has an education program that contributes to the improvement of the qualifications and expertise of individuals responsible for protecting the rights and welfare of research participants. Element I.1.F. The organization has and follows written policies and procedures for reviewing the scientific or scholarly validity of a proposed research study. Such procedures are coordinated with the ethics review process. Element I.1.G. The organization has and follows written policies and procedures that identify applicable laws in the localities where it conducts human research, takes them into account in the

review and conduct of research, and resolves differences between federal or national law and local laws. Standard I-2: The organization ensures that the HRPP has resources sufficient to protect the rights and welfare of research participants for the research activities that the organization conducts or oversees. Standard I-3: The organization’s transnational research activities are consistent with the ethical principles set forth in its HRPP and meet equivalent levels of participant protection as research conducted in the organization’s principal location while complying with local laws and taking into account cultural context. Standard I-4: The organization responds to the concerns of research participants. Element I.4.A. The organization has and follows written policies and procedures that establish a safe, confidential, and reliable channel for current, prospective, or past research participants or their designated representatives that permit them to discuss problems, concerns, and questions; obtain information; or offer input with an informed individual who is unaffiliated with the specific research protocol or plan. Element I.4.B. The organization conducts activities designed to enhance understanding of human research by participants, prospective participants, or their communities, when appropriate. These activities are evaluated on a regular basis for improvement. Element I.4.C. The organization promotes the involvement of community members, when appropriate, in the design and implementation of research and the dissemination of results. Standard I-5: The organization measures and improves, when necessary, compliance with organizational policies and procedures and applicable laws, regulations, codes, and guidance. The organization also measures and improves, when necessary, the quality, effectiveness, and efficiency of the HRPP. Element I.5.A. The organization conducts audits or surveys or uses other methods to assess compliance with organizational policies and procedures and applicable laws, regulations, codes, and guidance. The organization makes improvements to increase compliance, when necessary. Element I.5.B. The organization conducts audits or surveys or uses other methods to assess the quality, efficiency, and effectiveness of the HRPP. The organization identifies strengths and weaknesses of the HRPP and makes improvements, when necessary, to increase the quality, efficiency, and effectiveness of the program.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

THE ACCREDITATION STANDARDS

Element I.5.C. The organization has and follows written policies and procedures so that researchers and research staff may bring forward to the organization concerns or suggestions regarding the HRPP, including the ethics review process. Element I.5.D. The organization has and follows written policies and procedures for addressing allegations and findings of noncompliance with HRPP requirements. The organization works with the IRB or EC, when appropriate, to ensure that participants are protected when noncompliance occurs. Such policies and procedures include reporting these actions, when appropriate. Standard I-6: The organization has and follows written policies and procedures to ensure that research is conducted so that financial conflicts of interest are identified, managed, and minimized or eliminated. Element I.6.A. The organization has and follows written policies and procedures to identify, manage, and minimize or eliminate financial conflicts of interest of the organization that could influence the conduct of the research or the integrity of the HRPP. Element I.6.B. The organization has and follows written policies and procedures to identify, manage, and minimize or eliminate individual financial conflicts of interest of researchers and research staff that could influence the conduct of the research or the integrity of the HRPP. The organization works with the IRB or EC in ensuring that financial conflicts of interest are managed and minimized or eliminated, when appropriate. Standard I-7: The organization has and follows written policies and procedures to ensure that the use of any investigational or unlicensed test article complies with all applicable legal and regulatory requirements. Element I.7.A. When research involves investigational or unlicensed test articles, the organization confirms that the test articles have appropriate regulatory approval or meet exemptions for such approval. Element I.7.B. The organization has and follows written policies and procedures to ensure that the handling of investigational or unlicensed test articles conforms to legal and regulatory requirements. Element I.7.C. The organization has and follows written policies and procedures for compliance with legal and regulatory requirements governing emergency use of an investigational or unlicensed test article. Standard I-8: The organization works with public, industry, and private sponsors to apply the requirements of the HRPP to all participants.

67

Element I.8.A. The organization has a written agreement with the sponsor that addresses medical care for research participants with a research-related injury, when appropriate. Element I.8.B. In studies where sponsors conduct research site monitoring visits or conduct monitoring activities remotely, the organization has a written agreement with the sponsor that the sponsor promptly reports to the organization findings that could affect the safety of participants or influence the conduct of the study. Element I.8.C. When the sponsor has the responsibility to conduct data and safety monitoring, the organization has a written agreement with the sponsor that addresses provisions for monitoring the data to ensure the safety of participants and for providing data and safety monitoring reports to the organization. Element I.8.D. Before initiating research, the organization has a written agreement with the sponsor about plans for disseminating findings from the research and the roles that researchers and sponsors will play in the publication or disclosure of results. Element I.8.E. When participant safety could be directly affected by study results after the study has ended, the organization has a written agreement with the sponsor that the researcher or organization will be notified of the results to consider informing participants. Standard I-9: The organization has written policies and procedures to ensure that, when sharing oversight of research with another organization, the rights and welfare of research participants are protected.

Domain II: Institutional Review Board or Ethics Committee Domain II covers the IRB or EC, including its composition, review practices, documentation, and policies, such as those for handling unanticipated problems and protecting vulnerable participants. Standard II-1: The structure and composition of the IRB or EC are appropriate to the amount and nature of the research reviewed and in accordance with requirements of applicable laws, regulations, codes, and guidance. Element II.1.A. The IRB or EC membership permits appropriate representation at the meeting for the types of research under review, and this is reflected on the IRB or EC roster. The IRB or EC has one or more unaffiliated members; one or more members

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

68

5. ACCREDITATION OF HUMAN RESEARCH PROTECTION PROGRAMS

who represent the general perspective of participants; one or more members who do not have scientific expertise; one or more members who have scientific or scholarly expertise; and, when the IRB or EC regularly reviews research that involves vulnerable participants, one or more members who are knowledgeable about or experienced in working with such participants. Element II.1.B. The IRB or EC has qualified leadership (e.g., chair and vice chair) and qualified members and staff. Membership and composition of the IRB or EC are periodically reviewed and adjusted as appropriate. Element II.1.C. The organization has and follows written policies and procedures to separate competing business interests from ethics review functions. Element II.1.D. The IRB or EC has and follows written policies and procedures so that members and consultants do not participate in the review of research protocols or plans in which they have a conflict of interest, except to provide information requested by the IRB or EC. Element II.1.E. The IRB or EC has and follows written policies and procedures requiring research protocols or plans to be reviewed by individuals with appropriate scientific or scholarly expertise and other expertise or knowledge as required to review the research protocol or plan. Standard II-2: The IRB or EC evaluates each research protocol or plan to ensure the protection of participants. Element II.2.A. The IRB or EC has and follows written policies and procedures for determining when activities are exempt from applicable laws and regulations, when permitted by law or regulation and exercised by the IRB or EC. Such policies and procedures indicate that exemption determinations are not to be made by researchers or others who might have a conflict of interest regarding the studies. Element II.2.B. The IRB or EC has and follows written policies and procedures for addressing protection of participants in research that is exempt from applicable laws and regulations. These functions may be delegated to an entity other than the IRB or EC. Element II.2.C. The IRB or EC has and follows written policies and procedures for conducting meetings by the convened IRB or EC. Element II.2.D. The IRB or EC has and follows written policies and procedures to conduct reviews by the convened IRB or EC. Element II.2.D.1dInitial review. Element II.2.D.2dContinuing review.

Element II.2.D.3dReview of proposed modifications to previously approved research. Element II.2.E. The IRB or EC has and follows written policies and procedures to conduct reviews by an expedited procedure, if such procedure is used. Element II.2.E.1dInitial review. Element II.2.E.2dContinuing review. Element II.2.E.3dReview of proposed modifications to previously approved research. Element II.2.F. The IRB or EC has and follows written policies and procedures for addressing unanticipated problems involving risks to participants or others, and for reporting these actions, when appropriate. Element II.2.G. The IRB or EC has and follows written policies and procedures for suspending or terminating IRB or EC approval of research, if warranted, and for reporting these actions, when appropriate. Element II.2.H. The IRB or EC has and follows policies and procedures for managing multisite research by defining the responsibilities of participating sites that are relevant to the protection of research participants, such as reporting of unanticipated problems or interim results. Standard II-3: The IRB or EC approves each research protocol or plan according to criteria based on applicable laws, regulations, codes, and guidance. Element II.3.A. The IRB or EC has and follows written policies and procedures for identifying and analyzing risks and identifying measures to minimize such risks. The analysis of risk includes a determination that the risks to participants are reasonable in relation to the potential benefits to participants and society. Element II.3.B. The IRB or EC has and follows written policies and procedures for reviewing the plan for data and safety monitoring, when applicable, and determines that the data and safety monitoring plan provides adequate protection for participants. Element II.3.C. The IRB or EC has and follows written policies and procedures to evaluate the equitable selection of participants. Element II.3.C.1. The IRB or EC has and follows written policies and procedures to review proposed participant recruitment methods, advertising materials, and payment arrangements and determines whether such arrangements are fair, accurate, and appropriate. Element II.3.D. The IRB or EC has and follows written policies and procedures to evaluate the proposed arrangements for protecting the privacy interests of research participants, when appropriate, during their involvement in the research. Element II.3.E. The IRB or EC has and follows written policies and procedures to evaluate proposed

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

THE ACCREDITATION STANDARDS

arrangements for maintaining the confidentiality of identifiable data, when appropriate, preliminary to the research, during the research, and after the conclusion of the research. Element II.3.F. The IRB or EC has and follows written policies and procedures to evaluate the consent process and to require that the researcher appropriately document the consent process. Element II.3.G. The IRB or EC has and follows written policies and procedures for approving waivers or alterations of the consent process and waivers of consent documentation. Standard II-4: The IRB or EC provides additional protections for individuals who are vulnerable to coercion or undue influence and participate in research. Element II.4.A. The IRB or EC has and follows written policies and procedures for determining the risks to prospective participants who are vulnerable to coercion or undue influence and ensuring that additional protections are provided as required by applicable laws, regulations, codes, and guidance. Element II.4.B. The IRB or EC has and follows written policies and procedures requiring appropriate protections for prospective participants who cannot give consent or whose decision-making capacity is in question. Element II.4.C. The IRB or EC has and follows written policies and procedures for making exceptions to consent requirements for planned emergency research and reviews such exceptions according to applicable laws, regulations, codes, and guidance. Standard II-5: The IRB or EC maintains documentation of its activities. Element II.5.A. The IRB or EC maintains a complete set of materials relevant to the review of the research protocol or plan for a period of time sufficient to comply with legal and regulatory requirements, sponsor requirements, and organizational policies and procedures. Element II.5.B. The IRB or EC documents discussions and decisions on research studies and activities in accordance with legal and regulatory requirements, sponsor requirements, if any, and organizational policies and procedures.

Domain III: Researcher and Research Staff Domain III applies to the researchers and research staff, including their knowledge of and adherence to ethical standards, and government regulations, reporting requirements, the protocol, and organizational

69

policies. Domain III standards also focus on how well the researcher and research staff oversee the research and whether they are responsive to the questions and concerns of research participants. Standard III-1: In addition to following applicable laws and regulations, researchers and research staff adhere to ethical principles and standards appropriate for their discipline. In designing and conducting research studies, researchers and research staff have the protection of the rights and welfare of research participants as a primary concern. Element III.1.A. Researchers and research staff know which of the activities they conduct are overseen by the HRPP, and they seek guidance when appropriate. Element III.1.B. Researchers and research staff identify and disclose financial interests according to organizational policies and regulatory requirements and, with the organization, manage, minimize, or eliminate financial conflicts of interest. Element III.1.C. Researchers employ sound study design in accordance with the standards of their discipline. Researchers design studies in a manner that minimizes risks to participants. Element III.1.D. Researchers determine that the resources necessary to protect participants are present before conducting each research study. Element III.1.E. Researchers and research staff recruit participants in a fair and equitable manner. Element III.1.F. Researchers employ consent processes and methods of documentation appropriate to the type of research and the study population, emphasizing the importance of comprehension and voluntary participation to foster informed decision-making by participants. Element III.1.G. Researchers and research staff have a process to address participants’ concerns, complaints, or requests for information. Standard III-2: Researchers and research staff meet requirements for conducting research with participants and comply with all applicable laws, regulations, codes, and guidance; the organization’s policies and procedures for protecting research participants; and the IRB’s or EC’s determinations. Element III.2.A. Researchers and research staff are qualified by training and experience for their research roles, including knowledge of applicable laws, regulations, codes, and guidance; relevant professional standards; and the organization’s policies and procedures regarding the protection of research participants. Element III.2.B. Researchers maintain appropriate oversight of each research study as well as research

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

70

5. ACCREDITATION OF HUMAN RESEARCH PROTECTION PROGRAMS

staff and trainees, and appropriately delegate research responsibilities and functions. Element III.2.C. Researchers and research staff follow the requirements of the research protocol or plan and adhere to the policies and procedures of the organization and to the requirements or determinations of the IRB or EC. Element III.2.D. Researchers and research staff follow reporting requirements in accordance with applicable laws, regulations, codes, and guidance; the organization’s policies and procedures; and the IRB’s or EC’s requirements. To earn full accreditation, an organization must meet all applicable standards. For example, if an organization does not conduct research involving test articles and therefore is not regulated by the US Food and Drug Administration, Standard I.7 is not applicable.

STEPS TO ACCREDITATION The AAHRPP accreditation process is similar to those used by other accrediting bodies that evaluate specific competencies of education, animal care and research, and health-care organizations. The major steps include a self-assessment, an on-site evaluation, and review by the AAHRPP Council on Accreditation. The first step for those seeking accreditation is to conduct an in-depth self-assessment, or gap analysis, during which the applying organization compares its HRPP to the AAHRPP accreditation standards. The organization uses the AAHRPP Evaluation Instrument for Accreditation to perform an element-by-element assessment of HRPP policies, procedures, and practices and make revisions or improvements, as needed. This self-assessment typically is the most time-consuming phase of the process and can take several months or longer depending on the status of the HRPP and the resources available to devote to accreditation. For many organizations, however, the self-assessment also is the most valuable accreditation-related activity because it often results in the most comprehensive evaluation ever conducted of the organization’s entire HRPP. It is not unusual for organizations to identify unexpected areas of strength along with those in need of improvement. After completing the self-assessment, the organization submits what is known as the Step 1 Application. This document, which includes a program overview and the organization’s written policies and procedures, is reviewed by a trained, experienced peer from an AAHRPP-accredited organization or by a member of the AAHRPP staff. The Step 1 reviewer reads the application and works directly with the organization to make

the changes necessary to satisfy the accreditation standards. When revisions are complete, the organization submits a revised application, referred to as the Step 2 Application. Once the Step 2 Application is received, AAHRPP schedules a site visit by a team of trained peers from AAHRPP-accredited organizations. Teams are composed of two to six site visitors depending on the size and complexity of the HRPP. Teams generally include an IRB or compliance professional and a researcher. Team members are chosen, in part, for their experience with the research setting involved (i.e., university, hospital, IRB, etc.). The team reviews the application and conducts an on-site evaluation of the organization’s HRPP, assessing the program’s performance with respect to the accreditation standards as well as in practice. Following the site visit, the team provides the organization with a draft report identifying any discrepancies between the organization’s policies and practices and noting other areas of concern. The organization then has the opportunity to responddto point out any errors of fact and describe corrective actions that have been taken since the site visit. This response, the application, and draft site visit report are submitted to the Council on Accreditation. The Council meets quarterly to determine the accreditation status of applying organizations. In general, organizations are awarded one of four designations: full accreditation, qualified accreditation, accreditation pending, or accreditation withheld. Organizations that achieve accreditation have their names listed on the AAHRPP website and earn the right to display the AAHRPP seal. To maintain accreditation, organizations must be reevaluated 3 years following their initial accreditation, and every 5 years thereafter.

VALUE OF ACCREDITATION AAHRPP was founded on the belief that it would play a central role in strengthening protections for research participants and helping organizations improve regulatory compliance. Fifteen years later, AAHRPP accreditation has taken hold in the United States and made significant inroads overseas, and the results have echoed throughout the research enterprise. AAHRPP’s emphasis on a comprehensive, systematic approach to research protections has played a key role in the fundamental shift to organization-wide responsibility for research ethics and oversight. In addition, increasing acceptance of AAHRPP standards as the world’s standards is facilitating collaboration and laying

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

VALUE OF ACCREDITATION

the foundation for a global infrastructure built on a shared commitment to ethical practices. AAHRPP-accredited organizations value their accreditation and see the benefits to their HRPPs. In a 2015 survey of all AAHRPP-accredited organizations, a vast majority of respondents reported that achieving AAHRPP-accreditation was “very important” to their organization (Fig. 5.2). In addition, respondents were almost unanimous (91%) in their conviction that their overall HRPP has improved as a result of having achieved AAHRPP accreditation (Fig. 5.3). Since 2009, AAHRPP has been collecting metrics for HRPPs to help accredited and nonaccredited organizations benchmark their performance against others and target areas for improvement. Among the most telling results are those for IRB review times (Fig. 5.4). Although IRB review time has been criticized as a poor measure of quality, review times correlate with investigator and sponsor satisfaction. Among AAHRPP clients, in 2015 average review times from submission to initial review were as follows: 18 days for the convened IRB, 8 days by a reviewer using the expedited procedure, and 9 days for a determination of exempt. Other benefits of AAHRPP accreditation are less quantifiable but equally important to research organizations.

71

Accreditation status is a signal to sponsors and colleagues that an organization has a high-quality HRPP and, therefore, is a preferred partner. Pfizer5 was the first pharmaceutical company to announce that it would endeavor to use only accredited independent IRBs for review of multisite clinical trials and would choose an accredited research site over an unaccredited site, if other selection factors were equal. In addition, most industry sponsors set AAHRPP accreditation as a requirement before engaging independent IRBs to review multisite clinical trials. For the academic research community, accreditation has been helpful in facilitating collaborative relationships. Universities are confident in the standards of practice at other AAHRPP-accredited universities. Moreover, accreditation ensures that similar (or, sometimes, the same), policies and procedures are in place at each collaborating institution, that staff at collaborating institutions share similar levels of competence, and that each “speaks the same language.” In an era of increasing collaboration and reliance, including required use of single IRB review for multisite research (for examples, see NeuroNEXT6 and NIH StrokeNet7), the assurance of quality provided by AAHRPP accreditation has never been more critical.

FIGURE 5.2 Importance of achieving Association for the Accreditation of Human Research Protection Programs (AAHRPP) accreditation.

FIGURE 5.3 Results of Association for the Accreditation of Human Research Protection Programs (AAHRPP) accreditation.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

72

Calendar Days

5. ACCREDITATION OF HUMAN RESEARCH PROTECTION PROGRAMS

45 40 35 30 25 20 15 10 5 0

38

19

18

9

8

Time from Submission to Review by the Convened IRB

Time from Submission to Approval by the Convened IRB

Time from Submission to Review by the Expedited Procedure

Time from Submission to Approval by the Expedited Procedure

Time from Submission to Exempt Determination

FIGURE 5.4 Institutional review board (IRB) review times by type of review.

Today, AAHRPP is fulfilling its mission and promise as a leader in research protections and an advocate for the research volunteers, whose participation makes the entire enterprise possible. AAHRPP accreditation has become the gold standard for HRPPs around the globedand the symbol of quality and an organization’s commitment to safe, ethical research.

3. The benefits of accreditation are: a. Improvement in regulatory compliance b. Increased efficiency of the IRB c. Less unnecessary variation in the oversight system d. All of the above

References SUMMARY QUESTIONS 1. The AAHRPP accreditation standards apply to: a. Responsibilities of IRBs only b. Responsibilities of IRBs, researchers, and organizations c. Responsibilities of IRBs and researchers d. Responsibilities of researchers and organizations 2. The AAHRPP accreditation period for initial accreditation is: a. Three years b. Five years c. Two years d. Ten years

1. Institute of Medicine. Preserving public trust: accreditation and human research participant protection programs. Washington, DC: The National Academies Press; 2001. 2. National Bioethics Advisory Commission. Ethical and policy issues in research involving human participants, vol. I. Bethesda, MD: U.S. Government Printing Office; 2001. 3. AAHRPP accreditation standards. 2009. Retrieved from: https://admin. share.aahrpp.org/Website%20Documents/AAHRPP_Accreditation_ Standards.PDF. 4. Association of American Universities Task Force on Research Accountability. Report on university protections of human beings who are the subjects of research. New York, NY: AAU; 2000. 5. IRB Advisor Pfizer sets standard to require IRB accreditation. June 2009. Retrieved from:, http://www.ahcmedia.com/articles/113017-pfizersets-standard-to-require-irb-accreditation. 6. https://www.neuronext.org/about-us. 7. https://www.nihstrokenet.org.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

6 The Regulation of Drugs and Biological Products by the Food and Drug Administration Molly M. Flannery, Amy E. McKee, Diane M. Maloney, Jonathan P. Jarow U.S. Food and Drug Administration, Silver Spring, MD, United States

O U T L I N E Background

73

Mission and Terminology

74

Drug and Biological Product Life Cycle Discovery/Nonclinical Investigation Clinical Trials Responsibilities and Documentation Sponsors Investigators Clinical Protocol Institutional Review Board Food and Drug Administration Investigator Brochure

76 76 76 79 79 80 80 80 81 81

Investigational New Drug Safety Reports Marketing Approval/Licensure Pre-New Drug Application/Biologics License Application Submission Application Food and Drug Administration Review Postapproval

81 81 82 82

83

Compliance

84

Summary

84

Summary Questions

84

primarily in the Center for Drug Evaluation and Research (CDER) and the Center for Biologics Research and Evaluation (CBER).

The mission of the US Food and Drug Administration (FDA) is to protect and enhance the public health through the regulation of medical products, food, and tobacco and to spur innovation to address unmet medical and public health needs. The commissioner of the FDA is nominated by the President and confirmed by the Senate. The FDA’s organization consists of the Office of the Commissioner and four directorates overseeing the core functions of the agency: Medical Products and Tobacco, Foods and Veterinary Medicine, Global Regulatory Operations and Policy, and Operations. There are seven product review centers with oversight authority over specific types of products. In the United States, FDA-regulated products account for approximately 25% of spending by American consumers each year. This chapter provides an overview of the FDA and the regulation of human drug and biological products,

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00006-X

81

BACKGROUND The quality and safety of medical products have been major importance to the United States since the mid1880s. It was then that the US Congress passed the Drug Importation Act, which for the first time required the inspection and prevention of entry of adulterated medicines from abroad. In 1902 and 1906, two laws were passed that form the foundation of the FDA: the Biologics Control Act and the Food and Drug Act. Since that time, Congress has passed additional legislation enhancing FDA’s ability to protect the public health.

73

Copyright © 2018. Published by Elsevier Inc.

74

6. DRUG AND BIOLOGICS REGULATION

Many of the major laws that provide FDA with the authority for the regulation of drugs and biological products came about in response to significant public health problems and national drug tragedies. For example, in 1901, a diphtheria antitoxin adulterated with tetanus led to several deaths, which prompted the passage of the Biologics Control Act of 1902 (Virus, Serum, Antitoxin Act), designed to ensure the purity, potency, and safety of these and other biological products. In 1906, Upton Sinclair published The Jungle, an indictment of the meat packing industry. At the same time, Dr. Harvey Wiley, the chief chemist of the Bureau of Chemistry in the US Department of Agriculture (USDA), was pointing out that toxic adulterants could be found in foods and medicines. This led Congress to pass the Food and Drug Act, which was signed by President Theodore Roosevelt in 1906. This law prohibited interstate commerce of adulterated foods, drinks, and drugs. By 1933 the FDA, which had started as the Bureau of Chemistry under the USDA, had been established and recommended a complete revision of the Food and Drug Act of 1906. The momentum to pass a new act was accelerated by the elixir sulfanilamide tragedy. A new solvent, ethylene glycol, was used to formulate the elixir sulfanilamide that was put on the market without any testing in 1937. The new formulation led to the deaths of more than 100 people, many of whom were children. This led to the passage of the Federal Food, Drug, and Cosmetic Act (FD&C Act) in 1938. The FD&C Act extended FDA authority from food and drugs to cosmetics and devices. It also required that new drugs be shown to be safe before they could be marketed and authorized inspections of factories engaged in the manufacture of regulated products. In 1960, Dr. Francis Kelsey, an FDA medical officer, recommended that thalidomide not be approved for pregnancy-induced nausea in the United States due to insufficient safety data, despite its availability is much in Europe and Canada. It was subsequently discovered that thalidomide was responsible for severe birth defects in thousands of babies born in Europe and other countries. Dr. Kelsey became a national hero and was awarded the President’s Award for Distinguished Federal Civilian Service for her work; additionally, public support for stronger drug regulations resulted in the passage of the 1962 KefauvereHarris amendments to the FD&C Act, which strengthened the drug approval process. A key change in the statutory requirements was that drug manufacturers now were required to prove the effectiveness of a product before it could be approved for marketing. In 1971, the Public Health Service’s Bureau of Radiological Health was transferred to the FDA. Its mission was to protect the public from unnecessary radiation from electronic products in the home and the healing

arts. In the same year, the National Center for Toxicological Research was established to examine the biological effects of chemicals in the environment. The next year, the Division of Biological Standards, which was responsible for the regulation of biological products, was transferred from the National Institutes of Health (NIH) to the FDA to become the Bureau of Biologics. The FDA as we know it today was taking shape. The Prescription Drug User Fee Act (PDUFA) was passed in 1992 in response to a perceived lag in the approval of new drugs in the United States and allowed the agency to collect fees to support the process for the review of human drugs. In 1997, the Food and Drug Administration Modernization Act (FDAMA) was signed into law. This law reauthorized PDUFA and codified a number of FDA initiatives intended to speed the availability of new drugs for serious and life-threatening diseases. The Food and Drug Administration Amendments Act of 2007 (FDAAA) was signed into law. In addition to reauthorizing several critical FDA programs, FDAAA greatly increased the responsibilities of the FDA as well as provided the FDA with additional requirements, authorities, and resources relating to both pre- and postmarket drug safety. In 2010, the Biologics Price Competition and Innovation Act of 2009 was enacted, and one of its key provisions was an amendment to the Public Health Service Act to create a new regulatory pathway for biosimilar products. A few years later, the Food and Drug Administration Safety and Innovation Act of 2012 (FDASIA) was signed into law. Among other things, FDASIA provided the FDA with the authority to collect user fees for generic drugs and biosimilar products and to enhance the safety of the drug supply chain. FDASIA also gave the FDA a new tool, breakthrough therapy designation, intended to expedite the development and review of innovative new drugs that address certain unmet medical needs.

MISSION AND TERMINOLOGY The scope of the FDA’s mission to protect and enhance the public health is outlined in Table 6.1. The regulation of drug and biological products is based on science, law, and public health impact. The FDA is composed of scientists and experts of many disciplines, including physicians, biologists, chemists, pharmacologists, microbiologists, statisticians, consumer safety officers, and epidemiologists. The FDA is responsible for the review of regulatory submissions (e.g., applications for clinical research, marketing, and labeling), the development and implementation of regulatory policy, research and scientific exchange, product surveillance (e.g., adverse event reporting and product testing), compliance (e.g., inspections and enforcement actions),

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

75

MISSION AND TERMINOLOGY

TABLE 6.1 Food and Drug Administration’s Mission

TABLE 6.3

1. To promote the public health by promptly and efficiently reviewing clinical research and taking appropriate action on the marketing of regulated products in a timely manner.

Part 312

Investigational new drug application

2. With respect to such products, protect the public health by ensuring that foods are safe, wholesome, sanitary, and properly labeled; human and veterinary drugs are safe and effective; there is reasonable assurance of the safety and effectiveness of devices intended for human use; cosmetics are safe and properly labeled; and public health and safety are protected from electronic product radiation.

Part 3

Product jurisdiction

Part 50

Protection of human subjects

Part 56

Institutional review boards

Part 11

Electronic records; electronic signatures

3. Participate through appropriate processes with representatives of other countries to reduce the burden of regulation, harmonize regulatory requirements, and achieve appropriate reciprocal arrangements.

Part 58

Good laboratory practice for nonclinical laboratory studies

Part 314

New drug applications

Part 320

Bioavailability and bioequivalence requirements

Parts 600e680

Biologics

Part 54

Financial disclosure by clinical investigators

Part 25

Environmental impact considerations

Parts 201e202

Labeling and prescription drug advertising

Parts 210e211

Current good manufacturing practices

Parts 800e861

Medical devices and in vitro diagnostics

Parts 1270e1271

Human tissue intended for transplantation and human cells, tissues, and cellular- and tissuebased products

4. As determined to be appropriate by the Secretary, carry out paragraphs (1) through (3) in consultation with experts in science, medicine, and public health, and in cooperation with consumers, users, manufacturers, importers, packers, distributors, and retailers of regulated products. From the FDA modernization act of 1997 (PL105-115).

and outreach (e.g., education). As a science-based institution, the FDA strives to facilitate the development of new safe and effective medical products. The primary laws that govern the drug and biological products are shown in Table 6.2. Some important regulations for drugs, biologics, and medical devices in Title 21, Code of Federal Regulations (CFR), are shown in Table 6.3. These laws and regulations are intended to protect the public health. One of the FDA’s primary functions is to ensure compliance with these laws and regulations. The definitions of some of the terms used in this chapter’s discussion of the TABLE 6.2 Statutory Authorities

Federal food drug and cosmetic act

Drugs

Biologics

X

X

Public health service act

X

Interstate commerce

X

X

Foreign commerce

X

Generic equivalence

X

Orphan drug act

X

X

Prescription drug user fee act

X

X

Prescription drug marketing act

X

X

FDA modernization act of 1997

X

X

FDA amendments act of 2007

X

X

FDA safety and innovation act of 2012

X

X

Principal Regulations for Drug and Biological Products: Title 21, Code for Federal Regulations

FDA’s regulation of drugs and biological products are provided in Table 6.4. Another important role of the FDA is communication. The FDA strives to provide accurate information to health-care professionals and the public on product quality, effectiveness, and safety, including through our oversight of labeling, promotion/advertising, and compliance with good manufacturing practice. The FDA website (www.fda.gov) is an extremely valuable tool to access information. Among the documents available on the website are regulations and guidance documents. Guidance documents represent the agency’s current thinking, interpretation or policy, regarding a particular regulatory issue or product. These documents greatly facilitate the public’s understanding of laws, regulations, and policies applicable to FDA. In general, guidance documents are not binding and are updated as needed to provide accurate and timely information. Another important online resource is Drugs@FDA (http://www.accessdata.fda.gov/scripts/cder/drugsatfda/ index.cfm), which is a searchable catalog of FDA-approved

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

76 TABLE 6.4

6. DRUG AND BIOLOGICS REGULATION

Definitions and Terms

Law

A statute. An act of congress that outlines binding conduct or practice in the community.

Regulation

A rule issued by an agency under a law administered by the agency. A regulation interprets a law and has the force of law.

Code of Federal Regulations (CFR)

The compilation of all effective government regulations published annually by the US printing office. Food and Drug Administration’s (FDA) regulations are found in Title 21 of the CFR.

Guidance

FDA documents prepared for FDA staff, applicants/sponsors, and the public that describe the agency’s interpretation of or policy on a regulatory issue. In general, guidance documents are not legally binding.

Biologic

A virus, therapeutic serum, toxin, antitoxin, vaccine, blood, blood component or allergenic product, or analogous product, or arsphenamine or derivative of arsphenamine applicable to the prevention, treatment, or cure of a disease or condition of human beings. This includes immunoglobulins, cytokines, and a variety of other biotechnology-derived products, e.g., cell and nucleic acid products.

Drug

An article intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease in humans or other animals; an article recognized in the US Pharmacopoeia, the official homeopathic pharmacopoeia, or the official national formulary and their supplements; an article (other than food) intended to affect the structure or any function of the body of humans or other animals.

Device

An instrument, apparatus, implement, machine, contrivance, implant, in vitro reagent, which is intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treatment, or prevention of disease in humans or other animals; or is intended to affect the structure or any function of the body of humans or other animals; and does not achieve its primary intended purpose through chemical action within or on the body of humans or other animals and is not dependent on being metabolized for the achievement of its primary intended purpose.

Investigational New Drug Application (IND)

A request for authorization from the FDA to administer an investigational drug or biological product to humans.

New Drug Application (NDA)

A request for authorization from the FDA to sell and market a new pharmaceutical.

Biologics License Application (BLA)

A request for authorization from the FDA to introduce, or deliver for introduction, a biological product into interstate commerce.

drug and biological products, including approval letters, review documents, and labeling. The FDA also performs research regarding the products it regulates. Some examples of this research include research related to the establishment of standards and methods, toxicology, product safety, and basic mechanisms of actions or pathogenesis. This research advances FDA’s mission and is important for quality review of submissions, development of new policy and guidance, providing advice on product development and product safety.

DRUG AND BIOLOGICAL PRODUCT LIFE CYCLE The life cycle for new drug and biological products is divided into four stages: discovery/nonclinical investigation, clinical trials, marketing application/licensure, and postapproval.

Discovery/Nonclinical Investigation The earliest stage of product development involves the discovery and initial evaluation of an active moiety. In this period of drug development a production process sufficient to yield a consistent quality, clinical-grade material is required so that the drug product is adequately characterized. Tests and assays to characterize the product should be under development in this stage because they will be necessary to link the product to the outcome of animal or human clinical trials. At this time, the sponsor conducts animal safety studies to determine an appropriate starting dose in humans and to establish an initial toxicity profile for the product. These studies will assist in designing the first-in-human clinical trial to help ensure that the human participants are properly monitored for potential adverse events. This is the stage in which the biological rationale for the use of the product is proposed. If an animal efficacy model exists, studies in that model also should be performed to support the use of the product in humans. The FDA has developed a number of guidance documents on considerations in product development and nonclinical animal studies to help sponsors develop the necessary data to support an Investigational New Drug (IND) submission.

Clinical Trials The FD&C Act and the Public Health Service Act require that a new drug or biological product be approved before it can enter interstate commerce. Under

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

DRUG AND BIOLOGICAL PRODUCT LIFE CYCLE

its rulemaking authority, the FDA issued regulations found in 21 CFR Part 312, allowing an exception from the approval requirement for drugs and biologics for which an IND application is in effect. These regulations allow investigational products to be legally shipped in interstate commerce to conduct clinical investigations. A number of FDA guidance documents are applicable to the conduct of clinical trials. For example, FDA’s E6 Guidance for Industry on Good Clinical Practice is a harmonized guidance document developed as part of the International Council on Harmonisation (ICH) and provides a unified standard for the United States, the European Union, Japan, Canada, and Switzerland to facilitate the mutual acceptance of clinical data by the regulatory authorities in these jurisdictions. The term good clinical practice (GCP) refers to the design, conduct, recording, evaluating, monitoring, and reporting of clinical trials. The principles of GCP are provided in Table 6.5. During the clinical development of a product under an IND, additional product process development and testing/validation are performed. Also, additional nonclinical information is obtained regarding the safety and efficacy of the product. There are generally three phases of premarketing clinical research to examine the safety and efficacy of a drug or biological product in a “learn and confirm” model. Phase 1 trials include “first-in-human” studies that are normally small, dose-escalation trials that may include patients with a particular condition or normal volunteers, with the primary goal of assessing safety of the product using a particular route of administration. Phase 1 trials also examine the pharmacokinetics and metabolism of the investigational drug, which can include drugedrug interactions and food effect studies. The primary goal of these studies is to provide preliminary evidence of safety and dosing. Preliminary evidence of efficacy may be observed in Phase 1 studies but it is not the primary purpose. Phase 2 “proof of concept” trials consist of one or more moderately sized clinical trials for a particular patient population. Phase 2 trials are typically larger than Phase 1 studies and are designed to evaluate the effectiveness of a product for a particular indication in a patient population with the disease being studied. The primary purpose of Phase 2 trials is to detect efficacy and optimize dosing although safety information is continually collected and assessed. Phase 3 trials are often much larger trials that are designed to evaluate the benefits and risks of a product in a patient population with a defined clinical indication. The safety and efficacy data from these trials are generated to support marketing approval and to provide information to write the instructions for the use of the product for a particular indication. Some key issues for the design, conduct, and analysis of Phase 3 clinical

TABLE 6.5

77

Principles of Good Clinical Practice

• Clinical trials should be conducted in accordance with the ethical principles that have their origin in the declaration of helsinki, and that are consistent with good clinical practice and the applicable regulatory requirement(s). • Before a trial is initiated, foreseeable risks and inconveniences should be weighed against the anticipated benefit for the individual trial subject and society. A trial should be initiated and continued only if the anticipated benefits justify the risks. • The rights, safety, and well-being of the trial subjects are the most important considerations and should prevail over the interests of science and society. • The available nonclinical and clinical information on an investigational product should be adequate to support the proposed clinical trial. • Clinical trials should be scientifically sound and described in a clear, detailed protocol. • A trial should be conducted in compliance with the protocol that has received prior institutional review board/independent ethics committee approval/favorable opinion. • The medical care given to, and medical decisions made on behalf of, subjects should always be the responsibility of a qualified physician or, when appropriate, of a qualified dentist. • Each individual involved in conducting a trial should be qualified by education, training, and experience to perform his or her respective tasks. • Freely given informed consent should be obtained from every subject prior to clinical trial participation. • All clinical trial information should be recorded, handled, and stored in a way that allows its accurate reporting, interpretation, and verification. • The confidentiality of records that could identify subjects should be protected, respecting the privacy and confidentiality rules in accordance with the applicable regulatory requirement(s). • Investigational products should be manufactured, handled, and stored in accordance with applicable good manufacturing practice. They should be used in accordance with the approved protocol. • Systems with procedures that assure the quality of every aspect of the trial should be implemented. From ICH harmonized tripartite guideline for good clinical practice, Step 4, 1996 ICH secretariat c/o IFPMA, Geneva, Switzerland.

trials include the primary and secondary end points, trial population, randomization, stratification, blinding, sample size, participant adherence, and statistical analysis. It is important to note that the clinical development of a product may not proceed in a linear fashion from Phase 1 to Phase 3 trials, and a product may be simultaneous in all stages of clinical trial development for one or more indications. The content and format of the IND application is specified in the FDA regulations at 21 CFR 312.23. The IND application should include the following: a table of contents; an introductory statement including the

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

78

6. DRUG AND BIOLOGICS REGULATION

rationale for the drug or the research study and general investigational plan; chemistry, manufacturing, and control (CMC) information; pharmacology and toxicology information; previous human experience with the investigational product and other relevant information; protocols; and the investigator’s brochure (IB). Once the original IND is submitted, the FDA has 30 days to review and notify the submitter or sponsor whether the trial may proceed or has been placed on clinical hold. During those 30 days, the sponsor may not initiate the clinical trial. The IND review is aimed primarily at an evaluation of the safety of the product for human clinical trials. The IND is allowed to proceed if the agency has no safety concerns or if the sponsor does not hear from the FDA within 30 days. In contrast, a clinical hold notice is issued to the sponsor that the proposed clinical trial(s) may not begin (or an ongoing clinical trial is suspended) until certain stated deficiencies are resolved if there are safety concerns. Phase 1 trials may be placed on clinical hold for any of the following five reasons: 1. Human subjects are or would be exposed to an unreasonable and significant risk of illness or injury. 2. The clinical investigators are not qualified to conduct the study. 3. The IB is misleading, erroneous, or materially incomplete. 4. The IND does not contain sufficient information to assess the risks to subjects. 5. The IND is for the study of a life-threatening disease or condition that affects both genders, and men or women with reproductive potential who have the disease or condition being studied are excluded from eligibility (see Chapter 13). Once an IND is established, new studies may be initiated under that IND with submission of the protocol to the IND without a 30-day waiting period. Nevertheless, Phase 2 and 3 trials may be placed on clinical hold for any of the previously discussed reasons. They also may be placed on hold if the plan or protocol for the trial is clearly deficient in design to meet its stated objectives. If an IND is placed on clinical hold, the sponsor is notified by telephone or other means of rapid communication or in writing. This notification is followed with a letter that specifically states the deficiencies. Advice is available from FDA on appropriate corrective actions. It is then up to the sponsor to correct the deficiencies and notify the FDA of the corrections in a clinical hold response letter. Once the sponsor submits a complete response to the clinical hold, the FDA generally responds in writing within 30 calendar days to the clinical hold response letter. There is no automatic release from clinical hold. If the sponsor does not hear from

the FDA in 30 calendar days the clinical trial may not start. When FDA’s review of the clinical hold response is completed, the sponsor is notified that the trial(s) may proceed or that there are continuing deficiencies and the clinical hold is retained. A clinical investigation of a drug or biological product that is lawfully marketed in the United States may be exempt from the IND requirements set forth in the regulations. A clinical investigation is exempt if all of the following apply (21 CFR 312.2): 1. The investigation is not intended to be reported to FDA as a well-controlled study in support of a new indication for use nor intended to be used to support any other significant change in the labeling for the drug. 2. If the drug that is undergoing investigation is lawfully marketed as a prescription drug product, the investigation is not intended to support a significant change in the advertising for the product. 3. The investigation does not involve a route of administration or dosage level or use in a patient population or other factor that significantly increases the risks (or decreases the acceptability of the risks) associated with the use of the drug product. 4. The investigation is conducted in compliance with the requirements for institutional review set forth in 21 CFR Part 56 and with the requirements for informed consent set forth in 21 CFR Part 50. 5. The investigation is conducted in compliance with the requirements that a sponsor or investigator, or any person acting on behalf of a sponsor or investigator, (a) must not represent in a promotional context that an investigational product is safe or effective for the purposes for which it is under investigation or otherwise promote the drug, (b) must not commercially distribute or test market an investigational new drug, and (c) must not unduly prolong an investigation after finding that the results appear to establish sufficient data to support an application. FDA regulations require the sponsor to file an amendment to the IND if certain changes are made to the product, the nonclinical studies, or the clinical protocol. These include changes in product formulation and changes that affect the safety, scope, and scientific quality of the clinical protocol, including its data and analyses, or the addition of a new protocol. The sponsor also must file an annual report that includes, among other things, all changes in and results of the study. A sponsor may request to meet with FDA for advice on product development throughout the development life cycle. Often sponsors meet with the agency at the end of the nonclinical/discovery stage to discuss their data and their future plans prior to submitting an IND

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

DRUG AND BIOLOGICAL PRODUCT LIFE CYCLE

application. This meeting is referred to as a pre-IND meeting and may be very important in facilitating a successful IND. It is possible to request an end-of-phase-1 meeting to discuss the data obtained in a Phase 1 trial and the drug development plan, including selection of doses and patient populations in Phase 2 trials. Frequently, the sponsor meets with the FDA at the end of Phase 2 trials to discuss the outcomes of the trials as well as the design and analysis plan for the Phase 3 trials. At these end-of-phase-2 meetings, the sponsor and FDA frequently discuss the key end points, the specific patient population, and the statistical analysis plan that will provide a statistically significant and clinically meaningful result from the trial(s). Following completion of the Phase 3 or “pivotal” trials, the sponsor again may meet with the FDA to discuss a marketing application submission. These pre-BLA (Biologics License Application) or pre-NDA (New Drug Application) meetings focus on the content and format of a marketing application. Several mechanisms are available that accelerate the drug development process for life-threatening and severely debilitating illnesses, such as fast-track designation, breakthrough therapy designation, priority review (21 CFR 312 Subpart E), and accelerated approval (21 CFR 314.510 and 601.41). The purpose of these programs is to facilitate the development and expedite the review of new drug and biological products that are intended to treat serious and life-threatening conditions and that demonstrate the potential to address unmet medical needs. FDA issued a comprehensive guidance document in 2014 describing the qualifying criteria and features of each of these four programs (http:// www.fda.gov/downloads/drugs/guidancecompliance regulatoryinformation/guidances/ucm358301.pdf). The fast-track program offers actions to expedite development and review, including a rolling review of a marketing application. Breakthrough therapy designation also offers actions to expedite review, including intensive guidance from the FDA on drug development, an organizational commitment from the FDA to involve senior managers and experienced review and regulatory health project management staff in a proactive, collaborative, cross-disciplinary review, and rolling review of a marketing application. Priority review offers a 4-month shorter clock for review of an application. Finally, accelerated approval (21 CFR 314.510 and 601.41) is an FDA approval based on a surrogate end point that is reasonably likely to predict clinical benefit or based on a clinical end point other than survival or irreversible morbidity that is also likely to predict ultimate clinical benefit. Applicants of products approved under this pathway have been required to conduct Phase 4 (postmarketing) trial(s) to confirm and/or verify clinical benefit.

79

There are also a number of expanded access programs that are available for patients with a serious or immediately life-threatening disease under an IND (21 CFR 312 Subpart I). The three categories of expanded access uses are for individual patients (including emergency use), for intermediate-sized populations, and for large populations under a treatment IND or protocol. For all expanded access uses, FDA must determine that the patient has no comparable or satisfactory alternative therapy to treat the disease or condition. The FDA also must determine that the potential patient benefit justifies the potential risks and those risks are not unreasonable in the context of the disease or condition to be treated, and providing the drug for the requested use will not interfere with the initiation, conduct, or completion of clinical investigations that could support marketing approval of the expanded access use or otherwise compromise the potential development of that use (http://www.fda.gov/downloads/drugs/guidance complianceregulatoryinformation/guidances/ ucm351261.pdf).

Responsibilities and Documentation Sponsors Several groups, including the sponsors, investigators, Institutional Review Boards (IRBs), and the FDA, have responsibilities in clinical research that are described in the regulations and guidance documents. The responsibilities of the sponsor are found in FDA’s regulations at Subpart D of 21 CFR Part 312. The sponsor, generally the developer of the product, is the person or entity who submits the IND. The sponsor is responsible for selecting qualified investigators and providing them the necessary information to conduct the study properly. The sponsor also is responsible for the trial design, the trial management, data handling and record keeping, allocation of responsibilities, compensation to subjects and investigators, financing, and notification/submission to regulatory authorities (e.g., protocol submission). In addition, the sponsor is required to ensure that there is proper monitoring of the study and that it is conducted in accordance with the general investigational plan and protocols contained in the IND. The sponsor must ensure that all participating investigators and the FDA are promptly informed of significant new adverse effects or risks with respect to the product. The sponsor also is responsible for the quality assurance and quality control of the trial. Finally, the sponsor is accountable for maintaining and making available, as necessary, the information on the investigational product, including the manufacture of the product, supplying and handling the investigational product, record access, and safety information. A sponsor may transfer responsibility for

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

80

6. DRUG AND BIOLOGICS REGULATION

any or all of its obligations to a contract research organization; however, the sponsor is ultimately responsible for the quality and integrity of the trial. Investigators The sponsor is responsible for selecting qualified investigators. Investigators have multiple responsibilities, including following the protocol for the study and complying with all applicable regulations. It is their responsibility to protect the rights, safety, and welfare of subjects in their care. As part of the responsibility for protection of human subjects, an investigator must not involve a human being as a subject in research unless the investigator has obtained the legally effective informed consent of the subject or the subject’s legally authorized representative. In doing so, the investigator must assure that there is sufficient opportunity for the subject or the representative to consider whether or not to participate. The explanation of the study must be in language that the subject can be understood and presented in a manner that minimizes the possibility of coercion or undue influence. The consent form must not contain exculpatory language through which the subject or representation is made to waive or appear to waive the subject’s legal rights. The investigators must retain control of the investigational product and maintain records of the disposition of the product, records, and reports (e.g., progress and final reports, safety reports), case histories of the subjects, and termination or suspension of the trial. Investigators are required to report observed serious adverse events to the sponsor and to record nonserious adverse events and report them to the sponsor according to the timetable for reporting specified in the protocol. In addition, they are required to report to the IRB all changes in the research activity and all unanticipated problems involving risk to subjects or others. Investigators also must arrange for review of the IND protocols by the IRB and other communications with the IRB. Because of concerns of potential bias, they are required to supply sponsors with sufficient accurate financial information to allow the sponsor to report on financial interest to the FDA. If the sponsor and the investigator are the same individual, then that individual must carry out all of the responsibilities of the sponsor and investigator with appropriate safeguards or contracting arrangements to ensure the integrity of the trial and human subject safety. Clinical Protocol The clinical trial protocol and its amendments are critical elements of clinical research. The protocol should include general information, such as title, protocol number, names of sponsors, investigators, and

background information. The background information should include the name and description of the investigational drug product, nonclinical studies that impact the clinical trial, the population to be studied, known or possible risks and benefits to human subjects, and administrative information. The protocol should state the objectives and purpose of the trial, the trial design, the selection and withdrawal of subjects, the treatment of subjects, the assessment of efficacy/activity (where appropriate) and safety, and the statistical evaluation plan (where appropriate). It also should address the plan for quality control, monitoring and assurance, data handling, record keeping, and ethical considerations. A more detailed treatment of this subject may be found in Chapter 16. A special protocol assessment may be requested by an IND sponsor to evaluate the adequacy of certain proposed studies associated with drug development (see http://www.fda.gov/downloads/Drugs/Guidance ComplianceRegulatoryInformation/Guidances/UCM08 0571.pdf). Three types of protocols are eligible: animal carcinogenicity, final product stability, and Phase 3 trials whose data will form the primary basis for determination of efficacy. The submission of a clinical trial protocol should include the statistical analysis plan. The FDA has a 45-day review clock to respond to the submission with either a letter of agreement or suggested revisions to the protocol. Documented special protocol agreements are considered binding except when the sponsor does not follow the protocol, modifies the protocol without concurrence by FDA, or if there is a material change in the science. Institutional Review Board The constitution and responsibilities of the IRB are covered by the regulations in Part 56 of Title 21 of the CFR. The IRB is charged with reviewing and approving protocols that are to be carried out in the organization(s) that it serves. As described in Chapter 4, it is the IRB’s function to ensure that in each protocol the risks to human subjects are minimized and reasonable in relation to anticipated benefits, if any, to subjects, and the importance of the knowledge that may be expected to result. IRBs must assure that the selection of subjects is equitable and that informed consent is sought and appropriately documented. The regulations require that the IRB has at least five members with varying backgrounds to promote complete and adequate review of research activities at the institution(s). The IRB must have at least one member whose primary concerns are scientific, another whose primary concerns are nonscientific, and at least one member not otherwise affiliated with the institution. Chapter 4 provides a detailed explanation of the structure and function of the IRB.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

DRUG AND BIOLOGICAL PRODUCT LIFE CYCLE

Food and Drug Administration The FDA reviews all INDs and their amendments to determine whether they are in compliance with the appropriate laws and regulations. The regulations establish time frames for the performance for certain reviews and lay out the responsibilities of the FDA in communicating with the sponsors. The primary purpose of the review of the original IND submission and early amendments is to help assure that human subjects are not exposed to unreasonable risk. In the later phases of the IND process involving trials to support efficacy determinations, the FDA reviews also focus on whether the studies are constructed and carried out in a way that will yield valid data that can be considered for marketing approval. The FDA also interacts with sponsors through meetings and conference calls, starting at the pre-IND stage and continuing throughout the entire IND process, to address important product development, clinical trial design and analysis, and premarket submission issues. Investigator Brochure If the sponsor is not the investigator, there must be an IB. It is the sponsor’s responsibility to maintain and update the IB and give it to the investigators who are conducting the trial. This document generally includes information regarding the clinical and nonclinical data on the investigational product that are relevant to the use of the product in human subjects. Investigational New Drug Safety Reports Sponsors should submit IND safety reports to the FDA as described in 21 CFR 312.32 and 312.33. The reporting requirements for adverse events include expedited reports that consist of written reports and annual reports or information amendments. IND safety reports include any suspected adverse reaction that is both serious and unexpected, or any findings from other studies that suggest a significant risk in humans exposed to the drug, or tests in laboratory animals that suggest a significant risk in humans exposed to the drug, including reports of mutagenicity, teratogenicity, or carcinogenicity. A serious adverse event or serious suspected adverse reaction is one that, in the view of either the investigator or sponsor, results in any of the following outcomes: death, a life-threatening adverse event, inpatient hospitalization or prolongation of existing hospitalization, a persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions, or a congenital anomaly/birth defect. A life-threatening adverse event or life-threatening suspected adverse reaction is one that places the subject in the view of either the investigator or sponsor at immediate risk of death. The sponsor must notify the FDA and

81

all participating investigators as soon as possible, but not later than 15 calendar days, after the sponsor determines that the information qualifies for reporting in an IND safety report. The sponsor also shall notify the FDA of any unexpected fatal or life-threatening suspected adverse reaction as soon as possible, but not later than seven calendar days after the sponsor’s initial receipt of the information.

Marketing Approval/Licensure Section 351 of the Public Health Service Act requires that a biologics license be in effect for any biological product that is to be introduced into interstate commerce. The FD&C Act requires approval of a marketing application (NDA) for new drugs to be introduced into interstate commerce. The provisions of the IND regulations allow interstate transportation of drugs and biological products for clinical investigations. These investigations are intended to provide data to support a BLA or an NDA. Pre-New Drug Application/Biologics License Application Submission Although the IND phase is primarily directed at the collection of nonclinical data to support the safety of the clinical investigations, as well as clinical data, during this time much of the CMC information needed for a marketing application also is being developed. The formulation to be marketed should be identified and ideally should be used for the pivotal clinical trials. If the to-be-marketed formulation of the product differs from that used in the pivotal clinical trials, the sponsor will need to provide data to “bridge” the formulations. The product must be adequately characterized, and its stability demonstrated. Consistency of manufacture also must be proven. Although the specific approaches to the development of these data vary with the product area, there are a number of guidance documents available that provide insight into what information is important and how the information might be generated. During the pre-IND and IND stages, it is important that the potential applicant remain in contact with the FDA. It is far easier to address concerns, including both clinical trial and CMC issues, before the clinical protocol is under way. It is in the best interest of both the FDA and the sponsor to work out these details so that when the time comes for a marketing application to be submitted, there are no unexpected problems. After the sponsor compiles sufficient information, the sponsor will begin to plan the submission of the NDA or BLA. The FDA recognizes the value of pre-NDA/BLA meetings, and encourages sponsors to schedule a pre-NDA/BLA meeting well in advance of any planned

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

82

6. DRUG AND BIOLOGICS REGULATION

submission of an application. This meeting provides a forum for discussing the content, format, and timing of the proposed submission. While the sponsor is preparing to submit a BLA or NDA, the FDA is preparing to review it. A multidisciplinary review team is formed, and preliminary decisions concerning the handling of the submission are made. One of the first decisions is whether the review of the product should be handled under a standard review schedule or as a priority review. The standard and priority review schedules are based on goals agreed to and in conjunction with the PDUFA. Currently, the standard schedule for a new molecular entity requires a complete review in 10 months, whereas a priority review is to be completed in 6 months from the date of filing of the BLA or NDA. The review schedule decision is based on the use of the product (for severe or life-threatening illnesses) and whether it fills an unmet medical need. At this time, the review team also will decide which clinical study sites should be inspected and requests a bioresearch monitoring inspection. This inspection is focused on the verification of the data that are submitted to the FDA. The field investigators will help determine whether the studies were carried out according to regulations and appropriate informed consent was obtained. They also review the record keeping for compliance with the regulatory record keeping requirements and to determine whether protocols were followed. If deviations are observed, the field investigators will present the firm with a list of observations (FDA Form 483). The FDA Form 483 is considered, along with a written report called an Establishment Inspection Report, all evidence or documentation collected on-site, and any responses made by the company. The FDA considers all of this information and then determines what further action, if any, is appropriate to protect public health. The report of the bioresearch monitoring inspection is a key piece of the review of a BLA or NDA. Application The regulations prescribing the content of a BLA are found in 21 CFR 601.2 and those for the NDA in 21 CFR 314.50. The BLA/NDA must contain a signed cover sheet and the Form FDA 356h; this form provides information that enables the center to identify the type of submission, the applicant, and the reason for the submission. The bulk of the BLA/NDA submissions generally consist of nonclinical and clinical study reports that the applicant believes provide data supporting the safety and efficacy of the product. The applicant also must submit the proposed labeling for the product, which must be supported by the data. The BLA/NDA also must contain adequate CMC information to ensure that the product meets standards of

purity and potency. These data will include information on characterization, stability, the manufacturing process, and the facility in which the manufacturing is carried out. In the BLA/NDA, applicants include a statement about the nonclinical studies used to support the application being conducted in compliance with regulations on good laboratory practice for nonclinical laboratory studies (Part 58 of Title 21 CFR). If the studies were not conducted according to good laboratory practice, the applicant must explain why they were not. The applicant must certify that all clinical studies were conducted in compliance with the informed consent regulations in Part 50 of Title 21 of the CFR, and that each clinical study either was conducted in compliance with the IRB regulations in Part 56 or was not subject to those regulations. In addition, Part 54 of Title 21 of the CFR requires the submission of a financial certification or disclosure statement or both, for clinical investigators who conducted clinical studies submitted in the application. Every BLA/NDA also must include a statement regarding the effect of the product on the environment. Depending on the specific facts, the sponsor must provide either a claim for categorical exclusion or an environmental assessment. Under current regulations, most drug and biologic marketing applications are categorically excluded from the need to supply an environmental assessment; however, there are certain categories of products and processes that still require such an assessment. Sponsors should become aware of the need for an assessment during the IND process. Food and Drug Administration Review The receipt of the BLA/NDA at the FDA starts the “review clock.” The review team consists of the experts necessary to conduct a review of the submission. Generally, the team contains specialists in clinical and nonclinical data review, product area specialists, specialists in good manufacturing processes, biostatisticians, and a regulatory project manager. Reviewers in other specialty areas are added to the review team as necessary. The initial review of the BLA/NDA focuses on the suitability of the application for filing. If the application is significantly deficientdthat is, it lacks information necessary to permit a substantive reviewdthe FDA may refuse to file it. A “refuse to file” action terminates the review of that application. Although an applicant may elect to file over protest, the refuse to file action indicates a severely deficient submission that is unlikely to lead to an approval in the first review cycle. If the BLA/NDA is complete, the FDA files it and the substantive review of the application begins in earnest.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

DRUG AND BIOLOGICAL PRODUCT LIFE CYCLE

It is not uncommon for questions to arise during the review, for which the review team may send an “information request” to the applicant requesting further explanation or additional data. The responses to these are expected to be answered in a relatively rapid time frame to facilitate the review. As each discipline finishes its particular review, it prepares a review memo documenting what has been reviewed and any deficiencies that have been found. Inspections are part of the complete review of a BLA/NDA. One type of inspection is the bioresearch monitoring inspection mentioned previously. This inspection helps provide assurance that the review team can rely on the clinical data submitted to support the safety and efficacy of the product. The other inspection is a facility inspection in which product specialists and specialists in good manufacturing practice visit the manufacturing facilities. This inspection is aimed at assessing whether the product is made under appropriate conditions and the process for manufacture has been validated and is being followed. All aspects of the manufacture of the product are investigated during this inspection. The applicant is made aware of any significant observations at the end of the inspection. The inspectors complete an inspection report that becomes part of the review of the application. CBER/CDER sometimes present issues raised in the review of the application to an external advisory committee made up of experts in the disease as well as a patient and consumer representative. The use of an advisory committee allows the review team to bring specific questions or concerns to a broader forum of experts. For specific questions, FDA may include additional experts as part of the advisory committee meeting to provide advice in a particular area of concern. Not all BLAs/NDAs are presented at an advisory committee. A BLA/NDA may be presented if the product is a new molecular entity or if the review team has identified particular issues on which they need expert input. A critical part of the review process is the evaluation of the proposed labeling for the product. It is important that statements made in the labeling be supported by data. The ultimate goal of the review of the proposed labeling is to ensure that it clearly identifies the product and provides adequate information to allow the safe and appropriate use of the product. Patient labeling, when included, must be both clear and accurate, so that the patient will understand how to use the product properly. The review team will work with the applicant to obtain accurate and informative labeling. After the inspection reports are evaluated, the reviews completed, and any advisory committee

83

advice is considered, the review team makes a recommendation on the BLA/NDA and the division or office director with delegated signatory authority decides on the appropriate action. If the application is approved, the FDA issues a letter that serves as a license (BLA) or an approval (NDA), allowing the applicant to introduce the product into interstate commerce. If the review has resulted in questions or concerns, the FDA issues a “complete response letter.” This letter explains that the application cannot be approved and identifies all of the deficiencies that must be addressed to put the application in condition for approval. When the applicant responds to this letter, the review clock and the review begin again. The FDA publishes its approval letters, the approved labeling, and reviews on the website Drugs@FDA (http://www.accessdata. fda.gov/scripts/cder/drugsatfda/index.cfm). With rare exceptions, however, any complete response letters and reviews of unapproved applications remain confidential.

Postapproval Following marketing approval, the FDA is responsible for the review of changes to the NDA or BLA, including manufacturing changes, labeling changes, and new clinical indications, for the lifetime of the product. These changes must be submitted as supplements to the BLA or NDA. Supplements are reviewed and approved (or not) according to the timelines described by the PDUFA. There are two types of postmarket studies from a regulatory perspective: postmarketing commitments (PMC) and postmarketing requirements (PMR). PMRs include studies that sponsors are required to conduct under one or more statutes or regulations. These may include safety studies to assess a known serious risk or efficacy studies conducted to affirm clinical benefit for a product that was received accelerated approval based on a surrogate end point. In addition, pediatric studies can be required under the Pediatric Research Equity Act. PMCs are studies or clinical trials that a sponsor has agreed to conduct at the time of approval, but that are not required by a statute or regulation. Adverse events must be reported according to 21 CFR 600.80 for biological products and 21 CFR 314.80 for drug products. Postmarketing 15-day “alert reports” are submitted for adverse events that are both serious and unexpected as soon as possible but not later than 15 calendar days of initial receipt of the information by the applicant. These are generally reported through MedWatch for drugs and nonvaccine

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

84

6. DRUG AND BIOLOGICS REGULATION

biological products (http://www.fda.gov/safety/ medwatch/default.htm) or the Vaccine Adverse Events Reporting System for vaccines (https://vaers.hhs.gov/ index).

product is no longer safe or effective. The FDA also can pursue injunctive relief and criminal sanctions for violations when warranted.

SUMMARY

COMPLIANCE The FDA has the authority to disqualify clinical investigators from conducting clinical testing of new drugs and devices when the agency determines that the investigator has repeatedly or deliberately failed to comply with the requirements intended to protect study subjects and ensure data integrity. The FDA also can disqualify an investigator who has repeatedly or deliberately submitted false information to the agency or study sponsor in a required report. If disqualified, the investigator may no longer receive investigational products and will be ineligible to conduct any clinical investigation that supports an application for a research or marketing permit for products regulated by the FDA. The FDA reviews any marketing application that relies on data from studies performed by the disqualified investigator to determine whether the investigator submitted unreliable data that are essential to the continuation of an investigation or to the approval of a marketing application, or essential to the continued marketing of an FDA-regulated product. Depending on that determination, the FDA may decide that the investigation may not be considered in support of a research or marketing application or may withdraw approval of the product. Under its statutory debarment authority, the FDA also may ban or “debar” individuals and companies convicted of certain felonies or misdemeanors related to drug products. Once individuals have been subjected to debarment, they may no longer work for anyone with an approved or pending drug product application at the FDA, and the FDA will not accept or review NDAs submitted by debarred individuals. Even though this statutory authority was granted in the Generic Drug Enforcement Act of 1992, it applies to both innovator and generic drug manufacturers. Debarred companies may no longer submit, or assist others in submitting, NDAs. Following the approval of a product, the FDA performs biennial inspections to assess the firm’s compliance with current good manufacturing practice (21 CFR 210, 211 and 600e680). The FDA evaluates any observations listed in an FDA Form 483 and determines whether further regulatory action is needed. If the deficiencies are severe, the FDA can take appropriate regulatory actions, including steps to revoke the license (BLA) or withdraw approval (NDA) if FDA believes the

The FDA plays a vital regulatory role in the conduct of clinical drug research and drug product development. The regulations promulgated by the FDA are intended to help ensure human subject protection and data integrity. Even when a specific clinical investigation is exempt from the FDA’s IND requirements, it must be conducted in compliance with the requirements for IRB review (21 CFR part 56) and informed consent (21 CFR part 50). Thus, the principles of human subject protection must be maintained. A central role of the FDA is also to assist in the expeditious development of safe, effective, high-quality drugs for the US population. Thus, there are numerous programs in place to incentivize and expedite drug development for indications with a significant unmet medical need.

SUMMARY QUESTIONS 1. An investigator is responsible for: a. following the clinical protocol b. maintaining records of product disposition and case histories of subjects c. reporting adverse events d. protecting human subjects e. reporting financial interests f. all of the above 2. An IND may be placed on Clinical Hold if: a. Human subjects would be exposed to unreasonable risk of illness or injury b. A study of a life-threatening disease includes women with reproductive potential c. The investigator’s brochure is inadequate d. a and b e. a and c f. b and c 3. The FDA: a. will not meet with the sponsor during the IND phase b. will not inspect the clinical trial sites c. has programs to facilitate the development of new drugs and biologics d. has no time frame for review of a new IND e. publishes guidance documents that must be followed f. none of the above

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

SUMMARY QUESTIONS

4. The FDA has multiple expedited programs to accelerate development and review of new drugs and biologics. These programs include: a. Accelerated approval and priority review b. Accelerated approval and expanded access c. Breakthrough therapy and orphan product designation d. Breakthrough therapy and fast-track designation e. b and d f. a and c g. a and d

85

5. Principles of Good Clinical Practices include a. The rights, safety, and well-being of trial subjects b. Adequate nonclinical and clinical information on the investigational product c. Systems with procedures that assure the quality of every aspect of the trial d. a and b e. a and c f. a, b, and c

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

7 International Regulation of Drugs and Biological Products Theresa Mullin U.S. Food and Drug Administration, Silver Spring, MD, United States

O U T L I N E Introduction

87

Background Early Operations and Achievements of International Conference on Harmonisation Recent Evolution and Reforms Membership in the New International Council on Harmonisation Organization of the New International Council on Harmonisation Financing the New International Council on Harmonisation

88

Overview of the International Council on Harmonisation Technical Harmonization Process Nomination and Selection of Topics for Harmonization International Council on Harmonisation Five-Step Harmonization Procedure

88 91

93 94

International Council on Harmonisation Guidelines Most Relevant to Clinical Research

95

92

Future Work in Regulatory Harmonization

98

92

References

98

91

INTRODUCTION

of requirements across member states. The World Health Organization (WHO), an agency of the United Nations concerned with international public health, has established medicinal, clinical, and technical standards and promotes regulatory capacity building, training, and work sharing among regulatory authorities, including a biennial meeting of its International Conference of Drug Regulatory Authorities (ICDRA). The Pharmaceutical Inspection Convention and Pharmaceutical Inspection Cooperation Scheme (PIC/S) are two international instruments between countries and pharmaceutical inspection authorities intended to serve as a means to improve cooperation in the oversight of Good Manufacturing Practices between regulatory authorities. In addition to these global efforts, there are a number of regional harmonization initiatives. In the Asia Pacific region, the Asia-Pacific Economic Cooperation (APEC) and the Association of Southeast Asian Nations

The global drug regulatory environment is characterized by regulatory oversight conducted at the national and regional level by authorities, such as the US Food and Drug Administration (FDA) and regional health initiatives, such as the European Union (EU) European Commission (EC). The laws that govern each of these national or regional entities generally determine the specific regulatory requirements in each jurisdiction. However, recognizing the increasingly global nature of industry operations, from preclinical development through finished product manufacturing, drug regulatory authorities have undertaken a variety of efforts to harmonize regulatory requirements and cooperate in efforts to oversee drug industry operations. Within the EU, for example, the European Medicines Agency and the EC oversee a centralized process and harmonization

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00007-1

93

87

Copyright © 2018. Published by Elsevier Inc.

88

7. INTERNATIONAL REGULATION OF DRUGS

(ASEAN) have been established as venues for cooperation and regulatory harmonization. In addition, the Asian Economic Community (AEC) conducts regional harmonization of technical standards and regulatory requirements under the Pharmaceutical Product Working Group (PPWG). In South America, the Pan American Health Organization (PAHO) component of the WHO operates in collaboration with the Pan American Network for Drug Regulatory Harmonization (PANDRH). In Africa, the South African Development Community (SADC) works to strengthen regulatory capacity among its member countries, promote harmonized standards for pharmaceutical development and promote harmonized standards for treatment as well as access to essential medicines. The East African Medicines Regulatory Harmonization Program (EAC-MRH) also works to harmonize medicines regulation systems and procedures within the East African Community (EAC) in accordance with national and international policies and standards. In the Middle East, the Gulf Central Committee for Drug Registration (GCC-DR) similarly works to harmonize regional standards for drug marketing approval. Another major harmonization effort distinguished by its collaboration of technical experts drawn from both regulatory agencies and pharmaceutical industry, undertaken by drug regulators working with pharmaceutical industry stakeholders, has been the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (hereafter referred to as ICH), which was first established in 1990. ICH has been the primary engine for development of many of the harmonized technical standards that have been taken up and adopted and used as a basis for training by many of the global and regional harmonization efforts previously mentioned. This chapter will focus on the operations and contributions of ICH to support international regulation and the development of safe and effective medicines. It provides background on the establishment and earlier operations of ICH and describes more recent reforms to its structure and governance. It also provides the reader with an overview of the ICH technical harmonization process and finally provides a more detailed review of the guidelines primarily related to the planning and oversight of clinical research to inform regulatory decision-making.

BACKGROUND Early Operations and Achievements of International Conference on Harmonisation The development of new requirements by drug regulatory authorities in the major pharmaceutical markets in the United States, European Union, and

Japan in the 1970s and 1980s, related to demonstration of new drug efficacy in addition to the requirements for safety, resulted in better evidence generated by more comprehensive and often more complex drug development programs. This very positive public health impact also was accompanied by a less desirable increase in the time and cost of drug development. In addition, drug companies often found that differences in specific requirements related to drug safety, efficacy, and quality set by regulators in different regions of the world created extensive duplication of effort that further contributed to costs and delays in getting new drugs to patients. These economic and public health impacts provided key motivators to establish a process for international harmonization of regulatory standards for the marketing or registration of new drugs. The first meeting to explore the potential for international regulatory harmonization, held in Europe in 1990, was attended by representatives of regulatory authorities and the pharmaceutical industry from the United States, European Union, and Japan. This first meeting was to plan an international conference on harmonization but the meeting also discussed the wider implications and terms of reference of an ICH. At the first ICH Steering Committee (SC) meeting the Terms of Reference were agreed, and it was decided that the topics selected for harmonization would be divided into Safety (S), Quality (Q), and Efficacy (E) to reflect the three criteria, which are the basis for approving and authorizing new medicinal products. A fourth category of Multidisciplinary (M) guidelines emerged as it was determined that some needed areas for standard harmonization addressed more than one of the original three S, Q, and E categories, and a number of guidelines were identified as needed to address electronic standards for regulatory submission of new drug dossiers. The mission of the ICH established in 1990 was and remains to promote public health by making recommendations to achieve greater harmonization in the interpretation and application of technical guidelines and requirements for pharmaceutical product registration. The harmonization of these regulatory standards is considered to offer direct benefit to both regulatory authorities and to regulated industry. Major benefits commonly cited include the prevention of duplication of clinical trials in humans and more consistent protection of human subjects in clinical trials, accomplished primarily through the E guidelines. Another major benefit is considered to be the minimization of the use of animal studies without compromising drug safety and effectiveness, accomplished primarily through the S guidelines. ICH harmonization guidelines have been credited with streamlining the regulatory assessment process for new drug applications and producing

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

89

BACKGROUND

a combined impact of reducing the development times and resources needed for global drug development. Key elements contributing to the success of ICH include the involvement of both regulators and industry parties in the detailed technical harmonization work and the application of a science-based approach to harmonization through a consensus-driven process. That also is clearly outlined and closely managed by a senior governance body in ICH. The technical harmonization work is conducted by experts with comparable levels of expertise in a given topic area drawn from both regulatory agencies and the drug industry. In addition, the final steps of approval and adoption are controlled solely by the regulators, with a corresponding commitment to implement any approved guidelines within their region. When the ICH was initially established almost a quarter-century ago, it was comprised of six member parties who formed the ICH SC. This included three regulators: FDA, the EC, and the Japanese Ministry of Health, Labor and Welfare (MHLW)/Pharmaceutical and Medical Devices Agency (PMDA). The six parties also included corresponding pharmaceutical industry associations from these three regions: the Pharmaceutical Research and Manufacturers of America (PhRMA), the European Federation of Pharmaceutical Industries and Associations (EFPIA), and the Japanese Pharmaceutical Manufacturers Association (JPMA). In addition to these ICH SC members, there were several observer parties including Health Canada, the European Free Trade Association represented by Swissmedic, the WHO, and the International Federation of Pharmaceutical Manufacturers and Associations (IFPMA).1 Intensive work by these parties in the first decade or so following the establishment of ICH resulted in a number of initial seminal guidelines to harmonize requirements related to preclinical safety studies, clinical trial planning and oversight, and drug quality. Table 7.1 provides an overview or major topic areas currently addressed by ICH guidelines. One of the contributions in the early ICH work was the development of E6 Good Clinical Practices, a harmonized guideline for good clinical trial practices. The goals of ICH E6 include assurance of human subject protection and assurance of data quality. The guideline is intended to provide a standard guide so that drug developers know what they need to do both to comply with the regulations and document compliance, and the scope is limited to clinical research performed with regulatory intent. Developed with the particular aim of providing guidelines for the conduct of trials to explore the safety and effectiveness of investigational new drugs, E6 has provided critical guidance for both international regulators and clinical researchers. Investigational new drugs pose the greatest potential risks to study participants because of the limited safety and

TABLE 7.1

Sampling of Major Topic Areas Addressed by International Council on Harmonisation Guidelines

SAFETY • Carcinogenicity studies • Genotoxicity studies • Toxicokinetics and Pharmacokinetics • Toxicity testing • Reproductive toxicology

• Biotechnology products • Pharmacology studies • Immunotoxicology studies • Nonclinical evaluation for anticancer pharmaceuticals • Photosafety evaluation

EFFICACY • Clinical safety • Clinical study reports • Doseeresponse studies • Ethnic factors • Good clinical practice

• Clinical trials • Clinical evaluation by therapeutic cat • Clinical evaluation • Pharmacogenomics • Multiregional clinical trials

QUALITY • • • • •

Stability Analytical validation Impurities Pharmacopoeias Quality of biotechnology products • Specifications

• Good manufacturing practice • Pharmaceutical development • Quality risk management • Pharmaceutical quality system • Development and manufacture of drug substances

MULTIDISCIPLINARY • Medical Dictionary for Regulatory Activities terminology • Electronic standards • Nonclinical safety studies • Common Technical Document and Electronic Common Technical Document

• Data elements and standards for drug dictionaries • Gene therapy • Genotoxic impurities

effectiveness information available at the time of the study. The significant cost of obtaining study data, the prime source of evidence for assessment of drug safety and effectiveness, requires the sponsor of the research to have great confidence that the study will produce sufficient high-quality evidence that is acceptable to the regulators. E6 was developed in the mid-1990s as the international guideline to be followed when generating

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

90

7. INTERNATIONAL REGULATION OF DRUGS

clinical data intended to be submitted to regulatory authorities. Moreover, E6 aimed to provide enough specificity so as to minimize potential ambiguity and resulting inconsistency in the interpretation of the guideline across different global regions where approaches to health-care delivery and regulatory practice might be expected to vary. The specificity was thus intended to minimize the potential for E6 interpretation to be yet another source of variability across investigational sites in multiregional clinical trials. Over the past 20 years, E6 has played an essential role in enabling the continued growth and success of multiregional clinical trials of investigational new drugs,2 including critical guidance related to training, responsibilities, and expectations of investigators, sponsors, and institutional review boards. It has thereby supported the earlier submission of new drug applications (with data collected in conformance with the harmonized guidelines adopted by regulators in multiple regions), enabling earlier access to new medicines for patients who need them. E6 also has provided a valuable guide to clinical trial oversight for research conducted to inform regulatory decision-making at all stages of drug development. This includes clinical trials to investigate the safety and effectiveness of a drug for a new indication for marketing approval. In the United States, these data would be submitted to the FDA in an “efficacy supplement” to the original New Drug Application (NDA) or biologics licensing application (BLA). Other research may include postmarketing studies required of or agreed to by a sponsor that are conducted after regulatory approval of a drug for marketing. For example, FDA uses postmarketing study commitments and requirements to gather additional information about a product’s safety, efficacy, or optimal use. Another ICH harmonization accomplishment that is considered among the most significant is the development of a Common Technical Document (CTD) for regulatory submissions by industry sponsors.1,2 Prior to the development of the CTD, the regulatory submissions of marketing applications in different ICH regions each involved complex multiple submissions in differently organized and structured formats. Submissions even varied within regions from one application submission to the next and would even vary by company depending on the team who worked on assembling the application. This meant that reviewers in each region would effectively need to learn the structure of each new submitted application and hunt for the information that need to be available for a complete review. This general lack of international standards created a situation that was burdensome for industry to assemble dossiers to different standards or no existing standard at the regional level,

and equally cumbersome and difficult for regulators to access and navigate for review. The development and adoption of the harmonized CTD has revolutionized the submission of marketing applications by enabling sponsors to replace multiple divergent formats with a single technical dossier that can be submitted to all ICH regions and other regulatory authorizes who have adopted the standard, thus facilitating simultaneous submission for review and potential approval, and potentially earlier access for patients, in multiple world regions. Development of the original paper-based CTD standard also facilitated the development of a subsequent electronic standard (eCTD) that has further enabled regulators to do more automated checks on the completeness of applications and their readiness for review. The eCTD also has enabled the development of a suite of electronic review tools that have prompted better quality in submitted applications and enabled greater efficiency in regulatory review. Another area of major accomplishment for ICH concerns the extension and maintenance of the Medical Dictionary for Regulatory Activities (MedDRA).3 MedDRA is a highly specific standardized medical terminology originally begun by the UK’s Medicines and Healthcare Products Regulatory Agency (MHRA) and transferred to ICH for further development and support for broad international use. The MedDRA terminology is used for the marketing registration, documentation, and postmarketing safety monitoring of medical products. The medical products covered by the scope of MedDRA include pharmaceuticals, biologics, vaccines, and drugedevice combination products, and it is now being used by regulators, pharmaceutical companies, clinical research organizations, and health-care professionals. In addition to the original English version and Japanese translation, MedDRA has been translated and is maintained in languages including Chinese, Czech, Dutch, French, German, Hungarian, Italian, Portuguese, and Spanish. Each MedDRA term has an associated 8-digit numerical code, which remains the same irrespective of the language. While the initial focus of ICH effort was concentrated on development of harmonized standards to address the discrepancies between regulatory standards in the three ICH regions, it also became clear to the ICH parties that there was both considerable interest and value in having the ICH guidelines be considered for adoption beyond the three founding regions. This was true both because these non-ICH regions had growing pharmaceutical industry activities and the regulatory authorities wanted to gain efficiencies where possible by adopting existing ICH harmonized guidelines if they were deemed applicable and acceptable. This led to ICH creation of

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

BACKGROUND

the Global Cooperation Group (GCG)2 with the goal of promoting a better understanding of ICH and ICH guidelines. Through the GCG other non-ICH drug regulatory authorities (DRAs) and regional health initiatives (RHIs) were invited to attend the now-biannual ICH meetings to listen to ICH technical discussions at all levels from expert working group (EWG) who were actively engaged in the development or updating of ICH guidelines through SC meetings to oversee the guideline work of the experts. This forum helped to lay the groundwork for more recent reforms of the ICH governance and operations.

Recent Evolution and Reforms As noted earlier, since the establishment of ICH over 20 years ago, the pharmaceutical industry had evolved to be globally based. During this time, other non-ICH national and regional economies had grown rapidly and were emerging as global players with increasing pharmaceutical industry activities that include clinical development and manufacturing as well as marketing. Given these important changes, the ICH parties recognized the need to modernize the ICH, and identified several goals for a “reformed” ICH. These included the following: (1) establishing one major and preferred venue to focus global drug regulatory harmonization work that would be accessible to all key drug regulatory stakeholders; (2) creating a venue that would allow all these stakeholders the opportunity for input to the drug harmonization work; and (3) maintaining the efficiency and effective management of harmonization operations that had been key to ICH success and effectiveness throughout its past. ICH parties recognized the need for a single venue for global harmonization work noting that greater synergy that would be achieved by having drug regulatory authorities from emerging economies join the existing and well-established guideline process offered by ICH rather than seek to engage in bilateral national harmonization efforts or multilateral regional harmonization efforts. Considering the criticality of international harmonization efforts, there was also a strong desire to maximize the level of public health protection achieved with the limited resources typically available to drug regulators. There was a related concern that engaging in a variety of disconnected bilateral and regional harmonization efforts would ultimately lead to a fragmentary and suboptimal allocation of regulatory resources and less effective global oversight of regulated entities. The need to provide all regulatory stakeholders, particularly non-ICH drug regulators, an opportunity for greater harmonization input was based on

91

recognition that there was, at the time, no formal role for these non-ICH parties. In addition to the regular opportunity to adopt already-existing ICH guidelines, it was considered critically important to offer more formal and regular opportunities for input and engagement in the guideline development process and the new topics selected for harmonization. Finally, it was felt that new provisions to expand participation would need to be integrated with approaches to effectively manage what would become a larger and more diverse stakeholder engagement in the technical harmonization process. In addition to these overarching goals there were other objectives intended to bring ICH into the modern era of organizations involved in the public sphere. In addition to greater stakeholder inclusivity, this meant transitioning from a somewhat informal operation funded largely by industry contributions to a more formal and transparent organization and operations including establishing ICH as a legal entity, creating a more distributed and equitable approach to financing operations by member parties, and establishing more routine public sharing of information and outreach about ongoing ICH harmonization work and work products. Important milestones related to these goals and objectives have been reached over the past year. These include the establishment of the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) as a nonprofit association under Swiss law in October 2015, providing the specification of a new expanded structure for membership and governance.4

Membership in the New International Council on Harmonisation Under the new structure, ICH has five categories of membership: Founding Regulatory Members, Founding Industry Members, Standing Regulatory Members, Regulatory Members, and Industry Members. With the establishment of the new association, the parties that had been members of the earlier version of ICH became members of the new organization. Thus the Founding Regulatory Members include the FDA, EC, and MHLW/ PMDA. The Founding Industry Members include PhRMA, EFPIA, and JPMA. The Standing Regulatory Members include Health Canada and Swissmedic because these two drug regulatory authorities also served as members of the SC before the establishment of the new ICH.4 The new structure of ICH also provides for three categories of observership: Standing Observers, Observers, and Ad-Hoc Observers. The WHO and IFPMA, parties who had been observers in the earlier ICH, became

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

92

7. INTERNATIONAL REGULATION OF DRUGS

Standing Observers. Other parties who had participated in the GCG as DRAs and RHIs, or who had participated in the past as observers or interested industry parties, were invited to immediately join the new ICH as Observers and to consider becoming Members. To be eligible to become a Regulatory Member, -regulators need to have participated in at least three of the past four ICH biannual meetings, have appointed experts to at least two EWGs of ICH, and have implemented a minimum of the following three ICH guidelines: Q1: Stability Testing of New Drug Substances and Product; Q7: Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients, and E6: Good Clinical Practice Guideline. Regulatory Members have the right to attend meetings of the ICH Assembly, vote in the Assembly, elect Members to the ICH Management Committee and appoint experts to ICH Working Groups. The membership criteria for industry parties includes having the status of a global pharmaceutical industry association representing members from several countries in at least three continents, who are regulated or affected by ICH guidelines, have participated in at least three of the past four ICH biannual meetings, and have appointed experts to at least two EWGs of ICH. Industry Members also have the right to attend ICH Assembly meetings, vote on Assembly decisions (those not considered strictly the purview of regulatory Members), elect Members to the ICH Management Committee and appoint experts to ICH Working Groups.4 Eligibility criteria for Observers are rather minimal and designation of Observers is based on determination of their contribution or benefit to ICH. To be considered for formal Observer status, a party must be a government authority or an RHI representing government authorities responsible for regulating pharmaceuticals for human use, an international pharmaceutical industry organization or other international organization with a direct interest in pharmaceuticals. Observers have the right to attend Assembly meetings but without voting rights.4

membership, annual member fees and other finances, decisions related to the approval of new topics for ICH guidelines, adoption, amendment or withdrawal of topics, and the Association annual work plan and strategic plan of work. The ICH Assembly has jurisdiction over similar issues related to MedDRA Management Committee and the approval of cooperation between ICH and other organizations.4 The ICH Management Committee is comprised of representatives of the Founding Regulatory and Founding Industry Members, Standing Members, and Elected Members and is required to meet at least in conjunction with the Assembly meetings. Management Committee responsibilities are primarily administrative and financial and include, but are not limited to, planning logistics; preparing and convening Assembly meetings; preparing the ICH annual and multiyear strategic plan; exercising oversight over the EWG process and operations to ensure quality, efficiency, and timeliness of guideline development; making recommendations to the Assembly related to new topics; new membership or observership applications; changes to membership fees; and other financial matters.4 The new ICH Association actively participates in MedDRA through its Members, and the MedDRA Management Committee has the role of managing, supporting, and facilitating the maintenance, development, and dissemination of MedDRA. The MedDRA Management Committee is comprised of representatives from the Founding Regulatory, Founding Industry, and Standing Members of ICH, and representatives of MHRA, with the WHO serving as an observer. The MedDRA Management Committee is responsible for ensuring MedDRA’s integrity, viability, and sustainability as a harmonized standard. This extends to Committee oversight of contracted third parties serving as a maintenance and support services organization, oversight of the annual budget for MedDRA, and determination of subscription fees as part of that budget oversight.4

Organization of the New International Council on Harmonisation

Financing the New International Council on Harmonisation

The main deliberative bodies of the new ICH organization include the Assembly, the Management Committee, and the MedDRA Management Committee; and these bodies are provided administrative and project management support by the ICH Secretariat. The Assembly is composed of all categories of Members of the ICH association, and it is expected to hold at least one ordinary meeting each year. The scope of its governance includes, but is not limited to, decisions related to the ICH Articles of Association and Rules of Procedure, decisions about Membership status of various parties, decisions related to Management Committee

With the establishment of the new nonprofit association under Swiss law, ICH also has transitioned to a more formal approach to funding operations based on annual membership fees to cover costs associated with the ICH Secretariat and other regular operations and to finance the biannual meetings at which the Assembly and Management Committee meet and as importantly, EWGs meet face-to-face to engage in ICH technical guideline work. These in-person interactions typically span several days and really facilitate progress in technical harmonization discussions. The harmonization work of ICH is conducted through a highly

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

OVERVIEW OF THE INTERNATIONAL COUNCIL ON HARMONISATION TECHNICAL HARMONIZATION PROCESS

structured process that has proven to be quite effective and is further described in the section that follows.

OVERVIEW OF THE INTERNATIONAL COUNCIL ON HARMONISATION TECHNICAL HARMONIZATION PROCESS Nomination and Selection of Topics for Harmonization The ICH process begins with the proposal to undertake a new topic or update an existing ICH guideline related to drug Safety, Efficacy, Quality or a relevant Multidisciplinary issue. The proposal typically will include a brief statement of the identified problem caused by a lack of drug regulatory harmonization and would describe the main technical and scientific issues to be addressed in the proposed ICH technical harmonization work and the expected outcomes of that work. In view of the significant resource commitment required by all ICH parties to undertake regulatory harmonization work, new topic proposals also must provide a strong case for why a particular area of identified disharmony is so important for international harmonization. This might include evidence to suggest that the new topic proposal could potentially conserve regulatory agency or industry resources in the future or might potentially improve the timing of access of new drugs to patients. New topic proposals are also expected to address whether the identified technical issues for harmonization are feasible to address within the limits of current national laws and regulations in all of the ICH regions, whether the level of effort and length of time likely to be required of experts for the guideline work is feasible for at least the minimum set of required ICH Members, and whether the proposed topic might potentially compete for ICH resourcing within or across the topic categories (Q, S, E, M). Finally, new topic proposals are also expected to discuss the timing for when it would be anticipated that the benefits of the completed guideline would be realized and how the proposed topic relates to, and potentially complements or conflicts with, other existing guidelines.5 A topic proposal would be submitted to the ICH Management Committee and can be submitted by any ICH Member or Observer. The Management Committee will review and make an initial assessment of the mission relevance, urgency, and feasibility of all proposed topics, provide this assessment, offer recommendations, and seek endorsement and topic prioritization from the Assembly during a biannual meeting session.6 If a new topic proposal is endorsed by the Assembly, an informal Working Group will be established to develop a Concept Paper to further flesh out the

93

harmonization work that would be undertaken based on what was outlined in the original new topic proposal. ICH Members can nominate representatives to informal Working Groups, and all nominees are expected to have expertise relevant to the topic subject matter. Typically, unless otherwise specified by the Assembly, the official membership of an informal Working Group will be limited to two representatives per ICH Member, including one expert designated as the Topic Leader for that Member and the other designated as Deputy Topic Leader. The Topic Leaders and Deputy Topic Leaders are expected to participate in the informal Working Group discussions and to be the point of contact for any consultation carried out among experts between meetings. While all ICH Members may appoint expert representatives, every Founding Regulatory Member is required to nominate at least one expert to each informal Working Group. In addition, the presence of at least one expert from each Founding Regulatory Member and if nominated, one expert from each Founding Industry Member and Standing Regulatory Member nominated to the informal Working Group, is required to constitute a quorum for Group meetings. The informal Working Group functions to both develop and provide any further refinement needed to finalize the Concept Paper, and submit this to Management Committee for endorsement.5 Following the endorsement of a guideline Concept Paper, an EWG will be established. While all Founding Regulatory Members are required to appoint experts to all EWGs, the Founding Industry Members, Standing Regulatory Members, and other ICH Members are invited and encouraged to appoint technical experts to all EWGs. Unless otherwise specified by the Assembly, the Membership of an EWG will typically be limited to two representatives per ICH Member per Working Group (this limitation applies to both Regulatory Members and Industry Members) with one expert designated as Topic Leader and the other as Deputy Topic Leader. In addition, ICH Observers who would like to participate in the EWG may submit a request to appoint an expert observer to a specified Working Group.6 The EWG is responsible for developing a detailed Work Plan prior to initiation of the guideline work. The Work Plan will include anticipated milestones, a timeline for the completion of activities, a summary of any issues, and a justification for a future face-to-face meeting if that is requested by the EWG. The Work Plan for each EWG is posted on the ICH website. The work of the EWG is led by the appointed Regulatory Chair and Rapporteur. The Regulatory Members of the ICH Management Committee officially designate a Regulatory Chair from the Regulatory Members. The Rapporteur for the EWG, however, will be designated by the full Assembly and selected from among the Topic

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

94

7. INTERNATIONAL REGULATION OF DRUGS

Leaders designated by various ICH Members when the new topic was formally endorsed. In general, the Regulatory Chair and the Rapporteur will be from different regions. The role of the Rapporteur is to serve as the scientific cochair to the Regulatory Chair, and in that role he/she is expected to facilitate and manage scientific and technical activities of the EWG. This includes reconciling scientific differences of opinion to produce an ICH document with the scientific and technical content that is drafted in accordance with Assembly expectations and decisions. The Rapporteur typically works in close collaboration with the Regulatory Chair, whose role and major responsibility is to ensure that the initially proposed work plan time frames are met and work of the EWG remains within the scope of its Assembly-approved mandate.5,6

International Council on Harmonisation Five-Step Harmonization Procedure The Formal ICH Procedure is a five-step process that is used to develop the ICH harmonized guidelines for implementation within each Member’s region. Table 7.2 provides an overview of the five-step procedure, which is used for new guidelines and is initiated following the endorsement of a Concept Paper by the Assembly. In Step 1 of the Formal ICH Procedure, the EWG members work together to prepare a consensus draft of the technical document based on the objectives set out in the Concept Paper. The Rapporteur will typically prepare an initial draft of the technical document in consultation with the experts appointed to the EWG. The initial draft and successive revisions are discussed among the EWG and circulated with comments among the members of the group. Each ICH Member with experts appointed to the EWG is responsible for providing any comments within the timeframe allotted (typically

proposed by the Rapporteur and agreed to by the EWG members).5 The EWG will conduct some of this work in the week-long biannual face-to-face meetings of the ICH but typically performs much of the work in between the biannual meetings working via email and regular teleconference calls. When the EWG reaches consensus on the technical document, the consensus text approved by the ICH Members’ experts in the EWG is “signedoff” by those experts, making it the Step 1 Technical Document.5 Once the EWG signs off on the technical document, the Step 1 Technical Document with expert signatures is submitted to the ICH Assembly to request endorsement under Step 2a of the ICH process. In Step 2a, the Management Committee Regulatory Members and Industry Members will provide a recommendation to the Assembly on the decision to endorse the final Technical Document, based on the report of the EWG that there is sufficient scientific consensus on the technical issues for the Technical Document and recommendation to proceed to the next stage of regulatory consultation. The consensus text is endorsed by the Assembly as a Step 2a Final Technical Document either during a face-to-face meeting or through an electronic approval procedure that is organized by the ICH Secretariat.5 Recognizing that the ICH Regulatory Members, unlike Industry Members, are uniquely responsible for ultimate adoption, implementation, and potentially enforcement of new ICH guidelines as regulatory policy within their respective regions; Step 2b is a “Regulators only” step in which the ICH Regulatory Members will review the Step 2a Final Technical Document and take any actions, which might include revisions that they deem necessary to develop the draft “Guideline.” The consensus text of the Draft Guideline is then endorsed by the Regulatory Members of the ICH Assembly as the Step 2b Draft Guideline, and this allows the process to progress to Step 3 Regional Regulatory Consultation.5

TABLE 7.2 Overview of ICH 5-Step Procedure for Harmonized Regulatory Guidelines Before Step 1 Concept paper development

U

Work plan development

U

Consensus on draft of technical document All-party endorsement of final technical document Regulatory endorsement of draft guideline

Step 1

Step 2a

Step 2b

Step 3

Step 4

Step 5

U U U U

Regional regulatory consultation

U

Regulatory adoption

U

Regional implementation

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

INTERNATIONAL COUNCIL ON HARMONISATION GUIDELINES MOST RELEVANT TO CLINICAL RESEARCH

Step 3 of the Formal ICH Procedure begins with the public consultation process conducted by each of the ICH Regulatory Members in their respective regions. Step 3 continues with the collection and analysis of the public comments received across all regions participating in this process. After obtaining all regulatory consultation results, the EWG that organized the earlier Step 1e2 discussions for consensus building will reconvene. Although the reconvened group will include both Industry and Regulatory expert representatives, the leadership of the group may need to shift: if the Rapporteur was designated from an Industry Member until Step 2b, then a new Rapporteur will be appointed from a Regulatory Member typically from the same region as the previous industry Rapporteur. Step 3 concludes with completion and acceptance of any revisions that need to be made to the Step 2b Draft Guideline in response to public comments. The draft document generated as a result of the Step 3 phase is called Step 3 Experts Draft Guideline and is signed by the EWG experts of the ICH Regulatory Members and then submitted to the Assembly with a request for adoption.5 Adoption of the new Guideline occurs in Step 4. Adoption is based on a recommendation by the ICH Management Committee and the consensus of the ICH Assembly Regulatory Members affirming that the new Guideline is recommended for adoption by the Regulatory Members of the ICH regions.6 Following adoption the harmonized Guideline moves to Step 5, the final step of the process, and is implemented by each of the Regulatory Members in their respective regions. The harmonized Guideline is implemented according to the same national and regional procedures that apply to other regional regulatory Guidelines and requirements. In the United States, for example, ICH guidelines are treated as regulatory guidance to industry and made publicly available through a Federal Register Notice of Availability.

INTERNATIONAL COUNCIL ON HARMONISATION GUIDELINES MOST RELEVANT TO CLINICAL RESEARCH The Formal ICH Procedure has been used to develop E guidelines addressing Efficacy (E) topics and these are likely to have the greatest relevance to the planning and conduct of clinical research. Many of the currently available E guidelines were originally drafted in the first decade of ICH operations, under the earlier governance structure described in the Background section of this chapter. The E guideline numbering mainly reflects the chronological sequence of the development of these guidelines and is not intended to convey a particular priority or other dependency among the guidelines. In

95

addition, more than one E guideline will be relevant to the planning of a clinical study, particularly in the case of traditional interventional clinical trials performed to generate evidence to support drug regulatory review. Table 7.3 is provided to illustrate how multiple E guidelines may provide information and guidance relevant to a given subtopic. In this example, a set of critical to quality factors for clinical studies, based on a set of factors identified by the Clinical Trials Transformation Initiative,7 have been used to help develop the illustration. This section provides an overview of the content of this set of E guidelines. E1: The Extent of Population Exposure to Assess Clinical Safety for Drugs Intended for Long-Term Treatment of NonLife-Threatening Conditions. The goal of this guideline is to present an accepted set of principles for the safety evaluation of drugs intended for long-term treatment for non-life-threatening diseases, recognizing that safety evaluation during clinical drug development is expected to characterize and quantify the safety profile of a drug over a period of time consistent with the intended longterm use of the drug.3 E2A-E2F: Safety Data Management includes a set of six guidelines addressing different aspects of safety data management during clinical development and postapproval. The E2A guideline addresses clinical safety data management in terms of definitions and standards for expedited reporting. The E2B(R2) guideline focuses on data elements for transmission of individual case safety reports. The E2C(R2) guideline is concerned with clinical safety data management related to periodic safety update reports for marketed drugs (i.e., periodic benefiterisk evaluation reports), and E2D is focused on definitions and standards for expedited reporting of postapproval safety data. The E2E guideline addresses pharmacovigilance planning, and the E2F guideline is concerned with the development of safety update reports.3 E3: Structure and Content of Clinical Study Reports describes a single “core” clinical study report that can serve as an integrated complete report that would be acceptable to all regulatory authorities in ICH regions for any therapeutic, prophylactic, or diagnostic agent, including clinical and statistical presentations and analyses. Topics covered in this guideline on the structure and content of clinical study reports include study synopsis, ethics, the investigators and administrative structure, investigational plan, study patients, efficacy evaluation, safety evaluation, overall study conclusions, as well as other sections, including guidance throughout on the formats for presentation of the study data. The E4: DoseeResponse Information to Support Drug Registration guideline provides background on the purpose and use of doseeresponse information in clinical drug development and how this data should be

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

96 TABLE 7.3

7. INTERNATIONAL REGULATION OF DRUGS

International Council on Harmonisation Efficacy Guidelines

Critical to Quality Factors

E1

E2AeE2F

E3

E4

E5

E6

E7

E8

E9

E10

U

U

U

U

U

U

U

U

U

U

U

E11

E12

E14

E15

E16

E17

U

U

U

U

U

U

U

U

U

E18

PROTOCOL DESIGN Eligibility criteria (inclusion/ exclusion) U

Randomization Masking Types of controls

U

Data quantity

U

U

U U

U

Endpoints

U U

Procedures supporting study endpoints and data integrity

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

Investigational product handling and administration

U

FEASIBILITY U

Study and site feasibility

U U

Accrual

U

U

PATIENT SAFETY IRB consent

U

Informed consent

U

U

U

U

U

U

U

Withdrawal criteria and trial participant retention Signal detection and safety reporting

U (B)

U

Data monitoring committee stopping rules

U U U

U

U

U

U

U

STUDY CONDUCT Training (including investigator training)

U

Responsibilities among sponsor, investigator, and IRB

U

Data recording and reporting

U (B,C,F)

U

Data monitoring and management

U (A,B,D)

U U

Statistical analysis

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

U

STUDY REPORTING Dissemination of study results

U (D,F)

U

THIRD-PARTY ENGAGEMENT Delegation of sponsor responsibilities

U

Collaborations

U

IRB, institutional review boards.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

U

INTERNATIONAL COUNCIL ON HARMONISATION GUIDELINES MOST RELEVANT TO CLINICAL RESEARCH

obtained as an integral part of drug development. The guideline provides an overview of study designs for assessing doseeresponse, and advice on identification of a starting dose, titration steps, and other issues in dose-ranging or concentration response studies.3 E5(R1): Ethnic Factors in the Acceptability of Foreign Clinical Data provides a framework for evaluating the impact of ethnic factors on drug efficacy and safety and dose regimen, and guidance on regulatory and development strategies to allow evaluation of ethnic factors while minimizing duplication of studies across regions. This includes the use of bridging studies to extrapolate from the studied populations to a new global region to support acceptance of the data as a basis for drug registration in the new region.3 The E6(R1): Good Clinical Practice Consolidated Guideline is primarily focused on the assurance of human subject protection and assurance of data quality in clinical trials, including training that is needed for the investigators and others, and processes that should be followed both in study conduct and in documentation. Focused primarily on clinical research performed with regulatory intent, it provides a standard guide so that clinical researchers, including drug developers and clinical research staff, know what they need to do both to comply with the regulations and document compliance.3 E7: Studies in Support of Special Populations: Geriatrics. This guideline is primarily concerned with development of new molecular entities for the treatment of disease associated with aging, or new formulations or combinations of established drugs to treat conditions common among the elderly. The guideline addresses extension of the age range of elderly patients that would be desirable to include in studies and inclusion of sufficient numbers of elderly patients in the Phase 3 database. It also addresses the need for attention to pharmacokinetic differences between nonelderly and elderly patients, and circumstances when drugedrug interactions should be studied.3 The E8: General Considerations for Clinical Trials guideline offers a general overview of clinical trial topics including other ICH Guidelines concerning clinical trials. It contains, for example, a table classifying clinical studies according to objective and an annex crossreferencing other relevant ICH guidelines. Although the table includes examples of large simple trials, comparative effectiveness studies, and pharmacoeconomic studies, the guidance is primarily focused on studies intended to support drug regulatory submissions.3 The E9: Statistical Principles for Clinical Trials guideline, developed to harmonize the principles of statistical methodology applied to clinical trials to support marketing applications in all ICH regions, addresses key issues including considerations for overall clinical development such as trial context and scope, various

97

trial design considerations, trial conduct considerations, data analysis issues, the evaluation of safety and tolerability, and reporting. In addition, E9 is currently being revised to include considerations related to choosing appropriate estimands and defining sensitivity analyses in clinical trials.3 E10: Choice of Control Group and Related Issues in Clinical Trials. This guideline describes the general principles involved in choosing a control group for clinical trials intended to demonstrate efficacy and discusses related issues concerning trial design and conduct. Without explicitly addressing the regulatory requirements in various ICH regions, the guideline describes the purpose of a control group, different types of controls, what trials using different designs can demonstrate, and critical design and interpretation issues to be considered.3 The E11: Clinical Investigation of Medicinal Products in the Pediatric Population guideline provides an overview of critical issues in pediatric drug development. This guideline addresses topics including considerations when initiating a pediatric drug program, timing for the initiation of pediatric studies during drug development, the types of pharmacokinetic, pharmacokinetic/ pharmacodynamics, efficacy and safety studies to conduct, age categorizations for pediatric patients, and ethical considerations in pediatric clinical studies. E11 is currently being revised to update information for these topics and provide more discussion in several selected areas including formulation challenges in pediatric drug development and appropriate extrapolation of data from adult populations to pediatric populations and pediatric subgroups to other pediatric subgroups.3 The E12: Principles for Clinical Evaluation of New Antihypertensive Drugs guideline describes core principles for evaluation of hypertensive drugs accepted in all ICH regions that can be used in conjunction with any region-specific guidelines that may address other region-specific regulatory requirements. The E14: The Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs guideline provides recommendations concerning the design, conduct, analysis, and interpretation of clinical studies to assess the potential of a drug to delay cardiac repolarization. Measured in terms of the prolongation of the QT interval on the surface electrocardiogram, a delay in cardiac repolarization can indicate a potential increase in the risk of cardiac arrhythmias and is thus an important consideration in assessing drug safety.3 E15: Definitions for Genomic Biomarkers, Pharmacogenomics, Pharmacogenetics, Genomic Data and Sample Coding Categories. To facilitate the integration of the discipline of pharmacogenomics and pharmacogenetics into global drug development and regulatory review, the E15 guideline provides harmonized definitions for

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

98

7. INTERNATIONAL REGULATION OF DRUGS

key terms including genomic biomarkers, pharmacogenomics, pharmacogenetics, genomic data, and sample coding categories.3 E16: Genomic Biomarkers Related to Drug Response: Context, Structure, and Format of Qualification Submissions. Recognizing that use of biomarkers in drug development has the potential to help guide dose selection and thus enhance the benefiterisk profile of a new drug, the E16 guideline describes recommendations related to the context, structure, and format of regulatory submissions for qualification (assessment that a biomarker can be relied on) of genomic biomarkers as defined in E15.3 In addition to the aforementioned E guidelines, a Step 2 consensus draft version of E17: General on Principles for Planning and Design of Multiregional Clinical Trials (MRCTs) is also available and currently undergoing Step 3 regional regulatory consultation. The purpose of this guideline is to describe general principles for planning and design of MRCTs with the goal of increasing the acceptability of MRCTs in global regulatory submissions. This guideline addresses strategic issues as well as clinical trial design and protocol-related issues. The latter, for example, includes preconsideration of regional variability and its potential impact on efficacy and safety, subject selection, dose selection, estimation of overall sample size and allocation to regions, as well as other issues.3 A Step 2 consensus draft version of the E18: Guideline on Genomic Sampling and Management of Genomic Data guideline is also currently available. This draft guideline provides harmonized principles for genomic sampling and management of genomic data in clinical studies, to facilitate the implementation of such studies by enabling a common understanding of critical parameters for unbiased collection, storage and use of genomic samples and data. The guideline is also intended to increase awareness and provide guidance regarding subject data privacy, data protection, informed consent, and transparency, addressing the use of genomic samples and data regardless of the timing of analysis, considering both prespecified and nonprespecified use.3

FUTURE WORK IN REGULATORY HARMONIZATION It is a dynamic and exciting time for the international harmonization of drug regulatory standards. With the newly reformed ICH serving as the key central venue for regulatory harmonization work, the future direction is likely to be shaped by the needs of a diverse and growing global body of drug regulators, industry organizations, and other stakeholders, and the emerging areas of priority and consensus. One might expect a

diverse yet balanced portfolio of harmonization work that encompasses both new guideline work and major continuing renovation of earlier foundational guidelines originally developed decades ago. The new topics may address new areas of currently unmet need. The renovations and revisions may be undertaken to incorporate new regulatory science, new methodology, modern perspectives on patient engagement, or other important and recent advances in existing topic areas. This work is likely to span the set of Q, S, E, and M topics. It is also likely that future guideline work will address key issues for generic drug registration as well as innovator drug regulatory submissions. The resulting benefits of better quality development programs, more rigorous and complete regulatory submissions, and common scientific standards applied around the globe will increasingly be felt by patients around the globe. Expanding global adoption of common regulatory guidelines should improve patient access to safer, enable more efficient and less burdensome clinical trials, and expand availability of high-quality medicines on the market.

References 1. International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. The value and benefits of ICH to industry. 2000. http://www.ich.org/fileadmin/ Public_Web_Site/News_room/C_Publications/The_Value___ Benefits_of_ICH_to_Industry__January_2000.pdf. 2. International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. The value and benefits of ICH to drug regulatory authoritiesdadvancing harmonisation for better health. 2010. http://www.ich.org/fileadmin/Public_ Web_Site/News_room/C_Publications/ICH_20_anniversary_ Value_Benefits_of_ICH_for_Regulators.pdf. 3. International Council on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Please note that more information about: MedDRA; the latest versions of the Articles of Association, Rules of Procedure and SOPs of the ICH Working Groups, and the full text of all of the ICH guidelines can be found on the ICH website: www.ich.org. 4. International Council on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Articles of association. October 2015. 5. International Council on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Standard operating procedures of the ICH working groups. September 2016. 6. International Council on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Rules of procedure of the assembly. June 2016. 7. The Clinical Trials Transformation Initiative (CTTI). CTTI quality by design projectdcritical to quality (CTQ) factors principles document. 2015. https://www.ctti-clinicaltrials.org/files/principles_document _finaldraft_19may15_1.pdf.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

8 Clinical Research in International Settings: Opportunities, Challenges, and Recommendations Christopher O. Olopade, Michelle Tagle, Olufunmilayo I. Olopade The University of Chicago, Chicago, IL, United States

O U T L I N E Introduction

Develop and Enhance Local Institutional Review Board Capacity Develop Office for Sponsored Research/Office of Clinical Research Prepare Data Safety and Monitoring Plan for Adverse Events Provide Ancillary Care Use Technology for Effective Communication Have Long-Term Plans Integrate With Existing Infrastructure

99

Challenges Inadequate Human Resources Deficient Research Infrastructures Subpar Health-Care Systems Information Gaps Political Instability, Civil Disorders, and Natural Disasters Economic and Seasonal Migration Physical Barriers Study Participant Characteristics Ethical Issues

100 100 100 101 101

Recommendations Understand the Local Setting Train, Mentor, and Closely Supervise

103 103 104

101 102 102 102 102

INTRODUCTION

105 105 105 105 105 105

Conclusion

106

Summary Questions

106

References

107

systems, and widespread infections such as HIV, H1N1, Ebola, and more recently, Zika. In response, initiatives such as the United Nations Millennium Development Goals and Sustainable Development Goals have dramatically increased awareness of the state of global health and detailed a path toward health equity. In light of the global burden of disease and its disproportionate impact on resource-limited countries, there is tremendous opportunity to conduct standardsetting clinical research that could inform policy at international sites.

We live in a world of increasing interconnectedness and interdependence where people, goods, beliefs, ideas, and values are continually transcending national boundaries. While globalization may pose threats that demand our concerted efforts, it also promises opportunities for unprecedented growth and change. In the public health sphere, transnational approaches are being warranted by global concerns, including widening health disparities, poor health-care delivery

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00008-3

104

99

Copyright © 2018. Published by Elsevier Inc.

100

8. CLINICAL RESEARCH IN INTERNATIONAL SETTINGS

In taking a historical perspective on global health, we have learned that the health issues of low- and middleincome countries (LMICs) extend beyond infectious diseases following appreciable successes in global economic development, near eradication of polio, and improved control of HIV and malaria.10,68 The field of global health, which in the past had been almost synonymous with infectious disease, has since evolved to include chronic noncommunicable diseases (NCDs). Despite increases in life expectancy and improvements in living standards in LMICs, health disparities continue to widen and the prevalence of NCDs, such as hypertension, diabetes, and cancer, continues to rise.39 Since 2000, the number of NCD deaths has grown in every region of the world, and currently, NCDs are responsible for more deaths than all other causes added together. The World Health Organization (WHO) predicts that from the 38 million NCD deaths in 2012, the number will continue to grow to 52 million by 2030.75 To tackle the enormous burden of communicable and NCDs, there is a need for substantial global health research investment to address health problems affecting 90% of the world population.41 We must develop evidence-based interventions, including better characterization of disease patterns and their associated socioepidemiological factors. More efforts are needed to understand the pathogenesis of disease and to discover novel therapies targeted at neglected diseases.35 There is also an urgent need to improve health-care access, including access to current therapies against communicable diseases (e.g., HIV, tuberculosis) and emerging infectious diseases that pose major threats to lives in both resource-limited and resource-rich countries.30 The discovery and accessibility of these interventions will depend on the conduct of clinical research around the world, especially in LMICs where research activity is low and where evidence-based solutions may produce the largest impact on high early mortality rates.32 Conducting clinical research at international sites, especially in LMICs, are fraught with numerous challenges and therefore requires innovative approaches that place a high premium on understanding the local context and a fervent commitment to follow research regulations and ethical guidelines. In the proceeding sections, we briefly discuss some of the most common challenges experienced by the global research enterprise and offer recommendations that strengthen the prospects of novel solutions to our global health problems.

CHALLENGES Inadequate Human Resources One of the most significant challenges to conducting research in LMIC centers around human resources is

not only personnel are inadequately trained, but there is also a lack of a critical mass of investigators. The few who are highly competent are consequently in high demand and often recruited to work in highincome countries (HICs) where career advancement opportunities are greater.13 Among various factors that have contributed to the systematic decline of higher education and the clinical research workforce shortage in most resource-poor countries, some major ones include inadequate leadership and investment in academic infrastructures, a general lack of awareness of scientific benefits by administrators and the public, poor faculty development, and the “brain drain” of academic staff. This human capital flight has affected African countries, leaving an acute shortage of qualified academic staff within the higher education system to train the next generation of academic leaders. Every year since the 1990s, Africa has lost 20,000 professionals to the Westda phenomenon that the United Nations has identified to be one of the greatest challenges to the continent’s development.27 The outflow of health workers also has significantly impacted countries in the South East Asia Region, including Bangladesh, Bhutan, Myanmar, and Timor-Leste, leaving them below the global benchmark for health workforce population ratio.65 Further exacerbating the issues surrounding inadequate personnel, HIC-sponsored studies often lead to power imbalances where much of the research is primarily funded, conducted, and published by HIC investigators, and thereby leaving LMIC colleagues with little recognition of their work or benefits from the publications.13 Underscoring the discrepancy in human resources between countries, the World Development Indicators 2016da publication by World Bank Groupdreported that the per capita number of physicians (per 1000) was 4.9 in both Belgium and Spain whereas it was 0.20 in Kenya and 0.10 in Mauritania. This correlates with the health expenditures per capita of $4813, $2644, $70, and $47, respectively; and this huge disparity in health-care spending is, in turn, reflected in data on life expectancy at birthd81, 83, 61, and 63 years, respectively.73 Also, whereas spending has increased in HICs, it remains low in LMICs, especially in sub-Saharan Africa (SSA). This limited health-related spending is further worsened by corruption and poor management, which diverts even more funds away from the health sector.61

Deficient Research Infrastructures While scientific research has been considered a pillar for world development, there lies a disparity in the global distribution of the resources critical to building and sustaining scientific research capacity. Prime

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

CHALLENGES

examples of this phenomenon can be found in constrained cancer research programs in Africa2 and significant disparities in genomics research capacity in SSA, with South Africa exhibiting the greatest genomics research output due to greater investments into their genomics sector.1 In a majority of LMICs, economic constraints render research a luxury.24 Laboratories are underequipped due to insufficient funds, and the institutions that have purchasing power soon learn of the challenges associated with the installation, servicing, and maintenance of equipment.51 Supplies ranging from costly equipment replacement parts to inexpensive paper towels are a challenge to procure and often require steep overseas shipping fees. Damaged or malfunctioning equipment litter these resource-poor facilities due to the unavailability of maintenance services, thereby rendering the equipment useless and resulting in wasted time and funds.

Subpar Health-Care Systems Compared with LMICs, health-care delivery to the general populace is seamless in most HICs, but this has been possible due to learned experiences, capital investments, and opportunities to refine the processes over time. For example, the National Health Service (NHS) in the United Kingdom and the Servizio Sanitario Nazionale in Italy provide universal health coverage through public hospitals that is funded through tax policies. The systems in the Netherlands and Switzerland are based on compulsory insurance with in-built risk equalization to prevent setting premiums based on health status. Similarly, health care in France is offered by both private and public hospitals with a social security system that refunds most of the costs. The foundation of these systems is based on the revenue brought in through higher taxes for high-income earners, making quality health care physically and financially accessible to virtually the entire population. In contrast, most LMICs have meagerly funded and poorly structured health-care systems.39,40 Different levels of governments often fund the majority of health-care services with private and charitable organizations playing varying roles. Accessibility can be even further impaired for women and children as a result of religious, cultural, and economic factors.34 In addition, the coordination of service delivery through primary or secondary tiers is often poor, and the referral system is not well developed, as most people do not have specific primary care providers.57 These factors together impact the physical and financial accessibility to basic health care by the population. Moreover, these fragmented health systems result in duplicated efforts.7,18 Foreign donors who often run

101

nonoverlapping, disease-specific, vertical programs without sharing resources and experiences fund a majority of the functional programs. Both human and financial resources are wasted due to the lack of consolidation of clinical services and studies, and instead, separate but identical programs are being implemented.7,18,53

Information Gaps Within the scientific community, peer-reviewed literature is the official avenue through which communication takes place, and yet, researchers in LMICs have long experienced problems in accessing peer-reviewed content due to the high costs of journal subscriptions. While major biomedical publishers have invested in initiatives that offer such resources at little or no cost, other barriers remain, including limited computer access, scant Internet connections, unreliable electricity, and limited knowledge on how to use online resources.8 Furthermore, basic epidemiological data are not available to researchers, making it difficult to know, let alone prioritize, health needs in LMICs.55,77 Birth, death, and disease-specific registries are not maintained, as government policies regarding these data, even when they do exist, are not enforced due to inadequate facilities and trained staff. Lack of such vital data limits robust clinical research and critically impairs the capacity to project the kinds of interventions needed at specified locations.21,66

Political Instability, Civil Disorders, and Natural Disasters Many LMICs live in a constant state of flux, experiencing continuous change to strata and forms of leadership, which in turn affects field research.56 At the local level, changes in culture-based governance or political structure have a significant impact on the conduct of clinical research beyond that of granting permission and providing support. As most of these communities are tight-knit, the opinions of both spiritual and temporal authorities markedly influence the general population. Changes in such leadership or political mechanisms will invariably affect studies, especially those requiring long-term follow-up. Frequent changes in governance also impact morale in terms of basic job security of local research staff. Incessant turnover in management and its priorities can impair study continuity and undermine staff job satisfaction. Many of these societies too are ravaged by internal conflict from civil uprisings to full-blown warfare. From Kashmir and Afghanistan in Asia, to the Democratic Republic of Congo in Africa, to Colombia in South America and Chechnya in Europe, people live in

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

102

8. CLINICAL RESEARCH IN INTERNATIONAL SETTINGS

varying degrees of danger. Such settings create numerous obstacles to producing quality research, as the safety of both study participants and research staff may be compromised. There are concomitant changes in the needs and priorities in such communities, as the health needs will be largely first aid for physical injuries and the prevention of highly communicable diseases such as cholera and typhoid fever.63 In cases of natural disasters, such as the Indian Ocean tsunami in 2004 and Haiti’s earthquake in 2010, most of the surviving populations are displaced either temporarily or permanently.69 With both conflicts and natural disasters come the potential of loss to follow-up (due to death, displacement, etc.) and the likely disruption of transportation and other infrastructure components (e.g., hospitals) that are essential for the conduct of research. Also, unique ethical issues emerge when conducting research in countries marked by conflict and displacement, such as Syria and Turkey. This includes the potential harm imposed on participants by asking questions that may reactivate traumainduced distress. In light of the ever-shifting nature of refugee areas and associated safety issues, it is critical to design studies that are relatively flexible in their approach.17

Economic and Seasonal Migration In many of the rapidly developing economies, such as China, there is an unprecedented mass migration from rural to urban areas for economic reasons.46 This migration negatively impacts studies that require long-term follow-up and also creates confounding factors that complicate data analysis. Also, in agricultural-based economies, people move from one region to another whenever poor climate conditions affect land cultivation.62 In such cases, adults migrate temporarily in search of menial jobs. Similar to this group are pastoral nomads in SSA, Mongolia, and other developing countries. There are limited demographic and medical data on such populations, and cultural barriers further complicate the research process, as the willingness of such populations to talk to health-care workers is limited.62

Physical Barriers The physical distance between international project sites also can pose a challenge, as it often precludes the opportunity for consistent, regular dialog among scientists. Modern technology (e.g., phone, email, video conference) is only partially successful in creating a platform for conceptual exploration of the study, and such modes of communication may not always be as effective as in-person meetings. Travel expenses, including time

lost, also increase overall study costs. The differences in time zones among study sites may also hamper regular communication, leading to fragmented discussions via email.

Study Participant Characteristics A majority of potential research participants in resource-poor countries share unique characteristics that impact their health-seeking behavior and participation in research. Such factors include gender, level of education, income, and religious and traditional beliefs.50 Even the overall perception of disease itself determines the health-seeking behavior of individuals.14 Such beliefs on the causes, treatments, and prognosis of illness vary greatly across international borders and even within communities and families. Education, or lack thereof, significantly affects one’s occupation, level of income, quality of housing and access to medical care. It also impacts one’s health status and desire to seek treatment, and most importantly, determines one’s capacity to be a part of the decisionmaking process involved in health-care plans.3 Many Western medical terms have no equivalent translation, making understanding of the pathology and prognosis of the disease difficult to interpret or explain.54 In countries with prevalent interracial tensions, trust in both the health-care system and the race of the health-care provider can determine the choice of treatment (orthodox vs. alternative) and health-care facility.59

Ethical Issues Ethics in the conduct of clinical research in LMICs often generates controversy due to the peculiar socioeconomic factors found in such regions. The conceptual framework for clinical study design fits most perfectly into the sociodemographic characteristics of the predominant population in HICs. However, several factors (e.g., income, cultural beliefs, level of education) may preclude the direct transfer of clinical research frameworks from minorities in HICs to populations in LMICs. Several guidelines have been developed by a number of national and international organizations to guide the conduct of clinical research involving human subjects (Council for International Organizations of Medical Sciences16,48,49,59,67,74,76); Although covered in other chapters of this book, we will briefly highlight some salient aspects that are relevant to LMICs. Established guidelines, such as the Declaration of Helsinki developed by the World Medical Association, may be interpreted differently across international settings. This then leads to varying interpretations of concepts such as “informed consent,” “nonbeneficial”

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

103

RECOMMENDATIONS

studies, “vulnerable” populations, “equipoise” and when the use of controls are appropriate45 as well as of ethical issues such as injustice, coercion, and exploitation.29 For example, the capacity of participants to give “informed consent” is questionable because most modern terms have no exact translations in local languages. Furthermore, in the context of clinical research in lowresource settings, the absence of strong ethical and regulatory oversight leads to meager efforts to ensure participant safety or data integrity.31 For example, in India, it had been observed that ethical guidelines were violated, data were falsified, and impoverished illiterate citizens were exploited. There were increasing reports of deaths of clinical trial participants, and thereby resulting in the need for more stringent mechanisms in clinical trial regulation.11 Furthermore, since many in LMICs would not otherwise have access to treatment, provision of study drugs may raise ethical issues of vulnerability, inducement, and coercion, especially when coupled with payments to encourage participation.5 The use of treatments that are less effective than the “best proven intervention” is often times debated in the context of research in LMICs.45 Making drugs “reasonably available” to the community after the conclusion of a study also can add to a study’s projects costs, and some have argued that this may discourage the conduct of clinical trials.5,6,60,76 In cases of potentially fatal diseases such as HIV/AIDS, the use of placebo control has also been actively challenged on both ethical and moral grounds.72 Many believe that once an efficacious treatment has been identified, there is no justification for the use of placebo controls, and therefore, subsequent interventional studies should be done in comparison with the established standard regardless of local accessibility to the trial drug. Further complicating the research scene are the ethical requirements that may vary among institutional review boards (IRBs) within and between different countries,44 leading to debates if requirements should more closely resemble those found in HICs. Moreover, in some situations, nationalism and self-determination will influence acceptance of foreign research proposals by local IRBs, transfer of samples, and determination of intellectual property rights. Also, while many international guidelines stipulate the universality of standard of care, there are legitimate questions about its practicality, especially in resourcepoor settings.60,70,76 For example, in the case of prevention of mother-to-child HIV transmission, there are arguments if the accepted standard of care should resemble those offered in HICs or whether the local reality should be put in perspective.5,6,60,70 Often at the center of the standard of care controversy is the Declaration of Helsinki because its language implies use of a

universal standard. The Declaration, which has been amended nine times with the most recent revision in 2013, now mandates the testing of new interventions against the “best proven intervention” with only two exceptions: (1) when there is no existing “proven intervention” or (2) when “sound methodological reasons” justify deviating from the “best proven intervention”.42 A more recent event, specifically the 2014 West Africa Ebola outbreak, also has shed light on ethical issues surrounding the conduct of clinical research during an epidemic. While some argue that “unproven” interventions are acceptable and perhaps even compulsory under such extreme circumstances,71 others raise questions about the capacity of individuals inflicted with a fatal disease to make informed decisions about the use of unapproved drugs.23,25 Additionally, the outbreak had implications on clinical trial policies, as trials often take years to be approved by regulators and to be conducted according to the gold standards (i.e., randomized-controlled trial). Outbreaks are therefore often over before trials can begin. In the case of Ebola, a collaboration supported by WHO allowed researchers to bypass the “red tape,” leading to trial designs that helped provide useful data on how to control the outbreak.9 As seen, the conduct of clinical research studies in resource-poor countries is bound to raise several ethical and moral issues, especially where the current guidelines are not fully explicit in respect to prevailing socioeconomic factors in these regions or to extreme situations, such as an outbreak or epidemic.

RECOMMENDATIONS Understand the Local Setting The challenges faced by investigators when conducting clinical research in international settings often vary between and within regions. In most cases, lessons learned from one region cannot always be applied to another region without significant modification. Thus, investigators need to educate themselves on regionspecific information, such as cultural beliefs and values, geography and weather, and economic, political, and social climates and infrastructures. Such an understanding can inform different aspects of the research (e.g., study design, study population, setting, data collection methods, time frame) and ultimately help ensure a successful study. For example, population responses may vary by cultural context. In places like Costa Rica where health workers are revered, people are willing to participate in studies when approached. In contrast, rural residents in Kenya, Nigeria, and other parts of SSA will not

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

104

8. CLINICAL RESEARCH IN INTERNATIONAL SETTINGS

respond favorably to visitors (i.e., researchers) and are biased against public servants. Because a change in attitude may only happen when clinical studies are congruent with communities’ perceived needs,15 it is important for researchers to engage with communities in a way that allows for needs to be identified, concerns and ideas to be explored, and societal benefits to be highlighted.12,37 Additionally, ethical issues should be addressed comprehensively while being sensitive to local needs and cultures.38,43 There are other issues that should also be carefully considered before embarking on studies in developing countries. Of primary importance are the measures taken to protect the privacy of research participants, especially in studies involving stigmatized diseases (e.g., HIV/AIDS). Behaviors associated with these diseases, such as homosexuality, can carry risk of physical attack and in some cases, death. Also, repeated home visits by research staff may bring to the attention of the whole community the health problems of study subjects, again leading to stigmatization.28 Moreover, it is advisable to ensure that the consent process is easily understood, that consent documents are written at a sixth-grade reading level for increased understanding,20 and consent documents should be as informational as possible, going beyond directly translating them into the local language. Tests should be conducted to assess the level of understanding of the consent at the commencement of the study and also regularly throughout the study duration.20 Where surveys are involved, quality control methods should include using tape recorders, repeating questions, repeating visits, monitoring by supervisory staff, and rapidly reviewing completed questionnaires to identify errors and inconsistencies in subject responses. To avoid future conflicts, ownership of data and authorships also should be discussed as early as possible, especially with local collaborators. The adoption of a community-based or communityengaged participatory research model is a more sustainable solution to challenges posed by traditional research methods where investigators retain all the authority and seek interaction with communities only when recruiting subjects and collecting data. This entails negotiations at all stages of the research, where both parties highlight concerns, discuss issues, and collectively reach consensus. This has the potential to benefit both the researchers and community because it facilitates easier enrollment of participants and data collection as the community see themselves as equal partners in the process.58 In fact, documenting the terms of the research process in the form of a memorandum of understanding or similar document may be helpful. Further, this model may involve open discussion of research results with local stakeholders, which can lead to improvements in

the public health and well-being of the community. It is important to note, however, that the dissemination of results entails risks and challenges and must therefore be anticipated and properly managed as well.

Train, Mentor, and Closely Supervise Initial training geared toward researchers in developing countries needs to cover the basic principles of clinical research (e.g., study design and methods, analysis, ethical oversight) and may be better implemented at centers in LMICs rather than those in HICs. Efforts also should be made to develop independent local investigators and scientists through mentorship on grant writing and protocol development.33,47 At the initial stage, close supervision of the local study staff is required to ensure the ethical and judicious conduct of the study. Since in most instances many would lack the adequate clinical research skills and experience, technologies such as video conferencing and VOIP (voice over internet protocol) can be utilized to provide frequent feedback on the progress of the study and also troubleshoot emerging issues. Mentors should not expect work habits similar to their own, as planning, workloads, and sensitivity to deadlines and reporting vary with different cultures. Also, several approaches can be employed to address the “brain drain” of health care and research workers.64 These include providing incentives, strong mentorship, and grant, protocol and article writing support. At the federal level, there are a number of opportunities to engage in research capacity building efforts. The Medical Education Partnership Initiative (MEPI), for example, is a $130-million, 5-year award from the US NIH to 13 medical schools in Africa. Its aims are to bolster in-country medical education infrastructures and to increase retention of medical school faculty and clinical professors.52 In addition to MEPI and other similar initiatives, the NIH also offers research training and career development opportunities at the individual level, granting investigators a chance to enhance their knowledge and skill set according to their research background and interests.

Develop and Enhance Local Institutional Review Board Capacity The rapid increase in the number of international clinical trials based in developing countries, where research regulation is relatively weak, calls for the capacity building of local IRBs. Well-constituted and operational IRBs can accelerate research productivity at academic centers while ensuring human research subjects protection, as evidenced by an established IRB at the University of Ibadan in Nigeria that saw a 150% increase in number

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

RECOMMENDATIONS

of reviewed proposals and a 62% decrease in time to approval.19 However, despite the recognized importance of well-constituted ethics boards in regulating research, funding support to meet the mandate can often lag behind in many developing countries and must therefore also be addressed as a part of enhancing local IRB capacity.

Develop Office for Sponsored Research/Office of Clinical Research As the establishment of functional IRBs grows, there is a need for creating an office for sponsored research that can help investigators propose and manage grantfunded research while also serving as the key contact for external research sponsors. This office would be responsible for offering preaward and postaward services, including project proposal reviews, award negotiations, progress reports, amendments, and closeouts. Offices of clinical research ideally would ensure the proper conduct and management of clinical research that complies with relevant policies and regulations at the local, national, and international levels. They would be responsible for training all faculty, staff, and students involved in research on current guidelines and changes in regulatory issues. Activities should include establishing institutional research policies; providing resources and tools on the proper conduct of clinical research and clinical trials; and ensuring that all investigators, especially principal investigators, undergo required research training. Additional areas where the research enterprise may be enhanced include the handling of yearly declarations of conflict of interest, the development of management plans for situations where conflicts of interest exists, and the regulatory support for intellectual property rights.

Prepare Data Safety and Monitoring Plan for Adverse Events There is a need to put in place contingency plans to address adverse events before they occur. A standard operating procedure to deal with such eventualities should be devised, and training should be offered on initiation of the research study. Another essential aspect of preparing for clinical research that is often lacking is the development of a data safety and monitoring plan (DSMP) (see Chapter 10). Ideally, this should be developed prior to initiating any clinical trial so as to define activities that need monitoring, such as obtaining informed consent, ensuring high-quality data collection, and processing and reviewing plans for adverse events handling, protocol deviation, and violations. In clinical trials where death or morbidity are potential end points, the DSMP should involve creating a safety

105

monitoring committee to make timely decisions about the need to stop a study prematurely or to modify informed consent forms if necessary to deal with emerging risks.

Provide Ancillary Care A commentary to the Council for International Organizations of Medical Sciences’ Guideline 21 advises for the provision of health care to study participants beyond what is necessary for clinical research.16 Also, it is important to note that the Declaration of Helsinki now mandates both compensation and treatment for any research-related injury.45 Such will improve the commitment of research participants and the community to the project while providing benefits to the overall health status.

Use Technology for Effective Communication Effective communication among research sites is crucial and can considerably impact the success of a study. Field updates on the project status should be frequent so that potential problems are identified and quickly resolved. The adoption of the Internet as a medium for real-time communication is important and provides significant support for projects. Other technologies, such as solar energy, also have the potential to improve the execution and efficiency of clinical research projects through the provision of stable power supplies, and thereby lowering long-term costs.4,22,26 It is essential to invest in developing a computer and data management system with backup and maintenance plans due to frequent power fluctuations and surges, especially for studies conducted in settings where energy poverty is a problem.36 As such, it would be beneficial to leverage technological advancements to approach challenges posed by conducting research in low-resource settings.

Have Long-Term Plans Most studies in developing countries are often short in duration, goals, and scope. Lack of continuity usually leads to wasted resources and may negatively impact the community’s willingness to take part in future studies. Even for studies that are initially designed to be short in duration (e.g., interventional study), efforts should be made to maintain some form of contact with the local staff and community to help ensure continuity.

Integrate With Existing Infrastructure Despite the fact that in many LMICs’ health infrastructures are in deplorable conditions with inefficient

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

106

8. CLINICAL RESEARCH IN INTERNATIONAL SETTINGS

health systems, it is advised that efforts be made to integrate new studies into the system that is locally available. Such integration will prevent research workers from completely abandoning their primary duty of offering care to the general population. Also with this approach, investigators will be able to identify and build upon the strengths of local infrastructures, and thereby help strengthen the local health system and ensure sustainable clinical research practices.

CONCLUSION Improvements in global access and quality of health care are dependent on the discoveries of socioepidemiological risk factors for diseases, innovative evidencebased therapies, and their equitable distribution. Unfortunately, LMICs are critically deficient in their capacity to independently execute such efforts. They are confronted by innumerable challenges, including inadequate human resources, subpar research infrastructures and health systems, and economic and psychosocial factors of potential participants. These challenges are not insurmountable and should be no excuse for investigators and pharmaceutical companies to withhold conducting studies in the very populations that are most likely to benefit from patient-oriented studies. It is essential to address such challenges to help ensure the successful conduct of clinical research in resource-limited settings. Among a myriad of recommendations, a primary consideration is the setting in which the study will be conducted. Understanding the local setting (e.g., cultural beliefs and values; economic, social, and political climate and infrastructures) will help inform different aspects of the research, including the study design and effective participant recruitment and retention methods. It is also necessary to develop partnerships with communities as a means to engage leaders and residents in shared efforts to implement the study and disseminate findings, and thereby helping to create mutual trust, respect, and a sense of coownership of research projects. Furthermore, human resource development should entail comprehensive training programs with continual guidance and mentorship; and technological advancements (e.g., Web-based databases and communication platforms) should be utilized to enhance communication among research teams. Efforts also should be made to prevent fragmented or duplicated efforts by adapting to and leveraging the strengths of existing infrastructures. Also, the sharing of best practices between research teams working in similar settings may further improve efficiency and reveal initially unanticipated adverse events that need to be tackled preemptively.

In an increasingly interdependent world, global health issues are warranting transnational approaches that entail the conduct of clinical research in every region of the world, especially in resource-limited countries. These endeavors are rife with challenges, and it is essential to anticipate and address them appropriately to ensure that studies are adhering to research regulations and ethical guidelines. By engaging in such collective efforts, the international research enterprise can work toward implementing evidence-based interventions that alleviate global health disparities and enhance the quality of life for all nations.

SUMMARY QUESTIONS 1. Which of the following are recognized challenges to conducting quality clinical research in low- and middle-income country settings? a. Lack of resources and infrastructure for research b. Poor data collection methods c. Questionable ethical standards and poor subject protection d. All of the above 2. Use of placebos in clinical trials is justifiable under which of the following conditions? a. When there are approved and effective treatments for the condition b. If there is no disagreement about whether standard treatment is better than placebo c. When the additional risk posed by the use of placebo is minor and withholding the current standard therapy would not lead to serious or permanent harm d. If the study is being conducted in an international setting where standard therapy is unavailable 3. Which of the following is true regarding informed consent for studies in international settings? a. Societal benefit trumps individual risks in clinical research b. Risk to benefit ratio should always be in favor of the community c. The epidemic potential of a disease may shift risk benefit ratio in favor of the community d. Cultural values and norms can influence the informed consent process 4. Which of the following is/are important for the successful conduct of clinical trials in international settings? a. Knowledge of cultural beliefs and seasonal patterns b. Planning for alternative power supply c. Developing effective communication and data storage systems

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

d. Community engagement and partnership e. All of the above 5. Which of the following is/are the most important benefit(s) of international clinical trials? a. Contribution to improvement in global diplomacy b. Reduction of global health disparities c. Opportunity to study new drugs in a population with less likelihood of drugedrug interactions d. None of the above e. b and c

References 1. Adedokun BO, Olopade CO, Olopade OI. Building local capacity for genomics research in Africa: recommendations from analysis of publications in Sub-Saharan Africa from 2004 to 2013. Glob Health Action 2016;9:31026. http://dx.doi.org/10.3402/gha.v9.31026. 2. Adewole I, Martin DN, Williams MJ, Adebamowo C, Bhatia K, Berling C, et al. Building capacity for sustainable research programmes for cancer in Africa. Nat Rev Clin Oncol 2014;11(5): 251e9. http://dx.doi.org/10.1038/nrclinonc.2014.37. 3. Agency for Healthcare Research and Quality (AHRQ), & The Office of Behavioral and Social Sciences Research, NIH. Population health: behavioral and social science insights. 2015. Retrieved from: http:// www.ahrq.gov/professionals/education/curriculum-tools/ population-health/index.html. 4. Aviles W, Ortega O, Kuan G, Coloma J, Harris E. Integration of information technologies in clinical studies in Nicaragua. PLoS Med 2007; 4(10):1578e83. http://dx.doi.org/10.1371/journal.pmed.0040291. 5. Barry M. Ethical considerations of human investigation in developing countries: the AIDS dilemma. N Engl J Med 1988;319(16): 1083e6. http://dx.doi.org/10.1056/NEJM198810203191609. 6. Benatar SR, Singer PA. A new look at international research ethics. BMJ 2000;321(7264):824e6. 7. Biesma RG, Brugha R, Harmer A, Walsh A, Spicer N, Walt G. The effects of global health initiatives on country health systems: a review of the evidence from HIV/AIDS control. Health Policy Plan 2009;24(4):239e52. http://dx.doi.org/10.1093/heapol/czp025. 8. Burton A. Sharing science: enabling global access to the scientific literature. Environ Health Perspect 2011;119(12):A520e3. http:// dx.doi.org/10.1289/ehp.119-a520. 9. Butler D, Callaway E, Check Hayden E. How Ebola-vaccine success could reshape clinical-trial policy. Nature 2015;524(7563):13e4. http://dx.doi.org/10.1038/524013a. 10. Ceesay SJ, Casals-Pascual C, Erskine J, Anya SE, Duah NO, Fulford AJ, Conway DJ. Changes in malaria indices between 1999 and 2007 in the Gambia: a retrospective analysis. Lancet 2008;372(9649):1545e54. http://dx.doi.org/10.1016/S01406736(08)61654-2. 11. Chawan VS, Gawand KV, Phatak AM. Impact of new regulations on clinical trials in India. Int J Clin Trials 2015;2(3):3. 12. Chitambo BR, Smith JE, Ehlers VJ. Strategies for community participation in developing countries. Curationis 2002;25(3):76e83. 13. Chu KM, Jayaraman S, Kyamanywa P, Ntakiyiruta G. Building research capacity in Africa: equity and global health collaborations. PLoS Med 2014;11(3):e1001612. http://dx.doi.org/ 10.1371/journal.pmed.1001612. 14. Conrad P, Barker KK. The social construction of illness: key insights and policy implications. J Health Soc Behav 2010;51(Suppl.): S67e79. http://dx.doi.org/10.1177/0022146510383495. 15. Cornwall A, Jewkes R. What is participatory research? Soc Sci Med 1995;41(12):1667e76.

107

16. Council for International Organizations of Medical Sciences [CIOMS]. International ethical guidelines for biomedical research involving human subjects. 2002. Retrieved from Geneva, Switzerland: http://www. cioms.ch/publications/guidelines/guidelines_nov_2002_blurb.htm. 17. El-Khani A, Ulph F, Redmond AD, Calam R. Ethical issues in research into conflict and displacement. Lancet 2013;382(9894): 764e5. http://dx.doi.org/10.1016/S0140-6736(13)61824-3. 18. England R. The dangers of disease specific programmes for developing countries. BMJ 2007;335. 19. Falusi AG, Olopade OI, Olopade CO. Establishment of a standing ethics/institutional review board in a nigerian university: a blueprint for developing countries. J Empir Res Hum Res Ethics 2007; 2(1):21e30. http://dx.doi.org/10.1525/jer.2007.2.1.21. 20. Flory J, Emanuel E. Interventions to improve research participants’ understanding in informed consent for research: a systematic review. JAMA 2004;292(13):1593e601. http://dx.doi.org/10.1001/ jama.292.13.1593. 21. Gonzalez-Pier E, Gutierrez-Delgado C, Stevens G, BarrazaLlorens M, Porras-Condey R, Carvalho N, Salomon JA. Priority setting for health interventions in Mexico’s system of social protection in health. Lancet 2006;368(9547):1608e18. http://dx.doi.org/ 10.1016/S0140-6736(06)69567-6. 22. Gotch F, Gilmour J. Science, medicine and research in the developing world: a perspective. Nat Immunol 2007;8(12):1273e6. http://dx.doi.org/10.1038/ni1531. 23. Hantel A, Olopade CO. Drug and vaccine access in the Ebola epidemic: advising caution in compassionate use. Ann Intern Med 2015;162(2):141e2. http://dx.doi.org/10.7326/M14-2002. 24. Harris E. Building scientific capacity in developing countries. EMBO Rep 2004;5(1):7e11. http://dx.doi.org/10.1038/sj.embor.7400058. 25. Hayden EC, Reardon S. Should experimental drugs be used in the Ebola outbreak? Nature 2014. http://dx.doi.org/ 10.1038/nature.2014.15698. http://www.nature.com/news/ should-experimental-drugs-be-used-in-the-ebola-outbreak-1. 15698. 26. Heuck CC, Deom A. Health care in the developing world: need for appropriate laboratory technology. Clin Chem 1991;37(4):490e6. 27. Hofman K, Kramer B. Human resources for research: building bridges through the diaspora. Glob Health Action 2015;8:29559. http://dx.doi.org/10.3402/gha.v8.29559. 28. Hotez P, Ottesen E, Fenwick A, Molyneux D. The neglected tropical diseases: the ancient afflictions of stigma and poverty and the prospects for their control and elimination. Adv Exp Med Biol 2006;582: 23e33. http://dx.doi.org/10.1007/0-387-33026-7_3. 29. Johnatty RN. Clinical trials in developing countries: discussions at the ‘9th international symposium on long term clinical trials’, London, UK, 19e20 June 2000. Curr Control Trials Cardiovasc Med 2000; 1(1):55e8. http://dx.doi.org/10.1186/cvm-1-1-055. 30. Jones KE, Patel NG, Levy MA, Storeygard A, Balk D, Gittleman JL, Daszak P. Global trends in emerging infectious diseases. Nature 2008;451(7181):990e3. http://www.nature.com/nature/journal/ v451/n7181/suppinfo/nature06536_S1.html. 31. Kochhar S. Challenges and impact of conducting vaccine trials in Asia and Africa: new technologies in emerging markets, October 16the18th 2012; world vaccine congress, Lyon. Hum Vaccin Immunother 2013;9(4):924e7. http://dx.doi.org/10.4161/hv.23405. 32. Lang T, Siribaddana S. Clinical trials have gone global: is this a good thing? PLoS Med 2012;9(6):e1001228. http://dx.doi.org/ 10.1371/journal.pmed.1001228. 33. Lansang MA, Dennis R. Building capacity in health research in the developing world. Bull World Health Organ 2004;82(10):764e70. 34. Lawn JE, Lee AC, Kinney M, Sibley L, Carlo WA, Paul VK, et al. Two million intrapartum-related stillbirths and neonatal deaths: where, why, and what can be done? Int J Gynaecol Obstet 2009;107(Suppl. 1). http://dx.doi.org/10.1016/j.ijgo.2009.07.016. S5eS18, S19.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

108

8. CLINICAL RESEARCH IN INTERNATIONAL SETTINGS

35. Laxminarayan R, Mills AJ, Breman JG, Measham AR, Alleyne G, Claeson M, Jamison DT. Advancement of global health: key messages from the disease control priorities project. Lancet 2006; 367(9517):1193e208. http://dx.doi.org/10.1016/S0140-6736(06) 68440-7. 36. Lee PD. The role of appropriate medical technology procurement and user maintenance instructions in developing countries. J Clin Eng 1995;20(5):407e13. 37. Leung MW, Yen IH, Minkler M. Community based participatory research: a promising approach for increasing epidemiology’s relevance in the 21st century. Int J Epidemiol 2004;33(3):499e506. http://dx.doi.org/10.1093/ije/dyh010. 38. London L. Ethical oversight of public health research: can rules and IRBs make a difference in developing countries? Am J Public Health 2002;92(7):1079e84. 39. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJ. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet 2006;367(9524): 1747e57. http://dx.doi.org/10.1016/S0140-6736(06)68770-9. 40. Lu C, Schneider MT, Gubbins P, Leach-Kemon K, Jamison D, Murray CJ. Public financing of health in developing countries: a cross-national systematic analysis. Lancet 2010;375(9723): 1375e87. http://dx.doi.org/10.1016/S0140-6736(10)60233-4. 41. Luchetti M. Global health and the 10/90 gap. Br J Med Pract 2014; 7(4). 42. Marouf FE, Esplin BS. Setting a minimum standard of care in clinical trials: human rights and bioethics as complementary frameworks. Health Hum Rights 2015;17(1):E31e42. 43. McIntosh S, Sierra E, Dozier A, Diaz S, Quinones Z, Primack A, Chadwick G, Ossip-Klein DJ. Ethical review issues in collaborative research between us and low-middle income country partners: a case example. Bioethics 2008;22(8):414e22. http://dx.doi.org/ 10.1111/j.1467-8519.2008.00662.x. 44. McWilliams R, Hoover-Fong J, Hamosh A, Beck S, Beaty T, Cutting G. Problematic variation in local institutional review of a multicenter genetic epidemiology study. JAMA 2003;290(3): 360e6. http://dx.doi.org/10.1001/jama.290.3.360. 45. Millum J, Wendler D, Emanuel EJ. The 50th anniversary of the declaration of Helsinki: progress but many remaining challenges. JAMA 2013;310(20):2143e4. http://dx.doi.org/ 10.1001/jama.2013.281632. 46. Murphy R. How migrant labor is changing rural China. Cambridge, UK: Cambridge University Press; 2002. 47. Narasimhan V, Brown H, Pablos-Mendez A, Adams O, Dussault G, Elzinga G, et al. Responding to the global human resources crisis. Lancet 2004;363(9419):1469e72. http://dx.doi.org/10.1016/S01406736(04)16108-4. 48. National Commission for the Protection of Human Subjects of Biomedical, Behavioral Research. The Belmont report: ethical principles and guidelines for the protection of human subjects of research. 1979. Retrieved from: http://www.hhs.gov/ohrp/humansubjects/ guidance/belmont.html. 49. National Institutes of Health. Guidelines for the conduct of research involving humann subjects at the National Institutes of Health. 2001. Retrieved from: http://ohsr.od.nih.gov/guidelines/ GrayBooklet82404.pdf. 50. O’Donnell O. Access to health care in developing countries: breaking down demand side barriers. Cad Saude Publica 2007; 23(12):2820e34. 51. Oman CB, Gamaniel KS, Addy ME. Analytical chemistry and developing nations. properly functioning scientific equipment in developing countries. Anal Chem 2006;78(15):5273e6. 52. Omaswa FG. The contribution of the medical education partnership initiative to Africa’s renewal. Acad Med 2014;89(Suppl. 8): S16e8. http://dx.doi.org/10.1097/ACM.0000000000000341.

53. Pfeiffer J. International NGOs and primary health care in Mozambique: the need for a new model of collaboration. Soc Sci Med 2003;56(4):725e38. 54. Putsch III RW, Joyce M. Dealing with patients from other cultures. In: Walker HK, Hall WD, Hurst JW, editors. Clinical methods: the history, physical, and laboratory examinations. 3rd ed. 1990. Boston. 55. Ramanakumar AV. Need for epidemiological evidence from the developing world to know the cancer-related risk factors. J Cancer Res Ther 2007;3(1):29e33. 56. Rifkin SB. Lessons from community participation in health programmes: a review of the post Alma-Ata experience. Int Health 2009;1(1):31e6. http://dx.doi.org/10.1016/j.inhe.2009.02.001. 57. Rohde J, Cousens S, Chopra M, Tangcharoensathien V, Black R, Bhutta ZA, Lawn JE. 30 years after Alma-Ata: has primary health care worked in countries? Lancet 2008;372(9642):950e61. http:// dx.doi.org/10.1016/S0140-6736(08)61405-1. 58. Ross LF, Loup A, Nelson RM, Botkin JR, Kost R, Smith Jr GR, Gehlert S. Human subjects protections in community-engaged research: a research ethics framework. J Empir Res Hum Res Ethics 2010;5(1):5e17. http://dx.doi.org/10.1525/jer.2010.5.1.5. 59. Saha S, Komaromy M, Koepsell TD, Bindman AB. Patientphysician racial concordance and the perceived quality and use of health care. Arch Intern Med 1999;159(9):997e1004. 60. Shapiro HT, Meslin EM. Ethical issues in the design and conduct of clinical trials in developing countries. N Engl J Med 2001;345(2): 139e42. http://dx.doi.org/10.1056/NEJM200107123450212. 61. Sharma SP. Politics and corruption mar health care in Nepal. Lancet 2010;375(9731):2063e4. 62. Sheik-Mohamed A, Velema JP. Where health care has no access: the nomadic populations of sub-Saharan Africa. Trop Med Int Health 1999;4(10):695e707. 63. Sidel VW, Levy BS. The health impact of war. Int J Inj Contr Saf Promot 2008;15(4):189e95. http://dx.doi.org/10.1080/17457300802404935. 64. Stilwell B, Diallo K, Zurn P, Vujicic M, Adams O, Dal Poz M. Migration of health-care workers from developing countries: strategic approaches to its management. Bull World Health Organ 2004; 82(8):595e600. S0042-96862004000800009. 65. Tangcharoensathien V, Travis P. Accelerate implementation of the WHO global code of practice on international recruitment of health personnel: experiences from the South East Asia region: comment on “relevance and effectiveness of the WHO global code practice on the international recruitment of health personnel e ethical and systems perspectives”. Int J Health Policy Manag 2016;5(1): 43e6. http://dx.doi.org/10.15171/ijhpm.2015.161. 66. Timaeus I, Harpham T, Price M, Gilson L. Health surveys in developing countries: the objectives and design of an international programme. Soc Sci Med 1988;27(4):359e68. 67. U.S. Department of Health, Human Services. Nuremberg code: directives for human experimentation. 2013. Retrieved from: http://ori. hhs.gov/chapter-3-The-Protection-of-Human-Subjectsnuremberg-code-directives-human-experimentation. 68. United Nations Programme on HIV/AIDS [UNAIDS]. AIDS epidemic update 2009. 2009. Retrieved from: http://www.unaids. org/en/dataanalysis/epidemiology/2009aidsepidemicupdate/. 69. Waring SC, Brown BJ. The threat of communicable diseases following natural disasters: a public health response. Disaster Manag Response 2005;3(2):41e7. http://dx.doi.org/10.1016/ j.dmr.2005.02.003. 70. Wendler D, Emanuel EJ, Lie RK. The standard of care debate: can research in developing countries be both ethical and responsive to those countries’ health needs? Am J Public Health 2004;94(6): 923e8. 71. White BD, Gelinas LC, Shelton WN. In particular circumstances attempting unproven interventions is permissible and even

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

obligatory. Am J Bioeth 2015;15(4):53e5. http://dx.doi.org/ 10.1080/15265161.2015.1009566. 72. Wilfert CM, Ammann A, Bayer R, Curran JW, del Rio C, Faden RR, Sessions K. Science, ethics, and the future of research into maternal infant transmission of HIV-1. Perinatal HIV intervention research in developing countries workshop participants. Lancet 1999; 353(9155):832e5. 73. World Bank Group. World development indicators 2016. 2016. 74. World Health Organization. Handbook for good clinical research practice (GCP): guidance for implementation. 2002. Retrieved from: http://apps.who.int/prequal/info_general/documents/gcp/ gcp1.pdf.

109

75. World Health Organization. Global status report on noncommunicable diseases 2014. 2014. Retrieved from: http://apps.who.int/iris/ bitstream/10665/148114/1/9789241564854_eng.pdf. 76. World Medical Association [WMA]. World Medical Association Declaration of Helsinki: Ethical principles for medical research involving human subjects. Retrieved from: http://www.wma.net/en/ 30publications/10policies/b3/. 77. Yach D, Hawkes C, Gould CL, Hofman KJ. The global burden of chronic diseases: overcoming impediments to prevention and control. JAMA 2004;291(21):2616e22. http://dx.doi.org/10.1001/ jama.291.21.2616.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

9 The Role and Importance of Clinical Trial Registries and Results Databases Deborah A. Zarin, Rebecca J. Williams, Tony Tse, Nicholas C. Ide National Institutes of Health, Bethesda, MD, United States

O U T L I N E Introduction

111

Background Definitions Rationale for Clinical Trial Registration and Results Reporting History of ClinicalTrials.gov

112 112

118 118 118 119

Using ClinicalTrials.gov Data Intended Audience Search Tips for ClinicalTrials.gov Points to Consider When Using ClinicalTrials.gov to Study the Overall Clinical Research Enterprise

120 120 121

Looking Forward

121

Conclusion

123

Summary/Discussion Questions

123

References

123

112 113

Current Policies 115 Policies Affecting Clinical Trials in the United States 115 International Landscape 115 Registering Clinical Trials at ClinicalTrials.gov Data Standards and the Minimal Data Set Points to Consider Interventional Versus Observational Studies What Is a Single Clinical Trial? Importance of the Protocol Keeping Information Up-to-Date

116 116 117 117 117 117 117

Reporting Results to ClinicalTrials.gov Data Standards and the Minimal Data Set

118 118

INTRODUCTION The clinical research enterprise generates scientific data through the conduct of experiments in human volunteers. As described in other chapters, a key objective of the clinical research enterprise is obtaining generalizable knowledge to advance the medical evidence base and to inform clinical decision-making. Public access to information about individual research studies and their results is necessary to achieve this objective,

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00009-5

Points to Consider Data Preparation Review Criteria Relation of Results Reporting to Publication Key Scientific Principles and Best Practices for Reporting Issues in Reporting Outcome Measures Issues Related to Analysis Population

119 119 120

121

ethically, legally, and scientifically. Several high-profile cases, which are symptomatic of a deeper, more pervasive problem in the traditional model of clinical research information dissemination, indicate that lack of systematic access to information about ongoing and completed clinical studies can lead to a skewed view of available evidence regarding the safety or effectiveness of a medical intervention for a particular use (Table 9.1).1 This chapter focuses on trial registries and results databases that are designed to make summary clinical

111

Copyright © 2018. Published by Elsevier Inc.

112 TABLE 9.1

9. TRIAL REGISTRIES AND RESULTS DATABASES

Selected Examples of Distortion of the Evidence Base Caused by Incomplete Disclosure of Clinical Trial Information56e61

Issue

Description

Examples

Selective publication of studies

Publication limited to studies with favorable results

Antidepressants,56 Paxil (paroxetine) studies in children14,57

Selective reporting of outcomes

Publication limited to the most favorable prespecified outcomes; other less-favorable prespecified outcomes not acknowledged or reported

Cyclooxygenase-2 (COX-2) inhibitors58,59

Modification of prespecified outcome measures

Publication of outcome measures that differ from those prespecified in the protocol

Vytorin (ezetimibe þ simvastatin),60 Neurontin (gabapentin)61

research information publicly accessible and available. Although such databases serve multiple goals and audiences, one key goal is to mitigate the effects of bias from incomplete disclosure of clinical trials and their results by promoting full disclosure throughout the trial life cycle. The chapter also reviews recent trends and upcoming issues in promoting increased transparency to support public health and the scientific process. ClinicalTrials.gov (https://clinicaltrials.gov/), established and maintained by the National Library of Medicine (NLM) at the National Institutes of Health (NIH), is used throughout this chapter as a case study. It is the world’s largest publicly available clinical trials registry and results database.

BACKGROUND Definitions Clinical trial registration refers to the process of submitting and updating summary information about a clinical study protocol to a structured Web-based registry that is accessible to the public, such as ClinicalTrials.gov. A study record typically contains Stage of Study

summary information about the study, such as recruitment status, eligibility criteria, and contact information. Results reporting refers to the process of submitting and updating summary information about the results of a clinical study to a structured, publicly accessible, Web-based results database. The two processes parallel the life cycle of clinical trials (Fig. 9.1).

Rationale for Clinical Trial Registration and Results Reporting Different stakeholder groups have proposed a variety of reasons for registering clinical trials, and these reasons have been expanded over time to address new challenges (Table 9.2). Simes2 is widely recognized as the first to call for the comprehensive, prospective registration of clinical trials to address concerns that favorable results are published more often than unfavorable ones (i.e., publication bias). Simes observed that “a traditional review of the published literature could result in overly optimistic conclusions concerning new therapies. (p. 1538).” With access to a prospective trial registry that includes summary information about all clinical trials, it would be possible to identify all relevant

Steps in Clinical Trials Disclosure

IRB Review and Approval of Protocol Before

Study Initiation

During

After

Study Conduct & Protocol Amendments

Study Completion & Data Analysis

1. Initial Registration 2. Updates to the Registry (as necessary) • Recruitment Status • Enrollment • Start and Completion Dates • Key Protocol Changes 3. Initial Results Reporting 4. Updates to the Results Database and/or Registry (as necessary)

FIGURE 9.1 Steps in registration and results reporting parallel the research life cycle.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

BACKGROUND

113

TABLE 9.2 Ethical and Scientific Rationale for Clinical Trial Registration and Results Reporting Category

Reasons

Human subject protections

∙ Allow potential participants to find studies ∙ Assist ethical review boards and others to determine appropriateness of studies being reviewed (e.g., harms, benefits, redundancy) ∙ Promote fulfillment of ethical responsibility to human volunteersdresearch contributes to medical knowledge

Research integrity

∙ Facilitates tracking of protocol changes ∙ Enhances transparency of research enterprise

Evidence-based medicine

∙ Facilitates tracking of studies and outcome measures ∙ Allows more complete identification of relevant studies

Allocation of resources

∙ Promotes more efficient allocation of resources (e.g., investigators, institutional review boards, volunteers)

trials that were conducted (or are in progress) and to detect what proportions of initiated trials have provided published results. Others have cited registration and results reporting as important tools for helping to fulfill ethical principles underlying research in humans. For instance, medical research involving risk to humans generally is conducted on the basis of the promise that it will contribute to generalizable knowledge.3 When studies and their results are not publicly disclosed, the biomedical knowledge base cannot be advanced and promise to participants remains unfulfilled. Public disclosure through trial registration and results reporting promotes this ethical requirement.4,5 For instance, the Declaration of HelsinkidEthical Principles for Medical Research Involving Human Subjects (2013) requires prospective registration for “every research study involving human subjects” (Article 35), as well as dissemination of “negative and inconclusive as well as positive results” to the public (Article 36).6

History of ClinicalTrials.gov In 1988, the Health Omnibus Programs Extension (HOPE) Act7 required the NIH and the US Food and Drug Administration (FDA) with the Centers for Disease Control and Prevention (CDC) to provide public information about publicly and privately funded clinical trials of investigational drugs for human immunodeficiency virus (HIV)-related diseases. The AIDSTRIALS database, which contained summary protocol information for most NIH-funded and many industrysponsored acquired immunodeficiency syndrome (AIDS) trials, became available on the Web starting in 1996.8 The HOPE Act “envisioned medical providers and scientific researchers as the intended audience” (p. 129). In 1995, the National Cancer Institute (NCI) launched the CancerNet Website, which provided online access to a cancer trials registry as part of its

mandate under the National Cancer Act of 1971 to disseminate health information.9 A federally supported registry, ClinicalTrials.gov, was launched in February 2000 to implement Section 113 of the FDA Modernization Act (FDAMA) (Fig. 9.2).10 This law called on the NIH to “establish, maintain, and operate a data bank of information on clinical trials for drugs for serious or life-threatening diseases and conditions.” Specifically, it required this registry of clinical trials to include, among other things, “a description of the purpose of each experimental drug, .eligibility criteria for participation in the clinical trials, a description of the location of trial sites, and a point of contact for those wanting to enroll in the trial.” Because FDAMA was intended to provide information “[to] individuals with serious or life-threatening diseases and conditions, to other members of the public, to health care providers, and to researchers,” it required that information in the registry “shall be in a form that can be readily understood by members of the public.” Over time, other trial registration policies were implemented to enhance access to information about clinical trials. In 2004, the International Committee of Medical Journal Editors (ICMJE),11,12 an influential group of medical journal editors, issued a policy requiring registration of clinical trials before the enrollment of the first participant to document publicly the prespecified study design and to track any changes. The State of Maine13 enacted legislation and promulgated rules in 2005 (subsequently repealed in mid-2011) to ensure that information about clinical trials of FDA-approved drugs and biologics marketed in Maine was available to its citizens. State attorneys general also have incorporated mandatory clinical trial results reporting requirements into legal settlements with drug companies (e.g., the GlaxoSmithKline Clinical Study Registry [http://www.gskclinicalstudyregister.com/] following a lawsuit initiated by the New York State Office of the Attorney General in 2004 alleging that the company concealed results

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

114

9. TRIAL REGISTRIES AND RESULTS DATABASES

FIGURE 9.2 Cumulative number of registered studies from February 2000 to September 2014 and key events. Used with permission from Zarin DA, Tse T, Sheehan J. The proposed rule for U.S. clinical trial registration and results submission. New Engl J Med January 8, 2015;372(2):174e180. Copyright © 2016 Massachusetts Medical Society.

derived from studies of the antidepressant, Paxil14). In addition to its own statutory mandates, ClinicalTrials. gov accommodates such policies and encourages voluntary registration of trials that do not fit under any of these policiesdunder the assumption that a single comprehensive database containing standardized information about clinical trials serves the public good and supports a wide variety of user needs. In 2007, the FDA Amendments Act of 2007 (FDAAA) was enacted to expand the ClinicalTrials.gov registry and add a results database.15 Section 801, which amends the Public Health Service Act, requires a “responsible party”ddefined as the study sponsor or a designated principal investigator who controls the study datadto submit information to the ClinicalTrials.gov registry for a clinical trial of FDA-regulated drugs, biologics, and devices that meets the definition of “applicable clinical trial” (generally, phases 2 through 4 clinical trials of drugs, biologics, or devices with at least one site in the United States). In addition, section 801 requires the submission of summary results information for applicable clinical trials of drugs, biologics, and devices that have been approved, licensed, or cleared by FDA. Other provisions of the law specify what “clinical trial information” is required to be submitted, when it is to be submitted and posted publicly, and the timeline for

updating that information. FDAAA states that a goal of expanding ClinicalTrials.gov is “to enhance patient enrollment and provide a mechanism to track subsequent progress of clinical trials.” Enforcement provisions for noncompliance with this federal law include civil monetary penalties and withholding of federal grant funding. In 2014, the Department of Health and Human Services (DHHS) issued a notice of proposed rulemaking (NPRM) for public comment describing the draft requirements and procedures for the registration and results reporting of clinical trials on ClinicalTrials.gov, in accordance with FDAAA.16 Notably, the NPRM proposed requiring (1) results reporting for clinical trials of unapproved products (i.e., drugs, biological products, or devices that have not been approved, licensed, or cleared by FDA for any use) and (2) at registration, more structured reporting of all primary and secondary outcome measures specified in the study protocol. It also invited comments on requiring the submission of several types of adverse event information, such as the time frame during which adverse event data were collected, and providing an all-cause mortality table listing all participant deaths from any cause.17 In parallel, NIH proposed a separate policy indicating that “all NIH-funded awardees and investigators conducting clinical trials, funded in whole or in part by NIH,

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

115

CURRENT POLICIES

TABLE 9.3 Comparison of Summary Requirements for the Proposed Rule, Draft NIH Policy, and the ICMJE Policya Requirement

FDAAA NPRM (Proposed Rule)

Draft NIH Policy

ICMJE Policy

Type of study

Interventional studies considered “controlled”c

Interventional studies considered “clinical trials”d

Any interventional study

Intervention type

Drugs, biologics, and devices

Any type of intervention, including surgical, behavioral, or other interventions

Any type of intervention, including surgical, behavioral, or other interventions

Trial phase

Any phase, except phase 1 drug trials and small feasibility device trials

Any phase

Any phase

Funding source

Any, whether private or public

NIH-funded, in whole or in part

Any, whether private or public

Scope of reporting

Registration and summary results

Registration and summary results

Registration only

Responsibility for ensuring reporting

Responsible party (sponsor or designated principal investigator)

NIH awardee

Author of manuscript

Enforcement mechanisms

Up to $10,000 per day in civil monetary penalties; possible withholding of NIH and other federal grant funds

Possible suspension or termination of NIH funding; consideration of noncompliance in future funding decisions

Editor’s refusal to publish manuscript

b

a

FDAAA denotes Food and Drug Administration Amendments Act, ICMJE International Committee of Medical Journal Editors, NIH National Institutes of Health, and NPRM notice of proposed rule-making. b An investigational study is defined in the NPRM as “a clinical study or a clinical investigation [in which] participants are assigned prospectively to an intervention or interventions according to a protocol to evaluate the effect of the intervention(s) on biomedical or other health related outcomes.” c These include all multigroup interventional studies but may exclude some single-group studies that do not involve a nonconcurrent control. d These include all interventional studies, except those that do not meet the revised NIH definition of “clinical trial” (http://grants.nih.gov/grants/guide/notice-files/NOT-OD-15015.html). Used with permission from Zarin DA, Tse T, Sheehan J. The proposed rule for U.S. clinical trial registration and results submission. New Engl J Med January 8, 2015;372(2): 174e180. Copyright © 2016 Massachusetts Medical Society.

regardless of study phase, type of intervention, or whether they are subject to the FDAAA . are expected to ensure that their NIH-funded clinical trials are registered and summary results, including adverse event information, are submitted to ClinicalTrials.gov” (Table 9.3).18 The over 1000 public comments received for both the NPRM19 and the draft NIH policy were carefully reviewed and considered in drafting a final rule and the NIH policy, which are scheduled to be issued in Fall 2016. (NOTE: At the time of this writing in August 2016, HHS had submitted the draft final rule to the White House Office of Management and Budget (OMB) for regulatory review.) The effects of these two policies on transparency of the clinical research enterprise are anticipated to be far-reaching: An estimated 70% of registered trials sponsored by large US academic medical centers and other nonprofit organization will be subject to at least one of these requirements.20

CURRENT POLICIES

registration policy and US federal law (FDAAA). Additionally, several US organizations have implemented policies requiring registration and results reporting to ClinicalTrials.gov including the following: • Centers for Medicare and Medicaid Services (CMS): A general requirement of the CMS Coverage with Evidence Development program includes study registration and the public availability of study results within 12 months of the study’s primary completion date21; • Department of Veterans Affairs (VA): Requires registration and results reporting of VA Office of Research & Development-funded trials22; and • Patient-Centered Outcomes Research Institute (PCORI): Requires registration and results reporting for PCORI-funded comparative effectiveness research clinical trials and observational studies.23 States’ attorneys general and the HHS Office of the Inspector General have incorporated ClinicalTrials.gov reporting requirements into various legal agreements with pharmaceutical manufacturers.24,25

Policies Affecting Clinical Trials in the United States

International Landscape

Two major policies that currently affect the reporting of clinical trials in the United States are the ICMJE

Since 2004, the World Health Organization (WHO) has promoted trial registration internationally,26 including

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

116

9. TRIAL REGISTRIES AND RESULTS DATABASES

developing a standard set of minimal information required for registration.27 In 2015, WHO issued a Statement on Public Disclosure of Clinical Trials Results which, among other things, recommended that “the key outcomes [of clinical trial results] are to be made publicly available within 12 months of study completion by posting to the results section of the primary clinical trial registry.”28,29 WHO also established and operates the International Clinical Trials Registry Platform Search Portal.30 The ICMJE accepts prospective registration of clinical trials in the 15 primary registries (as of 8/5/16) of the WHO portal and ClinicalTrials.gov. However, substantial growth of trial registries around the world is associated with unresolved issues. For example, registration of a clinical study more than once (or duplicate registration) has increasingly become a challenge across the world’s registriesdit is often difficult to tell whether two registry entries represent the same or different studies.31 In 2004, in the European Union (EU), the European Medicines Agency (EMA) launched a legislatively mandated database for clinical trials of drugs and biologics subject to its jurisdictiondthe European Union Drug Regulating Authorities Clinical Trials database (EudraCT).32 Initially, EudraCT was accessible only to regulatory and legal entities in the EU. Provisions in subsequent regulations required that protocol- and results-related information for certain clinical trials submitted to EudraCT be made available publicly, regardless of the status of marketing approval.33,34 The EMA created the EU Clinical Trials Register (https:// www.clinicaltrialsregister.eu/ctr-search/), from which summary protocol information (starting in 2011) and summary results (starting in 2014) information is publicly available. Notably, the EudraCT summary results data requirements are “substantially aligned” with those of the ClinicalTrials.gov results database.35 In 2014, the EU adopted the Clinical Trial Regulations, which include requirements such as additional disclosure of clinical study reports submitted as part of marketing authorizations for investigational medicinal products.36

REGISTERING CLINICAL TRIALS AT CLINICALTRIALS.GOV In general, registration is the process of submitting summary protocol information for a clinical trial for public posting. ClinicalTrials.gov permits the registration of biomedical or health-related research studies in humans that meet the following two requirements37: 1. Conformance with any applicable human subject or ethics review regulations (or equivalent) (e.g., institutional review board [IRB] approval) and

2. Conformance with any applicable regulations of the national (or regional) health authority (or equivalent) Because ClinicalTrials.gov serves as a long-term public registry, posted records remain available to the public even after a trial is over. Registration data are submitted through a Web-based Protocol Registration and Results System (PRS) in one of two modes: interactive data entry or Extensible Markup Language (XML)-file upload. Once the entered or uploaded data have been reviewed and approved by the data provider, records are released to ClinicalTrials. gov for processing. The content of each record is reviewed for apparent validity, meaningful entries, logic and internal consistency, and formatting. If any major problems are detected, the record is sent back with comments. If no major problems are identified, the record is generally posted at ClinicalTrials.gov within two to five business days. The first time that a study is posted, a ClinicalTrials.gov identifier (also known as the “NCT number”) is assigned to that study. This NCT number will be used from that point on to uniquely identify the particular clinical trial in various contexts (e.g., manuscript submission to a journal,38 certification to FDA or NIH, PubMed citation index). Once posted, records are accessible to the public and maintained by the data provider through the PRS. Updates to data elements (e.g., recruitment status) are made through the PRS and updated records reposted after quality review. Updates should be made as soon as practicable to keep the information current and accurate (with some data elements requiring more rapid updates based on legal requirements). A history of all changes (and the dates on which they were made) can be viewed on the publicly available archive site.

Data Standards and the Minimal Data Set The ClinicalTrials.gov protocol registration data elements can be divided into four general categories: 1. Descriptive information (e.g., study type, phase, study design, outcomes) 2. Recruitment information (e.g., eligibility criteria, overall recruitment status) 3. Location and contact information (e.g., sponsor name, facility name, and contact) 4. Administrative data (e.g., organization’s unique protocol identifier, secondary identifiers) These data elements are intended to provide the basic information needed to describe a study. Although not all are required by ClinicalTrials.gov or other policies, all data providers are encouraged to complete all data

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REGISTERING CLINICAL TRIALS AT CLINICALTRIALS.GOV

elements. Note that the ClinicalTrials.gov data elements include the ICMJE/WHO minimum 20-item Trial Registration Data Set (Version 1.2.1).27

Points to Consider Interventional Versus Observational Studies ClinicalTrials.gov uses the following definitions to distinguish between interventional and observation studies39: • Interventional: A clinical study in which participants are assigned to receive one or more interventions (or no intervention) so that researchers can evaluate the effects of the interventions on biomedical or healthrelated outcomes. The assignments are determined by a study protocol. Participants may receive diagnostic, therapeutic, or other types of interventions. • Observational: A clinical study in which participants identified as belonging to study groups are assessed for biomedical or health outcomes. Participants may receive diagnostic, therapeutic, or other types of interventions as part of their routine care, but the investigator does not assign participants to specific interventions (as in an interventional study). The key factor in differentiating between these study types is whether the individual received a specific intervention based on assignment by an investigator according to a research protocol. A common misconception is that studies of diagnostic interventions are not interventional. For example, some studies that investigators consider to be observational or epidemiologic use experimental diagnostic tests as part of the design. Use of these tests makes the study interventional because such participants are being exposed to an intervention as a consequence of the research protocol; they would not have received that intervention had they not been in the study. This would also be true if the study involves the use of an approved diagnostic test (e.g., a positron emission tomography scan) with greater frequency or in a different manner than would have occurred had the individual not been included in the study. What Is a Single Clinical Trial? Each clinical trial should be registered once at ClinicalTrials.gov. In general, a clinical trial has a defined group of human subjects who are assigned to interventions, and the collected data are grouped together for analysis, based on a protocol. However, it sometimes is difficult to determine what represents a single clinical trial. Many trials occur at multiple sites but follow a single core protocol (although a site may modify the

117

protocol, e.g., based on local IRB review), and the data from each site are intended to be combined and analyzed in aggregate. For purposes of registration and results reporting at ClinicalTrials.gov, such trials represent a single clinical trial and should be registered only once. On occasion, multiple individuals associated with a trial at different study sites inadvertently register it separately (i.e., duplicate registration). These separate registrations often will contain slightly different content for key data elements. The outcome of this is ambiguity about whether the two (or more) registry records represent the same study. This undermines one important function of the databasedto provide a unique denominator of all trials conducted on a given condition. Therefore, systems are in place at ClinicalTrials.gov to help avoid these “unintended” duplicates. Study sponsors or trial personnel can avoid such problems by identifying one person who has the responsibility for submitting information to ClinicalTrials.gov. This designation must occur at the very beginning and ideally should be noted on the protocol and within any clinical trial agreements. Another challenge in determining what is a single clinical trial involves follow-on study designs that occur after completion of the “initial” study in which participants are tracked for an extended period of time to allow for additional adverse event and secondary outcome data to be collected, typically using an open-label, nonrandomized design. Because these data collections often are defined within the original protocol and analysis plan, and include the same participants as the original protocol, such studies generally are considered a single clinical trial. In contrast, a follow-on study that requires reconsent and/or includes subjects who were not part of the original study is generally considered to be a separate trial. Importance of the Protocol As in much scientific research, the development of a research plan or protocol that includes prespecified hypotheses and methods, including explicitly defined variables of interest, is critical. The validity of any statistical analyses or conclusions is based on adherence to those prespecified methods. Registration must provide a description of the study/protocol with sufficient specificity to allow viewers to make a meaningful determination on whether a report of the results (either in a results database or in a publication) is consistent with the original (or amended) protocol.40 Keeping Information Up-to-Date Data must be kept up-to-date to best serve the needs of researchers, potential participants, and the general public. Some data elements (e.g., recruitment status, anticipated start and completion dates) will predictably

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

118

9. TRIAL REGISTRIES AND RESULTS DATABASES

change over time; other data elements may change if the protocol has been amended (e.g., modification of a primary outcome measure). In the case of ClinicalTrials. gov, all changes are tracked in a publicly available archive site (https://clinicaltrials.gov/archive/), although the default public view always will contain the most recent version for each data element.

REPORTING RESULTS TO CLINICALTRIALS.GOV Data providers may create the results section of a ClinicalTrials.gov record after data collection for that study is complete for at least one primary outcome measure. To do so, a protocol registration (i.e., a record with a ClinicalTrials.gov identifier) must already exist in general. Similar to the protocol registration process, results information must be submitted through the Web-based PRS, either interactively or by XML-file upload, using (required and optional) data elements.41 The results database as implemented was designed to satisfy key legal requirements, including the need for certain search features. The design of the ClinicalTrials. gov results database also was informed by current standards and best practices in results reporting (e.g., for journal publication42 or regulatory submission43), discussions with numerous experts, and comments from stakeholders.

Data Standards and the Minimal Data Set The Results section of the record consists of administrative information and four scientific modules: Participant Flow, Baseline Characteristics, Outcome Measures and Statistical Analyses, and Adverse Event Information. The data tables submitted in each module must TABLE 9.4

be sufficiently informative by themselves because they are displayed at ClinicalTrials.gov without detailed, supporting narrative text (Table 9.4).

Points to Consider Data Preparation Summarizing results information is typically a complex cognitive task that requires familiarity with the study and the data, and experience with presenting summary results data in a tabular format. Just as in the preparation of results for journal publication or submission of data to regulatory authorities, it is likely that the principal investigator and/or the study biostatistician will need to be involved in the preparation of a submission to the ClinicalTrials.gov results database. The information entered must be accurate, precise, and informative to an educated reader of the medical literature who is not already familiar with the specific study. After data entry but before the time of submission, it may be helpful to have a colleague who was not part of the study team review the submission for clarity and comprehension. Review Criteria Records are reviewed before public posting for apparent validity, meaningful entries, logic, internal consistency, and formatting (Table 9.5). Following automated validation, which alerts data providers to missing or internally inconsistent information, all submissions are manually reviewed. As of 2016, about a third of all records submitted by results data providers could be posted without revision. For example, some invalid data can be detected by ClinicalTrials.gov staff, but other data cannot be verified because ClinicalTrials.gov does not have an independent source of study data (e.g., “832 years” is clearly an invalid entry for mean age,

Summary of Scientific Modules in the ClinicalTrials.gov Results Database

Module (in Tabular Format)

Brief Description

Participant flow

Progress of research participants through each stage of a trial, including the number of trial participants who dropped out (identical in purpose to a CONSORT flow diagram,42 but represented as tables)

Baseline characteristics

Demographic and baseline data for the entire trial population and for each arm or comparison group

Outcome measures and statistical analyses

Data for each outcome measure by arm (i.e., initial assignment of groups to interventions) or comparison group (i.e., groups receiving interventions regardless of initial assignment). Accommodates categorical, continuous, and time-to-event data types and a variety of statistical analyses

Adverse events

Listing of (1) all serious adverse events and (2) other adverse events exceeding a specified frequency threshold within an arm or group. Both tables include anticipated and unanticipated adverse events by arm and are grouped by organ system.

Adapted from Tse T, Williams RJ, Zarin DA. Update on registration of clinical trials in ClinicalTrials.gov. Chest July 2009;136(1):304e305.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

119

REPORTING RESULTS TO CLINICALTRIALS.GOV

TABLE 9.5 ClinicalTrials.gov Quality Review Criteria Quality Review Criterion

Description

Example

Comment

Lack of apparent validity

Data are not plausible on the basis of information provided

Outcome measure data: mean value of 263 h of sleep per day

Measure of mean hours per day can have values only in the range of 0 to 24, so value of 263 is not valid

Meaningless entry

Information is too vague to permit interpretation of data

Outcome measure: description states “clinical evaluation of adverse events, laboratory parameters, and imaging,” data reported as 100 and 96 participants in each group

Data are uninformative; unclear what counts of 100 and 96 participants refer to; description of outcome measure not sufficient for an understanding of the specific outcome

Data mismatch

Data are not consistent with descriptive information

Outcome measure is described as “time to disease progression,” data reported as 42 and 21 participants in each group

A time-to-event measure requires a unit of time (e.g., days or months)

Internal inconsistency

Information in one section of record conflicts with or appears to be inconsistent with information in another section

Study type is “observational,” but study title includes the word “randomized”

Randomized studies are interventional, not observational

Trial design unclear

Structure of tables and relevant group names and descriptions do not permit a reader to understand the overall trial design or do not accurately reflect the design

Results modules: participant flow and baseline characteristics entered as a two-group study with a total of 400 participants; outcomes entered for three comparison groups with 600 participants

If there is a third group, this should be reflected in the description of participant flow and baseline characteristics

Used with Permission From Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results databaseeupdate and key issues. New Engl J Med March 3, 2011; 364(9):852e860. Copyright © 2016 Massachusetts Medical Society.

whereas “83.2 years” may or may not be the true mean age). Thus, meeting the ClinicalTrials.gov review criteria does not guarantee that the submission is valid or fully compliant with various policy or legal requirements. Relation of Results Reporting to Publication It is important to note that the goals of reviewing and reporting summary protocol and results information at ClinicalTrials.gov are not identical to those of editorial and peer review for journal publication. Although both seek to ensure accurate and informative data reporting, ClinicalTrials.gov does not reject studies based on the perceived quality or significance of the research and allows for disclosure of all prespecified primary and secondary outcomes as well as other prespecified and post hoc outcomes. In contrast, journal peer review focuses on selecting quality research likely to be of particular interest to its readers, and authors and editors may limit the focus of an article to critical aspects of the research. Journal review also aims to ensure that the narrative is consistent with the data and provides appropriate context and conclusions; because ClinicalTrials.gov does not permit significant narrative portions, these functions are not applicable. In November 2010, an analysis of a sample of 150 entries posted in the ClinicalTrials.gov results database revealed that only 78 (52%) were associated with a PubMed citation.44 Thus, despite the different goals, the results database

complements journal publication by providing information in a structured format that might not otherwise be available.

Key Scientific Principles and Best Practices for Reporting Issues in Reporting Outcome Measures Prespecification of primary and secondary outcome measures in the protocol and registration record is critical to the integrity of the conduct of the study as well as results reporting. ClinicalTrials.gov defines primary outcome measure as “a planned measurement described in the protocol that is used to determine the effect of interventions on participants in a clinical trial.”39 The Consolidated Standards of Reporting Trials (CONSORT) statement, an international standard for reporting the results of randomized clinical trials, requires that prespecified primary and secondary outcome measures be “completely defined., including how and when they were assessed (p. 2).”42 In the ClinicalTrials.gov results database, a fully specified outcome measure involves several components: four levels of specificity and a time frame (Fig. 9.3). Further, CONSORT defines the primary outcome measure to be “the pre-specified outcome considered to be of greatest importance to relevant stakeholders .and is usually the one used in the sample size

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

120

9. TRIAL REGISTRIES AND RESULTS DATABASES

FIGURE 9.3 An example of the four levels of specification in reporting outcome measures. Used with permission from Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results databaseeupdate and key issues. New Engl J Med March 3, 2011;364(9):852e860. Copyright © 2016 Massachusetts Medical Society.

calculation.” Analysis of 2178 clinical trial records with results posted revealed that the median number of primary outcome measures reported per trial was 1, although many include multiple primary outcome measures per trial (up to 71).44 This raises questions not only about the way in which clinical trialists use this term but also about the degree of selective reporting of outcome measures that may occur in the publication process because publications do not generally report such large numbers of primary outcome measures. ClinicalTrials.gov defines secondary outcome measures as “A planned Outcome Measure in the protocol that is not as important as the Primary Outcome Measure but is still of interest in evaluating the effect of an intervention.”39 Based on the same analysis (n ¼ 2178), the median number of secondary outcome measures reported per trial was 3 with up to 122 per trial.44 Because registration requires the reporting of all primary and secondary outcome measures, it has become necessary to consider the boundary between “secondary” and “other prespecified” outcome measures, which some call tertiary outcome measures. As suggested in the proposed rule,17 for purposes of registration and results reporting at ClinicalTrials.gov, outcome measures for which a specific analytic plan is prespecified are typically considered “secondary,” and other, more

exploratory outcome measures are considered “other prespecified.” Issues Related to Analysis Population In general, the outcome measure can be thought of as the numerator and the analysis population as the denominator when trial results are reported. For example, if the outcome measure of interest is the number of participants with a myocardial infarction (MI), outcomes may be compared across arms as number of participants with MI/number of participants in arm (or number of participants analyzed). Structured trial reporting has led to greater awareness of the degree to which different analysis populations may be used across analyses within a given trial. ClinicalTrials.gov does not mandate any particular method of analyzing results (e.g., intention to treat vs. per protocol). The goal is simply to ensure transparency in what was done. Results should enable other observers to make their own judgments about the validity of the reported analysis.

USING CLINICALTRIALS.GOV DATA Intended Audience The general public is ultimately the beneficiary of the ClinicalTrials.gov system, but different parts of

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

121

LOOKING FORWARD

ClinicalTrials.gov are likely to be of more or less interest to different audiences. For example, although summary protocol and eligibility information may be of use to members of the general public (and their advisors), results information in its current form may be of greatest interest to researchers and systematic reviewers, who understand the strengths and limitations of summary results from individual trials. Researchers may find the data posted in ClinicalTrials.gov to be useful when planning their own research projects. For example, a researcher may want to know what other trials have been completed (whether results have been published) or are ongoing that use the same intervention or patient population. Similarly, IRB members (and investigators) may want to ensure that they are aware of all prior relevant research, so that their assessment of risks and benefits for a proposed clinical study is as complete as possible. Researchers, systematic reviewers, and policy-makers use data from the ClinicalTrials.gov database to explore various aspects of the overall clinical research enterprise. For example, the registry and results database have been used to assess issues such as time to publication among completed registered trials,45 percentage of registered trials subject to federal human research oversight regulation,46 reasons for premature termination of trials,47 and discrepancies between ClinicalTrials.gov results database entries48,49 and other sources including publications and documents publicly available on the Drugs@FDA website.

Search Tips for ClinicalTrials.gov ClinicalTrials.gov provides a basic free text search of all registered studies. Advanced search options allow a user to conduct a more granular search that takes advantage of the database structure. All searches make use of the Essie concept-based search engine.50 For example, the search engine incorporates a bank of synonyms, so that a search for “paxil” also will find trials that use any known synonym (e.g., paroxetine, brl-29060, fg-7051, Seroxat); similarly, a search for “heart attack” will find trials that use the term “myocardial infarction.” Overall, the ClinicalTrials.gov website provides users with a flexible interface that accommodates searches of varying complexity for users with varying needs and levels of skill/experience. ClinicalTrials.gov allows for downloading of the data from retrieved studies or the full data set for analysis by data scientists. The Clinical Trials Transformation Initiative (CTTI) group also provides an “analysis-ready” data set derived from

ClinicalTrials.gov data that is intended to make it easier for researchers to assess the data.51

Points to Consider When Using ClinicalTrials. gov to Study the Overall Clinical Research Enterprise Although ClinicalTrials.gov is by far the largest source of information about ongoing and completed clinical trials, it does not include all studies. The data set will reflect prevailing policies, both in the United States and internationally. As these policies have changed over time, any sampling biases have likely changed as well. In general, the data set is likely to be most complete for drug and device trials sponsored by multinational or US companies. NIH-funded studies of drugs and devices are also likely to be well represented. However, care should be taken before making inferences about the overall worldwide population of clinical trials.

LOOKING FORWARD The Institute of Medicine52 and the ICMJE53 and others have called for greater sharing of individual participant data (IPD) for secondary analysis. The trialtransparency movement argues that providing greater third-party access to IPD, along with the protocol and other associated documents, would provide insights into decisions underlying the original analysis and provide accountability and allow for exploration of novel research questions.54 We believe that this new frontier in trial transparency can best be understood as part of an overall three-level trial reporting system (TRS) framework (Fig. 9.4).55 In this construct, registration and results reporting serve as the base of the TRS. The sharing of IPD and related documents sits atop these foundational layers. Without an overview of the clinical research landscape that the base provides, availability of IPD for certain trials but not others would result in another type of selective publication or reporting bias. In Table 9.6 we illustrate how registration, results reporting, and IPD sharing complement each other in increasing transparency, in the context of a case study. Thus, while systematic sharing of IPD is most certainly the newest frontier in trial transparency, this one component should not divert attention and/or resources from the ongoing need for more accurate, complete, and consistent registration and results reporting.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

122

9. TRIAL REGISTRIES AND RESULTS DATABASES

FIGURE 9.4 Schematic depicting the functions of the three key components of the trial reporting system. Reprinted under a CC-0 public domain dedication from Zarin DA, Tse T. Sharing individual participant data (IPD) within the context of the trial reporting system (TRS). PLoS Med January 2016; 13(1):e1001946.

TABLE 9.6

Key Issues With Trials of Antidepressant Use in Children for Depression and the Role of the trial reporting system (TRS)

Key Issue

Relevant TRS Component

Comment

Lack of prospective public information about all trials of Paxil and other selective serotonin reuptake inhibitors (SSRIs) in depressed children

Prospective registration

Registration would have provided a public list of all ongoing and completed trials of Paxil/SSRIs in depressed children

Alleged suppression of “negative” results from certain Paxil trials in depressed children62

Prospective registration

Registration would have allowed the detection of trials without disclosed results

Summary results reporting

Results database entries would have provided access to “minimum reporting set” including all prespecified outcome measures and all serious adverse events

Prospective registration

Archival registration information would have allowed for the detection of unacknowledged changes in prespecified outcome measures and detection of nonprespecified outcome measures reported as statistically significant

Summary results reporting

Structured reporting devoid of interpretation or conclusions would have made summary data publicly available while avoiding the possibility of spinning the results

Sharing highly granular individual participant data (IPD) and documents (e.g., case report forms)

Access to high-granularity IPD enabled the elucidation of data analytic decisions that had not been publicly disclosed; reanalysis was possible with different methods of categorizing adverse events

Detection of selective reporting bias of efficacy and safety findings in the published results of Study 329, unacknowledged changes in outcome measures, and other issues63

Invalid and unacknowledged categorization of certain adverse events, resulting in the underreporting of suicidality64

Reprinted under a CC-0 public domain dedication from Zarin DA, Tse T. Sharing individual participant data (IPD) within the context of the trial reporting system (TRS). PLoS Med January 2016;13(1):e1001946.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

CONCLUSION Because of a series of policy and legal actions, clinical trial registration has become standard practice internationally. Public reporting of summary results represents a new and evolving area. As clinical trial disclosure requirements continue to evolve, novel uses of data from the registry and results database will become clearer as more information is posted publicly. Ultimately, however, the quality and accuracydand thus utilitydof the registry and results database depend on the diligence and integrity of trial sponsors and investigators.

SUMMARY/DISCUSSION QUESTIONS 1. Which of the following is a rationale for registering and reporting results of clinical trials? a. Mitigating the effects of selective reporting and publication bias b. Fulfilling ethical principles underlying human research c. Facilitating assessment of research integrity, such as tracking protocol changes d. All of the above 2. The ClinicalTrials.gov Results Database consists of a. A repository of deidentified patient-level data for certain registered clinical trials b. Narrative abstracts from publications reporting clinical trial results c. Summary data displayed in a tabular format for certain registered trials d. Full-text articles submitted to ClinicalTrials.gov 3. Which of the following entities has not required the reporting of trial results to an online database? a. Medical journal editors b. US federal government c. State attorneys general d. The EU 4. ClinicalTrials.gov submissions are not a. Reviewed by both automated validation checks and human experts b. Provided through either interactive data entry or file upload c. Systematically verified against external, objective data sources d. Assigned a unique identifier that may be used to track that particular trial 5. When ClinicalTrials.gov data are used to perform aggregate analyses of a large sample of clinical trials, which caveat should be heeded by researchers? a. The database is not comprehensive: it does not include all clinical trials. b. The database is static: records do not need to be updated after registration.

123

c. The database is not comprehensive: it includes only recruiting clinical trials. d. The database is not comprehensive: it includes only drug and device clinical trials.

References 1. Zarin DA, Tse T. Medicine. Moving toward transparency of clinical trials. Science March 7, 2008;319(5868):1340e2. 2. Simes RJ. Publication bias: the case for an international registry of clinical trials. J Clin Oncol October 1986;4(10):1529e41. 3. Emanuel EJ, Wendler D, Grady C. What makes clinical research ethical? JAMA May 24e31, 2000;283(20):2701e11. 4. Levin LA, Palmer JG. Institutional review boards should require clinical trial registration. Arch Intern Med August 13e27, 2007; 167(15):1576e80. 5. Mann H. Research ethics committees and public dissemination of clinical trial results. Lancet August 3, 2002;360(9330):406e8. 6. World Medical Association. WMA declaration of Helsinki e ethical principles for medical research involving human subjects. 64th WMA general assembly, Fortaleza, Brazil. 2013. Available at: http:// www.wma.net/en/30publications/10policies/b3/. 7. Health omnibus programs extension act of 1988. Public law 100e607. 8. Katz DG, Dutcher GA, Toigo TA, Bates R, Temple F, Cadden CG. The AIDS Clinical Trials Information Service (ACTIS): a decade of providing clinical trials information. Public Health Rep MareApr 2002;117(2):123e30. 9. Grama LM, Beckwith M, Bittinger W, et al. The role of user input in shaping online information from the National Cancer Institute. J Med Internet Res July 1, 2005;7(3):e25. 10. Food and drug administration modernization act of 1997. Public law 105e115. 11. De Angelis C, Drazen JM, Frizelle FA, et al. Clinical trial registration: a statement from the international committee of medical journal editors. Ann Intern Med September 21, 2004; 141(6):477e8. 12. Laine C, De Angelis C, Delamothe T, et al. Clinical trial registration: looking back and moving ahead. Ann Intern Med August 21, 2007; 147(4):275e7. 13. Maine State Law, 22 MRSA c605, x2700-A An act regarding advertising by drug manufacturers and disclosure of clinical trials. 2005. Repealed in 2011 by H.P. 530 e L.D. 719. 14. Rennie D. Trial registration: a great idea switches from ignored to irresistible. JAMA September 15, 2004;292(11):1359e62. 15. Food and drug administration amendments act of 2007. Public law 110e185. 16. Notice of proposed rulemaking: clinical trials registration and results submission. Fed Regist November 21, 2014;79:69566e680. https://federalregister.gov/a/2014-26197. 17. Zarin DA, Tse T, Sheehan J. The proposed rule for U.S. clinical trial registration and results submission. New Engl J Med January 8, 2015;372(2):174e80. 18. National Institutes of Health. Request for public comments on the draft nih policy on dissemination of nih-funded clinical trial information. Bethesda, MD: NIH, Department of Health and Human Services; November 19, 2014. Available at: http://grants.nih.gov/grants/ guide/notice-files/NOT-OD-15-019.html. 19. Clinical trials registration and results submission. Docket ID: NIH-2011-0003. Available at: https://www.regulations.gov/#! docketDetail;D¼NIH-2011-0003. 20. Zarin DA, Tse T, Ross JS. Trial-results reporting and academic medical centers. New Engl J Med June 11, 2015;372(24):2371e2.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

124

9. TRIAL REGISTRIES AND RESULTS DATABASES

21. Centers for Medicare, Medicaid Services. Guidance for the public, industry, and CMS staff: coverage with evidence development. November 20, 2014. Available at: https://www.cms.gov/ medicare-coverage-database/details/medicare-coveragedocument-details.aspx?MCDId¼27. 22. Department of Veterans Affairs. ORD sponsored clinical trials: registration and submission of summary results. 2015. Available at: http:// www.research.va.gov/resources/ORD_Admin/clinical_trials/. 23. Patient-Centered Outcomes Research Institute. PCORI’s process for peer review of primary research and public release of research findings. February 24, 2015. Available at: http://www.pcori.org/sites/ default/files/PCORI-Peer-Review-and-Release-of-FindingsProcess.pdf. 24. Oregon Department of Justice. Attorney general’s medicaid fraud unit settles medicaid rebate cases with Merck. February 7, 2008. Available at: http://www.doj.state.or.us/releases/pages/2008/rel020808.aspx. 25. Office of Inspector General, U.S. Department of Health, Human Services. Johnson & Johnson Corporate integrity agreement. October 31, 2013. Available at: http://oig.hhs.gov/fraud/cia/agreements/ Johnson_Johnson_10312013.pdf. 26. Ghersi D, Pang T. From Mexico to Mali: four years in the history of clinical trial registration. J Evidence-Based Med February 2009;2(1): 1e7. 27. World Health Organization. Trial registration data set (version 1.2.1). 2011. Available at: http://www.who.int/ictrp/network/trds/en/ index.html. 28. World Health Organization. WHO statement on public disclosure of clinical trial results. April 14, 2015. Available at: http://www.who. int/ictrp/results/reporting/en/. 29. Moorthy VS, Karam G, Vannice KS, Kieny MP. Rationale for WHO’s new position calling for prompt reporting and public disclosure of interventional clinical trial results. PLoS Med April 2015;12(4):e1001819. 30. World Health Organization. International clinical trials registry platform (ICTRP): about the ICTRP search portal. 2011. Available at: http://www.who.int/ictrp/search/en/. 31. van Valkenhoef G, Loane RF, Zarin DA. Previously unidentified duplicate registrations of clinical trials: an exploratory analysis of registry data worldwide. Syst Reviews 2016;5(1):116. 32. European Commission. Directive 2001/20/EC of the European Parliament and of the Council of 4 April 2001 on the approximation of the laws, regulations and administrative provisions of the member states relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use. Off J Eur Communities April 4, 2001;L121:34e44. 33. European Commission. Communication from the commission regarding the guideline on the data fields contained in the clinical trials database provided for in article 11 of directive 2001/20/EC to be included in the database on medicinal products provided for in article 57 of regulation (EC) No 726/2004. Off J Eur Union July 3, 2008;C168:3e4. 34. European Commission. Communication from the commission d guidance on the information concerning paediatric clinical trials to be entered into the EU Database on Clinical Trials (EudraCT) and on the information to be made public by the European Medicines Agency (EMEA), in accordance with article 41 of regulation (EC) No 1901/2006. Off J Eur Union February 4, 2009;C 28:1e4. 35. European Medicines Agency. European medicines agency launches a new version of EudraCT: summary results of clinical trials soon to be available to the public. October 11, 2013. Available at: http://www. ema.europa.eu/ema/index.jsp?curl¼pages%2Fnews_and_events %2Fnews%2F2013%2F10%2Fnews_detail_001918.jsp. 36. European Commission. Regulation (EU) No 536/2014 of the European Parliament and of the council of 16 April 2014 on clinical trials on medicinal products for human use, and repealing directive 2001/20/EC. Off J Eur Communities May 27, 2014;L158:1e76.

37. Tse T, Williams RJ, Zarin DA. Update on registration of clinical trials in ClinicalTrials.gov. Chest July 2009;136(1):304e5. 38. Zarin DA, Tse T. Unambiguous identification of obesity trials. New Engl J Med February 7, 2013;368(6):580e1. 39. National Library of Medicine. ClinicalTrials.gov glossary of common site terms. 2012. Available at: https://clinicaltrials.gov/ct2/aboutstudies/glossary. 40. Zarin DA, Tse T. Trust but verify: trial registration and determining fidelity to the protocol. Ann Intern Med July 2, 2013;159(1):65e7. 41. Tse T, Williams RJ, Zarin DA. Reporting “basic results” in ClinicalTrials.gov. Chest July 2009;136(1):295e303. 42. Schulz KF, Altman DG, Moher D, Group C. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med June 1, 2010;152(11):726e32. 43. ICH Harmonised Tripartite Guideline E3: Structure and Content of Study Reports. 30 International conference on harmonisation of technical requirements for registration of pharmaceuticals for human use. November 1995. 44. Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results databaseeupdate and key issues. New Engl J Med March 3, 2011;364(9):852e60. 45. Ross JS, Mocanu M, Lampropulos JF, Tse T, Krumholz HM. Time to publication among completed clinical trials. JAMA Intern Med May 13, 2013;173(9):825e8. 46. Zarin DA, Tse T, Menikoff J. Federal human research oversight of clinical trials in the United States. JAMA March 5, 2014;311(9): 960e1. 47. Williams RJ, Tse T, DiPiazza K, Zarin DA. Terminated trials in the ClinicalTrials.gov results database: evaluation of availability of primary outcome data and reasons for termination. PLoS One 2015;10(5):e0127242. 48. Becker JE, Krumholz HM, Ben-Josef G, Ross JS. Reporting of results in ClinicalTrials.gov and high-impact journals. JAMA March 12, 2014;311(10):1063e5. 49. Hartung DM, Zarin DA, Guise JM, McDonagh M, Paynter R, Helfand M. Reporting discrepancies between the ClinicalTrials.gov results database and peer-reviewed publications. Ann Intern Med April 1, 2014;160(7):477e83. 50. Ide NC, Loane RF, Demner-Fushman D. Essie: a concept-based search engine for structured biomedical text. J Am Med Inform Assoc JAMIA MayeJun 2007;14(3):253e63. 51. Clinical Trials Transformation Initiative. State of clinical trials: AACT database. 2016. Available at: http://www.ctti-clinicaltrials.org/ what-we-do/analysis-dissemination/state-clinical-trials/aactdatabase. 52. Institute of Medicine. Sharing clinical trial data: maximizing benefits, minimizing risks. Washington DC: The National Academies Press; January 14, 2015. 53. Taichman DB, Backus J, Baethge C, et al. Sharing clinical trial data: a proposal from the international committee of medical journal editors. Ann Intern Med April 5, 2016;164(7):505e6. 54. Zarin DA. Participant-level data and the new frontier in trial transparency. New Engl J Med August 1, 2013;369(5):468e9. 55. Zarin DA, Tse T. Sharing individual participant data (IPD) within the context of the trial reporting system (TRS). PLoS Med January 2016;13(1):e1001946. 56. Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective publication of antidepressant trials and its influence on apparent efficacy. New Engl J Med January 17, 2008;358(3): 252e60. 57. Lancet. Is GSK guilty of fraud? Lancet June 12, 2004;363(9425):1919. 58. Juni P, Rutjes AW, Dieppe PA. Are selective COX 2 inhibitors superior to traditional non steroidal anti-inflammatory drugs? BMJ June 1, 2002;324(7349):1287e8. 59. Krumholz HM, Ross JS, Presler AH, Egilman DS. What have we learnt from Vioxx? BMJ January 20, 2007;334(7585):120e3.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

REFERENCES

60. Mitka M. Controversies surround heart drug study: questions about Vytorin and trial sponsors’ conduct. JAMA February 27, 2008;299(8):885e7. 61. Vedula SS, Bero L, Scherer RW, Dickersin K. Outcome reporting in industry-sponsored trials of gabapentin for off-label use. New Engl J Med November 12, 2009;361(20):1963e71. 62. GSK assurance of discontinuance. August 26, 2004. Available at: http://www.ag.ny.gov/sites/default/files/press-releases/ archived/aug26a_04_attach2.pdf.

125

63. Jureidini JN, McHenry LB, Mansfield PR. Clinical trials and drug promotion: selective reporting of study 329. Int J Risk Saf Med 2008;20:76e81. 64. Leslie LK, Newman TB, Chesney PJ, Perrin JM. The food and drug administration’s deliberations on antidepressant use in pediatric patients. Pediatrics July 2005;116(1):195e204.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

C H A P T E R

10 Data and Safety Monitoring 1

Paul G. Wakim1, Pamela A. Shaw2

National Institutes of Health, Bethesda, MD, United States; 2University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, United States

O U T L I N E Why Monitor?

127

Who Monitors? Data and Safety Monitoring Board History of Data and Safety Monitoring Boards When Is a Data and Safety Monitoring Board Needed?

128 128

What to Monitor? Monitoring Participant Safety Monitoring Trial Conduct Participant Flow Participants’ Baseline Characteristics Randomization Outcome Regulatory Compliance Trial Performance Data Quality

130 130 131 131 131 131 132 132 133

133 134 135

128

When and How Often to Monitor?

136

129

Special Topics General Structure of Data and Safety Monitoring Board Meetings Masking of the Data and Safety Monitoring Board

137

Summary

138

Summary Questions

139

Acknowledgments

139

References

139

WHY MONITOR? The primary reason for monitoring a clinical trial is to ensure that it does not compromise the safety of its participants. Increased risk of participation could be a result of the experimental treatment(s), control treatment, or any other trial-related procedure. Investigators’ legal and ethical responsibilities are to ensure that the participants in their trial are not subjected to unnecessary physical harm and/or suboptimal care, that is, providing a treatment to some or all participants when there is evidence of better treatments. This evidence could either come internally, from accumulated study data, or externally from clinical trials conducted by other

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00010-1

Interim Analyses Sample Size Recalculation Interim Analyses for Efficacy, Futility, and/or Harm

137 137

groups. Unnecessary risk also can result from continuing the trial when there is little chance of gaining additional scientific information. In situations like these, stopping the trial may be the only course of action. The second most important reason for trial monitoring is to ensure data integrity. At regular time points during the trial, investigators and sponsors should ask themselves the following questions: when the trial is completed, will the quality of the data collected be high enough to produce meaningful results? Will the data analysis be able to answer the primary research question with reasonable accuracy and minimal bias? From a data validity perspective, is continuing the trial worth the additional effort and expense? Threats to data integrity

127

Copyright © 2018. Published by Elsevier Inc.

128

10. DATA AND SAFETY MONITORING

could result from very slow enrollment, poor randomization process, inaccurate data entry, inadequate masking mechanism, or extensive missing data due to missed visits or loss to follow-up. Another important reason to monitor a trial is to validly answer the primary research question with the minimum amount of resources. For example, if it is found, while the trial is ongoing, that it can produce valid results with a smaller sample size, recruiting the larger sample is unethical to the yet-to-be-randomized participants (exposing them to unnecessary risk) and wastes the sponsor’s time and money. It is also unethical to future patients who would have benefited earlier from possible positive trial results. One bonus benefit of vigilant and objective monitoring is that it increases the validity and credibility of the results in the eyes of the scientific community and regulatory agencies. In this chapter, we discuss key elements of data and safety monitoring (DSM), the frequency of the activity, and the groups responsible for monitoring the clinical trial, including the roles and responsibilities of the trial’s Data and Safety Monitoring Board (DSMB).

WHO MONITORS? Data and Safety Monitoring Board Because investigators are, by definition, very closely involved in their own clinical trial, and therefore may not be completely objective with regard to their own trial’s progress and performance, a group of independent experts is called to monitor the trial’s conduct. To effectively monitor the trial and determine whether it is ethical to continue, the external monitoring board frequently relies on examining unmasked (unblinded) data on safety and other study outcomes. The review of unmasked data needs to be done with utmost confidentiality and should not be shared with study investigators or the sponsor while the trial is ongoing. The research team’s knowledge of results accumulated so far may influence their decisions and behavior with participants, potentially biasing future outcomes and final results. This group of outside experts goes by several names. The most common names are: DSMB, Data and Safety Monitoring Committee (DSMC), and Data Monitoring Committee (DMC). Less common names are: External Safety Monitoring Committee (ESMC) and Treatment Effects Monitoring Committee (TEMC). DSMBs are typically comprised of an independent group of subject-matter experts in the disease setting and type of interventions being studied. For example, a DSMB for a study of a novel imaging tool to guide

treatment in breast cancer would include a breast cancer specialist and a radiologist. An ethicist and a biostatistician are other key members of the board. In many settings, it has become standard also to include a patient advocate. It is also helpful to have members with extensive DSMB experience, who can offer guidance on conventions and conduct of the meeting and who can help guide the board through what may be difficult or complex decisions about the trial. The roles and responsibilities of the DSMB are outlined in the DSMB charter, which lists board members and their affiliation, and identifies key roles, such as the DSMB chair, ethicist, and biostatistician (see for example, Ref. 1). The DSMB chair runs the meetings and ensures that the guidelines of the charter are followed. DSMB chairs are the main liaison between the board, trial sponsor, and trial investigators. They are responsible for communicating the DSMB recommendations to the trial sponsor at the end of each meeting. The charter also specifies the voting members of the DSMB. Voting members are independent of the trial leadership and sponsor because they may need to make difficult decisions about the trial, such as recommending closing a study arm, closing enrollment at a poor-performing site, or even stopping the trial altogether. Credibility of trial results could be put into question if there is a perceived conflict of interest for one or more DSMB members. For example, if a trial is stopped early for efficacy, a journal editor or regulatory body reviewing the main results could call into question this decision if it perceives that there were DSMB members or someone closely related to the member who would stand to financially gain from such an action. For this reason, DSMB members need to be free of any significant conflict of interest, such as financial holdings or professional relationship with a company that manufactures a product that could be impacted in some way by results of the study. Regularly during the study, DSMB members should be solicited to provide financial disclosures and other potential conflicts of interest to trial leadership. Keeping DSMB members independent and free of conflicts helps maintain the confidentiality of the trial, as relationships and/or regular contact with the sponsor or trial team may create an environment where confidential information about interim trial results could be accidentally revealed, or where DSMB members could be subject to pressure to disclose confidential information discussed at DSMB meetings.

History of Data and Safety Monitoring Boards The concept of an independent board overseeing the conduct of a randomized clinical trial surfaced in the 1950s. The US National Institutes of Health (NIH) was

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

WHO MONITORS?

probably the first organization to put the concept into practice, in the early 1960s, in large trials on improving survival from acute myocardial infarction.2,3 As early as 1962, one of the first NIH DSMBs (called Policy Board at the time) was established to act in a senior advisory capacity to the participating investigators of the Coronary Drug Project (CDP) of the National Heart Institute (currently the National Heart, Lung, and Blood Institute).4 A few years later, the same Policy Board recommended that CDP personnel no longer have access to study endpoints. It also recommended the formation of a Safety Monitoring Committee to review confidential safety data.4 In 1967, an advisory board to the National Heart Institute developed recommendations on the organization, review, and administration of cooperative studiesdlarge multisite clinical trials. These recommendations were published 20 years later in the influential Greenberg Report.5 One of the recommendations was a cooperative study organizational structure with a “Policy Board or Advisory Committee of senior scientists” who would “review the overall plan, make recommendations on any possible changes (including changes in protocol and operating procedures), adjudicate controversies that may develop, and advise the National Heart Institute on such matters as the addition of new participants or the dropping of nonproductive units.” Members of this Advisory Board should be “experts in the field of the study but not data-contributing participants in it”.5 Since then, it became more common for large randomized clinical trial sponsors to form independent committees to monitor the safety, data integrity, and ethics of their trials. Other US federal agencies started to establish independent boards for their trials: the National Eye Institute in the early 1970s; the Department of Veterans Affairs in the mid-1970s; the National Cancer Institute and National Institute of Allergy and Infectious Diseases in the early 1980s.2 In the early 1990s, the pharmaceutical industry started to use DSMBs, particularly in trials on cardiovascular diseases.2 However, despite the growing use of DSMBs, little was published on the operational aspects of trial monitoring and the functioning of monitoring committees. So in 1992, the NIH organized a workshop on practical issues in data monitoring of clinical trials, and in 1993, a whole issue of the journal Statistics in Medicine was dedicated to the workshop proceedings.6 The articles in that issue presented experiences from a wide variety of disciplines, industries, countries, and disease areas. They addressed ethical, logistical, and operational considerations in data monitoring and interim analyses. As of today, many books, journal articles, and guidance documents on DSMBs have been published. In addition to the references cited in the body of this

129

chapter, we list a few additional ones in the Summary section, at the end of this chapter.

When Is a Data and Safety Monitoring Board Needed? Clinical trialists, regulatory agencies, and sponsoring organizations would all agree that (1) every clinical trial needs oversight and monitoring and (2) not every clinical trial needs an independent DSMB. But then, how do investigators determine whether they should have a DSMB for their trial? Current guidelines are vague and general, perhaps rightly so, since it is difficult to classify clinical trials generically into those that do and those that do not need a DSMB, without knowing the details of each trial. A World Health Organization (WHO) report states that a DSMB is considered “relevant” in studies that focus on mortality and/or severe morbidity; involve high-risk interventions; test novel interventions with potential serious adverse outcomes; of long duration; where interim analyses could justify early termination; in emergency situations; and those that involve vulnerable populations.7 The report then states that not all studies that fall in these categories necessarily require a DSMB and that there may be trials that do not fall in any of these categories that may still need a DSMB. Similar guidelines were developed by the European Medicines Agency (EMA), which give examples of when a DSMB should be set up and when it should not.8 For example, the guidelines state that “clinical studies in non-critical indications where patients are treated for a relatively short time and the drugs under investigation are well characterized and known for not harming patients” might not need a DSMB. The EMA adds that with such trials, a DSMB may even be counterproductive because the additional preparations for the DSMB may delay the closure of the trial.8 US federal government agencies also have developed their own guidelines (e.g., Refs. 3,9e11). Others also have published recommendations on the need for a DSMB. For example, the Society for Clinical Trials Working Group on Data Monitoring developed guidelines for DSM of early-phase (Phase I and II) trials that do not require an outside, independent monitoring board12; and Ellenberg et al.2 list four general criteria for determining the need and value of a DSMB. The gist of these guidelines is fairly similar, namely, that a DSMB is needed when the trial (1) involves interventions with relatively high or unknown risks (e.g., gene therapy in advanced stage cancer patients); (2) concerns a disease/condition that has serious health implications (e.g., coronary heart disease); (3) is conducted on a fragile or vulnerable population (e.g., pregnant

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

130

10. DATA AND SAFETY MONITORING

women, the elderly in nursing homes, and children); (4) is large enough (i.e., of long duration and/or costly) that it can be stopped early; or, (5) is controversial (e.g., showing that an existing standard of care that has been used for many years is not better than placebo, or perhaps even inferior to placebo when considering its side effects). Since late-phase (Phase III and IV) and multicenter clinical trials generally fall under at least the fourth category, they typically need independent monitoring. Trials that do not need DSMBs are earlyphase (Phase I and II), single-site, open-label, low-risk trials. For trials that fall between these two ends of the spectrum, investigators should consult with the regulatory agencies and sponsoring organization to discuss each specific case.

Interim analyses examine data on clinical or biological outcomes collected so far. They are conducted while the trial is ongoing. They are particularly useful when there is potential to appreciably shorten the duration of a lengthy trial. Interim analyses can be divided into two types: (1) sample size recalculation and (2) interim analyses for efficacy, futility, and/or harm. The next three sections address each monitoring activity in more detail. Although participants’ safety is always most important, it may not be the primary clinical outcome of interest. For example, if it has been shown in previous studies that a medication is generally safe, the primary research question may focus instead on whether the medication reduces blood pressuredthe primary research outcome of interest. When a riskebenefit balance is not clear, both efficacy and safety outcomes should be monitored during an interim analysis.

WHAT TO MONITOR? There are three main components of clinical trial monitoring: (1) participant safety; (2) trial conduct; and (3) interim analyses (see Fig. 10.1). Monitoring participant safety examines whether there are safety concerns. It generally includes consideration of adverse events (AEs) that occur during the trial, with a key focus on frequency and occurrence of unexpected events, as well as any severe events. Monitoring trial conduct assesses whether the trial is being successfully carried out as planned. It typically evaluates (1) participant flow; (2) participants’ baseline characteristics; (3) randomization outcome; (4) regulatory compliance; (5) trial performance; and (6) data quality.

Monitoring Participant Safety Again, first and foremost, investigators should be concerned about the safety and well-being of their trial’s participants. The more risky the trial interventions, the more scrutinized safety monitoring should be. AE is the most commonly used measure of safety. The NIH defines AEs as any “untoward medical occurrence in a human subject, including any abnormal sign (for example, abnormal physical exam or laboratory finding), symptom, or disease, temporally associated with the subject’s participation in research, whether or not considered related to the subject’s participation in the research.”13 The NIH also specifies that some AEs are considered serious adverse events (SAEs) if they

FIGURE 10.1 Most common clinical trial monitoring activities.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

WHAT TO MONITOR?

result in death; are life-threatening; require either inpatient hospitalization or the prolongation of hospitalization; result in a persistent or significant disability/ incapacity; or result in a congenital anomaly/birth defect. Based on appropriate medical judgment, other important medical events also may be considered SAEs if an intervention on a trial participant is required to prevent one of the outcomes mentioned above.13 The International Conference on Harmonisation (ICH) defines an SAE in similar terms.14 For each AE/SAE occurrence, the following information is typically reported: participant identification code number, AE/SAE description, start date, severity, relationship to trial, outcome, and resolution date (when available). Since many treatments have expected and acceptable off-target effects, such as a standard cancer therapy causing nausea, pain, or hair loss, AEs/SAEs need to be discussed in terms of whether or not they are expected or unexpected. The severity of the disease being studied is a factor when considering the acceptability of the AE. For instance, for a life-threatening condition, such as an advanced-stage cancer, an increased occurrence of a treatable AE, such as nausea or pain may be acceptable, whereas in another setting, such as a new antihistamine given to healthy people for a mild allergy, it may not. This can be a complex evaluation, and one that deserves robust discussion by an interdisciplinary committee in terms of what is and is not acceptable. The issue of unanticipated risk is given further discussion in Food and Drug Administration (FDA)15 and in Chapter 11 (Unanticipated Risk in Clinical Research).

Monitoring Trial Conduct This activity focuses on the general conduct of the trial and not on the research results of the trial. It is performed without any look at data related to the trial’s endpoints or research outcomes. What follows is a list of typical trial information that helps assess how well the trial is proceeding. Participant Flow Participant flow, sometimes called participants’ disposition, is the procedural flow of individuals going from being approached to participate in the clinical research study, to being prescreened, then screened for inclusion/exclusion criteria, consented, randomized, receiving and completing the corresponding treatment, staying in the trial for the whole duration of the trial, and finally being included in the primary analysis. Typically, participant flow is a visual explanation of how the large number of individuals who were approached becomes the much smaller number of

131

participants who are included in the primary analysis. This is important because it shows whether the declining number of individuals at every trial stage from beginning to end reflects realistic expectations in a clinical trial on such a disease and target population. It is used to flag any potential selection biases in the group of participants who are eventually analyzed. It also indicates whether the inclusion/exclusion criteria are too stringent (i.e., including in the trial only a narrow subset of the population of interest). In summary, it gives a sense of the representativeness of the group of participants who will be included in the final primary analysis. For visual examples of participant flow diagrams, the reader is referred to guidelines from the ICH16 and Moher et al.17 Participants’ Baseline Characteristics Baseline characteristics represent information on study participants’ important traits collected before randomization. Examples of baseline characteristics include demographics (e.g., gender, race, ethnicity, age, socioeconomic status, education), basic body parameters (e.g., height, weight, body mass index), vital signs (e.g., blood pressure), and disease-related information (e.g., severity, onset). The objective of this exercise is to assess whether the participants recruited so far are representatives of the target population described in the protocol and to ensure that the trial results can be generalized to the broader target population. Randomization Outcome In addition to establishing causality by attributing differences between treatment groups to efficacy, randomization is performed in clinical trials to achieve balance in known and unknown factors that could influence the primary response. The question here is whether randomization has been properly conducted and whether the results of randomization match the intended randomization strategies. One way to address it is to examine the profiles of study participants who have been randomized so far to each intervention group, including their stratification factors. For example, if gender is a stratification factor, there should be roughly an equal number of women in each intervention group, and similarly for men. Other baseline characteristics also should be balanced across intervention groups. Although statistical tests and the reporting of corresponding P-values to compare baseline characteristics between intervention groups are to be avoided after the trial is completed,18,19 they can be used while the trial is proceeding to check whether randomization is properly conducted.20 One should keep in mind though that if a relatively large number of baseline characteristics are being examined, one can expect by chance a few imbalances between treatment groups, typically

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

132

10. DATA AND SAFETY MONITORING

for about 5% of the baseline characteristics. However, repeated imbalances over the course of the trial, and patterns of imbalances should be flagged and investigated further. Regulatory Compliance Complying with regulatory laws, rules and regulations is not a trivial undertaking. Keeping up-to-date with all regulatory issues and ensuring that they are followed and implemented can be daunting. Reporting deadlines, specially related to reporting AEs and SAEs, need to be met. In addition to regulatory safety reporting requirements, here are some examples of specific regulatory compliance items to monitor: Institutional Review Board (IRB) approval/renewal; Federalwide Assurance number and expiration date; FDA Investigational New Drug (IND) application; and any other significant regulatory issues. Ensuring regulatory compliance is ultimately the responsibility of the sponsor. In this book, Chapters 4 (Institutional Review Boards), 6 (The Regulation of Drugs and Biological Products by the Food and Drug Administration), and 12 (Legal Issues) cover different aspects of this topic in more detail. Trial Performance What follows is a list of key trial performance criteria. Protocol Compliance by Research Staff In general, deviations from, and violations of, operating procedures described in the protocol should be documented and reported to the IRB, DSMB, and applicable regulatory agencies. Such violations could include

improper informed consent procedures, inclusion/ exclusion criteria not met, visits conducted outside the permissible time window, or inadequate record keeping. Clinicians’ adherence to treatment (treatment fidelity) is another example of protocol compliance. The question is whether research staff involved in providing treatment to participants are following the procedures and approaches described in the protocol. This is particularly germane with clinical trials assessing the efficacy of psychosocial therapies. Although quantifying clinicians’ treatment adherence may be challenging, it is important to capture and report that information when it is directly related to data integrity as defined above. One approach is to videotape a sample of psychosocial treatment sessions and evaluate their conformity to the protocol. Recruitment Is the trial recruiting at the expected pace specified in the protocol? If not, why not? Are there any actions the investigator or sponsor can take to resolve recruitment issues? Recruitment performance can be graphed as in Fig. 10.2 by showing the actual and expected number of randomizations over time. An accompanying table can show the expected and actual number of randomizations. Actual divided by expected number of randomizations (expressed in percent) is a quick way to check on recruitment, particularly if the trial involves multiple sites. In such cases, the graph and table can be presented for each site, as well as overall. In trials where a sufficient number of individuals are being screened, but a higher than expected proportion are found to be

FIGURE 10.2 Monitoring recruitment.

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

WHAT TO MONITOR?

ineligible, presenting additional graphs for the number screened and the number randomized can be informative with respect to whether the recruitment problem is related to finding potentially eligible and interested participants, or that the eligibility criteria are difficult to satisfy. Participants’ Treatment Adherence (Treatment Exposure) It is equally important that participants get the treatment they are assigned to receive, for example, taking the medication and/or attending counseling sessions as often as described in the protocol. Otherwise, the treatments being compared at the end of the trial are no longer the ones intended, and the results no longer reflect the primary objective of the trial. One way to monitor treatment exposure is to quantify the expected treatment that each participant should receive and the treatment actually received. For example, based on the number of participants randomized, one can calculate the number of medication pills expected to have been taken by all participants combined, or the number of psychosocial therapy sessions expected to have been attended by all participants combined. One also can calculate the number of medication pills actually taken, or number of sessions actually attended. The “actual” divided by the “expected,” expressed in percent, gives a general sense of treatment adherence. Data Completeness (Availability of Primary and Other Key Endpoints) This is not about revealing the value of the primary or other key endpoints, whether for each participant or in aggregate. It is about determining whether the endpoint data are collected and entered in the database, that is, the extent of “missingness” of the primary and other key outcomes of interest. The ratio of the number of primary outcome values actually collected (nonmissing) at a particular point in time to the number of primary outcome values expected to be collected at the same point in time, expressed in percent, is one way to quantify the extent of “nonmissingness.” It represents the amount of data that will be available for the primary analyses. The higher the extent of “missingness,” the less reliable the final results. Indeed, the main advantage of randomizationdinference on causalitydis compromised when the database includes too many missing primary endpoints. Outcomes that are not of primary interest but are burdensome to collect (e.g., biopsy samples for a nonprimary biomarker endpoint) also are worth monitoring for completeness to justify the excess burden on participants. A sponsor also may wish to have expensive endpoints monitored to assure that they are collected in a manner that will yield meaningful scientific results.

133

Attendance at Follow-Up Visits (Retention) Results from follow-up visits are meaningless, and resources are wasted, if few participants come back for follow-up visits. The number of follow-up visits actually attended by all participants randomized so far, and who have reached that stage of the clinical trial, divided by the number of follow-up visits expected to have been attended at this point in time, expressed in percent, is one way of quantifying retention. If retention rates are worrisome, the DSMB makes suggestions for strategies to improve retention, or more often, asks the research team to come up with a plan to try to improve retention. Data Quality Data captured in the trial’s database need to be valid and accurate. Investigators need to monitor the quality of the data on a regular basis to ensure that data entry is performed as accurately as possible and that data from other sources (e.g., laboratories) are transferred to the study database without errors. An error rate (total number of discrepancies divided by the total number of fields audited), expressed in percent, is one way of quantifying and monitoring data quality. Flags and Triggers Some of the trial conduct indicators listed above can be summarized in color-coded tables. For example, the criteria shown in Fig. 10.3 may be used to color-code the performance of the trial overall and that of each participating clinical site. This gives a quick visual and objective way of identifying good- and poorperforming sites in a multicenter clinical trial. The thresholds indicated in the figure are only examples for illustration purposes. They vary according to the research field, treatments being tested, and trial phase (IeIV). In summary, monitoring trial conduct is an exercise that examines descriptive statistics related to trial performance and not participant outcomes or trial results.

Interim Analyses Interim analysis typically refers to a statistical analysis of the primary endpoint performed while the trial is proceeding. The values of the primary outcome (as opposed to whether they are missing) are usually the focus of primary analyses. They may be used to calculate the treatment effect (the difference between the effects of the experimental and control treatments), or some other key statistical parameters such as variability. Trial designs that include interim analyses fall under the umbrella of “adaptive designs” because the final trial design depends on the results of interim looks at the data (Ref. 21; see also Chapter 27, Intermediate Topics

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

134

10. DATA AND SAFETY MONITORING

FIGURE 10.3 Flags and triggers.

in Biostatistics). As is the case with all adaptive designs, interim analyses are prespecified in the protocol. The most commonly used interim analyses can be grouped into two categories: (1) sample size recalculation (or reestimation) and (2) interim analyses of participant outcomes to test for efficacy, futility, and/or harm. These two types are addressed in the next two sections. The result of an interim analysis is only one of several pieces of information that is considered in deciding how the trial should continue or whether it should be stopped.18 For example, Anand, Wittes, and Yusuf22 have argued that one may want to continue a trial even after an interim analysis shows futility, in order to collect as much safety information as possible, and to produce robust safety estimates. The number of AEs may be significantly different between two treatment conditions with similar effects, in which case the treatment with the higher AEs is to be avoided since it has no benefit and only more safety concerns. Sample Size Recalculation In statistics, a nuisance parameter is a parameter whose value is not of particular interest but that does affect the distribution of other parameters that are of interest. For example, investigators are not directly interested in the variance of the primary endpoint; but the magnitude of the variance does affect the results of the study, such as the width of the confidence interval and the corresponding P-value. Other examples of nuisance parameters, say in a setting where the mean response is of interest, are correlation between responses within a cluster, drop-out rate, and proportion of events in the overall trial cohort.

During the protocol development phase, values for the minimum treatment effect to be detected and for nuisance parameters are needed for power and sample size calculations. They are typically based on educated guesses. It is therefore advisable to assess at some point during the trial whether these guesses were realistic. Sample size recalculation is about revisiting the initial sample size that was determined at the trial design stage before the trial started. Sample size recalculation is generally planned upfront and described in the protocol. It is timed so that the sample size can be changed before recruitment is completed. If the initial sample size is small, or if recruitment is fast, the initial sample size may be reached by the time a new sample size is calculated and considered. For example, if the plan is to recalculate the sample size at 50% recruitment, it is important to estimate where actual recruitment will be by the time a final decision about the new sample size is made. Starting and stopping recruitment can be problematic (e.g., staffing issues, lack of consistency in type of patient recruited); so it is important to consider a sample size recalculation with enough time to implement such a change without any interruption in enrollment, or before enrollment exceeds the newly calculated sample size. Sample size recalculation can be divided into two distinct approaches: one that involves nuisance parameters only and one that also involves an estimate of the treatment effect.21 Each approach is discussed next. Sample Size Recalculation Based Only on Nuisance Parameters The question is whether the values of variance, correlation, drop-out rate, or proportion of events in the

I. ETHICAL, REGULATORY, AND LEGAL ISSUES

WHAT TO MONITOR?

control group, which were assumed at the beginning of the trial, are consistent with what is actually observed so far; and consequently, whether the sample size calculated initially is still adequate based on these observed nuisance parameter values. From such an exercise, there are three possible outcomes: (1) the current sample size is adequate and therefore there is no need for change; (2) the sample size should be increased, unless cost is prohibitively raised; or (3) a lower sample size would be adequate. This last scenario occurs, for example, when the assumed variance at the design stage is higher than what is observed from the data. In this case, the appropriate decision is debatable. Some recommend to decrease the sample size, not to subject future participants to unnecessary risk, to save resources, and to publish results earlier. Others recommend to keep the sample size as initially planned, so that more data give more accurate results for both primary and secondary outcomes, as well as for safety and subgroup analyses; and since resources were approved for the initial sample size, there is no need to reduce it. They also argue that the nuisance parameters calculated midstream are themselves uncertain and may change when all the data are collected and analyzed, and may end up closer to what was initially planned. Recalculating the sample size based only on nuisance parameters does not involve any “statistical penalty,” since by definition nuisance parameters contain no information about the outcome of interest. Thus there is relatively little downside (other than additional analyst time) to including sample size recalculation in the initial trial design, unless the trial length or rate of recruitment does not allow for any sample size modification. Sample Size Recalculation Based on Nuisance Parameters and Observed Treatment Effect Here, we consider the case of whether the sample size should be changed based on the values of the nuisance parameters and the treatment effect observed so far. The reasoning motivating this kind of analysis is that if the observed treatment effect is smaller than initially anticipated, the current sample size will not provide enough power to detect the observed, smaller treatment effect; and consequently, the sample size should be increased. This is a controversial issue. Some believe that power analysis should be based on the minimum clinically meaningful treatment effect to be detected and not on the observed treatment effect regardless of how small it is. Criticism of performing this type of interim analysis relates to concerns about potential bias, loss of efficiency, and the possibility of increasing the sample size to detect clinically meaningless differences.21,23

135

Unlike the previous sample size recalculation exercise, recalculating the sample size based on nuisance parameters and observed treatment effect does involve a “statistical penalty.” This is because multiple looks at the treatment effect, just like multiple comparisons, do increase the chance of claiming statistical significance when in fact there is no difference. The nature of the statistical penalty is to make each of the repeated looks at the data at a significance level that is stricter than the usual 0.05 (i.e., 1) and protective associations (relative risk < 1). When a relative risk or OR is equal to 1, it means that risk of exposure (or odds of exposure) is the same in those with or without disease; that is, there is no association between disease and exposure. Both the relative risk and OR are measures of the strength of the association; they do not answer, for instance, how much disease may be prevented if we could eliminate an exposure.

243

The difference in the risks or incidence rates is simply that disease risk in those exposed or with a certain characteristic minus the disease risk in those not exposed. Sometimes we are interested in absolute risk, risk differences, or attributable risks. Attributable risk is the amount of disease that can be attributed to a certain characteristic or exposure. While many times attributable risk is presented as a proportion of the absolute risk in the comparison group, the goal is still to look at excess risk. If an observational study reports that people with a new strain of influenza (Group A) have 25% chance of death and people with a mutation (Group B) have 50% chance of death, we generally jump to the conclusion that Group B is twice as bad off. We intuitively jump to the relative risk. Few people will immediately say there is an OR of 3; while the OR is 3, few people take that from the simple data without a computer calculating the number. We will report the OR in the manuscript, though, and unfortunately, someone will see the OR and incorrectly interpret it that people are three times more likely to die in Group B. They might say the odds are three times higher but that too will be misinterpreted by the person reading the news and worried about dying. In fact the risk difference has an even different interpretation. Perhaps, the probability of being in Group A is 0.005% and the probability of being in Group B is 0.00001%. Should we be worried about dying from this and take every preventive step possible? Absolute risk would tell us no. If we or someone close to us dies, we likely care and wish the preventive steps were taken. If we are less close to the situation, we likely complain about all the measures being taken or worry needlessly that we are going to die. The concepts of relative and attributable risk are essential for policy and when prevention is being decided. They are not only used in observational studies but also are frequently used to interpret study data. Several common misinterpretations of the data are discussed in the next section.

MISTAKES, MISCONCEPTIONS, AND MISINTERPRETATIONS None of us are perfect, and we frequently make mistakes. Below we discuss a few of the common mistakes, misconceptions, and misinterpretations involving observational studies.

Always Trusting Bivariate Associations Based on Observational Study Data At times, our analyses ignore the richness of observations and simply compare two variables to each other. These are bivariate associations. An example of this is

II. STUDY DESIGN AND BIOSTATISTICS

244

17. DESIGN OF OBSERVATIONAL STUDIES

when we use chi-square tests and correlation as the only statistical analyses. While these tests are useful to explore data, they have severe limitations in observational studies. The correlation of two continuous variables evaluates if the two variables have a straight line association. Two variables may be strongly associated in a U or X shape or another shape and not have a strong correlation but have a very strong association. Additionally, if other variables are added to the modeling the association seen with a chi-square or correlation test may disappear, or what looked to be nothing becomes a strong association. We discuss methods to investigate associations between variables in Chapter 24. Since correlation and bivariate chi-square tests may be misleading because many univariate and bivariate associations disappear when a multivariate method such as regression is used, correlation and chi-square tests are not recommended for definitive analyses of observational data. They are useful for exploratory analyses and provide a simple way to present data.

Assuming Odds Ratios and Relative Risks Will Have a Similar Magnitude Relative risk can be computed in two possible ways. We have to define “risk of what?” A simple way to explain this is each time we can consider if a characteristic increases the probability of success; we also could have considered if the characteristic decreases the probability of failure. A small relative change in the probability of one event’s occurrence is usually associated with a large relative change in the event not occurring. Schulman and colleagues ran a controlled experiment and published an article reporting potential bias by physicians when recommending cardiac catheterization for patients with chest pain.29 A simplification of the data is presented in Table 17.3. Using the data in Table 17.3, we can compute the OR to be 0.57 or 1.74. The paper (using multivariate logistic regression) concluded that physicians make different recommendations for male patients than for female patients. Schwartz and colleagues30 wrote a critique and said the OR overstated the effect and that the RR was appropriate. Using the data in Table 17.3 the RR is only 0.93 (the reciprocal would be 1.07). The associated responses are quite illuminating and worth the time to read. Study Example Data

TABLE 17.3

No Catheterization

Catheterization

Total

Male

34 (9.4%)

326 (90.6%)

360

Female

55 (15.3%)

305 (84.7%)

360

Total

89

631

720

While ORs can be more easily adjusted for covariates, the relative risk may be clinically the more important association to consider here. Is it appropriate to look at 90.6% versus 84.7%? Yes. It is also appropriate to compare the rates for recommending a less aggressive intervention (9.4% vs. 15.3%) where the relative risk is 1.63 (reciprocal 0.61), quite a large value. Relative risks seem more intuitive but that does not mean that they are easier to interpret or even possible to calculate. Care must always be taken when interpreting results. In this case everyone was correct, more information could be provided, and interpretation was incomplete.

Misinterpreting Relative Measures The Schwartz piece lists several useful communication guidelines.30 It is hard to correctly interpret an OR. Many times, to be efficient with words, we misinterpret ORs as relative risks or risk ratios. Sometimes ORs can be converted to risk ratios, but this is tricky and should be done carefully by a statistician or epidemiologist. Another common problem is ensuring the appropriate comparison is made. Which group is in the denominator? Sometimes findings about one specific group are accidentally attributed to a broader group. Finally, sometimes we report relative rates when we could simply report absolute rates. Reporting a group is 50% less likely to receive treatment sounds useful, but reporting the actual percent treated in each group is more useful and improves the ability of the reader to draw conclusions.

Implying Causation (Even When We Do Not Mean to Do It) “The impact is huge, just look at the large OR/relative risk/absolute risk!” Focusing on the association’s magnitude or size and assuming causation is a common misstatement. In a study with multiple variables, there often will be several variables that appear to be related to the outcome or exposure of interest. We may think that a temporal sequence seems biologically reasonable. While several studies may demonstrate significant evidence that the effect of an exposure on an outcome is nonzero, observational studies, especially a single observational study, cannot allow us to assume causation. At most we can assess associations. Not everyone agrees with this and Chapter 27 will go in to more details about causal inference from observational data. Table 17.4 lists the evidence that researchers believe is necessary to “prove” that observational study evidence supports causation. The table is brief, but in short all of the evidence needs to point the same way, the observational evidence needs to be strong, clinically or

II. STUDY DESIGN AND BIOSTATISTICS

MISTAKES, MISCONCEPTIONS, AND MISINTERPRETATIONS

TABLE 17.4 1. 2. 3. 4. 5. 6. 7. 8.

Evidence in Observational Studies That Supports Causation

Statistical significance Strength of the association (odds ratio, relative risk) Doseeresponse relationships Temporal sequence of exposure and outcome Consistency of the association (internal “validity”) Replication of results (external “validity”) Biologic plausibility Experimental evidence

biologically meaningful and plausible, statistically significant, shows some type of dose response, and ultimately some experimental evidence is needed. This experimental evidence might be in animals or laboratory experiments to supplement observational data collected in humans. In short, it is not one study’s worth of evidence; an entire body of evidence from many studies conducted by many different groups is needed.

Confusing Causation, Prediction, Association, and Confounding Consider the following example. A mother’s genome has a causal relationship to her daughter’s height because the mother gives part of her genes that influence height to her daughter. However, if the mother also has a son, that son’s genome is merely associated with the daughter’s (his sister’s) height because each sibling receives part of their mother’s genes, but the son does not give any of his genes to the daughter (except in Greek tragedies). Finally, a mother’s genome and a nutrition program are confounded because their effects on a daughter’s height are mixed together. (An aside: randomized studies are not immune to confounding.) We may use the mother’s height as part of an algorithm to predict the height of her children; prediction is discussed in Chapter 27. Consider another hypothetical example. Suppose a team of researchers designs a cohort study to address the question of whether smoking causes premature death. They may construct two groups of middle-aged men (50e55 years old) who are smokers and nonsmokers, with 2500 subjects in each group. The subjects may be examined at baseline, followed prospectively and longitudinally, and their ages at death were recorded. Suppose the median time to death is 8 years earlier for smokers than for nonsmokers, and that this difference is statistically significant. Are the researchers justified in concluding from this study that smoking causes premature death? No. The tobacco companies can respond that smokers are inherently different from nonsmokers. Perhaps there are some genetic, socioeconomic, or behavioral factors that cause (or predispose)

245

people to smoke and that also cause them to die at an earlier age. Are the researchers, nevertheless, justified in concluding from this study that smoking is associated with premature death? Yes, that is the precise function of observational studiesdto propose associations. Was 50e55 the right age group for the study? Perhaps for seeing the events of interest and the study question, but it does exclude men who could have been in the study if they had not died prior to age 50. We may say it will be hard to generalize the information to other age groups, or we feel comfortable generalizing the results if the percentage of men dying prior to age 50 is small or the reasons well characterized.

Assuming Observational and Randomized Studies Never Agree A large set of observational studies led to a set of ambitious clinical trials and observational studies in women’s health. The Women’s Health Initiative (WHI) was launched in 1991 and consisted of a set of trials in postmenopausal women motivated by several prevention hypotheses.31 Women were enrolled into a randomized controlled trial (RCT) or an observational study. The hormone replacement therapy (HRT) hypothesis assumed women assigned to estrogen replacement therapy would have lower rates of CHD and osteoporosisrelated fractures. Progestin and estrogen were to be used in women with a uterus, and breast and endometrial cancers would be monitored. The hypothesized cardioprotective effects of HRT in postmenopausal women could not be proven in observational studies but had become widely accepted over time due to the adverse affects of menopause on the lipid profile. Prior to WHI, epidemiologic evidence, the majority of 30 observational studies, reported a benefit in ageadjusted all-cause mortality among estrogen users. Questions remained about the demographic profile associated with these observational studies’ participants: rather healthy and younger with little pertinent data on women beginning hormones after age 60 years; the use of the combination treatment estrogen plus progestin instead of unopposed estrogen; and the overall risk and benefit trade-off. Observational studies had noted a modest elevation in the risk of breast cancer with long-term estrogen use; however, adverse event data on progestin were inconsistent at the time. At the inception of the WHI, it was to be the study with the longest follow-up. The questions addressed in the clinical trials were posed based on epidemiological evidence. The diet portion of WHI also was based on epidemiological evidence.32 When approaching a RCT from a base of cohort studies, several important points must be addressed. If

II. STUDY DESIGN AND BIOSTATISTICS

246

17. DESIGN OF OBSERVATIONAL STUDIES

the motivation for a cohort study is to evaluate risks associated with a treatment or an exposure, then the study needs not only long-term users but also sufficient number of newly exposed participants to assess shortand long-term intervention effects. Time variation also must be taken in to account, and exposure effects may need to be evaluated over defined exposure periods.33 Confounding due to unmeasured factors has an important role in observational research, one that results in misleading observational studies. The estrogen plus progestin portion of WHI stopped early after finding estrogen plus progestin did not confer cardiac protection and could increase the risk of CHD, especially during the first year after the initiation of hormone use.34 The take-home messages were not simple yes/no answers and women were advised to talk with their doctors about their personal health and family history. Whereas some believed the WHI hormone therapy results were surprising, others did not.35e37 What the experience has taught us is to pay close attention to observational study design and analysis. We have to remember that a potential for publication bias, changes in populations under study, and incorrect analyses of prior study data (even if due to unavoidable circumstances) may lead to varying results. Additionally, populations shift over time. Hypothesis development, particularly in prevention which many times is based on cohort data, is vital and far more difficult than imagined at first glance. As researchers, we must always think ahead while planning our current study to the next several trials and studies that may result from our anticipated (and unanticipated) findings.

Trying to Design a Randomized Study When We Need an Observational Study Consideration of observational studies as alternatives to RCTs provides insight into the advantages and disadvantages of the latter from a scientific perspective. Suppose we want to study the effects of oral contraceptives on the risk of breast cancer over 30 years in women who began to use the pill in their early 20s. From the scientific perspective the ideal way to address this question would be through a clinical trial. The researchers would randomly assign women in the trial to either treatment or placebo groups and then follow them prospectively for 30 years and observe which group experiences more cases of breast cancer. Such a study would present many ethical challenges and also would prove impractical since it would be impossible to blind the subjects and researchers as to the treatment assignment, at least after the first pregnancy. From the ethical and practical perspectives, the best way to address some questions would be through an

observational study. An example of a cohort study: women would choose whether or not to use certain types of oral contraceptives in their early 20s, and the researchers would merely follow them prospectively and longitudinally over 30 years to observe who develops breast cancer. We would need to consider whether the group of women who chose to use the medication differed systematically from the women who chose not to do so. In a case-control study, researchers would construct groups of women in their 50s who had or had not developed breast cancer and then retrospectively look into the past to determine which women had used oral contraceptives and determine which other life events may have influenced the risk of breast cancer. We would need to consider how well the women selected for this study reflect the original population of women who began to use the pill in their early 20s. We also would need to consider if other risk factors were well understood and if the data had been reliably collected on all risk factors on all women in the study. We would have to realize that by sampling women in their 50s, we would likely miss early, aggressive cancers where women died before they reach age 50, and that our results would not be generalizable to all types of breast cancer. In a cross-sectional study, researchers would collect a sample of women in their 50s and then simultaneously classify them on contraceptive use and breast cancer occurrence. We again would need to consider how well we could use this sample to make inference back to the population of women who began to use the medications in their early 20s. Regardless of how well an observational study is constructed, questions may arise about applicability and unknown risk factors. Currently, different oral contraceptive pills, devices, and doses are used compared to 30 years ago, so can the previous study shed light on today’s women in their early 20s and their future health? Are there unknown or unmeasurable risk factors that might be playing a role in the study results? In observational studies, we can only control for known and measured variables.

Assuming an Observational Study Is “Safe” and Does Not Need External Monitoring Too frequently, investigators conducting observational research (and those overseeing it) assume it is minimal risk simply because of the lack of intervention, and they further assume that external oversight by a committee such as Data and Safety Monitoring Board would not be useful. However, many reasons exist to consider having an Observational Study Monitoring Board (OSMB) or similar external committee to look at the same elements described in Chapter 10 on data

II. STUDY DESIGN AND BIOSTATISTICS

247

REFERENCES

and safety monitoring. Regular external review of overall study conduct and data procedures, including protection of the confidentiality of participant data, can help studies have more reliable results. Such committees also provide recommendations related to overall scientific direction, proposed ancillary studies, participant burden, center performance and study progress, analyses and quality control, and issues related to referral for abnormal findings, informed consent, and safety. Trials with high risk, including risk to the public health if poorly performed, with multiple sites or registries, measures with potential safety concerns, and others should consider convening such a board. The US National Heart, Lung and Blood Institute has information online further describing responsibilities of an OSMB (https://www.nhlbi.nih.gov/research/funding/ human-subjects/data-safety-monitoring-faq#14).

CONCLUSIONS Observational studies are valuable alternatives, predecessors, and follow-ups to clinical studies. They may be used to chart the natural history or extent of a disease in a population. They may be useful in providing preliminary evidence of an effect, which ethics and practicality permitting can subsequently be studied with a well-designed RCT. They may be used to follow up changes in a population over time after the findings from a RCT are released. All studies have weaknesses; observational studies have the scientific weakness that they can be used only to find associations between risk factors and responses, but alone they cannot establish causation. That does not diminish their importance. Observational studies may seem easy to some clinical researchers. That is a mistake. It is hard to do good observational research. Because they are not controlled experiments, in observational studies many factors may be varying across subjects simultaneously, and hence we are required to measure many things in many different and potentially changing ways, very accurately, at times rather often to not miss fleeting changes in the data. We cannot lose any data, and we need to measure everything the same way every time. We will never know what we do not measure, and we cannot measure everything. Researchers doing experimental studies are supposed to do the same thing but try to fall back on introducing a single change and randomization to help determine causation and say that what was unmeasured should be balanced between groups. Good observational studies are extremely useful clinically and in research. A single case report may lead to the discovery of a new disease, and another might lead to a cure or a way to focus on a population most at risk for the disease. Without large surveys we cannot assess

large populations and compare where they are now to where they were in the past. Many associations cannot be otherwise studied, as a randomized study would not be ethical or feasible. These study methods are appropriately used not only in human studies but also are used by veterinarians, agriculturalists, police, and others. Good observational studies are vital to inform medical, public health, policy, and regulatory decisions.

QUESTIONS 1. Does epidemiology assume that human disease occurs at random? a. Yes b. No 2. Which of the following is most likely a case-control study? a. Report of five cases of pneumocystis pneumonia in previously healthy homosexual men b. National survey of health and nutrition c. Association study of maternal use of stilbestrol with tumor appearance d. Observational study of cardiovascular health in men and women over 65 3. For rare disease does the odds ratio (OR) estimate the relative risk (RR)? a. No b. Yes c. Depends

Acknowledgments The author would like to thank Jack M. Guralnik and Teri A. Manolio for their work on previous editions of this book and course materials over the years. She also would like to thank her many students over the years for their great questions, challenging research projects, and many examples.

Disclosures This chapter reflects the views of the author and should not be construed to represent FDA’s views or policies.

References 1. Altman DG. Practical statistics for medical research. Boca Raton (FL): Chapman & Hall; 1991. 2. Agresti A. Categorical data analysis. 2nd ed. Hoboken (NJ): Wiley; 2002. 3. Gordis L. Epidemiology. Philadelphia (PA): Harcourt Brace & Company; 1996. 4. Hulley SB, Cummings SR, Browner WS, Grady D, Hearst N, Newman TB. Designing clinical research. 2nd ed. Philadelphia (PA): Lippincott Williams & Wilkins; 2001. 5. Elwood M. Critical appraisals of epidemiological studies and clinical trials. 2nd ed. Great Britain: Oxford University Press; 1998.

II. STUDY DESIGN AND BIOSTATISTICS

248

17. DESIGN OF OBSERVATIONAL STUDIES

6. Fleiss JL. The design and analysis of clinical experiments. New York: Wiley; 1999. 7. Vandenbroucke JP, von Elm E, Altman DG, Gøtzsche PC, Mulrow CD, Pocock SJ, Poole C, Schlesselman JJ, Egger M, STROBE Initiative. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. PLoS Med 2007;4(10). 8. Friedman GD. Cigarette smoking and geographic variation in coronary heart disease mortality in the United States. J Chronic Dis 1967;20:769e79. 9. Joo JB, Cummings AJ. Acute thoracoabdominal aortic dissection presenting as painless, transient paralysis of the lower extremities: a case report. Emerg Med 2000;19:333e7. 10. CDC. Pneumocystis pneumonia, Los Angeles. MMWR 1981;30: 250e2. 11. Hedley AA, Ogden CL, Johnson CL, Carroll MD, Curtin LR, Flegal KM. Prevalence of overweight and obesity among US children, adolescents, and adults, 1999e2002. JAMA 2004;291: 2847e50. 12. Klungel OH, Kaplan RC, Heckbert SR, Smith NL, Lemaitre RN, Longstreth Jr WT, Leufkens HGM, de Boer A, Psaty BM. Control of blood pressure and risk of stroke among pharmacologically treated hypertensive patients. Stroke 2000;31:420e4. 13. Flegal KM, Graubard BI, Williamson DF, Gail MH. Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA 2007;298:2028e37. 14. Lilienfeld AM, Lilienfeld DE. Foundations of epidemiology. 3rd ed. New York: Oxford University Press, Inc.; 1980. 15. Strong JP, Malcom GT, McMahon CA, Tracy RE, Newman WP, Herderick EE, Cornhill JF, Pathobiological Determinants of Atherosclerosis in Youth Research Group. Prevalence and extent of atherosclerosis in adolescents and young adults: implications for prevention from the pathobiological determinants of atherosclerosis in youth study. JAMA 1999;281:727e35. 16. Newman AB, Naydeck B, Sutton-Tyrrell K, Edmundowicz D, Gottdiener J, Kuller LH. Coronary artery calcification in older adults with minimal clinical or subclinical cardiovascular disease. J Am Geriatr Soc 2000;48:256e63. 17. Schlesselman JJ. Case-control studies: design, conduct, and analysis. New York: Oxford University Press, Inc.; 1982. p. 17e9. 18. Sackett DL. Bias in analytic research. J Chronic Dis 1979;2:51e63. 19. Herbst AL, Ulfelder H, Poskaner DC. Adenocarcinoma of the vagina: association of maternal stilbesterol therapy with tumor appearance in young women. N Engl J Med 1974;284:878e81. 20. Plassman BL, Havlik RJ, Steffens DC, Helms MJ, Newman TN, Drosdick D, Phillips C, Gau BA, WelsheBohmer KA, Burke JR, Guralnik JM, Breitner JCS, et al. Documented head injury in early adulthood and risk of Alzheimer’s disease and other dementias. Neurology 2000;55:1158e66. 21. Doll R, Hill AB. The mortality of doctors in relation to their smoking habits: a preliminary report. Br Med J 1954;228(i):1451e5. 22. Doll R, Peto R, Boreham J, Sutherland I. Mortality in relation to smoking: 50 years observations on male British doctors. Br Med J 2004;328:1519e33. 23. Brackbill RM, Hadler JL, DiGrande L, Ekenga CC, Farfel MR, Friedman S, Perlman SE, Stellman SD, Walker DJ, Wu D, Yu S,

24.

25.

26.

27. 28. 29.

30.

31.

32.

33. 34.

35. 36.

37.

Thorpe LE. Asthma and posttraumatic stress symptoms 5 to 6 years following exposure to the World Trade Center terrorist attack. JAMA 2009;302:502e16. Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics 1988;44: 1049e60. Ridker PM, Hennekens CH, Miletich JP. G20210A mutation in prothrombin gene and risk of myocardial infarction, stroke, and venous thrombosis in a large cohort of US men. Circulation 1999; 99:999e1004. Roest M, van der Schouw YT, de Valk B, Marx JJM, Tempelman MJ, de Groot PG, Sixma JJ, Banga JD. Heterozygosity for a hereditary hemochromatosis gene is associated with cardiovascular death in women. Circulation 1999;100:268e73. Barlow WE, Ichikawa L, Rosner D, Izumi A. Analysis of casecohort designs. J Clin Epidemiol 1999;52:1165e72. Laurion JP, Troponin I. An update on clinical utility and method standardization. Ann Clin Lab Sci 2000;30:412e21. Schulman KA, Berlin JA, Harless W, Kerner JF, Sistrunk S, Gersh BJ, Dube´ R, Taleghani CK, Burke JE, Williams S, Eisenberg JM, Ayers W, Escarce JJ. The effect of race and sex on physicians’ recommendations for cardiac catheterization. N Engl J Med 1999;340:618e26. Schwartz LM, Woloshin S, Welch HG. Misunderstandings about the effects of race and sex on physicians’ referrals for cardiac catheterization. N Engl J Med 1999;341:279e83. Women’s Health Initiative Study Group. Design of the Women’s Health Initiative clinical trial and observational study. Control Clin Trials 1998;19:61e109. Prentice RL, Sheppard L. Dietary fat and cancer: consistency of the epidemiologic data, and disease prevention that may follow from a practical reduction in fat consumption. Cancer Causes Control 1990; l:81e97. Prentice RL, Pettinger M, Anderson GL. Statistical issues arising in the Women’s Health Initiative. Stat Med 2005;61:899e910. Women’s Health Initiative Investigators. Estrogen plus progestin and the risk of coronary heart disease. N Engl J Med 2003;349: 523e34. Prentice RL. Observational studies, clinical trials, and the Women’s Health Initiative. Lifetime Data Anal 2007;13:449e62. Prentice RL, Langer RD, Stefanick ML, Howard BV, Pettinger M, Anderson GL, Barad D, Curb JD, Kotchen J, Kuller L, Limacher M, Wactawski-Wende J, Women’s Health Initiative Investigators. Combined analysis of Women’s Health Initiative observational and clinical trial data on postmenopausal hormone treatment and cardiovascular disease. Am J Epidemiol 2006;163: 589e99. Prentice RL, Langer RD, Stefanick ML, Howard BV, Pettinger M, Anderson GL, Barad D, Curb JD, Kotchen J, Kuller L, Limacher M, Wactawski-Wende J, Women’s Health Initiative Investigators. Combined postmenopausal hormone therapy and cardiovascular disease: toward resolving the discrepancy between observational studies and the Women’s Health Initiative clinical trial. Am J Epidemiol 2005;163:404e14.

II. STUDY DESIGN AND BIOSTATISTICS

C H A P T E R

18 Design of Clinical Trials and Studies 1

Catherine M. Stoney1, Laura Lee Johnson2

National Institutes of Health, Bethesda, MD, United States; 2U.S. Food and Drug Administration, Silver Spring, MD, United States

O U T L I N E Design of Clinical Trials

Usual and Standard Care Controls Multiple Control Groups

250

The Purpose of Clinical Trials and Clinical Studies 250 Understanding the Spectrum of the Research Continuum Phase I Studies Phase II Studies Phase III Studies Phase IV Studies Dissemination and Implementation Studies Comparative Effectiveness Research Explanatory Versus Pragmatic Trials Quasiexperimental Studies

251 252 252 253 253 253 254 254 255

Clinical Trial Designs Crossover Designs Enriched Enrollment Designs Factorial Designs Parallel Groups Designs Sequential Trial Designs and Interim Analyses Group-Randomized Trial Designs Adaptive Treatment Designs

255 255 256 256 256 257 257 257

Critical Issues in Clinical Study Design Blinding or Masking Intervention Development Choosing the Comparison Group

258 258 258 258

Control Groups Wait-List Control Time and Attention Control Placebo Control Sham Control

258 259 259 259 260

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00018-6

260 260

Placebo Responses Background Identifying Placebo Responders

261 261 261

Mistakes and Misconceptions Not Looking at the CONSORT Statement Before, During, and After a Study Waiting Until the Large Definitive Study to Worry About the Details Failing to Increase the Treatment Effect Failing to Decrease the Variance Not Taking Care When Choosing a Control Group Always Assuming Placebo Groups Are Unethical Assuming Placebo Treatment Is (Im)Possible in Long-Term Studies Confusing Placebo Response and Regression to the Mean Using a Factorial or Partial Factorial Design Instead of a Parallel Group Design Assuming Small, Open-Label, Nonrandomized, Uncontrolled Studies Offer No Evidence

262

Conclusions

266

Summary Questions

266

Acknowledgments

266

Disclosures

266

References

266

Further Reading

268

249

262 262 263 263 263 264 264 264 265 265

Copyright © 2018. Published by Elsevier Inc.

250

18. CLINICAL TRIAL DESIGNS

DESIGN OF CLINICAL TRIALS Randomized controlled trials (RCTs), when correctly designed and rigorously conducted, provide the most definitive answers regarding intervention effects, but other clinical trial designs and observational investigations can be appropriately employed depending on resources and the specific questions of interest.1 While nonrandomized experimental studies may inform the potential promise of clinical trials, it is the RCT that provides the strongest evidence both for the causal nature of a modifiable factor and for the impact on disease outcomes in modifying that factor. The essential purpose of the RCT is to determine whether a particular intervention or treatment can reasonably be inferred to cause a change in health, disease progression, or risk factor(s) associated with a disease. The nature of the interventions could be pharmacologic, surgical, behavioral, device, strategy-based or could consist of multiple components. Randomization is used as part of many studies, but when people refer to RCTs they usually are focusing on a portion of randomized studies that are large and expected to have a hopefully definitive result. When such studies are undertaken depends in large measure on the existing literature and needs in the area. While there may be frequently promising observations that suggest a clinical trial is warranted, other essential factors must be established before undertaking an RCT. These might include such things as development of an intervention or treatment, which has been shown to alter some relevant outcomes, determination of the appropriate dose of the intervention, establishment of feasibility in a particular patient population, showing evidence for changes in surrogate markers of a disease of interest, and gathering other essential information necessary for optimizing design. A good understanding of the existing literature in the area of interest, the strength of the designs and findings of the literature, and clarity regarding where the bulk of science for a particular intervention or treatment sits on the research continuum are essential for formulating an appropriate question and design for a clinical trial. A note to readers: all of the study designs and related issues discussed can be used in many types of experiments. This chapter focuses on studies in humans related to disease, but the general principles apply in other settings as well. Careful planning is important to avoid errors and misleading results.

THE PURPOSE OF CLINICAL TRIALS AND CLINICAL STUDIES Clinical trials provide a unique source of evidence. As discussed in Chapter 17, observational or epidemiologic

studies are critical for hypothesis generation, discovering associations, identifying areas of health research that are “low-hanging fruit,” and providing evidence when a clinical trial is not feasible or ethical. But clinical trials are unique because they are experiments that allow researchers to infer causality. While consistency in the findings of a large number of observational studies can lead to the belief that the associations are causal, this belief is a fallacy. One of the best examples of this phenomenon was demonstrated with the results of the Women’s Health Initiative (WHI), which was launched after a large number of observational studies suggested that hormone replacement therapy (HRT) in postmenopausal women was cardioprotective. The findings across many observational and epidemiological studies at the time were considered consistent in large part. Most studies reported cardiovascular benefits in age-adjusted all-cause mortality for estrogen users. Consequently, HRT was in widespread clinical use for symptom management among peri- and postmenopausal women and, in some cases, for cardiovascular protection. The WHI was launched in 1991 and consisted of a set of clinical trials in postmenopausal women to definitively test the effect of HRT on cardiovascular risk and osteoporosis-related fractures, in addition to a new set of observational studies. Previous observational studies had noted a modest elevation in the risk of breast cancer with long-term estrogen use; however, adverse event data on progestin were inconsistent at the time. The estrogen plus progestin portion of WHI stopped early after finding estrogen plus progestin did not confer cardiac protection and in fact could increase the risk of coronary heart disease (CHD), especially during the first year after the initiation of hormone use.2 These findings were in direct contrast to the interpretations of many, but not all, epidemiological and observational study findings from the previous 30 years, and underscored the importance of the clinical trial in establishing causality. Also important is that the WHI was conducted in diverse populations of interest to investigators, but these populations were different than many of the previous epidemiological studies. Trials are not only typically conceptualized by design characteristics but also can be categorized according to the objectives of the investigation. Treatment trials are quite common, and they are designed to test specific interventions. The main objective is to understand the efficacy and effectiveness of an intervention on a specific outcome or set of outcomes in a specific patient population. Prevention trials aim to test ways to not treat, but rather prevent, diseases. Primary prevention trials are focused on preventing initial occurrence of a disease; secondary prevention trials many times are focused on preventing recurrence or exacerbation of a disease in asymptomatic persons with risk factors or positive

II. STUDY DESIGN AND BIOSTATISTICS

UNDERSTANDING THE SPECTRUM OF THE RESEARCH CONTINUUM

screening; tertiary prevention trials are focused on those who have symptomatic disease in an attempt to prevent further deterioration. Screening trials test strategies for identifying the detection of disease or risk for a disease. Diagnostic trials examine strategies for the correct diagnosis of a disease or disorder. Health-related quality of life studies are focused on strategies to promote quality of life, either within the context of a disease or treatment or by itself. Comparative effectiveness trials are concerned with identifying which of two or more treatments are superior in some way. The strength of the well-designed RCT is its ability to establish causality, thus overcoming the major weakness of all other types of study designs. However, it is important to note that the RCT design is sometimes not feasible for some clinical questions. Ethical considerations, costs, resources, or time may all prove prohibitive in certain cases, making the conduct of an RCT not feasible or optimal. For example, today a longitudinal RCT studying the health effects of cigarette smoking on adolescents into adulthood would not be feasible because it would be unethical to randomize adolescents to a smoking condition. A question about what is ethical may change over time, however. While previously considered unethical in some cases and countries, randomized trials of pregnant women and infants have become increasingly important to provide evidence-driven recommendations. We know that new does not always equate to better, and sometimes the more efficacious and safer study arm to be on is the placebo or active control arm. The International Conference on Harmonisation E10 guideline discusses choice of control group and related issues, and it also clarifies that without well done randomization there is a recognized inability to control bias, and the resulting consequences of potential persuasiveness of the findings are deeply problematic.3 It is unethical to conduct a study that will not, due to its design alone, result in interpretable findings. Whenever possible, the RCT is the strongest study design for establishing causality.

UNDERSTANDING THE SPECTRUM OF THE RESEARCH CONTINUUM The most optimal design, analytic strategy, and end points for any research study are dependent, more than anything else, on both what question is being addressed and where that specific question fits on the spectrum of the research continuum. Failure to consider these factors can lead to violations of both internal and external validity. External validity refers to the extent to which a research finding can be generalized to other situations, individuals, measurement instruments, and across time. Internal validity refers to the strength to which the independent variable, many times a treatment

251

or exposure but, in general, whatever is being manipulated or changed, can be said to be responsible for the outcome or change. Factors that decrease the generalizability or that compete with the independent variable to explain the findings are termed threats to validity. Very early phase investigations may be hypothesis generating and, while sometimes randomized, are most typically observational and epidemiological designs. Exploratory studies are most appropriate in treatment development and do not require rigid control, but are far from definitive and allow few conclusions to be made. Such studies are useful in establishing the possibility of a signal that the intervention may have an effect on the outcome of interest. Phase I clinical studies often follow exploratory studies and are primarily focused on questions of mechanism and safety. Because of the importance of high internal validity in questions of safety, Phase I studies require tight experimental control. Midphase investigations, such as Phase II (efficacy) RCTs, are hypothesis confirming and require control of known and unknown sources of error to minimize violating internal validity. Later phase studies, such as Phase III (effectiveness) and Phase IV RCTs and implementation trials, are more translational in nature and are primarily concerned with whether a particular intervention can be effective when implemented in increasingly real-world settings. These latter trials, while still requiring control of some sources of error, are more concerned with whether a particular intervention can be effective across situations outside of the tightly controlled setting of an earlier phase RCT. For example, late-phase trials may address questions of whether findings of a Phase II trial can generalize across various patients and health-care provider situations as they exist in the real world. Thus, the type of control most important for this phase of study is control for external, rather than internal, validity. Understanding not only the specific question that one wishes to address but also the state of the science in that particular area is critical to developing an appropriate study design for that particular question at a particular point in time. Inherent in this discussion is that the first stage in developing an appropriate design and analytic strategy for any given study is being able to clearly articulate the precise research question to be addressed. The more carefully and fully the research question can be expressed, the clearer the choices of design, end points, control groups, and analytic strategies become, and various elements of how to develop and conduct the study under question are clarified. It is important to be highly conversant in the relevant literature in the area to understand not only the questions that have been addressed and answered but also how they are addressed, what is known, where the gaps are, and where the knowledge base sits.

II. STUDY DESIGN AND BIOSTATISTICS

252

18. CLINICAL TRIAL DESIGNS

Phase I Studies Phase I studies, which include dose-ranging and safety studies, traditionally (but not always) are nonrandomized. The fundamental goal of these studies is to find appropriate dose levels and to detect potential toxicities due to the investigational intervention or treatment. When feasible, a dose-limiting toxicity (DLT) threshold or physical event must be defined to create a stopping rule. Usually the definition of a DLT is based on criteria, such as a certain grade of toxicity as defined by the National Cancer Institute Common Toxicity Criteria for Adverse Events (NCI CTCAE). For interventions that do not result in toxicity regardless of dose, investigators must establish criteria for toxicity or define something other than toxicity for the DLT. It is important to note, however, that dose-ranging studies are no less important for these types of interventions. For example, the appropriate dose of a particular psychotherapy for the treatment of major depressive disorder (MDD) is important to determine, and such a dosing study needs to establish safety, stopping rules, and specific definition of toxicity for that particular intervention. In this example, toxicity might be conceptualized as diminishing return in relationship to measured side effects and patient burden. Different interventions, study designs, and patient populations will require different stopping rules, especially since some treatments are nontoxic at all dose levels.3e5 In all cases, however, carefully defining the DLT prior to the onset of the dose-ranging study is imperative. Classically for pharmacologic treatment studies, a few dose levels or categories are selected and a small number of participants are treated at each dose level, typically escalating through the dose levels in the following manner. A few participants are enrolled at the lowest dose level in the protocol, and if none develops a DLT then the study escalates to the next dose. If a DLT is observed in one of the participants, then a few additional participants are enrolled, such that now all enrolled participants will receive the current dose. If none of the additional participants develops a DLT, then the study escalates to the next dose. Participants are not entered at a new dose level until all participants in the previous levels remain free of toxicities for a specified period of time. In addition, the maximum tolerated dose (MTD) usually is defined as the dose level immediately below the level at which a certain percent, usually 33% (which may be as low as two study participants) experienced a DLT. Usually, the study aims to find a safe dose defined as the MTD or the study finishes at the maximum dose that is prespecified in the protocol. There are many variations on this type of study.6,7 For nonpharmacologic treatment studies, dose-ranging investigations are less common and instead doses often

are chosen on the basis of feasibility, cost, typical practice (for those interventions which are in active clinical use but not empirically established as efficacious), and patient burden. For drug and nondrug studies, doses are sometimes determined as part of a Phase II trial. Whenever dose ranging is determined, it is critical for establishing safety and especially for identifying the optimal treatment intensity. Failure to include doseranging studies, including in nonpharmacologic investigations, can lead either to the premature conclusion that a treatment is not efficacious (if the dose is too low) or that a treatment is efficacious but not accessible (due to cost, burden, etc.). Regarding safety, the use of this methodology from pharmacology treatment studies to broader contexts has resulted in the definition of MTD also to be expanded to include nontoxic but nevertheless undesirable events (i.e., expanded definition of toxicity). Randomized Phase I studies are also conducted for pharmacologic and nonpharmacologic interventions. These studies may include a small number of participants receiving a control intervention at each dose level.

Phase II Studies The early screening of new therapies in the past often was conducted with single arm or nonrandomized studies. In treatments for cardiovascular disease, for example, Phase II studies often were conducted in which patients were treated and their responses (not necessarily a clinical response but often responses on a biomarker) were observed. The purposes of these studies was not to prove that a new therapy was efficacious on the ultimate clinical end point of interest (e.g., survival) but only that it has sufficient activity (e.g., reduction in a clinically relevant biomarker) to be tested in a larger RCT. These types of early Phase II studies provide the necessary signal that some activity is occurring to justify larger and more expensive trials. These, as well as earlier phase studies, are often referred to as proof of concept studies. Randomized or not, early Phase II designs typically require a relatively small number of patients; when the evidence shows that the benefits of the new therapy are small or nonexistent, the designs prevent large numbers of patients from being exposed to useless or even potentially harmful treatments. The disadvantage of the nonrandomized and open-label strategies sometimes considered for Phase II is that there is less experimental control imposed than is optimal. Thus, internal validity is sacrificed and can result in prominent placebo effects, investigator and other sorts of bias due to lack of masking, regression to the mean, and other threats to internal validity.17 For example, an early Phase II treatment study of patients with relapsing/

II. STUDY DESIGN AND BIOSTATISTICS

UNDERSTANDING THE SPECTRUM OF THE RESEARCH CONTINUUM

remitting multiple sclerosis (RRMS) might screen participants for moderate magnetic resonance imaging (MRI) activity and then follow them longitudinally. Because the natural process is relapsing/remitting, there is the potential that patients will be screened when in a relapse and naturally move into a remitting phase over time. Consequently, one may see a reduction in disease activity over time, even if the experimental therapy is ineffective. This design would be improved with a more rigorous design, incorporating an appropriate control group and perhaps multiple measurements. However, this would require designing a larger and more expensive trial with little evidence that the experimental treatment might be effective. Thus, early Phase II studies play an important role in helping to direct the next steps in a research program, but the findings should not lead, without many additional subsequent and more rigorous studies, to changes in clinical practice, guidelines, or health-care policy. Another early Phase II design is an external or historical control study. These studies have been used in rare diseases and cancer research.8 Here, instead of creating a randomized comparison group, a single group of patients is treated, and their responses are compared with controls from previous studies or a registry. These studies have the advantage of using only half the number of patients, and all the current study subjects receive the new therapy. They also have serious disadvantages. In addition to the other problems with a nonrandomized study, these controls often do not provide a good comparison with the new treatment patients. For example, controls often are taken from studies conducted years ago for conditions with rapidly changing profiles and treatment strategies; this is especially problematic because diagnosis, treatments, technology, and patient care can change over even relatively short periods of time. Measurements that may be present in the intervention arm may not exist in the control’s available data or not on the time schedule used by the intervention arm. In addition, the patient population characteristics may change. These changes, which are often not recognized or reported, can result in serious biases for assessing treatment efficacy. All of these designs may consider employing a version of an optimal two-stage design for statistically controlled interim analyses,9e11 but randomized study designs are usually preferred. Randomized Phase II clinical trials can avoid the problems of the nonrandomized studies described previously, employ greater control, and enable stronger causal inferences to be made. These designs aim to select the superior treatment arm from two or more arms. Some designs use a binary outcome describing failure or success along with statistical selection theory to determine sample size, and others rely on continuous or ordinal outcomes. Some are based on frequentist

253

methods and others on Bayesian methods. The goal is to have a low probability of choosing an inferior arm out of the total number of arms of the study. In the randomized Phase II design, an interim analysis to stop the study early for sufficient biological or clinical activity is possible, allowing the agent to be moved forward more quickly to another, larger study. This is in contrast to the optimal stage 2 design, which cannot be stopped early for efficacy. Thus, it is important to choose the design that is best for the study question.12

Phase III Studies Phase III studies, which can be efficacy or effectiveness studies, are large prospective trials designed to compare an experimental and control (or standard) intervention. Phase III trials can be designed to demonstrate superiority, noninferiority, or equivalence. They are typically longer in duration than Phase II trials and employ less control on participant characteristics, delivery of the intervention, and characteristics of the study environment. Phase III studies are more “real world” than are earlier phase studies, and while internal validity is lower, external validity is much higher. As with Phase II trials, interventions tested in Phase III trials include drug, behavioral, devices, surgical procedures, and more. Phase III trials may be used for many types of investigations, including evaluating an intervention for the purposes of treatment, prevention, or diagnosis.

Phase IV Studies Phase IV studies are large, generally postmarketing surveillance studies of population safety, effectiveness, and generalizability. Phase IV clinical studies are designed to study longer-term effects of treatments on populations.

Dissemination and Implementation Studies Dissemination and implementation studies typically are carried out after interventions or treatments have been identified as effective or efficacious in reasonably sized studies. These studies are essential to the process of translating new information to public health use. Although the terms dissemination and implementation are sometimes used interchangeably to refer to all translational science, the terms are distinct and refer to specific steps in the translation of evidence-based practices into general use. Implementation science tests ways of integrating proven treatments and practice patterns into specific contexts and settings. Dissemination science tests ways of informing health-care providers and the public regarding evidence-based treatments. Both

II. STUDY DESIGN AND BIOSTATISTICS

254

18. CLINICAL TRIAL DESIGNS

types of research are important steps in evidence-based research. The main focus of dissemination and implementation studies is to test ways of taking treatments that are efficacious or effective and evaluate their uptake, reach, sustainability, and spread in real-world settings. Rather than solely providing evidence of treatment efficacy or effectiveness, dissemination and implementation studies identify ways of implementing treatments in real-world settings, such as schools, worksites, primary care settings, and community clinics, which have great variability in clinical practice and patient characteristics. Outcomes of this type of research include training programs, outreach programs, cost-effectiveness and cost analysis, uptake and spread by providers, and adherence by both the clinical practice and patient communities. For many, although not all types of interventions, measures of implementation barriers and facilitators are incorporated into all phases of clinical research. In particular, it is useful to carry out investigations to better understand training needs, how the proposed treatment can (or will not) be integrated in to people’s lives, practice patterns and settings, and similar information in preparation for Phase III trials.

Comparative Effectiveness Research Comparative effectiveness research is a strategy that usually compares two or more evidence-based treatments, strategies, or diagnostic procedures on specific outcomes for clinical benefit and risk. Since the main focus is on two or more active and effective treatments, often “no-treatment” control groups are excluded. The ultimate purpose is to inform public policy, coverage policy, clinicians, and patients. Because treatments with empirically demonstrated efficacy and effectiveness are ideally the main focus of comparative effectiveness research, these types of studies are typically conducted at a late phase of research in a particular area. Comparative effectiveness research does not consist of a particular design strategy per se, but rather describes the goals of identifying what particular treatment is most effective for specific patients and conditions. This type of research emphasizes the value of understanding individual differences in response, and strives to identify optimal ways of tailoring treatments and procedures for specific patient needs. Although RCTs can be effectively employed to answer many comparative effectiveness questions, many other strategies are available that employ different methods, assumptions, and models.13 For example, systematic reviews, nonrandomized studies using data from electronic medical records or registries, and metaanalytic techniques are often more cost-effective than large RCTs, can sometimes answer questions regarding how best to tailor treatments to specific patient needs, and

often can be carried out efficiently if appropriate data exist and quality methods are followed. These techniques are discussed in part in Chapters 19e22.

Explanatory Versus Pragmatic Trials Schwartz and Lellouch14 characterized two different purposes of clinical trials, explanatory and pragmatic, articulating one of the most useful distinctions for the design of clinical trials. An explanatory trial is conducted under ideal circumstances, is generally highly controlled, and often seeks to elucidate a biological mechanism or establish idealized efficacy. The study population is relatively homogeneous and can be used as a model from which one may learn principles of pharmacology, physiology, or behavior that are likely to shed light on a variety of clinical issues. Explanatory clinical studies are focused on generally smaller populations and targeted outcomes to understand mechanisms of an intervention and maintain rigorous control over experimental factors. The findings are not highly generalizable to other settings, populations, and practitioners. They are particularly useful in understanding whether an intervention is efficacious and can inform larger effectiveness trials. A pragmatic trial approach focuses on whether an intervention, when practiced in real-world settings, demonstrates the expected benefit in a variety of unselected patient populations. Pragmatic trials usually focus on how an intervention shown to be efficacious in an explanatory trial works in real-world settings. Pragmatic trials typically focus on larger, more heterogeneous populations, with designs that have less control over the specific interventions and how and when they are delivered. Pragmatic trials are most useful for understanding generalizability of effects and for helping to inform health-care policy and decision-making. While an oversimplification, the dichotomous explanatory versus pragmatic reason for conducting a clinical trial may provide a useful perspective for making design choices in complex cases. Neither approach is superior to the other, but it is important to understand the advantages of each and to know which circumstances are most appropriate for which type of trial. Understanding the basic purpose of the trial will have a significant impact not only on the question addressed but also on nearly all aspects of the study design. Often these questions are addressed at different points in the research continuum. Useful tools for helping in the design of trials and in identifying where studies fall on the exploratorye pragmatic continuum are the PragmaticeExplanatory Continuum Indicator Summary (PRECIS) and PRECIS-2 tools. Originally developed by 25 clinical trialists,15 the original PRECIS tool was modified in 201516 to improve

II. STUDY DESIGN AND BIOSTATISTICS

255

CLINICAL TRIAL DESIGNS

reliability and validity. The PRECIS-2 tool allows investigators to score potential (or existing) studies on nine domains, indicating where the design falls on the explanatoryepragmatic spectrum for each domain. Domains include eligibility criteria, recruitment, setting, organization, flexibility in intervention delivery, flexibility in prompting adherence, follow-up, primary outcome, and analysis. The purpose is to assist investigators in matching the decisions made at the design stage with the desired purpose of the trial and to evaluate published trials.

Quasiexperimental Studies Quasiexperimental designs are those design strategies that do not contain all the essential elements of a true experiment but can be made powerful enough to make strong causal inferences by controlling potential threats to internal validity.17,18 Typically, quasiexperimental designs are employed because the investigator is not able to randomize study arms. When a treatment of interest cannot be randomized or a control group is not feasible, other factors can be incorporated into the design to increase internal validity. Quasiexperimental studies often are conducted in more naturalistic settings; these designs may exploit natural phenomenon such as investigating the effects of health outcomes after an earthquake or other natural disaster. Such designs are also used commonly in studying behavioral interventions in the context of real-world settings.18 In conducting such studies, the key responsibility of the researcher is to carefully think through all sources of _potential bias and threats to validity and impose appropriate and available design and analytic strategies to control such threats as much as possible. In the examples above, incorporating an available control (nonexposed) group to the study is a way to improve and strengthen the design. In some cases, it may be possible to access information of the exposed population prior to the disaster or intervention. An example is a study of the health effects of the 1980 earthquake in southern Italy.19 The affected population had been involved in a longitudinal epidemiological investigation of major CHD risk factors prior to the earthquake. At the time of the earthquake, a portion of the population had been examined in a follow-up health screen and the other portion had not. Within 2 weeks following the earthquake, health screenings were resumed for the group that had not yet been followed up. Both groupsd the exposed (screening after the earthquake) and unexposed (screening prior to the earthquake)dwere compared on a variety of cardiovascular disease biomarkers such as heart rate, cholesterol, and triglycerides. Although randomization was not possible, some control was (serendipitously) available to enable the researchers

to make some fairly strong causal inferences about the differences between the two groups attributable to exposure to the earthquake. The investigators found short-term elevated cardiovascular risk factors in the group exposed prior to data collection; subsequent follow-up data collected 7 years after the earthquake demonstrated that these elevations did not persist.

CLINICAL TRIAL DESIGNS Choosing an appropriate study design depends on whether the design matches (i.e., will be able to answer) the question that is being posed, the specific end points of interest, and whether the question being posed is the most appropriate one, given where the state of science stands on the research continuum for that particular question. In general, simple designs with a targeted and well-characterized question, clearly defined end points, and patient characteristics that will allow a clear and definitive answer are optimal.

Crossover Designs In crossover designs, each study participant receives all treatments that are being investigated but at different times. The order in which a study participant receives the treatments is randomized. For example, patient A is randomized to receive Treatment #1 for a period of time. After completing Treatment #1, the patient then “crosses over” and receives Treatment #2. Usually between treatments is a period of time called a washout when no treatment is delivered. Outcomes are examined during and/or after each treatment. In some crossover designs, particularly ones with more than two treatments, patients may not receive all treatments under investigation (partial crossover or incomplete block) but would receive more than one. The advantage of such a design is that each patient serves as his or her own control and this significantly reduces between-subject variability, allowing the detection of smaller effect sizes with reduced sample sizes. Crossover designs for the right patient population and treatment can therefore have considerably more power than other design strategies because of the reduced variation in nonspecific (nontreatment-related) factors. Crossover designs can be particularly advantageous for studying patients with conditions that have symptoms, which are relapsing/ remitting or episodic in nature if the relapsing/remitting cycles are short, such as migraine and functional pain disorders.20 The major disadvantage of crossover designs is apparent when the treatment being investigated has a sustained effect on the outcome of interest. In the example above, if Treatment #1 has an effect that is maintained long after the treatment is over, then in a

II. STUDY DESIGN AND BIOSTATISTICS

256

18. CLINICAL TRIAL DESIGNS

crossover study the impact of Treatment #2 cannot be clearly separated from Treatment #1; a crossover design in such a situation would be problematic. Such carryover effects are problematic because they sometimes cannot be seen or measured. In addition, crossover designs may be unethical in some circumstances. For example, if one of the treatments is efficacious and the condition is a serious threat, it may be unethical to cross participants in the efficacious treatment arm to the other, potentially nonefficacious treatment arm. Another potential disadvantage to this design is that patients must be enrolled in crossover studies for longer periods of time than are typical. This induces added burden on the patient and could result in greater dropout rates. For conditions that are episodic, crossover designs can be useful if the episodic nature is somewhat predictable. However, for unstable or progressive conditions, crossover designs may be problematic because added variability due to change in disease will be introduced over the course of the study. Finally, in addition to greater patient burden because of the length of the study, crossover trials often can be unpalatable to patients because they involve participation in two or more interventions rather than just one. However, for treatments that are perceived by patients as potentially equally efficacious, crossover designs can be appropriate, particularly for conditions that have stable or predictable clinical characteristics over time. A variation on crossover designs is n-of-1 studies.

Enriched Enrollment Designs A variant of the crossover design, the enriched enrollment design, may be useful in studying treatments to which only a minority of patients respond.21 If the results are not statistically significant in a conventional clinical trial but an intervention appears effective for subpopulations of patients, it is not possible to retrospectively point at the responders and claim that the treatment accounted for their relief. A potentially useful strategy is to enter responders into a second prospective comparison trial. If the results of the second trial considered alone are statistically significant, this suggests that the patients’ initial response was not just due to chance. Although at times statistically defensible, enriched enrollment designs are open to the criticism that prior exposure to the treatment may defeat a double-blind procedure (particularly with treatments that have distinctive side effects) and sometimes result in spurious positive results. Another caveat is that positive results from an enriched population of responders can no longer be generalized to the entire patient population, but rather just to a subpopulation of similarly defined responders. Enriched enrollment studies may

be of interest in treatment intervention studies because they demonstrate some limited evidence for a treatment response22 and may therefore suggest further investigation.

Factorial Designs In a factorial design, each level of a factor (treatment or condition) occurs in combination with every level of every other factor. Experimental units are assigned randomly to treatment combinations rather than individual treatments. The Fourth International Study of Infarct Survival23 was a large, multisite RCT designed as a factorial study with three treatments: oral captopril, oral mononitrate, and intravenous magnesium sulfate. The purpose of the trial was to assess the effectiveness of one of these three treatments among patients with suspected myocardial infarction (MI) on 35-day survival. Each of the three treatments could be delivered at one of two levels (e.g., placebo, standard dosage). Therefore, for this study there are eight (2
2
2 ¼ 8) possible treatment combinations. Each patient was randomized to one of the eight combinations with a probability of 1/8 (12.5%). Classically, each intervention should have independent effects; in other words, there is no interaction between any of the interventions. However, this assumption often is not valid. In such cases, a parallel arm study may be appropriate because every treatment combination is tested on a different group of participants, enabling an estimate of interactions or synergistic effects between various treatments on the response (e.g., 35-day mortality). A major challenge of factorial designs is to (1) meet the independence assumption or (2) choose a sufficiently large sample size to be able to detect meaningful interactions with high power or a good statistical chance of seeing an interaction, if it in truth is present. The main reason factorial designs are used is to examine multiple hypotheses with a single study. For example, the ISIS-4 study was designed to simultaneously examine the role of three treatments in reducing 35-day mortality in treating acute MIs. Designing a factorial study saved resources compared to designing three separate parallel group studies for each of the experimental treatments. Note that if some particular treatment combinations are not of interest, a partial or fractional factorial design that omits the less interesting combinations may be used.

Parallel Groups Designs In parallel group designs, participants are randomized to one of several possible treatments. Interest focuses on comparing the effects of the treatments on a common response or outcome. One of these groups

II. STUDY DESIGN AND BIOSTATISTICS

CLINICAL TRIAL DESIGNS

may be a placebo group (a group assigned to a placebo pill) or a control group (a group assigned to a standard or an alternative treatment). The effect on the response could be adjusted for baseline measurements of patient characteristics. A clinical trial of felbamate monotherapy for the treatment of intractable partial epilepsy was conducted with a parallel groups design, with two groups.24 The response in this trial was the average daily seizure frequency over the 2-week follow-up period. Furthermore, the double-blind randomized parallel groups design is the “gold standard” to which all other designs should be compared. It is the ideal study design to arrive at a definitive answer to a clinical question and is often the design of choice for large-scale definitive clinical trials. The challenge of this design is that it often requires large sample sizes and thus requires large amounts of resources.

Sequential Trial Designs and Interim Analyses In sequential trials, the parallel groups are studied not for a fixed period of time but, rather, until either a clear benefit from one treatment group appears or it becomes highly unlikely that any difference will emerge. These trials tend to be shorter than fixed-length trials when one treatment is much more effective than the other treatments. In group sequential trials,25 the data are analyzed after a certain proportion of the observations have been collected, perhaps after one-fourth, one-half, and threefourths of the expected total number of participants or events, and once more at the end of the study. Data analyses of the primary outcome variables during a study are called interim analyses. Group sequential trials are easier than sequential trials to plan regarding duration and resources, and they also can be stopped early if one treatment is much more effective than the others. All trials must have a mechanism for stopping early if evidence of harm due to the treatment emerges. Trials also may be stopped for futility, where futility is defined as the unlikelihood that a positive treatment effect will emerge as the result at the end of the trial. Chapter 27 covers this topic in more depth. It is particularly important to have statistical expertise on the team for designing, conducting, and interpreting interim analyses for efficacy or futility.

Group-Randomized Trial Designs Group-randomized (also known as cluster randomized) trials are those trials where the unit of randomization is one of the several types of groups rather than an individual. Such groups might include schools, clinics, worksites, communities, or other units. Group

257

randomization to treatment can be an efficient strategy when an intervention is difficult to implement on an individual level without the risk of contamination, such as interventions that affect environments. Consider, for example, an intervention seeking to change eating patterns by altering supermarket environments. Because the treatment cannot be delivered to individuals separately, group randomization of communities utilizing those supermarkets would be appropriate for such a study. Although the members of the groups are the individual units, which are observed and measured, the number of groups randomized to each condition is typically small because they often contain large numbers of members. The small number of randomized groups introduces a greater potential for threats to internal validity to operate (because randomizing a small number of groups is less likely to control potential bias), which is one major disadvantage of grouprandomized trials. Some of these threats can be decreased by utilizing appropriate analytic strategies, adherence to stringent design strategies, anticipating and measuring potential confounding variables, and by increasing retention rates. Another disadvantage of this design strategy is the need for access to adequate numbers of groups. For a study of changes within a health-care system, for example, conducting a grouprandomized trial on 30 or more health-care systems can be a significant and expensive undertaking,26 although costs can sometimes be mitigated if data from electronic health records or other available resources can be appropriately accessed.

Adaptive Treatment Designs Adaptive treatment (also called adaptive intervention or stepped-care or dynamic treatment) trials are designs that allow changes in the dose or components of an intervention after the onset of the study, as a function of individual or environmental factors or characteristics. Decision rules are established before the onset of the study regarding the characteristics of interest (e.g., gender, outcome of interest) and how they will determine assignment to specific intervention components or dose, and individuals can be randomly assigned to condition several times. These designs, when carefully constructed, can be efficient and cost-effective and are increasingly used because they allow development of individually tailored treatment strategies. One effective tool for developing adaptive interventions is the use of the Sequential, Multiple Assignment, Randomized Trial (SMART) design27, which facilities data-based planning of decision rules in the adaptive design (see Intervention Development section below).

II. STUDY DESIGN AND BIOSTATISTICS

258

18. CLINICAL TRIAL DESIGNS

CRITICAL ISSUES IN CLINICAL STUDY DESIGN

behavioral strategies that are maximally potent while also efficient by allowing identification of component parts.

Blinding or Masking Blinding or masking investigators, study staff, and participants to treatment condition, when possible, may be almost as important as randomization itself. The basic concept is to ensure that those who are delivering or receiving an intervention or measuring the outcome of an intervention do not know which treatment group a participant is part of; by doing this there is reduced conscious and unconscious bias and use of information from other sources. Individuals to be masked include the study participants, investigators, anyone conducting assessments, and those with contact with any of these people. Sometimes an intervention cannot be masked; in such cases, the study team must make all attempts to minimize potential sources of bias. For example, those involved in outcome assessments can be masked regarding study arm and/or study hypotheses to prevent a biased assessment of the outcomes or biased behaviors such as probing more deeply about potential side effects. In the case of behavioral interventions, it may not be possible to mask participants to study arm but it is possible to mask regarding study hypotheses. While masking to study arm can be more difficult and costly, it can have significant impact on study credibility. In general, an open study, or one that does not have elements of masking, is less credible than one with some or complete masking. These study design elements were discussed in more detail in this chapter, and randomization is the topic of Chapter 23.

Choosing the Comparison Group Comparison groups serve the important purpose of allowing an evaluation of what outcomes would result in the absence of the experimental or intervention condition for clinical studies, or, for example, in case-control studies how factors differentiate those who do and do not have the disease in question. In all studies, including case-control studies, it is important to describe carefully both the experimental (or case) and comparison (or control) groups. Comparison groups can be participants randomized to a placebo control, usual care, standard of care, attention control, or alternative treatment. In the latter case, if both treatment arms have demonstrated efficacy and/or are guideline based, such a study design can be called a comparative effectiveness trial, which as described above essentially compares two efficacious or effective treatments to each other. In all other designs, at least one of the study arms is an experimental arm, and one or more additional arms serve to control one or more factors. For every study, it is important to have a clear understanding of what factors should be controlled with the comparison group because this will typically dictate what conclusions can be drawn from the data. Given their importance to clinical study design and trial interpretation, several common control groups are described in the next section.

CONTROL GROUPS Intervention Development Intervention development strategies vary enormously according to the particular type of intervention being developed. For pharmacologic studies, intervention development is a long process that can take years and is quite costly, but occurs prior to any Phase I studies. For nonpharmacologic studies, intervention development is more likely (although not necessarily) to be iterative. In all cases, however, taking the time to systematically and empirically develop an effective and well-articulated intervention helps to ensure that future clinical trials are sound, and it decreases the opportunity for random error variance or weak interventions to move forward to large clinical trials. Adaptive treatment designs such as the Multiphase Optimization Strategy (MOST), SMART, and the Just-in-Time Adaptive Interventions designs are innovative new ways to develop and optimize behavioral interventions in an efficient manner.28e30 These methods were developed to design

One of the most complicated issues in clinical trial design is how to choose and design the most appropriate control group for a specific treatment and outcome.31 The purpose of the control group is to control potential threats to internal validity so the dependent variable(s) of interest, rather than any other nontreatment-related factors, can be said to be more likely associated with the active ingredient of the experimental treatment. The primary driving force for the choice of a control group should be the specific question being addressed; different control group conditions will allow different conclusions to be made. In other words, to choose the most appropriate control group mandates that one has a good grasp of what needs to be controlled in the experimental setting, including how the treatment of interest is defined and what the outcomes of interest are. Thus, there is no control group that can be said to be “correct” in all cases. Some of the factors that control groups are meant to control include the following: expectations

II. STUDY DESIGN AND BIOSTATISTICS

CONTROL GROUPS

(both patient and provider); time and attention (for example, from the provider, from groups in group interventions, as a result of measurement of outcomes or observation, or as a result of diagnosis); practitioner effects; social support (from practitioner and other sources); compensation (outside of potential treatmentrelated benefits); demand or burden; risk; disease progression; and nonspecific effects, including contextual effects. This last factor, context, is often thought of as that group of nonspecific effects that may come from environmental, social, and structural characteristics. Nonspecific effects are any effects of the complex treatment package on the outcomes of interest that are not due to the specific mechanisms being tested. For example, patients receiving treatment in an outpatient hospital setting might have particular responses unrelated to the treatment itself, and thus the contextual factor of where treatment is received would be important to control in such a circumstance. Control groups can take many different forms, from wait-list control, placebo control, sham control, time and attention control, and active comparator control groups. Because of the importance of choosing the most appropriate control group for any specific study, the advantages and disadvantages of each of these will be addressed.

Wait-List Control A wait-list control group is an unmasked, untreated group. In other words, participants in a wait-list control group are denied the experimental treatment but are aware that they are not receiving treatment. In such cases negative expectancies may occur; since there is no expectation for getting better sometimes patients do more poorly. This can then inflate the apparent effects of the experimental group. Wait-list groups really are not untreated because they are contacted, consented, randomized, diagnosed, and measured. For treatments that carry low expectancy for benefit and high risk, wait-list participants may actually fare better than those in the experimental group. Wait-list control groups do not represent a real natural history group. Wait-list groups, while providing some limited information particularly in early phase studies, are not sufficient for allowing definitive assessment of the clinical utility of a particular intervention strategy. One way that wait-list control groups can be most effectively utilized is in studies that employ additional control groups. In this case, information from a wait-list group is valuable because when used in conjunction with other control groups, it can allow some conclusions about effects due to natural history.

259

Time and Attention Control Time and attention control groups are most commonly employed for nonpharmacologic interventions and are meant to control contextual and nontreatment-related variance, so that any group differences can be ascribed to the active “ingredient” of the experimental treatment.32 Attention control groups are particularly useful in controlling effects of practitioner attention, as well as the attention and social support that might result for some nonpharmacological trials run in a group setting. Inclusion of a time and attention control group results in a conservative estimate of the effect size of the treatment, because many outcomes are significantly influenced by these nonspecific factors.

Placebo Control A placebo control group is typically considered to be one that receives an inert treatment. The conventional control group in a pharmacologic study is the placebo control because it is easy to mask both the experimenter and patient to condition, as long as the placebo is well matched on all key characteristics to the experimental drug. Thus, if the experimental drug is a pill, a good placebo pill will look, smell, and taste the same as the experimental drug and will have as many of the same side effects as possible (e.g., changes in urine color will be similar) while having no active ingredient. In addition, placebo controls should be given in the same setting as the experimental group, with the same dosing schedule, and be the same on all other sensory components and delivery of the treatment. For some patients and under some conditions, however, placebo responses occur and such responses may occur in as many as 30% or more of study participants. A variety of factors are widely known to influence the extent of placebo responding, including expectancy, social factors, and certain conditions such as chronic pain and Parkinson’s disease. For nonpharmacologic interventions, placebo control groups are more complex because there are domains other than sensory ones to be considered, and some of these domains would be expected to produce some effects in some patients under some conditions. The design of a placebo condition for such studies can be challenging because often these interventions cannot be easily masked. Interestingly, numerous factors can modify the extent of a placebo response. For treatments involving patiente practitioner interaction, Hawthorne effects (the phenomenon that individuals will change their behaviors when they believe that they are being observed) can be present. Expectation for success (or failure) of a treatment on the part of either the patient or provider can

II. STUDY DESIGN AND BIOSTATISTICS

260

18. CLINICAL TRIAL DESIGNS

impact placebo responding. Later in the chapter we will address placebo response in more detail.

Sham Control A sham control group is a group that experiences the same procedures as the experimental group, without the active portion of the intervention. Sham controls are most common among interventions involving devices. Just as with other control groups, sham controls reduce the potential for bias, particularly bias related to group differences in adherence and expectations. Sham controls also control treatment-related factors and procedures that are not considered to be active. For example, a sham control group for a surgical intervention would include all aspects of the surgery except whatever aspect is thought to be beneficial. Despite the benefits and strength of sham control groups, they are relatively infrequent for surgical interventions because they are invasive and share a similar degree of risk than the experimental group without benefit. However, they are somewhat more frequent for other procedurally based interventions when the risk of a sham procedure is low.

Usual and Standard Care Controls Usual care control groups (sometimes referred to as treatment as usual groups or standard care groups) are particularly relevant and important to clinical research. Studies using these control groups vary somewhat. All essentially consist of patient populations who are enrolled in a research study but who receive treatment in their usual care settings as they would if they were not enrolled in the study. Typically, the treatments that usual care control groups receive can be as varied as is typical clinical care. Closely related are enhanced usual care groups, which consist of some common and minimal level of care to address ethical issues, but which have some additional enhancements. These groups are often enhanced with elements of the active intervention that are considered to be inert or not of interest. Usual care control groups can be quite informative in effectiveness trials because they reflect what care is typically provided for a given condition and thus what occurs in real-world situations. However, there is tremendous variability in what usual care consists of; for example, usual care varies tremendously according to insurance status, socioeconomic status, and geographic region. Standard of care (in contrast to standard care) groups are somewhat less frequently employed as control groups. Standard of care groups can be thought of as groups that employ best practices, guideline based or the most commonly agreed on effective treatment for the condition of interest. It is a generally higher level and more

consistent level of care than usual care. These control groups require careful documentation and evaluation of the content of the control, since it may vary over time and across study sites. This variability may increase the needed sample size for studies using these types of control groups.

Multiple Control Groups Although the simplest of the classic designs consists of two treatmentsdintervention and control, often a placebo controldmany trials also include additional comparison groups. For example, a standard “positive control” that has previously been shown effective for that condition serves as a yardstick against which to compare the magnitude of the response produced by the experimental treatment. Without the positive control, a failure of the experimental treatment to produce a greater response than the placebo could render the study inconclusive. Although tempting to conclude that the treatment was ineffective, it is possible that the assessment instruments were insensitive, the procedures of the experimental observer were variable or confusing, the patient population has particularly high placebo response, or merely because of random variation. If a positive control were included and shown superior to both placebo and the experimental treatment, this would strengthen the conclusions, which could be drawn regarding the failure of the experimental treatment to have an effect. Alternatively, if all three arms produced similar responses, one could conclude that the study methods were inadequate to show the effects of even an efficacious treatment. In addition to a design testing an experimental treatment, a standard treatment, and a placebo, other multiarm designs without a placebo are also possible. Many clinical trials include additional treatment or control groups that are chosen to further elucidate the major research question. Other trials might include multiple control groups, each of which might control for different aspects of a multicomponent intervention. Alternatively, two or more doses of the same intervention could be compared to any control group, which would strengthen the causal inferences that could be made. A dosee response curve showing no or a small response with the control group and escalating responses with greater doses of the experimental intervention could convincingly demonstrate the positive benefit of the experimental treatment. Whatever the disease area of interest, one may wish to test the soundness of proposed research designs by graphing the possible outcomes of the trial. If the conclusion given a particular outcome is ambiguous, consider additional treatment groups that would distinguish among the alternative explanations. The addition of

II. STUDY DESIGN AND BIOSTATISTICS

PLACEBO RESPONSES

treatment or control groups is costly, however. One must either recruit more patients or reduce the size of each treatment group, lessening the statistical power of the comparisons. In many cases, particularly where negative results will not be of great interest, researchers may choose to omit controls whose main value is to clarify the interpretation of the negative result. In all cases, as with so many design decisions, which need to be made, the major factors that should drive the particular design of study arms should be the question being addressed and the end points and outcomes of interest.

PLACEBO RESPONSES Placebo, which means “I will please” in Latin, is a term applied to a presumably inactive intervention. The placebo response is the favorable response that the inactive treatment often elicits. Scientists and philosophers have wrestled with this concept for generations.33 For the purposes of clinical trial design an understanding of what constitutes a placebo group and how to understand responses of placebo groups is of special interest and importance.

Background Placebos clearly can affect subjective ratings of symptoms and function among both patients and clinicians. There is little debate about this phenomenon, and it has typically been ascribed to patient’s beliefs or expectations regarding their individual randomization in the active treatment group. In clinical research, such responses have been considered biased and primarily due to changes in symptom reporting rather than indications of positive physiological benefit. However, placebos also have been shown to influence a variety of physiological measurements, including blood pressure, airway resistance, neural functioning, and gastrointestinal motility.34,35 While various studies have shown levels of “placebo responding” among Parkinson’s patients,36,37 Alzheimer’s disease patients,38,39 and depressed and schizophrenic patients40 that were larger than expected due to spontaneous visit-to-visit fluctuations or due to the natural history of the disorder, other studies have examined the mechanisms for higher than typical placebo responses among these various patient groups. There are no simple answers. Initial reports suggesting that placebo analgesic responses after surgery can be reduced to endorphin secretion41 have been refuted by the finding that placebo analgesia is not reduced in magnitude by pretreating patients with large doses of naloxone.42 Placebo responses undoubtedly involve brain centers for language, sensation, mood, movement, and anticipation of the future, that

261

is, most of the brain and every bodily system are under its control.35 This has led to a rich and complicated literature.43 The recent psychopharmacology literature offers a revealing debate about placebo responses because in recent years large placebo effects in particularly highresponding patient populations have caused many trials of novel antidepressants and anxiolytics to fail. Some experts warn investigators to avoid psychotherapeutic intervention and to keep warm contact with the patient to the minimum needed to ensure patient compliance to counteract these effects, with the notion that placebo responses are not only due to expectations of benefit but also to patienteprovider interactions. To counteract the desire of the patient to please the practitioner with a positive report, it can be helpful to emphasize that the value of the experimental treatment is unknown. Sullivan44 explored the paradox that when clinical investigators dismiss the placebo response as a nuisance to be contained they impoverish scientific conceptions of healing because placebo responding is likely to be a part of all active treatments. An alternative view is that a better understanding of placebo responses will reveal “specific mechanisms” of the healing interaction, which may include cognition, expectations, environmental factors, and patienteprovider relationships impacting or interacting with an active pharmacological treatment. One important goal of the clinical investigator is to maximize the ratio of the specific treatment effect to the experimental variation. Large placebo responses work against this goal in two respects. First, the “specific treatment effect” is inferred to be the difference between improvement shown by study participants on the treatment and those on a placebo. To the extent that patient or practitioner expectations influence treatment, a portion of the treatment response can be considered a placebo responsedthat is, responding due to expectations rather than “active” treatment. In cases in which the placebo effect is large, a “ceiling effect” may limit the amount of incremental difference that can be seen with a specific treatment. Second, placebo responses, and the nature of the interaction between placebo and specific treatment responses, may vary greatly among individuals with different backgrounds, cognitive styles, expectations, and relationships with their health-care providers. Therefore, as the mean size of the placebo response increases, the experimental variance may increase, with corresponding loss of power.

Identifying Placebo Responders Several clinical investigators view placebo responses as nuisance variables because active treatment effects become difficult to differentiate from placebo effects. One response has been to identify high placebo

II. STUDY DESIGN AND BIOSTATISTICS

262

18. CLINICAL TRIAL DESIGNS

responders prior to randomization, and exclude them from clinical trials. This has resulted in mixed conclusions. In analgesic studies carried out in the early 1950s, several leading research teams concluded that they were unable to sort out such a subgroup45; given repeated single doses of placebo interspersed with doses of opioids, more than 80% of patients with surgical or cancer pain reported analgesia from at least one dose of placebo. In other disease areas, however, the quest to identify placebo responders has continued in the form of single-blind placebo run-in periods preceding randomization. Patients with MDDs tend to have higher than typical placebo responding. An analysis of several clinical trial cohorts of depressed patients46,47 identified similar responses of patients on both tricyclic antidepressants and placebo. Initial mood improvements that fluctuate and eventually relapse are common in both drugtreated and placebo-treated patients and are inferred to be placebo responses. In contrast, those showing steady improvements with onset after 2 weeks are virtually limited to the drug groups. These investigators have argued for using a short placebo run-in period to exclude patients with a marked placebo response and to stratify and statistically correct the outcomes of patients with lesser degrees of improvement during the run-in.48 Other psychiatric investigators consider placebo runins unhelpful.49,50 They object that this maneuver wastes time, is deceptive in intent, and does not work. They believe clinicians emit subliminal cues that the placebo run-in offers no real treatment, which dampen patients’ response, whereas a much larger placebo effect occurs at the time of the real randomization. Montgomery49 and Schweizer and Rickels51 propose the alternative of a longer baseline observation period to exclude patients with mild or rapidly cycling mood disorders. In a review of methods in irritable bowel syndrome trials, Hawkey52 points out another liability of placebo run-in periods in spontaneously fluctuating disorders. By excluding patients whose symptoms have decreased by chance during the run-in period, one tends to be left with patients whose symptoms may have worsened by chance. Other investigators have suggested that because placebo responses are less durable than specific therapeutic responses, lengthening trial duration might increase the treatment-placebo difference.47,53 However, lengthening a study increases the cost and potentially the number of dropouts. Moreover, some placebo responses are durable. A variety of major surgical procedures that later proved to be useless, including gastric freezing for duodenal ulcers and actual or sham internal mammary artery ligation for angina pectoris, were initially reported to improve or eliminate the pain in the majority of patients for 1 year after surgery. Both patient and clinician expectations contribute to the

placebo effect although are most certainly not the entirety of what makes up the placebo response.43 Many studies have shown that subjects who notice side effects after taking a pill will report more improvement than those who feel no side effects. There are two ways to minimize such bias. First, placebos should be as similar to active treatments on as many possible variables except for the presumed active ingredient in the experimental intervention. Second, investigators should strive to maximize the effectiveness of blinding procedures and assess if patients and practitioners can guess study assignment by the outcomes, appearance, taste, or side effects of the treatments.54 Assigning placebo control groups for nonpharmacological interventions is more challenging than for drug studies31 as described in the section on control groups.

MISTAKES AND MISCONCEPTIONS Not Looking at the CONSORT Statement Before, During, and After a Study The original Consolidated Standards of Reporting Trials statement was developed to improve the reporting and conduct of randomized clinical trials.55 Although the statement primarily consists of a checklist of essential elements to assist researchers report research findings consistently, careful consideration of the 25 items in the CONSORT checklist during the design phase of a future trial can be valuable. The included items refer to standards for reporting the title, introduction, methods, analyses, results discussion, and other important sections of a research report. Understanding these critical elements can help researchers ensure that the appropriate data are collected in the appropriate manner for reporting according to CONSORT guidelines. Although first published in 1996, a number of updated guidelines for specific trial designs have subsequently been published.56,57

Waiting Until the Large Definitive Study to Worry About the Details The focus of this chapter is primarily on the design of relatively small clinical trials; design and analytic issues related to large clinical trials are covered in several other chapters and books. Small and moderately sized clinical trials, although seldom definitive, are essential in establishing estimates of treatment effects and feasibility while identifying critical patient characteristics, outcomes, and components of the intervention. They usually are more practical to conduct than are larger-scale trials. Small and moderately sized clinical trials may be single or multisite trials. In all trial designs, but particularly in

II. STUDY DESIGN AND BIOSTATISTICS

MISTAKES AND MISCONCEPTIONS

trials with modest sample sizes, the design must carefully consider the estimated effect of the treatment and the estimated (nontreatment-related) variance. Thus, research design strategies for maximizing treatment effects and minimizing error variance, while important for all experimental research, are especially critical for small and moderately sized clinical trials with more limited sample sizes. Failing to Increase the Treatment Effect The treatment effects detected in a study may be increased in several ways including maximizing the treatment dose or intensity, choosing comparison or placebo interventions with minimal expected impact on outcomes, or choosing to study (and treat) patients most likely to be responsive and adhere to the intervention and least likely to be responsive to the control or comparison treatment. For example, certain patient populations are known to be particularly responsive to placebo conditions. Clinical studies of these populations, when the design includes a placebo group, must have a powerful intervention to discriminate the responses of the placebo control from those of the intervention groups. Failing to Decrease the Variance Variance due to treatment is otherwise termed treatment effects, but all other sources of variance decrease the ability to identify those treatment effects. There are several approaches to decreasing clinical trial variance. Significant variance can occur as a function of the experimental conduct of the study. Decreasing the variability in measuring the primary outcome is often a powerful and inexpensive way to stimulate the pace of therapeutic advance in an entire field. For example, careful consideration of the psychometric properties of the scales used in trials can help identify sound instruments and will result in outcomes with decreased error variance. Consideration of the primary outcomes may identify known fluctuations in those outcomes according to time of day, seasonality, and other variations that are not of interest but which will necessarily increase measurement variance. Blood pressure, for example, demonstrates a daily rhythm, and for studies including blood pressure as an important outcome, the time of day in which blood pressure is taken (as well as many other variables such as posture, instrumentation, practitioner effects, etc.) will influence nontreatment variance. Variations also may occur normally for some outcomes such as pain, which can fluctuate widely day to day in chronic pain patients. Similarly, certain conditions are associated with relapsing and remitting symptoms. For these types of outcomes especially, repeated measurement can often decrease variability and enhance the ability to identify treatment effects.

263

In the sample size formula (see Chapter 25) only unexplained sources of variation contribute to the variance, s2. If there are adequate data to identify the predictors of the outcomes of interest and those predictors can be measured, those components can be removed from this unexplained error term. From a design perspective, removing these components is accomplished by controlling the sources of variance in one or more control groups or by careful selection of patient characteristics. This also sometimes can be accomplished analytically. For example, Jung and colleagues58 reported that rash duration, age, sex, the presence of a prodrome, and the severity of pain and acute rash explain 23% of the variance in the occurrence of postherpetic neuralgia. Covariates such as surgical trauma and prior opioid exposure have been reported to improve the sensitivity of analgesic clinical trials.59 Assessment of genetic polymorphisms that affect individuals’ treatment response offer promise in explaining part of the outcome variance in many disease areas.60,61 Since sample size is proportional to variance, using this type of additional knowledge could reduce the size of a study or at least lead to a more specific study design.

Not Taking Care When Choosing a Control Group As with many aspects of study design, the question of determining the appropriate control group depends on the question being addressed, available resources, the outcomes of interest, and where the state of the sciences lies on the research continuum. For example, for dissemination and implementation studies, which address questions of how a treatment can be disseminated to real-world settings (community, worksites, schools, community-based clinics), control groups may be formed to control factors related to the dissemination or implementation plan rather than the intervention itself. Such studies may compare treatments disseminated by specifically trained community health workers (experimental group) to those provided by hospital staff. For late-stage effectiveness or pragmatic trials, control groups may be primarily usual care groups, because the focus of the question is whether the intervention is effective in the community setting, which puts little or no constraints on the patient entry criteria, the way the intervention is practiced, or strategies to enhance adherence or retention. The primary focus in designing the control group is to have a good understanding of the factors one wishes to control. This requires a concise articulation of the question to be answered and the most critical outcomes and end points to be measured. In some cases, the primary question is related to the mechanisms of action of the experimental treatment. In these cases, the

II. STUDY DESIGN AND BIOSTATISTICS

264

18. CLINICAL TRIAL DESIGNS

hypothesis to be tested is in regard to if and how the experimental treatment effects change, the outcomes are efficacy and mechanistic, and the control group should be designed to control all other presumably “nonactive” elements of the treatment. In other cases, the primary question is whether or not a new treatment effects any change through any means. In such cases, the question is whether a signal for efficacy may be apparent and the outcomes are broad. Here a usual care or waitlist control group may be appropriate. It is a common mistake to adhere in all studies to a single “correct” control group. While a researcher may certainly employ a control group that is inadequate for the question being addressed, the problem is a lack of harmony between the question and the control group rather than the choice of control group itself.

Always Assuming Placebo Groups Are Unethical It is sometimes argued that the inclusion of a placebo control when an effective treatment is available is unethical62 and that the more appropriate design should compare new treatments to a standard treatment. The standard treatment typically is defined as one of two things: standard of care, the optimal available treatment for a given condition; or standard care, how the typical treatment is given for that condition in real-world practice. Although this may be true in cases in which withholding the known treatment poses major risks of irreversible harm (e.g., studies of treatments of aggressive cancers, serious infections, or any condition for which withholding immediate effective treatment causes permanent damage), for other cases the inclusion of a placebo condition is justifiable and necessary to advance clinical science. The exclusion of placebo groups in all conditions where a treatment exists could impair the early development of many treatments, when proof of principle for a weak treatment is needed to continue efforts to improve the treatment.63e66 This is especially the case when known treatments do not have proven effectiveness but are efficacious in small, well-controlled trials only. Enrollment in a placebo arm is also acceptable when the trial is relatively short in duration, withholding a treatment does not cause permanent harm or excess distress, and when existing treatments are not well accepted by patients. The best case for the inclusion of a placebo group is when comparing a new treatment to a standard treatment, where the addition of a placebo arm can help to clarify the possibility that neither active treatment was effective in that particular trial and that natural history nor placebo effects explain the results. This is particularly the case if early stage research in the area failed to include placebo controls. As discussed previously, a

study that excludes a placebo group may produce spurious evidence for the new drug’s efficacy and lead to widespread use of an ineffective medication. Many variations of when and how to use placebo control groups have been well articulated.67

Assuming Placebo Treatment Is (Im)Possible in Long-Term Studies In short-term investigations of treatments focused on patient symptoms, placebos are often ethically justified because patients understand that they can terminate the study and take additional medication at any time.68 In actual practice, many patients experience some placebo relief, and most tolerate the study for a period of time. Chronic disease studies can be more difficult, however, for both practical and ethical reasons. Patients have a more difficult time tolerating unrelieved severe symptoms for a long period of time, and this is especially the case if effective or efficacious treatments exist. This can lead to potential differential dropout rates between patients in the placebo and intervention arms, and make it extremely difficult to make causal inferences. Ethical issues also are a concern for placebo studies, particularly if a potentially efficacious treatment is available. For certain conditions that progress without treatment, it is obvious that researchers cannot ethically give a placebo alone if it could cause permanent harm. Therefore, in these situations, one ethically feasible way to conduct placebo-controlled studies is to give both placebo and active treatment groups as an add-on treatment to patients already on optimal doses of a standard treatment.

Confusing Placebo Response and Regression to the Mean To what extent can we differentiate placebo responding from high responses changing due to the phenomenon of regression to the mean? In two large placebo-controlled doseeresponse studies of irbesartan,68 an antihypertensive, diastolic pressure initially dropped by a mean of 4 mmHg in patients treated with placebo capsules and 5e10 mm in patients treated with irbesartan. Was the 4 mm drop a “placebo response?” A plausible alternative explanation was that this improvement reflects the phenomenon of “regression to the mean.” In chronic disorders with fluctuating symptoms and signs, patients are more likely to volunteer for studies and qualify for entry when their disease is in a worse period. Conversely, after study entry, there will be a tendency for them to improve just by random variation. (See Chapter 27 for more details.) One way to distinguish a

II. STUDY DESIGN AND BIOSTATISTICS

MISTAKES AND MISCONCEPTIONS

placebo response from regression to the mean is to observe the outcome over a longer period of time and after treatment (including placebo) has ceased. A subsequent increase in the outcome would suggest the response was due to patients’ expectations of a drug effect during treatment. Another albeit not perfect way to distinguish placebo response from regression to the mean is to include a no treatment group or wait-list control group as well as a placebo group. One may infer that improvement in the no treatment group is regression to the mean, and the additional improvement in the placebo group is the placebo response. Hrobjartsson and Gotzche69 used this strategy to measure placebo responses in 156 published clinical trials that included both a placebo group and a no treatment group. They concluded that most or all of what is commonly considered placebo response is really regression to the mean, except perhaps in studies of pain, anxiety, and other outcomes reported by the subject. This finding fails to explain the physiological and neurological changes that occur with placebos, however.

Using a Factorial or Partial Factorial Design Instead of a Parallel Group Design One of the advantages of using a factorial design is savings in sample size. Chapter 25 discusses power and sample size in general, but there are situations when factorial or partial factorial designs do not allow smaller sample sizes relative to a parallel group design. As mentioned earlier, there are several assumptions that come with the sample size savings of a factorial design. One is that each of the interventions has an independent effect on the outcomes measured in the trial, which is not always true. In fact, one of the study goals may be to investigate the interaction of the treatments under study. Studies wishing to look at potential interactions of the treatments will need to be powered to do so, which will increase the sample size, sometimes (but not always) to the size of a parallel group design. One alternative is to consider each possible treatment combination as a separate arm in a parallel group design, although there are other methods that can be used, such as the MOST and others that allow smaller sample sizes for partial factorial designs even when postulating and testing interaction effects.70

Assuming Small, Open-Label, Nonrandomized, Uncontrolled Studies Offer No Evidence The ISIS-4 study,23 a multisite randomized study with approximately 58,000 participants was discussed in the factorial design section. Another study was alluded to in the early Phase II design section of this chapter. Several

265

years ago, the intramural research program of the National Institute of Neurological Disorders and Stroke (NINDS) conducted a series of studies to evaluate images of contrast-enhanced lesions as a measure of disease activity in early RRMS. The contrasting agent gadolinium causes areas of bloodebrain barrier breakdown to appear on MRI images as bright spots or lesions. Traditional clinical measures of disease activity, such as those based on assessing physical or mental disability, are known to be very insensitive during the early phase of the disease. By comparison, it is thought that the number and area of these lesions as measured by serial monthly MRI images may be a more sensitive measure of disease activity during this phase.71,72 A series of Phase II (safety/ efficacy) studies were conducted at NINDS to screen new agents, including beta-interferon, for effectiveness. One study examined the effect of beta-interferon on lesion activity during the early phase of RRMS.73,74 The beta-interferon study was designed to have 14 patients followed for 13 months. Patients remained untreated during the first 7 months (seven serial MRI images) and then were treated with beta-interferon during the last 6 months (six serial MRI measurements). The primary outcome or response in this study was the average monthly number of lesions on treatment minus the corresponding average number during the untreated baseline period. The study results showed that beta-interferon significantly reduced the number of lesions compared to baseline. This study is a nonrandomized study in which all patients were switched over to the investigational treatment after 6 months. This type of nonrandomized design is one of many used to screen for new therapeutic agents. The intramural research program of NINDS also conducted the previously mentioned clinical trial to study the efficacy of felbamate monotherapy for the treatment of intractable partial epilepsy.24 The patients in this study had partial and secondary generalized seizures and were undergoing presurgical monitoring. The effectiveness of felbamate monotherapy was compared to that of a placebo. Forty patients were randomized to either felbamate (n ¼ 19) or placebo (n ¼ 20) and followed in the clinical center for 2 weeks. The patients’ numbers and types of seizures were recorded daily for 2 weeks. The primary outcome of this study was daily seizure rates for patients on treatment or placebo. The study results showed that felbamate monotherapy significantly reduced the number of seizures compared to the placebo. This type of randomized design is often used to test promising new treatments for efficacy. The diverse study designs of these examples illustrate fundamental issues. The study evaluating the effect of beta-interferon is a nonrandomized study, whereas the felbamate monotherapy trial and the ISIS-4 trials are randomized clinical trials. Their designs varied as did

II. STUDY DESIGN AND BIOSTATISTICS

266

18. CLINICAL TRIAL DESIGNS

their control groups and sample sizes. In each study the investigators wished to determine whether a given treatment or treatments were effective in the care of patients with a specific disease or medical risk. Each study was conducted because the investigators wanted to be able to treat not only the patients in that study but also all patients with similar characteristics, diseases, and medical risks. Each study, though, was started at a different point in the research continuum. The goal of introducing these and other study examples is to be clear that not all studies are right at every single time. Different design elements are needed at different points. At times information and evidence can be borrowed from different populations or fields, but research must start somewhere.

3. Quasiexperimental designs a. Are one of the most powerful design strategies used in clinical trials b. Lack randomization, a control group, and an independent variable c. Can be used when randomization is not possible, such as in naturalistic settings d. Typically cannot lead to strong causal inferences

Acknowledgments We thank Jack M. Guralnik, Teri A. Manolio, Paul S. Albert, Craig B. Borkowf, and the late Mitchell B. Max who contributed material to the previous editions of this textbook that informed this chapter.

Disclosures CONCLUSIONS The goal of this chapter has been to outline the essential questions researchers must address in choosing a clinical research design and designing a control group for a clinical trial. The most fundamental of these questions is carefully considering the specific goal(s) of the study in terms of the study questions being addressed. Understanding the components of the new treatment or intervention that are considered to be active, identifying the outcomes of particular interest, and careful consideration of how the study fits in the context of the body of existing knowledge in the area are fundamental elements of optimal clinical trial design and evidence-based research.

SUMMARY QUESTIONS 1. What might be considered to be the first stage in developing an appropriate design and analytic strategy for a research study? a. Developing an effective and feasible intervention strategy b. Articulating the precise research question to be addressed c. Locating an appropriate patient population of interest d. Conducting a dose-ranging study 2. What type of control group is the most appropriate to use in most clinical trials? a. Either a placebo or wait-list control group b. A usual care control group c. A time and attention control group d. There is not a single correct answer to this question because the most appropriate control group depends on the question and outcomes of interest

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health (NIH) or the US government. This chapter reflects the views of the author and should not be construed to represent FDA’s views or policies.

References 1. Evans SR. Clinical trial structures. J Exp Stoke Transl Med 2010;3: 8e18. 2. Manson JE, Hsia J, Johnson KC, Rossouw JE, Assaf AR, Lasser NL, Trevisan M, Black HR, Heckbert SR, Detrano R, Strickland OL, Wong ND, Crouse JR, Stein E, Cushman M, Women’s Health Initiative Investigators. Estrogen plus progestin and the risk of coronary heart disease. N Engl J Med 2003;349:523e34. 3. International Conference on Harmonisation. Choice of control group and related issues in clinical trials. Fed Regist 2001;66. 4. Korn EL. Nontoxicity endpoints in phase I trial designs for targeted, non-cytotoxic agents. J Natl Cancer Inst 2004;96:977e8. 5. Parulekar WR, Eisenhauer EA. Phase I trial design for solid tumor studies of targeted, non-cytotoxic agents: theory and practice. J Natl Cancer Inst 2004;96:990e7. 6. Hunsberger S, Rubinstein LV, Dancey J, Korn EL. Dose escalation trial designs based on a molecularly targeted endpoint. Stat Med 2005;24:2171e81. 7. Thall PF, Cook JD. Dose-finding based on efficacy-toxicity tradeoffs. Biometrics 2004;60:684e93. 8. Thall PF, Simon R. Incorporating historical control data in planning phase II clinical trials. Stat Med 1990;9:215e28. 9. Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin Trials 1989;10:1e10. 10. Simon R, Wittes RE, Ellenberg SS. Randomized phase II clinical trials. Cancer Treat Rep 1985;69:1375e81. 11. Steinberg SM, Venzon DJ. Early selection in a randomized phase II clinical trial. Stat Med 2002;21:1711e26. 12. Green S, Benedetti J, Crowley J, et al. Clinical trials in oncology. Data monitoring committees for Southwest Oncology Group clinical trials. 2nd ed. 2002. 13. Sox HC, Greenfield S. Comparative effectiveness research: a report from the Institute of Medicine. Ann Intern Med 2009;151: 203e5. 14. Schwartz D, Lellouch J. Explanatory and pragmatic attitudes in therapeutical trials. J Chronic Dis 1967;20:637e48.

II. STUDY DESIGN AND BIOSTATISTICS

REFERENCES

15. Thorpe KE, Zwarenstein M, Oxman AD, Treweek S, Furberg CD, Altman DG, Tunis S, Bergel E, Harvey I, Magid DJ, Chalkidou K. A pragmatic-explanatory continuum indicator summary (PRECIS): a tool to help trial designers. J Clin Epidemiol 2009;62:464e75. 16. Loudon K, Treweek S, Sullivan F, Donnan P, Thorpe KE, Zwarenstein M. The PRECIS-2 tool: designing trials that are fit for purpose. BMJ 2015;350:h2147. 17. Campbell DT, Stanley JC. Experimental and quasi-experimental designs for research. Boston: Houghton Mifflin Co.; 1963. 18. Shadish WR, Cook TD, Campbell DT. Experimental and quasiexperimental designs for research. Boston: Houghton Mifflin; 2002. 19. Trevisan M, Jossa F, Farinaro KV, Panico S, Giumetti D, Mancini M. Earthquake and coronary heart disease risk factors: a longitudinal study. Am J Epidemiol 1991;135:632e7. 20. Lipton RB, Bigal ME, Stewart WF. Clinical trials of acute treatments for migraine including multiple attack studies of pain, disability, and health-related quality of life. Neurology 2005;65(12 Suppl. 4): S50e8. 21. Byas-Smith MG, Max MB, Muir J, Kingman A. Transdermal clonidine compared to placebo in painful diabetic neuropathy using a twostage ‘enriched enrollment’ design. Pain 1995;60:267e74. 22. Temple RJ. Special study designs: early escape, enrichment, studies in non-responders. Commun Stat Theor Methods 1994;23: 499e531. 23. ISIS-4 (Fourth International Study of Infarct Survival) Collaborative Group. ISIS-4: a randomised factorial trial assessing early oral captopril, oral mononitrate, and intravenous magnesium sulphate in 58,050 patients with suspected acute myocardial infarction. Lancet 1995;345:669e85. 24. Theodore WH, Albert P, Stertz B, et al. Felbamate monotherapy: implications for antiepileptic drug development. Epilepsia 1995; 36:1105e10. 25. Friedman LM, Furberg CD, DeMets DL. Fundamentals of clinical trials. New York: Springer; 2015. 26. Murray DM, Varnell SP, Blitstein JL. Design and analysis of group-randomized trials: a review of recent methodological developments. Am J Public Health 2004;94:423e32. 27. Lagoa CM, Bekiroglu K, Lanza S, Murphy SA. Designing adaptive intensive interventions using methods from engineering. J Consult Clin Psychol 2014;82:868e78. 28. Almirall D, Nahum-Shani I, Sherwood NE, Murphy SA. Introduction to SMART designs for the development of adaptive interventions: with application to weight loss research. Transl Behav Med 2014;4: 260e74. 29. Collins LM, Murphy SA, Strecher V. The multiphase optimization strategy (MOST) and the sequential multiple assignment randomized trial (SMART): new methods for more potent eHealth interventions. Am J Prev Med 2007;32(5 Suppl.):S112e8. 30. Nahum-Shani I, Smith SM, Spring BJ, Witkiewitz K, Tewari A, Murphy SA. Just- in-time adaptive interventions (JITAIs) in mobile health: key components and design principles for ongoing health behavior support. Ann Behav Med 2016. http://dx.doi.org/ 10.1007/s12160-016-9830-8. 31. Mohr DC, Spring B, Freedland KE, Beckner V, Arean P, Hollon SD, Ockene J, Kaplan R. The selection and design of control conditions for randomized controlled trials of psychological interventions. Psychother Psychosom 2009;78:275e84. 32. Freedland KE, Mohr DC, Davidson KW, Schwartz JE. Usual and unusual care: existing practice control groups in randomized controlled trials of behavioral interventions. Psychosom Med 2011; 73:323e35. 33. White L, Tursky B, Schwartz GE, editors. Placebo: theory, research, and mechanism. New York: Guildford; 1985.

267

34. Soiro HM. Doctors, patients, and placebos. New Haven (CT): Yale University Press; 1986. 35. Benedetti G, Carlino E, Pollo A. How placebos change the patient’s brain. Neuropsychopharmacology 2011;36:339e54. 36. McRae C, Cherin E, Yamazaki TG, Diem G, Vo AH, Russell D, Ellgring JH, Fahn S, Greene P, Dillon S, Winfield H, Bjugstad KB, Freed CR. Effects of perceived treatment on quality of life and medical outcomes in a double-blind placebo surgery trial. Arch Gen Psychiatry 2004;61:412e20. 37. Shetty N, Friedman JH, Kieburtz K, Marshall FJ, Oakes D. The placebo response in Parkinson’s disease. Parkinson Study Group. Clin Neuropharmacol 1999;22:207e12. 38. Spencer CM, Noble S. Rivastigmine. A review of its use in Alzheimer’s disease. Drugs Aging 1998;13:391e411. 39. Kawas CH, Clark CM, Farlow MR, Knopman DS, Marson D, Morris JC, Thal LJ, Whitehouse PJ. Clinical trials in Alzheimer disease: debate on the use of placebo controls. Alzheimer Dis Assoc Disord 1999;13:124e38. 40. Montgomery SA. The failure of placebo-controlled studies: ECNP consensus meeting. Eur Neuropsychopharmacol 1999;9:271. 41. Levine JD, Gordon NC, Fields HL. The mechanism of placebo analgesia. Lancet 1978;2:654e7. 42. Gracely RH, Dunbar R, Wolskee PJ, Deeter WR. Placebo and naloxone can alter post-surgical pain by separate mechanisms. Nature 1983;306:264e5. 43. Guess RH, Kleinman A, Kusek JW. The science of the placebo. London: BMJ Books; 2002. 44. Sullivan MD. Placebo controls and epistemic control in orthodox medicine. J Med Philos 1993;18:213e31. 45. Houde RW, Beaver WT. Clinical measurement of pain. New York: Academic Press; 1965. 46. Nierenberg AA, Quitkin FM, Kremer C, Keller MB, Thase ME. Placebo-controlled continuation treatment with mirtazapine: acute pattern of response predicts relapse. Neuropsychopharmacology 2004;29:1012e8. 47. Quitkin FM, Stewart JW, McGarth PJ, Nunes E, Ocepek-Welikson K, Tricamo E, Rabkin JG, Klein DF. Further evidence that a placebo response to antidepressants can be identified. Am J Psychiatry 1993; 150:566e70. 48. Quitkin FM, McGrath PJ, Stewart JW, Ocepek-Welikson K, Taylor BP, Nunes E, Delivannides D, Agosti V, Donovan SJ, Ross D, Petkova E, Klein DF. Placebo run-in period in studies of depressive disorders. Clinical, heuristic and research implications. Br J Psychiatry 1998;173:242e8. 49. Montgomery SA. Alternatives to placebo-controlled trials in psychiatry. In: ECNP consensus meeting, September 26, 1996, Amsterdam. European College of Neuropsychopharmacology. Eur Neuropsychopharmacol, vol. 9; 1999. p. 265e9. 50. Trivedi M, Rush J. Does a placebo run-in or a placebo treatment cell affect the efficacy of antidepressant medications? Neuropsychopharmacology 1994;11:33e43. 51. Schweizer E, Rickels K. Placebo response in generalized anxiety: its effect on the outcome of clinical trials. J Clin Psychiatry 1997; 58(Suppl. 11):30e8. 52. Hawkey CJ. Irritable bowel syndrome clinical trial design: future needs. Am J Med 1999;107:98Se102S. 53. Spiller RC. Problems and challenges in the design of irritable bowel syndrome clinical trials: experience from published trials. Am J Med 1999;107:91Se7S. 54. Moscucci M, Byrne L, Weintraub M, Cox C. Blinding, unblinding, and the placebo effect: an analysis of patients’ guesses of treatment assignment in a double-blind clinical trial. Clin Pharmacol Ther 1987;41:259e65.

II. STUDY DESIGN AND BIOSTATISTICS

268

18. CLINICAL TRIAL DESIGNS

55. Begg C, Cho M, Eastwood S, Horton R, Moher D, Olkin I, Pitkin R, Rennie D, Schulz KF, Simel D, Stroup DF. Improving the quality of reporting of randomized controlled trials. The CONSORT statement. JAMA 1996;276:637e9. 56. Moher D, Hopewell S, Schulz KF, Montori V, Gøtzsche PC, Devereaux PJ, Elbourne D, Egger M, Altman DG. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ 2010;340:c869. 57. Schulz KF, Altman DG, Moher D. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. Trials 2010;11:32e57. 58. Jung BF, Johnson RW, Griffin DRJ, Dworkin RH. Risk factors for postherpetic neuralgia in patients with herpes zoster. Neurology 2004;62:1545e51. 59. Max MB, Portenoy RK, Laska EM. The design of analgesic clinical trials. Advances in pain research and therapy, vol. 18. New York: Raven Press; 1991. 60. Askmalm MS, Carstensen J, Nordenskjold B, Olsson B, Rutqvist LE, Skoog L, Sta˚l O. Mutation and accumulation of p53 related to results of adjuvant therapy of postmenopausal breast cancer patients. Acta Oncol 2004;43:235e44. 61. Lee DK, Currie GP, Hall IP, Lima JJ, Lipworth BJ. The arginine-16 beta2-adrenoceptor polymorphism predisposes to bronchoprotective subsensitivity in patients treated with formoterol and salmeterol. Br J Clin Pharmacol 2004;57:68e75. 62. Rothman KJ, Michels KB. The continuing unethical use of placebo controls. N Engl J Med 1994;331:394e8. 63. Charney DS, Nemeroff CB, Lewis L, Laden SK, Gorman JM, Laska EM, Borenstein M, Bowden CL, Caplan A, Emslie GJ, Evans DL, Geller B, Grabowski LE, Herson J, Kalin NH, Keck Jr PE, Kirsch I, Krishnan KR, Kupfer DJ, Makuch RW, Miller FG, Pardes H, Post R, Reynolds MM, Roberts L, Rosenbaum JF, Rosenstein DL, Rubinow DR, Rush AJ, Ryan ND, Sachs GS, Schatzberg AF, Solomon S. National Depressive and Manic-Depressive Association consensus statement on the use of placebo in clinical trials of mood disorders. Arch Gen Psychiatry 2002;59:262e70. 64. Fleischhacker WW, Czobor P, Hummer M, Kemmler G, Kohnen R, Volavka J. Placebo or active control trials of antipsychotic drugs? Arch Gen Psychiatry 2003;60:458e64. 65. Loder E, Goldstein R, Biondi D. Placebo effects in oral triptan trials: the scientific and ethical rationale for continued use of placebo controls. Cephalalgia 2005;25:124e31.

66. Temple RJ. When are clinical trials of a given agent vs. placebo no longer appropriate or feasible? Control Clin Trials 1997;18:613e20. 67. Miller FG, Shorr AF. Unnecessary use of placebo controls: the case of asthma clinical trials. Arch Intern Med 2002;162:1673e7. 68. Pool JL, Guthrie RM, Littlejohn TW. Dose-related antihypertensive effects of irbesartan in patients with mild-to-moderate hypertension. Am J Hypertens 1998;11:462e70. 69. Hrobjartsson A, Gotzsche PC. Is the placebo powerless? An analysis of clinical trials comparing placebo with no treatment. N Engl J Med 2001;344:1594e602. 70. Collins LM, Baker TB, Mermelstein RJ, Piper ME, Jorenby DE, Smith SS, Christiansen BA, Schlam TR, Cook JW, Fiore MC. The multiphase optimization strategy for engineering effective tobacco use interventions. Ann Behav Med 2011;41:208e26. 71. Albert PS, McFarland HF, Smith ME, Frank JA. Time series for modelling counts from a relapsing-remitting disease: application to modelling disease activity in multiple sclerosis. Stat Med 1994; 13:453e66. 72. McFarland HF, Frank JA, Albert PS, Smith ME, Martin R, Harris JO, Patronas N, Maloni H, McFarlin DE. Using gadolinium-enhanced magnetic resonance imaging lesions to monitor disease activity in multiple sclerosis. Ann Neurol 1992;32:758e66. 73. Stone LA, Frank JA, Albert PS, Bash C, Smith ME, Maloni H, McFarland HF. The effect of interferon-beta on blood-brain barrier disruptions demonstrated by contrast-enhanced magnetic resonance imaging in relapsing-remitting multiple sclerosis. Ann Neurol 1995;37:611e9. 74. Stone LA, Frank JA, Albert PS, Bash CN, Calabresi PA, Maloni H, McFarland HF. Characterization of MRI response to treatment with interferon beta-1b: contrast-enhancing MRI lesion frequency as a primary outcome measure. Neurology 1997;49:862e9.

Further Reading 1. Levine RJ. The need to revise the Declaration of Helsinki. N Engl J Med 1999;341:531e4. 2. Max MB, Schafer SC, Culnane M, Smoller B, Dubner R, Gracely RH. Amitriptyline, but not lorazepam, relieves postherpetic neuralgia. Neurology 1988;38:1427e32. 3. Max MB, Schafer SC, Culnane M, Dubner R, Gracely RH. Association of pain relief with drug side effects in postherpetic neuralgia: a single-dose study of clonidine, codeine, ibuprofen, and placebo. Clin Pharmacol Ther 1988;43:363e71.

II. STUDY DESIGN AND BIOSTATISTICS

C H A P T E R

19 The Role of Comparative Effectiveness Research Joe V. Selby, Evelyn P. Whitlock, Kelly S. Sherman, Jean R. Slutsky Patient-Centered Outcomes Research Institute (PCORI), Washington, DC, United States

O U T L I N E Introduction

The Role of Engagement in Specifying Research Questions

269

A History of Comparative Clinical Effectiveness Research

270

The Patient-Centered Outcomes Research Institute 271 The Role of Comparative Clinical Effectiveness Research in the Nation’s Medical Research Enterprise The Methods of Comparative Clinical Effectiveness Research Getting the Research Question Right Choosing the Study Population Selecting Appropriate Interventions and Comparator(s) Choosing Clinical Outcomes to Be Measured

273 275 275 276 276 277

INTRODUCTION Individuals making decisions about their health and health care face many choices. So do others involved directly (i.e., caregivers, family members, clinicians) or indirectly (i.e., delivery systems, employers, health plans, policy-makers) in these decisions. Choices are the result of continuous efforts to improve the prevention, detection, and treatment of illness through basic science and clinical research, system-level interventions to improve health-care delivery and the public’s health, and marketing initiatives to promote new products and services. Alternative health-care choices may differ in

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00019-8

277

Study Designs for CER Studies Experimental Study Designs for CER Observational Study Designs for CER Cohort Designs Adjusting for and Avoiding Confounding in Observational CER Studies Assessing Treatment Heterogeneity

278 279 281 282

Evidence Synthesis in CER

285

283 284

Building a National Infrastructure for the Conduct of Comparative Effectiveness Research 287 Conclusions

290

References

290

effectiveness, in safety, in side-effects, costs, or convenience. Comparing viable alternative approaches using scientific methods to determine whether one leads to better results (i.e., a preponderance of desirable over undesirable outcomes) is an important activity of clinical and health services research. Once comparative information is generated and disseminated, individual patients may consider it, often in partnership with a clinician, apply their personal values or preferences to trade-offs between desirable and undesirable outcomes, and make informed decisions. Health systems, payers, and policy-makers may use the comparative information to make decisions at system and population levels.

269

Copyright © 2018. Published by Elsevier Inc.

270

19. COMPARATIVE EFFECTIVENESS

A HISTORY OF COMPARATIVE CLINICAL EFFECTIVENESS RESEARCH Comparative clinical effectiveness research (CER) is the scientific search for and quantification of differences in the full range of benefits and harms of two or more approaches to preventing, diagnosing, treating, or managing disease. CER is not a new activity. Published articles in which the term “comparative effectiveness” appears in the title are found among the earliest entries in PubMed (Fig. 19.1). Growth in the number of such publications has accelerated rapidly over the past 10 years. The closely related concept of “pragmatic” research was first defined in 1967 as clinical research intended to support decision-making and distinguished from “explanatory” research, which is intended to build knowledge alone.1 CER is pragmatic research in that it aims to generate evidence that can support decisionmaking and improve health care. Research questions are related not to whether or how an intervention works in an ideal situation but to whether interventions work in routine clinical settings and in the broad range of patients found in real-world clinical practice. In 2003, Tunis et al.2 described the “practical” clinical study as one that compares alternatives that are meaningful to decisionmakers, in typical patient populations and settings, that proactively engages decision-makers in generating the research questions, and that considers a wider range of outcomes. Publication of this paper coincided with the passage of the Medicare Prescription Drug, Improvement, and Modernization Act (also called the

Medicare Modernization Act or MMA) of 2003, which included Section 1013 authorizing up to $50 million for the Agency for Healthcare Research and Quality (AHRQ) to conduct a program of comparative clinical effectiveness research. Interest in CER grew rapidly in the United States during the first decade of this century, with the allocation of $1.1 billion through The American Recovery and Reinvestment Act of 2009 (ARRA, or the “stimulus package”) to fund CER across the federal government and the generation of the top 100 CER topics by the Institute of Medicine (IOM) (now the National Academy of Medicine) in 2009 (see Text Box for details). CER was increasingly seen as an important tool for informing choices made by patients and others, for improving the quality of decision-making in the face of possible harms as well as benefits and, importantly, for reducing ineffective or wasteful care, thereby helping to control or lower the costs of care while improving patient outcomes. By 2006, the possibility of establishing a national institute that would fund, conduct, and synthesize comparative effectiveness research was being proposed by persons within or close to the US government.3,4 Concerns that any concentrated CER effort could fall prey to inconsistent funding appropriations and political discord if placed within the federal government increased support for establishing an independent institute with mandatory funding charged with a broad CER mandate. In 2010, the Patient-Centered Outcomes Research Institute (PCORI) was created as the independent entity as part of the Patient Protection and Affordability Act.

Number of 'Comparative Effectiveness' Articles in PubMed over Time Comparative Effectiveness Articles per 100,000

50 45 40 35 30 25 20 15 10 5 0 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Year

FIGURE 19.1 Number of articles in PubMed with “comparative effectiveness” in the title per 100,000 total published articles, 1950e2015.

II. STUDY DESIGN AND BIOSTATISTICS

THE PATIENT-CENTERED OUTCOMES RESEARCH INSTITUTE

A BRIEF HISTORY OF EVENTS LEADING TO CREATION OF THE PATIENT-CENTERED OUTCOMES RESEARCH INSTITUTE 2003 The Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (MMA), Section 1013, authorized up to $50M for the AHRQ to conduct outcomes research and evidence synthesis for 10 high-priority chronic conditions through its Evidence-based Practice Centers and its newly formed Developing Evidence to Inform Decisions about Effectiveness (DEcIDE) Network. 2006 Wilensky3 discussed a range of options for creating and housing a national center for the conduct and synthesis of comparative effectiveness research. 2007 Congressional Budget Office4 proposed options for creating a center of comparative effectiveness research, funding the center, and using the CER results. 2009 The ARRA economic stimulus allocated $1.1 billion specifically for CER to the AHRQ ($300 million), the National Institutes of Health (NIH, $400 million), and the Department of Health and Human Services (DHHS, $300 million).5 Funds were dedicated to funding large CER studies,6 improving the quality and availability of national data systems for conducting CER,7,8 creation of sustainable disease-specific registries,9 and workforce training.5 ARRA also funded the IOM (now the National Academy of Medicine) to develop an initial set of high-priority CER topics.10 The committee engaged stakeholders via a website and day-long public meeting to collect nominations for topics that could be addressed using CER and produced a set of 100 high-priority topics across a wide range of conditions, grouped into four-ranked quartiles. The topics included general questions related to prevention, diagnosis, treatment, and monitoring of a wide range of conditions; fully half of the 100 topics, research topics focused on interventions at the health-care system level.

Numerous definitions of CER were published during this period (Table 19.1) reflecting a very high degree of interest in the evolving concept. All definitions agree that CER’s purpose is to support decision-making by patients as well as other stakeholders, and all suggest that the CER approach may be applied to both clinical and health system interventions. Most specify that CER compares a broad range of interventions (not simply pharmaceuticals) and also can be applied to prevention, diagnosis, and monitoring. Some definitions mention that CER takes place in real-world settings and diverse populations, stressing that it is “effectiveness” and not efficacy research. At least two definitions state that CER is conducted in response to the “expressed needs” of decision-makers, pointing researchers toward

271

engaging with the ultimate users of the research before and during its conduct. Several definitions note that CER should attend to differences in comparative effectiveness of interventions between patient subgroups or care settings. That is, CER seeks to learn “what works best for whom.” Several definitions clarify that CER includes randomized, controlled trials as well as observational studies. Just one definition includes costeffectiveness; the remainder restricts the definition to assessment of comparative clinical effectiveness. We take the approach of not including cost-effectiveness analysis in this chapter. With good clinical effectiveness information, others can apply locally relevant costs, fill in remaining assumptions, and model costeffectiveness.

THE PATIENT-CENTERED OUTCOMES RESEARCH INSTITUTE As part of the Patient Protection and Affordable Care Act of 2010, the US Congress established the PCORI as an independent agency with the sole purpose of conducting and funding comparative clinical effectiveness research.15 Direct funding of approximately $3.5 billion was provided for the period from 2010 to 2019 from the US Treasury and from mandated fees levied on health insurers, including Medicare, commercial health plans, and large self-insured employers. PCORI was created as a not-for-profit, nongovernmental agency, overseen by a 21-person multistakeholder Board of Governors. A 17-member Methodology Committee also was created to develop and improve the science and methods of comparative clinical effectiveness research. Both Board and the Methodology Committee members are appointed by the Comptroller General, US Government Accountability Office. PCORI established, through a public process, five broad national priorities (Table 19.2) to guide its research agenda.16 In each priority area, PCORI funds both investigator-initiated projects and specific, usually larger scale, projects that focus on high-priority topics. To identify high-priority topics, PCORI solicits research questions from a range of stakeholders using multiple ongoing strategies. Questions may be submitted via PCORI’s website. PCORI convenes meetings and maintains active relationships with a broad range of organizations representing patients and consumers, caregivers, clinicians and delivery systems, health plans and large employers, and the research community. These organizations are invited to submit lists of CER questions in their areas of interest and to participate as stakeholders on PCORI-funded research. PCORI also supports the convening of researchers with patient, consumer, and clinician communities through engagement

II. STUDY DESIGN AND BIOSTATISTICS

Recent Definitions of Comparative Effectiveness Research Year

Definition

Agency for Healthcare Research and Quality7

2006

Comparative effectiveness research is designed to inform health-care decisions by providing evidence on the effectiveness, benefits, and harms of different treatment options. The evidence is generated from research studies that compare drugs, medical devices, tests, surgeries, or ways to deliver health care.

Congressional Budget Office

2007

As applied in the health-care sector, an analysis of comparative effectiveness is simply a rigorous evaluation of the impact of different options that are available for treating a given medical condition for a particular set of patients. Such a study may compare similar treatments, such as competing drugs, or it may analyze very different approaches, such as surgery and drug therapy. The analysis may focus only on the relative medical benefits and risks of each option, or it may also weigh both the costs and the benefits of those options. In some cases, a given treatment may prove to be more effective clinically or more cost-effective for a broad range of patients, but frequently a key issue is determining which specific types of patients would benefit most from it.

Medicare Payment Advisory Commission (MedPAC)11

2008

Comparative-effectiveness analysis evaluates the relative value of drugs, devices, diagnostic and surgical procedures, diagnostic tests, and medical services. By value, we mean the clinical effectiveness of a service compared with its alternatives. Comparativeeffectiveness information has the potential to promote care of higher value and quality in the public and private sectors.

Institute of Medicine

2009

Comparative effectiveness research is the generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat, and monitor a clinical condition or to improve the delivery of care. The purpose of CER is to assist consumers, clinicians, purchasers, and policy-makers to make informed decisions that will improve health care at both the individual and population levels.

Federal Coordinating Committee on Comparative Effectiveness Research12

2009

Comparative effectiveness research is the conduct and synthesis of research comparing the benefits and harms of different interventions and strategies to prevent, diagnose, treat, and monitor health conditions in “real-world” settings. The purpose of this research is to improve health outcomes by developing and disseminating evidence-based information to patients, clinicians, and other decision-makers, responding to their expressed needs, about which interventions are most effective for which patients under specific circumstances. • To provide this information, comparative effectiveness research must assess a comprehensive array of health-related outcomes for diverse patient populations and subgroups. • Defined interventions compared may include medications, procedures, medical and assistive devices and technologies, diagnostic testing, behavioral change, and delivery system strategies. • This research necessitates the development, expansion, and use of a variety of data sources and methods to assess comparative effectiveness and actively disseminate the results. The definition above is not meant to exclude randomized trials; however, these trials would need comparator arms other than placebo and be representative of populations seen in “real-world” practice.

RAND Corporation13

2009

Comparative effectiveness research examines the degree to which alternative treatments for the same health problem produce equivalent or different health outcomes. The products of comparative effectiveness research can be used in a variety of ways, including to provide information to physicians and patients in choosing appropriate treatments, as well as input into insurance benefit design, coverage determination, and payment.

National Pharmaceutical Council14

2016

Comparative effectiveness research (CER) is the conduct and synthesis of research comparing the benefits and harms of different interventions and strategies to prevent, diagnose, treat and monitor health conditions. The goal of CER is to improve health outcomes by developing and disseminating evidence-based information to patients, clinicians, and other decision-makers, responding to their expressed needs, about which interventions are most effective for which patients under specific circumstances. CER uses a wide range of research methods, including randomized controlled trials, observational studies, and systematic reviews, a structured assessment of evidence available from multiple primary studies.

19. COMPARATIVE EFFECTIVENESS

II. STUDY DESIGN AND BIOSTATISTICS

Source

272

TABLE 19.1

273

THE ROLE OF COMPARATIVE CLINICAL EFFECTIVENESS RESEARCH IN THE NATION’S MEDICAL RESEARCH ENTERPRISE

TABLE 19.2 1

The National Research Priorities of the PatientCentered Outcomes Research Institute

Assessment of Prevention, Diagnosis, and Treatment OptionsdComparing the effectiveness and safety of alternative prevention, diagnosis, and treatment options to see which ones work best for different people with a particular health problem.

2

Improving Healthcare SystemsdComparing health system elevel approaches to improving access, supporting patient self-care, innovative use of health information technology, coordinating care for complex conditions, and deploying workforce effectively.

3

Communication and Dissemination ResearchdComparing approaches to providing comparative effectiveness research information, empowering people to ask for and use the information, and supporting shared decision-making between patients and their providers.

4

Addressing DisparitiesdIdentifying potential differences in prevention, diagnosis or treatment effectiveness, or preferred clinical outcomes across patient populations and the health care required to achieve best outcomes in each population.

5

Accelerating Patient-Centered Outcomes Research and Methodological ResearchdImproving the nation’s capacity to conduct patient-centered outcomes research, by building data infrastructure, improving analytic methods, and training researchers, patients, and other stakeholders to participate in this research.

awards. Many of these projects develop and submit CER questions. In consultation with PCORI’s Board of Governors and multiple stakeholder groups, PCORI processes topics through a prioritization pathway (Fig. 19.4). Standing multistakeholder advisory panels, one for each of PCORI’s national priorities, are appointed to 2-year terms by PCORI from among volunteers responding to annual calls. Advisory panels meet regularly and provide guidance on topics in the pipeline and new issues that PCORI should consider. Submitted questions and their progress through prioritization can be followed on PCORI’s website. Following Advisory Panel input, PCORI’s Board of Governors may assign questions prioritized highly either for development as specific (i.e., “targeted”) funding announcements or for placement on a list of highpriority questions included in PCORI’s Pragmatic Clinical Studies initiative. Both the targeted announcements and Pragmatic Clinical study awards fund relatively large studies. Awardees are expected to actively involve relevant stakeholder organizations in the research process and to attend to possible differences in relative effectiveness and harms across patient subgroups in their analyses. The entire process, from identification of study topics to the conduct of research, is uniquely stakeholder-driven.

THE ROLE OF COMPARATIVE CLINICAL EFFECTIVENESS RESEARCH IN THE NATION’S MEDICAL RESEARCH ENTERPRISE CER addresses genuine uncertainty at the point of clinical or health system decision-making; it aims to provide the clinical information needed to move forward on an individual, practice, system, and/or policy level. As such, CER is not the first design employed in a new area of inquiry (Fig. 19.2) and only becomes relevant when there are at least two viable options being considered and an important choice to be made with respect to preventing, diagnosing, treating, or managing illness. By contrast, “discovery” or traditional biomedical research covers the initial steps in a new area and is the domain of the NIH in the United States along with the life sciences industry. These activities range from understanding the genetic, molecular, and environmental mechanisms underlying health and illness to the more applied areas of developing new pharmaceuticals, devices, procedures, or diagnostics based on this understanding. Epidemiologic inquiry at this stage can provide insights that lead back to the basic sciences or forward to new approaches for preventing, diagnosing or treating illness. Once new approaches are identified, further knowledge acquisition develops in a relatively orderly, although not always predictable sequence, depending on the type of innovation. For drugs and medical devices, the path to approval involves oversight and review by the US Food and Drug Administration (FDA). For pharmaceuticals, next steps after initial proof-of-concept and safety/dosing studies involve placebo-controlled efficacy studies sponsored largely by industry under FDA guidance and scrutiny. Efficacy studies tend to focus on one or more “primary outcomes” and often leave other

NIH Industry Academia

FDA CMS Discovery

Patients Specialties Payers

Clinical and Healthcare Policy

PATIENT-CENTERED OUTCOMES RESEARCH INSTITUTE

Regulation/ Approval

Comparative Clinical Effectiveness Research (CER)

3

FIGURE 19.2 PCORI’s role in National Clinical Research Enterprise.

II. STUDY DESIGN AND BIOSTATISTICS

274

19. COMPARATIVE EFFECTIVENESS

outcomes unstudied. Sample sizes and duration of follow-up are calculated for the primary outcomes. Studies are planned to reach a conclusion as early as possible. Study populations are typically narrowly selected, often for having a high risk for the primary outcome, but relatively limited potential for harms. Extremes of the age spectrum and individuals with comorbidities often are excluded from the study population. Left unanswered is how well study findings might generalize to the broader range of patients to whom they will be applied in practice, who may be at much lower risk of disease or higher risk of adverse effects. Patients with comorbid illnesses along with the disease being studied also are often excluded, further reducing the generalizability of preapproval trials. Smaller sample size and relatively short follow-up commonly fail to detect rarer adverse effects or those which occur only after longer periods of time. For new medical devices, pathways also typically involve industry sponsorship with FDA oversight and approval, but the pivotal studies are unlikely to involve placebo-controlled trials and often do not involve randomization. New procedures and health system interventions develop even less systematically and without FDA involvement. New procedures may spread into practice with or without case series publications. System-level interventions often incorporate multiple components, some proven effective previously and others unproven, into new efforts such as populationmanagement programs, but full programs are rarely evaluated before being implemented and it is difficult to know what components are critical for the success of the intervention. The final step in reaching practice for most new interventions (pharmaceuticals, devices, procedures, and system-level programs) is the decision on whether health-care payers, both public (Centers for Medicare and Medicaid Services [CMS]) and private (commercial insurers and self-insured employers) will cover the costs for these products. For legislative reasons, CMS currently enjoys less leeway than other payers in deciding on whether to cover new pharmaceuticals. For all other types of interventions, all payers must make these decisions. In situations where there are existing treatments or approaches, CER can be a valuable tool for making informed decisions about new interventions that help to maximize health benefits, and reduce harms as well as waste. Occasionally, findings from preapproval efficacy studies are so strong that they provide a sufficient basis for making evidence-based clinical and coverage decisions. A recent example is the emergence of the new direct-acting antiviral agents for the treatment of hepatitis C (hepatitis C virus (HCV)). In preapproval efficacy studies among patients with moderate to severe

HCV-related liver disease, these agents were so superior to older agents in terms of eradicating the virus, preventing progression of disease, and freedom from serious adverse effects that clinical guidelines changed immediately17 and insurers quickly covered and promoted treatment for this group of hepatitis C patients. In other cases, the comparative effectiveness of new interventions (vs. earlier approaches) is not clear when initial coverage decisions must be made. There may be little or no evidence of how new interventions work, in terms of either effectiveness or safety, in patients who differ from those in the preapproval trials, such as older patients, those with comorbid illness, those with milder illness, or those in different clinical settings than where the approval studies were conducted. There is rarely any direct evidence comparing the new interventions with previously available therapies in these patient groups. Thus, new interventions in a clinical area with established therapies or practices are a prime subject area for CER. These practical questions often require larger, longer CER studies, considering a greater range of outcomes and may require several years to complete. Payers, as well as patients and clinicians, face a period where decisions must be made on minimal evidence. Systematic reviews based on the multiple, usually small, preapproval trials or that indirectly compare findings from such studies to findings from studies of older, available therapies, may provide some assistance. However, the lack of more complete evidence on how the new interventions will perform in broad populations and everyday settings or on how they affect a range of outcomes not included in preapproval efficacy studies inevitably leaves many unknowns at this point. Decision-makers also turn to modeling of the natural history of illness and of the effectiveness and safety of new versus older treatment options to estimate comparative effectiveness and support initial decisions regarding coverage and pricing. Simulation models may help inform patient and clinician decision-making by presenting quantitative information on possible trade-offs. However, the utility of these models depends greatly on the model structure and the appropriateness of its assumptions. A second useful by-product may be a clearer appreciation of the critical “evidence gaps,” the CER needed to reduce uncertainty in preliminary assessments of comparative benefits and harms. In situations where marginal differences in effectiveness (benefits and harms) are either demonstrated or estimated to be small but costs differences are considered to be large, cost-effectiveness, and cost-utility models are sometimes useful.18 These models relate differences in overall effectiveness to differences in costs across comparative options using a common metric, such as cost per quality-adjusted life year gained.

II. STUDY DESIGN AND BIOSTATISTICS

THE METHODS OF COMPARATIVE CLINICAL EFFECTIVENESS RESEARCH

Cost-effectiveness approaches have more recently been adapted and relabeled as “value assessments,” and used to recommend pricing levels linked to the perceived value of a treatment.19 Again, unless these models have accurate and complete CER information, the value estimates will remain tenuous. The interface between those who fund and conduct CER, those who build and report on simulation models, including costeffectiveness and value, and those who must make clinical and policy decisions is a critical locus for meaningful dialog and for generation of applied CER in the interest of patients.

THE METHODS OF COMPARATIVE CLINICAL EFFECTIVENESS RESEARCH The research methods of CER are not fundamentally different from those used in etiologic research, safety research, or efficacy studies. However, the goals of CER and of pragmatic research place novel demands on traditional clinical research methodologies. To support decision-making, at either the individual or population level, more attention must be paid to identifying the right research question, the question that could change decision-making and practice. To obtain study results that can be generalized to broad patient populations, more heterogeneous populations must be studied, with fewer exclusion criteria and greater participation rates than in most clinical research. Studies must be conducted under as near real-world conditions as possible, blending research more seamlessly into clinical care rather than isolating it in specialized research units. Larger sample sizes often are needed in CER. The differences to be detected (or ruled out) between two active treatments are likely to be small but nonetheless important. The level of certainty needed to justify changes in clinical care choices demands a precision that depends in part on larger sample size. The broader, more heterogeneous study populations of CER usually represent a broader range of expected treatment effects; thus CER studies demand sample sizes large enough to examine treatment heterogeneity and allow for detection of possible subgroup differences, meaningful benefits and significant harms. The challenges of making research more pragmatic create a fertile area for methods research and development, in both observational and experimental CER. Recognizing this need, the AHRQ has funded a substantial body of methods research, beginning with work conducted by its Evidence-based Practice Centers and strengthened in its DEcIDE Program.20 This work has created a strong foundation for further CER development in the United States and beyond (see Text Box). AHRQ convened a series of symposia beginning in

275

2006 to build the methods research community and disseminate findings. Proceedings of three symposia have been published21e23 and stand as important collections of methodological work aimed at advancing CER methods. These symposia focused almost exclusively on observational CER methods. More recently, PCORI’s Methodology Committee met its legislative mandate by developing, posting for public comment, revising, and publishing in 2012 an initial set of Methodology Standards24 for conducting patientcentered comparative clinical effectiveness research. PCORI’s Methodology Standards are regularly updated or revised by the Methodology Committee, which also is engaged to develop new standards as needed to support innovations in CER. Standards (Table 19.3) address a broad set of challenges to decision-support research and clarify that experimental as well as observational approaches to primary data collection as well as evidence synthesis are fundamental components of CER methodology. PCORI requires all applicants to adhere to these Methodology Standards in developing and submitting research proposals and in conducting the research. Merit reviewers evaluate submitted research proposals against the Methodology Standards and Peer Review of final research reports assesses adherence to the standards in the conduct of PCORI-funded CER. Training materials and continuing medical education materials have been developed to aid in dissemination of these standards and are available on PCORI’s website.25

Getting the Research Question Right Research that can provide guidance for clinical decision-makers, change practice, reduce practice variation, and increase the quality of health-care decisions demands research questions that are carefully considered and constructed in collaboration with those who will use the study’s findings. Poorly constructed comparisons may be “interesting” and may add to clinical knowledge incrementally, but they are unlikely to support improved decision-making and may confuse choices, even if the studies are well conducted. PCORI’s Methodology Standards (summarized in Table 19.3) include six recommendations intended to aid in specifying research questions. The standards first emphasize a careful and systematic approach to identify true research gaps, including use of systematic reviews and clinical guidelines as sources of important unanswered questions. Both systematic reviews and clinical guidelines development efforts typically conclude with a set of unanswered questions. Especially in the case of clinical guidelines, the answers to these questions would likely be used to refine guidelines and change

II. STUDY DESIGN AND BIOSTATISTICS

276

19. COMPARATIVE EFFECTIVENESS

TABLE 19.3

PCORI Methodology Standards

CROSS-CUTTING STANDARDS FOR PCOR 1

Standards for Formulating Research Questions Six standards that specify what to include in research protocols as a means of increasing study quality as well as transparency in research

2

Standards Associated with Patient-Centeredness Four standards that promote effective patient engagement and the explicit incorporation of patient needs, values, and preferences into research

3

Standards for Data Integrity and Rigorous Analyses Six standards that describe necessary documentation of key decisions and tests of the assumptions made in analyses

4

Standards for Preventing and Handling Missing Data Five standards outlining proper statistical methods for handling missing data

5

Standards for Heterogeneity of Treatment Effects Four standards on how to account for the fact that different people do not always respond the same way to the same treatment

STANDARDS FOR SPECIFIC STUDY DESIGNS AND METHODS 6

Standards for Data Registries Three standards to help ensure that registries contain relevant, highquality data that are used appropriately

7

Standards for Data Networks as Research-Facilitating Structures Two standards to help ensure that key components are included in network design and considered when network data are used in studies

8

Standards for Causal Inference Methods Six standards on accounting for possible sources of bias and addressing them to produce valid conclusions about the causal effect of an intervention

9

Standards for Adaptive and Bayesian Trial Designs Five standards providing guidance on the design and conduct of studies that use such designs

10

Standards for Studies of Diagnostic Tests Five standards that address studying the impact of diagnostic tests on subsequent care and patient outcomes

11

Standards for Systematic Reviews One standard that outlines the application of standards for systematic reviews

practices. The standards emphasize the importance of specifying all aspects of the question necessary for it to be relevant and actionable, generally represented by the PICOTS acronym (Population/patients, Interventions, Comparators, Outcomes, Timing, Setting). We comment further on each of these in the following paragraphs.

Choosing the Study Population The Methodology Standards stress the importance of selecting the appropriate study population given the

question. A study population too narrowly defined (e.g., only very high-risk patients) offers little useful information for the broader, more typical patient populations in whom the treatment will subsequently be considered and promoted. Yet, many such narrowly focused studies are conducted and reported, leaving patients, clinicians, and insurers to wonder how well the study results apply to themselves or the majority of patients. Conversely, a study population can be defined too broadly and fail to answer useful questions for anyone. If it includes substantial numbers of patients in whom the treatment would rarely be considered or who would not be expected to benefit from treatment, it is likely to underestimate or fail to detect true benefits even for large patient subgroups. The Methodology Standards also urge CER investigators to consider key subgroups of interest in advance when defining the study populations and to power studies so that comparative effectiveness may be assessed within these subgroups. A critical role of CER is, in fact, to define the optimal prevention, diagnosis and treatment strategies for individuals within heterogeneous populations, as defined by demographic, clinical, and genetic characteristics. There will nearly always be trade-offs to consider between the breadth of the study population, the utility of findings to various audiences, and the costs or feasibility of the study. These trade-offs will require careful consideration and prioritization among stakeholders and may ultimately mean that multiple studies will be needed to fully address a clinical or policy question.

Selecting Appropriate Interventions and Comparator(s) Selecting the most appropriate comparator(s) in CER studies is among the most difficult choices to be made. CER aims to address gaps in knowledge about the relative benefits and harms of selected interventions compared with other available and legitimate approaches. For clinical treatment with pharmaceuticals, devices and procedures, the decision has usually been made that something needs to be done. Comparing the treatment to doing nothing would not be helpful. Thus, comparisons are almost always with “active” alternatives. Sometimes the need for direct head-to-head comparisons is obvious (e.g., two new drugs intended for the same patient population with potential differences in effectiveness, acceptability, adverse effects or costs; a new agent vs. the current standard of care; or surgery vs. optimal nonsurgical care). But in many more instances, careful thought is required to ensure that the comparator represents a realistic treatment option for the study population. Comparators that will be

II. STUDY DESIGN AND BIOSTATISTICS

THE METHODS OF COMPARATIVE CLINICAL EFFECTIVENESS RESEARCH

outdated by the time study results are available as well as comparators considered to be substandard care are not useful in CER. Alternatives must correspond to genuine choices faced by patients, clinicians, or delivery systems and should not include “straw man” choices with known or suspected imbalances (i.e., the highest effective dose of one agent vs. lower doses of another). For prevention and screening measures, it is more common to compare preventive interventions versus no or placebo interventions, simply because the “higher standard” of proven to be better than no action (first do no harm) must be met first. CER has less commonly been a part of traditional considerations in clinical prevention. However, the United States Preventive Services Task Force (USPSTF) has found some instances, such as colorectal cancer, cervical cancer, or breast cancer screening, where there are now two or more viable screening choices (e.g., fecal occult blood tests vs. endoscopy, HPV testing vs. pap smears) that may be compared. CER also is useful for comparing rescreening intervals and alternative follow-up diagnostic approaches in those who screen positive. System-level interventions are common concerns among those who nominate and prioritize CER questions, since failure to consistently deliver what is already known to be beneficial is widespread in the United States and other countries. However, the relevant interventions to be compared do not align strictly with either treatment-related or preventive services. Systems may need to know whether organizing care (differently), changing the mix of health-care personnel, altering hours of operations or modifying payment schemes will improve the delivery of, outcomes from, or satisfaction with clinical services. Some system-level interventions are composed of multiple components. Single components may have been shown to be effective previously, but the combined intervention may not have been tested. Comparisons of different combinations may then be most useful. In other cases, the optimal direct comparisons are of two “active” interventions (e.g., telecare vs. face-to-face mental health care; email vs. telephone health coaching; multispecialty referral clinics vs. integrated primary care services). However, the most relevant and viable comparator in many instances is continuing business as usual versus implementing a system change. “Usual care” then becomes the most appropriate comparator. Because “usual care” has been shown to vary widely between settings, it is essential to carefully define its components in advance and to monitor the delivery of that care during the study. A CER study that cannot fully describe what happened in both comparator arms fails to answer the “compared-towhat” question and is difficult to interpret, replicate, or act upon.

277

Choosing Clinical Outcomes to Be Measured To support decision-making, CER studies usually include a set of clinical outcomes that is larger and measured at longer, more clinically meaningful timepoints than in efficacy studies. In efficacy studies, simply proving that a new treatment is better than placebo over the short term for a primary outcome may fully meet study goals of obtaining FDA approval. In practice, patients and clinicians often need information for a much broader range of legitimate outcomes, and over longer periods of time. These often include outcomes reported directly by patients such as fatigue, depression, functional capacity, or symptoms specific to the illness. Individual patients vary in the outcomes of most importance for decision-making. In any CER study, it is highly desirable to use standard outcome measures, measures that have been validated and reported on in previous studies, whenever possible so that study results may be compared with those of earlier studies and easily included in research syntheses and metaanalyses. For example, the PatientReported Outcomes Measurement Information System (PROMIS) is a part of a set of publicly available, standardized, general, and symptom-specific item lists developed for use in clinical research.26 Many diseasespecific standard instruments also are emerging. An important frontier in CER and clinical medicine is identifying the full set of meaningful outcomes that should be collected for research on a specific condition.27,28

The Role of Engagement in Specifying Research Questions Specifying all these aspects of CER questions presents trade-offs between desire for rich information, study costs, and burden to participants. Various stakeholders often disagree initially on the optimal research question and some compromise may be needed. A steady focus on the most urgent needs of patients and clinicians and on practical aspects of studying, the research question is important for reaching agreement. Studies designed with key research stakeholders absent during the planning phase are much more likely to face disagreement on the implications, relevance, or utility of findings when they become available. PCORI has stressed the need for “engaging” all relevant stakeholders in every aspect of the research process, beginning with the selection of research questions, but extending to the preparation and review of study proposals, and conduct of the research. PCORI’s Methodology Standards include guidance on engaging stakeholders “as appropriate.” The goals of engagement are to ensure that the study will ask a

II. STUDY DESIGN AND BIOSTATISTICS

278

19. COMPARATIVE EFFECTIVENESS

question relevant to stakeholder needs, be conducted with fidelity to the original question, and be reported in ways that prove useful for these key stakeholders and serve dissemination. Both individuals and organizations add value and relevance to the research process. Individual patients and caregivers bring the “lived experience” of those with the condition, providing qualitative insight into the relative importance of various symptoms, treatment side effects, longer term outcomes, and feasibility and acceptability of proposed treatment approaches. Advocacy organizations for specific conditions bring a broader awareness of the entire spectrum of the condition, from prevention and early detection to palliative and end-of-life care for conditions that are lifethreatening. They also can bring awareness of clinical and public policy issues that raise important CER questions or affect the feasibility of new treatment approaches. Clinician organizations responsible for developing evidence-based clinical care guidelines in a specialty area and payer organizations responsible for making coverage decisions for new treatments are especially well-positioned to recognize important comparative questions. Representatives of the life-sciences industry can raise questions and provide insights about the comparative effectiveness of their products. Other funding agencies, particularly AHRQ, introduced elements of stakeholder engagement before PCORI was created, especially for the prioritization of research questions and dissemination activities.29 The FDA, the NIH, and industry research have all followed PCORI’s lead and increasingly are emphasizing engagement, especially patient engagement, in setting research priorities. However, none but PCORI currently require engagement consistently in all aspects of the research process or provide resources to facilitate the experience. An engagement rubric has been developed by PCORI to assist applicants and other researchers in selecting and implementing suitable engagement strategies and taxonomy of stakeholder engagement has identified the stages of research where engagement may be employed.30,31 Many engagement strategies are described in the rubric and found among PCORI-funded projects. Although the practice of engaging patients and other stakeholders appears sensible, the evidence supporting its benefits and defining its methods is just beginning to appear. It is likely that many strategies may be interchangeable, and best strategies have not been identified. A systematic review covering 66 papers published before 2009 found a positive impact of engagement of patients or the public on all stages of health-related research. Improvement areas included development of user-focused research objectives, user-relevant research questions, user-friendly study information, questionnaires and interview schedules, as well as

more appropriate recruitment strategies, consumerfocused interpretation of data, and enhanced implementation and dissemination of results.32 Most of the studies reviewed were qualitative studies. A more recent systematic review of engagement in CER considered 70 studies published from 2003 to 2012.33 In this review, engagement with patients was much more frequent than with clinicians, payers, or other stakeholders. Engagement also has been reported less frequently for the later stages of the research processdconducting, interpreting, and disseminating researchdthan for earlier stages such as identification and prioritization of research questions.34 It is likely that gaps in the continuity of engagement within projects would decrease chances that the study’s findings would be accepted and implemented by all relevant stakeholders. Further evidence is needed to establish the broad effectiveness of stakeholder engagement, to compare strategies and identify those that are most effective and efficient, and to understand the impact of engaging stakeholders other than patients, particularly clinicians, systems leaders, and payers (i.e., insurers and employers). Evaluation outcomes should include recruitment and retention rates in CER studies and ultimately the dissemination and implementation of study findings.

STUDY DESIGNS FOR CER STUDIES A range of analytic study designs can be used for conducting CER (Fig. 19.3). Limitations related to validity, feasibility, utility, or costs affect each design, and no single approach can adequately address the full range of CER questions faced by patients, clinicians, payers, policy-makers, and industry. However, the possible study designs can be used complementarily to build useful information in a specific research area. Choosing the most appropriate study design for a specific question requires consideration of the question itself, the previously available evidence, each design’s known weaknesses as well as its possible advantages in terms of time, cost, validity, and acceptability to patients. Methods development in CER often seeks to reduce or ameliorate or offset known deficiencies of a particular study design or enhance its benefits. The first and most critical decision in selecting the study design is whether the CER study will require random allocation of the intervention or whether a nonrandomized (observational) study of the question could adequately adjust for the biases of confounding by indication or self-selection.35 Clinicians select treatments for individual patients based on their estimate of the patient’s risks, higher-risk patients getting earlier or more aggressive preventive, diagnostic, or treatment

II. STUDY DESIGN AND BIOSTATISTICS

279

STUDY DESIGNS FOR CER STUDIES

Is a randomized design necessary?

Individual Randomized Controlled Trial

Observational Study Designs Cohort Studies Prospective Retrospective

Cluster Randomized Trial

Case-Control Studies

Stepped Wedge Designs??

Area Variation Studies* Instrumental Variable Analyses*

Chance of Residual Confounding

Chance of Residual Confounding

Experimental Study Designs

* Although confounding at the level of the individual patient is greatly reduced, the possibilities of confounding at the level of the small area or the instrument must be considered.

FIGURE 19.3 Experimental and observational study designs in CER.

strategies while lower-risk patients get less aggressive choices. Patients also are selective. Healthier patients, those with healthy behaviors, or those already concerned about certain risks may each be more likely than others to use preventive measures, undergo more intensive or frequent screening or choose one treatment approach over another. These tendencies are strong, and they bias observational comparisons of outcomes when treatments have assigned by clinicians or chosen by patients (as in most nonrandomized studies). Patients in the comparator groups already differ at baseline in ways that predict their outcomes, even if the interventions are exactly equivalent. Efforts to measure and adjust outcomes analyses for known differences between patient groups are a central part of observational research, but in many instances remaining concerns about patient differences undermine confidence in their findings and the impact of findings on clinical practice. When observed differences in outcomes are “small” (e.g., m0 at the a ¼ 5% significance level, we compare the observed z test statistic value to the critical value 1.645 (1.645 if HA:mx < m0), which is associated with the upper (lower) 5% of the normal distribution, and if the test statistic exceeds this value we reject the null hypothesis. Many statisticians will use a tail probability of 2.5% to define the rejection region, using the associated critical value of 1.960 (or 1.960), even if the test is one-sided to not make one-sided tests “easier.” This approach has the advantage of keeping the hypothesis test consistent with the associated 95% CI for the statistic. Confidence Intervals Another way to evaluate evidence is by using a CI. When a ¼ 0.05, we use a 95% CI. For general a, a 100  (1 e a)% CI for a population parameter is formed around the point estimate of interest. The most basic CI is that for the mean, m. If variance is known, the CI has the following formula:   z1a=2 s z1a=2 s x pffiffiffi ; xþ pffiffiffi . (24.9) n n By contrast, if the variance is unknown, then the sample standard deviation sx is used instead and the t critical value is used instead of the corresponding z value. In most cases, the variance is not known and must be estimated. Hence, the t-statistic is commonly reported for continuous data and the Student t-distribution is used to determine the critical value. For large sample sizes the critical values for the Student t and the normal distribution will be nearly identical, and so sometimes in practice the latter is used even for a t-statistic. There is an important parallelism between hypothesis testing and CI construction. Specifically, if the hypothesized population parameter falls within the CI, we do not reject the null hypothesis. For a 95% CI this is similar to performing a two-sided test at the a ¼ 5% significance level. z Tests or t Tests The choice between the t and z tests can be important. Although some people will switch to the normal z critical value as soon as sample size looks slightly large (e.g., n > 30), doing so can be problematic. If we look at the Z and T values at the 0.975 percentile, the upper end of a 95% CI, at df ¼ 30 the T value is 4% larger than the normal value. At df ¼ 120, there is still a 1% difference between the t distribution and the normal distribution. This may seem silly, and many times the

347

difference between these two distributions and their associated tests will not matter; a test is highly significant or nonsignificant, but in general it is best to use a Student’s t distribution if indeed that is what the test and data warrant. If the variance is unknown, which it almost always is, and we have continuous data, the t-test is recommended. Suppose an investigator wishes to determine whether pediatric anesthesiologists have unusually high serum Erp 58 protein levels. This protein is associated with industrial halothane hepatitis. Suppose that he/she collects n ¼ 9 blood samples, and the sample mean and standard deviation of the protein levels are x ¼ 0.35 and sx ¼ 0.12 (optical density units), respectively. If the mean protein level is over 0.28, it will suggest that further study is needed. He/She chooses a one-sided a ¼ 5% significance level. This hypothesis test corresponds to H0:mx  0.28 versus HA:mx > 0.28. Using formula Eq. (24.8) one can calculate T¼

0:35  0:28
pffiffiffi ¼ 1:75; 0:12 9

(24.10)

which is less than the 95% percentile, t1ea,8 ¼ 1.860, of Student’s t distribution with n e 1 ¼ 8 df (p-value ¼ .06). Thus, he/she does not reject the null hypothesis, although he/she may wish to collect a larger sample to explore this question further. Note, had the normal percentile z0.95 ¼ 1.64 been used, the null would have been rejected. In practice, one would collect a larger sample and use a more advanced method such as multiple regression to adjust this hypothesis test for important covariates such as age, gender, work experience, body mass, and medical history.9

Binary Data Just as we can perform hypothesis tests on continuous data, we can perform them on proportions estimated from binary data. Binary or dichotomous outcomes are common in medical research. Binary or dichotomous data have two possible outcomes, such as success or failure, presence or absence of disease, or survival or death. When each observation is scored as a 1 (success) or a 0 (failure), then the average across of the data is simply the proportion of successes. Typically for binary data, a hypothesis test will be performed for the sample proportion(s). A quick note on notation: Eq. (24.4) and most statistical textbooks use Greek letters for population parameters, such as p for the true population proportion. Several former course participants have mentioned the mathematical constant p (w3.14) is what jumps to mind when seeing p in this section. With apologies to statisticians, in an attempt to help these readers the

II. STUDY DESIGN AND BIOSTATISTICS

348

24. HYPOTHESIS TESTING

rest of this chapter uses p for population proportions and b p for sample estimates of a proportion. Developing a Test There are a variety of different tests that can be used with binary data, including the z test, continuity corrections to the z test, and exact tests. Let p1 denote a population proportion, let b p 1 denote a sample estimate of that proportion, and let p0 denote that proportion’s value under the null hypothesis. To test the two-sided hypothesis H0: p1 ¼ p0 versus HA: p1 s p0 (or a corresponding onesided hypothesis), we can consult the reference distribution for a sample size n (the binomial) and write down the test statistic following the formula in Eq. (24.1), b p 1  p0 ffi Z ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p0 ð1 p0 Þ=n

(24.11)

which statistical theory (the central limit theorem) tells us has approximately the standard normal distribution for large enough sample sizes (n > 25, np0 > 5, and n(1 e p0) > 5). If this test statistic falls in the extreme percentiles of the standard normal distribution (or beyond the appropriate lower or upper percentiles for a one-sided test) we can reject the null hypothesis. For small to modest sample sizes, the normal distribution will only give an approximate p-value for the test statistic in Eq. (24.11) due to the discreteness of the data. We can improve the approximation by adding a small sample continuity correction or by performing an exact test. See Altman5 for details. Exact Tests If we have sufficient computing power we can perform an exact binomial test rather than using the normal distribution to approximate the sampling distribution for the binary test statistic and the resulting pvalue. In an exact binomial test, we enumerate the true binomial probabilities for each of the possible numbers of successes or events (0, 1, 2,., n) and then reject the null hypothesis when the sum of the probabilities for values as extreme or more extreme than the observed value is less than the significance level. For example, suppose that the null hypothesis is H0: p ¼ 0.35 and n ¼ 6; under this hypothesis, the true binomial probabilities of observing exactly 0, 1, 2, 3, 4, 5, and 6 events out of six trials are 0.075, 0.244, 0.328, 0.235, 0.095, 0.020, and 0.002, respectively. Thus, if the alternative hypothesis is HA:p < 0.35 and we observe five events, the one-sided pvalue is 0.020 þ 0.002 ¼ 0.022. There are a few ways to define the two-sided p-value even just using exact probabilities. Perhaps the simplest and most common method is to calculate the one-sided p-value for the tail in the direction of the observed data and double it. Thus, if the alternative hypothesis is HA: p s 0.35, the

two-sided p-value is 0.044. In both of these examples, we would reject the null hypothesis at the 5% significance level but not at the 1% significance level. Confidence Intervals Similar to constructing the test statistic, binomial CI construction can make use of a normal approximation but improvements can be made. The normal approximation methods for binomial CI construction tend to produce CIs that are too small on average and thus have lower coverage rates than the specified confidence levels. In other words, even though we calculate something called a 95% CI, in truth less than 95% of the time that interval contains the true value of interest. For binomial data we need a better interval method. One classical approach for obtaining better binomial CIs is the ClopperePearson method,10 which uses exact binomial probabilities to give CIs which are appropriate for all sample sizes. The ClopperePearson CIs consist of all proportion parameters that are consistent with the observed binomial data at a particular significance level using the exact binomial test with a two-sided hypothesis. Most statistical software can easily provide the ClopperePearson exact confidence bounds for proportions. Several other methods also exist and which one is used can make a difference. For example, with n ¼ 60 trials and x ¼ 15 successes or events, the (1) Wald or normal approximation, (2) ClopperePearson, (3) AgrestieCoull,11 and (4) SAIFS12 methods give 95% CIs of (1) (0.140, 0.360), (2) (0.147, 0.379), (3) (0.157, 0.373), and (4) (0.137, 0.374), respectively. These CIs are all close, but if we are near one of the boundaries the method used may really matter. A statistician can help implement these improved methods for binomial CI construction, which essentially build on the conceptual framework presented in this chapter.

Example Suppose that in response to complaints about allergies, a large hospital changes the standard brand of rubber gloves that it supplies to a new but more expensive brand. An administrator wishes to know what proportion of nurses in that hospital prefers the new gloves, p1, and if that proportion is at least p0 ¼ 40%, he/she will consider the change worthwhile. He/She chooses a onesided significance level of a ¼ 5%. This hypothesis test corresponds to H0:p1  0.4 versus HA:p1 > 0.4. He/She finds that out of a sample of 30 nurses, 18 prefer the new brand, and the rest are indifferent. Hence n ¼ 30, b p 1 ¼ 18/30 ¼ 0.6, and using the previous formula, 0:6  0:4 Z ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 2:24; ð0:6 0:4Þ=30

II. STUDY DESIGN AND BIOSTATISTICS

(24.12)

TWO-SAMPLE HYPOTHESIS TESTS WITH APPLICATIONS TO CLINICAL RESEARCH

349

which exceeds the 95% percentile, Z1ea ¼ 1.645, of the standard normal distribution (p-value ¼ 0.01). By comparison, using an exact binomial test the p-value is 0.02. Thus, he/she rejects the null hypothesis and decides to adopt the new brand of gloves. Since this is a one-sided test, using a one-sided 0.025 level test and comparing the exact binomial p-value to 0.025 (not 0.05) or the result of Eq. 24.12 to Z1ea ¼ 1.96 (as opposed to Z1ea ¼ 1.645) would have been preferred to stay consistent with using a two-sided 95% CI. Indeed, examining two-sided 95% CIs he/she finds similar results using the normal approximation or Wald (0.425, 0.775), ClopperePearson (0.406, 0.773), AgrestieCoull (0.423, 0.754), and SAIFS (0.404, 0.787) methods. Although these improved CI methods may seem to create a little extra work, they can be crucial when the binomial test statistic lies near the boundary of significance.

We need to make modeling assumptions to set up a hypothesis test. This will allow us to develop a sampling distribution for the test statistic under the null hypothesis that the mean difference is 0 (i.e., there is no effect of beta-interferon). We assume that the differences for each patient are independent and normally distributed from a population with mean md and variance s2, where d is the mean of the differences on all n subjects. When s2 is known, the test statistic

TWO-SAMPLE HYPOTHESIS TESTS WITH APPLICATIONS TO CLINICAL RESEARCH

This test statistic has Student’s t distribution with n  1 df under the null hypothesis. Before we begin the study, we choose a significance level (i.e., what type I error amount or cut off will we use at the end of the study to determine if the amount of evidence we need to reject the null hypothesis is or is not present in the data). If the test statistic’s value is in the lower or upper a/2  100 percentiles of the reference distribution we reject the null hypothesis and conclude that the means in the two groups (in this case the pretreatment and treatment groups) are not equal. If the test statistic is not in the extreme tails of the distribution we conclude that we fail to reject the null hypothesis, and hence that there is insufficient evidence to conclude that the means in the two groups are different. The p-value is the probability of observing a test statistic value larger (in magnitude or absolute value) than what one observed. Suppose the observed value is Tobs, and let T denote a random Student t variable. Then the p-value is P(t < eTobs) þ P(t > Tobs) for a twosided test. The p-value for a one-sided test with alternative hypothesis HA:md > 0 is P(t > Tobs). Tests based on Z and T test statistics values are called paired z-tests and paired t-tests, respectively. Similar to what is discussed in the z Tests or t Tests section, a paired z test is used when s2 is known, a paired t-test is used when s2 needs to be estimated from the data, and we choose the t-test when wondering which of the two to use.

The goal of many clinical studies is a comparative one, such as comparing the treatment response between two groups. Here, we develop hypothesis tests for comparing the means of two normal populations in both paired and unpaired analyses. We also discuss hypothesis tests for comparing two population proportions. These tests then will be used to analyze the data from the motivating examples in the next section.

Tests for Comparing the Means of Two Normal Populations Paired Data We first consider the hypothesis test appropriate for a pair of outcomes. This analysis corresponds to the betainterferon/MRI trial in which measurements on each patient are observed both before and during treatment. In this situation, we have two observations on every patient from which we can compute the difference di ¼ xi e yi. The data consist of n differences: d1, d2,., dn where n is the number of subjects in the study. In the beta-interferon/MRI study, n ¼ 14. The observations xi and yi correspond to suitably transformed individual mean monthly lesion counts during the baseline period and during the active treatment period for the ith subject. As discussed in general in Hypotheses for the Beta-Interferon/Magnetic Resonance Imaging Study section, the hypothesis we will be testing is H0 : md ¼ 0 vs. HA : md s0:

(24.13)

Z¼

d pffiffiffi s= n

(24.14)

has the standard normal distribution under the null hypothesis. When s2 is unknown (as is common in most situations in medical statistics), we need to estimate the variance s2 from the data. When s2 is unknown, the test statistic is vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u n  u 1 X 2 d T ¼ pffiffiffi; where s ¼ t di  d . (24.15) n  1 i¼1 s= n

Unpaired Data We next consider tests of two normal population means for unpaired data. We discuss the cases of equal

II. STUDY DESIGN AND BIOSTATISTICS

350

24. HYPOTHESIS TESTING

variances and different variances separately. We begin with a discussion of the equal variance case. The example that corresponds to this test is the felbamate monotherapy trial, and it is similar to many other parallel groups designs. We assume that we have observations from two groups of subjects, with sample sizes n and m. We assume that the observations x1, x2, x3,., xn and y1, y2, y3,., ym come from two independent normal distributions with a common variance s2 and population means m1 and m2, respectively. The hypothesis test for this situation is H0 : m1 ¼ m2 vs. HA : m1 sm2 :

(24.16)

We calculate the difference in the two sample means and follow Eq. (24.1) again to write down the test statistic using the sampling distribution for the difference of two independent sample means. When s is known, the test statistic of interest is xy Z ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi s 1=n þ 1=m

(24.17)

which has the standard normal distribution. When s2 needs to be estimated from the data we calculate the pooled variance estimator to put in to the test statistic, xy T ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi; where s 1=n þ 1=m vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u m  n X uP 2 u ðx1  xÞ2 þ y1  y u ti¼1 i¼1 . s¼ nþm2

(24.18)

which has the Student’s t distribution with n þ m e 2 df under the null hypothesis. The preceding estimate of s is the pooled sample standard deviation and is based on the assumption of equal variances in the two groups. As in the previous hypothesis test, if the Z and T test statistics values are in the lower or upper a/2  100 percentiles of this reference distribution, we reject the null hypothesis. These tests based on the Z and T test statistics values are called two-sample z-tests and two-sample t-tests, respectively. Two-sample z-tests are used when s2 is known, and two-sample t-tests are used when s2 needs to be estimated from the data. In many situations, the assumption of equal variance in the two treatment groups is not a good assumption. Since treatments may be effective in only a fraction of subjects, often the variability of the outcome in the treatment group is larger than that of the placebo group. The test statistic to use in this situation is xy Z ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi . ffi;
s2x n þ s2y m

(24.19)

where both the sample sizes in each group are large or when the variances are known. The Z test statistic has the standard normal distribution when the null hypothesis is true. When the variance estimates are unknown and need to be estimated using the data, the test statistic is xy T ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

. s2x n þ s2y m

(24.20)

Under the null hypothesis that the means in the two groups are equal, the preceding test statistic has a distribution that is approximately the Student’s t distribution with u df (determined by Satterthwaite’s formula), where 

2 s2x nþ s2y m degrees of freedom ¼ u ¼ 
  2
2 (24.21) 2 sy m s2x n þ n1 m1 When this result is not an integer, u should be conservatively rounded downward. Luckily, there is software to compute these quantities for us. Generally the applied statistician today only needs to choose the right test and let the computer do the rest of the work. As with the other hypothesis tests we discussed, if the test statistics values are in the lower or upper a/2  100 percentiles of the reference distribution, we reject the null hypothesis and conclude that the means in the two groups are unequal. The t-test with unequal variances often is called Welch’s t-test. Because this test will be valid for both the case of unequal or equal variances, it is the commonly preferred test for the two-sample setting.

Tests for Comparing Two Population Proportions In the ISIS-4 study, the primary outcome was a binary variable signifying whether a randomized patient was alive or dead at 35 days after randomization. Our interest focuses on comparing participants who were randomized to receive magnesium and those randomized to not receive magnesium. The typical data structure for this two-sample problem involves the number of positive responses in n subjects from group 1 and the number of positive responses in m subjects from group 2. The null hypothesis is that the two groups (magnesium, no magnesium) have the same 35-day survival probability. The hypothesis for this test is H0 : p1 ¼ p2 vs. HA : p1 sp2 :

(24.22)

The assumptions for the test are that (1) the data are binary, (2) observations are independent, and (3) there is a common probability of a “yes” response for each

II. STUDY DESIGN AND BIOSTATISTICS

351

HYPOTHESIS TESTS FOR THE MOTIVATING EXAMPLES

of the two groups. For large sample sizes (typically considered n and m both greater than 25, and np1, n(1  p1), mp2, and m(1  p2) each greater than 5), we usually can approximate the discrete sampling distribution with the normal distribution and use a two-sample z-test for comparing the two population proportions. The test statistic is b p1  b p Z ¼ vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   2  ffi u ub b p 2 1 b p1 p2 t p 1 1 b þ n m

(24.23)

which, for large sample sizes, has approximately the standard normal distribution under the null hypothesis that the population proportions are equal in the two groups. Sometimes a pooled variance is used. Other tests have been developed for small samples. For example, the Fisher’s exact test enumerates all possible values of the test statistic under the null hypothesis and can be a valid test for values of the proportions and sample sizes, no matter how small.13

HYPOTHESIS TESTS FOR THE MOTIVATING EXAMPLES We now conduct hypothesis tests to analyze the data from the three motivating examples.

Hypothesis Tests for the Beta-Interferon/ Magnetic Resonance Imaging Study The beta-interferon/MRI study consisted of 14 patients followed for 13 months, 7 months on baseline and 6 months on treatment. The outcome was the average number of monthly contrast-enhanced lesions on treatment minus the corresponding average number during baseline. Table 24.1 summarizes the data from the trial. A total of 13 of 14 patients had decreased lesion frequency on treatment compared with their baseline frequency. This result suggests that beta-interferon lowers disease activity in early RRMS. The inferential question is this: do the data provide enough evidence to make a statement about the population of all RRMS patients? A hypothesis test is used to address this question. We conducted a two-tailed test of whether there is a difference between lesion frequency during baseline and lesion frequency after treatment. We chose a significance level of 0.05 before the study began. First, note that the structure of the data suggests that a paired t-test is appropriate. Data are paired since observations on different patients are independent. The variance of the difference in lesion activity for each subject is unknown. In addition,

TABLE 24.1

Beta-Interferon and Magnetic Resonance Imaging Study

Patient Number

Baseline (Mean Lesions/Month)

6-Month Treatment (Mean Lesions/Month)

1

2.43

0

2

1.71

0.67

3

3.14

1.00

4

1.29

0.33

5

0.57

1.67

6

2.00

0

7

6.00

0.33

8

0.43

0

9

12.86

0.17

10

6.42

0.67

11

0.57

0

12

0.71

0

13

1.57

0.17

14

3.17

1.67

the data transformed to the log scale appeared to be approximately normally distributed. The data were transformed so that di ¼ log[(7-month baseline mean) þ 0.5] e log[(6-month treatment mean) þ 0.5]. The constant 0.5 was added to all numbers, a common practice, since the log of 0 is undefined. We use a paired t-test with a test statistic computed as T¼

d pffiffiffi ¼ 4:8: s= n

(24.24)

The test statistic has a t distribution with 14 e 1 ¼ 13 df when the null hypothesis is true. The a/2  100 (2.5%) lower and upper percentiles of the reference distribution are 2.16 and 2.16, respectively. Since 4.8 is less than 2.16, we reject H0 and conclude that there is a difference between lesion frequency during baseline and lesion frequency on beta-interferon. The p-value for the two-sided test can be computed as P(t13 < 4.8) þ P(t13 > 4.8) ¼ 0.0004, where T13 denotes a random variable with the t distribution on 13 df. This means that if the null hypothesis of no effect was true, there would only be a 1 in 2500 chance of observing a test statistic as large (in absolute value) as the one we observed. The test used was a two-sided test for several different reasons. We care if there is a decrease or an increase in lesions. Investigators should be cautious about using one-sided tests, which are only appropriate when there is interest in detecting a beneficial effect from treatment and there would be no interest in detecting a

II. STUDY DESIGN AND BIOSTATISTICS

352

24. HYPOTHESIS TESTING

harmful effect (or the opposite). This is very rare. Onesided tests done at the 0.05 level also would be awkward when the one-sided test was significant, but the twosided 0.05 test was not significant. This is because the commonly reported two-sided 95% CI is consistent with a two-sided test, thus a nonsignificant two-sided test means the CI would contain values consistent with no difference. This would be in contradiction to the one-sided test at a 5% level that declared the two groups significantly different. Always performing tests, one sided or two-sided, with tail probabilities of a/2 avoids this problem. When debates arise about the use of a oneor two-sided test, typically the use of a/2 two-sided tests wins.

Hypothesis Tests for the Felbamate Monotherapy Trial The felbamate monotherapy trial was designed as a parallel group design with 19 patients randomized to the felbamate arm and 21 patients randomized to the placebo arm. Seizure frequency was monitored during the 2-week follow-up period in the hospital or until a patient dropped out of the study. The outcome was daily seizure rates over follow-up period. The test was a two-tailed test of whether there is a difference in seizure frequency between the felbamate and placebo arms. We chose a significance level of 0.05 before the study began. The hypothesis is H0 : mtreatment ¼ mplacebo vs: HA : mtreatment smplacebo : (24.25) The appropriate test is an unpaired t test. The data are independent and approximately normally distributed on the square root scale (by taking square roots of the mean daily seizure counts on all patients). On the square root scale, the mean seizure rates are x ¼ 1.42 in the placebo group and y ¼ 0.42 in the treatment group. The sample standard deviations were sx ¼ 1.3 and sy ¼ 1.0, suggesting that there are higher amounts of variation in the placebo arm. We begin by performing a test under an assumption of equal variances in the two groups. This test is not commonly used in practice, but it does appear in some software packages and is the basis for a commonly used sample size formula discussed in the next chapter. Using formulas Eqs. (24.15) and (24.18), we find the common variance s ¼ 1.17. The test statistic assuming that both populations have a common variance is xy T ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 2:71 s ð1=nÞ þ ð1=mÞ

(24.26)

When the null hypothesis is true, the test statistic has a t distribution with n þ m e 2 ¼ 38 df. The a/2  100

(2.5%) lower and upper percentiles of the t distribution with 38 df are 2.02 and 2.02, respectively. Because 2.71 is greater than 2.02, we reject the null hypothesis and conclude that there is a difference in seizure frequency in the placebo and felbamate arms. The p-value which equals P(t38 > 2.71) þ P(t38 < 2.71) ¼ 0.01, means that the chance is approximately 1 in 100 of getting a test-statistic this large (either positive or negative) if the null hypothesis is true. Thus, we can reasonably reject the null hypothesis at this significance level. By comparison a Welch’s t-test, which does not assume an equal variance for the two populations, was conducted. The test was done on the square root scale and resulted in T ¼ 2.74, df ¼ 37.09 rounded down to 37, and a p-value of 0.009, which is similar to the result from the test assuming a common population variance.

Hypothesis Tests for the ISIS-4 Trial: Comparing the Magnesium and No Magnesium Arms The ISIS-4 study was a factorial design of three treatments. We focus on comparing participants receiving magnesium to those not receiving magnesium. A total of 58,050 MI patients were randomized: 29,011 received magnesium and 29,039 did not receive magnesium. The inferential question was whether the proportion of participants dying during the first 35 days after an MI differed between the two groups. The hypothesis is H0 : pMgþ ¼ pMg vs. HA : pMgþ spMg :

(24.27)

The test was two-sided and conducted at the 0.05 significance level. We assume that individual binary outcomes are independent with a common probability of dying in each group, and we note that the sample sizes are large, so we can test this hypothesis with a twosample z-test. The data from the study are presented Table 24.2. The proportion dead at 35 days after randomization (35-day mortality) can be estimated as b p Mgþ ¼ 2216/ 29,011 ¼ 0.0764 and b p Mg ¼ 2103/29,039 ¼ 0.0724. The mortality rate is slightly larger in the magnesium arm. We can formulate the hypothesis test with the test statistic, b p Mg  b p Mgþ Z ¼ vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi    ffi u ub b p Mgþ 1 b p Mg p Mgþ t p Mg 1 b þ n m

(24.28)

The test statistic, at least approximately, has the standard normal distribution when the null hypothesis is true. The 2.5% lower and upper percentiles of the

II. STUDY DESIGN AND BIOSTATISTICS

MISSTATEMENTS AND MISCONCEPTIONS

TABLE 24.2

ISIS-4 Trial 2  2 Table MgD

Mge

Dead

2,216

2,103

Alive

26,795

26,936

Total

29,011

29,039

normal distribution are 1.960 and 1.960, respectively. Using the data provided we find Z ¼ 1.82, and since 1.82 falls between 1.960 and 1.960, we do not reject the null hypothesis and we do not have enough evidence to conclude that the population proportions are unequal. The p-value is P(Z < 1.82) þ P(Z > 1.82) ¼ 0.07.

COMMON MISTAKES IN HYPOTHESIS TESTING As the previous section shows, hypothesis testing requires one to make appropriate assumptions about the structure and distribution of a data set, especially the relationships between the sample observations. There are a number of mistakes that researchers commonly make in hypothesis testing due to ignoring the structure of the sample data or failing to check the assumptions of the hypothesis test. These types of mistakes can lead to faulty conclusions. Some of these common mistakes, illustrated in the context of the t-test, follow. A mistake often committed by somewhat inexperienced researchers is to ignore the pairing between observations within subjects. Although the unpaired t-test remains valid (correct type I error rate), there could be a substantial loss of power or efficiency, making it more difficult to identify true effects. Testing paired continuous data with a two-sample unpaired t-test is a common mistake and we pay for it, but the opposite of this mistake is far worse. Incorrectly assuming a paired structure between two independent samples is worse than ignoring pairing that truly exists. Testing unpaired continuous data with a paired t-test is more serious and could lead to the wrong inference. Along these lines, too frequently researchers ignore the dependence that occurs when multiple observations are made on each subject. For example, if there are five subjects and 3, 2, 1, 2, and 2 measurements are made on these subjects, respectively, there are not 10 independent observations. In this case, more complicated methods, such as mixed models regression, must be used to analyze the data. This mistake is both very common and serious because observations on the same subject tend to be more similar (positively correlated) than those on different subjects. Use of a simple test that

353

ignores the dependence will tend to give p-values that are too small compared to the correct values, which in turn will lead one to conclude that the data provide more evidence against the null hypothesis than they actually do. This topic will be further discussed in Chapter 27. At times our analyses ignore the apparent sample distribution of observations, especially features such as skewness, outliers or extreme values, and lower or upper limits on measurement accuracy. Performing a t-test on highly skewed data without appropriate adjustments can lead to the wrong conclusions; although, the t-test is generally robust against mistakes such as ignoring moderate amounts of skewness for sufficiently large samples. Lastly, when we look at the sample distribution we should not forget the variance. Frequently when analyzing data we make a simple mistake and assume equal variances in two groups and perform pooled t-test without knowing from external sources that the variance is the same in the two groups or examining the data either graphically or numerically. We should instead at least perform a Welch’s t-test for two samples. In most cases this is not too serious a mistake. Indeed, the felbamate monotherapy example showed that the two-sample t-test is robust to ignoring the differences between the variances of the two samples.

MISSTATEMENTS AND MISCONCEPTIONS The following are some of the major misstatements and misconceptions that arise when performing hypothesis tests and reporting the results. 1. “Failing to reject the null hypothesis (H0) means that it is true.” On the contrary, failing to reject the null hypothesis may merely indicate that there is not enough evidence to state that it is false at a particular significance level. Failing to reject the null is not equivalent to accepting the null hypothesis. The null hypothesis may be true or it may be false, but we do not have the evidence in the sample being studied to reject it. 2. “The p-value is small so the two sample means (x and y) are significantly different from each other.” This approach is incorrect because the p-value is a statistical tool for making inferences about the true population means. The goal of hypothesis testing is to make statements about population parameters and not the samples themselves. People misstate this frequently, sometimes by accident and sometimes not. 3. “The impact is huge, just look at the tiny p-value!” Focusing on the statistical significance of an effect (its

II. STUDY DESIGN AND BIOSTATISTICS

354

24. HYPOTHESIS TESTING

p-value) but ignoring the effect’s magnitude or size is a common misstatement. In a study with multiple explanatory variables, there often will be several variables that appear to be related to the outcome of interest. While a small p-value may demonstrate significant evidence that the effect of the variable on the outcome is nonzero, the point estimate and CIs for the magnitude of the effect demonstrate how much of an impact that variable has on the magnitude of the response. 4. One of the most commonly discussed misconceptions is confusing statistical significance with clinical significance. This relates to misstatement 3. In the ISIS-4 trial, the participants who received intravenous magnesium sulfide had a 35-day unadjusted mortality rate of 7.64%, whereas those who did not receive that treatment had a corresponding mortality rate of 7.24%. If the two-sided p-value had been equal to 0.007 (it was actually p ¼ 0.07), we would need to ask ourselves this: even though the p-value was quite significant at 0.007, was the increase in mortality of 0.40% on the treatment clinically troubling? Possibly, it is troubling if 0.4% is equal to many lives per year. Possibly, such a small difference is not troubling in some studies or is weighed with other issues such as side effects. Just because a finding is statistically significant does not make it clinically significant. Both types of significance are needed.

SPECIAL CONSIDERATIONS The following topics extend the concepts presented in the previous sections. Most of this chapter was devoted to establishing a conceptual framework for statistical hypothesis testing. We focused primarily on tests for comparing two populations because these tests are the most common types of tests used in clinical research. Here, we briefly describe other methodology that is commonly used in analyzing the data from medical studies. More details on all these subjects can be found in the references.

Comparing More Than Two Groups: One-Way Analysis of Variance The analysis of variance (ANOVA) framework extends the methodology for comparing the means of two populations to more than two populations. This method may be applicable in multiarm clinical trials in which interest focuses on detecting any differences among the various treatments. The hypotheses for comparing k population means with ANOVA can be written as

H0 : m1 ¼ m2 ¼ / ¼ mk vs. HA : Some mi smj

(24.29)

Here, the null hypothesis is that the means in each of the k groups are equal. The assumptions for this test are that the data are normally distributed with a constant population variance across the k groups. In addition, it is assumed that the data for each of the subjects are statistically independent. The test statistic used is the ratio of the between-subject variance to the within-subject variance. Under the null hypothesis of equal population means, the test statistic has an F distribution, and one can obtain a p-value to assess the significance of this test (see Altman5 for more details).

Simple and Multiple Linear Regression Simple linear regression is a technique used to examine the strength of a linear relationship in a set of bivariate or paired data, where one variable acts as the predictor and the other as the response. For example, one may be interested in examining whether there is a linear increase in blood pressure with age for a certain range of ages. The simple linear regression model for blood pressure (y) as a function of age (x) is yi ¼ b0 þ bi xi þ εi ;

(24.30)

where b0 and b1 are the intercept and slope for the regression line, respectively, and the index i denotes the value for the ith individual. In addition, εi is an error term (assumed to be normally distributed with mean ¼ 0 and variance ¼ s2) that characterizes the scatter around the regression line. The intercept (b0) and slope (b1) parameters are estimated using least squares fitting. Least squares fitting involves choosing the line that minimizes the sum of the squared vertical differences between the responses and the points predicted by the fitted line at values of the predictor variable. Hypothesis testing also plays an important role in regression. We often wish to test whether there is a significant change in one variable with each unit increase in a second variable, not only with the data we observed in the sample but also in the population from which the sample data were drawn. In other words, we wish to test whether there is a linear trend present in the data. The null hypothesis would be that there is no linear trend. The hypotheses for linear regression can be stated as H0 : b1 ¼ 0 vs. HA : b1 s0:

(24.31)

The assumptions for this test are that response observations are independent and normally distributed (with constant variance) around the regression line. The test statistic for a significant linear relationship is the ratio of the variance of the data points around the average y value (y) relative to the variance around the regression

II. STUDY DESIGN AND BIOSTATISTICS

SPECIAL CONSIDERATIONS

line. A large test statistic of this type reflects either a steep slope or small variability around a slope. This test statistic has an F distribution under the null hypothesis that the slope is zero (i.e., a horizontal line), and one can obtain a p-value to assess the significance of this test. Multiple or multivariate regression is an extension of simple linear regression, which allows for more than one variable or covariate. We may be interested in examining for a linear increase in blood pressure with age (xi) after adjusting for weight (zi). The multiple regression model can be written as yi ¼ b0 þ b1 xi þ b2 zi þ εi :

(24.32)

The hypotheses, one for each b, for multiple regression are formulated in a similar way as for simple linear regression. For multiple linear regression, each variable that has a slope (b) coefficient found to be significantly different from 0 is commonly interpreted as a variable that has an independent effect on y.

Multiple Comparisons When making many statistical comparisons, i.e., performing multiple hypothesis tests, a certain fraction of the test statistics will be statistically significant even when the null hypothesis is true. In general, when a series of tests is performed at the a significance level, approximately a  100% of tests will be significant at the a level even when the null hypothesis for each test is true. For example, even if the null hypotheses are true for all tests, when conducting many independent hypothesis tests at the 0.05 significance level, on average (in the long term) 5 of 100 tests will be significant by chance alone. Issues of multiple comparisons arise in various situations, such as in clinical trials with multiple end points and multiple looks at the data. By doing multiple tests, you naturally increase your chances of making a type I error if no adjustment is made to the usual testing framework for a single test statistic. Pairwise comparison among the sample means of several groups is also an area in which issues of multiple comparisons may be of concern. For k groups, there are k(k e 1)/2 pairwise comparisons, and just by chance some may reach significance. Our last example is with multiple regression analysis in which many candidate predictor variables are tested and entered into the model. Some of these variables may result in a significant result just by chance. With an ongoing study and many interim analyses or inspections of the data, with no adjustment for performing multiple comparisons, we have a high chance of rejecting the null hypothesis at some time point even when the null hypothesis is true. There are various approaches to the multiple comparisons problem. First, consider if multiple comparisons is actually a problem. If we ask multiple questions we

355

expect multiple answers. If we ask related questions we expect related answers. Looking at the totality of the evidence when interpreting results is far more useful than overzealous correction for multiple comparisons (or ignoring all but the single significant p-value out of 50). One rather informal approach to multiple comparisons is to choose a significance level a lower than the traditional 0.05 level (e.g., 0.01) to prevent many falsepositive conclusions or to “control the false discovery rate.” The number of comparisons should be made explicit in the article. More formal approaches to control the “experimentewise” type I error using corrections for multiple comparisons have been proposed. An example is the Bonferroni correction, in which the type I error rate is taken as a/n, where n is the number of comparisons made. Another class of methods has been developed to correct for multiple comparisons that result from monitoring trial results during the trial. Interim monitoring methods that control the type I error rate are available for various study designs and are discussed further in Chapter 27.14 The classic reference by Hochberg and Tamhane provides a broader discussion of methodology to adjust for multiple comparisons.15 It is best to address the issue of multiple comparisons during the design stage of a study. One should determine how many comparisons will be made and then explicitly state these comparisons. Studies should generally be designed to minimize the number of statistical tests at the end of the study. Ad hoc solutions to the multiple comparisons problem may be done for exploratory or epidemiologic studies. Multiple comparison adjustments should be made for the primary analyses of definitive studies (such as phase III confirmatory studies) to rigorously maintain the type I error rate, i.e., the probability of falsely rejecting any null among those tested, at the chosen a level. Studies that focus on a single primary outcome and data analyzed at the end of study avoid the issue of multiple comparisons. The topic of multiple comparisons is expanded in Chapter 27.

Nonparametric Versus Parametric Tests Inferential methods that rely on assumptions about the underlying distributions from which the data originate are called parametric methods, whereas those that make no distributional assumptions are called nonparametric methods. Nonparametric methods often are used when data do not meet the distributional assumptions of parametric methods, such as asymmetric distributions or unusual numbers of extreme values. Nonparametric methods are usually based on the ranks of observations as opposed to their actual values, which lessens the impact of skewness and extreme outliers in the raw data. Hypotheses are usually stated in terms of distributions instead

II. STUDY DESIGN AND BIOSTATISTICS

356

24. HYPOTHESIS TESTING

of means. Corresponding to the two-sample hypothesis tests of means discussed in this chapter are the following parametric:nonparametric pairs: • Paired t-test: Wilcoxon signed rank test or the sign test • Two-sample t-test: Wilcoxon rank sum test • Analysis of variance: KruskaleWallis test • One proportion z-test: Exact binomial test

2.

In general, nonparametric tests have somewhat lower power than their parametric counterparts, when the parametric method is making the right assumption for the underlying distribution for the data. This is the price one pays for making fewer assumptions about the data. Fewer assumptions, however, does not necessarily mean no distributional assumptions. For large sample sizes, parametric and nonparametric tests generally lead to the same inferences. More information about nonparametric approaches can be found in van Belle et al.16 3.

CONCLUSION Study design is part science and part art. Statistical hypotheses need to match the question of interest and many choices are available. Guidance for clinical trials protocols such as Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) highlight the need for care in this area.17 It is wise to consult with a statistician at the earliest stages of planning a study to obtain help with study design and hypothesis generation. Many times scientific hypotheses can be finetuned as one finds a compromise between the most interesting question to answer and the question that will be practical to address given the constraints of available resources. The chosen study design will limit which statistical hypotheses can be tested and which test statistics are appropriate for the data.18 Both the test statistic and the level of significance should be specified in advance. Once a test statistic is chosen, its sampling distribution can be used to calculate the p-value for the observed data. The p-value summarizes the evidence in the data against the null hypothesis. While statistical software is widely available for data analysis and calculations of p-values, the researcher is required to understand which test statistics are appropriate for the hypothesis of interest and the study design that generated the data.19 Timely collaboration with an eager statistician will help you and your studies succeed.

SUMMARY QUESTIONS 1. In the Women’s Health Initiative (WHI) Clinical Trial of Calcium and Vitamin D Supplementation, 36,382 postmenopausal women aged 50e79 were

4.

5. 6.

7.

randomized to either 1000 mg of calcium with 400 International Units (IU) of vitamin D3 daily or placebo and followed for an average of 7 years. Investigators were interested in whether supplementation of calcium and vitamin D would reduce hip fractures. State the null and alternative hypothesis. Jackson et al.20 on behalf of the WHI investigators, reported the results of the Calcium and Vitamin D trial described in problem 1. a. The hip bone density was 1.06 percent higher in the supplementation group compared to placebo, p ¼ 0.01. Interpret this result. Was the null hypothesis rejected? b. These authors also reported the ratio of hip fractures for the supplement to placebo group. The intention-to-treat (ITT) analysis reported the rate ratio (95% CI) as 0.88 (0.72, 1.08), whereas for adherent women only the rate ratio was 0.71 (0.52, 0.97). Interpret these two results. Hint: the ratio of two numbers that are equal is 1. Suppose investigators were interested in comparing the efficacy of two different antibiotic regimens for the treatment of sepsis in a randomized phase II trial. The primary outcome is the binary outcome of 30-day survival (yes/no). State the null and alternative hypothesis. Define the following terms: a. p-value b. type I error c. type II error Why is hypothesis testing an important part of any analysis from a phase III clinical trial? Does failing to reject H0 mean that H0 is true? a. Yes b. No, failing to reject the null hypothesis means there is insufficient evidence to reject the null hypothesis. The null hypothesis for a protocol is the mean systolic blood pressure in the control group is equal (the same) as the mean systolic blood pressure in the experimental intervention group. Which one of the following statements is FALSE? a. The p-value is the probability that the null hypothesis is true b. The p-value is a measure of the strength of evidence in the data that the null hypothesis is not true c. The p-value is the probability that data generated under the null hypothesis will produce a test statistic as extreme or more extreme than the value we actually observed d. The p-value is the probability of observing a difference between the groups in the samples’ mean systolic blood pressures as large as or larger than that observed if the two groups have the same true mean systolic blood pressure

II. STUDY DESIGN AND BIOSTATISTICS

REFERENCES

Acknowledgments The authors wish to thank Paul S. Albert for his contributions to the earlier versions of this chapter in previous editions of this book. His worked examples and dedication to training live on in this text.

Disclaimers This chapter reflects the views of the author and should not be construed to represent FDA’s views or policies. The findings and conclusions in this chapter are those of the authors and do not necessarily represent the official position of the Centers for Disease and Control Prevention.

References 1. Stone LA, Frank JA, Albert PS, et al. Characterization of MRI response to treatment with interferon beta lb: contrast enhancing MRI lesion frequency as a primary outcome measure. Neurology 1997;49:862e9. 2. Theodore WH, Albert P, Stertz B, et al. Felbamate monotherapy: implications for antiepileptic drug development. Epilepsia 1995; 36:1105e10. 3. ISIS-4 Collaborative Group. ISIS-4: a randomized factorial trial assessing early oral captopril, oral mononitrate, and intravenous magnesium sulphate in 58,050 patients with suspected acute myocardial infarction. Lancet 1995;345:669e85. 4. Piantadosi S. Clinical trials: a methodologic perspective. 2nd ed. Hoboken, NJ: John Wiley & Sons; 2005. 5. Altman DG. Practical statistics for medical research. Boca Raton, FL: Chapman & Hall; 1991. 6. Moore DS, Notz WI. Statistics: concepts and controversies. 5th ed. New York: Freeman; 2005. 7. Moore DS. Introduction to the practice of statistics. 5th ed. New York: Freeman; 2005.

357

8. Armitage P, Berry G, Matthews JNS. Statistical methods in medical research. Blackwell: Oxford; 2001. 9. Draper NR, Smith H. Applied regression analysis. 3rd ed. New York: Wiley; 1998. 10. Clopper CJ, Pearson ES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 1934;26: 404e13. 11. Agresti A, Coull BA. Approximate is better than “exact” for interval estimation of binomial proportions. Am Stat 1998;52: 119e26. 12. Borkowf CB. Constructing confidence intervals for binomial proportions with near nominal coverage by adding a single imaginary failure or success. Stat Med 2006;25:3679e95. 13. Agresti A. Categorical data analysis. 2nd ed. Hoboken, NJ: Wiley; 2002. 14. Friedman LM, Furberg CD, DeMets DL. Fundamentals of clinical trials. 4th ed. New York: Springer; 2010. 15. Hochberg Y, Tamhane AC. Multiple comparison procedures. New York: Wiley; 1987. 16. van Belle G, Fisher LD, Heagerty PJ, Lumley TS. Biostatistics: a methodology for the Health sciences. 2nd ed. New York: Wiley; 2004. 17. Chan A-W, Tetzlaff JM, Gøtzsche PC, Altman DG, Mann H, Berlin J, Dickersin K, Hro´bjartsson A, Schulz KF, Parulekar WR, Krleza-Jeric K, Laupacis A, Moher D. SPIRIT 2013 explanation and elaboration: guidance for protocols of clinical trials. BMJ 2013;346:e7586. 18. U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research (CDER) and Center for Biologics Evaluation and Research (CBER). Multiple endpoints in clinical trials: guidance for industry. Draft. January 2017. http://www.fda.gov/ucm/groups/fdagovpublic/@fdagov-drugs-gen/documents/document/ucm536750. pdf. 19. Wasserstein RL, Lazar NA. The ASA’s statement on p-values: context, process, and purpose. Am Statistician 2016;70:129e33. 20. Jackson RD, LaCroix AZ, Gass M. Calcium plus vitamin D supplementation and the risk of fractures. NEJM 2006;357:669e83.

II. STUDY DESIGN AND BIOSTATISTICS

C H A P T E R

25 Power and Sample Size Calculations 1

Craig B. Borkowf1, Laura Lee Johnson2, Paul S. Albert3

Centers for Disease Control and Prevention, Atlanta, GA, United States; 2U.S. Food and Drug Administration, Silver Spring, MD, United States; 3National Institutes of Health, Rockville, MD, United States

O U T L I N E Introduction Basic Concepts Notational Conventions Review of the Normal and t-Distributions

Sample Size Calculations for Precision in Confidence Interval Construction 361 Confidence Intervals for Means of Continuous Data 361 Confidence Intervals for Binomial Proportions 362 Sample Size Calculations for Hypothesis Tests: One Sample of Data Calculations for Continuous Data Regarding a Single Population Mean Calculations for Binary Data Regarding a Single Population Proportion Two-Stage Designs for a Single Population Proportion

362 362 363 363

Sample Size Calculations for Hypothesis Tests: Paired Data 364 Calculations for Paired Continuous Data 364 Calculations for Paired Binary Data 365 Sample Size Calculations for Hypothesis Tests: Two Independent Samples 366

INTRODUCTION This chapter introduces several fundamental concepts related to power and sample size calculations. We first review some key questions that should be considered when developing research studies. We then introduce the concept of statistical power and explain why having adequate power is essential for designing successful studies. Next, we present some basic sample size Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00025-3

Calculations for Continuous Data With Equal Variances and Equal Sample Sizes 366 Calculations for Continuous Data With Unequal Variances or Unequal Sample Sizes 367 Calculations for Two Independent Samples of Binary Data 367

359 360 360 361

Advanced Methods and Other Topics Alternative Statistics and Sample Size Calculation Methods Several Advanced Study Designs Retention of Subjects Statistical Computing

368

Conclusion

369

Exercises

370

Acknowledgments

371

Disclaimers

371

References

371

368 368 369 369

formulas for when we plan to collect a sample of either continuous or binary data and then wish to construct a confidence interval for the population mean or the population proportion with a certain degree of precision. Subsequently, we present some basic sample size formulas for when we plan to collect one sample, a paired sample, or two independent samples of either continuous or binary data and then wish to test hypotheses about specific characteristics of the populations from which the data

359

Copyright © 2018. Published by Elsevier Inc.

360

25. POWER AND SAMPLE SIZE CALCULATIONS

came. Finally, we discuss several advanced topics related to sample size calculations and the collaborative process of study design.

Basic Concepts One of the most common yet challenging collaborations between statisticians and researchers is the sample size calculation. They must work together to gather established knowledge and to elicit beliefs about the research questions to construct reasonable hypotheses about what future data may show. Some typical questions are as follows: What are the study groups or treatment arms of interest (e.g., new treatment, placebo)? What is the expected distribution or type of data to be collected (e.g., bell-shaped, continuous, count, ordinal, categorical, binary)? What are the anticipated means, medians, and ranges of the future data? What is the variation in repeated measurements for each subject, and what is the variation between measurements for different subjects? How much does it cost to set up the study and then to enroll each additional subject or site? What are the maximum numbers of subjects and sites that can realistically be enrolled in the study? What are the consequences of a Type I error, namely, rejecting the null hypothesis (H0) when it is indeed true? Conversely, what are the consequences of a Type II error, namely, failing to reject the null hypothesis when it is false and instead a particular alternative hypothesis (H1) is true? The desire to control the probability of making either a Type I or a Type II error leads to the concept of statistical power. Recall that hypothesis tests are designed so that, for a particular test statistic (e.g., z-test, t-test), the probability of making a Type I error, or the significance level, denoted a, is held at some fixed value. Sample size calculations are designed so that, for the same test, the probability of making a Type II error, denoted b, for a particular alternative hypothesis also is held at some fixed value. In turn, (1
b) equals the statistical power of a particular test, which is the probability of rejecting the null hypothesis when that alternative hypothesis is indeed true. When designing studies, we need to consider the statistical power because power indicates the chance of detecting a statistically significant effect (of any magnitude) when in truth an effect of a certain magnitude exists. Studies with low power are unlikely to answer key scientific questions with sufficient precision or to produce statistically significant results even when meaningful effects do exist and are thus an inefficient use of precious scientific resources. It is important to remember that the absence of evidence for an effect is not the same as evidence for the absence of an effect.1 Although there is no hard-and-fast rule,2 the consensus is that it is

desirable to have at least 80% power to test credible scientifically or clinically meaningful hypotheses.3,4 One can take various approaches to sample size and power calculations for a given set of null and alternative hypotheses, a chosen significance level (a) and a particular test statistic. Suppose we wish to design a parallelgroups trial in which subjects are randomly assigned to receive either a new treatment or a placebo. First, one may calculate power for a fixed sample size. For example, one may ask, “If we enroll 30 subjects, what is the power, or chance, to detect a statistically significant effect (of any magnitude) when the truth is that the new treatment produces a 20% reduction in the main outcome?” Second, one may calculate the required sample size for a fixed power. For instance, one may ask, “What is the required sample size in each of two groups to have 80% power to detect a statistically significant effect (of any magnitude) when the truth is that the new treatment produces a 20% reduction in the main outcome?” This chapter is focused on the latter approach, namely, estimating the minimum required sample size for a fixed power. When publishing the results of clinical and scientific research, it is standard practice to report how the sample size calculations were performed. For example, the CONSORT 2010 Statement (item 7a)5,6 requires authors to explain how the sample size was determined. This information enables the reader to evaluate the credibility of the null and alternative hypotheses, the appropriateness of the test statistics, and the plausibility of the design parameters, and thus whether the sample size and power calculations adequately reflect the balance of clinical, scientific, and statistical concerns. Note that many of the sample size formulas presented in this chapter are approximate, and some tend to underestimate the required sample size. To illustrate how well these more basic methods perform, we will periodically mention advanced methods that are beyond the scope of this chapter. It is advisable to consult a statistician about the best methods to use for a particular set of research questions. Furthermore, because sample size calculations depend heavily on the assumptions made by the investigators, it also is wise to perform a series of calculations under a variety of null and alternative hypotheses.

Notational Conventions One central scientific aim is to collect a sample of data from a population or group and then calculate sample statistics from those data to estimate the corresponding parameters, or characteristics, of that population or group. In this chapter, we use Greek letters to represent population parameters and their hypothesized values, such as the population mean (m), variance (s2), standard

II. STUDY DESIGN AND BIOSTATISTICS

SAMPLE SIZE CALCULATIONS FOR PRECISION IN CONFIDENCE INTERVAL CONSTRUCTION

deviation (s), or proportion (p). By contrast, we use Latin letters to represent sample statistics calculated from data, such as the sample mean (x), variance (s2), standard deviation (s), or proportion (p). In addition, we use subscripts to distinguish the parameters and the statistics that correspond to various hypotheses or study groups.

Review of the Normal and t-Distributions Recall that we often model continuous data (perhaps after an appropriate transformation) by the Normal (Gaussian) distribution, or a bell-shaped curve, with mean m and variance s2. In particular, the standard Normal distribution has a mean of zero and a variance of one. The z-statistic calculated for the z-test (described in the preceding chapter) follows this distribution. We use the symbol Zc to represent the c  100 percentile of the standard Normal distribution, such that c equals the proportion of this distribution less than Zc (or, equivalently, the fraction of area under the standard Normal curve to the left of Zc), where c ranges from 0 to 1. Some key percentiles of the standard Normal distribution are Z0.5 ¼ 0, Z0.8 ¼ 0.842, Z0.9 ¼ 1.282, Z0.95 ¼ 1.645, Z0.975 ¼ 1.960, and Z0.99 ¼ 2.326. Furthermore, because this distribution is symmetric about zero, Zc ¼
Z1
c. Additional values can be readily obtained from tables in most introductory statistics textbooks (e.g., Altman7) and from most statistical software packages. By comparison, when the population variance is unknown, we use the sample variance to calculate the t-statistic for the t-test (described in the preceding chapter). The t-statistic follows the t-distribution, the shape of which depends on the degrees-of-freedom parameter, a function of the sample size. We use the symbol Tf,c to represent the c  100 percentile of the t-distribution with f degrees of freedom. For example, for selected values of f, the 97.5 percentiles of the tdistribution are T10,0.975 ¼ 2.228, T20,0.975 ¼ 2.086, T30,0.975 ¼ 2.042, T60,0.975 ¼ 2.000, and T100,0.975 ¼ 1.984. Note that as f increases, the c  100 percentiles of the t-distribution decrease in magnitude toward the corresponding percentile of the standard Normal distribution; hence Tf,c > Tfþ1,c > Zc for c > 0.5. Furthermore, because the t-distribution is symmetric about zero, Tf,c ¼
Tf,1
c. Consult an introductory statistics textbook for more details about this important distribution.

SAMPLE SIZE CALCULATIONS FOR PRECISION IN CONFIDENCE INTERVAL CONSTRUCTION In this section, we consider sample size calculations for the construction of confidence intervals of a desired width or precision. These calculations depend only on the chosen significance level (a) but not on power.

361

Confidence Intervals for Means of Continuous Data First, we can calculate the sample size required to construct a confidence interval of width w for an unknown population mean. We assume that the data come from a population that is well approximated (possibly after some transformation) by a Normal distribution with unknown mean m and known variance s2. We may plan to collect a sample of data from this population and then calculate the sample mean x and the sample variance s2. Recall that we may construct a (1
a) 100% confidence interval for the population mean as follows:
pffiffiffi x  Z1
a=2 s n : (25.1) For this confidence interval to have width w (or less), we need to solve for the smallest sample size n such that pffiffiffi (25.2) 2Z1
a=2 s n  w; or, equivalently,8

 n  4Z21
a=2 s2 w2 :

(25.3)

By comparison, if the variance is unknown, we may construct a (1
a) 100% confidence interval for the population mean using the t-distribution with (n
1) degrees of freedom and the sample standard deviation s:
pffiffiffi x  Tn
1;1
a=2 s n : (25.4) Since the sample standard deviation s needs to be calculated from the future data, we substitute a hypothesized value sh for s and solve numerically for the smallest sample size n such that pffiffiffi 2Tn
1;1
a=2 sh n  w; (25.5) or, equivalently,

 2 s2h w2 : n  4Tn
1;1
a

(25.6)

Assuming the hypothesized variance s2h is close to the true variance s2, the future confidence interval will have approximately the desired width. Example 1. Suppose a clinician wishes to estimate the mean serum albumin level in a specific population of patients with primary biliary cirrhosis of the liver. Because serum albumin is an important indicator of the synthetic function of the liver, she wants to obtain a tight confidence interval around the estimated mean. An earlier study found a mean of 35 g/L and standard deviation of 6 g/L. If she wishes to estimate the mean in this new population of patients with a 95% confidence interval (a ¼ 0.05) of width w ¼ 4 g/L, how many patients should she enroll in this study?

II. STUDY DESIGN AND BIOSTATISTICS

362

25. POWER AND SAMPLE SIZE CALCULATIONS

To construct a 95% confidence interval, we use Z1
a/2 ¼ 1.960. Using Eq. (25.3), we calculate . n ¼ 4ð1:960Þ2 ð6Þ2 ð4Þ2 ¼ 34:6z35: (25.7) Thus, the minimum required sample size in this study is n ¼ 35 patients. Furthermore, because the true variance of the new population is really unknown, she may plan to construct a 95% confidence interval using the t-distribution and the future sample standard deviation s. Solving Eq. (25.6) numerically, we obtain n ¼ 38, a slightly larger sample size. By way of comparison, using an advanced method beyond the scope of this chapter,9 we can strictly control the probability that the width of a confidence interval constructed using the t-distribution does not exceed a certain threshold, assuming that the true standard deviation does not exceed its hypothesized value, i.e., s  sh. In the above example, if s ¼ 6 g/L, the probability is only 50% that the width of a 95% confidence interval constructed using the t-distribution will be less than or equal to w ¼ 4 g/L for a sample size of n ¼ 38. This probability increases to 80%, 90%, and 95% for sample sizes of n ¼ 44, 47, and 50, respectively.

Confidence Intervals for Binomial Proportions Likewise, we can calculate the sample size required to construct a confidence interval of width w for an unknown binomial proportion p. Let p denote a sample proportion. Recall that we may construct a (1
a) 100% binomial confidence interval for a population proportion as follows: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi.pffiffiffi
p  Z1
a=2 pð1
pÞ n : (25.8) Because the sample proportion p must be estimated from the future data, we substitute a hypothesized value of ph for p and solve for the smallest sample size n such that pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi.pffiffiffi 2Z1
a=2 ph ð1
ph Þ n  w; (25.9)

desired width w. By way of comparison, a more advanced exact binomial method, beyond the scope of this chapter, shows that when the true proportions are p ¼ 0.1, 0.2, 0.3, 0.4, and 0.5, we need somewhat larger sample sizes of n ¼ 189, 291, 359, 395, and 402, respectively, for the probability to be 90% that the width of the 95% confidence interval will indeed equal w ¼ 0.1 or less. More generally, we can use any confidence interval formula to solve directly or numerically for the smallest sample size n that gives the interval of the desired width. Remember that when the width of an interval depends on sample statistics that need to be estimated from the future data (e.g., the sample standard deviation), the sample size formula depends on the hypothesized values of the corresponding population parameters. Thus, it is wise to consider a range of plausible parameter values when performing sample size calculations.

SAMPLE SIZE CALCULATIONS FOR HYPOTHESIS TESTS: ONE SAMPLE OF DATA In this section, we discuss sample size calculations in the context of testing hypotheses about the population mean or proportion when we plan to sample data from a single population. Recall that for hypothesis testing, the alternative hypothesis is the general negation of the null hypothesis (e.g., H0: m ¼ m0 vs. H1: m s m0). By contrast, for sample size calculations, a specific alternative hypothesis is required (e.g., H1: m ¼ m1), where the difference between the null and alternative hypotheses represents a scientifically or clinically meaningful difference (e.g., d ¼ m1
m0). Note that the expressions on the left side of each hypothesis are the unknown parameters about which we wish to make inference, whereas the expressions on the right side are the hypothesized values that we use in the sample size calculations. In the next two sections, we extend these basic principles to sample size calculations for paired data and two independent samples of data.

or, equivalently,8

 n  4Z21
a=2 ph ð1
ph Þ w2 :

(25.10)

For example, to obtain a 95% confidence interval of width w ¼ 0.1 for true proportions p ¼ 0.1, 0.2, 0.3, 0.4, and 0.5, we need sample sizes of n ¼ 139, 246, 323, 369, and 385, respectively. By symmetry, results for p and (1
p) are the same. It is important to remember that these methods for binomial proportions are approximate and that depending on the sample proportion p actually obtained, the interval can be wider than the

Calculations for Continuous Data Regarding a Single Population Mean Consider the following null and alternative hypotheses about the mean m of a population that has (approximately) a Normal distribution with known variance s2: H0 : m ¼ m0 vs. H1 : m ¼ m1

(25.11)

Let d ¼ m1
m0 denote the scientifically or clinically meaningful difference, and let x denote the sample

II. STUDY DESIGN AND BIOSTATISTICS

SAMPLE SIZE CALCULATIONS FOR HYPOTHESIS TESTS: ONE SAMPLE OF DATA

mean. Suppose we plan to conduct a two-sided hypothesis test by using the z-statistic,  pffiffiffi (25.12) z ¼ ðx
m0 Þ s n at the a significance level. To calculate the required sample size, we specify the power (1
b), and use the following formula10,11:
2  n ¼ Z1
a=2 þ Z1
b s2 d2 (25.13) In this and all other sample size formulas, if n is not an integer, it should be rounded up. For a one-sided test, replace Z1
a/2 by Z1
a. Eq. (25.13) shows that the required sample size increases as power (1
b) and the population variance (s2) increase and as the significance level (a) and the meaningful difference (d) decrease. Also, for a given significance level, sample sizes are smaller for one-sided hypothesis tests than for comparable two-sided hypothesis tests. Example 2. Patients with hypertrophic cardiomyopathy (HCM) have enlarged left ventricles (mean, 300 g) compared with the general population (mean, 120 g). A cardiologist studying a particular genetic mutation that causes HCM wishes to determine whether the mean left ventricular mass of patients with this particular mutation differs significantly from the mean for other patients with HCM. If the true difference equals or exceeds the meaningful difference of d ¼ 10 g in either direction, it is important to reject the null hypothesis of equality (m ¼ 300 g). If past laboratory measurements suggest that s ¼ 30 g and he chooses a 5% significance level (a ¼ 0.05) and a power of 90% (b ¼ 0.1), what sample size does he need? This hypothesis is two-sided, so Z1ea/2 ¼ 1.960 and Z1eb ¼ 1.282. Using the previous formula, we calculate . n ¼ ð1:960þ 1:282Þ2 ð30Þ2 ð10Þ2 ¼ 94:6z95 (25.14) Thus, the minimum required sample size is n ¼ 95 subjects for this study. By contrast, if the population variance is unknown, as in the case where prior studies of that population have not been conducted, we must plan to use a t-test (with the sample standard deviation s) rather than a z-test (with the population standard deviation s). Because the t-distribution is wider than the standard Normal distribution, the above calculation based on the Normal percentiles Z1ea/2 and Z1eb (instead of their t-distribution counterparts) underestimates the required sample size. Hence, the convention is to increase the sample size slightly to compensate for this underestimation.12 By comparison, if the cardiologist in Example 2 plans to use a t-test (with the sample standard deviation s), the minimum required sample size is n ¼ 97.

363

Calculations for Binary Data Regarding a Single Population Proportion Next, consider the following null and alternative hypotheses about the proportion p of a population that has some binary characteristic: H0 : p ¼ p0 vs. H1 : p ¼ p1 :

(25.15)

Let d ¼ p1
p0 denote the scientifically or clinically meaningful difference, and let p denote the sample proportion. Suppose we plan to conduct a two-sided hypothesis test by using the z-statistic, .pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi z ¼ ðp
p0 Þ p0 ð1
p0 Þ=n; (25.16) at the a significance level. To calculate the required sample size, we specify the power (1
b), and use the following formula13: h pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi i2 . 2 n ¼ Z1
a=2 p0 ð1
p0 Þ þ Z1
b p1 ð1
p1 Þ d : (25.17) Example 3. Suppose an oncologist wishes to conduct a Phase II (safety/efficacy) clinical trial to test a new cancer drug. If only 20% of patients will benefit from this drug, she does not wish to continue to study it because drugs with comparable efficacy are already available. Conversely, if at least 40% of patients will benefit from this drug, she wishes to have an 80% chance to reject the null hypothesis and consequently to continue to study the drug. Using a one-sided z-test at the 5% significance level (a ¼ 0.05) and 80% power (b ¼ 0.2), how many participants should she enroll in this clinical trial? This hypothesis is one-sided, so Z1ea ¼ 1.645 and Z1eb ¼ 0.842. The null proportion is p0 ¼ 0.2, the alternative proportion is p1 ¼ 0.4, and the difference is d ¼ 0.2. Using the previous formula, we calculate h pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi i2 . n ¼ 1:645 ð0:2Þð0:8Þ þ 0:842 ð0:4Þð0:6Þ ð0:2Þ2 ¼ 28:6z29: (25.18) Thus, the minimum required sample size is n ¼ 29 patients for this single-arm clinical trial. Similarly, with 90% power (b ¼ 0.1, Z1eb ¼ 1.282), the required sample size is n ¼ 42. By way of comparison, an advanced exact binomial method gives required sample sizes of n ¼ 35 and 47 for 80% and 90% power, respectively.

Two-Stage Designs for a Single Population Proportion For early Phase II (safety/efficacy) trials, an alternative approach for testing hypotheses about a single population proportion is to use a two-stage design.10,14 In a

II. STUDY DESIGN AND BIOSTATISTICS

364

25. POWER AND SAMPLE SIZE CALCULATIONS

two-stage design, one sequentially enrolls up to a certain number of subjects in a first stage, and then, if the evidence is sufficiently promising, one sequentially enrolls up to a certain number of subjects in a second stage. If the evidence is not sufficiently promising during either the first or the second stage, one may stop enrollment immediately without completing that stage. For selected sets of hypotheses, significance levels, and powers, tables in the above references indicate the maximum number of subjects to enroll in each stage and the corresponding stopping rules. These rules are chosen so that “optimal” two-stage designs have the smallest expected, or average, sample size under the null hypothesis. Example 4. Consider a two-stage trial to test the same hypotheses as in the preceding example, namely H0: p ¼ 0.2 versus H1: p ¼ 0.4, with a 5% significance level and 80% power. In the first stage, one would enroll up to m1 ¼ 13 participants sequentially, and if r1 ¼ 3 or fewer study participants respond positively to the drug, one should stop enrollment and abandon the drug. In the second stage, one would enroll up to 30 additional participants sequentially, for a maximum of m2 ¼ 43; then, if r2 ¼ 12 or fewer study participants respond, one should abandon the drug, whereas if 13 or more participants respond, the drug should be considered for further study. With a 5% significance level and 80% power, if the null hypothesis is true (i.e., p ¼ 0.2), one would need to enroll, on average, 21 participants in the trial (i.e., substantially less than n ¼ 29 or m2 ¼ 43) to conclude that the drug should be abandoned. By comparison, with a 5% significance level and 90% power (m1 ¼ 19, r1 ¼ 4, m2 ¼ 54, r2 ¼ 15), if the null hypothesis is true, one would need to enroll, on average, 30 participants in the trial (i.e., substantially less than n ¼ 42 or m2 ¼ 54) to reach the same conclusion.

SAMPLE SIZE CALCULATIONS FOR HYPOTHESIS TESTS: PAIRED DATA In this section, we discuss sample size calculations for hypothesis tests about characteristics of paired data. The pairing may result from measuring the same characteristics of each subject before and after treatment or from taking similar measurements of opposite paired limbs or organs of the same subject (e.g., arms, legs, eyes, ears). The pairing also may result from taking measurements of twins or two persons with similar or “matched” characteristics, one of whom is assigned to the treatment group and the other to the control group. Note that the methods appropriate for paired data more

closely resemble those for one sample of data (see the preceding section) than those for two independent samples (see the next section).

Calculations for Paired Continuous Data Suppose we plan to collect two measurements, x and y, of each subject or pair of subjects. We assume that the measurements (x, y) come from populations with means mx and my and variances s2x and s2y , respectively. Let d ¼ y
x for each subject. We assume that the difference d has (approximately) a Normal distribution with unknown mean md and known variance s2d . Consider the following null and alternative hypotheses about the mean md: H0 : md ¼ m0 vs. H1 : md ¼ m1 :

(25.19)

Let d ¼ m1
m0 denote the scientifically or clinically meaningful difference, and let d ¼ y
x denote the sample difference in means. Suppose we plan to conduct a two-sided hypothesis test by using the paired z-statistic,   pffiffiffi z ¼ d
m0 sd n ; (25.20) at the a significance level. To calculate the required sample size, we specify the power (1 e b), and use the following formula:15
2  n ¼ Z1
a=2 þ Z1
b s2d d2 : (25.21) Observe that Eq. (25.21) resembles Eq. (25.13) for sample size calculations for one sample of continuous data. When the sum of variances of the two original measurements are larger than the variance of the difference
i.e.; s2x þ s2y > s2d , a design based on a paired sample

requires a smaller sample size than a design based on two independent samples. Example 5. Suppose an investigator wishes to design a pilot study to investigate the effect of a new medication on diastolic blood pressure in hypertensive patients. He plans to take two measurements of each subject, one measurement at baseline when the subject has not yet taken the medication (x), followed by a second measurement when the subject has been taking the medication for 12 weeks (y). He then plans to compute the difference between these measurements for each subject (d). Past laboratory measurements suggest that the standard deviations of the original measurements are sx ¼ sy ¼ 20 mm Hg, respectively. The investigator wishes to perform a two-sided paired z-test at the 5% significance level (a ¼ 0.05) regarding whether there is a change in average diastolic blood pressure on the new medication. He wants a 90% chance to reject the

II. STUDY DESIGN AND BIOSTATISTICS

SAMPLE SIZE CALCULATIONS FOR HYPOTHESIS TESTS: PAIRED DATA

null hypothesis of equality if the true difference is d ¼ 3 mm Hg in either direction (90% power, b ¼ 0.1). If past measurements suggest that the standard deviation of the difference is sd ¼ 15 mm Hg, what sample size does he need? This hypothesis is two-sided, so Z1ea/2 ¼ 1.960 and Z1eb ¼ 1.282. Using the previous formula, we calculate . n ¼ ð1:960þ 1:282Þ2 ð15Þ2 ð3Þ2 ¼ 262:7 z 263: (25.22) Thus, the minimum required sample size is n ¼ 263 subjects for this paired-sample study. Remember that sample size calculations depend heavily on the particular parameters chosen, so it is wise to repeat these calculations under different assumptions. Suppose the investigator can reduce the within-subject variation by taking three repeated measurements of each subject while not taking medication and while taking the new medication. If the standard deviation of the difference in the means of the three repeated measurements is sd ¼ 12 mm Hg, what sample size does he need? Using Eq. (25.21), we calculate . n ¼ ð1:960þ 1:282Þ2 ð12Þ2 ð3Þ2 ¼ 168:2 z 169: (25.23) Thus, the minimum required sample size is n ¼ 169 subjects, about two-thirds of the previous sample size calculation. By comparison, if the meaningful difference is instead d ¼ 6 mm Hg, what sample size does he need? Using Eq. (25.21), we calculate . n ¼ ð1:960þ 1:282Þ2 ð12Þ2 ð6Þ2 ¼ 42:0 z 43: (25.24) Thus, the minimum required sample size is n ¼ 43 subjects, about one-fourth of the previous sample size calculation. Note that this smaller sample size comes with an implicit price, namely, the decreased ability to detect differences smaller than this new meaningful difference. The design used in this example is a nonrandomized baseline-versus-treatment design, which may be subject to various problems, such as regression to the mean over time (e.g., participants who were selected because they were especially sick at screening may naturally progress to a less severe state, even in the absence of treatment efficacy), bias due to external trends over time (e.g., seasonal changes may affect participants in an asthma study), and investigator bias in evaluating treatment efficacy due to knowing when each participant took the active treatment (i.e., treatment assignments were not masked). An alternative design that mitigates these problems is the parallel-groups design, with or without baseline measurements (see “Calculations for continuous data with equal variances and equal sample sizes,” below).

365

Calculations for Paired Binary Data Next, consider the case of paired binary data (a, b). Such data may arise, for example, in a study in which the same condition affects both eyes and each eye is given a different treatment, labeled A and B. When outcomes for each treatment are coded as 0 (i.e., no, failure) and 1 (i.e., yes, success), there are four possible joint outcomes (0, 0), (0, 1), (1, 0), and (1, 1) with corresponding population proportions p00, p01, p10, and p11, respectively (where p00 þ p01 þ p10 þ p11 ¼ 1). The success rate of treatment A is p10 þ p11, whereas the success rate of treatment B is p01 þ p11. In turn, the difference between the success rates of these treatments is p10
p01. One possible approach for paired binary data is to test whether the difference in success rates p10
p01 equals zero (i.e., both treatments are equally effective) or a specific meaningful difference d (i.e., treatment A is better than treatment B if d > 0), conditional on the sum of the two discordant joint outcomes p01 þ p10 being equal to q. This approach translates into the following null and alternative hypotheses about the discordant difference p10
p01: H0 : p10
p01 ¼ 0 vs. H1 : p10
p01 ¼ d;

(25.25)

given p10 þ p01 ¼ q: Let p00, p01, p10, and p11 denote the sample proportions in the four joint outcome categories. Suppose we plan to conduct a two-sided hypothesis test by using the z-statistic corresponding to McNemar’s test, pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi p10 þ p01 : (25.26) z ¼ nðp10
p01 Þ at the a significance level. To calculate the required sample size, we specify the power (1
b), and use the following formula:15,16
pffiffiffiffiffiffiffiffiffiffiffiffiffi 2 . pffiffiffi n ¼ Z1
a=2 q þ Z1
b q
d2 (25.27) d2 : Example 6. Suppose a pediatrician wishes to determine whether the right-handed children in her practice are more likely to get infections in their right ear than in their left ear over the course of a year. The four possible joint outcomes are (0, 0), no ear infections; (0, 1), at least one infection in the right ear only; (1, 0), at least one infection in the left ear only; and (1, 1), at least one infection in each ear, where infections may occur concurrently or sequentially. She observes that, in a typical year, about 50% of infants never have an ear infection, 40% have at least one infection in either their left or right ear (but not both ears), and 10% have at least one infection in each ear. The null hypothesis is that among right-handed children, 20% have at least one infection in their left ear only, and 20% have at least one

II. STUDY DESIGN AND BIOSTATISTICS

366

25. POWER AND SAMPLE SIZE CALCULATIONS

infection in their right ear only. The alternative hypothesis chosen by the pediatrician is that 18% have at least one infection in their left ear only, and 22% have at least one infection in their right ear only (q ¼ 0.40, d ¼ 0.04). She wants to perform a one-sided McNemar’s test at the 5% significance level (a ¼ 0.05). If she wants a 90% chance to reject the null hypothesis of equality when the alternative hypothesis is true (90% power, b ¼ 0.1), how many medical records should she review for this study? This hypothesis is one-sided, so Z1ea ¼ 1.645 and Z1eb ¼ 1.282. Using the previous formula, we calculate  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2

pffiffiffiffiffiffiffi n ¼ ð1:645Þ 0:4 þ ð1:282Þ 0:4
ð0:04Þ2 ð0:04Þ2 ¼ 2138:1 z 2139: (25.28) Thus, the minimum required sample size is n ¼ 2139 right-handed children for this paired binary data study.

SAMPLE SIZE CALCULATIONS FOR HYPOTHESIS TESTS: TWO INDEPENDENT SAMPLES In this section, we discuss sample size calculations for hypothesis tests based on two independent samples when we wish to estimate the difference of two population means or proportions. For continuous data, we consider populations with equal and unequal variances and groups with equal and unequal sample sizes.

Calculations for Continuous Data With Equal Variances and Equal Sample Sizes Suppose we plan to collect the same number of measurements from two independent populations or groups, labeled X and Y, with (approximately) Normal distributions with unknown means mx and my and common known variance s2x ¼ s2y ¼ s2c . Consider the following null and alternative hypotheses about the difference in the population means: H0 : my
mx ¼ d0 vs. H1 : my
mx ¼ d1 :

(25.29)

Let d ¼ d1
d0 denote the scientifically or clinically meaningful difference. Suppose we plan to conduct a two-sided hypothesis test by using the z-statistic, .
pffiffiffiffiffiffiffiffi  z ¼ ½ðx
yÞ
d0  sc 2=n ; (25.30) at the a significance level. To calculate the required common sample size in each group, ne, we specify the power (1
b), and use the following formula15:
2  (25.31) ne ¼ 2 Z1
a=2 þ Z1
b s2c d2 :

In turn, the total sample size equals twice the sample size in each group, i.e., n ¼ 2ne. Example 7. Let us revisit the scenario considered in Example 5. Suppose an investigator wishes to design a pilot study to investigate the effect of a new medication on diastolic blood pressure in hypertensive patients by using a parallel-groups design. He plans to randomly assign patients to receive either the treatment or a placebo. Measurements will be collected at baseline and after 12 weeks’ follow-up. The investigator plans to use a two-sided z-test to determine whether the change in the treatment group is different from that in the placebo group at the 5% significance level (a ¼ 0.05). He wants a 90% chance to reject the null hypothesis of equality if the true difference is d ¼ 3 mm Hg in either direction (90% power, b ¼ 0.1). If past measurements suggest that the common standard deviation of the change in both groups is sc ¼ 15 mm Hg, what sample size does he need for each group? This hypothesis is two-sided, so Z1ea/2 ¼ 1.960 and Z1eb ¼ 1.282. Using the previous formula, we calculate . ne ¼ 2ð1:960þ 1:282Þ2 ð15Þ2 ð3Þ2 ¼ 525:6 z 526: (25.32) Hence, the minimum required sample size is ne ¼ 526 subjects for each group, and thus n ¼ 2ne ¼ 1052 subjects in total. As before, suppose the investigator can reduce the variation by taking three repeated measurements of each subject while not taking medication and while taking either the new medication or a placebo. If the standard deviation of the difference in the means of the three repeated measurements is sc ¼ 12 mm Hg, what sample size does he need for each group? Using Eq. (25.31), we calculate . ne ¼ 2ð1:960þ 1:282Þ2 ð12Þ2 ð3Þ2 ¼ 336:3 z 337: (25.33) Hence, the minimum required sample size is ne ¼ 337 subjects for each group, about two-thirds of the previous sample size calculation, and thus n ¼ 2ne ¼ 674 subjects in total. By comparison, if the meaningful difference is instead d ¼ 6 mm Hg, what sample size does he need for each group? Using Eq. (25.31), we calculate . ne ¼ 2ð1:960þ 1:282Þ2 ð12Þ2 ð6Þ2 ¼ 84:1 z 85: (25.34) Hence, the minimum required sample size is ne ¼ 85 subjects for each group, about one-fourth of the previous sample size calculation, and thus n ¼ 2ne ¼ 170 subjects in total. The randomized parallel-groups design presented here requires four times as many subjects as the nonrandomized baseline-versus-treatment design with comparable design parameters (presented in “Calculations for Paired Continuous Data,”). The advantage of

II. STUDY DESIGN AND BIOSTATISTICS

SAMPLE SIZE CALCULATIONS FOR HYPOTHESIS TESTS: TWO INDEPENDENT SAMPLES

randomized designs is that they avoid many of the biases inherent in nonrandomized designs, such as investigator bias due to the lack of masking of treatment assignments. One strategy, often used in cancer research, is to screen many potential treatments on a small number of patients by using nonrandomized designs and then to test the most promising treatments more definitively by using randomized designs. This strategy makes efficient use of limited resources and potentially minimizes the exposure of patients to toxic, but ineffective, treatments.

Calculations for Continuous Data With Unequal Variances or Unequal Sample Sizes We can also calculate the required sample size for testing hypotheses about the difference between two population means when the variances in the two groups, s2x and s2y , are unequal. Suppose we plan to conduct a two-sided hypothesis test by using the z-statistic,

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
. s2x þ s2y ne ; z ¼ ½ðx
yÞ
d0  (25.35) at the a significance level. To calculate the required common sample size in each group, we use the following formula:11
2
. ne ¼ Z1
a=2 þ Z1
b (25.36) s2x þ s2y d2 : In some situations we may wish to design trials with different numbers of subjects in each group. For example, in placebo-controlled parallel-groups trials, we may wish to randomly assign a larger proportion of subjects to the new treatment than to the placebo. Let l ¼ ny =nx denote the desired ratio of sample sizes in the two groups. Suppose we plan to conduct a two-sided hypothesis test by using the z-statistic,

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 . ffi z ¼ ½ðx
yÞ
d0  s2x nx þ s2y ny ; (25.37) at the a significance level. To calculate the required sample size in the x-group, we use the following formula:11
2
. . nx ¼ Z1
a=2 þ Z1
b s2x þ s2y l d2 : (25.38) In turn, ny ¼ lnx . Note that the choice of l ¼ sy =sx minimizes the total sample size, which is n ¼ nx þ ny. Example 8. Continuing the preceding example, suppose the standard deviation in the new medication (treatment) group is 16 mm Hg and the standard deviation in the placebo group is 8 mm Hg, and all other design parameters are the same (d ¼ 3 mm Hg, a ¼ 0.05, b ¼ 0.1). Using Eq. (25.36), we calculate

i. ne ¼ ð1:960þ 1:282Þ2 ð16Þ2 þ ð8Þ2 ð3Þ2

367

h

¼ 373:7z374:

(25.39)

Thus, the minimum required sample size is ne ¼ 374 subjects for each group, and thus n ¼ 2ne ¼ 748 subjects in total. Next, suppose we wish to enroll twice as many subjects in the treatment group as in the placebo group. Let l ¼ 0.5. Using Eq. (25.38), we calculate h i nx ¼ ð1:960 þ 1:282Þ2 ð16Þ2 þ ð8Þ2 =ð0:5Þ =ð3Þ2 ¼ 448:5z449:

(25.40)

In turn, ny ¼ 225. Hence, the required sample sizes in the treatment and placebo groups are nx ¼ 449 and ny ¼ 225 subjects, respectively, and thus n ¼ nx þ ny ¼ 674 subjects in total.

Calculations for Two Independent Samples of Binary Data Next, we consider sample size calculations for comparing two independent binomial proportions calculated from binary data. Suppose we plan to sample equal numbers of binary data from two populations or groups, labeled A and B, with underlying population proportions pa and pb, respectively. One possible approach is to test whether the difference between the proportions pb
pa equals zero (i.e., both proportions are the same) or whether the proportions have specific alternative values pa ¼ qa and pb ¼ qb , where d ¼ qb
qa equals the scientifically or clinically meaningful difference. This approach translates into the following null and alternative hypotheses about the difference in the population proportions, pb
pa: H0 : pb
pa ¼ 0 vs. H1 : pb
pa ¼ d;

(25.41)

where d ¼ qb
qa : Let pa and pb denote the sample estimates of pa and pb,  respectively, and let p ¼ ðpa þ pb Þ 2 denote their average. Also, the difference between the sample proportions is pb
pa . Suppose we plan to conduct a twosided hypothesis test by using the z-statistic,

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2pð1
pÞ=n: z ¼ ðpb
pa Þ (25.42) at the a significance level. For computational convenience, let q ¼ ðqa þ qb Þ=2 denote the average of the   two alternative proportions, and let u0 ¼ 2q 1
q and u1 ¼ qa ð1
qa Þ þ qb ð1
qb Þ denote the variances of pb
pa under the null and alternative hypotheses,

II. STUDY DESIGN AND BIOSTATISTICS

368

25. POWER AND SAMPLE SIZE CALCULATIONS

respectively. To calculate the required common sample size in each group, we specify the power (1
b), and use the following formula:13
pffiffiffiffiffiffi pffiffiffiffiffiffi 2 . 2 ne ¼ Z1
a=2 u0 þ Z1
b u1 (25.43) d : Example 9. Suppose an immunologist wishes to compare mumps vaccination rates in two communities, one affected by a mumps outbreak among male adolescents and one unaffected. The null hypothesis is that the vaccination rates in the two communities are equal. The particular alternative hypothesis of interest is that the vaccination rates in the affected and unaffected communities are 80% and 90%, respectively. The immunologist plans to use a two-sided z-test to determine whether the vaccination rates differ in the affected and unaffected communities at the 5% significance level (a ¼ 0.05). He wants a 95% chance to reject the null hypothesis of equality if the true difference is d ¼ 10% (95% power, b ¼ 0.05). How many unrelated male adolescents should he enroll in each community for this study? This hypothesis is two-sided, so Z1
a/2 ¼ 1.960 and Z1
b ¼ 1.645. First, we calculate the average vaccination rate in both populations, q ¼ (0.80 þ 0.90)/2 ¼ 0.85, and the variances of pb
pa under the null and alternative hypotheses, u0 ¼ 2(0.85) (0.15) ¼ 0.255, and u1 ¼ (0.80) (0.20) þ (0.90) (0.10) ¼ 0.25. Using the previous formula, we calculate h pffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffi i2 . ne ¼ ð1:960Þ 0:225 þ ð1:645Þ 0:25 ð0:10Þ2 (25.44) ¼ 328:4z329: Thus, the minimum required sample size is ne ¼ 329 unrelated male adolescents in each community, for a total of n ¼ 2ne ¼ 658 in the entire study. By way of comparison, using an advanced method based on Fisher’s exact test, the minimum required sample size is 343 unrelated male adolescents in each community, for a total of 686 in the entire study.

ADVANCED METHODS AND OTHER TOPICS In this section, we consider various advanced methods that are beyond the scope of this chapter to put the basic methods presented here in perspective. We also discuss issues related to subject retention and modern statistical computing so that the reader will be familiar with these topics and understand how they may apply to a particular research study.

Alternative Statistics and Sample Size Calculation Methods We have emphasized that many of the methods for sample size calculation presented in this chapter are

approximate and that more advanced methods may be available. For instance, for continuous data, there are numerical algorithms based on the t-distribution, and for binary data, there are exact methods based on the binomial distribution, which provide more accurate sample size calculations. Likewise, in addition to the methods based on means and proportions (and their differences) presented here, methods exist for many other parametric statistics (e.g., correlation coefficient, odds ratio) and nonparametric statistics (e.g., Fisher’s exact test, Wilcoxon signed rank test for paired data, Wilcoxon rank sum test for two independent samples), which may be better suited to answering a particular research question.8,11 Standard statistical software can easily perform many of these advanced calculations.

Several Advanced Study Designs It is important to be familiar with several types of study designs and analytic methods for which sample size calculations are more complicated and for which special methods are needed. Some of these designs are as follows: • Multiarm clinical trials.17 Multiarm trials are an efficient way to compare several active treatments with a common placebo. The more hypothesis tests performed, however, the greater the overall chance of making a false discovery, that is, rejecting some null hypothesis and thus concluding that a particular active treatment is better than the placebo, when in truth the new treatment is ineffective. The Bonferroni correction can be used to control the overall experiment-wise or family-wise error rate in a single trial. Specifically, the required sample size for each of k comparisons between an active treatment and the placebo can be calculated using a significance level of a/k, giving an experiment-wise or family-wise error rate not to exceed a. • Group sequential trials.18 In this class of designs, the clinical trial is divided into several stages with specified (usually equal) numbers of subjects to be enrolled during each stage. After each stage is completed, the data are analyzed, and a decision is made to continue the trial and collect more data or to stop the trial early for either clear efficacy (i.e., the study shows that the new treatment is remarkably successful and that further data are unlikely to reverse this decision) or futility (i.e., the study, even if continued to its planned end, is unlikely to produce a decision in favor of the new treatment). The required sample size for the entire study can be obtained by calculating the sample size for a regular design without interim analyses and then inflating this number, by using one of several common methods,11,18 to reflect the multiple times at which the data are scheduled to be examined. The actual number of subjects enrolled can be significantly

369

CONCLUSION

smaller than the maximum sample size if the study is stopped early for either efficacy or futility. • Survival analysis.8,11,15 Survival analysis methods aim to estimate the distribution of the time to a particular event (e.g., relapse, infection, death) in each study group, possibly adjusting for baseline covariates of interest. A special challenge is that some subjects may not experience an event; instead they may be censored at a particular time, after which they are no longer observed in the study. Survival data are often analyzed using KaplaneMeier methods, Cox proportional hazards regression, and parametric regression modeling. For survival analysis studies the required sample size depends on both the expected rate of events in each study group and the expected rate of censoring. • Group randomized trials (GRTs)19 and cluster randomization trials.20 In this class of designs, groups or clusters of subjects, rather than individual subjects, are randomly assigned to the various treatment groups, although measurements of individual subjects are still made. Subjects in the same group tend to be more similar than subjects in different groups. Let s2b denote the variance between the means of the different groups at baseline (the between-group variance), and let s2w denote the variance between subjects in the same group (the within-group variance). To account for the dependence between subjects in the same group, we compute the intraclass correlation (ICC), denoted rICC ¼ s2b

 2  sb þ s2w :

(25.45)

Then, to obtain the required sample size for a particular GRT (nGRT), the calculated sample size for the regular design without groups (n) must be inflated to reflect the common, or average, group size (m) and the ICC: nGRT ¼ nð1þ ðm
1ÞrICC Þ:

(25.46)

• Generalized linear models (GLMs) (e.g., regression, repeated measures, and longitudinal data models).21 Suppose we wish to model the outcome, or response (whether continuous or binary), as a function of the treatment assignment or the main risk factor, plus other baseline covariates of interest. Although we may plan to use GLM methods to analyze the future data, the reality is that sample size calculations for such methods tend to be quite complicated and, in general, explicit formulas do not exist. Therefore, to obtain approximate sample sizes, we tend in practice to perform sample size calculations appropriate for a simplified method of analysis (e.g., a t-test for two independent samples instead of a multiple linear regression model).

Retention of Subjects Regardless of the simplicity or complexity of the study design, a major challenge in almost any clinical trial or study is retention, especially when the followup period for each subject lasts for many months or even years. Some subjects drop out (e.g., because of treatment side effects) or are lost to follow-up (e.g., because they move out of the study area). Furthermore, some subjects do not adhere to their assigned treatment (e.g., persons assigned to the placebo group who take the active treatment, i.e., they “drop in” to the treatment group). The intention-to-treat principle, discussed in Chapter 23, is an analytic strategy for dealing with these likely challenges. In many studies it is prudent to plan for a 10% to 20% rate of retention loss. As a rule of thumb, Lachin22 suggested inflating the calculated sample size by a factor of 1/(1
r)2, where r is the combined rate of drop-out, loss to follow-up, and other nonadherence to the study protocol. This formula adjusts for the loss of subjects plus the bias that results when key characteristics of the lost subjects differ from those of the study group as a whole.

Statistical Computing Modern statistical computing has become a great asset for sample size calculations. Most simply, statistical software facilitates the rapid calculation of numerous sample sizes (or statistical powers) for testing various hypotheses under a variety of assumptions about the underlying population parameters. Results can be arranged in tables to show how the calculated sample size varies as a function of changes in these assumptions. Also, numerical methods can provide solutions to sample size calculations when explicit formulas do not exist. Furthermore, for complex and otherwise intractable problems, data can be simulated under the null and alternative hypotheses for various sample sizes, and the empirical rejection rates of the null hypothesis can be compared with the desired Type I and Type II error rates, respectively. In addition, to help optimize the choice of sample size for a particular study design, special software can be used to graph power as a function of sample size for various parameter choices.

CONCLUSION In this chapter we have introduced the concept of statistical power, a key element in sample size calculations. We have discussed some basic sample size formulas for precision in confidence interval construction and for hypothesis testing, for both continuous and binary data. For hypothesis testing, we have considered formulas

II. STUDY DESIGN AND BIOSTATISTICS

370

25. POWER AND SAMPLE SIZE CALCULATIONS

for when we plan to collect one sample, a paired sample, and two independent samples of data. In addition, we have mentioned several advanced topics related to the choice of test statistics and hypotheses, study design, subject retention, and statistical computing. Performing sample size calculations can be one of the most challenging aspects of study design, but when statisticians and researchers collaborate closely, this process can be highly rewarding. Collaboration provides the opportunity to review the current state of knowledge, to clarify research objectives and hypotheses, and in turn to refine study designs. We advocate early and close collaboration, because once major design decisions have been made, it may be difficult to correct problems that may be found later. The goal of collaboration is to design powerful clinical trials and studies with sufficiently large sample sizes that will have a good chance of providing strong evidence to answer scientifically and clinically meaningful questions.

5.

6.

EXERCISES 1. Confidence interval for a mean. Suppose an infectious disease specialist wishes to estimate the mean CD4 counts among a population of HIVinfected pregnant women before starting treatment. He expects the data to have (approximately) a Normal distribution with a mean of 500 cells/mm3 and a standard deviation of 50 cells/mm3. If he wishes to obtain a 95% confidence interval with a width of 20 cells/mm3 for the true mean, show that he should enroll at least n ¼ 97 subjects in this study. 2. Confidence interval for a binomial proportion. Suppose a hematologist wishes to estimate the prevalence of Factor V Leiden among patients treated for a deep vein thrombosis. On the basis of past studies, she expects this prevalence to be approximately 25%. If she wishes to obtain a 95% confidence interval with a width of 0.1 (on average) for this prevalence, show that she should enroll at least n ¼ 289 subjects in this study. 3. Power and parameter choices. Use Eq. (25.13) to show that Z1eb and thus power (1
b) increases as the sample size (n), the significance level (a), and the meaningful difference (d) increase, and as the population variance (s2) decreases. Power is also greater for one-sided hypothesis tests than for comparable . two-sided hypothesis tests. (Hint: pffiffiffiffiffi Z1
b ¼ nd s
Z1
a=2 :) 4. One sample, continuous data. Suppose a biochemist wishes to study homocysteine levels in

7.

8.

blood specimens from men older than 50 who have cardiovascular disease. The mean serum homocysteine level among these men is 14 mmol/L before treatment, and she wants an 80% chance to reject the null hypothesis of no change if the mean serum homocysteine level drops to 12 mmol/L after these men take folate tablets for 10 weeks. She plans to use a one-sided z-test at the 5% significance level. If the standard deviation is 4 mmol/L, show that she needs to enroll at least n ¼ 25 patients in this study. One sample, binary data. Suppose an endocrinologist wishes to design a Phase II trial to test a new drug to reduce fatigue in diabetic patients. If the new drug reduces symptoms in only 30% of patients, its effect is not sufficient to merit further development. If the new drug reduces symptoms in 50% of patients, he wants a 95% chance to reject the null hypothesis of insufficient effect. He plans to use a one-sided z-test at the 5% significance level. Show that he needs to enroll at least n ¼ 63 patients in this clinical trial. Two-stage designs. Using a six-sided die, simulate data under the null and alternative hypotheses in Example 4, namely H0: p ¼ 0.2 versus H1: p ¼ 0.4, with a 5% significance level and 80% power. First, under the null hypothesis, treat rolls of 1 as treatment success, 2e5 as treatment failure, and reroll 6s. This procedure gives a 20% success rate per subject. Second, under the alternative hypothesis, treat rolls of 1e2 as treatment success, 3e5 as treatment failure, and reroll 6s. This procedure gives a 40% success rate per subject. Using the stopping rules in Example 4, how many subjects in total are enrolled in these simulated trials under each hypothesis? Repeat each simulation five times and compare the totals obtained to the fixed sample size, n ¼ 29 (Example 3) and the maximum two-stage sample size, m2 ¼ 43 (Example 4). Paired continuous data. Suppose a nutritionist wishes to study the weight change among obese men (BMI  30) on a 16-week low-fat diet, complemented by daily exercise. Assume that the standard deviations of the before and after weights are both 25 kg, whereas the standard deviation of the difference is 15 kg. She plans to use a two-sided paired z-test at the 5% significance level. She wishes to have a 90% chance to reject the null hypothesis of equality when the true change in weight (in either direction) is 8 kg. Show that if she correctly uses the paired sample size formula, she obtains a minimum required sample size of n ¼ 37 subjects. Paired binary data. Suppose an obstetrician wishes to compare the rates of testing for HIV and

II. STUDY DESIGN AND BIOSTATISTICS

371

REFERENCES

hepatitis B among pregnant women arriving in labor at a large hospital. One blood specimen should be taken, but for various reasons (e.g., imminent labor, patient’s refusal), one or both of the tests may not be performed. The obstetrician is concerned about a significant missed opportunity: some patients receive a hepatitis B test but fail to receive an HIV test at the same time. He estimates that 30% of pregnant women receive both tests but that 50% of women receive only one of the two tests. The null hypothesis is that 25% of women receive only an HIV test, whereas 25% of women receive only a hepatitis B test. The alternative hypothesis is that 20% of women receive only an HIV test, whereas 30% of women receive only a hepatitis B test (q ¼ 0.50, d ¼ 0.10). The obstetrician wants to perform a one-sided McNemar ’s test at the 5% significance level. Show that if he wants a 90% chance to reject the null hypothesis of equality when the alternative hypothesis is true, he needs to review n ¼ 425 medical records for this study. 9. Two independent samples, continuous data. Suppose a psychologist wishes to design a randomized parallel-groups trial to compare the impact of white noise and classical music (e.g., Mozart) on the performance of college students on a 200 question problem-solving test. On the basis of her knowledge of past students, she expects the white noise group to have a bell-shaped distribution with a mean of 120 points and a standard deviation of 15 points. She plans to compare the performance of the students in each group with a two-sided z-test at the 5% significance level. She wants a 90% chance to reject the null hypothesis of equality between the two groups when the performance in the music group is better or worse by 10 points. (a) If the standard deviation for the music group is also 15 points, show that she should enroll ne ¼ 48 students in each group of this study (total ¼ 96). (b) If the standard deviation for the music group is 30 points, show that she should enroll ne ¼ 119 students in each group (total ¼ 238). (c) If the standard deviation for the music group is 30 points, and she wishes to enroll twice as many students in the music group as in the white noise group, show that she should enroll nwn ¼ 71 and nm ¼ 142 students in the white noise (wn) and music (m) groups, respectively (total ¼ 213). 10. Two independent samples, binary data. Suppose a geneticist wishes to study the prevalence of sickle cell trait among two geographically separated populations in sub-Saharan Africa. He wants a 90% chance to reject the null hypothesis of equality if the

true prevalence is 10% in one population and 25% in the other. He plans to use a two-sided z-test at the 5% significance level. Show that he should enroll ne ¼ 133 subjects in each population for this study (total ¼ 266).

Acknowledgments The authors wish to thank Timothy A. Green, Lillian S. Lin, Marie S. Morgan, Philip J. Peters, Travis H. Sanchez, and Ryan E. Wiegand for helpful comments on this chapter.

Disclaimers The findings and conclusions in this chapter are those of the authors and do not necessarily represent the official position of the Centers for Disease and Control Prevention. This chapter reflects the views of the authors and should not be construed to represent FDA’s views or policies.

References 1. Altman DG, Bland JM. Absence of evidence is not evidence of absence. Br Med J 1995;311(7003):485. 2. International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. Guidance for industry: E9 statistical principles for clinical trials. Fed Regist 1998;63(179):49583e98. 3. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. p. 56. 4. Lee SJ, Zelen M. Clinical trials and sample size considerations: another perspective. Stat Sci 2000;15(2):95e110. 5. Schulz KF, Altman DG, Moher D, CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. Br Med J 2010;340. 698e702:c332. 6. Moher D, Hopewell S, Schulz KF, Montori V, Gøtzsche PC, Devereaux PJ, Elbourne D, Egger M, Altman DG. CONSORT 2010 Explanation and Elaboration: updated guidelines for reporting parallel group randomised trials. Br Med J 2010;340. 1e28:c869. 7. Altman DG. Practical statistics for medical research. Boca Raton, FL: Chapman & Hall/CRC; 1991. 8. Machin D, Campbell MJ, Tan SB, Tan SH. Sample size tables for clinical studies. 3rd ed. West Sussex, England: John Wiley & Sons Ltd.; 2009. 9. Beal SL. Sample size determination for confidence intervals on the population mean and on the difference between two population means. Biometrics 1989;45(3):969e77. 10. Piantadosi S. Clinical trials: a methodologic perspective. 2nd ed. Hoboken, NJ: John Wiley & Sons; 2005. 11. Chow S-C, Shao J, Wang H. Sample size calculations in clinical research. 2nd ed. Boca Raton, FL: Chapman & Hall/CRC; 2008. 12. Guenther WC. Sample size formulas for normal theory T-tests. Am Statistician 1981;35(4):243e4. 13. Fleiss JL, Levin B, Paik MC. Statistical methods for rates and proportions. 3rd ed. Hoboken, NJ: John Wiley & Sons; 2003. 14. Simon R. Optimal two-stage designs for phase II clinical trials. Controlled Clin Trials 1989;10(1):1e10. 15. Friedman LM, Furberg CD, DeMets DL. Fundamentals of clinical trials. 4th ed. New York: Springer; 2010. 16. Connor RJ. Sample size for testing differences in proportions for the paired-sample design. Biometrics 1987;43(1):207e11.

II. STUDY DESIGN AND BIOSTATISTICS

372

25. POWER AND SAMPLE SIZE CALCULATIONS

17. Freidlin B, Korn EL, Gray R, Martin A. Multi-arm clinical trials of new agents: some design considerations. Clin Cancer Res 2008; 14(14):4368e71. 18. Jennison C, Turnbull BW. Group sequential methods with applications to clinical trials. Boca Raton, FL: Chapman & Hall/CRC; 2000. 19. Murray DM. Design and analysis of group-randomized trials. New York: Oxford University Press; 1998.

20. Donner A, Klar N. Design and analysis of cluster randomization trials in health research. London: Arnold; 2000. 21. Chow S-C, Liu J-P. Design and analysis of clinical trials: concepts and methodologies. 3rd ed. Hoboken, NJ: John Wiley & Sons, Inc.; 2013. 22. Lachin JM. Introduction to sample size determination and power analysis for clinical trials. Controlled Clin Trials 1981;2(2):93e113. Correction 2(4):337.

II. STUDY DESIGN AND BIOSTATISTICS

C H A P T E R

26 An Introduction to Survival Analysis Laura Lee Johnson U.S. Food and Drug Administration, Silver Spring, MD, United States

O U T L I N E Introduction

373

Features of Survival Data

374

Survival Function KaplaneMeier and Product-Limit Estimators Calculation and Formula for an Estimate Calculation of Variance Comparing Two Survival Functions Comparing Two Survival Functions at a Given Time Point Comparing Two Survival Functions Using the Whole Curve: Log-Rank Test Example 1: Chronic Active Hepatitis Study

375 375 375 376 377 377 377 378

This chapter introduces some commonly used statistical methods for the analysis of survival time data in medical research. Survival data consist of two pieces of information for each subject: the time under observation and the ultimate outcome at the end of that time. Analysis of survival time or time-to-event data is complicated because the follow-up length often is different for each participant, and the event of interest, such as myocardial infarction (MI), often is not observed in all the subjects by the end of the study. For those participants in whom the event of interest is not observed, what is known is that their survival times are longer than their time spent in the study, but their exact survival times are unknown. This chapter describes features of survival time data, defines the true or underlying survival function, and introduces the KaplaneMeier or product-limit estimator for the survival function. It also presents several approaches for comparing two survival curves, a summary

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00026-5

Stratified Log-Rank Test

379

Proportional Hazards Model Calculation and Formulas

379 379

Common Mistakes

380

Conclusion

380

Questions

381

Acknowledgments

381

Disclaimer

381

References

381

of stratified analysis methods, and Cox’s proportional hazards regression analysis.

INTRODUCTION In survival analysis the main interest focuses on the time taken for some dichotomous event to occur. Although the term survival is used, the event of interest is not limited to death or failure. It can be any dichotomous event, such as nonfatal MI, adverse events, computer crashes, or bursting of a balloon filling with air; essentially it can be any definable event. Several different fields have different names for similar types of analyses. Survival, time to event, and failure time analysis are common names in the medical world; reliability theory and reliability analysis are common names in engineering, and duration modeling and

373

Copyright © 2018. Published by Elsevier Inc.

9 8 Patient 5 6 7 4 3 2 1 0

5

10

15

Time (months)

Patient 5 6

7

8

9

10

FIGURE 26.1 Diagram of patient accrual and follow-up from the data from Table 26.1. Solid circles, uncensored observation; open circles, censored observation.

4

Survival time is defined as the time from some fixed starting point (time origin) to the onset of the event of interest. In experimental animal studies often the starting point is the same time for all subjects. In controlled clinical trials the starting point frequently is the actual time a participant enters the study, thus the starting point may vary for each participant. In epidemiology, the time origin may be birth, time of first exposure, or another point in time. There are two key features of survival data. First, the length of follow-up time varies among participants. For example, for a study with a fixed end date, participants entering the study later on would have shorter followup time than those entering the study earlier. Second, the event of interest is almost never observed in all subjects by the end of a study. The survival time is called censored if the event is not observed by the end of the study; this indicates the period of observation ended before the event occurred. This type of censoring is referred to as right censoring, the most common type of censoring in clinical studies. Censoring may occur for various reasons. One common reason is that the study ends before the event occurs. Such censoring is called administrative censoring. Other reasons for censoring include the withdrawal of participants from the study and loss of contact with participants who move out of the study area. Censoring for reasons unrelated to the outcome for each participant (i.e., the occurrence of the event of interest or not) is called independent censoring. In all the methods presented in this chapter, the assumption of independent censoring is required. The diagrams in Figs. 26.1 and 26.2 and Table 26.1 are commonly used to illustrate the features of survival data. In this example, patient accrual occurs during the first 6 months of the study. After that, participants are monitored for a minimum of 12 months. The total length of the study is 18 months. This is an example of a study in which the total possible follow-up time will vary among study participants, in this case between 12 and 18 months, based on when each participant enters the study. The patients accrued earliest are observed for the longest time. Fig. 26.1 illustrates the staggered entry of participants into the study during the 6-month accrual period. Many survival studies have this pattern of participant accrual. The standard statistical methods for survival analysis assume that those participants

3

FEATURES OF SURVIVAL DATA

2

duration analysis are among the synonyms found in sociology and economics. Nevertheless the questions being answered and the statistical methods and concepts are similar. For this chapter we will take the clinical research frame of reference and use the term “survival.”

10

26. AN INTRODUCTION TO SURVIVAL ANALYSIS

1

374

0

FIGURE 26.2

5 10 Survival Time (months)

15

Diagram of the survival times for Table 26.1.

who enter the study at any given time are a representative random sample of those in the population still at risk at that time. Furthermore, the assumption of population homogeneity over time is made, namely that the characteristics of the population available for sampling remain essentially constant over time, at least to a reasonable approximation. These assumptions are particularly important in choosing how to estimate the hazard function, discussed later in the chapter. Looking at Fig. 26.1 more closely, we see the first participant was recruited at the beginning of the study (time 0) and had an event at approximately month 10. The second participant also was recruited at the beginning of the study but censored at approximately month 11. The survival time for each participant is obtained by subtracting the time of entry into the study from either the time of the event or the time of last follow-up without observing the event. Fig. 26.2 provides a modified presentation of the data in Fig. 26.1, moving the lines so all the survival times start from time 0.

II. STUDY DESIGN AND BIOSTATISTICS

375

SURVIVAL FUNCTION

TABLE 26.1

Data From First Hypothetical Example

Patient No.

Time at Entry (Months)

Time at Death or Censoring (Months)

Dead (D) or Censored (C)

Survival Time (Months)

1

0.0

10.6

D

10.6

2

0.0

11.5

C

11.5

3

0.4

16.0

C

15.6

4

1.1

6.2

D

5.1

5

1.3

7.1

C

5.8

6

3.5

10.2

D

6.7

7

3.9

18.0

C

14.1

8

4.5

16.1

D

11.6

9

5.2

18.0

C

12.8

10

5.9

18.0

C

12.1

Fig. 26.2 illustrates the survival time of each participant and provides a simpler picture than the first figure for comparing the survival times among the participants. This figure would not necessarily be a better way to examine the data if the assumption of time homogeneity did not apply. Also note that even though the study could follow participants for a minimum of 12 months, only 4 of the 10 participants were followed for 12 or more months; two participants were censored and six others died prior to being followed for 12 months. For additional reading, similar examples can be found in introductory textbooks.1e4

SURVIVAL FUNCTION The survival function, denoted by S(t), is the probability of an individual surviving at least until time t, where 0
S(t)
1. If the survival function is known from theory or empirical observations, then we can use it to analyze the survival experience of a population at various time points. For example, if S represents the survival experience of post-MI patients, then we can estimate the probability of surviving 3 years post-MI. However, information about the survival experience is usually limited to a sample of individuals rather than the whole population, meaning we do not observe the true survival function except with a complete census. We use our sample to estimate the survival function and make inferences from this estimate to the population of interest. Note that in statistics the phrase “survival experience” is used to denote a statistical function of the time to an event, not as a phrase to describe the quality of life among participants in a clinical study.

KaplaneMeier and Product-Limit Estimators The standard estimator of the survival function, proposed by Kaplan and Meier, is called the KaplaneMeier (K-M) or product-limit estimator.5 This estimator is obtained by taking the product of a sequence of conditional probabilities to create the KaplaneMeier curve, an estimate of the true survival function. Similar methods such as life tables or actuarial analysis are used in other fields. Although the names are used interchangeably in some fields, these three items differ based on how precisely times are recorded in the data set. For our purposes, we will discuss only the K-M. We first formally define this estimator and then consider a simple example in Table 26.2. Calculation and Formula for an Estimate Formally, the product-limit estimator is defined as: Y b SðtÞ ¼ ð1  fi =ri Þ; (26.1) i:ti
t

where the ti’s are the ordered observed survival times or censoring times from the sample, fi is the number of events that occur at a time ti, and ri is the number of individuals at risk at ti. In Table 26.2, at time 0 we started following 20 individuals and wait for a defined event to occur. At 5 months there are 20 individuals at risk, of whom 2 have the event of interest. Thus, the estimate of the probability of surviving beyond 5 months is equal to 1  (number having the event at 5 months/number at risk at 5 months), which is 1  2/20 ¼ 0.90. At 6 months, 18 individuals are still at risk but no one has an event. Among those at risk, the probability of survival beyond 6 months is equal to the product of the probability of surviving 5 months and the

II. STUDY DESIGN AND BIOSTATISTICS

376

26. AN INTRODUCTION TO SURVIVAL ANALYSIS

TABLE 26.2

Construction of a Product-Limit Estimator

Time, ti (Month)

No. at Risk, ri

No. of Events, fi

Product-Limit Estimator

Cumulative No. Censored, mc(ti)

Cumulative No. of Events, me(ti)

0

20

0

1.00

0

0

5

20

2

1 e 2/20 ¼ 0.90

0

2

6

18

0

(1 e 0/18)  0.90 ¼ 0.90

0

2

10

15

1

(1 e 1/15)  0.90 ¼ 0.84

3

3

13

14

2

(1 e 2/14)  0.84 ¼ 0.72

3

5

placed under the X-axis of the graphs at each time point to clarify the sample size at each time point. Calculation of Variance Since Sˆ is an estimator, it varies from sample to sample. Such variability is measured by the variance of Sˆ that, when the sample size is large, can be approximated using the Greenwood estimator   X fi Var b SðtÞ z b SðtÞ2 : (26.2) r ðr  fi Þ i:t
t i i i

With the product-limit estimate and its variance, one can make inferences about the survival probability at various time points. For example, we can construct a 95% confidence interval for S(13) from the preceding example. The variance of Sˆ(13) is approximately equal to 0.0115. Thus, p the 95%ffi confidence interval at ffi t ¼ 13 is ffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffi (0.72 e 1.96  0:0115, 0.72 þ 1.96  0:0115) ¼ (0.51, 0.93). The K-M curve or product-limit estimate of the data in Table 26.1 is plotted in Fig. 26.4. There are six tick marks on the curve corresponding to the six times at which censorings occur. We can use the data to find the probability of surviving beyond 6 months is 0.9 with variance 0.009.

0.8 0.6 0.2

0.4

Proportion Surviving

0.8 0.6 0.4 0.0

0.0

0.2

Proportion Surviving

1.0

1.0

conditional probability of surviving 6 months given having survived 5 months, which in this data set is 0.90  1 ¼ 0.90. Next, three individuals are censored between 6 and 10 months, and thus they are not at risk at 10 months. That leaves 15 individuals at risk at 10 months: 20 minus 2 who had an event before 10 months, minus 3 more who were censored before 10 months. The estimate of the probability of surviving beyond 10 months, conditional on having survived up to that point, is equal to 1  (number having the event at 10 months/ number at risk at 10 months), or 1  1/15 z 0.67. The probability of survival beyond 10 months is equal to the product of the probability of surviving up to 10 months and the conditional probability of surviving 10 months, given having survived 9 months, written as 0.90  (1  1/15) z 0.90  0.67 z 0.84 in Table 26.2. The later survival probabilities were obtained the same way. The K-M curve or product-limit estimate from this example is plotted in Fig. 26.3. It is a step function in which steps occur at the observed survival times. Notice that the survival curve remains flat at the times when censorings occur. Many statistical graphing packages place a tick mark on the curve to indicate that a censoring has occurred. Sample sizes also should be

0

5 10 Survival Time (months)

15

0

5 10 Survival Time (months)

15

FIGURE 26.3 Product-limit survival curve from the data in

FIGURE 26.4 Product-limit survival curve from the data in

Table 26.2.

Table 26.1.

II. STUDY DESIGN AND BIOSTATISTICS

377

SURVIVAL FUNCTION

The Greenwood variance estimator tends to underestimate the variance in the tails of the survival distribution. Other variance estimators, including those by Borkowf, attempt to alleviate this problem.6 The interested reader may consult a statistician for more information.

null hypothesis and conclude that the two survival probabilities at the prespecified time point are not equal. Otherwise, we fail to reject the null hypothesis, and hence there is insufficient evidence to conclude that the two survival probabilities are different at that time t. In addition, one can calculate the p-value for the test statistic. If the p-value is smaller than a, then we conclude that the two survival probabilities are different.

Comparing Two Survival Functions We may use hypothesis tests to compare the survival functions for two independent groups. These groups need to be defined based on information from before the follow-up period, in other words information available at baseline. For example, in a clinical trial, patients are randomized to either treatment A or treatment B. There are two methods of comparison. The first is to compare the two survival functions at a prespecified time point. One might be interested in comparing the survival experience between two groups at 1 year postrandomization. Alternatively, instead of comparing the survival at a fixed time point, one may want to compare the overall survival experience. In this case, the comparison is made across the entire range of the survival times. We first consider methods for the fixed time point comparison. Then we introduce the log-rank test statistic to compare the overall survival experience. Comparing Two Survival Functions at a Given Time Point Formally, the null hypothesis to be tested is written as H0 : S1 ðtÞ ¼ S2 ðtÞ;

(26.3)

where t is a prespecified time point, and subscripts on the two survival functions correspond to the two study groups. The alternative hypothesis is that the two survival curves are different at that time point. To test these hypotheses, we calculate the product-limit estimate at time t for each sample, namely Sˆ1(t), Sˆ2(t), and form the test statistic Z¼ 

b S 2 ðtÞ S 1 ðtÞ  b  0 SE b S 1 ðtÞ  b S 2 ðtÞ

 where SE b S 1 ðtÞ b S 2 ðtÞ ¼

(26.4)

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi     var b S 1 ðtÞ þ var b S 2 ðtÞ :

As with any test statistic, Z varies from sample to sample, and thus it has a sampling distribution. When the two samples are independent and the sample size is large, Z has approximately a standard normal distribution under the null hypothesis. If the Z statistic is beyond the lower or upper a/2  100 percentile (for a two-sided test) or a  100 percentile (for a one-sided test) of the standard normal distribution, we reject the

Comparing Two Survival Functions Using the Whole Curve: Log-Rank Test The comparison at a specific time point presented in the previous section is not completely satisfactory for the following reasons, given by Pocock.7 First, the time point chosen for comparison is usually, although not always, arbitrary. Second, there is always a danger that one may choose the time point post hoc, at the time when the largest difference occurs, which would tend to exaggerate the differences between the two survival functions. If someone chose to measure differences at 3.69 years, it would look suspicious but at 1, 2, 4, 6 years, it would look and actually might be perfectly reasonable. Third, one is not making full use of the precise survival times for each patient. The log-rank test is the most common method for comparing the overall survival experience between two groups.8 The null hypothesis to be tested is H0: S1($) ¼ S2($) where the dot means the whole range of the survival times on the curve. The alternative hypothesis is the negation of this null hypothesis. Thus, rather than an arbitrary fixed time point, we will test whether the two independent groups come from populations with the same overall survival experience. The log-rank test also does not make use of the actual times at which events and censoring occur but rather uses the ranks of those times. The log-rank test essentially compares the observed number of events with the number expected to occur if the two groups have the same underlying true survival functions. Specifically, we arrange the distinct survival times from the two groups in an ascending order, excluding censored survival times. Let {t1, t2,., tK} denote these ordered survival times. As illustrated in Table 26.3 at each time tj, j ¼ 1,., K, we construct a 2  2 table.

TABLE 26.3

Log-Rank Test Example 2  2 Table No. Dead

No. Surviving

Total

Group 1

aj

bj

aj þ bj

Group 2

cj

dj

cj þ dj

Total

aj þ cj

bj þ dj

nj

II. STUDY DESIGN AND BIOSTATISTICS

378

26. AN INTRODUCTION TO SURVIVAL ANALYSIS

If the null hypothesis is true, then the expected number of deaths in group 1 at time j, denoted by E(aj), is equal to total # at risk in group 1 Eðaj Þ ¼ total # of events at tj  total # at risk    ¼ aj þ cj aj þ bj =nj ;

2, 3, 4, 7, 10, 22, 28, 29, 32, 37, 40, 41, 54, 61, 63, 71, 127þ, 140þ, 146þ, 158þ, 167þ, 182þ

2, 6, 12, 54, 56þ, 68, 89, 96, 96, 125þ, 128þ, 131þ, 140þ, 141þ, 143, 145þ, 146, 148þ, 162þ, 169, 173þ, 181þ Censored survival times are indicated by a plus sign (þ)

1.0

(26.6)

We form a 2  2 table for each of the ordered survival times (thus we have K tables), and we then calculate the test statistic using the results from these tables:   K  P aj  E aj

J¼1

The Z statistic has approximately the standard normal distribution under H0. If the Z statistic is beyond the lower or upper a/2  100 percentile (a  100 percentile for one-sided test) of the standard normal distribution, then we reject H0 and conclude that the two samples are not from populations with the same survival experience. In addition, we can calculate the p-value for this test. If the p-value is smaller than a, then the null hypothesis is rejected. Otherwise, there is insufficient evidence to conclude that the two survival functions are different. Variations to the log-rank test exist, and some differentially weight events at different parts of the time scale. Different variations used on the same data set may lead to different answers. Additionally, log-rank methods should be cautiously interpreted if the observed survival curves cross. Example 1: Chronic Active Hepatitis Study Kirk et al. randomized 44 patients with chronic active hepatitis to either prednisolone or an untreated control group.9 Pocock analyzed these data in detail. Their survival times are listed in Table 26.4. The productlimit estimates are plotted in Fig. 26.5. It shows that the patients treated with prednisolone have a better survival experience overall. We consider two-tailed hypothesis tests with a ¼ 5%. We first compare the survival proportions at 5 years (60 months), where the product-limit estimate is 0.41 in the control group and 0.82 in the prednisolone group. The standard error of

0.0

0.2

(26.7)

Control Prednisolone

Proportion Surviving 0.4 0.6 0.8

and the variance of aj is equal to      a j þ b j a j þ c j b j þ dj c j þ dj   Varðaj Þ ¼ nj  1 n2j

J¼1

Survival times (months) in the control group

Survival times (months) in the prednisolone group

(26.5)

Z ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : K P Varðaj Þ

TABLE 26.4 Survival Data for 44 Patients With Chronic Active Hepatitis

0

50 100 Survival Time (months)

150

FIGURE 26.5 Product-limit survival curves from the chronic active hepatitis study.

the difference of these two estimates is 0.1322. Using Eq. (26.4), the Z statistic equals (0.41 e 0.82)/ 0.1322 ¼ 3.08. Because the Z statistic is lower than 1.96, the 2.5th percentile of the standard normal distribution, we reject the null hypothesis and conclude that patients receiving prednisolone have better survival probability at 5 years than untreated patients. Next, we compare the overall survival experience between the two groups using the log-rank test. To illustrate how the log-rank test works, consider one survival time at 10 months. At this time point, there are 18 patients at risk in the control group, of whom 1 dies, and there are 20 at risk in the prednisolone group, of whom none dies. The expected number of events in the control group, under the null hypothesis, is 0.47 with variance 0.25. We need to repeat this calculation for every observed death time and then use these results to form the Z statistic. The Z statistic equals 2.16, which is greater than 1.96, the 97.5th percentile of the standard normal distribution. Thus, we conclude that patients treated with prednisolone have better overall survival than those untreated.

II. STUDY DESIGN AND BIOSTATISTICS

379

SURVIVAL FUNCTION

Stratified Log-Rank Test In a clinical study, it is sensible to collect information on each participant’s personal characteristics (e.g., age, sex, and ethnicity) and history of disease and treatment. Some of these factors may be related to the participant’s survival experience. For example, young people usually have a better survival experience than old people. If the prognostic factors are related to survival and they are balanced between the two treatment groups, then the log-rank test presented in the previous section is appropriate. Otherwise, the log-rank test may be biased. Specifically it is possible that the two populations have the same survival experience, yet the logrank test declares they are different. Confounding is a change in the relationship between the outcome and treatment due to an imbalance in another variable. When confounding occurs, the difference in survival experiences observed between the two groups actually comes from the differences in the prognostic factors rather than the treatment. One solution to alleviate this bias is to calculate the difference in survival experience between the two groups within each level of the prognostic factors. For example, if the survival experience is dramatically different for patients with a history of MI compared to patients without a history of MI, then we may want to compare the survival difference among patients with and without the history of MI separately. By doing the analysis within each level of the prognostic factors, any observed difference in survival can be attributed to the treatment effect. The stratified log-rank test is the log-rank test that accounts for the difference in the prognostic factors between the two groups. Specifically, we divide the data according to the levels of the significant prognostic factors and form a stratum for each level. At each level we arrange the survival times in ascending order and calculate the observed number of events, expected number of events, and variance at each survival time as we would do in the regular log-rank test. Let aij denote the number of events in group 1 in the ith stratum at time tij. Then the stratified log-rank test has the following form:   M P K  P aij  E aij i¼1 j¼1

Z ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi ; PP Var aij i

(26.8)

j

where m is number of strata, Ki is number of survival times in the ith stratum, E(aij) is the mean of aij, and Var(aij) is the variance of aij. Under the null hypothesis of no difference in survival between the two groups, and when the sample size in each stratum is large, the

Z statistic has approximately the standard normal distribution.

Proportional Hazards Model The stratified log-rank test is a useful method for comparing the survival between two treatment groups while accounting for the effects of prognostic factors. However, it has some limitations. First, quantitative prognostic factors must be categorized to form strata. Second, if there are many prognostic factors, each with several levels, the number of strata can quickly become large with few patients in each stratum. This results in loss of power in the stratified test. Finally, whether we use the unstratified (regular) or stratified log-rank test, it is primarily a significance test and it does not estimate the magnitude of the difference between the survival experiences. For these reasons, we need a method that adjusts for both categorical and continuous prognostic factors in providing an estimate of the treatment difference. The proportional hazards model proposed by Cox aims at achieving these goals.10 Specifically, it uses regression methods to model the shape of the hazard function with one empirical part that depends on time and a second exponential part that depends on the other covariates. In the following section we describe the concepts behind the basic Cox proportional hazards model. This model is highly flexible and, with appropriate caution, it is possible for the proportional hazards assumption to be dropped. Dropping this assumption and other more advanced survival topics are discussed in Chapter 27. Calculation and Formulas Consider a patient at risk of an event after being followed for time t. The hazard at time t represents the instantaneous probability that the patient will have an event before any subsequent time t þ d, d > 0. Mathematically, the hazard function, h(t), is the derivative of log{S(t)}, so the hazard function for a patient at risk at time t is defined as hðtÞ ¼ lim

d/0

probability of having an event before time t þ d . d (26.9)

The hazard can be interpreted as an instantaneous event rate. A somewhat more intuitive, simpler interpretation is to remember the following: higher hazardsd worse survival; lower hazardsdbetter survival. The proportional hazard has the representation:   hðtÞ ¼ h0 ðtÞexp b1 x1 þ . þ bp xp ; (26.10)

II. STUDY DESIGN AND BIOSTATISTICS

380

26. AN INTRODUCTION TO SURVIVAL ANALYSIS

where h0(t) is called the baseline hazard; bk, k ¼ 1,., p are regression coefficients; and xk, k ¼ 1,., p are prognostic factors. If there is no prognostic factor present in the hazard function, then h(t) is the same as the baseline hazard. A proportional hazard means that the change in a prognostic factor results in a proportional change of the hazard on a log scale. To demonstrate this, consider only one prognostic factor, x with the associated coefficient b. The log of the hazard at time t for x ¼ a is log h0 (t) þ b  a, and when x ¼ b is log h0 (t) þ b  b. The difference of the hazards on the log scale is (a e b)  b, which does not change with t. The difference of the hazards is proportional to the change in x, in this case a e b, and the “proportion” is equal to the regression coefficient b. In this type of modeling, a function of the regression coefficient b, exp(b), is referred to as the relative risk. Commonly, x represents the treatment indicator and is given a value of 1 if a patient is assigned to the investigational treatment, and x is given a value of 0 if the patient is assigned to the control treatment. Then b is used to measure the magnitude of the treatment difference because exp(b) ¼ h0 (t) exp(b  1)/h0 (t) exp(b  0) represents the hazard ratio in the two treatment groups. If b ¼ 0, then the hazard ratio is 1, and thus the two groups have the same survival experience. The regression coefficients are usually unknown and need to be estimated from the data. It is complex to estimate these regression coefficients. However, statistical software to estimate the hazard function is commonly available. When the sample size is large, the estimate of each regression coefficient approximately follows a normal distribution. In addition, we can test the effect of prognostic factors. Let b and SE(b) denote the estimate of b and standard error. The Z statistic Z¼

b b   SE b b

COMMON MISTAKES In this section we briefly discuss three common mistakes people make when first using survival analysis in hopes of helping the reader avoid them. The introductory books mentioned previously have more information on each of these items.1e4 In this chapter we never mention mean survival. We cannot sensibly look at mean survival time when some survival times are censored. The same is true when looking at the proportion who survived a given amount of time, such as 1 year, when not all participants were followed for 1 year. We might be able to read the median survival derived from the KaplaneMeier if the sample’s curve drops below 0.5. Additionally, when comparing two survival curves, caution is in order. The log-rank test will use all of the data and is appropriate in many cases. Only using visual inspection of curves is not appropriate. What may appear to be a large difference between two curves at later time points may be an illusion. If confidence intervals are drawn, it will become clearer; due to typically smaller amounts of data on the right-hand side of the curve, the confidence intervals will widen. As discussed in Chapter 24, a large difference in point estimates does not automatically mean a statistically significant difference. Lastly, we want to reemphasize that survival analysis comparison groups need to be defined on factors known before treatment. Investigators may wish to compare responders to nonresponders. Such analyses, creating comparison groups based on factors known only after the treatment has begun, may give misleading results. It is strongly recommended that expert advice be obtained before pursuing such analyses.

(26.11)

CONCLUSION

has approximately the standard normal distribution under the null hypothesis H0:b ¼ 0dthat is, where the prognostic factor has no effect. For consistent unbiased estimates the Cox model requires independent censoring, either directly or given covariates in the Cox model. Cox models can handle data sets with right censoring, staggered entry, and more difficult tasks that K-M cannot such as recurrent events, competing risks, and much more as long as we have available representative risk sets at each time point to allow us to model and estimate event rates.

Survival analysis makes inference about event rates as a function of time. The two primary methods to estimate the true underlying survival curve are the KaplaneMeier estimator and Cox proportional hazards regression. The KaplaneMeier estimator is simple and supports stratification factors but cannot accommodate covariates. The Cox model does provide a framework for making inferences about covariates and some versions require proportional hazards, although all versions are quite flexible when used and interpreted correctly. Independent censoring, either directly in the KaplaneMeier estimator or given covariates in the

II. STUDY DESIGN AND BIOSTATISTICS

REFERENCES

Cox model, is a requirement for consistent unbiased estimates. Survival analysis can handle right censoring, staggered entry, recurrent events, competing risks, and much more as long as we have available representative risk sets at each time point to allow us to model and estimate event rates. Statistical methods for survival analysis remain an active area of research and collaboration among statisticians and their colleagues, and the reader will benefit from joining in this process. Additional topics in survival analysis are discussed in Chapter 27.

QUESTIONS 1. What information do we need to define the survival outcome variable? a. The event b. The time origin c. The time scale d. The time at which an event occurs e. All of the above 2. You are conducting a randomized trial in patients with liver cirrhosis comparing Drug A to placebo where the outcome is survival. You want to know how this drug performs in the liver cirrhosis population. Some patients in this population have ascites, a condition that results in a worse prognosis and lower survival compared to liver cirrhosis patients without ascites. At randomization your trial collects information on prothrombin time and the presence or absence of ascites. Physicians are initially reluctant to enter patients with ascites into the trial because of potential toxicity concerns in this subpopulation. However, after 1½ years, physicians are no longer reluctant to enter patients with ascites into the trial and recruitment begins to mimic the clinic population. What type of analysis can be used to look at survival in this case? [Remember, the KaplaneMeier is group (e.g., placebo group) specific] a. KaplaneMeier, Cox regression any type, it does not matter which type of survival analysis is used

381

b. Simple KaplaneMeier. Easy to understand and in this case the censoring is independent c. Use a Cox model. You expect the KaplaneMeier estimator will be wrong

Acknowledgments Dr. Joanna H. Shih of the Biometric Research Branch, Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, was the original author of this chapter and coauthor of the chapter for editions 2 and 3 of this book. Her work is the basis for this chapter and almost all examples are derivatives of her work. Dr. Craig Borkowf from the CDC also provided extensive edits. This chapter should not be construed to represent the US Government, FDA, NIH, or CDC’s official positions, views, or policies.

Disclaimer This chapter reflects the views of the author and should not be construed to represent FDA’s views or policies.

References 1. Friedman LM, Furberg CD, DeMets DL, Reboussin DM, Grander CB. Fundamentals of clinical trials. 5th ed. New York: Springer; 2015. 2. Altman DG. Practical statistics for medical research. New York: Chapman & Hall; 1991. 3. Hosmer DW, Lemeshow S. Applied survival analysis: regression modeling of time to event data. 2nd ed. New York: Wiley; 2008. 4. Lee ET. Statistical methods for survival data analysis. New York: Wiley-Interscience; 1992. 5. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457e81. 6. Borkowf CB. A simple hybrid variance estimator for the Kaplane Meier survival function. Stat Med 2005;24:827e51. 7. Pocock SJ. Clinical trials: a practical approach. New York: Wiley; 1983. 8. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: 719e48. 9. Kirk AP, Jain S, Pocock S, Thomas HC, Sherlock S. Late results of the Royal Free Hospital prospective controlled trial of prednisolone therapy in hepatitis B surface antigen negative chronic active hepatitis. Gut 1980;21:78e83. 10. Cox DR. Regression models and life tables. J R Stat Soc Ser B 1972; 34:187e220.

II. STUDY DESIGN AND BIOSTATISTICS

C H A P T E R

27 Intermediate Topics in Biostatistics 1

Pamela A. Shaw1, Laura Lee Johnson2, Michael A. Proschan3

University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, United States; 2U.S. Food and Drug Administration, Silver Spring, MD, United States; 3National Institutes of Health, Bethesda, MD, United States

O U T L I N E Special Topics in Trial Design Interim Monitoring and Alpha Spending Introduction Efficacy Boundaries Futility Summary Adaptive Designs Superiority, Noninferiority, and Equivalence Special Considerations for Sample Size Considerations for Early Phase Studies Unequal Sample Sizes

384 384 384 384 386 387 387 389 390 391 391

Special Considerations in Data Analysis A Trick for Confidence Interval Estimation When No Events Occur Data Dependencies Correlation Relationships in Organization, Space, and Time Essential Issues in Microarrays, Functional MRI, and Other Applications With Massive Data Sets

392

Regression to the Mean Introduction What Is Regression to the Mean? Examples Example 1 Change After Exceeding a Threshold Example 2 Placebo Effect Example 3 Screening Period Versus Trial Event Rates Ways to Address Regression to the Mean Summary

394 394 394 395 395 395

Diagnostic Testing Measures of Accuracy Considerations for Study Design

396 396 398

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00027-7

392 392 392 393 393

395

396 396

Common Mistakes and Biases Summary

399 400

Special Considerations in Survival Analysis Changes Over Time in Coefficients and Covariates Time-Varying Coefficients or Time-Dependent Hazard Ratios Time-Dependent Covariates Dependent or Informative Censoring Changes in Inclusion/Exclusion Criteria and Nonindependent Censoring Competing Risks Left and Interval Censoring Recurrent Events Analysis Sample Size

400 400 400 401

401 401 402 402

402 402

Missing Data Introduction Types of Missing Data Methods for Handling Missing Data Common Mistakes Summary

403 403 403 404 404 405

Causal Inference in Observational Studies

405

Concluding Remarks

406

Summary Questions

406

Acknowledgments

407

Disclaimers

407

References

407

383

Copyright © 2018. Published by Elsevier Inc.

384

27. INTERMEDIATE TOPICS IN BIOSTATISTICS

SPECIAL TOPICS IN TRIAL DESIGN In this chapter, we consider a few special topics in trial design. These topics include interim monitoring for efficacy and futility, adaptive designs, noninferiority and equivalence designs, and special considerations in sample size determination.

Interim Monitoring and Alpha Spending Introduction There are many reasons an investigator or sponsor may wish to look at interim data before the prespecified completion time of a trial. We focus on two such reasonsd(1) convincing early evidence that a treatment is beneficial, and (2) the realization that continuation of the trial is futile. Monitoring interim data allows the opportunity to stop the trial early when indicated. Early stopping can not only save time and resources but also speed up the timeline for getting an efficacious drug approved and available to those who need it. As discussed in Chapter 10, the Data and Safety Monitoring Board (DSMB) monitors many aspects of an ongoing trial, and the ability to monitor outcome data in a double-blind trial is one of the main motivations for having a separate independent monitoring board. Efficacy monitoring needs to follow a formal analytical plan. This monitoring generally takes the form of a prespecified boundary for efficacy such that if this boundary is exceeded, significant early evidence for efficacy is declared. Futility monitoring, on the other hand, is usually less formal and less binding. We give a very brief review of efficacy boundaries, and then we present useful tools for futility monitoring. Further statistical aspects can be found in Jennison and Turnbull1 or Proschan et al.2 Efficacy Boundaries Some clinical trials in the 1970s monitored for efficacy several times and declared treatment beneficial if the p-value ever dropped below a. To see why this approach is problematic, think about how your assessment of a dart thrower’s skill would change if you found out that the bull’s eye he bragged about hitting occurred only after 10 prior attempts. Whether it is hitting the bull’s eye or reaching a small p-value, these events will always happen eventually by chance alone if we are allowed enough tries. The adage “If you torture data long enough, they will eventually confess” applies. Another way of expressing this is that the type I error ratedthe probability of falsely declaring the treatment beneficialdis inflated when we use level a to declare significance for each of several tests.3 For instance, even with only one analysis at the halfway point and a final analysis, the chance of falsely declaring benefit at

least once with level 0.05 tests is about 8% instead of 5%. With more interim analyses the error rate can be much higher. To avoid inflating the type I error rate, we must require stronger evidence in the form of smaller p-values at each time point to ensure the desired overall error rate, a. This is what monitoring boundaries are designed to do. The HaybittleePeto method4 is based on the Bonferroni inequality stating that the probability of making at least one error among several tests can never be higher than the sum of the error probabilities for each test (because summing individual error probabilities counts multiple errors multiple times). We encountered the Bonferroni inequality when we noted above that the error rate from two tests at level 0.05 is 0.08 and not 0.05 þ 0.05 ¼ 0.10. The HaybittleePeto method, slightly modified, requires an interim p-value to be 0.001 or less to be declared significant, and then adjusts the final p-value cut point by subtraction. For example, with three interim analyses, the interim error rate is at most 0.001 þ 0.001 þ 0.001 ¼ 0.003, so we are allowed to use 0.05  0.003 ¼ 0.047 at the final analysis. More generally, if there are k  1 analyses before the final one, and none of these reaches the P < .001 requirement for early termination, the treatment is declared beneficial if the p-value at the final analysis is less than 0.05  (k  1) * (0.001). The HaybittleePeto method has four desirable properties: (1) substantial evidence is required to stop very early in the trial (good because early in a trial, uncertainty is high); (2) the level of evidence required at the final analysis is close to what it would have been with no monitoring; (3) the looks need not be equally spaced; and (4) it is always conservative with no further assumptions needed about the distribution of the test statistic over time. Despite being one of the earliest methods, it remains popular today. For example, it was used in a recent trial of immediate versus deferred antiretroviral therapy (ART) for HIV-infected infants.5 Its only real drawback is that there is a precipitous drop-off in the level of evidence required from interim analyses to the final analysis. This can lead to a logical inconsistency whereby the interim evidence is deemed insufficient to declare benefit, then the efficacy trend reverses (i.e., the placebo arm does better than the treatment arm for the remaining data), yet the efficacy boundary at the end of the trial is still crossed. How could the evidence be convincing after a reversal of trends, but not before? The HaybittleePeto boundary is able to control the error probability with no further assumptions because it is overly conservative in the sense that the error probability is always less than, rather than equal to, a. To improve it by making the type I error rate equal to a requires knowledge of the distribution of the test

II. STUDY DESIGN AND BIOSTATISTICS

SPECIAL TOPICS IN TRIAL DESIGN

statistic over time. One might speculate that this distribution would be different for different test statistics. Remarkably, it is about the same for the most popular standardized test statistics used in clinical trials, including (but not limited to) the t-statistic for continuous outcomes, the z-statistic comparing proportions for binary outcomes, the logrank z-statistic for survival outcomes, and variants of these statistics that adjust for baseline covariates (see Chapter 2 of Proschan et al.2). Therefore, the monitoring boundaries discussed below are valid for this broad class of test statistics (while the HaybittleePeto boundary is valid for any test statistic). Pocock’s (1977) monitoring method uses the same level of evidence (in terms of p-value) at each of several equally spaced analyses.6 For example, with a twotailed a ¼ 0.05 test, the p-value needed to declare significance at any of five equally spaced looks is 0.0158. The bad news is that the final analysis also requires this stringent level of evidence. This is a big penalty to pay for monitoring and a sharp contrast with the Haybittlee Peto method, whose p-value cutoff at the last of five analyses is 0.05  4(0.001) ¼ 0.046 (Table 27.1). This drawback of the Pocock method led Pocock later to recommend against his own procedure. Still, the Pocock boundary is sometimes used for safety monitoring because a safety signal can often be seen early, and we do not want extremely high boundaries to stop for harm. Much more popular than Pocock’s boundary for efficacy monitoring is that of O’Brien and Fleming.7 Like Pocock’s boundary, it requires the analyses to be equally spaced. Unlike the Pocock boundary, it requires quite stringent evidence to stop very early, but the evidence required at the final analysis is close to the level required with no monitoring (Table 27.1). The AIDS Clinical Trials Group Protocol 076 trial, which examined the efficacy of Zidouvudine to prevent mother to child transmission of HIV,8 used the O’Briene Fleming boundary. This trial had two planned interim looks for efficacy, after each 1/3 of the projected infant infections. At the time of the first interim analysis, a large beneficial effect of treatment was observed. After

TABLE 27.1 P-value Required to Declare Significance at Each of Five Equally Spaced Analyses Analysis

HaybittleePeto

O’BrieneFleming

Pocock

1

0.001

0.000

0.016

2

0.001

0.001

0.016

3

0.001

0.008

0.016

4

0.001

0.023

0.016

5

0.046

0.041

0.016

385

careful review of the data, including review of data quality and completeness, toxicity, and HIV transmission rates, the DSMB recommended stopping the trial. Trial leaders and sponsor NIH agreed and the trial was stopped, leading to an earlier adoption of this successful treatment in clinical practice than if no monitoring had been in place. The Pocock and O’BrieneFleming boundaries require prespecification of the number of looks, and these looks must be equally spaced. DSMBs prefer to have more flexibility. Unpredictable schedules of DSMB members preclude having equally spaced analyses. Moreover, the board may want to change the number of looks after the trial has started. Lan and DeMets (1983) proposed a method that readily accommodates changes in number and timing of looks.9 It is based on the information fraction t, a measure of the fraction of the trial that has been completed thus far; t ¼ 0 and t ¼ 1 at the beginning and end, respectively. For trials with a continuous or binary outcome, t is the ratio of the number of patients evaluated so far to the number who will be evaluated by trial’s end. Thus, if 50 of the trial’s planned total of 200 patients have had their primary outcome measured by the time of an interim analysis, then t ¼ 50/200 ¼ 0.25. For survival trials, t is the number of patients with an event so far divided by the number of patients who will have an event by trial’s end. Thus, if a mortality trial plans on 100 deaths by trial’s end and 40 have occurred thus far, t ¼ 40/100 ¼ 0.40. Before the trial begins, one specifies a spending function a*(t) dictating the cumulative error rate used by information fraction t, with a*(0) ¼ 0 and a*(1) ¼ a. Whenever an interim analysis arises, one simply estimates the current information fraction t and uses a computer program to determine the boundary that makes the cumulative error rate a*(t). Boundary properties depend on the spending function selected. The most popular ones spend very little alpha early, and then spend rapidly near the end of the trial. One such spending function generates boundaries very similar to those of O’Brien and Fleming if the analyses are equally spaced in terms of information. A different spending function, producing boundaries similar to Pocock’s, spends more alpha early than does the O’BrieneFleming-like spending function (Fig. 27.1). It should be avoided for efficacy monitoring but may be appropriate for safety monitoring. It is important for the investigator to interact with the statistician to choose the most appropriate boundary. Strictly speaking, the spending function approach does not allow changing look times on the basis of data trends (e.g., looking more frequently because one is close to reaching a boundary). Nonetheless, different authors have shown that the popular spending functions maintain good properties even if the look times are changed because of data trends.10,11

II. STUDY DESIGN AND BIOSTATISTICS

386

27. INTERMEDIATE TOPICS IN BIOSTATISTICS

0.05

Cumulative Alpha

0.04

0.03 Pocock-Like Spending Function

0.02

0.01

O-F-Like Spending Function

0.0 0.0

0.2

0.4

0.6

0.8

1.0

Information Fraction

FIGURE 27.1 Spending functions specify the cumulative type I error rate to spend by information fraction t. The most popular ones, like the bottom curve yielding boundaries similar to O’Brien and Fleming,7 spend very little early and then rise steeply at the end. The upper curve spends too aggressively early in the trial and yields boundaries similar to Pocock.23

Because of their flexibility, spending functions have been used in many important trials, including the Cardiac Arrhythmia Suppression Trial,12 a well-known example of interim monitoring that had enabled harmful treatment arms to be identified and terminated early. Futility It may become apparent at an interim analysis that continuation is futile because there is little hope of showing that the treatment is beneficial. One way this could happen is that the assumptions underlying the original sample size/power calculation are wrong. For instance, the event rate might be much lower than anticipated in a binary outcome trial, or the variance might be much higher than expected in a continuous outcome trial, or recruitment may be much worse than expected. Under any of these scenarios, power could suffer to such an extent that the original question cannot be reliably answered. If we had known this ahead of time, we never would have started the trial. A useful tool under these circumstances is a revised power calculation; we redo the original power calculation but use revised estimates of things such as the event rate, variance, or realistic sample size (if recruitment is lagging). The trial data are used only to update estimates of parameters in the power calculation. If revised power is low, then a null result will not mean much because it will not rule out the originally hypothesized treatment effect. This suggests that continuation may be futile. On the other hand, a better than expected treatment effect could attenuate this diminution of power. Therefore, it is attractive to take into account the observed

treatment effect and compute conditional power (CP)dthe conditional probability of obtaining a significant result at the end, given the data observed thus far. If CP is low (say, under 20%), then we are unlikely to obtain a statistically significant result at the end of the trial. CP is often computed under a number of different assumptions about the treatment effect, including the originally hypothesized effect (which ought to be the primary consideration) and the effect observed thus far in the trial. Revised power and CP are very different. CP uses trial data not only to revise estimates of event rates, variances, etc., but also to compute the interim treatment effect. Therefore, only a statistician with access to data by arm can compute CP. Revised power and CP also tell us different things. Revised power tells us whether a null result will be useful; if revised power is high, we can still reliably answer the original question because a null result will still rule out the originally hypothesized treatment effect. CP, on the other hand, tells us whether we are likely to get a significant result. Often CP is the main futility driver, but when a treatment is in widespread use prior to your trial, you may want to continue, if ethically feasible, in the face of low CP if revised power is high. For instance, the combination of glucosamine/chondroitin was widely used for arthritis before there was any randomized trial evidence supporting it. If your trial of glucosamine/chondroitin had low CP but high revised power at an interim analysis, you might want to continue to the end to prove that this treatment does not work. This would be ethically justifiable in a relatively short duration trial whose end point is not dire.

II. STUDY DESIGN AND BIOSTATISTICS

SPECIAL TOPICS IN TRIAL DESIGN

While all boundaries are advisory, DSMBs often regard futility boundaries as less binding than efficacy boundaries. Nonetheless, for both efficacy and futility monitoring, boundaries are only one of many factors considered by the DSMB (see Chapter 10). Summary The most popular methods of monitoring are the ones that are the most flexible. Despite being one of the earliest proposals, HaybittleePeto is popular because of its flexibility, simplicity, applicability to any test statistic, and conservatism. The O’Briene Fleming boundary is also popular. It requires even stronger evidence than HaybittleePeto to stop very early, and the level of evidence required at the end is close to what would have been required with no monitoring. The drawback of O’BrieneFleming is that one must specify the number of analyses, and they must be equally spaced in terms of information. Spending functions eliminate these requirements. The O’Briene Fleming-like spending function generates boundaries very similar to those of O’Brien and Fleming if the looks are equally spaced yet can be used even if they are not. The flexibility of the spending function approach makes it the preferred monitoring method in clinical trials. It is important to discuss carefully the monitoring plan with the study statistician and the DSMB before the study begins. Futility monitoring is also important because it conserves resources in cases when there is little hope of demonstrating that the treatment works. Two useful tools, revised power and CP, have complementary roles. CP tells us whether a null result is likely to occur, while revised power tells whether a null result would still be meaningful; i.e., whether it would effectively rule out the hypothesized treatment benefit.

Adaptive Designs Sample size and power calculations in a clinical trial depend on things such as the event rate, variability of a continuous outcome, the dropout rate, etc. We use our best available information to estimate these, but sometimes that information is wrong. No matter how relevant we think our pretrial data are, they cannot be as relevant as our trial data. The trial data might show that the event rate is lower than we thought or the variability is higher than we thought. If we had known this at the beginning, we might have made the trial larger. Adaptive methods offer the opportunity to change the design after seeing data from our trial. We focus on methods specifically designed to control the type I error rate, which is important for regulatory purposes. Nonetheless, we mention a different class of adaptive procedures not specifically designed to control

387

this error rate at the end of this section. Our intent in this section is to give a very brief overview of adaptive methods to help you differentiate potentially useful ones from those that could potentially jeopardize a trial’s credibility. Potential adaptations include changing: sample size, population, number of arms, allocation probabilities to different arms, even the primary outcome. Some of these may frighten you, and that is good! Any adaptation will receive close scrutiny. If it scares you, it will likely scare the scientific community, including reviewers of your outcome paper. Other adaptations may seem benign. Questions to ask in assessing whether an adaptation is benign or potentially malignant are 1. Was the change drastic? 2. Was the decision to change something made after breaking the treatment blind? 3. Was the change unplanned? 4. Was the change made based on the observed treatment effect? Answering yes to any of these questions should raise your level of concern. With regard to question 1, we note that modifications after a trial starts are actually commonplace, but the changes are usually relatively minor. For instance, in response to lagging enrollment, we might relax entry criteria slightly or increase the recruitment period. If we find problems in measuring the primary end point, we might institute additional training of personnel or increase the number of measurement days we average over to diminish variability. Other changes are much more dramatic like dropping an arm, terminating the trial, or changing the primary end point. Some adaptive methods allow such drastic changes because they are based on combining independent p-values; their validity depends solely on the assumption that the p-values are independent and uniformly distributed under the null hypothesis. For instance, Fisher’s method of combining independent p-values p1 and p2 refers 2 ln(p1p2) to the critical value of a chi-squared distribution with 4 degrees of freedom. Bauer and Kohne (1994) recognized that this method could be used with a two-stage procedure whereby one uses the first stage data to determine whether the original assumptions are valid and to possibly change the design, and the p-value from the second stage would be based on the new design.13 For instance, one might decide, after seeing data from 50 of the originally planned 100 participants, that the sample size should be increased to 150. One would then make the second stage sample size 100 instead of 50 and compute the second stage p-value from those 100. Even though the original design was changed, 2 ln(p1p2) still has a chi-squared distribution with 4 degrees of freedom

II. STUDY DESIGN AND BIOSTATISTICS

388

27. INTERMEDIATE TOPICS IN BIOSTATISTICS

under the null hypothesis. In fact, the same would be true if one had decided to change the primary end point after the first stage, and therein lies the rub. The procedure has the correct type I error probability, but there is no guarantee that it will be interpretable or acceptable to the scientific community. It may be that neither the p-value using the original end point nor the p-value using the new end point is statistically significant, yet the curious combination of p-values reaches statistical significance. How do we interpret the result? Do we say the treatment improved outcome? Another potential problem with adaptive methods is that we may be only fooling ourselves when we notice trends that cause us to change the design. After all, the changes we make are based on only a subset of data. Bauer and Kohne’s groundbreaking work opened a Pandora’s Box highlighting the trade-off between flexibility and the potential for criticism on scientific grounds. Nonetheless, such methods can be quite useful when one restricts attention to reasonable adaptations. Question 2 is relevant because adaptations made before breaking the blind are considered by many, including regulatory agencies, as more pristine. One example is modification of sample size in a binary outcome after observing that the overall event rate (i.e., combined across arms) is lower than anticipated. The basic idea is that from the overall proportion of patients with events and the prespecified treatment effect (e.g., a 25% relative reduction), we can estimate the event probabilities in the separate arms, from which we can determine an appropriate sample size.14 Because we peeked only at the overall event proportion, we glean virtually no information about the treatment effect. Similarly, in a trial with a continuous outcome, we can estimate the variance by computing a “lumped” variance of all participants (lumped across arms) and incorporating information about the postulated treatment effect.15 These methods contrast sharply with that of the preceding paragraph, where we were allowed to peek at the data by arm and see the observed treatment effect. Even more extreme changes are sometimes made on the basis of blinded data. One trial of lung disease found that the originally planned primary outcome variable could not be measured reliably, so they changed it. The change was made on the basis of examining blinded lung scans, which convey no information about the observed treatment effect under the null hypothesis. Therefore, even though the answer to question 1 was yes, the change was considered acceptable because it was made before breaking the treatment blind. One must be very careful to ensure that the information used to make a change really does not contain any information that would break the blind. For instance, it would not be valid to examine in a blinded fashion the level of study drug in the blood of trial

participants; such information would clearly unveil the treatment assignments. The third question concerning whether there was a plan for possible adaptation in place at the beginning of the trial is often considered crucial. It is essential to have a prespecified plan for adaptations that could occur after unblinding, such as sample size changes after looking at the observed treatment effect. There is no valid way to do this without prespecifying an accepted method that controls the type I error rate. Nonetheless, unplanned things sometimes happen, as evidenced by the lung trial example. Such unplanned changes can still be acceptable if they were made before breaking the blind. Question 4 concerns whether the method is based on the observed treatment difference. The Bauer and Kohne procedure is one example, but there are now many others.16e18 Such methods are subject to the greatest scrutiny and criticism, so they should be used only in the rare situation when very little pretrial data are available to determine the likely treatment effect. Some have argued that the treatment effect we specify in a clinical trial should be the minimum clinically relevant effect and therefore should not be subject to reestimation. Nonetheless, it may not be obvious what the minimally clinically relevant effect is. Even if we can specify it, it may be much smaller than the effect we are likely to see, which means that we may choose a sample size that is much larger than necessary. Others have argued that we can always use monitoring boundaries to stop early in that case, but the prespecification of a huge sample size upfront may cause the sponsor to opt out. On the other hand, if the sponsor began a trial whose originally planned sample size was smaller, but there was a promising trend, they might be willing to increase the sample size. In summary, adaptive methods based on looking at the observed treatment difference (other than the usual group-sequential monitoring to stop early) are the least often needed and are subject to the greatest criticism. They should be used only when there is very little information about the likely treatment effect. There is a netherworld lying somewhere between adaptations made before breaking the blind and those that overtly use the observed treatment effect to make decisions. Sometimes we need better information about nuisance parameters such as event rates (but not treatment differences) or variances. We have seen that one can use blinded data to estimate these, but the danger is that blinded data will include any treatment effect. This can lead us to the wrong conclusion if the actual treatment effect differs from what was anticipated. For example, in a binary outcome trial, the overall proportion of patients with an event may be lower than expected, causing us to consider a sample size increase. But one possible reason for a smaller than

II. STUDY DESIGN AND BIOSTATISTICS

SPECIAL TOPICS IN TRIAL DESIGN

expected overall event rate might be that the treatment is much better than expected. Similarly, in a trial with a continuous outcome, the overall variance might be large because the treatment effect is greater than expected. These are relatively infrequent occurrences; nonetheless, we may want to avoid them by estimating nuisance parameters using by-arm data. By and large, this causes very few problems. One subtle issue is that a statistical sophisticate with access to summary data by arm and overall can actually reconstruct the observed treatment effect. Adaptive methods have been criticized by some for being (1) inefficient, (2) uninterpretable, or (3) unsound. Adaptations that are not well thought out could suffer from any of these problems, just as an exercise program that is not well thought out could cause health risks, but that does not mean that exercise in general is bad. The fact that there could be problems underscores the need to carefully evaluate the properties of an adaptive procedure before the trial begins. Our focus has been on methods that control the type I error rate at level a, but there is another class, called Bayesian adaptive methods, which are not specifically designed to control this error rate.19 Bayesians formally quantify their prior information about the treatment effect in terms of a prior distribution. They then update it by considering the conditional distribution of the treatment effect given the observed data from the trial. This conditional distribution, called the posterior distribution, is used to make inferences about the treatment effect. A major advantage of Bayesian methods is that they incorporate prior information in a natural way, but one disadvantage is that the conclusion depends on the selected prior. This is closely related to the fact that these methods do not necessarily control the type I error rate; e.g., if our prior distribution reflects confidence that the treatment works, then even a small observed effect may lead us to declare the treatment beneficial. Bayesians recognize this potential and often demonstrate that the chosen prior still leads to a reasonable type I error rate. The Ebola Virus Disease Medical Countermeasures Trial in Liberia, Sierra Leone, Guinea, and the United States is an example of a Bayesian design with a noninformative prior distribution intended to have little influence on the ultimate conclusion (see Dodd et al.112 or Proschan et al.113 for more details on the design and statistical aspects). Patients with Ebola virus disease were randomized to optimized standard of care (oSOC) or oSOC plus ZMapp, a triple monoclonal antibody mixture having very promising preclinical results. The primary outcome was 28-day mortality. A uniform distribution was used in both arms for the prior distribution of the probability of mortality by 28 days. This conservative prior assumes not only that ZMapp

389

has no benefit, but that harm is just as likely as benefit. The posterior probability that the oSOC þ ZMapp arm is superior to oSOC was computed given the observed data. A value of 0.99 or greater or 0.975 or greater would constitute compelling evidence at an interim or final analysis, respectively. This procedure is similar to, but less conservative than, the HaybittleePeto monitoring boundary applied to Fisher’s exact test. A credible interval of likely values for the difference in proportions also was computed. The trial was stopped early because the epidemic ended before the anticipated 100 patients per arm could be recruited. See Partnership for Research in Ebola Virus in Liberia II (PREVAIL II) Writing Team 2016 for results.114 In summary, flexibility is very desirable in clinical trials, but there is always a price to be paid. For example, monitoring boundaries not only allow us to stop early if the interim data warrant it but they also decrease power; it is always more efficient and powerful to conduct a single test at the end of the trial. Adaptive methods offer even more flexibility. We can not only stop early but also we can change many design features. The more flexibility we have, the greater the potential loss of efficiency and the greater the potential for problems of interpretation. That is why one should carefully consider these factors in the design phase of the trial by evaluating power and other properties of adaptive procedures. In some cases, the adaptation was completely unplanned. Such changes may still be acceptable if the decision to change was made on the basis of blinded data. If it was made after seeing data by arm and therefore could have been a function of the observed treatment effect, there is no valid way to protect the type I error rate.

Superiority, Noninferiority, and Equivalence The goal of most parallel, randomized controlled trials is to show that a new treatment is better than a placebo or active control treatment. These trials are referred to as superiority trials. There are some settings where, instead, investigators are interested in showing that a new treatment is roughly the same as an existing treatment. A trial with this aim is referred to as an equivalence trial. Furthermore, there are other settings where it may be enough that a trial demonstrates a new treatment is nearly as good as an existing treatment. For instance, an existing drug could be prohibitively expensive, have associated toxicities or have a certain amount of inconvenience, say due to frequent or IV dosing. If a new drug has added benefits, such as less cost, less toxicity or a simpler dosing regimen, then it could still be preferred even if there was some loss in efficacy relative to a comparison treatment. In this setting, a trial would aim to demonstrate that any loss in

II. STUDY DESIGN AND BIOSTATISTICS

390

27. INTERMEDIATE TOPICS IN BIOSTATISTICS

efficacy associated with the new treatment was within some tolerable margin (D) of the standard treatment. This type of trial is referred to as a noninferiority trial. The tolerable difference in efficacy (D) is known as the equivalence or noninferiority margin. Superiority, noninferiority, and equivalence trials have three very different goals in terms of the question they are designed to answer. We learned from Chapter 24 that a trial’s question of interest has an associated hypothesis test and that this test, together with specified accuracy in terms of the type I and type II error, will inform the sample size calculation (see Chapter 25). The testing and sample size methods presented in Chapters 24 and 25 are generally for the setting of superiority trials, but similar methodology can be applied to noninferiority and equivalence trials. In equivalence and noninferiority trials, as compared to superiority trials, the null and alternative hypotheses are essentially switched. The null hypothesis for an equivalence trial is that the treatments under the study have different efficacy, which is the magnitude of the efficacy difference jm0  m1j is greater than some acceptable margin (D), and the alternative hypothesis is that the difference is smaller than D. In the formal hypothesis testing framework of Chapter 24, one has H0 : jm0  m1 j  D vs. HA : jm0  m1 j< D

(27.1)

The noninferiority trial can be seen as a “one-sided” version of this. That is, the null hypothesis is that the efficacy for the control treatment (m0) exceeds that for the novel treatment by more than an acceptable amount; the alternative is that the deficit in efficacy for the novel treatment is acceptably small. Formally, one has H0 : m0  m1  D vs. HA : m0  m1 < D

(27.2)

Note that the alternative hypothesis includes the possibility that the novel treatment has greater efficacy than the current treatment; however, in the above noninferior alternative, this possibility is not distinguished from the novel treatment being slightly worse than the active control. If a treatment is declared noninferior, then superiority could be subsequently tested using a separate, prespecified test.20 To demonstrate noninferiority typically, once a study is completed, a two-sided confidence interval for the estimated difference in outcome between two treatments will be constructed. If the entire interval is on the desired side of the noninferiority margin, then noninferiority can be declared. Similar methods are described for equivalence studies. Special consideration has to be given to the bounds that are used to determine if one can declare equivalence or noninferiority of a novel treatment to an active control. How does one define noninferiority? A noninferiority study wants to demonstrate that the investigational

treatment is not worse than the comparison treatment by more than the acceptable margin, usually a small amount. Will the size of the margin depend on the frequency and severity of side effects for the comparison treatment? Sometimes. This margin has to be prespecified and tailored for each study and justified clinically and statistically. It is important to keep in mind the size of the efficacy benefit the established comparison treatment has over placebo, as seen in previous trials. If the noninferiority margin is too large, by comparison to the efficacy advantage the active control has over placebo, a treatment that is declared noninferior to the active control could be close to being equivalent to a placebo, or worse. This possibility becomes an even more serious risk if the active control was itself proven effective in a noninferiority trial. For an equivalence trial, the goal is to show the current treatment is within some acceptably small margin of an established treatment. This type of trial will tend to be larger, and thus more expensive, than having to establish superiority over a placebo. Noninferiority and equivalence trials have a number of issues that complicate their design, analysis, and interpretation relative to superiority trials. For example, the evidence for efficacy of a treatment declared noninferior or equivalent becomes problematic if the reference product is itself later called into question. Despite the reasonable and common scientific goal of establishing that a treatment is good enough or better than a standard treatment, inferiority, and equivalence trials are seldom done in most settings.20 This may be due to associated challenges with these designs. The reader is encouraged to read more about noninferiority and equivalence trials and their associated advantages and challenges, with the reference section below providing just a few examples from the vast literature on this topic.20e24 Specific guidelines for reporting the results of equivalence and noninferiority trials also have been developed which extend the Consolidated Standards of Reporting Trials standards for reporting results from randomized controlled clinical trials.20,25 At the time of this writing the Food and Drug Administration has released a guidance document for the conduct of noninferiority clinical trials in the United States.26 Investigators are best off consulting current regulatory agency guidance documents for details as the acceptability of methods is changing rapidly.

Special Considerations for Sample Size Here we consider a few special topics in the area of sample size. These include considerations for planning the sample size in small, early phase studies and tricks for determining sample size for unbalanced designs.

II. STUDY DESIGN AND BIOSTATISTICS

SPECIAL TOPICS IN TRIAL DESIGN

More details can be found elsewhere; this section is to provide quick reference for a few common issues. Considerations for Early Phase Studies In the sample size and power calculations presented previously in Chapter 25, we assumed particular values for the detectable differences of interest and the variances pertinent to the test statistic to be estimated on a trial’s conclusion. Often these values are chosen based on previous experience, but sometimes there exists much uncertainty about appropriate choices for these values. In this case, while planning the study, one should construct a table showing the sample size or power values for a wide variety of possible assumptions about parameters, such as the differences of interest and the true variances. If possible, one should aim to have a sufficiently large sample size to preserve good power across a wide variety of possible parameter values, rather than merely have satisfactory power for a narrow set of assumptions. What if the study is really preliminary? Many textbooks focus on designing larger studies, while researchers are frequently interested in designing smaller studies, such as a dosing study, an early phase study, or other early pilot work. What if you want an initial estimate of the variance? Many groups have developed study designs for phase I and II studies.27e30 Because space is limited, we will not summarize them here, but we will summarize several of the common discussions that arise. First and foremost, before a study begins make sure the data the study plans to collect will provide enough information to move the science forward. This means asking, will this study provide us with data to help us better estimate the sample size or dose for future studies? Will this study help us decide if there should even be another study? Someone needs to get an answer from the study we are conducting, and that answer may not be what is desired and it may only be an incremental step, but it should be an answer of some sort. If not, then it is questionable why the work is being done at all. In many early studies we want to find the maximally tolerated dose or find at which dose a certain threshold of toxicity is hit. Defining toxicity and when to stop is fairly difficult. Definitions need to be applicable across study participants and reasonable in the real world. When to stop is dependent on the people and diseases under investigation, in addition to the other treatment options available. Some interventions will not cause toxicity and other dose stopping rules may also apply.30,31 While classically many early studies used a 3 þ 3 type of study design, many new methods can be used to establish one or several doses to move forward for further study, and several of these study designs include 6e15 people per dose, randomization to

391

multiple arms, multiple control arms, or some blinding.32 A few key elements remain, though. When a first in human study is conducted, entry of patients to a new dose level does not occur until all patients in the previous level are beyond a certain timeframe where we are looking for or expecting toxicity. Investigators could be wrong about the timeframe, however, and this is why long-term follow-up is key even in early studies. Sometimes we do randomize to multiple doses in early studies, though, if investigators and regulators feel it is safe. Sample size estimates are available for early studies, but typically they are unlike what is found in Chapter 25. In these cases, a methodology that relies on exact inference, rather than methods that rely on large sample properties (e.g., normal distribution), often is applied. Screening of new therapies will at times look at biomarkers and surrogate outcomes but the focus is usually on safety and toxicity, not to prove “final” efficacy. Biomarkers and surrogate outcomes are typically able to provide preliminary evidence of efficacy, more quickly and with smaller sample sizes than a study aiming to establish a benefit for a true outcome of interest, such as a mortality benefit. Several designs are based on finding sufficient activity to decide to move research forward and test the intervention in a larger randomized study. Thus, early studies usually include small numbers of patients, may be in the hundreds total, but may be less than 100/arm and even as small as 6e15/arm. Early studies may be small but also they should be treated with as much care as later phase studies. The designs, sample sizes, and implementation often are dictated by safety and cost considerations. Regardless, we need to keep in mind the trade-offs of having an “underpowered study.” Due to the variability of results in a small study, an increased risk exists of a spurious result that could lead investigators to move a useless treatment forward or drop a truly promising treatment. Too many times decisions are made relying on one small study done at a single site instead of multiple small studies done at multiple sites in the early stages of development of a novel treatment. Dose finding studies are notoriously small, despite the fact that dose is likely one of the most important decisions to make in designing studies for any treatment in the early phase of the development.33 Some have proposed using a formal framework of decision analysis, incorporating costs to these two different types of errors, into the sample size calculation.27,28,34 Unequal Sample Sizes In general, it is easiest to estimate sample size for balanced designs in which the number of individuals assigned to each treatment group is the same.

II. STUDY DESIGN AND BIOSTATISTICS

392

27. INTERMEDIATE TOPICS IN BIOSTATISTICS

Unfortunately, for practical reasons, we cannot always obtain the sample size we desire for one of the groups of interest, especially in observational studies. For example, in a caseecontrol study there may be a fixed number of cases or devices that are available, and yet the standard sample size calculation may indicate that more observations are required. If we want to obtain a specific value of power with a fixed number of cases, we may use the following formula to determine the ratio of controls to cases, namely, n k¼ ; (27.3) 2n1  n where n is the number of subjects in each group required for the given power under the balanced design, n1 is the fixed number of cases, and, in turn, kn1 will be the number of controls.35 For example, if we originally calculated a balanced study would need n ¼ 13 subjects per arm, but then decided we would only have access to n1 ¼ 11 cases, then k ¼ 1.44, and thus we need kn1 ¼ 16 controls and n1 ¼ 11 cases to achieve the same power as the balanced design with n ¼ 13 controls and n ¼ 13 cases. Sometimes the ratio of cases to controls is in question. In general there is little reason to use more than 4 or 5 controls per case because the gain in precision or power by adding only controls eventually becomes negligible as the ratio of controls to cases increases. Even in randomized studies, people may want to use an unequal allocation, perhaps randomizing two people to the investigational agent for every one person randomized to the control intervention. The section “Calculations for Continuous Data with Unequal Variances or Unequal Sample Sizes” in Chapter 25 has formulas and an example determining the sample size when a study wants k patients randomized to the new investigational treatment arm for every patient randomized to the placebo arm. However, a fast trick uses the total sample size alone. This trick assumes the total sample size for the study increases by a factor off ðkþ 1Þ2 4k

(27.4)

For example, if the total sample size for two equal groups is 26 and we want to randomize using a 2:1 ratio, then by Eq. (27.4) one can calculate a new total sample size 2

26  ð2 þ 1Þ 26  9 ¼ 29:25z30; ¼ 8 42

(27.5)

indicating that we need 20 people in one group and 10 in the other. Of course, a small sample size like this might have other issues; this is only an illustration. Will the arm with 10 people yield enough information for the study to make decisions about what steps should come next after the study is complete? In general sample size decisions need to balance what is minimally

necessary, what practically can be implemented, what information is needed to make decisions about planning future studies, and what is needed to support the planned data analyses. A good reference for general and practical rules of thumb is van Belle.35

SPECIAL CONSIDERATIONS IN DATA ANALYSIS Many different issues can come up in data analysis. In this section, we discuss first a common problem in small trials, how to put a confidence interval on an event rate when no events were observed. Then we move to data dependencies, first describing correlation and repeated measures then transitioning to the discussion of microarrays, functional MRI (fMRI), and other massive data sets where the data are rarely independent and simple statistical analyses may lead to incorrect inference. Issues related to multiplicity and hypothesis testing also are discussed.

A Trick for Confidence Interval Estimation When No Events Occur One trick related to estimating the upper confidence interval limit is particularly useful in experiments with a binomial outcome. The “rule of three” states that the 95% upper bound for a binomial proportion when no events occur is approximated by 3/n.35,36 The improved “rule of three”2 is 3/(n þ 1) and is a uniformly better approximation than 3/n.36 For example, if we conduct 25 rodent experiments and have no fatal outcomes, then the 95% upper confidence bound on the true rate of mortality is approximately equal to 3/25% or 12% using the older rule. The lower bound is trivially 0 since the lowest possible proportion was observed. Using the improved rule of three, the upper bound may be better approximated by 3/26% or 11.5%. The exact calculation for the upper bound in this case is 11.3%, slightly less than the two quick approximations.

Data Dependencies Correlation Correlation coefficients are measures of agreement between paired variables (xi, yi), where there is one independent pair of observations for each subject. The general formula for the sample (Pearson) correlation is n P

r¼

II. STUDY DESIGN AND BIOSTATISTICS

n P i¼1

  ðxi  xÞ yi  y

i¼1

ðxi  xÞ2

n  P i¼1

yi  y

2

(27.6)

SPECIAL CONSIDERATIONS IN DATA ANALYSIS

The sample correlation r lies between the values 1 and 1, which correspond to perfect negative and positive linear relationships, respectively. A value of r ¼ 0 corresponds to no linear relationship, but other nonlinear associations may exist. Also, the statistic r2 describes the proportion of variation about the mean in one variable that is explained by the second variable. One may compute p-values for the hypothesis of zero correlation, although the magnitude of the change in one variable induced by a change, say by one unit, in the other is often more important than the fact that the correlation is nonzero. In addition, if one replaces the raw data for each variable by the respective ranks of that data, one obtains Spearman’s rank correlation. Spearman’s rank correlation is a nonparametric statistic, as its value is unchanged by transformations of the data, such as taking the logarithm, which preserve the ranks. For further details, see Altman.37 Relationships in Organization, Space, and Time It is important to recognize the various structures and relationships that may exist among the data in organization, space, and time. In some studies, there may be hierarchical relationships among subjects. For example, in a community-wide observational study on the health of school-aged children, we may have children nested within classrooms, within schools, within school districts, etc. Similarly, in a study with a geographical or spatial component, measurements on locations that are closer together may tend to be more similar, and those nearer large cities may tend to have different traits than more rural locations. Furthermore, in a longitudinal study in which repeated measurements are made on each subject, the measurements made closer together in time may be more highly related to each other than those made at more distant times. It is important to recognize these various structures and relationships because they need to be considered appropriately in the statistical design of studies and the analysis of data. Typically this structure is captured by the choice of a statistical model or by choosing an analytical method that can otherwise account for the correlation between related measurements in the data. Failure to properly account for the correlation will generally lead to incorrect inference, such as incorrect estimates of standard errors and p-values. For paired data, accounting for correlation could be as simple as choosing to do a paired test, such as the paired t-tests or the signed-rank Wilcoxon test. For longitudinal data, there are a variety of useful methods, including generalize estimating equations and mixed models for more general repeated measures data.38 It is important to work with a statistician to capture appropriately the structure of the data so that valid analysis and proper inference can be performed.

393

Essential Issues in Microarrays, Functional MRI, and Other Applications With Massive Data Sets Structural relationships in data sets are important for all studies, but they are especially important for the massive data sets we find in microarray and fMRI experiments. The basic rules and assumptions described previously pertain to these data sets and slight variations from those assumptions may result in large changes in inference, such as the difference between the t and z distributions at a given cut point for a certain sample size. When trying to draw inference to a population we need to look at numerous independent samples. Looking at 100,000 pieces of information for five people means we know those five people very well but does not mean we can extrapolate that knowledge to the entire human race. The structure inherent in a data set may make many of the observations correlated or otherwise related. Furthermore, multiple comparisons issues abound; data mining, hunting for patterns and useful information in massive data sets are common; and many other issues arise not only in analysis but also in computing a reasonable sample size estimate for a study. The simple hypothesis testing methods discussed in previous chapters relied on having independent samples and measurements. The only exception we saw was for the paired t-test, but there we looked at independent differences within each pair of subjects. If we take multiple measures over time in the same person, as in a longitudinal study, these measures are not independent, and indeed measurements taken closer in time may be more similar than those taken at distant times. We also may take a single sample (e.g., biopsy) from a person but use it to report multiple outcomes; this is commonly done in microarray and fMRI experiments. If we use a microarray, the gene expression we see from some probe sets may be associated with the gene expression seen in other probe sets because the probe sets are either for the same gene or for genes that are associated with each other. Indeed, since many microarray chips routinely test for 10,000 þ genes at a time, and some have multiple tests for the same gene on a chip, the importance of the correlation structures and multiple comparisons cannot be underestimated. Although correlation is occasionally discussed in the analysis of microarray data, it is a hot topic in fMRI data analyses. The voxels, defined registered areas on the fMRI image, have a correlation structure. Currently, some analysis methods ignore it, some methods impose a simple uniform structure that is not modified for different parts of the brain, and other methods attempt to fully model the correlation structure of the voxels. Although the later methods may be the most accurate,

II. STUDY DESIGN AND BIOSTATISTICS

394

27. INTERMEDIATE TOPICS IN BIOSTATISTICS

they also require computing capabilities and impractical sample sizes not commonly found at the point of this writing. Another common problem that arises in microarray or other high-throughput experiments with large numbers of tests is the choice of error rates for sample size calculation. This calculation requires the consideration of the impact and consequences of multiple testing, a topic discussed in Chapter 24. Many researchers choose to use a significance level of 0.001 and power of 0.95 or higher to control the false discovery rate and select a candidate pool on which to follow up.39 Especially in the high-throughput designs, a Bonferroni correction may make it impossible to have any statistically significant items in a study with a small sample size. Several methods that aim to limit the false discovery rate may be employed, and new methods are frequently described in journals.40e42 It also is important to remember that technical replicates help reduce laboratory variability, but independent biological replicates (i.e., samples from many different people) are important. The sample size calculations presented previously may be used with special care given to the choice of significance level and power, but in choosing which samples to collect and use, it is important to remember that the sample size estimates are for the number of independent samples. Statistical genetics, including microarray and genome-wide association studies, and fMRI are a few of several areas with massive data sets for each individual sample that can hide the fact that often there are few independent samples.39 Data mining is common, but it too must take into account the correlation, multiple comparisons, and many other issues described previously. Likewise, we can perform sample size calculations in studies designed to check for clustering of genes, single-nucleotide polymorphisms (SNPs), etc. Currently, there are no hard and fast rules about the methods, except for the consensus that most methods are attempted with too few independent samples to uncover potentially complex biological structure. In the end, how many independent samples do we need to make a reasonable inference about a population? Consult your local statistician who specializes in the type of data in your study for details, new updates, and guidance. These are not just computing issues; in fact, with this much data for every specimen plenty of numbers can come roaring out of a computer, but we need to ensure they are the answers to the scientific questions of interest and that the experiment and analysis can be replicated.

REGRESSION TO THE MEAN Introduction In Part II of this book, Biostatistics and Epidemiology, many topic sections include a subsection on common mistakes. The entire section that follows could carry this title, as regression to the mean (RTM) is a very slippery topic with the best practitioners falling into traps set by this statistical phenomenon. In this section, we explain the principle of RTM, present some classic examples where RTM can be misleading, and discuss approaches for analysis and study design that avoid some of the common mistakes.

What Is Regression to the Mean? RTM is the tendency of observations to become less extreme on repeat measurement. In clinical research, the trap is set when at baseline one observes a measurement that is extreme compared to some health standard, and then after applying an intervention, takes a followup measure to see whether the treatment helped. Take, for example, blood pressure. Because of RTM, we expect blood pressure values that were higher than some cutoff at baseline to be lower on second measurement even without an efficacious intervention. This is because blood pressure is inherently a noisy measurement, with variability in both the instrument measuring blood pressure and true within-person (biological) variability. Individuals with a higher than average blood pressure are a mix of those whose underlying blood pressure is truly high and those who by chance had a more extreme measurement than usual. If there were no variability, there would be no RTM and the blood pressure measure would be identical when repeated. The more variation in an outcome, the bigger the effect of RTM could be. The further away from the mean that a measurement is, the more one would expect it to drop (even assuming no intervention) when the measurement is repeated. See Bland and Altman43 or Davis44 for further mathematical details. The name RTM comes from the well-known 1886 paper by Francis Galton, “Regression toward Mediocrity in Hereditary Stature.”45 In this paper Galton examined the association of children’s heights with their parents. He remarked that for taller than average parents, the height of their children was closer to the overall mean height of children than the height of their parents and he referred to this as “regression to mediocrity.” The term regression, in fact, gets its name from this landmark paper.43

II. STUDY DESIGN AND BIOSTATISTICS

REGRESSION TO THE MEAN

Regression theory tells us that for two measurements with correlation r, if the first measurement x is sx standard deviations away from its mean, then the second measurement y is on average rsy away from its mean. If the two measures are not perfectly correlated, i.e., r is less than one in magnitude, then y is expected to be closer to its mean (in standard deviation units) than x to its mean.

Examples Here we cover a few classic instances where RTM can confound or otherwise adversely affect study results. Example 1 Change After Exceeding a Threshold A recent trial, comparing the efficacy of multiple active ART regimens on disease progression for individuals with advanced HIV infection, provides an example of how extreme RTM can be. In this trial, which we will refer to as Trial A, it was observed that individuals observed with renal dysfunction at baseline, defined by an estimated glomerular filtration rate (eGFR) less than 60, had a dramatic increase in their eGFR at their 3-month measurement. Investigators wanted to conclude that patients improved because of interventions given to the patient. Luckily, there were statisticians around to prevent this faulty conclusion. Though investigators were aware of RTM, the size of the change observed for patients with baseline renal dysfunction was thought to be too clinically significant to be a statistical artifact and not a true change in patient status. To convince investigators that RTM, and not the intervention, was playing a large role in the increase in eGFR, study statisticians ran an experiment backward in time. Instead of using the baseline value to classify patients as having renal dysfunction, they used the 2-month value. After identifying participants whose value 2 months postbaseline was less than 60, they looked backward in time at the 1-month values for those same participants. In the absence of RTM and the presence of a treatment benefit, we would expect their mean values to be somewhat smaller at 1 month because they had only 1 month of treatment instead of two. But instead the opposite was true; the means were 46.2 at month 2 and 61.6 at month 1. This is clear evidence that the observed increases are being caused by regression to the mean. Furthermore, the people with eGFR 25%) reduction in the mortality of patients with myocardial infarction.107 However, the large Fourth International Study of Infarct Survival (ISIS-4) trial found no significant effect on mortality with either treatment.108 A meta-analysis of the heart failure treatment nesiritide found excess mortality and renal dysfunction, but a definitive large trial found neither excess mortality nor renal dysfunction.109 Although many causes have been posited for these discrepancies, a definitive explanation does not exist. The major message seems to be that large numbers of patients are needed to be certain of the effects of a therapy.

UNDERSTANDING COVARIATES AND SUBGROUPS Because of the insatiable curiosity of clinicians and patients about whether different responses to treatment may be seen in different types of patients, an analysis of trial results as a function of baseline characteristics is

inevitable. Traditionally, such questions have been addressed using a subgroup analysis, in which the treatment effect is estimated as a function of baseline characteristics taken one at a time (e.g., age, sex, weight). This approach has been called a “false-positive result machine” but might just as well be referred to as a “false-negative result machine.” False positives are generated because of the problem of multiple comparisons; by chance alone, a significant difference will be apparent in at least 1 in 20 subgroups even if there is absolutely no treatment effect. In 1980, Lee and colleagues96 randomly split a population of 1073 into two hypothetical treatment groups (the treatments were actually identical) and found a difference in survival in a subgroup of patients with a p-value less than 0.05. At the same time, given the large number of patients needed to demonstrate an important treatment effect, dividing the population into subgroups markedly reduces the power to detect real differences. Consider a treatment that reduces mortality by 15% in a population equally divided between men and women, with a p-value for the treatment effect of 0.03. If the treatment

II. STUDY DESIGN AND BIOSTATISTICS

432

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

effect is identical for men and women, the approximate p-value will be 0.06 within each subgroup because each group is half as large. It would obviously be foolish to conclude that the treatment was effective in the overall population but not in men or women. A more appropriate and conservative method would be to develop a statistical model to predict outcome with regard to the primary endpoint for the trial, and then to evaluate the effect of the treatment as an effect of each covariate after adjusting for the effects of the general prognostic model. This type of analysis, known as a treatment by covariate interaction analysis, assumes that the treatment effect is homogeneous in subgroups examined unless a definitive difference is observed. Recent research on this topic is producing effective methods of assessing heterogeneity of treatment effect.18,19 An example of this approach occurred in the Prospective Randomized Amlodipine Survival Evaluation (PRAISE) trial,110 which observed a reduction in mortality with amlodipine in patients with idiopathic dilated cardiomyopathy but not in patients with ischemic cardiomyopathy. This case was particularly interesting because this subgroup was prespecified to the extent that the randomization was stratified. However, the reason for the stratification was that the trial designers expected that amlodipine would be ineffective in patients without cardiovascular disease; in fact, the opposite occurred. Acting responsibly, the trial organization mounted a confirmatory second trial. In the completed follow-up trial (PRAISE-2), the special benefit in the idiopathic dilated cardiomyopathy group was not replicated. In the BARI trial,100 post hoc analysis showed a significant benefit of bypass surgery in patients with treated diabetes mellitus but not in other patients. This analysis had not been specified before the trial started enrollment nor had the randomization been stratified. However, the data and safety monitoring committee had asked for an analysis of this issue based on concerns raised in an acute revascularization trial. Subsequent trials have confirmed the signal that surgery seems especially indicated in patients with diabetes and coronary disease.111

THERAPEUTIC TRUISMS A review of recent clinical trials points out that many commonly held beliefs about clinical practice need to be challenged. If these assumptions are shown to be less solid than previously believed, a substantial change in the pace of clinical investigation will be needed. Medical trainees have often been taught that variations in practice patterns are inconsequential. The common observation that different practitioners treat the same problem in different ways has been tolerated

because of the general belief that these differences do not matter. Clinical trials have demonstrated, however, that small changes in practice patterns for common diseases can have a sizable impact.112 An example is the extreme variation in recommendations regarding the preferred Activated Partial Thromboplastin Time (aPTT) for patients treated with unfractionated heparin anticoagulation therapy. Based on the pathophysiologic surrogate of arterial patency,113,114 GUSTO investigators adjusted the recommended aPTT upward in the transition from GUSTO-I to GUSTO-IIa. The average 8-s increase in aPTT resulted in a doubling of the rate of intracranial hemorrhage in patients treated with thrombolytic therapy and heparin.67 When the heparin dose was reduced in GUSTO-IIb, the intracranial hemorrhage rate reproduced that observed in GUSTO-I.70 Another ingrained belief of medical training is that observation of the patient will provide definitive evidence for changing treatment. Although no one would dispute the importance of following symptoms, many acute therapies have effects that cannot be judged in a short time, and many therapies for chronic illness prevent adverse outcomes in patients with very few symptoms. For example, in treating acute congestive heart failure, inotropic agents improve cardiac output early after initiation of therapy but lead to a higher risk of death. Beta-blockers cause symptomatic deterioration acutely but appear to improve long-term outcome. Mibefradil was effective in reducing angina and improving exercise tolerance, but it caused sudden death in an alarming proportion of patients, leading to its removal from the market. Similarly, the standard method of determining the dose of a drug has been to measure physiologic endpointsda technique that in a sense represents a surrogate endpoint approach. No field has more impressively demonstrated the uncertainties of this method in multiple circumstances than has heart failure therapy. Various vasodilator and inotropic therapies improve hemodynamics in the acute phase but have been shown to increase subsequent mortality. Experience with heparin and warfarin has taught us that large numbers of patients are required to understand the relationship between the dose of a drug and clinical outcome. On the other hand, in genetic diseases with a culprit gene or protein that can be reliably modified, the measurement of a biomarker can provide excellent prediction of therapeutic benefit. In fact, the accelerated approval pathway for serious and life-threatening diseases is founded on the concept that biomarkers that are “reasonably likely” to predict clinical benefit can be used to allow marketing of therapies under development. The evidentiary basis and skill needed to discern which biomarkers fit into this category and which are

II. STUDY DESIGN AND BIOSTATISTICS

OPERATIONAL ORGANIZATION FOR LARGE-SCALE CLINICAL RESEARCH

more likely to join the many failed candidate biomarkers is an area in need of intense exploration. The BEST compendium, a joint effort by the NIH, the FDA, and the research community (described above), offers a publicly available compilation of what is known about biomarkers and other tools for medical product development.74 Finally, the maxim “Do no harm” has been a fundamental tenet of medical practice. However, most biologically potent therapies cause harm in some patients while helping others. Emphasis on the neurologic complications of bypass surgery provides ample demonstration that a therapy that saves lives can also lead to complications in particular patients.115 Intracranial hemorrhage resulting from thrombolytic therapy exemplifies a therapy that is beneficial for populations but has devastating effects on some individuals. As noted earlier, beta blockade causes early deterioration in many patients with heart failure, but the longerterm survival benefits are documented in multiple clinical trials. Patients who are harmed can be detected easily, but those patients whose lives are saved cannot be detected.

OPERATIONAL ORGANIZATION FOR LARGE-SCALE CLINICAL RESEARCH Whether the investigator or sponsor is contemplating a large or a small trial, the general principles of study organization should be the same (Fig. 28.6). A balance of interest and power must be created to ensure that after the trial is designed, the experiment can be performed without bias and interpretation will be generalizable. A comprehensive public knowledge

FIGURE 28.6

433

base on these issues is being compiled at the NIH Collaboratory’s “Living Textbook” for pragmatic clinical researchdthe Rethinking Clinical Trials website (http://www.rethinkingclinicaltrials.org).

Executive Functions The Steering Committee In a large trial, the steering committee is a critical component of the study organization. This group designs, executes, and disseminates the study. A diverse steering committee, providing multiple points of view that encompass biology, biostatistics, and clinical medicine, is more likely to organize a trial that will withstand external scrutiny. This same principle holds for small trials; an individual investigator, by organizing a committee of peers, can avoid egocentric thinking about a clinical trial. Increasingly, patients who either represent trial participants (or are themselves study volunteers) are invited to participate in steering committee functions. While this is not yet the norm, it will likely become so over the next decade. The principal investigator plays a key role in the function of the trial as a whole, and a healthy interaction between the principal investigator and the steering committee can provide a stimulating exchange of ideas on how best to conduct a trial. The principal trial statistician is crucial in final decisions regarding study design and data analysis. An executive committee can be useful, providing a small group to make real-time critical decisions for the trial organization as the trial is implemented. This committee typically should include the sponsor, the principal investigator, the statistician, and key representatives from the steering committee and the data coordinating center (DCC).

Principles of study organization.

II. STUDY DESIGN AND BIOSTATISTICS

434

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

The Data and Safety Monitoring Committee The data and safety monitoring committee (DSMC) (also called a data and safety monitoring board) is constructed to oversee the safety of the trial from the point of view of the patients being enrolled (see Chapter 10). The DSMC should include clinical experts, biostatisticians, and sometimes medical ethicists; these individuals should have no financial interest, emotional attachment, or other investment in the therapies being studied. Committee members have access to otherwise confidential data during the course of the trial, allowing decisions to be made on the basis of information that, if made available to investigators, could compromise their objectivity. The DSMC carries an increasingly scrutinized ethical obligation to review the management of the trial in the broadest sense, in conjunction with each institutional review board, to ensure that patients are treated according to ethical principles. The role of the DSMC has become a topic of significant global interest. Little has been published about the function of these groups, yet they hold considerable power over the functioning of clinical trials. The NIH116 and the FDA117 have published guidelines for DSMCs, and a textbook is also available.118 The Institutional Review Board The institutional review board (IRB) continues to play a critical role in the conduct of all types of clinical research (see Chapter 4). Approval by the IRB generally is required for any type of research, even if the research is not funded by an external source. The IRB should comprise physicians and scientists with expertise in clinical trials, representatives with expertise in medical ethics, and representatives of society in the community in which the research is being conducted. As with the DSMC, the IRB function has come under scrutiny, especially by government agencies charged with ensuring the protection of human subjects. Several types of studies are typically exempted from the IRB process, including studies of public behavior, research on educational practices, and studies of existing data in which the research data cannot be linked to individual subjects, although following recent changes to the Common Rule, limited IRB review of such studies may be required to ensure privacy safeguards.119 Surveys and interviews may be exempted when subjects are not identified and the data have a very low likelihood of leading to a lawsuit, financial loss, or reduced employability of the subject. For multicenter trials, increasing consensus indicates that the practice of multiple reviews of the protocol and consent form at the local level is detrimental to public health, and that use of central IRBs and reciprocal

arrangements is far more sensible. The Office for Human Research Protections endorsed this view based on evidence that most consent form changes were not helpful and that redundant reviews led to huge amounts of work that undermine the ability of the trial to be organized on a common basis,120 and it has now been incorporated into the Common Rule.119 Furthermore, the NIH has now instituted central IRBs as a national policy for NIH-funded research,121 and recent legislation extended central IRB function to trials evaluating medical devices for the first time.122 Another critical consideration is the need for IRBs, especially those in academic medical centers, to focus on small investigator-initiated studies for which there is no other ethical oversight. Generally, it is hoped that as institutions move toward broader human research and participants’ protection programs, the IRB will be able to focus more on the ethics of the actual conduct of studies rather than on the paperwork. Regulatory Authorities Government regulatory authorities have played a major role in the conduct of clinical research. Requirements by the FDA and other national health authorities provide the rules by which industry-sponsored clinical trials are conducted. In general, regulatory requirements include interpretation of fundamental guidelines to ensure adherence to human rights and ethical standards. The FDA and equivalent international authorities are charged with ensuring that drugs and devices that are marketed are safe and effective (a charge with broad leeway for interpretation). It is important to note that in the United States, there is no mandate to assess comparative effectiveness or cost effectiveness. Although efforts to harmonize regulations have been made through the International Conference on Harmonisation, which has many documents describing best practices, the globalization of trials123 raises additional questions about how to reach a more efficient consensus. The Good Clinical Practice (GCP) guidances provide a starting point for a consensus on appropriate clinical trial conduct across national boundaries. However, they also seem to have encouraged a tendency toward literal interpretation. This in turn may promote bureaucracy that raises costs substantially rather than encouraging best clinical research efforts to answer critical questions for patients and the public.124,125 The Clinical Trials Transformation Initiative (CTTI)126 represents a publiceprivate partnership intended to improve the conduct of clinical trials, both in the United States and globally. The FDA and many other entities representing industry, academia, contract research organizations (CROs), law firms, and patient advocacy groups belong

II. STUDY DESIGN AND BIOSTATISTICS

OPERATIONAL ORGANIZATION FOR LARGE-SCALE CLINICAL RESEARCH

to this organization, whose Quality by Design approach127 articulates principles for implementing quality measures appropriate to the particular type of trial. GCP is currently undergoing an update to take these issues into account. A similar effort called the Medical Device Innovation Consortium (MDIC; http://mdic.org/) is developing approaches applicable to trials and observational studies for medical devices. Industry or Government Sponsors Having provided funding for the study, the sponsor of a clinical trial understandably prefers to be heavily involved in its conduct. Worldwide, most clinical investigation is now funded directly by the pharmaceutical or medical device industry, with many coordinating and oversight functions performed by for-profit CROs. This approach seems reasonable and desirable for explanatory trials, but pragmatic trials, if not overseen by an independent steering committee, run a greater risk of bias because the sponsor of such a study has a large financial stake in the success of the therapy being tested. Even in the case of government sponsorship, trials are frequently performed as a result of political agendas, with much to be gained or lost for individuals within the scientific community depending on the result. All of these issues speak to the importance of a diverse steering committee with a balance of conflicts capable of managing the general functioning of a large pragmatic clinical trial.

435

difference between a successful trial and a failure. In large trials, a small change in protocol or the addition of one more visit or testing procedure can add huge amounts to the total costs. The larger the trial, the greater the economy of scale in materials, supplies, and organization. For example, a simple protocol amendment can take months and can cost hundreds of thousands of dollars (and even more in terms of the delay) to successfully go through multiple national regulatory authorities and hundreds of IRBs. For these reasons, the selection of intellectual leaders of a trial who are in touch with the practical implications of their decisions is an important activity. Data Coordinating Center The DCC is responsible for coordinating the collection and cleaning of data for the clinical trial. In this role, the DCC must comply with an increasing number of regulations governing both the quality of the data and their confidentiality. Furthermore, the DCC should produce routine reports that allow the trial organization and the DSMC to oversee the conduct of the trial and to ensure that the question the human subjects volunteered to answer is being addressed properly. The DCC must be able to harness data from multiple formats, including standard paper data records, remote data entry, and increasingly information from existing digital data sources, including EHRs, claims data, information from personal devices and social media, and health system data warehouses. Site Management Organization

Coordinating Functions The coordinating functions of large pragmatic trials may be viewed as a whole as in Fig. 28.6. The fundamental functions are intellectual and scientific leadership, site management, and data management. These core functions are supported by various administrative functions, including information technology, finance, human resources, contracts management, pharmacy and supplies distribution, and randomization services. Given the magnitude of large trials, each project depends on the administrative and leadership skills of a project manager and a principal investigator. A major weakness in any one of these functions can lead to failure of the entire effort, whereas excellence in all components creates a fulfilling and exciting experience. Intellectual Leadership The roles of the principal investigator and the chief statistician are critical to the success of the trial organization. Not only must these leaders provide conceptual expertise, but their knowledge of successful approaches to operational concepts in the real world may be the

In large trials, managing the conduct of research sites is a major effort. It requires training and communications programs and regulatory affairs expertise to ensure compliance with federal and nonfederal guidelines. Increasingly, international enrollment is needed either for adequate sample size or to meet regulatory needs in different countries and cultures, and the organization must be able to provide in-service education and study monitoring in multiple languages while complying with regulations from multiple national authorities. Given the imperative to initiate and complete trials efficiently, site management groups are increasingly concerned with maintaining good relations with clinical sites that perform well in clinical trials. These relationships are often fostered by ongoing educational programs aimed at increasing the quality of participation at the sites and rewarding personnel by supplementing their knowledge of conducting and interpreting clinical trials. In addition, metrics are being implemented to measure functions such as recruitment rates, protocol deviations, data quality, and personnel turnover. Sites that perform well are selected for future trials to enhance efficiency.

II. STUDY DESIGN AND BIOSTATISTICS

436

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

Supporting Functions

enterprise under control by attempting to standardize contracts into a common operating model.

Information Technology Large trials are increasingly dependent on a successful information platform. A competitive coordinating center is dependent on first-rate information technology expertise to maintain communication, often on a global basis. This function will be stressed in the coming years, as the field, which has already moved away from paper forms, will now migrate from stand-alone electronic data capture systems to EHRs and disease registries with interoperable data. Managing this transition will require expertise in informatics rather than simple implementation of stand-alone IT systems.128 Finance Even in relatively simple, low-paying trials, an effective financial system is critical to success. Study budgets typically are divided, with approximately half of funds going to the sites performing the study and half going to coordinating efforts, with this money frequently split among multiple contractors and subcontractors. Because payments to the sites typically depend on documented activities at the sites, the flow of cash needs to be carefully regulated to avoid overpayment or underpayment. Furthermore, the coordinating staff needs to be carefully monitored to ensure that study funds are appropriately allocated to get the work done without overspending. See Chapter 33 for further details. Human Resources The staff required to conduct large pragmatic trials comprises a diverse group of employees with different needs. Statisticians, informaticists, and information technology experts are difficult to attract and retain in this very competitive environment. The second most difficult group of employees to find and retain is qualified project leaders. The knowledge base required and the skills needed are extraordinary. Contracts Management For better or worse, specific activities in our global society are increasingly directed by legal contracts. In a typical large clinical trial, a huge number of contracts must be in place, and an entourage of lawyers is busily looking out for the interests of each entity. The sponsor typically will contract with entities to coordinate portions of the trial. The number of coordination contracts depends on whether the primary coordination is done internally within an industry or government sponsor or is contracted out to one or more CROs. Each participating site then has a contract with the sponsor, the coordinating organization, or both. Multiple efforts are underway to bring this chaotic and redundant

Pharmacy and Supplies The production and distribution of study materials, including those required for in-service training, and actual supplies, such as investigational drugs and devices, require considerable expertise. The knowledge required ranges from practical skills such as knowing how to package materials for maximum understanding by the sites, to expertise in “just-in-time” distribution across international boundaries and working knowledge of the mountains of regulations concerning good clinical practice and good manufacturing practice for clinical trials. Randomization Services A fundamental principle of large pragmatic trials is that proper randomization will balance for baseline risk, including both known and unknown risk factors, to allow for an unbiased comparison of treatments. In large multicenter trials, this issue takes on tremendous complexity. Because sealed envelopes are notoriously prone to tampering in large, geographically distributed trials, central randomization has been viewed as superior. This can be accomplished by telephone randomization or by an interactive voice randomization service, which offers the advantage of providing instantaneous access to global networks of investigators and automatic recording of patient characteristics at the time of randomization. Project Management and Regulatory Affairs Within the context of the sponsoring organization with its ongoing priorities, the coordinating entities with their ongoing priorities, and the sites with their ongoing priorities, someone must ensure that the individual project is completed on time and on budget. This responsibility typically is shared by the principal investigator, the project manager, and a sponsor representative (a project officer for government grants and contracts and a scientific or business manager for industry trials). This task ideally should fall to people with skills in organizational management, finance, regulatory affairs, medical affairs, leadership, and personnel management. Individuals with the skills to carry out these responsibilities are difficult to find. It is interesting to note that relatively few educational programs are in place to train these people, despite the huge shortage of qualified individuals. Similarly, when significant human research is done, it is necessary to have a deep understanding of regulations governing research and to ensure that someone is responsible for designing and overseeing the implementation of appropriate procedures. Furthermore,

II. STUDY DESIGN AND BIOSTATISTICS

CONTROVERSIES AND PERSONAL PERSPECTIVE

documentation of quality and the need to undergo inspections and external reviews must be managed.

INTEGRATION INTO PRACTICE Because the ultimate goal of clinical investigation is to improve the care of patients, integrating the findings of a clinical investigation into practice must be undertaken carefully. The old method whereby each practitioner reads the literature and makes individual decisions is inadequate. Recognition of this deficit has led to a variety of efforts to synthesize empirical information into practice guidelines (Fig. 28.7). These guidelines may be considered as different paths to the top of the mountain, with several different routes acceptable as long as the difficulty and likelihood of success are known. In addition, large efforts such as the Cochrane Collaboration129 are attempting to make available systematic overviews of clinical trials in most major therapeutic areas. This effort has been integrated into a cycle of quality construct in which disease registries form the basis for capturing continuous information about the quality of care for populations.130 Within these populations, clinical trials that are of adequate size and are performed in relevant study cohorts can lead to definitive clinical practice guidelines. These guidelines then can form the basis for performance measures that are used to capture the quality of care delivered. Ultimately, gaps in clinical outcomes in this system can be used to define the need for new technologies and behavioral approaches. Increasingly, the linkage of interoperable electronic health records, professional societyedriven quality efforts, and patient/payer-driven interest in improving outcomes is leading to a system in which clinical trials

FIGURE 28.7 The mountain of evidence.

437

are embedded within disease registries so that the total population can be understood and the implementation of findings into practice can be measured.131 Research indicates that only about 15% of major clinical practice guideline recommendations in cardiology are made on the basis of adequately conducted randomized trials, and more than half are based only on expert opinion.132,133 Furthermore, a review of more than 108,000 clinical trials in the ClinicalTrials.gov registry demonstrates that 96% of trials have fewer than 1000 participants and 58% have fewer than 100 participants, so that the vast majority of effort is going into trials that by design could not inform medical decisionmaking with a high degree of confidence.

CONTROVERSIES AND PERSONAL PERSPECTIVE Governmental Regulation Versus Professional Responsibility to Drive the Creation of Evidence During the past several years, as the US health-care system has evolved into a business model, controversy has emerged over the role of government regulation of clinical research. Regulation comes in two forms: (1) requirements that stem from authorities charged with ensuring the safety and efficacy of products sold for medical purposes and (2) ethical regulation of investigation done for the public good. Although the usual business approach is to assume that the marketplace will sort out beneficial therapies from those that do not work, history belies this belief in terms of medical therapies. Determining medical benefit is complex, and observation made in ignorance of the principles of controlled clinical trials is inadequate; without requirements for controlled trials, many detrimental therapies would likely still be in use.134 In areas in which outcome studies have not been mandated, the potential public risk is unknown and may be substantial. The creation of the forerunner to today’s FDA occurred because of the sale of pharmaceutical products that were lethally contaminated, and many similar increments in regulatory oversight and authority have resulted from similar tragedies. The mandate of the FDA extends only to the point of determining that a therapy is safe and effective compared with no therapy, however, and does not include the determination of which therapy is more effective when two effective therapies are available unless there is a specifically identified safety issue that would distinguish one product from another. Additionally, the mandate with regard to pharmaceuticals and biologics is more stringent, with a general requirement for substantial evidence of clinical benefit from randomized controlled trials, whereas device regulations

II. STUDY DESIGN AND BIOSTATISTICS

438

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

allow less definitive evidence (“reasonable assurance”) of clinical benefit. The avalanche of new nutritional supplements and behavioral therapies has been exempted from the scientific demonstration of efficacy. Once a pharmaceutical or device product is on the market, the FDA has some control over the advertising and promotion that can be done with that product, but the FDA can deal only with errors of commission and not with errors of omission, and recent court challenges in the United States have weakened the FDA’s authority. Thus, a therapy that lowers blood pressure and has been shown in adequate controlled trials to lower mortality and stroke rates can advertise that benefit. A comparative therapy that lowers blood pressure but has not been shown to reduce mortality and stroke can be advertised as being effective in lowering blood pressure without mention of the absence of clinical outcomes trials. My belief is that the major onus is on the medical profession to ensure that adequate clinical trials are done to demonstrate which therapies are most effective. The industry that develops new therapies is obligated only to meet the regulations for marketing the product. Managed care organizations and insurers are focused on cost rather than outcome, and in many cases incentives are in place to impede new therapies from clinical use when they increase expense. Only with the active involvement of health-care providers will the necessary studies be done to make the best information available pertaining to choices among active therapies. Given the interests of multiple sectors, coupled with dramatic progress in the scaling of clinical research by individual initiatives such as FDA’s Sentinel System and NEST, the NIH’s HCS Research Collaboratory and Precision Medicine Initiative, and the National PatientCentered Clinical Research Network (PCORnet), a strategy is needed to accelerate the progress toward a comprehensive approach to generating scientific evidence through clinical research (EvGen).135 Such an approach includes clinical trials and observational studies and focuses on large-scale efforts to incorporate research into practice.136 The National Academy of Medicine has developed an extensive rationale for a “learning health system” in which learning would be an essential part of routine clinical practice.137

Composite and Surrogate Endpoints The number of important clinical questions far exceeds our ability to address them. In addition, we are entering an exciting era in which the convergence of advances in genetics, genomics, molecular biology, information technology, and engineering holds out the promise of significant increases in the number of effective therapies. For this reason, there is a great temptation

to perform clinical trials using composite endpoints (a combination of several clinical outcomes in one endpoint) and surrogate endpoints. Composite endpoints have the advantage of creating an outcome scale that allows for categorical or continuous measures that add power to the study. However, less important endpoints can drive the difference when small differences in a more important endpoint such as mortality cannot possibly be detected. Similarly, surrogate endpoints are hampered by constant uncertainty about whether a therapy-induced change in the surrogate will truly predict clinical benefit, particularly because the difference between therapies might relate to an outcome dictated by an unexpected effect of the therapy. For these reasons, surrogates and composites should be regarded as intermediate approaches to determining which therapies need to be definitively tested. Important exceptions occur when a minor modification is made to a drug or device and consensus exists that the modification will not fundamentally alter the effect of the treatment or in the case of a rare, lifethreatening disease with no alternative treatments when a culprit biomarker can be improved with a new therapy. However, even in those cases, given the power of evidence generation from emerging technologies and operational structures, practitioners and investigators need to eschew the temptation to substitute putative surrogates for adequate measures of health outcomes in sufficiently large populations over the long haul. Instead, the clinical community should focus on more efficient methods of aggregating large amounts of information at a lower cost.

Randomized Trials Versus Observational Studies Randomized trials and observational studies represent different approaches to answering questions. Traditionally, the randomized trial has been hampered by enrollment of selected patients who perhaps are not representative of clinical practice and by concern about the ethics of human experimentation. Given the current organization of medicine, the number of therapeutic questions unavoidably will far exceed the number of randomized trials that can be performed. Having been involved in numerous randomized trials and observational treatment comparisons, I believe that much more effort is needed to expand the capacity to perform randomized trials. Until the technological revolution that allowed the rapid accumulation of information from multiple sites and the introduction of the large pragmatic trial method, it seemed that observational studies were needed to fill in the huge gaps in clinical therapeutics.138 With current capabilities, the major role of observational studies should be to generate

II. STUDY DESIGN AND BIOSTATISTICS

439

THE FUTURE

hypotheses, to fill in information about small variations in practice, and to deal with the extrapolations necessary to inform decisions about acute and chronic disease management. Because most patients live for years with chronic disease, questions such as what type of lipidlowering agent to use, how to lower blood sugar in diabetes, how to treat depression in the teenage years, and whether to use antibody-based therapies on a long-term basis for arthritis will have long-term implications requiring extrapolation beyond time frames that can reasonably be measured. The technology half-life is now so short that long-term randomized trials run the risk of becoming historical artifacts. As discussed earlier, much of this will be driven by a well-recognized need for comparative effectiveness studies139,140 but that also will be hotly debated from multiple angles. It is hoped that PCORI, CTTI, MDIC, and the new NIH approach to clinical trials will provide a forum for consensus on these difficult issues. In the modern mode of evidence generation discussed above, observation will essentially provide continuous learning opportunities as data from different systems are pooled to provide measures of outcomes and quality. When an important question arises that require a large sample size because the causal effect size is small, randomization can be used without the need to construct an entirely new entity to conduct the trial.

Sharing of Information Given the for-profit nature of the medical products and pharmaceutical industry and the increasingly financial orientation of the health-care delivery system, there is reason for concern that important medical information is not being made available.141 The financial incentive in the medical products industry has stimulated tremendous creativity and should be preserved. A potential conflict exists, however, between the professional ethic of the health-care provider to share information to improve the delivery of health care in broader terms and the need for financially driven delivery systems to maintain a competitive edge. Ironically, this potential conflict emerges at a time when capabilities for rapidly sharing and interpreting data have improved in an exponential fashion. Patients expect that when medical research is done, the information will be shared; one could argue that information sharing to advance health, in general, is a critical part of the ethical contract between provider and patient when informed consent is obtained. Recent US government legislation mandating registration of trials, results, and adverse events for the vast majority of clinical trials is producing a huge change in

the landscape.142 Not only will results be publicly accessible, but they will be available in a form that can be used to aggregate information across trials.143 This mandate has the potential for making all clinical trials public enough that results will be made available to avoid damaging reflections on the sponsor, even if the results are unfavorable to a particular product or belief. Further sunshine on clinical trials conduct and data is the result of pressure from patient advocacy groups and independent alliances of patients with diseases who are collecting their own data. The Cystic Fibrosis Foundation and the Multiple Myeloma Research Foundation have taken control of much of the research to ensure that it is accessible to their constituents. Not only does this increase the relevance of clinical trials to their constituents’ needs, speed the conduct of trials, and make those trials more efficient and feasible, but it also ensures that the results will be available to potentially affected patients.

THE FUTURE Because of the continuing explosion of knowledge about human biology, the increasing sophistication of patients and clinicians, and the inability of society to increase the financial outlay for the unfettered use of biology and technology and realization that the best choices for individual patients and populations require deep information and analytical decision support, the future of clinical investigation is bright. Without quantitative information, there is no rational method to make decisions about what will be supported in medical practice and what will be eschewed. In coming years, practitioners will make increasing use of EHRs that will generate computerized databases to capture information at the point of care. Early efforts in this area, focusing on procedure reports to meet mandates from payers and quality reviewers, will be replaced by systems aimed at capturing information about the entire course of the patient’s encounter with the health-care system. The proliferation of social media and wearable devices will provide another major source of rich data to be integrated into the system. Although the impetus for this approach will come from those who pay for medical care, practitioners will find the information useful for justifying rational medical care and improving the overall efficiency of care delivery. Multimedia presentations will allow the clinician to view medical records and imaging studies simultaneously in the clinic or in the hospital. To promote efficient exchange of information, the nomenclature of diagnosis, treatments, and outcomes will progressively become standardized. A further significant issue that will need to be addressed is the recognition of uncertainty in medical

II. STUDY DESIGN AND BIOSTATISTICS

440

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

practice. The system will only begin to operate at the needed pace when providers and patients acknowledge uncertainty and the importance of definitively resolving this uncertainty with quantitative information derived through a combination of observational studies and randomized trials. The computerized management of digital information will be part of an inevitable coalescence of practitioners into integrated health systems. To efficiently care for populations of patients at a reasonable cost and to coordinate care for individuals, practitioners will work in large, geographically linked, economically interdependent groups. This integration of health systems will propel outcomes research into a new era in which strategies of care can be tested over time and refined in continuous learning processes for healthcare providers. Although integrated health-care systems will provide the structure for medical practice and research, global communications will provide mechanisms for quickly answering questions about diagnosis, prevention, prognosis, and treatment of common and uncommon diseases. The ability to aggregate information about thousands of patients in multiple health systems will change the critical issues facing clinical researchers. Increasingly, attention will be diverted away from efforts to obtain data and toward the development of efficient means of analyzing and interpreting the types of information that will be available. Within the United States, efforts will continue to focus on improving trial operations, using integrated health systems as the fundamental units within which clinical research takes place. The thousands of independent clinical research sites typical of past trial settings will continue to coalesce into hundreds of integrated health systems, each owning a collection of sites and systems for collecting, storing, and analysis of data. Data standards and common terminology and nomenclature will continue to converge so that data will be interoperable across systems. Large government-funded coordinating centers will be linked with private efforts to create capacity in which systems will contribute in different ways to different research projects using shared approaches to operations and data storage and exchange. Ultimately, leading practitioners will band together in national and global networks oriented toward treating illnesses of common interest. When a specific question requiring randomization is identified, the studies will be much simpler because the randomization can simply be added to the computerized database, and information that currently requires construction of a clinical trials database will be immediately accessible without additional work. Information systems will be designed to provide continuous feedback of information to

clinicians, supporting rational decisions about therapy. In essence, a continuous series of observational studies will be in progress, assessing outcomes as a function of diagnostic processes and therapeutic strategies with the goal of optimizing quality of care. When questions arise about the best technologies, drugs, devices, or practices, randomization will be used.

SUMMARY QUESTIONS 1. When expressing results from a clinical trial, researchers should: a. Describe both absolute and relative risk reductions b. Report the number needed to treat/number needed to harm c. Describe the relative benefits of the therapy, expressed as odds ratios or risk ratios d. All of the above 2. Which of the following is not a truism of clinical trial design? a. Unintended biological effects are relatively uncommon b. Therapeutic interactions are unpredictable c. Quantitative interactions are common d. Qualitative interactions are uncommon 3. Which of the following are possible sources of error in metaanalyses? a. Publication bias b. Lack of comparability among included studies c. The assumption that trials included in the analysis are homogeneous with regard to the estimate of the outcome d. All of the above 4. Which of the following is primarily responsible for protecting the safety and well-being of research participants enrolled in a clinical trial? a. Institutional review board (IRB) b. Trial steering committee (SC) c. Data and safety monitoring committee (DSMC) d. Principal investigator (PI) 5. Requirements for obtaining informed consent from study participants may be relaxed under what circumstances? a. Phase 1 studies b. Nonindustry-funded research c. Research conducted in emergency contexts d. None of the above

References 1. Fisher RA, Mackenzie WA. Studies of crop variation: II. The manurial response of different potato varieties. J Agric Sci 1923; 13:315. 2. Lilienfield AM. Ceteris paribus: the evolution of the clinical trial. Bull Hist Med 1982;56:1e18.

II. STUDY DESIGN AND BIOSTATISTICS

REFERENCES

3. Medical Research Council. Streptomycin treatment of pulmonary tuberculosis. Br Med J 1948;2:769e82. 4. ISIS-1 (First International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous atenolol among 16,027 cases of suspected acute myocardial infarction: ISIS-1. Lancet 1986;2:57e66. 5. Horvath MM, Winfield S, Evans S, Slopek S, Shang H, Ferranti J. The DEDUCE Guided Query tool: providing simplified access to clinical data for research and quality improvement. J Biomed Inform 2011;44:266e76. 6. Kohane IS. Using electronic health records to drive discovery in disease genomics. Nat Rev Genet 2011;12:417e28. 7. National Cardiovascular Data Registry. Available at: http:// cvquality.acc.org/ncdr-home.aspx. 8. Society of Thoracic Surgeons Database. Available at: http://www. sts.org/national-database. 9. US Food and Drug Administration. FDA’s Sentinel Initiatived background. Available at: http://www.fda.gov/Safety/FDAs SentinelInitiative/ucm149340.htm. 10. US Food and Drug Administration. National evaluation system for health technology. Available at: http://www.fda.gov/ aboutfda/centersoffices/officeofmedicalproductsandtobacco/ cdrh/cdrhreports/ucm301912.htm. 11. Roberts R, Rodriguez W, Murphy D, Crescenzi T. Pediatric drug labeling: improving the safety and efficacy of pediatric therapies. JAMA 2003;290:905e11. 12. Alexander KP, Peterson ED. Evidence-based care for all patients. Am J Med 2003;114:333e5. 13. Forrow L, Taylor WC, Arnold RM. Absolutely relative: how research results are summarized can affect treatment decisions. Am J Med 1992;92:121e4. 14. Bobbio M, Demichelis B, Giustetto G. Completeness of reporting trial results: effect on physicians’ willingness to prescribe. Lancet 1994;343:1209e11. 15. Naylor CD, Chen E, Strauss B. Measured enthusiasm: does the method of reporting trial results alter perceptions of therapeutic effectiveness? Ann Intern Med 1992;117:916e21. 16. Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet 1994;343:311e22. 17. Collins R, Reith C, Emberson J, Armitage J, Baigent C, et al. Interpretation of the evidence for the efficacy and safety of statin therapy. Lancet 2016;388:2532e61. 18. Harrell Jr FE, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluation assumption and adequacy, and measuring and reducing errors. Stat Med 1996;15:361e87. 19. Kent DM, Nelson J, Dahbreh IJ, Altman DG, Hayward RA. Risk and treatment effect heterogeneity: re-analysis of individual participant data from 32 large clinical trials. Int J Epidemiol 2016; 45:2075e88. 20. Califf RM, Zarin DA, Kramer JM, Sherman RE, Aberle LH, Tasneem A. Characteristics of clinical trials registered in ClinicalTrials.gov, 2007e2010. JAMA 2012;307:1838e47. 21. Olson CM, Rennie D, Cook D, Dickersin K, Flanagin A, Hogan JW, et al. Publication bias in editorial decision making. JAMA 2002; 287:2825e8. 22. Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. ClinicalTrials.gov results databasedupdate and key issues. N Engl J Med 2011;364:852e60. 23. The GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673e82. 24. Collins FS, Varmus H. A new initiative on precision medicine. N Engl J Med 2015;372:793e5.

441

25. The National Institutes of Health. The precision medicine initiative cohort programdbuilding a research foundation for 21st century medicine. September 17, 2015. Available at: https://www.nih.gov/sites/ default/files/research-training/initiatives/pmi/pmi-workinggroup-report-20150917-2.pdf. 26. Ge D, Fellay J, Thompson AJ, Simon JS, Shianna KV, Urban TJ, et al. Genetic variation in IL28B predicts hepatitis C treatmentinduced viral clearance. Nature 2009;461:399e401. 27. Davidson NE. HER2-targeted therapies: how far we’ve comed and where we’re headed. Oncology (Williston Park) 2011;25:425e6. 28. Solanezumab: too late in mild Alzheimer’s disease? Lancet Neurol 2017;16:97. 29. Siemers ER, Sundell KL, Carlson C, Case M, Sethuraman G, et al. Phase 3 solanezumab trials: Secondary outcomes in mild Alzheimer’s disease patients. Alzheimers Dement 2016;12:110e20. 30. Gross R, Strom BL. Toward improved adverse event/suspected adverse drug reaction reporting. Pharmacoepidemiol Drug Saf 2003;12:89e91. 31. Davidson MH. Rosuvastatin safety: lessons from the FDA review and post-approval surveillance. Expert Opin Drug Saf 2004;3:547e57. 32. Topol EJ. Arthritis medicines and cardiovascular eventsd“House of coxibs”. JAMA 2005;293:366e8. 33. Bennett CL, Silver SM, Djulbegovic B, Samaras AT, Blau CA, Gleason KJ, et al. Venous thromboembolism and mortality associated with recombinant erythropoietin and darbepoetin administration for the treatment of cancer-associated anemia. JAMA 2008;299:914e24. 34. Pfeffer MA, Burdmann EA, Chen CY, Cooper ME, de Zeeuw D, Eckardt KU, et al. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med 2009;361:2019e32. 35. Singh AK, Szczech L, Tang KL, Barnhart H, Sapp S, Wolfson M, Reddan D, CHOIR Investigators. Correction of anemia with epoetin alfa in chronic kidney disease. N Engl J Med 2006;355:2085e98. 36. US Food and Drug Administration website. FDA Drug Safety Communication: Modified dosing recommendations to improve the safe use of Erythropoiesis-Stimulating Agents (ESAs) in chronic kidney disease. Available at: http://www.fda.gov/Drugs/DrugSafety/ ucm259639.htm. 37. US Food and Drug Administration. 22 case studies where phase 2 and phase 3 trials had divergent results. January 2017. Available at: https://www.fda.gov/downloads/AboutFDA/ReportsManuals Forms/Reports/UCM535780.pdf. 38. Lewis JD, Ferrara A, Peng T, Hedderson M, Bilker WB, Quesenberry Jr CP, et al. Risk of bladder cancer among diabetic patients treated with pioglitazone: interim report of a longitudinal cohort study. Diabetes Care 2011;34:916e22. 39. Ferrara A, Lewis JD, Quesenberry Jr CP, Peng T, Strom BL, Van Den Eeden SK, et al. Cohort study of pioglitazone and cancer incidence in patients with diabetes. Diabetes Care 2011; 34:923e9. 40. Piccinni C, Motola D, Marchesini G, Poluzzi E. Assessing the association of pioglitazone use and bladder cancer through drug adverse event reporting. Diabetes Care 2011;34:1369e71. 41. US Food and Drug Administration website. FDA alert (updated 6/15/2011). Actos (pioglitazone): ongoing safety review: potential increased risk of bladder cancer. Available at: http://www.fda. gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationfor PatientsandProviders/ucm109136.htm. 42. US Food and Drug Administration. Guidance for industry. Diabetes mellitusdevaluating cardiovascular risk in new antidiabetic therapies to treat type 2 diabetes. December 2008. Available at: http://www.fda. gov/downloads/drugs/guidancecomplianceregulatoryinformation/ guidances/ucm071627.pdf. 43. Hirshberg B, Katz A. Cardiovascular outcome studies with novel antidiabetes agents: scientific and operational considerations. Diabetes Care 2013;36(Suppl. 2):S253e8.

II. STUDY DESIGN AND BIOSTATISTICS

442

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

44. Smith RJ, Goldfine AB, Hiatt WR. Evaluating the cardiovascular safety of new medications for type 2 diabetes: time to reassess? Diabetes Care 2016;39:738e42. 45. Sherman RE, Li J, Shapley S, Robb M, Woodcock J. Expediting drug developmentdthe FDA’s new “breakthrough therapy designation. N Engl J Med 2013;369:1877e80. 46. Califf RM, Sugarman J. Exploring the ethical and regulatory issues in pragmatic clinical trials. Clin Trials 2015;12:436e41. 47. Tunis SR, Stryer DB, Clancy CM. Practical clinical trials: increasing the value of clinical research for decision making in clinical and health policy. JAMA 2003;290:1624e32. 48. PRECIS-2 website. Available at: https://www.precis-2.org/. 49. NIH Collaboratory Living Textbook of Pragmatic Clinical Trials. About NIH Collaboratory. Available at: http://www.rethinkingclinicaltrials.org/about-nih-collaboratory/. 50. Majumdar SR, Roe MT, Peterson ED, Chen AY, Gibler WB, Armstrong PW. Better outcomes for patients treated at hospitals that participate in clinical trials. Arch Intern Med 2008;168:657e62. 51. I-SPY Trials. About I-SPY trials. Available at: http://www. ispytrials.org/about. 52. Bhatt DL, Mehta C. Adaptive designs for clinical trials. N Engl J Med 2016;375:65e74. 53. Lo B, Fiegal D, Cummins S, Hulley SB. Addressing ethical issues. In: Hulley SB, Cummings SR, editors. Designing clinical research. Baltimore: Williams & Wilkins; 1988. p. 151e7. 54. Peto R, Collins R, Gray R. Large-scale randomized evidence: large, simple trials and overview of trials. J Clin Epidemiol 1995; 48:23e40. 55. Shah BR, Drozda J, Peterson ED. Leveraging observational registries to inform comparative effectiveness research. Am Heart J 2010;160:8e15. 56. Peterson ED, Roe MT, Chen AY, Fonarow GC, Lytle BL, Cannon CP, et al. The NCDR ACTION Registry-GWTG: transforming contemporary acute myocardial infarction clinical care. Heart 2010;96:1798e802. 57. Olsen LA, Aisner D, McGinnis JM, editors. Institute of Medicine Roundtable on Evidence-Based Medicine. The Learning Healthcare System (workshop summary). Washington, DC: National Academies Press; 2007. 58. Whicher DM, Miller JE, Dunham KM, Joffe S. Gatekeepers for pragmatic clinical trials. Clin Trials 2015;12:442e8. 59. US Food and Drug Administration. Guidance for institutional review boards, investigators, and sponsors. Exception from informed consent requirements for emergency research. March 2011. Available at: http://www.fda.gov/downloads/RegulatoryInformation/ Guidances/UCM249673.pdf. 60. Collins R, Peto R. Introducing new treatments for cancer: practical, ethical and legal problems. In: Williams CJ, editor. The ethics of clinical trials. New York: Wiley; 1992. 61. Karlowski TR, Chalmers TC, Frenkel LD, Kapikian AZ, Lewis TL, Lynch JM. Ascorbic acid for the common cold: a prophylactic and therapeutic trial. JAMA 1975;231:1038e42. 62. Henkin RI, Schechter PJ, Friedewald WT, Demets DL, Raff M. A double-blind study of the effects of zinc sulfate on taste and smell dysfunction. Am J Med Sci 1976;272:285e99. 63. Transcelerate Biopharma. Placebo and standard of care data sharing. Available at: http://www.transceleratebiopharmainc.com/ initiatives/placebo-standard-of-care/. 64. Farr BM, Gwaltney Jr JM. The problems of taste in placebo matching: an evaluation of zinc gluconate for the common cold. J Chronic Dis 1987;40:875e9. 65. Harrington RA, Lincoff AM, Califf RM, Holmes Jr DR, Berdan LG, O’Hanesian MA, et al. Characteristics and consequences of myocardial infarction after percutaneous coronary intervention: insights from the Coronary Angioplasty versus Excisional Atherectomy Trial (CAVEAT). J Am Coll Cardiol 1995;25:1693e9.

66. Mahaffey KW, Granger CB, Tardiff BE, Woodlief L, Lee KL, Weaver WD, et al. Califf RM for the GUSTO-IIb investigators. Endpoint adjudication by a clinical events committee can impact the statistical outcome of a clinical trial: results from GUSTO-IIb [abstract]. J Am Coll Cardiol 1997;29(Suppl. A):410A. 67. GUSTO IIa Investigators. Randomized trial of intravenous heparin versus recombinant hirudin for acute coronary syndromes. Circulation 1994;90:1631e7. 68. Antman EM. Hirudin in acute myocardial infarction. Safety report from the thrombolysis and thrombin inhibition in myocardial infarction (TIMI) 9A trial. Circulation 1994;90:1624e30. 69. Molhoek GP, Laarman GJ, Lok DJ, Luz CM, Kingma JH, van den Bos AA, et al. Angiographic dose-finding study with r-hirudin (HBW 023) for the improvement of thrombolytic therapy with streptokinase (HIT-SK) Interim results. Eur Heart J 1995;16(Suppl. D):33e7. 70. The GUSTO IIb Investigators. A comparison of recombinant hirudin with heparin for the treatment of acute coronary syndromes. N Engl J Med 1996;335:775e82. 71. Antman EM. Hirudin in acute myocardial infarction. Thrombolysis and thrombin inhibition in myocardial infarction (TIMI) 9B trial. Circulation 1996;94:863e5. 72. Connors Jr AF, Speroff T, Dawson NV, Thomas C, Harrell Jr FE, Wagner D, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996;276: 889e97. 73. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009;361:1139e51. Erratum in: N Engl J Med 2010;363:1877. 74. NIH-FDA Biomarker Working Group. BEST (Biomarkers, EndpointS, and other Tools) Resource. Available at: https://www.ncbi. nlm.nih.gov/books/NBK326791/. 75. Pratt CM, Moye L. The cardiac arrhythmia suppression trial: implications for anti-arrhythmic drug development. J Clin Pharmacol 1990;30:967e74. 76. Packer M. Calcium channel blockers in chronic heart failure: the risks of “physiologically rational” therapy [editorial]. Circulation 1990;82:2254e7. 77. The ALLHAT Investigators. Major cardiovascular events in hypertensive patients randomized to doxazosin vs chlorthalidoned the antihypertensive and lipid-lowering treatment to prevent heart attack trial (ALLHAT). JAMA 2000;283:1967e75. 78. Solomon SD, Uno H, Lewis EF, Eckardt KU, Lin J, Burdmann EA, et al. Erythropoietic response and outcomes in kidney disease and type 2 diabetes. N Engl J Med 2010;363:1146e55. 79. Micheel CM, Ball JR, editors. Evaluation of surrogate endpoints and biomarkers in chronic disease. Institute of Medicine consensus report. Washington (DC): National Academies Press; 2010. 80. Angell M, Kassirer JP. Editorials and conflicts of interest [editorial]. N Engl J Med 1996;335:1055e6. 81. Califf RM, Nissen SE, DeMaria AN, Ohman EM, Pitt B, Willerson JT, et al. Task force 2: investigator participation in clinical research. J Am Coll Cardiol 2004;44:1729e36. 82. Shuren J, Califf RM. Need for a national evaluation system for health technology. JAMA 2016;316(11):1153e4. 83. GUSTO Angiographic Investigators. The effects of tissue plasminogen activator, streptokinase, or both on coronary artery patency, ventricular function, and survival after acute myocardial infarction. N Engl J Med 1993;329:1615e22. 84. Rouleau JL, Moye LA, Pfeffer MA, Arnold JM, Bernstein V, Cuddy TE, et al. A comparison of management patterns after acute myocardial infarction in Canada and the United States. N Engl J Med 1993;328:779e84. 85. Mark DB, Naylor CD, Hlatky MA, Califf RM, Topol EJ, Granger CB, et al. Use of medical resources and quality of life after

II. STUDY DESIGN AND BIOSTATISTICS

REFERENCES

86.

87. 88. 89. 90. 91. 92. 93.

94.

95. 96.

97.

98.

99.

100.

101.

102.

103. 104.

105. 106.

107.

acute myocardial infarction in Canada and the United States. N Engl J Med 1994;331:1130e5. Pilote L, Califf RM, Sapp S, Miller DP, Mark DB, Weaver WD, et al. Regional variation across the United States in the treatment of acute myocardial infarction. N Engl J Med 1995;333:565e72. Pocock SJ, Stone GW. The primary outcome failsdwhat next? N Engl J Med 2016;375:861e70. Pocock SJ, Stone GW. The primary outcome is positivedis that good enough? N Engl J Med 2016;375:971e9. Yusuf S, Wittes J. Interpreting geographic variations in results of randomized, controlled trials. N Engl J Med 2016;375:2263e71. Dickersin K, Min YI. Publication bias: the problem that won’t go away. Ann NY Acad Sci 1993;703:135e46. Simes RJ. Publication bias: the case for an international registry of clinical trials. J Clin Oncol 1987;4:1529e41. Dickersin K, Chan S, Chalmers TC, Sacks HS, Smith Jr H. Publication bias and clinical trials. Controlled Clin Trials 1987;8:343e53. Kirkham JJ, Dwan KM, Altman DG, Gamble C, Dodd S, Smyth R, et al. The impact of outcome reporting bias in randomised controlled trials on a cohort of systematic reviews. BMJ 2010; 340:c365. Dwan K, Gamble C, Kolamunnage-Dona R, Mohammed S, Powell C, Williamson PR. Assessing the potential for outcome reporting bias in a review: a tutorial. Trials 2010;11:52. ClinicalTrials.gov. Trends, charts, and maps. Available at: https:// clinicaltrials.gov/ct2/resources/trends. Lee KL, McNeer JF, Starmer CF, Harris PJ, Rosati RA. Clinical judgment and statistics: lessons from a simulated randomized trial in coronary artery disease. Circulation 1980;61:508e15. Gibbons RJ, Christian TF, Hopfenspirger M, Hodge DO, Bailey KR. Myocardium at risk and infarct size after thrombolytic therapy for acute myocardial infarction: implications for the design of randomized trials of acute intervention. J Am Coll Cardiol 1994;24:616e23. Frieman JA, Chalmers TC, Smith Jr H, Kuebler RR. The importance of beta, the type II error and sample size in the design and interpretation of the randomized control trial: survey of 71 negative trials. N Engl J Med 1978;299:690e4. Kaul S, Diamond GA. Good enough: a primer on the analysis and interpretation of noninferiority trials. Ann Intern Med 2006;145(1): 62e9. BARI Investigators. Comparison of coronary bypass surgery with angioplasty in patients with multivessel disease. N Engl J Med 1996;335:217e25. The International Joint Efficacy Comparison of Thrombolytics (INJECT) Investigators. Randomised, double-blind comparison of reteplase double-bolus administration with streptokinase in acute myocardial infarction: trial to investigate equivalence. Lancet 1995;346:329e36. Van De Werf F, Adgey J, Ardissino D, Armstrong PW, Aylward P, Barbash G, et al. Single-bolus tenecteplase compared with frontloaded alteplase in acute myocardial infarction: the ASSENT-2 double-blind randomised trial. Lancet 1999;354:716e22. Begg C, Berlin J. Publication bias: a problem in interpreting medical data. J R Stat Soc A 1988;151:419e45. Detsky AS, Naylor CD, O’Rourke K, McGeer AJ, L’Abbe´ KA. Incorporating variations in the quality of individual randomized trials into meta-analysis. J Clin Epidemiol 1992;45:255e65. Berkey C, Hoaglin D, Mosteller F, Colditz GA. A random-effects regression model for meta-analysis. Stat Med 1995;14:395e411. Lau J, Antman EM, Jimenez-Silva J, Kupelnick B, Mosteller F, Chalmers TC. Cumulative meta-analysis of therapeutic trials for myocardial infarction. N Engl J Med 1992;327:248e54. Antman E. Randomized trial of magnesium for acute myocardial infarction: big numbers do not tell the whole story. Am J Cardiol 1995;75:391e3.

443

108. ISIS-4 (Fourth International Study of Infarct Survival) Collaborative Group. ISIS-4: a randomised factorial trial assessing early oral captopril, oral mononitrate, and intravenous magnesium sulphate in 48,050 patients with suspected acute myocardial infarction. Lancet 1995;345:669e85. 109. O’Connor CM, Starling RC, Hernandez AH, Armstrong PW, Dickstein K, Hasselblad V, et al. Effect of nesiritide in patients with acute decompensated heart failure. N Engl J Med 2011;365: 32e43. 110. Packer M, O’Connor CM, Ghali JK, Pressler ML, Carson PE, Belkin RN, et al. Effect of amlodipine on morbidity and mortality in severe chronic heart failure. N Engl J Med 1996;335:1107e14. 111. Frye RL, August P, Brooks M, Hardison RM, Kelsey SF, MacGregor JM, et al. A randomized trial of therapies for type 2 diabetes and coronary artery disease. N Engl J Med 2009;360:2503e15. 112. Arnout J, Simoons M, de Bono D, Rapold HJ, Collen D, Verstraete M. Correlation between level of heparinization and patency of the infarct-related coronary artery after treatment of acute myocardial infarction with alteplase (rt-PA). J Am Coll Cardiol 1992;20:513e9. 113. Hsia J, Kleiman NS, Aguirre FV, Chaitman BR, Roberts R, Ross AM. Heparin-induced prolongation of partial thromboplastin time after thrombolysis: relation to coronary artery patency. J Am Coll Cardiol 1992;20:31e5. 114. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988;2:349e60. 115. Roach GW, Kanchuger M, Mangano CM, Newman M, Nussmeier N, Wolman R, et al. Adverse cerebral outcomes after coronary bypass surgery. N Engl J Med 1996;335:1857e63. 116. National Institutes of Health. NIH policies and IC guidance for data and safety monitoring of clinical trials. Available at: https:// humansubjects.nih.gov/data_safety. 117. US Food and Drug Administration. Guidance for clinical trial sponsors. Establishment and operation of clinical trial data monitoring committees. March 2006. Available at: http://www.fda.gov/ downloads/RegulatoryInformation/Guidances/ucm127073. pdf. 118. Ellenberg SS, Fleming TR, DeMets DL. Data monitoring committees in clinical trials: a practical perspective. Chichester, UK: John Wiley & Sons; 2002. 119. Federal Policy for the Protection of Human Subjects (45 CFR part 46). Available at: https://s3.amazonaws.com/public-inspection. federalregister.gov/2017-01058.pdf 120. Menikoff J. The paradoxical problem with multiple-IRB review. N Engl J Med 2010;363:1591e3. 121. National Institutes of Health. Final NIH policy on the use of a single institutional review board for multi-site research. June 21, 2016. Available at: https://grants.nih.gov/grants/guide/notice-files/NOTOD-16-094.html. 122. H.R. 6. 21st Century Cures Act. Title II: development. Subtitle Me medical device regulatory process improvements. Section 2262. Available at: https://www.congress.gov/bill/114th-congress/housebill/6. 123. Glickman SW, McHutchison JG, Peterson ED, Cairns CB, Harrington RA, Califf RM, Schulman KA. Ethical and scientific implications of the globalization of clinical research. N Engl J Med 2009;360:816e23. 124. Yusuf S, Bosch J, Devereaux PJ, Collins R, Baigent C, Granger C, Califf R, Temple R. Sensible guidelines for the conduct of large randomized trials. Clin Trials 2008;5:38e9. 125. Califf RM. Clinical trials bureaucracy: unintended consequences of well-intentioned policy. Clin Trials 2006;3:496e502. 126. Clinical Trials Transformation Initiative website. Available at: www.ctti-clinicaltrials.org.

II. STUDY DESIGN AND BIOSTATISTICS

444

28. LARGE CLINICAL TRIALS AND REGISTRIESdCLINICAL RESEARCH INSTITUTES

127. Meeker-O’Connell A, Glessner C, Behm M, Mulinde J, Roach N, et al. Enhancing clinical evidence by proactively building quality into clinical trials. Clin Trials 2016;13:439e44. 128. Bernstam EV, Hersh WR, Johnson SB, Chute CG, Nguyen H, Sim I, et al. Synergies and distinctions between computational disciplines in biomedical research: perspective from the Clinical and Translational Science Award programs. Acad Med 2009;84:964e70. 129. Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions, version 5.1.0 [updated September 2011]. London, UK: The Cochrane Collaboration; 2011. Available at: http://handbook.cochrane.org/. 130. Califf RM, Peterson ED, Gibbons RJ, Garson Jr A, Brindis RG, Beller GA, Smith Jr SC, American College of Cardiology, American Heart Association. Integrating quality into the cycle of therapeutic development. J Am Coll Cardiol 2002;40:1895e901. 131. Welke KF, Ferguson Jr TB, Coombs LP, Dokholyan RS, Murray CJ, Schrader MA, et al. Validity of the society of thoracic surgeons national adult cardiac surgery database. Ann Thorac Surg 2004;77: 1137e9. 132. Tricoci P, Allen JM, Kramer JM, Califf RM, Smith Jr SC. Scientific evidence underlying the ACC/AHA clinical practice guidelines. JAMA 2009;301:831e41. Erratum: JAMA 2009;301:1544. 133. Han H, Chao H, Guerra A, Sosa A, Christopoulos G, et al. Evolution of the American College of Cardiology/American Heart Association clinical guidelines. J Am Coll Cardiol 2015;65:2726e34. 134. Califf RM. Benefit-risk assessments at the US Food and Drug Administration: Finding the balance. JAMA 2017;317 [Epub ahead of print].

135. Califf RM, Robb MA, Bindman AB, Briggs JP, Collins FS, et al. Transforming evidence generation to support health and health care decisions. N Engl J Med 2016;375:2395e400. 136. Sherman RE, Anderson SA, Dal Pan GJ, Gray GW, Gross T, et al. Real-world evidencedWhat is it and what can it tell us? N Engl J Med 2016;375:2293e7. 137. National Academies Press. The learning health system series. Available at: https://www.nap.edu/catalog/13301/the-learninghealth-system-series. 138. Califf RM, Pryor DB, Greenfield Jr JC. Beyond randomized clinical trials: applying clinical experience in the treatment of patients with coronary artery disease. Circulation 1986;74:1191e4. 139. VanLare JM, Conway PH, Sox HC. Five next steps for a new national program for comparative-effectiveness research. N Engl J Med 2010;362:970e3. 140. Lauer MS, Collins FS. Using science to improve the nation’s health system: NIH’s commitment to comparative effectiveness research. JAMA 2010;303:2182e3. 141. Whitehouse.gov. Report to the President from the Vice President. Cancer Moonshot. October 17, 2016. Available at: https:// obamawhitehouse.archives.gov/sites/default/files/docs/finalvp_ exec_report_10-17-16final_3.pdf. 142. Zarin DA, Tse T, Williams RJ, Carr S. Trial reporting in ClinicalTrials.govdthe final rule. N Engl J Med 2016;375: 1998e2004. 143. Clinical Trials Transformation Initiative website. AACTddatabase for aggregate analysis of ClinicalTrials.gov. Available at: https:// www.ctti-clinicaltrials.org/aact-database.

II. STUDY DESIGN AND BIOSTATISTICS

P A R T I I I

TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

29 Intellectual Property and Technology Transfer Bruce Goldsteina National Institutes of Health, Rockville, MD, United States

O U T L I N E Introduction

Part One: Intellectual Property Generally 448 Background: Intellectual Property Defined 448 PatentsdHistorical Overview 449 First Steps: Before the American Revolution 449 United States Constitution 449 United States, 1789e1951: Systemic Adjustments 450 United States: The Modern Framework 450 Patent Treaties 451 Modern Philosophy of Patent Law 453 Fairness and the “Quid Pro Quo” 453 Incentives for Product Development 453 Economic Engine 454 Core Concepts of US Patent Law 454 What Is a Patent? 454 Substantive Criteria for Patentability 455 Other Key Terms Defined 464 Patent Infringement (United States) 469 Basic Elements of the Patent Application Process 473 Content of a Patent Application 473 US Applications: Types and Filing Procedures 475 International Applications and Filing Procedures 479 Current Major Efforts to Alter US Patent Laws 481 Copyrights, Trademarks, and Trade Secrets 483 Copyrights 483 Trademarks 485 Trade Secrets 485 Part Two: Patents and Technology Transfer Critical Laws Concerning Patents and Federally Supported Research

Federal Funding of Private “Extramural” Research: The BayheDole Act 487 Federal “Intramural” Research: The StevensonWydler Act and the Federal Technology Transfer Act 493 Patenting and Licensing by Federal Agencies 495

448

487 487

Part Three: Technology Transfer Agreements 503 Background: Hypothetical Scenario 503 The First and Biggest Mistake: Signing the Agreements 504 Contract Execution in General 504 Scope of Actual Authority of Government Laboratories 506 Agreements to Protect Confidentiality 506 Background: Trade Secrets 507 Secrets and the Government 507 Anatomy of a Confidential Disclosure Agreement 508 Agreements to Transfer Materials 509 The Basic Material Transfer Agreement 509 Collaboration and Inventions: The Cooperative Research and Development Agreement 514 Background 514 Cooperative Research and Development Agreement Basics 515 Conclusion

518

Brief Glossary of Critical Terms in Patenting

519

Review Questions

519

References

520

a

Mr. Goldstein is a patent attorney serving as the Assistant Director for Monitoring & Enforcement Unit in the NIH Office of Technology Transfer. This chapter reflects the personal views of Mr. Goldstein, not of his employer. No official support or endorsement by the National Institutes of Health or the United States Government is intended or should be inferred.

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00029-0

447

Copyright © 2018. Published by Elsevier Inc.

448

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

INTRODUCTION Scientific research, ranging from the most basic to the most applied, is no longer (if it ever was) a loose enterprise of individuals conducting research as solitary islands, self-sufficient, and independent. More and more, research requires substantial interaction and interconnection with various colleaguesdcolleagues at universities, at nonprofit organizations, at government labs, and at companies. Many of these organizations care about patent rights in some way, whether acquiring or avoiding them. Thus, the pool of researchers whose careers remain untouched by patents has been shrinking for decades, and the trend is not slowing. One consequence of this trend is that intellectual property (“IP”) rights have a larger impact on how research is conducted, such that a researcher’s ignorance of them can be costly. Of the major types of IP, patents in particular play several roles. For many scientists, being listed as inventors on patents represents a means besides publication in peerreviewed journals to advance in their careers. Some scientists get involved in the patenting process when the lawyers handling a patent need technical advice or information about an invention. For those managing lab budgets, controlling patent costs without sacrificing valuable properties requires at least a modest understanding of patent principles and rules. Patents affect many scientists (albeit indirectly) when their colleagues are reluctant to collaborate or share materials unless invention rights are protected. Other scientists investigate patents so as not to waste effort researching a topic in which so many patents exist that the fruits of their labors are likely to be blocked. Finally, for those who want to see their ideas and discoveries become something more than just an item of academic interest, patents are a critical tool for enticing companies to invest in developing their discoveries into new products. As important as knowledge of IP may be, however, the goal of this chapter is limited to covering only the very basics, for several reasons. IP laws are complex and shift rapidly, and they carry enough specialized rules and nuanced interpretations that expert help from a qualified attorney should always be sought to address specific situations. Moreover, an in-depth knowledge of IP law and procedure is not necessary to conduct research. Nonetheless, scientists should have a functional understanding of what IP is generally and particularly how patents operate, to recognize problems and to be ready and able to help the experts address the issues. This chapter is divided into three major parts. Part One discusses IP in general terms. This part will define and differentiate the four major types of IP (e.g., history, philosophy), explain core concepts (e.g., substantive criteria, infringement), and briefly describe the process of

acquiring them (e.g., content, sequence of events, international elements). Part Two discusses patents in the specific context of technology transferdwhich will address in particular inventions created by, or with support from, the federal government. Part Three will focus on transactional agreements other than patent licenses, all of which are designed enable the movement of ideas from federal and academic labs into the commercial sector.

PART ONE: INTELLECTUAL PROPERTY GENERALLY Background: Intellectual Property Defined Property, in its broadest sense, means anything a person can directly control or can lawfully block (or “exclude”) others from controlling it. A property-owner’s exclusivity may take different forms, such as a physical barrier, a secret stash, or most importantly, a specific law that a court will enforce. When someone violates the property-owner’s exclusivity, the property owner can go to a court and secure an order to the violator to remedy the intrusion, perhaps returning the property or paying for the damage caused. This definition holds for land and things attached to it, for tangible personal property, and for intangible personal property (such as a bank account). “IP” follows the same principle but is subtly different. The key aspect for all types of IP is that a violation of exclusivity does not deplete or destroy the property; rather, the violation (called “infringement”) can be repeated infinitely. The damage caused by IP infringement, however, is realdoften measurable in the form of lost sales. Beyond that unifying principle, however, each type of IP is unique. Applying the laws, rules, policies, and processes designed for one type in the context of another type is a recipe for disaster. In the United States, four forms of IP dominate: patents, copyrights, trademarks, and trade secrets. In certain specialized fields, a person may run across some of the other, more esoteric forms of IP. For example, plant breeders may utilize a “Plant Variety Protection Act” certificate, issued by the US Department of Agriculture, and computer-chip makers may acquire “Mask Works” rights. But in the context of biomedical research focused on developing new drugs, biologics, and medical devices, only the four main types typically arise. To understand the context of each, a brief and very broad-view summary of each is helpful before diving into the details. A patent covers useful embodiments of discoveries, including machines and devices, chemical compositions, certain plants, and methods of making and using them. Such patents generally last about 20 years from the filing date of the initial patent

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

application. A copyright covers works of authorship that have been fixed in a tangible medium of expression, such as ink-on-paper, magnetic computer media, and film. Copyrights in works created today last the life of the author plus 70 years (older works may be governed by different times). A trademark is any form of unique identifier that helps consumers differentiate the sources of two or more similar goods or services. Trademark rights persist for as long as the mark is actually used in commerce and continues to help consumers. A trade secret is any commercially valuable, secret information that the owner has protected with reasonable security measuresdbut a trade secret only lasts for as long as the information remains secret and valuable, and only as long as the owner continues actually protecting it. Of these four types, patents play an overwhelmingly outsized role in the process of developing new biomedical products. Accordingly, the predominance of this part in the chapter will focus on the details of patent law, and it will close with a brief summary of the structure of the other three types.

PatentsdHistorical Overview First Steps: Before the American Revolution Patenting is not a new development nor even an American ideadthe concept first arose hundreds of years ago. To be sure, legal scholars have long argued over the “true” beginning of patent law. Some cite Greek writings 2500 years old describing certain exclusive rights that vaguely resemble something analogous to modern patents; other scholars set a much higher bar, accepting only those statutes that clearly resemble the current patent system. The majority of scholars, however, appear to agree that the first formal statute on patents was enacted by the Venetian Senate in 1474. That law protected any “new and ingenious contrivance” for up to 10 years, and it allowed the creator to sue for money and to have the infringing work destroyed. The rights provided by this statute were very similar to those given to book authors (printers being one of the industrial guilds the Venetian Senate wanted to foster). Beginning at least as early as King Henry VI in 1449, the English Crown began granting “Letters Patent,” which granted a favored person, guild, or company a commercial monopoly on certain products or services (usually, the Crown received a stream of income in exchange). Unhappy with perceived abuses of this practice, as well as with the pernicious effects monopolies had on commerce, the Parliament enacted the Statute of Monopolies in 1623. This statute prohibited further grants of monopolies, except for those granted on new inventions and for a limited term of years. Over the next 150 years, subsequent

449

acts and court cases eventually evolved these original Letters Patent on inventions to a system recognizable under modern patent standards. United States Constitution Before the American Revolution, the economies of the colonies were centered on the export of raw materials and agriculture, with little labor or capital to spare. Nonetheless, as early as 1641, some colonies created patent systems as a means to foster particular industries locally, which the Crown tolerated as long as Britain’s centralized manufacturing control was not threatened. With the start of the American Revolutionary War, imports of manufactured goods from England ceased, spurring the States1 in the newly created United States to develop home-grown manufacturing capabilities. As the Revolutionary War progressed and chances of success improved, the States began issuing patents in earnest. Unfortunately, by the time the Second Constitutional Congress met in 1787, the problems of State-specific patent systems had become obvious: getting patents in all States was difficult, patents were unenforceable outside the issuing State, and the scope of protection, rights, conditions, and duration varied. The most prominent example of the problems of Statespecific patent systems was the fight over the steamboat patents. Although Robert Fulton is widely credited with developing the first commercially successful steamboat design, it was not the first steam-powered boat invented in America. John Fitch and James Rumsey each claimed he was the first to invent the application of steam power (previously and unquestionably invented by James Watt) to navigating boats. Each sought patents in several States, sometimes the same State, and each spent years trying to disprove the other’s claims. As the steamboat represented a crucial means for opening up trade between the newly settled interior territories and the coastal ports, members of the Constitutional Congress, including George Washington, personally followed the fight closely. Another issue noted at the time was the dominance of British manufacturing. American inventors feared that their discoveries would be appropriated by manufacturers in England, who could overwhelm American capacity and underprice American manufacturers. Fostering and protecting domestic manufacturing industries was increasingly recognized as another means to support independence. Moreover, with the growth of lucrative interstate tradedand the taxes on that tradedStates recognized that local patent systems posed a barrier. As a direct result of these experiences, the framers of the Federal Constitution included a specific enumerated power for the Congress in Article I, Section 8: Congress shall have the Power . To promote the Progress of Science and the useful Arts, by securing for limited Times to

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

450

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

Authors and Inventors the exclusive Right to their respective Writings and Discoveries ..

This clause simultaneously reserves the patenting power exclusively to Congress, rather than the States, while it limits Congress’ authority to grant exclusive rights to: (1) a limited scope (writings and discoveries only); (2) a limited duration; and (3) a limited purpose (to the extent doing so promotes the progress of science and the useful arts). United States, 1789e1951: Systemic Adjustments One of the first laws passed by the first Congress after adoption of the Federal Constitution was the US Patent Act of 1790. It authorized any two of the Secretary of State, Secretary of War, and the Attorney General to grant patents if they found the invention to be “sufficiently useful and important.” The patents could be up to 14 years in duration, provided that the inventor offered a specification describing the invention to the Secretary of State at the time the patent was granted. The first US patent granted went to Samuel Hopkins for a new apparatus and process for making potash and pearl ash. It was approved by Secretary of State Thomas Jefferson and Attorney General Edmund Randolph, and then signed by President George Washington. As the rate of filings rapidly outpaced expectations, the original system was quickly replaced by the US Patent Act of 1793. This act authorized the creation of a formal patent office within the Department of State, where it remained until 1925, when it was moved to the Department of Commerce (DoC).2 The act provided the first formal definition of patentable subject matter, close to the one currently in use: “any new and useful art, machine, manufacture or composition of matter and any new and useful improvement on any art, machine, manufacture or composition of matter.” Also, the act clarified that the inventor’s specification must be “in such full, clear, and exact terms, as to distinguish the same from all other things before known and to enable any person skilled in the art or science of which it is a part, or with which it is most nearly connected, to make, compound and use the same.” The law changed the review by the patent office from a substantive examination to the formality of mere registration. Because many of the newly issued patents concerned “inventions” that were not new at all, Congress overhauled the patent system in 1836 to provide for substantive review of all patent applications. This act imposed on inventors a duty to distinguish the claimed invention from that which came before. It also established the concept of a statutory bar for failing to file the application timely. The act included means for an applicant to amend the application if the patent office raised any objections to the application as initially filed.

Additionally, it clarified that patents are to be awarded to whoever actually invented first, not merely to whoever filed the patent application first. Finally, it allowed for an extension of term of up to 7 years in certain circumstances (the standard term of new patents was formally extended in 1861 to 17 years from the date of issuance, with no extensions available). A late addition to the 1836 patent overhaul, arising after a fire destroyed the official record of patents (many of which were never recovered), initiated the current patent-numbering system and authorized the creation of several patent depository libraries around the country. These depository libraries served to increase public access to patents, as well as to prevent obliteration of the patent record should a disaster strike again. Several disparate efforts to reform the patent system brewing in the late 1860s were consolidated into a single overhaul in 1870, accomplishing several structural changes. Specifically, the new law required that every patent application “particularly point out and claim the part, improvement, or combination” constituting the invention, it codified the judicial doctrine of “best mode” (discussed below), it created a mechanism for figuring out which competing application claiming the same invention should prevail, it clarified that any public use or salednot just use or sale by the inventordwould trigger the statutory bar, and it authorized the patent office to issue regulations establishing processes for the orderly review and approval of patent applications. United States: The Modern Framework The 1952 Patent Act The Patent Act of 1952 fundamentally changed the structure of US patent law in three key ways. The first way changed the standard for patentability: The new law codified the judicial doctrine that inventions must be “not obvious,” which was in addition to the requirement for “novelty.” Second, the law changed the enforcement of patents: Before 1952, the courts were left to decide for themselves whether someone had infringed a patent; the new law crafted a formal definition for “infringement,” which raised the significance of the “claims” of a patent. Third, the law made a series of structural changes, the most significant of which was to clarify that patents can still contain enforceable claims even after some of its other claims have been held invalid. The concepts of “novelty,” “nonobviousness,” and “claims” are all discussed in more detail later. The “Federal Circuit” In 1982, Congress created a special court to consolidate and resolve questions on patent law. To explain its significance, a brief background on litigation generally under US federalism principles is required.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

Most legal cases, such as contract disputes, divorces, and crimes, are tried in State courts; federal courts generally only hear cases if the dispute arose under federal law, though they also may hear cases arising between citizens of different States. Under the Constitution, Congress has exclusive power to make patent law; the States are forbidden to pass any patent laws of their own. Moreover, State courts are likewise forbidden from deciding any issue of substantive patent lawdall such cases must be brought in a federal district court. Since the 1890s, all appeals from the federal district courts, including patent cases, have been heard by regional “Circuit” Courts of Appeals; requests for further appeals go to the US Supreme Court, which has discretion whether or not to hear any given case. In 1909, Congress created the first Court of Appeals with jurisdiction over fixed subject matter (as opposed to over territories). This court originally only had authority to hear appeals from the Customs Office. This jurisdiction expanded in 1929 to include appeals from the US Patent Office, becoming the US Court of Customs and Patent Appeals (“CCPA”).3 As is true for many areas of law, the various Circuit Courts of Appeals did not agree with each other on major tenets of patent law. Compounding the problem was the fact that the regional circuits, which heard patent cases in the context of infringement litigation, often disagreed with the CCPA, which heard patent cases in the context of the application proceedings before the US Patent Office. To clarify patent law, in 1982 Congress created the Court of Appeals for the Federal Circuit (“CAFC”), abolishing the CCPA in the process.4 While it is still possible for a substantive patent-related issue to be heard in one of the regional circuits, that event is rare, and when any regional circuit court considers such an issue, that circuit court will rely on and defer to decisions of the CAFC on all substantive principles of patent law. US Patent Reform of 2011 On November 16, 2011, the Leahy-Smith America Invents Act,5 more commonly and informally called the Patent Reform Act, became effective. To those whose professions are deeply enmeshed in patent lawd whether through applying for, licensing of, enforcing, or challenging the validity of patentsdthe Patent Reform Act was certainly the most significant overhaul of U.S. patent law since 1952, perhaps since 1836. To everyone else, the changes were more subtle, and to those directly engaged in biomedical research and product development, the changes were even more subtle. The goals of the new law included making patent law more consistent with international standards, making the process of acquiring a patent more coherent and more predictable, and making the value and validity

451

of patents more secure by making the processes for eliminating or fixing poor patents easier. The most high-profile part changed who has the right to receive a patent: from the first person to invent to the first inventor to file a patent application. Other major components of the law included changing the standard of what constitutes “prior art,” simplified formalities in filing, reforming the mechanisms through which others can challenge patents and patent applications before the US Patent and Trademark Office (USPTO), and removing the “best mode” standard as a basis for invalidating a patent. These changes will be discussed below in more detail. How well the law will serve its goals will not be known for a long time. Some of the terms of the Patent Reform Act took effect immediately, but most of its terms started operating during an 18-month transition period. Patents filed before March 16, 2013 (and a small set of those filed later) continue to be governed by the patent laws and procedures in effect before the Patent Reform Act; so the old rules will remain relevant until 2033. Patent Treaties Patents have substantial international significance, so the fact that several major treaties address patents should surprise no one. The following discussion summarizes some of the most important treaties and conventions applicable to modern patent practice, arranged in chronological order of signature. 1. The Paris Convention for the Protection of Industrial Property (1883) The first treaty governing IP rights was the Paris Convention.6 Essentially, it specified that the filing date for a patent application in one country would be recognized in every other member’s patent office, provided the applicant files parallel applications in those other members’ patent offices within 12 months. As of December 2011, the Paris Convention has 174 members. Kuwait was the latest to join, in 2015. 2. The Strasbourg Patent Convention (signed 1963, effective 1980) To harmonize substantive standards for granting patents, the European Union negotiated a treaty called the “Convention on the Unification of Certain Points of Substantive Law on Patents for Invention,” commonly called the Strasbourg Patent Convention. The treaty became effective in 1980, once eight nations’ legislatures had ratified it (including France and Germany). Currently, thirteen countries are members.7 While the treaty harmonized substantive patent law, procedural laws remained outside its scope.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

452

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

3. The WIPO Convention (signed 1967, effective 1970) In 1967, the “Convention Establishing the World Intellectual Property Organization (WIPO)” was signed in Stockholm, Sweden. WIPO, operating under the auspices of the United Nations and headquartered in Geneva, Switzerland, currently has 184 contracting members (each member has one vote). WIPO administers 24 multinational treaties, proposes various international initiatives, holds various events discussing emerging IP issues, and provides a number of IPrelated services. 4. The Patent Cooperation Treaty (PCT) (signed 1970, effective 1978) The PCT, administered by WIPO, was designed to streamline international filing of patents on a single invention by crafting a common application design and common filing procedure. Any signatory to the Paris Convention is automatically eligible to join the PCT. Initially adopted by 18 contracting countries, as of July 2017 the treaty has 152 members. Nonmembers include Paraguay, Uruguay, Venezuela, Iraq, and Pakistan. The procedures for filing a patent application under the PCT are discussed in more depth below. 5. Regional Patenting Systems On liberation from French colonial rule in the early 1960s, twelve newly founded African nations banded together to harmonize several legal institutions, including their patenting systems. They created the first regional patenting authority, the African and Malagasy Patent Rights Authority (OAMPI), born on September 13, 1962, on the adoption of the Libreville Agreement. That treaty required member nations to conform their laws such that the issuance of a patent by a centralized authority would have uniform legal status as a patent in all member nations. In 1977, this organization became the Organisation Africaine de la Proprie´te´ Intellectuelle (OAPI).8 Europe soon decided to follow OAMPI’s example, to minimize the burden of having to file separate patent applications in every European nationdbut Europe’s challenge was larger because each patent application had to be translated into the official national (i.e., unique) language, had to be filed by local counsel on exactly the same day, and otherwise had to follow all nation-specific laws and procedures. In 1973, the European Union adopted the “Convention on the Grant of European Patents,” commonly called the “European Patent Convention” or simply the EPC. The EPC required member nations to harmonize their patent laws and authorized creation of a single, regional

patent office that would handle intake and initial processing of patents, which later would be transferred to national patent offices in those European nations selected by the applicant. While the European Patent Office (EPO) does grant a “patent,” which in theory is directly enforceable throughout all member nations, each member retains the independent authority to require that every patent must be filed in the national patent office translated into the local language (“validated”) before its owner can enforce the patent in that nation. Three other regional systems have been founded since the EPO. The first is the African Regional Intellectual Property Organization (ARIPO),9 which was founded in 1976 and uses English as the common official language. The second, the Eurasian Patent Organization (EAPO), was founded in 1994 and uses Russian as the common official language.10 Finally, the Gulf Cooperation Council (GCC) operates a regional patent registration office for six Arab states, which became operational in 1998.11 Unlike the other regional offices, however, the GCC office does not participate in the PCT system, even though all of its member states participate in the PCT system. 6. Updates to the GATT of 1947dUruguay Revisions and TRIPS (signed 1994, effective June 5, 1995); Doha Declaration (signed 2001) In 1947, the United Nations negotiated the General Agreement on Tariffs and Trade. This treaty created a body to review and resolve trade disputes among its members. The members continue to update the underlying treaty through a series of “Rounds” of negotiations. The Uruguay Round, held from 1986 to 1993, which culminated in the formation of the World Trade Organization, also expanded topics eligible for discussion, including IP. At the end of the Uruguay Round, the members signed the “Agreement on Trade-Related Aspects of Intellectual Property Rights,” commonly known as “TRIPS.” TRIPS required its members to harmonize certain major elements of their patent, copyright, and trademark laws. In the United States, Congress had to change patent laws (as applied only to those applications filed after June 5, 1995, when the treaty entered effect) in several key ways, the most important of which were: (1) calculation of patent term; and (2) publication of pending patent applications. The TRIPS treaty quickly became controversial because of the perception that emerging-economy countries would be dis-empowered to protect the health and welfare of its citizens merely to preserve the patent rights (i.e., profits) of multinational pharmaceutical

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

453

As a general proposition, free-market economies thrive best when every participant has an equal ability to compete. Nothing is wrong with people trying to earn the position as the sole provider of a good or service through low pricing, superior quality, or other characteristic, as long as others can try to match or beat them. So why would any free-market country go out of its way to create a legal mechanism that allows someone to suppresses other people’s ability to enter the market?

inventions secret for as long as possible to prepare for the competition. The patent systemdand any market distortion resulting from a given patent12dactually serves the public good. First, it provides the innovator with an incentive to try to innovate, by shielding the innovator against the copier for a limited time. Collaterally, this benefit provides an incentive to those with spare capital to invest in research and development, hoping a valuable invention will arise. Second, patents expire on a fixed date, meaning that the public (particularly copiers) knows when they can start competing. Third, patents are public documents, meaning that the public can begin to learn from a patent the moment it is published, even if no one may compete yet. Fourth, patents provide the public with an incentive to design something new that is not blocked by the patent. At the same time, the patent system imposes a specific “price” to the innovator: The innovator must disclose everything. As discussed in more detail below, the innovator must tell the public how to practice the invention in such clear and detailed terms that someone having ordinary skills in the field of the invention could practice it. Additionally, the inventor must alert the USPTO about any prior publications or other public disclosures related to the invention (to the extent the inventor is aware of them), to help the USPTO narrow the scope of the patent appropriately. In essence, both the inventor and the public gain from the patent system: by conceding something valuable, each gains something else valuable. This critical concept in patent law is generally called the “quid pro quo,” a Latin phrase meaning “this for that.” Those who object to the benefits conferred on patent owners should bear in mind the benefits that the public enjoys.

Fairness and the “Quid Pro Quo”

Incentives for Product Development

An old adage states, “build a better mousetrap and the world will beat a path to your door.” That adage is generally respectable, but standing alone, it has a flaw: How can a person or a small business hope to compete with a large corporation that already has established manufacturing facilities in countries with cheap labor and economies of scale, already has well organized transportation logistics, and already has marketing units and dozens of distribution centers primed to sell the product? Copiers are merely free riding on others’ creativity; by definition, they do not add to the body of knowledge or otherwise create anything new, so their contribution to society (lower pricing) is less valuable than the innovators’ invention. As such, basic fairness suggests that the innovator should have the first opportunity to reap profits. Far worse to society: without a patent, innovators have an incentive to keep their

For many products, once they are invented, they are ready for marketdbut not all. Sometimes, a prototype or discovery works in the lab, but requires substantial experimentation and modification to be turned into a market-ready product. No one will invest the resources to convert that good idea into an actual productdunless they can be assured that, once the product reaches the market, they will have the opportunity to earn back the original investment plus a profit. Patents provide that assured opportunity (though only the market will determine if the inventor will actually see any financial return). While some economic studies have shown this prediction of behavior is indeed a general trend,13 a case in point is a new pharmaceutical product. A researcher might discover today that a class of molecules having a common chemical backbone all appear to treat a

companies. Accordingly, the members of the World Trade Organization (WTO) Ministerial Conference of 2001, meeting in Doha, Qatar, adopted the “Doha Declaration on the TRIPS Agreement and Public Health.” This declaration asserted that TRIPS includes flexibility among member states to circumvent patent rights to protect public health in the event of national emergencies or other circumstances of extreme urgency (each member state may determine what qualifies based on its own criteria). 7. Patent Prosecution Highway Technically not a treaty, the Patent Prosecution Highway (“PPH”) is a series of agreements between patent offices to cooperate on substantive review of patent applications. It permits the various participating offices to share information, to harmonize practices, and most importantly, to fast-track examination of claims that other participating offices have allowed. Applicants seeking to utilize this program do not have to pay an additional fee to the patent offices, though the complexity might entail additional lawyers’ fees.

Modern Philosophy of Patent Law

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

454

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

specific bacterial infection in a Petri dish, but that does not mean that that researcher can start selling the chemical as a drug. The researcher must first discover, among other things, all of the following: (1) which chemical subtype has the optimal therapeutic benefit with lowest toxicity; (2) how to get the drug into the body so enough of it will reach the bacteria; (3) how the drug moves through the body, and what side effects it causes; (4) what the minimum, optimal, and maximum effective doses are, and how long the drug stays effective in the body; and (5) whether there are any hidden risks that will only become apparent after wide-scale testing. Moreover, the researcher needs enough data on each of these points to convince the US Food and Drug Administration (FDA) and its international counterparts that the drug is safe and effective. Pharmaceutical testing requires years and hundreds of millions of dollars, and, for most candidate drugs, likely will never earn back what was spent developing it. Drug-development companies rely on one success to pay for many failures. If a drug that reaches the market can be manufactured at pennies a pill, copiers will get a free ride on the inventor’s hard work and investment. Without a patent, no one will invest in developing the inventor’s molecule into a drug, the drug will never reach the market, and the public health will not improve. To be sure, the facts that make the “incentive to develop” rationale for the patent system so compelling for FDA-regulated biomedical products are not present in other fields, and even where the critical facts are present, they are not necessarily as persuasive. Nonetheless, this rationale remains relevant in any field requiring a significant translational step between inventing a technology and marketing a product that embodies that technology. Economic Engine A hotly debated economic theory supporting a strong patent system argues that nations with stronger patent systems have more robust economies than nations with weaker patent systems. The idea is that nations become committed to strong protection of IP when they are already investing a substantial fraction of their disposable wealth or gross domestic product (GDP) on research and development, and on infrastructuredin other words, nations that decide to rely on innovation as a source of prosperity (as opposed, say, to relying on raw resources or cheap labor) will grow faster than those that do not. The data on this point is not uniform, but the majority of empirical studies show a positive correlation between strengthening patent systems and economic growth14 While none of these studies proves that strong patent systems cause growth, the weight of the data does suggest some relationship exists.

Core Concepts of US Patent Law What Is a Patent? Patents Internationally As noted above, patents are national documents: a patent issued by the USPTO has legal effect only within the borders of the United States; a patent issued by the French Patent Office is effective only within the borders of France; etc. Each nation decides what subject matter can be patented within its borders, and what the applicant must do before that nation’s patent office will issue a patent. The exact scope of rights granted by each nation also varies. Additionally, enforcement of patents can require navigating administrative offices, specialized courts, and other elements unique to each nation. Some nations will respect patents issued by other nations,15 but this is the exception rather than the norm. Utility, Plant, Design The United States offers three categories of patents. The first and by far most common is the “utility” patent. “Utility” patents cover new and useful embodiments of ideas and innovations. According to the USPTO’s data for 2000e14, about 95% of all patent applications filed in the United States are utility patents. As a result, most unspecified references to “patents” actually refer to utility patents. Other than in this section, all references in this chapter to patents are about utility patents. Another type of patent is the “plant” patent. These protect strains of plants that the inventor both (1) invented or discovered and (2) propagated asexually.16 Examples of invention or discovery of plants include cultivated sports, mutants, hybrids, and newly found seedlings (other than a tuber propagated plant or a plant found in an uncultivated state). Asexual propagation techniques include rooting of cuttings, layering, budding, grafting, and inarching. Even for patent practitioners, plant patents are rare, with just 0.2%e0.3% of all US patent applications being for plant patents. The final type of patent is the “design” patent. Design patents protect novel, nonfunctional, ornamental elements.17 For example, in a newly designed bicycle rack, improvements in the way it allows bicycles to be stacked and locked (a function) would be protected by a utility patent, while its unique appearance might be protected by a design patent. On sneakers, stitching having no structural function often is the subject of a design patent.18 Litigation between Apple, Inc. and Samsung over cell phone design patents has reached the popular press. That said, design patents remain a relative niche, at a little over 5% of all US patent applications.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

Specific Rights Conveyed by Patents The USPTO publishes formal notice to the public each time the USPTO issues a patent, proclaiming that the government has granted certain rights to the invention described in the patent, which will persist through a specific end date. The powers attached to a patent are the right to exclude others from making, using, selling, and importing the claimed invention in the United States. The recipient of the newly issued patent, called the “patentee,”19 may ask the government to enforce those rights (i.e., sue for patent infringement), may enter contracts with others in which the patentee agrees not to enforce the patent (a “license”), may sell the patent, or may simply do nothing. A question commonly arises at this point: “so, can the patentee practice the invention described in the patent?” The answer is, “not always.” Consider a hypothetical case, ignoring history: Inventor A patents the first aircraft, which uses a powered propeller attached to a fixed wing. The following year, Inventor B patents the first jet-powered engine. For as long as these patents remain active, no one may combine these two technologies to make a fixed-wing aircraft powered by jet engines, because Inventor B’s efforts would be blocked by Inventor A’s patent, and A’s efforts would be blocked by B’s patent. Inventor B could, however, try to sell a jet-powered car without worrying about Inventor A’s aircraft patent. To sell a product covered by multiple patents, someone must secure permissions from all patent owners to make it either by licensing or by acquiring ownership of the relevant patents. Alternatively, someone can design a new version of the product that is not blocked (in the hypothetical, perhaps by inventing a helicopter). Indeed, this incentive to “design around” others’ patents is a critical feature of patent law. Substantive Criteria for Patentability Having now described the broadest outlines of patents, the next step is to focus on the core criteria for acquiring patents. In other words, this section will try to address the circumstances under which a discovery qualifies as “patentable.” Under US law, an inventor is presumptively entitled to a patent on appropriate “subject matter,” unless the invention lacks “utility,” is not “novel,” is “obvious,” or is not adequately “described in writing.” Each of these quoted words and phrases carries special meaning in patent law, as elaborated below. Patentable Subject Matter General Principles Under US law, specifically 35 U.S.C. x 101, an inventor may apply for a patent claiming new or improved versions of (1) compositions of matter;

455

(2) machines; (3) articles of manufacture; and (4) processes. “Compositions of matter” generally include chemicals (including biological products), alloys, and other mixtures. “Articles of manufacture” include all objects that are not machines or compositions of matter; an easy example is a golf ball with a unique pattern of dimples. “Processes” refers to activitiesdmainly methods of making or using a thing. Also, an inventor may claim a combination of these elements. As the Supreme Court made clear, these categories cover “anything under the sun that is made by man.”20 Over several cases, however, the court has observed that not everything under the sun is, in fact, made by man; therefore, not everything can be patented. First, mathematical formulas, abstract ideas, and properties of physics cannot be patented because humans merely observe or discover them, we do not make them.21 So, Einstein could not have patented his famous equation “E ¼ mc2,” but if someone built a fusion-powered electric generator that relies on that formula, that machine clearly could be patented.22 Likewise, mere strategies and vague concepts are not patentable standing alone, rather, they must be actually embodied in a prototype or practical process. Second, items as they occur in nature cannot be patented, again because humans did not make them, we just found them. So, if someone walks through a forest, chews on a leaf from a tree, and notices her headache has disappeared, she would not be able to patent the leaf, the tree, or the tree’s genome, because these existed in their current form before any human realized their health effects. If, however, she took some of those leaves back to a laboratory and extracted an active chemical from those leaves, she could patent the isolated biochemically active ingredient. Third, patents are inappropriate for works of authorship, such as text and visual arts, which are the domain of copyright law, and for product-identifying logos, slogans, and monikers, which are the domain of trademark law. To be sure, some items overlap and so can be simultaneously protected by more than one form of IP. For example, computer software can incorporate a new method of processing data (patentable) as well as the text of the software coding (copyright). Likewise, a sneaker’s stitching can include both novel ornamentation (design patent) and marketing logos (trademark). These instances of overlap, however, are the exception rather than the rule. These general rules may appear straightforward, but applying them to specific situations has proven difficult, particularly in the last few decades. Three areas in particular have generated considerable debate, all of which have the potential to impact biomedical research: “mere associations,” living organisms and DNA, and

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

456

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

software and algorithms. Each of these deserves a brief discussion to show how the concept of statutory subject matter has evolved recently. “Mere Associations:” LabCorp v. Metabolite After much painstaking research, a scientist notices that a change in a blood-borne biomarker correlates with a medical condition or disease. In that situation, where is the line between a new diagnostic tool (patentable) and the mere recognition of a natural phenomenon (unpatentable)? Trying to draw that line has triggered a substantial debate, especially after the lightning-rod case known as LabCorp v. Metabolite.23 Deficiencies in folic acid and vitamin B12 are known to cause several diseases, but detecting the precise level of each of these vitamins in the body is difficult. In 1986, researchers at the University of Colorado and at Columbia University recognized that because these vitamins are critical to the metabolism of homocysteine (an amino acid), elevated homocysteine levels in a person’s blood would reveal a deficiency in one or both vitamins. Additionally, they discovered that simply measuring the amount of homocysteine circulating freely in the blood was not an accurate indicator; rather, they figured out how to measure total homocysteine in the body based on a blood sample. A patent application was filed, assigned to University Patents Inc. (UPI, a for-profit company established to market universities’ patents), then assigned to Competitive Technologies (UPI’s successor), then licensed exclusively to Metabolite Laboratories, Inc. (“Metabolite”), and finally sublicensed to Roche Biomedical Laboratories (which was later spun off to become Laboratory Corporation of America (“LabCorp”). The patent had 34 claims, two of which are critical: Claim 1 describes in detail a new method for measuring total homocysteine, and it was not controversial. Claim 13, however, reads as follows: A method for detecting a deficiency of cobalamin or folate in warm-blooded animals comprising the steps of assaying a body fluid for an elevated level of total homocysteine; and correlating an elevated level of total homocysteine in said body fluid with a deficiency of cobalamin or folate.

LabCorp changed its testing methods and decided not to continue paying royalties to Metabolite, and Metabolite sued LabCorp for patent infringement. A jury found willful infringement, the CAFC affirmed, and LabCorp appealed to the Supreme Court. Initially, the Supreme Court agreed to hear the case, but on different grounds than had been discussed by any of the parties in any of the prior proceedings: whether a claim directed merely to “correlating” test results with a condition is patentable (such that a doctor necessarily infringes the patent merely by thinking about the

relationship after looking at a test result). Although the court ultimately dismissed the appeal as “improvidently granted,” leaving the lower decision intact, three Justices dissented, arguing that the court should have decided this issuedpresumably by striking down the patent.24 Living Organisms and DNA: From Chakrabarty to Mayo to Myriad Following the discovery of the helical structure of DNA, a question has repeatedly arisen, albeit in different forms: Under what circumstances can one patent DNAdthe stuff of life? On the one hand, DNA is a chemical, subject to all the same principles and dynamics of any other chemical. Patent law, in general, is carefully constructed to avoid creating classes of patents with different rules for different technologies. On the other hand, DNA is something more than just any chemical, inextricably tied at some level to an organism’s very identity. Moreover, besides the purely logical positions, the mere notion of tinkering with DNA triggers powerful fears and other emotional reactions. Drawing a suitable line between patentable and unpatentable, therefore, remains a difficult and controversial exercise. The first major foray began with the 1980 Supreme Court case of Diamond v. Chakrabarty.25 Dr. Chakrabarty genetically engineered a bacterium capable of breaking down crude oil, to be used to clean up oil spills. The USPTO rejected his patent application, asserting that living organisms could not be patented even if they were engineered. The CCPA reversed this decision, and the Supreme Court agreed with the CCPA in a 5-4 decision. The majority concluded that it is up to Congress to set the bar for what is patentable, and Congress did not limit “compositions” and “articles of manufacture” to nonliving ones. As this organism was clearly created by a human through engineering, rather than either found in nature or bred using natural evolution, it qualifies for patent protection. While the Chakrabarty decision has been criticized on policy grounds (i.e., that living organisms should not be patentable), the decision remained legally unchallenged for 30 years. In substantial reliance on that decision,26 hundreds of biotechnology companies have been formed,27 many of which rely on patents claiming genetic elements in one way or another. Also, the USPTO has issued tens of thousands of patents claiming DNA or RNA, including patents on nonhuman genetic elements. Finally, by some estimates, US patents claim as much as 20% of the human genome.28 In this context, two court cases represent a potentially major shift: Mayo Collaborative Svcs. v. Prometheus Labs., Inc.,29 and Association for Molecular Pathologists v. USPTO and Myriad Genetics.30 Each case’s complex facts

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

are worth examining in some detail as they highlight several intertwined policy issues. The Mayo case concerned two thiopurine drugs, 6-MP and AZA (AZA is a precursor that converts to 6-MP in the body), both used for years to treat certain inflammatory bowel diseases. The body eventually metabolizes the drugs into toxic products, such that the doctor must give enough of a dose to be therapeutically effective without causing too much toxicity. Unfortunately, each person metabolizes thiopurines differently. In the late 1990s, the Hopital-Sainte-Justine (Montreal, Canada) discovered that metabolite levels correlated to therapeutic dosing, patented the method, and exclusively licensed rights to Prometheus Labs. Initially, the Mayo Clinic purchased Prometheus’s test, but in 2004 Mayo ceased purchasing the test, announcing its intent to begin using its own test and to begin selling that test to others. Prometheus sued for patent infringement. The district court found that Mayo’s test infringed the patents, but ruled that the patents were invalid because they essentially claimed correlations involving natural phenomena. On appeal, the CAFC reversed, holding that the claims as a whole included “administering” the drug and “determining” metabolite levels, both of which were specific, physically transformative acts. The Supreme Court initially vacated this decision for reconsideration in light of a recently issued patent decision on the subject,31 but the CAFC decided that its prior holding should not be changed. Mayo again appealed to the Supreme Court. The Supreme Court, in contrast, unanimously held that the patent, as written, actually claimed a property of nature. The court focused on the thing that made the discovery important: the relationship between certain metabolites in the blood and the likelihood that the drug would prove ineffective or cause harm. The way the body metabolizes thiopurines is an entirely natural process, and the claims merely recite that natural process and add some unpatentable additional stepsdadministering the drug and taking measurements, both of which were already known. Even if the court viewed the collection of individually unpatentable steps as a unitary “invention,” the only part that distinguishes it from prior testing is the recognition of the significance of a measurement, not the mechanism for measuring. The Myriad case concerns gene mutations associated with susceptibility to breast cancer. In the early 1990s, scientists at a company called Myriad Genetics, working with a variety of colleagues,32 sequenced two genes critically important to a large fraction of breast cancers, dubbed BRCA1 and BRCA2. Those people who inherit a deleterious mutation in either gene have a dramatically higher likelihood of developing cancer, particularly breast, ovarian, and prostate cancers.

457

Currently, the medical community has still not agreed on a single set of criteria for when to get tested. The NIH National Cancer Institute (NCI) suggests that, when someone is diagnosed with breast or ovarian cancer, the patient should be tested for the BRCA genes, and if the test is positive, their family should be screened as well.33 Those with the gene can better select the best therapies for current cancer, know to increase their vigilance for future cancer, and can better gauge whether or not removing tissue preemptively is wise. Myriad filed for and received several patents claiming specific genetic sequences and methods of diagnosing cancer using these sequences. Myriad immediately began offering testing services, at a cost on the order of about US$3000 per test; unfortunately, the cost of the test was not covered by insurance and Medicaid for many years.34 Other sources, however, ignored the patents and began offering the test: the regional public health plan of Ontario, Canada, offered the same test for about a third the price; also, several pathologists and other clinicians offered the test out of their own labs. Myriad sent “cease and desist” letters to them, offering license terms that these other testers say were offensive. Myriad sued the University of Pennsylvania, after which most testers stopped offering BRCA tests, albeit under protest. In May 2009, a large and diverse group filed suit against Myriad, the University of Utah, and the USPTO. The plaintiffs included the Association for Molecular Pathology (AMP), American Civil Liberties Union (ACLU), several medical groups,35 several clinicians,36 genetic counselors,37 patient advocacy groups,38 and individual patients claiming injury based on Myriad’s testing policies.39 The case drew considerable national attention, with over 30 “friend of the court” briefs filed by interested third parties urging various outcomes. On March 30, 2010, the federal district court held in favor of the plaintiffs against Myriad and the University of Utah (claims against the USPTO were dismissed). The district court ruled in favor of the plaintiffs, and Myriad appealed. The CAFC reversed in part and affirmed in part. Favoring the views of Myriad, the court ruled that isolated DNA can be patentable subject matter, precisely because these molecules are different from those that occur in nature directly because a human changed the molecule. More specifically, one must consider whether human intervention has given “markedly different,” or “distinctive,” characteristics. In doing so, the court suggested that limiting the analysis to the “informational content” of the molecule is not appropriate. Favoring the views of the plaintiffs, the court held that the “comparing” and “analyzing” claims are ineligible for patent protection. Looking at the language of Myriad’s claims, the court focused on the fact that they entirely relied on mental processes occurring

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

458

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

independent of the physical steps taken to gather the data. AMP appealed to the Supreme Court. The Supreme Court largely sided with AMP and the other plaintiffs, but drew a distinct line. For raw genomic DNA and resulting mRNA, these are items found in nature. The court was unpersuaded by arguments that the isolated BRCA DNA does not naturally occur anywhere, only with human intervention. The chromosomal location and information content of the genes existed before any human found them, and even Myriad conceded that it did not change the isolated oligonucleotides. The groundbreaking nature and publichealth impact of the discovery, while substantial, do not convert what is already in existence into a patentable invention. Where there is patentable subject matter is where a human has changed the form or sequence. In particular, cDNA sequences (as long as they are not so short as to be indistinguishable from the original) truly are not found in nature, they are creations of human intervention. The court noted explicitly that none of the method claims, new applications of newly discovered sequences, and inventions involving the rearrangement of genetic elements were being considered. In the wake of this holding, Myriad effectively ceased efforts to enforce the remaining claims of its patents. Algorithms and Software: Benson-Flook-Diehr, State Street, and Bilski Algorithms put the patent system in a quandary: at what point is a mathematical formula transformed from merely being the description of a natural phenomenon to something “made by man”? That proposition, while easy to pose, has repeatedly proven difficult to answer, as shown by a series of court cases. Following the 1908 court case of Hotel Security Checking Co. v. Lorraine Co.,40 the Circuit Courts of Appeals ruled consistently that methods of doing business, even if truly new, are unpatentable subject matter. For decades, the Supreme Court declined to take any case on the issue, so commentators generally believed the issue was settled. Then, a triad of Supreme Court cases in the 1970s changed the landscape. The first in the triad was Gottschalk v. Benson.41 Mr. Benson filed a patent application on a method of using a general-purpose computer to convert “binary-coded decimals” into pure binary numbers. The Patent Office rejected the application, but this was reversed by the CCPA. The Supreme Court, however, agreed with the Patent Office that the invention was not patentable because it claimed a pure mathematical operation. “Transformation and reduction of an article ‘to a different state or thing’ is the clue to the patentability of a process claim that does not include particular machines.”42 The court carefully chose not to hold that a process patent must either be tied to a particular

machine or apparatus or must operate to change articles or materials to a different state, and explicitly indicated that the decision did not preclude patents on software. Nonetheless, the court clarified that the software in question was unpatentable because the software was not limited to a particular computer, such that a patent would preclude anyone from using any computer to calculate that mathematical formula. Four years later, the Supreme Court faced the case of Parker v. Flook.43 Mr. Flook filed an application claiming an improved catalytic converter in which the improvement was that the computer monitoring conditions in the catalytic chamber used a new algorithm to update the “alarm limits” (the designated normal operating range) for the converter. As in Benson, the USPTO rejected the application and the CCPA reversed. Citing Benson, the Supreme Court again reversed. While this time the algorithm was tied to a specific piece of hardware, the court indicated that the mere fact that some incidental physical event happens after a computer has used the algorithm does not necessarily make the algorithm itself patentable.44 Three years later, the Supreme Court heard the case of Diamond v. Diehr.45 The invention was directed to a machine for molding uncured synthetic rubber, using a computer with software that used a new algorithm to determine when the rubber was cured.46 The USPTO rejected the patent application, and again, the CCPA reversed. This time, the Supreme Court affirmed the CCPA’s decision, indicating that the invention was patentable. While acknowledging that an algorithm by itself is not patentable, a machine or process that uses an algorithm can be. Here, the process of curing rubber was itself was new; the fact that it depended on a mathematical formula did not make the process as a whole unpatentable.47 The Benson-Flook-Diehr triad laid a framework for analyzing process claims that relied on math. Unfortunately, that framework was notoriously difficult to interpret; indeed, many commentators said that these three cases were splitting hairs very finely, and others said they were altogether irreconcilable.48 Moreover, the cutting edge of technology had begun expanding beyond macromechanical devices toward computers and other microelectronic arts. Accordingly, the uestions of patentability raised by software-dependent inventions continued to crop up. In 1998, the CAFC, in case of State Street Bank & Trust Co. v. Signature Financial Group,49 apparently relaxed the standards for software patenting, and in so doing, opened the floodgates to a new group of patent applications. Signature Financial Group received a patent for a “hub and spoke” system of financial services, where various funds (spokes) pooled their assets in a central computer system (hub) that tracked and allocated shares and profits

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

according to rules established by the Internal Revenue Service (the US government agency for taxation) for how to avoid taxes on a partnership. While the claim was carefully limited to a machine, the only novel element was the method of arranging finances. Signature then sued State Street Bank for infringement and won at trial. On appeal, State Street argued that the invention should not have been patented because software is, in essence, a “business method.” In a surprising move, rather than decide whether or not software was a “business method,” the CAFC took the opportunity to disagree with the underlying logic of Hotel Security, and concluded there is no “business methods exception” to the Supreme Court’s blanket statement in Chakrabarty that “anything under the sun made by man is patentable.” Rather, the CAFC concluded, any novel, nonobvious method of doing anything that produces a “useful, concrete, and tangible result” is patentable. That conclusion necessarily includes methods of doing business and software. The State Street case opened the way for patent applications unabashedly claiming business methods. From 2002 to 2009, filings rose from 7400 to over 15,000 per year, with issued patents rising from about 500 to over 1700 per year.50 This influx of newly issued patents naturally led to several high-profile efforts to enforce them. One of these efforts involved a lawsuit by Amazon. com against Barnes and Noble over the “1-click” online-shopping patent.51 Another case involved a lawsuit between RTP, which owned a portfolio of patents on methods relating to wireless e-mail and RF antennas, and a company called RIM, which owns and operates the then-wildly popular BlackBerry smart phone. Although that case settled, an injunction issued in the lawsuit threatened to shut down BlackBerry services nationwide.52 While representing a tiny fraction of all patents, business-method and software patents had grabbed national attention. Then came the Bilski case. In 1997, Mr. Bilski and Mr. Warsaw filed a patent application claiming a method for managing risks associated with commodity trading. The application merely disclosed the mathematical and theoretical means for addressing the problem, without any associated machine or tangible transformation of any physical item. Accordingly, the USPTO rejected the application as not drawn to statutory subject matter. The inventors appealed to the CAFC.53 In a complex decision,54 the CAFC, looking at the prior Supreme Court cases, held that to be patentable subject matter, a process must pass the so-called “machine or transformation” test: A claimed process is patent-eligible if “(1) it is tied to a particular machine or apparatus or (2) it transforms a particular article into a different state or thing.”55 The CAFC carefully did not overrule State Street,56 indicating that there is

459

no “business method exception” and no broad exclusion of software from patent eligibility. Nonetheless, the CAFC indicated that any such invention would face the method-or-transformation test as the sole test for eligibility. On appeal, the Supreme Court unanimously affirmed the conclusion of the CAFC’s opinion, but rejected the logic.57 Indeed, the Supreme Court disagreed with the CAFC’s assertion that prior Supreme Court cases required that the machine-or-transformation test be the only test. While that test might be a useful construct in many cases, and while only rarely will a patenteligible invention fail that test, the Supreme Court explicitly repudiated the idea that the test should be the sole means for deciding. The Supreme Court also took the opportunity to reject the State Street standard of “useful, concrete, and tangible result” test as overbroad (though the Supreme Court did not overrule the case altogether). Rather, the Supreme Court elected to decide the Bilski case narrowly, on its facts, without setting forth a new test or other bright-line rule. Since then, the USPTO has issued guidance suggesting that while the legal foundation of any rejection of business method and software patents must rest on the BensonFlook-Diehr triad, any invention that clearly passes the “machine-or-transformation” test will be patentable.58 To sum up this line of cases, the range of patentable subject matter remains broadly open: any method can be patentable subject matter, regardless of the field in which it arises, and regardless of whether it is tied to a tangible object, as long as the method as a whole is something “made by man,” and also provided it otherwise meets the criteria of patentability. Unfortunately, no bright-line rule or clear test currently exists to determine whether or not a given process fits this description. For the present, the Benson-Flook-Diehr triad continues to be the core framework for analysis, though the “machine-or-transformation” test from In re Bilski may be used as a preliminary means for evaluation. Apparently, the Supreme Court anticipates continued evolution of this area of law, so more cases are likely in the near future. “Utility” (“Industrial Applicability”) An often overlooked requirement for obtaining a patent is that the invention must have a use. This “utility” requirement, derived directly from the Patent and Copyright clause of the US Constitution (“to promote the progress of science and the useful arts”), is embodied in the patent statute at 35 U.S.C. x 101. Abroad, other nations have a similar requirement, which is usually called “industrial applicability.” The utility requirement is often overlooked because it is so easy to overcome. The Supreme Court has described the utility requirement as including “anything

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

460

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

under the sun that is made by man.”59 At the same time, a minimum threshold does exist. To pass muster, an invention must have a “specific and substantial” use that is “credible” to a person having ordinary skill in the field of the invention. “Specific and substantial” means that the invention must be useable in a particular, “real world” way, other than in a way that just about anything could also serve regardless of its individual properties.60 For example, any heavy object can serve as a “doorstop” or “landfill,” regardless of its shape or structural engineering, but not every object can be an efficient, comfortable, and affordable modular desk. Likewise, any manufactured edible ingredient can be a “food additive” regardless of what it does once eaten, but not every such ingredient can abate a vitamin-E deficiency or reduce uptake of dietary cholesterol. A pet rock, regardless how much a person might enjoy having one, does nothing. “Credible” basically means that the invention is capable of doing what is promised.61 For example, a perpetual-motion machine cannot function as promised, because doing so would violate a fundamental law of thermodynamics. Likewise, no drug can be a “cure for cancer,” because “cure” entails a range, permanence, and completeness that is impossible given the very nature and variations of cancer. At the same time, because inventions inherently push the boundaries of what society believes is possible, patent applications are rarely rejected on this basis unless the inherent impossibility of the use is readily apparent. Indeed, even if the invention only works crudely or in part, that is enough.62 Note, however, that the “utility” standard is not meant to stand in for other issues, particularly market-based ones. For example, an invention that serves an immoral, or even illegal, purpose is still “useful” as far as patent law is concerned (the law and social norms restricting the use may change).63 Also, an invention may have suitable utility even if it is unsafe: as an example, many patents have been issued for explosives and firearms. Moreover, other agencies are responsible for ensuring safety of products on the marketdfor instance, the FDA must approve of a new drug before it may be marketed in the United States. Finally, in contrast to scientific publications, the utility of an invention need not offer any improvement at all over existing products, or otherwise provide a public benefitdlet alone represent a breakthrough. Markets will decide whether an invention is worthy of being purchased. To show this, consider two patents granted since 2000: The first is US Patent 6,321,753, issued November 27, 2001, which covers a “tanning restraint apparatus,” an adjustable tube designed to orient the

feet of a sunbather properly by holding the big toes together. The second is US Patent 6,905,430, issued June 14, 2005, which covers an engineered “water skipping” stone, designed to skip along a curved path when thrown across a body of water. No inventor won a scientific prize for either invention and neither product became a market sensation, but these ignoble outcomes have no bearing on utility. “Novelty” General Principles “Novelty” in patent law is at once both a simple, bright-line test, and a highly abstract concept. In essence, “novelty” is about whether an invention in its entirety was previously made available to the public in a single disclosure. While that concept may appear simple, the actual rules for conducting the test are far from it. Under the 1952 patent law, the test for novelty compares two dates, one relating to the invention and the other to the reference date of the prior disclosure; if the reference date is earlier than the invention date, the invention is not patentable. The tougher question is figuring out what the dates are. Under the Patent Reform Act, the test is, essentially, whether the inventor filed a patent application before anyone else disclosed the invention. All of the standards may be found in 35 U.S.C. x 102. The first test of novelty is known generally as “anticipation,” and it looks to the date on which the inventor conceived of the invention. In essence, the idea here is that you may not claim to be an inventor if someone else disclosed the invention before you thought of it. The fact that you never actually saw that prior disclosure does not matter, as long as you could have found it had you tried harder. Under the 1952 patent law, two types of prior disclosures apply here: (1) the invention is described in a patent or other printed publication anywhere in the world or (2) the invention was “known or used by others” in the United States. To be “known or used” basically means any nonsecret application of the invention that can be readily discerned by members of the public, should they try to do so.64 Most of the time, “anticipation” arguments arise in a straightforward manner: an applicant files a patent application, and the USPTO rejects it on the basis of an earlier publication or US patent. This situation is typically labeled by its statutory source: a “x 102(a)” anticipation. The 1952 statute has two other, relatively minor forms of anticipation. The first is if the applicant is not the true inventor (x 102(f)). The USPTO has no way to know an applicant is not the true inventor unless the applicant says so, but infringers often find out when sued, leading to embarrassing (and often costly) developments for the shamed patentee. The other form of

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

anticipation arises when someone else has filed a patent applicationdwhich is initially secretdbefore an applicant invented, and the earlier patent application is discovered later (x 102(e)). Despite complexity, this pattern of events happens, though infrequently. The second test of novelty is known generally as “statutory bar.” The idea here is that inventors may lose their rights to a patent if they do not move fast enough to file the patent application. For most of the world, the moment an invention has been publicly disclosed by anyone anywhere, patent rights are lost. Nations that use this standard are called “absolute novelty” jurisdictions. In the United States, patent law is more forgiving: an inventor has a 1-year grace period after a public disclosure in which to file a patent application. A few other nations also provide a limited grace period though the exact term and rules vary.65 Like anticipation, statutory bar under the 1952 law includes two types of disclosures that matter, but their description is subtly different from anticipation-based disclosures: (1) the invention is described in a patent or other printed publication anywhere in the world or (2) the invention was “in public use or on sale” in the United States. The “in public use” prong requires two factual subelementsdthe invention was accessible or discernable by the public, and the use must have had a commercial character, even if technically not a “sale.”66 The existence or absence of a confidentiality agreement is evidence, but not dispositive proof, of whether a use was a public use.67 The “on sale” prong can be satisfied, based on specific facts, by any bona fide sale, offer to sell, license, or offer to license (even if done in secret), provided the technology is sufficiently developed that it is ready for patenting at the time.68 Assignment of the invention rights, however, does not trigger the “on sale” bar.69 Also like anticipation, statutory bar under the 1952 law has some less common forms. The first, and most straightforward, is that the inventor may not secure a patent if the inventor has already abandoned the technology (x 102(c)). As far as the USPTO is concerned, “abandonment” means intentional abandonment; while it is not necessary for the inventor to explicitly state an intention, delay alone is not sufficient to infer abandonment.70 The other form of statutory bar is where the applicant first files an application in a patent office outside the United States and waits more than 12 months to convert the application to a US application (x 102(d)). Under the Patent Reform Act, the concepts of anticipation and bar continue but under a new standard called “first inventor to file.” Essentially, for any patent application filed on or after the first public disclosure of the invention by someone other than the inventor, the application will be denied for anticipation, and an application filed on or after the one-year anniversary of the inventor’s own public disclosure of the invention (or by

461

someone who received the invention from the inventor), the application will be denied because of statutory bar. The Patent Reform Act applies these new rules to patent applications filed on or after March 16, 2013, but the 1952 Patent Act rules apply to those filed before then. Competing Claims of First-To-Invent: The “Interference” Scientists race against each other all the time to be the first to discover things. Sooner or later, cases are bound to arise where two (or more) people claim to have the right to own the patent on an invention because each thinks he invented it first. In most of the world, this argument is settled by looking at whose patent application was received by the patent office firstd known as a “first to file” system. Under the 1952 law, the United States used a “first to invent” system: in theory, whoever can prove the earlier date of invention wins the patent. This concept was captured in 35 U.S.C. x 102(g). Under this structure, the USPTO approaches contests by initiating a specialized and exceptionally complex procedure called an “interference.” An interference can be declared between two pending applications or between an application and an issued patent.71 Normally, either the applicant or the USPTO’s patent examiner notices overlap with some other inventor’s case, but in theory, the USPTO could declare an interference based on information provided by a third party (no third party may participate after an interference has been declared). Procedurally, interferences progress through several stages. First, the USPTO decides the essence of the invention(s) being contested and identifies which inventor is the “senior” party, who enjoys the presumption of having invented first (the other is the “junior” party); next, the parties engage in litigation-style discovery of each other’s documents, conduct depositions, and so forth; then the parties file and argue various motions to try to shape the actual interference hearing (exact description of the invention, scope of evidence permitted, etc.). A panel of administrative patent judges then holds a formal interference hearing; their decision may be appealed to the USPTO Board of Patent Appeals, then to the CAFC, and perhaps the Supreme Court. Instead of appealing to the CAFC, however, the losing party may alternatively file a complaint in the federal district court against the USPTO, asking the court to order the USPTO to issue the patent according to the complaint.72 At each stage, a ruling could reset the procedural clock, perhaps all the way to the beginning. Interferences are complicated and expensive undertakingsdand, fortunately, rare. For example, the USPTO declared 66 in FY2008 and 55 in FY2009, or about 0.02% of the nearly 400,000 applications that the USPTO disposed of in that time.73 Most resolve within 2 years, but complex cases concerning valuable technology can last up to a decade or sometimes even

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

462

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

longer.74 Companies keep their exact costs confidential, but very rough and anecdotal estimates suggest that interferences can easily run several hundred thousand dollars, perhaps into the millions of dollars for the worst cases. One of the major reforms of the Patent Reform Act was to eradicate interferences. For applications filed on or after March 16, 2013, the only basis for a latecomer to challenge an already-filed application is to assert that the applicant merely “derived” (i.e., learned of) the invention from the true inventor. Interferences will not disappear quickly, however, because any issued patent or pending application containing even one claim that was originally in an application filed before March 16, 2013, is still subject to the interference procedures. “Nonobviousness” General Principles While “novelty” looked at whether the entire invention is described in a single reference, what about inventions that are not entirely described in a single reference? This is where the concept of “obviousness” comes to play, which appears in 35 U.S.C. x 103. For neophytes to patent law, the word “obviousness” might be the term most likely to generate confusion because its legal meaning is substantially different from what appears in nonlegal dictionaries. The core concept that the word “obvious”75 tries to capture is the minimum quantum of innovation necessary to justify granting a patent. That minimum quantum cannot be rigidly defined in advance of the creation of an invention; it must be analyzed fresh, case by case, in the context of the specific field of the invention, and at the field’s level of scientific maturity on the date of the invention. The formal test for “obviousness” reads in a simple, formulaic way but is tricky to apply. Essentially, an invention is “obvious” if a person having “ordinary skill” in the field of the invention, looking at all of the pertinent prior references, would find the claimed improvement to be insufficiently different. Generally, obviousness involves comparing the claimed invention against two or more prior references, with each prior reference containing a piece of the claimed invention, and the sum of the elements in the prior references essentially covering the same ground as the claimed invention. A single prior reference may be enough if the difference is truly trivial. The obviousness test gives rise to two concepts: the hypothetical colleague and the hypothetical room. The colleague is an imaginary “average” person having normal, ordinary skill in the “art” (the field relating to the invention)dnot the genius, not the novice. This imaginary colleague stands in an imaginary room of potentially infinite space containing every patent,

publication, item on the market, and so forth. The colleague may look at any item and may combine them as needed. Back in the real world, the USPTO and the applicant (or applicant’s attorney) then argue over what this hypothetical colleague would think. Later, judges and juries may weigh in too. The following are common examples of ways the USPTO may find an invention “obvious:” Combining previously disclosed elements according to known methods to yield predictable results; simple substitution of one known element for another to obtain predictable results; use of known techniques to improve similar devices (methods or products) in the same way; and applying a known technique to a known device (method or product) ready for improvement to yield predictable results.76 Note that the determination of obviousness should not be mere hindsight, as every invention seems both apparent and understandable (“obvious” in the nonpatent sense) in hindsight. Rather, one must engage in a mental exercise of looking at the prior references while pretending that the invention has not yet been made. Likewise, one must resist the temptation to be swayed by the manner in which the inventor came to recognize the invention (painstaking drudgery of data collection, flash of genius, blow to the head, etc.). Also, the range of prior references available for selection and combination is not infinite. Rather, one must have a reason to select a given reference and combine it with another. For example, in trying to figure out how to get fat-soluble drugs to be absorbed by the body, one might look at references that talk about salad dressings made with edible emulsifiers selected to improve the uptake of vitamins, but one would probably not look at references concerning the likely inedible (and possibly toxic) emulsifiers used in the manufacture of plastics. “Secondary Considerations” Even if the USPTO finds two or more references that, when combined, completely disclose the entire invention, that situation does not automatically kill the invention, provided the applicant can provide additional facts to rebut the inference of obviousness. These facts are collectively called “secondary considerations.” Basically, a secondary consideration is any fact that would not be true if the prior references actually made the invention obvious.77 For example, events such as commercial success, efforts by competitors to design around the invention, infringement by competitors, and strong interest shown by others to license the invention, all suggest the invention is not obviousd because, if the invention were really as obvious as the USPTO claims, greed would have long ago pushed these others to get their own versions of the invention

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

on the market. The fact that they had not done so means the invention was harder to discover or make than it looks. Another major secondary consideration is the “unexpected result.” Evidence that the invention has some superior or unexpected quality that transcends the mere combination of prior references is usually sufficient to overcome an obviousness rejection.78 For instance, the combination of two drugs might have a synergistic effect not otherwise suggested in prior trials of each drug alone. That said, someone who merely recognizes a hidden property that always has been present will not be allowed to claim that property as an “unexpected result.”79 Finally, an uncommon but potent secondary consideration is a “negative” reference, i.e., one suggesting that the invention is impractical. The most powerful form of negative reference is where an authority figure in the field asserts that the essence of the invention will not work (this type of reference is said to “teach away” from the invention). The other, more typical form of negative reference is one describing repeated failures to achieve the invention. In both types, the persuasive weight of the negative reference is potent evidence that the invention is truly not obvious. “Obvious to Try” A frequent trap in an obviousness analysis is the question of “obvious to try” for any given invention: Is the “next logical step” patentable? The proposition that obviousness should not be analyzed merely in hindsight bears repeating, as every invention is inherently the next step from that which came before it (and success makes the invention appear logical in hindsight). Even so, sometimes a given improvement truly is obvious precisely because it is indeed the logical progression.80 Consider, for example, a hypothetical new drug, to be taken orally. As is true for many new drugs, this hypothetical one is hydrophobic, and so it is not absorbed well by the body. There are only a few common ways of packaging hydrophobic drugs to improve oral absorption, such as ensconcing the drug in a biodegradable coating, mixing the drug with an excipient designed specifically to enhance absorption, or reducing particle diameter to a few microns or less. Successfully doing one of these strategies would be “obvious to try” because the available choices are so severely constrained. In contrast, redesigning the chemical structure of the drug molecule itself to make it less hydrophobic would not be “obvious to try” because the range of ways in which a molecule can be modified has no clear limitdand the inventor has no assurance that any particular other formulation will retain the functional profile of the original molecule.

463

Written Description, Enablement, and Best Mode As mentioned, the cornerstone of the quid pro quo is full disclosure: in exchange for exclusivity rights, the inventor must divulge everything known about the invention at the time of the application. The statutory source of this requirement can be found in 35 U.S.C. x 112, first paragraph. More specifically, the patent application must include three elements: a description of the invention, the manner of making and using it, and the “best mode” contemplated by the inventor.81 Written Description The CAFC recently confirmed that the simple description of the invention itself is a separate requirement from the more difficult description of how to make and use it.82 The applicant may satisfy this obligation by terms of showing a prototype was built or a process was actually practiced, by terms of indicating that a sample of a biological invention has been deposited in a repository, or by terms that show the invention was ready to be made or practiced exactly as described. The “description of the invention” obligation may seem trivial, and in most cases it should be, but sometimes it can trip up an unwary applicant, particularly for complex and cutting-edge technologies. For example, consider a hypothetical invention of a chemical extract from a plant that suppresses human cell growthdpotentially very useful in treating cancer. If the inventor cannot identify what the extract is, others likely will be unable to tell whether their extracts from other plants are actually the same as the inventor’s. To patent the chemical composition successfully, the inventor must be able to describe it in detaildcomplete or partial structure, other physical/chemical properties (e.g., gas chromatography, spectral characteristics, binding affinity/specificity, molecular weight/length), functional aspects coupled with known functione structure correlations (function alone is not enough83), or some combination of these.84 Enablement Judging by the volume of litigated cases, the task of crafting the text of how to make and use the invention is the most difficult part of satisfying the written description requirement. Specifically, the application must describe the invention in detail, with methods, examples, drawings, and other elements that together demonstrate how to practice the invention. This text must be clear enough that a colleague of ordinary skill could, in fact, get the invention to work without undue experimentation. A patent application with that level of detail is said to be “enabled.” Note that the “enablement” rule does not require that the application’s description be so precise and thorough

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

464

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

to guarantee that the invention will work on the first try, just that it will work without “undue experimentation.” Exactly how much experimentation is “undue” depends on the circumstances. Factors include the range and content of prior references, the level of “ordinary skill” in the field of the invention, the level of predictability in the field, the amount of direction provided by the inventor, and the examples provided.85 Also, as long as the invention as described works at all, the fact that the quality of the final product as described in the patent is so bad that it is not commercially viable does not defeat patentability.86 In the biomedical arena, enablement plays a key role when an invention concerns a new drug or other therapeutic to treat humans. For example, consider a pharmaceutical candidate that has shown successful in both in vitro and in vivo (animal) testsdif the candidate has not yet been tested in humans, does the inventor have enough information to know how to use it to treat a disease? The answer is, “sometimes.” For some diseases, such as bacterial infections, in vitro data may be sufficient: chances are high, it will work on or in a human, such that no further showing is needed for patenting purposes (FDA approval is another matter). For drugs alleviating symptoms, perhaps in vivo data might be necessary. For some psychoactive drugs and for preventative vaccines for certain diseases such as HIV, however, successful tests in animal models have proven a poor predictor of success in humans; the USPTO may reject patents relating to such diseases until the inventor demonstrates the composition actually works in humans. Best Mode The “best mode” requirement of the 1952 patent law was crafted to be a safeguard against an applicant’s desire to obtain patent protection without making a full disclosure as required by the statute. Specifically, the requirement forbids inventors from solely disclosing what they know to be their inferior embodiments, while retaining the best for themselves.87 In prosecuting a patent application, the best mode requirement is an easy threshold: simply say which of the various ways of practicing the invention is the “preferred embodiment” or similar phrasing. As long as the application clearly indicates which of the several modes the applicant says is best, the USPTO will assume the applicant is telling the truth. If only one means of practicing the invention is included in the application, the USPTO will presume it to be the best mode.88 The mere failure to include a method that happens to be better than those in the application is not grounds to invalidate a patentdas long as the inventor, at the time of filing the application either did not know about the better method or did not recognize it as the best.89

Where the best mode requirement comes into play is in litigation. Those accused of infringing a patent, as part of a broader effort to invalidate a patent, often try to convince the judge and/or jury that the patent’s assertion of best mode was false, and that the applicant made that assertion of best mode with the intent to mislead the USPTO and the public. Because of perceived abuses of this doctrine in litigation, a major aspect of patent reform has focused on changing how the best mode requirement can be used in litigation, if not eliminating it entirely. The Patent Reform Act has removed the best mode requirement in all but name. Effective immediately, no patent can be found invalid for failure to satisfy the best mode requirement. Oddly, despite removing all enforcement mechanisms, Congress left in the requirement that the applicant must include the best mode in the application. Perhaps Congress will revisit the matter in a technical correction to the statutory language, or the USPTO will issue clarifying rules, but in the meantime, patent applicants likely will continue the practice of specifically identifying which of the several embodiments listed in any given application is believed to be the best. Other Key Terms Defined One of the major reasons many people find patent law arcane and tedious is that patent laws (and lawyers) rely heavily on formally defined terms of art, odd jargon, stilted phrasing, and other terms having meanings that are several degrees off the way the rest of society uses those same terms. This section of the chapter will examine some of the most significant terms widely used in patent law. “Prior Art” Previous sections of this chapter have included mention of “references” and “disclosures” that the USPTO uses to reject a patent application. Patent law has a formal term to capture all of these references and disclosures: “prior art.” Exactly what qualifies as prior art, however, depends on the circumstanced specifically, the particular subsection of 35 U.S.C. x 102 and x 103. For the rest of this section, only United States law will be considereddfirst looking at the 1952 law’s version, then at how the Patent Reform Act changed matters. • § 102(a)d“Anticipation” For classic “anticipation” under the 1952 version of x 102(a), prior art includes (1) all issued patents, all published patent applications, and all other “printed publications” in any language anywhere in the world; and (2) all instances where the invention was publicly known or used in the United States. These references

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

will defeat patentability if they were made public before the date that the inventor invented. The logic of this is easy to see: they demonstrate that the inventor was not the first to think of the idea. How does this work? If a hypothetical inventor, Sally, conceived of an invention on Thursday January 1, 2011, a scholarly article published by David in a Brazilian journal (in Portuguese) disclosing the same invention would not anticipate Sally’s patent application if published on Wednesday January 7, 2011, but would anticipate it if published on Wednesday December 14, 2010. Also, if George had independently made the same invention in November 2008 and incorporated it into his farm equipment, which all the neighbors could borrow, George’s use would not anticipate if his equipment were located in New Zealand but would anticipate if it were in Montana. Note that this rule applies even though Sallydor even the US public generallydnever knew the prior art in question existed.90 Two items should be noted concerning a “printed publication.” First, the phrase does not necessarily require paper to be involved. Since the 1952 Patent Act, court decisions have clarified that the point of the phrase is to capture any material reasonably accessible by interested members of the public regardless of the medium of expression. Thus, television broadcasts and electronic publications available online qualify.91 As for the “date” of a publication, the key is when a member of the public could have found it through a reasonably diligent search.92 Note also that “publication” does not require distribution; depositing a single copy of a graduate thesis in a university’s library is enough as long as the public reasonably could find it.93 Under the Patent Reform Act, any disclosure of the invention anywhere countsdincluding a description in a patent or printed publication, a public use of the invention anywhere, a product or service embodying the invention that is on sale anywhere, or any other disclosure that is “otherwise available to the public.” If a disclosure of the invention became publicly available before the application was filed, the disclosure qualifies as anticipatory prior art. • § 102(b)d“Bar” Under the 1952 law, the rules for prior art under the “statutory bar” provision of x 102(b) are, with certain exceptions, comparable to those under x 102(a). Prior art under x 102(b) includes (1) all issued patents, all published patent applications, and all other “printed publications” in any language anywhere in the world and (2) all instances where the invention was “in public use or on sale” in the United States. These references will defeat patentability if they were made public more than 1 year before the date that the inventor filed the patent application.

465

The big difference between the scope of prior art under x 102(a) and x 102(b) is that the former involves “known or used” while the latter involves “in public use or on sale.” “Public use” refers to commercial exploitation, even though no actual sale takes place or the event does not occur in full view of the publiceprivate product demonstrations, for example, to prospective investors or customers, are common triggers of the bar, along with allowing third parties to test-run the product in the absence of a duty of confidentiality.94 “On sale” includes bona fide offers to sell, current sales for future delivery, and certain licenses of the patent rightsd though not assignment of the patent rights.95 Finally, as discussed above, experimental uses do not qualify as prior art under x 102(b). As for timing, the statutory bar provision affords a 1year grace period in which to file a patent application. So, consider again the hypothetical timeline used in the discussion on anticipation: if Sally waits 13 months to file her patent application (say, Thursday February 14, 2012), David’s article would bar Sally’s patent application if published on Wednesday January 7, 2011, even though it did not anticipate her invention. If Sally were to stumble on David’s article in June 2011, she would immediately know she has until January 7, 2012, to file her patent application in the United States. Likewise, if George started selling his equipment in New Zealand on January 7, 2011, Sally’s patent is safe, but if George were to sell his equipment in Montana, that sale would bar her February 2012 application. Under the Patent Reform Act, as mentioned above, the 1-year grace period is specifically limited to the inventor’s own disclosures along with disclosures by those who received the invention from the inventor. The new law essentially eliminated the distinction between x 102(a) and x 102(b) (as well as certain rules that created a class of art called “secret prior art”) in favor of any disclosure that is “otherwise available to the public.” A few commentators have wondered whether the new law creates an incentive for the inventor to publish before filing the patent application, because the inventor’s own disclosure bars competitors from filing patent applications, but will not serve as a bar against the inventor’s own application. Such a strategy is likely unwise, however, given that the preapplication disclosure would likely destroy patent rights in most other nations besides the United States. • Collaborationsd§ 102(e, f, & g)/§ 103 Consider a hypothetical case involving three toy companies, A Corp., B Inc., and C, L.L.C. Two of them, A and B, collaborate to design a new toy under a detailed research agreement. A shows B an unfiled patent application for their latest idea. B realizes that a substantial

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

466

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

revision would make the toy much better, so B files a new patent application, naming B as sole inventor (A agrees and decides not to file its own application). B makes the toy, and A and B share royalties. C decides to make a cheap knock-off of B’s toy, and B sues C for patent infringement. C argues that B’s patent should be voided as punishment for B’s having violated the duty to disclose everything to the USPTO, namely, A’s draft patent application. B contends that there was no duty to disclose the application, as it was never available to the public in any form, so it does not fit the definition of “prior art” under either x 102(a) or x 102(b). Roughly, this hypothetical tracks the key facts from the case of OddzOn Products, Inc., v. Just Toys, Inc.96 The trial court held that the patent owned by OddzOn was valid and enforceable, but Just Toys did not infringe it. Both parties appealed, and the CAFC affirmed the district court’s decision on all issues. The CAFC concluded that any failure by OddzOn to supply documents to the USPTO was harmless because the differences between the two versions of the toy were so substantial that the OddzOn application was patentable nonetheless. That result is not what makes this case significant; rather, a small side discussion in the court’s opinion made the case much more important than its holding would have suggested.97 The court began its analysis by noting that the doctrine behind anticipation relies on the premise that the inventor may only get a patent for things the inventor actually made, rather than for things previously made by others. So, hypothetically, if the patent application filed by OddzOn had been identical to the drafts OddzOn says “inspired” the invention, then x 102(f), which bars granting a patent if the applicant did not invent, clearly would have pertained to the OddzOn patent application. Moreover, an applicant has a duty to disclose information to the USPTO that bears on whether or not the applicant is entitled to a patent, which includes x 102(f) references. Likewise (reasoned the CAFC), if the OddzOn application were insubstantially different from the “inspiring” references, then the OddzOn application should have been denied because it would be “obvious” under x 103 in light of those references (continuing the logical theme that OddzOn did not actually “invent” anything). Thus, for purposes of the inventor’s duty to provide information to the USPTO, references the inventor saw before conceiving of the claimed invention that would be relevant under x 102(f) (or, for similar reasons, relevant under x 102(e and/or g)) were “prior art” for x 103 analyses. As a last-nail-in-the-coffin point, the CAFC observed that Congress had explicitly noted that communications between two coinventors are not to be considered “prior art” under x 103 if all coinventors shared a duty to assign

their rights in inventions to a common employer. Because the statute did not extend to employees working for two different employers, OddzOn would not have been able to claim the benefit of that exemption. As a direct result of the OddzOn decision, Congress responded to the concerns of nonprofit research organizations, universities, and companies that depended on interinstitutional collaborations to develop their technologies. So, in 2004, Congress passed the “CREATE Act.” This law explicitly amended 35 U.S.C. x 103 by adding a subsection to exempt x 102(f) art from considerations of obviousness, provided two conditions are met. First, the invention had to have been made during and within the scope of a “joint research agreement,” and second, the patent application includes the names of the parties to that joint research agreement.98 The Patent Reform Act explicitly makes the rules of the CREATE Act applicable to all future patent applications. Going forward, the bottom line is that researchers involved in a patent application must think broadly when asked to collect and report on all prior communications and references. Most researchers will only think about public communications, such as talks, posters, and the likes; few naturally think about casual exchanges with colleagues that, based on friendship, honor, courtesy, or other unwritten understanding, are believed to be confidential. Indeed, some might even overlook exchanges made under a formal nondisclosure agreement. Theoretically, all of these might qualify as “prior art” that must be reported to the USPTO because the failure to do so can risk jeopardizing the validity of the patent. “Conception” Versus “Reduction to Practice” One of the most critical principles in patent law is “conception,” from which several important related terms arise. Unfortunately, “conception” is a metaphysical construct, and as such, difficult to prove or disprove whether and when it occurred. Even so, it can easily make or break a multimillion-dollar lawsuit. “Conception” is defined as “the formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention as it is thereafter to be applied in practice.”99 This idea must include both a present recognition of its inventive features and an appreciation that the idea represents something different from that came beforehand.100 To be sure, however, “conception” does not require that the inventor realize that the invention is, in fact, patentable.101 Closely related to “conception” is the term “reduction to practice.” Reduction to practice is the direct implementation of the invention: building a working prototype, synthesizing a chemical, physically carrying out a method, etc. The inventor does not have to reduce the invention to practice personally; instructing someone

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

else do it on behalf of the inventor is sufficient.102 Also, whether or not the invention is commercially viable does not matter. The legal significance of these closely related terms lies mainly in their implications. “Conception” is critical because “inventorship” (discussed in the following section) turns on conception, and ownership turns on inventorship. “Reduction to practice” is critical because it provides a factual basis for enablement; thus, reduction to practice is a nominal prerequisite to filing a patent application under x 112(a). Also, reduction to practice often has a major impact on the outcome of interferences. So what happens if someone conceived of an invention, but neither physically built a prototype nor otherwise actually reduced the invention to practice? There is another, intangible way to get to the same point: the inventor can describe how to make and use the invention in such precise detail that one of ordinary skill in the art can do it (without “undue experimentation” to get it to work, of course). Imagine the first person to conceive of a clock. If that person put on paper instructions and drawings so clear that someone could manufacture the right gears and other components to build one and has step-by-step procedures for putting it together, then whether or not the inventor ever built one personally does not matter. This sort of description is called “constructive reduction to practice.” Indeed, the patent application itself, assuming it includes enough instructions and data, constitutes constructive reduction to practice.103 Note that “constructive reduction to practice” satisfies the quid pro quo of patent philosophy: the inventor has fully taught the public how to make and use the invention, even if the inventor never made or used the invention personally. The public can learn from any patent, design around it, and when the patent expires, freely practice the invention. What if the description turns out to be faulty? The risk is mostly carried by the patentee. If the invention works but the description is substantially incomplete or flawed, that fact is relatively easy to prove at trial, so the patent is vulnerable to being declared void. If the description fails to work because the invention itself does not function, the patentee is the only one who suffers for it, not the public, because the patentee spent a lot of money on a worthless patent (who could possibly infringe an invention that does not work?). “Prophetic Conception” Versus “Simultaneous Conception and Reduction to Practice” When someone envisions a potential invention but does not yet have experimental data, at what point has that person “conceived” for purposes of patenting? Answering this question requires exploring a complex

467

nook of the law, bracketed by two concepts: On one side, “prophetic conception” is enough, while on the other, “simultaneous conception and reduction to practice” is the minimum requirement. Drawing the line between them is extraordinarily difficult, as the following discussion shows. “Prophetic conception” was the subject of a court case concerning the anti-HIV drug, AZT. On the discovery of the HIV virus in the mid-1980s, the NCI called on pharmaceutical companies to look through their inventory for candidate drugs, which the NCI would screen for possible anti-HIV activity. BurroughseWellcome (“BW”) joined those that answered the call, providing eleven compounds to NCI, one of which was AZT. Originally, AZT had been made in the 1960s, tested first as an anticancer agent, and then in the 1970s as a broadspectrum antiviral, but the drug was unsuccessful at both. When NCI’s preliminary screens showed positive results, NCI jump-started the clinical research; by 1987, the compound was ready for the market. Meanwhile, BW applied for and received a patent on the use of AZT to treat HIV. BW initially offered to sell the drug at prices up to $10,000 per patient per year, so two generic drug makers raced to make cheaper versions, which prompted a patent infringement suit. In the courts, the generics argued that BW’s patents were invalid because BW was not the true inventor; NCI was at least was a joint inventor. After all, they argued, BW never tested the drug on live HIV virus in any form, and without that data, BW could not know whether the invention would work. The CAFC, however, sided with BW.104 Although actual reduction is sometimes the only way to be sure that an invention will work in fact, in this case, BW had shown it had a clear idea of the invention, with “reasonable expectation of success.” This conclusion was based mainly on the fact that BW, before sending samples of AZT to NCI, had drafted a patent application with detailed instructions on how to make and use AZT to treat HIV, which turned out to be correct. The accurate detail in the draft patent application reflected the fact that BW had more than a mere hope of success; NCI’s contribution in this instance was simply the actual reduction to practice that inured to the benefit of BW.105 The other concept, “simultaneous conception and reduction to practice,” is highlighted by the race to perfect a prospective vaccine against Hepatitis B. In the following summary, some of the facts are disputed and others omitted for brevity, but it is close enough to elucidate the concept. In the late 1970s, scientists were still trying to coax bacteria to generate human proteins by introducing DNA. At the time, no one really understood why the process sometimes worked but usually failed. Shortly

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

468

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

after Dr. Heitzmann induced a yeast cell to make the human interferon protein in 1979, he decided to turn his attention to making vaccines. In particular, he wanted to find a way to produce a protein that normally sits attached to the surface of the Hepatitis B virus. He described his detailed plans in his lab notebook in February 1981, but he was delayed in showing they worked until July. Meanwhile, starting in March, his competitor, Dr. Rutter, planned very similar experiments but was able to complete the experiments in June. The USPTO declared an interference on the ensuing patent applications, which finally reached the CAFC in 2001. On appeal, the CAFC decided that Dr. Rutter had won the race to conceive, not Dr. Heitzmann.106 In contrast to the BW case, Dr. Heitzmann indeed merely hoped his experiments would work, despite all the prophetic detail in his lab notebooks. The CAFC carefully pointed out that conception must be viewed from the perspective of the day. Back in 1981, the posttranslational modification machinery of eukaryotic cells had not yet been discovered, so no one knew why yeasts worked and bacteria did not. For all Dr. Heitzmann knew, yeast could have been equally as frustrating a system as bacteria. In this instance, conception (the moment of a reasonable expectation of success) did not occur until the moment of actual reduction to practice.107 “Inventorship” and “Joint Inventorship” In the United States, determining who invented an invention is critical because the inventor owns the invention (at least initially). The 1952 patent law stated that, with very limited exceptions, only the inventor may apply for a patent.108 The inventor may, of course, transfer ownership of the invention to anotherdfor example, the inventor’s employerdexactly like any other property, but this extra step is cumbersome, particularly for large research organizations. Also, other nations do not necessarily follow this pattern, leading to inconsistent practices and confusion. The Patent Reform Act simplified the process by which a party other than the inventor can apply for a patent simply by demonstrating that the inventor has a duty to assign the invention to the applicant. Even so, inventorship remains the determining factor of initial ownership of the invention. Joint inventorship is well described as “one of muddiest concepts in the muddy metaphysics of patent law.”109 For two or more individuals to be joint inventors, each must have contributed to the inventive concept. The patent statute does not require that the inventors physically worked together or at the same time that each contributed in the same way or amount, or that each contributed to every element of the invention.110 Rather, to be a coinventor, each must contribute in some manner to the conception or reduction to

practice of the invention, and that contribution is not trivial when measured against the full invention.111 Some situations are clearly not enough to support “joint inventorship.” First, one who merely supervises work or suggests an idea of a result to be accomplished, rather than a specific means of accomplishing it, is not a joint inventor.112 Likewise, someone who merely suggests adding something that was entirely already in the prior art does not gain joint inventor status.113 Also, one who mechanically transfers key research materials to the inventor or who simply carries out the instructions of the inventor does not become a joint inventor despite the person’s substantial physical contribution to the research.114 All of these rules apply even where these colleagues of the inventor have duly earned coauthorship on the scientific paper disclosing the invention. In short, coauthorship is not coinventorship. Under the 1952 law, incorrect naming of inventors is easy to fix at any time as long as the error was not made with deceptive intent. If a patentee, however, knowingly and with the intent to deceive, incorrectly identified the inventorsdwhether by adding someone who should not be included or omitting someone who shouldda court can rule that the patent is unenforceable by the patentee or even invalid altogether. The Patent Reform Act removed the “deceptive intent” standard from the statute, though it kept in the obligation to name the inventors. The significance of this change is that a patent can no longer be held void merely for an incorrect listing of inventors, even with clear intent to deceive (other remedies, however, may still apply). Inventorship status is a legal determination. Consequently, when a dispute arises concerning whether or not someone qualifies as a joint inventor, it must be resolved with the help of qualified patent attorney (or, at worst, by a court). It cannot be “settled” through quiet compromise; patent law has no sympathy for bruised egos. Of course, any court will be reluctant to take any drastic actions over an easily corrected error, but especially under the law before the Patent Reform Act, a defendant could easily raise a large cloud of smoke at trial when a patentee was less than diligent during the application phase. One of the typical ways this situation plays out will be discussed in more detail below in the context of litigation. Transfers of Ownership: “Assignment” Versus “License” Two additional terms should be briefly mentioned here, mainly because they are often confused and misused by those not familiar with US property law principles. They are “assignment” and “license.” An “assignment” is a transfer of title to a property, whether or not in exchange for value. Once title is transferred, the recipient (“assignee”) becomes sole

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

owner. Assignments should never be treated lightly because the assignee can even enforce the patent against an inventor. This principle, on reflection, should come as no surprise: Few would think that after they sell their house they retain the perpetual right to reenter that house at any time. Unfortunately, some scientists unfamiliar with US patent laws carry the belief that they can continue to have a say in developing “their” invention even after assigning away their rights. Even among those scientists who know what “assignment” means, too many do not bother to read the documents others give them to sign and end up inadvertently giving away their rights. A “license” is, at its heart, simply an agreement not to sue for infringement. Patent licensing is a complex field deserving its own textbook,115 but as a general statement, a patentee has broad latitude in how to structure licenses. Licenses come in all levels of complexity: they can be formal or informal, written or implied, contingent or immediate, optioned, exclusive or coexclusive or nonexclusive, royalty-bearing or royalty-free, limited in scope, limited in time, or limited by a vast range of terms and conditions. Complex licenses can subdivide technologies by field of use, such that one field (say, therapeutic applications) has exclusive rights while another (say, diagnostics) will be nonexclusive. As a general proposition, however, while a license does confer certain rights, it does not transfer direct ownership.116 Patent Infringement (United States) By itself, a patent is little more than a document with a pretty ribbon on it. What makes a patent worth the time, expense, and trouble to obtain it is the fact that a patent offers its owner the opportunity to stop others from practicing the invention claimed in the patent in the country that issued the patent. Often, to realize this opportunity, a patentee simply needs to bring the patent to the attention of an alleged infringer, who then either signs a license or stops infringing. Sometimes, however, the alleged infringer does not cooperate, so the patentee must ask a court to enforce the patent. Civil Liability: In General As mentioned above, the owner of a United States patent has the right to enforce the patent against any party that makes, uses, sells, or imports the claimed invention in the United States. This is called “direct” liability, and it is a “civil” wrong, which means it is purely a matter between private parties (in contrast, “criminal” wrongs are those that carry the possibility of incarceration and/or fines, which are enforced exclusively by the government). To enforce a patent, the patentee (“plaintiff”) must file a lawsuit in federal district court, and prove that the accused (“defendant”)

469

actually made, actually used, actually sold, or actually imported the invention somewhere within the country. The patentee does not have to prove that the defendant was even aware the patent existed, let alone intended to infringe it, though such evidence can lead to enhanced damages. A product or service can only infringe an issued patent, not a pending application (while pending, no one knows the exact scope of rights). For pending applications that have already been published, however, a patentee has the right to receive a reasonable royalty from infringers after the infringers have actual knowledge of the pending applications (and only to the extent the infringer’s product or service is “substantially identical” to the invention described in the application).117 Owners of pending and issued patents are not under a formal duty to mark their products with notice of pending or issued patents, but if they choose to do so there are strict rules for it, and the failure to mark may result in a substantial reduction of the damages the patentee can collect.118 Co-ownership presents a major complication in this neat picture. Not surprisingly, each co-owner of a patent cannot be held liable for infringing the patent, and each co-owner can license that freedom from being charged with infringement to others. Less intuitively, co-owners do not owe each other any duty to account for profits or royalties they collect by virtue of their share of ownership (unless they negotiate an agreement otherwise). This twist means that it is effectively meaningless for a co-owner to say, “I co-own 50% of a patent”dor 1% or 99% for that matterdbecause the co-owner has the exact same right to be free from infringing the patent regardless of the percentage. Co-ownership is directly tied to joint inventorship: poorly confirming who the inventors are can easily ruin a major business strategy. In 1985, the USPTO issued a patent with 55 claims to Dr. Yoon on a surgical device. Dr. Yoon granted an exclusive license to Ethicon, Inc., which in turn sued its competitor, US Surgical Corp., for infringement. In preparing for trial, US Surgical discovered that Dr. Yoon had collaborated with Dr. Choi in developing the device and arranged for Dr. Choi to grant US Surgical a retroactive license. At trial, the court concluded that Dr. Choi indeed contributed to 2 of the 55 claims and so should have been named as a joint inventor. Once Dr. Choi became recognized as a coinventor, Ethicon’s litigation had to be dismissed.119 Civil Liability: Contributory and Induced Infringement Even if someone does not directly infringe a patent, there are two other ways that person can be found liable. These ways are “induced infringement” and “contributory infringement.” In both types, the defendant will

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

470

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

be equally liable for damages as if the defendant were the party that directly infringed the patent. These types of lawsuits are particularly useful when the direct infringers are the consumers of the plaintiff’s patented product. Induced infringement occurs where one party takes a “positive act” to spur another to infringe a patent.120 A positive act could be teaching others how to infringe, directing or commanding someone to infringe, or advising someone to do something that infringes a patent.121 Also, exporting all of the components of a kit (which, if assembled in the United States, would infringe a patent) constitutes induced infringement.122 Contributory infringement (sometimes inaccurately called “indirect infringement”) occurs where someone sells, offers for sale, or imports a component of a larger patented product, such that even though the component itself does not infringe, it is specifically designed or adapted for use in the infringing product and has no other substantial noninfringing use.123 By necessary implication of the adaptation requirement, “contributory infringement” requires a showing that the contributory infringer actually knew of the patent.124 Major Defenses When hit with a lawsuit claiming patent infringement, the first line of defense always is that the accused product or service does not infringe the patent. In essence, the defendant must prove that the accused product or service lacks at least one limitation in each of the claims of the patent. This process often requires the court to interpret the meaning of words and phrases appearing in the claims. The second major defense is to assert that the patent should not have issued because the USPTO’s Examiner failed to find a key piece of prior art, erred, or similar reasoning.125 The patent carries a presumption that it is valid, which the defendant must overcome with clear and convincing evidence.126 Where the nature of the defense is that the USPTO’s substantive interpretation of the prior art was flawed, courts are reluctant to substitute their judgment for the USPTO,127 but sometimes this approach is successful. The third major defense is a series of “equitable” argumentsdi.e., those appealing to a sense of general fair play. These include “inequitable conduct” before the USPTO (discussed later), “laches” (the plaintiff, who could have sued anytime, waited years until the defendant had racked up extra damages),128 and “estoppel” (the patent applicant made certain statements to the USPTO about the limits on the scope of the claims in this patent, so the patentee should be bound by those limits). One other important equitable defense is “patent misuse:” the patentee has tried to use the patent as a

lever to gain exclusive control over more than the patent actually covers. For example, the patentee might require customers to buy unpatented Product Y as a condition to getting patented Product X, where Product Y is a staple good that has nothing to do with Product X. Alternatively, the patentee might require its prospective licensees to sign a contract extending royalty payments and/or forbidding competition beyond the expiration date of the patent. Where any of the equitable defenses succeed, courts have great latitude in fashioning a remedy to fix the problem. Specific Exemptions and Immunities Research-Use Exemption: Madey v. Duke University Since at least 1861,129 and certainly since 1935,130 courts have declined to find that universities infringe patents when the universities’ activities are confined to research. This exception to infringement is not explicitly in the US patent statute, rather judges inferred its existence based on the underlying philosophy of the patent system itself: research on a patented technology is critical to verify that it works, to learn how to design around the claims, and to enhance the body of public knowledge. By this logic, only uses by which the alleged infringer sought to profit or to harm the patentee should be deemed “infringement.”131 In the late 1980s, Dr. John Madey left Stanford University to join Duke University, bringing with him a laserbased machine on which Madey personally owned a patent. Later, Madey had a falling out with Duke; when Madey left, however, he did not take his device with him. Madey sued Duke for several reasons, one of which concerned Duke’s continued use of Madey’s patented device. The trial court dismissed that claim, but the CAFC reversed the dismissal,132 ruling (among other things) that: [U]niversities, despite their non-profit tax status, can and do operate in a for-profit manner, for example in competing for endowments/grants and in licensing future inventions; and the historic, common-law “research” exemption from infringement is limited to acts done for pure idle curiosity.133

The Madey decision, and this quoted passage in particular, served to call into question the generally held view that a “university infringement shield” existed134 and gave impetus to an effort calling on Congress to pass a statutory research-use exemption. While the CAFC made clear that activities by universities could constitute patent infringement, the court did not provide clear guidance on what might be acceptable, noninfringing activities. This uncertainty and ambiguity as to what constitutes patent infringement may have caused researchers greater apprehension when practicing their craft. Therefore, rather

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

than stimulating the innovative activities that the patent system is intended to foster, this decision may have had the opposite effect. Although as a practical matter the Madey case may be more an exception than rule (infringement suits against universities are extremely rare), the loss of the common perception of academic insulation from patent infringement has alarmed many in the academic and public sectors. In addition, many fear that even truly noncommercial research will now require “due diligence”135 to avoid patent infringement or at least enhanced damages. Generic Drugs: The “Bolar Amendment” and Merck vs. Integra Developing a new drug from a mere test molecule into a dose-controlled pill or injection that a doctor may lawfully prescribe requires the investment of years of work and hundreds of millions of dollars. Pioneer drug developers, therefore, have a powerful incentive to maximize their exclusive position against later copiers. What rights does a patentee have against a prospective competitor who is merely preparing to sell copies? In the early 1980s, Roche was nearing the end of its patent on valium, and a competitor, Bolar Pharmaceuticals, started work developing a generic copy using some of Roche’s drug to verify bioequivalence. Even though Bolar did not try to sell anything before Roche’s patent expired, Roche sued for patent infringement anyway. The Federal Circuit held that Bolar’s research on making a generic version was not an “experimental use” because Bolar intended to sell the copy thus made.136 Congress, which had already been working on an overhaul of the process by which the FDA approved drugs (both pioneer and generic), took the opportunity to restructure the corner of the patent system that overlapped with the FDA’s jurisdiction. This new statute, called the Hatch-Waxman Act, facilitated entry of generic versions of patented drugs into the marketplace as soon as possible after patent expiration through the use of the Abbreviated New Drug Application. The law included a limited exemption to patent infringement for certain uses of the invention as long as the activity is “reasonably related” to, among other things, an FDA submission.137 This limited exception is known as the “Bolar Amendment.” The phrase “reasonably related” came into sharper focus in the mid-2000s with the Supreme Court case of Merck KGaA v. Integra Life Sciences, Inc.138 Integra co-owned five patents concerning a peptide that promotes cell adhesion by attaching to certain receptors on the outside of specific endothelial cells. Merck used these peptides in its research on angiogenesis inhibitors and with it, discovered a drug that blocks the peptide. As Merck’s drug reached the market, Integra sued for

471

infringement. The Supreme Court unanimously decided that Merck’s use of the patented peptide could be covered by the Bolar Amendment, even though Merck’s use of the peptide was not intended to develop a generic version of the peptide. The court reasoned that the potentially infringing use was “reasonably related” to an FDA submission because the resulting data might eventually be part of a later IND, IDE, NDA, BLA, 510(k), or PMA. The Medical Practitioner Exemption (“Frist-Ganske Amendment”) One specific defense available in the medical field is the “medical practitioner” exemption.139 If a medical practitioner infringes a patent by performing a medical or surgical procedure on a body140 neither the practitioner nor any “related health care entity” (e.g., hospital, clinic) can be held liable for that infringement. This exemption does not, however, protect medical practitioners to the extent they work for an organization (including a nonprofit one) that designs, develops, manufactures, and sells products intended for the practice of medicine by the medical community.141 US Government as Infringer The United States Government has its own special defense against patent infringement liability.142 A patentee’s only recourse is to sue the government in the US Court of Claims in Washington D.C., and the patentee’s remedy is limited to a “reasonable royalty.” The patentee cannot, under any circumstances, be awarded “actual” damages, augmentation of damages, injunctions, costs, or attorneys’ fees. This defense extends to infringement by contractors and agents acting on behalf of the government. While a few suits are filed every year against the government, this exemption presumably dissuades many others from bothering, though it is impossible to gauge how many. Remedies: Types and Measures If the patentee wins on liability, the patentee has the right to receive money (“damages”), which may be increased by up to a factor of three as punishment for intentional infringement.143 In exceptional cases, the court may require the loser to pay the winner’s attorney’s fees.144 Also, for extraordinary cases, the patentee may ask the court to issue an order (an “injunction”) forbidding the defendant from further infringement and perhaps ordering the destruction of any existing infringing inventory.145 Calculating damages is a notoriously complex topic. Briefly, the least a successful patentee can expect to receive is an amount “adequate to compensate for the infringement, but in no event less than a reasonable royalty.” A “reasonable” royalty is what the court thinks would have been accepted in an unemotional

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

472

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

negotiation between independent parties on equal footing (known as an “arm’s length” transaction).146 If the invention conveniently arose in an industry where such licenses are common, the court can use the industry average as a benchmark; otherwise, each party will bring in experts to offer opinions as to the technology’s commercial value. Alternatively, a patentee can try to prove actual injury, such as lost profits, lost sales, lost market opportunity, and/or disgorgement of the defendant’s profits.147 “Declaratory Judgment” Actions Under the Constitution, courts have power to decide “cases and controversies,” which means they avoid issuing any decision on hypothetical questions and on disputes that have not yet matured into a present, bona fide controversy. Consequently, parties who wanted to challenge the validity of a patent had to wait until the patentee sued for infringement, and then was stuck litigating wherever the patentee chose to file suit.148 Since the 1920s, Congress has authorized courts in limited circumstances to grant “declaratory judgments” on issues that are on the cusp of being ripe controversies.149 One of the circumstances that permit someone to file a Declaratory Judgement (“DJ”) suit included is potential patent infringement. Anyone who faces the reasonable apprehension of being sued for infringement may file a request for a declaration as to whether or notd(1) the patent is valid, and if so, (2) the plaintiff’s good/service infringes the defendant’s patent.150 Notice that in a DJ lawsuit, the plaintiff is the alleged infringer, the defendant is the patentee. This means the alleged infringer has an opportunity to select a venue before the patentee does so. Of course, this strategy is not risk-free: once someone files a DJ lawsuit, the patentee is almost always effectively forced to countersue for patent infringement, even where the patentee might not have bothered to do so otherwise. Importation and the International Trade Commission One of the four basic rights of a patent is to exclude others from importing goods into the United States that infringe a US patent. Obviously, a patentee may try to sue an infringer in the US district court for importation in the same way as for making, using, and selling, but the power of US courts ends at the border, so as long as foreign infringers do not come to the United States or otherwise maintain assets here, the courts are effectively not available. Fortunately for patentees, another venue is available to enforce their patents: the US International Trade Commission (“ITC”). In parallel to US patent law, section 337 of the Tariff Act of 1930151 makes illegal the importation of any

article that infringes a patent, including items made abroad using a process patented in the US and authorizes the ITC to investigate allegations, provide a patentee temporary relief while the case is underway, determine whether or not infringement has occurred, and if so, order that the US Customs Office exclude the offending goods from entering any US port. The ITC has certain major advantages over the US courts. First, while the courts require the effective presence of the defendant before liability can be found (called “jurisdiction over the person”), the ITC merely requires that the goods in question have arrived in a US port (called “in rem jurisdiction”).152 That difference in jurisdiction is critical where the alleged infringer never sets foot here and has no other assets here to seize. Second, as an arm of the Executive Branch, the ITC has nationwide subpoena power (courts are local by nature, so nationwide subpoenas are extremely limited). Third, the ITC has the power to issue a “temporary exclusion order” while the case is underway, which requires far less proof of harm than a temporary injunction from a court.153 Fourth, the patentee does not need to have suffered any injury to be entitled to a permanent exclusion order unlike in the courts.154 Finally, because the evidence needed to receive relief is lower, ITC cases are often faster than courts. Courts do have one huge advantage over the ITC: they are the only mechanism for recovering money damages. Practical Issues of Litigation Just because someone can sue does not necessarily mean doing so is a good idea. A very old adage says that anyone who goes into court a pig comes out sausage. This saying wryly and accurately captures the overall practical reality of litigation. First and foremost, litigants in almost every lawsuit must pay their own costs. Litigation generally is expensive, but patent litigation is extraordinarily so. A 2009 study found that the average cost to litigate a patent infringement case ranged from $650,000 for a small case (damages under $1 million) to $2.5 million for a medium case (damages under $25 million) to over $5 million for a large case (damages over $25 million).155 Another study suggests that the average (mean) cost is about $3 million.156 These figures do not include loss of opportunity to use those resources in the litigants’ businesses, any loss of market opportunities, or the mental toll litigation exacts from the participants. Second, while an issued patent carries a legal presumption that it is valid, having an issued patent does not assure automatic victory in court. On the contrary: every year, a large fraction of cases end with a ruling that the litigated patents are void (the actual fraction varies dramatically by year and by technology). In addition, a court might find that the patent applicant secured

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

issuance through inequitable conduct, which comes with its own severe penalties (not to mention paying the other side’s litigation costs and attorneys’ fees). Finally, a court might issue a ruling on a major question of patent law, which casts serious doubt on the validity of many other patents owned by the plaintiff. Third, litigation takes time to resolve. One study of lawsuits filed between 1995 and 2008 found that the median time between the initiation of a lawsuit and the first day of trial ranged from about 1 to 3 years, with a median of 2 years.157 If the case is completely resolved before trial (for instance, on a motion for summary judgment that is not appealed), resolution is typical in 1e2 years and rarely more than 3 years, but if the case goes to trial or on appeal, 3e4 years is more typical, with complex cases lasting a decade or more.158 Finally, even a successful outcome in the court is not necessarily a victory in the end. For instance, many infringers are “judgment proof,” meaning that they lack the resources to pay any damages, let alone the full amount awarded by the court. Second, a court might rule that the defendant infringed, but the plaintiff suffered little damage (the cost of litigating was higher than the reward). Third, public relations costs must be factored in: a company might lose more business from sour press coverage than it recovers from the litigation. These three reasons are often major factors in why companies rarely enforce their patents against individuals (particularly researchers at universities).

Basic Elements of the Patent Application Process For the purposes of this section, all discussions relate to US laws and procedures except as specifically noted. Due to harmonization efforts, many of these topics have close analogs in other nations’ patent systems, but the variations are so numerous and substantial that a useful examination of other systems is beyond the scope of this chapter. Content of a Patent Application As a general proposition, a patent application includes three major elements known as the “specification,” the “claims,” and assorted technical items. Each application may contain only one invention. As mentioned, the applicant owes the USPTO (and the public) a duty to disclose everything material to patentability, and efforts to hide information or otherwise deceive the USPTO is known as “inequitable conduct,” which can lead to an unenforceable and/or void patent. The following discussion elaborates on each of these points.

473

Specification The specification is the narrative describing the invention. The application must include a formal title, an abstract of the invention, and a detailed description of how to make and use the invention. Normally, the application includes a section describing the background of the technology, showing how the existing tools do not address a specific problemdwhich, of course, the invention solves. It usually includes several examples of how to make and use the invention, such as assorted ways to put the pieces together, or different synthesis protocols, or alternative options if one or another component is unavailable; the application then must, as noted above, identify which example the inventor believes is the best mode. It must include diagrams or drawings whenever they would be necessary to understand the invention, along with a description of each drawing.159 Finally, it must include a formal title, a list of cross-references to other pending applications, and an abstract of the invention. Each word in an application is normally accorded its ordinary, dictionary definition, and technical jargon that is truly standard in a field will be given its field-specific meaning, but the applicant is encouraged nonetheless to define specialized terminology as needed for clarity. Indeed, an axiom of patent law is that the applicant has the right to be his or her own lexicographer.160 Of course, an applicant (and, later, patentee) will be held to whatever definition is used. One critical point about the specification: Once the application has been filed, while the text may be (and often is) altered or amended, no new substantive material may be added. An improvement, a newly crafted embodiment, or a new field of use necessarily broadens the scope of the original filing. As these new developments clearly arose after the filing date of the original invention, including them would muddy the question of patentability in light of intervening disclosures. Instead, these additional elements must be captured by an entirely new patent application. The mechanism for doing so is discussed in the following section. Claims Every patent must include at least one “claim.” The importance of the claims to a patent cannot be overstated. Like a deed to land, the claim articulates the legal boundaries of the “property” from which the public is excluded. This function is critical, as it notifies the public of what may and may not be done without infringing the patent. A patent therefore lives and dies by the exact wording of its claims: phrasing that is too broad will capture technologies already public (and thus renders the claim invalid), while phrasing that is too narrow may miss the most commercially valuable form (which

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

474

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

might not be apparent before the product reaches the marketplace). In terms of format, a claim is a single sentence that describes the subject matter of the invention together with all necessary elements. Typically, a patent contains multiple claims (each of which must have some substantive variation from all of the others). Each claim must be numbered separately. So, a hypothetical patent on a chair might include the following three claims: I claim: 1. An object-supporting device, comprising a surface in a horizontal orientation relative to the floor, and at least one vertical support member depending from the horizontal surface and attached thereto. 2. The object-supporting device of claim 1, wherein the number of vertical depending support members is at least three. 3. The object-supporting device of claim 1, and further comprising a second surface connected to the first surface, wherein the edge of the second surface is situated near the edge of the first surface and aligned substantially at a right angle to the first surface. Notice that claim 1 includes just two elements: a surface for supporting (presumably a person, but perhaps books, laundry, or anything else one might put on the seat of a chair) and at least one supporting leg (some chairs and stools use only a center post). Likewise, claim 2 adds a minimum number of legs, and claim 3 adds a backrest. These minimum requirements are referred to as “limitations.” The more limitations a claim includes, the fewer variations will fall within the scope of the claim. Claims with more limitations are said to be “narrow,” and claims with few limitations are “broad.” Claims 2 and 3, each of which adds limitations to a prior claim, are “dependent” claims; claim 1, naturally, is an “independent” claim. Claim 1 in any patent should be the broadest claim161dbut it is possible for another, later claim to end up being the broadest. Each claim in an application is reviewed by the USPTO separately. If claims 1 and 2 are found invalid (say, for instance, because one-legged stools and tables with four legs were well known), then claim 3 might still be valid if no one had ever attached a backrest. Similarly, after the patent issues, a court may be asked to review the patent to determine whether it is valid; at that point, the court will likewise do a claim-by-claim analysis, though will give some deference to the USPTO as the experts. As long as one claim survives the court’s review, the patent as a whole is valid, though the patentee will not be permitted to enforce those particular claims that a court has found to be invalid. Inventorship (and coinventorship) is determined on a claim-by-claim basis. Therefore, once the USPTO and applicant agree on a final version of the claims but

before the patent issues, the applicant must reevaluate the final claims to ensure that inventorship has not changed (or must amend the application accordingly). Infringement also is determined on a claim-by-claim basis. Specifically, the court162 will compare the allegedly infringing article to the words in the claim. If all of the limitations in a claim are present in the article, then even if the article also includes additional elements (e.g., the “seat” and “backrest” surfaces each includes a cushion), the article infringes the claim. On the other hand, if even one element in the claim is missing from the article (e.g., a “chair” having a seat and backrest but lacking legs, for picnics), the article does not infringe the claim. Technical Items In addition to the items above, the application must include several technical and administrative items. These include an oath/declaration by the applicant(s) concerning inventorship of the claimed invention; evidence of assignment, if the applicant is not also an inventor; an Invention Disclosure Statement (discussed in the following section); certain forms; and of course, the filing fee. With the Patent Reform Act, the inventor’s oath/declaration is now far less critical. For inventions involving nucleotide or protein sequences, the application must include a sequence listing.163 This is now accomplished by including a separate electronic file along with the application, rather than directly in the text of the specification. Likewise, for inventions comprising software, the code may be submitted in a separate file. Finally, for biological materials capable of self-replication (whether directly, like cell lines, or indirectly, like viruses), the applicant may deposit a sample in an approved repository and then rely on that deposit in crafting the language of the application.164 Normally, models and other physical exemplars are not required for issuance of a patent, but an applicant may submit one if it will help the USPTO understand the invention better; in unusual cases, the USPTO may require that the applicant submit a physical exemplar, but generally the USPTO only requires drawings or electronic images.165 If a patent claims a biological material that cannot be made or isolated without substantial experimentation, deposit of a sample in a repository is required. One Invention per Application (“Unity”) Each patent application must contain exactly one invention.166 If the various claims in an application actually cover two or more independent and distinct inventions, the USPTO may divide the claims into two or more groups (each representing a single invention)

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

and require the applicant to pick which group to pursue first. This process is discussed in greater depth below. Also, if an applicant files two independent applications both claiming the same invention, the USPTO will reject the later-filed application on the grounds of “double patenting.” The applicant can overcome this rejection by ceding any patent life beyond the earliestto-expire patent (known as a “terminal disclaimer”). This ensures that the applicant will not arrange to extend the effective life of the invention beyond the term to which the applicant would be entitled from the first-filed application. The Duty of Disclosure and “Inequitable Conduct” As mentioned, the applicant has the duty of disclosure to the USPTO.167 A willful failure to comply with this duty can result in a court finding that the applicant (and, if complicit, the applicant’s patent attorney or patent agent168) engaged in “inequitable conduct.” The court’s remedy for inequitable conduct is either to rule that the patentee will not be allowed to enforce the patent or else rule that the patent is void altogether. Accordingly, the application must include all information material to patentability. The scope of information obviously includes all prior art related to the invention of which the inventor is aware (including, as discussed above, certain secret art). In addition, the applicant must disclose the “best mode,”169 other patents and pending applications related to the invention,170 and anything else “material,” i.e., that reasonably might influence the USPTO’s decision.171 Note that the duty of disclosure is a continuing one; therefore, if the applicant discovers material information during the life of the patent application (or patent), the applicant may submit it at any time. Typically, the applicant reports prior art using an “Information Disclosure Statement,” or IDS. The IDS includes patents, published patent applications, journal articles, slides from presentations, posters, abstracts, and other items that might qualify as prior art (merely including something in an IDS does not constitute an admission that the listing is relevant prior art). Any item not in English must be translated or must provide an English-language summary. While the application remains pending, the applicant may submit supplemental IDSs to report newly discovered prior art. US Applications: Types and Filing Procedures Those who have never participated in the patenting process, along with those on the periphery of colleagues’ inventions, often wonder: why does “the system” take so long to produce just one patent? Why do the patent attorneys seem to want to speak with the inventors repeatedly and spend lots of time talking about the invention? why could not the lawyers write a decent

475

patent application on the first try? And why do patents cost so much? These concerns can be alleviated by understanding what goes into getting a patent. Basic Types of Applications The USPTO offers applicants four types of patent applications that will eventually lead to an issued patent plus two additional types that do not. The four are the original (often called the “parent”) application, the “Continuation,” the “Continuation-In-Part (CIP),” and the “Divisional.” The two additional types are the “Provisional” and the now-abolished “Statutory Invention Registration.” The following discussion simply identifies each type briefly; a more detailed discussion for most types appears later in this chapter. The original application is the classic start of the review and approval process, called “patent prosecution.” That application fixes the substantive benchmark for the scope of patent rights that the applicant might receive, assuming all other criteria for patentability are met. Subsequent (sometimes called “daughter”) applications rely heavily on the exact content of the original (“parent”) application. The USPTO has two main ways to file a parent application: The US-primacy version, under 35 U.S.C. x 111 (known as the “111” application), and the international-primacy version, under 35 U.S.C. x 371 (known as the “371” application). The 371 application is generally used in the context of a set of international patent applications filed under the “PCT”, while the 111 is typically used when the applicant only intends to pursue US rights. While formats, forms, procedures, and timing dates differ, both versions accomplish the same goal and must contain the same core information. A Continuation application (sometimes indicated as “CON”) is used after the USPTO has rejected an application to “reset” the patent-prosecution clock, which will be discussed below. An applicant may use a CON rewrite the claims but may not add any new substantive content or otherwise expand the scope beyond what was in the parent. In theory, both the parent and CON could issue as separate patents, but given the expense, generally the parent is abandoned once the CON has been filed. A “CIP” is the primary means of capturing improvements and other additions to an invention after a parent has been filed. For purposes of defeating potential prior artdincluding, in particular, the inventor’s own disclosuresdthe CIP claims “priority” to the filing date of a still-pending patent application and then adds whatever expansion the inventor discovered later. To the extent the CIP overlaps with the parent, the CIP cannot be defeated by art arising after the filing date of the parent.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

476

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

The fourth substantive patent application is a Divisional (“DIV”). This type is used when the USPTO recognizes that the original application contains multiple inventions. If, for example, the USPTO says that the original application contains three distinct inventions, the applicant would narrow the parent to cover one invention and then later (while the parent application was still pending) file one or two DIVs, each capturing one of the two remaining inventions. A “Provisional” (“PRV”) application is fundamentally different from the other types because the USPTO will not examine it substantively. It is used as a placeholder to defer the filing of the parent. The applicant has up to 1 year to file a fully enabled utility application based on the PRV; the parent application then gains the benefit of the earlier filing date (called the “priority date”) of the PRV against prior art covered by the PRV. A PRV does not need to contain claims, and many of the formalities of filing are waived until the filing of the parent. The PRV has another benefit though it carries a risk. Because many formalities are waived, an inventor may simply select raw items like manuscripts and slides, attach some basic USPTO forms, and submit it, as is, to the USPTO. This approach is known among the patent community as a “coversheet Provisional,” and it is mainly used in emergency-filing situations, such as when the manuscript will be published or the slides will be presented within a few days. The primary risk is that the coversheet Provisionaldand, in the worst case, the patent relying on itdwill prove to be worthless because it lacks key information or uses ill-considered phrasing. Finally, under the 1952 law, the Statutory Invention Registration (“SIR”) can be used to report an invention to the USPTO without claiming exclusive rights to it. In essence, it is used to make it easy for the USPTO’s examining corps to find the invention when looking at subsequent applications. The SIR system was designed to encourage disclosure in cases where a company merely needed access to a technology with no need for exclusivity, but feared the possibility that a competitor would seek a patent on that exact technology and the USPTO would not find out about the earlier company’s discovery. SIRs look like regular original applications without claims; they are reviewed by the USPTO only to see if it meets the written description requirement.172 The SIR was never widely used and the Patent Reform Act abolished the SIR process, but a person searching the background of a technology might run across one. Timing Considerations Export Control Under US law, the mere act of filing a patent application outside the United States on an

invention made here constitutes an “export,” and as such it is subject to certain controls.173 The applicant must wait 6 months after filing a US-based application (including a PRV) before filing in any other country. The USPTO reviews all applications to see if they contain information critical to national security. If so, the USPTO will refer the case to the federal agency having an interest in the technology for review. For those few applications covering inventions having nationalsecurity implications, the USPTO will issue a “secrecy order,” which will block publication and issuance until the concerned agency allows the USPTO to lift the order. If the USPTO does not issue a secrecy order within 6 months of the applicant’s initial filing, the applicant has an automatic license to file internationally. Publication With two major exceptions, the USPTO will publish all applications in its Official Gazette 18 months after the applications are filed. This means the USPTO will publish 18 months after the filing date of a PRV not from the subsequently filed utility application, though the content will be the utility application not the PRV. Furthermore, the official file associated with the earliest application (called the “file wrapper”)dwhich includes, among other things, all written communications to and from the USPTO, and written summaries of telephone conversations and inperson interviewsdare made available online for free by the USPTO.174 The first exception to the publication rule is where the applicant (1) indicates in the original application that the parent application is a US-only filing; (2) promises not to file anywhere else; and (3) explicitly requests that the USPTO not publish the application and file wrapper.175 Second, the USPTO will not publish an application that an applicant has expressly abandoned before the 18-month deadline (i.e., to keep the invention as a trade secret). Patent Life On issuance, patents expire 20 years from the filing date of the parent application. That means all daughter applications (CONs, CIPs, and DIVs) all expire on the same date that the parent willd this is the price for gaining the benefit of the parent’s filing date against later-arising prior art. Note that expiration is calculated from the filing date of the parent, such that a PRV has the effect of extending the term by up to 1 year. While most patents do indeed expire on the twentieth anniversary of the parent application, many patents actually have their terms adjusted. First, the USPTO may require that an applicant agree to a “terminal disclaimer” to avoid effectively extending the life of an identical, earlier-filed application (this only applies where the second-filed application does

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

not claim priority to the earlier-filed application). A terminal disclaimer means the later-filed patent will expire the same day as the earlier-filed patent. Second, unreasonable delay by the USPTO in prosecuting a patent application is grounds for extending the life of the patent; this extension, however, will be reduced, day-for-day, to the extent delay was the fault of the applicant.176 Third, for certain biomedical products requiring approval by the FDA, the patentee may petition the USPTO to extend the patent life up to 14 years; again, the extension will be reduced by the patentee’s own tardiness.177 In Europe, a patent owner can obtain a “Supplementary Protection Certificate” (“SPC”) on the expiration of a patent for regulated biomedical agents if the premarketing regulatory review lasts more than 5 years. The SPC normally has a maximum life of 5 years or 5½ years with an approved pediatric investigation plan. Prosecution of a Patent Application While the statute is written as though the applicant has an affirmative right to a patent unless the USPTO proves otherwise, the process of getting a patent is not as simple as sending in a form with a check and then waiting around for the certificate to arrive in the mail. Instead, the USPTO will examine applications critically, challenge any and all flaws seen, and essentially force the applicant to demonstrate why the applicant is entitled to a patent. This involved, back-and-forth process is the heart of patent prosecution. Here is the rough sequence for a typical application: • The applicant files an application (parent, CIP, or DIV); • If applicable, the USPTO notes missing parts and/or identifies multiple inventions, and the applicant responds accordingly; • The USPTO issues a “first office action on the merits” (abbreviated “FAOM”), typically noting informalities and other matters of form that must be fixed, “rejecting” certain claims as substantively unpatentable, and “allowing” any claims that are patentable as is; • The USPTO publishes the application (which often happens before the FAOM); • The applicant responds to the office action; • The USPTO issues a second office action, which may include a “final rejection” of some or all of the previously rejected claims or may “allow” some or all of the still-pending claims; • The applicant responds to the second office action, assuming some clams were rejected in it; • Assuming the USPTO agrees that all pending claims are patentable, the USPTO “allows” the application to proceed toward issuance;

477

• The applicant addresses any remaining informalities, rechecks inventorship, and corrects other nonsubstantive flaws; and • The USPTO “issues” the patent. The typical reason why the USPTO rejects a claim178 is that the prior art suggests that the invention is not novel, is obvious, or both. Sometimes, the USPTO will reject a claim because it fails to meet the test of enablement. Enablement rejections are more common in technological niches where finding a viable solution to a particular problem has proven especially difficultdlike finding an effective prophylactic vaccine against HIV. In rare cases (perhaps when an applicant is not represented by a patent attorney), the rejection is based on the fact that the applicant has tried to claim unpatentable subject matter or failed to specify a suitable utility.179 On receiving an office action containing objections to technical matters and rejection of some or all claims, the applicant has three tasks in crafting the response. First, the applicant addresses technical flaws and informalities (though some of these may or must be deferred to the end of prosecution). Second, the applicant addresses any rejections that are not based on prior art. This can be handled by, for example, providing additional data, testimony from witnesses and experts, additional images, or sometimes a sample or prototype. Also, the applicant can try to amend the specification to clarify the language (though may not add new matter that expands the scope). Third, the applicant addresses all art-based rejections. Here, the applicant presents any arguments against the USPTO’s logic, cites favorable authoritative references, and for rejections based on obviousness, submits evidence of secondary factors. Also, the applicant may (often must) amend the claims to include additional limitations. As is apparent, a “final rejection” by the USPTO is not the death sentence it seems to bedall it means is that all rejections have been identified, and at least one has not (yet) been fixed.180 The USPTO cannot issue a final rejection in the FAOM of a parent, CIP, or DIV but may do so on the FAOM of a CON. A final rejection will, however, close the prosecution if the applicant has nothing more to say or add to the record. When the USPTO issues an office action containing a final rejection, the applicant has five choices. First, the applicant may amend the application as required by the USPTO. Second, the applicant may file a “Request for Continued Examination” (“RCE”), which permits the applicant to note mistakes made by the USPTO or raise additional arguments, though the claims may not be amended in an RCE. Third, the applicant may file a CON, which removes the finality of the “final rejection,” permits the applicant to add or amend

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

478

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

claims, and also includes any further arguments the applicant wants to include. Fourth, the applicant may appeal the case.181 Finally, the applicant may abandon the application. As mentioned above, CONs, CIPs, and DIVs must be filed while the parent is still pending. Thus, after the application issues into a patent, the applicant may no longer file any CONs, CIPs, or DIVs. Moreover, the parent applicationdwhich was published by the USPTO 18 months after the earliest filingdbecomes prior art against any subsequent applications that the applicant might file. Note that filing a CIP and/or DIV has no impact on the prosecution of the parent. Also note that delays in filing CIPs and DIVs reduce the life of any patents that issue from them, as their life is calculated from the filing of the parent. Options “After Issuance” Once a patent has been issued, the USPTO’s role in the process officially ends, after which any further official actions on patent validity take place exclusively in the courts. There are two major ways, however, that the patentee may ask the USPTO to reopen a patent for another look: Reissue and Reexamination. Whenever a patent is inoperative or invalid (in whole or in part) due to a defect in the specification or drawings, or because the patentee failed to claim as much as could be claimed under the specification, the patentee may surrender the patent and file a “Reissue” application, provided the problem was caused by an error introduced with no deceptive intent. (Minor clerical and typographical errors that do not render the patent inoperable or invalid are fixed by filing a petition for a Certificate of Correction.182) No new matter may be added, of course, and the claim scope may only be enlarged if the patentee files within 2 years of issuance.183 The other major way to reopen a case is through a process called “Reexamination,” in which someone asks the USPTO to take another look at the patent in light of prior art not previously considered.184 Anyone may request a Reexamination, though requests by the USPTO staff are considered inappropriate.185 Reexamination historically has been exclusively between the patentee and the USPTO, called an “Ex Parte” Reexamination. Members of the public may not participate beyond the initial submission of the additional prior art. On November 29, 1999, the law changed to allow the USPTO to offer a new type of Reexamination proceeding in which members of the public could participate, called an “Inter Partes” Reexamination. The goal was to find a mechanism that was less expensive and faster than litigation by which members of the public could actively

challenge patents by means other than unadorned submission of additional prior art. For those interested in trying the Inter Partes mechanism, the law imposed some stiff restrictions. First, the petitioner must bring up all possible additional prior art in one proceeding; a petitioner may not file a second Inter Partes Reexamination request to submit prior art that could have been raised in the first proceeding.186 Second, the petitioner must present all issues at once; if the patentee later sues the petitioner in court for infringement, the petitioner will be barred from presenting issues that could have been raised in the Reexamination.187 While the number of people using the mechanism is slowly rising,188 the USPTO’s data reveal that the mechanism has not been utilized as much as hoped.189 The Patent Reform Act implemented some major changes to the Reexamination procedures. In particular, the law abolished the old Inter Partes Reexamination and replaced it with two new mechanisms: the “Post-Grant Review” (PGR) and the “Inter Partes Review” (IPR). The law also created a third post-grant process, called “Supplemental Examination.” The old Ex Parte Reexam procedures remain essentially intact. For the PGR, any third party may, within 9 months of the issuance of a patent, file a petition with the USPTO asserting that the patent is invalid. The PGR applies only to patent applications filed on or after March 16, 2013, and the third-party petitioner does not participate in the proceedings. Unlike an Ex Parte Reexam, where the only basis for challenging a patent is prior art not considered by the USPTO, the PGR petition may assert any grounds of invalidity except best mode and inequitable conduct. The IPR may be filed against any issued patent, including patents already issued before the Patent Reform Act, though it must be filed at least 9 months after the patent has issued, up to the expiration of the patent. Again, like the PGR and unlike an Ex Parte Reexam, the IPR petition may assert any grounds of invalidity except best mode and inequitable conduct. Finally, like its predecessor Inter Partes Reexam, the IPR carries the consequences of estoppels for any issue that was or could have been handled in the IPR proceeding. The Supplemental Examination is a proceeding designed to allow a patentee (or, more frequently, a later assignee) to “clean up” a messy or inadequate prosecution to strengthen a patent against later challenges to validity. Specifically, a patentee/assignee may submit prior art that the USPTO had not previously considered, and the USPTO has 90 days to determine whether or not that prior art matters enough to start a Reexamination. If the USPTO declines to do so, the patent is insulated against the charge that it was obtained through “inequitable conduct.”

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

International Applications and Filing Procedures The process for securing a patent in any one country is complex enough, but coordinating the simultaneous filing of patents in a collection of countries is dauntingly difficult and expensive. Accordingly, countries have arranged several treaties to make international patenting simpler, less expensive, and less burdensome. The key treaties are mentioned above; the following discussion represents current practice under them. Patent Cooperation Treaty Applications Under the PCT, an application filed in any member country that conforms to certain rules and conditions will be honored in all member countries. The PCT process does not eliminate the need for an applicant to prosecute nation-specific patent applications in each national or regional patent office, and the PCT application itself will never become an “international patent,” but it does facilitate the effort to acquire patents from multiple countries in a variety of ways. The PCT process consists of two main phases: The “international phase” and the “national phase.” Essentially, the international phase begins when an application is filed in one (member) patent office called the “Receiving Office,” and the Receiving Office performs certain preliminary functions. The national phase begins up to 30 months after the priority date190 by which point the applicant must convert the single international filing into regional or nation-specific applications. The applicant selects which national and/or regional offices to enter. The international phase consists of five specific actions by the Receiving Office,191 three of which are mandatory and two are optional at the election of the applicant.192 The three mandatory actions are (1) checking the application for completeness and matters of form, (2) performing a preliminary search of prior art and then issuing an “International Search Report” (ISR), which contains a written opinion on patentability, and (3) publishing the application and the ISR. The optional actions are (4) performing supplementary search report and (5) writing an International Preliminary Examination Report (IPER). The optional actions are a mixed blessing for applicants. On the positive side, the IPER is not binding on the national patent offices, and the supplemental search easily may reveal useful prior art, thus suggesting which strategies are worthwhile and which are fruitless to pursuedextremely valuable information, given the costs of international patenting. On the negative side, any remarks by the Receiving Office questioning patentability will almost certainly be picked up by the national offices; if those remarks were poorly considered or

479

mistaken, fixing the damage will cost resources and could complicate the prosecution needlessly. The key benefits of the PCT system is that an applicant starts by filing one applicationdin one language, with one patent officedand paying one set of initial filing fees and one set of legal bills, all in one currency. All of the remaining applications and fees can be deferred by months or even years, and in the meantime, the applicant finds local counsel and translators, arranges for licensing, drums up investment, and perhaps conducts some further research and development into the product. All national patent offices enjoy the benefit of the uniform format, which leads to more efficient reviews. The applicant also gains the benefit of the ISR, which may reveal prior art that the applicant had not found prior to filing the PCT application, and this can help the applicant decide how best to proceed before making a large investment in national stage prosecutions. Finally, the IPER offers a window into a likely course for the prosecution, information that reduces risk and encourages investment. Regional Patent Offices Regional patent offices (EPO, ARIPO, OAPI, EAPO, and GCC) centralize patent prosecution for their member states. At national phase, the applicant files a single application in the regional office, which then carries out a single prosecution on the substance of the application. When the regional office issues a patent, that patent is valid in all member states, though each member state may have certain prerogatives before the patent may be enforced there. For the EPO, the national phase application must be filed in English, French, or German. On allowance (called the “Grant”) of the EPO patent, the applicant must select the member countries in which the applicant wants to acquire patents (called the “Validation” stage). After Grant but before Validation, anyone may challenge the patent in an inter partes administrative proceeding. Assuming no one challenges the patent, or the challenger loses, the case proceeds to Validation, at which point the applicant must have the patent translated into the national language of each country and must hire local counsel to file the EPO patent in each local patent office. As of the writing of this chapter, European Patents are still national documents, and the EPO patent is not directly enforceable in any court. Each member of the EPO reserves the right to require that the EPO patent conform to local formatting. More importantly, patents are still enforced nationally; in other words, for allegedly infringing acts occurring in Greece, the patentee must sue in the Greek courts based on the Greek version of the patent and so on. Inventorship, ownership, validity,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

480

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

renewal, revocation, and infringement are determined independently under each state’s laws. To simplify the process further, 15 member states of the EPO further agreed to limit the need for translations. Under the “London Agreement” of October 2000, seven agreed to accept applications written in English, French, or German with no translation into the local language, and eight agreed to accept them as long as the claims are translated.193 Of the eight requiring translation of the claims, five also require that the description be translated into English.194 Of the seventeen remaining states, three accept applications written entirely in English, one accepts French, and one accepts German, without requiring translation into the national language. Only 12 states still require full translation, and some of them are likely to join the London Agreement eventually. Historically, an applicant could choose to ignore the EPO entirely and file national phase applications directly in the patent offices for specific member countries. Eleven members, however, have closed that route.195 For these countries, local patenting is only available following the Grant by the EPO. The legal landscape for European patents is in the process of changing, however. In February 2013, 25 members of the European Union signed a new agreement creating a unitary patent system, which will come into force once the legislatures of at least 13 states, including France, Germany, and the United Kingdom, have ratified and implemented it (as of the writing of this chapter, 9 have done so, including France). Italy and Spain both sought to annul the treaty over language issues, but the last of their petitions was dismissed in May 2015. Once the treaty has been fully implemented, patent applications may be filed in English, French, or German without the need to translate into any other language. Once issued by the EPO, a patent will be enforceable throughout all participating European countries without proceeding through Validation. Moreover, all litigation will be handled by specialty trial courts in each of the United Kingdom, France, and Germanydeach of which expected to receive about 1/3 of the filingsdplus local trial courts in the Scandinavian countries, Ireland, and the Netherlands. There will be only one appeals court, in Luxembourg. To lower caseload, there will also be Patent Arbitration and Mediation Centers in Slovenia and Portugal. All judges will be required to have substantial technical training. As of this writing, the United Kingdom is preparing to withdraw from the European Union, so it is still unclear how that will impact existing treaties and processes. Combining US and Patent Cooperation Treaty Filings As mentioned above, there are four major paths for pursuing US patents in the context of the PCT system. Each path has substantial consequences, and under the right

circumstances, each has its own merits. Thus, having a basic understanding of their rough contours is a good idea. In the first strategy, the applicant ignores the PRV and files a PCT application in the USPTO. Thirty months later, at national stage, the applicant files a “371” application in the USPTO (along with the national applications in other countries). Recall that the 371 form of the parent application is used to “convert” the PCT into a national-stage US application. Thereafter, the applicant may later file CONs, CIPs, and/or DIVs just as with the standard “111” application. The patent(s) issuing from the 371 filing will expire 20 years after the earliest application was filed. This strategy maximizes the time between the PCT and national-phase applications, and it accelerates issuance of the US patent by a year but loses a year of patent life. If the technology is likely to be obsolete well before the patent expires, this strategy might save some upfront costs. In the second strategy, the applicant files a PRV, PCT, and 371 applications with the USPTO at the appropriate times. As with the first strategy, the applicant may file CONs, CIPs, and/or DIVs; these patents that result will expire 21 years from the filing of the PRV. This strategy maximizes the deferral of costs except that for the minimal extra expense of filing the PRV, and the applicant gains a year of patent life but results in lengthy prosecutions (during which time the applicant cannot sue infringers). This is critical for technologies that might be reaching full profitability just toward the end of the patent life, such as in the field of pharmaceuticals. In the third strategy, the applicant files a PRV, but instead of filing the PCT at 12 months and the 371 at 30 months, the applicant files the PCT application together with the 111 application. While the 111 application is normally used when the applicant does not file a PCT application, it may be used together with the PCT. The reason for filing the 111 is that the USPTO will begin substantive prosecution of the case years earlier than with the 371. The advantage is that the patents resulting from that 111 are likely to issue earlier, meaning that the patentee will have a larger window in which to enforce the US patent(s). The main disadvantage is that costs are accelerated. In the fourth strategy, speed is everything. When filing the original PCT application in the USPTO, the applicant requests “fast-track” designation, which comes with a very large fee but promises accelerated reviewdunder 1 year, perhaps in just a few months. Simultaneously, the applicant requests an immediate interview with whoever is assigned to examine the application for the USPTO. The goal there is to shape the examiner’s understanding of the invention and prior art before the FAOM, which greatly increases the chance that the invention will be allowed immediately. Once the USPTO allows commercially useful claims, the applicant

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

proceeds to file national-stage applications as quickly as possible, claiming the benefit of the PPH to convince other countries’ patent offices to grant claims of comparable. This strategy requires a major cash outlay in the beginning of the prosecution, and it demands highquality drafting and advocacy by the patent attorneyd but by avoiding a lengthy prosecution, the applicant ultimately saves more money on attorney’s fees than is spent on this strategy. General Strategy Notes The first major strategy note concerns costs. Patenting is an extraordinarily expensive endeavor. Fees charged by the USPTO add up quickly; for example, before the Patent Reform Act, each original filing (including PCT, 111, 371, CON, CIP, DIV, and reissue) ran about $300, with a surcharge for “excess” claims ($50 per claim over 20 plus $250 per independent claim over the first 3), the USPTO’s ISR/IPER fee was about $1000, and miscellaneous prosecution fees cost between $100 and $2000 per item. At issuance, the USPTO charged a $1500 fee and then charged “maintenance fees” after each of the third, seventh, and eleventh anniversaries of the issuance date, which total another $8000 or so.196 These and all other USPTO fees were increased by 15% in the Patent Reform Act, and the USPTO received authority to reset its fees in the future. Useful estimates are very hard to offer because each case is truly unique, but everyone apparently agrees they are expensive.197 Webpages for several law firms suggest that they will charge in the range of $10,000 for an average US case, plus $20,000 per foreign country, but the fine print always includes the warning that costs can go higher.198 One study in 1996 found that cradle-tograve costs for a patent ranged from $2000 (South Africa) to $40,000 (Japan)199; costs have only risen since. Assuming a reasonable estimate is $15,000 per country, then filing national applications just in the United States, Canada, Australia, and the EPO (selecting all 37 countries) would run $600,000. In the EPO, selecting just the London Agreement countries plus those that accept English applications would save about $250,000. Data collected by the American Intellectual Property Law Association suggest these figures are in the right order of magnitude though the exact selection of countries and the technology in question make a substantial difference in price.200 The second major strategy note concerns timing. Accelerating the prosecution accelerates when fees are due, but it also hastens the point where the patentee can try to enforce the patent. Delaying prosecution also carries the potential harm of reducing the amount of patent term extension available. For some technologies, such as pharmaceuticals, loss of patent life might outstrip the upfront costs of acquiring the patent.

481

The third major strategy note involves the dilemma over CIPs and DIVs. An applicant might be tempted to cram as many variations as possible into an original filing, and then piggyback with later CIPs to take advantage of an earlier filing date. This strategy is particularly appealing where the field is full of fierce competitors publishing papers on very similar subjects. On the down side, the strategy results in a shorter life for the portfolio as a whole and a more complicated prosecution overall. Alternatively, an applicant might try to tease apart each unique embodiment, to file as many applications as possible, independently. This strategy is particularly appealing where the applicant wants each case to have a clean prosecution or, perhaps, the applicant wants to create a “thicket” of patents to deter competitors from trying to invalidate any particular patent. On the down side, this strategy takes the substantial risk that intervening prior art may arise in between filings, and that the strategy will provoke the USPTO to issue a “double patenting” rejection in the later cases. Current Major Efforts to Alter US Patent Laws Public pressure to reform certain abuses (real and perceived), as well as to harmonize US patent law with international standards, has been building for years. The Patent Reform Act of 2011 was the most recent major move, but it surely will not be the last. Moreover, US patent law continuously evolves through court cases. Chances are exceptionally high, therefore, that at least one major change to the patent law as described in this chapter will have occurred by the time this chapter has published. Bearing this situation in mind, the following represents a snapshot of efforts currently underway. International Harmonization As stated above, patents are national documents, each valid only in the nation that issued it. Variations in national patent laws increase the cost of acquiring patents, and in some cases, the cost of doing business itself. In the bygone era where international trade was not as important as domestic, variations in patent law from nation to nation were a hindrance but not a critical one. As international trade has grown from a minor to major to critical element in commerce between the United States and its major trading partners, the added cost caused by variations in patent laws has grown alongside. Several major treaties have been signed to streamline and harmonize patent laws and practices (discussed below). Additionally, the USPTO, the EPO, and the Japanese Patent Office have established a trilateral commission to examine ways of harmonizing the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

482

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

respective systems. Also, patents are a recurring topic for the regular international trade-treaty negotiations. One major, common discussion concerns eligibility to file applications. The United States’ 1952 law follows a first-to-invent system, with a complex mechanism for resolving disputes over who was first to invent, and the law is transitioning to a “first inventor to file” system; most other nations use a simple first-to-file system. Another major difference involves the “statutory bar:” the United States permits applicants to file patent applications for up to 1 year after a disclosure, while most other nations would bar patents immediately, with no grace period, even if the disclosure was made by the inventor. Patents on Genes and “Mere Associations” Health-care patents commonly inspire passionate views, but two areas are particularly intense at this time. Both were discussed in depth earlier in this chapter, in the discussion of patent eligibility, but any review of major changes now underway requires at least a brief overview. The first issue is that patenting discoveries relating to genetics remains highly controversial despite (or because of) 30 years’ experience with such patents. Several times in the last two decades, members of Congress have submitted bills that would limit or prohibit patents on genes. The USPTO has issued guidelines specifically directed toward gene-based patents, which were upheld by the CAFC.201 The Supreme Court’s decision in Myriad largely favored the views of opponents to gene patents, and while the court’s decision officially was limited to the particular facts of the case, the logic used cast doubt over the validity of many biotechnology patents.202 The second intensely discussed topic is in patents claiming rights to “mere associations.” Here, the discovery in contention is the pure recognition that a certain naturally occurring circumstance has medical significancedfor example, noticing that an elevated level of a biomarker indicates a deficiency in a vitamind but both the circumstance itself and the method of detecting the circumstance are already known. In that instance, all that remains to be patented is the mental step a doctor performs on receiving test results, namely, associating the test results to the condition. In the Mayo decision, the US Supreme Court signaled distaste for diagnostic patents to the extent they relied on such associations.203 Abusive Tactics: “Patent Trolls” and “Inequitable Conduct” Every commercial venue and legal system has its abusers, and efforts to rectify abuses are always ongoing; patent law is no exception. Presently, two

major areas are the focus for reform: Aggressive use of patents by nonpracticing entities (NPEs) (so-called “patent trolls”), and unreasonable allegations of “inequitable conduct.” By way of background, a basic tenet of property law in the United States (whether it is land, personal property, or IP) is that the owner of property has the right to do with it as he/she pleases. With some noted exceptions, there is nothing illegal about the owner of property deciding not to develop or otherwise commercially exploit the property. Simultaneously, the owner has every right to stop others from taking (“trespassing” on) that property or to demand royalties in exchange for others’ commercial use. This general philosophy has provoked a particular dispute in the patent community concerning “NPEs.” NPEs are businesses that are created around a patent portfolio but have no intention of making any product themselves any time soon. Exactly how an NPE chooses to use its portfolio can lead either to products or to problems for the marketplace. Most NPEs, including universities and government labs, license their portfolios to larger companies as an incentive to develop and distribute new products. The success of that strategy may depend on aggressive enforcement of the patent portfolio, especially during development. Nonetheless, this strategy is consistent with the underlying social and economic purposes of the patent system. Alternatively, some NPEsd pejoratively called “patent trolls”duse their portfolios simply as a threat to extract money from others whose products probably infringe the patents. While such a strategy may be legal, it runs counter to the social and economic goals of the patent system. The social harms from strategy are compounded by the use of patents of dubious validity or ridiculously narrow scope, and by threats levied against companies that are clearly not infringing but for various reasons cannot litigate. True “troll”-like practices clearly would put a drag on the marketplace greatly adding business risk and wasting resources with no public benefit, but describing theoretical culprits is far easier than accurately picking out the real ones. Moreover, targets of patent enforcement by either kind of NPE face a tough choice: fighting the accusation (which will cost of millions of dollars and will consume years of staff time with no guarantee of success) or paying a royalty despite the fact the target believes the money is not owed. From this situation, two tough questions arise: What meaningful, objective test will distinguish the socially harmful NPEs? If none exists, which is better overall: to tolerate the harmful NPEs or to quash all NPEs? Often, for any given NPE, the label fits or fails based solely on one’s perspective.204 Without an objective test to distinguish market-harming NPEs from beneficial

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

NPEs, any limits on patent enforcement risk crimping legitimate business activities. Unfortunately, there are no easy answers to those two tough questions. Another zone of contention surrounds allegations of “inequitable conduct.” As discussed below, patent applicants (and their attorneys) owe a duty of candor and disclosure to the USPTO while their applications wind their way through the system. At worst, violations of this duty can lead to the patent being held void/unenforceable, and the patent attorney handling the case faces disbarment and malpractice suits. When someone is sued for patent infringement, one possible defense is to assert that the defendant should not be held to have infringed a patent that issued because of a violation of this duty. Even though there is no evidence that the frequency of violations of this duty have increased over the past 20 years or so, the level of evidence that the defendant must produce has been ratcheted down, and the penalties for raising this defense frivolously have not been vigorously imposed, such that these allegations are now approaching standard fare for infringement suits rather than a rarity.205 Worse, they substantially raise the complexity, duration, uncertainty, and costs of litigating. Two major events have reduced the threat of inequitable conduct charges. First, the Patent Reform Act has limited the scope of this problem substantially by removing the phrase “without deceptive intent” from several places in the patent statute. Second, the CAFC issued a major decision in the case of Therasense, Inc. v. Becton, Dickinson & Co.206 In the Therasense case, the CAFC clarified that a finding of inequitable conduct requires clear and convincing evidence of both that a piece of prior art omitted from prosecution was “material” to patent eligibility (i.e., the patent would not have issued but for the omission) and that the applicant had a specific “intent” to deceive through the omission (i.e., the deception was truly knowing and deliberate, rather than the applicant should have known better). These two developments have substantially limited the range of patents vulnerable to charges of inequitable conduct. Nonetheless, the doctrine survives. Compulsory Licensing and Breaking Patents Several countries and nongovernmental organizations have long taken the stance that access to critical medicines is a basic human right, which should not be subordinated to the interests of corporations’ profits from selling those medicines. They blame patents as the key cause (if not the only cause) of the fact that certain medicines are too expensive for access by poorer nations. Consequently, they have spent years trying to convince the international community to accept the premise that each nation has the right to force patent owners to license their medical technologies

483

nonexclusively. In a few high-profile instances, Brazil and Thailand have threatened to revoke (“break”) certain patents they had previously issued on valuable drugs. More controversially, Brazil has also threatened to break US patents and copyrights wholesale if the United States does not change its stance on certain unrelated trade issues.207

Copyrights, Trademarks, and Trade Secrets Copyrights As discussed above, patents protect new and useful embodiments of discoveries and innovations; copyrights protect original works of authorship. Copyrights are intended mainly to protect new items that are creative in nature rather than useful. Inventing something truly new and useful on demand is extremely difficult, but there are an infinite number of ways of creating an original work of authorship around a given idea. Because of this divergence, the laws of patent and copyright are fundamentally different. As the following summary will make clearer, carelessly lumping copyrights together with patents can lead to problems; each type of property deserves independent thought and handling. To be entitled to a copyright, the author must create an “original” work. “Originality” for copyright is substantially different than the “novelty” standard for inventions. An “invention” is not novel if it was ever disclosed, even if the inventor never saw the disclosure. To be “original,” in contrast, the new work merely must be “not copied.” So, in theory, two people who independently create identical works get independent copyrights in what appears to be the exact same thingdas long as neither one copied from the other. Also, the work must be “fixed in a tangible medium of expression.” Copyright exists the moment the author lifts pen from page or hits “save” on the computer (magnetic media counts) or shoots a picture (chemical film or digital memory). Thus, a person standing at a podium making a speech gets no copyright in the speaker’s own ephemeral performance, yet anyone in the audience using a video camera owns a copyright in that person’s particular recording. Note that copyright exists automatically on fixation; no later act of registration is required for copyright to exist (though it is a good idea). Certain items cannot be copyrighted. Raw facts and raw data cannot be copyrighted, as they were not works of “authorship.” Just because a person is the first to write a fact down on paper does not mean the person authored the fact itselfdthe author must add some element of creativity, such as a unique selection, arrangement,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

484

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

or coordination of data.208 Also, mere names, slogans, or phrases cannot be copyrighted, as they fall under the domain of trademark law (and, in addition, they are too small to be deemed “works of authorship”). Finally, functional aspects of an artistic work, as mentioned above, are protected by patents, while the artistic elements are protected by copyright. As already mentioned, in the case of software, a point worth noting is that both copyright and patent may apply. The particular code that a programmer types into a computer (“source code”) and compiles into binary (“object code”) can be copyrighted, just like a literary work. Simultaneously, it can be patented, if the architecture of the software presents a novel and nonobvious method of processing data. The owner of a copyright has several specific rights. Among other things, the owner has the affirmative, exclusive right to do and authorize others to do the following: (1) to reproduce, including into another medium of expression; (2) to prepare new, “derivative works” based on the original; (3) to distribute copies to the public, including lending and at no charge; and (4) for certain types of works, to perform or to display the work publicly.209 Copyright ownership generally follows authorship but not always. In the employment context, the employee normally owns works made while working unless one of three major situations applies. First, if the employee is specifically hired to write, such as a newspaper reporter, or if the employee is hired under a contract containing a “work for hire” clause, the employer is the owner and the author (despite doing all the creative work, the employee is deemed not the author).210 Second, if a contractor creates a work under a contract that requires assignment of copyrights to the employer, the employer owns the copyright though the contractor continues to be the author. Third, if the author is an employee of the US government and wrote the work as part of the author’s official duties, the work is not eligible for copyright protection in the United States.211 “Joint authorship” is unlike joint inventorship in that the coauthors must create their respective parts of the joint work with the conscious intent of creating a unitary final product. Also, joint ownership of copyright is different from joint ownership of a patent in that copyright co-owners owe each other a duty to share the net royalties they collect from licensing, litigating, or using the copyright, whereas patent co-owners may use the jointly owned patent independently, without any accounting to the co-owners. Distinct from copyright in a “joint work,” one can copyright a “compilation.”212 A compilation copyright arises where a new work is created by collecting and recasting other prior works. The compilation copyright only extends to whatever the compiler’s contribution

to the work might be, e.g., layout and editing. So, two scientists collaborating to write a singular manuscript would jointly own the copyright in their joint work; meanwhile, the scientific journal in which the manuscript is published would acquire an independent copyright in the compilation of all of the articles in it but would not own the copyright to any particular article unless the manuscript authors assign their rights to the journal. While registration of a copyright is not required, doing so is a good idea. A copyright cannot be enforced in court until it has been registered in the US Copyright Office.213 Also, for acts of infringement that occur before registration, the copyright owner must prove actual economic harm to be entitled to receive any money, but for acts occurring after registration, the copyright owner is entitled to receive between $750 and $30,000 per infringing act, regardless of the economic impact (if the owner prefers, the court may award actual damages). If the infringer can prove he had a genuine and good-faith (if ultimately false) belief that the copying did not infringe the copyright, the court can lower this threshold to $200 per act of infringement but not lower.214 Placing notice of copyright215 on the work is not required but is also a good idea. If the author has duly placed a visible copyright notice on the work, chances are much better that a court will decide later acts of infringement were willful, the maximum damages can rise to $150,000 per act of infringement.216 Unlike in patents, copyright infringement is a felony if done willfully and for purposes of commercial advantage or personal financial gain. Punishments can run as high as 10 years imprisonment and fines up to $250,000 per incidence of criminal copying. The law does place limits on copyright owners’ ability to enforce against copying by others, and of these, three are particularly important to most people. The first is the concept of “first sale”donce a customer buys a copy of a work, the copyright owner’s rights in that particular copy are extinguished; that customer may resell that copy of the work without infringing the copyright.217 The second is that certain works come with automatic “compulsory” licenses218; arguably, the most important of these for research is the compulsory license on software, which allows consumers to make a backup archival copy.219 The third and most famous limitation on copyright infringement is the defense of “fair use.” By law, an act of copying that might technically infringe a copyright is not infringing if the use is “fair”dfor example, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research.220 That judgment, however, is not based on the perspective and opinion of

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART ONE: INTELLECTUAL PROPERTY GENERALLY

the copier but of the jury at trial, and only after the act of copying has already been established.221 As a result, relying on “fair use” as a shield against liability is risky. Trademarks Trademarks serve a fundamentally different function from the other major forms of IP, namely, trademarks function to protect the market not the owner of the IP. As a result, trademarks have radically different aspects, rights, and rules. Trademark law in the United States derives from the government’s basic duty and power to regulate commerce. Each State has its own trademark law for intra-State commerce, while the federal government has a national trademark law to regulate inter-State and international commerce. These laws are intended to complement each other operating simultaneously and in harmony (although any conflicts are resolved in favor of federal law). A mark222 is any word, phrase, logo, symbol, shape, number, letter(s), color, sound, scent, or other device (or combination of these) that serves to identify the source of specific goods or services in commerce and to distinguish them from similar goods or services sold by others. A mark serves this function whether or not the consumer knows the name of either the maker of the good or the provider of the service, as long as the consumer knows that any two items bearing the same mark come from the same source. Thus, consumers know that the mark Crest identifies the source of all toothpaste bearing that same starburst logo behind the word, even though most consumers have no idea that Procter and Gamble is that source. If, hypothetically, a new tube of nutmeg-flavored toothpaste bears the “Crest” mark, consumers will assume it was made by whoever made the original version. The primary function of a mark is to make shopping easier. Consider how easy it is to find a soap that has specific characteristics in every bar based on the marks (e.g., a mild soap from Ivory, a skinsoftening soap from Dove, and a deodorant bar from Irish Spring). Consider also how much more likely a person is to buy new products from the same company that makes that person’s current favorites. Now, imagine buying soap in the manner most people buy bulk tomatoesdtesting each bar for scent and washing characteristics rather than simply by looking at the label. The mark makes the whole process much more efficient. Alongside this primary function is the incentive effect a mark has on the manufacturer to ensure consistent quality. A consumer who has a bad experience in a restaurant is not likely to return to that restaurant, and if the restaurant is a franchised chain, the consumer

485

will be turned off from ever returning to any store in that chain. Whether one loves or hates the fast food served at McDonalds restaurants, everyone who has eaten at any branch at least once knows what to expect each time after at any other branch. That effect is a direct consequence of the power of a mark to influence purchasing. Unlike with patents, copyrights, and trade secrets, a mark-owner does not acquire ownership as a result of creating the mark or for any quid pro quo. Rather, ownership is the result of the recognition conferred by the purchasing public. If the consumers recognize and use the mark to identify and differentiate goods and services, then the vendor owns the mark. If the consumers cease using the mark to identify the source of goods and services, and start using the mark to identify the products themselves, the identifier ceases to be a mark and becomes a generic word. Examples of this include aspirin, cellophane, escalator, and trampolinedall of which used to be registered trademarks. Ownership of a mark begins with actual use in commerce not with registration.223 Before registration, the owner may use the symbol “TM” on its goods (or “SM” in ads for its services) as a placeholder to indicate that the user claims ownership of the mark; once the mark has been registered in the USPTO, the owner may use the symbol “Ò”. State registration requires use in commerce in that State; federal registration requires commercial use in more than one State. As between two parties claiming the same mark, ownership of a mark goes to the senior user not necessarily to the creator of the mark (in other words, if you abandon your mark, someone can step in and claim it). Trade Secrets General Principles In a limited sense, the philosophy behind trade secret law is the reverse of patents. As noted above, a critical part of the “quid pro quo” of the patent system is that the applicant must disclose everything about the invention known to the inventor. If, however, an inventor were willing to give up the potential for courtenforced market exclusivity afforded by a patent, the inventor remains free to keep the invention secret. To a limited extent, such secrets are recognized by the law as valuable IP, albeit a disfavored form of property given that society gains almost no benefit from protecting it. In the United States, trade-secret law is set by each State, so exact details and nuances vary.224 For present purposes, a simple general definition of a trade secret is any information that (1) is actually a secret, or at least not readily available to someone motivated to find the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

486

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

secret; (2) derives from that secrecy some commercial value, whether through reducing costs, through increasing in profits, or through the fact that the secret would be difficult and expensive to reverse engineer; and (3) the owner takes steps reasonable under the circumstances to keep it secret (usually through efforts like marking of key documents as “confidential,” segregating them from other commercial papers with highly limited access, and alerting and/or training staff about how to handle it). There are two big problems with trade secrets. The first is value: the state-of-the-art technology in any given field changes rapidly, so every technology becomes obsolete sooner or later. Keeping a technology secret sometimes can accelerate its obsolescence if the rest of the field adopts another solution as the industry standard. The second is secrecy itself: successfully preserving a secret is difficult, and few companies take the time and trouble to actually put any meaningful measures in place. The law has little sympathy, however, for those who cannot be bothered to take reasonable measures to protect their secrets; if a secret were truly valuable, the owner would take those steps. For new inventions, companies face an immediate choice: whether to seek a patent on it or keep it as a trade secret. US patent applications remain confidential for the first 18 months, after which the US Patent Office publishes them, even if no patent ever issues.225 If the applicant withdraws the application before it is published, the information can be kept secret, assuming it remains secret otherwise. Assuming a technology is a bona fide trade secret, the owner can stop others from misappropriating the secret technology (if the owner moves quickly enough) or can try to recover the value of the lost secrecy in the form of money damages. Relying on this strategy comes with substantial risks, however, because the courts will not allow the owner to stop others from reverseengineering the invention (as long as they do not have unfair access to the secret). Also, assuming the tradesecret owner has not publicly disclosed the secret through the sale of a product, someone else who successfully designs a new product that replicates the technology might even be able to patent it, despite not being the first to discover it!226 Key Statutes Relating to Trade Secrets and Federal Employees As a general statement, governments and academic organizations do not generate any trade secrets of their own (they do not sell anything after all), but they routinely possess trade secrets owned by others. For agencies of the US government, three major statutes are particularly relevant in the handling of trade secrets known as Freedom of Information Act (FOIA), the

Federal Trade Secrets Act, and the Economic Espionage Act (EEA). Following the Watergate scandal, Congress passed a series of laws designed to change how the government handles its own records. This law is the “FOIA.”227 Basically, any record in the possession of the government must be produced on request, unless there is some exemption in the law. One major exemption is for trade secrets.228 As a practical matter, agencies are sensitive about releasing trade secrets and strive not to do so, but if the private party that gave the secret to the agency failed to indicate whether or not the information is a trade secret and the status is not otherwise obvious, chances are much higher that the agency will release the information. Another exemption to FOIA is for inventions in which the government owns an interest.229 Unlike other FOIA exemptions, however, this particular exemption only lasts for “a reasonable time for a patent application to be filed.” The exemption does extend to “copies of any document which is part of an application for patent,” but given that the USPTO publishes patent applications 18 months after filing (and makes them available for free online), there is rarely any need for a FOIA request for such documents. The second major law applicable to federal employees is the “Federal Trade Secrets Act.”230 This law bars any employee of the US government from disclosing any information acquired through the employee’s official duties that qualifies as a trade secret,231 except to the extent the employee is required to do so by law or court order. Violators are subject to a fine and/or up to 1 year in prisondnot to mention being fired. The third law is the EEA of 1996.232 Unlike laws governing other forms of espionage, which turn on classified or national defense information, this law focuses on commercially valuable secrets. The EEA creates two classes of crimes, summarized as follows: misappropriation to benefit a foreign power233 and misappropriation intended to cause the original owner injury.234 Penalties for the former class include fines of up to $500,000 and 15 years imprisonment for individuals, and fines of up to $10 million for organizations. Penalties for the latter class are limited to imprisonment up to 5 years for individuals, and fines of up to $5 million for organizations. In addition, for both classes, those convicted must forfeit any proceeds from the crime and all property used to commit the crime. The law does not, however, authorize a private citizen to bring a federal civil lawsuit. Prosecutions under the EEA are exceptionally rare. Based on data from the Department of Justice, prosecutors brought about 34 cases under x 1832 (causing the original owner injury) the EEA from 2000 through the end of 2005,235 and just 5 cases under x 1831 (benefitting a foreign power).236 Given the low fanfare, most of these

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

cases apparently resolved without a conviction. Indeed, the US Department of Justice announced its very first conviction under x 1831 (benefitting a foreign power) in 2006, when two California residents pleaded guilty to trying to steal microprocessor designs from Sun Microsystems and Transmeta Corp., specifically to create a competing company in China, where the Chinese government would contribute seed capital and share in the profits. More common are prosecutions of scientists changing jobs who take their research materials with them, knowing that the former employer had intended to develop that work into a product. Two high-profile examples occurred in 2001e02. In one, the FBI dropped charges against two former Harvard postdocs when they arranged for the return of the materials237; in the other, arising from the Cleveland Clinic, the exemployee had already reached Japan, and the FBI charged a colleague only with lying to law enforcement about the arrangement (the Japanese courts refused to extradite the ex-employee).238 While cases are rare, they can have a devastating effect on a researcher’s career. To sum up, researchers working in federal laboratories should be at least minimally aware of trade secrets and their obligations to protect them. While the federal researchers do not generate trade secrets, they may come to possess the trade secrets of others. When that happens, the consequences for mishandling those trade secrets can be substantial.

PART TWO: PATENTS AND TECHNOLOGY TRANSFER Critical Laws Concerning Patents and Federally Supported Research According to one study, about one third of the funds spent in the United States each year on research and development comes from the US government, the largest single source of R&D funds in the world.239 These funds are allocated primarily through the use of grants and contracts to a wide array of entities, ranging across academia, not-for-profit centers, other governments, and for-profits. Some funds are committed to in-house research facilities, such as the National Institutes of Health’s (NIH) Intramural Research Program. While the recipients of federal funds have relatively broad latitude in how they spend the money, the money comes with certain rules, rights, and restrictions. The following section of this chapter discusses the rules, rights, and restrictions specific to IP and technology transfer.

487

Federal Funding of Private “Extramural” Research: The BayheDole Act Researchers seeking funds from the US government have major three mechanisms to do so: a contract, a grant, and a cooperative agreement.240 These funded researchers (collectively called “Contractors,” regardless of which funding mechanism was used) are bound by the “University and Small Business Patent Procedures Act of 1980,” better known as the “BayheDole Act.” This relatively short law, spanning just 13 sections and about 30 pages of text, fundamentally changed the philosophical and operational effects of federal funds on innovation. History and Philosophy By 1980, the US government spending on R&D had reached about $7e8 billion per year, representing about 1% of the US GDP, and about half of all R&D funding in the United States.241 As Congress had not restricted or directed the agencies to handle invention rights arising from the use of those funds, each agency set its own policies and procedures, and all agencies except the Department of Defense chose to require that Contractors assign their rights to the US government (the taxpayer paid for it after all).242 From 1950 to 1980, the government collected just under 30,000 patents from its Contractors. Lacking clear statutory authority to grant exclusive commercialization licenses, less than 4% were licensed in any fashion, and no products embodying these inventions had reached the market.243 Starting in the early 1970s, the Department of Health Education and Welfare (predecessor to the Department of Health and Human Services) and the National Science Foundation began allowing universities and nonprofits to enter “Institutional Patent Agreements.” These agreements, among other things, authorized those Contractors having preapproved patent policies conforming to agency stipulations to retain title to any inventions funded by the agency. While highly controversial, these agreements, along with the patent policy of the Department of Defense, laid the key groundwork for Congressional hearings in the late 1970s. After much debate, Congress passed a bill sponsored by Senators Birch Bayh (Indiana) and Robert Dole (Kansas), which President Carter signed into law on December 12, 1980. Federal funding under the BayheDole Act is a critical component of the US research enterprise and economy. According to the National Science Foundation, for each of fiscal years 2014e16 the government spent about $130 billion on research and development, of which over $110 billion was used to fund extramural research.244 The Association of University Technology Managers (“AUTM”) reported that in the 35 years following passage of BayheDole, the rate of patents issued to

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

488

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

Contractors per year grew from 250 in 1980 to over 6300 in 2014, universities spun off more than 4200 startups based on federally funded inventions that were still operational in 2015 and around 10,000 new products reached the market (914 in 2014 alone)dabout 1.25 new products per daydbased on federally funded inventions. From 1996 to 2013, patent licensing by universities to industry bolstered US gross industry output by an estimated $1 trillion, increased US GDP by $500 billion, and supported up to nearly 4 million US jobs. About 70% of this growth took place in the biomedical arena.245 About 42% of all US graduate students in scientific fields are supported by federal funds. While the BayheDole Act cannot take full credit for all of this activity, it clearly played a major role. Despite objective measures of economic growth largely attributable to the BayheDole Act, the law has faced continual criticism. The challenges typically fall into four categories. The first is that the act mainly rewards high-return research and does not adequately incentivize research into rare and neglected diseases. Second, other mechanisms to incentivize research exist, and some of these do not carry the deleterious effects blamed on patents (such as higher drug pricing). Third, the focus on patenting has shifted academic research away from “pure” research toward for-profit activities, leading to conflicts of interest, suppression of research data for commercial gain, and cumbersome restrictions on access to research tools. Fourth, the provisions designed to remedy abuses are too limited and too slow to deter anything. Unfortunately, a thorough discussion of the merits and weaknesses of the BayheDole Act and the related criticisms is beyond the scope of this chapter. Organization of Clauses The BayheDole Act is codified at 35 U.S.C. xx 200-212, though some of the sections interspersed in it do not apply to Contractors. The critical operative sections cover the Congressional policy and objectives (x 200), defined terms (x 201), disposition of rights (x 202), “march in” (x 203), and preference for US industry (x 204). Regulations implementing the BayheDole Act may be found at 37 C.F.R. Part 401. Key Conceptsd§§ 200 and 201 In establishing the BayheDole Act, Congress offered a laundry list of policy goals that the agencies must follow in issuing any funding agreement. Quoting directly from x 200, these goals are to promote the utilization of inventions arising from federally supported research or development; to encourage maximum participation of small business firms in federally supported research and development efforts; to promote

collaboration between commercial concerns and nonprofit organizations, including universities; to ensure that inventions made by nonprofit organizations and small business firms are used in a manner to promote free competition and enterprise without unduly encumbering future research and discovery; to promote the commercialization and public availability of inventions made in the United States by United States industry and labor; to ensure that the Government obtains sufficient rights in federally supported inventions to meet the needs of the Government and protect the public against nonuse or unreasonable use of inventions; and to minimize the costs of administering policies in this area.

The fulcrum of the BayheDole Act is the defined term “Subject Invention,” as all operative terms only apply to Subject Inventions. A Subject Invention is “an invention of the contractor conceived or first actually reduced to practice” under a funding agreement for “experimental, developmental, or research work.” Two critical points should be noted from this definition: First, not all funding agreements are governed by the BayheDole Act; second, an invention is a Subject Invention if it was either conceived or first actually reduced to practice under the funding agreement. Under xx 202-204, all funding agreements for experimental, developmental, or research work must allow a Contractor to elect to retain title to any Subject Invention, with three sets of critical limitations. First, the funding agency may change or remove the core clauses in “exceptional circumstances.” Second, the Contractor must agree to several specific duties of reporting, development, and licensing of every Subject Invention. Third, the government has certain rights in any Subject Invention.

Core Terms Required in BayheDole Funding Agreementsd§ 202 Section 202(c) lists the major clauses that must appear in any funding agreement governed by the BayheDole Act.246 Note that only some of these clauses apply to all Contractors in all circumstances; some are more limited in scope. The following discussion represents a brief summary of these terms; generally, only those people working directly on either inventions or the terms of funding agreements need greater detail. All Contractors are subject to eight major elements: 1. The Contractor must report all Subject Inventions within 2 months of when the Contractor learns of the invention. 2. The Contractor must make a written “election of title” within 2 years of notifying the funding agency of the existence of the Subject Invention (or less, if the invention has been publicly disclosed). 3. If the Contractor elects title, the Contractor must file and maintain patents.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

4. The funding agency may require periodic reporting from the Contractor on efforts by the contractor or his licensees or assignees to use or to obtain utilization of the Subject Invention. 5. The Contractor must arrange for any patent on the Subject Invention to include a notice that the invention was made with government support. 6. The government has a nonexclusive, worldwide, royalty-free license in every Subject Invention. 7. The obligations of x 203 (March-In) and x 204 (US manufacturing) apply to all Subject Inventions. 8. If the Contractor fails to report the invention timely, rejects title, fails to elect timely, fails to file patent applications timely, or abandons any pending patent applications or issued patents, the government may take title (if the Contractor abandons only selected international patents while retaining others, the government may only take title to those abandoned patents). For nonprofit Contractors, funding agreements also must include four more restrictions247: 9. The Contractor may not assign Subject Inventions without permission of the funding agency (except to an organization primarily managing inventions); 10. The Contractor must share royalties with the inventor(s); 11. Income from the Subject Invention (after costs and the inventors’ share) must be used to support scientific research or education248; and 12. Except where infeasible after reasonable inquiry, licenses shall go to small businesses. Where a joint invention arises with at least one inventor working for a Contractor and another working for the government, x 202(e) authorizes the funding agency and the Contractor to consolidate ownership either in the Contractor or in the government to expedite development, as long as the funding agreement does not require the Contractor to assign title to the government. Where a nonprofit Contractor either declines title or intends to abandon a case, sometimes another entity is interested in picking up the invention. If so, the Contractor may ask the funding agency for a waiver of the prohibition against assignment. Also, the Bayhe Dole Act explicitly allows the inventor(s) to ask the Contractor and funding agency to allow the inventor(s) to “retain title” to a Subject Invention (this part of the law, x 202(d), can be invoked either before or after the Contractor has made an election of title, as long as the Contractor has indicated disinterest in the Subject Invention for some reason). For either type of request, granting is discretionary, so the funding agency is free to impose any additional conditions on the assignee the funding agency deems fit. Typically, the assignee

489

must assume all of the BayheDole obligations that had applied to the Contractor. § 202dReporting Obligations (iEdison and RePORT) As discussed above, every recipient of federal funds to conduct research and development work must report every invention made using those funds, as well as patent filings, steps toward utilization of the invention, and other items. In 1999, the NIH led an interagency effort to create a streamlined system for reporting inventions: the Interagency Edison system, known as “iEdison.”249 The NIH continues to maintain and update iEdison on behalf of the participating agencies. The iEdison webpage starts with links to each participating technology transfer office. Current participants include components from the Departments of Defense, Health and Human Services, Agriculture, Commerce, Homeland Security, Energy, and Transportation, plus the Environmental Protection Agency (EPA) and the National Science Foundation. Some participating agencies, including the NIH, require reporting to be done using the iEdison system, while the others merely use the iEdison as a portal to relay the user to the appropriate office. For NIH’s contractors, the iEdison system includes a secure login that permits reporting of inventions, supply a confirmatory government license, report on patent filings and other steps toward utilization of Subject Inventions, request that the government allow the Contractor to assign a Subject Invention, request a waiver of the US manufacturing obligation, query the database to generate reports and update and edit the electronic file as needed. While reports on Subject Inventions through the iEdison system are confidential, the NIH has made some data public. The NIH maintains much of its funding data in a searchable system called the Research Portfolio Online Reporting Tools or RePORT.250 The RePORT system provides a central point of access to data, reports, and analyses of NIH expenditures, intramural research, and the results of all NIH-supported research. While the RePORT system is a powerful tool for examining many aspects of NIH’s funding patterns, users can search the system by an inventor’s name to see if the inventor ever received an NIH grant, and then find out what publications and issued patents the inventor reported as associated with each grant found. § 202dDetermination of Exceptional Circumstances As mentioned, every federal funding agreement must include standard provisions allowing the Contractor to elect title to Subject Inventions, except that under certain circumstances, the funding agency may change or

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

490

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

remove these terms. In particular, the most controversial situation where this may arise251 is where the agency finds that, in light of exceptional circumstances, restricting or eliminating the Contractor’s rights in Subject Inventions “will better promote the policy and objectives” of the BayheDole Act, which are articulated in 35 U.S.C. x 200. That finding is called a “Determination of Exceptional Circumstances,” or more frequently, a “DEC.” Some misconceptions surround the DEC mechanism. First and foremost, the DEC should not be used simply because a federal lab’s own programmatic needs would be better served by changing the standard BayheDole rights. Also, silence by the lab contemplating a DEC is not golden; the lab should alert the community of those who might be interested in bidding on a BayheDole contract that the lab is contemplating a DEC and offer them an opportunity to comment on it, to ensure that the agency has fully considered whether or not the use of the DEC is appropriate and its scope sufficiently tailored to the need. And, where feasible, the lab should utilize a so-called “Greater Rights” clause, which gives the contractor a pathway to ask the lab on a case-bycase basis to give the contractor greater rights to an invention than it normally would have under that DECdmore specifically, to reinstate the contractor’s usual BayheDole rights. An agency invoking the DEC must comply with several important regulatory restrictions. First, the agency must prepare a written determination that the exceptional conditions exist, including a statement of facts supporting the conclusion, and an analysis justifying it. This analysis should address with specificity how the alternate provisions will better achieve the objectives in x 200. A copy of the determination must be promptly provided to the Contractor along with information on how to appeal the DEC. Once the agency has approved a DEC, the agency must then write deviated clauses for the funding agreement (for procurement contracts, which are governed by the Federal Acquisition Regulation, the deviated clauses must be approved by the head of the department or agency).252 The NIH has used the DEC mechanism several times. The NCI, for example, has used the DEC to alter the contract of a company operating a facility in Frederick, Maryland, owned by the NCI. Under this altered contract, the NCI operates, among many other projects, a highly successful program originally called “Rapid Access to Intervention Development” (RAID), now sponsored by the National Cybersecurity Assessment and Technical Services (NCATS) under the current name “Bridging Interventional Development Gaps” (“BrIDGs”). BrIDGs offers academia the opportunity to use the contracted infrastructure to help them overcome a specific problem in the development of a new drug,

such as through in vivo screening, initial production of cGMP materials, scale-up of synthesis, and the like. Without the DEC, no one would submit their materials to BrIDGs if the commercial contractor would have rights in any subsequent invention. Despite having used DECs successfully, the NIH still treats them as the name suggests: truly exceptional. Congress clearly believed that Contractors should be entitled to these rights, even if that arrangement frustrates a government program’s ability to conduct its research or otherwise offends anyone’s personal philosophy on whether Contractors should get those rights. The only justification for a DEC is that changing the rules better promotes the policy and objectives of the Bayhe Dole Act than not changing the rules. “March-In”d§ 203 The BayheDole Act includes one major threat against Contractors who fail to ensure their patents on Subject Inventions are properly utilized: the government’s “March-In” rights. Under 35 U.S.C. x 203, the funding agency may either force the Contractor (or its assignee or exclusive licensee) to offer (sub)licenses in a Subject Invention to responsible applicants in any field of use, or the funding agency may grant those licenses directly, if the funding agency finds one of four conditions have been met. These conditions are as follows: 1. the contractor “has not taken or is not expected to take within a reasonable time, effective steps to achieve practical application of the Subject Invention in such field of use;” 2. March-In is “necessary to alleviate health or safety needs which are not reasonably satisfied by the contractor;” 3. March-In is “necessary to meet requirements for public use specified by federal regulations and such requirements are not reasonably satisfied by the contractor;” or 4. either an exclusive license does not contain the obligation for US Manufacture (x 204) or the exclusive licensee has breached that obligation. The funding agency cannot simply invoke this clause and start issuing licenses; rather, 37 C.F.R. x 401.6 establishes detailed procedures that the agency must follow. These procedures include giving the Contractor time to comment on any proposed invocation of March-In, holding a formal fact-finding hearing, issuing a written agency opinion, and allowing the Contractor to appeal adverse decisions to the courts. From 1980 to 2016, formal petitions for an agency to exercise March-In rights have been filed only a handful of times and only involving one federal research agency, the NIH. Each time the petitions were denied, and the reasons illustrate the difficulties petitioners face.253 As

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

of the writing of this chapter, one additional petition remains open, filed in January 2016 with both the NIH and the Army, concerning the expensive prostate cancer drug Xtandi. A review of three of the petitions resolved by NIH will help explain the legal contours of March-In. Case Study: CellPro In the early 1980s under a grant from the NIH, the Johns Hopkins University (JHU) invented (and later received patents on) an antibody-based stem-cell separator to process peripheral blood and bone marrow for diagnostic and therapeutic uses. JHU exclusively licensed the technology to Becton-Dickenson & Co., which exclusively sublicensed the therapeutic uses to Baxter Healthcare Corp.; Baxter developed a prototype marrow reconstitutor in 1991, began clinical trials in 1993, and received regulatory approval in Europe in 1995. By 1997, when Baxter finally filed a PMA254 with the US FDA, Baxter’s separator had been installed in numerous transplant centers across the United States for use in clinical studies. Shortly after JHU had received its patents, researchers at the Fred Hutchinson Cancer Center developed their own antibody-based stem-cell separator, formed a spin-off company (CellPro, Inc.) in 1989, which began selling a prototype device for marrow reconstitution for research purposes in 1991. CellPro sought and secured FDA approval to market its product in December 1996 less than 3 months before Baxter filed its NDA. News articles indicate that Baxter initially offered to sublicense to CellPro,255 but whatever may have happened in the negotiations, CellPro decided the company did not need a license from JHU. So, Baxter sued CellPro for patent infringement in 1995. In March 1997, a jury found CellPro had willfully infringed JHU’s patents; in a scathing ruling, the district court tripled the damages (more than enough to put CellPro out of business) and ordered CellPro to cease selling its units once Baxter’s units are approved by the FDA. The Court of Appeals later affirmed this decision. Days before the jury handed down its verdict, CellPro, working with former Senators Birch Bayh and Lloyd Cutler (a cosponsor of the BayheDole Act) filed a petition with the Department of Health and Human Services invoking the March-In clause of the BayheDole Act. The petition raised two grounds: (1) Baxter had failed to take reasonable steps to bring the technology to practical application and (2) March-in is necessary to alleviate a health and safety need not reasonably satisfied by Baxter. In support of both arguments, CellPro relied heavily on the fact that Baxter had not yet received FDA approval to market, while CellPro had. What would happen, CellPro asked, to all the cancer patients in need of this technology if CellPro were shut down?256

491

The Secretary delegated the petition to the NIH, as the source of the original grant, to resolve. In addition to briefs and supporting documentation from the two parties, the NIH also received letters, briefs, and other items from others interested in the outcome. After considering all of this information, the NIH concluded that holding an evidentiary hearing to investigate invoking the March-In clause was not warranted. Regarding CellPro’s first argument, the NIH noted that the statute’s standard is that JHU, BectonDickinson, and Baxter “[have] not taken, or [are] not expected to take within a reasonable time, effective steps to achieve practical application of the Subject Invention.” The law does not require success, merely “effective steps.” The NIH recited all of the actions taken by JHU, Becton-Dickinson, and Baxter since the invention was first discovered, culminating with the fact that Baxter’s prototype was already in use in numerous centers across the country, that Baxter was actively pursuing FDA approval and was already selling the product in Europe. Regarding CellPro’s second argument, the NIH noted the fact that the medical community had not yet settled the debate over whether this new technology produced substantially better clinical results than existing marrow transplantation techniques. As such, the NIH reasoned that imposing its judgment on the medical community would be unwarranted and inappropriate. Observing that Baxter refrained from asking the court to order the immediate withdrawal of CellPro’s product before Baxter’s product received FDA approval, and that Baxter’s device would be available widely through the clinical trials, the NIH concluded that cancer patients truly needing this technology would have a viable product available to them continuously. Therefore, Baxter had reasonably met the health and safety need. In concluding the discussion on the second argument, the NIH sounded a clear, cautionary note to the general public that the March-In system is not a fallback mechanism for those who lose a patent infringement lawsuit. The NIH stated: We are wary, however, of forced attempts to influence the marketplace for the benefit of a single company, particularly when such actions may have far-reaching repercussions on many companies’ and investors’ future willingness to invest in federally funded medical technologies. The patent system, with its resultant predictability for investment and commercial development, is the means chosen by Congress for ensuring the development and dissemination of new and useful technologies. It has proven to be an effective means for the development of health care technologies. In exercising its authorities under the BayheDole Act, NIH is mindful of the broader public health implications of a March-In proceeding, including the potential loss of new health care products yet to be developed from federally funded research.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

492

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

Case Study: Abbott and Pfizer This case study actually involved two petitions filed together but concerning two independent drugs, Norvir and Xalatan. According to the petitioners, the two cases shared the common theme that they were too expensive. Between 1988 and 1996, Abbott Laboratories invented, patented, developed, and received FDA approval to sell a new anti-HIV protease inhibitor, Norvir (ritonavir), partly using NIH funds to support the research. In December 2003, Abbott was nearing the point of seeking FDA approval for its new combination anti-HIV drug, Kaletra. In an apparent effort to boost sales of Kaletra, Abbott raised the wholesale price of Norvir from $1.71/day to $8.57/day, making Kaletra considerably less expensive by comparison. Meanwhile, in 1996, a scientist at Columbia University, working under a grant from the NIH and in collaboration with a scientist at Pharmacia, invented latanaprost, a topical treatment for glaucoma. Columbia licensed the invention to Pharmacia (later Pfizer), which then developed Xalatan. When Pfizer began selling Xalatan abroad, however, the price in Canada and Europe was substantially lower than the price in the United States. In 2004, an advocacy group called Essential Inventions submitted two petitions to the Secretary of Health and Human Services (HHS) to invoke March-In, one for Norvir and one for Xalatan. These petitions both invoked the “effective steps to achieve practical application” and the “necessary to alleviate health and safety needs” bases for invoking March-In, and in both cases, argued that pricing was the cause. As in the CellPro case, HHS delegated the petitions to the NIH for determination. The NIH invited speakers from the petitioner, the companies, the public, and Congress to come to the NIH to discuss the two cases in a public forum. Also, as with the CellPro case, many parties submitted written testimony, letters, briefs, and other documents for the NIH to consider. After reviewing all of this material, the NIH decided to deny both petitions. Regarding the first March-In basis, “effective steps to achieve practical application,” the NIH noted that Abbott and Pfizer each had indeed put their respective drugs on the market. Regardless of price, actual marketing is a “practical application.” Regarding the second basis, “necessary to alleviate health and safety needs,” the NIH noted that the drugs are available in sufficient quantity to meet demand, that it is being widely used, and that no evidence was presented that invoking March-In would in fact alleviate any unmet health or safety need. For either drug, the NIH was unwilling to use March-In merely as a tool for interfering in market dynamics and for controlling pricing.

The NIH’s decision did not, however, dissuade the advocates for aggressive use of March-In for expensive HIV drugs. In October 2012, Knowledge Ecology International (KEI), along with three other groups, filed a petition with the NIH to grant open licenses to the public to the Norvir patents based on the fact that the drug is between 4x and 10x as expensive in the United States as it is in other high-income countries. In addition, the petition asked NIH to set two rules: (1) any drug utilizing an NIH-funded invention cannot exceed certain averages of these high-income countries and (2) if the drug exceeds that price and the manufacturer will not grant a license to others, the NIH will do so. In November 2013, the NIH rejected that later petition. Regarding most of the petition, the NIH concluded it relied on the same information or arguments, and without anything new, there was no basis to reconsider the prior outcome. The one new argument made in that petition, that the Americans with Disabilities Act (“ADA”) and the Patient Protection and Affordable Care Act (“PPACA”), both of which imposed certain obligations relating to insurance coverage, did not apply because neither statute required the use of ritonavir. As for granting open licenses, the NIH noted that Hatch-Waxman already provides a means for achieving those ends, which generic companies are already using. Case StudydGenzyme Fabry disease is a rare genetic disorder in which the body is unable to make a key metabolic enzyme causing severe symptoms and death. The only treatment is regular injections of that enzyme. In the early 1990s, researchers at Mount Sinai School of Medicine, working under an NIH grant, discovered and patented a way to make pure recombinant enzyme efficiently. The technology was exclusively licensed to Genzyme, which introduced Fabrazyme to the market in 2003. Unfortunately, in 2009, manufacture was interrupted because of facility shortcomings, and in May 2010, Genzyme entered a Consent Decree with the FDA detailing how Genzyme will restore quality compliance, which included building a new facility. Because of that interruption, patients had to make do with severely reduced dosing, which proved ineffective in staving off progression of the disorder. Based on the criteria that action is necessary to alleviate health needs, and that Mount Sinai and Genzyme are not reasonably able to satisfy that need, patients suffering from Fabry disease petitioned the NIH in August 2010 to exercise March-In and grant an open nonexclusive commercial license to anyone who wanted it. The NIH decided in December 2010 to deny that petition reluctantly. While the facts clearly supported an unmet health need, the problem in granting the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

petition is that it would not help. Manufacturing a biologic is difficult and the process for securing FDA approval lengthy, but that process is exempt from patent infringement (see the discussion of the Bolar amendment, above). The patents, meanwhile, were set to expire in September 2015. Therefore, even if NIH granted the petition immediately based on available information no one would be able to reach the market in time for Mount Sinai and Genzyme to enforce the patents against them. This time, however, the NIH took one additional step. Acknowledging that this decision was purely based on specific facts, the NIH indicated it will require Mount Sinai to provide regular reports of Genzyme’s progress, watch for relevant legal and regulatory developments, and encourage anyone to submit information suggesting someone could reach the market more quickly than believed. If anything suggested a substantial change in circumstance, the NIH would reopen its decision and modify it accordingly. No such change happened, however, and so in February 2013, as Genzyme restored full manufacturing, the NIH fully closed the petition. Duty of US Manufactured§ 204 Under 35 U.S.C. x 204, Congress imposed on every Contractor (and, if the Subject Invention is ever assigned, on any assignee) a “Preference for United States Industry.” Specifically, the Contractor may not grant anyone the exclusive right to use or sell a Subject Invention in the United States unless the licensee “agrees that any products embodying the Subject Invention or produced through the use of the Subject Invention will be manufactured substantially in the United States.” The phrasing of x 204 includes two major zones where the law does not have any effect. First, the law is limited to exclusive licensing: nonexclusive licensees of the Contractor, and the Contractor itself, may manufacture a product embodying or made using a Subject Invention anywhere, even if foreign-made products later will be used and sold in the United States. Second, the Contractor is free to grant exclusive rights to use and sell a Subject Invention without requiring that the product be made in the United States, as long as those products will never enter the United States, nor used to provide services in the United States. Where the US manufacturing obligation does apply, the funding agency may, on a case-by-case basis, waive it if the Contractor demonstrates that either (or both) of two situations are true: 1. “reasonable but unsuccessful efforts have been made to grant licenses on similar terms to potential licensees that would be likely to manufacture substantially in the United States” or

493

2. “that under the circumstances domestic manufacture is not commercially feasible” The lone paragraph in the regulations implementing x 204 (37 C.F.R. x 401.14(i)) merely repeats the statutory phrasing; it does not shed any light on what efforts are “reasonable,” what percentage of total manufacturing represents “substantially in the United States,” or what circumstances demonstrate commercial infeasibility. Although such ambiguity makes the task of demonstrating either prong challenging, it nonetheless helps Contractors by allowing the agency to be flexible. Funding Agreements Outside the BayheDole Act Involving Patent Rights Two specific kinds of agreements for funding private R&D are not governed by the BayheDole Act based on specific statutory authorization. The first is the Space Act Agreement, which only the National Aeronautics and Space Administration (NASA) may utilize. The other is called the “Other Transactions” agreements, which NASA, Departments of Defense (DoD), Energy (DoE), Transportation (DoT), and Homeland Security (DHS) may utilize. The terms of the BayheDole Act do not apply to these two types of agreements, so in theory, an agency and its private-sector partner can dispose of patent rights, however, they please. Each agency has established its own rules governing the use of these agreements. Federal “Intramural” Research: The StevensonWydler Act and the Federal Technology Transfer Act By Executive Order257 and by regulation,258 every employee of the federal government must assign ownership of any invention made within the scope of employment.259 Also, as discussed above, the government may (depending on the decisions of the Contractor) receive title to inventions made under a funding agreement subject to the BayheDole Act, and the government will take title if that funding agreement has been made subject to a DEC. As a result, the government over time has collected a portfolio of inventions.260 While the BayheDole Act created the basic authority for federal agencies to patent and license their intramural inventions, federal technology transfer embraces a broader scope reflected in two additional laws: the Stevenson-Wydler Technology Innovation Act of 1980 (better known simply as the “Stevenson-Wydler” Act) and the Federal Technology Transfer Act of 1986 (FTTA).261 They, together with the federal patenting and licensing clauses of the BayheDole Act, form the core of the structure of technology transfer for federal agencies.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

494

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

History and Philosophy of Stevenson-Wydler and Federal Technology Transfer Act Just 2 months before Congress passed the BayheDole Act, Congress passed the Stevenson-Wydler Act.262 Almost exactly 6 years later, Congress passed a major amendment to the Stevenson-Wydler Act known as the FTTA. Even more directly than BayheDole, Stevenson-Wydler and the FTTA were designed to unlock the commercial potential of federal research funding, but this time focusing more on tapping the informational, intellectual, and technical resources created by agencies from their in-house activities. These two acts, along with subsequent amendments, are together codified as Chapter 63, “Technology Innovation” (15 U.S.C. xx 3701-3718). As Stevenson-Wydler arose in an era where many senior federal employees believed the government should not participate in industrial development at all, it was designed to change the culture of the agencies by requiring agency staff to participate in technology transfer activities. Specifically, the following policy applies to all agencies: 1. It is the continuing responsibility of the federal government to ensure the full use of the results of the Nation’s Federal investment in research and development. To this end the federal government shall strive where appropriate to transfer federally owned or originated technology to state and local governments and to the private sector. 2. Technology transfer, consistent with mission responsibilities, is a responsibility of each laboratory science and engineering professional. 3 .Each laboratory director shall ensure that efforts to transfer technology are considered positively in laboratory job descriptions, employee promotion policies, and evaluation of the job performance of scientists and engineers in the laboratory.263 The FTTA, in its turn, joined the effort of changing culture by explicitly requiring agencies to share their royalties with the inventors and by authorizing agencies to use royalty funds to grant cash awards to staff. Key Concepts and Major Clauses Stevenson-Wydler accomplished several key tasks. First, it required each agency to establish Offices for Research and Technology Applications (ORTAs), or the equivalent. ORTAs are tasked with, among other things, assessing the research portfolio of its agency, promote dissemination of information about federally developed technologies that might be useful to industry, and cooperate with and provide technical assistance to state and local governments. Second, it established the Federal Laboratory Consortium to enable agencies with research

labs to have a forum in which to work together. Third, it authorized agencies to create Cooperative Research Centers in universities or nonprofits, which can combine federal funds with university funds to promote the development of a local infrastructure able to train a pool of talent, start companies, and engage in collaborative research. Unfortunately, government scientists still were either unable or unmotivated (often both) to collaborate with industry counterparts. In particular, inventors received no more than a token reward for reporting inventions (which consumed time from research), and industry would not touch a collaborative project without a binding promise that they would be able to license government inventions arising from the project. In light of these issues, Congress passed the FTTA. This law primarily created a new form of agreement called the Cooperative Research and Development Agreement, or “CRADA,” discussed at length later in this chapter. In short, a CRADA allows an agency to promise a collaborator that the collaborator at a minimum will have certain rights in any invention made under the CRADA that is owned in whole or in part by the government. Also, the FTTA allows the Contractor to contribute funds, personnel, services, and property to the agency, and to allow the agency to contribute personnel, services, and property (but not funds) to the collaborator for use in fulfilling the CRADA research. The FTTA also required agencies to share their royalties with the inventor(s); currently, the inventors’ share of the royalties on any license (after patent expenses) is no less than the first $2000 collected in a calendar year, plus 15% of everything above that, but no inventor may receive more than $150,000 in any calendar year regardless of the number of inventions or licenses associated with that inventor. Finally, the agency is entitled to keep royalties for up to two full fiscal years, and to use its royalties to reward staff, and to support further research, scientific exchange, and educational activities. Subsequent Supporting Acts Since 1986, Congress amended Stevenson-Wydler several more times but two particularly stand out. These are the National Technology Transfer Advancement Act of 1995 (NTTAA)264 and the Technology Transfer Commercialization Act of 2000 (TTCA).265 The NTTAA, better known as the “Morella Bill,” accomplished several things, two of which had broad effect across agencies. The more important of these two changes was an overhaul of the terms governing CRADAs, expanding the scope of authority the federal laboratory has in contributing and accepting resources, and

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

clarifying the rights a collaborator has in inventions made under the CRADA in whole or in part by federal employees. The other key change was to increase the inventors’ share of royalties collected by an agency raising the cap to $150,000. The TTCA mainly changed how the federal government licenses its patents but also included some small but important changes to CRADAs. On licensing, the TTCA added substantial detail to the obligations of an agency before it may grant a license. Mainly, these obligations include requiring a development plan (for both exclusive and nonexclusive licenses) and ensuring that every exclusive license meets the listed policy factors. Also, the TTCA subtly changed the scope of what the agencies may license, changing “patents and patent applications” to “inventions.” On CRADAs, the statute authorized an agency, if it chose to do so, to explicitly extend the collaborator’s license rights to capture patent applications filed before the CRADA was signed as long as the invention is directly within the scope of work under that CRADA. Patenting and Licensing by Federal Agencies Patenting and Licensing by Agency Various Agency Missions Across the government, 25 agencies fund R&D; 9 departments and 2 independent agencies conduct and support enough research to be required to report on their technology transfer activities. These are the Departments of Agriculture (USDA), Commerce (DoC), Defense (DoD), Energy (DoE), Health and Human Services (HHS), Homeland Security (DHS), Interior (DoI), Transportation (DoT), and Veterans Affairs (VA), plus the National Aeronautics and Space Administration (NASA) and the Environmental Protection Agency (EPA).266 Each is unique, with a specific mission assigned to it by Congress, and some have subagencies with their own statutory mission. That mission, in turn, drives the research the federal labs conduct, as well as the products the lab’s agency deems suitable for patent protection. For example, the respective missions of the NIH,267 USDA,268 NASA,269 DoE,270 and EPA271 look nothing like each other. Differences in mission also change the patenting strategy by considering the sorts of inventions each agency makes available for licensing. For example, currently the NIH and USDA mostly offer very similar types of inventions, such as methods of diagnosing pathogens, vaccines, and recombinant genes and proteins. Yet, each of the two agencies has a segment that the other does not significantly pursue, for example, the USDA has some inventions directly aimed at increasing industrial output or efficiency, while the NIH has a portfolio of

495

inventions on infectious diseases that have nothing to do with agricultural products. Naturally, the USDA’s focus is on livestock health and environmental technologies but also aspects of human health to the extent it affected by agricultural products, while the NIH’s focus is on human health but also aspects of animal biology and environmental technologies to the extent they relate to human health. Broadening the view to look at EPA, DoE, and NASA, the overlap becomes substantially narrower. EPA’s range of technologies, while strongly influenced by human health, concerns technologies directly focused on the environment itselfdair, water, and soil sampling techniques, for example. There is partial overlap in that some of the techniques for testing the environment are also useful for diagnosing humans and vice versa. DoE’s portfolio covers the entire span of physical sciences from new materials to high-energy physics to environmental sciences. A substantial subset of DoE’s patented work includes life sciences (DoE was a founder of the Human Genome Project), but the focus tends to be on clean energy, effects of radiation, the environmental basis for biological functions, carbon sequestration, and environmental cleanup. NASA, in turn, mostly investigates materials and devices useful in space flight, tools for use in its various missions, and technologies relating to the environment and weather. Scope of Licensing Authority All federal agencies rely on the same core legal authority to license patentable inventions: 35 U.S.C. xx 207 and 209 along with the implementing regulations at 37 C.F.R. Part 404. Agencies may grant exclusive, coexclusive, or nonexclusive licenses, which may be royalty-free or royalty bearing. The agency may include any terms the agency determines are appropriate in the public interest including granting the licensee the power to enforce the patent. Agencies must normally impose on its licensees a requirement that goods that are intended to be used or sold in the United States must be manufactured substantially in the United States similar to the requirement placed on Contractors under BayheDole. All agencies must include certain standard terms in every license. First and foremost, the license must include a retained nontransferrable, irrevocable, paidup license for any federal agency to practice the invention or have the invention practiced throughout the world by or on behalf of the government of the United States. In addition, every license must include a requirement of periodic reporting on utilization efforts and rights of the agency to terminate the license.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

496

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

Exclusive and Coexclusive LicensingdAdditional Considerations Before granting any exclusive or coexclusive licenses (except for those that arise under a CRADA), an agency must notify the public of its intent to grant the license a minimum of 15 days before granting the license and must consider all comments received before the comment period expires. Also, the applicant for such a license must submit a plan for developing and/or marketing the invention. This plan is kept confidential as a trade secret. An agency may grant exclusive or coexclusive licenses to its inventions only if the following conditions are met: • granting the license is a reasonable and necessary incentive to call forth the investment capital and expenditures needed to bring the invention to practical application, or otherwise promote the invention’s utilization by the public; • the agency finds, based on the applicant’s plan for the technology, that the public will be served by the granting of the license; • the proposed scope of exclusivity is not greater than reasonably necessary to provide the incentive for bringing the invention to practical application, as proposed by the applicant, or otherwise to promote the invention’s utilization by the public; • the applicant makes a commitment to achieve practical application of the invention within a reasonable time, which time may be extended by the agency on the applicant’s request and the applicant’s demonstration that the refusal of such extension would be unreasonable; • granting the license will not tend to substantially lessen competition or create or maintain a violation of the federal antitrust laws; and • in the case of an invention covered by a foreign patent application or patent, the interests of the federal government or US industry in foreign commerce will be enhanced. An agency deciding which company will receive an exclusive or coexclusive license must preference small businesses having equal or greater likelihood as other applicants to bring the invention to practical application within a reasonable time. Results Every year, the 11 federal agencies listed above provide data in a consolidated report to the President and Congress. The report includes both survey data and recent illustrative success stories. While these numbers in these reports should be viewed with caution, as the figures swing substantially from year to year, they do give a partial glimpse of the general level of activity.

The agencies’ FY2014 Report,272 which covers 2010e14 data, shows that the participating agencies collectively tallied an average of about 5100 new invention reports per year, filing about 2300 US patent applications and receiving 1800 issued patents (per the Congressional Budget Office, cumulatively from 1990 to 2005, the government received about 15,000 patents273). By volume of inventions, NASA, DoE, and DoD were the top three agencies with HHS a distant fourth. Looking at the licensing of inventions,274 however, the overall average rate was just over 470 newly executed licenses per year, with HHS leading the others, followed by DoE and USDA, the three of which together accounting for about 90% of all invention licenses. As of FY2014, about 11,000 invention licenses remain active. For royalties on its inventions, on average the government received an average of over $180 million per year, with HHS collecting nearly $110 million per year, DoE over $40 million, and DoD over $13 million. And all of that money (less a substantial fraction shared with the inventors) was plowed back into additional research that the US taxpayers did not have to support. Of course, patents and royalties are both useless benchmarks of how the public has actually benefitted from research by federal scientists. Unfortunately, no good metric has been developed to measure public benefit. So, agencies must rely on narratives of success stories. Examples of these include the following: • The National Wildlife Research Center (USDA) collaborated with the aviation industry to develop external lighting systems for aircraft both to enhance the pilots’ ability to detect birds and to help birds better see aircraft. • The National Institute of Standards and Technology (DoC) set up a consortium of 79 members to initiate a public-safety broadband network designed to benefit emergency-services agencies nationwide to improve communication among first responders. • An Army R&D lab (DoD) invented the “HyperX” multicore parallel processor for soldier-wearable devices to enable storing, processing, and retrieving massive amounts of data using low power, and through a series of SBIR contracts, converted the technology for civilian uses. • The Pacific Northwest National Laboratory (DoE) invented a novel vanadium redox battery that has dramatically improved operating temperature range, higher energy density, and lower cost than prior redox technologies. • Discoveries at the Centers for Disease Control and Prevention, CDC (HHS) led to a rapid, highly sensitive diagnostic panel for the H1N1 influenza virus (the flu epidemic of 2009) affordable in countries with developing economies.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

• The Federal Aviation Administration (DoT) created “Concept Development and Design Description of Electronic Flight Data Interfaces for Airport Traffic Control Towers,” which promises to increase a controller’s ability to acquire, track, and record information, as well as communicate and coordinate that information with others. • Out of EPA research into byproducts of water disinfectants, agency scientists discovered that magnets can stimulate tissue growth in a manner that could improve bone density and help in wound healing. • Researchers at the Kennedy Space Center (NASA) developed an electromigration technique that sends corrosion-inhibiting ions into rebar to prevent rust, corrosion, and separation from the surrounding concrete. This excerpted list, obviously, represents only the smallest fraction of the results of technology-transfer work done at the agencies, but it offers a glimpse of the range and potential impact of that work. Inventions by the National Institutes of Health In the context of clinical research, the federal agency making most of the inventions is the NIH. While other agencies also engage in clinical research, to a large extent their views are informed by the NIH’s policies. Accordingly, this chapter will look at how the NIH manages its invention portfolio. The primary mission of the NIH’s research laboratories is to seek fundamental knowledge about the nature and behavior of living systems and the application of that knowledge to enhance health, lengthen life, and reduce the burdens of illness and disability. Among the NIH’s core goals that are most pertinent to technology transfer are the following: 1. To foster fundamental creative discoveries, innovative research strategies, and their applications as a basis for ultimately protecting and improving health; 2. To develop, maintain, and renew scientific human and physical resources that will ensure the Nation’s capability to prevent disease; and 3. To expand the knowledge base in medical and associated sciences to enhance the Nation’s economic well-being and ensure a continued high return on the public investment in research. The NIH views all of its patent-related decisions through this lens. Patent and Patent-Related Policies General Most technologies developed at the NIH are made available to the public without securing patent protection. This “transfer” is done mainly through

497

publications in scientific journals, public presentations, online resources, and, for most tangibles like software and biological materials, through licenses for commercial development or research. In some cases, however, the NIH determines that seeking patent protection on its inventions is necessary to facilitate the commercial development of products or services that will benefit the public health. The most obvious instance is where no one will invest in developing the technology without the exclusivity of a patent. Primarily, this situation arises with candidate drugs and biologics, cutting-edge medical devices, and vaccines. Occasionally, other technologies (e.g., diagnostic kits) also require exclusive marketing rights as an incentive to invest in turning a discovery into a product. In deciding whether to pursue a patent, the NIH also typically looks at a variety of specific common factors. For example, even if the NIH determines that a technology is patentable, the NIH may decline to seek patent protection due to low public health or commercial priority; however, programmatic goals relating to small markets (for example, rare and neglected diseases) may suggest that patent protection is warranted. In other cases, a patentable technology might be sufficiently developed that it can be utilized immediately through publication alone, such that patents might actually impede broad dissemination or rapid utilization. The decision to file is also informed by other NIH policies, including, for example, the NIH policies on research tools and genomic inventions discussed below. Additionally, in accordance with longstanding traditions of scientific freedom, the NIH’s research results are to be published rapidly; publication will not be significantly delayed merely for the purpose of filing patent applications. Research Tools In December 1999, the NIH published its policy on “Research Tools.”275 While not the first statement describing the NIH’s views on the importance of disseminating the results of research supported by the NIH, the Research Tools policy represented the most specific statement to date, and it remains a benchmark for both the NIH and institutions funded by the NIH. The Research Tools policy is intentionally broad in scope,276 articulates four core principles, and urges widespread adoption of a one-page “Simple Letter Agreement” to authorize the transfers of research materials for nonclinical research. The first principle articulated by the Research Tools policy is to ensure academic freedom and publication. When a researcher’s work is supported by the NIH, the NIH expects that researcher to avoid imposing restrictions on others’ work or even to avoid accepting such restrictions when imposed by others. Therefore,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

498

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

restrictions and prohibitions on when, where, and what research may be published are discouraged. Obviously, many collaborations, particularly those with industry, must be handled to accommodate specific needs, but these accommodations must be minimal and limited. For example, a term delaying publication long enough to file a patent application is fine, but a term delaying publication indefinitely is not. The second principle is to ensure appropriate implementation of the BayheDole Act. While recognizing that the BayheDole Act gives Contractors full discretion to use patents to develop their technologies as appropriate to the case, Contractors should recognize that patents and exclusive licensing are not the only, and in some cases not the most appropriate, means for using and disseminating research results. Indeed, Contractors must realize that these tools can sometimes thwart, rather than promote, utilization. For inventions comprising research tools in particular, exclusive licensing of patents (such as to a for-profit sponsor for exclusive internal use) can be antithetical to the goals of the BayheDole Act. The third principle is to minimize administrative impediments to academic research. Contractors are urged to harmonize their practices as much as feasible, for example, by adopting the Simple Letter Agreement or other standardized material transfer agreement (MTA). Also, Contractors should try to develop and implement clear policies to articulate acceptable conditions for acquiring resources and refuse to yield on unacceptable conditions. For-profit companies,277 in turn, should minimize the encumbrances they seek to impose on nonprofit organizations for the academic use of their tools. This means not imposing terms to control publication, not over-valuing (and, hence, overcharging for) a tool and not demanding rights (even just a royalty stream) in the nonprofit’s future inventions. Companies may legitimately impose terms to avoid losing control of a proprietary tool to a competitor but not much more. Finally, the fourth principle is to ensure dissemination of research resources developed with NIH funds. Progress in science depends on prompt access to the unique research resources that arise from biomedical research laboratories throughout government, academia, and industry. Limiting the use of research tools to one or a small number of institutions creates a great risk that fruitful avenues of research will be neglected. Accordingly, Contractors are expected to share their tools and to manage interactions with third parties that have the potential to restrict Contractors’ ability to disseminate research tools developed with NIH funds. When sharing with for-profits, Contractors may distinguish

internal use from the right to commercial development and sale or provision of services. Sharing of Data and Model Organisms In the early 2000s, the NIH decided to issue more specific policies on the obligations of Contractors to share the results of their NIH-supported research. So, on February 26, 2003, the NIH issued its “Final Statement on Sharing Research Data,”278 which was followed on May 7, 2004, by the “Policy On Sharing of Model Organisms for Biomedical Research.”279 Both statements clearly indicated that they were consistent with all prior statements of NIH’s sharing policies and merely provided certain specific extensions. The major extension in the Data Sharing policy was that Contractors asking for more than $500,000 of direct costs in any year must include a plan for sharing data or else explain why sharing is not possible. The Data Sharing policy recognizes that data sharing may be complicated or limited in some cases by institutional policies, by local IRB rules, and by local, state, and federal laws and regulations. When the sharing of data will be limited, applicants should explain such limitations in their data sharing plans. The major extension in the Model Organisms policy is that those researchers submitting proposals that specifically anticipate the creation of an organism (whether it is a mammal, a fish, a fly, a fungus, or a bacteria) that will serve as a model of a human disease or condition, the researcher must include a plan for sharing the organism or else explain why sharing is not possible. This policy carefully notes that it is not subject to the $500,000 threshold; all applications must comply. When the applicant submits a plan, it must address three topics: (1) Will the terms of any MTA be no more restrictive than the Simple Letter Agreement? (2) How will terms that are inappropriate under the NIH’s Research Tools policy be addressed? (3) What will be done to ensure the model organism will remain widely available and accessible to the research community? For both policies, the NIH will evaluate and rank applications without regard to the particular content of the plans submitted, but if an applicant fails to submit a sharing plan or if the applicant refuses to share without providing an explanation, the NIH may require one before awarding the funding agreement. To help applicants, the NIH developed some examples of sharing plans and made them available on the relevant webpages. While these policies and procedures only directly apply to Contractors funded by the NIH, the principles they embodydsharing of data and research resourcesd applies to Intramural research equally. Consequently,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

the terms of any licenses to NIH-owned technologies will be influenced to the extent those principles apply. National Institutes of Health Portfolio Size and Scope The NIH directly employs roughly 18,000 people, plus over 3000 nonemployees (fellows, visiting scientists, and trainees) at any given timedany of whom might be an inventor of something even if they are not normally engaged directly in research. Topics of scientific investigation cover the entire spectrum of biomedical study, including such topics as drug discovery, bioinformatics software, animal research, imaging devices, environmental issues, behavior, pain, and infectious diseases. These 21,000 people typically generate about 400 new invention disclosures every year, and many of these invention disclosures name more than one inventor (not all of whom work at the NIH). As of October 2015, the responsibility for prosecuting patents and negotiating licenses on inventions owned by the NIH falls to the Institute or Center (IC) from which the invention arose. Once a patent license has been executed, the NIH Office of Technology Transfer (NIH OTT) monitors, enforces, and (as needed) amends those agreements on behalf of the ICs, as well as processes royalty payments. All told, the NIH has about 200 staff dedicated to technology transfer activities. In the course of fiscal years 2003e15, the NIH filed about 4400 original applications, including PRVs, PCTs, 111s, and CIPs; adding DIVs, the annual total rises to about 5000 applications. As of the end of fiscal year 2015, the NIH owned or co-owned approximately 2000 issued US patents and a worldwide total of about 6400 patents. Based on 2011e15 activity, the NIH OTT annually negotiates and executes about 250 license agreements (including amendments that changed the substantive scope of the license), monitors about 1500 active licenses, and collects over $100 million in royalties. While the exact figures fluctuate from year to year, the trends have remained relatively mostly steady, with the main exception of royalties, which is volatile. The National Institutes of Health Licensing Program National Institutes of Health General Licensing Policies Just as in deciding whether or not to apply for patents, the NIH strives to license280 its intramural inventions in ways designed to facilitate further research and commercial development of the inventions. As noted, having exclusive rights is clearly critical where premarketing regulatory burdens are substantial, and it also could be critical to encourage a company to keep important research materials available that would not be commercially viable through mere nonexclusive licensing. Where such facts are present, exclusivity could be best. At the same time, nonexclusive licensing has several advantages over exclusive in many cases. First,

499

nonexclusive licensing often enhances public benefit by providing each licensee an incentive to develop its product expeditiously and to make better quality or lower cost versions. Second, nonexclusive licensing promotes the development of second-generation products by multiple competitors. Third, in the biomedical arena, it may represent alternative products and services from which doctors and patients can select the optimal version for a given situation. Fourth, having multiple providers can increase consumer access to the product. Each case must be examined to see whether any of these advantages applies and, if so, whether it is outweighed by the licensee’s need for exclusive rights as an incentive. Where a patented technology requires investment of effort and money into further development, the question of whether exclusivity is necessary to encourage that investment becomes critical. For a disease clearly caused by a single mutated gene (e.g., cystic fibrosis), very little further development is needed, and broadest nonexclusivity is sufficient to encourage widespread commercial use. For a highly complex testdsuch as one involving many genes but not every mutation matters equally (e.g., hereditary hearing loss)dproving the clinical utility of any given gene or mutation might require a substantial investment. Also, where the FDA requires clinical studies before a gene-based diagnostic test can be marketed, that cost increases dramatically. Finally, some genes are involved in multiple but unrelated disease states, such that many different diagnostic tests could require access to a single gene. Even where granting an exclusive license is appropriate, the logic supporting exclusivity rarely applies to all elements or uses of an invention. For example, a technology might need investment to develop an FDAapproved product, but once published, no investment is needed simply to encourage further research on or using the discovery. Accordingly, in granting any exclusive license, the NIH does so only after a critical analysis of the licensee’s application. Additionally, the NIH always reserves certain rights and imposes certain limitations. First, the NIH reserves the right to offer some aspects in a coexclusive or nonexclusive manner to ensure that each distinct embodiment has the best chance of being developed in a timely way. Also, the NIH includes “antishelving” terms in all exclusive licenses, such that if the licensee fails to diligently develop and widely distribute the NIH’s technology, the NIH can look for another licensee. Furthermore, a critical component of all of the NIH’s intramural licenses, including those for gene-based inventions, is the reservation of the right for the NIH to continue to engage in research on its technologies, and to license that right to anyone else who wants to conduct further research, including both for-profit and nonprofit

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

500

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

institutions. Reserving research rights is a bedrock NIH licensing policy, consonant with letter and spirit of the laws authorizing licensing,281 and rooted in fundamental principles of research freedom. Because a single invention often can be utilized in multiple ways, the NIH narrowly tailors its licenses (particularly exclusive licenses) by “field of use,” a phrase referring to a particular set of activities authorized under a license. For example, a drug that could treat a variety of solid tumors might be licensed only for use in treating breast cancer, reserving other fields of use for other licensees. The field of use in any given license is based first on the nature of the prospective licensee’s request, second on the justification for exclusivity, and third according to the ability of the prospective licensee to develop the invention as promised and within a reasonable period of time. If a prospective licensee can justify and develop more than one field of use, the NIH may include multiple fields in the license. Research-related uses and clinical uses are carefully segregated, such that each can be licensed using fundamentally different strategies suited to their respective needs. In deciding on an appropriate scope within a broad field of use, a crucial fact to bear in mind is that many companies will not be able to develop the full range of applications within that field. To find out, every company requesting a license from the NIH must submit a credible and detailed commercial development plan (“CDP”), and the scope is then tailored to the licensee’s CDP in light of its abilities. For example, a small company might want a license for all cancers but realistically have resources to develop an FDAapproved diagnostic kit for only a few types (say, breast and ovarian cancers). If so, the field of use of the license will be limited to these types, leaving the NIH with the ability to license the invention to another company for other cancers (e.g., leukemia, brain, skin, liver, and bone). For “platform” technologies (i.e., technologies that can be integrated into many others), a common way of properly tailoring exclusivity is to offer a field of use limited to joining the NIH’s invention to the licensee’s own proprietary technology. An example of this is a vaccine that can be adapted to any number of delivery systems, where an exclusive license is limited to the licensee’s proprietary, disease-specific system. Finally, an exclusive license can be limited in time. A technology that has low-to-moderate risk and cost associated with its development might be licensed exclusively for a certain number of years after the effective date of the license or after-market launch. This strategy offers another way to calibrate the scope of exclusivity to the level necessary to encourage the

investment needed to achieve practical application of the technology. Every commercial license requiring product development, including every nonexclusive commercialization license, includes a series of research and commercial stages that the NIH expects the licensee to meet. The exact details always turn on the particular CDP submitted by the applicant. These research and commercialization milestones are intended to be simultaneously challenging and achievable to ensure the reasonable development of the NIH’s technologies. The negotiation of the scope of a license is linked to the CDP and appropriate milestones. If the licensee desires a broader scope to include multiple indications, the milestones and diligence requirements are tailored to address these as well. The NIH experience has been that tailoring these elements is considerably more complex than royalties, such that negotiations on CDP and milestones take more time than on the financial structure. Milestones in licenses can be tailored depending mainly on the field of use and stage of development. For licenses having therapeutic fields of use, milestones typically cover preclinical/animal studies, filing an Investigational New Drug application (IND) or Investigational Device Exemption application (IDE) with the FDA, completing Phase 1, Phase 2, and/or Phase 3 clinical trials, and filing the appropriate application requesting approval to begin marketing. For diagnostic fields of use, milestones may address preclinical/animal studies (if any are needed), regulatory filings, clinical trials (if any are needed), availability of second opinions and confirmatory testing from sources other than the licensee, and ensuring reasonable insurance coverage. Even for well-developed technologies, milestones are included in the license, which may be as simple as starting to sell the product. Sometimes, business facts underlying milestones that were true during the negotiation of a license change over time, and the license should include terms that anticipate changes. For example, what if, several years after executing the license, a larger company acquires the licensee, and the purchaser is not interested in continuing the development of the NIH’s invention? Even where corporate mergers are intended to stoke product pipelines, sometimes the acquiring company has to decide which of the licensee’s products to pursue first. Also, even in the absence of a merger, a licensee’s business model may change, such that they decide to readjust the priorities of its product development efforts. Whatever the reason, sometimes a licensee reduces its activity to a minimum, far from enough to succeed in a timely way. While the NIH expects that each licensee will reach each milestone in a timely way, the NIH is well aware

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART TWO: PATENTS AND TECHNOLOGY TRANSFER

that product development is rarely straightforward and problem-free. Consequently, where a licensee reports its activities timely, allocates resources appropriate for the technology, and otherwise makes a commercially reasonable effort, the NIH endeavors to work with the licensee to adjust milestones as needed to facilitate getting the product to the market as expeditiously as possible. In case the licensee fails to demonstrate sufficient commitment to moving the technology to market, the NIH’s licenses include terms related to termination of the license allowing the NIH to try to license the development of those indications to others. Finally, the NIH may include licensing terms designed to support its underlying mission. While companies develop products simply for profit, the NIH’s effort is aimed to improve public health. These goals are mostly consistent with each other, but to ensure the public reaps specific health benefits from a given technology, the NIH may add certain terms to some of its licenses, sometimes called “Public Return” clauses. Public Return clauses run from simple to complex and can be tailored to achieve unique public benefits from the licensed invention. In the NIH’s first effort to craft a Public Return clause, the NIH’s licensee agreed to donate some of its product (gowns coated in an antiviral barrier) to the CDC for use overseas during an Ebola outbreak. In another example, a licensee’s scientific staff participated with Alaska Ride to raise money for HIV/ AIDS research raising US$4.1 million in 1 year. Overall, the NIH’s experience with Public Return clauses has been highly successful both for the NIH and for the licensees. Best Practices for Licensing Genomic Inventions In 2004, the NIH developed a guidance document concerning gene-based technologies. This document, titled the Best Practices for the Licensing of Genomic Inventions (or, more commonly, the “Best Practices”),282 is similar to guidance documents on the same topic issued later by other organizations,283 but more importantly, it carries the same themes previously articulated by the NIH regarding other situations applied in the particular context of genomic technologies. The Best Practices describes the NIH’s preferred approach to licensing gene-based inventions. The Best Practices recognizes that the range of inventions covered is broad,284 and that any given nucleotide-based invention may have a range of uses, including therapeutic, diagnostic, prognostic, and research uses. Acknowledging all of this, the Best Practices describes various considerations on whether or not to file for patents, on general licensing strategies, and on exclusivity. As a bottom line, the Best Practices encourages licensing policies and strategies that maximize access, as well as commercial and research utilization of the technology to benefit the public health.

501

Scope of Licensing Authority As with all federal agencies, the NIH primarily relies on 35 U.S.C. xx 207 and 209, along with the implementing regulations at 37 C.F.R. Part 404, for the authority to license its inventions. The NIH, however, has another statutory source on which it may draw authority: the Public Health Services Act authorizes the NIH to “license” certain materials, regardless of whether or not they are patentable inventions. Specifically, 42 U.S.C. x 282(c) authorizes the Director of the NIH (or a designee) to “make available to individuals and entities, for biomedical and behavioral research, substances and living organisms. Such substances and organisms shall be made available under such terms and conditions (including payment for them) as the Secretary determines appropriate.” This provision provides part of the basis for the “Biological Materials License,” discussed below. Types and Structure of National Institutes of Health Licenses The NIH has crafted several categories of model licenses for general use: patent commercialization licenses (exclusive, coexclusive, and nonexclusive); patent internal-use licenses (nonexclusive only); Biological Materials Licenses (nonexclusive; may be commercial or internal-use); and Commercial Evaluation Licenses. Also, the NIH has several specialized forms of a license to handle other situations. The NIH mainly uses the Biological Materials License where the rights being granted to the licensee include access to a sample or supply of physical specimens, even if the materials are not “biological” in nature (e.g., software). A typical NIH license covers the same ground as other licenses after the statutorily required terms are factored in. Major elements include the exact grant of rights, including sublicensing and enforcement; royalties; patent prosecution (if the case is still pending); record-keeping and reporting; progress and performance milestones; term and termination rights; restrictions (if any) on fields of use and territory; and auditing. The NIH attaches certain items to most of its licenses, including instructions on how to make royalty payments, the specific agreed benchmarks and the licensee’s development plan. National Institutes of Health Licensing Processd Overview Generally, the IC managing the patenting of an invention (or, for inventions co-owned by multiple ICs, the lead IC) announces its existence as soon as practical after a patent application has been filed and notifies the public that the invention is available for licensing. At a minimum, the IC posts its marketing abstracts in the Federal Register, but ICs also uses a variety of other venues to get the word out, including public bulletin boards, an IC webpage, e-mail distribution lists, RSS feeds, and others. When someone expresses interest in taking a license, the individual is directed to complete an application,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

502

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

which in turn helps guide the selection of the appropriate model license. Depending on what is in the application, the IC may insert tailored provisions into the first draft. For nonexclusive licenses, negotiations will continue until both parties are ready to sign; for exclusive and coexclusive licenses, the IC publishes its “Notice Of Intent To Grant,” normally allowing comments and competing applications to be submitted for a specified period, typically 15 days, before engaging deeply in negotiations. Occasionally, the IC receives competing applications for an exclusive license to the same technology at about the same time. If so, the IC has several options, based on what the applicants want, including the following: carve out unique fields of use for each licensee; divide exclusivity by different countries; accept coexclusive rights; or grant exclusivity to one applicant while denying the other’s application (perhaps encourage the licensee to sublicense to the other). When an invention made in whole or in part by IC staff arose under a CRADA, this public-announcement process is eliminated; the point of using the CRADA mechanism is to grant the Collaborator the first option to negotiate a license (exclusive, partially exclusive, or nonexclusive) in the IC’s rights immediately. To be clear, exercising that option does not guarantee a license will be executed. Also, ICs still require CRADA Collaborators to submit credible and detailed CDPs and will tailor each license to the CDP. If the CRADA Collaborator declines to exercise that option, or if the license covers less than all fields of use, then the NIH will advertise the remaining rights in the invention. Of course, even if a Collaborator has rejected the option, the Collaborator retains the right under the standard patent-licensing regulations to submit an application later though perhaps in competition with other applicants. After SignaturedRoyalty Management, Monitoring, and Enforcement As the technology-transfer community widely knows execution of a license is not the end of the process but the beginning of a long-term relationship with the licensee. Accordingly, every NIH licensed including every nonexclusive licensedhas terms requiring the licensee to report on progress developing the invention. The process of overseeing license compliance is managed centrally at the NIH OTT. Reports must be provided to the NIH OTT at regular intervals at least annually. To the extent a license has milestones, the license terms also require reporting on reaching each milestone, in addition to the standard periodic reporting. Also, the NIH reserves the right to audit its licensees typically after sales volumes are large enough that errors are likely. In addition to checking financial compliance, the audit may include examining other licensee records related to the technology, including those showing whether the licensee has met

its other due diligence commitments. The NIH pays for that audit (flat fee only, rather than a percentage of recovered royalties and competitively bid each audit), unless the audit reveals a substantial short payment in which case the licensee must pay the cost. In the NIH’s view, the purpose of this oversight is to help keep the development of the technology on track. As the work progresses, the terms of a license may require adaptation to reflect events occurring after signature and lessons learned from those events. While the NIH expects its licensees to fulfill all terms of the licenses, the NIH’s oversight is targeted toward making sure the product reaches and stays on the market, rather than in simply punishing licensees for failing to meet strict timelines or royalty payments. In terms of allegations of patent infringement, the NIH believes that ability to use its IP portfolio as a tool to develop technology is diminished by infringement for several reasons. First, infringement discourages the licensing and development of government technologies if nonlicensees can practice the invention without cost. Second, competition by infringers reduces sales of licensed products and the consequent royalty income to the government. Third, infringement reduces the inventors’ incentive to report inventions. On the other hand, the high cost of litigation or the modest value of the technology can outweigh the value of vigorous enforcement of patent rights. Each case deserves careful thought and balanced actiondneither too passive nor too aggressive. Accordingly, the policy of the NIH is to enforce its patents in the manner that the NIH believes is best suited to advance the mission of the agency. Toward this end, the NIH will, as appropriate, encourage alleged infringers to license or sublicense the technology, grant its other licensees the right of enforcement as provided by 35 U.S.C. x 207(a)(2) or refer the matter to the Department of Justice for possible civil action. Success The core purpose of the technology transfer enterprise is to get technologies out of the lab and into the marketplace. Toward that end, vast sums of money and millions of person-hours of labor have been spent on the endeavor over the past two or three decades. What results? The Association of University Technology Managers, representing mainly the nonprofit and academic institutions whose research is supported by the US government, reported in 2015285 that 6363 US patents issued, 965 new commercial products were introduced to the market, 914 new companies were formed, and 4688 startup companies from prior years were still operating as of the end of 2014. The NIH has played its part in this process proudly. In fiscal years 1996e2015, the NIH collected over $1.3 billion in royalties on over 800 different products

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

on the market, which collectively represent about $75 billion of gross sales. While most of these products are research tools, 41 are products approved by the FDA for human use, 17 are FDA-approved veterinary vaccines, and one is a veterinary drug. Particular successes have included: • Synagis, a monoclonal antibody against Respiratory Syncytial Virus based on a monoclonal developed by the NIH; • Havrix, the first vaccine against Hepatitis A, was based on a collaborative project involving SmithKline Beecham, the US Army, the CDC, and the NIH National Institute on Allergy and Infectious Diseases; • Vitravene, the first and currently only antisense drug approved by the FDA, treats cytomegalovirus; • Thyrogen, the first recombinant Thyroid-Stimulating Hormone approved by the FDA, was discovered by the NIH National Institute of Diabetes and Digestive and Kidney Disorders; • Velcade, a revolutionary drug invented by Millennium Pharmaceuticals, uses a method of freeze-dry packaging invented by the NIH NCI as part of a collaborative clinical trial; and • Taxus, the drug-coated stents for use in coronary balloon angioplasty, was invented at the NIH National Institute on Aging (which later was adapted for noncoronary balloon angioplasty made by another company). The NIH also has taken a lead-by-example role in using its portfolio to ensure that the developing world has meaningful access to NIH’s inventions for diseases and conditions that strike them particularly hard. First, the NIH requires its exclusive licensees to provide in their development plans a description of how they will bring their products to these developing countries’ markets, and where appropriate, tailors the terms of the license to make that plan feasible. In addition, the NIH has directly licensed technologies to companies in the developing countries to enhance accessdtechnologies including ddI (an anti-HIV drug) and vaccines for rotavirus, dengue fever, meningococcus, typhoid fever, and varicella virus. Also, the NIH has invited nonprofit institutions to list their inventions relating to neglected diseases alongside those owned by the NIH in a free, searchable database to facilitate licensing activities.

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS Background: Hypothetical Scenario Gillian Niher, MD, PhD, has developed a stellar reputation as an up-and-coming neuronal researcher. Her

503

focus has been on therapies for neural injuries primarily peripheral nerves. From a teaching position at Smallville Medical School, she found a tenure-track position at the NIH in a lab with facilities in the NIH Clinical Center. Unfortunately, she was stuck for ideas for her next blockbuster study; although generally interested in a variety of cutting-edge technologies, she had not yet settled on one. Then, her very close college friend, Alan Prophet, PhD., came to Bethesda on a business trip and visited her. Over lunch, Alan told Gillian about his gene-therapy research at Tate State University (a private institution in Maryland that does not rely on grants from NIH to support its bioscience research, but several projects are funded by industry). Alan mentioned that Tate State sponsored “spin-off” companies for professors who invent new bioscience products. He also mentioned that he was named as a coinventor on a recently issued patent on the genetic sequence of a newly discovered neuronal growth factor. With support from Tate State, Alan and his colleagues created a small company called Polemian Dynamics to develop this gene. They had found some support from a group of venture capitalists, who received a large share of corporate control in exchange for financing. The company had already succeeded using the gene in several in vitro models. They also had recently done some toxicity and efficacy tests in injured rats and rabbits, but the results were not yet public. Alan invited Gillian to visit Polemian’s facilities, and Gillian excitedly agreed. Two weeks later, she went to the Polemian Dynamics facilities near the Tate State campus. When she arrived, Alan told her that before he could give her a tour, she would have to sign a form the lawyers drafted to make sure trade secrets stayed secret, and Gillian agreed to comply. Then, Alan showed her preliminary data that demonstrated the growth factor was surprisingly effective in stimulating neuron regrowth, either when the growth factor protein was delivered directly to the site of neuronal injury or when a plasmid incorporating the gene was applied to the extracellular matrix. Impressed with these results, Gillian saw an opportunity to establish a collaboration: Polemian’s growth factor entering clinical trials at NIH. She consulted her scientific director about the project and was pleased that he was very interested. Alan and his partners in Polemian were equally excited when she made that suggestion to them. Alan and Gillian quickly drafted a protocol for human trials, which was favorably received by Gillian’s laboratory chief and scientific director, as well as by the venture capital group. After Gillian signed some of Polemian’s forms, Polemian sent large amounts of good medical practice-grade materials for Gillian to use at NIH. The process of establishing the study appeared to be on the fast track to success.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

504

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

Soon thereafter, while reviewing the final animal study data Alan had provided, Gillian noticed two things Polemian Dynamics had missed. First, the rabbits in the “control” group (those given only blank plasmid) had no noticeable neuronal growthdthat is, the number of nerve endings was unchanged with the injection of the plasmiddbut they seemed to be improved in terms of muscle movement and strength. On closer examination of the rabbits, she found that the original injured nerve endings had in fact regrown. In contrast, all of the rabbits that received the gene had completely new nerves growing in addition to the original ones, and all the rabbits that received nothing had no neuronal stimulation at all. Something in the plasmid appeared to have activity. Second, she noticed that those rabbits receiving the gene had exuberant growth of neurons, even in regions in which all the original neurons were dead. Alan was naturally excited to hear about these observations but told Gillian to keep them quiet just long enough so that Polemian Dynamics could file a patent application. Reluctantly Gillian agreed; however, she quietly sent samples of the plasmid with and without the gene, to John Rogers, MD, a colleague of hers at Smallville, for careful analysis of the plasmid’s sequence. The clinical trials began, and over the following weeks the pair began collecting data. Then the major problems began. Alan and Gillian continued to prepare the manuscript for the paper disclosing Gillian’s discoveries, but Polemian Dynamics insisted Alan delay his efforts, telling Gillian that the delay was needed because the patent application was not yet ready. This created a problem for Gillian, who was obligated to publish her results as soon as possible. Then, while on a visit to Alan’s office at Polemian, Gillian saw some documents indicating that a patent application had already been filed by Polemian describing her discoveries, but she was not named as an inventor. Furious, Gillian quickly finished the rough draft and submitted the manuscript immediately. On learning of this act, Polemian demanded that Gillian retract the publication, return all remaining stores of the gene, and terminate the study, but Gillian refused. To make matters worse, 10 subjects in the clinical trial were experiencing something very strange. The regions of tissue receiving the gene were experiencing hypersensitivity to the point of severe pain. Histological analysis of the tissue revealed that the neurons were growing far more exuberantly in humans than in either rats or rabbits. The stimulating factor was out of control. As if matters were not bad enough, John used Gillian’s sample plasmids to generate a large quantity of gene-bearing plasmid, which he had injected into 10 undergraduate volunteers at Smallville College without securing approval from the institutional review board, acquiring

informed consent, or even controlling the quality of the materials he had injected. Six of these students have experienced the neuronal hyperplasia. Gillian is now being sued by Polemian Dynamics for breaches of their contracts, misappropriation of trade secrets, and patent infringement. Although the injured patients and students are suing Polemian for making the dangerous materials, Polemian has asked the court to order Gillian to pay Polemian’s legal bills and any judgment associated with that product liability suit on the grounds that Gillian had agreed to do so in her various contracts. The media, having heard of the Smallville incident, has placed the whole story on national news. Congress has issued subpoenas to her entire lab, asking why NIH is sponsoring secret clinical trials of unproven, dangerous genes in our nation’s children. The scientific director personally has asked her to resign. Finally, Gillian’s attorney has told her the Assistant US Attorney is investigating whether to charge her with criminal sanctions. What went wrong, and how could the tools of technology development have helped avoid these problems? By unraveling the complicated mess LISTNUM and reviewing each piece, we will illuminate the traps and show the tools that would help avoid them.

The First and Biggest Mistake: Signing the Agreements Contract Execution in General By the time most Americans have reached adulthood, they have been scolded to read all contracts before signing them, no matter how long and confusing the fine print may be. Indeed, in many cases, the documents we are asked to sign are so complicated and difficult to read that common sense demands hiring a lawyer. Nonetheless, because hiring lawyers is expensive and time-consuming, and because many are unaware of the actual risk of something going wrong, people ignore that risk and signdoften without even readingdhappy to have saved the time and money. Only later, when we need the lawyer’s equivalent of a root canal, do we ruefully ask for help to clean up the mess. Yet even if the document is simple and the person being asked to sign it has taken the time to read it, major pitfalls still lurk. For starters, if something goes wrong, who is on the hook? As a general rule, a person who signs a contract is promising to fulfill the terms of the contract.286 That promise means Gillian could be liable if the promises in the contracts she signed are not satisfied. This risk is especially dangerous if the agreement purports to make promises that the signer cannot keep, such as a promise to keep something secret that by law must be disclosed.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

A bigger problem here is “agency,” or the power to act on someone else’s behalf. If the signer purports to bind another party (typically the signer’s employer) to perform a promise, the signer truly must have authority from that party to bind it in order for the party to be bound.287 Moreover, the authority must extend to the particular type of contract: If person A has limited authority to buy groceries for person B, A may not use B’s money to buy investment bonds. Though these rules appear simple on their face, they are less simple in practice. Although the owner of a private entity can bind that entity, generally, individual employees do not have authority to bind their employers. Certain employees, namely those who occupy key offices in a corporation (e.g., president or chief executive officer) or a university (e.g., provost or dean), typically have formal written authority to bind their employers to the contracts they sign on their behalf.288 The formal authority typically appears in charters, articles of incorporation, bylaws, or employment contracts. Other times, authority is expressly delegated in a memo or other writing, such as through a power of attorney. In government agencies, each statute passed by Congress that created each agency specify which offices can bind that agency, and actual authority below that level must be formally delegated in writing. In each case, this express grant of power is called “actual” authority. Occasionally, authority to act as an agent reasonably can be inferred from the circumstances even if no actual authority exists. If the general counsel, associate dean, or senior vice president of a company or university signs a contract, others might be justified in relying on the signature, even if the individual has no written delegation to display.289 This is a narrow exception, however, and one cannot reasonably assume that any employee of a company (even a senior one) has authority to bind that company. Because Gillian did not have any indicia she had authority to bind her IC290 (e.g., being the IC’s director or its technology-transfer officer), Polemian Dynamics had a poor basis for assuming her signature alone would bind anyone at NIH other than herself, and so would have weak grounds at best for asserting that the IC breached any contracts. This conclusion is cold comfort for Gillian. Normally, if an agent acts within the scope of the authority delegated by the principal, the agent will not be liable if the principal later breaks the contract.291 That immunity, however, rests on whether the agent acted within the scope of the authority. Because Gillian’s signature was not authorized by NIH, she will not be protected by the fact that she signed the agreements, even if she did it in an attempt to carry out her official duties. Finally, even if a scientist who signs an agreement clearly lacked authority to bind the employer, the

505

employer may still be placed in the position of facing an irate company. Two particularly high-profile cases highlight the problem. According to an article published in The Scientist,292 Dr. David Kern, a medical professor at Brown University, was asked by a local fabric company called Microfibers to consult on two cases involving a rare syndrome called interstitial lung disease. He discovered it was due to conditions in Microfiber’s factories and also discovered cases in other employees of Microfibers working at two specific facilities. Immediately, he began the process of publishing his results. Microfibers, however, threatened to sue both Kern and his employer on the basis of certain nondisclosure agreements signed by students in Kern’s department, who had come to Microfibers for a visit 2 years before on an unrelated matter. Apparently, neither Kern nor his employer had ever ratified the agreements, and it is unclear whether either was even aware of the agreements’ existence. Even so, Kern’s employer placed in the highly awkward position of having to face litigation or restraining Kern, elected the latter. Another high-profile example of an attempt to suppress research, reported in major newspapers,293 occurred between the former Boots Pharmaceuticals294 and the University of California at San Francisco (UCSF). In 1987, Dr. Betty Dong, a scientist at UCSF, signed Boots’s research funding agreement personally to conduct a study on whether Synthroid (a synthetic drug for the treatment of hyperthyroidism) was superior to generic equivalents. The study, completed in 1990, indicated that the generics were bioequivalent to Synthroid. Dr. Dong handed copies of the data to Boots, but by 1995, Boots had not released any of the information, so Dr. Dong submitted a manuscript to the Journal of the American Medical Association (JAMA). Boots asserted the study was flawed and refused permission to publish, and the original research agreement stated that permission was required before the results could be made public. Despite the fact that the provision violated UCSF policy, UCSF’s attorney told Dr. Dong that UCSF would honor the term, and if she wanted to publish on her own, she would have to defend herself against Boots’s threatened litigation without UCSF support. Faced with this threat, Dr. Dong asked JAMA to halt the article. Only after intervention by Dr. Louis Sullivan, then Secretary of the US Department of Health and Human Services, did Boots relent and allow publication,295 but not before Boots had published a scathing critique, reinterpreting the data in a manner that cast a more favorable light on Synthroid.296 Clearly, the overwhelming majority of companies do not normally behave in the manner experienced by Drs. Dong and Kern, or else more recent examples would appear in the scientific press. Even so, the two

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

506

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

cases serve well as a cautionary tale of how badly this sort of relationship can go when someone lacking adequate training and experience in negotiating such agreements decides to sign one personally. Scope of Actual Authority of Government Laboratories The context of government laboratories adds an additional twist. For most people, laws are disabling: In other words, you can do whatever you want unless it is prohibited by law. For the government as an acting entity with few exceptions, laws are enabling: A government agency (and its authorized representative) can do only what the law has expressly authorized. In the establishment of relationships between government agencies and nongovernment parties, this divergence of point of view often causes substantial problems. In particular, companies, nonprofits, and private universities, all accustomed to crafting essentially whatever terms their internal institutional policies will allow, simply do not understand why the agency says “No, we cannot do that.” The enabling character of law as it applies to government action stems from the Constitutiondthe very foundation for both federal and the state governmentsd which lists those specific things the government can do. Ultimately, the written authority for an agency to take a given action must be directly traceable from a provision in the Constitution, to a law passed by the legislative branch (or, occasionally, an order issued by the executive branch), through regulations promulgated by the secretary of the agency, and a written trail of delegations down the chain of command within that agency. At each delegation, the authority to act may be restricted further. The scope and meaning of these documents may be illuminated by opinions of courts, the Attorney General, and perhaps the general counsel of the agency. Finally, each agency may establish its own policies of implementation, which generally stem from the mission set out in the original legislation. As a consequence, even if a given person has the raw potential to receive authority to act on behalf of the agency, the scope of authority actually delegated may be severely circumscribed by these various layers of government. In certain circumstances, a particular office in an agency may want to take an action that is still within the law but exceeds existing delegations of authority. Unfortunately, circumventing a given authority may require so much review at so many levels and may precipitate so much political fallout, that only an extraordinary case would justify the attempt. Occasionally, the law also acts on agencies in a disabling way. For example, agencies of the government are directly forbidden to take an action that would incur on the agency a debt that exceeds its appropriated

budget without express statutory authorization to do so.297 Thus, in the Polemian Dynamics scenario, NIH could not agree to protect Polemian Dynamics from the product liability lawsuits brought by the injured students because the possible judgments against Polemian Dynamics (not to mention Polemian Dynamics’s legal fees) might well exceed the agency’s appropriated budget. At best, Polemian Dynamics may feel cheated, having entered an agreement in good faith, and will be reluctant to enter future agreements with anyone at NIH. At worst, if any government employee purports to incur such a liability on behalf of the governmentdas Gillian did in the agreements she signeddthe employee risks, in theory at least, going to jail.298

Agreements to Protect Confidentiality One political extreme holds the view that the government is engaged in the systematic suppression of information that the public has a need to know. The other extreme asserts that the government is not capable of keeping information secret without being forced to do so by law. Reality lies somewhere between these extremes. Since the passage of the FOIA, a lively debate has ensued over the proper balance between these two opposing positions. Sometimes, the government must reveal the information on which its actions and policies are based; other times, release of information in government possession would cause injury without providing any public benefit. In the arena of scientific research, the debate is as strong as anywhere. Occasionally, government scientists need access to confidential information in the hands of private parties to do their jobs. By the same token, these same government scientists must publish their research results. The challenge is to find a way to accommodate the legitimate needs of industry to protect trade secrets and of individuals to protect their privacy without giving a private party the power to restrict the government scientist’s duty to publish results or the public’s right to know. The reach of FOIA is not limited to federal laboratories and offices. In 1997, Congress extended the reach of FOIA to nongovernmental researchers receiving federal funds.299 Specifically, Congress ordered the Office of Management and Budget to amend Circular A-110 “to require federal awarding agencies to ensure that all data produced under an award will be made available to the public through the procedures established under the Freedom of Information Act.” Effective for all grants (new and continuing) awarded after March 16, 2000, data that are (1) first produced in a project that is supported in whole or in part with federal funds and (2) cited publicly and officially by a federal agency in

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

support of an action that has the force and effect of law are subject to disclosure under FOIA.300 Background: Trade Secrets As discussed above, a trade secret can be any piece of information that (1) is more or less exclusively known by the party claiming it (i.e., it is truly a secret in the field), (2) is protected by measures that are reasonable under the circumstances, and (3) has some economic valued either because the owner of the secret experiences a direct and tangible economic benefit (e.g., a cheaper way of making a formulation) or because the competitors of the owner would have to expend considerable resources to discover the secret through lawful means (e.g., by reverse engineering).301 For as long as the information actually remains a secret, the legal right to protect the secrecy of that information continues. If the basic criteria are met, the owner of a trade secret has grounds to ask a court to protect that secret against “misappropriation” by assessing money damages and sometimes by imposing an injunction.302 The difficulty in trade secret litigation, typically, lies in proving that all the initial criteria are met. For example, assuming your confidante wrongly disclosed your secret, how do you prove that your information was actually a secret at the time you disclosed it to the confidante? Were the steps you took to keep your information secret “reasonable” (and will a jury agree)? These are difficult facts to prove even in the best of conditions. Moreover, as a purely practical matter, the likelihood is low that an injured party will recover through the legal process the true value of what was lost when the secret was revealed, even if misappropriation has been proved. And this does not begin to consider defenses the trade-secret recipient might have, such as where disclosure was authorized. That is where some kind of agreement can help. For several reasons, using some form of confidential disclosure agreement (CDA) is a good idea for all concerned. First and arguably foremost, a signed agreement proves the people involved actually knew that a disclosure of a trade secret may occur, and merely putting a signature on paper often has the psychological effect of making those involved treat the terms of the written agreement more seriously than they would with a mere handshake. Second, clear terms can help avoid disagreements and can clarify which information should be treated as confidential, as well as what acts are or are not appropriate. Third, where a patent application has not yet been filed, a written confidentiality agreement reduces the risk that a patent office or court will later deem the prefiling disclosure to be a bar against patenting. Finally, even where there is a wrongful disclosure, if it is a minor disclosure, the party owning the trade secret still has a chance of getting legal protection for the

507

information in the future because the party can point to the agreement as evidence that the party took every reasonable step under the circumstances. Secrets and the Government Under FOIA303 (and its various state counterparts), all government records must be disclosed on request, unless the government can demonstrate that the information in the record falls into a specific, narrow exception on a short list set out by Congress; even then, the government must disclose a redacted version if feasible. Of the exceptions on that list, five are routinely relevant to the federal government’s biological and medical research. They are exceptions for trade secrets and “commercial and financial information,”304 internal decision-making,305 personal information of a private nature,306 unfiled patent applications in which the government owns an interest,307 and certain research information generated under a “Cooperative Research and Development Agreement” (CRADA) (discussed in more detail later).308 This arrangement presents a dilemma for NIH. On the one hand, from a scientific perspective, data should be meticulously collected, organized, and carefully analyzed before drawing conclusions; releasing preliminary conclusions could be irresponsible if they have not been grounded in properly collected data, particularly if the conclusions have not undergone some substantive review. This is especially true where the premature release of “unsifted” information would be misleading. Furthermore, NIH acknowledges that private research facilities have a legitimate need to protect their trade secrets and individuals have the right to privacy; NIH understands that these parties will not cooperate with NIH if the confidentiality of their information will not be protected. On the other hand, even apart from the commands of FOIA, NIH has strong reasons to support disclosure of all research results as quickly as possible. For example, the bedrock mission of NIH is to uncover new knowledge that will lead to better health for everyone. NIH depends on the rapid communication of research results to advance that mission. Also, because the most talented scientists cannot advance their careers if impediments block their ability to publish important results in a timely manner, they will instead work in a more publication-friendly environment. For these and other reasons, NIH is strongly committed to the principle that scientific advancement relies on the unfettered and rapid dissemination of information. NIH will never approve any agreement in which a private entity has substantive control or veto power over the research publication of one of its scientists, lest valuable information that was developed by taxpayer funds be stifled to further private interests. On this point, NIH will not

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

508

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

negotiate and encourages the academic community to follow its lead. As a compromise, NIH strives to draw a line between the information provided to NIH and the research results derived from that information. NIH will work with collaborators to protect legitimate trade secrets from inadvertently being disclosed in publications. Specifically, NIH will delay disclosures enough to give collaborators a reasonable opportunity to file patent applications on discoveries. Also, NIH will seriously consider any requests by collaborators to redact or edit manuscripts and other disclosures before they are made public. Nonetheless, NIH must retain final authority to decide whether to go ahead with a given disclosure. Anatomy of a Confidential Disclosure Agreement A normal CDA309 addresses four major points,310 in one form or another. First, it identifies the information to be disclosed. Second, it names the parties. Third, it states how the confidential information will be handled. Finally, it specifies the duration. Occasionally, some agreements discuss rights to IPdthat which exists prior to any disclosure under the agreement and that which is discovered because of the disclosure, should any arised but this is not a necessary term. The information to be disclosed defines scope and reach of the agreement. Consequently, this is the single most important part, and a well-crafted CDA will clearly identify the information to be disclosed. Unfortunately, there is a tension between the “provider” of the information, who typically wants the definition to be as broad as possible, and the “recipient,” who wants it as specific as possible. Also, the provider will not want the CDA’s description of the information to include the confidential information itself. Nevertheless, some description should be fashioned that will make clear to the recipient exactly what the provider expects the recipient to keep confidential. Note that the agreement can accomplish this task in one of two ways, either by identifying the nature of the information with specificity (e.g., “the investigator’s brochure for company’s study drug”) or by obliging the provider to mark all documents with the legend “confidential” and reducing oral disclosures to writing (and marking them) within a set time. Although providers may dislike agreeing to accept the duty to mark, doing so is in their interests: As a rule, courts will not impose a greater duty on a recipient to identify and segregate a provider’s confidential information than a provider imposed on itself.311 In other words, if the information were truly valuable, a provider would have marked it. Also, as a matter of reasonableness, the agreement should specify those situations in which information

ostensibly provided under the agreement will not be deemed confidential, such as (1) information that is or becomes public through no misdeed by the recipient; (2) information that the recipient lawfully receives from a third party, that the recipient already knows, or that the recipient independently creates; and (3) information that must be disclosed by force of law. Next, identifying the parties is simple, yet surprisingly often it is botched by making the individual receiving the information sign as the party rather than the individual’s employer. One reason this is a mistake is the question of agency: Providers have essentially no protection if they ask individuals to sign agreements on behalf of their recipient employers, unless the individual’s authority to do so is starkly apparent. Even if agency is not an issue, another problem lies in the hidden trap that caught Gillian Niher when she signed Polemian Dynamics’s CDA in her personal capacity: She breached her CDA merely by telling her scientific director and lab chief about Polemian’s informationd not to mention by telling John Rogers at Smallvilledand any remedies specified in the CDA could be invoked against her. How the parties will handle the confidential information is usually where the most substantial negotiations occur because the possibilities are virtually endless. For example, what measures will be taken to control who at the recipient’s lab will have access to documents? When the agreement ends, what will be done with the documents, and for how long will the provider’s rights survive? If the recipient wants to publish, what steps will the recipient have to take to ensure the publication does not contain the provider’s confidential information? What will the provider’s rights be if the recipient is ordered by a court to disclose the confidential information? Each of these issues could be negotiated within the policies of the parties. Finally, the agreement should have a clear, specified ending point. Some providers ask for (and sometimes receive) promises to keep information confidential indefinitely. Nevertheless, as Benjamin Franklin once wrote in Poor Richard’s Almanac, “three can keep a secret, if two are dead;” in other words, the more who know a secret, the shorter its secret status will be. In addition, the dizzying pace at which biomedical technology is advancing strongly implies that the commercial value to a piece of confidential information depreciates rapidly, even if competitors never learn the secret. Consequently, a reasonable term to keep a secret should reflect the true life of the secret and little more. This is particularly important in the academic world, in which the act of dissemination is the source of value for information. NIH policy is to keep information given to it as confidential for 3 years, which can be extended for an additional 2 years on requestdsubject, of course, to the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

limitations imposed by the Freedom of Information Act and other laws. Even for nongovernment parties, only in the most unusual circumstances is it even meaningful to promise to maintain a secret for more than 5 years. IP is only occasionally a true issue. Most parties appreciate the unlikelihood that the recipient will invent something immediately and directly on seeing the provider’s confidential information. Others, comfortable with the strength of their background patent position, do not concern themselves with what might happen if someone improves on the technology. In both of these cases, the agreement will state (at most) that patent law will govern ownership of patentable discoveries, and no licenses are promised. Still, some providers (usually small companies having a single core technology in a competitive market) will insist that they be promised certain rights in anything invented by the recipient as a direct consequence of learning the confidential information. Companies and universities may, in the circumstances of the moment, decide that the benefit is worth the risk and agree to such a term. The government can never do so under a CDA. With the singular exception of a CRADA (discussed later), any term in an agreement that purports to promise rights in future government inventions, including even the option to negotiate a license, lacks authority under the law.

509

the transferred material. Occasionally, the unique nature of the material to be transferred genuinely demands special treatment. Other times, the value of the material to the provider will justify added consideration. Nonetheless, the MTA should be an easy agreement to establish even taking care to avoid the major pitfalls and accommodate the needs of an unusual case. In principle, each pending MTA represents a set of experiments that are not being done because of paperwork. In practice, MTAs can get bogged down by posturing, by the overburden of negotiators, by unrealistic expectations of one of the parties, or perhaps by the sluggishness of a provider who is cooperating only out of courtesy and cannot be bothered to hurry. Perennially, delays caused by MTAs are the single most common complaint by scientists about technology transfer. Still, no matter how tempting cutting corners or bypassing procedure may seem, a failure to take care can create problems such as those suffered by Gillian.

Background

Anatomy of the Material Transfer Agreement A normal MTA will address the following separate topics: (1) identifying the provider and recipient; (2) identifying the material; (3) how the material will (or will not) be used; (4) how confidential information regarding the material, passed to the recipient incidental to the material transfer, will be maintained; (5) recipient’s rights with respect to the material; (6) the term of the agreement; (7) indemnification and warranties; and (8) inventions derived from the use of the material. The MTAs now in circulation have particular terms that range from the truly innocuous to the truly outrageous. Each has its pitfalls for the unwary.

A widely acknowledged axiom of academia is that the widest possible circulation of research materials is crucial to maintaining the pace of research. For years, and even today, little more than packing documents, cover letters, or even bills of lading provide the only evidence of transfers of materials. NIH has long searched for constructive methods of transferring materials without any kind of documentation or at least to minimize the amount of paperwork required.312 Companies and a few universities, however, believe profits might be reaped by controlling the flow of the unique and useful things they have made. Others, moreover, have realized their vulnerability to product liability lawsuits (not to mention accusations of theft of trade secrets and patent rights, in addition to theft of the material). Accordingly, agreements to document the transfer of materials have begun to proliferate tremendously. For the foreseeable future, the MTA is here to stay. Fundamentally, a routine MTA should be a simple, innocuous agreement, essentially promising that the recipient will not do anything unethical or stupid with

Parties As with every agreement, the MTA should identify everyone involved, namely, the providing and receiving organizations. Many MTAs also name the provider’s scientist and/or recipient’s scientist, but where this is done, the MTA should clarify that the scientists are not the actual parties to the agreement. Again, this serves the very clear purpose of specifying who has agreed to be bound by the agreement and who is responsible if it is not carried out. Thus, when Dr. Niher signed Polemian Dynamics’s MTA in her personal capacity, she was personally bound by whatever terms Polemian had demanded reasonable or unreasonable. Increasingly, providers of material are demanding that all people who will handle the provided materials must actually sign an agreement in their personal capacities. To be sure, there is some wisdom in requiring that the recipient scientist acknowledge, in writing, having received the MTA, having read it, and having understood the terms under which the materials were transferred. Even so, in the overwhelming majority of cases, forcing the recipient scientist to be bound personally is

Agreements to Transfer Materials The Basic Material Transfer Agreement

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

510

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

pointless overkill because the recipient scientists are already bound by employment agreements, because other tort-based remedies exist regardless of whether the recipient scientist signed the MTA, and because the maximum damages for the breach of a contract such as this rarely will rise anywhere near a lawyer’s litigation fee. Materials The MTA must also specify the materials to be transferred. Although this also is obvious, not all descriptions of materials are created equal. For example, some MTAs define the “materials” to include all “derivatives,” regardless of whether the derivative incorporates any part of the original material. If the original material is a plasmid and the derivative is the plasmid incorporating an inserted oligonucleotide, this term may be understandable, but what if the original material is a cell line to be used to screen candidate drugs? Arguably, any drugs discovered or designed using the screening cell line could be construed as a “derivative.” Everyone should watch for this subtle attempt to reach into future inventions (i.e., defining the “material” as including anything invented with it). Government labs must be particularly careful because rights to future inventions cannot be promised under the MTA; such a “back-door” transfer of invention rights would be unlawful. One issue of particular concern to government laboratories is the status of the materials: Are they for sale? The MTA is authorized for the purpose of enabling research and no other purpose. If a private party recipient can buy a particular material from the marketplace, the recipient should pay for it; NIH is not a manufacturer or retailer, let alone a free supplier of commercial materials. Likewise, if NIH scientists can buy materials from competitive retailers, the use of the MTA to circumvent the procurement laws and regulations would be inappropriate (and possibly illegal). Uses The MTA should include a brief research plan (RP) and clearly state prohibited activitiesdin particular, that the research materials should not be used in humans. Essentially, these provisions serve two purposes; namely, they put the provider on notice of the nature of experiments the recipient plans to do, and they instruct the recipient not to do anything else. If Gillian had sent the plasmid to John Rogers under a formal MTA (assuming she was not prohibited from doing so by a prior MTA with Polemian Dynamics), then she would have had a clear, easy answer to the congressional inquiry: John agreed in writing not to test the plasmid in humans; if he broke the agreement by doing just that, Congress should be asking him why he did it.

Confidentiality In certain cases, confidentiality should be addressed but rarely does this present a problem. If documents containing trade secrets about the material are transferred with the material and to the extent the material constitutes a trade secret, confidentiality should be preserved; if the provider is still worried, the provider simply should not send those documents. Occasionally, however, companies will insist that certain limitations be placed on the recipient’s ability to publish results. These limitations vary, from a mere 30-day delay (but only to permit the filing of patent applications on discoveries) at one end of the spectrum, to the right to review and redact in the middle of the spectrum, to the absolute right to prohibit any disclosures of any kind in perpetuity at the far end. Although private parties may negotiate whatever terms match their policies, NIH has a strict, essentially nonnegotiable policy never to permit any private party to control or limit the NIH scientist’s prerogative to publish. Because NIH wants to collaborate, however, NIH will seriously consider any comments collaborators have and will accommodate any reasonable request to redact confidential information not absolutely necessary to publish. Rights in the Materials As a general principle, the typical MTA creates, in legal terms, a “bailment.” In other words, the relationship between the parties, the scientists, and the materials is analogous to the relationship between a restaurant, the restaurant’s coat-check host, a guest, and the guest’s coat. If the guest, 5 minutes later, demands the coat back, the host cannot refuse to deliver it. The host may not do with the guest’s coat as the host sees fit, regardless of whether the guest paid for the coat-check service and even if the host’s actions confer a benefit to the guest. Likewise, the recipient of research materials under an MTA may hold the materials, must return or destroy the materials on demand, and may use the materials only as the provider says the recipient may use them. The recipient under an MTA does not have any ownership rights in the physical material transferred, even after the provider has asked the recipient to destroy the material. The bailment relationship should be (and normally is) detailed in a term in the MTA. This term usually states that the recipient will have a limited license to use the materials, but that the provider retains title. The MTA often will state that the recipient will keep control over the materials and will not permit anyone to handle or use the materials other than those under the recipient’s direct supervision. The MTA should state that the recipient will not transfer the materials to any third party without the written consent of the provider. All of this is routine and recommended even if not required.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

Termination Every contract should have a clear ending point. That event could be mutual consent, unilateral request by provider, the delivery/consumption of goods, the creation of a joint work product, or a simple expiration date. This is purely a matter of practicality. It addresses, for example, how long information must be kept confidential; how long the recipient has to track the MTA; and which rights, if any, continue after the material has been consumed, and if some do, for how long. Although parties certainly can agree to make an MTA last indefinitely, the absence of a formal termination event could cause bad feelings if each party’s understanding is inconsistent with the other’s. This is especially important where materials may sit in storage for years, long after the original recipient scientist (who understood the limitations imposed on the provider’s materials by the MTA) has moved on to another position elsewhere. Typical MTAs arranged by the NIH state simply that the recipient of materials will protect confidential information relating to the materials for a term of 3 years, which may be extended by another 2 years on written request by the party providing the materials. Warranties and Indemnification Often, private parties to contracts make certain promises to each other that are beyond such matters as quantity, delivery date, price, etc. Promises such as these can constitute warranties and indemnification. These terms should be approached with great caution and under the advice of an attorney because such terms can create liability beyond the “four corners” of the agreement. A warranty is a special promise, above the promises normally included in a contract, that a certain relevant fact is true.313 In the ordinary sale of retail products, for instance, the merchant provides the consumer with the promise that the product in the box is what the label on the box says it is (called a “warranty of merchantability”) and does what the merchant claims it will do (a “warranty of fitness for a particular purpose”). The warranty may be expressly stated, implied by the context, or imposed by law. If not forbidden by a law, parties may agree to waive certain warranties that ordinarily would apply automatically. In the absence of a warranty, if the merchant breaches a contract, the other party gets the cash value of the contract as damagesd you get your money backdand no more. If a warranty is provided, and the promised fact turns out not to be true, the warrantor may be held liable for all foreseeable, consequential damages above the value of the contract, provided the damages can be shown caused by the breach of warranty.314 In essence, a warranty is an agreement to shift risk. Research-related contracts often disclaim any warranty of merchantability and fitness for any particular

511

purpose. These warranties were created to protect consumers against shady merchants selling shoddy goods. Such warranties, however, are rarely necessary to protect researchers handling materials of unknown properties and hazardsdresearchers are normally expected to be careful with such items. Also, agreements in the research arena routinely disclaim any warranty that materials being transferred do not infringe some third party’s IP rights. Sometimes, however, a provider of material will insist that the recipient warrant such things as that the investigator will comply with the laws of a certain country (other than the recipient’s own) or that the terms of the MTA do not conflict with any other agreement entered by the recipient. Facts such as these may be impossible to ascertain, and so a warranty regarding these facts could be disastrous. Indemnification essentially is a promise in the other direction: The customer promises the merchant that if the customer does something stupid with the product and injures someone who then sues the merchant, the customer will “step into the shoes” of the merchant for the purposes of defending the litigation, including paying lawyer’s fees, as well as paying any judgments against the merchant if the merchant loses. Indemnification essentially is another way parties can shift risks. Suppose in Gillian’s case, for example, when she signed Polemian Dynamics’s agreements, she agreed to indemnify Polemian against any third-party lawsuit concerning the materials she got from Polemian or arising from her use of them. If so, then even though she did not manufacture the materials and she did not tell anyone that the materials were safe or would work properly, she could be forced to pay any judgments imposed on Polemian for making an unsafe product. Indemnification creates a particular problem for agencies of the federal government and of many states. Companies and universities routinely acquire liability insurance specifically to cover litigation expenses, and though individuals often do not do so, they can; government agencies, in contrast, cannot indemnify anyone unless the law expressly states otherwise. Under the Adequacy of Appropriations Act315 and the Antideficiency Act,316 for example, a federal agency may not incur a debt or liability greater than the amount of money Congress has appropriated to that agency. Indemnification is an open-ended promise to pay whatever is assessed, even if that assessment exceeds the agency’s budget. In the worst case, any federal employee purporting to incur such a liability on behalf of the federal government could be subject to criminal sanctions.317 At best, when a company that thought it had secured indemnification from the government learns the truth, the company may believe that the scientist and the government negotiated in bad faith.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

512

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

Inventions: “Reach-Through” Rights The terms in MTAs relating to IP are often the most nettlesome because they directly address the diverging views regarding how research material should be treated. Generally, a consensus has arisen that the clinical uses of materials (i.e., diagnostic, prognostic, or therapeutic applications) may be restricted by those who invented them to enable the inventor to recoup its investment and perhaps make a profit. For example, if a new, patented chemical is found to treat a disease, the inventor/patent owner should be able to control who can sell this new drug. The question is the extent to which pure research uses should be similarly restricted. In other words, if the new drug were being used to explore the mechanism of action of a cellular process unrelated to the condition the drug was invented to treat, should the inventor/patent owner be entitled to extract large royalties for each experiment or perhaps claim rights in discoveries made from those experiments? Industry traditionally views all of its creations as things that required a capital investment and that can provide a source of revenue. Some even believe that all discoveries made using the creation, which could only have been made using the creation, are really part and parcel to the original creation. In various forms, some in industry now ask for so-called “reach-through” rights. Specifically, in exchange for the use of the materials, the provider would get some kind of rights in anything the recipient invents. Sometimes, the provider asks merely for an “option” to a license to be negotiated later; other times, the provider asks for a prenegotiated license, often royalty-free, occasionally exclusive (i.e., no one can develop the invention but the provider). A few ask for total assignment of any inventions. Academia views inventions as the practical consequence of theoretical discoveries, and that the former should serve the latter not the other way around. Otherwise stated, any use of an invention that serves purely to investigate facts should be free and unfettered. Exorbitant fees or powerful reach-through rights, therefore, create barriers to research and learning to the free flow of ideas. If a particular road to the development of a technology contains too many toll booths, the researcher will be forced to search for other, probably less efficient routes and so may miss important discoveries. Additionally, at least from academia’s point of view, the mere fact that someone has asked for reachthrough does not necessarily mean granting it would be fair or reasonable. Put to the absurd extreme, if person A loans person B a screwdriver, should A be allowed to claim ownership of every piece of equipment, and perhaps every building, B builds with it? Aggressive reach-through by industry creates an even larger barrier for government researchers because the government is extremely limited in its authority to grant license rights,

even when the grant is appropriate. In fact, the only mechanism now existing for a government laboratory to promise a private party present rights to the laboratory’s future inventions is through a CRADA (discussed later). The Uniform Biological Material Transfer Agreement In the early 1990s, various nonprofit research organizations, universities, and NIH together realized that the MTA was an annoying, bureaucratic nuisance. All agreed on the major principles governing the transfer of materials among each other; all agreed not to do anything unethical or stupid with each other’s research materials. So, they wondered, why must every MTA be renegotiated? To avoid the unnecessary extra paperwork, the academic community created the Uniform Biological Material Transfer Agreement (UBMTA)318da “treaty,” for lack of a better descriptiondto which any nonprofit organization or university could become a member. Under the UBMTA, any signatory could transfer materials to any other signatory using a prenegotiated form that could be signed directly by the scientists doing the transfer rather than an administrator. The UBMTA is not mandatory, so if the provider has a special interest in the transferred materials (e.g., because the technology is exclusively licensed to a company), the provider could revert to the standard MTA process. To the extent it has been utilized, the UBMTA process has dramatically streamlined the process and decreased the time needed to arrange for the transfer of materials among members. Unfortunately, the UBMTA has not been used as much as it might be. Part of the reason appears to be a lack of awareness that the mechanism exists, and another part seems to be that the UBMTA, crafted as a compromise among a committee of diverse parties, is confusingly written. The most visible part, however, appears to be the fact that universities and nonprofit organizations are marketing their technologies more aggressively, signing exclusive arrangements with companies more often, and thus finding that the UBMTA is not adequate. Still, it remains a valuable tool. The Clinical Trial Agreement Obviously, Gillian Niher could not have brought Polemian Dynamics’s materials to NIH under the MTA because MTAs expressly prohibit using transferred materials in humans. To address this limitation in the MTA, some of the NIH ICs and academic institutions have developed a variant, which would permit them to use received materials for clinical purposes. The Clinical Trial Agreement (CTA) is, at its heart, an expanded MTA. In addition to all the topics arising under the MTA, the CTA addresses other issues specific to clinical trials. A well-crafted CTA should reflect, at a minimum,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

special consideration relating to protocol drafting, regulatory filings, interactions with regulatory agencies, use of data, and how the agreement might be terminated in the middle of the clinical trial without endangering the patients enrolled in the trial. Because the provider does not have to participate in research under a CTA, the CTA should make clear the provider’s role. Some providers are pleased to be passive, particularly those who have little or no experience in running clinical trials or interacting with the FDA; other providers want at least an equal role as the recipient in drafting, reviewing, and approving any protocols, as well as in analyzing the data. NIH is flexible, provided that no outside party has the authority to command NIH personnel, restrict NIH research, or veto NIH publications. Additionally, the CTA must clearly state who will be responsible for filing any regulatory documents with the FDA, such as an IND application, necessary to enable the research to begin. Because INDs are expensive and complicated, companies often are happy to let NIH bear responsibility for filing the IND if the NIH is so inclined. If NIH is going to accept that responsibility, however, the provider should agree to send NIH the necessary formulation data or, at least, the provider must give NIH access to a drug master file. As a matter of law, the holder of the IND is responsible for reporting adverse events,319 for monitoring the conduct of the trial,320 and for participating in any direct interactions with the FDA.321 When NIH holds the IND, some providers want to participate in this process, and some do not; the term is negotiable. If the provider holds the IND, however, NIH must have the right to file its own adverse event reports and must be permitted to participate in any meetings with the FDA. This is to ensure that information negatively affecting the product being tested will be timely disclosed to the FDA in proper context. Almost all companies would never suppress such data, but the temptation for a company, which may be depending on the success of the product, to put a misleading spin on damaging information can be enormous. Physicians who are participating in the trial have a legal duty to report adverse events; the failure to do so could lead to administrative, or even criminal, penalties.322 Consequently, NIH would rather risk insulting a company and insist on retaining this right. Normally, a CTA will state that each party will share with the other all raw data generated under the clinical trial, except to the extent necessary to protect the confidentiality of the patients. Furthermore, each party normally has the right to use the data for its own purposes (reserving to each party, of course, the right to file patents on the inventions of its own employees). The parties may, if they like, agree to publish jointly;

513

however, NIH will always reserve the right to publish independently if the provider declines to join in a particular publication. For studies involving the NIH or one of its grantees, if the provider of a study drug must have direct access to identifiable private information of any study subject, then the provider may inadvertently become regulated under the Common Rule.323 If so, the parties would be well advised to seek legal help in navigating this situation. Finally, some term should address what happens if one or both of the parties determines that the agreement should be terminated before the protocol has been fully carried out. As a matter of medical ethics, a doctor should not be forced to abandon a viable course of therapy already being administered to a patient due solely to a provider’s refusal to continue providing the therapeutic agent. On the other hand, providers do not want to be forced to continue squandering significant resources on a project they have determined will not be profitable. Fortunately, there are several mechanisms to protect both parties’ needs. For example, the provider could agree to provide enough agents at the beginning of the trial to supply the entire protocol. Alternatively, the provider could give a license, plus information on the manufacture of the materials, to hire a contractor to make enough agents to complete the trial (if the recipient cannot make the materials). The mechanism is negotiable even if the principle is not. Other Key Specialized Material Transfer Agreements Materials in Repositories The point of a repository is to enable researchers to access samples of research materials, typically biological materials, from a centralized source. Some of the ICs at NIH maintain repositories of biological materials, including transgenic animals, cDNA clones, and viruses. The NCI, in addition, maintains a special repository of natural products collected from sites around the world. Private entities, such as the American Type Culture Collection and Jackson Laboratories, maintain repositories for public access. Use of repositories raises one common issue relating to MTAs, specifically relating to “background rights.” When the creator of the materials places a supply in the custody of a repository, the creator may have filed patent applications on the materials and may demand that the repository put restrictions on the further distribution of the materials. Normally, these restrictions are similar to those that would appear in a standard MTA (i.e., do not do anything stupid or unethical with the materials). Occasionally, the creator demands that the repository extract reach-through rights from any recipient for the benefit of the creator. Those who would access a private repository should be vigilant for such terms.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

514

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

The NCI natural products repository has a unique twist, which is serving as a model for transnational research in other arenas. NCI’s authority under the law to control what happens to materials it sends out of a repository is severely limited. Because most of the materials were collected from developing countries, NCI negotiated agreements with these countries, trying to find ways within US law to ensure that a significant portion of any economic benefits derived from materials collected would flow back to the country of origin. Ultimately, NCI established a memorandum of understanding with each source country, which has resulted in the favorable cooperation of, and even collaboration with, the local scientists and universities in these countries. Software Transfer Agreements Suppose a scientist at NIH wants to work on software now under development. If the software was written by a potential collaborator, can an MTA be used to allow the collaborator to transfer the software? Alternatively, what about transferring the software out? The answer to both is a qualified “yes.” On a superficial level, the use of an MTA should be legally sufficient to permit the transfer of the physical magnetic media containing the code. On a deeper, more theoretical level, the issue is somewhat more complicated. Specifically, it is not clear whether NIH’s authority to transfer biological materials324 includes the intangible essence of software code (separated from the physical media on which it is written). Regardless, an agreement to transfer software must always conform to all laws and NIH policies, such as that the software is not commercially available and that the provider does not demand reach-through to NIH inventions. The NIH office of the General Counsel has approved use of various software transfer agreements by some of the ICs, many of which have been streamlined into a simple “click-wrap” version (i.e., clicking on the “accept” button before downloading software is sufficient to create a legally binding agreement). Note that other federal agencies have greater authorities granted by law to transfer and foster development of software made in whole or in part by their labs. For purposes of this chapter, however, those are beyond the scope.

Collaboration and Inventions: The Cooperative Research and Development Agreement Background Uncounted collaborations occur every year that are never formally documented and that are never

embodied in any kind of contract. Obviously, some kind of written agreement is required when the collaboration becomes complicated, the nature of the research requires the employers of the collaborating scientists to commit significant materials, or one or both parties is worried about how rights to inventions will be handled. For private parties, the possible terms are essentially limited only by each party’s policies and available resources. For the government, matters are not so simple. When a government employee invents something, the employee must assign ownership rights in that invention over to the government.325 Yet, the core mission of NIH is to conduct research to improve the public health not to sell products and make profit. Therefore, when someone at NIH discovers a new prognostic/diagnostic tool or a new therapy, NIH is unable to commercialize products embodying the invention (i.e., engineer mass production, tap distribution channels, market, and sell)donly private parties can do that. The law requires the government to offer the opportunity to license government inventions to all interested parties in open competition. In a sense, the public owns each government invention, so everyone (the public) should have fair access to every opportunity to acquire rights in each invention. This arrangement is appropriate for NIH inventions made purely by NIH personnel working exclusively with NIH personnel at NIH-owned facilities, but what about inventions through a collaboration with someone outside NIH? Indeed, these laws made companies nervous about collaborating with government scientists or laboratories because the companies had no assurance that they would have rights in inventions their work enabled. For example, a company probably would be reluctant to collaborate with the government on an improved analog to the company’s main drug if it feared the government would license the analog to another company to increase competition. In particular, small companies worried that larger companies could outbid them, even though the small companies’ collaborative contributions made the invention possible. Therefore, in 1987,326 and through updates in the ensuing years,327 Congress further authorized government laboratories to enter a CRADA that provided the laboratories a measure of flexibility in arranging such collaborations. For this purpose, each IC of the NIH constitutes a “laboratory.” Currently, the CRADA is the only legal mechanism by which a government laboratory can, in the present, promise a collaborator certain rights in inventions yet to be created by the government as a consequence of the collaboration (while not a statutedefined term as it is in the Bayh-Dole Act, agencies generally refer to CRADA inventions as “Subject Inventions” because the substance of the definition under the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

CRADA is nearly identical to that in the Bayh-Dole Act). The CRADA discussed in this chapter, therefore, is unique to governmenteprivate collaborations (although the principles involved may have applicability beyond this particular scope). Cooperative Research and Development Agreement Basics Foremost, the keystone of a CRADA is collaboration.328 Each party must contribute some intellectual effort toward a specific research project. That collaboration drives the process of developing the agreement, and, in turn, that process is designed to authorize the negotiation of terms in the agreement suitable to enable the project. Under a CRADA, the government laboratory may • contribute physical resources to a collaborator; • dedicate staff time to a project; • permit a collaborator’s staff to work in government facilities without requiring that staff member to assign all inventions to the government (as is usually required); and • promise the collaborator an exclusive option to elect an exclusive or nonexclusive license (collaborator’s choice) in any government rights in any invention that will be conceived or first reduced to practice in the conduct of research under the CRADA. The CRADA is not a grant, procurement contract, or other “funding mechanism”329; in other words, the government laboratory is prohibited from transferring congressionally appropriated funds to a CRADA collaborator under any circumstances. Under a CRADA, the collaborator may • contribute resources to the government laboratory; • dedicate staff time to a project; • permit government researchers to perform their CRADA-related research in the collaborator’s facilities; and • transfer funds to the government for the laboratory’s use in carrying out the CRADA. In addition, essentially all the issues pertinent to CDAs, MTAs, and CTAs can arise in the negotiation of a CRADA. Finally, the CRADA has some additional, administrative twists unique to the nature of the agreement, which will be discussed in more detail later. As is obvious, the CRADA involves resolution of a wide variety of important issues. Consequently, an understanding of what a CRADA comprises can smooth the process greatly. The fastest NIH can establish a CRADA is approximately 1 month. Complicated cases have required a year of negotiations and occasionally more. A rough estimate for the time needed to establish a new CRADA is between 4 and 8 months, depending in

515

large measure on the speed and flexibility of the collaborator’s review process. For the NIH, the major stages include selecting a collaborator, negotiating the agreement, institutional review of the agreement, and, finally, execution by the partiesdeach of which is discussed in turn. Selecting the Collaborator In the vast majority of cases, the selection of a CRADA collaborator is one of the simplest of the four main phases. Occasionally, however, this process presents serious hurdles. These hurdles can be classified as either fair access or conflict of interest. By law, a federal laboratory must provide every possible collaborator “fair access” to any opportunity to enter a CRADA.330 In the vaguely related context of selecting contractors to perform a service or selecting merchants to sell goods to the government, the Federal Acquisition Regulations thoroughly specify the procedure for ensuring that any interested party can apply for the opportunity. For CRADAs, in contrast, this process is not so well defined with good reason. In the overwhelming majority of cases, a given research collaboration can only be done with a single collaborator. For instance, a CRADA to develop the collaborator’s patented new drug cannot be done by anyone but the owner (or licensee) of the patent. In such cases, no purpose would be served by opening the selection process to a competitive bid. Still, the government is not permitted to pick collaborators in an arbitrary or capricious waydthe selection must always be reasonable under the circumstances. As a general rule, if research under a CRADA genuinely depends on access by the government to a prospective collaborator’s proprietary technology, unique expertise, or unique facilities, “fair access” is deemed satisfied without any effort having been made to find someone else (because no one else would suffice). This is not as beneficial for collaborators as it might appear at first blush, however, because the CRADA research would be circumscribed by that uniqueness. The laboratory would be free to initiate CRADAs on similar themes utilizing other technologiesdprovided, of course, that the laboratory can satisfy all the requirements of each CRADA, and that the RP of each CRADA does not overlap any other. For instance, a laboratory having a new cDNA library may initiate one CRADA with a gene array maker using its propriety chip technology and another CRADA with a company with unique protein analysis technology to create an expression profile for this cDNA library. Indeed, in principle, if the RPs were written specifically enough and the research carefully segregated, the laboratory could engage in more than one CRADA to analyze different proteomic aspects of

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

516

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

the library, limiting each CRADA to research utilizing that collaborator’s unique technology. For those cases in which access to a particular technology is not a necessary prerequisite, the laboratory may announce to the world that a CRADA opportunity exists and permit anyone interested to submit a research proposal. Again, unlike the Federal Acquisition Regulations, the law governing CRADAs provides no formal guidance or specific mechanism for making such announcements. At a minimum, publication in the Federal Register should suffice, but there is no limit to the types or number of venues that may be used for announcing. Thereafter, if one collaborator is selected on the basis of a proposal submitted under that announcement, others who did bother to respond would have little grounds for complaining they did not have fair access. A question often arises in the selection of collaborators, namely, whether a federal laboratory can enter a CRADA with either a nonprofit entity or a company based outside the United States. The answer to this question is “yes” for both kinds of collaborators with certain caveats. For example, in a collaboration with a nonprofit entity, particularly universities, the parties must consider how the products that might be developed under the CRADA will be commercialized. Also, unlike private parties, the federal laboratory has limited authority to control the flow of money, which makes sharing royalties a tricky endeavor. These are issues the nonprofit entity should consider before embarking on the negotiation for a CRADA because the terms will have to be carefully crafted. For a foreign company, the law governing CRADAs requires only the following: (1) if two parties apply for the same opportunity, and if one is a US company and the other is a foreign company, the federal laboratory must give preference to the US company331 and (2) collaborators promising to substantially manufacture in the United States any products embodying Subject Inventions licensed to the collaborator will receive preference over those who do not so promise.332 Assuming the collaborator is appropriately and fairly selected, the other hurdle to cross before negotiations can begin is to confirm that the NIH’s Principal Investigator (PI) will not have a conflict of interest. For example, if the PI owns stock in the prospective collaborator or is in the process of negotiating employment with the prospective collaborator, the PI’s independence could be questioned even if not actually compromised.333 To avoid such problems, NIH has designed a questionnaire for its PIs to complete and submit to their ethics officers for review. This process protects the PIs from accusations of unfairly steering opportunities to favored companies. Furthermore, the review uncovers subtle problems in the selection process before the

negotiations become too involved usually in time to address them to the satisfaction of everyone. For clinical projects, NIH has implemented two variations on its standard CRADA, one for “intramural” studies (human studies to be conducted exclusively within NIH) and one for “extramural” studies (some or all to be conducted at grantee or contractor sites). For both types, as with the CTA, a clinical CRADA normally should reflect, at a minimum, special consideration relating to protocol drafting, regulatory filings, interactions with regulatory agencies, use of data, and how the agreement might be terminated in the middle of the clinical trial without endangering the patients enrolled in the trial. Unlike the CTA, however, the collaborator will always participate in a clinical CRADA contributing intellectual effort to portions of the research, if not to all of it. Negotiating the Agreement Once the collaborator has been appropriately selected, the negotiations may begin. A complete CRADA should have at least three parts: (1) the RP, which includes specific commitments of particular actions by each party; (2) the commitment of specific resources by each party; and (3) the agreed legal terms. Other items can be included if the parties see fit. Modifications to the Cooperative Research and Development Agreement Language The NIH, CDC, and FDA jointly developed several template (“Model”) CRADAs for use in particular circumstances; all NIH CRADA negotiations begin with one of them. A particular proposed CRADA should clearly indicate any specific changes to the Model language. Some of the language in the Model is little more than a restatement of existing law. For example, the mandatory government licenses to collaborator’s Subject Inventions derive from a specific statutory command334; these cannot be removed. Others reflect critical NIH policies and are rarely changed without solidly logical, objectively reasonable justifications. The remainder of the terms can be, and often are, negotiated to accommodate the unique needs of each collaborator and the planned research. Appendix A: The Research Plan The RP should serve three functions. First and foremost, it should lay out exactly what each party will do. The more specific these allocations of work are, the less likely confusion over responsibilities will be. Second, it should circumscribe the activities so that activities “outside” and “inside” the scope of the RP can be readily distinguished; these, in turn, define which inventions are governed by the agreement and which are not. For example, if the RP contemplates incorporating

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

PART THREE: TECHNOLOGY TRANSFER AGREEMENTS

an antigen into a vaccine, the accidental discovery of a new and improved excipient would be a Subject Invention if discovered under the CRADA, while the discovery of a wonderful shoe polish probably would not be a Subject Invention, even if made by the CRADA research team during the term of the CRADA. Third, if NIH invents something and the collaborator elects the option to a license, the collaborator is entitled under the law 53 to a prenegotiated field of use; NIH’s normal prenegotiated field of use is “the scope of the RP.” Although not absolutely required, an RP may also incorporate additional information to serve other functions. For example, the RP presents a useful opportunity to explain the background of the technology, to highlight the experience and interests of the parties’ PIs, and to explain in detail why the selected collaborator is particularly suited to the project. Also, the RP can contain an agreed abstract for public release, which each party understands up front may be freely disclosed to the public at any time by the other. Having such an abstract is especially important for NIH, which must often answer regular FOIA requests for routine data relating to CRADAs. Companies also appreciate the reduced risk offered by such an abstract because they no longer have to worry about reviewing every proposed disclosure for these routine FOIA requests. Finally, the RP can include such other useful information as the parties deem appropriate, such as a list of the most relevant publications, background patents owned by each party, and any pertinent prior agreements between the parties. Financial and Material Contributions In NIH CRADAs, the commitment of physical and financial resources are collected in Appendix B. Specifically, this part of the CRADA spells out exactly what materials, facilities, equipment, and staff will be committed by each party and the funds (if any) that the collaborator will provide to NIH. Each Appendix B is unique; there is no requirement that every CRADA involve the commitment by either party to any particular one of these items. Ultimately, the resources to be committed by each party will depend on the research that each party wants to perform. If, for example, the collaborator wants NIH to perform an experiment using a particular piece of equipment neither party owns, the collaborator may choose to buy the equipment and loan it to NIH, to hire a contractor with the equipment to run the experiment, or to give the NIH laboratory money to buy one; alternatively, the NIH lab may decide to purchase the equipment directly. If neither the collaborator nor the NIH laboratory can afford the project’s cost or if each could pay but is unwilling to bear the expense for other reasons, the RP would have to be modified or scaled back.

517

The funding aspect of CRADAs offers a particularly useful source of opportunities to government laboratories. First, funds transferred by the collaborator to the government may be used to hire personnel who will not be subject to the hiring ceilings otherwise imposed by law. Second, unlike appropriated money, funds transferred to the government under a CRADA may be kept by the laboratory for the duration of the CRADA, and it will never revert to the US Treasury. Third, subject to routine ethics review, the funds can be used for the travel-related expenses of government researchers in carrying out the CRADA. Furthermore, receipt of CRADA funds and materials allows the PI to explore additional, perhaps costly experiments that would not otherwise be supported by the laboratory’s budget. Of course, the laboratory must regularly account to the collaborator how the funds are spent, the funds may be used only to pay for CRADA-related materials or activities, and any unobligated funds at the end of the CRADA must be returned to the collaborator. The funding aspect of the CRADA also benefits companies. For example, a collaborator can leverage a relatively small contribution into a scientific project of far larger value. In addition, the CRADA presents a way for a company to support particular government research that is of interest to the company, without running afoul of the ethical concerns implicated in the gift process. Also, companies that do not have large budgets may be able to fund CRADA research with money received under a federal grant, such as the Small Business Innovative Research program. As long as the research project of the CRADA is distinct from the research project under the grant, such grant money can be used in this manner. In exchange, the company receives a wealth of expertise not available from any other source in the worlddnot just in a particular scientific field but also in regulatory filings. With respect to this funding aspect of the CRADA in particular, one point should be clearly reemphasized: The foundation of every CRADA is intellectual collaboration. Although the CRADA mechanism offers NIH laboratories the opportunity to supplement the resources they receive through routine channels, this aspect should not dominate the CRADA. If the only reason a laboratory has for entering a CRADA is the material support, the use of the CRADA mechanism is inappropriate. Reciprocally, if the CRADA collaborator is only interested in acquiring a “pair of hands” for the collaborator’s benefit and has no interest in the intellectual contributions of the NIH scientists, there is no collaboration and the CRADA is not appropriate, even if the laboratory is willing to assist the collaborator.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

518

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

National Institutes of Health Review of the Agreement Once any conflict of interest questions have been resolved, the scope of the research has been clearly identified in the RP, resources have been promised, and legal language has been hashed out, the complete agreement must be reviewed. Overall, this process currently requires four formal approvals within the IC, informal review and comment by the NIH, and then final execution by the director or deputy director of the IC. Technically, the law provides that the Secretary of the department or agency in which the federal laboratory resides may void CRADAs within 30 days of execution.335 For the NIH, the Department of Health and Human Services has delegated that authority to the NIH Deputy Director for Intramural Research (DDIR). Because of the high level of experience with CRADAs across the NIH tech-transfer community, DDIR has elected to waive the prerogative to invoke the authority as long as the ICs have their new CRADAs reviewed for comment by a select committee to ensure that CRADAs are within NIH policies and are generally consistent with each other. Execution by the Parties and the Effective Date By its terms, the CRADA becomes effective on the day when the last signature is inked. Could the parties agree that the agreement will be effective on a date after final signature? Certainly. How about making the agreement retroactively effectivedin other words, setting the effective date to a point before the final signature? By itself, backdating is apparently not authorized by the law; NIH cannot promise IP rights without anything having been signed by the collaborator and the institute. Unfortunately, this inability to make CRADAs retroactive put prospective collaborators and NIH in a quandary; because CRADA negotiations take months, and because the NIH approval process takes weeks (sometimes more than 1 month), either the scientists must remain idle or the collaborator must risk losing rights to any NIH inventions that are invented just before the CRADA is signed. Several CRADA opportunities were lost because of this problem. To solve it, NIH developed the Letter of Intent (LOI). The LOI is a simple promise that if a CRADA is signed, its effective date will be retroactive to the effective date of the LOI. Unfortunately, the mechanism has certain limitations. First, because the LOI is not a promise that a CRADA will ever be signed, some collaborators are unwilling to begin a project under an LOI. Also, some projects depend on the transfer of funds to begin; however, no funds may pass to NIH under an LOI because it is not a promise that the full CRADA will be signed. Furthermore, because the LOI was originally intended

solely to allow research to begin while the paperwork is completed, it is limited to a short, 6-month life, which may be extended for cause. Nonetheless, many collaborators are satisfied with this mechanism, and the LOI has proven to be a valuable mechanism for facilitating collaborations. Possibilities CRADAs have enabled a large, and growing, number of exciting projects. NIH laboratories and companies have been able to study therapies for rare diseases, new (perhaps high-risk) uses of existing drugs for new indications, and therapies and vaccines for diseases primarily occurring in poor countriesdtechnologies most companies would consider too high a financial risk to invest resources developingdby pooling their resources and expertise. Beyond this, NIH laboratories have been able to access manufacturing channels and unique research materials, often that would be prohibitively expensive to procure without the CRADA, especially for the smaller ICs. Companies, in turn, have found they can access a unique source of expertise and can tap a research entity whose bedrock interest is to help successful products reach the bedside without having to rely on the assistance of a competitor. In one specific and successful example, when NCI needed a tool to perform microdissection of cells for clinical pathology of cancerous tissue, NCI and Arcturus Engineering agreed to enter a CRADA to develop one. Laser capture microdissection was created, and it is now on the market. In the case of Gillian Niher, a clinical CRADA would have enabled her project and protected her interests in publishing, receiving material and financial support, and handling regulatory filings. It would also have guaranteed NIH’s interest in protecting the patients enrolled in the clinical trial. Additionally, it would have protected Polemian Dynamics’s interest in ensuring compliance by Gillian with the terms of their agreement and perhaps secured rights in Gillian’s invention involving the bare plasmid. In short, a clinical CRADA would have established the ground rules by which the parties would act, ensured no one operated on a misconception and authorized them to do what they wanted to do.

CONCLUSION Scientists who began their research careers 20 years ago or more may have expected never to become involved in the “filthy lucre” of business preferring the oasis of the ivory tower, no longer. To survive, modern researchers, particularly in the biomedical arena, must be able to understand the fundamentals of product development, and in particular, the role patents play. Worse, ignorance and naivete´ about the rules of this

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

519

REVIEW QUESTIONS

new and undesired world or mistakenly believing in popular myths about patents, all risk causing a researcher to fall into traps that could destroy a major project and even a career. For those who might mourn the loss of their dream of what science should be, there is an equal bright side. Today’s savvy scientist will recognize that patentsd while expensive, complicated, arcane, and timeconsumingdpotentially can be a tremendous resource in the effort to turn basic discoveries and good ideas into tangible outcomes that actually benefit society. Not all discoveries require patents, and patents and patent licensing do not solve all problems, but where they do, they are extraordinarily useful tools. Therefore, even though some researchers will never invent anything, every researcher should know at least what patents are and how they work.

Novelty

Bright-line test: Was the entire invention described in a single prior art reference before the relevant critical date?

Prior Art

Any reference regardless of form (publication or prototype) available to the public if one were motivated to look for it

Provisional (PRV)

A “place-holder” application that allows delaying the filing of a full application by up to a year

Patent Cooperation Treaty (PCT)

An international patent application filed in one patent office but claiming the right to file national applications later

Reduction to Practice (Actual)

Physical construction of a working prototype or successful performance of a process

Reduction to Practice (Constructive)

Written instructions and data showing how to make and use the invention so that one of ordinary skill can practice it without undue experimentation

Statutory Bar

Novelty-defeating prior art disclosed more than 1 year before the inventor’s earliestfiled patent application

Statutory Subject Matter Novelty-defeating art disclosed before the date of invention

The invention is a machine, composition, article of manufacture, or process that was created by a human

Utility

Assignment

A transfer (sale or otherwise) of ownership of property

The invention has a credible, substantial, and specific use

Written Description

Conception

Formulation in the mind of the inventor(s) of the complete, operative ideadincluding specific means not just the desired end

The invention is described clearly enough for others to know what it is

Continuation (CON)

A daughter patent application that “resets” the prosecution clock and does not add new material

Continuation-In-Part (CIP)

A daughter patent application that adds new material

Divisional (DIV)

A daughter patent application filed later, where the original application contained more than one invention

Enablement

Sufficiently clear description of the invention so that others can make and use it without undue experimentation

Interference

An administrative proceeding before the USPTO to determine which inventor is entitled to receive a patent

License

An agreement not to sue another for infringement of intellectual property rights

National Stage Application

Nation-specific patent applications claiming priority to a single prior PCT filing

Nonobviousness

One of ordinary skill in the field of the invention would not find it insignificantly different from the collected prior art

BRIEF GLOSSARY OF CRITICAL TERMS IN PATENTING

Anticipation

REVIEW QUESTIONS 1. You, in your first week working in a technology transfer office, have been assigned to handle an invention that your predecessor had started before he left. Reviewing the file, you notice that the inventor submitted her data to your predecessor well before publishing, and your predecessor had timely arranged to file a US Provisional 8 months ago. The invention has not yet been exclusively licensed, but likely will be someday, and the technology is important enough to your institution that funding the patent prosecution is not an issue here. Which of the following set of patent applications do you expect you will have to arrange? a. A PCT in 4 months and a US-371 in 22 months. b. A US-111 in 4 months and National Phase applications in major markets in 22 months.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

520

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

c. A US-Provisional and US-111 to be filed in 4 months and National Phase applications in major markets in 22 months. d. A US-371 to be filed in 4 months and not to pursue any international filings unless in the next 22 months we find a licensee who agrees to pay for them. e. A PCT in 4 months, and in 22 months, to file a US-371 and National Phase applications in major markets. 2. By statute, in deciding whether or not to grant any applicant’s request for an exclusive license, the government must consider each of the following except: a. Whether the technology is a CRADA “Subject Invention.” b. Whether competition will be enhanced or reduced. c. Whether exclusivity benefits the public interest. d. The breadth of the proposed field(s) of use. e. Whether nonexclusive licensing could achieve practical development. 3. Under BayheDole, a Determination of Exceptional Circumstances is appropriate when: a. Programmatic goals of an agency’s intramural program cannot be met without restricting or eliminating the extramural inventors’ right to elect title to their inventions. b. The agency is planning on using a DEC in a contract for services other than research. c. The contracting agency has determined that changing the standard patent clauses will better promote the policy and objectives of the BayheDole Act. d. Prospective contractors have been notified that a DEC is being considered and have been given the ability to veto its use. e. The contracting agency has “Greater Rights” to the invention over the contractor.

References 1. For clarity through this chapter, the capitalized term “State” refers specifically to one of the fifty States of the United States, rather than a generic nation-state. 2. Later, in 1975, the US Patent Office became the US Patent and Trademark Office (“USPTO”). 3. Pub. L. No. 914, 70th Cong. (March 2, 1929). 4. The CAFC was given jurisdiction over several topics: substantive patent law (whether arising in the district courts or the U.S. Patent and Trademark Office (the “USPTO”)), appeals from the U.S. Court of Claims, the U.S. Court of Veterans Claims, and the U.S. Court of International Trade. 5. P.L. 112-129. 6. Following closely in the footsteps of the Paris Convention, most of the world quickly adopted a sister treaty known as the Berne Convention on Artistic Works of 1886. Under it, each member nation agreed to confer to owners of copyrights granted by other member nations the same rights that the member grants under its

7.

8.

9.

10.

11.

12.

13. 14.

15.

16.

17. 18.

19.

20. 21. 22.

own copyrights. The United States initially refused to join, but a century later became a member in 1988. Specifically, Belgium, Denmark, France, Germany, Ireland, Italy, Liechtenstein, Luxembourg, the Former Yugoslav Republic of Macedonia, Netherlands, Sweden, Switzerland, and United Kingdom have ratified or acceded to it. The official website is currently maintained by the WIPO, available here. The member states are: Benin, Burkina Faso, Cameroon, Central African Republic, Chad, Congo, Coˆte d’Ivoire, Gabon, Guinea, Guinea-Bissau, Equatorial Guinea, Mali, Mauritania, Niger, Senegal, and Togo. Not all member states use French as an official national language. http://www.aripo.org/. The organization includes 16 member states and 14 observer states. Not all member states use English as an official national language. The official website is available here. The member states are: Turkmenistan, the Republic of Belarus, the Republic of Tajikistan, Russia, the Azerbaijan Republic, the Republic of Kazakhstan, Kyrgyzstan, the Republic of Armenia, and the Republic of Moldova. The official website is available here. The member states are: United Arab Emirates, Bahrain, Saudi Arabia, Oman, Qatar, and Kuwait. Note: A patent, by itself, is not a “monopoly,” though many people carelessly treat these two terms as synonyms. To have a monopoly, one must actually sell something. Patent owners need not sell anything for a patent to be enforceable; also, as is often the case, impractical products do not become worth buying just because they are patented. Consider, for example, the “Marshmallow System” (US Patent No. 6,800,312), in which a marshmallow destined to become a s’more is manufactured with a hollow core shaped to hold a piece of chocolate, which hopefully will minimize the loss of melted chocolate once the combination has been cooked. See, e.g. Mansfield E, Schwartz M, Wagner S. Imitation costs and patents: an empirical study Econ J December 1981;91(364):907e18. See, e.g., Park W, Ginarte J. (Published Online: June 29, 2007), Intellectual property rights and economic growth. Contemporary Econ Policy 15(3), 51e61; Griliches Z. Patent statistics as economic indicators: a survey. J Econ Lit December 1990;28(4):1661e707; Ginarte J, Park W. Determinants of patent rights: a crossnational study. Res Policy 1997;26:283e30; Maskus K, Penubarti M. How trade-related are intellectual property rights?. J Int Econ 1995;39:227e48. For example, under the right circumstances, the Canadian and Israeli patent offices will issue a patent having the identical scope as a newly issued U.S. patent. 35 U.S.C. x 161; 37 C.F.R. x 1.161. Newly bred varieties of plants can be legally protected through a system unrelated to patent law, called the “Plant Variety Protection Act”, which is administered by the US Department of Agriculture. A discussion of this system is outside the scope of this chapter. 35 U.S.C. x 171; 37 CFR xx 1.152-1.155. Design patents have so much overlap with copyrights and trademarks that they are easily confused and difficult to clarify. As all three of these are not especially relevant to biomedical research activities, clarifying this topic is beyond the scope of this chapter. Technically, the patent statute defines “patentee” to include both the initial owner and anyone who later becomes an owner of that patent. 35 U.S.C. x 100. Diamond v. Chakrabarty, 447 U.S. 303, 308 (1980). Chakrabarty, 447 U.S. at 309. Another example is the case of O’Reilly v. Morse, 56 US 62 (1853), over Samuel Morse’s telegraph patent application. Morse tried to claim “the use of the motive power of the electric or galvanic current, which I call electro-magnetism, however developed for

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

23.

24.

25. 26.

27.

28. 29. 30. 31. 32.

33. 34.

35.

36.

37.

marking or printing intelligible characters, signs, or letters, at any distances”. The Supreme Court held that this claim was overbroad, as it encompassed an entire force of nature. The full name and citation is Laboratory Corporation of America Holdings v. Metabolite Laboratories, Inc., et al., 370 F.3d 1354 (Fed. Cir. 2004), cert. granted in part, 546 U.S. 975 (2005), cert. dismissed, 548 U.S. 124(2006). LabCorp, 548 U.S. at 125 (Breyer, Stevens, Souter, JJ, dissenting). Justice Stevens has since questioned his own dissent, in oral argument for the case of Mayo Collaborative Svces v. Prometheus Lab., Inc. (No. 10-1150, oral argument 7 December 2011). 447 U.S. 303 (1980). To be sure, several other events occurred from 1980 to 1982, each of which arguably may have had a bigger impact on the growth of the US biotechnology industry than the Chakrabarty decision would have had by itself. For example, from 1980 to 1982, ground-breaking patents were awarded for making proteins using recombinant DNA (the famous Cohen/Boyer patent), for the first automated DNA-synthesizing machine, for a gas-phase protein sequencer, and for the discovery of how to induce yeast (rather than E. coli) to synthesize a human protein, specifically interferondeach of which helped make biotechnology into a commercially feasible business. Also in 1980, Congress passed the Bayh-Dole Act and Stevenson-Wydler Act, which are both discussed below; these laws changed the rules on inventions made using Federal funds. Additionally, the FDA approved recombinant human insulin in 1982, demonstrating a viable business model for a blockbuster biotech “drug”. Finally, the creation of the CAFC in 1982 gave the business community confidence that the legal status of patents would be more predictable. There is no way to know for certain which events had the bigger impacts. The Biotechnology Industry Organization reports that, as of 31 December 2006, nearly 1500 biotech companies were operating in the United States, employing some 180,000; the stocks in 336 of these companies were publicly traded, with a combined market capitalization of US $360 billion. That report is available online here. Jensen K, Murray F. Policy forum: intellectual property landscape of the human genome. Science October 14 2005;310:239e40. 132 S.Ct. 1289 (2012) (hereinafter, “Mayo”). 133 S.Ct. 2107 (2013) (hereinafter, “Myriad”). 130 S. Ct. 3543 (2010) (citing Bilski v. Kappos, 130 S. Ct. 3218(2010)). Collaborators included scientists at the NIH National Institute of Environmental Health Sciences, the University of Utah, McGill University, Eli Lilli and Company, the University of Laval (Quebec), the Hospital for Sick Children in Toronto, the University of Pennsylvania, and the Institute for Cancer Research in London. NCI BRCA fact sheet (rev. 29 May 2009), available online here. According to the court’s opinion, as well as the 2010 Report of the HHS Secretary’s Advisory Committee on Genes, Health and Society (SACGHS Report), while the price of Myriad’s test seems expensive, the per-amplicon cost is comparable to genetic tests that are unpatented. The final version of the report is available online here. Specifically, the Association for Molecular Pathology, American College of Medical Genetics, American Society for Clinical Pathology, and College of American Pathologists. Specifically, Drs. Haig Kazazian and Arupa Ganguly, both of the University of Pennsylvania; Dr. Wendy Chung, of Columbia University; Dr. Harry Ostrer, of New York University; and Drs. David Ledbetter and Stephen Warren, of Emory University. Specifically, Ms. Ellen Matloff, of Yale University; and Ms. Elisa Reich, of New York University.

521

38. Specifically, Breast Cancer Action and Boston Women’s Health Book Collective. 39. Specifically, Lisbeth Ceriani, Runi Limary, Genae Girard, Patrice Fortune, Vicki Thomason, and Kathleen Raker. 40. 160 F. 467 (2d Cir. 1908). To be sure, the USPTO and some district courts considered the issue before then, but the Circuit Court cases before Hotel Security are ambiguous on this particular topic. See, e.g., Ex parte Abraham, 1869 Dec. Comm’r Pat. 59 (1869), and United States Credit Sys. Co. v. Am. Credit Indem. Co., 53 F. 818 (C.C.S.D.N.Y. 1893), aff’d on other grounds, 59 F. 139 (2d Cir. 1893); see Cincinnati Traction Co. v. Pope, 210 F. 443, 444e46 (6th Cir. 1913), and Rand, McNally & Co. v. Exchange Scrip-Book Co., 187 F. 984, 984e85 (7th Cir. 1911). The 1983 MPEP, x 706.03(a), codified the USPTO’s long-standing view that, “though seemingly within the category of process or method, a method of doing business can be rejected as not being within the statutory classes.” 41. 409 U.S. 63 (1972). 42. Id., at 70. 43. 437 U.S. 584 (1978). 44. Id., at 590-91. See also In re Grams, 888 F.2d 835, 839-41 (Fed. Cir. 1989) (simply including a general “data-gathering step,” without any critical limitation on how the data is collected, does not render an algorithm patentable subject matter under Flook because every algorithm requires gathering data as input). 45. 450 U.S. 175 (1981). 46. At the time the invention was made, no one knew how to obtain an accurate measure of the rubber’s current temperature without opening the press. The invention solved this problem by using embedded thermocouples to constantly check the temperature, and then fed the measured values into a computer. The computer then used the algorithm in an iterative way to calculate when the molding machine should open the press. 47. Id., at 184e185. 48. See, e.g., Chisum D. The Patentability of Algorithms. U Pitt L Rev 1986;47:959e1022; Bruzga C. Review of the Benson-Flook-Diehr Trilogy: Can the Subject Matter Validity of Patent Claims Reciting Mathematical Formulae Be Determined under 35 U.S.C. Section 112? J. Pat. & Trademark Off. Soc’y 1987;69:97e216; see also Risch M. Everything is Patentable. Tenn L Rev 2008;75:591, at pp. 646e47 (discussing the Federal Circuit’s efforts to reconcile Benson, Flook, and Diehr in the case of in re Bilski). 49. 149 F.3d 1368 (Fed. Cir. 1998); confirmed by AT&T Corporation v. Excel Communications, Inc., 172 F.3d 1352 (Fed. Cir. 1999). 50. USPTO detailed statistics are available here. 51. Hall B. Business Method Patents, Innovation, and Policy. U.C. Berkeley Department of Economics Working Paper, No. E03-331, May 2003, Table 2, p. 24, available online here. Amazon.com won a four-year long ex parte reexamination before the USPTO (story available here). 52. McKenna B, et al. Patently Absurd: The inside story of RIM’s wireless war,” The Globe and Mail, February 21, 2006, available online here. 53. Normally, the CAFC hears cases by a panel of three judges; here, on its own initiative, the CAFC elected to have the entire court hear the case (“en banc”). 54. While nine of twelve judges joined in the majority opinion, the case provoked one concurring opinion and three dissenting opinions. In a rare compliment, the Supreme Court later suggested that “[s]tudents of patent law would be well advised to study these scholarly opinions.” Bilski v. Kappos, 561 U.S. ___, No. 08-964, Slip Op. at 3 (June 28, 2010), aff’g on other grounds, In re Bilski, 545 F.3d 943 (CAFC 2008). 55. In re Bilski, 545 F.3d at 954 (relying primarily on Benson, 409 U.S. at 70).

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

522

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

56. In re Bilski, 545 F.3d at 960 & n.23; see also, In re Ferguson, 558 F.3d 1359, 1364 n.3. (Fed. Cir. 2009). 57. To be precise, the two concurring opinions each agreed with the result the rejection of the CAFC’s tests, though each concurring opinion would have applied different logic than the majority. 58. “Interim Guidance for Determining Subject Matter Eligibility for Process Claims in View of Bilski v. Kappos,” 75 Fed. Reg. 43922 (21 July 2010). The Guidance clarifies that the “machine or transformation” test may not be the sole test for making these decisions. 59. Diamond v. Chakrabarty, 447 US 303, 309 (1980) (citing S. Rep. No. 1979, 82d Cong., 2d Sess., 5(1952); H. R. Rep. No. 1923, 82d Cong., 2d Sess., 6 (1952)). 60. US Patent and Trademark Office Manual of Patent Examination Procedure, x 2107.01 (rev. 18 Dec 2008) (hereinafter, “MPEP”). 61. Id. 62. Id. 63. MPEP, x 706.03(a) (citing Juicy Whip Inc. v. Orange Bang Inc., 185 F.3d 1364, 1367e1368 (Fed. Cir. 1999)). 64. Carella v. Starlight Archery, 804 F.2d 135 (Fed. Cir. 1986). 65. Examples of countries with grace periods include Australia (12 months), Canada (12 months), Japan (6 months), South Korea (6 months), Mexico (12 months), and Brazil (12 months). Of course, each country’s rules on when the grace period applies is unique; for instance, Japan only provides the protection of its grace period for the inventor’s own disclosures made through an experiment, publication, or presentation at a study meeting or exhibition, as well as for those disclosures that happen against the inventor’s will. See Japanese Patent Law (Law No. 121 of 1959, as amended), Section 30 (an English translation is maintained online by the WIPO, at the following webpage). 66. Invitrogen Corp. v. Biocrest Manufacturing L.P., 424 F.3d 1374 (Fed. Cir. 2005). 67. Bernhardt, L.L.C. v. Collezione Europa USA, Inc., 386 F.3d 1371, 1380e1381 (Fed. Cir. 2004). 68. See In re Kollar, 286 F.3d 1326, 1330 n.3, 1330e1331, (Fed. Cir. 2002) (distinguishing licenses which trigger the on-sale bar (e.g., a standard computer software license wherein the product is just as immediately transferred to the licensee as if it were sold), from licenses that merely grant rights to an invention which do not per se trigger the on-sale bar (e.g., exclusive rights to market the invention or potential patent rights)). 69. Moleculon Research Corp. v. CBS, Inc., 793 F.2d 1261, 1267, (Fed. Cir. 1986). 70. Moore v. United States, 194 USPQ 423, 428 (Ct. Cl. 1977); Petersen v. Fee Int’l, Ltd., 381 F. Supp. 1071, 182 USPQ 264 (W.D. Okla. 1974). 71. Interferences can be declared between more than two parties, but for purposes of this chapter, two is plenty. 72. 35 U.S.C. x 146. 73. Crouch, D. Is Novelty Obsolete? Chronicling the Irrelevance of The Invention Date in U.S. Patent Law. Mich Telecomm Tech L Rev 16, __ (2010). Data from 1980 to 1994 suggest that a little over 200 interferences were declared each year, of which about 5% proceed to a full hearing. Kingston W. Light on simultaneous invention from US Patent Office ‘Interference’ records. World Patent Information September 2004;26(3):209e20. 74. One study of interferences declared between 1988 and 1994 found that the mean time to resolution was about 21 months, with a range of 16 days to 10 years. Cohen L, Ishii J. An Empirical Investigation of Patent Races: Evidence from Patent Priority Disputes at the U.S. Patent and Trademark Office, (2005) working draft available online here. 75. In the patent systems of other countries, the parallel concept is called “Inventive Step.”

76. MPEP x 2141, “Examination Guidelines For Determining Obviousness Under 35 U.S.C. 103,” Part III. 77. Graham v. John Deere Co., 383 U.S. 1 (1966). 78. In re Dillon, 919 F.2d 688, 692 (Fed. Cir. 1990). 79. In re Baxter Travenol Labs., 952 F.2d 388 (Fed. Cir. 1991) (Appellant argued that the presence of DEHP as the plasticizer in a blood collection bag unexpectedly suppressed hemolysis, however, the prior DEHP-plasticized bag achieved same result, though this property went undetected). 80. KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398 (2007). 81. MPEP, x 2161. 82. Ariad Pharmaceuticals, Inc., v. Eli Lilly & Co., ___ F.3d ___, No. 2008-1248 (Fed. Cir., Mar. 22, 2010) (en banc). 83. Regents of the University of California v. Eli Lilly, 119 F.3d 1559, 1566, (Fed. Cir. 1997), cert. denied, 523 U.S. 1089 (1998). 84. MPEP, x 2163(II)(A)(3). 85. In re Wands, 858 F.2d 731, 737 (Fed. Cir. 1988). 86. CFMT, Inc. v. Yieldup Int’l Corp., 349 F.3d 1333, 1338 (Fed. Cir. 2003). 87. In re Nelson, 280 F.2d 172 (CCPA 1960). 88. MPEP xx 706.03(c), 2165-2165.4. 89. Benger Labs. Ltd. v. R.K. Laros Co., 209 F. Supp. 639 (E.D. Pa. 1962). 90. In re Hall, 781 F.2d 897 (Fed. Cir. 1986). 91. MPEP, x 2128; see, e.g., In re Epstein, 32 F.3d 1559 (Fed. Cir. 1994). 92. For example, a journal distributed to the public only in paper form and through the mail becomes prior art when the public first receives the journal through the mail, not the date on which the publishers deposit the journal in the mail. In re Schlittler, 234 F.2d 882 (CCPA 1956). 93. In re Hall, 781 F.2d at 900 (Fed. Cir. 1986); see also MPEP x 2128.01. 94. TP Labs., Inc. v. Professional Positioners, Inc., 724 F.2d 965, 972 (Fed. Cir. 1983) (commercial exploitation in secret was still a “public use”); In re Smith, 714 F.2d 1127, 1134 (Fed. Cir. 1983). 95. MPEP, x 2133.03(b). 96. 122 F.3d 1396 (Fed. Cir. 1997). The case also included other claims besides patent infringement. 97. Legally speaking, only the formal holding, along with the portion of logic necessary to reach that holding, have the effect of “law”. When a court includes general commentary or a logical argument that is not necessary to support the holding of the case, such commentary is called “obiter dicta” (or, more simply, “dicta”). Courts offer dicta to give the legal community a clue as to how related cases might be decided. Dicta is not law, and so is technically not binding in future cases, but ignoring dicta is unwise. 98. A “joint research agreement” is defined as “a written contract, grant, or cooperative agreement entered into by two or more persons or entities for the performance of experimental, developmental, or research work in the field of the claimed invention.” x 103(c)(3). Congress made clear that it had no interest in being too specific, in order to give the people participating in a collaborative agreement the flexibility to decide whether or not it qualifies as a “joint research agreement.” 99. Townsend v. Smith, 36 F.2d 292, 295 (CCPA 1930); see also Hitzeman v. Rutter, 243 F.3d 1345 (Fed. Cir. 2001). 100. Silvestri v. Grant, 496 F.2d 593, 596 (CCPA 1974) (“an accidental and unappreciated duplication of an invention does not defeat the patent right of one who, though later in time was the first to recognize that which constitutes the inventive subject matter”); Invitrogen, Corp. v. Clontech Laboratories, Inc., 429 F.3d 1052, 1064 (Fed. Cir. 2005) (In situations where there is unrecognized accidental duplication, establishing conception requires evidence that the inventor actually made the invention and understood the invention to have the features that comprise the inventive subject matter at issue).

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

101. Dow Chem. Co. v. Astro-Valcour, Inc., 267 F.3d 1334, 1341 (Fed. Cir. 2001). 102. De Solms v. Schoenwald, 15 USPQ2d 1507, 1510 (Bd. Pat. App. & Inter. 1990). 103. MPEP, x 2138.05. 104. Burroughs Wellcome Co. v. Barr Labs, Inc., 40 F.3d 1223 (Fed. Cir. 1994), cert. denied, 516 US 1070 (1996). 105. Id., 40 F.3d at 1231. Interestingly, the Canadian Supreme Court reached the same conclusion. Apotex Inc. v. Wellcome Foundation Ltd., (2002) 4 S.C.R. 153 (scientists do not become patentees of an invention merely by executing tests; to qualify as inventors, researchers must contribute to the discovery of the inventive concept, not just verify that another person’s idea works). 106. Hitzeman v. Rutter, 243 F.3d 1345 (Fed. Cir. 2001). 107. Id. 243 F.3d at 1357-59. 108. 35 U.S.C. x 102(f); see also MPEP xx 2137 and 2137.01. This rule appears to derive directly from the U.S. Constitution, Art. I, Section 8 (“To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries” (emphasis added)). 109. Mueller Brass Co. v. Reading Industries, Inc., 352 F.Supp. 1357, 1372 (E.D.Pa. 1972), aff’d, 487 F.2d 1395 (3d Cir. 1973). 110. 35 U.S.C. x 116. 111. Pannu v. Iolab Corp., 155 F.3d 1344, 1351 (Fed. Cir. 1998). 112. Garrett Corp. v. U.S., 422 F.2d 874, 881 (Ct. Cl. 1970). 113. Nartron Corp. v. Schukra USA Inc., 558 F. 3d 1352 (Fed. Cir. 2009). 114. MPEP x 2137.01. See also Burroughs Wellcome, 40 F.3d at 1230e1231 (discussing NCI’s possible status as co-inventor based on their work). 115. See, e.g., Brunsvold B., et al. Drafting Patent License Agreements. 6th ed. BNA Books; 2008; Wright B. Drafting Patents For Litigation And Licensing. BNA Books; 2008. Licensing principles and procedures for federally owned inventions, particularly those of the NIH, are discussed in greater depth below. 116. A license could include terms under which the licensee has a right to receive assignment at a later date, and a very poorly written exclusive license can effectively transfer so much that a court will deem it to be an assignment despite designation as a “license.” Both of these situations, however, are uncommon. 117. 35 U.S.C. x 154(d). 118. 35 U.S.C. x 287. 119. Ethicon, Inc., v. U.S. Surgical Corp., 135 F.3d 1456 (Fed. Cir. 1998). 120. 35 U.S.C. x 271(b), Met-Coil Sys Corp. v. Korners Unlimited, 803 F.2d 684 (Fed. Cir. 1986). 121. Goodwall Construction v. Beers Construction, 210 USPQ 272 (N.D. Ga. 1981). 122. 35 U.S.C. x 271(f)(1). 123. 35 U.S.C. x 271(c). 124. Amersham International PLC v. Corning Glassworks, 618 F.Supp. 507 (D. Mich. 1985). 125. Patents are presumptively valid (see 35 U.S.C. x 282), but sometimes the USPTO never saw critical prior art, or perhaps the patentee failed to disclose some material information not otherwise publicly available. 126. 35 U.S.C. x 282. 127. In the case of Dickinson v. Zurko, 119 S.Ct. 1816 (1999), the Supreme Court required all courts to review factual determinations by the USPTO using the standards set in the Administrative Procedures Act. Specifically, an appellate court may only set aside a finding of fact if the USPTO’s decision was “clearly erroneous,” i.e., the finding was “unsupported by substantial evidence” submitted to the USPTO. Id., citing 5 U.S.C. x 706.

523

128. The statute of limitations requires that any lawsuit be brought within six years (35 U.S.C. x 286), however, laches can apply even if this time has not yet run out. 129. Poppenhusen v. Falke, 19 F. Cas. 1048 (No. 11279) (CCSDNY 1861). 130. Ruth v. Stearns-Roger Mfg. Co., 13 F. Supp. 697 (D.Colo 1935), rev’d on other grounds, 87 F.2d 35 (10th Cir. 1936). 131. Id. 132. Madey v. Duke University, 307 F.3d 1351 (Fed. Cir. 2002), cert. denied, 539 U.S. 958 (2003). 133. The court continued: “Our precedent clearly does not immunize use that is in any way commercial in nature. Similarly, our precedent does not immunize any conduct that is in keeping with the alleged infringer’s legitimate business, regardless of commercial implications. For example, major research universities, such as Duke, often sanction and fund research projects with arguably no commercial application whatsoever. However, these projects unmistakably further the institution’s legitimate business objectives, including educating and enlightening students and faculty participating in these projects. These projects also serve, for example, to increase the status of the institution and lure lucrative research grants, students and faculty” 307 F.3d at 1362. 134. Resnik DB. “Patents and the research exemption. Science 2003; 299(5608):821e2. 135. “Due diligence” is a context-specific term of art, generally referring to the process of avoiding obvious risks associated with a particular endeavor by exercising the particular level of diligence that is reasonable and appropriate for the circumstances. For purposes of this chapter, “due diligence” refers to the process in which a patent attorney determines whether or not a particular line of research, development, or other activity would be covered by any patents (also called “freedom to operate” analyses). This task is often more difficult in basic-research settings than applied-research settings, because the activities performed by a basic researcher may implicate a more diverse range of patents. 136. Roche Products Inc. v. Bolar Pharmaceutical Co., 733 F.2d 858 (Fed. Cir. 1984). 137. 35 U.S.C. x 271(e)(1) states that “[i]t shall not be an act of infringement to make, use, offer to sell, or sell within the United States or import into the United States a patented invention (other than a new animal drug or veterinary biological product (as those terms are used in the Federal Food, Drug, and Cosmetic Act and the Act of March 4, 1913) which is primarily manufactured using recombinant DNA, recombinant RNA, hybridoma technology, or other processes involving site specific genetic manipulation techniques) solely for uses reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs or veterinary biological products.” 138. 545 U.S. 193 (2005). 139. 35 U.S.C. x 287(c). 140. For this purpose, a “body” is defined in the statute to mean “a human body, organ or cadaver, or a nonhuman animal used in medical research or instruction directly relating to the treatment of humans.” Id., x 287(c)(2)(E). 141. Id., x 287(c)(3). 142. 28 U.S.C. x 1498. 143. 35 U.S.C. x 284. For design patents only, the patentee is entitled to receive the infringer’s profits on top of any other damages (such as a reasonable royalty). 35 U.S.C. x 289. 144. 35 U.S.C. x 285. 145. 35 U.S.C. x 283; eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006).

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

524

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

146. Georgia-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116, 1120 (S.D.N.Y. 1970), modified and aff’d, 446 F.2d 295 (2d Cir. 1971) (widely cited list of fifteen nonexclusive factors for evaluating what royalty would be “reasonable”); Barnhardt J, Revisiting A Reasonable Royalty as a Measure of Damages for Patent Infringement. 5 Wake Forest I.P.L.J. 2005;1. 147. Pincus L. The Computation of Damages in Patent Infringement Cases. Harvard J L & Tech 1991;5:95e143; Blair R, Cotter T. Working Paper: Rethinking Patent Damages. February 23, 2001 available online here. 148. While changing the venue has always been possible, courts have discretion whether or not to do so, and they typically defer to the plaintiff’s choice unless the reasons for changing clearly outweigh the reasons for leaving the case where it was filed. In re Volkswagen of America, Inc., 545 F.3d 304 (5th Cir. 2008) (en banc decision that the district court abused its discretion by not changing venue). 149. 28 U.S.C. xx 2201-02; Maryland Casualty Co. v. Pacific Coal & Oil Co., 312 U. S. 270, 273 (1941) (“Basically, the question in each case is whether the facts alleged, under all the circumstances, show that there is a substantial controversy, between parties having adverse legal interests, of sufficient immediacy and reality to warrant the issuance of a declaratory judgment.”). 150. MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118 (2007). 151. 19 U.S.C. x 1337(a)(1)(B). 152. Santos N, et al. What IP holders ought to know about the ITC and the district courts. J High Tech L 2007;7:173e9. at p. 174 & nn.10-11. 153. Id. at 176. Exclusion orders are often “limited” to a set of respondents, but can block both infringing products and downstream products that contain an infringing component. Id. at 175. 154. eBay v. MercExchange, LLC., 547 U.S. 388 (2006) (injunctions should not be granted automatically upon finding of patent infringement; patentee still has to prove the same factual standards as anyone else seeking an injunction). 155. “2009 Report of the Economic Survey,” American Intellectual Property Law Association, p. 29 (2009). The report is available online only to members of the AIPLA (contact information is available at http://www.aipla.org for requests for copies). 156. “Cost and Duration of Patent Litigation,” Managing Intellectual Property, “Briefings” column (01 Feb 2009), available online here. 157. Levko A, et al. Patent Litigation Study: A Closer Look. PriceWaterhouseCoopers, LLC., 2009, available online here. Another study found similar results; see Zura P. Patent Litigation Statistics. The 271 Blog, May 30, 2007, available online here. 158. Id. 159. 35 U.S.C. x 113; MPEP x 601(f). 160. MPEP x 2173.05(a); Process Control Corp. v. HydReclaim Corp., 190 F.3d 1350, 1357 (Fed. Cir. 1999). 161. 37 C.F.R. x 1.75(g); MPEP x 608.01(j). 162. The meaning of claim phrasing is always determined by the judge; in most cases, the question of whether the defendant’s product is covered by a claim is decided by a jury, though sometimes it is done by the judge. 163. 37 C.F.R. x 1.821. 164. 37 C.F.R. x 1.802. 165. 35 U.S.C. x 114; 37 C.F.R. x 1.91; MPEP x 608.03. 166. 35 U.S.C. x 121; 37 C.F.R. x 1.141; MPEP xx 802-802.02. 167. 37 C.F.R. x 1.56 (broadly imposing a duty of candor and good faith). 168. In the United States, only the inventor (or whoever the inventor has assigned it to, such as the inventor’s employer) may file the application, unless the inventor is represented by either a patent agent or patent attorney. A “patent agent” is someone registered

169. 170. 171.

172. 173. 174. 175. 176. 177. 178.

179.

180. 181.

182. 183. 184. 185. 186. 187.

188.

189.

with the USPTO after demonstrating a minimum amount of knowledge of patent law and of at least one recognized scientific discipline. A “patent attorney” is a patent agent who is also licensed to practice law in at least one state. To the extent this chapter refers to patent attorneys, it is understood to include patent agents for simplicity. Again, after the Patent Reform Act, failure to disclose the best mode can no longer be grounds for invalidating a patent. McKesson Information Solutions, Inc. v. Bridge Medical, Inc., 487 F.3d 897 (Fed. Cir. 2007). Mammen C. Controlling the “Plague”: Reforming the Doctrine of Inequitable Conduct. 24 Berkeley Tech L J 1329 (April 2010). One case even said that the false remark by patent counsel that his search of the prior art was “careful and thorough,” intended to speed review by the USPTO, was inequitable conduct. Gen. Electro Music Corp. v. Samick Music Corp., 19 F.3d 1405, 1411e1412 (Fed. Cir. 1994). 37 C.F.R. x 1.293; MPEP x 1103. See, e.g., 35 U.S.C. xx; 184-187. The Public PAIR system is available online here. 37 C.F.R. x 1.14, MPEP xx; 103-104. 35 U.S.C. x 154; MPEP x 2710. 35 U.S.C. x 156; MPEP x 2750. By USPTO rule, a “rejection” is always “on the merits” of the patentability of a claim (i.e., the claim is unpatentable under xx 101, 102, 103, and/or 112). MPEP xx 706 et seq. The USPTO might object to other flaws in the application, and these flaws might ultimately prove fatal to an application, but such concerns are not called “rejections.” As noted at various points above, several court cases are brewing in which some parties want to expand the range of rejections based on x 101 to include business methods, software, and processes involving only “mental steps” or “mere correlations” (mainly diagnostic inventions). If these parties are successful, the number of rejections under 35 U.S.C. x 101 likely will increase. 37 C.F.R. x1.113; MPEP x 706.07. Procedurally, the appeal goes first to the USPTO’s Board of Patent Appeals and Interferences (“BPAI”), then generally to the CAFC, then the US Supreme Court. Alternatively, after losing before the BPAI, the applicant may file suit in federal district court asking the court to order the USPTO to issue a patent (this case, in turn, gets appealed to the CAFC). 35 U.S.C. xx 254-255; MPEP xx 1480 et seq. 35 U.S.C. x 251; MPEP xx 1401 et seq. 35 U.S.C. x 301; MPEP xx 2205-2296. MPEP x 2203. In rare cases, the Commissioner of Patents may open a reexamination case on his own initiative. 35 U.S.C. x 314(b)(2). 35 U.S.C. x 315(c). If a prior art reference was truly unavailable at the time of an inter partes reexamination, it may be raised in a later infringement suit. A. Baluch and S. Maebius, “The Surprising Efficacy of Inter Partes Reexaminations: An Analysis of the Factors Responsible For Its 73% Patent Kill Rate And How To Properly Defend Against It,” Working Paper (2008), available online here. USPTO data on all reexaminations are available here. Some commentators have suggested that these flaws are so bad that a patent attorney’s recommendation that a client use the inter partes reexamination would be tantamount to malpractice. See, e.g., Knowles S, et al. Inter Partes Reexamination in the United States. J Pat & Trademark Off Soc’y 2004;86:611, at p. 614; Kunin S, Fetting A. The Metamorphosis of Inter Partes Reexamination. Berkeley Tech L J 2004;19:971e88, at pp. 978e79.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

190. Note that in the United States, the priority date may be set by the Provision application, rather than the PCT application. Also, it is worth noting that member nations are free to allow national phase applications be filed later than 30 months; the EPO, among several others, allows filing up to 31 months after the priority date, and Canada permits national filings as late as 42 months with payment of a late fee. 191. To be clear, only fourteen patent offices, which include the USPTO and EPO, are certified to perform two of the five actions, namely the ISR and the IPER, discussed later in this paragraph. The full list of eligible patent offices is maintained by WIPO, and is available online here. An applicant is allowed to designate any office on this list to perform the ISR and IPER, even if the applicant’s Receiving Office is among them. As such, an American filing a PCT application in the USPTO may designate the Korean Patent Office to conduct its ISR and IPER. 192. Before 2004, the three mandatory stages were called the “Chapter I” segment, and the optional stages were known as the “Chapter II” segment. This designation used to be more significant, because until deadlines for filing national phase depended on whether or not the applicant made a demand under Chapter II. 193. No translations: UK, France, Germany, Luxembourg, Monaco, Liechtenstein, Switzerland. Claims-only: Croatia, Denmark, Iceland, Latvia, Lithuania, Netherlands, Slovenia, Sweden. 194. Croatia, Denmark, Iceland, Netherlands, Sweden. 195. These are Belgium, Cyprus, France, Greece, Ireland, Italy, Latvia, Malta, Monaco, the Netherlands and Slovenia. WIPO maintains the official list online here. 196. Small inventors get a 50% break on most of these fees. 197. See, e.g., Berrier E. Global Patent Costs Must Be Reduced. Franklin Pierce Law Center IDEA: The Journal of Law and Technology 1996: 473e511, available online here. 198. See, e.g., Oppedahl Patent Law Firm webpage, Law Office of Jerry R. Potts webpage, and Neustel Law Firm webpage. See also, Quinn G. Cost of Obtaining a Patent. I P Watchdog, December 31, 2007, available online here. 199. Berrier, supra n. 212, at p. 474. 200. Slides of a presentation containing this data are available to AIPLA members online here. 201. In re Fisher, 421 F.3d 1365 (CAFC 2005). 202. Ass’n for Molecular Pathology v. USPTO, Case 1:09-cv-04515-RWS (S.D.N.Y. 29 March 2010). 203. Laboratory Corp. of Amer. Holding v. Metabolite Laboratories, Inc., cert. dismissed, 548 U. S. 124, 125e139 (2006) (Breyer, Souter, Stevens, JJ, dissenting) (hereinafter, LabCorp). 204. For example, in 2006, the company Research In Motion (RIM), maker of the famous BlackBerry phones, settled a patent infringement case filed by NTP, with RIM paying NTP over US$612 million. Shortly before settlement, the case attracted international attention when RIM was temporarily forced to shut down all BlackBerry services. During the fight, RIM and others publicly accused NTP of being a “patent troll” because NTP never had any intention of developing its patented technologies. Others countered that RIM was free-riding on others’ discoveries. 205. See, Crouch D. Measuring the Plague of Inequitable Conduct. Patently-O Blog, June 02, 2010, available online here. 206. 593 F.3d ____ (Fed. Cir. No. 08-1511, May 25, 2010). The USPTO has issued draft rules for public comment in which the USPTO proposes to enshrine the Therasense standards in the rules for applying for patents. See Federal Register 76(140):43631 (Thursday, July 21, 2011).

525

207. “Brazil to Break Patents on U.S. Films, Books, Drugs (Update 2),” Bloomberg.com (first posted online 15 March 2010), available online here. 208. See Feist Publications, Inc., v. Rural Telephone Service Co., 499 U.S. 340 (1991) (telephone “white pages” cannot be copyrighted, as it is merely a functional collection of data). 209. 17 U.S.C. x 106. The owner also has so-called “moral rights” in the work (17 U.S.C. x 106A), and may control importation (17 U.S.C. xx 601-603). 210. See 17 U.S.C. xx 101 (definition of “work for hire”) and 201(b) (ownership of works for hire). 211. 17 U.S.C. x 105. Note that the Government may own a copyright if it was created privately and then assigned to the Government. 212. 17 U.S.C. x 101. Note that a “collective work” and a “compilation” are effectively synonyms. 213. The Copyright Office is situated in the Library of Congress. Note that the Library of Congress is a part of the legislative branch, not the executive branch, of the U.S. Government. 214. 17 U.S.C. x 504(c). 215. Copyright notice includes: (1) the “©” symbol or the word “copyright”; (2) the owner’s name; and (3) the year in which the copyright in the work was established. 17 U.S.C. x 401(b). 216. 17 U.S.C. x 504(c)(2). 217. 17 U.S.C. x 109. Patent law has a similar doctrinedcalled “exhaustion”dbut it appears in court cases, not the statute. See Bauer & Cie. v. O’Donnell, 229 U.S. 1 (1913) (patents could not be used to control resale prices). The continued viability of this doctrine recently has been called into question; see, e.g., Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700 (Fed. Cir. 1992) (doctrine of exhaustion was only a unilaterally disclaimable “implied license”). 218. See 17 U.S.C. xx 109-122. Note that these compulsory licenses are not necessarily royalty-free. 219. Id., x 117. 220. 17 U.S.C. x 107. 221. Campbell v. Acuff-Rose Music, 510 U.S. 569 (1994) (Court rejected conclusion that parody was presumptively unfair just because the work was being sold; remanded for further proceedings). In particular, the jury must consider, at least, the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes, the nature of the copyrighted work, the amount and substantiality of the portion used in relation to the copyrighted work as a whole, and the effect of the use upon the potential market for or value of the copyrighted work. 17 U.S.C. x 107. 222. There are four types of “marks”: trade marks, for goods (like LegoÒ toys); service marks, for services (like GreyhoundÒ bus lines); collective marks, for members of a group (like AAAÒ garages and other automotive service providers); and certification marks, for certifying that certain products meet certain standards (like the Florida Sunshine TreeÒ for juice made from oranges grown in the State of Florida). 223. One can file an “intent to use” application in the USPTO for registration of a mark that is not yet in use, but the applicant must prove the mark is actually in use before the USPTO will issue a registration. 224. As of December 2011, 43 States and the District of Columbia have adopted some form of the Uniform Trade Secrets Act (UTSA), but the legislatures of some may have tweaked the language and courts in the States have interpreted its clauses in subtly different ways. The remainder use common-law principles that are largely consistent with the UTSA, but again have subtle differences.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

526

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

225. Currently, an inventor can ask the Patent Office to keep the application unpublished throughout the process, provided the inventor promises never to apply for patent protection on that invention in any other country. If a patent ever issues on that application, it is published upon issuance. 226. Under the Patent Reform Act, the original inventor (and tradesecret owner) might not be liable for patent infringement, but will lose any battle over who is entitled to the patent. 227. 5 U.S.C. x 552. Similar State laws are often called “open records” acts and, more colorfully, “sunshine” laws. Also, FOIA applies to records held by grantees and contractors to the extent those records were used by the funding agency to formulate policy. See 2 C.F.R Part 215; see also OMB Circular A-110, “Uniform Administrative Requirements for Grants and Other Agreements with Institutions of Higher Education, Hospitals and Other Non-Profit Organizations” (as amended 09/30/1999), available online here. 228. x 552(b)(4). Specifically, it exempts “trade secrets and commercial or financial information obtained from a person and privileged or confidential.” 229. 35 U.S.C. x 205. 230. 18 U.S.C. x 1905. (Civil penalties are also possible under certain circumstances; see 12 U.S.C. x 417.) 231. More precisely, the relevant information includes “trade secrets, processes, operations, style of work, or apparatus, or to the identity, confidential statistical data, amount or source of any income, profits, losses, or expenditures of any person, firm, partnership, corporation, or association.” Id. 232. 18 U.S.C. xx 1831-1839. 233. 18 U.S.C. x 1831. 234. 18 U.S.C. x 1832. 235. Millen P. Commentary: The Economic Espionage Act - is it finally catching on?. St. Louis Daily Record & St. Louis Countain, March 19, 2006, available online here; see also generally, Danielson M. Economic Espionage: A Framework for a Workable Solution. 10 Minn J L Sci & Tech 2009;10:503e48, at pp. 516e18 (discussing why prosecutions are rarely brought). 236. Department of Justice Press Release (June 18, 2008), available online here. Exact and up-to-date figures for pure x 1832 prosecutions are hard to find because these cases are often consolidated with other IP crimes, such as criminal copyright infringement and violation of the Digital Millennium Copyright Act. 237. US v. Zhu, discussed in Biotech spies arrested in Harvard case. Nature Biotech 2002;20:760e1. 238. US v. Okamoto, discussed in Espionage charges threaten to undermine research relations. Nature May 17, 2001;411:225e6. 239. NSF “Science and Engineering Indicators,” Chapter 4 (January 2008), available online here. According to a 2009 commentary in the New England Journal of Medicine, the twenty most researchintensive medical schools receive 80e85% of their research revenues, and a third of total revenues, from federal research grants. Capmbell E. The Future of Research Funding in Academic Medicine. NEJM April 9, 2009;360:1482e83 (citing proprietary data from the Association of American Medical Colleges). 240. For those unfamiliar with these terms, essentially speaking, a contract is used when the goal is to produce something for the benefit of the Government, while a grant and cooperative agreement are used when the goal is for the benefit of the recipient; the difference between a grant and a cooperative agreement is that the latter anticipates substantial involvement of the funding agency. See 31 U.S.C. xx 6303e6305.

241. For contrast, in 2008, the Government spent $398 billion on R&D, which represented 0.73% of GDP and 26% of all domestic funding of R&D. National Science Foundation, Science and Engineering Indicators: 2010, (available online here). 242. “The Bayh-Dole Act at 25,” White Paper by BayhDole25, Inc., p. 14 (April 17, 2006) (available online here). 243. Id., at 16, see also, “Technology Transfer, Administration of the Bayh-Dole Act by Research Universities” U.S. Government Accounting Office (GAO) Report, p. 3 (May 7, 1978). Removing military patents, where normal market forces are overridden by national security concerns, the GAO report notes that fewer than 10% of the 12,000 unexpired patents were licensed. 244. The NSF publishes data annually, with occasional updates midyear. As of June 2016, the most recent data is available here. The NIH by itself provides about $27 billion in grants, contracts, and cooperative agreements, supporting over 100,000 investigators (over 300,000 research staff) across 3000 institutions. 245. “The Role of Federally-Funded University Research in the Patent System,” Statement of the Association of University Technology Managers to the Senate Committee on the Judiciary (October 24, 2007), available online here. 246. The implementing regulations, 37 C.F.R. x 401.14, contain greater detail, and the Federal Acquisition Regulation, 48 C.F.R. Part 1 x 52.227-11 contains the standard patent rights clauses that appear specifically in procurement contracts. 247. Actually, a fifth clause must be added for funding agreements concerning government-owned, contractor-operated facilities. 248. This clause does not apply to funding agreements concerning government-owned, contractor-operated facilities. 249. The public access webpage can be found here. 250. The RePORT system can be accessed here. 251. In addition, the statute permits restricting or eliminating a Contractor’s rights if the Contractor is situated outside the United States, if necessary to conduct or protect foreign intelligence or counterintelligence, or where the funding agreement applies to certain Government-owned, contractor-operated facilities of the Department of Energy. 35 U.S.C. x 202(a)(i-iv). 252. 37 C.F.R. x 401.3(e,f). 253. One article that nicely discusses the nuances of the issues is McGarey B, Levey A. Patents, Products, and Public Health: An Analysis of the CellPro March-In Petition. Berkeley Tech L J Fall 1999;14:1095. 254. A “PMA,” or “Pre-Market Approval” application, is a formal request to the FDA for permission to market a new or high-risk medical device. 255. See, e.g., Baxter, Hopkins Ask for Rejection of CellPro Petition. Biotechnology Law Report MayeJune 1997;16(3):374e80. 256. In dramatic public testimony, the CEO of CellPro told how his own life had been saved using CellPro’s separator to treat his leukemia. 257. Executive Order 10096 (Janurary 23, 1950). 258. 37 C.F.R. Part 501; for the Department of Health and Human Services in particular, see also 45 C.F.R. Part 7. 259. More precisely, EO 10096 and the implementing regulations apply to any invention “made by any Government employee (1) during working hours, or (2) with a contribution by the Government of facilities, equipment, materials, funds, or information, or of time or services of other Government employees on official duty, or (3) which bear a direct relation to or are made in consequence of the official duties of the inventor.”

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

260. Technically, all patents are owned by the Government as a whole, regardless of the agency in which the inventor might have worked, but for administrative convenience in tracking responsibilities and spending royalties, the agency employing the inventor(s) is treated as the “owner” of the invention. Patent applications for Government-owned inventions typically indicate that the patent assignee is “The United States Government, as represented by the Department of Health and Human Services,” or similar phrasing. 261. PL 99-502 (October 2, 1986). 262. Formally, the “Stevenson-Wydler Technology Innovation Act of 1980,” PL 96-480 (October 21, 1980). 263. 15 U.S.C. x 3710(a). 264. PL 104-113. 265. PL 106-404. 266. The top seven in FY2007, which accounted for 96% of all R&D funds spent that year, were DoD ($56 billion), HHS ($29 billion), NASA ($8 billion), DoE ($8 billion), NSF ($4 billion), USDA ($2 billion), and DHS ($1 billion). NSF “Science and Engineering Indicators,” supra n. 237. 267. NIH’s mission is “to seek fundamental knowledge about the nature and behavior of living systems and the application of that knowledge to enhance health, lengthen life, and reduce the burdens of illness and disability;” a broader discussion of its goals can be found online here. 268. USDA’s mission is to “provide leadership on food, agriculture, natural resources, and related issues based on sound public policy, the best available science, and efficient management;” a broader discussion of its goals can be found online here. 269. NASA’s mission is “to pioneer the future in space exploration, scientific discovery and aeronautics research;” a broader discussion of its goals can be found online here. 270. DOE’s mission is “to advance the national, economic, and energy security of the United States; to promote scientific and technological innovation in support of that mission; and to ensure the environmental cleanup of the national nuclear weapons complex; ” a broader discussion of its goals can be found online here. 271. EPA’s mission may be found on its home page, and a further discussion can be found online here. 272. “Federal Laboratory Technology Transfer Fiscal Year 2014: Summary Report to the President and Congress,” available online here. 273. Congressional Budget Office Report, “Federal Support for Research and Development,” Figure 9 (June 2007), available online here. 274. To be sure, some agencies do not rely solely on “invention licenses,” and use in addition “other IP” licenses, which include copyright (mainly software), Plant Variety Protection Act, and other forms of intellectual property. 275. “Principles And Guidelines for Recipients of NIH Research Grants and Contracts on Obtaining and Disseminating Biomedical Research Resources: Final Notice,” 64 FR 72090 (December 23, 1999). 276. The term “research tool” (or “unique research resource”) “is used in its broadest sense to embrace the full range of tools that scientists use in the laboratory, including cell lines, monoclonal antibodies, reagents, animal models, growth factors, combinatorial chemistry and DNA libraries, clones and cloning tools (such as PCR), methods, laboratory equipment and machines.” 277. To be sure, the Research Tools policy does not directly apply to anyone whose work is not supported by the NIH, but the NIH hoped that others would heed the policy anyway.

527

278. Available online here. 279. Available online here. The NIH also issued a brochure, available online here, and a webpage, available here. 280. The NIH licenses, rather than assigns, its technologies, because licensing enables the NIH to have a continuous, meaningful role in the development of its technologies. 281. See, e.g., 35 U.S.C. x 209(d)(1), 37 C.F.R. xx 404.5(b) and 404.7(a)(2). 282. 70 Fed. Reg. 18413 (April 11, 2005), available online on the NIH OTT webpage. 283. See, e.g., OECD, “Guidelines for the Licensing of Genomic Inventions,” 2006, available online here; AUTM, “Nine Points to Consider in Licensing University Technology,” March 2007, available online here; European Commission, “Report from the Commission to the Council and the European Parliament - Development and implications of patent law in the field of biotechnology and genetic engineering,” July 2005, available online here; National Academies of Science, “Reaping the Benefits of Genomic and Proteomic Research: Intellectual Property Rights, Innovation, and Public Health,” December 2006, available for purchase online here. 284. Examples include: expressed sequence tags (ESTs); cDNAs; haplotypes; antisense RNAs; small interfering RNAs (siRNAs); full-length genes and their expression products; as well as methods and devices for the sequencing of genomes, quantification of nucleic acid molecules, detection of single nucleotide polymorphisms (SNPs), and genetic modifications. Since the publication of the Best Practices, the field has moved towards whole-genome sequencing and analysis. 285. Highlights from the “AUTM U.S. Licensing Activity Survey: FY2014” available online here. The full report is available for a fee. 286. See, for example, Restatement (2d) Agency, x 322 (an agent who fails to disclose existence of agency or identity of the principal is personally liable) and x 329 (an individual lacking agency authority may be liable for breach of warranty of agency). 287. See, for example, id. x 1 (definition of agency, principal, and agent), x 26 (creation of actual agency), x 140 (principal liability for the acts of an authorized agent), and x 159 (principal liable for acts of agent with apparent authority). 288. See, for example, id. x 1 comment c (attorney at law) and x 14C (although individual members of the board of directors are not agents of the corporation, officers hired to conduct the company’s business are). In theory, actual authority does not have to be written, see id. x 26 (creation of agency relationship may be oral) and x 27 (creation of agency by apparent authority may be by oral statements of principal), but as a matter of practical reality, agency relationships based on oral statements are difficult to prove. 289. Id. x 27. 290. NIH has 27 subdivisions under its aegis, each of which is either an “institute” or a “center.” The institutes and centers of the NIH, together with the Food and Drug Administration and the Centers for Disease Control and Prevention (see infra, n. 47), are all referred to as “ICs” for simplicity’s sake. 291. Id. x 320. 292. Gwynne P. Corporate collaborations: scientists can face publishing restraints. The Scientist May 24, 1999:1e6. 293. Weiss R. Thyroid drug study reveals tug of war over privately financed research. Washington Post April 16, 1997:A03; King Jr R. Bitter pill: How a drug firm paid for university study, then undermined it. Wall Street Journal April 15, 1997:A01; Rennie D. Thyroid storm [Editorial]. JAMA 1997;277(15):1238e43. 294. Boots Pharmaceuticals was purchased by BASF AG in April 1995.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

528

29. INTELLECTUAL PROPERTY AND TECHNOLOGY TRANSFER

295. Dong BJ, et al. Bioequivalence of generic and brand-name levothyroxine products in the treatment of hypothyroidism. JAMA 1997;277:1205e13. 296. Mayor GH, et al. Limitations of levothyroxine bioequivalence evaluation: analysis of an attempted study. Am J Ther 1995;2: 417e32. Dr. Mayor was also an associate editor of this journal at the time. 297. The Adequacy of Appropriations Act, 41 U.S.C. x 11, and the Antideficiency Act, 31 U.S.C. x 1341. 298. 31 U.S.C. x 1350: “An officer or employee of the United States Government or of the District of Columbia government knowingly and willfully violating section 1341(a) or 1342 of this title shall be fined not more than $5000, imprisoned for not more than 2 years, or both.” Based on the fact that the author is unaware of any case in which the U.S. Department of Justice has even attempted to prosecute anyone for this crime on the basis of an unauthorized indemnification clause appearing in a research-related agreement, jail appears to be an extraordinarily remote possibility. 299. Omnibus Budget Reconciliation Act, Public Law 105-277 (1997). 300. See 65 Fed. Reg. 14406 (March 16, 2000). 301. See, for example, Restatement (2d) Torts x 757 comment b; Roger Milgrim, Milgrim on Trade Secrets x 101 (discussing the Uniform Trade Secret Act); cf. Economic Espionage Act of 1996, 18 U.S.C. x 1839(4) (1997) (definitions). Each state in the United States has its own trade secret law. As a result, there are many overlapping definitions and rules concerning trade secrets. Specific matters should be addressed by attorneys who have particular familiarity with the laws of the jurisdiction in question. 302. Milgrim, x 16.01[7]. 303. 5 U.S.C. x 552. 304. Id., x 552(b)(4). Information generated by a government scientist under a CRADA is also exempt, provided the information is such that it would be deemed a trade secret if it had been given to the government by the collaborator. 15 U.S.C. x 3710a(c)(7). 305. Id., x 552(b)(5). 306. Id., x 552(b)(6). See also Privacy Act of 1974, 5 U.S.C. x 552a. 307. 35 U.S.C. x 205. This exemption only applies for a “reasonable time in order for a patent application to be filed.” 308. 15 U.S.C. x 3710a(c)(7). Of particular note, subparagraph (7)(B) extends the “trade secret” exemption of the Freedom of Information Act to cover data generated by government scientists under a CRADA, provided that the data so generated would qualify as a trade secret if it had been provided by the CRADA collaborator. However, this extra exemption only lasts five years from the development of that data. 309. The CDA goes by other descriptions, including “Non-Disclosure Agreement” or “NDA.” In the context of biomedical products development, however, the acronym “NDA” could be confused with the “New Drug Application,” which is why this that acronym is not used here. 310. Depending on the parties negotiating the agreement, it often, but not always, contains some additional terms. Examples of such provisions include those that specify the law of the agreement (e.g., “Federal law shall govern”), certification provisions (e.g., certification by signer of authority to bind the party), indemnification provisions, and disclaimers of warranties. An attorney should be consulted before any of these provisions are accepted. Although these terms may be common, they do not necessarily have to appear in an agreement to make the agreement valid and binding. 311. See, for example, Weigh Systems South, Inc. v. Mark’s Scales & Equipment, Inc., 68 S.W.3d 299 (Ark. 2002) (although the extent of measures taken to guard secrecy of information is only one of the factors a court will consider in determining its status as a trade secret, it is a prominent factor); Tyson Foods, Inc. v. ConAgra, Inc., 79

312. 313. 314.

315. 316. 317. 318. 319. 320. 321.

322. 323. 324. 325. 326. 327. 328.

S.W.3d 326 (Ark. 2002) (where the employer did not restrict access to secret documents, stamp documents “Confidential,” or notify staff which data the company considered to be trade secret, the company cannot rely on broad, nonspecific secrecy obligations in employment contracts and employee handbooks); Capsonic Group, Inc. v. Plus-Met Corp., 361 N.E.2d 41,44 (1st Dist. Ill. 1986), cert, denied, 505 N.E.2d 353 (1987) (lack of guards or secure zones, no controls over nonemployees visiting the plant, and failure to mark documents or lock them away all suggest company does not consider its know-how a valuable trade secret); ConAgra, Inc. v. Tyson Foods, Inc., 30 S. W.3d 725, 729-30 (Ark. 2000) (“If Tyson did not consider it necessary to preclude the dissemination of pricing information by its customers, why should this court on de novo review enforce the secrecy of that same information?”); Engineered Mechanical Svcs., Inc. v. Langlois, 464 So.2d 329,335-37 (La. App. 1984), cert. denied, 467 So.2d 531 (1985) (despite employment contract terms against disclosure of trade secrets, company failed to communicate to employees which data or documents it considered secret, thus no data were really secret); Electro-Craft, 220 U.S.P.Q. 820-21; MBL (USA) Corp. v. Diekman, 445 N.E.2d 418, 424-425 (1st Dist. Ill. 1983) (despite employment agreement requiring that employees maintain trade secrets, the fact that the company failed to label which documents contained secretsdand that employees were unaware of which was whichdundercut claims of trade secret); Dynamics Research Corp. v. Analytic Sciences Corp., 400 N.E.2d 1274, 1287 (Mass. App. 1980) (the fact that company information is commercially valuable does not make it a trade secret, so the company’s failure to distinguish its secret information from information of general knowledge voids claims of trade secret, despite an agreement promising blanket nondisclosure of trade secrets). Cf., Tele-Count Engineers, Inc. v. Pacific Tel. & Tel. Co., 168 Cal.App.3d 455, (1st Dist. Cal. App. 1985) (in “breach of confidence” tort, plaintiff bears burden of proving that the defendant knew with particularity which information is secret). Federal Register Notice published on May 25, 1999 (64 FR 28205). Black’s Law Dictionary, 6th ed., 1586e89; 1990. See also Corbin A. Corbin On Contracts, x 14 (single-volume edition). Williston S. Williston On Contracts 3d x 1364C (buyer’s consequential damages under the Uniform Commercial Code), x 1394 (general consequential damages for breach of warranty). 41 U.S.C. x 6301. 31 U.S.C. x 1341. 31 U.S.C. x 1350. See the Web page of the Association of University Technology Managers at www.autm.net. 21 C.F.R. x 310.305 and x 312.32. 21 C.F.R. xx 312.50 (general duties of sponsor), 312.53 (selecting investigators and monitors). See, for example, 21 C.F.R. x 312.47(meetings with FDA), x 312.50 (general duties of sponsor), x 312.58 (FDA inspection of sponsor’s records), and x 312.68 (FDA inspection of records of sponsor’s clinical investigator). 21 U.S.C. xx 335a, 335b. 45 C.F.R. Part 46, Subpart A. 42 U.S.C. x 282(c) (“substances and living organisms”). See, e.g., Executive Order No. 10096 (1952), as amended. The Federal Technology Transfer Act, P.L. 99-502 (1986) (amending 15 U.S.C. x 3710a). For example, The National Technology Transfer and Advancement Act, P.L. 104-113 (1995) (amending 15 U.S.C. x 3710a). To be sure, not every agency of the U.S. government views the minimum degree of “collaboration” equally. The only case pertaining to this point is Chem Service, Inc. v. Environmental Protection Agency, 12 F.3d 1256 (3d Cir. 1993). In this case, the Third Circuit suggested

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

that the CRADA statute must be viewed together with procurement and grant statutes, such that if the primary purpose of the interaction is to procure goods for the benefit of the government, the government must use a procurement contract, not a CRADA. Implicitly, then, the CRADA is appropriate where the primary purpose is collaborative research and development. 329. Some confusion occasionally arises between a “cooperative agreement” (15 U.S.C. x 3706), which is a mechanism analogous to a grant by which federal funds can be legally transferred to a private party, and a “cooperative research and development agreement,” which is not a funding mechanism.

529

15 U.S.C. x 3710a(c)(4). Id. x 3710a(c)(4)(B). Id.; see also 35 U.S.C. x 209, 37 C.F.R. Part 404. Although rarely exercised, in instances posing an apparent (but not actual) conflict of interest, the NIH institute has the power to elect to waive that conflict if the research is of overriding importance to the institute and no other PI could carry out the research. 334. 15 U.S.C. x 3710a(b)(l). 335. 15 U.S.C. x 3710a(c)(5).

330. 331. 332. 333.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

30 Data Management in Clinical Trials Diane C. St Germain, Marjorie J. Good National Cancer Institute, National Institutes of Health, Rockville, MD, United States

O U T L I N E The Research Team Principal Investigator and Subinvestigators Research Director/Manager Clinical Trials Nurse Clinical Research Associate Database Administrator Statistician

531 532 532 533 533 533 533

Data Management Data Elements Case Report Forms Choosing a Database System Data Collection Sources of Data Quality Control of Data

533 534 535 536 537 538 538

Conducting clinical research is a time-consuming yet rewarding experience. There are many facets to consider during the development of a clinical trial, all of equal importance. First is having a knowledgeable, adequately staffed research team well versed in the conduct of clinical research with access to the appropriate resources (space, computers, patient access, etc.). Second, a clearly written protocol document that provides the necessary background information and details outlining all aspects of conducting the study. It is during the development phase of a clinical trial that all aspects of data collection and management should be established: what, how, when data are collected, and by whom. Third, rigorous data collection and analysis are crucial, as data reveal the trial’s progress, toxicities, and determine the outcome of a trial. The reported results of a clinical trial should reflect verifiable, accurate data that have been collected in a methodical and judicious

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00030-7

Auditing

538

Unanticipated Problems and Adverse Event Monitoring and Reporting

540

Legal and Regulatory Issues Related to Data Reporting

542

Follow-Up and Analysis

543

Record Retention

543

Conclusion

544

Summary Questions

544

References

544

fashion. Lastly, each member of the research team must be knowledgeable of and comply with the legal and regulatory aspects of clinical research. Conducting clinical research is a significant commitment of time, resources, and money; hence it is vital to have these elements be in place for the trial to succeed. This chapter will discuss the essentials of data management in the conduct of a successful clinical trial.

THE RESEARCH TEAM Teams are defined as two or more people who interact dynamically, interdependently, and adaptively to achieve a common goal.1 The team’s goal(s) should be collectively and clearly defined and should include clarification of the skills needed for each task, who is responsible for each task and who will monitor task

531

Copyright © 2018. Published by Elsevier Inc.

532

30. DATA MANAGEMENT

completion. This effort requires team members to interact dynamically while concurrently investing a significant amount of time and effort toward coordination and communication. Assembling an experienced research team to conduct a clinical trial that interacts dynamically, interdependently, and adaptively is an essential element to ensure its success. Collectively, the team’s responsibilities include protecting the rights and ensuring the safety of research participants. Each staff member is responsible for conducting clinical research ethically, maintaining patient confidentiality, following the Code of Federal Regulations (CFR) and complying with the Health Insurance Portability and Accountability Act (HIPAA) regulations. All research team members are responsible for informing the principal investigator (PI) or research director if they suspect data fraud, manipulation, or other misconduct.2 Each member plays a pivotal role, and attention must be given to ensure adequate training, staffing, and resources that are available to conduct the trial. Staffing requirements and responsibilities should be clearly delineated during the planning phase of the trial. Identification of training needs and who will conduct the training should be planned in advance to ensure adequate preparation by the staff. There are many members of the team including the PI, subinvestigators, research director/manager, clinical trials nurses (CTNs), clinical research associates (CRAs), data managers, quality assurance staff, statistician, pharmacist, radiologist, pathologist, bioethicist, social worker, contract and billing compliance specialists, dieticians, and potentially many other disciplines depending on the nature of the trial. Job titles may vary according to geographic location or the institution and may be shared; for example, one person might fulfill several roles or one role might involve the efforts of several individuals; however, the overall responsibilities generally remain the same. The responsibilities with which each staff member is tasked will depend on how well the team is staffed, the clinical trial infrastructure of the institution, and the type and size of the trial.

Principal Investigator and Subinvestigators The PI carries the primary responsibility for all aspects of the clinical trial including the ethical conduct of research, complying with federal regulations, ensuring all research is submitted to and approved by an Institutional Review Board (IRB), obtaining and documenting informed consent and assent, reporting progress of approved research to the IRB, verifying eligibility of enrolled patients, adhering to the protocol, reporting adverse events, and submitting timely and accurate data.3 The PI determines the efficacy of the study agent

or intervention and determines whether an adverse event is study related. The PI can delegate some of these tasks to qualified personnel, but she/he is the person ultimately responsible and accountable to the sponsor and funding agency. Subinvestigators, also referred to as coinvestigators, may also perform PI functions but under varying circumstances may work under the supervision of the PI. They remain fully responsible for fulfilling all of the obligations of an investigator including responsibility for conducting studies in compliance with all federal and local regulations. Subinvestigators may include associate physicians, research fellows, residents, and CTNs4 if the extent of their involvement includes performing critical study functions and making direct and significant contributions to the clinical data.5 When a sponsor selects investigators to participate in a trial that is being conducted under an investigational new drug application (IND), the sponsor must obtain a completed and signed Statement of Investigator, Food and Drug Administration (FDA) Form 1572 from each investigator before permitting the investigator to begin participation. Under 21 CFR 312.3(b), each subinvestigator is functioning as an investigator and therefore each must sign a separate 1572. The PI and subinvestigator signatures on this legally binding document constitutes the investigator’s affirmation that he/she is qualified to conduct the research and is committed to abide by FDA regulations.5 The form is to be completed as accurately as possible. Any deliberate falsification of information is a criminal offense, which, depending on the severity, can result in disbarment.5 The FDA website maintains a list of physicians prohibited from participating in clinical trials as well as a list of noncompliance with the CFR.

Research Director/Manager The research director or manager is responsible for managing the clinical research department, oversees all clinical research activities and works closely with the PI to modify processes as needed, maintains the integrity and quality of the research being conducted, and anticipates the workload associated with the trial. Individuals in this position ensure every person involved is informed of his/her role and staff is thoroughly trained to maintain the integrity of the data collected. They also are responsible for quality assurance, compliance with regulations and standard operating procedures, proper documentation, developing a plan for addressing research misconduct, and data mismanagement and budgeting for the department.6,7 The director/manager may also work closely with financial and billing compliance specialists in the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

533

DATA MANAGEMENT

development of coverage analyses for individual trials to verify whether components of the trial will be either billable to insurance or covered by the sponsor.

Clinical Trials Nurse In addition to having strong knowledge of the disease or disorder under study, a CTN has solid comprehension of how clinical research is conducted and can apply the basic nursing tenets of clinical and critical thinking skills, bedside experience, interpersonal skills, and patient advocacy.8 The CTN has intimate knowledge of the protocol, data, and regulatory requirements in addition to an understanding of how the clinical trial process impacts the study participants in whose care they may be involved.8 There are many facets to the CTN role including identifying and screening patients, reviewing eligibility criteria, ensuring the initial and ongoing consent process, registering and randomizing patients, scheduling study visits and required tests, conducting staff and patient/family education, and communicating and collaborating with caregivers, investigators, ancillary research staff, primary care physicians or their staff, and trial sponsors.8 In addition, a CTN interviews the patient for toxicities and adverse events, assists in the reporting of serious adverse events and the assignment of attribution, makes dose adjustments per protocol, and works with the PI to develop and evaluate patient management strategies.8 She/he may obtain serum samples including pharmacokinetic sampling and prepare serum and tissue samples for shipping. The CTN also may complete case report forms, input data or may provide oversight to the person doing so. A CTN should be involved during protocol development to review the design of the case report forms and address issues of protocol feasibility regarding patient accrual and data collection. A CTN typically has a bachelor’s degree and often a master’s degree in nursing. In some settings, a portion of the CTN role may be filled by a CRA who is generally not a registered nurse and may have a diverse educational background. Depending on the setting and trial, a CRA may work collaboratively with a nurse or research manager who provides oversight. Regardless of which entity fills the role, this role is critical in the management of a clinical trial. The CTN or CRA are often the go-to person as they are intimately involved with the protocol and will know the protocol requirements at any given timed what, when, and how the data need to be collected.

Clinical Research Associate CRAs are responsible for data collection and documentation as well as are often responsible for managing

trial-related supplies and other research-related support tasks.7 The person in this role should be adept with electronic data capturing systems and analyses. Responsibilities include abstracting data from the source documents into the research record, performing quality checks on data, preparing reports, interim monitoring of the trial, and monitoring patient accrual.4 He/she may complete case report forms or enter data electronically. The CRA reports missing data and discrepancies to the research manager or PI. The CRA will note a lack in source documentation and will follow up with the research team to complete. Because the CRA follows the data so closely, she/he may be the first to note study trends and plays a crucial role in informing the study team of these findings. Finally, the CRA also will play a significant role in an external audit and may coordinate internal audits.

Database Administrator The database administrator (DBA) is an information technology (IT) professional who uses specialized software to store and organize data. The role may include capacity planning, installation, configuration, database design, migration, performance monitoring, security, troubleshooting, as well as backup and data recovery.9 Two related roles include the systems analyst, responsible for all aspects related to trials software including design, development, testing, validation, and problem solving,10 and the programmer who is responsible for writing and maintaining the computer programs.

Statistician The statistician is involved in the design of the study and works closely with the PI and clinical investigators to determine the study end points and to determine if the sample size is sufficient and therefore provides enough power to address the study’s aims.11 The statistician will determine the statistical methodology used in the trial and analyses. Lastly, the statistician is responsible for analyzing the data, including an interim analysis when indicated, the handling of missing data, monitoring the progress of the trial, and communicating trial-specific information to the Data Safety and Monitoring Committee (DSMC).12

DATA MANAGEMENT Data are the cornerstones of a clinical trial and provide safety information as well as study results. It is critical that the data are of high quality and are collected in a timely manner. Determining when, how, and who will collect the data in advance of the study initiation are

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

534

30. DATA MANAGEMENT

essential. Planning ahead and engaging the entire research team during the development of the clinical trial and in writing the protocol document as it relates to their area of expertise will minimize problems once the study is activated. The protocol is the primary source of information to conduct a clinical research study, guiding its users to select and enroll patients, implement the study treatment, and collect data in a consistent manner. The protocol should be written in a clear and concise manner and provide sufficient detail to ensure that the entire clinical research team is able to understand and carryout the research study. Key elements included in the protocol are the background, which provides information and data to support the study design and hypotheses; the study objectives; and the study plan for how the trial will be conducted to meet these objectives. The protocol also should contain detailed information regarding the type and source of data to be collected, how the data will be collected and the schedule for collection. The PI will work closely with the statistician to determine the study end points, sample size, how the data will be analyzed, and what data will be included in the final analysis. Refer to Chapter 16 for more information on writing a protocol. During the design and development of a clinical trial, it is vital to establish the data elements to be collected, the design of the data collection instruments, and the design of the computer database.13 The process for data management should result is “an error-free, valid, and statistically sound database.”14 Data management plans (DMPs) are developed to document processes and procedures for how data is to be handled in a given trial.14,15 The DMP may vary according the type of study being conducted (prospective, observational, registries).15 According to Krishnankutty et al.,14 “The DMP describes the database design, data entry and data tracking guidelines, quality control measures, serious adverse event reconciliation guidelines, discrepancy management, data transfer/extraction, and database locking guidelines.”

Data Elements Data elements to be collected are determined by the objectives of the study. Table 30.1 shows typical data elements captured. Every effort should be made to use common data elements (CDEs) to facilitate data sharing across research projects. CDEs describe the type of data to be collected and use standardized language for the question as well as the associated values.16 The National Cancer Institute (NCI) Cancer Therapy Evaluation Program (CTEP) has developed CDEs in all major cancers and those related to pathology and banking of

TABLE 30.1 Typical Data Elements Captured On-study Demographic data Eligibility criteria Family history Patient history and physical examination including performance status Surgical history Prior treatment Concomitant medications Laboratory and radiology results Pathology Review of current symptoms During the study Study treatment (dosing, frequency) Laboratory and radiology results Concomitant medications including over-the-counter medications and indications Adverse events Toxicities Hospitalizations Treatment response Study termination Treatment stop date and reason Follow-up Disease status Nonprotocol treatment Long-term adverse events Date of death Cause of death Autopsy results if performed

biospecimens to meet the following objectives as outlined by NCI16: • • • • •

To identify discrete, defined items for data collection To promote consistent data collection in the field To eliminate unneeded or redundant data collection To promote consistent reporting and analysis To reduce the possibility of error related to data translation and transmission • To facilitate data sharing NCI CDEs are available for public use and are stored in the Cancer Data Standards Registry and Repositories (caDSR), a registry developed and maintained by the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

535

DATA MANAGEMENT

TABLE 30.2

Study Parameters/Study Calendar

Required Studies

Baseline

Prior to Cycle 1

Prior to Cycle 3

Prior to Cycle 6

Off-Study

History and physical exam

x

Wt, performance status

x x

x x

x x

x x

x x

CBC, differential, platelets

x

x

x

x

x

AST/ALT

x

x

x

x

x

Chest X-ray

x

x

x

x

CT scan abdomen

x

x

x

EKG

x

FACT-QOL

X

x

x x

x

AST/ALT, aspartate aminotransferase/alanine transaminase; CBC, complete blood count; EKG, electrocardiogram; FACT-QOL, Functional Assessment of Cancer Therapy-Quality of Life; Wt, weight.

NCI Center for Bioinformatics and Information Technology (CBIIT). The CDE Browser is used to search, browse, and export CDE from the caDSR and provides information needed to users.16 The required data collection for a clinical trial is often provided in table format and is referred to as study arameters or study calendar (Table 30.2) in the protocol. The CTN or CRA and/or the data manager frequently refer to the study parameters or study calendar, hence it needs to be clearly written with explicit instructions on when and how the data are collected. Data may consist of patient demographic information; physical measurements (e.g., vital signs, weight); laboratory, pathology or diagnostic results; and treatment-related adverse events. Data also may consist of patient reporting of disease and treatment-related symptoms and toxicities and quality of life, which are collected using validated measurement tools. Symptoms and toxicities may be reported by patients using automated reporting via the telephone or via handheld devices such as smartphones or tablets. Forms also may be mailed to participants for completion. The NCI has developed Patient Reported Outcomes-Common Terminology Criteria for Adverse Events (PRO-CTCAE), a patient-reported outcome measure used to report treatment-related toxicities for patients on a clinical trial. PRO-CTCAE is used as a companion to Common Terminology Criteria for Adverse Events (CTCAE) reporting.17 Issues related to feasibility of the data collection should be addressed during the development of the clinical trial. Consideration needs to be given to the amount of data being requested, the frequency of the collection, and the burden placed on the staff as well as the participant. It is important to collect only data that are germane to the aims of the study and will be included in the analysis. The data elements to be collected and method of collection will determine the resources and

staffing required to conduct the trial. Issues that will require additional resources include pharmacokinetics, centralized radiology or pathology review, and correlative studies, including those with tissue and serum acquisition.

Case Report Forms Case report forms (CRFs) are preprinted forms developed by the sponsor or the PI to collect participant data according to protocol and regulatory requirements. The CRF design impacts data quality and the accuracy of its collection. Appropriately designed CRFs are easy to complete, collect consistent and valid data, avoid collection of unnecessary and duplicative data, minimize missing data, and collect all of the data necessary for analysis. If designed properly, audits and data analysis will be conducted with greater ease.18 The CRF can be designed once the protocol is complete, which is ideal, or can be designed during protocol development. If the later approach is chosen, many versions are likely to be created; hence, it needs to be version controlled.18 Case report forms can be either paper based or electronic. Other than the actual design of the form, the issues to consider during development are generally the same for either paper or electronic. There are several advantages to electronic data entry. There are less error rates, as there is no need to transfer data from the primary source document to the paper CRF to the database. There are built-in edit checks to avoid/minimize errors, and if errors in input occur or erroneous data are entered, they can be identified in real time. The trial site as well as the sponsor can rapidly analyze data. Lastly, automatically generated reports can alert the investigator to issues and trends such as an unanticipated number of adverse events or toxicities, data discrepancies, missing

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

536

30. DATA MANAGEMENT

data or slow accrual.18,19 Awareness of these issues can prompt the development of a timely plan to overcome them. The disadvantage of eCRFs is the need for user training and system validation.19 Designing a paper CRF can be tedious and requires attention to detail to avoid inconsistencies and duplication.19 Data should correlate with the statistical software so there is little need to convert the data.13 To minimize transcription and conversion errors, data obtained via the CRF should be clear and concise. Most trials will have multiple forms so it is important to be consistent, as much as possible, in the layout, design, and coding of each form. Lengthy text should be avoided, as this will be more difficult to interpret. The use of CDEs will standardize how the questions are asked. The forms should flow in an organized manner and be easy to complete and submit. The volume of CRFs will vary from trial to trial depending on the type and phase of the trial as well as the trial design. When multiple paper-based forms are used, they can be placed in a binder that is organized according to the flow of the protocol and study visits: on-study, treatment, follow-up, and off-study. Electronically, the flow is the same with access to data forms in the order dictated by the protocol. The sponsor often develops procedure manuals to instruct the participating sites on how to complete the forms and how to meet the data submission requirements. In addition, the sponsor usually has an investigators’ meeting to review all aspects of the protocol including data collection. Sponsors also may conduct a site initiation visit at which time the PI’s responsibilities are reviewed and the protocol is reviewed in detail, including review of the CRFs, data entry, and submission and reporting requirements. All of these efforts are intended to ensure uniformity in the way the protocol is carried out, especially for multisite studies.

Choosing a Database System After the data elements have been determined and the CRFs have been designed, the next step is choosing a computer database system. The system chosen should be designed so that it has the required functionality to meet the needs of the institution’s clinical trial activities. In addition, the chosen system should be integrated into the normal workflow of the institution’s clinical research activities.20 It is important to choose a system that is user-friendly, with training and support materials readily available. Based on these needs, an institution may develop a customized system or may purchase a commercially available system. Considerations when selecting a system include the scope, scalability, interoperability, security, and the underlying structure of the

system.15 “Scalability is the ability of an application or business approach to grow without a significant loss of efficiency or effectiveness.”15 It is important to choose a system that meets the changing data management needs of the institution over time. Interoperability allows data to flow from one system to another to avoid manual data entry. Interoperability applies to data as well, so that the data have shared meaning and can be aggregated.15 Many organizations encourage or require data sharing. The National Institutes of Health (NIH) encourages data sharing and requires a data-sharing plan for some trials.21 When data are to be shared, policies of data governance and maintenance need to be established.15 Regardless of the system chosen, it must meet the requirements, as stated in 21 CFR, part 11, which address electronic records and electronic signatures. Persons using electronic records are required to employ mechanisms to ensure the data are accurate and reliable and have not been altered; to create accurate and complete copies of the records for inspection and review; to protect the records and retrieve when necessary; to limit access to authorized individuals; to readily identify who has entered data and to clearly see when data have been modified, i.e., maintain an audit trail; to hold individuals accountable and responsible for the data under their electronic signature; and to provide appropriate training.22 It also is important to develop standard operating procedures (SOPs), bearing in mind the CFR, when using an electronic database system. According to Tompkins,18 the SOP should include the following: • A coding system: A well-defined coding system with prescribed terminology and business rules to facilitate consistent data entry resulting in easier data analysis. • Security: Computer security includes such basics as a unique password assigned to each individual, scheduled password changes, secured computers, backup tapes stored in a separate location, defined user roles, firewalls, and virus protection. In addition, encryption and secure transfer mechanisms are necessary to protect data. • Edit checks: The database should have edit checks for data entry where an alert or constraint is triggered when data are entered that do not conform to programmed parameters or mandatory fields are not completed. • On-site computer support and help desk availability: Providing technical expertise to solve problems and maintain current systems is essential for increasing efficiency in the conduct of a clinical trial for the investigative site, the sponsor, and the regulatory agency.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

537

DATA MANAGEMENT

Source Document(s) Case Report Forms Sponsor’s Database Audit Sponsor’s Data Report

Patient’s Medical Record (Source Document)

Sponsor’s Database

Site’s Database

FIGURE 30.1 Electronic data capture. Courtesy: Troy Budd.

Regardless of the system chosen, it is important to develop processes and standard operating procedures to ensure the security, integrity, and quality of data.15

Data Collection A clinical trial generates a significant amount of data and requires a meticulous approach to its collection, storage, and analysis. It is critical that staff assigned to a trial is commensurate with the anticipated workload and the staff is appropriately trained. Staff members should not be overtaxed and should have the ability to do the job with adequate time and resources. Equally important is that the staff responsible for data collection and protocol adherence be extremely organized with “an eye for detail.” This includes planning ahead and anticipating potential barriers to obtaining the data. For example, do the operation hours of the lab coincide with when the participant needs labs drawn? Does the date of a required test fall on the weekend or a holiday? Methods used to enter data online include entry via a site-to-site connection or a virtual private network (VPN) data entry via a Web application and electronic gathering at the site.23 More and more handheld devices

are used for data input by both research staff and patients. There are several approaches to electronic data collection (Fig. 30.1). The sites may have a paper version of the CRF which they complete and then enter the data into the computer. Alternatively, there is an eCRF where data are entered directly into data fields. Designs of the data fields include coded values, multiple choice with drop-down menus or text.18 Common problems encountered with data management include lack of source documentation, errors in protocol adherence, missing data, and transcription errors. In addition, a lag in data entry is common but should be addressed since delinquent or data captured retrospectively can impact data accuracy. Every effort should be made to anticipate these problems. Tools that can be useful to stay organized include a data tracking tool, phone logs, and a system to remind treating physicians of protocol requirements.23 Clinical trials management systems (CTMS) are commonly used to aid in the conduct of trials (screening/eligibility, patient schedules, adverse event monitoring, data submission requirements), administrative procedures (billing), and regulatory requirements (IRB documentation, informed consent, and HIPAA authorization).19 No CTMS system is likely to meet all of the institution’s clinical research

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

538

30. DATA MANAGEMENT

needs, particularly if they conduct trials from a variety of sponsors who use proprietary tools and web-based tools for data collection. Ideally, the CTMS should be integrated with the institution’s electronic medical record, grants management system, billing system, and adverse event reporting.19 Patient-related issues contributing to challenges with data collection include poor recall of adverse events and poor compliance. Tools to assist patients to report adverse events and toxicities include diaries and more recently automated IT-based telephone monitoring systems or handheld devices such as smartphones or tablets. Compliance can be enhanced by the use of study calendars, either paper or electronic versions, to remind patients to take study medication or of their scheduled tests and appointments. Creating a clinical research folder for each patient is useful to house all study-related documentation including copies of completed CRFs, signed informed consent, eligibility checklist, laboratory, radiology and pathology results, and the study calendar. The folder can be referred to readily as questions arise, when patient information is required quickly and during an audit.

Sources of Data Source documents are where data are first recorded.24 Data entered in an eCRF are considered the source data. FDA can request additional documentation during an inspection to support the data entered electronically.20 The source document verifies the protocol tests and results, that the study participants are real, and that the treatment is given in accordance with the protocol. The source document can be the original or a certified copy. Examples of source documents are included in Table 30.3. The source documents should reflect a patient’s participation on a trial including eligibility and the events that occurred once enrolled on the trial. An auditor should be able to reconstruct the patient’s course of treatment on a study by piecing together all of the data obtained from the original source. The only way to determine protocol adherence is to refer to the source document. Data that are not documented or verifiable do not exist in the eyes of the sponsor or auditor. Data may be captured using patient diaries, calendars, questionnaires, phone logs, and mail surveys. These tools can be considered source documents when signed and dated and are used to document side effects, adverse events, and patient compliance with study medications that are administered orally.23 Sponsors may vary regarding their acceptance of these as source documents so it is important to verify what is acceptable and the sponsor’s requirements for their use.

TABLE 30.3 Source Documents Hospital records Clinical and office charts Lab reports Pathology reports Surgical reports Radiology reports Physician progress notes Nurses notes Letters from referring physicians Original radiological films Pathology slides, tissue blocks Tumor measurements Participant diaries, medication logs Participant interviews

Quality Control of Data Data submitted in paper form to the sponsor are checked for completeness as well as accuracy. In return, the site will receive queries to address inconsistencies in the data or provide clarification. If corrections are required on the CRF, this is done with a single line through the data entered, initialed, and dated. Erasing, using whiteout, or otherwise obliterating an error is prohibited.24 Data added in response to a query also are initialed and dated. Copies of all queries and responses are kept as part of the trial record. Data queries are a normal process of conducting a clinical trial. Despite the best efforts of the staff, they are unavoidable. Data submitted electronically tends to have fewer queries, as there are edit checks at the time of entry. In addition, these checks limit missing data.

AUDITING Quality assurance measures such as audits and monitoring visits are required to ensure that clinical trials are carried out as designed, that good clinical practices (GCP) and regulations governing the conduct of clinical trials are followed, and that the data used to assess treatment effects are accurate and complete.25 More importantly, these efforts are necessary to assure that the rights, welfare, and safety of clinical trial participants are protected.26 Audits and monitoring visits are conducted by various entities including federal (e.g., NCI, FDA, Office of Human Subjects Protections [OHRP]), nonfederal sponsors (e.g., pharmaceutical companies,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

539

AUDITING

Recommended Regulatory File Documentsa

contract research organizations [CRO]), accreditation bodies (e.g., Association for the Accreditation of Human Research Protection Programs [AAHRPP]) as well as by internal institutional research and/or compliance offices.27 Audits associated with clinical research endeavors generally consist of the review of three key components: (1) IRB and informed consent; (2) pharmacy/drug storage and accountability; and (3) patient case review. Assessing these components requires the review of many documents such as study-specific regulatory forms, consent documents, drug accountability logs, case report forms, and participant medical records. Maintaining accurate and up-to-date trial-related regulatory documents in a regulatory file or study binder controlled by a protocol information or research office is a recommended component of GCP guidelines and key to undergoing an efficient and successful audit. Table 30.4 provides examples of what documents should be maintained in a regulatory file. During the audit, data entered into case report forms will be checked against the source documents for accuracy. Data include verification of informed consent, study eligibility, compliance with the study parameters, radiologic and laboratory results, disease outcome and response, concomitant medications, toxicities, and adverse events. The auditor also will verify that the study agent and/or treatment was given according to schedule using the proper dose and administration dictated in the protocol. Additionally, the quality of the data will be assessed for overall completeness, accuracy, and timeliness. The pharmacy will be audited regarding where and how the study agent is stored, dispensed, and disposed. Investigational agent shelf inventory records will be compared to the drug inventory logs, and the drug accountability records will be compared to the participant medical records.28 The auditor also will verify procedures for randomization and for maintaining the blind of treatments. The FDA conducts either announced and unannounced inspections or audits of clinical research sites under the following circumstances:29

TABLE 30.4

• To verify the accuracy and reliability of the data that has been submitted • As a result of a complaint to the FDA about the conduct of a study at a particular site • In response to sponsor concerns • Upon termination of a clinical site • During ongoing clinical trials to provide real-time assessment of the conduct of the trial and protection of human subjects • At the request of an FDA review division • Related to certain classes of investigation products that FDA has identified as products of special interest.

Clinical study reports

Form FDA 1572 for each investigator Current curriculum vitae (CV) for principal investigator (PI) and each subinvestigator Copy of medical licenses for PI and each subinvestigator Training certificates Study staff responsibilities and signature form Financial disclosure statements Clinical trial agreement or other contract/legal agreements Investigator’s brochure and updates Protocol and amendments Study manual of procedures and updates (if applicable) Informed consent documents Advertisement/recruitment materials Federalwide Assurance number(s) and expiration date(s) IRB/Institutional ethics committee membership roster IRB correspondence Institutional Review Board (IRB) approvals Copy of serious adverse event reports sent to sponsor Investigational new drug safety report received from sponsor Safety and data monitoring committee reports and correspondence Laboratory certification and/or accreditation(s) Lab normal reference ranges Subject screening and enrollment logs Subject identification code list (if applicable) Treatment allocation and decoding documentation (if applicable) Blank set of case report forms (CRFs) Site initiation visit report Final trial closeout monitoring report Monitoring log and reports Monitor correspondence

Notes to file (if applicable) a

Grant N, White T. Creating and maintaining a regulatory file. ONS Manual for Clinical Trials Nurses. 2016. Oncology Nursing Society.

Though unannounced inspections are typically generated by issues related to a specific study or participant, all of the site’s clinical trial activities will be at risk to be reviewed to obtain a broad view of clinical research conducted at the site. At the conclusion of an audit, the sponsor audit team or FDA inspector will conduct an exit interview with the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

540

30. DATA MANAGEMENT

PI and research staff to discuss general audit findings. This is then followed by a formal written report in which deficiencies found will be described and audit ratings will be assigned. For audit ratings reported as “acceptable needs follow-up”, “unacceptable”, or “an official action is indicated”,27 the research site must provide a corrective and preventive action plan (CAPA). The CAPA must address each major deficiency within a designated timeframe describing the actions the research site plans to take to correct and prevent recurrence of future deficiencies.30 If the audit reveals evidence of serious noncompliance, sponsor actions can include probation, issuance of warning letter(s), suspension of patient enrollment privileges, repeat audit, denial of access to investigational agents, and/or termination of a grant or contract.27 Some deficiencies found during FDA and sponsor audits include failure to follow the investigational plan and agreement, protocol deviations, inadequate record keeping, inadequate drug accountability for investigational products, and inadequate subject protection, including informed consent issues.29 Audits can be a source of stress and concern for research staff; however, they can and should be considered as an educational opportunity that will ultimately improve a research program’s quality. Preferably, the best preparation for an audit is performed prospectively.28 Research sites should not rely on external auditors or reviewers to assess their programs but rather should conduct periodic self-review of their research records and the research team, identify problems and correct them internally via the process of CAPAs. Someone other than the person who has submitted the data should do quality control checks. Internal audits should be done routinely, not just to prepare for an impending external audit. At the time of an external audit, the deficiencies found during an internal audit may still be rated as deficiencies; however, they may be rated less severely because a CAPA has already been instituted.28 Consistent internal review of a research program’s data management processes and addressing deficiencies with well thought-out corrective action plans can stimulate improvements in their overall conduct of clinical trials,28 which can translate into improved future external audits.

UNANTICIPATED PROBLEMS AND ADVERSE EVENT MONITORING AND REPORTING A critical issue in the conduct of clinical research is the maintenance of safety and protection of the participants. One measure to ensure patient safety that has been mandated by regulatory agencies is for all clinical

studies to be reviewed and approved by an IRB before it is initiated as well as for all clinical studies to undergo continuing review at intervals appropriate to the degree of risk presented by the study, but at least annually.31 For the IRB to meet its obligation for continuing review, it must have information concerning interim unanticipated problems (UPs), including adverse events (AEs) and/or other problems that may have warranted substantive changes in the protocol, informed consent process/document, or other corrective actions undertaken to protect the safety welfare or rights of subjects.32,33 Hence, it is imperative to understand what constitutes a UP, to monitor for them and report them appropriately. According to the OHRP33 unanticipated problems are incidents, experiences, or outcomes that meet all of the following criteria: 1. Unexpected (in terms of nature, severity, or frequency) given (1) the research procedures that are described in the protocol-related documents, such as the IRB-approved research protocol and informed consent document, and (2) the characteristics of the subject population being studied; 2. Related or possibly related to participation in the research (in this guidance document, possibly related means there is a reasonable possibility that the incident, experience, or outcome may have been caused by the procedures involved in the research); and 3. Suggests that the research places subjects or others at a greater risk of harm (including physical, psychological, economic, or social harm) than was previously known or recognized. It is important to understand unanticipated problems are not isolated only to AEs. There are other types of incidents, experiences, and outcomes that may occur during the conduct of a study such as dosing errors, individual identifiable data loss, and drug contamination or manufacturing errors that also warrant prompt reporting to the IRB, appropriate institutional officials, study sponsor, and OHRP.33 The FDA defines an AE as any untoward medical occurrence associated with the use of a drug in humans, whether or not considered drug related.34 An adverse event or suspected adverse event (SAE) is considered “serious”34 if, in the view of the investigator or sponsor, it results in any of the following outcomes: • Death • Life-threatening event • Hospitalization or prolongation of existing hospitalization • Persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

UNANTICIPATED PROBLEMS AND ADVERSE EVENT MONITORING AND REPORTING

• Congenital anomaly or birth defect • Other conditions that may jeopardize the patient and may require medical or surgical intervention to prevent one of the outcomes listed above. Not all AEs are required to be reported to the IRB. In a 2009 guidance document, the FDA advises an AE observed during the conduct of a study should be reported to the IRB only if it was unexpected, serious, and would have implications for the conduct of the study.35 Similarly, OHRP provides three questions that should be asked when determining if an AE should be reported: (1) Is the AE unexpected? (2) Is the AE related or possibly related to participation in the research? (3) Does the AE suggest the research places subjects and others at a greater risk than previously known or recognized? If the answer to all three questions is “yes,” then the AE is an unanticipated problem and must be reported.33 OHRP has provided a visual diagram that further clarifies which AEs and UPs need to be reported (Fig. 30.2). Other entities beyond the IRBs also may need to be informed of study-related UPs, AEs, and SAEs including the study sponsor, the FDA, and other institutional biosafety or research oversight committees. Reporting requirements vary depending on whether the study medication is commercially available or investigational. Each protocol document should provide specific information pertaining to whom, what, and when to report these occurrences. The time frame for reporting will vary depending on the drug or device and the event. If the sponsor is applying for an IND, they will be required to provide a safety report to the FDA and to all other investigators conducting trials using that agent. The data provided in the IND safety reports may lead to protocol amendments, revisions in the consent document, and/ or changes to the investigator’s brochure.36

541

It is important to use standard terms and definitions to represent clinical concepts when reporting AEs as this will lead to better communication and sharing of data among investigators, sponsors, drug manufacturers, and regulatory agencies such as the FDA.36 Standardization also facilitates safety monitoring, analysis, and drug development. There are several standard systems of reporting AEs and the protocol will cite the system and version to be used. The CTCAE,37 located at http://ctep.cancer.gov/protocol development/electronic_applications/ctc.htm, is a standardized method of reporting adverse events which is required in clinical trials sponsored by the National Cancer Institute and has become standard in industrysponsored cancer trials and drug labels.38 Version 4.0 is the currently available version; however, there are several subcategories and older versions available. Research staff should therefore be aware of the trialspecific CTCAE version being used. As the CTCAE is updated older protocols may use lower versions. The CTCAE is a subset of the Medical Dictionary for Regulatory Activities (MedDRA) terms that are pertinent to AEs seen in oncology clinical trials.36 For each AE within CTCAE, clinical grades can range from 0 to 5, representing severity levels of none, mild, moderate, severe, life-threatening, or death, although the maximum possible grade and specific grade level description/criteria can vary depending on the AE.39 It is imperative that the clinical trials nurse and others caring for the study participant are well versed in the expected and potential AEs of the study treatment. The protocol and the investigator’s brochure describe the expected AEs and serve as resources to the research team.36 The informed consent document serves this purpose for the participants. Participants and family members should be educated regarding the expected AEs, including instructions on reporting to the research

Unanticipated Problems

A= Adverse Events that are not Unanticipated Problems

C= B= Unanticipated Adverse Events Problems that are that are not Unanticipated Adverse Events Problems

Adverse Events Under 45 CFR part 46: Do not report A, Do report (B+C)

FIGURE 30.2 General relationship between adverse events and unanticipated problems.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

542

30. DATA MANAGEMENT

TABLE 30.5

National Cancer Institute Adverse Event (AE) Reporting Requirements

Relationship

Attribution

Description

Unrelated to investigational agent/intervention

Unrelated

The AE is clearly not related to the intervention

Unlikely

The AE is doubtfully related to the intervention

Possible

The AE may be related to the intervention

Probable

The AE is likely related to the intervention

Definite

The AE is clearly related to the intervention

Related to investigational agent/ intervention

team and guidelines regarding what are deemed a medical emergency. All AEs experienced by the participant in a clinical trial must be documented in the participant’s medical record (source document). Once a severity of the event is established, the attribution or relationship of the event to the research needs to be determined and a category assigned. Table 30.5 describes AE attribution categories. Even though the CTN may be responsible for gathering information about the AE and collaborating with the investigator to assign attribution, the clinical investigators and ultimately the clinical trial PI have the primary responsibility for AE identification, documentation, grading, and assignment of attribution to the investigational agent/intervention.40

LEGAL AND REGULATORY ISSUES RELATED TO DATA REPORTING All individuals participating in the conduct of clinical research, regardless of their role, are responsible for having the requisite knowledge and understanding of the legal and regulatory aspects of research. This may be a challenge at times because the regulations change periodically. However, all required information and changes are easily accessible on the FDA and OHRP websites. Regulations, guidance documents, and standards for conducting research are living documents that are often edited and updated. There are options available on the websites to request notification of any changes and updates via email (Table 30.6). Feigning ignorance is not accepted as the information is readily available, and it is expected that all individuals participating in research will comply with the standards and rules governing the conduct of clinical research. Some consider the regulatory requirements to be daunting; however, knowledge and compliance of the regulations establish a solid

TABLE 30.6 Office for Human Research Protections and U.S. Food and Drug Administration Mailing Lists Organization

Description

Office for Human Research Protections (OHRP) http://www.hhs.gov/ohrp/ news/sign-up-forannouncements/index.html

Join OHRP’s News distribution list by sending an email to: [email protected], with the following text message body: SUBSCRIBE OHRP-L First name Last name, replacing First name and Last name with your name. Your email address will automatically be captured from the “From” address of your message.

U.S. Food and Drug Administration (FDA) www.fda.gov/emaillist.html

Free email alert service allows you to receive important FDA news and information as they become available. When signing up you will be asked to select the topics that interest you. To subscribe, all you need is a valid email address.

foundation to conduct research with the protection and well-being of the study participants the overarching aim. In the United States the conduct of clinical trials is governed by regulations enacted by the Department of Health and Human Services (DHHS) with the two primary DHHS regulatory groups being OHRP and FDA. Both OHRP and FDA have regulations (CFR) (see Chapters 2e5) and several guidance documents with which individuals participating in the conduct of clinical research should be familiar.41 Any research that receives federal funds will be regulated by the federal government and managed by OHRP. The FDA regulates research that involves the licensing of a drug or product. The CFR is to be followed during the conduct of clinical research in the United States. The CFR provides a

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

543

RECORD RETENTION

guide to the design, conduct, performance, monitoring, analysis, and reporting of clinical trials. The regulations ensure sound data and the protection of subjects participating in research.42 In addition to these regulations there are the International Conference on Harmonization (ICH) Guidelines, state laws, and standard operating procedures of the sponsor and institution conducting research.

FOLLOW-UP AND ANALYSIS The trial enters the follow-up phase once the accrual goals are met. No new patients will be entered on the trial though data still are collected. The length of this phase will vary according to the trial design and study end points. For example, if survival is the primary end point in an oncology trial, the follow-up phase could be quite lengthy.42 Typically, the data collected during the follow-up phase is kept to a minimum. Once the study is closed at a site, the sponsor may conduct a closeout visit at which time all CRFs will be reviewed and queries completed, all extra CRFs are returned or destroyed, the site’s regulatory binder is reviewed to ensure consistency with sponsor’s master file, all study supplies are returned or destroyed, all biologic samples have been shipped or backup samples destroyed, and the PI has provided the IRB with the final study report.27 In addition, records will be reviewed to verify all significant protocol deviations (study procedures not conducted according to protocol, enrollment of inappropriate subjects, dosing errors, consenting errors, unblinding, subjects developing withdrawal criteria yet continuing in study, etc.); AEs and SAEs have been properly recorded and the sponsor/IRB has been notified as appropriate.43 A significant number of data queries may be generated at this time as the sponsor prepares for final analysis. Arrangements will be made to store the records for the duration determined by the sponsor.27 When all of the data are submitted and are complete, the database will be locked to prepare for final data analysis. The report generated from this analysis will be used to report the results of the trial in one or several publications.

RECORD RETENTION Data and records from a clinical trial must be maintained. This will include copies of completed case report forms, source documentation, all correspondence, pharmacy records, and all regulatory documents. The ICH GCP, OHRP, and FDA have varying requirements for

the retention of records (Table 30.7). The sponsor and local institutions also may have their own policies, which may be more stringent than federal requirements. Research sites should comply with the strictest policy.41 Studies conducted internationally will need to follow the rules for the country conducting the study. At a minimum, these records must be maintained for 2 years following the date that the marketing application is approved for an IND, or if application is disapproved, 2 years after shipment and delivery of the drug for investigational use is discontinued and the FDA is notified. IRB records are to be kept for at least 3 years after study completion.44e46

TABLE 30.7

Records Retention Policies

Regulatory Agency

Records Retention Policy

International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH GCP) http://www.ich.org/fileadmin/ Public_Web_Site/ICH_Products/ Guidelines/Efficacy/E6/E6_R1_ Guideline.pdf

E6dSection 5.5.11: The sponsor specific essential documents should be retained until at least 2 years after the last approval of a marketing application in an ICH region and until there are no pending or contemplated marketing applications in an ICH region or at least 2 years have elapsed since the formal discontinuation of clinical development of the investigational product. These documents should be retained for a longer period however if required by the applicable regulatory requirement(s) or if needed by the sponsor.

Office of Human Research Protections (OHRP) http://www.hhs.gov/ohrp/ regulations-and-policy/ regulations/45-cfr-46/index. html#46.115

45 CFR 46.115: The records required by this policy shall be retained for at least 3 years, and records relating to research, which is conducted, shall be retained for at least 3 years after completion of the research. All records shall be accessible for inspection and copying by authorized representatives of the department or agency at reasonable times and in a reasonable manner.

U.S. Food and Drug Administration (FDA) http://www.accessdata.fda.gov/ scripts/cdrh/cfdocs/cfcfr/ cfrsearch.cfm? cfrpart¼56&showfr¼1

21 CFR 56.115: The records required by this regulation shall be retained for at least 3 years after completion of the research, and the records shall be accessible for inspection and copying by authorized representatives of the Food and Drug Administration at reasonable times and in a reasonable manner

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

544

30. DATA MANAGEMENT

CONCLUSION Data management plays a critical role in the outcome of a clinical trial and depending on how well it is done, contributes to its success or failure. Key elements that contribute to the successful conduct of a clinical trial include a clearly written protocol containing a detailed statistical and data management plan, and an adequately trained and staffed research team who are knowledgeable about the regulatory issues and ethical conduct of research. All research team members must be included in the design and development of a clinical trial and writing of the protocol. Conducting clinical research can be very rewarding; it provides an opportunity for the team to be at the forefront of emerging science and new discovery.

SUMMARY QUESTIONS 1. The principal investigator of a clinical trial is responsible for: a. The eligibility of enrolled patients b. Adverse event reporting c. Timely reporting of data d. All of the above 2. Source documents include all of the following except: a. Participant diaries b. A telephone call c. Photographs d. Medication logs 3. An adverse event is defined as: a. A known toxicity of the study agent b. A side effect caused by the study agent c. An unexpected event related to the study agent d. An untoward event that may or may not be related to the study agent 4. A for-cause audit conducted by the Food and Drug Administration is best described as: a. A routine audit to review the regulatory documents, assess the trial’s progress and accrual b. An audit performed in preparation for a drug or device to be marketed c. An audit performed to investigate possible fraud or misconduct d. A prearranged audit to review the storage, dispensing, and disposal of an investigational agent used in a clinical trial

References 1. Taplin SH, et al. J Oncol Pract May 2015;11(3):231e8. http://dx.doi.org/ 10.1200/JOP.2014.003376. Epub 2015 April 14. 2. https://ori.hhs.gov/images/ddblock/data.pdf.

3. Office for Human Research Protections. Investigator responsibilities e Frequently asked questions. Retrieved from: http://www. hhs.gov/ohrp/regulations-and policy/guidance/faq/investigator -responsibilities/. 4. Schmotzer GL, Ness E. The research team. In: Klimaszewski AD, Bacon M, Eggert JA, Ness E, Westendorp JG, Willenberg K, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2016. p. 77e87. 5. U.S. Food and Drug Administration (FDA) Information Sheet Guidance for Sponsors, Clinical Investigators and IRBs e Frequently Asked Questions e Statement of Investigator (Form FDA 1572). Retrieved from: http://www.fda.gov/downloads/ regulatoryinformation/guidances/UCM214282. 6. https://ori.hhs.gov/images/ddblock/data.pdf. 7. http://www.payscale.com/research/US. 8. ONS CTN Competency Statement. https://www.ons.org/sites/ default/files/ctncompetencies.pdf. 9. https://en.wikipedia.org/wiki/Database_administrator. 10. http://hiring.monster.com/hr/hr-best-practices/recruiting-hiringadvice/job-descriptions/systems-analyst-job-description.aspx. 11. Madsen LT, Ness E. Protocol development and response assessment. In: Klimaszewski AD, Bacon M, Eggert JA, Ness E, Westendorp JG, Willenberg K, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2016. p. 127e40. 12. http://www.bls.gov/ooh/math/statisticians.htm#tab-2. 13. McFadden E. Data definition, forms, and database design. In: McFadden E, editor. Management of data in clinical trials. 2nd ed. Hoboken: Wiley-Interscience; 2007. p. 33e55. 14. Krishnankutty B, et al. Data management in clinical research: an overview. Indian J Pharmacol 2012;44:168e72. 15. Reeves D, Walden A. Data management and electronic data management systems. In: Klimaszewski AD, Bacon M, Eggert JA, Ness E, Westendorp JG, Willenberg K, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2016. p. 369e95. 16. https://wiki.nci.nih.gov/display/caDSR/ CTEPþCommonþDataþElements. 17. http://healthcaredelivery.cancer.gov/pro-ctcae/. 18. Tompkins A. Data management in clinical trials. In: Gallin JI, Ognibene FP, editors. Principles and practice of clinical research. 2nd ed. Burlington: Academic Press; 2007. p. 67e76. 19. Bellary S, Krishnankutty B, Latha MS. Basics of case report form designing in clinical research. Clin Data Manage 2014;5(4):159e66. 20. Ochs MF, Casagrande JT. Information systems for cancer research. Cancer Invest 2008;26:1060e7. 21. https://grants.nih.gov/grants/policy/data_sharing/data_sharing _guidance.htm. 22. https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFR Search.cfm?CFRPart¼11. 23. Razvillas B. The need for data management tools. In: Klimaszewski AD, Bacon M, Deininger HE, Ford BA, Westendorp JG, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2008. p. 273e85. 24. Hall TD. Documentation and forms submission. In: Klimaszewski AD, Bacon M, Deininger HE, Ford BA, Westendorp JG, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2008. p. 287e8. 25. Morrison, et al. Monitoring the quality of conduct of clinical trials: a survey of current practices. Clin Trials 2011;8:342e9. 26. GCP ICH E6 (Section 2.0) Located at: http://www.ich.org/ fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/ E6/E6_R1_Guideline.pdf. 27. Patricia EA, Meadows SA. Preparing for audits, inspections and monitoring visits. In: Klimaszewski AD, Bacon M, Eggert JA, Ness E, Westendorp JG, Willenberg K, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2016. p. 403e11.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

28. You YN, et al. The audit process and how to ensure a successful audit. In: Cancer Clinical Trials: Proactive Strategies. US: Springer; 2007. p. 179e97. 29. http://www.fda.gov/downloads/RegulatoryInformation/Guidances/ UCM126553.pdf. 30. http://www.fda.gov/downloads/Training/CDRHLearn/UCM421767 .pdf. 31. 21CFR part 50, 56, 312 and 812. 32. U.S. Food, Drug Administration. Guidance for clinical investigators, sponsors and IRBs: adverse event reporting to IRBs e Improving human subject protection. January 2009. Retrieved from: http://www.fda. gov/downloads/regulatoryinformation/guidances/UCM126572. pdf. 33. OHRP Guidance (2007). Retrieved from: http://www.hhs.gov/ ohrp/regulations-and-policy/guidance/reviewingunanticipated-problems/. 34. U. S. Food and Drug Administration. 21 CFR 312.32. Retrieved from: http://www.ecfr.gov/cgi-bin/text-idx?SID¼b30935cc48333 b7d3c646a4e34b2cac5&mc¼true&node¼se21.5.312_ 132&rgn¼div8. 35. U.S. FDA Guidance (2009). Retrieved from: http://www.fda.gov/ downloads/regulatoryinformation/guidances/UCM126572.pdf. 36. Ness E, Lau Clark AM. Adverse events in ONS manual for clinical trials nurses. In: Klimaszewski AD, Bacon M, Eggert JA, Ness E, Westendorp JG, Willenberg K, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2016. p. 271e89. 37. Located at: http://ctep.cancer.gov/protocoldevelopment/electronic_ applications/ctc.htm.

545

38. Basch E, et al. Development of the National cancer Institute’s patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE). J Natl Cancer Inst 2014; 106(9):1e11. 39. Basch E, Rogak LJ, Dueck C. Methods for implementing and reporting patient-reported outcome (PRO) measures of symptomatic adverse events in cancer clinical trials. Clin Ther 2016 (in press). 40. National cancer Institute guidelines for investigators: adverse event reporting requirements for DCTD (CTEP and CIP) and DCP INDs and IDEs. 2013. Located at: http://ctep.cancer.gov/protocolDevelopment/ electronic_applications/docs/aeguidelines.pdf#search¼% 22attribution%22. 41. Brown S, Markus S, Bales CA. Legal, regulatory, and legislative issues. In: Klimaszewski AD, Bacon M, Eggert JA, Ness E, Westendorp JG, Willenberg K, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2016. p. 51e64. 42. Wood LS. Adverse events. In: Klimaszewski AD, Bacon M, Deininger HE, Ford BA, Westendorp JG, editors. Manual for clinical trials nursing. Pittsburgh: Oncology Nursing Society; 2008. p. 197e214. 43. ClinOps Toolkit. Clinical trial site Close-out visit. 2008. Retrieved at: http://clinopstoolkit.com/2008/07/close-out-visits.html. 44. ICH GCP (E6 e Section 5.5.11). http://www.ich.org/fileadmin/ Public_Web_Site/ICH_Products/Guidelines/Efficacy/E6/E6_ R1_ Guideline.pdf. 45. OHRP (45 CFR 46.115). http://www.hhs.gov/ohrp/regulationsand-policy/regulations/45-cfr-46/index.html#46.115. 46. FDA (21 CFR 56.115). http://www.accessdata.fda.gov/scripts/ cdrh/cfdocs/cfcfr/cfrsearch.cfm?cfrpart¼56&showfr¼1.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

31 Clinical Research Data: Characteristics, Representation, Storage, and Retrieval James J. Cimino University of Alabama School of Medicine, Birmingham, AL, United States

O U T L I N E Introduction

547

Data as Surrogates The Indirect Nature of Clinical Research Data Objectivity and Subjectivity of Clinical Data Transparency, Rigor, and Reproducibility Metadata

547 548 548 549 549

Types of Data

550

Data Standards Data Capture, Storage, and Retrieval

552 552

Responsible Stewardship of Data

553

Cooperative Sharing Efforts

555

Summary

556

550

Summary Questions

556

551

References

557

INTRODUCTION Clinical research data are the basic building blocks of clinical research. Indeed, clinical research phases can be defined as either involving data collection or data analysis. Once in the latter phase, the data are all that remain for analysis, publication, and hypothesis generation. Although good data collection and management practices cannot make up for poor study design, poor data collection and management can render a perfectly executed trial useless. The failure of a clinical study to produce generalizable knowledge because of bad data practices carries both resource and ethical costs. The investigator or sponsor can be held accountable for the former, but nothing can compensate subjects if the research that exposed them undue harm is worthless. It is incumbent on every investigator to understand the statistical rules of study design and the rules and regulations that govern the submission of data and the

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00031-9

Clinical Trials Data Management Systems Clinical Data Repositories

protection of subject privacy. This chapter deals with the practical aspects of data: what they are, how they are represented, and how they are stored to support their retrieval and use.

DATA AS SURROGATES The process of clinical research typically involves many objects and events, such as subjects and their body parts, drugs and devices, and diagnostic and therapeutic procedures. Except for specimens retained in a biobank, very few of these objects and events remain with the researcher for analysis. What is left, instead, are the data collected about these objects and events. It is therefore imperative to recognize that the data are not “real” but merely surrogates of the “truth” of what has taken place in the real world. This perspective is required to understand that how the data are captured

547

Copyright © 2018. Published by Elsevier Inc.

548

31. CLINICAL RESEARCH DATA

and recorded will affect how they can be interpreted. If we pretend that the data themselves represent the truth, we will be caught off guard by biases and fail to interpret the data in ways that can predict future events in the real world.

The Indirect Nature of Clinical Research Data Typical measurements made in the course of clinical research are often simple markers of complex states or are single endpoint assessments that capture the outcome of some complex sequence of events. For example, hematocrit is often used as an indicator of the oxygen-carrying capacity of blood. It also can be used as an indicator of the health of the bone marrow where red blood cells are produced, or as an indicator of response to an acute intervention such as a red blood cell transfusion. However, the actual number reported as “hematocrit” is none of these things: it is the percentage of the blood volume occupied by red blood cells after blood is centrifuged in a heparinized microcapillary tube under controlled forces and for a prescribed period of time. Thus, the hematocrit measurement tells us nothing directly about how much oxygen is actually being transported, nor how much is released to the tissues; this depends on other factors such as chemical alterations of the hemoglobin molecule or the concentration of hydrogen ions in the blood. It is not even always a reliable measure of the amount of blood in the subject’s body; for example, it can be falsely normal in the setting of an acute hemorrhage. The interpretation of even as simple and reproducible a datum as a hematocrit measurement therefore is only as accurate as the attendant contextual information (has the subject been exposed to carbon monoxide or is he/she bleeding?). Similarly, endpoint assessments that seem straightforward, such as mortality or the occurrence of an event, mask many assumptions that have consequences for the statistical interpretation of data and for the subsequent understanding of study outcomes. For example, hidden in a simple classification of survival are decisions regarding how to statistically handle patients who died from causes unrelated to the condition or treatment being studied, how to account for patients who are lost to follow-up, how to adjust for normal age-associated mortality, etc. These decisions need to be identified at the time of study design so that the data necessary to make appropriate corrections are collected. It also is easy to lose track of the larger context of the study and the assumptions underlying the choice of data to collect, especially when these assumptions are part of the prevailing wisdom at the time of the study’s design. a

These observations lead to practical recommendations for data management: recognize when observations are really surrogates for outcomes of interest, plan to capture all the data needed to make necessary interpretations and statistical corrections, make explicit all the steps that lead from the surrogate to the outcome of interest, and then examine those steps to see if they may be introducing undue bias.

Objectivity and Subjectivity of Clinical Data An important part of interpreting data is an understanding of whether the data are objective or subjective. There are a number of definitions of these terms,a but the essence of the distinction relevant to clinical research data collection is the degree to which recorded data (as opposed to the later interpretation of data) may be influenced by the individual thought of the observer. In the case of data reported by subjects themselves (such as symptoms of a disease), a researcher can objectively record the subjects’ statements but the statements themselves are necessarily subjective. This is in contrast to observations (such as signs of a disease) that can be made by outside observers. There is an important difference, for example, between an episode of bleeding reported by a subject and an episode witnessed by a researcher. While some subjective data, such as how a research subject feels after some intervention, can play very important roles in clinical trials, their subjective nature should be understood, and all possible efforts to capture corroborating objective data should be made as well. The observer of the status or event is only one aspect of data objectivity. Equally important is the process by which the observation is represented in the data that are ultimately recorded. A human observer is never completely free of cognitive and psychological influences that can affect the recording of data, as there is almost always some level of interpretation of what is perceived. For example, the observation that a subject “seemed weak” may be more subjective than if the subject reported weakness himself. One test of the subjectivity of such observations is to have multiple researchers report on the same observation. In the above example, if three people observe the same subject and independently agree that the subject seemed weak, then there is some increased reassurance that the data are objective, although there is a 25% random chance that they would agree, one way or the other. Requiring three researchers for each observation, on the other hand, triples the effort required to conduct the research

http://www.merriam-webster.com/dictionary/.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

DATA AS SURROGATES

or reduces by two-thirds the amount of research data that can be collected. We can increase the objectivity of observations through a variety of means. One method, of course, is to use an unbiased device to record information. For example, we can use a mechanical scale to weigh a subject, or we could measure muscle strength by asking the subject to lift an object of known weight. However, even supposedly objective devices, e.g., sphygmomanometers to measure blood pressure, can be misinterpreted.1

Transparency, Rigor, and Reproducibility There are natural concerns about the validity of study results that are reported by the researchers who design and conduct research studies and then provide the analysis of their own data.b Scientific rigor is the use of formal, well-tested research methods to ensure, to the extent possible, that results will be robust and unbiased. Transparency demands that these methods are clearly explained, including full descriptions of resources, materials, study subjects, reagents and the like, such that the results can be verified by others who wish to reproduce the analyses or event reexecute the experiment.2 Where human observation is required for data collection, researchers can employ rating systems and train the observers to use these scales so that can be used in a rigorous manner. For example, a five-point scale for muscle strength (ranging from fivednormal down to zerodno movement) can be used by multiple observers to produce consistent results.3 There are two other aspects to consider when assessing the rigor of an observation. First, the limits of the precision of data should be explicitly recognized, especially if the data are to be used for further calculation. In blood pressure measurements, for example, even reliable observers will have some variation in the measurements made between observers and even by the same observer. Second, rigor extends beyond the scope of the immediate research. Other researchers wishing to corroborate findings of a prior study will need to know how observations were collected in that study to produce comparable results in their own study. This extends not only to the methods used to record values but also to the measurements themselves, such as names of tests, diagnoses, symptoms, etc., so that when results are published, the writer and reader share a common vocabulary and understanding of the problem. A rigorous approach to data capture and analysis is a basic requirement for high-quality research. Scientific reporting today, however, must include not only the b

549

results but also a clear, transparent description of all that was done and all that occurred, including source data and adverse outcomes, for research to be considered truly reproducible.

Metadata Metadata simply means “data about the data.” The accurate interpretation of research data requires a rich set of information about the data, such as the subject, the method of collection, and the relationship of the data to the events in the research protocol, e.g., “initial weight on entry into the study” or “the white blood cell count measured 5 days after the third dose of chemotherapy.” Temporal metadata require particular attention. Data collection always occurs at some point in time. Even with continuous data monitoring, we must extract features for analysis that will have some kind of “time stamp” associated with them. Temporal information can be a point in time, a time period, or an openended time range; for example, a condition that began at a certain day and time and has continued through to the present. Time can be precise, e.g., “the pain started 1 h ago” or imprecise, e.g., “the patient started Drug X about a year ago but stopped a few months ago.” Even when time stamps are precise, care must be taken to understand the implications of a time. If a blood specimen is drawn to measure a drug level, not only must we know the time that the specimen was drawn (as opposed to the time the specimen was received in the laboratory or the time the result was reported), we also may need to know the time that the drug was administered. The recognition that data are observations collected at particular points in time, and that these points of observation will later represent the entire research study, makes explicit that it is up to the investigator to choose not only what to measure but also when and how frequently to measure it. Both of these characteristics of a measurement, the “what” and the “when,” must be chosen not only to best reflect the states and processes under observation but also to be amenable to statistical analysis. The statistical methods will depend on knowing the expected distribution under the null hypothesis. In many cases, this distribution is dependent on the accuracy and precision of the data being analyzed. It is essential, therefore, that the investigator consider the statistical methods that will be used to analyze the data at the same time that the choice is made of what and when to measure. Naturally, the choice of methods and any modifications to generally

http://grants.nih.gov/grants/guide/notice-files/NOT-OD-21-103.html.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

550

31. CLINICAL RESEARCH DATA

accepted techniques must be reported with the results in the interests of transparency and, ultimately, reproducibility.

TYPES OF DATA Data come in many different forms, some of which have already been mentioned and can be thought of in general terms as to whether they are quantitative or qualitative. Quantitative data refer to numeric measurements (including time) that can be manipulated mathematically. Measurements of blood pressure and hematocrit produce quantitative data. Precision is an important aspect of such data. For example, while a body temperature may be measured to a relatively precise level (e.g., “101.2” with four significant digits), a serum chloride measurement would be less precise (e.g., “101” with three significant digits), and an absolute eosinophil count even less so (e.g., “100” with only one significant digit). Qualitative data, on the other hand, represent conceptual entities rather than numeric values. Examples include subject gender and race, signs and symptoms, and diagnoses. They also may represent concepts that relate to quantitative data. For example, while “blood pressure” and “hematocrit” are associated with numeric results, the methods used in the procedures themselves (e.g., “blood pressure measured by arterial catheter” or “blood pressure measured by sphygmomanometer”) are a type of qualitative data. Somewhere between quantitative and qualitative lie the measurements that produce ordinal results. Ordinal data often look like numbers, but the discrete values are only relative to each other in terms of the order of their size or intensity but not through any specific distance between them. For example, urine protein measurements are often reported as “0,” “1þ,” “2þ,” and so on. These correspond roughly to mass concentrations but not with a precision that tells us that the difference between 1þ and 2þ and the difference between 2þ and 3þ have any specific relationship (equal, linear, geometric, or exponential, for example). This distinction is important because it impacts on the types of statistics that can be applied to analyze the data. For example, if we have a subject whose daily urine proteins are 1þ, 1þ, 1þ, and 2þ, we can say that he usually has a þ1 and that he seems to be getting worse, but we cannot say that his average urine protein is 1.25þ. Another type of data that has both quantitative and qualitative characteristics is signal data. Signal data are typically quantitative in nature but may be treated in a qualitative way. For example, electrocardiogram tracings are measurements of voltage over time from

various perspectives on the subject’s body, but rather than trying to analyze point-by-point changes in voltage, we abstract the data into qualitative concepts, such as P waves and QT intervals. Similarly, images, such as chest X-rays, are sets of points on film or in a computer image, each of which has a location and intensity, but we abstract these points into recognizable shapes, such as anatomic or pathologic features. To be sure, we still consider quantitative aspects of signal data, such as length of the QT interval or size of a tumor, but these measurements are secondary interpretations of qualities that have been extracted from the original numbers. It is worth noting that the reliability of these qualitative data is a function of the reliability of the original quantitative data and reproducibility of the interpretation method.

DATA STANDARDS Data standards provide ways of representing data to support their usefulness and exchange. Standards exist for representing both quantitative and qualitative data as well as for organizing the data into formats that can be communicated among researchers and information systems. Standards for quantitative data are usually related to the units of measure applied to the data. Some of these are from long-standing convention, such as the representation of blood pressure in terms of millimeters of mercuryda technically outdated concept. Others may be specified by, or even required by, standards setting organizations such as the International Organization for Standards (ISO) (Ref. 4; https://www.iso.org/aboutus.html), which provides standard units and symbols for quantities and units used within the different fields of science and technology.5 Standards for qualitative data usually take the form of controlled terminologies, which are finite, enumerated sets of well-defined terms that are intended to represent and convey information unambiguously. Standard terminologies typically cover specific domains, such as the laboratory test names, medical diagnoses, medications, and almost any other area of qualitative data. Using standard terminologies provides a reproducible way to collect clinical research data; those recording the data use a common language for their data, making data easier to aggregate and compare. However, care must be taken to make sure that the chosen terminology has sufficient breadth to cover all possible situations, and the terms in the terminology must be clearly understandable so that they can be chosen consistently. For example, a list of countries should be complete and up-to-date for recording origin and travel history (a person born in Dhaka could be recorded as being Indian, East Pakistani or Bangladeshi,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

DATA CAPTURE, STORAGE, AND RETRIEVAL

depending on birthdate), while a list of races may need instructions for proper use (e.g., is a person of European descent who was born in Japan considered Caucasian or Asian?). Terminologies may be created by individual researchers seeking to label their own data to aggregate similar observations; however, using standard terminologies to record data facilitates the sharing of data and their aggregation from multiple studies and is essentially mandatory for achieving the goal of reproducibility of clinical trials. Standard terminologies are produced through a variety of means, including ad hoc (created by a cooperative group of interested parties and accepted by mutual agreement), de facto (created by one party that is dominant enough to overcome competition from other standards), government mandate (required for use by a government agency), and consensus (created through an open method that considers input from all interested parties).6 Although no specific terminology standards are generally required for clinical research, some terminologies are required for federal reporting purposes, such as the Medical Dictionary for Regulatory Activities (MedRA) for reporting to the US Food and Drug Administration (FDA) all adverse events occurring in drug trials,7 and the Common Terminology Criteria for Adverse Events for reporting adverse events to the Cancer Therapy Evaluation Program at the National Cancer Institute.8 One terminology that has frequently been used in studies involving assessment of morbidity and mortality is the International Classification of Diseases, Ninth Edition, with Clinical Modifications (ICD-9-CM).9 This terminology is maintained by the US Department of Health and Human Services for the purpose of coding patient information, and for the dual purpose of billing for health care and reporting causes of death to the World Health Organization. In 2012, ICD-9-CM was replaced with a larger, updated version, ICD-10-CM. Because virtually every patient record in the United States has data coded with ICD-9-CM or ICD-10-CM, researchers who wish to collect large data sets will be tempted to use such data to stand in for actual clinical statements about research subjects. However, researchers repeatedly find that ICD codes assigned to patient records have an undesired level of reliability, when compared to the actual content of the records.10 Limitations on the use of ICD-coded data are complicated by the fact that meanings of codes changed over time11 and that the mapping of data from ICD-9-CM to ICD-10-CM is not straightforward.12 Table 31.1 shows examples of similar ICD-9-CM and ICD-10-CM codes. Although researchers continue to make use of ICD data, they rarely choose it as a way to capture their own data, for example, to record patient diagnoses.

551

This is generally because it lacks the level of detail needed for representing clinical concepts. For example, ICD-9-CM has the term “Primary Neoplasm of the Lung” but does not provide differentiation into clinically important subtypes, such as small cell carcinoma and squamous cell carcinoma. When researchers do turn to existing standard terminologies, they find a wide selection, such as the Logical Object Identifiers Names and Codes (LOINC), which is best known for its coverage of laboratory test terms,13 RxNorm and its clinical subset RxTerms, which cover medication terms,14 the Gene Ontology (GO),15 which provides terms for the cellular structures, molecular functions, and biologic processes related to gene products, and the Systematized Nomenclature of Medicine-Clinical Terms (SNOMED-CT), which covers a wide variety of other terminology domains.16 Several institutes at the National Institutes of Health (NIH) have established Common Data Elements, which are specific data elements that are typically collectedd sometimes as a requirement of the funding agencyein research studies. Many elements include specific definitions on their meaning, their allowed values, and even the (reproducible) methods by which they should be collected.17 Standards also exist for organizing clinical research data to enable sharing and aggregation. The Clinical Data Interchange Standards Consortium (CDISC),18 a group with predominantly pharmaceutical company representation, is creating a standard for submitting regulatory information to the FDA. Health Level Seven (HL7),19 an open standards development organization that establishes consensus standards for all manner of clinical and administrative data, has created a standard Clinical Document Architecture (CDA) and is now working to extend it to model clinical research data. All of these groups have come together to form the Biomedical Research Integrated Domain Group (BRIDG) to coordinate and perhaps even unify their efforts.20,21

DATA CAPTURE, STORAGE, AND RETRIEVAL When the decision is made on what data to collect, and how to collect them, the next decision must be about where to put them. A great deal of clinical research data are collected on encounter forms or case report forms, which are simply templates (either paper- or computer-based) into which data may be entered. These forms help assure that the data collection is complete and recorded (where appropriate) with the correct controlled terminology. Sometimes the data move on further than these forms until an analyst copies them into a spreadsheet or statistical software program to perform analysis.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

552

31. CLINICAL RESEARCH DATA

TABLE 31.1

Changes in International Classification of Diseases Codes Over Time

Term Name

ICD-9-CM (2007)

ICD-9-CM (2008)

ICD-10-CM (2021)

Shock following incomplete spontaneous abortion

785.9

785.9

O03.31

Shock following complete or unspecified spontaneous abortion

785.9

785.9

O03.81

Shock following (induced) termination of pregnancy

785.9

785.9

O04.81

Shock following failed attempted termination of pregnancy

785.9

785.9

O07.31

Shock following ectopic and molar pregnancy

785.9

785.9

O08.3

Shock during or following labor and delivery

785.9

785.9

O75.1

State of emotional shock and stress, unspecified

785.9

785.9

R45.7

Cardiogenic shock

785.1

785.1

R57.0

Hypovolemic shock

785.9

785.2

R57.1

Other shock

785.9

785.9

R57.8

Shock, unspecified

785.9

785.9

R57.9

Severe sepsis without septic shock

785.9

785.9

R65.20

Severe sepsis with septic shock

785.9

785.9

R65.21

Shock due to being struck by lightning, initial encounter

785.9

785.9

T75.01XA

Shock due to being struck by lightning, subsequent encounter

785.9

785.9

T75.01XD

Shock due to being struck by lightning, sequela

785.9

785.9

T75.01XS

Anaphylactic shock, unspecified, initial encounter

785.9

785.9

T78.2XXA

Anaphylactic shock, unspecified, subsequent encounter

785.9

785.9

T78.2XXD

Anaphylactic shock, unspecified, sequela

785.9

785.9

T78.2XXS

Examples shown are all codes for shock, excluding those caused by trauma or medical procedures. Note that most terms would be coded in ICD-9-CM as simply “785.59 Other Shock,” since more specific terms did not exist. Note also that the 2008 version of ICD-9-CM included a code “785.2 Septic Shock,” thus changing the meaning of code 785.9. In addition, the meaning of “Other Shock” in ICD-10-CM is completely different because it does not include the other 16 terms shown, nor others not shown.

Clinical Trials Data Management Systems More sophisticated approaches to data capture and storage are provided by clinical trials data management systems (CTDMs). These systems assist researchers in designing their trial, including defining the data to be captured and provide sophisticated user interfaces with which researchers and even the subjects themselves can enter the research data directly. Some systems provide features such as data validation (which uses rules to make sure the data make sense) and data verification (having the person entering the data double-check them by reading or even reentering them to minimize data entry errors) to help improve the data quality. CTDMs also can interface with other systems such as electronic health records systems (EHRs) to import additional data, e.g., laboratory test results that might already be captured in the subjects’ medical records. A wide variety of CTDMs are available, ranging from no-cost “open source” systems to commercial systems. c

One that has become especially popular is the Research Electronic Data Capture (REDCap) system, which is available free of charge and currently has over 1875 installations in over 100 countries.c REDCap users can install the software at their own institutions and then set up individual databases and data capture forms for each study.22 Fig. 31.1 shows an example of a REDCap screen.

Clinical Data Repositories Beyond the CTDMs lie larger, more comprehensive data repositories, sometimes referred to as data warehouses or data marts. These systems collect data from disparate smaller systems to support clinical research in several ways.23 First, they can provide a means for archiving data for prolonged storage, independent from smaller systems that may have more limited resources and longevity. Second, they may be better suited for data

http://projectredcap.org/.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

RESPONSIBLE STEWARDSHIP OF DATA

553

FIGURE 31.1 Sample screen from Research Electronic Data Capture showing a simple case report form that captures patient laboratory test results and weight.

analysis functions than the CTDMs, whose primary function is data collection. Third, they provide a platform for integrating data from multiple sources to support the needs of existing studies that have data collected in a distributed manner. Fourth, they can make the aggregated data available for reuse in new studies. This is done, for example, by pooling data from multiple studies to create larger cohorts or to identify potential subjects who might be eligible for a new study. The NIH has created such a repository to permit the aggregation of clinical research data from across its 27 institutes and centers to support translational research (see Fig. 31.2).24 Informatics for Integrating Biology and the Bedside (i2b2) is a project being developed under a grant from the National Center for Advancing Translational Services (NCATS) at NIH.d Based at Partners HealthCare System and Harvard University, i2b2 is developing an information system framework to allow clinical researchers to reuse existing clinical data for discovery research. The i2b2 platform includes a workflow framework and a data repository as well as tools for terminology management and natural language processing.25 Fig. 31.3 shows a sample screen from i2b2.

d

RESPONSIBLE STEWARDSHIP OF DATA Data will stand in for the actual observations made during a clinical research project. The circumstances of its recording must be clear, including who made the observation, whether the record of the observation was amended or deleted, and when and for what reason such amendments and deletions were made. These requirements are very much like those that are legally required for medical records, for comparable reasons. Legal requirements on data arise because they may be used as evidence in an argued court case. Likewise, every research publication is an “argument” for the particular interpretation advanced by its author(s), and the “rules of evidence” are similar. In the case of paper records, striking through rather than erasing the original data when making changes, and signing and dating each change, meet such requirements for transparency. In electronic environments, the need to reconstruct the history of any datum means that systems must be designed not simply to store data elements but also to record all changes or deletions and keep the original observations available. This is usually achieved through a combination of security rules that limit access and

https://www.i2b2.org/.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

554

31. CLINICAL RESEARCH DATA

FIGURE 31.2 Sample screen from the Biomedical Translational Research Information System, National Institutes of Health’s clinical data repository. In this instance, the use has requested all diagnoses listed for all patients with a vitamin B12 level greater than 999. The system found 231,704 diagnoses on 5834 subjects.

activity to authorized individuals and that treat a username and password, or other identification mechanism such as a fingerprint, as equivalent to a signature, combined with a log of what records were accessed and what changes were made, by whom, and when. Implicitly, such a system only works if there are administrative controls that forbid the sharing of usernames and passwords and define penalties for unauthorized use. In the current environment, in which most databases are on computers that are directly or indirectly connected to the Internet, there needs to be additional safeguards to ensure that access to and alteration of data are prevented from outside the institution. These safeguards typically take the form of hardware or software that can block or detect unauthorized access. Just as important paper records often are stored in fireproof vaults and have physical copies maintained, so electronic databases also must be “backed up” regularly, preferably with at least one set of backups stored at a secure off-site location. In addition to considerations of physical storage and maintenance, a researcher should collect and store data so that they are as reusable as possible. Data are lost far too frequently when an investigator leaves an

e

institution, taking his or her computer password or knowledge of how he or she collected data with him or her and leaving data that are inaccessible or uninterpretable. Loss of data represents loss of an investment of institutional time, and it also means that to answer questions that potentially could have been answered with existing data, new subjects will need to be recruited and exposed to risk.e As a corollary to the need to preserve data, they also should be collected in as standard a form as possible, and metadata (including definitions of what each data element means, how it was defined and collected) should be stored alongside the data. Storing data for reuse is generally a good idea, but it can be difficult to enforce in an academic environment that creates incentives for individual investigators to control and limit access to research data. Such a structural environment limits the long-term value of any particular piece of information and blocks the research community from taking advantage of the ability of computers to store and combine very large amounts of data. It will be one of the major challenges of the coming years to reconstruct the academic research environment to recognize and reward researchers for contributing

For an amusing but all-to-real example of the challenges of data sharing, see https://www.youtube.com/watch?v¼N2zK3sAtr-4.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

COOPERATIVE SHARING EFFORTS

555

FIGURE 31.3 Sample screen from the i2b2 (Informatics for Integrating the Biology and the Bedside) user interface. In this example, the user has asked for a list of all patients who have received insulin and who have the diagnosis diabetes mellitus.

data to repositories outside their personal or programmatic control. The habitual accumulation of data has the potential to transform clinical research from an activity focused on the individual study to an enterprise concerned with the evolving interpretation of a growing body of knowledge. There also are other types of data that need to be collected by the clinical researcher. Most of these are used either to validate or establish the context of the experimental observations. Some examples of these types of data include the research protocol document, the subject consents, the statistical analysis plan, the credentials of the investigators, the files documenting who entered/modified electronic data and when, and the privacy policy of the research institution. The guidelines and regulations for data management apply to these types of data as well as to data documenting traditional research endpoints.

COOPERATIVE SHARING EFFORTS The technical challenges of data modeling, representation, storage, and use are being addressed every day by f

researchers around the world who are often “reinventing the wheel.” Opportunities exist, however, for researchers wishing to make use of ideas, technologies, and even data that are being shared through a number of cooperative efforts. For example, the Advanced Clinical Trials (ACT) projectf is creating a federated network of National Clinical and Translational Science Award Consortium sites to support the accrual of research subjects into clinical trials. Many of the 60þ institutions receiving Clinical and Translational Science Awards (CTSA), also funded by the National Center for Advancing Translational Sciences (NCATS), are adopting i2b2 technologies to support research and collaboration by joining their i2b2 instances to Shared Health Research Information Networks (SHRINE), which allow a query to a single i2b2 instance to be repeated at each instance on the network, thereby providing information about the numbers of patients at each site who are potential subjects for a particular research study.26 For those seeking to share data, and to avail themselves of data shared by others, the National Center for Biotechnology Information (NCBI) at the NIH’s National Library of Medicine (NLM) is creating a public repository of individual-level data, including exposure

https://www.act-network.org.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

556

31. CLINICAL RESEARCH DATA

history, signs, symptoms, diagnostic test results, and genetic data. Called the Database of Genome and Phenome (dbGAP), the project provides stable data sets that allow multiple researchers to reference the same samples in their publications of secondary analyses of the data.27 Additional data from clinical trials, currently limited to summary results, also are being made available by the NLM through the ClinicalTrials.gov resource, which is a repository of data from clinical trials.28

SUMMARY Every investigator should consider data management before beginning a clinical research project, whether it is an interventional clinical trial or a natural history study. Establishing standard operating procedures for data management will simplify future audits and ensure that adequate data are collected for planned statistical analyses. During the conduct of a clinical protocol, the researcher should put herself or himself in the role of someone trying to discredit the study’s conclusions and question every procedure that could cast doubt on the accuracy, validity, or relevance of the data collected. Data should be collected using standard methods and instruments, and the definitions and explanations that describe the data and provide context should be stored along with them. There is a growing recognition that reporting rigorous methods in a transparent manner is as important as reporting the results themselves. Researchers must maintain a constant awareness that the data they accrue are proxies for the actual events and objects in the real world and be on guard for biases that the collection process may introduce.

SUMMARY QUESTIONS 1. A researcher interviews a patient who says he is having pain. The researcher asks “How many hours have you had the pain?” A later review of the data collection form shows that the number “14” was recorded. Which of the following statements is not true about this datum? a. If the subject did not appear, to the researcher, to be in pain, then he probably was not really having pain b. The “14” might be inaccurate because the subject may not be able to tell time properly c. The “14” might be inaccurate because the subject actually said “40” but the researcher did not hear properly d. If two researchers observed the subject and both heard him say 14, then the reliability of the datum is increased e. All of the above are true

2. In the above example, which of the following is true about the subject’s answer to the pain duration question? a. It is subjective and quantitative b. It is objective and quantitative c. It is subjective and qualitative d. It is subjective and qualitative e. None of the above is true 3. Which of the following is not metadata about the subject’s answer to the pain duration question? a. The subject’s identifier (e.g., name or medical record number) b. The name of the researcher who recorded the answer c. The fact that the answer was “14” d. The fact that the units of measure was “hours” e. All of the above are metadata 4. Recording a serum potassium laboratory result with a controlled terminology would not be appropriate for which of the following? a. The name of the test (e.g., “serum potassium”) b. The unit of measure that is used to report the result (e.g., “mEq/L”) c. A numeric test result (e.g., “4.0”) d. A comment result (e.g., “specimen hemolyzed”) e. All of the above could be recorded with controlled terminology 5. Which of the following information about an observation does not need to be recorded when performing clinical research? a. The accuracy of the measurement procedure (e.g., the number of significant digits) b. The relationship if the observation to the research plan (e.g., whether the observation was made before or after a treatment being is studied) c. The consent form that the patient signed to participate in the study d. The details of the research protocol e. All of the above need to be recorded 6. Which of the following would be an example of something that is not required for conducting reproducible research? a. Using a standard pain scale for eliciting symptoms from a research subject b. Describing the search terms used for retrieving citations from PubMed for a systematic literature review c. Adhering closely to the Vancouver citation style when submitting a manuscript for publication d. Describing the statistical method used for analyzing study results e. Submitting detailed outcomes data to ClinicalTrials.gov

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

References 1. Gosse P, Coulon P. Ambulatory or home measurement of blood pressure? J Clin Hypertens 2009;11(4):234e7. 2. Ioannidis JP. Why most published research findings are false. PLoS Med 2005;2(8):e124. 3. Florence JM, Pandya S, King WM, Robison JD, Baty J, Miller JP, et al. Intrarater reliability of manual muscle test (Medical Research Council scale) grades in Duchenne’s muscular dystrophy. Phys Ther 1992;72(2):121e2. 4. Schadow G, McDonald CJ, Suico JG, Fo¨hring U, Tolxdorff T. Units of measure in clinical information systems. J Am Med Inform Assoc 1999;6(2):211e62. 5. International Organization for Standardization. ISO 2955. Information processing: representation of SI and other units in systems with limited character sets. 1983. Geneva. 6. Hammond WE, Jaffe C, Cimino JJ, Huff SM. Standards in biomedical informatics. In: Shortliffe EH, Cimino JJ, editors. Biomedical informatics: computer applications in health care and biomedicine. New York: Springer; 2014 [Chapter 7]. 7. Brown EG, Wood L, Wood S. The medical dictionary for regulatory Activities (MedDRA). Drug Saf 1999;20(2):109e17. 8. Trotti A, Colevas AD, Setser A, Rusch V, Jaques D, Budach V, et al. CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 2003; 13(3):176e81. 9. United States National Center for Health Statistics. International classification of diseases, ninth revision, with clinical modifications. 1980. Washington. 10. Iezzoni LI. Using administrative diagnostic data to assess the quality of hospital care. Pitfalls and potential of ICD-9-CM. Int J Technol Assess Health Care 1990;6(2):272e81. 11. Yu AC, Cimino JJ. A comparison of two methods for retrieving ICD-9CM data: the effect of using an ontology-based method for handling terminology changes. J Biomed Inform April 2011;44(2):289e98. 12. Cimino JJ, Remennick L. Adapting a clinical data repository to ICD-10-CM through the use of a terminology repository. AMIA Annu Symp Proc November 2014;14:405e13. 13. Forrey AW, McDonald CJ, DeMoor G, Huff SM, Leavelle D, Leland D, et al. Logical observation identifier names and codes (LOINC) database: a public use set of codes and names for electronic reporting of clinical laboratory test results. Clin Chem 1996;42(1):81e90. 14. Fung KW, McDonald C, Bray BE. RxTermsda drug interface terminology derived from RxNorm. AMIA Annu Symp Proc 2008;6:227e31. 15. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 2000;25(1):25e9.

557

16. International Health Terminology Standards Development Organization. The systematized nomenclature for medicine-clinical terms. 2010. Copenhagen. 17. Rubinstein YR, McInnes P. NIH/NCATS/GRDRÒ Common Data Elements: a leading force for standardized data collection. Contemp Clin Trials May 2012;42:78e80. 18. Kuchinke W, Wiegelmann S, Verplancke P, Ohmann C. Extended cooperation in clinical studies through exchange of CDISC metadata between different study software solutions. Methods Inf Med 2006;45(4):441e6. 19. Klein A, Prokosch HU, Mu¨ller M, Ganslandt T. Experiences with an interoperable data acquisition platform for multi-centric research networks based on HL7 CDA. Methods Inf Med 2007; 46(5):580e5. 20. Fridsma DB, Evans J, Hastak S, Mead CN. The BRIDG project: a technical report. J Am Med Inform Assoc 2008;21(2):130e7. 21. Laleci GB, Yuksel M, Dogac A. Providing semantic interoperability between clinical care and clinical research domains. IEEE J Biomed Health Inform March 2013;17(2):356e69. 22. Harris PA, Thielke R, Taylor R, Payne J, Gonzalez N, Conde JG. Research Electronic Data Capture (REDCap) e a metadatadriven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42(2): 377e81. 23. Prokosch HU, Ganslandt T. Perspectives for medical informatics. Reusing the electronic medical record for clinical research. Methods Inf Med 2009;48(1):38e44. 24. Cimino JJ, Ayres EJ, Remennik L, Rath S, Freedman R, Beri A, Chen Y, Huser V. The National Institutes of Health’s Biomedical Translational Research Information System (BTRIS): design, contents, functionality and experience to date. J Biomed Inform December 2014;52:11e27. 25. Weber GM, Murphy SN, McMurry AJ, Macfadden D, Nigrin DJ, Churchill S, et al. The Shared Health Research Information Network (SHRINE): a prototype federated query tool for clinical data repositories. J Am Med Inform Assoc 2009;16(5):624e30. 26. Reis SE, Berglund L, Bernard GR, Califf RM, Fitzgerald GA, Johnson PC. National Clinical and Translational Science Awards Consortium. Reengineering the national clinical and translational research enterprise: the strategic plan of the National Clinical and Translational Science Awards Consortium. Acad Med 2010; 85(3):463e9. 27. Mailman MD, Feolo M, Jin Y, Kimura M, Tryka K, Bagoutdinov R, et al. The NCBI dbGaP database of genotypes and phenotypes. Nat Genet 2007;39(10):1181e6. 28. Tse T, Williams RJ, Zarin DA. Update on registration of clinical trials in ClinicalTrials.gov. Chest 2009;136(1):304e5.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

32 Management of Patient Samples 1

Karen E. Berliner1, Amy P.N. Skubitz2

National Institutes of Health, Bethesda, MD, United States; 2University of Minnesota, Minneapolis, MN, United States

O U T L I N E Introduction

559

Successful Research Rests on a Foundation of Careful Planning

560

Specimen Handling

565

Specimen Transit

565

Specimen Storage

566

The Role of Pre-analytic Variables in Research Using Patient Specimens

560

Access to Patient Samples

567

Training and Accreditation

561

The Importance of Good Record Keeping

562

Specimen Culling, Transfer of Collections, and Repository Closings

567

Specimen Tracking

562

Summary Questions

567

Specimen Collection

563

References

568

INTRODUCTION For many years, most of the attention in performing clinical research has been placed on three general areas: (1) identification of biological markers or “biomarkers” that may indicate the onset of disease, (2) identification of pharmaceuticals that may reverse disease states, and (3) development of approaches to prevent the disease state from evolving in the first place. While these aims continue to serve as the guiding forces for conducting clinical research, new attention has been placed on the importance of the biological specimen or “biospecimen” as the source material for scientific investigation. Until recently, the handling of biospecimens has relied on assumptions such as “colder is better,” an idea which implies that biological specimens are best preserved when handled and stored at colder temperatures, or that fixation times are not critical as long as the specimens are exposed to the appropriate chemicals and reagents. While

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00032-0

these generalizations may hold true for many material types, they do not hold true for all. “Biospecimen Science” is the relatively new field of scientific investigation that aims to examine these assumptions and clarify the optimal conditions for specimen collection, processing, storage, and dissemination.1,2 In recognition of the importance of improving the state of the science as to how biospecimens are collected and handled, several governmental and nonprofit organizations emerged to exchange information on effective practices and research findings. In 1999, a small group of professionals working with biorepositories (i.e., centers for biospecimen processing, storage, and dissemination) gathered to share their experience and offer each other assistance to improve and enhance consistency in biospecimen handling. Since its inception, the International Society for Biological and Environmental Repositories (ISBER) has brought together professionals associated with biorepositories from all over the world and has

559

Copyright © 2018. Published by Elsevier Inc.

560

32. MANAGEMENT OF PATIENT SAMPLES

produced three editions of its “Best Practices for Repositories”3e5 (available in PDF form at www.isber. org), with a fourth edition under development as this book goes to press. The US National Cancer Institute (NCI) of the National Institutes of Health (NIH) has also long recognized that the lack of standardized, high-quality biospecimens has resulted in one of the most significant roadblocks to the progress of cancer research. In 2002, the NCI initiated a due diligence process to understand the state of its funded biospecimen resources and the quality of biospecimens used in cancer research. To achieve this aim, the NCI conducted surveys and created community forums that led to the establishment of the Office of Biorepositories and Biospecimen Research (OBBR) to address issues pertinent to the development of practices that promote the isolation and utilization of high-quality specimens in cancer research. In 2006, OBBR (which later evolved into the Biorepositories and Biospecimen Research Branch, BBRB) developed its “First-Generation Guidelines for NCI-Supported Research” that was subsequently revised based on public comment and was retitled, “NCI Best Practices for Biospecimen Resources” or “NCI Best Practices.” The most recent version of the NCI Best Practices was released in 2016 and can be found at http://biospecimens.cancer.gov/practices. Another important OBBR initiative includes the creation of the Biospecimen Research Network that offers investigators a quick glance at important scientific publications in support of this new field (https://brd.nci.nih.gov/brd/). This chapter will focus on practices relevant to the management of patient samples in clinical research. The reader is encouraged to review the documents included as references to learn more about the effective handling of biological specimens.

SUCCESSFUL RESEARCH RESTS ON A FOUNDATION OF CAREFUL PLANNING Important scientific findings result from the use of a well-planned research strategy and require the development of protocols that support the isolation and preservation of patient samples so that they closely resemble their status prior to their removal from the patient. The optimal situation is one in which the analytes to be measured have been well characterized, and specimen collection and processing protocols are welldocumented in the scientific literature. However, since new methods and technologies are continuously being developed, it may not be possible at the time of biospecimen collection to determine which downstream assays may be available in the future. When specific target analytes are known, it is important that the specimens used are “fit for purpose” or, in other words, have

been demonstrated to support the integrity of the particular analyte as the biospecimen makes its way through the collection and processing pathway. When specific targets are not known at the outset of a study, general practices must be implemented in a timely manner according to information that is available in the literature.

THE ROLE OF PRE-ANALYTIC VARIABLES IN RESEARCH USING PATIENT SPECIMENS In recent years, the scientific community has come to the realization that all patient samples are not created equal. Evidence-based protocols are defined as protocols that have been scientifically optimized for a specific measurement through the use of quality control measures with standards that allow for the reproducibility of the assay under defined conditions. In addition, it is also important for the investigator to understand that the molecular profile of a tissue or other biological material can be influenced by a variety of factors. These factors may include exposures of the patient to certain medications, food or beverages consumed prior to the collection of the biospecimen,6,7 or certain environmental exposures.8 The time of day at which a specimen is collected also may influence the molecular profiles of some analytes under certain circumstances.9 Additional influences that should be considered are the time that elapses between when the tissue is separated from the patient’s blood supply and when the tissue is removed (the ischemic time); the time that elapses between when specimens are removed from the patient and when they are processed10; and the temperature at which the specimen is kept prior to processing. The combined effects of these types of exposures or events prior to the time at which a biospecimen is used in an assay have been described as “pre-analytic” variables (Fig. 32.1). These factors are often some of the most difficult to capture and yet can have significant effects on research outcomes.11 The reader is directed to reviews of the burgeoning field of literature on pre-analytic variables that have been summarized by BBRB.12 As much information as is possible about the condition of the patient from whom the specimen has been taken (e.g., disease status, drug profile, timing of the collection, and food exposures) should be recorded and tracked throughout the life cycle of the specimen. This information may become critical when evaluating the future suitability of the biospecimen for a particular investigation. Managing patient samples to ensure their suitability for specific tests can best be achieved when these preanalytic variables can be traced. A system for documenting pre-analytic variables has been developed in which a Sample PREanalytic Code is assigned to each

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

TRAINING AND ACCREDITATION

561

FIGURE 32.1 Variables affecting the molecular integrity of biospecimens. Both the pre- and postacquisitions variables listed here are covered under the category of “pre-analytic variables.”

biospecimen.13,14 The code is equivalent to a “specimen barcode” that provides details about pre-analytic processing for that specimen. The code is comprised of seven elements that contain information such as the material type, processing steps, and storage conditions that may be important in understanding a sample’s “fit for purpose” for a particular assay (see examples in Fig. 32.2).

TRAINING AND ACCREDITATION Not only must the research plan be based on carefully described protocols that consider pre-analytic variables, but also those involved in handling and processing the specimens must be trained on the myriad of activities that support the effective isolation, transport, processing,

FIGURE 32.2 Sample PREanalytic Code (SPREC) nomenclature for documenting pre-analytical variables. Two examples are provided for how biospecimens would be labeled using the SPREC.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

562

32. MANAGEMENT OF PATIENT SAMPLES

storage, and dissemination of specimens for testing. To this end, the Canadian Tissue Repository Network (CTRNet) has developed training modules that were based on ISBER’s “Best Practices for Repositories.”3e5 CTRNet’s “Introduction to Biobanking Course” was designed to give repository personnel a general overview of the key issues in establishing, maintaining, and accessing a biospecimen repository or “biobank.” The education program includes nine online modules designed to provide “how-to” knowledge for researchers and biobankers. Modules cover the topics of biobank basics, governance, ethics, privacy and security, facility design and safety, quality management and process improvement, informed consent, biospecimen collection and processing, storage and distribution, as well as data systems and records management. ISBER has developed training modules for technicians on the topics of receiving specimens into a repository, specimen tracking, dissemination, shipping, selecting the right storage environment for biospecimens, and process improvement in repositories. As this chapter goes to press, ISBER is exploring a partnership with the American Society for Clinical Pathology to develop a qualification program for repository technicians and technologists. Today training courses are available worldwide to educate repository personnel, many of which have been endorsed by ISBER. The College of American Pathology (CAP) now offers an accreditation program for biorepositories that is designed to improve the quality and consistency of facilities that collect, process, store, and disseminate biospecimens for research. To qualify for accreditation, biorepositories are provided with a checklist of requirements that range from sample accrual and processing to database security. Repositories undergo a site visit by individuals with expertise in biorepositories, and the items on the checklist are verified. As of July 2016, 58 biorepositories in the United States have been accredited by CAP.

which results can be reproduced by others, to establish the scientific validity of findings. Careful record keeping also may contribute to the establishment of intellectual property claims or assist with the defense against allegations of research misconduct. Records should be kept in a manner that is legible and would allow others to understand and reproduce the steps taken to obtain the same findings. Records should capture: • the name of the person making the record • the purpose of the research • the names of all individuals involved in the collection and processing efforts • lists of Standard Operating Procedures (SOPs) • information about the participant (name, age, appropriate health statistics) • the time and date on which the specimens were collected and all subsequent processing steps Records can be kept in bound or unbound notebooks or in electronic formats such as those found associated with ordinary software or specialized software packages. Notebooks should be used for the specific purpose of recording laboratory and other aspects integral to specimen collection, processing, storage, and dissemination. All records should be kept in a secure manner to ensure that access to the records is limited to authorized users and that personal health information is kept confidential. The US Department of Health and Human Services (DHHS) regulations (www.hhs.gov/ ohrp/regulations-and-policy/guidance/guidance-onwritten-irb-procedures) require that consent and other Institutional Review Board (IRB) records be retained for at least 3 years, and records relating to research be retained for at least 3 years after compilation of the research. Records may be kept indefinitely as records may become invaluable resources for future research endeavors. All records must be accessible for inspection and copying by authorized representatives of DHHS at reasonable times and in a reasonable manner.

THE IMPORTANCE OF GOOD RECORD KEEPING SPECIMEN TRACKING The NIH has stated in its publication, “Guidelines for Scientific Record Keeping in the Intramural Research Program at the NIH” (https://oir.nih.gov/sourcebook/ ethical-conduct/research-ethics/nih-guidelines/guidehuman-biospecimen-storage-tracking), that “the progress and excellence of NIH research are dependent on our vigilance in maintaining the highest quality of records for every aspect of the science conducted here.” Good record keeping is a requirement of all scientific endeavors and contributes to the successful analyses of research results, as well as for the review of those results once submitted for publication. It provides the basis for

The ability to effectively manage patient samples requires that the original specimen collected, as well as all derivatives of the original specimen generated through various processing protocols, be tracked. Each specimen should be given a unique identifier (ID) that can be followed throughout its life cycle. These numbers allow for traceability to an ID linked to a particular patient, unless the linkage between the patient ID and specimen ID has been irreversibly interrupted, such that the identity of the patient has been anonymized. Derivatives of the original specimen (e.g., DNA, RNA, or

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

563

SPECIMEN COLLECTION

FIGURE 32.3 The images on the top and bottom represent a onedimensional (1D) barcode and a two-dimensional (2D) barcode, respectively. The 1D barcode is comprised of a single row of bars, and the data encoded are presented across the horizontal width of the label. Data encoded in the 2D label are contained in both the vertical and horizontal directions and are able to contain a greater amount of information.

protein extractions) should always be traceable to the original specimen. All specimens should have a printed label that contains the biospecimen ID that is encoded in either a one-dimensional (1D) or two-dimensional (2D) scannable barcode (Fig. 32.3). 1D barcodes have an advantage in that they contain a high degree of redundancy, which means that the barcode can be read even with considerable degradation. However, 1D barcodes are limited in the amount of information they contain, since all of the data are encoded in the horizontal width. 2D barcodes have the advantage that they contain information that is read in a vertical direction as well as a horizontal direction and so the amount of information in the code is greater. There are advantages to including human readable information on the label (i.e., the specimen ID) to facilitate manual processing, but this may not always be necessary, especially if barcode readers are available. Many specimen containers are now manufactured with 2D barcodes already printed on the bottom of the container, which can facilitate automated specimen handling without the need or expense of adding a physical label. Effective use of these containers without additional printed labels requires that a scanner is always available should the identity of a specimen need to be verified. Tubes with 2D barcodes on the bottom have the advantage that if they are stored in a box with holes on the area underneath the 2D barcode, specimens can be scanned rapidly using plate readers that read all specimens in the box within a matter of seconds. Selection of

the best labels and collection and storage tubes for a particular study and purpose should be performed well in advance of the commencement of specimen collection to make sure that the selections are well suited for downstream needs. Integral to the effective management of patient specimens is having an inventory system that tracks the specimens as they move from collection through processing, storage, and dissemination. Computer-based systems allow for the most accurate, dependable tracking of samples. Although this may be accomplished by using simple spreadsheet programs, the best way to track specimens and characteristics of those specimens is to use an inventory system that is created specifically for this purpose. Many inventory system software programs are commercially available and can be used easily without significant customization. These programs have the ability to track personal health information, subject IDs, specimen IDs, specimen processing events, and specimen location. In addition, many programs allow for the uploading and downloading of batch records, which facilitates specimen accession, distribution, and the sharing of data under carefully defined conditions. All inventory management systems or other computer-based systems that contain personal health information must be operated under strict security practices. Regardless of whether records are electronic or paper, access to patient information and information about patient samples must be tightly controlled. Access to view and modify information should be allowed according to defined privileges. Changes or modifications to records should be traceable to the time, date, and the identity of the individual making the changes. Records should be backed up daily to a format and location that allows for their secure maintenance according to a defined management plan that includes consideration of a wide variety of emergency situations.

SPECIMEN COLLECTION The effective collection of patient samples begins with the receipt of a signed consent form from the patient (or appropriate guardian) that documents permission to use the samples under defined conditions (see Chapter 2). Consent details must be tracked throughout the life cycle of the sample and its derivatives to ensure that any testing performed falls within the parameters to which the patient has agreed. Study design and the use of patient samples are governed by IRBs (Chapter 4). Investigators are responsible for ensuring that appropriate approvals are in place for the use of biospecimens and biospecimen-related data and should allow sufficient

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

564

32. MANAGEMENT OF PATIENT SAMPLES

time for appropriate review when planning for the implementation of research activities. The actual collection of the patient samples should be well orchestrated as it may involve a number of individuals and circumstances around which careful planning must be executed to ensure that the highest-quality specimens are available for downstream processing. In addition to the biorepository technicians, some of the personnel involved in this process on a daily basis may include the study coordinator, admissions nurse, anesthesiologist, operating room nurse, surgeon, pathologist, and/or pathologist assistant. All individuals involved in the collection process should have the appropriate expertise and be well trained in the tasks that they are expected to perform as stated above. Where possible and appropriate, the success of research efforts can be enhanced by having a joint meeting(s) among those involved in the collection of patient samples to ensure a smooth transition from one phase of the collection effort to another. When multiple collections sites are involved, it is even more critical that meetings are held in advance of specimen collection to ensure that protocols are implemented in a uniform manner. To ensure that biospecimens for clinical trials are collected and processed correctly, it is important that detailed SOPs are provided to the personnel involved in the collection effort and that they have a clear understanding of what needs to be done. As a first step, a research nurse or study coordinator should contact the biorepository staff in advance to apprise them of any necessary information pertaining to the patient who has been enrolled in the clinical trial and the specific date and time that the surgery is scheduled. The SOP should outline which consent form should be used as well as the type, number, and processing steps for each biospecimen to be collected. For a typical clinical trial study, a biorepository technician should review the operating room schedule a week in advance, then again 1 day in advance, and finally on the day of surgery. The technician should notify the operating room nurses of the cases that are expected to be collected that day and which tissues are to be obtained so that specimens will be handled properly. This technician is responsible for ensuring that each patient who is scheduled for surgery and will be contributing specimens signs the appropriate consent forms. Separate IRB-approved patient consent forms may need to be designed for the acquisition of waste tissue following surgery, as well as for saliva, urine, ascites fluid, and blood. In addition, the patient consent form should be translated into languages other than English as needed, to make sure that the patient can understand the information contained therein. A translator should be present if the patient does not understand the language in which the consent form has been written.

A research study that is linked to a clinical trial may involve the collection of presurgical specimens or ones that are collected during the surgery itself. Samples collected prior to surgery may include (but are not limited to) blood, urine, saliva, or a buccal smear. Samples collected during surgery may be resected from the patient or may be collected through the use of a catheter. Prior to surgery, a biorepository technician may post signs in the operating room to remind the operating room staff that tissues from a specific patient are required for the clinical trial study. All specimens should be handled in as sterile a manner as possible. Surgeons should be requested to place specimens into a sterile container (provided by the biorepository) or in a sterile basin covered with sterile towels. Sterile instruments for this purpose should be maintained by the biorepository. The technician should be notified when a requested specimen has been surgically removed from a patient. The technician should then transport the specimen to the accessioning area and wait until the appropriate pathology staff member has taken the portion of the specimen that they require for a complete diagnosis. The surgical pathologist then will give the technician the waste tissue and sign a “specimen inventory” form indicating their approval for the biorepository to procure the waste tissue. The technician should process the tissue under the guidance of the surgical pathologist according to the specifications of the clinical trial. Tissues collected may be snap frozen (either on dry ice or in liquid nitrogen), frozen in Optimal Cutting Temperature (OCT), or fixed in buffered formalin and then embedded in paraffin. Sections can be cut onto slides from either the OCT-embedded tissue or the formalin-fixed paraffin-embedded blocks, and then used for pathology review. Some clinical trials may require that “touch preps” (whereby a tissue sample is gently dabbed onto a microscope slide to distribute a thin layer of cells for microscopic examination) be made from fresh wet tissues, while in other cases fresh tissues may need to be placed in a special tissue culture media or buffered saline solution so that the cells can be tested in functional studies. Blood taken from patients for clinical trial studies may be requested as whole blood or may need to be processed into sera or plasma. Some studies may require further processing of the blood such that white blood cells or red blood cells need to be isolated. Sometimes the availability of patient samples may be realized with short notice for the collection staff (e.g., transplant, death/autopsy material). Planning efforts should anticipate the need for trained personnel, materials, and supplies (e.g., chilled specimen containers, sterile instruments, etc.) without the need for last minute arrangements that may compromise specimen integrity.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

565

SPECIMEN TRANSIT

Most collections will likely be performed under more controlled conditions, but preparation for all possible scenarios will increase the likelihood of successful specimen collections.

SPECIMEN HANDLING As has been underscored throughout this chapter, handling specimens according to evidence-based protocols for a particular research design is always the best approach. However, it is not always possible to know exactly which assays will be performed. In those cases, specimens should be collected according to protocols that are most likely to support the effective maintenance of the biospecimen so that it most closely resembles its state prior to removal from the patient. Having the correct cryovials, vacutainers, and other containers available at the collection site is critical, and those containers should support anticipated downstream uses for the samples. Agents that stabilize biological samples such as anticoagulants may be essential for some downstream applications while having a deleterious effect on others. In addition, some supplies may carry expiration dates that set a defined period of acceptable use (e.g., vacutainers, tissue culture media, and antibiotics). Every effort should be made to track expiration dates and make sure that supplies and reagents are used within the allowable time frame. Container selection is another important area of consideration in the management of patient samples. Containers should be composed of materials that will not interfere with the downstream assays; this is especially important if testing will include measurements of trace compounds. In addition, containers must be stable under subsequent storage environments and be of a size appropriate to the volume of the material stored. In general, vials with threaded caps do better when specimens are stored at ultracold temperatures (i.e., 80 C and cooler). Selecting the appropriate containers may reduce sublimation of the specimen over time and will facilitate subsequent storage and handling. Smaller tubes take up less storage volume and thus allow for efficient use of environmental storage containers (e.g., refrigerators and freezers). Even when automated specimen handling is not anticipated in the immediate life of the study, the use of automated specimen handling equipment in clinical research is on the rise. Vials and other containers that are amenable to automated handling (e.g., liquid handling equipment for aliquot preparation or plates used in immunofluorescence or other multiplex assays) can save labor and improve the accuracy and quality of specimen handling activities down the road. Finally, investigators should be sure that containers sent to outside

laboratories for testing are appropriate for the protocols utilized by those laboratories. Another area that is important to consider when processing biological specimens is the amount of the specimen that is being stored. Frequently it is necessary to divide, or prepare aliquots of, the material collected to allow for different uses for the specimens in question, or to be able to repeat tests when necessary. Increasing the number of aliquots is likely to increase the amount of storage needed, adding to the costs of implementing the research. Certain molecular species, such as proteins, may be irreversibly modified by exposure to cycles of freezing and thawing, whereas other species such as DNA are thought to be relatively resistant to such temperature cycles. When sensitivity to cycles of freezing and thawing is anticipated, aliquots should be made that will most likely reflect the volume of material needed for anticipated assays. In contrast, source vials containing molecular species that are not affected by freezing and warming cycles can be stored and aliquots made according to the schedule of the testing to be performed. Delaying the preparation of aliquots offers the advantage that as new testing platforms are developed, the amount of material required can be specifically removed and the specimen can be used in the most efficient manner possible.

SPECIMEN TRANSIT Whether a sample is being transferred down the hall or across the globe, there are particular considerations that should be made to ensure specimen integrity. Even for short transfers, care should be taken to ensure that the temperature and other environmental conditions needed for the stabilization of the biospecimen are met. Sometimes specimens must be immediately cooled by storage on frozen cold packs (approximating 4 C storage), on dry ice (approximating storage at 80 C), or in liquid nitrogen vapor ( 150 C) to stop molecular and/or enzymatic activity within the biospecimen. The rate at which a biospecimen is cooled also may play an important role in the subsequent viability of the cells contained therein or the structural integrity of the proteins. Other times, cooling the specimen to temperatures any lower than ambient may have a deleterious effect on some molecular components. This is where it can be advantageous to measure the analytes of interest as a function of different handling conditions during the specimen collection and processing before the study is initiated, to be sure that transit temperatures are appropriate. In cases where specimens are transported over periods longer than an hour, a monitoring device may be included within the shipment to allow investigators to document

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

566

32. MANAGEMENT OF PATIENT SAMPLES

whether deleterious temperature excursions occurred during transit. Transport outside of an institution or over greater distances carries additional requirements for adherence not only to good practices that will ensure the stable transport of the sample but also to national and international regulations. Transfer of specimens and/or data that are associated with a patient sample to a research entity outside of the one responsible for its collection may require the use of a Material Transfer Agreement (MTA) (see Chapter 29). MTAs document the allowed uses for the samples and/or data, ensure that uses are consistent with permissions afforded in patient consent documents and may address issues pertaining to intellectual property rights associated with the material. Depending on institutional policies, MTAs may require approval by IRBs or other bodies that govern the use of human subjects in research. Consequently, efforts to have MTAs established and obtain all needed approvals should be initiated as early as possible in the research plan. All specimens transferred from one research institution to another must be packaged according to national and international regulations. Most couriers require that the standards created by the International Air Transportation Association be met even when the transit may not involve transportation by air. Shippers begin by determining which regulatory requirements must be met and the physical requirements needed to ensure proper shipping. The shipper must determine how the specimens should be classified (e.g., infectious substance, Category A or Category B, patient specimens, or genetically modified organisms). In addition, the shipper must understand what preservatives may have been used that could be considered gaseous, toxic, flammable, or corrosive; all of these are “dangerous goods.” Strict adherence to packaging and documentation requirements as outlined in the standards for each dangerous good being transported must be met. Shippers are required to receive training in this area before they perform any shipping activities. This training must include general familiarization with the regulations on the specifics relating to the type and number of specimens being transported and concerning safety matters relating to the preparation and implementation of the shipment. The shipper has sole responsibility for adhering to local, national, and international regulations governing the shipment of biological specimens.

SPECIMEN STORAGE Issues pertaining to specimen storage begin when a sample is received in the laboratory setting or specimen repository in which the specimen will be stored. SOPs

should be written that govern the tracking of a specimen on its arrival to document the time it arrives and its condition. Receipts (as well as shipments out of the laboratory or repository) should be recorded in a log that also records the name of the person who receives the shipment and any interim storage location used. Specimens should always arrive accompanied by a specimen manifest that documents the specimen IDs contained within the shipment. Any damage to the specimens (and the packaging) should be documented and communicated to the shipper. Under some adverse situations, specimens may need to be replaced. Initiating this process as early as possible can minimize the effect of the occurrence on the overall implementation of the research effort. Once receipt of the specimens is recorded in the log, the specimens should be evaluated to ensure that all of the specimens in the container match what is included with the manifest. Any discrepancies should be immediately communicated to the shipper. If the samples have been labeled with barcodes, inventory of the specimens can be made rapidly using a barcode scanner (or a plate reader). If specimen labels do not meet the specifications of the protocols for the study, appropriate labels should be generated and applied so that subsequent specimen handling can be expedited. After confirmation of the manifest has been completed, specimens should be transferred to the appropriate storage locations and the locations should be entered into the specimen inventory system. In addition to ensuring that specimens stored have proper labels, consideration should be given to the strategy used to place the specimens into their storage environments. Specimens should be stored according to a plan that reflects how they are likely to be removed at a later time for testing. Specimens are usually retrieved from storage for a particular study using a defined material type (e.g., plasma, serum, or urine). Storing specimens according to anticipated retrieval patterns will expedite the retrieval process and minimize the labor needed to retrieve specimens. There are times when the order of the specimens in a container must be maintained, so the investigator should be consulted before specimens are rearranged. When selecting the storage environment, factors to consider should include temperature, humidity, and sensitivity to light, as appropriate. In the absence of data to the contrary, one may assume that the colder the storage environment, the greater the stability of the biological material. This is certainly true for the cryopreservation of cell suspensions, wherein the maintenance of viability is required. Intact cells are generally stored in liquid nitrogen vapor temperatures (< 150 C) that are below what is referred to as the Tg or glass transition temperature (< 135 C) where all

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

567

SUMMARY QUESTIONS

molecular motion is believed to cease. Many biological materials are stored at what are considered “ultra-low” temperatures that range from 70 C to 80 C. In contrast, specimens that have been fixed with formalin or other fixatives may be stored at room temperature but under conditions that control for humidity. Tissue blocks that have been fixed and then embedded in paraffin may be vacuum sealed in plastic packets, then stored at 4 C in the dark.15 The storage conditions selected should reflect practices reported in the scientific literature for similar types of specimens, and for the performance of assays similar to those to be included in the planned study.16e18 Recently, new storage matrices have been developed that allow for room temperature storage of DNA19 and RNA.20 Use of these products may facilitate storage in situations where freezers or power supplies may not be available or when long-term storage is needed at a low cost. If biospecimens are to be stored in mechanical refrigerators or freezers, investigators should evaluate new products as they become available. New designs include equipment that requires less energy and can significantly reduce facility costs for specimen storage. All equipment, whether mechanical or cooled by liquid nitrogen or other coolants, should be tested before specimens are added to ensure that the equipment performs under the manufacturer’s specifications. Emergency power (or reserves of liquid nitrogen or other coolant) should be available to allow for interruptions in power and should be maintained to allow at least 48e72 h of emergency power supply. In addition, should individual units fail, some cold, empty backup storage units should be available to hold specimens from the failing unit. In large biorepositories with dozens of storage units, a rule of thumb is to maintain approximately 10% of the total volume needed for ultracold and approximately 3% backup for liquid nitrogen. All equipment in specimen repositories should be included in preventive maintenance programs to extend the time-effective operation of the equipment for as long as possible. Monitoring systems should be implemented to alert appropriate staff on a 24-h basis to any temperature excursions that might affect specimen integrity.

ACCESS TO PATIENT SAMPLES Use of patient samples for a particular study is under the guardianship of a custodian, who ensures that those uses are implemented according to the scientific goals for the study. While custodians are generally individuals named as contacts for the biospecimen resource, decisions on use and requests to use the biospecimens for other allowable research may be made by a group of experts

and may include participation from laypersons with interests in the particular area of research. Guidelines on how specimen resources should be used need to be well defined and clearly articulated and be processed in a timely manner. Requests should be based on sound scientific research questions and must fall within parameters established in patient consent documentation. At the time of consent, patients should be given contact information and guidance as to what steps should be taken if they later wish to have their specimens or data withdrawn from a study. Repositories should have procedures in place to identify the specimens and ensure that the specimens and any associated data are not used for subsequent research activity.

SPECIMEN CULLING, TRANSFER OF COLLECTIONS, AND REPOSITORY CLOSINGS Because of the expense of maintaining collections of patients’ samples, it should be a regular practice to make sure that the specimens stored are ones that are suitable for research. Sometimes conditions arise when the integrity of a specimen has been compromised (e.g., when its label has been lost or cannot be read or because of environmental problems such as a freezer failure). Periodic reviews of collections should be performed to make sure that the specimens stored are appropriate and are “fit for purpose” for the anticipated research. There are other situations, such as when research funds are no longer available, that collections must be destroyed or “culled” because they can no longer be maintained by the original biospecimen resource. Under these conditions, it may be possible to have the specimens transferred to another research entity that may be able to use the specimens in new research activities (which are consistent with consent documentation). All culling activities, whether by destruction or transfer to a new institution, should be well-documented. Under conditions wherein specimen collections are transferred to a new institution, all requirements mentioned earlier regarding the establishment of MTAs, ensuring that other approvals are in place, and that specimen transport regulations have been met, must be considered prior to the transfer of the specimens.

SUMMARY QUESTIONS 1. What is a biorepository? a. A storage center for biological specimens b. An educational center for high school biology majors

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

568

2.

3.

4.

5.

6.

7.

32. MANAGEMENT OF PATIENT SAMPLES

c. A library of medical records d. A storage site for repossessed vehicles What does “fit for purpose” mean with respect to the development of scientific protocols? a. That residual specimens from prior testing efforts are used in the development of specimen handling protocols b. That the type of biospecimen used has been optimized for a particular type of analysis c. That individuals performing the testing are trained appropriately to perform the assays needed d. That multiple laboratories will perform testing simultaneously to assess consistency in the results What are “evidence-based” protocols? a. Protocols that have been developed within the past year b. Protocols that yield reproducible results c. Protocols that have been used consistently in a laboratory setting d. Protocols that have been optimized using highquality specimens and other reference standards to ensure that the procedures developed will yield scientifically significant results Which of the following would be considered “preanalytic variables”? a. Medications taken prior to specimen collection b. Time at which the specimen is collected c. The type of specimen fixative used and the time the specimen spent in the fixative d. All of the above Which of the following types of records should be maintained? a. The name of the person making the record b. The list of SOPs c. The time and date when the specimen was collected d. All of the above Which of the following statements are true regarding 2D barcodes? a. They contain information stored in a vertical as well as a horizontal direction on the label b. They can contain less information than a 1D barcode c. They usually take up a smaller space on a label d. They may be printed on the bottom of a vial What is an MTA and what is its purpose? a. An MTA is a Management Transfer Agreement that describes what type of shipping must be used to send biospecimens to their next destination b. An MTA is a Material Transfer Agreement that it is used to document how a biospecimen may be used once it is transferred to its next destination c. An MTA is a Missing Tariff Analysis that is used to investigate biospecimens detained in customs during international shipments

d. An MTA is a Manual Temperature Analysis that is performed when an environmental temperature chamber exceeds its allowable range of operation

References 1. Vaught JB. Biorepository and biospecimen science: a new focus for CEBP. Cancer Epidemiol Biomarkers Prev 2006;15(9):1572e3. 2. Vaught J. Biobanking comes of age: the transition to biospecimen science. Annu Rev Pharmacol Toxicol 2016;56:211e28. 3. Pitt KE, Campbell LD, Skubitz APN, Aamodt RL, Anouna A, et al. Best practices for repositories. I: collection, storage, and retrieval of human biological materials for research. Cell Preserv Technol 2005; 3(1):5e48. 4. Pitt KE, Campbell LD, Skubitz APN, Somiari SB, Sexton KC, Pugh RS. Best practices for repositories: collection, storage, retrieval, and distribution of biological materials for research. 2nd ed. Cell Preserv Technol 2008;6(1):3e58. 5. Campbell LD, Betsou F, Gariua DL, Grini JG, Pitt KE, Pugh RS, et al. Best practices for repositories: collection, Storage, reterival and distribution of biological materials for research. 3rd ed. Biopreserv Biobank 2012;10(2):79e161. 6. So¨derholm PP, Alfthan G, Koskela AH, Adlercreutz H, Tikkanen MJ. The effect of high-fiber rye bread enriched with nonesterified plant sterols on major serum lipids and apolipoproteins in normocholesterolemic individuals. Nutr Metab Cardiovasc Dis 2012;22(7):575e82. 7. Lippi G, Lima-Oliveira G, Salvagno GL, Montagnana M, Gelati M, Picheth G, et al. Influence of a light meal on routine haematological tests. Blood Transfus 2010;8(2):94e9. 8. Hernandez ML, Harris B, Lay JC, Bromberg PA, Diaz-Sanchez D, Devlin RB, et al. Comparative airway inflammatory response of normal volunteers to ozone and lipopolysaccharide challenge. Inhal Toxicol 2010;22(8):648e56. 9. Mazzoccoli G, De Cata A, Greco A, Carughi S, Giuliani F, Tarquini R. Circadian rhythmicity of lymphocyte subpopulations and relationship with neuro-endocrine system. J Biol Regul Homeost Agents 2010;24(3):341e50. 10. Spruessel A, Steimann G, Jung M, Lee SA, Carr T, Fentz A-K, et al. Tissue ischemia time affects gene and protein expression patterns within minutes following surgical tumor excision. Biotechniques 2004;36:1030. 11. Betsou F, Barnes R, Burke T, Coppola D, DeSouza Y, Eliason J, et al., International Society for Biological and Environmental Repositories (ISBER) Working Group on Biospecimen Science. Human biospecimen research: experimental protocol and quality control tools. Cancer Epidemiol Biomark Prev 2009;18:1017e25. 12. Betsou F, Gunter E, Clements J, De Souza Y, Goddard KAB, Guadagni F, Yan W, Skubitz A, Somiari S, Yeadon T, Chuaqui R. Identification of evidence-based biospecimen quality-control tools. A report of the international Society for biological and environmental repositories (isber) biospecimen science working group. J Mol Diagn 2013;15:3e16. 13. Betsou F, Lehmann S, Ashton G, Barnes M, Benson EE, Coppola D, DeSouza Y, Eliason J, Glazer B, Guadagni F, Harding K, Horsfall DJ, Kleeberger C, Nanni U, Prasad A, Shea K, Skubitz A, Somiari S, Gunter E, International Society for Biological and Environmental Repositories (ISBER) Working Group on Biospecimen Science. Standard preanalytical coding for biospecimens: defining the sample PREanalytical code (SPREC). Cancer Epidemiol Biomark Prev 2010;19(4):1004e11. 14. Lehmann S, Guadagni F, Moore H, Ashton G, Barnes M, Benson E, Clements J, Koppandi I, Coppola D, Demiroglu SY, DeSouza Y, De Wilde A, Duker J, Eliason J, Glazer B,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

Harding K, Jeon JP, Kessler J, Kokkat T, Nanni U, Shea K, Skubitz A, Somiari S, Tybring G, Gunter E, Betsou F, International Society for Biological and Environmental Repositories (ISBER) Working Group on Biospecimen Science. Standard preanalytical coding for biospecimens: review and implementation of the Sample PREanalytical Code (SPREC). Biopreserv Biobank 2012;10(4):366e74. 15. Srinivasan M, Sedmak D, Jewell S. Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol 2002;161(6):1961e71. 16. Unhale SA, Skubitz APN, Solomon R, Hubel A. Stabilization of tissue specimens for pathological examination and biomedical research. Biopreserv Biobank 2012;10(6):493e500.

569

17. Hubel A, Aksan A, Skubitz APN, Wendt C, Zhong X. State of the art in preservation of fluid biospecimens. Biopreserv Biobank 2011; 9:237e44. 18. Hubel A, Spindler R, Skubitz APN. Storage of biospecimens: selection of the optimal storage temperature. Biopreserv Biobank 2014; 12(3):165e75. 19. Muller R, Betsou F, Barnes MG, Harding K, Bonnet J, Kofanova O, Crowe JH. Preservation of biospecimens at ambient temperature: special focus on nucleic acids and opportunities for the biobanking community. Biopreserv Biobank 2016;14(2):89e98. 20. Mathay C, Yan W, Chuaqui R, Skubitz A, Jeon J, Fall N, Betsou F, Barnes M. Short-term stability study of RNA at room temperature. Biopreserv Biobank 2012;10(6):532e42.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

33 Evaluating a Protocol Budget Phyllis Klein Washington University, St. Louis, MO, United States

O U T L I N E Overview

571

Institutional Review Board Fees

572

Overhead or Indirect Cost

572

Determining the Hourly Rate

572

The “Per Patient” Budget

573

Start-Up Cost and Invoiced Items

577

OVERVIEW There are many types of protocol budgets, but the one thing they all have in common is that they summarize how one will be paid for work when participating in a clinical trial. Research comes in all shapes and sizes, but unless “money is no object,” you need to discover if the cost of participating is worth the effort. To do this you need to prepare a study budget. This entails an independent review of the protocol by turning each study-related task into a cost for the site. Specifically, each task should be broken down into its smallest detail to estimate the “time and materials” needed to complete the job. To separate the different parts of a study for estimating purposes, we will divide our budget into two main categories: • Start-up costs (and ongoing administrative and invoiceable activities) • “Per patient” budget

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00033-2

Submitting Your Budget to the Sponsor for Approval

582

Areas of Concern

585

Walking Away

586

Wrapping Up

586

Just as there are many types of budgets, there also are different types of sponsors. For the sake of simplicity, this chapter will divide sponsors into two types: • For-profit • Not-for-profit The difference between the two is that “not-forprofit” studies will be calculated at your site’s cost plus overhead. “For-profit” study budgets will allow a higher hourly rate in an effort to recover (some) coordinator down time. In addition, all budgets include some level of overhead to cover the cost of site operations. Prior to preparing a protocol’s budget, consider some housekeeping items: • Institutional Review Board (IRB) fees • Overhead or indirect cost (and how it is applied) • Employee hourly rates

571

Copyright © 2018. Published by Elsevier Inc.

572

33. EVALUATING A CLINICAL TRIAL BUDGET

INSTITUTIONAL REVIEW BOARD FEES There are two kinds of IRBs (also known as an independent ethics committee, ethical review board, or research ethics board), and both charge differently for their reviews: • Commercial IRB • Local IRB Commercial IRBs review protocols for a sponsor and are able to add all sites under one main review. The main part of the application is completed by the sponsor, and the commercial IRB charges the sponsor directly for this initial review. Each site added to the initial application will generate an additional fee paid directly by the sponsor. Each site is responsible for a small part of the application and consent, as applicable, to any local laws, ordinances, or preferences. Local IRBs are generally associated with academic centers or hospitals. A member of the study team will be required to complete the local IRB application and the informed consent form (ICF) for their site. Since the site bears the burden of this responsibility, they also cover the cost for the employee (at their site) who does this work. As a site, you need to know which type of IRB application you are obligated to utilize. Any fees related to your IRB application responsibilities will be predicated on the amount of effort required to complete these services and will be included in the start-up and ongoing administrative activities portion of the budget.

OVERHEAD OR INDIRECT COST Overhead is meant to cover the expense of operating your site. These costs are not directly related to the research you are doing, but they indirectly affect your site’s cost to participate. The rent, electric bill, desks, computers, phones, and administrative help are all components of this expense. A sponsor expects you to have an office from which to operate and the tools/equipment required to do the job to participate in their study. If you do not know what the overhead rate is at your site, ask the business manager or financial manager in your office or principal investigator’s (PI’s) department. In addition to the rate, you need to know what costs will be subject to overhead. Is it all costs or only some costs (e.g., the clinical aspects that include the “per patient” budget, as opposed to all costs that would include start-up fees, stipends, travel expenses)? Or are their two different rates, one for the administrative activities and the other for the clinical portions? Once

you understand the overhead policy at your site, you will be able to apply the rate(s) appropriately.

DETERMINING THE HOURLY RATE The next step is to determine the hourly rate of the employees who will be involved in the day-to-day conduct of the study. There may be many ways to do this, but by determining the number of days per year that the employee is available to work is a good place to start. Let us break this down step by step using Table 33.1 to illustrate this process. (Refer to the Employee Handbook at your site to substitute the holidays, vacation hours, paid sick time, etc., as a difference in these numbers will affect the hourly rate.) Now that you know the number of hours the employee is available to work, you can calculate the hourly rate. Start with the annual salary of the first employee. For purposes of this example, let us assume you are using a study coordinator at an annual salary of $60,000. If you refer back to Table 33.1, you will see how this calculation was done, and the resulting rate is $51.21 per hour. This is your cost for this particular employee. You may repeat the process for each employee to get accurate rates, or you can take the average salary of all of your employees in a particular role to determine an hourly rate for anyone that assumes that role for any study. When used in the budget as the hourly rate, this will give you your cost to conduct this trial. You would use this rate for “not-for- profit” studies. “For-profit” studies, like those sponsored by industry (pharmaceutical companies), can support a higher hourly rate to cover the down time in the everyday life of a typical study coordinator. It is impossible to plan a daily work flow that will keep any clinical research employee busy exactly 8 h per day, 52 weeks per year (less their time off), given any number of research studies. That said, your employees expect to be paid for every scheduled hour of work per year as indicated by their annual salaries. If an investigator keeps adding work to ensure their study coordinators generate enough funds to cover their salaries, it will not take long before the employees are overworked and feel underappreciated. Just how many studies does it take to cover a coordinator’s salary is a mystery that is impossible to predict. It takes trial and error. So, to determine an hourly rate that can be used for industry-sponsored trials also will take some effort and calculation. In Table 33.1, the percentage rate was calculated using the same premise that determined the hourly rate. If our assumptions are correct, then the best possible percentage of time a coordinator has available to work is 73%, but this does not cover the time between studies when the coordinator is at work, but has little to dodtimes when work is slow. To cover this

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

573

THE “PER PATIENT” BUDGET

TABLE 33.1

Hourly Rates

How to Compute Your Hourly Rate

Hours/Year

Calculations

Full-time employee works 261 days/year at 8 h day

2080

(52 weeks  5 days ¼ 260 days ¼ 2080 h)

Less vacation days (e.g., 22 days at 8 h day)

176

(22  8 ¼ 176 hrs)

Less unverified sick days (e.g., 5 days at 8 h day)

40

Less paid holidays (e.g., 8 days at 8 h day)

64

Actual workable hours per year

1800

Less breaks, email, etc. (e.g., 1 h per work day)

225

Yielding realistic work hours per year

1575

Less one mandatory meetings per/week per/year at 1 h each

52

Weeks per year

Less time when work is slow and nothing is done

To be determined

This quantity is unknown and may differ at each time point in a year. It depends on the number of studies open and number of patients enrolled at any given time during the year. It is helpful to be aware that this category exists even if you can’t quantify it

Approximate possible billable hours/year

1523

Maximum percentage of hours that can be billed each year

73%

Annual average salary of coordinators (e.g., $60,000/year)

$60,000.00

Add for benefits paid by employers (e.g., 30%)

$18,000.00

Annual cost of employee to employer

$78,000.00

Hourly rate

$51.21

“down time,” we will augment the coordinator hourly rate only when budgeting for the “for-profit” studies. For our example, let us start by adding 50% (of the hourly rate) to the predetermined rate to use in “for-profit” budgets. In our example, $51.21 was our cost rate, add 50% by multiplying $51.21 by 1.5 ¼ $76.82. If this rate suits your site, you can round up to $80 per hour or down to $75 to get a number easier to work with than $76.82. Now you are set with an hourly rate to use in either “not-for profit” studies ($51.21) or “for-profit” studies ($75e80). Next utilize these rates in the budget for the protocol of your choice. [Later if you decide this is not an adequate rate to cover your coordinator expenses, can you add 60% or 75% (of the hourly rate) and see if this calculation is a better fit for your site. This approach can be used until you find a comfortable fit for your research needs.]

Since all studies are developed to improve patient care in some way by providing an alternate treatment option (drug), a new test (device), or by simply collecting data, computing the cost based on the “per patient” visit schedule is the best place to start. In your protocol you may find a visit schedule already prepared. If your study does not have one, you need to create the time line and procedure list from the details in the protocol. An Excel spreadsheet can be used to construct not only the timeline and procedure list, but the budget as well. By creating the timeline, you will indicate the number of visits your patients will be required to attend once they consent to participate. The next step is to determine what procedures will take place at each patient visit in the timeline. When working with a new investigator for a PI-initiated study, a good budget person can help them develop their protocol with prompts such as:

THE “PER PATIENT” BUDGET

• Will there be a full physical exam at each visit, or only at the first or last visit? • A brief (directed) exam at other visits or no exams? Are special exams needed: neurology, ophthalmology, dermatology, etc.?

Before you begin, it is important to have a detailed description of what is going to take place during the study. The more detailed the description, the better the budget.

(1800/8 ¼ 225 h)

1523/2080 ¼ 73%

$60,000  30% ¼ $18,000

$78,000/1523 ¼ $51.21

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

574

33. EVALUATING A CLINICAL TRIAL BUDGET

• Will labs be obtained at some visits? All visits? Central lab or local lab? Research only or use of standard of care (SOC) labs? • Will the patient complete questionnaires? • Will any other measurement of patient safety such as electrocardiogram or magnetic resonance imaging be used? • Will there be investigational drugs in this study? Will a pharmacy be needed? Or will the coordinator (or PI) be able to store and dispense the meds? It may not occur to the PI that you are not familiar with a routine clinic visit in his/her therapeutic area. Asking the PI to verbalize exactly what is expected at each visit will help him/her think through each visit and help you better understand what is needed to prepare the budget. The cost of all procedures at all visits will determine the amount of the “per patient” portion of the budget. Therefore, the protocol needs to include everything that will happen during this study. But let us not get ahead of ourselves. Obtaining the cost of all procedures done by others (persons or departments not part of the immediate study team) is needed. Create a contact list containing the names of individuals that can help with pricing for particular procedures in different departments. This will provide TABLE 33.2 Sponsor: PI: overhead rate:

Procedure MRI lumbar puncture

an easy reference for future budgets. When requesting pricing, be sure to let your contact know if the study is “for-profit” or “not-for-profit.” If a lumbar puncture is required for your study, determine where the procedure will take place and who will do it. Is there a technical fee that needs to be paid to cover the use of the hospital (or another room that contains a sterile environment)? The technical fee covers the place and supplies, while a professional fee covers the physician or other qualified personnel doing the procedure. If pathology is involved, technical and professional fees for that work are additional. If anesthesia is involved, include the cost of medication and use of the equipment needed to deliver it (the technical fee), as well as a fee for the anesthesiologist or nurse anesthetist to administer it (the professional fee). You could potentially have a single procedure that will require costs for all of the above examples (consider a colonoscopy where biopsies are obtained and sent to pathology). To track all of the many parts for the total cost of one procedure, use of an Excel spreadsheet like the one displayed as Table 33.2 may help. The effort of the coordinator and investigator does not end with totaling the cost of the procedures. There is work to be done preparing for each visit, cleaning up after each visit, and completing the case report forms

Cost Breakdown Budgeted Costs 36% CPT code

CPT LABS: code CBC w/diff CMP Creatinine Kinase FSH serum hcg qual hepatic function panel LDH (lactate dehydrogenase) lipase magnesium phosphorous PT/INR PTT Urinalysis, macro Urinalysis, macro only subtotal

Technical fees

Prof Fee

Contrast/ Supplies

CRC

Subtotal $ $ $ $ $

cost

-

5% COL ↑ $ $ $ $ $

-

total cost with overhead $ $ $ $ $ -

Amount 5% COL ↑ Budgeted Reminders: don’t forget anesthesia don’t forget pathology

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

THE “PER PATIENT” BUDGET

(CRFs) documenting each visit. Later, CRFs may be queried when monitors visit the site to verify the data collected with the source documents (after patients are enrolled). Together, all of these activities need to be included in the budget. To obtain the costs of all study-related activities, estimate how long it takes to do each task. If you assign time by quarter hours, you can estimate the cost by dividing your hourly rate into quarter-hour segments. For example (and using the predetermined “forprofit” hourly rate), if a task takes an hour, the cost will be $75e$80; if 45 min, the cost will be $55e$60; if 30 min, $35e$40; and if 15 min, $15e$20. If a procedure is quick (like a weight or height measurement), estimate the cost as a smaller fraction of your rate, such as $10. Budgeting is not an exact science, sometimes it is important to use best judgment to come up with a starting place for a cost. Ask someone who has done the procedure previously to help determine if the cost selected is an accurate one. When signing the Statement of Investigator Form 1572, the PI accepts full responsibility for the research done at his/her site. Besides capturing the cost of any study specific procedures (e.g., physical exams) done by the PI, it is allowable to include a fee per visit for PI oversight. Start with a small fee per visit (e.g., $50), of which the total accumulation of this fee for all visits will represent the “per patient” fee to cover PI effort to sign off on all study paperwork (including review of all labs and decisions regarding patient safety) during each subjects study participation. Whatever amount is used, the fee should be reasonable so it can be absorbed into the cost of each visit. In addition, there are patient-related expenses that can be covered in the budget to allow the patient to participate without personal expense. Covering the cost of parking and transportation can be reimbursed in the form of a stipend per visit. Your sponsor will tell you if using a stipend is an option. Travel expenses can be requested in your budget for patients traveling a great distance (such as >100 miles) from the site. This expense would be reimbursed outside the “per patient” budget since it will not be applicable to all patients. Other related travel reimbursements also are possible (e.g., taxi fare, bus fare, hotel costs, meals, etc.). To obtain additional funds such as these, all you need is prior approval from the sponsor. Only costs that apply to all potential participants are included in the “per patient” budget. So that you can see what a completed “per patient” budget looks like, Table 33.3 provides a generic example. Please note in our template example that coordinator services requiring an hour are listed as $75 or $80 (e.g.,

575

informed consent; inclusion/exclusion criteria, etc.); services requiring ½ hour are indicated by $35e$40 (e.g., inclusion/exclusion criteria at visit 2; monitoring visits); services requiring 15 min are listed as $20 (e.g., vital signs; Borg Dyspnea Index, etc.); and services taking less than 15 min are listed as $10 (e.g., weight). For procedures requiring a physician, it is appropriate to ask the physician for the name of someone in his/her department to contact for professional fee research pricing. Determine whether the business manager or someone in billing can help identify the correct research rates for this study. Many procedures done by physicians are bundled into a current procedural terminology code (CPT) that is then used to standardize pricing for these procedures nationwide. When you are quoted an amount, ask if they are quoting the Medicare rate or the rate based on the fee used for insurance billing. Be sure to let them know who is funding your study (industry, National Institutes of Health [NIH], department) as there may be different rates for different kinds of studies. Many industry sponsors use Medicare rates for their budget templates, but sites may not agree with this choice as it can be lower than the actual cost. Sites may choose to use a percentage over the Medicare rate or a percentage under the insurance billing rate. Knowing how your research rate compares to insurance rates or Medicare rates will help you understand how your budget compares to the sponsors template. Remember to add overhead to each visit of the total “per patient” budget! Now that the “per patient” budget for your site has been assembled, it is time to compare it to the sponsor’s budget template. Usually sponsors have an idea what they think the total budget range should be for a clinical trial they are supporting. To gain perspective, create a grid in Excel to do this comparison. Be sure to note the overhead rate the sponsor applied in their template because sometimes the difference between overhead rates can account for a big discrepancy between the two budgets. Make sure the timelines match. Sometimes a sponsor shows a single visit that can be repeated many times (until all enrolled patients reach a particular milestone). You understand that each time this visit repeats, the sponsor intends to pay the site the agreed upon amount. However, if you estimate this visit to be repeated six times in your budget, while the sponsor repeats it twice in their budget template, then your total budget will be higher than the sponsor’s total budget because the visits are not aligned equally (your budget will have four more visits). The proposed grid method can assist with this problem. It allows for a comparison of the costs per visit, as the visits and amounts are listed side by side. An example of a budget comparison grid can be found in Table 33.4.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

576 TABLE 33.3

33. EVALUATING A CLINICAL TRIAL BUDGET

Study Budget Template

Investigator Name: Sponsoring Agency:

Unit cost

Week -6 to -8 Week 0

Week 8

Week 16

Visit #5 Week 24 or ET

50.00

50.00

50.00

50.00

100.00

Informed consent

75.00

75.00

Incusion/exclusion criteria check

80.00

80.00

40.00

$120.00

Medical history

75.00

75.00

40.00

$115.00

Vital signs

20.00

20.00

20.00

20.00

20.00

20.00

Weight

10.00

10.00

10.00

10.00

10.00

10.00

$50.00

ECG

100.00

100.00

$200.00

6-minute walk

30.00

30.00

30.00

30.00

30.00

30.00

$150.00

Borg dyspnea index

20.00

20.00

20.00

20.00

20.00

20.00

$100.00

WHO functional class

50.00

50.00

50.00

50.00

50.00

50.00

$250.00

Concomitant medications

20.00

20.00

20.00

20.00

20.00

20.00

20.00

$120.00

Adverse event monitoring

20.00

20.00

20.00

20.00

20.00

20.00

20.00

$120.00

Medication dispensing

20.00

20.00

SF-36 survey

30.00

30.00

Phlebotomy and central lab processing

75.00

75.00

75.00

75.00

Case report form (CRF) completion

150.00

150.00

150.00

Monitoring visit

35.00

35.00

Coordinator services

150.00

150.00

Per visit subtotal Procedure listdcustomize to your study

Screen Visit #1

Baseline #2

Visit #3

Visit #4

Follow-Up #6 4 Weeks post tx

Per procedure Y

50.00

$400.00

Total

ADMINISTRATIVE Patient stipend

100.00

COORDINATOR SERVICES $75.00

100.00

20.00

$120.00

$20.00 30.00

30.00

$90.00

75.00

75.00

75.00

$450.00

150.00

150.00

150.00

75.00

$825.00

35.00

35.00

35.00

35.00

35.00

$210.00

150.00

150.00

150.00

150.00

75.00

$825.00

150.00

150.00

150.00

150.00

75.00

$675.00

50.00

50.00

50.00

50.00

$300.00

21.00

21.00

$84.00

INVESTIGATOR SERVICES Physician evaluation

150.00

Principal investigator fees (for patient oversight)

50.00

50.00

50.00

Basic metabolic panel (BMP)

21.00

21.00

21.00

Complete blood count (CBC)

24.00

24.00

24.00

24.00

24.00

24.00

24.00

$144.00

Liver function Tests (monthly)

40.00

40.00

40.00

40.00

40.00

40.00

40.00

$240.00

Biomarkers

200.00

200.00

$400.00

Pregnancy Test

25.00

25.00

25.00

25.00

$125.00

700.00

$1,400.00

LOCAL LABS

200.00 25.00

25.00

PROCEDURES Echocardiogram

700.00

Per visit Subtotal Overhead Total per visit cost with OH

36%

700.00 1,020.00

2,120.00

919.00

919.00

2,020.00

610.00

$7,608.00

367.20

763.20

330.84

330.84

727.20

219.60

$2,738.88

$1,387.20

$2,883.20

$1,249.84

$1,249.84

$2,747.20

$829.60

$10,346.88

577

START-UP COST AND INVOICED ITEMS

TABLE 33.4

Budget Comparison Grid

Sponsor: Site: PI: date of transaction:

Visits 1 2 3 4 5 6 7 8 9 10 11 12 13 total OH

Sponsor original budget template

→

Site estimate

Site budget to sponsor

Sponsor’s counter

Site counter

Approved Budget

$1,387.20 $2,883.20 $1,249.84 $1,249.84 $2,747.20 $829.60

$

20%

$10,346.88 $ 36%

36%

Once you have the opportunity to compare your estimated budget against the sponsor’s template, you will be able to determine if you have missed something (your budget is too low) or overestimated your costs (your budget is too high). A big discrepancy in one direction or the other will give you the insight to dig deeper and determine why. If still in doubt, before you submit your budget to the PI and coordinator for review, then return to the protocol and review what is needed at any particular visit. Double check your work to make sure you read the protocol correctly. Be sure to explain your budget to the PI and study team so that they understand how you estimated the budget. They may be able to address your questions if the protocol was not clear. Please note that it is also possible that the sponsor underestimated the procedure costs or coordinator effort required to complete the study visits, and therefore their budget template is too low.

START-UP COST AND INVOICED ITEMS The next area of the budget to complete is the grid for the start-up costs, annual costs, and any costs that will be invoiced as they are done. If you recall, in the overview it was mentioned that there are two parts of every budget. We tackled the “per patient” part first, because it is where the bulk of the activities that require pricing from different sources takes place. But, a lot of work also goes into the start-up activities, not to mention

$

36%

$ 36%

36%

preparing and negotiating the budget. During this section of the chapter, we will discuss how to arrive at fees for the following: • Regulatory and IRB application preparation (initial, annual, and amendment processing) • Study close-out fees • Document archiving • Investigational New Drug safety report processing fees This section is based on fees that are paid for services rendered, rather than itemizing procedure or labor costs. This effort is sometimes referred to as administrative work. It can be done by the coordinator, PI, or by a centralized group that specializes in these activities. These fees are obtained by estimating how long the activity will take to complete, and then creating a fee to cover this effort. You will be paid the agreed-upon fee regardless of how long it actually takes to perform this activity. If you are not sure, ask someone at your site who has done this work in the past to approximate how long it usually takes to do the tasks we are about to describe. For example, if you are submitting to a local IRB, preparing the initial IRB application, can take many hours. Not only do you complete the application, you also edit the ICF template by adding site specific language. Sometimes you transfer the template consent language into a local consent template. Upon completion of this task, the ICF is sent to the sponsor for review; revisions can be requested. Ultimately, you need sponsor

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

578

33. EVALUATING A CLINICAL TRIAL BUDGET

approval prior to submission to the IRB. This scenario can go back and forth a couple of times until both sides are satisfied with the results. Once the application and consent are submitted to the IRB, it is possible that the IRB may have questions about the application and/or request revisions to the ICF, too. Once those changes are made, the ICF is sent back to the sponsor for final approval before you return the consent to the IRB. Needless to say, the total number of hours spent prior to IRB approval can be more than you expect. For the purpose of this example, let us estimate 24 h for completing the initial IRB application including consent. Annually, the IRB application needs review and approval. This process requires less effort than it did initially because the renewal application and ICF only require updates since the last approval. You may also need to update regulatory documents if licenses (or any other documents) have expired. For the annual effort, let us estimate 10e15 h. The time will vary based on the number of subinvestigators and the number of participants enrolled at your site. To submit protocol amendments to the IRB for review and approval, changes need to be made to the IRB application prior to implementation of the protocol changes. Since you do not have to update regulatory documents or report on patient enrollment (these are done annually), this effort can be less than the annual estimate. For amendment processing, we will estimate 7e10 h (which includes changes to the consent if required by the amendment). Some of the initial regulatory documents prepared, need to be signed/dated by the PI. These include: Protocol and Investigator Brochure (IB) signature pages, as well as the Form FDA 1572 if an investigational drug is part of the protocol. To support the Form FDA1572, the regulatory coordinator will need to collect: • Curriculum vitae (CV) and appropriate license for the principal investigator and all subinvestigators listed on the 1572 • Financial Disclosure for all investigators listed on the 1572 • IRB federal wide assurance letter • College of American Pathologists (CAP), Clinical Laboratory improvement Amendments (CLIA), and CV/license for the lab director if using a local lab (if a Central Lab is used, the sponsor will supply these documents for your regulatory binder/file as they are also required) All of these documents illustrate the qualifications of the site to conduct this study. Let us estimate the effort for the regulatory coordinator to prepare and collect these documents at 6 h. Preparation and negotiation of a study budget can take time. This chapter is summarizing everything to

consider. Let us estimate the time to prepare and negotiate a budget at 16 h (you may find this number to be a little low if the study is a complicated one, but it is a starting place). Coordinators also have preparation work to do before the study begins and the first patient walks through the door. • • • • • •

creating source documents setting up local laboratory or radiology accounts unpacking lab kits used for central labs attendance at an investigator meeting conference calls from study monitors setting up and attending the site initiation visit before enrollment can begin

Let us allow 40 h for study coordinator effort. The PI also prepares for the study. Before being selected as a site, the PI needs to review the protocol and decide if this study is feasible for the patient population at his/her site. Questions the PI may consider are as follows: • Are the inclusion/exclusion criteria too strict that most patients will not qualify? Or are they too loose so that anyone can qualify? • Are the labs too restrictive? • Will the sponsor offer an open label extension study so that patients who respond to the investigational product can continue on the study medication until U.S. Food and Drug Administration’s (FDA’s) approval? For the Principal Investigator’s Time, let us allow 5 h. Most sponsors allow you to request a study start-up fee. All of the set-up time we discussed in the previous paragraphs in this section represent study start-up activities. (Please note that the effort related to the IRB application in this section is referencing use of a local IRB. Effort required when using a commercial IRB will be less for the site.) Let us tally the hours we estimated and figure what fee we should request: • IRB preparationd24 h • Regulatory document preparation and collectiond6 h • Budget preparation and negotiationd16 h • Coordinator source document preparation and account set-up timed40 h • PI feasibility and protocol reviewd5 h For this exercise, we will assume the initial IRB preparation, regulatory preparation, budget preparation, and coordinator preparation share an hourly rate of $75 (using our “for-profit” rate). Therefore the total hours spent on these activities are (24 þ 6 þ 16 þ 40 ¼ 86) 86 h. 86  $75 ¼ $6450. Plus 5 h of PI time at $200 per

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

START-UP COST AND INVOICED ITEMS

hour ¼ $1000. So, our total start-up fee based on the estimated effort discussed in this section should be $7450. It is likely the sponsor will ask what services this fee encompasses. It saves time to prepare memos in advance that illustrate the services covered by each fee (and detail the hours you are estimating). The total of all these memos should equal your initial start-up fee request. An example of memos explaining the start-up fee of $7450 can be found as Memo Examples #1e3 (notice that the total of the three memos equals the requested start-up fee). All memos should be typed on your site’s business stationery, dated and signed by the person at your site that approved the fee amounts. If

579

you do not reference any particular study in your memo, you can use these memos for all budgets you prepare to justify your fees. Sponsors want to know that the fees you request are charged for all industry-sponsored studies and not just the current study you are negotiating with them. Other departments involved in your study may also have start-up fees that need to be itemized in this section of the budget. • Local IRB fees (commercial IRB fees are paid directly by the sponsor) • To justify these fees, provide a link to the IRB website

Memo Example #1 (Print on Business Stationary)

Memorandum for Record Date

Regulatory and IRB Preparation fee – Funding detail In order to allow the coordinator adequate time to screen and enroll patients, and not to be buried with constant paperwork, our regulatory coordinators provide this service for the PI. We charge a $3,450 fee for this service which is passed along to the sponsor as part of the Start-up fee. This fee includes all paperwork related to: • preparing the initial IRB submission paperwork • preparing the local ICF • preparing any other applicable committee applications • providing liaison services between the sponsor, the site, and the IRB to get this study approved • working with the sponsor to complete all of the regulatory paperwork, including: o Submission of signature pages o Preparing the 1572 and all supporting documents (FDs, CVs, licenses, etc.) o Preparing the budget (which includes getting research costs from other university and hospital departments supplying a study related procedure (ie. x-ray, MRI, etc.) o Medicare coverage analysis, and anything else required before the site initiation This fee supports ~46 hours work. After the first year, there is a renewal fee of $1,000 which covers preparing the IRB renewal application and all updates to the regulatory paperwork for the next year. The annual fee supports approximately 13-15 hours of regulatory coordinator effort. An amendment fee of $700 is to cover the cost of processing the IRB paperwork for sponsor related protocol amendments. This fee includes changing the consent form, if needed, as well. It covers about 7-9 hours of effort for each amendment. Please feel free to contact me at (xxx) -XXX-XXXX to discuss this further.

Sincerely, Name of Decision-making Authority Title

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

580

33. EVALUATING A CLINICAL TRIAL BUDGET

Memo Example #2 (Print on Business Stationary)

Memorandum for Record Date

Coordinator Start-Up fee – Funding detail The Coordinator Start-Up fee is just that - a one-time fee for preparing for this study before the first patient walks through the door. It includes preparing for the site initiation, pre-study conference calls; preparing source documents; unpacking CRFs (if applicable, or training for electronic CRFs), unpacking and sorting lab kits, and any other study related materials that are shipped to our site; preparing research requisitions for study procedures ( ie. x-rays, biopsies) that need to be set-up before the study begins, etc. This fee of $3,000 supports about 40 hours of coordinator prep work, plus overhead. Please feel free to contact me at (xxx)-XXX-XXXX to discuss this further.

Sincerely,

Name of Decision-making Authority Title

Memo Example #3 (Print on Business Stationary)

Memorandum for Record Date

Investigator Feasibility & Protocol Review fee – Funding detail Before a study is considered at our site, the Investigator must review the protocol to see if it is appropriate to offer to his/her patients. Once that decision is made, the investigator then must review the budget, compare it to the protocol, and determine if the study is financially feasible for our site. Many times negotiations must occur to make this determination. The dedicated Investigator hours to this study are many more that our request. Our fee of $1,000 for the investigator’s time covers ~4-5 hours. Please feel free to contact me at (xxx)-XXX-XXXX to discuss this further.

Sincerely,

Name of Decision-making Authority Title III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

581

START-UP COST AND INVOICED ITEMS

TABLE 33.5

Start-Up and Invoiceable Items

Sample Study Start-Up Fees

Initial Fees

Annual Fees

Institutional review board (IRB) review fee (industry-sponsored trial)

$2,500.00

$1,500.00

Regulatory and IRB preparation fee

$3,450.00

$1,000.00

Coordinator start-up fee

$3,000.00

Principal investigator feasibility and protocol review

$1,000.00

PRN Items

Comments for PRN Fees

$700.00

Amendment processing

Study closeout fee

$1,500.00

Document archiving

$1,000.00

Research pharmacy fees

$1,500.00

Prescreening

$2,000.00

$1,000.00 For chart reviews

Screen failures

Invoiced per procedure completed

Serious adverse event at our site

$700.00

Each occurrence

Reconsenting fee

$75.00

Each occurrence

Review/submission of investigational new drug safety reports:

$35.00

Each

Pregnancy test

$35.00

Each

Chest X-ray

$225.00

Each

Total

$16,450.00

$3,500.00

• Pharmacydinitial and annual fees • Ask your pharmacy to prepare a formal estimate for each study on letterhead (or a service agreement form) that references the study protocol. They can list their initial and annual fees, as well as their dispensing fees. This document can then be provided to the sponsor in support of the fees you are requesting. • Radiology fees to set-up for special imaging requirements • Facility fees for research units • Prescreening and/or recruitment fees Any other service provided for some participants and not others (e.g., pregnancy tests; travel expenses; serious adverse event follow-up) can be listed as an invoiceable item so that the sponsor has an opportunity to review and approve these costs. While this may seems like a never-ending list, these fees will only be charged if they apply. Refer to Table 33.5 where these costs are listed as an example of what the start-up fees and invoiceable items table can look like. Once the entire budget has been created and priced from trustworthy sources at your site, then you are ready to have your budget approved by the PI, coordinator, and/or the business manager of the department that employs your PI. If you are the PI preparing the budget, ask your coordinator and business manager to

review the budget to make sure you are not biased in your pricing. Involving others to provide input can validate your budget and allow you to make adjustments if necessary. The only way the PI and study team can evaluate the budget is to ensure that they have all of the necessary documents. The attachments should include: • “Per patient” budget and start-up grids that you developed • Sponsor’s budget template (if the sponsor provided one) • Budget comparison grid that compares the “per patient” budget (that you developed) visit-by-visit with the template provided by the sponsor • Cost breakdown spreadsheet that shows the pricing per procedure and how you arrived at the total amount Request that each reviewer reply to you with any suggested changes or approval of the internal budget you provided. If changes are made, resubmit the revised documents back to the same reviewers, and ask for a second reply. Continue in this manner until you have received approval from everyone and no other changes are requested. This will provide you with the documentation that those involved in this research at your site (other than you) reviewed the documents, approved them, and asked you to prepare the site budget for the sponsor on behalf of your site.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

582

33. EVALUATING A CLINICAL TRIAL BUDGET

Memo Example #4 (Print on Business Stationary)

Memorandum for Record Date

Pre-screening fee – Funding detail The Pre-screening funds are to pay for coordinator time to identify patients that may be eligible for the study by searching the internal/clinic/hospital database for appropriate candidates by matching their diagnosis ICD-10 codes with the one required for the study. Then, they check the participant’s chart against the inclusion/exclusion criteria to confirm an appropriate candidate. This effort can be very time-consuming. This fee is to help offset this cost as it begins before the patient signs a consent, and before the ‘per patient’ budget provides funding. In studies that require patient charts to be pulled and reviewed for eligibility in a clinical trial, we request funds to cover the cost of reviewing these patient charts. The nurse coordinator cost per hour is $75. We request $2,000 in advance to spend ~26 hours reviewing appropriate charts. We may request an additional amount if this process is fruitful. This exercise will search for patients in advance so that we can contact them and arrange for a study visit. The alternative is to not prepare in advance and wait until patients can be identified through a scheduled clinic visit.

Please feel free to contact me at (xxx)-XXX-XXXX to discuss this further.

Sincerely,

Name of Decision-making Authority Title

SUBMITTING YOUR BUDGET TO THE SPONSOR FOR APPROVAL The best way to submit your site budget to the sponsor is to modify the template they provide for that purpose. You may edit their template so that it better represents your internal budget requirements. It is likely that you will need to adjust the overhead rate, and perhaps the rates for individual procedures. The sponsor will let you know what they allow (and what they do not) after they have reviewed your total site budget. They may ask you to provide a rationale for a procedure rate change. Remember that you will not get what you do not request. So, be reasonable, but ask for everything you need. The worst that can happen is that the sponsor will say “no.” Be prepared to justify the fees you request. You can do this by creating memos explaining each fee as you would for the start-up fees. Make these memos generic so that you can reuse them, if needed, for future budget justifications. You may be surprised at how well these

memos are received by sponsors. To give you an idea of other memos to create and use, see Memo Examples #4e8. You may combine “like” fees into one memo so that the sponsor can see how they are related (as we did with Memo Example #1). In Memo Example #1, we included the annual regulatory and IRB preparation fee and the amendment processing fee with the initial regulatory and IRB preparation fee; this memo is used to illustrate that this effort continues after the initial work is approved. The IRB will review the study annually, and amendments need to be submitted to the IRB as they occur. All of this work needs to be recognized, accounted for, and reimbursed. Our memo describes this process and funding detail. Using the same rationale as applied above, for the annual regulatory and IRB preparation fee, you could request w$1000 (10e15 hours at $75 per hour), and for each protocol amendment submitted to the IRB, you could request w$700 (7e10 hours at $75 per hour).

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

SUBMITTING YOUR BUDGET TO THE SPONSOR FOR APPROVAL

Memo Example #5 (Print on Business Stationary)

Memorandum for Record Date

SAE at our Site fee – Funding detail The $700 "Serious Adverse Event" fee at our site is for the coordinator/PI time that is needed to complete the paperwork that the sponsor/IRB requires if, and only if, there is a serious adverse event at our site. If none occur, then we will not charge this fee. ‘Adverse event monitoring’ is the responsibility of the PI during the course of any study. However, once a patient is identified as having a serious adverse event, the effort to collect and report the data surrounding this event is not reflected in the per patient budget, nor is it an expectation of the PI or patient. Since the study drug is investigational, the PI may not have experience with this drug, and even if the PI has participated in a previous study, the patient has not. Each patient reacts differently to every drug, so each event is uniquely documented. This additional effort can be very burdensome for the study team, and is often associated with extra hours of work and weeks of follow-up that need to be accounted for in the budget. This fee is to cover the additional effort required when a SAE occurs. If we do not have an SAE, then we have not unduly charged the sponsor for study coverage that did not occur. This task is not administrative, but directly related to the drug the sponsor is researching. Do not confuse this task with adverse event monitoring which occurs at each visit between the study patient and the PI/coordinator and is, of course, included in our budget. This fee accounts for 7-10 hours @ $75 per hour of effort to properly document this event to the IRB and to the sponsor. Please feel free to contact me at (xxx) XXX-XXXX to discuss this further.

Sincerely,

Name of Decision-making Authority Title

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

583

584

33. EVALUATING A CLINICAL TRIAL BUDGET

Memo Example #6 (Print on Business Stationary)

Memorandum for Record Date

Re-consenting fee – Funding detail The re-consenting fee is to cover the coordinator/PI time needed to re-consent patients following sponsor initiated changes to the consent, such as a protocol change, a change in risks, or any other sponsor specified change. The rationale is that consenting a participant is a process not a document to sign. We will set aside time that was not specifically captured in the budget to explain these changes to the participant, and to be available to answer their questions. We estimate this task to take anywhere between ¼ hour to 1 hour of coordinator and/or PI time. We are asking for $75 per occurrence to cover this effort when re-consenting is required. Please feel free to contact me at (xxx) XXX-XXXX to discuss this further.

Sincerely,

Name of Decision-making Authority Title Memo Example #7 (Print on Business Stationary)

Memorandum for Record Date

IND Safety Reporting Reimbursement fee – Funding detail The processing of IND Safety reports costs us more than we are requesting, as there is considerable effort behind the scenes involved in reviewing and perhaps reporting these documents. The coordinator and the PI each have to review them, initial & date them, then decide if they are eligible for submission to the IRB by determining if they meet the 3 criteria of: related, unexpected, and serious. Once the decision is made to submit, they need to be scanned, and electronically submitted to the IRB, and/or filed in the regulatory binder for later review by a sponsor’s monitor. The IND safety reporting is not included in our per patient budget because we have no way of knowing how many of these reports we will receive since that volume could be impacted by other studies in which this investigational drug is being used. The charge is $35 for each report. This fee represents only 10-15 minutes of coordinator and PI combined time per report. Please feel free to contact me at (xxx) XXX-XXXX to discuss this further. Sincerely,

Name of Decision-making Authority Title

AREAS OF CONCERN

585

Memo Example #8 (Print on Business Stationary)

Memorandum for Record Date Study Closeout fee – Funding detail Due to increased activity at the close of a clinical trial to meet with the monitor for a final visit, to finalize queries, provide missing documents, and ultimately close out the study with the IRB, we need to cover this additional effort on the part of the study coordinator and regulatory coordinator as the ‘per patient’ budget is no longer providing any funding. Our $1,000 fee covers 10-14 hours of study coordinator/regulatory coordinator time to account for this additional effort. Please feel free to contact me at (xxx) XXX-XXXX to discuss this further. Sincerely,

Name of Decision-making Authority Title

AREAS OF CONCERN We would be remiss if we did not discuss areas that impact our clinical trial budget, even though they are not actually covered by our budget. • Mixing research with routine care, or sometimes referred to as standard of care (SOC) • Precertifying patient insurance for research participation • Billing errors • Insurance denial language The bullets listed above may add coordinator and/or investigator effort to the clinical trial that is not generally included in the clinical trial budget, mainly because these problems are rarely discussed prior to their occurrence and are forgotten shortly afterward. Streamlining patient care is a definite advantage for the patient when mixing research with routine care. It also is absolutely possible that this method will result in a cheaper “per patient” budget for the sponsor. But the risks can outweigh this value for the sites. It is entirely possible that patient insurance will be denied if their healthcare insurance policy does not

include research participation. If denied, your research patient may start receiving bills for the services you thought were SOC. Preventing this from happening will require additional coordinator time to protect the patient by precertifying their insurance. Since it is likely that your research coordinator has never had to precertify patient insurance before, they may need to discuss this with the precertification team located within your department. It often happens that a person doing a precertification can be on hold with an insurance company for hours! I was informed by the supervisor of a precertification team in one department that they are so busy (with their usual daily assignment) that they do not have time to help study coordinators with research precertifications. So, coordinators are left to complete this task on their own. This is a time-consuming process that is not included in the sponsor’s budget template. The sponsor thought using SOC would save money. It may for them, but not for the site. Please note that if the patient’s insurance carrier informs you that your patient’s policy does not allow for research participation, you will not be able to enroll this patient in your study unless they are willing to pay for their treatment out-of-pocket (unlikely), or the sponsor agrees to cover this cost (possibly).

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

586

33. EVALUATING A CLINICAL TRIAL BUDGET

There is a greater risk for billing errors if research is mixed with routine care. Double billing occasionally happensdthis occurs when a procedure is billed to patient insurance as well as being covered in the study budget. It takes significant administrative time to straighten out the billing and refund any payments made to patient’s insurance and/or Medicare. Any billing not corrected (if found) could be considered Insurance or Medicare fraud. Some sponsors are willing to cover the cost of a procedure if the patient insurance denies payment. This is a “slippery slope” that can result in complications for your site. Being aware of the Medicare’s Secondary Payer (MSP) rule and their most-favored nation status (https:// www.cms.gov/Medicare/Coordination-of-Benefits-andRecovery/Coordination-of-Benefits-and-RecoveryOverview/Medicare-Secondary-Payer/MedicareSecondary-Payer.html) may be helpful in determining if your organization is at risk when a situation such as this arises.

WALKING AWAY If you cannot reach a budget agreement with the sponsor because your estimate exceeds an amount they are willing to pay, you should consider not participating in the study. Not every study offered is a good fit for a particular site. One should discuss this option with the study team or business manager before making the final decision. Are there areas/costs in your budget that can be trimmed to meet the sponsor half way? If the department really wants to participate and can afford to break even or lose a little money, then this decision will become clear during the discussion. The reasons for participation in any study can be different based on each person’s perspective. Considering all perspectives is the best way to make such a decision. Cutting your losses before you actually start can save you time and money in the long run.

WRAPPING UP Once the study is complete and the IRB application is closed, you look at the study fund account balance and determine if the sponsor was supporting the clinical trial or if you were supporting the sponsor! wouldn’t you like to know before you enrolled patient #3 or #4 that your

budget was inadequate to cover all the costs associated with the study? There is no problem obtaining the costs associated with procedures that occur outside your office (labs, endoscopies, biopsies, etc.). The real wild card is estimating coordinator time. If there was a way for you to determine if the coordinator time estimated in your budget was adequate (or not), would you take the time to investigate? Hopefully, you would. This final step can be accomplished simply by adding the total funding (in the approved budget) just for the coordinator effort at each visit. In Table 33.3, you will find the itemized study coordinator services grouped together, so it will be easy for you to total the sums per visit. Divide the total of all coordinator dollars per visit by the coordinator hourly rate to determine the number of hours estimated for each visit. In Table 33.6, the coordinator services were pulled from the budget and separated for this illustration. By providing this information to the study coordinator (for each study they are coordinating), the coordinator now has the target number of hours estimated per visit for his/her effort, and it is now possible to get feedback on a regular basis on those visits once they are completed. Ask the coordinator to inform you if the actual hours exceed the estimated hours on a regular basis. Every patient is different, and some patients take more time than others. If on average the number of estimated hours is within range, then it is possible to assume that the approved budget was accurately estimated. On the other hand, if the actual hours habitually exceed the estimated hours, you need to find out why. Is the reason the patients are sicker and demand more coordinator time? Or do the procedures need more time to complete? If the protocol demands a pharmacokinetics schedule that requires two coordinators to execute, while the budget clearly provides funding for one coordinator, then this is a problem. For any of these scenarios, you can go back to the sponsor and request additional funding. Explaining in detail why the study is costing your site more than anticipated, while still providing the sponsor with the valuable information they need, offers you an opportunity to renegotiate the budget per visit. Remember that participating in clinical trials is a team effort. The responsibility of evaluating each portion of the budget should be shared among the team members. This will make your site budgets better in the future. Better budgets mean better funding to cover your site costs. A win/win scenario, don’t you think?

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

587

WRAPPING UP

TABLE 33.6

Coordinator Hours

EXAMPLE TIME LINE–Add column for additional visits Example visit list–customize to your study

Unit Cost

Screen Visit #1

Baseline #2

Visit #3

Visit #4

Visit #5

FollowUp #6

Week -6 to -8

Week 0

Week 8

Week 16

Week 24 or ET

4 Weeks per procedure post tx ↓

TOTAL

Coordinator services Informed Consent Incusion/Exclusion Criteria Check Medical History Vital Signs Weight ECG 6-minute Walk Borg Dyspnea Index WHO Functional Class Concomitant Medications Adverse Event Monitoring Medication dispensing SF-36 survey Phlebotomy & Central Lab processing Case Report Form (CRF) Completion Monitoring visit Coordinator services

75.00 80.00 75.00 20.00 10.00 100.00 30.00 20.00 50.00 20.00 20.00 20.00 30.00 75.00 150.00 35.00 150.00

75.00 80.00 75.00 20.00 10.00 30.00 20.00 50.00 20.00 20.00

75.00 150.00 35.00 150.00

40.00 40.00 20.00 10.00 100.00 30.00 20.00 50.00 20.00 20.00 20.00 30.00 75.00 150.00 35.00 150.00

20.00 10.00

20.00 10.00

20.00

30.00 20.00 50.00 20.00 20.00

20.00 10.00 100.00 30.00 20.00 50.00 20.00 20.00

30.00 20.00 50.00 20.00 20.00

75.00 150.00 35.00 150.00

75.00 150.00 35.00 150.00

30.00 75.00 150.00 35.00 150.00

30.00 75.00 75.00 35.00 75.00

20.00 20.00

hourly rate

total funding per visit divide by hourly rate

$75 $810.00 $810.00 $580.00 $580.00 $710.00 $350.00 hours 10.8 7.7 7.7 9.5 4.7 estimated 10.8 per visit

$75.00 $120.00 $115.00 $120.00 $50.00 $200.00 $150.00 $100.00 $250.00 $120.00 $120.00 $20.00 $90.00 $450.00 $825.00 $210.00 $825.00 per patient total hours 51.2 $3,840.00 51.2

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

34 Getting the Funding You Need to Support Your Research: Navigating the National Institutes of Health Peer Review Process Valerie L. Prenger National Institutes of Health, Bethesda, MD, United States

O U T L I N E Overview of National Institutes of Health Mission and Organization of National Institutes of Health Responsibilities of National Institutes of Health Staff National Institutes of Health Extramural Funding Mechanisms National Institutes of Health Funding Announcements Funding Opportunity Announcements Requests for Applications and Program Announcements in the National Institutes of Health Guide Electronic Submission of Applications Through Grants.gov Multiple Principal Investigators

590 590 591 592 592 592 593

593 594

The National Institutes of Health Peer Review Process for Grants 594 The National Institutes of Health Dual-Review System 594 National Institutes of Health Review “Cycles” 594 Assignment of Applications to a Review Group and Funding Institute 595 How Are Reviewers Selected? 595 How Does the Review Proceed? 596 Review Criteria for Research Project Grant Applications 596 Core Review Criteria 596 Additional Review Criteria 597 Additional Review Considerations 597 Research Project Grant Applications From New/ Early-Stage Investigators 598 Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00034-4

Possible Scientific Review Group Actions Overall Impact/Priority Score and Percentiles The Summary Statement Tells You What the Reviewers Thought About Your Application Review by National Advisory Councils and Boards What Determines Which Applications Are Awarded? Confidentiality and Conflict of Interest

598 599 599 599 600 600

Hints for Preparing Better Grant Applications 600 Planning Your Application 600 Allow Sufficient Time to Prepare the Application 600 Get Help 600 Follow the Instructions CloselydSubmit a Complete and Carefully Prepared Application 601 Hints and Suggestions for Preparing Each Part of Your Application 601 SF424 (R&R) Project Summary/Abstract 601 PHS 398 Specific Research Plan Component 601 PHS 398 Specific Human Subjects Sections 603 Vertebrate Animals 603 Budget and Justification 604 Senior/Key Personnel Profiles Component and Biosketches 604 Facilities and Other Resources 604 Appendix 605 Recent Changes to Application Procedures for National Institutes of HealtheFunded Clinical TrialsdMore to Come

605

Revising Unsuccessful Applications How to Decide Whether to Revise Your Application How to Revise and Resubmit Your Application

605 605 606

589

Copyright © 2018. Published by Elsevier Inc.

590

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

What if It Appears That the Study Section Was Inappropriate or Biased? What if It Appears That There Was a Procedural Error During Peer Review? National Institutes of Health Grant Programs for Clinical Researchers at Various Stages in Their Careers Individual Career Development (“K”) Awards Mentored Career Development Awards Mentored Clinical Scientist Development Award (K08) Mentored Patient-Oriented Research Career Development Award (K23) Career Transition Awards K99/R00 Pathway to Independence Award

606 606

607 607 607 607 607 607 607

Once an investigator has selected a general research topic, determined what others have already done, identified promising new research areas, planned a project, and estimated the budget required for it, she/he will need to obtain funding to perform the study. There are more opportunities than ever for clinical researchers to develop new diagnostic and therapeutic approaches to many diseases. Many public and private organizations support biomedical research and clinical trials, but most biomedical research conducted at research centers and academic institutions in the United States is funded by the National Institutes of Health (NIH). Translational and clinical research can be broader in scope than laboratory research and more difficult to plan, describe, and carry out. Most clinician scientists also receive little training in grant writing. Recent studies at NIH show that the overall success rate of applications proposing clinical research and clinical trials is somewhat lower than the success rate for applications for basic or laboratory-based research. This was not due to the review panel assignment, the composition of the review committee, the cost of the proposed research, or whether clinical applications were reviewed in the same review group as basic research applications. It is likely that the applications were just not as well prepared. Rejected grant applications can challenge the ego, cause critical delays, or result in loss of research opportunities, particularly if the research project is linked to ongoing clinical trials. A thorough understanding of the NIH peer review process will help both new and established clinical investigators be more competitive in applying for research funds. The purpose of this chapter is to provide (1) an overview of the NIH and the NIH peer review process, (2) suggestions for writing more

K22 Career Transition Awards Independent Scientist Awards Midcareer Investigator Award in Patient-Oriented Research Exploratory/Development Grant (R21) Applications Small Research Grant (R03) Applications Loan Repayment Program

How to Stay Informed About National Institutes of Health Peer Review “About Grants” Page (https://grants.nih.gov/grants/ about_grants.htm) National Institutes of Health Institute/Center Home Pages The Center for Scientific Review Home Page (www. csr.nih.gov)

608 608

608 608 609 609 609 609 609 609

competitive grant applications, (3) brief descriptions of some NIH grant programs for clinical researchers at various stages of their careers, and (4) ways to find current information about the NIH grants process when you are ready to apply.

OVERVIEW OF NATIONAL INSTITUTES OF HEALTH Mission and Organization of National Institutes of Health NIH is a federal agency that is part of the U.S. Department of Health and Human Services (DHHS). The mission of NIH is to seek fundamental knowledge about the nature and behavior of living systems and the application of that knowledge to enhance health, lengthen life, and reduce the burdens of illness and disability. NIH accomplishes this mission by conducting and supporting research, research training and career development, research infrastructure and resources, and conferences related to • the causes, diagnosis, prevention, and cures of human diseases; • the processes of human growth and development; • the biological effects of environmental contaminants; • the understanding of mental, addictive, and physical disorders; and • the collection, dissemination, and exchange of information in medicine and health. The structure of the NIH is shown in Fig. 34.1. Most NIH components have both intramural programs with laboratories and clinics staffed by NIH employees and

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

591

OVERVIEW OF NATIONAL INSTITUTES OF HEALTH

Naonal Instutes of Health Office of the Director

Naonal Instute on Aging

Naonal Instute on Alcohol Abuse and Alcoholism

Naonal Instute of Allergy and Infecous Disease

Naonal Instute of Arthris and Musculoskeletal and Skin Diseases

Naonal Cancer Instute

Naonal Instute of Child Health and Human Development

Naonal Instute on Deafness and Other Communicaon Disorders

Naonal Instute of Dental and Craniofacial Research

Naonal Instute of Diabetes and Digesve and Kidney Diseases

Naonal Instute on Drug Abuse

Naonal Instute of Environmental Health Sciences

Naonal Eye Instute

Naonal Instute of General Medical Sciences

Naonal Heart, Lung, and Blood Instute

Naonal Human Genome Research Instute

Naonal Instute of Mental Health

Naonal Instute of Neurological Disorders and Stroke

Naonal Instute of Nursing Research

Naonal Instute on Minority Health and Health Disparies

Naonal Center for Complimentary and Integrave Health

Fogarty Internaonal Center

Naonal Center for Advancing Translaonal Sciences

Naonal Library of Medicine

Naonal Instute of Biomedical Imaging and Bioengineering

Clinical Center

FIGURE 34.1

Center for Informaon Technology

Center for Scienfic Review

Institutes, centers, and offices of the National Institutes of Health.

extramural programs through which research is supported in institutions worldwide. It is important to note that although each NIH institute has specific scientific areas of primary interest, there are many areas of shared interest. For example, asthma is a shared interest of the National Institute of Allergy and Infectious Diseases (NIAID) and the National Heart, Lung, and Blood Institute (NHLBI); and the National Cancer Institute (NCI) and NIAID are both interested in transplantation biology and the life cycle of oncogenic viruses. Applicants should discuss potential research projects with program staff in all relevant NIH components before preparing a grant application. The total budget of the NIH was approximately $33.1 billion in fiscal year 2017 (Table 34.1). About 85% of the NIH budget is used to support extramural research through different types of grants and contracts.

TABLE 34.1

Responsibilities of National Institutes of Health Staff

oversees scientific aspects of the program. Grants and contracts management staff provide financial stewardship and oversight of an institute’s extramural research programs. Each NIH funding component also has a review office that manages the scientific peer review of

Each of NIH’s extramural research programs is managed by a program official, an NIH scientist who

Distribution of National Institutes of Health’s (NIH’s) $33.1 Billion Fiscal Year 2017 Budgeta % of NIH Budget*

Component Extramural research

Research project grants

55

R&D contracts

10

Research centers

8

Other research grants

9

Research training

3

Intramural research Research management and support

11 5

* Adds to more than 100% due to rounding. a https://www.hhs.gov/about/budget/fy2017/budget-in-brief/nih/index.html.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

592

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

TABLE 34.2

Roles and Responsibilities of National Institutes of Health (NIH) Extramural Staff

NIH Staff

Role and Responsibilities

Scientific Review Officer (SRO)

In Center for Scientific Review and in each NIH Institute/Center scientific review office Organizes, manages, conducts, and reports scientific peer review of grant applications and/or contract proposals Liaison between applicants and reviewers

Program Officer/Program Director (PO/PD)

In NIH institutes and centers Manages a portfolio of awarded grants/contracts Monitors scientific progress made on grants/contracts

Grants Management Officer (GMO)/Contracts Management Officer (CMO)

In NIH institutes and centers Fiscal stewardship of portfolio of awarded grants/contracts Negotiates fiscal aspects of awards Monitors financial progress made on grants/contracts

contract proposals and highly mission-oriented grant applications. Table 34.2 summarizes the responsibilities of NIH review, program, and grants and contracts management staff.

approaches for the project. The grantee files a yearly progress report, but NIH staff is not involved in carrying out the work. Cooperative agreements are similar to grants in that the purpose is to assist and support research or related activities. However, they include a substantial scientific and/or technical role by NIH staff, such as coordination of awardee activities or approval of phases or processes in the project. Cooperative agreement applications usually are solicited via a specific Request for Applications (RFA), which describes the activities that will be supported as well as NIH staff involvement. In contrast, NIH uses contracts to purchase a service or product, and the awarding NIH component establishes the plans, parameters, and detailed requirements for projects supported by contracts. Contract proposals are almost always solicited through specific requests for proposals that are published in Federal Business Opportunities (www.fbo.gov). Contracts have special submission and review processes, review criteria, mechanisms for reimbursement of costs, involvement of the funding institute, and delivery of the end product. Because most NIH support of extramural research is via grants and cooperative agreements, this chapter does not address contracts. For more information about NIH support of research and development contracts and contract opportunities, see https://oalm.od.nih.gov/.

National Institutes of Health Funding Announcements Funding Opportunity Announcements

National Institutes of Health Extramural Funding Mechanisms NIH uses three types of funding mechanisms to support extramural research and development: grants, cooperative agreements, and contracts. The relationship between NIH and the awardees in each of these funding mechanisms is summarized in Table 34.3. Most NIH grants are “investigator initiated”dwhich means that the principal investigator (PI) is responsible for developing the ideas, concepts, methods, and

TABLE 34.3

National Institutes of Health (NIH) Extramural Award Mechanisms

Award Mechanism

NIH Role

NIH Provides

Grant

Patron

Assistance, encouragement

Cooperative agreement

Partner

Assistance, with substantial program staff involvement

Contract

Purchaser

Direction

All NIH grant applications are submitted electronically on the Standard Form 424 Research and Related (R&R) application form either through the federal Web portal Grants.gov or through the use of the Application Submission System and Interface for Submission Tracking (ASSIST). The use of ASSIST is suggested when submitting complex multicomponent applications. The Grants.gov portal allows you to search all federal grant programs through the “Find” option and to apply for grants through the “Apply” option. Applicant institutions also must register their project directors/principal investigators (PD/PIs) in the NIH electronic research administration (eRA) Commons. Registration volume peaks close to submission dates, so start the registration process at least 4 weeks before your target submission date. All applicationsdincluding those that are “investigator initiated”dmust be submitted in response to an open NIH Funding Opportunity Announcement (FOA). The NIH has posted generic (or “parent”) FOAs in Grants.gov for the most common types of grants. Specific FOAs are published for each RFA and

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

OVERVIEW OF NATIONAL INSTITUTES OF HEALTH

Program Announcement (PA) and for NIH institute/ center-specific grant mechanisms. Each FOA has an opening date and a submission deadline. Many FOAs are active for several years and include multiple submission/receipt dates. Applicants should read the entire FOA carefully to determine the specific submission/receipt deadlines. Applications submitted after the submission deadline may be rejected or may be held over to the next submission date. Requests for Applications and Program Announcements in the National Institutes of Health Guide The NIH may invite submission of grant applications to address areas of special interest to an awarding institute by issuing an RFA or PA in the NIH Guide for Grants and Contracts (https://grants.nih.gov/funding/ index.htm) with a link to the correct FOA in Grants.gov. Table 34.4 summarizes the key features of RFAs and PAs. RFAs generally have specific funds set aside to support research, training, or infrastructure on a given topic. RFAs also list NIH staff contacts, and it is a good idea to discuss your potential project with them (prior to submission) to ensure that you meet all responsiveness and eligibility criteria. Peer review of applications for an RFA usually is managed by the peer review office in the TABLE 34.4

Key Features of National Institutes of Health (NIH) Program Announcements and Requests for Applications

Program Announcement (PA)

Addresses a relatively broad field/category of research Usually no set-aside budget Usually submit on regular receipt dates Regular review criteria for type of applications requested Frequently more than one NIH component involved Applications may be reviewed by CSR or the issuing NIH component

Request for applications (RFA)

Addresses a well-defined area of research Set-aside budget Submit on special, usually one-time-only receipt dates Often special eligibility and/or review criteria Often special application format and/or submission instructions Usually only one NIH component involved Applications usually reviewed by the issuing NIH component

593

issuing NIH component. In contrast, a PA usually indicates NIH’s interest in supporting research in a broad area without specific set-aside funds. Applications responding to PAs usually are reviewed with other “investigator-initiated” applications on similar topics through the usual channels in study sections organized by the NIH Center for Scientific Review (CSR). Both RFAs and PAs may have special eligibility requirements, application preparation procedures, receipt dates, and/ or conditions of award, so it is important to read the announcement carefully before preparing an application.

Electronic Submission of Applications Through Grants.gov The electronic grant submission process through Grants.gov involves several steps that must be completed in sequence. 1. Search for and identify an open FOA on Grants.gov. 2. Download and complete the specific grant application package for that FOA. Some parts of the application (i.e., investigator and budget information) require entering information into specific templates, and other parts (Biosketches, Research Plan, Bibliography) require Portable Document Format (PDF) attachments with narratives prepared by the applicant. 3. An authorized official at the applicant organization submits the completed application package through Grants.gov. (Be sure to save a local copy!) 4. NIH eRA software retrieves the application from Grants.gov and checks the application against NIH’s business rules. 5. NIH notifies both the PI and signing official (SO) for the applicant institution by email to check the NIH eRA Commons for results of the NIH validations check. If the application passed NIH validations checks, there will be a grant image in the eRA Commons. If the image is not accurate, the PI and SO may reject the application in the eRA Commons, make the necessary changes, and submit the changed/corrected application again via Grants.gov. However, submission of the changed/corrected application must be completed by the deadline stated in the FOAdthere is no “grace period” for correction of applicant errors! If the application failed NIH validation checks, the eRA Commons will show errors and warnings. The PI and SO must fix the errors and submit the entire corrected application again through Grants.gov by the deadline in the FOA. 6. eRA Commons saves the grant image, and NIH begins processing the application.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

594

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

Each NIH grant application package in Grants.gov will include both SF424 (R&R) components and NIHspecific PHS 398 components as well as an application guide with complete instructions for submitting an application. The NIH logo indicates fields in the SF424 (R&R) components that are not mandatory on the federal-wide form but are required by NIH. The NIH eRA website (https://grants.nih.gov/ grants/how-to-apply-application-guide.htm) has a number of resources for applicants, including the following: • SF424 (R&R) application guides, sample application packages, and related resources • eRA Commons registration training • Steps for writing an application • How to submit, track, and view your grant application • What happens to your grant after it is submitted

Multiple Principal Investigators For most types of grant applications, NIH allows applicants and their institutions to identify more than one PD/PI. The multiple PI option is to encourage interdisciplinary and team approaches to biomedical research. Each FOA will indicate if the multiple PD/PI option is available for that type of application. The multiple PD/PI option is not appropriate for individual career awards, individual fellowships, Dissertation Grants, Director’s Pioneer Awards, and Shared Instrumentation Grants. Investigators and the applicant organization are responsible for deciding whether to apply for a single or a multiple PI grant based on the scientific goals of the project. See the Multiple PI website (https://grants.nih.gov/grants/multi_pi/) for more information. Applications with multiple PIs must include a leadership plan that describes the rationale for choosing a multiple PD/PI option and the governance and organizational structure of the leadership team and the research project, including the roles and responsibilities of each PD/PI and plans for communication, decisionmaking, and resolving conflicts.

The National Institutes of Health Dual-Review System The cornerstone of the NIH grants process is the “dual-review system,” with two sequential levels of review that separate scientific assessment of the projects from policy decisions about scientific areas to be supported and the resources to be allocated (Fig. 34.2). The first level of review is by panels of experts organized by scientific discipline or research area to evaluate the scientific and/or technical merit of the applications. These scientific review groups (SRGs) also are commonly called “study sections.” Each SRG is managed by a scientific review officer (SRO), an NIH scientist with expertise in the SRG’s area of science. The second level of review is by an NIH institute’s or center’s national advisory board or council, which is composed of both scientific and public representatives noted for their expertise or interest in matters related to the mission of the institute. Council recommendations are based on scientific merit as judged by the SRGs plus relevance to an institute’s mission and programs.

National Institutes of Health Review “Cycles” The NIH receives 70,000 to 80,000 applications per year for about 200 different types of grants. To handle this load, each type of grant application has a designated receipt date(s) indicated in the FOA, with three receipt dates per year for most types of applications (https:// grants.nih.gov/grants/how-to-apply-application-guide/ due-dates-and-submission-policies/due-dates.htm). RFAs and some PAs have special receipt dates. Table 34.5 shows the three overlapping review cycles for grant applications that result from the standard NIH receipt dates. The review cycle for a grant application begins when an investigator submits an application to NIH and concludes when the applicant organization and the PI are notified of the recommendation of the council (Fig. 34.3).

THE NATIONAL INSTITUTES OF HEALTH PEER REVIEW PROCESS FOR GRANTS NIH draws on the national pool of scientists actively engaged in research to assist in evaluating the tens of thousands of grant applications received annually. These scientific “peers” advise NIH which applications are most likely to have a high impact in each field.

FIGURE 34.2 Schematic of the dual-review system employed by the National Institutes of Health.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

595

THE NATIONAL INSTITUTES OF HEALTH PEER REVIEW PROCESS FOR GRANTS

TABLE 34.5

National Institutes of Health Grant Application Receipt, Review, and Award Cycles

Application Receipt Dates

Scientific Review Group Meetings

National Advisory Council Meetings

Earliest Possible Award Date

January 25eMay 7

JuneeJuly

AugusteOctober

September 30 or December 1

May 25eSeptember 7

OctobereNovember

JanuaryeFebruary

April 1

September 25eJanuary 7

FebruaryeMarch

MayeJune

July 1

FIGURE 34.3 The review cycle for a research grant application submitted to the National Institutes of Health.

Assignment of Applications to a Review Group and Funding Institute All grant applications are processed centrally by the NIH CSR Division of Receipt and Referral, which determines if the application is appropriate for NIH and where it should be reviewed. Based on the type of application and the mission areas of the various NIH components, an application is assigned to a potential awarding NIH component and to either a CSR study section or an institute/center (IC) review committee for scientific merit review. The rosters and scientific areas of the CSR study sections and links to other review committees are available at https://public.csr.nih.gov/StudySections/ Pages/default.aspx. Applicants are notified via the NIH eRA Commons (https://public.era.nih.gov/commons/ public/login) of the review group and the SRO within a few weeks after receipt of the application. The PD/PI of an application may provide suggestions about appropriate review groups and/or scientific expertise areas needed to evaluate the application in the cover letter attachment that accompanies the application. The SRO may invite additional temporary members to serve as reviewers if specialized expertise is required to review an application. If the research objectives and approaches of an application cannot be reviewed appropriately by an existing SRG, a Special Emphasis Panel may be constituted for the review.

Assignment of an application to an NIH funding component is based on the institute’s congressionally mandated program responsibilities. If the subject matter of an application relates to two or more institutes, a dual, or multiple, assignment may be made. The CSR has no responsibility for either decisions about funding or the management of funded grant programs.

How Are Reviewers Selected? The primary requirement for serving on an SRG is demonstrated achievement as an independent investigator. Service also requires mature judgment, balanced perspective, objectivity, ability to work effectively in a group, and a commitment to complete review assignments, maintain confidentiality of applications and discussions, and avoid conflicts of interest. NIH also considers geographic distribution and representation of ethnic minority and female scientists in the selection of SRG members. Members usually are appointed to an SRG for 4 years, with staggered terms, so about one-fourth of the members of each SRG change each year. Several NIH institutes also include patient representatives or advocates on SRGs reviewing clinical research applications, especially those involving clinical trials. Patient advocates have expertise in the impact of the disease on patients and caregivers, strategies and approaches

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

596

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

likely to succeed in patient recruitment and follow-up, and quality of life issues. Patient advocates are full voting members of the review panels.

How Does the Review Proceed? Standing SRGs (“study sections”) normally meet three times a year for 1e3 days each time, depending on the number and types of grant applications to be reviewed. Typically, a CSR study section is responsible for the review of 60e100 research project grant applications at each meeting. The SRO assigns each application to three or more members of the SRG, who provide detailed written reviews listing specific strengths and weaknesses of the application related to each of the review criteria. Each member may be assigned to prepare detailed critiques for 5e10 applications and as a discussant (reader) on additional applications. Reviewers receive the applications and instructions for preparing their reviews 4e 6 weeks before the SRG meeting. The SRO is the designated federal official in charge of the meeting and handles all communications with applicants and reviewers during the review process. The SRO and the chairperson, one of the members of the SRG, conduct the meeting. The chairperson calls on the assigned reviewers and discussants to present their critiques and then moderates a discussion among all SRG members. Other members of the SRG question assigned reviewers about the application or their critiques. SRG members score each application on the basis of their own assessment of its strengths and weaknesses, scientific merit, and potential impact on the field. To use the time at the review meeting most effectively, many SRGs use a streamlined review process in which only the top half of all applications to be reviewed are discussed at the SRG meeting; the rest of the applications are not discussed. Applications may be discussed in the order of the preliminary overall impact scores provided by the assigned reviewers.

Review Criteria for Research Project Grant Applications In 2016, NIH updated the instructions and review criteria for research grants and mentored career development award applications to emphasize the four elements of rigor and transparency (https://grants. nih.gov/reproducibility/index.htm). The four elements are as follows: • Scientific premisedkey data introduced by the applicant to justify the project; • Scientific rigordstrict application of the scientific method to ensure robust and unbiased experimental

design, methodology, analysis interpretation, and reporting of results; • Consideration of relevant biological variables, such as sexdrefers to critical biological factors affecting health and disease in studies involving human subjects or vertebrate animals; • Authentication of key biological and/or chemical resourcesd resources that may differ from lab to lab or over time, could influence the research data, and are integral to the proposed research (i.e., cell lines, specialty chemicals, antibodies, and other biologics, not standard laboratory reagents). Core Review Criteria Reviewers consider each of the following five core review criteria in their determination of the scientific merit of the project, and assigned reviewers give a separate score for each. 1. Significance: Does the project address an important problem or a critical barrier to progress in the field? Is there a strong scientific premise for the project? If the aims of the project are achieved, how will scientific knowledge, technical capability, and/or clinical practice be improved? How will successful completion of the aims change the concepts, methods, technologies, treatments, services, or preventative interventions that drive this field? 2. Investigators: Are the PD(s)/PI(s), collaborators, and other researchers well suited to the project? If EarlyStage Investigators (ESIs)or New Investigators, or in the early stages of independent careers, do they have appropriate experience and training? If established, have they demonstrated an ongoing record of accomplishments that have advanced their field(s)? If the project is collaborative or multi-PD/PI, do the investigators have complementary and integrated expertise; are their leadership approach, governance and organizational structure appropriate for the project? 3. Innovation: Does the application challenge and seek to shift current research or clinical practice paradigms by utilizing novel theoretical concepts, approaches or methodologies, instrumentation, or interventions? Are the concepts, approaches or methodologies, instrumentation, or interventions novel to one field of research or novel in a broad sense? Is a refinement, improvement, or new application of theoretical concepts, approaches or methodologies, instrumentation, or interventions proposed? 4. Approach: Are the overall strategy, methodology, and analyses well-reasoned and appropriate to accomplish the specific aims of the project? Have the investigators presented strategies to ensure a robust

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

THE NATIONAL INSTITUTES OF HEALTH PEER REVIEW PROCESS FOR GRANTS

and unbiased approach, as appropriate for the work proposed? Are potential problems, alternative strategies, and benchmarks for success presented? If the project is in the early stages of development, will the strategy establish feasibility and will particularly risky aspects be managed? Have the investigators presented adequate plans to address relevant biological variables, such as sex, for studies in vertebrate animals or human subjects? If the project involves human subjects and/or NIH-defined clinical research, are the plans to address (1) the protection of human subjects from research risks and (2) inclusion (or exclusion) of individuals on the basis of sex/ gender, race, and ethnicity, as well as the inclusion or exclusion of children, justified in terms of the scientific goals and research strategy proposed? 5. Environment: Will the scientific environment in which the work will be done contribute to the probability of success? Are the institutional support, equipment, and other physical resources available to the investigators adequate for the project proposed? Will the project benefit from unique features of the scientific environment, subject populations, or collaborative arrangements? Some RFAs and PAs may list additional review criteria and/or additional elements under each of the standard core review criteria that relate to the specific requirement(s) of the program announced in the FOA. Additional Review Criteria The following items also are considered in the determination of scientific merit and the impact/priority score for research project grant applications but do not receive individual scores. 1. Protection of human subjects from research risk: Most clinical research projects will involve human subjects, either living persons with whom you will interact directly or identifiable specimens from them. The reviewers evaluate the proposed use of human subjects, the risks to the subjects, the plans to protect them from risks, whether the risks are reasonable in relation to the anticipated benefits to the subjects, and the importance of the knowledge that may result from the research. For projects that claim an exemption, the reviewers evaluate the justification for the exemption. Serious deficiencies in the plans to protect human subjects may be considered weaknesses under the “approach” review criterion. In addition, there is a bar to award until all reviewer concerns about protection of human subjects have been resolved. 2. Inclusion of women, minorities, and children in research: Reviewers evaluate the proposed plans for

3.

4.

5. 6.

7.

597

the inclusion of minorities and members of both genders, as well as the inclusion of children. Serious deficiencies in the plans to include women, minorities and/or children will be considered weaknesses under the “approach” review criterion and may result in a bar to award. Vertebrate animals: Reviewers evaluate the proposed use of live vertebrate animals, including the strains, ages, sex, and number of animals; justification for using animals; adequacy of veterinary care; procedures for limiting distress, pain, and injury; and methods for euthanasia. Serious deficiencies in the plans to protect vertebrate animals will be considered weaknesses under the “approach” review criterion and may result in a bar to award. Resubmission applications: Reviewers evaluate the application as now presented, considering the responses to the previous review and changes to the project since the previous submission. Renewal applications: The reviewers consider progress made in the last funding period. Revision applications (formerly called competing supplement applications): The reviewers consider the appropriateness of the proposed expansion of the scope of the project. Biohazards: Reviewers assess whether materials or procedures proposed are potentially hazardous to research personnel and/or the environment, and, if so, whether adequate protection is proposed. An award cannot be made until all concerns about hazardous conditions have been resolved. Serious deficiencies in the plans to protect against biohazards may be considered weaknesses under the “approach” review criterion.

Additional Review Considerations As applicable for the proposed project, reviewers address each of the following items but do not give scores for them and do not consider them in the overall impact/priority score. 1. Applications from foreign organizations: If the application is from an institution outside of the United States or involves a substantial foreign component, the reviewers comment on whether the research could be done in the United States or if the foreign location offers specific advantages, such as resources, equipment, or study populations that are not available in the United States. 2. Select agent research: Federal policy requires applicants proposing research with certain biological agents and toxins, termed “select agents,” that have the potential to pose a severe threat to public, animal, or plant health; to disclose the proposed use in NIH

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

598

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

grant applications; and for reviewers to evaluate the plans to handle and contain such agents (see https:// www.selectagents.gov/ and https://grants.nih.gov/ grants/policy/select_agent/ for policies related to grants involving select agents). Although this item is not scored, serious deficiencies in the experience of the research team or the plans for handling such agents may be addressed under the “investigators,” “approach,” or “environment” review criteria. 3. Resource Sharing Plans: NIH expects awardees to share research data and to make the results of the projects that it funds available to the public. Reviewers comment on whether the following Resource Sharing Plans are reasonable: a. Data Sharing Plan. Applications requesting more than $500,000 direct costs in any year must include a data sharing plan (https://grants.nih.gov/ grants/policy/data_sharing/data_sharing_ guidance.htm). Some RFAs and PAs may request a data sharing plan for all applications regardless of costs. b. Sharing Model Organisms. All NIH grant applications are expected to include a plan for sharing and distributing unique model organism research resources generated using NIH funding or to state why sharing is restricted or not possible (https://grants.nih.gov/grants/guide/noticefiles/NOT-OD-04-042.html). Model organisms include both mammalian models, such as mice and rats, and nonmammalian models, such as yeast, amoebae, worms, flies, zebra fish, etc. Investigators may request funds to defray costs associated with sharing. For some special initiatives and grant programs specifically directed toward the development of model organisms, reviewers’ evaluation of the plan for sharing model organisms may factor into the overall score. c. Genomic Data Sharing Plan (GDS). All applications that include large-scale human and nonhuman genomic data generated from NIHfunded research are expected to include a plan for submission of genomic data to the NIH-designated data repository or provide an appropriate explanation for why submission to the repository will not be possible (https://grants.nih.gov/ grants/guide/notice-files/NOT-OD-08-013.html). 4. Authentication of key biological resources: Reviewers evaluate the plans for identifying and ensuring the validity of key biological and/or chemical resources. 5. Budget and period of support: Reviewers evaluate whether the requested budget and period of support are adequately justified and reasonable for the proposed research.

Research Project Grant Applications From New/ Early-Stage Investigators New Investigators and ESIsda new investigator who is within 10 years of completing his/her terminal research degree or 10 years of completing medical residencydare encouraged to apply for traditional NIH research project grant (R01) applications to accelerate their transition to an independent scientific career. Wherever possible, the NIH will cluster applications from New Investigators and ESIs for discussion during initial peer review. When reviewing applications from New Investigators and ESIs, reviewers consider the experience of and the resources available to the investigator and apply the five review criteria in a manner appropriate to the expectations for, and problems likely to be faced by, a New Investigator. Specifically: • Investigator: More emphasis is placed on the applicant’s training and research potential than on his or her track record and number of publications. • Approach: More emphasis is placed on demonstrating that the techniques and approaches proposed are feasible than on the presentation of preliminary results. • Environment: Reviewers look for evidence of institutional commitment in terms of space and time to perform the research. NIH also has special receipt dates that allow resubmission of R01 applications from New/ESIs for the very next review meeting, saving approximately 4.5 months in the overall resubmission process. Consultation with NIH program staff and senior colleagues about the specific weaknesses cited in the summary statement, and whether they are amenable to a “quick fix,” is very important. For more information on New Investigators and ESIs, see https://grants.nih.gov/grants/guide/notice-files/ NOT-OD-09-013.html and https://grants.nih.gov/ grants/new_investigators/.

Possible Scientific Review Group Actions SRGs have several options for each application that is discussed. • Score: If the SRG members have sufficient information to make a final recommendation about the application, they will score the application. • Deferral: In the rare instance that an SRG cannot make a recommendation without additional information, it may defer the application to the next review cycle. The SRO will contact the applicant to obtain the necessary information, or, if the information can be obtained only by discussion with the applicant or by direct observation, a telephone conference with the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

THE NATIONAL INSTITUTES OF HEALTH PEER REVIEW PROCESS FOR GRANTS

applicant or a project site visit may be scheduled. Deferred applications are usually reviewed again by the same SRG during the next review cycle. Deferral usually is not an option for applications received in response to a one-time RFA. • Not recommended for further consideration: In reviews that are not streamlined, applications may be “not recommended for further consideration” if they lack significant and substantial merit or if they involve procedures that are gravely hazardous or pose very serious risks to human subjects or vertebrate animals. The advisory councils do not consider these applications and they cannot be awarded.

Overall Impact/Priority Score and Percentiles Reviewers provide an overall impact/priority score to reflect the likelihood that the project will exert a sustained, powerful influence on the research field(s) involved. Each of the review criteria is considered in assigning the overall impact/priority score, with reviewers weighting them as appropriate for each application. An application does not need to be strong in all categories to be judged likely to have a major scientific impactdfor example, a project that is not innovative may be essential to advance a field. However, it is unlikely that projects with low significance or serious problems in the approach will have a high impact. Reviewers provide a one-paragraph overall impact evaluation summarizing the factors that led to the impact score. Each reviewer who is not in conflict with an application assigns an impact/priority score ranging from 1 (exceptional) to 9 (poor). A score of 1 indicates an application with essentially no weaknesses that will have an exceptionally high impact, whereas a score of 9 indicates an application with serious and substantive weaknesses, very few strengths, and little likelihood of making an impact. Reviewers consider not only the number of strengths and weaknesses but also their relative importance; for example, a major strength may outweigh many minor weaknesses (See https://grants.nih.gov/ grants/peer/guidelines_general/scoring_system_and_ procedure.pdf.) After the meeting, the individual reviewers’ numeric ratings for each scored application are averaged and multiplied by 10 to provide the final two-digit impact/ priority score. The best possible score is 10 and the worst is 90. In addition to the impact/priority score, percentile ranks also are calculated for most research project grant (R01) applications. The percentile represents the relative position or rank of a priority score on a 100 percentile scale; the 1st percentile is the best and the 100th is the worst. Percentiles are calculated using a reference base

599

of R01 applications reviewed by a study section at three consecutive meetings. Percentile ranking is currently the primary factor used by most NIH institutes in deciding which applications to fund, although each may set a different percentile “payline” for applications.

The Summary Statement Tells You What the Reviewers Thought About Your Application Immediately after the SRG meeting, the SRO prepares a summary statement for each application that documents the deliberations of the SRG and is the official record of the review. Summary statements generally include the assigned reviewers’ essentially unedited written comments (prepared by reviewers before the peer review meeting) and the scores they gave for each of the five core review criteria. For applications that were discussed, the SRO also prepares a “Resume and Summary of Discussion” to convey highlights of the discussion that occurred during the peer review meeting (i.e., major strengths and weaknesses of the application). The Resume and Summary of Discussion should explain how the SRG arrived at the final score of the grant. Summary statements for scored applications also may include budget recommendations and administrative notes with reviewer concerns about research involving human subjects or animals or potential overlap with other ongoing projects. Applicants should expect to be able to access the summary statement through the NIH eRA Commons within 4 to 6 weeks after the review meeting. Summary statements for applications from New/ESIs are generally available within 2 weeks after the review meeting, since there are special deadlines for resubmission of amended applications for New Investigators and ESIs.

Review by National Advisory Councils and Boards The second-level review for grant applications is by each institute/center’s advisory council or board, which assesses the quality of the scientific merit review by the SRG, considers the relevance of the proposed research to the institute’s programs and priorities, and advises the institute/center on policy issues. Generally, councils review only scored applications. Most types of NIH grants cannot be awarded without consideration by a council/board. Advisory council members use summary statements as the main source of information about applications. For most applications, councils concur with the recommendation of the SRG. If the council disagrees with an SRG recommendation, it may recommend rereview of the application. In addition, the council may advise the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

600

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

institute, based on the relevance of the project to the institute’s mission, that an application should receive more or less favorable consideration for funding than that would be indicated by the impact/priority score and/or percentile rating.

What Determines Which Applications Are Awarded? Awards are made based on the scientific and technical merit of the application, as reflected by its impact/priority score and/or percentile rating, the relevance of the application to the mission and programs of the institute, and the availability of funds. Each NIH institute/center generally sets a payline for each of the different types of applications. Paylines may differ considerably among the NIH institutes/centers, depending on their overall budgets, their portfolio of award mechanisms, and the advice of their advisory councils about portfolio balance. Paylines also may differ from program to program within an institute/center.

HINTS FOR PREPARING BETTER GRANT APPLICATIONS After you have decided on your research area and designed a specific project, the most important element in increasing the chances that you will be successful in getting funding for your project is a well-prepared grant application. The reviewers assigned to your application will be scientists working in the general area of your research project. Consider them “informed strangers.” The application is your marketing tooldit must convey a large amount of information while generating excitement about the project. After reading the application, the reviewers must understand the rationale for and objectives of the project, see where it fits in the “big picture” and why it is important, and feel confident that you can actually design, carry out, and interpret the proposed experiments to have an impact on the field. A well-prepared application leads the reviewers through the project logically and says much about you as the PD/PI, particularly that you “think like a scientist.” Therefore, the process for preparing an NIH application requires a significant amount of time, a high level of organization, and attention to detail.

Confidentiality and Conflict of Interest Confidentiality of review proceedings is essential to maintain the integrity of the peer review system. Under no circumstances may reviewers discuss the review proceedings outside of the SRG meeting or advise applicants or others of SRG recommendations. The SRO in charge of the SRG handles all inquiries from applicants and reviewers. In addition, reviewers may not solicit opinions or reviews from experts outside the SRG. Review materials and the proceedings of review meetings are privileged communications for use only by reviewers and NIH staff. Conflict of interest in scientific peer review occurs when a reviewer has a personal or financial interest in an application or when an application involves a close relative or professional associate of the reviewer, such as a collaborator on any research project. The SRO for the review identifies conflicts of interest among the reviewers before the review, and reviewers sign a certification before the review meeting stating that they will not participate in the discussion of any application with which they are in conflict. At the beginning of each SRG meeting, the SRO again explains the NIH confidentiality and conflict of interest policies, and review staff keep a record of which members leave the room during the meeting because of conflicts of interest. At the end of the meeting, reviewers sign a second confidentiality and conflict of interest certification, leave all review materials with the SRO, and erase any electronic review materials they have received.

Planning Your Application Allow Sufficient Time to Prepare the Application A first-time applicant for a traditional “R01” research grant should allow at least 3 months to prepare and finalize the application. Establish deadlines for the preparation of each part of the application, particularly when collaborating investigators are involved. Be aware of administrative deadlines for sign off within your institution, and be sure to leave enough time to correct validation errors through Grants.gov before the submission date, if necessary. Prepare a draft of the application early enough that objective experts (e.g., successful grantees, a mentor, or an institutional panel) can review it and provide extremely frank feedback and suggestions for revisiondfriends and close associates are rarely as critical as reviewers on an NIH study section. Get Help Find someone in your institution to help you understand the NIH grant application process and forms. Incomplete applications may not pass eRA validation checks or may be returned without review. Ask colleagues for copies of successful recent NIH grant applications to get a more concrete idea of what each section should include. NIH recently updated its application guides and supplemental instructions based on applicant feedback (https://grants.nih.gov/grants/

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

HINTS FOR PREPARING BETTER GRANT APPLICATIONS

guide/notice-files/NOT-OD-17-023.html), so be sure to look at an application in the new format! Talk to program representatives from the NIH institutes with interests in your research area about whether your project falls within the scope of an existing RFA or PA or an area of special program emphasis.

Follow the Instructions CloselydSubmit a Complete and Carefully Prepared Application Before you begin writing your application, download the application package linked to the FOA of interest and read the SF424 (R&R) instructions carefully to become familiar with all the requirements and certifications necessary. If you are submitting your application in response to a specific PA or RFA, read the announcement in detail for special eligibility requirements, formatting instructions, and/or submission deadlines. It is important that you download the complete application package from the FOA each time you applyddo not use previous versions. NIH frequently updates policies, procedures, and application requirements, so check the NIH website and do not rely on “hearsay”. Your application must be complete for review as submitted. If several people are contributing to writing the application, decide who will do the final editing. Reread your application. Have someone else read it. Proofread it again before submission. You, and only you, are responsible for making sure your application is written with good grammar, that the flow of experiments is clear, that the references and figure legends are accurate, and that information is organized and displayed legibly and in the correct order. Once the submission deadline has passed, NIH accepts update materials (a.k.a., postsubmission materials), such as revised budgets or additional information, only in a limited number of specific circumstances and cannot “change pages” to correct mistakes in the application. See https://grants.nih.gov/grants/guide/notice-files/ NOT-OD-16-130.html for examples of acceptable and unacceptable postsubmission materials.

Hints and Suggestions for Preparing Each Part of Your Application This section should be used in conjunction with the Grants.gov Application Guide for the SF424 (R&R) forms package. The items discussed here are important parts of the application on which reviewers focus; many first-time applicants have problems with them. (Note that they are not listed here in the order in which they are attached to the SF424 (R&R) application).

601

SF424 (R&R) Project Summary/Abstract The purpose of the Project Summary/Abstract is to convey succinctly the major aspects of the proposed project. It is used in the application referral process to determine what study section is most appropriate to review the application and to what NIH funding component(s) is most relevant. Members of the review committee who are not primary reviewers may rely heavily on the Project Summary/Abstract to understand your project. If an award is made, the Project Summary/ Abstract will be available to the public, so do not include any proprietary information. • View the Project Summary/Abstract as your onepage advertisement. • Be complete but concise. Include the rationale for the project and its potential impact on the field. Summarize the specific aims and hypotheses, the short- and long-term objectives, the unique features of the project, the types of methods (i.e., genetic, immunologic, genomic, proteomic, population surveys, etc.) you will use, and the expected results and how you will evaluate them. • Do not exceed the space allotted or your application may fail validation in Grants.gov. • Write the Project Summary/Abstract last so that it reflects the entire project. PHS 398 Specific Research Plan Component Table 34.6 summarizes the key features of the PHS 398 Specific Research Plan component in successful grant applications. This is the most important part of the application and will largely determine whether the application receives a highdand potentially fundabled impact/priority score. The Research Plan includes the Specific Aims and Research Strategy of the project. The Research Plan as a whole should answer the following questions: • Why is this work important and what impact will it have on this and related fields? • How is the work innovative? • How will the research be accomplished and the results be analyzed? • What have you and/or others done to establish the feasibility of what you are proposing? A strong and clear Research Plan establishes your credibility as the PI. One person should revise and edit the final draft. Make sure that all sections are internally consistent and support each other. The thought processes behind the project should be clear. Emphasize how the combination of a strong hypothesis, important preliminary data, and a new experimental system and/or approach will enable important progress to be made. Emphasize biological mechanisms in your

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

602

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

TABLE 34.6

Key Features of Successful Research Grant Applications

Hypothesis

• Make sure that all aims are related to the stated hypothesis or problem. • Focus on aims where you have good supporting preliminary data and scientific expertise.

A meaningful hypothesis and a means of testing it A sound rationale for the hypothesis A set of related aims focused on the hypothesis Preliminary data Shows proper training for the research proposed and the ability to interpret results Include alternative interpretations of results and address limitations of methods Well-organized research plan Aims focused, not diffuse Clear experimental plan, with rationale for methods chosen, criteria for proceeding from one aim to the next, alternatives if experiments do not work Research priorities clearly indicated Emphasize mechanismdavoid “descriptive data gathering” Clear plans for data analysis, with alternative interpretations addressed Access to key reagents, patients, specimens, facilities, etc., well documented

hypotheses, experiments, and interpretation of results as much as possible. Use a numbering system and/or subheadings to lead reviewers through the Research Plan. Use diagrams for complex processes, relationships, or organizational schemes. Specific Aims The Specific Aims should concisely state the goals of the proposed research and summarize the expected outcome(s) including the impact that the results could have on the field. The Specific Aims section may not exceed one page. Succinctly list broad, long-term goals (e.g., the hypothesis to be tested) and provide a list of specific time-phased research objectives. • State the hypotheses or problem(s) to be solved clearly and definitively. Make sure the hypotheses are understandable and testable within the proposed time frame. • Be brief and specific. State each aim in one sentence. Use a brief paragraph under each aim if more detail is needed. • Do not “bite off more than you can chew.” A small, focused project with a feasible timetable is generally better than a diffuse, multifaceted “fishing expedition.” Most successful applications have two to four specific aims.

Research Strategy The Research Strategy section describes how the research actually will be carried out and is crucial to how favorably an application is reviewed. This section is organized to address the Significance, Innovation, and Approach review criteria (listed previously in “Core Review Criteria”) and is limited to 12 pages for an R01 application. The page limits for other types of applications are listed at https://grants.nih.gov/ grants/how-to-apply-application-guide/format-andwrite/page-limits.htm. In the Significance section, you must make a compelling case for your proposed research project and excite the reviewers about your project. Explain why your objectives and specific research questions are important and what barriers to progress in the field the project addresses, given the current state of the art in the field. How will the results advance the field? In the Innovation section, show how your project challenges or seeks to shift current paradigms in the field or has novel concepts, approaches, or methods. In the Approach section, describe the overall strategy, methodology, and analyses that you will use to accomplish the specific aims. Emphasize the rationale for selecting the proposed methods and approaches, how the data will be collected and analyzed, and the criteria for moving from one aim to the next. Discuss alternative approaches that will be pursued if the proposed approach is not successful. Include a projected sequence or timeline for the project. Address precautions to be used for work that may be hazardous to project personnel or human subjects. Explain how relevant biological variables, such as sex, are factored into your research design and analysis when studying vertebrate animals and humans. The approach for clinical projects should be developed with input from a statistician. Be sure to address statistical issues in study design and data analysis, with appropriate power calculations. Include data from any preliminary studies you have done, as well as citations to published reports that establish the feasibility of the proposed work in the Research Strategy section. • Number the sections to correspond to the numbered Specific Aims. Be sure to explain how the results from the experiments proposed address your stated hypotheses! • Use active voice and be specific. For example, it is not sufficient to state, “A variety of viruses will be grown in cells in standard systems.” The reviewers will want

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

HINTS FOR PREPARING BETTER GRANT APPLICATIONS

•

• • •

to know which viruses, cells, and systems; your rationale for selecting them over others; how they will be used; and if you have ever done work like this previously. You will not have room for all of the experimental details within the 12-page limit. Cite publications for standard methods and explain new or unusual methods in more detail. Be realistic about how much work can be done with the time and resources proposed. Use diagrams or flow charts to convey the flow of experiments, samples, and/or data. Include specific, definitive letters of collaboration for critical patients, populations, specimens, equipment, or reagents that will be provided by others.

PHS 398 Specific Human Subjects Sections You must include these sections if you answer “yes” to the question “Are human subjects involved?” on the SF424 (R&R) Other Project Information formdfailure to do so will result in validation errors in Grants.gov. Applications that do not adequately address research on human subjects may be returned without review, or the review or award may be delayed. Therefore, before preparing an application, you should review the instructions in the SF424 (R&R) Supplemental Grant Application Instructions (https://grants.nih. gov/grants/how-to-apply-application-guide/formsd/supplemental-instructions-forms-d.pdf) and the Office for Human Research Protections’ (OHRP) website (https://www.hhs.gov/ohrp/) for guidance on and decision charts for research involving human subjects. There is no page limit for these sections; be thorough but succinct. You may not use these sections to get around the page limits in the Research Strategy section, but you may include statistical analyses and power calculations justifying the numbers or types of patients or specimens required. Deficiencies in any of these sections will be noted as weaknesses under the “approach” review criterion, are likely to affect the impact score for the project negatively, and also may result in an administrative bar to award until all reviewer concerns are resolved.

603

Data Safety Monitoring Plan For each proposed clinical trial, NIH requires a data and safety monitoring plan that describes the oversight and monitoring to ensure the safety of participants and the validity and integrity of the data. Large-scale (Phase III) trials must have a full Data and Safety Monitoring Board. NIH policy allows submission and review of grant applications before they are approved by your Institutional Review Board (IRB). If your application is likely to be funded, NIH will ask for documentation of IRB approval. Work with human subjects cannot begin without documentation of approval. Inclusion of Women and Minorities As described in Chapter 13, Public Law 103-43 requires that women and members of minority groups and their subpopulations be included in all NIHsupported clinical research projects involving human subjects, unless a clear and compelling rationale establishes that inclusion is inappropriate with respect to the health of the subjects or the purpose of the research. All applications proposing clinical research must contain explicit plans for including both genders (except for projects on gender-specific conditions such as prostate cancer, ovarian cancer, or pregnancy) and minorities in the subject population. You also must complete an “Inclusion Enrollment Report” (see Section R.400 in Research Instructions for NIH and Other Public Health Service (PHS) AgenciesdForms D) for each clinical study. Cost is not an appropriate justification for limited representation, and it is not sufficient to state that no one will be excluded on the basis of gender or race. Inclusion of Children NIH also requires that children (individuals under the age of 18 years) must be included in all clinical research supported by the NIH, unless there are scientific or ethical reasons for excluding them. To determine if inclusion of children applies to your application, follow the instructions in the Supplemental Grant Application Instructions for Preparing the Human Subjects Section of the Research Plan. Justification is required if there is limited or no representation of children.

Protection of Human Subjects Provide a complete description of the proposed involvement of human subjects as it relates to the work outlined in the Research Plan. If an exemption is claimed, explain why it is justified. If no exemption is claimed, address all items listed in Part II, Supplemental Instructions for Preparing the Human Subjects section of the Research Plan.

Vertebrate Animals The purpose of this section is to document the humane treatment of live vertebrate animals in the proposed research. There is no specified length but be succinct. Provide a complete description of the proposed use of vertebrate animals in the project, and address all three criteria listed in the SF424 (R&R) Application

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

604

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

Guide. Deficiencies will be noted as weaknesses under the “approach” review criterion and also may result in a bar to award. NIH policy allows submission and review of grant applications before Institutional Animal Care and Use Committee (IACUC) approval of the use of vertebrate animals. If your application is likely to be funded, NIH will ask for documentation of IACUC approval. Work with vertebrate animals cannot begin without IACUC approval. Budget and Justification The purpose of the Budget section is to present and justify the costs you are requesting to accomplish the project aims and objectives. For multiinstitutional applications, there must be a separate subaward/consortium budget component for each subawardee/consortium organization that will perform a substantive portion of the project. The application package for most NIH FOAs will include two budget forms: the SF424 (R&R) Budget Form and the PHS 398 Modular Budget Form. Each NIH application will use one of these budget forms. The modular budget format is applicable for research grant applications requesting $250,000 or less per year for direct costs in all years; consortium/contractual facilities and administrative costs may be requested in addition to the $250,000 direct costs per year limit. For modular budgets, applicants estimate the total research budget required for each year in multiples of $25,000 (e.g., $125,000, $150,000, or $225,000) and do not itemize categories such as glassware, reagents, animals, equipment, and travel. The budget justification should specify the roles and person-months of effort proposed for each of the listed project personnel and explain any large costs, unusual items, or unapparent costs that contribute to the overall estimate for the first year of the project. The budget for future years of the project should be similarly estimated, and any increases or decreases in the number of requested budget modules should be explained. For grants with modular budgets, the award will not be increased by an inflation factor each year. Many clinical research studies will require more than $250,000 in any year. Such applications must include a complete SF424 (R&R) Budget Form, with a detailed budget for each year of support requested. The SF424 (R&R) Budget Form includes three separate data screens. Read the instructions carefully and include all required fields. The form will generate a cumulative budget for the total project period. The budget should include costs for all personnel, consultants, equipment, supplies, travel, patient care, and other expenses (e.g., animal maintenance, equipment service contracts, and off-site space rentals). The Budget Justification attachment should explain the

roles of the proposed personnel and the need for items requested. • Be realistic. Both “padding” and deliberately underbudgeting reflect naı¨vete´ or lack of appreciation of the scope of the work proposed. They undercut your credibility as the PI and are viewed as weaknesses by the reviewers. • If possible, identify specific individuals for each position requested. “To be named” positions can be viewed as noncritical and are often deleted by reviewers. • Justify all requested equipment. Acquisition of major equipment is scrutinized carefully, especially equipment that is not project specific, such as fax machines and computers. • Break out supply costs into major categories (e.g., reagents, disposables, or animals). • Explain any year-to-year fluctuations in the budget, particularly the level of effort of personnel. Changes should parallel the research plan and project aims. • If there is a coinvestigator at another institution who will require salary and/or supplies to work on the project, be sure to include her or him in your budget. Senior/Key Personnel Profiles Component and Biosketches The biosketches showcase the expertise and experience of you and the research team involved in your project. Reviewers use the information in the biosketches to address the “investigators” review criterion and evaluate whether the proposed research team has the qualifications and experience to carry out the proposed work and overcome any problems that may arise. Each biosketch is limited to five pages in length and must not contain any graphics or embedded or attached files. Be sure that each biosketch includes a Personal Statement that is customized to the role of the investigator in this particular project. Facilities and Other Resources Reviewers use the information in this section to evaluate Environment. This section should show the reviewers that you have all of the equipment and space, including clinic and clinical laboratory space, necessary to perform the proposed project successfully. Do not assume the reviewers will know what is in your institution or what is actually available for your use. For applications from ESIs, this section also should describe the institutional investment in the success of the investigator, such as resources for training or travel, career enrichment programs, protected time for research, etc. • Make sure this section addresses all of the requirements of the proposed research plan.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REVISING UNSUCCESSFUL APPLICATIONS

• Justify any reliance on resources external to your research laboratory. Include strong and definitive letters of collaboration from the providers of those resources. • Show that all subcontractors/consortium members have the capability to perform the tasks assigned to them. • Make certain your resources and budget requests are consistent. Do not request funds for equipment listed in the Resources section as already available to you. Appendix Beginning with the January 25, 2017 receipt date, NIH has eliminated most Appendix materials for all applications (https://grants.nih.gov/grants/guide/ notice-files/NOT-OD-17-035.html). However, some FOAs will allow or require specific documents to be placed in the Appendix section. Please read the FOA carefully before submitting your application.

RECENT CHANGES TO APPLICATION PROCEDURES FOR NATIONAL INSTITUTES OF HEALTHeFUNDED CLINICAL TRIALSdMORE TO COME Over the past few years, NIH has been taking steps to enhance its management and oversight of clinical trials. In an effort of be transparent, NIH has been publishing Notices in the NIH Guide explaining changes to standard procedures or existing policies. Three recent Notices should be of particular interest to you. • Policy on Good Clinical Practice Training for NIH Awardees Involved in NIH-funded Clinical Trials (https://grants.nih.gov/grants/guide/notice-files/ NOT-OD-16-148.html)dAll NIH-funded investigators and staff who are involved in the conduct, oversight, or management of clinical trials should be trained in Good Clinical Practice (GCP). • NIH Policy on the Dissemination of NIH-Funded Clinical Trial Information (https://grants.nih.gov/ grants/guide/notice-files/NOT-OD-16-149. html)dNIH expects that all investigators conducting NIH-supported (in whole or in part) clinical trials will register their trials at ClinicalTrials.gov. NIH issued this policy to promote broad and responsible dissemination of information from NIH-funded clinical trials to the community. • Update on Clinical Trial Funding Opportunity Announcement Policy (https://grants.nih.gov/ grants/guide/notice-files/NOT-OD-17-043. html)dEffective January 25, 2018, if your application involves one or more clinical trials, you must submit

605

it through an FOA specifically designed for clinical trials. Please note that these changes will result in clinical trialespecific review criteria, application forms, and modified application instructions. Stay tuned and watch the NIH Guide for more information!

REVISING UNSUCCESSFUL APPLICATIONS Competition for NIH research and career development awards is tough, and it is common for applicants not to succeed on the first attempt. Table 34.7 lists some of the most common reasons for unsuccessful applications. Although a rejected grant application can be hard on the ego, the reality is that most investigators have to resubmit applications before securing funding for their research. Revising an application provides an opportunity to rethink weaknesses in your design, approach, and methods and to address the reviewers’ concerns.

How to Decide Whether to Revise Your Application Read and reread the summary statement. Look for the main problems identified by the reviewers. Discuss the summary statement with the NIH program officer (PO) responsible for your application. If the reviewers thought the main ideas and research question are worthwhile and important, then it is worth revising the application. Common fixable problems include unclear organization of the project, insufficient preliminary TABLE 34.7

Most Common Problems With Unsuccessful National Institutes of Health Grant Applications

Problem Not Significant Hypothesis is not scientifically sound Poor scientific rationale for the project or proposed approach Diffuse, superficial, or unfocused research plan Poor choice of experimental methods, models, or technologies Inadequate controls Insufficient statistical power for clinical studies Lack of knowledge of published relevant work Lack of new or original ideas Unrealistically large amount of work Uncertainty concerning future directions Lack of experience in the essential methodology

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

606

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

data, diffuse or unclear aims, too much work for the project period requested, inadequate experience of the proposed personnel, inadequate controls, and insufficient attention to potential problems or how the data will be interpreted. If the reviewers identified fundamental problems in the significance, scientific rationale, hypothesis, or approach, then it may be best to begin with a new idea and develop a new project. For example, the critiques may state that the objective is not very important, the project will add little to advance the field, the hypothesis is not sound, the work has already been done by others, or the proposed design or methods are not appropriate for testing the hypothesis.

How to Revise and Resubmit Your Application NIH allows one resubmission of an unsuccessful application within 37 months of the original application (https://grants.nih.gov/grants/guide/notice-files/ NOT-OD-10-140.html). Resubmission applications must include significant changes in the Research Plan and a one-page Introduction that summarizes the changes made in the application since the original review. If you disagree with the weaknesses noted in the previous review, you may use the Introduction to try to convince the reviewers of your point of view. However, regardless of how you feel, be courteous and do not insult the reviewersdsome of them may still be on the review committee! The key to revising and resubmitting your application successfully is to address the reviewers’ main concerns. Clarify the objectives and rationale for the project. Add preliminary data or an experienced collaborator. Delete weak and peripheral aims or experiments, and refocus diffuse projects tightly on the hypothesis. Change the approaches or methods that will be used if necessary. Rethink the design of a clinical trial to address concerns about statistical power. Ask a colleague who is experienced in your field and in grantsmanship, but who is not involved in your project, to read your application and the summary statement and provide advice. Even if you respond to all of the reviewers’ comments, your resubmitted application may still not receive a fundable score. This may happen for several reasons. When you make changes to the project, you risk introducing new problems. In addition, science “moves on,” so a project with high significance when first submitted may not be as important by the time the resubmission is reviewed. The membership of review committees also changes, so new reviewers with different perspectives may review your resubmitted application. Still, do not be discouraged. You can still

submit a new application (https://grants.nih.gov/ grants/guide/notice-files/NOT-OD-14-074.html).

What if It Appears That the Study Section Was Inappropriate or Biased? If it appears that there was not sufficient expertise on the review panel (e.g., a molecularly oriented study section reviewing a clinical trial), or you have reason to believe that there was a bias in the review, you should revise and resubmit the application and request a different study section for the review in the cover letter attachment. Real bias in the review is very rare. Reviewers are alert to potential bias among competitors on the review group and argue against it vigorously. SROs also are alert to potential bias among reviewers.

What if It Appears That There Was a Procedural Error During Peer Review? NIH has an appeals process designed to resolve procedural issues surrounding the initial peer review of an application. If you are concerned about any procedural aspects of the initial peer review of your application, you should first consider the comments in your summary statement and then contact your PO. The PO may be able to answer your questions about the summary statement. If your concerns remain unresolved, you may send a letter of appeal to your PO who will then forward it to his/her Institutes’ council for consideration. NIH accepts appeal letters for four distinct reasons: • Evidence of bias on the part of one or more peer reviewers; • Conflict of interest on the part of one or more peer reviewers; • Lack of appropriate expertise within the SRG; and • Factual error(s) made by one or more reviewers that could have altered the outcome of the review significantly. Appeal letters based solely on differences of scientific opinion are not accepted (https://grants.nih.gov/ grants/guide/notice-files/NOT-OD-11-064.html). A word of caution about timingdwhile you are waiting for council to consider your appeal, you may not submit another version of your application to NIH. Even if the council concurs with your appeal, your original grant submission (unedited) will be sent back for rereview. Therefore, in most cases it is more advantageous to revise and resubmit a new or resubmission application rather than appeal the initial review.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

NATIONAL INSTITUTES OF HEALTH GRANT PROGRAMS FOR CLINICAL RESEARCHERS AT VARIOUS STAGES IN THEIR CAREERS

NATIONAL INSTITUTES OF HEALTH GRANT PROGRAMS FOR CLINICAL RESEARCHERS AT VARIOUS STAGES IN THEIR CAREERS Although the R01 research project grant is the most well known and popular of NIH’s grants, NIH has several types of awards specifically designed to support clinical researchers at various stages in their careers. In addition, other types of grants, such as career transition awards and small (R03) and exploratory (R21) grants, are useful ways for new clinical investigators to obtain the preliminary data and proof of concept that are needed to prepare a competitive R01 application. Finally, the Loan Repayment Program (LRP) can help clinicians repay educational debts in return for a commitment to research.

Individual Career Development (“K”) Awards Detailed information about career development awards can be found at the NIH K Kiosk at https:// researchtraining.nih.gov/programs/career-development. There are a number of different types of career development awards, and not all NIH institutes and centers participate in all of them. In addition, each participating NIH component may have its own guidelines and requirements for a particular career development award to accommodate the career needs of researchers working in different fields. Therefore, you should contact the training and career development office in the NIH institute closest to your research interests before preparing an application. At the time of award, candidates for most NIH career development awards must be citizens or noncitizen nationals of the United States or permanent residents. The exception is the K99/R00 Pathway to Independence Award. Note that the review criteria for career development awards are different from the review criteria for research project grants discussed previously. Review criteria for the various career development awards vary somewhat but generally focus on the following: • Qualifications of the candidate • Career development plan • Research project to be conducted as part of the career development plan • Qualifications of the sponsor(s) for mentored awards • Environmental and institutional commitment to the candidate

Mentored Career Development Awards The candidate must identify a mentor with extensive research experience and must devote at least 75% effort

607

to career development research activities during the period of the award. Mentored Clinical Scientist Development Award (K08) The Mentored Clinical Scientist Development Award (K08) provides support for clinical professionals to develop into independent investigators. In general, K08 awards support more laboratory-oriented, translational, or preclinical research projects; clinicians who wish to pursue patient-oriented research (POR) training should see the section on the K23 award. There is substantial variability among the sponsoring NIH institutes in eligibility requirements, allowable costs, and application procedures. Applicants should contact the individual institutes for specific guidelines. Candidates should hold a clinical doctoral degree and should have initiated postgraduate clinical training. The requested project period may be for 3, 4, or 5 years, depending on the candidate’s prior research experience, additional experiences needed, and the policy of the awarding NIH institute. Awards are not renewable. Mentored Patient-Oriented Research Career Development Award (K23) The purpose of the Mentored Patient-Oriented Research Career Development Award (K23) is to support the career development of investigators who have made a commitment to POR and have the potential to develop into productive clinical investigators. For the purposes of this award, POR is defined as research conducted with human subjects (or on material of human origin, such as tissues, specimens, and cognitive phenomena) for which an investigator directly interacts with human subjects on research on mechanisms of human disease, therapeutic interventions, clinical trials, or development of new technologies. Candidates must have a clinical degree or its equivalent and must have completed clinical training, including specialty and, if applicable, subspecialty training, before award. Candidates may request 3e5 years of support, depending on their previous training and experience.

Career Transition Awards K99/R00 Pathway to Independence Award One of the most challenging transitions in any research career is from postdoctoral trainee to independent scientist. This award is designed to help the most promising, exceptionally talented, new investigators make the transition from trainee to independent investigator. Candidates must have no more than 4 years of postdoctoral research training experience and be in mentored, postdoctoral training positions. The K99/

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

608

34. GETTING NIH SUPPORT FOR YOUR RESEARCH

R00 award will provide up to 5 years of support in two phases. The initial mentored K99 phase will provide support for salary and research expenses for up to 2 years to complete research, publish results, and bridge to an independent research position. Candidates must commit at least 75% effort to the grant during the mentored phase. (Physician-scientists in surgical specialties may request a minimum of 50% effort.) The candidate must propose a research project that will be pursued as an independent investigator during the second R00 phase of the award. The R00 phase may be up to 3 years to support transition, as an independent scientist, to an extramural sponsoring institution where the candidate will be a tenure-track assistant professor or equivalent. This support will allow the awardee to continue to work toward establishing an independent research program and prepare an application for regular research grant (R01) support. Support for the independent phase, however, is not automatic and is contingent on being accepted by an extramural institution and NIH programmatic review of progress during the mentored phase of the award. For more information about this program, see https:// grants.nih.gov/grants/guide/pa-files/PA-16-193.html and the NIH K Kiosk. K22 Career Transition Awards K22 Career Transition Awards are intended to facilitate the transition of investigators, particularly clinical investigators, from the mentored to the independent stage of their careers. K22 awards provide “protected time” for newly independent investigators to develop their initial research programs in a research institution of the candidate’s choice. The unique feature of this award is that individuals may apply without a sponsoring institution while they are still in a “mentored” position. Because policies about the K22 awards differ markedly among NIH institutes, potential applicants should contact the training office in the NIH component most closely associated with their research interests before preparing an application.

Independent Scientist Awards The Independent Scientist Award (K02) provides up to 5 years of salary support for newly independent scientists who can demonstrate the need for a period of intensive research focus as a means of enhancing their research careers and enabling them to expand their potential to make significant contributions to their field of research. A candidate must have a doctoral degree and independent, peer-reviewed research support at the time the award is made; some NIH institutes and centers require the candidate to have an NIH research

grant at the time of application, whereas others will accept candidates with peer-reviewed, independent research support from other sources. The candidate must devote at least 75% effort to conducting research and research career development during the award. In addition, the candidate must be able to demonstrate that the requested period of 3e 5 years of salary support and protected time will foster his or her career as a highly productive scientist in the indicated field of research.

Midcareer Investigator Award in PatientOriented Research The purpose of the Midcareer Investigator Award in POR(K24) is to provide protected time for clinician investigators to devote to POR and to act as research mentors primarily for clinical residents, fellows, and/or junior faculty. This award is primarily intended for clinician investigators at the associate professor (or equivalent) level who have an established record of independent, peerreviewed federal or private research grant funding in POR. It is expected that investigators will obtain new or additional independent peer-reviewed funding, assume leadership roles in collaborative POR programs, and increase efforts and commitment to mentor beginning clinician investigators in POR, thereby increasing the pool of well-trained clinical researchers.

Exploratory/Development Grant (R21) Applications The R21 award mechanism is intended to encourage exploratory and developmental research projects in innovative new research areas by supporting early and conceptual stages of these projects. R21 applications are suitable for novel ideas that break new ground and need feasibility testing or high-risk/high-reward studies. Not all NIH institutes and centers accept investigator-initiated R21 applications under the R21 Parent FOA. However, those that do not may have RFAs or PAs to solicit R21 applications to meet specific program needs. Consult the NIH R21 website at https://grants.nih.gov/grants/funding/r21.htm before preparing an application. The review criteria are the same as for R01 research project grant applications described previously. However, the Research Plan for an R21 application is limited to six pages. Accordingly, reviewers will focus on the conceptual framework, the level of innovation, and the potential to significantly advance knowledge or understanding. Because this type of grant is designed to support innovative new ideas, preliminary data as evidence of feasibility are not required. Justification for

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

HOW TO STAY INFORMED ABOUT NATIONAL INSTITUTES OF HEALTH PEER REVIEW

the proposed work can be provided through literature citations, data from other sources, or, when available, investigator-generated data. R21 grants are generally limited to a total budget request of $275,000 for 2 years of support and generally are not renewable.

Small Research Grant (R03) Applications Small research grants provide research support that is limited in time (usually 1 or 2 years) and amount (usually $50,000 direct costs per year) and are nonrenewable. R03s are generally for support of preliminary studies or short-term projects. The results of an R03 grant often provide the preliminary findings for an R01 grant application. The review criteria for R03s are the same as those for R01s. Not all NIH institutes support R03 awards, and of those that do, different institutes have different objectives, guidelines, and requirements for their small grant programs. Not all NIH institutes and centers accept investigator-initiated R03 applications under the R03 Parent FOA. However, those that do not may have RFAs or PAs to solicit R03 applications to meet specific program needs. Consult the NIH R03 website at https://grants.nih.gov/grants/funding/r03.htm before preparing an application.

Loan Repayment Program The NIH LRPs were initiated in 2002 to attract health professionals to careers in clinical, pediatric, health disparity, or contraceptive and infertility research. There is also a Loan Repayment Program for Clinical Researchers from Disadvantaged Backgrounds. In exchange for a 2- or 3-year commitment to a research career, NIH will repay up to $35,000 per year of your educational debt from qualifying types of student loans. In addition, NIH will make corresponding federal tax payments to your Internal Revenue Service tax account to cover your increased federal taxes and also may reimburse other tax increases incurred as a result of your LRP benefits. For more information, see the LRP website at www.lrp.nih.gov. To qualify, you must have a doctoral or equivalent degree and be a US citizen, national, or permanent resident, and your research must be funded by a domestic nonprofit or US government (federal, state, or local) entity. You must commit 50% of your time (at least 20 h per week) for 2 years to the research and your educational debt must equal at least 20% of your “institutional base salary.” NIH issues payments directly to lenders on a quarterly basis. To remain eligible for the LRP program, student loans must remain segregated from noneducational loans and loans held by another person, such as a spouse or a child.

609

HOW TO STAY INFORMED ABOUT NATIONAL INSTITUTES OF HEALTH PEER REVIEW NIH periodically updates specific policies, forms, and procedures regarding the peer review process. Therefore, before you prepare a grant application, you should visit the NIH website (www.nih.gov) to obtain the latest information and discuss current application and review procedures with NIH program or review staff. The following are a few of the “starting points” for finding current information when you are ready to apply for research support.

“About Grants” Page (https://grants.nih.gov/ grants/about_grants.htm) This page covers many topics, from the basics of the grants application and review process to management of awards.

National Institutes of Health Institute/Center Home Pages Each of the NIH component institutes also has a home page. The general format for the Internet addresses is “www.Institute acronym.nih.gov” (e.g., www.nci.nih. gov, www.nhlbi.nih.gov, etc.) Each institute’s home page will have a way for you to find a contact for each general area of science that the institute supports, as well as the office responsible for managing the institute’s training and career development portfolio. The NIAID website has a particularly well-developed section on the grants process, with flow diagrams for investigators at different career stages and different types of grants (https://www.niaid.nih.gov/grants-contracts/chooseaward-career-stage).

The Center for Scientific Review Home Page (www.csr.nih.gov) Potential applicants are encouraged to visit the CSR home page for additional information about CSR and peer review. You can find the schedules for CSR study section meetings, study section rosters, information about review criteria and review procedures, and resources for applicants, including advice to investigators submitting clinical applications. Particularly instructive for new applicants is a video of a mock study section that illustrates the peer review process and how reviewers discuss applications.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

C H A P T E R

35 Philanthropy’s Role in Advancing Biomedical Research 1

Elaine K. Gallin1, Maryrose Franko2, Enriqueta Bond3

QE Philanthropic Advisors, Potomac, MD, United States; 2Health Research Alliance, Research Triangle Park, NC, United States; 3QE Philanthropic Advisors, Warrenton, VA, United States

O U T L I N E Introduction

Biomedical Imaging and Bioengineering Neuroinflammation Biomarkers Stem Cell Research

611

Organization of the Philanthropic Sector and Terminology Foundations Public Charities Alliances and Umbrella Organizations Serving the Philanthropic Sector

613 613 614

Supporting Institutions Stimulating Innovation Translating Discoveries into Cures, Therapeutics, and Preventions of Disease Establishing Product Development Partnerships Fostering Dissemination of Information, Data Sharing, and Patient Engagement Advocating for Resources and Policy Changes

614

History of the Philanthropic Sector 615 Private Foundations 615 Public Charities and Patient-Oriented Organizations 616 Areas of Contribution Philanthropic Sector: Areas of Contribution Developing Human Capital Building Knowledge and Expanding Scientific Disciplines

620 621 623 624 624 626

617 617 617

Conclusions and Future Directions

628

Summary Questions

629

619

References

629

INTRODUCTION This chapter reviews the philanthropic sector’s contributions to biomedical research with an emphasis on translational and clinical research. The philanthropic sector can be defined in various ways; in this chapter it is defined broadly to include the following types of organizations supporting biomedical research and related training activities: private foundations; not-for-profit voluntary health organizations (VHOs), and other charities that raise funds from the public; and private notfor-profiteendowed research institutions, such as the Howard Hughes Medical Institute (HHMI), which Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00035-6

619 619 620 620

support biomedical research and training. The United States outpaces other countries in terms of the number of philanthropic organizations supporting biomedical research and the magnitude of total support provided. However, many other countries also have vibrant philanthropic sectors that contribute significantly to their nations’ available biomedical research funds. While this chapter focuses primarily on the US philanthropic sector, it also includes examples of contributions made by noneUS based organizations. In 2014 philanthropic giving in the United States exceeded $358 billion. The bulk of this support consisted of charitable gifts from individuals (72%, or $258.51

611

Copyright © 2018. Published by Elsevier Inc.

612

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

billion). Foundation grant making (including family foundations) accounted for about 15% of the total, with the remainder coming from bequests and corporations. The greatest share of the $358 billion supported religious institutions and causes (31%), with the next highest share funding education.1 Philanthropy’s contribution to biomedical or health research in the United States is only a small percentage of the $358 billion and can only be estimated in part because it is often tracked with the philanthropic funds devoted to the provision of health services. Fig. 35.1 below taken from a 2012 article by Moses et al. plots US medical research funds by sector from 1994 through 2012.2 The contribution from private foundations and VHOs is estimated to have been about $4.2 billion in 2012. This is less than 4% of the total expenditures on US health research, an amount eclipsed by $67.9 billion invested by the pharmaceutical, biotechnology, and device industries, and $44.3 billion invested by the federal government and state and local governments. Comparisons across countries are difficult because the definition of and regulations governing philanthropic and charitable organizations differ from country to country, and comprehensive country-level information is often lacking. In the United Kingdom, the Association of Medical Research Charities,3 a national membership organization of leading medical and health research charities, reported that its 138 members invested 1.3 billion pounds (almost $2 billion) on research, which represented 15% of the total 8.5 billion pounds invested by all sectorsdgovernment, industry, and charities.4 Thus, in terms of the percentage of total

health research support, the philanthropic sector appears to provide a larger percentage of available research funding in the United Kingdom than in the United States. Despite country-level variability and limited data, the philanthropic sector has been growing in many developed countries, and even in emerging market countries such as Brazil, Russia, India, and China. This growth is likely to continue.5 This is good news because the philanthropic sector occupies a vital and distinctive niche in biomedical and health research, differing from government and for-profit support of research in a number of ways. First, government support for research is determined primarily by politicians who must consider many competing priorities and the greatest public good, avoid issues their constituencies deem too fractious, and in some cases, support issues that facilitate their reelection. In contrast, research support by for-profit companies is driven by the potential for timely financial gains, often limiting their investments in high risk, long-term research. Nonprofit, nongovernment philanthropic organizations are unconstrained by these boundaries. Thus, they are freer to invest in high-risk research, investigate unconventional hypotheses, work on difficult and seemingly intractable problems that require long-term investments, and support research in areas with little potential for large financial returns such as research on very rare diseases. Without these constraints, they can play a critical role in elevating the profile of neglected problems. Second, philanthropic organizations are often less bureaucratic, enabling them to act quickly and flexibly in response to emerging opportunities and needs. They

FIGURE 35.1 US funding for medical research by source, 1994e2012. Moses H, Matheson DHM, CairnsSmith S, George BP, Laisch C, Dorsey ER. The Anatomy of Medical Research: U.S. and International Comparisons. J Am Med Assoc 2015;313:174e89. http://dx.doi.org/10.1001/jama. 2014.15939.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

ORGANIZATION OF THE PHILANTHROPIC SECTOR AND TERMINOLOGY

also can facilitate product development by helping navigate regulatory hurdles, connecting innovators with entrepreneurs, and/or providing funds to help derisk projects. Finally and perhaps most importantly, the breadth and diversity of philanthropic organizations in the United States as well as in many other developed countries provides a myriad of opportunities for those seeking funding, making it likely that with perseverance committed investigators and excellent institutions will find support for their work. This chapter illustrates the distinctive role the philanthropic sector plays in biomedical research by briefly reviewing its history and providing examples or case studies of the contributions of different philanthropic organizations. The examples cover the continuum of biomedical research (depicted in Fig. 35.2) from basic discovery through clinical sciences and into public health, although special emphasis is given to translational and clinical research.

ORGANIZATION OF THE PHILANTHROPIC SECTOR AND TERMINOLOGY The terminology used to describe organizations in the philanthropic or not-for-profit, nongovernment sector can be confusing because terms often lack a precise meaning, and the tax codes and regulations relating to nonprofit organizations may differ between countries. With that caveat, the following terminology and

Translation of basic discovery into human studies

Basic Biomedical Research (molecules, cells, animals)

What is the molecular defect? What are possible drug targets? Does the potential therapeutic work in animals?

T-1 Block

613

organizations are part of the lexicon of the US philanthropic sector that supports health/biomedical research and related training activities.

Foundations The US Council on Foundations defines a “foundation” as an entity that supports charitable activities by making grants to unrelated organizations or institutions or to individuals for scientific, educational, cultural, religious, or other charitable purposes.6 In 2014 there were more than 81,000 foundations in the United States, but many of them are quite small, and even among the larger foundations only some of them fund biomedical research. The largest US foundationdthe Bill and Melinda Gates Foundation (with $44 billion in reported assets as of December 2014)ddevotes a substantial portion of its grant making to health research with many of its grants awarded outside the United States.7 Foundations that receive their funds from a single, main source of funding are referred to as private foundations, as opposed to “public foundations” or “grantmaking public charities” that obtain funding from numerous sources and must continue to seek support from diverse sources to retain their public charity status in the United States. The Foundation Center differentiates the following kinds of private foundations in the United States: • Independent or family foundations receive endowments from individuals or families (and, in the

Translation of new clinical knowledge into practice

Clinical Science and Knowledge (patients)

Is the intervention safe and efficacious?

Goal: T-2 Block

Improved Health (populations)

Is it effective? How does it compare with existing therapy? What is the evidence base?

Knowledge flow from clinical and population studies generates new hypotheses for basic biomedical research

FIGURE 35.2 Conceptualization of the continuum of clinical investigation.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

614

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

case of family foundations, they continue to show measurable donor or donor-family involvement). Examples of independent and family foundations include the Burroughs Welcome Fund, the Bill and Melinda Gates Foundation, and the Leona M. and Harry B. Helmsley Charitable Trust. • Company-sponsored or corporate foundations receive funds from their parent companies, although they are legally separate entities. Many pharmaceutical and device companies have foundations. Examples include the Merck Company Foundation, the Avon Foundation for Women, and the Beckman Coulter Foundation. • Operating Foundations run their own programs and services and typically do not provide much grant support to outside organizations. The Kaiser Family Foundation and the Pew Charitable Trust are examples of prominent operating foundations that support activities relevant to health and/or biomedical research. In addition to these three types of private foundations, the US Congress has charted several “hybrid” organizations referred to as “foundations” that serve as a bridge between federal agencies supporting research and the private for-profit and nonprofit sectors. They are the Foundation for the National Institutes of Health (FNIH)8; the ReaganeUdall Foundation for the United States Food and Drug Administration (FDA)9; and the Centers for Disease Control and Prevention (CDC) Foundation serving the CDC.10 These entities, which can receive funding from the US government, as well as from private philanthropic and for-profit organizations, serve to stimulate cross-sector collaborations.

Public Charities In addition to the types of private foundations listed above and the “hybrid” foundations affiliated with the US federal government, numerous “grantmaking public charities” also fund biomedical research and related training activities. These organizations must obtain money from diverse sources to retain their status as US public charities, according to US internal revenue rules. Approximately 950,000 organizations were classified as public charities in the United States in 2013 and estimates indicate that close to 38,000 or 13% were involved in health-related activities. However, it is unclear how many of these organizations support research versus health services and other activities.11 These organizations incorporate any of the following terms in their official name: foundation, alliance, society, or association. Many of them are disease-focused VHOs or patient-driven groups that can vary considerably from large established organizations such as the American

Cancer Society (ACS), the American Heart Association, and the Cystic Fibrosis Foundation (CFF), which support a variety of research programs both in the United States and in other countries to organizations focused on a geographical region such as the New York Community Trust or the Cleveland Community Foundation to very small organizations with one or two staffers and limited budgets. Throughout this chapter we refer to organizations that are driven by patients and their families as well as other volunteers as “voluntary health organizations.” Understanding the differences between these types of organizations is important when developing a strategy for seeking research support. Moreover, it is important to emphasize that VHOs can vary considerably in the rigor of their management and decision-making processes. The following characteristics are hallmarks of effective VHOs and should be considered when evaluating collaborations and possible opportunities.12 • They have clear missions and governance structures and strong connections to the communities they serve; • They provide well-researched, independent perspectives and are thus trusted sources of information; • If they support research and training programs, those programs are targeted to the needs of their communities, vetted by independent outstanding experts, and metrics of success collected; and • They have effective management teams that attract donors by keeping overhead costs low, staying focused on their goals, and directing most of their funds to supporting their programs. A good starting point for collecting information on a VHO is GuideStar, a clearinghouse for information on more than 1.8 million US nonprofits.13

Alliances and Umbrella Organizations Serving the Philanthropic Sector As the philanthropic sector has grown, alliances and umbrella groups have been established to foster collaborations and address complex, system-wide issues, inform policy, advocate for resources, and/or define and share best practices. Some of these organizations include members from different sectors (academia, industry, and philanthropies), while others are limited to representatives of the philanthropic sector. A sampling of these organizations is described below: • The Health Research Alliance (HRA) founded in 2005 is a member organization of nonprofit funders of biomedical research committed to maximizing the impact of biomedical research to improve human

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

HISTORY OF THE PHILANTHROPIC SECTOR

health.14 Today, with 73 members, HRA is a growing consortium of private foundations and public charities funding biomedical and health research across the research continuum. The HRA has collected grantee information on HRA memberefunded grants since 2006.15 To date, there are over 31,000 records in the database representing over $8.8 billion in awards. HRA has used this comprehensive database to analyze and disseminate a snapshot of funding of biomedical research and training by HRA members. HRA also convenes its members biannually to address issues key to accelerating research discovery and its translation. The HRA network has been effective at identifying or creating innovative approaches to address challenges in biomedical research across research fields, disease areas, career stages, etc., and has spawned numerous collaborations among funders. • Grantmakers in Health launched in 1982, provides resources for grant makers and others seeking expertise in health-related philanthropy. Its member organizations include foundations, VHOs, and other groups interested in health issues including how to deliver medical care, access to quality care, and disease prevention and wellness.16 Thus, much of the research funded by its member organizations falls within the second translational block depicted in Fig. 35.2. • Research!America is advocacy organization committed to making research to improve health a higher national priority. To inform policy, it conducts polls and uses the data it collects to inform Congress and the executive branch of public opinion on support for public health research and infrastructure, basic and medical research, and the American public’s support for global health.17 Research! America’s members represent philanthropic, academic, and for-profit organizations. Their goals include advocating for funding for medical and health research from the public and private sectors and informing the public and decision makers on the benefits of medical and health research. • The Genetic Alliance is a health advocacy alliance that includes more than 1200 disease-specific advocacy organizations, as well as universities, private companies, government agencies, and public policy organization.18 The network is a dynamic and growing open space for shared resources, creative tools, and innovative programs. Among its various initiatives has been strong support for “Open Access”dchampioning the public policy changes designed to provide consumers with “public access” to published articles and patient-led registries and biorepositories.

615

• FasterCures, which is part of the Milken Institute, works to reduce the roadblocks that slow progress in the development of new cures and treatments for diseases.19 Its annual Partnering for Cures meeting convenes patient advocates, researchers, funders, investors, and policymakers from every sector of the medical research and development system to share best practices, spur collaborations, and stimulate innovation and venture philanthropy. • The International Cancer Research Partnership (ICRP)20 is an alliance of over 100 cancer organizations working together to enhance global collaboration and strategic coordination of research. ICRP’s aim is to improve access to information on cancer research projects by sharing data in a common format. It is known as the Common Scientific Outline (CSO). Use of the CSO across funders maximizes the impact of the member organizations’ efforts and benefits researchers and cancer patients worldwide.

HISTORY OF THE PHILANTHROPIC SECTOR Private Foundations It was not until the beginning of the 20th century that the philanthropic sector as we know it today was first launched. Leading US industrialists were amassing great fortunes and it was an opportune time to invest portions of their fortunes for the public good. Of the small group of industrial tycoons who helped shape the US philanthropic enterprise, John D. Rockefeller is the most notable for his contributions to medicine and public health.21 The Rockefeller Institute for Medical Research, known today as Rockefeller University, was created in 1901 and was the first institution in the United States devoted solely to understanding the underlying causes of human diseases through biomedical research. John D. Rockefeller wanted to establish a US research center similar to France’s Pasteur Institute, which had been created to apply laboratory science to advance knowledge of human disease such as tuberculosis, diphtheria, and typhoid fever. Notably, the Rockefeller Institute was the home of the world’s first clinical research hospital, which is still functioning today. Despite fears by some that creating a large private foundation might undermine American democracy, the Rockefeller Foundation was chartered in New York State in 1913 with the mandate to “promote the well-being of mankind throughout the world.”22 Within a decade, it was the largest philanthropic organization worldwide concentrating its efforts primarily on public health and medical education. Its grant making was heavily influenced by

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

616

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

the 1910 Flexner Report commissioned by the Carnegie Foundation for the Advancement of Teaching, which argued that medical education should become a more academic rigorous program grounded in basic science and research.23 The Rockefeller Foundation helped implement some of the Flexner Report recommendations by providing funds to modernize medical schools and build their faculty.24 John D. Rockefeller and the Rockefeller Foundation also established the first school of public health at Johns Hopkins University and launched two notable public health campaigns in the early 1900sdone to eradicate hookworm disease in the United States and another to tackle yellow fever in the Western hemisphere. In 1909 the Rockefeller Sanitary Commission for the Eradication of Hookworm Disease was established to address the fact that over 25% of children in some southern US states were estimated to be infected with hookworm. Five years later, hookworm, while not eradicated, was much less prevalent because of the Commission’s efforts, which included a campaign to build sanitary privies. The Rockefeller Foundation’s Yellow Fever Commission was similarly successful, almost wiping out the disease in South and Central America by the 1920s. It is also noteworthy that at a time when the United States was very insular, the Rockefeller Foundation extended its grant making beyond the borders of the United States. One of its seminal achievements was establishing the China Medical Board to help develop a modern system of medicine in China and oversee the reorganization of the Peking Union Medical College that is a leading research and training institution in China today. The Rockefeller Foundation’s extraordinary accomplishments in its first 50 years set a high bar for subsequent foundations whose numbers continued to grow throughout the 20th century. In the United States, there were almost 80,000 foundations at the beginning of the 21st century. Today two of the world’s wealthiest foundationsdthe Bill and Melinda Gates Foundation and the Wellcome Trustddevote a significant portion of their resources to biomedical research and training, as do many other foundations established by wealthy philanthropists and their families over the last century. Many wealthy philanthropists have invested heavily in leading US academic health institutions, enabling these institutions to expand their research capabilities and build new research centers. The first two decades of the 21st century also have been notable for a rise in young, newly affluent entrepreneurs who are intent on translating lessons learned in business into high-impact philanthropy addressing critical global and societal issues. Terms such as “venture philanthropy” and “philanthrocapitalism” are proliferating as this new generation of philanthropists is

entering the sector and experimenting with ways to stimulate innovation and ensure that their philanthropic investments have impact.25 Some of these individuals are participating in the Giving Pledge, a movement initiated by Bill and Melinda Gates and Warren Buffet where individuals and their families commit at least half of their wealth to charity. This is likely to translate to more philanthropic funds devoted to biomedical research. As of 2016, more than 141 of the world’s billionaires have signed on to the Gates/Buffet Giving Pledge. The Foundation Center has started the Glasspockets program, which is now tracking the philanthropic activities of the 141 participants (16 reside outside the United States) worth more than $707 billion.26

Public Charities and Patient-Oriented Organizations A few decades after the Rockefeller Foundation was established, another type of philanthropic organization da public charitydwas formed that also would serve as an important role modeldthe National Foundation for Infantile Paralysis (known as the National Foundation). The National Foundation was established in 1938 by President Franklin D. Roosevelt, who had contracted polio in 1921 at the age of 39. With the support of millions of Americans, it quickly grew to be the largest VHO in the United States.27 The National Foundation not only advocated for and cared for polio patients; it also funded research focused on the bold goal of developing a polio vaccine. Understanding that the field of virology needed to be developed and that more basic knowledge of the polio virus was critical before a vaccine could be developed, the National Foundation invested $55 million in research and $33 million in fellowships and other training programs between 1938 and 1959. Its funds supported the research of Albert Sabin and Jonas Salk, as well as many other investigators, including the Nobel Prizeewinning work of John Enders, Thomas Weller, and Frederick Robbins demonstrating the polio virus could infect tissue other than nerve cells and be grown in culture.28 In 1953, at a time when the federal government’s research investment in polio was less than $75,000, the National Foundation invested $2 million in research to develop a polio vaccine. That research investment laid the groundwork for launching the world’s biggest public health experiment a year laterdtesting the efficacy of the Salk vaccine in more than 1.8 million school children. While some of the work supported by the National Foundation had its scientific controversies,29 the National Foundation’s willingness to take risks was critical to the development of an effective polio vaccine in less than two decades,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

AREAS OF CONTRIBUTION

and to building the field of virology. In an era without significant investment from industry or government, the National Foundation demonstrated the power of voluntary patient organizations to catalyze the full spectrum of disease-focused researchdfrom discovering the basic science of a disease to identifying therapeutic targets and then developing and testing interventions through clinical studies. Building on the success of the National Foundation (today known as the March of Dimes) and other early VHOs like the American Cancer Society, founded in 1913, and the American Heart Association, founded in 1924, the number of US VHOs steadily grew throughout the 20th century and continues to grow. Similar trends, while starting later, are evident in other developed countries.30

AREAS OF CONTRIBUTION The philanthropic sector’s numerous contributions to biomedical research can be grouped into the areas listed in the sidebar. These are discussed briefly below:

Philanthropic Sector: Areas of Contribution • • • • •

Developing Human Capital Building Knowledge and Expanding Scientific Disciplines Supporting Institutions Stimulating Innovation Translating Discoveries into Cures, Therapeutics, and Preventions of Disease • Establishing Product Development Partnerships • Fostering the Dissemination of Information, Data Sharing, and Patient Engagement • Advocating for Resources and Policy Changes

Developing Human Capital A hallmark of many philanthropic funders of biomedical research is their emphasis on nurturing talented individuals, often at early stages in their careers. Investments in junior level researchers can be considered high risk, essentially granting significant resources to individuals who show promise but do not yet have a lengthy track record of success. These programs tend to be highly competitive and thus enjoy a high level of prestige. In addition, they vary in the type of research fundeddbasic, clinical, disease-focuseddand the mechanisms by which the programs support early-career scientists along the path to becoming independent investigators. Since awarding its first grants in 1998, the goal of the Doris Duke Charitable Foundation (DDCF) Medical

617

Research Program31 has been to advance the translation of biomedical discoveries into applications that improve human health through the support of clinical research. DDCF supports physician-scientists in the United States at different stages of their careers through many competitive award programs, several of which specifically target early-career physician-scientists working in academia. Like DDCF, the Burroughs Wellcome Fund (BWF)32 and the HHMI33 also have supported earlycareer investigators working in any area of biomedical research. In contrast, VHOs interested in supporting young investigators target their grants to their specific diseases to build research capacity in those areas. For example, the Leukemia and Lymphoma Society, established in 1949 by the family of a child who died of leukemia, has prioritized supporting young investigators, and over the years, researchers they supported have achieved important clinical advances including the development of the first effective drugs for childhood leukemia.34 Because the transition from a postdoctoral fellow to an independent investigator can be a vulnerable period in a researcher’s career, several philanthropic organizations also offer “bridging awards,” which support an investigator for a year or two at the senior postdoctoral stage, and then enable them to transition to an independent investigator with a faculty position for up to an additional 3 years of support. BWF was one of the first funders to develop and rigorously evaluate this funding mechanism, offering its first bridging Awardsdthe BWF Career Awards for Medical Scientistsdin 1994. BWF’s program was modeled after the Lucille P. Markey Charitable Trust Scholars Program, offered between 1985 and 1991, which supported 113 researchers transitioning from fellowships to faculty positions. An evaluation of BWF’s bridging awards concluded that they are particularly valuable in facilitating the transition to a career as an independent investigator and providing physicianscientists with much needed protected time for research.35 Recognizing the success of the early bridging awards provided by the Markey Trust and BWF, other philanthropic funders and the National Institutes of Health (NIH) also established bridging award programs. Examples include the NIH K99/R00 awards and Pancreatic Cancer Action NetworkeAmerican Association for Cancer Research (AACR) Pathway to Leadership Grants. Table 35.1 lists a sample of the organizations offering early-career or career development programs and is not meant to be comprehensive. Most of these organizations are members of the Health Research Alliance described earlier in this chapter. An early-career award from a philanthropic funder can be especially valuable because it often is more flexible than government funding in terms of the activities

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

618

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

TABLE 35.1

Organizations With Career Development Awards

Organization Name

URL

Alzheimer’s Association

www.alz.org

American Association for Cancer Research

www.aacr.org

American Cancer Society

www.cancer.org

American Diabetes Association

www.diabetes.org

American Epilepsy Society

www.aesnet.org

American Federation for Aging Research

www.afar.org

American Heart Association

www.professional.heart.org

Arthritis National Research Foundation

www.curearthrisis.org

Autism Science Foundation

www.autismsciencefoundation.org

Avon Breast Cancer Crusade

www.avonfoundation.org

Bladder Cancer Advocacy Network

www.bcan.org

BrightFocus Foundation

www.brightfocus.org

Burroughs Wellcome Fund

www.bwfund.org

Cancer Research Institute

www.cancerresearch.org

Conquer Cancer Foundation of ASCO

www.conquercancerfoundation.org

CURE (Citizens United for Research in Epilepsy)

www.cureepilepsy.org

Damon Runyon

www.damonrunyon.org

Doris Duke Charitable Foundation

www.ddcf.org

Foundation Fighting Blindness

www.blindness.org

Gerber Foundation

www.gerberfoundation.org

Howard Hughes Medical Institute

www.hhmi.org

JDRF

www.jdrf.org

Leukemia & Lymphoma Society

www.lls.org

LUNGevity Foundation

www.lungevity.org

Lymphoma Research Foundation

www.lymphoma.org

March of Dimes Foundation

www.marchofdimes.org

Melanoma Research Alliance

www.curemelanoma.org

Multiple Myeloma Research Foundation

www.themmrf.org

New York Stem Cell Foundation

www.nyscf.org

Pancreatic Cancer Action Network

www.pancan.org

Parkinson’s Disease Foundation

www.pfd.org

Rheumatology Research Foundation

www.rheumresearch.org

Rita Allen Foundation

www.ritaallen.org

Sarnoff Cardiovascular Research Foundation

www.sarnofffoundation.org

Simons Foundation

www.simonsfoundation.org

St. Baldrick’s Foundation

www.stbaldricks.org

Susan G. Komen

www.komen.org

The Medical Foundation at HRiA

www.hria.org

V Foundation for Cancer Research

www.jimmyv.org

This is not a comprehensive list.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

AREAS OF CONTRIBUTION

supported and its ability to provide grantees with cost and no-cost extensions. Such flexibility can be particularly important for clinical research, where delays in patient recruitment and regulatory approvals can slow progress. Another key element of many philanthropic career development award programs, which distinguishes them from federally funded programs, is their practice of convening awardees at regular intervals to build research networks and provide other types of professional development. These sessions are viewed by philanthropic funders as a way not only to monitor their grantees’ progress but also as a way to enable the grantees to interact with leaders in their fields and to ensure that a lack of critical skills (in areas like laboratory management and grant writing) does not jeopardize the success of investigators in whom the funders have made a significant investment. Thus, although obtaining federal funding is a significant early-career milestone, the extra features and flexibility associated with philanthropic support can influence a researcher to choose a foundation award instead of government funding. An example of these professional development activities is the “Grantee Weekend” for awardees of The Pancreatic Cancer Action NetworkeAACR’s Pathway to Leadership Grant mentioned earlier. The weekend involves scientific sessions during which grantees provide updates of research, receive input from the Pancreatic Cancer Action Network’s Scientific and Medical Advisory Board, and attend networking and other activities to enhance the awardees’ professional development skills. Another example of philanthropy-supported mentoring was a laboratory management and leadership course, jointly organized by HHMI and BWF in 2002, for investigators holding their organizations’ earlycareer awards. The course content formed the basis for a book, Making the Right Moves: A Practical Guide to Scientific Management for Postdocs and New Faculty, which is available to download.36 The companion book (also available on the HHMI website) Training Scientists to Make the Right Moves: A Practical Guide to Developing Programs in Scientific Management has enabled academic institutions and professional societies to provide similar professional development opportunities for their earlycareer investigators. These activities formed the basis of Excellence Everywhere, a resource for scientists in emerging science centers prepared by the BWF in 2009.37

Building Knowledge and Expanding Scientific Disciplines In addition to fostering the careers of early-career investigators, philanthropic sector support has helped

619

create and build scientific disciplines. In some cases, this has meant establishing an academic home, a training program, or infrastructure for a new research field, or nucleating a critical mass of investigators in underresourced, emerging areas, or attracting scientists from a wide range of disciplines to tackle seemingly intractable problems. Despite the fact that in the United States the federal government’s expenditures in the life sciences dwarfs that of philanthropy’s, many examples exist where philanthropic investments in underfunded and newly emerging areas have helped develop and/ or expand scientific disciplines and supported critical early stage proof of principle research. Four research areas where philanthropic support has made, or is making a difference, are biomedical imaging and bioengineering, neuroinflammation, biomarkers, and stem cell work. Biomedical Imaging and Bioengineering The Whitaker Foundation, founded in 1975, contributed more than $700 million during its 30 year history, primarily to build the interdisciplinary field of biomedical engineering.38 At the time the Whitaker Foundation began its grant making, support for engineering research with medical applications had fallen between the purviews of the NIH and the National Science Foundation, and few formal training programs were available. The Whitaker Foundation supported research, curriculum development, education programs, faculty hires, building construction, textbooks, and conferences, essentially establishing 30 departments of bioengineering at universities across the United States. This early investment by the Whitaker Foundation has been leveraged by many subsequent funders, including the NIH when it established the National Institute of Biomedical Imaging and Bioengineering (NIBIB). In 2005, HHMI and NIBIB formed a partnership to support the development of new training opportunities in emerging interdisciplinary research environments.39 The goal of the program was to help biomedical research institutions develop a cadre of PhD scientists trained to conduct interdisciplinary research at the intersections of the biomedical, physical, computational, and mathematical disciplines. The initiative consisted of two phases: HHMI-funded Phase I, which supported the establishment of new interdisciplinary training programs at research institutions, and Phase II, funded by NIBIB, sustained the training programs through their critical early years. Neuroinflammation Neuroinflammation is prevalent in many diseases of the brain including Alzheimer’s disease, traumatic brain injury, amyotrophic lateral sclerosis (ALS), Parkinson’s disease, Huntington’s, and many others. Research on

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

620

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

neuroinflammation is challenging not only because it is difficult to study in humans but also because neuroinflammation in animal models manifests very differently from the human disease. Though a great deal of research points to neuroinflammation as having a contributory role in Alzheimer’s disease and other brain disorders, this area of investigation has been severely understudied and underfunded. Moreover, while Alzheimer’s disease affects a large share of the population and with an aging population its prevalence in the United States is increasing, it receives very little federal support compared to cancer or heart disease. In fiscal year 2015, the federal government committed some $5.4 billion to cancer research, about $1.2 billion to heart disease, and $3 billion to research on HIV/AIDS. Research funding for Alzheimer’s reached only about $566 million.40 Several US and UK organizations, including the Alzheimer’s Association and the Welcome Trust, have identified this funding gap and have targeted significant resources toward the study of neuroinflammation. The Alzheimer’s Association’s Part the Cloud Challenge on Neuroinflammation aims to accelerate understanding of immune processes involved in neurodegeneration and accelerate therapeutics and discovery of compounds to be tested in early clinical trials.41 Since its inception, Part the Cloud has raised over $6.4 million and continues to support both US and international researchers conducting studies in North America. In 2014 the Wellcome Trust committed 5 million pounds to support a publice private “Consortium for Neuroimmunology of Mood Disorder and Alzheimer’s Disease.” Seven universities and two companies are members of the consortium, which aims “to pioneer a new scientific roadmap for delivering immunological medicines to patients with major brain and mental health disorders.”42 Biomarkers The identification of valid biomarkers that are well characterized and shown to correctly predict relevant clinical outcomes is critical, but has been challenging. While a few well-validated biomarkers exist, in many instances the promise or “validity” of biomarkers has not yet been shown. The Biomarker Consortium was established to advance progress in this critical area by combining the forces of the public and private sectors.43 The consortium is made up of nonprofit organizations, government agencies, and for-profit companies and is managed by the FNIH. It aims to discover, develop, and seek regulatory approval for biomarkers to support new drug development, preventive medicine, and medical diagnostics. With more highly predictive biomarkers markers that have an impact during a patient’s illness or lifespan, it will help create a new era of precision medicine.

Cancer Research UK, a charity funding only UK cancer research, also funds biomarker research through their Biomarker Project Awards.44 Another noteworthy biomarker effort is the Global Consortium for Biomarker Standardization, which was established by the Alzheimer’s Association to help make sense of the measurements being done in Alzheimer’s patients using magnetic resonance imaging, measurements of brain volume, beta-amyloid imaging, etc.45 The consortium brings together key researchers, clinicians, and industry, regulatory and government leaders in Alzheimer’s disease to achieve consensus on the best ways to standardize and validate biomarker tests for use in clinical practices around the world. Stem Cell Research Fortunately for science, the philanthropic sector is free to engage in areas that may be politically unpopular, or for which there are federal restrictions, such as reproductive science and stem cell research. In 2001, President George W. Bush announced restrictions on federal support for research involving human embryonic stem cells. This policy was revoked by an Executive Order from President Barack Obama in 2009, but expanded federal funding for stem cell research remains uncertain, pending the outcome of a federal lawsuit challenging the new policy. Private foundations, however, are free to support stem cell research, and many of them have chosen to devote considerable resources to this area of research. In some cases, foundations have simply continued to support embryonic stem cell research as part of their ongoing grants portfolio. In other cases, foundations such as the Broad Foundation46dwhich supported Centers for Regenerative Medicine and Stem Cell Research at both the University of California, San Francisco (UCSF) and the University of California, Los Angeles (UCLA)dhave helped to build the needed infrastructure. In a few cases, foundations have made stem cell research their core mission. For example, the New York Stem Cell Foundation, founded in 2005, opened a stem cell research facility in New York City in 2006 and has funded research projects and postdoctoral fellowships.47 Lastly, a number of philanthropic organizations including the Juvenile Diabetes Research Foundation International,48 the Christopher and Dana Reeve Foundation,49 and the Michael J. Fox Foundation50 have devoted time and resources to public education and advocacy in support of stem cell research.

Supporting Institutions Philanthropic donors have contributed a significant portion of infrastructural support at many of the leading US academic health centers. In some cases, the support

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

AREAS OF CONTRIBUTION

has taken the form of very large grants that provide general undirected support, which can be used in a variety of ways to strengthen existing institutions. As the size of these pledges or grants has increased, many academic institutions have been renamed for major benefactors including the Keck School of Medicine at the University of Southern California, Weill Cornell Medical College, and the Raymond and Ruth Perelman School of Medicine at the University of Pennsylvania. In other cases, philanthropic organizations and wealthy donors have established or helped to fund new, freestanding research institutes focused on new technologies or new research models. For example, the Broad Foundation created the Eli and Edythe L. Broad Institute.51 The Broad Institute is an interinstitutional research entity linked to the Massachusetts Institute of Technology and Harvard University, as well as its affiliated major teaching hospitals, Dana Farber Cancer Institute and Massachusetts General Hospital. Building on these linkages, it has created a collaborative interdisciplinary research model to study the molecular basis of inherited diseases, the mutations that underlie different cancer types, and other challenging research areas. The Janelia Research Campus, established in 2006 by the HHMI, similarly was designed to provide an interdisciplinary, collaborative environment to pursue challenging scientific problems and long-term projects difficult to address in less-supportive environments.52 Provided with outstanding core facilities, Janelia scientists work on information processing by neural circuits and the development of new imaging and computational technologies. With no tenure, no teaching or administrative responsibilities, and no need to seek external research support, their time can be devoted solely to advancing knowledge in these challenging and critically important areas. Research institutions in Africa and other regions have also benefited from investments by the philanthropic sector. This is particularly true for South Africa’s easternmost provincedKwaZulu-Nataldknown not only for Africa’s largest port in Durban but also for its high HIV/AIDS incidence and the first reported cases of extreme drug-resistant tuberculosis in South Africa.53 Between 1996 and 1998, two philanthropic organizations made significant investments that have helped the province deal with these critical issues. In 1996 the Wellcome Trust established the Africa Centre for Population and Reproductive Health to conduct population surveillance and epidemiological studies in the rural Hlabisa area.54 Three years later, DDCF partnered with the University of KwaZulu-Natal (UKZN) to build the Doris Duke Medical Research Institute (DDMRI)55dthe first standalone research building at the UKZN medical school, which housed the HIV Pathogenesis Program.56 The DDMRI provided much needed research space and

621

research funding to support the province’s leading HIV/AIDS investigators, house clinical laboratories for the Wellcome Trust’s Africa Centre, and the HIV Pathogenesis Program. This investment enabled researchers to attract additional support from NIH and other funders to expand their work. About a decade later, the province’s growing need for additional research on HIV/AIDS and tuberculosis, and the successful investments of the Wellcome Trust and DDCF, led HHMI to commit funds to build the KwaZulu-Natal Research Institute for Tuberculosis and HIV (K-RITH) and provide 10 years of research support for investigators working at K-RITH.57 Most recently, HHMI and the Welcome Trust provided additional support and partnered with the University of KwaZulu-Natal to further expand research capacity in the province merging the Africa Center and K-RITH to form the Africa Health Research Institute.58 The Wellcome Trust invested in building research institutions throughout sub-Saharan Africa. Most recently, together with the UK Department for International Development, it launched the Developing Excellence in Leadership, Training, and Science in Africa (DELTAS Africa) program.59 DELTAS aims to establish worldclass research environments at African universities with a strong focus on creating training opportunities for the next generation of researchers. The DELTAS Africa scheme will award over 46 million pounds ($70 million US dollars) over an initial period of 5 years to programs led from universities and research institutes in Ghana, Kenya, Mali, South Africa, Uganda, and Zimbabwe. This program builds on the Wellcome Trust’s earlier African Institutions Initiative, which supported seven capacity-building consortia, involving 18 African countries and 54 institutions across the continent, as well as 20 institutions from Northern (predominantly European) institutions.

Stimulating Innovation Many philanthropic organizations think their primary role is to provide “risk capital” needed to stimulate innovation, increase the pace of scientific breakthroughs, and ensure a robust pipeline of drugs and diagnostics in development. These organizations may focus all their efforts on stimulating innovation in a given area of research, or they may invest just a portion of their resources on such activities, reserving other funds to develop human capital, provide institutional support, or other research-related activities. A range of different approaches has been used to stimulate innovation including providing pilot grants for exploratory projects, supporting cross-disciplinary research teams, and utilizing social media and prizes to attract

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

622

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

researchers to work on difficult problems and mine existing data. Examples of these approaches are provided below. • Pilot Grants A commonly used approach to stimulate innovation is to provide modest levels of funding for relatively short periods (often 1e2 years) to stimulate new investigators as well as established investigators to explore unproven directions that, if successful, would address an important need. This approach, while risky, has been successfully used by many organizations. DDCF, which funded clinical research on HIV/AIDS care and treatment in Africa between 1999 and 2007, awarded nine 2-year grants called Innovation in Clinical Research Awards in 2003 to stimulate development of low-cost point-of-care diagnostics of AIDS care in low-resource regions.60 One of these grants produced the early stage feasibility data leading to a Bill and Melinda Gates Foundation Grand Challenges award and ultimately to a disposable semiquantitative rapid CD4 test now being marketed as [email protected] Launched in 2003 by the Gates Foundation, Grand Challenge awards are the most visible and numerous pilot grants funded by a single philanthropic organization. The Grand Challenges Program was designed to encourage creative minds anywhere in the world to work on major global health problems. Today, it has morphed into a family of initiatives supported by a consortium of partners including Brazil, Canada, India, Norway, South Africa, the United Kingdom and the United States. According to the website, 1896 recipients in 83 countries have received Grand Challenges awards.62 • Cross-Disciplinary Research Awards Innovations often occur at the intersections of disciplines where individuals are stimulated to think “out-of-the-box.” Moreover, some critical questions can only be addressed when teams of researchers with both biomedical and computational skills work together to collect, aggregate, and analyze large data sets. Below are two examples of philanthropic activities designed to facilitate these types of interactions: • The Prostate Cancer Challenge Awards63 funded by the Prostate Cancer Foundation are designed to support transformational prostate cancer research by funding cross-disciplinary teams of investigators addressing strategic issues. These awards range from $300,000 to $1,500,000 per year for 3 years. The Challenge Awards encourage partnerships of investigators doing “first in field” research currently not working on prostate cancer to team up with established prostate investigators.

• Stand Up to Cancer (SU2C) Dream Team Awards64 bring together top researchers at different institutions to develop new and improved approaches to cancer research. Each Dream Team receives $6 to $22 million in funding for a period of 3e4 years from the AACR, the association administering the program. AACR launched the SU2C program to eliminate barriers to creativity and collaboration by enabling scientists with different areas of expertise from different institutions across the country and around the world to work together and bring new therapies from the laboratory to the clinic. • Challenges and Prizes Galvanized by the potential of harnessing social networks to solve difficult problems that were significant roadblocks to research progress, the philanthropic sector has experimented with issuing challenges and offering prizes to address some of these research questions, which if solved would remove significant roadblocks to progress. Typical government- or philanthropic-funded research grants support researchers to work on proposed, future studies. In contrast, prizes or challenges fund the “winner” of a competition retrospectively, when the specific goal defined by the funder is achieved. This mechanism, while potentially very cost-effective in stimulating a broad array of researchers to address difficult problems, requires the articulation of a specific problem and clear metrics to determine the successful applicant. It should be noted that while such competitions are referred to as “prizes” they differ from awards such as the Nobel Prize or Lasker Award, which recognize an outstanding scientific achievement or body of scientific work in a field. The following are two creative funding mechanisms or reward structures involving social networking that have been used by philanthropies to advance biomedical research: • Prize4Life was founded in 2006 by a group of Harvard business school students when one of them was diagnosed with ALS.65 Prize4Life is one of the first philanthropic organizations to use prizes to accelerate the discovery of new treatments and a cure for ALS. Thirty-seven groups from around the world competed for Prize4Life’s first $1 million prizedthe ALS Biomarker Prize. The prize was awarded in 2011 for development of electrical impedance myography, a biomarker that could assess ALS disease progression with more sensitivity and consistency than the existing approaches that were inadequate. This biomarker has been critically important in assessing the effectiveness of new ALS treatments. Prize4Life’s second competition,

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

623

AREAS OF CONTRIBUTION

the Avi Kremer ALS Treatment Prize, was launched in 2012 and will be awarded to a team that identifies a treatment that extends the lives of an ALS mouse models by 25%. While an official winner has not yet been announced, the Prize4Life website (prize4life.org) indicates that one team was identified in 2015 as having a treatment that if validated could meet the criteria of the competition. • The Seizure Detection and Prediction Challenge sponsored by the American Epilepsy Society, the Epilepsy Foundation, and the National Institute of Neurological Diseases and Stroke is an excellent example of a philanthropic sector and government working together on a challenge.66 Contestants were charged with analyzing retrospective intracranial electroencephalogram (EEG) data from four dogs with naturally occurring epilepsy and eight patients with medication-resistant seizure during evaluation for epilepsy surgery, and identifying the earliest EEG changes leading to seizures with the fewest false alarms for an $8000 prize. The second phase of the competition will award $20,000 to analyze the same data set to predict seizures in advance of their clinical onset with the highest accuracy. Accurate seizure prediction and detection is critical to building devices that might prevent those seizures. Five hundred and four teams competed in these two challenges. An Australian software engineer won the first competition, and the second competition was won by a team of investigators that included an engineer, a mathematician, and other scientists.

Translating Discoveries into Cures, Therapeutics, and Preventions of Disease The translation from basic science discoveries into new therapeutics and products is often a long and

TABLE 35.2

difficult process. To deliver better health outcomes to their constituent patient groups impatient with the pace of progress and the inattention to development of therapeutics, many of the disease-specific organizations have a laser-like focus on accelerating this process. In many cases, the organization’s founders have significant business expertise that they turned to philanthropic purposes. Recognizing the time lost between completion of an experiment and the publication of results in scientific journals, more and more voluntary health agencies and philanthropic groups require the researchers they support to collaborate and share basic discovery data prior to publication as well as sharing tissue samples and other patient data.67 A number of VHOs have created separate nonprofit arms focused on drug development that work with biotechnology and pharmaceutical companies. By absorbing some of the costs and risks of early stage development, helping to accrue patients for clinical trials, and facilitating regulatory reviews, they create conditions encouraging commercial investment. Not surprisingly these organizations have been particularly important in neglected rare disease areas, where they are beginning to invigorate anemic or empty drug pipelines.68 Examples of these organizations are presented in Table 35.2 and a few examples are described below. The CFF69 was started in 1955 by a group of parents of affected children. At that time most children diagnosed with cystic fibrosis (CF) did not reach school age, little was known about its causes, and most physicians were unfamiliar with its treatment. As a rare disease, there was no concerted research effort and little research funding. Over its 55-year history the CFF established specialized treatment centers and funded research leading to the discovery of the CF gene in 1989. Notably, in 2000 CFF pioneered a nonprofit therapeutics development model, establishing the Cystic Fibrosis Foundation Therapeutics, Inc. as a nonprofit drug discovery and development arm supporting milestone-driven projects across all stages of drug development. One of those

Examples of Voluntary Health Organizations and Their Drug-Development Affiliates

Date Established

Parent Foundation

Date Affiliate Established

Drug Development Affiliate

1946

National Multiple Sclerosis Society

2007

Fast Forward

1950

Muscular Dystrophy Association

2009

MDA Venture Philanthropy

1955

Cystic Fibrosis Foundation

2000

Cystic Fibrosis Foundation Therapeutics, Inc.

1971

Foundation Fighting Blindness

2002

National Neurovision Research Institute

1998

Multiple Myeloma Research Foundation

2004

Multiple Myeloma Research Consortium

1949

Leukemia and Lymphoma Society

2007

Therapy Acceleration Program

This is not a comprehensive list.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

624

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

projects involved an investment of $150 million in the development of the drug Kalydeco, which is now sold by Vertex Pharmaceuticals. In November 2014, CFF sold its Kalydeco royalty rights for a cash payment of $3.3 billion.70 While no other VHO has duplicated the level of financial payback achieved by CFF, other VHOs have targeted a significant portion of their research funds to develop new therapeutics and diagnostics. The Leukemia and Lymphoma Society’s Therapy Acceleration Program (TAP) identifies and funds innovative projects related to therapies, supportive care, and diagnostics that have the potential to change the standard of care for patients with blood cancer.71 Currently the TAP pipeline has 10 products listed in Phase I, II, or III stages of clinical testing. Similarly, the Multiple Myeloma Research Foundation (MMRF)72 founded in 1998 by identical twin sisters Kathy Giusti and Karen Andrews soon after Kathy’s diagnosis of multiple myeloma, works to find new treatments. Kathy realized that the medical research model built on a system of competition for funding and publishing was not working to develop cures. MMRF was created to replace this system with a new patientcentered model of collaboration that is milestone driven and provides the needed infrastructure and collaboration across academic-industry sectors to drive innovation and develop new therapeutics. MMRF’s results have been impressive, with 10 drugs receiving FDA approval in the time it normally takes for one. In addition to establishing a patient registry, tissue bank, genomics initiative, and a research network, MMRF also created the MMRF Researcher Gateway to make key genomic and clinical data accessible to all scientists and clinicians.73

Establishing Product Development Partnerships Product Development Partnerships (PDPs) are another model designed to decrease the lag time between discovery and product development. PDPs, which can vary in their organization and processes, are nonprofit organizations created to enable public, private, academic, and philanthropic sectors to aggregate funding for and collaborate on the development of drugs, vaccines, and other health tools. Many PDPs have targeted developing therapeutics for neglected diseases that disproportionately affect the poor where commercial incentives are lacking and where the need is great. By investing in a portfolio of potential products, aggregating funds from different organizations and removing barriers to collaboration, PDPs minimize the risk and cost to each partner organization and increase the efficiency of the R&D process. The Bill and Melinda Gates Foundation has been a leader in developing this model by providing start-up funds to a variety of

PDPs. It is very likely that the PDP model will continue to play an important role in developing new drugs and diagnostics, particularly in areas for which a robust commercial market is absent, although a major challenge has been obtaining sustained funding for the relatively long time periods needed to develop and validate new products.74 Two examples of successful PDPs are the Drugs for Neglected Disease Initiative75 and the Medicines for Malaria Venture:76 • The Drugs for Neglected Disease Initiative (DNDi)77 was launched by Medecins sans Frontieres, also known as Doctors without Borders, in 2003 as a partnership with five public sector institutions. DNDi has delivered, along with its partners, combination therapies for malaria, late-stage sleeping sickness, and visceral leishmaniasis, and a pediatric dosage of Bensnidazole for Chagas disease. Recently, DNDi launched a 5-year project to develop affordable treatments for the millions of people suffering from Hepatitis C, with little or no access to the new highly effective, but very costly Hepatitis C drugs developed by Gilead Pharmaceutical and other companies. To address this inequality, DNDi’s Hepatitis C virus project aims to develop a short course of affordable easy-to-use, highly efficacious, safe, and all-oral pan genotypic regime that can be delivered widely as a public health intervention to the Hepatitis C epidemic. In 2012 the Wellcome Trust awarded DNDi a grant to support a Phase II clinical trial in Bolivia of a promising treatment for Chagas disease. And in 2013 DNDi received the Rockefeller Foundation’s “Next Century Innovators Award” as one of 100 innovators promoting positive and innovative social change worldwide. • The Medicines for Malaria Venture (MMV) aims to reduce the burden of malaria in disease endemic countries by discovering, developing, and delivering new effective and affordable antimalarial drugs.78 Its donors have included the Bill and Melinda Gates Foundation, the Rockefeller Foundation, the Wellcome Trust, the Australian government, Irish Agency for International Development (AID), the US NIH, and others. An early notable PDP success was MMVs partnership with Novartis to develop a much needed pediatric formula of the malaria drug Coartem in 2009.

Fostering Dissemination of Information, Data Sharing, and Patient Engagement Foundations, patient groups, advocacy organizations funded by philanthropy, and individual philanthropists have played increasingly important roles in: communicating science to the public; disseminating information

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

AREAS OF CONTRIBUTION

to inform patients and policy makers; raising the profile of the need for medical research; and fostering open access journals, and the timely sharing of data. The Kaiser Family Foundation, which evolved from a private grantmaking foundation to a private operating foundation and now to a public charity, is totally dedicated to providing trusted health policy analysis and information. In the words of its President, Drew Altman, we basically are a producer of policy analysis on domestic and global health policy; we are a producer of public opinion and survey research.,; we are a “go to” clearinghouse, synthesizer and translator of all the best information on the issues we work on, whether we produce it or others do; and we operate a major health care news service dedicated to producing the best in-depth coverage of health care and health policy called Kaiser Health News (KHN) the nation’s first not-for-profit health policy news service.79

In the United States today, the partisan political debate about the health system fueled by the influence of the $3 trillion dollar health industry has made independent, unbiased reporting and rigorous analysis of health data more critical than everdand the Kaiser Family Foundation has filled this important gap. Similarly, other philanthropies have advanced the movement for open access to research publications and the timely sharing of primary data. Two philanthropic grantsda 2001 grant from the Gordon and Betty Moore Foundation to support The Public Library of Science (PLoS) and a 2002 grant from the William and Flora Hewlett Foundation to support Creative Commonsd were seminal in launching the open access publishing movement. This movement has continued to grow with assistance from other philanthropic organizations. For example, the HHMI, the Wellcome Trust and the Max Planck Society helped establish the open access journal eLife,80 which does not charge authors fees and uses a unique collaborative review process where referees work together to identify a paper’s strengths and weaknesses. Authors also are given the option to make their accepted manuscripts openly available shortly after receiving a final decision. The HRA has formed a partnership with the National Library of Medicine to enable the grantees funded by more than 73 different HRA-member organizations to deposit their publications into PubMedCentral, enabling adoption of Public Access policies by HRA members. In addition, funders such as the Multiple Myeloma Research Foundation, the Leona and Harry Helmsley Foundation, and the Simons Foundation Autism Research Initiative are supporting central data warehouses to store data in interpretable formats that enable aggregation and data sharing. Data sharing, which more and more VHOs and other philanthropic funders are requiring, has the added benefit of helping

625

to ensure the reproducibility of research findings.81 Recognizing the need to develop new tools for analyzing large data sets, in 2013 the Simons Foundation also established the Simons Center for Data Analysis (SCDA). While its initial focus is genomics and neurosciences, SCDA intends to collaborate with researchers across other disciplines.82 Importantly, in reaction to the crisis of irreproducible research,83 philanthropic organizations that support biomedical research (along with other funders) are devoting more resources to evaluating approaches and to assuring the research they fund is reproducible, including in some cases a requirement for open access to primary data. As VHOs focus more attention on developing products faster and for less money, they are mandating sharing more information “precompetitively” to improve the efficiency of the product development process. Additionally, to target patient and family member needs, VHOs have created and operated “patientpowered” patient registries and research networks since as early as 1995. These registries and networks are distinguished from researcher-generated registries in that they are managed by patients and family members working with VHOs, instead of researchers and academic centers. Patient-powered registries and patient-powered research networks (PPRNs) offer new directions for patient-centered outcomes research, providing patients and family members another way to become engaged in research beyond the role of advisor or informant to researcher-generated studies.84 One example of a PPRN research project is the Parkinson’s iPhone Study that utilizes the iPhone appd mPowerdwhich was developed with support from the Michael J. Fox Foundation, the Robert Wood Johnson Foundation, and Parkinson’s Net.85 The iPhone study, which is managed by the nonprofit research organization Sage Bionetworks, collects data from patients via cell phones on the health of the patient and symptoms of Parkinson disease progression, including information on changes in balance, gait, memory, and speech. Patients also input information on when they take medicine to help determine the effects of the medicine on the symptoms being monitored. The goal of the program is to learn about variations in Parkinson’s disease, to improve the way these variations are managed, and to determine if mobile devices and sensors can help measure the progression of the disease. Launched in March 2015, this patient-centered path-breaking endeavor also enables participants to choose who can access their data. As of March 2016, over 75% of the over 12,000 mPower participants have chosen to share their data broadly with researchers and the first 6 months of data has been released by Sage Bionetworks.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

626

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

Advocating for Resources and Policy Changes The voice of committed patients and their families speaking through VHOs has steadily become more powerful as these organizations increasingly use networking and social media approaches to raise funds and awareness that spurs action. A prominent example is the ice bucket challenge sponsored in 2014 by the ALS Association that turned “viral” demonstrating the potential impact of social media in raising research funds and awareness. The ALS Ice Bucket challenge funds helped support a multinational study analyzing the DNA of 15,000 ALS patients (Project MinE) that recently identified a new ALS-associated genedNEK1.86 Nonetheless, even before “social networking” became a powerful tool to build public awareness, organizations like the Lasker Foundation, the Juvenile Diabetes Research Foundation, and the ACS devoted significant resource to direct public attention to their disease areas. Perhaps the most powerful example of advocacy is the seminal role VHO’s and foundations played at the end of the 20th century in galvanizing public attention on the exploding HIV/AIDS epidemic and advocating for policy changes and research funds. Better known by its acronymdACT UPdthe AIDS Coalition to Unleash Power started in 1987 in the United States and made history as a passionate advocacy group that raised public awareness to the unfolding tragedy of AIDS. ACT UP’s often disruptive approaches led to some important victories including decreasing the FDA approval time for some antiretroviral drugs (the antiretroviral AZT was approved in only 108 days) and persuading the BurroughseWellcome Company to lower the price of AZT.87 By the mid-1990s, responding to ACT UP and other advocacy groups, as well as the alarming expansion of the HIV/AIDS epidemic, government support of HIV research had substantially increased and a number of foundations, including the Bill and Melinda Gates

Interacting With the Philanthropic Sector Research scientists should be aware that philanthropic organizations are not just funders of scientific research, but also can be a scientist’s valuable ally in advancing their research and scientific career. Of course having an individual research grant not only makes you more competitive for academic positions, but it also increases your ability to secure protected time for your research. In addition, grants from the philanthropic sector are usually more flexible than government grants. Awardees of nonprofit organizations often have the option to use the award funds for purposes not allowed in federal grants, as well

Foundation and the DDCF, began committing funds to support both basic and clinical HIV research in the United States and abroad. However, three philanthropic organizationsdthe Aaron Diamond Foundation, the Foundation for AIDS Research (amfAR),88 and the Elizabeth Glaser Pediatric AIDS Foundation (EGPAF)89d stand out for providing significant support for HIV research in the late 1980s and early 1990s when the epidemic was still in its early days, there was relatively little government support for HIV/AIDS research, and few philanthropic funders were focused on HIV/ AIDS. While the Aaron Diamond Foundation and the related Irene Diamond Fund spent down their endowments and no longer exist,90 both EGPAF and amfAR still support AIDS research today. As the research system has become increasingly politicized over issues such as stem cell research or privacy issues that constrain sharing of data, the independent voice and support of VHOs is more essential than ever in protecting public health, increasing support for underfunded research areas, improving and streamlining clinical research regulations, and ensuring patient engagement in all aspects of translational and clinical research. Cancer research has benefited greatly from the voices of many different VHOs including the National Breast Cancer Coalition (NBCC). NBCC’s campaign to increase US government funding for breast cancer resulted in an unprecedented multimillion dollar US Department of Defense breast cancer research project, which has committed over $2.5 billion to breast cancer research since it was launched in 1991.91 Another noteworthy example is the aggressive campaign to make every state smoke-free and to increase tobacco taxes and funding for cancer research that has been launched by the Cancer Action Network, an affiliate of the ACS.92

as make mid-course corrections, which are not easy as an awardee of the federal government. Learning how to interact effectively with the nonprofit sector can be a very valuable skill. Savvy grant seekers include philanthropy in their search for research support and often a research scientist’s first interaction with a nonprofit funder is during the application process. Granted the nonprofit sector does not contribute nearly the amount of total grant dollars as the NIH; however, in many cases the success rate is much higher. To be successful, investigators must target the application to the organization’s mission and the Requests for Applications (RFA’s) stated goals. Study the

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

AREAS OF CONTRIBUTION

organization’s website and the RFA carefully. If your research aligns with the mission and the RFA’s goals, then apply, making sure to follow instructions exactly. Most funders, especially VHOs, have very targeted instructions. For instance, many ask for a lay summary that will be primarily reviewed by lay stakeholders (patients, family members) but applicants often completely ignore these funder-specific instructions and target the summary toward scientific reviewers. Another key piece of advice is to pick up the phone or send an email to the program officer listed if you have any questions, especially if you are unsure of the eligibility criteria or alignment of your research area to the funder’s mission. It’s a better use of a program director’s time and yours to identify a mismatch before submitting a nonresponsive application. On receiving the award, the same rules apply. Follow the instructions in the award package exactly and make sure you communicate with your funder. Once you are an awardee, it is up to you to make the most of this valuable opportunity. You can enhance the impact of the funding on your career as well as advance the mission of the funding organization by following five simple pieces of advice:

Communicate Be Appreciative Build Relationships Help Advance the Mission Follow the Rules

627

need to showcase awardees and their research advances to demonstrate the impact of their funding and to raise money.

Be Appreciative When you communicate your finding to the scientific community and the world, be sure to acknowledge grant support in publications, presentations, etc. Use social media to tout your findings and give a nod to your funding source as often as appropriate. Suggest competitive colleagues as potential reviewers and encourage colleagues and mentees to consider this funder for support. Let your funder know that you are someone who appreciates the support.

Build Relationships It is of great benefit to both you and the funder to establish personal relationships with those administering your grant. If you will be attending a scientific conference let your program officer know, and ask if there is a way you can help them while you are there. Attend if they sponsor a reception. Is there an opportunity for you to recruit applicants or talk to junior scientists while perusing posters? Funders also need awardees to do media briefings and to serve on committees. If you are asked and do not have the time, offer names of appropriate colleagues. Look for opportunities provided by funders that not only keep you engaged with them but also provide professional benefits such as serving on review committees. Being a reviewer, participating in media briefings and engaging in the other activities mentioned can provide early-career scientists with a platform to springboard their career by giving them visibility and experience. Just like in most aspects of life, funders often give first priority to those with whom they have established relationships.

Help to Advance the Mission Communicate Your funder can be your biggest advocate throughout your career. Let your program officer know about a publication or significant finding as soon as you candeven while embargoed. Funders are in a great position to generate publicity and increase the impact of your publications by effectively communicating your findings not only to the scientific community but also to a much wider audience as well. If funders know early about scientific discoveries they are often in an excellent position to help with patent protection, finding a licensee, etc. Of course your funder wants you to get recognition for your success, but making sure your discoveries are appropriately followed up also helps them advance their mission. Funders

Patients, caregivers, politicians, and the “lay public” in general often say that talking directly with scientists about their research and hearing the excitement in their voices about the impact of their research is inspirational. Funders recognize this and many bring in their earlycareer scientist awardees to fundraising events. Donors love to hear directly from the awardee about what it means to be a young investigator working on the front lines in research. Funders appreciate awardees offering to go out into the community or to talk to donors about their research. Philanthropic organizations that also do advocacy often turn to their scientists to talk about the science to science administrators and politicians. Not only can this advance the organization’s mission but can also boost a researcher’s profile. Thus, if you receive Continued

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

628

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

an invitation to be a speaker or panelist at a funder’s conference you should accept.

Follow the Rules Following the rules may seem obvious, but it is rare to find an awardee who actually complies with all funder’s policies. Submit progress reports and financial reports as requested and in a timely manner. Work closely with your host institution (including the tech transfer office when appropriate) and follow their policies as well. Make sure both you and your institution know and comply with your funders’ policies with respect to intellectual property issues and issues of open access publications and data. If something is not clear, contact your

CONCLUSIONS AND FUTURE DIRECTIONS The philanthropic sector’s contributions to biomedical research have grown throughout the 20th century and the first decades of the 21st century. Even though the philanthropic sector’s resources are dwarfed by the resources of either the public (government) or forprofit sectors, the sector has played a critical role in supporting biomedical research and related training activities and its relevance is likely to continue. Thus, there is a high probability that most clinical investigators will interact with philanthropic organizations at some point during their careers. This chapter provides examples of different activities supported by philanthropic organizations as well as general advice to help investigators interact with this sector. The following are some current trends in the sector’s support for biomedical research: • An emphasis on strong patient engagement in all aspects of the research endeavor led by VHOs. • A move toward milestone-driven or venture philanthropy that aims to accelerate the translation of knowledge into new therapies, preventions, and cures. • More funding mechanisms that foster collaborations, create networks, and establish patient registries to be shared beyond a single department or institution. • Increased partnerships across all sectors (for-profit, government, and philanthropic) to accelerate progress. • Mega gifts from higher worth individuals and families that transform research institutions and advance knowledge. • Increasing globalization with more philanthropic organizations willing to work internationally to support the best science and create global

program officer. Funders may be understaffed making it hard for them to keep up with the research advances stemming from their support. So help them (and help them to help you) by keeping them updated even outside the normal progress report cycle. The funders need this information to evaluate their programs and measure the impact of their funded research. Given the myriad of organizations and individuals that make up the philanthropic community, researchers seeking support have many opportunities to secure funding. Paying attention to the advice above should improve your success in identifying additional sources of support as well as landing and maintaining a grant.

interdisciplinary networks that address complex problems and share data. • Adaptation of successful entrepreneurial business approaches to philanthropic biomedical research programs. • A commitment by more not-for-profit funders to open science and data sharing that accelerates research and product development and helps to ensure the reproducibility of data. • Continued innovation using social networks and advances in information technology to help fund projects, collect patient data, and advance knowledge about what works and what does not. These trends are likely to continue, albeit they probably will be modified and leveraged in ways yet to be imagined. Several underfunded and/or politically fractious areas also may attract more philanthropic attention. Stem cell research and issues related to reproductive health where abortion and fetal research controversies have limited research support will continue to be politically difficult for government agencies to fund and thus are in need of more support from the philanthropic sector. In the United States, research on public health approaches to reducing gun violence have similarly been politicized leading to the elimination of funds from the US Congress to support the Centers for Disease Control’s research on the epidemiology of gun violence. A chronically underfunded research area that also needs increased attention by the philanthropic community as well as the government and for-profit sectors (including health insurance companies) is health services or implementation research, which seeks to determine how to efficiently deliver health interventions both in the developed and the developing world. Neglected diseases largely found in the developing world will continue to require philanthropic support to help overcome market forces that prevent industry from investing in treatments for these diseases.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

REFERENCES

As biomedical research involves more and more analysis of large data sets, investigators will need to be trained in informatics and computational skills and new career paths including mid-career training opportunities will need to be supported. Students, fellows, and active investigators seeking to develop skills in emerging and often cross-disciplinary areas should include VHOs and private foundations that are committed to developing human capital as potential providers of training opportunities. These trends and gaps are sure to influence the way that philanthropic organizations conduct their business and the projects that they choose to fund in the future. Investigators should be mindful of them as they consider philanthropic sources of support for their work.

SUMMARY QUESTIONS 1. What types of organizations make up the philanthropic sector, and what kinds of support do they provide to the clinical research enterprise? 2. Philanthropic support for biomedical research has occupied a distinct niche when compared with forprofit and government sectors. Why? 3. Discuss three major contributions that the philanthropic sector has made in advancing biomedical research.

References 1. Giving USA Annual Report on Philanthropy for Year 2014. Giving USA Foundation (website). www.givingusa.org. 2. Moses H, Matheson DHM, Cairns-Smith S, George BP, Laisch C, Dorsey ER. The anatomy of medical research: U.S. and international comparisons,. J Am Med Assoc 2015;313:174e89. http:// dx.doi.org/10.1001/jama.2014.15939. 3. Association of Medical Charities (website). www.amrc.org.uk. 4. U.K. Health research analysis. 2014 (website), http://www. hrcsonline.net/pages/uk-health-research-analysis-2014. 5. Spero JE. Charity and philanthropy in Russia, China, India and Brazil, the foundation center, ISBN 978-1-59542-476-1. 6. The Council on Foundations (website). www.cof.org/content/ foundation-basics. 7. The Foundation Center (website). www.foundationcenter.org. 8. The Foundation for the National Institutes of Health (website). www.fnih.org. 9. Reagan-Udall Foundation for the Food and Drug Administration (website). www.reaganudall.org. 10. CDC Foundation serving the Centers for Disease Control and Prevention (website). www.cdcfoundation.org. 11. McKeever BS. The non-profit sector in brief 2015: public charities, giving and volunteering. Urban Institute. National Center for Charitable Statistics (website). nccs.urban.org. 12. Gallin EK, Bond E, Califf RM, Crowley Jr WF, Davis P, Galbraith R, Reece EA. Forging stronger partnerships between academic health centers and patient-driven organizations. J Am Med Assoc 2013;88: 1220e4. http://dx.doi.org/10.1097/ACM.0b013e31829ed.2a7. 13. Guidestar (website). www.guidestar.org.

629

14. Health Research Alliance (website). www.healthra.org. 15. Health Research Alliance. Grants in the Health Research Alliance Shared Portfolio (gHRAsp) (website). www.healthra.org/ghrasp. 16. Grantmakers in Health (website). www.gih.org. 17. Research!America (website). www.researchamericainc.com. 18. Genetic Alliance (website). www.geneticalliance.org. 19. Faster Cures (website). www.fastercures.org. 20. International Cancer Research Partnership (website). www. icrpartnership.org. 21. Chernow R. Titan: the life of John D. Rockefeller, Sr. (NY): Random House; 1997. p. 561e70. 22. Rockefeller Democracy, Philanthropy, The Rockefeller Foundation and the American Experiment, page 24, The Rockefeller foundation (website). Rockerfellerfoundation.org. 23. Flexner A. Medical education in the United States and Canada. (NY): Carnegie Foundation for the Advancement of Teaching; 1910. 24. Markel H. Abraham Flexner and his Remarkable report on medical education: a century later. JAMA 2010;303:888e90. 25. The birth of philanthrocapitalism. Economist February 23, 2006 (website), www.economist.com/node/5517656. 26. Glass Pockets. A service of the Foundation Center (website). www. glasspockets.org/philanthropy-in-focus/eye-on-the-givingpledge/a-closer-look. 27. Oshinsky DM. Polio: an American story. (NY): Oxford University Press; 2005. 28. The Nobel Prize in Physiology or Medicine 1954. Nobelprize.org. Nobel Media AB 2014. Web. May 17, 2016 (website). www. nobelprize.org/nobel_prizes/medicine/laureates. 29. Jacobs CD. Jonas Salk: a life. (NY): Oxford University Press; 2015. 30. Hall Dobkin P. Historical perspectives on nonprofit organizations in the United States, Chapter 1. In: Renz D, editor. The Jossey-Bass Handbook of NonProfit management and leadership. 4th ed. Assoc John Wiley and Sons, Inc.; 2016, ISBN 978-1-118-85296-5. 31. Doris Duke Charitable Foundation (website). www.ddcf.org/ what-we-fund/medical-research. 32. Burroughs Wellcome Fund (website). www.bwfund.org. 33. Howard Hughes Medical Institute (website). www.hhmi.org. 34. Leukemia and Lymphoma Society (website). www.leukemialymphoma.org. 35. Pion G, Ionescu-Pioggia M. Bridging postdoctoral training and faculty position: initial outcomes of the Burroughs Wellcome fund career awards in the biomedical sciences. Acad Med 2003;78: 177e86. 36. Resources for Early Career Scientists Howard Hughes (website). www.hhmi.org/programs/resources-early-career-scientistdevelopment. 37. Excellence Everywhere, The Burroughs Wellcome fund, 2008 (website). www.excellenceeverywhere.org. 38. The Whitaker Foundation (website). http://www.whitaker.org/ about-us/iie-program-history. 39. HHMI News (2004) (website). http://www.hhmi.org/news/ hhmi-nibibnih-invest-35-million-interdisciplinary-phd-programs. 40. Reid TR. Where’s the war on Alzheimer’s? AARP Bull 2015 (website), www.aarp.org/health/brain-health/info-2015/alzheimersresearch. 41. Alzheimer’s Association (website). www.alz.org. 42. The Wellcome Trust (website). https://welcome.ac.uk. 43. The Biomarker Consortium (website). http://www. biomarkersconsortium.org. 44. Cancer Research UK (website). www.cancerresearchuk.org. 45. Global Biomarker Consortium, Alzheimer’s Association (website). http://www.alz.org/research/funding/global_biomarker_ consortium.asp. 46. The Broad Foundation (website). www.broadfoundation.org/ scimed.html. 47. New York Stem Cell Foundation (website). www.nyscf.org.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

630

35. PHILANTHROPY’S ROLE IN BIOMEDICAL RESEARCH

48. The Juvenile Diabetes Research Foundation (website). www. jdrf.org. 49. Christopher Reeve Foundation (website). www.christopherreeve. org. 50. Michael J. Fox Foundation (website). www.michaeljfox.org. 51. The Broad Institute (website). www.broadinstitute.org. 52. Janelia Research Campus. Howard Hughes Medical Institute (website). www.janelia.org. 53. Andrews JR, Shah NS, Gandhi N, Moll T, Friedland G. Multidrugresistant and extensively drug-resistant tuberculosis: implications for the HIV epidemic and antiretroviral therapy rollout in South Africa. J Infect Dis December 1, 2007;196(Suppl. 3): S482e90. http://dx.doi.org/10.1086/521121. Review PMID: 18181698. 54. The Africa Center (website). www.africacentre.ac.za. 55. Doris Duke Medical Research Institute Opens in South Africa, (2003) Philanthropy News Digest (website). philanthropy newsdigest.org. 56. HIV Pathogenesis Program at the University of KwaZulu-Natal (website). hpp.ukzn.ac.za. 57. The KwaZulu-Natal Research Institute for Tuberculosis, HIV (KRITH) at the University of KwaZulu-Natal (website). www.krith.org. 58. African Health Research Institute launches in South Africa, July 18, 2016. HHMI News (website). hhmi.org. 59. Developing Excellence in Leadership, Training and science in Africa (DELTAS Africa) program. Wellcome Trust (website). www. wellcome.ac.uk/Funding. 60. Innovations in Clinical Research Awards 2003, Doris Duke charitable foundation (website). www.ddcf.org/. 61. VISITECT@CD4. Omega Diagnostics, (website). www. omegadiagnostice.com/cd4. 62. Global Grand Challenges. Bill and Melinda Gates foundation (website). www.gcgh.grandchallenges.org/about. 63. Prostate Cancer Challenge, Prostate cancer foundation (website). www.pcf.org. 64. Dream Team Awards. Stand up to cancer (website). www. standup2cancer.org. 65. Prize for Life (website). http://www.prize4life.org/. 66. The Seizure Detection, Prediction Challenge. American epilepsy society (website). www.aesnet.org. 67. Groopman J. Buying a Cure: What business know-how can do for disease. The New Yorker January 2, 2008. 68. Rare Diseases, Orphan Products: Accelerating Research and Development. Institute of medicine. (Washington, DC): National Academies Press; 2010. 69. The Cystic Fibrosis Foundation (website). www.cff.org. 70. Deal by Cystic Fibrosis Foundation Raises Cash, Some Concerns, Andrew Pollack, November 19, 2014 New York Times (website).

71. 72. 73. 74.

75. 76. 77. 78. 79. 80. 81.

82. 83. 84.

85. 86. 87. 88. 89. 90.

91. 92.

http://www.nytimes.com/2014/11/19/business/for-cystic-fibrosisfoundation-venture-yields-windfall-in-hope-and-cash.html?. Therapy Acceleration Program, Leukemia and Lypmphoma society. (website). www.lls.org/therapy-acceleration-program. The Multiple Myeloma Research Foundation (website). www. themmrf.org. Muliple Myeloma Research Foundation (website). http://www. themmrf.org/about-the-mmrf. Financing, Incentives for Neglected Disease R&D: Opportunities and challenges. Drugs for neglected diseases initiative, June 19, 2011 (website). www.who.int/phi/news/phi_3_DNDi_submission_ CEWG_190611_en.pdf. Drugs for Neglected Diseases Initiative (website). http://www. dndi.org/. Medicines for Malaria Venture (website). www.mmv.org. Drugs for Neglected Diseases (website). www.dndi.org. Medicines for Malaria Venture (website). mmv.org. Kaiser Family Foundation; President’s message (website). http:// kff.org/about-us. eLIFE (website). www.elifesciences.org/about. AlQuraishi M, Sorger PK. Data sharing: reproducibility will only come with data liberation. Sci Translational Med 2016;8(338): 339ed.7. http://dx.doi.org/10.1126/scitranslmed.aafo968. Simons Center for Data Analysis. Simons foundation (website). www.simonsfoundation.org/simons-center-for-data-analysis. Reality check on reproducibility. Nature May 25, 2016;533(437). http://dx.doi.org/10.1038/533437a. Workman TA. Engaging patients in information sharing and data collection: the role of patient-powered registries and research networks. In: AHRQ community Forum white paper. Publication No. 13-EHC124-EF. Rockville, MD: Agency for Healthcare Research and Quality; 2013. mPower Mobile Parkinson Disease Study (website). http:// parkinsonmpower.org. Ice Bucket Challenge. The Amyotrophic lateral Sclerosis (ALS) association (website). www.alsa.org. Engel J. The epidemic, a global history of AIDS. Smithsonian Books, Books (NY): Harper Collins Publisher; 2016. The American Foundation for AIDS Research (website). www. amfar.org. Elizabeth Glaser Pediatric AIDS Foundation (website). www. pedaids.org. Irene Diamond Fund to close with final grants totaling $40 million, Philanthropy News Digest, December 31, 2012. (website): philanthropynewsdigest.org. The National Breast Cancer Coalition (website). www. breastcancerdeadline2020.org/about-nbcc. The American Cancer Society Cancer Action Network (website). www.acscan.org.

III. TECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH

P A R T I V

CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

36 Identifying, Understanding, and Managing Patient Safety and Clinical Risks in the Clinical Research Environment Laura M. Lee, David K. Henderson National Institutes of Health, Bethesda, MD, United States

O U T L I N E Identifying and Managing Clinical Risk in the Clinical Research Environment

Patient Safety and Clinical Event Reporting Systems 640 633

Building a Road map to Safe and High-Quality Care and Research Support: Applying the Principles of High Reliability in the Clinical Research Environment 635 Leveraging Patient Safety and Quality Improvement Techniques in the Conduct of Clinical Research 635 Proactively Assessing Clinical and Operational Risk 638 Continually Monitoring the Clinical Research Environment for Risk 639

IDENTIFYING AND MANAGING CLINICAL RISK IN THE CLINICAL RESEARCH ENVIRONMENT Clinical research is most often conducted in an environment that is part of a larger health-care system. Both the conduct of clinical research and the practice of clinical medicine involve risk. The extent to which risk is present in the hospital environment has been well documented in recent years. The Institute of Medicine’s groundbreaking report “To Err is Human” characterized the magnitude of the occurrence of medical errors during patient care in the United States, estimating that between 44,000 and 98,000 hospital-based deaths per year could be attributed to medical errors.1 In more than 15 years since this report was released, Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00036-8

Electronic Surveillance for Errors and System Failures

641

Patient Safety and Clinical Quality Measures

641

Assessing Clinical Research Participants’ Perceptions of the Clinical Research Experience

642

Conclusion

642

Summary Questions

643

References

643

multiple studies have provided supporting evidence for these staggering findings.2 In a 2016 study by Makary and colleagues that analyzed medical death data from four studies conducted from 2000 to 2008, the authors estimated that more than 250,000 deaths each year in the United States result from medical errors.3 In the study, authors highlighted that these results place medical errors third on the Center for Disease Control and Prevention’s list of leading causes of death in the United Statesdbehind the 611,105 deaths that are associated with heart disease and the 584,881 deaths attributable to cancer, and just above the 149,205 patients who die from respiratory diseases. Findings from a study by Van Den Bos et al. estimate that the annual cost of measurable medical errors, in 2008, was over $17 billion.4 According to data from the Centers for Disease

633

Copyright © 2018. Published by Elsevier Inc.

634

36. MANAGING PATIENT SAFETY AND RISK

Control and Prevention, nearly 2 million health careassociated infections occur each year; costing the United States an estimated $20 billion dollars.5 A 2011 study by Magill and colleagues estimated the number of health care-associated infections to be 721,800.6 These figures are unquestionably daunting; however, they do not include either the scores of near misses or latent errors (errors that never reach the patient or result in harm) or the system failures that influence a patient’s healthcare experiences negatively. Based on these striking mortality and health care-associated infections data, the fact that the health-care environment is fraught with risks and potential errors that must be identified and deftly managed if patients are to be cared for safely and appropriately seems incontrovertible. Traditional clinical research is also conducted in the context of these health-care risks. As in the case for clinical care, and as noted by several authors in this text, the conduct of clinical research is inherently associated with risk. During both the scientific and human subjects’ protection review processes, great effort is expended to estimate, calculate, and articulate the relative risk associated with each study drug, device, and intervention. This intense scrutiny at the protocol level works to improve the safety of subjects relative to the risks associated with the study question. Murff et al. describe additional risks that rarely are considered formally during the review of a clinical research protocol, including the clinical environment in which the research will be conducted, as well as system failures that are inherently associated with clinical medicine.7 The health-care environmentdwhether an inpatient unit, an ambulatory care clinic, or a community health centerdis a complex system that is influenced by multiple factors that contribute to, or mitigate, risk in the conduct of clinical research. Nolan describes a system as “a collection of interdependent elements that interact to achieve a common purpose.”8 If one applies this definition to the clinical care environment, examples of interdependent elements that one might consider include such factors as the institution’s culture (especially with respect to safety), the skill mix and competence of the staff, the availability of state-of-the-art equipment, and the quality of information systems, to name but a few. The health-care literature is rife with examples of system failures resulting in harm to patients; for example, significant medication errors (e.g., the death of the Boston Globe reporter, Betsy Lehman, who died of a massive chemotherapy overdose and the tragic loss of young Josie King),9e12 wrong site surgeries that often top the Joint Commission’s list of reported sentinel events,13 and the relative epidemic of health care-associated infections.6,14e16 Whereas any one of these events could be considered an error resulting from an individual provider’s negligence, systems

thinking compels us to consider these adverse events as failures in a series of interrelated and/or codependent processes or systems. In truth, in the current complex health-care environment such events almost invariably involve a series of missed opportunities to correct the errordhence they truly represent system failures.14 This shift in focusdfrom the individual to the systemd forces organizations to broaden their analyses of incidents and, thereby, broaden the impact of any improvements. Because of the endemic nature of errors and system failures in clinical care, investigators and review bodies must collaborate with the health-care practitioners with whom they entrust their participants’ safety to assure that the system/environment in which clinical research is conducted is safe and has the necessary infrastructure in place to support the study. Further, the research team must have strategies in place to monitor the clinical research environment to identify risks and clinical events that could contribute to adverse events and/or protocol deviations and to assure that processes are in place to prevent, mitigate, and manage risks. Clinical research programs must, therefore, embrace a system’s approach to managing risk associated with the conduct of clinical research. For the purposes of this chapter, the term “conduct of clinical research” refers both to activities outlined in the research study as well as those intrinsic clinical care activities that are essential to the successful implementation of a study but that may not be explicitly described in the research protocol. Examples of such clinical care activities include infection control measures, medication management procedures, the design of the physical environment, and information management. These critical clinical care functions often are assumed to be present and functioning at an optimal level to support the investigator’s study; however, investigators and/or institutional review boards may lack formal processes to assess the capacity of the clinical environment to support the study under review. The research team must engage the clinical care organization proactively to assure that the appropriate infrastructure is in place to provide care safely and efficiently to study participants. In collaboration with the health-care team, the research team should have processes in place to • identify clinical care functions that are critical to the success of the protocol; • identify and assess critical risk points of the clinical care processes that might place participants at undue clinical risk and/or compromise the integrity of the study; • monitor the clinical environment continually for adverse events, errors, near misses, and process failures; • assess systematically and thoroughly errors that occur;

IV. CLINICAL RESEARCH INFRASTRUCTURE

LEVERAGING PATIENT SAFETY AND QUALITY IMPROVEMENT TECHNIQUES IN THE CONDUCT OF CLINICAL RESEARCH

• establish an armamentarium of process improvement tools with which to manage process and system issues when they are identified; and • develop robust processes to communicate and learn from untoward events that occur in the clinical care environment in the spirit of organizational learning and continuous improvement. These processes and the tools and techniques described in this chapter can be implemented and managed at a variety of levels of an organization. If the research program resides in a large health-care system, many of these activities can be managed by the hospital’s patient safety and clinical quality enterprise, in collaboration with the research teams. However, these processes and tools effectively can be applied on a much smaller scale (i.e., an individual research unit) with the same degree of success. Regardless of where in the organizational structure these functions reside, the findings from these performance measure, risk mitigation, and improvement strategies should be communicated directly to the leadership and across the organization.

BUILDING A ROAD MAP TO SAFE AND HIGH-QUALITY CARE AND RESEARCH SUPPORT: APPLYING THE PRINCIPLES OF HIGH RELIABILITY IN THE CLINICAL RESEARCH ENVIRONMENT Nuclear power, aviation, chemical manufacturing, and aerospace are all complex industries in which a single error can result in catastrophic consequences; however, remarkably, these industries are considered incredibly safe. These enterprises adhere to a set of high reliability principles that are aimed at identifying and managing risks in their respective cultures. Sutcliffe describes high reliability organizations (HROs) as having a “collective mindfulness” that supports and promotes a culture in which all staff are encouraged to seek out and share all unsafe conditions or problem before the event compromises operations or service delivery.15 The five principles of high reliability are 1. 2. 3. 4. 5.

Preoccupation with failure Sensitivity to operations Resistance to simplify Commitment to resilience Deference to expertise

The National Institutes of Health Clinical Center (NIH CC) applies these principles to the design and management of complex patient care processes as well as to the design and conduct of clinical research support. An error or lapse in safety in the aforementioned industries’ processes can result in tragic outcomesdmuch like a lapse

635

in proper infection control in the care of a highly immunocompromised patient or a patient with Ebola virus infection can be catastrophic from a personal as well as an organizational perspective. The NIH CC staff learned a great deal from experiences preparing for and providing care for Ebola patients during the recent epidemic. Table 36.1 provides an overview of each of the principles of high reliability and how the NIH CC leveraged these concepts in the care of acutely ill Ebola patients.

LEVERAGING PATIENT SAFETY AND QUALITY IMPROVEMENT TECHNIQUES IN THE CONDUCT OF CLINICAL RESEARCH A first step in assuring that appropriate clinical care infrastructure is in place to support a planned study is to examine the research protocol to identify the clinical functions that will be required to support the conduct of the study safely. This process begins with the active engagement of the study investigators, the research team, and the patient care staff who identify, objectively and prospectively, steps in the research process that may place participants at risk. Health-care performance improvement tools such as flowcharting,16e18 failure mode and effects analysis (FMEA),18e24 and clinical quality performance measures can be applied effectively to the analysis and management of risk in the context of planning clinical research. These performance improvement tools are used for this purpose in our institution (NIH CC). We provide this specific information about our own institution to provide context for this discussion. The NIH CC occupies a unique position in the nation’s biomedical research establishment. The NIH CC is a distinctive and complex hospital whose primary mission is the support of science. High-quality clinical care is provided at the NIH CC in the context of clinical research, but the primary driver of that care is science. The NIH CC provides clinical research support for, and clinical care to, the research participants enrolled in the more than 1600 active clinical research protocols ongoing at the NIH CC. The NIH CC’s research portfolio differs substantively from most academic medical centers. Of the NIH Clinical Center’s approximately 1600 clinical research protocols nearly half are designed to study the natural history and pathogenesis of rare, often genetically determined rare diseases. The other half of the NIH Clinical Center’s clinical research portfolio is comprised of clinical trials. More than 90% of these clinical trials are phase I or phase II “proof-of-principle” or “first-in-human” translational trials. This unique intersection of clinical care and clinical research mandates that the NIH CC use myriad health-care performance

IV. CLINICAL RESEARCH INFRASTRUCTURE

636 TABLE 36.1

36. MANAGING PATIENT SAFETY AND RISK

Relating the Principles of High Reliability to the Management of Patients With Ebola Virus Disease

HRO Principle

Description

Application in the Care of Ebola Patients

Preoccupation with failure

All staff are aware of the potential for risk and harm and are encouraged continually to scan the environment for potential and real threats to safety.

• Staff were empowered to ask about, and looking for, the untoward outcomes that could result from our care processes. • The team used several tools from high reliability industries to help identify risky processes and behaviors before a catastrophic event occurs (e.g., Failure Mode and Effects Analysis, Root Cause Analysis). • The care team built in multiple layers of redundancy for high risk activities. For instance, the role of “Wat-San” was developed to assure that the critical tasks of donning and doffing personal protective gear were accomplished in a safe manner. The role of the “Wat-San” was to actively and meticulously direct the care providers during each step of the process. The Wat-San had the authority to stop all activity on the spot. • The essential nature of drilling and conducting “Day in the Life” exercises is borne of this “preoccupation with failure”ddrills were a central aspect of the care team’s preparation and continual readiness.

Sensitivity to operations

Maintaining a “situational awareness” is a hallmark of HROs. Leaders and staff need to be constantly aware of how processes and systems affect the organization.

• The care team used “safety huddles” liberally. In safety huddles staff gather briefly (5e10 min) to discuss issues/ concerns that have developed over the course of their tour of duty • The larger team, led by the hospital and unit medical leadership, “huddled” every day at 3 p.m. to review the events of each day. Unit leaders and staff used this real-time information to drive decisions/process changes. • Leaders and staff evaluated the effectiveness of their care processes continuously to identify opportunities for improvement.

Resistance to simplify

Staff appreciate that health care is complex, ever-changing and fraught with interdependencies. When faced with challenges, errors, or untoward events HROs seek to understand the “root cause and contributing factors” of these events rather than settle for a more superficial or expedient explanation. And whereas standardization of processes may be useful, HROs understand and actively manage complexity.

• Staff were encouraged to resist the risks associated with “painting with broad strokes” when evaluating lapses and failing to dig deeply to find the real source of a particular problem. • When issues were identified in the shift and daily huddles staff and leaders were encouraged to ask “WHY?” at least five times when investigating events and errors to assure that the true nature of the event was identified.

Commitment to Resilience

This principle is based on the assumption that errors and lapse will occur and organizations must actively design processes that allow an organization to “bounce back” from errors.

• Every organization must develop strategies to sustain operations and “bounce back” when (not if) an untoward event occurs • The question, “What if?” needs to end every process step designed in the care of high risk patients. • “What if” a staff experiences an occupation exposure?” • “What if” the public has a negative reaction to our work with Ebola patients?” • What if..

IV. CLINICAL RESEARCH INFRASTRUCTURE

LEVERAGING PATIENT SAFETY AND QUALITY IMPROVEMENT TECHNIQUES IN THE CONDUCT OF CLINICAL RESEARCH

TABLE 36.1

637

Relating the Principles of High Reliability to the Management of Patients With Ebola Virus Diseasedcont’d

HRO Principle

Description

Application in the Care of Ebola Patients

Deference to expertise

Staff closest to, and responsible for, the workflow and care processes is the most knowledge about how best to manage their work. HROs actively solicit input from frontline staff when designing processes of care and, in particular, at time of crisis or during an emergency.

• The frontline nursing, infection control, laboratory, transport, housekeeping and other ancillary staff were key members of the planning team from the beginning of the initiative. • Frontline staff were active participants at the daily leadershipled huddles. • A challenge was managing the well-intentioned “directives” of leadership staff who were removed from day to day bedside care but have a vested interest in the care of Ebola patients.

improvement tools creatively to manage the safe implementation of a broad spectrum of clinical research protocols effectively. One example of how performance improvement tools have been used to enhance the conduct of clinical research occurred in our institution as the severe acute respiratory syndrome (SARS) epidemic evolved in 2004. At that time, investigators at the NIH Clinical Center authored two protocols designed to gain insight into the epidemiology, pathogenesis, and natural history of this new disease as well as to assess strategies for the clinical evaluation and management of patients with SARS. These protocols perhaps positioned the NIH CC somewhat uniquely as one of the few health-care facilities in the world that was actively recruiting patients with SARS. The research protocols received rigorous scientific and human subject protections vetting and approval, and the principal investigators were poised to enroll their first participant/patient. However, several circumstances caused the NIH CC, as an organization, to pause before the decision was made to open recruitment. Because of the nature of the studies conducted at the NIH CC, many of the patients/participants recruited to participate in its clinical research protocols are highly immunocompromised, either as a direct result of their underlying disease or due to the interventions associated with the research studies in which they are enrolled. Further, at the time of the SARS outbreak, the NIH CC’s clinical environment was in a buildingd constructed in the 1950sdthat posed significant infrastructural hurdles to providing safe care to patients infected with highly infectious (and in this instance, possibly airborne) pathogens. In this complex clinical context, investigators from the National Institute of Allergy and Infectious Diseases submitted protocols requiring the provision of care for highly infectious SARS patients. Subsequently, these proposals were presented to the NIH community. As one might have anticipated, several investigators who provide care for

patients who have severely compromised immune function were adamantly opposed to admitting patients with SARS or other highly contagious respiratory illnesses electively to the NIH CC. However, because of the clearly urgent public health need, as well as the potential unprecedented scientific opportunities, the leadership of the NIH CC approached the issue not by asking: “Can we safely provide care to patients with SARS?” but rather by asking the question, “How can we care safely for all of our patients?” As the protocols were being reviewed for human subjects’ protection, the leadership of the NIH CC set out to identify the critical clinical functions that they felt must be present and operating at optimal levels to admit and care for SARS patients safely. This assessment required the collaboration of the research team, the leadership of the NIH CC, and the active participation of key clinical departments such as hospital epidemiology, nursing, critical care medicine, pharmacy, and housekeeping. Using flowcharting techniques,17,18 the team carefully cataloged each step in the research process and identified clinical care functions necessary to support the research requirements. Clinical and operational functions that were identified as being essential to the successful care of these patients are outlined in Table 36.2. This exercise was eye-opening in that both the clinical care providers and the investigators were astounded by the breadth of hospital functions that required flawless orchestration to assure that the protocols could be implemented safely. Following the identification of the key clinical and operational requirements, a team was charged with assuring that appropriate policies, procedures, staff, equipment, and physical infrastructure were in place and functioning optimally and efficiently prior to the admission of the first SARS patient. The clinical care team worked closely with the research team as well as with the community of NIH investigators to assure alignment with the study requirements and to time, appropriately, the admission of the first protocol participant.

IV. CLINICAL RESEARCH INFRASTRUCTURE

638 TABLE 36.2

36. MANAGING PATIENT SAFETY AND RISK

Critical Clinical and Operational Functions Required for the Safe and Effective Management of the Patient With Severe Acute Respiratory Syndrome

• Infection control • Isolation rooms - Availability - Capabilities • Appropriateness of current isolation precautions • Personal protective equipment (PPE) - Availability - Staff competence • Critical care medicine • Intensive care unit capacity • Staff competence • Respiratory therapy • Ventilator availability • Staff competence • Staff, patient, and family education • Personal Protective Equipment • Isolation precautions • Managing containment/isolation/quarantine • Laboratory medicine • Availability of tests/assays for clinical research and clinical care • Medication management • Supply chain issues • Equipment and supplies • Ventilators • Personal protective equipment • Social work/emotional support • Participant and family support • Patient transport • To the NIH Clinical Center • Within the NIH Clinical Center • Security • Transportation assistance • Crowd control • Housekeeping • Infection control training • Code blue • Exposure/transmission mitigation • Public relations/Communication • For staff, participants, families, public NIH, The National Institutes of Health.

PROACTIVELY ASSESSING CLINICAL AND OPERATIONAL RISK The development of a complete listing of the essential clinical processes that need to be in place is a critical first step to prepare for the implementation of a new research protocol. As was the case with the SARS protocol, this process can be daunting. Focusing an organization’s finite resources to assure that attention is paid to the most critical and potentially riskiest care processes is a challenge for most organizations. A variety of tools exist to guide the objective prioritization of what may ultimately be a very long list of critical issues that must be addressed. FMEA is a powerful risk assessment tool that provides a systems-based, human-factors-focused,

and objective methodology for identifying and prioritizing risk, with the ultimate aim of reducing patient harm and enhancing clinical research.19,21e28 Historically, FMEA has been used in the engineering environment to assess high-risk processes associated with power generation in the commercial nuclear power industry; in aviation to assess the acceptability of aircraft designs; and in the automotive industry to establish manufacturing requirements for cars and trucks.19,22,24 DeRosier and colleagues at the Veteran’s Administration’s Center for Patient Safety are credited with moving the techniques of FMEA to the bedside, applying the concept of prospective risk analysis to health-care processes.29,30 In 2002, the use of FMEA in health care further expanded with the issuance of a Joint Commission requirement that all health-care organizations seeking accreditation conduct at least one proactive risk assessment on a high-risk clinical process every 18 months.31 Clinical care practitioners as well as clinical research professionals can use FMEAs to identify risk and to avert adverse events, errors, and other system failures in a variety of health-care settings.19,20,22,23,27,28 In a complex care environment, where risk is compounded by the interplay of clinical medicine and clinical research, FMEA is a useful tool to guide risk mitigation by identifying critical risk points in clinical care and clinical research processes. In 2016, investigators at the National Institute of Mental Health initiated a clinical research protocol at the NIH CC to study the neurobiology of suicide and to identify risk factors for short and long-term suicidality. One of the study phases involved admitting actively suicidal patients. In response to this research protocol, the NIH CC set in motion a rigorous risk mitigation initiative. An interdisciplinary team comprised of patient safety professionals, research investigators, nurses, social workers, and hospital leadership used FMEA to evaluate potential risk points throughout the clinical research and patient care processes. The analysis was facilitated using QI Path, an FMEA software package. The team segregated the research protocol into three overarching care delivery processes as well as several subprocesses for analysis as illustrated in Fig. 36.1. Using a consensus decision-making model, Failure Modes were identified for each subprocess and assigned a Hazard Score. Hazard Scores were the product of three factors: the probability of failure occurring (range 1e5), the severity of the outcome (range 1e5), and the failure’s detectability (range 1e4). The maximum Hazard Score was 100. The Failure Modes were ranked by Hazard Score to guide prioritization for risk mitigation. Thirty-nine Failure Modes were identified. Hazard Scores ranged from 0 to 60 with a mean of 18.8. The majority of the Failure Modes skewed to “low risk”dlikely due to the NIH CC’s existing heightened safety

IV. CLINICAL RESEARCH INFRASTRUCTURE

PROACTIVELY ASSESSING CLINICAL AND OPERATIONAL RISK

639

FIGURE 36.1 Neurobiology of suicide protocol process and subprocess steps. NIH CC, The National Institutes of Health Clinical Center.

measures that were implemented in response to a past in-hospital suicide. However, nine Failure Modes had Hazard Scores of >30 (the top 20% of all Failure Modes) and were targeted for risk mitigation. Immediate interventions focused on the following Failure Modes with Hazard Scores of 60:

FMEA has been, and continues to be, a highly effective tool for identifying, characterizing, and prioritizing risk associated with complex patient care and clinical research processes. Findings from this consensus-driven, objective, and quantitative analysis can be used successfully to leverage organizational change and resource allocation.

• “Presence of Environmental Risk” • “Inadequate Nurse Staffing” • “Undetected Suicidality at Discharge”

Continually Monitoring the Clinical Research Environment for Risk

Other high-risk Failure Modes included “Inadequate Staff Training” (Hazard Score 45), “Patient Coerced to Enroll” (Hazard Score 36), “Patient Harms Self” (Hazard Score 36), and “Patient Elopement” (Hazard Score 30). Informed by the data from the FMEA, the patient care and research teams collaboratively developed risk mitigation strategies aimed at reducing the likelihood of patient harm or research lapses: a rigorous environmental assessment process was implemented to identify hazards posed by equipment, sharps, and the physical environment; the nurse skill mix as well as staffing plans were evaluated and adjusted based on evidence from the FMEA; and the medical and research teams developed strategies to assure that patients were thoroughly assessed for suicidality prior to discharge. In addition, two positive “unintended consequences” resulted from this analysis: • the process provided an effective forum for focused, deliberate discussions between the research team and clinical care staff regarding protocol requirements that otherwise might not have occurred; and • the care team’s “preoccupation with failure” and focus on the culture of patient safety were reinvigorated.

Using FMEA to identify the clinical risks points associated with the implementation of a clinical care or clinical research process is a critical first step in mitigating patient/participant risk. FMEAs, or other risk assessments, identify process points that are associated with increased risk to the patient/participant, to involved providers, and/or to the scientific integrity of the study. As illustrated in the example above, these risks can be addressed by myriad clinical and organizational interventions that are aimed at reducing those risks. The next step in mitigating risk is the deployment of strategies to assess the effectiveness of interventions, and to survey the research and care environments continually for other risks to the participants and to the study. Measurement is fundamental to assessment efforts and improvement in the quality of care. The history of health-care quality improvement and measurement dates to Florence Nightingale’s collection of mortality data and infection rates during the Crimean War, as well as to the work of Ernest Codman in establishing standards for hospitals in the early 1900s, including his provocative “end results hypothesis”

IV. CLINICAL RESEARCH INFRASTRUCTURE

640

36. MANAGING PATIENT SAFETY AND RISK

that led to a Semmelweis-like estrangement from the health-care establishment (see Chapter 1).32,33 In a classic paper in 1968 Avedis Donabedian recommended measuring health-care quality in three areas: structure (the characteristics of a health-care setting); process (what is done in the health-care setting); and outcomes (the status of the patient resulting from specific interventions).34 This paradigm remains the mainstay of modern health-care performance measurement programs and is the basis for the local, state, and federal programs designed to measure the quality of clinical care and identify health care-associated risks and adverse events. Clinical research programs, too, should implement processes to assess the performance of the clinical research enterprise systematicallydprimarily to assess for risks to the participants, investigators, and care providers, as well as for threats to the integrity of the study. Collecting and reporting adverse events that occur during the course of a research study is a mandatory component of both the research process and the protection of human subjects. However, event reporting in clinical research focuses on individual protocols, not on how the system of clinical research is performing as a whole. This “protocolcentric” focus fails to identify clinical care and clinical research system failures that potentially might impact participant safety across multiple studies. Murff and his colleagues35,36 have described the need for research teams to develop reporting systems that collect data about reportable adverse events, as well as “near misses” or “latent failures” in the clinical care and clinical research environments. Near misses or latent failures are errors that do not result in patient/participant harm; however, these events do have the potential to do harm if the circumstances of the event were somewhat different. Identifying and analyzing near misses or system failures provide the care and research teams the luxury of designing and implementing interventions to interrupt the error cycle prior to the occurrence of a serious error. Surveillance for errors, adverse events, and latent errors in the clinical care or research setting can be accomplished using a variety of strategies including1 event reporting systems2; electronic surveillance systems that utilize clinical triggers to identify errors; and3 analysis of clinical performance measurement data.

Patient Safety and Clinical Event Reporting Systems Since the 1980s most health-care patient safety and clinical quality programs have relied on voluntary or mandatory occurrence reporting systems (ORSs) as a critical source of data regarding clinical care errors and/or latent errors and near misses. These reporting systems are readily accessible and have the capacity to

provide detailed information about these kinds of events.18,34,36e46 The NIH CC has had a hospital-wide electronic ORS since the early 1980s. This voluntary electronic event reporting system captures more than 5000 reports per year. Events entered into the ORS span the spectrum of clinical care and clinical research eventsd from serious harmful errors to reports of service quality. The NIH CC has found the ORS to be particularly useful as a surveillance tool for identifying trends in latent failures in clinical care and clinical research processes that otherwise would likely not be identified. The following is an example of a potentially harmful near miss or latent failure that could have had a negative impact on clinical research, had the issues not been identified by using data from the NIH CC ORS. On review of data from the ORS, the Pharmacy Department Quality Officer noted a trend in administration events that was occurring associated with a specific investigational drug in a phase I clinical trial. The reports indicated that drug delivery was delayed on several occasions because the infusion pumps had inexplicably suspended infusion. Each time the infusion would stop, the nursing staff would troubleshoot the problem, requiring the infusions to be restarted several times during a single delivery, potentially resulting in a delay of study drug administration, and potentially adversely affecting drug levels and pharmacokinetics. Alerted to these administration errors via the ORS, the Pharmacy Quality Officer met with the study investigators, the clinical care staff, and the nursing staff. Collectively the group conducted an intensive review of the events. The common factor identified in each incident was that the infusion pumps stopped due to an “air in line” alert, although no air was noted in the tubing. All efforts to determine the cause of the alert generation were futile. Finally, the team contacted the research team who had developed and conducted the initial laboratory testing of the drug in an effort to identify a reason for the alerts. Following the review of the current medication administration procedure for this study, the research team noted that the initial safety testing for the drug was performed using a different brand of intravenous tubing than the brand stocked and used in the NIH Clinical Center. The team changed the procedure for administering the study drug, mandating a change in the brand of intravenous tubing used. No additional reports of misadministrations were reported during the remainder of the study. Whereas these incidents do not appear to have caused any harm to the participant or the study, the potential for harm to the participant and the study are obvious, and any future potential adverse events or protocol deviations were averted as a result of identifying this series of events via the NIH CC’s ORS. The success of voluntary event reporting is dependent on the organizational culture in which the

IV. CLINICAL RESEARCH INFRASTRUCTURE

PATIENT SAFETY AND CLINICAL QUALITY MEASURES

reporting system is deployed, as well as on the manner in which the staff and leadership of the organization use the data to drive improvements in care and research. Establishing a nonpunitive “just culture” that encourages the reporting of events, free of reprisal, is essential to maintaining a robust and meaningful reporting system. Equally important is an organizational commitment to using the data provided by staff to understand system and process errors and failures and to develop strategies to mitigate risk and improve care. Finally, organizations should be committed to sharing performance measurement data with the staff to keep them informed and aligned with institutional performance improvement strategies.

ELECTRONIC SURVEILLANCE FOR ERRORS AND SYSTEM FAILURES The nearly universal deployment of electronic clinical information systems in health-care settings provides a robust platform for identifying adverse events in clinical care as well as in clinical research. Electronic surveillance for adverse events has proven effective in identifying, in real-time, events such as adverse drug toxicities and interactions, health care-associated infections, and other iatrogenic injuries or events. This technology uses clinical triggers to signal the presence of potential errors or adverse events. Clinical triggers can include high-risk medications, select abnormal laboratory values (e.g., abnormal serum potassium levels, microbiology culture results), treatment interventions such as antidotes (e.g., Naloxone (Narcan), vitamin K), corrective procedures (e.g., chest tube insertions and dialysis) and unplanned intensive care unit admissions.9,47e52 Electronic event surveillance for clinical care and clinical research errors and latent failures provides a tool that administrative institutional leadership, clinicians, and research investigators can use to identify, mitigate, and report these events in a much timelier manner than traditional, voluntary incident reporting systems.

PATIENT SAFETY AND CLINICAL QUALITY MEASURES Another excellent source of information about the capacity of a hospital or other health-care organization to provide a safe environment in which to conduct clinical research is the organization’s patient safety and clinical quality performance measurement program. These measurement programs collect data that are used to assess the quality of the care and services provided to patients.

641

All hospitals and other health-care facilities accredited by the Joint Commission must have systems in place to measure, continually, a proscribed list of clinical activities.53 Most hospitals also participate in a variety of national and/or state clinical performance measurement activities, often as a condition of funding and certification.54 Regardless of the type of performance indicators used by an organization to monitor patient care processes, these measures, if well designed and appropriately implemented, provide valuable insight into the health-care organization’s management of critical patient care processes. Table 36.3 provides a list of frequently monitored processes of care. Performance measurement activity in the clinical research setting also should take into consideration important clinical metrics associated with the scientific protocol under study. In addition to national and state benchmarks for clinical performance metrics, clinical researchers should be mindful of the specific processes on which their studies depend and create appropriate performance metrics to track the success of those systems. Data from performance measurement indicators provide investigators with critical information to guide study planning and preparation. For instance, if a clinical research study intervention will be conducted on a highly immunocompromised patient population, the effectiveness of the hospital’s infection control processes becomes highly relevant. Quantitative and objective

TABLE 36.3

Examples of Clinical Care Performance Indicators

• Medication management • Medication errors • Pharmacist Interventions • Pain management • Reassessment for pain postintervention • Treatment delivery • Delays in treatment • Patient wait times • Invasive procedures • Complication rates • Returns to the operating room • Readmissions following outpatient procedures • Wrong site/person surgery • Infection control • Infection rates (e.g., central line-associated bloodstream infection, catheter-associated urinary tract infections, surgical site infections) • Hand hygiene compliance • Timing of surgical antimicrobial prophylaxis • Health-care worker vaccination rates • Patient falls • Transfusion management • Disease/diagnosis-specific measures • Acute myocardial infarction • Pneumonia • Heart failure • Stroke

IV. CLINICAL RESEARCH INFRASTRUCTURE

642

36. MANAGING PATIENT SAFETY AND RISK

data about the organization’s health care-associated infection rates and hand hygiene practices provide valuable information about system and process issues that might need to be addressed prior to recruiting and enrolling patients/participants. Assuring that the clinical care environment in which clinical research participants will receive safe and high-quality care that supports clinical research is a shared responsibility of the health-care organization’s leadership, the care providers, and the research team. Basic quality improvement tools such as flowcharting, prospective risk assessment methodologies (e.g., FMEA), and clinical quality and patient safety performance measures provide objective data to guide protocol planning and implementation strategies.

ASSESSING CLINICAL RESEARCH PARTICIPANTS’ PERCEPTIONS OF THE CLINICAL RESEARCH EXPERIENCE One aspect of quality that has been far less intensely addressed in the literature is the assessment of the quality of the care and services provided to research participants. In particular, few studies have addressed participants’ perceptions of their experiences with clinical research processes. In this section, we focus on the assessment of participants’ perceptions of their clinical research experiences. The clinical research community can look to the healthcare industry for guidance in determining how to assess participants’ experiences as research subjects. Although the processes of providing patient care in a hospital and the conduct of a clinical research study often differ, investigators and the care teams supporting the research must be mindful that at the center of these processes is a human being interacting with a health-care delivery system whether in a community hospital, a Clinical Translational Science Award unit, or an outpatient setting. How research subjects perceive the clinical research experience provides valuable insights for future improvement of both the clinical and scientific processes. The NIH CC has surveyed patient/participants using a Picker-derived survey instrument55 since the mid-1990s. The NIH CC is motivated to understand how patient/ participants perceived the care they receive during their participation in studies conducted at the NIH CC as one method to assure that the needs and expectations of this special group of individuals who volunteer to contribute to scientific discovery are met. In 1995, the Clinical Center partnered with the Picker Institute (the National Research Corporation (NRC) acquired the Picker Institute in 2001) to develop a method of eliciting patient/participant feedback about critical aspects of their clinical research experience at the NIH CC. The Picker Institute’s philosophy

of eliciting information from patients about their experiences was used to develop the NIH CC’s survey. The survey was tailored to the unique clinical research environment of the Clinical Center and was designed to include several questions addressing the experience of participating in clinical research. For the past 15 years, the NIH CC used these data to identify opportunities to improve our patient/participants’ experiences. Issues such as communication with clinical staff, attention to emotional support, and the participant’s understanding about the point at which he or she can cease participation in a study were identified as areas requiring focused review and attention. These issues and others were addressed using the NIH CC’s organizational performance improvement structure. Interventions were implemented and improvements were measured. Many of these issues would not have been identified as problematic had the NIH CC leadership not actively queried their research participants about their perceptions of these processes. In 2003, the leadership of the Rockefeller University Hospital Center for Clinical and Translational Science expressed interest in the NIH Clinical Center’s survey and, subsequently, partnered with the NIH Clinical Center and the NRCePicker develop a valid and reliable survey instrument specifically to measure participants’ perceptions of their clinical research experiences.56

CONCLUSION Patient safety, clinical quality, and efficient and effective processes of care delivery are of equal import to clinical care and clinical research. Irrespective of the approach taken, we believe that researchers and institutions involved in clinical research must collect data from a variety of sources, including the solicitation of perceptions from participants and staff input about their research experiences, to improve the conduct of clinical research continually. Developing clinical research programs that include structured approaches to collecting reliable information about factors that contribute to process failures and adverse events (including careful root cause analyses); mechanisms for assessing trends in process and outcome failures, structured approaches to identifying risk points prior to study implementation (e.g., FMEA); and obtaining participant insights about their perceptions of the clinical research experience will provide the necessary data to allow institutions and investigators to improve the clinical research experience. We believe these approaches to patient safety, clinical quality, and clinical research quality and safety will increase substantially the likelihood of successful completion of clinical studies.

IV. CLINICAL RESEARCH INFRASTRUCTURE

REFERENCES

SUMMARY QUESTIONS Which of the following is not a characteristic of an HRO? 1. 2. 3. 4. 5.

Preoccupation with failure Punitive approach to managing untoward events Resistance to simplify Commitment to Resilience Deference to expertise

Which of the following tools is used prospectively to identify risk in processes, procedures, and protocols? 1. 2. 3. 4.

Root Cause Analysis FMEA Ishikawa Diagram Pareto Chart

The evaluation of a hospital’s Central LineAssociated Bloodstream Infection rate is considered to be what type of measure? 1. Structure measure 2. Process measure 3. Outcome measure The person credited with establishing standards for evaluating hospitals is 1. 2. 3. 4.

Donald Berwick Florence Nightingale Ernest Codman Ignaz Semmelweis

References 1. Kohn L. To Err is Human. Building a safer health system. In: Io M, editor. National Academy Press; 1999. 2. James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf 2013;9(3):122e8. 3. Makary MA, Daniel M. Medical error-the third leading cause of death in the US. BMJ 2016;353:i2139. 4. Van Den Bos J, Rustagi K, Gray T, Halford M, Ziemkiewicz E, Shreve J. The $17.1 billion problem: the annual cost of measurable medical errors. Health Aff 2017;36(4). 5. Scott R. The direct medical costs of healthcare-associated infections in U.S. Hospitals and the benefits of prevention. In: Prevention CfDCa, editor. Atlanta, GA: Department of Health and Human Services; 2009. p. 10. 6. Magill SS, Edwards JR, Bamberg W, Beldavs ZG, Dumyati G, Kainer MA, et al. Multistate point-prevalence survey of health care-associated infections. N Engl J Med 2014;370(13):1198e208. 7. Murff HJ, Pichert JW, Byrne DW, Hedstrom C, Black M, Churchill L, et al. Research participant safety and systems factors in general clinical research centers. IRB 2006;28(6):8e14. 8. Nolan TW. System changes to improve patient safety. BMJ 2000; 320(7237):771e3. 9. Bates DW, Cullen DJ, Laird N, Petersen LA, Small SD, Servi D, et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. ADE Prevention Study Group. JAMA 1995;274(1):29e34.

643

10. Leape LL, Bates DW, Cullen DJ, Cooper J, Demonaco HJ, Gallivan T, et al. Systems analysis of adverse drug events. ADE Prevention Study Group. JAMA 1995;274(1):35e43. 11. Niedowski E. How medical errors took a little girl’s life. Baltimore Sun December 14, 2003;2003. 12. Roush W. Dana-Farber death sends a warning to research hospitals. Science 1995;269(5222):295e6. 13. Stahel PF, Sabel AL, Victoroff MS, Varnell J, Lembitz A, Boyle DJ, et al. Wrong-site and wrong-patient procedures in the universal protocol era: analysis of a prospective database of physician selfreported occurrences. Arch Surg 2010;145(10):978e84. 14. Reason J. Human error: models and management. BMJ 2000; 320(7237):768e70. 15. Sutcliffe KM. High reliability organizations (HROs). Best Pract Res Clin Anaesthesiol 2011;25(2):133e44. 16. Lau CY. Quality improvement tools and processes. Neurosurg Clin N Am 2015;26(2):177e87. viii. 17. Spath PL. Flow charting for quality improvement. J Qual Assur 1991;13(5):20e4. 18. Varkey P. Medical Quaity Management: Theory and Practice. Sudbury: Jones and Bartlett. 2010. 19. Ashley L, Armitage G, Neary M, Hollingsworth G. A practical guide to failure mode and effects analysis in health care: making the most of the team and its meetings. Jt Comm J Qual Patient Saf 2010;36(8):351e8. 20. Bramstedt KA. Failure mode and effects analysis as an informed consent tool for investigational cardiothoracic devices. ASAIO J 2002;48(3):293e5. 21. Coles G, Fuller B, Nordquist K, Weissenberger S, Anderson L, DuBois B. Three kinds of proactive risk analyses for health care. Jt Comm J Qual Patient Saf 2010;36(8):365e75. 22. Duwe B, Fuchs BD, Hansen-Flaschen J. Failure mode and effects analysis application to critical care medicine. Crit Care Clin 2005; 21(1):21e30. vii. 23. Sheridan-Leos N, Schulmeister L, Hartranft S. Failure mode and effect analysis: a technique to prevent chemotherapy errors. Clin J Oncol Nurs 2006;10:393e8. 24. Wreathall J, Nemeth C. Assessing risk: the role of probabilistic risk assessment (PRA) in patient safety improvement. Qual Saf Health Care 2004;13(3):206e12. 25. Browne A. Common cause analysis: focus on institutional change. In: Henriksen KBJ, Keyes MA, editors. Advances in patient safety: new directions and alternative approachesvol. 1; 2008. Rockville (MD). 26. Cody RJ. Anticipating risk for human subjects participating in clinical research: application of failure mode and effects analysis. Cancer Invest 2006;24(2):209e14. 27. Linkin DR, Sausman C, Santos L, Lyons C, Fox C, Aumiller L, et al. Applicability of healthcare failure mode and effects analysis to healthcare epidemiology: evaluation of the sterilization and use of surgical instruments. Clin Infect Dis 2005;41(7):1014e9. 28. Senders JW. FMEA and RCA: the mantras of modern risk management. Qual Saf Health Care 2004;13(4):249e50. 29. DeRosier J, Stalhandske E, Bagian JP, Nudell T. Using health care failure mode and effect analysis: the VA national center for patient safety’s prospective risk analysis system. Jt Comm J Qual Improv 2002;28(5):248e67. 09. 30. Stalhandske E, DeRosier J, Patail B, Gosbee J. How to make the most of failure mode and effect analysis. Biomed Instrum Technol 2003;37(2):96e102. 31. Commission TJ. Leadership (LD), LD-2eLD-3. In: Resources JC, editor. Oakbrook terrace, IL: Joint commission resources; 2011. p. LD-1eLD-3. 32. Loeb JM. The current state of performance measurement in health care. Int J Qual Health Care 2004;16(Suppl. 1):i5e9. 33. Luce JM, Bindman AB, Lee PR. A brief history of health care quality assessment and improvement in the United States. West J Med 1994;160(3):263e8.

IV. CLINICAL RESEARCH INFRASTRUCTURE

644

36. MANAGING PATIENT SAFETY AND RISK

34. Donabedian A. Measuring and evaluating hospital and medical care. Bull N Y Acad Med 1976;52(1):51e9. 35. Murff HJ, Forster AJ, Peterson JF, Fiskio JM, Heiman HL, Bates DW. Electronically screening discharge summaries for adverse medical events. J Am Med Inform Assoc 2003;10(4):339e50. 36. Murff HJ, Patel VL, Hripcsak G, Bates DW. Detecting adverse events for patient safety research: a review of current methodologies. J Biomed Inform 2003;36(1e2):131e43. 37. Anderson JE, Kodate N, Walters R, Dodds A. Can incident reporting improve safety? Healthcare practitioners’ views of the effectiveness of incident reporting. Int J Qual Health Care 2013;25(2): 141e50. 38. Berwick DM. The science of improvement. JAMA 2008;299(10): 1182e4. 39. Blumenthal D, Ferris TG. Safety in the academic medical center: transforming challenges into ingredients for improvement. Acad Med 2006;81(9):817e22. 40. Donabedian A. The evaluation of medical care programs. Bull N Y Acad Med 1968;44(2):117e24. 41. Kaplan HS, Fastman BR. Organization of event reporting data for sense making and system improvement. Qual Saf Health Care 2003;12(Suppl. 2):ii68e72. 42. Mitchell I, Schuster A, Smith K, et al. Patient safety reporting: a qualitative study of thoughts and perceptions of experts 15 years after ‘To Err is Human’. BMJ Qual Saf 2015;25(2):92e9. 43. Nuckols TK, Bell DS, Liu H, Paddock SM, Hilborne LH. Rates and types of events reported to established incident reporting systems in two US hospitals. Qual Saf Health Care 2007;16(3): 164e8. 44. Olsen S, Neale G, Schwab K, Psaila B, Patel T, Chapman EJ, et al. Hospital staff should use more than one method to detect adverse events and potential adverse events: incident reporting, pharmacist surveillance and local real-time record review may all have a place. Qual Saf Health Care 2007;16(1):40e4. 45. Pronovost PJ. Improving the value of patient safety reporting systems. In: Henriksen KBJ, Keyes MA, editors. Advances in patient

46. 47. 48.

49.

50.

51.

52.

53.

54. 55.

56.

safety: new directions and alternative approachesvol. 1. Rockville (Maryland): Agency for Healthcare Research and Quality; 2008. Thomas EJ, Petersen LA. Measuring errors and adverse events in health care. J Gen Intern Med 2003;18(1):61e7. Bates DW. Using information technology to reduce rates of medication errors in hospitals. BMJ 2000;320(7237):788e91. Handler SM, Altman RL, Perera S, Hanlon JT, Studenski SA, Bost JE, et al. A systematic review of the performance characteristics of clinical event monitor signals used to detect adverse drug events in the hospital setting. J Am Med Inform Assoc 2007;14(4): 451e8. Hripcsak G, Bakken S, Stetson PD, Patel VL. Mining complex clinical data for patient safety research: a framework for event discovery. J Biomed Inform 2003;36(1e2):120e30. Resar RK, Rozich JD, Classen D. Methodology and rationale for the measurement of harm with trigger tools. Qual Saf Health Care 2003; 12(Suppl. 2):ii39e45. Sammer C, Miller S, Jones C, Nelson A, Garrett P, Classen D, et al. Developing and evaluating an automated all-cause harm trigger system. Jt Comm J Qual Patient Saf 2017;43(4):155e65. Szekendi MK, Sullivan C, Bobb A, Feinglass J, Rooney D, Barnard C, et al. Active surveillance using electronic triggers to detect adverse events in hospitalized patients. Qual Saf Health Care 2006;15(3):184e90. Commission TJ. Performance improvement (PI), PI-4ePI-5. Comprehensive accreditation manual for hospitals. Oakbrook Terrace (IL): The Joint Commission; 2017. Sorian R. Measuring reporting, and rewarding performance in health care. 2006. p. 4e10. Jenkinson C, Coulter A, Bruster S. The Picker Patient Experience Questionnaire: development and validation using data from inpatient surveys in five countries. Int J Qual Health Care 2002;14(5): 353e8. Kost RG, Lee LM, Yessis J, Wesley RA, Henderson DK, Coller BS. Assessing participant-centered outcomes to improve clinical research. N Engl J Med 2013;369(23):2179e81.

IV. CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

37 Clinical Pharmacology and Its Role in Pharmaceutical Development 1

Sue Cheng1, Konstantina M. Vanevski2, Juan J.L. Lertora3

Bayer HealthCare Pharmaceuticals, Inc., Whippany, NJ, United States; 2Bayer HealthCare Pharmaceuticals, Inc., Basel, Switzerland; 3Duke University School of Medicine, Durham, NC, United States

O U T L I N E Functional Imaging Tools Related to Phase 0 Personalized Medicine Design and Conduct of Improved and Rigorous Phase IeII Studies With Adequate Exploration of the ExposureeResponse Relationship Modeling and Simulation and Model-Based Drug Development Advent of Pharmacogenetics and Pharmacogenomics

Clinical Pharmacology as a Translational Discipline 645 Definition and Scope 645 Overview of Drug Development

646

Current State of Affairs in Drug Development

647

Contribution of Clinical Pharmacology First in Human Study Starting Dose in First in Human Study Dose Escalation in First in Human Study Identification, Development, and Qualification of Biomarkers and Utilization of Functional Imaging Tools Qualifying New Biomarkers Safety Biomarkers Efficacy Biomarkers and Surrogate End Points

649 649 649 649

649 650 650 650

CLINICAL PHARMACOLOGY AS A TRANSLATIONAL DISCIPLINE Definition and Scope Almost 80 years after Paul Martini introduced the term “clinical pharmacology”1 there still is not a universally accepted definition of clinical pharmacology, and as a result, most take a broad view.2 For the purposes of this chapter, we can define clinical pharmacology as the translational discipline that deals with the study of drugs in humans in the context of clinical science and drug development.

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00037-X

652 652

652 653 654

The Role of the Regulatory Agency FDA and Clinical Pharmacology FDA and Drug Safety FDA and the Special Populations

654 655 655 656

Summary Questions

656

References

656

The American Society for Clinical Pharmacology and Therapeutics has adopted a working definition of the discipline that encompasses the spectrum of activities in drug discovery, development, regulation, and utilization.3 The American College of Clinical Pharmacology published a position paper in 1999 in which it stated that “optimization of therapeutic response through continual monitoring of drug therapy, is the ultimate goal of the discipline of clinical pharmacology.” This publication also defined the clinical pharmacologist as “.a person with a doctoral degree, who is actively and persistently involved in activities pertaining to optimizing therapeutics.”4

645

Copyright © 2018 Elsevier Inc. All rights reserved.

646

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

Breckenridge et al. pointed out that by virtue of their training, clinical pharmacologists bring detailed expertise on the mechanism of action of drugs, doseeresponse relationships, adverse effects, drug disposition, and pharmacokinetics (PK), as well as knowledge of their therapeutic use in medical practice.5 Furthermore, adequately trained clinical pharmacologists are indispensable for their ability to integrate basic and clinical biomedical sciences and they contribute to optimizing the design of clinical trials, particularly in the early phases of drug development (Phase IeII). Some also specialize in pharmacoepidemiology, biostatistics and pharmacometrics, and pharmacoeconomics, which are all important to the “development, testing, and practical use of medicines.”6 Thus, clinical pharmacologists play an important role in pharmaceutical industry research and development, academic translational research, and regulatory government agencies.

OVERVIEW OF DRUG DEVELOPMENT Drug discovery and development broadly can follow two different paradigmsdphysiology-based drug discovery, which follows compound screening and profiling based on physiological readouts, and target-based drug discovery, which begins by identifying the function of a possible therapeutic target and its role in disease.7 The new drug development process is carried out through a sequence of developmental and evaluative steps. In the United States, this is done under an Investigational New Drug Application (IND), which ultimately leads to submission of a New Drug application (NDA) (Fig. 37.1). The process includes preclinical

research and development and clinical trials, commonly divided into phases 1, 2A, 2B, and 3, and NDA review by the US Food and Drug Administration (FDA) [see Chapter 6]. For drugs that are shown to be effective and that can be administered with acceptable toxicity, the process results in NDA approval and marketing of the drug.8 In Europe, the drug evaluation is carried out through the Committee for Medicinal Products for Human Use [see Chapter 7]. If the data on quality, safety, and efficacy of the medicinal product are sufficient and accepted as valid by the Committee, a recommendation is sent to the European Commission to be transformed into a marketing authorization valid for the whole of the European Union.9 One of the recommendations in the FDA’s 2004 Critical Path Initiative (CPI) report suggests that new tools are needed earlier in the process to help identify promising candidate molecules and reduce the time and resources expended on the development of such candidate products. As a result the agency issued a new guidance in 2006, the FDA Guidance on Exploratory IND Studies, in which early Phase 1 exploratory approaches are described.10 These are consistent with the regulatory requirements for human subject protection but involve fewer resources; this accelerates the development of compounds that have shown promising results during the preclinical development by establishing very early if these compound behave in human subjects as expected from preclinical studies. These studies are known as “Phase 0” or “microdosing” trials, referring to the exploratory, first-in-human (FIH) studies conducted early in Phase I involving limited human exposure and are without therapeutic or diagnostic value. The goal of these is to collect

IND

NDA

Chemical Synthesis and Formulation Development Animal Models for Efficacy Assay Development Animal PK and PD

Dose Escalation and Initial PK

Proof of Concept and Dose Finding

Large Efficacy Trials with PK Screen

Animal Toxicology PK and PD Studies in Special Populations PHASE I

Preclinical Development

PHASE II

PHASE III

Clinical Development

FIGURE 37.1 The process of new drug development in the United States. Adapted from Atkinson Jr AJ. Introduction to clinical pharmacology. In: Atkinson AJ, Abernethy DR, Daniels CE, Dedrick RL, Markey SP, editors. Principles of Clinical Pharmacology, 2nd ed. San Diego: Academic Press pp. 1e7.

IV. CLINICAL RESEARCH INFRASTRUCTURE

647

CURRENT STATE OF AFFAIRS IN DRUG DEVELOPMENT

Basic Research

Industry - FDA Interactions During Development

Prototype Design or Discovery

Preclinical Development

Clinical Development Phase 1

Pre-IND Meeting

Phase 2

Phase 3

End of Phase 2a Meeting

Initial IND Submissions

End of Phase 2 Meeting

Ongoing Submission

FDA Filing/ Approval & Launch Prepartaion

Safety Update

Market Application Submission

Pre-BLA or NDA Meeting

IND Review Phase

Application Review Phase

FIGURE 37.2

IndustryeFDA Interactions During Drug Development. This figure depicts the extensive industryeFDA interactions that occur during product development, using the drug development process as a specific example. Developers often meet with the agency before submitting an Investigational New Drug (IND) to discuss early development plans. An IND must be filed and cleared by the FDA before human testing can commence in the United States. During the clinical phase, there are ongoing submissions of new protocols and results of testing. Developers often request additional meetings to get FDA agreement on the methods proposed for evaluation of safety or efficacy, and also on manufacturing issues. PD, pharmacodynamic; PK, pharmacokinetics. Adapted from FDA Critical Path Initiative e Challenges and Opportunities Report e March 2004. http://www.fda.gov/ScienceResearch/SpecialTopics/CriticalPathInitiative/CriticalPathOpportunitiesReports/ucm077262.htm.

preliminary PK/PD (pharmacodynamic) data of the investigated compound(s), including receptor-binding imaging data, and do not address safety or efficacy since they utilize subtherapeutic doses. The preliminary data that a Phase 0 trial produces greatly aid the “go/no go” decision-making process and the ranking of candidate drugs.11e13 Phase I studies are conducted to evaluate the safety and PK and PD properties of the compound(s), and to determine the maximum tolerable dose limits. These studies are conducted in a limited number (20e100) of either healthy volunteers or patients. Phase II studies are subdivided as Phase IIA, those aimed at exploring drug efficacy and “proof of concept,” and Phase IIB, those aimed at establishing optimal doses to be used in the target population; this provides dosing regimens to be evaluated in Phase III. The objective of Phase III in the clinical drug development process is to demonstrate the safety and efficacy of the compound(s) for clinical use in a randomized controlled trial (RCT) enrolling larger patient populations. Additionally, they are aimed at assessing dosing ranges in different patient strata as well as in special populations (children, patients with impaired renal and/or hepatic function, etc.).14 Phase I and II studies are part of the “learning” phase that ideally defines the target clinical population and the dose or dose range to be used in Phase III trials, the “confirming” phase of the “learn-and-confirm” paradigm of

drug development. Safety continues to be evaluated throughout and includes postmarketing surveillance (Phase IV) after FDA approval and marketing, so that the “learning phase” continues throughout the life cycle of the drug.15 An overview of the clinical phases of drug development is provided in Fig. 37.2 and in Table 37.1.16

CURRENT STATE OF AFFAIRS IN DRUG DEVELOPMENT Over the past couple of decades, the FDA has increased its requirements for drug testing, mainly to address safety issues. A new drug requires an average of 15 years to reach the market today and approaches a billion dollars in research and development.17 Unfortunately, much of the expense is spent on drugs that never reach the market. Only one in 10 drugs that enter clinical testing receives eventual FDA approval, in spite of millions of dollars spent on preclinical testing.18 It is of great concern that the failure rate has increased up to 50% during phase III for drugs that have shown evidence of effectiveness in Phase II.19 This very high attrition rate, together with the enormous costs of drug development and the impending patent expirations, is the main driving forces behind the increased tendency to conduct clinical trials in the developing world in an

IV. CLINICAL RESEARCH INFRASTRUCTURE

648 TABLE 37.1

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

Overview of Clinical Drug Development

Phase

Objectives

Questions

Phase 0 Exploratory Studies Microdosing

To collect preliminary PK/PD data of the investigated compound(s) To identify a lead candidate(s) based on desired attributesdaid in “Goeno Go” decision

Will the drug bind to the target receptor population (PET scanning)? What dose range should be studied in early clinical trials given the uncertainty in the predicted dose required for efficacy and safety?

Phase I Traditional First in Human Studies Dose Ranging

To determine the maximum tolerable dose limits To assess toxicity and to collect safety data to support initial dosing in humans and dose escalation strategies To evaluate the safety and PK and PD properties of the compound(s)

What is the maximum tolerable dose? What are the PK attributes of the new compound(s) in initial human studies? What are the PD effects of the new compound(s) in initial human studies?

Phase IIA Proof of Concept (PoC)

To demonstrate efficacy in the intended patient population (per indication) and establish PoC

What are the attributes of the drug in target population compared to the existing therapy (i.e., standard of care)?

Phase IIB

To establish optimal doses to be used in target population, thus establishing dosing regimens to be evaluated in Phase III To aid in decision-making regarding design of Phase III trials

What trial design for a PoC/Phase IIB study will clearly demonstrate efficacy in the target population? Consider: patient populationestratification and number and strength(s) of doses What are the critical attributes that may favor PoC study over a larger Phase II study? Consider: duration of the study, optimal sample collection schemes, and types of biomarkers, surrogates, or end points responses that need to (should) be measured.

Phase III

To demonstrate safety and efficacy for clinical use

Do the selected dose(s) demonstrate the desired safety and efficacy in the population? What are optimum dose ranges in different patient strata as well as in the special populations?

Phase IV

To conduct safety surveillance (pharmacovigilance) during the postmarketing phasedmay or may not be required by the regulatory agencies.

What, if any, are the rare or long-term adverse effects in a much larger and diverse patient population?

PD, pharmacodynamic; PET, positron emission tomography; PK, pharmacokinetics

effort to contain costs; at the same time this approach raises new safety and ethical concerns. The US FDA has recognized these issues and addressed them in the CPI, highlighting the slowdown instead of the expected acceleration in innovative medical therapies reaching patients.20 FDA has published an “opportunities list” of issues that lie along the critical path of drug discovery, translation, development, and approval that greatly reemphasizes the utility of clinical pharmacology in modern drug development.21 The industry experts are in agreement with this approach, reiterating that the principles of clinical pharmacology ought to be “more aggressively employed and integrated into all stages of drug development supporting the push to decrease costs and improve timelines.”22 The goal is to promote more efficient drug development and evaluation as well as improved quality, safety, and effectiveness of the products FDA regulates. The critical path achievements are in three main areas: developing biomarkers and other scientific tools, streamlining clinical trials, and ensuring product safety.23 The work to identify and

qualify biomarkers that can personalize therapy and make clinical trials more efficient and safer included the discovery of genetic markers associated with adverse events (AEs) and predictive safety testing. To streamline clinical trials, FDA also launched the Clinical Trials Transformation Initiative, sponsored training courses, and issued draft guidance documents of adaptive and noninferiority clinical trial designs. To ensure product safety prescription drug labels are available electronically on the Internet and ongoing cooperative efforts with the Centers for Medicare and Medicaid Services and the Assistant Secretary for Planning and Evaluation have been expanded to include other federal partners to develop active surveillance methodologies and conduct “medical product-adverse outcome” queries. For example, as result of this collaboration, FDA’s existing safety surveillance capacity was enhanced to provide close to real-time vaccine safety monitoring of seasonal and H1N1 influenza vaccines, and the Dual Antiplatelet Therapy Collaboration to design and conduct a postmarketing study of drug-eluting stents was established. 23 In

IV. CLINICAL RESEARCH INFRASTRUCTURE

CONTRIBUTION OF CLINICAL PHARMACOLOGY

addition, FDA has created the Sentinel System for medical product safety surveillance, which includes the Active Postmarket Risk Identification and Analysis system and Postmarket Rapid Immunization Safety Monitoring system for vaccine safety.24 Further analysis demonstrated that sentinel analytic tools can produce findings similar to those produced by a highly customized protocol-driven assessment.25

CONTRIBUTION OF CLINICAL PHARMACOLOGY In the context of the opportunities list, the science and principles of clinical pharmacology remain at the forefront of the strategies and tactics being implemented across the pharmaceutical industry to improve the odds of success. After knowledge is acquired in preclinical phases including molecular studies in single-cell preparations, research is expanded to investigations in animal models followed by human clinical trials. By its very nature clinical pharmacology is translational.26,27

First in Human Study The process of clinical development of new drugs or biologic agents follows a set of preclinical studies that have been conducted in animals. Before a new chemical entity (compound) can be administered to humans, many important properties, such as its physiologic effects, mechanism of action, toxicity, absorption, distribution, metabolism, and elimination, have to be characterized through experimental pharmacological analysis during the preclinical phase of drug development. Various in vitro and in vivo systems are utilized to achieve these goals including isolated tissue, purified enzymes and receptor systems, isolated whole organs, and intact animals. FDA’s guidance on Safety Testing of Drug Metabolites encourages the identification of any differences in drug metabolism between animals used in nonclinical safety assessments and humans as early as possible during the drug development process.28 The discovery of additional drug metabolites that could be pharmacologically active and/or toxic in the late phases of drug development can potentially cause delays in clinical development and marketing. Starting Dose in First in Human Study The objectives of the FIH study are to assess the safety, tolerability, PD, and PK profile of the candidate drug or biologic agent. The maximum recommended starting dose (MRSD) for a FIH study should be determined based on the relevant preclinical information, including the toxicology profile of the compound in all the tested

649

species, the pharmacologically active dose (PAD), and the PK (absorption, distribution, metabolism, and excretion (ADME)) of the investigational agent. According to the FDA Guidance of Estimating the Maximum Safe Starting Dose in Initial Clinical Trials for Therapeutics in Adult Healthy Volunteers,29 there are five steps to be considered for determining the MRSD, which include the following: defining the No Observed Adverse Events Level (NOAEL in mg/kg) in toxicity studies, converting each animal species NOAEL to a human equivalent dose (HED), picking the HED from the most appropriate animal species, choosing a safety factor (generally 10) to reduce the HED, and further lowering the dose based on the observed PAD in animals and other relevant factors that may impact the MRSD level. This FDA guidance also provides an alternative PKguided approach that uses plasma concentration (instead of dose) for animal-to-human extrapolation and defines a target systemic exposure (e.g., assess systemic exposure (AUC)) coupled to predicting clearance (CL) in humans using allometric scaling or physiologically based PK modeling. Dose Escalation in First in Human Study Dose escalation schemes should be decided on a case by case basis. Real-time PK data to define AUC and compare to NOAEL AUC and real-time PD results are essential in making a decision about next steps. In general, rapid dose escalation (2
the MRSD) is done initially until the target PK, PD, or safety end point is reached and then more cautious (e.g., 1.5
) escalation steps are implemented. Additional elements that should be considered when developing the dose escalation scheme include steepness of dose/concentration-response curve, seriousness and reversibility of toxicities, ability to detect a potential AE in humans, nonlinear PK with dose escalation, and systemic exposure compared to NOAEL AUC. This approach is generally consistent with the guidance from European Medicines Agency (EMA).30

Identification, Development, and Qualification of Biomarkers and Utilization of Functional Imaging Tools One of the greatest challenges in the process of drug development is the ability to predict the performance of a new (potential) drug as early as possible and with the highest degree of certainty. Some of the most important signposts along the development pathway are the quantitative measures of the characteristics that reflect “normal biological processes, pathogenic processes, or pharmacological responses to a therapeutic intervention” in animals or humans, which are known as biomarkers.31

IV. CLINICAL RESEARCH INFRASTRUCTURE

650

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

Categories of biomarkers include the following: susceptibility/risk, diagnostic, monitoring, prognostic, predictive, PD/response, safety, and surrogate end point. The safety and surrogate end point biomarkers are discussed in more detail in sections Safety Biomarkers and Efficacy Biomarkers and Surrogate End Points.32 Qualifying New Biomarkers The vast majority of the biomarkers used in drug development today have been in use for many years. However, they have been empirically derived and often lack predictive and explanatory power. Additionally, there are a large number of potential new biomarkers that have been proposed, but their utility has not been evaluated yet. This evaluation work identified as biomarker qualification was addressed in the CPI Opportunities List, which led to the development of a formal biomarker qualification process by the FDA and the respective legal framework provided in the Qualification Process for Drug Development Tools guidance first issued in 2010.33 In this document the agency identifies possible types of drug development toolsdbiomarkers that assess various biological characteristics, including “genetic composition, receptor expression patterns, radiographic or other imaging-based measurements, blood composition measurements (e.g., serum enzyme levels, prostate-specific antigen), electrocardiographic parameters, or organ function (e.g., creatinine clearance, pulmonary function tests, cardiac ejection fraction).” The FDA guidance Qualification Process for Drug Development Tools (updated in January 2014) includes, but is not limited to, biomarkers, clinical outcome assessments (COA), and animal models for drug development. Biomarkers can be used to select patients for inclusion in a clinical trial, identify safety problems related to a candidate drug, or predict an eventual benefit from treatment. COA can be used to determine the treatment benefit of the drug. The COA should be a well-defined and reliable assessment of a targeted concept(s) in a specified context of use based on adequate and well-controlled investigations. The Animal Model Qualification program focuses on the animal models intended for use in the suitable efficacy studies that serve as substantial evidence of effectiveness for drugs developed under the Animal Rule.34 Safety Biomarkers Following the failure of major promising drugs and their consequent withdrawal from the market (Table 37.4), it has become clear that there is an urgent need for development of predictive safety biomarkers for use during the drug development process. The animal-to-human test sequence provides an ideal setting in which to evaluate the predictive value of new markers of organ toxicity. Their use in animal toxicology studies

would contribute to greater effectiveness of the safety screening process prior to introducing new drugs into humans, provide for more adequate selection of initial human doses, and help target toxicity monitoring in early trials. Further, they can be utilized as predictors of interindividual variability, drug disposition, and acceptable drug toxicity during clinical development and in the postmarketing phase. For example, postmarketing safety monitoring increasingly utilizes genomic biomarkers in an effort to identify targets that may lead to significant toxicity and other life-threatening adverse reactions (Table 37.2).35e37 A major step forward in this field was the conclusion of the first pilot and joint qualification process for biomarkers between the United States FDA and the EMA in 2009 that resulted in qualification of seven renal biomarkers for drug-induced kidney injury, namely: KIM-1, Albumin, Total Protein, b2-microglobulin, Cystatin C, Clusterin, and Trefoil Factor-3.37e40 Other safety biomarkers of interest include toxicogenomic analysis, translational biomarkers of hepatotoxicity, and biomarkers of skeletal muscle toxicity. While reactive drug metabolites that may induce clinical adverse effects in humans have been well established (e.g., acetaminophen and isoniazid hepatotoxicity) and studied, another area of safety dealing with assessment of nonreactive toxic drug metabolites has gained attention in the recent years. Only recently adequate regulatory guidance has been implemented requiring the efficient identification and profiling of circulating human metabolites in the early stages of clinical development thus contributing to the overall efforts to improve the safety profile of the compounds before Phase III trials begin. A good example is the Metabolites in Safety Testing Committee of the Pharmaceutical Research and Manufacturers of America and the dialogue it conducted with the US FDA.41 Efficacy Biomarkers and Surrogate End Points Efficacy biomarkers are to be used for improving the predicting of doseeresponse characteristics as well as for monitoring response during treatment. PD (or activity) biomarkers are used to assess the change in disease status that has occurred in a patient following a therapeutic intervention. They may be treatment specific or more broadly informative of disease status, i.e., if there is a response to the given pharmacologic intervention (e.g., blood pressure, cholesterol, HbA1C, intraocular pressure, radiographic measures, and Creactive protein). A surrogate end point is a biomarker that is used to predict clinical benefit. A clinical end point, on the other hand, is the final or defined outcome that is used to measure the drug effect; it is a direct measurement of how a patient feels, functions, or survives. Changes in

IV. CLINICAL RESEARCH INFRASTRUCTURE

CONTRIBUTION OF CLINICAL PHARMACOLOGY

TABLE 37.2

651

Examples of Pharmacogenetic Information in FDA-Approved Drug Labels

Drug Name

Pharmacogenetic Test

Label Information (Relevant Excerpts)a

Abacavir (Ziagen)

HLA-B*5701

Patients who carry the HLA-B*5701 allele are at high risk for experiencing a hypersensitivity reaction to abacavir.

Irinotecan (Campostar)

UGT1A1*28

When administered in combination with other agents, or as a single-agent, a reduction in the starting dose by at least one level of CAMPOSTAR should be considered for patients known to be homozygous for the UGT1A1*28 allele.

Warfarin

CYP2C9 VKORC1

The maintenance dose needed to achieve a target PT/INR is influenced by: • Clinical factors including age, race, body weight, sex, concomitant medications, and comorbidities, and • Genetic factors (CYP2C9 and VKORC1 genotypes).

Carbamazepine Phenytoin

HLA-B*1502

Patients with ancestry in genetically at-risk populations should be screened for the presence of HLA-B*1502 prior to initiating treatment with Tegretol. Patients testing positive for the allele should not be treated with Tegretol unless the benefit clearly outweighs the risk.

Trastuzumab (Herceptin)

HER2

Detection of HER2 protein overexpression is necessary for selection of patients appropriate for Herceptin therapy because these are the only patients studied for whom benefit has been shown. Assessment for HER2 overexpression and of HER2 gene amplification should be performed by laboratories with demonstrated proficiency in the specific technology being utilized.

Tetrabenazine (Xenazine)

CYP2D6

Patients should be genotyped for CYP2D6 prior to treatment with daily doses of tetrabenazine over 50 mg. Patients who are poor metabolizers should not be given daily doses greater than 50 mg.

Afatinib (Gilotrif)

EGFR exon 19 deletion or exon 21 substitution (L858R) positive

The first-line treatment of patients with metastatic non-small cell lung cancer (NSCLC) whose tumors have epidermal growth factor receptor (EGFR) exon 19 deletions or exon 21 (L858R) substitution mutations as detected by an FDA-approved test Limitation of Use: Safety and efficacy of GILOTRIF have not been established in patients whose tumors have other EGFR mutations.

Ibrutinib (Imbruvica)

del (17p)

Indicated for the treatment of Chronic lymphocytic leukemia (CLL)/Small lymphocytic lymphoma (SLL) with 17p deletion (1.3).

Imatinib mesylate (Gleevec)

BCR-ABL1

Indicated for the treatment of newly diagnosed adult and pediatric patients with Philadelphia chromosome positive chronic myeloid leukemia (PhþCML) in chronic phase (1.1).

Carisoprodol (SOMA)

CYP2C19

Patients with Reduced CYP2C19 Activity: SOMA should be used with caution in patients with reduced CYP2C19 activity. Published studies indicate that patients who are poor CYP2C19 metabolizers have a fourfold increase in exposure to carisoprodol, and concomitant 50% reduced exposure to meprobamate compared to normal CYP2C19 metabolizers. The prevalence of poor metabolizers in Caucasians and African Americans is approximately 3%e5% and in Asians is approximately 15%e20%.

a

All labels were last accessed on 12-28-2016 at http://www.fda.gov/Drugs/ScienceResearch/ResearchAreas/Pharmacogenetics/ucm083378.htm.

surrogate end point biomarkers can often be detected earlier than the clinical end points, and the use of a qualified surrogate end point can accelerate the development process for a treatment breakthrough. Before a biomarker can be accepted as a surrogate end point, however, there are two major issues to be addressed. There has to be a high level of confidence that changes in the marker consistently predict the desired clinical end points, and there also must be a comprehensive and thoughtful discussion of possible

long-term risks and safety issues (e.g., trials using a surrogate end point for effectiveness can be shorter and may not evaluate longer-term risks or adequately address potential safety issues).36,42 FDA has provided examples of biomarkers used as outcomes in development of approved new molecular entities and new biological therapeutics (e.g., hepatitis C virus RNA, human immunodeficiency virus RNA, sputum culture conversion to negative, and parasite count resolution for infectious disease; CD34 positive cell count,

IV. CLINICAL RESEARCH INFRASTRUCTURE

652

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

complete blood count, tumor burden by Bcr-Abl, tumor burden by Philadelphia chromosome positive cells, plasma methotrexate, splenic volume, serum asparaginase, serum testosterone, and tumor burden by imaging for oncology).43 Functional Imaging Tools Related to Phase 0 Imaging modalities long have been crucial to the researcher in observing changes, either at the organ, tissue, cell, or molecular level, both in animals and humans, responding to physiological, environmental, or pharmacological stimuli. Functional imaging tools that are noninvasive and are used in vivo (e.g., optical imaging, positron emission tomographydPET, and single photon emission computed tomographydSPECT) have become important especially in the preclinical and clinical development phases to assess animal and human target engagement and to inform dose selection for proof-of-concept and later-phase trials. They hold vast potential for use as biomarkers for many purposes during the drug development process, such as establishing and measuring treatment efficacy, aiding in patient stratification, and improving diagnosis.44,45

Personalized Medicine The notion of “personalized medicine” has become an integral part of modern drug development and delivery. The term describes targeted therapeutic approaches to individual patients and patient groups based on precise classification of the disease status and application of pharmacogenomics and other targeted diagnostic tests to optimize drug dosing, therapeutic benefit, and safety.46 For therapies directed at molecular targets, personalized approaches to therapy help identify the “responders” versus “nonresponders” and who should receive the therapy (e.g., breast cancer and tamoxifen). Pharmacogenomics is defined as the study of drug exposure and/or response as related to variation in DNA and RNA characteristics. Drug exposure is estimated from the PK profile (e.g., AUC) following administration, and drug response is PD (effect of a drug on physiologic and pathologic processes including those related to effectiveness and those related to adverse reactions). Pharmacogenetics is a subset of pharmacogenomics and is defined as the study of variations in DNA sequence as related to drug response. Pharmacogenetics can help identify and distinguish poor versus rapid metabolizers and help predict suboptimal dosing resulting in lack of efficacy (rapid metabolizers) or overdosing causing serious side effects in the poor metabolizers (e.g., warfarin therapy).47 Utilizing this knowledge for development of rigorous dosing protocols based on a patient’s unique genetic profile may prevent dose-related clinical

toxicity. Early drug development typically focuses on gene variants relevant to the drug’s PK profile for absorption, distribution, metabolism (including formation of active metabolites), and excretion (ADME processes), and not on gene variants directly related to the drug’s PD that can predispose to toxicities such as immune reactions, or that code for intended or unintended drug targets and other biological pathways related to the drug’s pharmacologic effect. The new -omic technologies (genomics, proteomics, and metabolomics) have contributed to advances in clinical pharmacology and hold great promise as sources of new and powerful diagnostic, prognostic, and predictive biomarkers of drug efficacy and toxicity.48 For example, FDA has approved several in vitro genomic diagnostic tests for drug-metabolizing enzymes that detect specific genetic variations which may influence an individual’s response to treatment and can identify patients who are at high(er) risk for serious toxicity or other life-threatening side effects from certain therapies because the recommended doses are too high for them (Table 37.2). As a result, an increasing number of diagnostic laboratories carry out pharmacogenetic tests while the FDA already has approved modifications to many drug labels that now contain pharmacogeneticrelated information and/or appropriate warnings.49 FDA’s regulatory processes have evolved in response to, and in anticipation of, scientific developments that are critical for the development of personalized therapeutics and diagnostics.50

Design and Conduct of Improved and Rigorous Phase IeII Studies With Adequate Exploration of the ExposureeResponse Relationship One of the potential shortcomings in clinical drug development relates to the current approach to clinical trials. As a result, there is an increased interest in changing the approaches to conducting clinical trials from the simply “assessing the outcomes and producing better ones” paradigm, to making an assessment of what can be learned from successes and failures. Over the past 60 years, in parallel with the high number of newly developed drugs, the number of RCTs has grown tremendously, providing a cornerstone for evidence-based medicine. At the same time, the needs for application of new trial designs have increased as a result of both new knowledge and new technologies that provide for more efficient application of clinical pharmacology in the modern drug development process.51 In fact, FDA has recognized these needs in the CPI 2004 and in the Opportunities List, which resulted in issuing new guidance documents, for example, the Draft

IV. CLINICAL RESEARCH INFRASTRUCTURE

CONTRIBUTION OF CLINICAL PHARMACOLOGY

Guidance for Adaptive Design Clinical Trials for Drugs and Biologics published in February of 2010. Adaptive designs: In recent years, the use of adaptive design methods in clinical research and drug development based on accumulated data has become very popular due to its flexibility and efficiency; it allows for changes in design or analyses to be made based on the accumulated data at any point of the trial. This approach can contribute to improvement of Phase II trials to find the right dose for confirmatory Phase III clinical trials. One example for application of the adaptive trial design is the “I SPY 2 TRIAL” (Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2) conducted in an effort to study a “rapid and focused clinical development of paired oncologic therapies and biomarkers.”52,53 Further, as a result of the now common practice of obtaining plasma concentration measurements during clinical trials for PK/PD analysis of efficacy and adverse effects, a “randomized concentration controlled trial” (RCCT) has evolved in which a deliberate attempt is made to hold the plasma concentration within a defined range. However, RCCTs have not been widely adopted because they are complex to manage in comparison to the traditional fixed-dose trials and are difficult to implement in the postmarketing phase.54,55 In recent years, there has been an increased interest in the Bayesian approach to clinical trials, that is the “Learn and Apply” approach,56 as it permits adjustments during the trial as new information becomes available; this allows for updates on the prior assumptions on which the trial was based. Exposureeresponse (or dose response) relationship describes the change in observable effect in an organism (response) that occurred as a result of administration of differing doses or concentrations of a given drug (exposure). Phases 1 and 2 are ideal settings in which this relationship should be explored, and the observed responses (e.g., nonclinical biomarkers, potentially valid surrogate end points, or short-term clinical effects) can be used to collect critical information pertinent to the following: 1. Extrapolating findings from animal to human studies 2. Providing primary evidence for the efficacy and safety of a given compound 3. Providing evidence for the proof of concept 4. Guiding the design of initial clinical end point trials that use a plausibly useful dose range Furthermore, well-designed population studies of the exposureeresponse relationship can provide for an integrated understanding of dose, exposure, patient characteristics, and response as related to the efficacy and tolerability of a given compound.57 These improved approaches to clinical trials, along with a sufficient exploration and critical interpretation

653

of the exposureeresponse relationship, will certainly contribute to improving the odds of choosing more adequate dosing strategies for Phase III trials, and increasing the success rate along the pipeline.58

Modeling and Simulation and Model-Based Drug Development PK and PD modeling and simulation are wellrecognized powerful tools that enable effective implementation of the learn-and-confirm paradigm in drug development.59 These tools can be used to expedite the identification and testing of appropriate drug doses, particularly drugs that have a narrow range of safe and effective exposure. This in turn, provides for more informed dose/regimen selection, which could lead to increased trial success; it can eliminate one or more dosing arms from the protocol, reduce total number of patients recruited, and possibly reduce the length of the clinical trial. For example60: 1. Using qualified biomarkers of patient response, new PK sampling times and frequencies can be identified. 2. The knowledge of the optimal PK/PD properties of one compound can be used to optimize dosing of a backup candidate compound. 3. Can support conduct of more efficient trials of shorter duration for future compounds. There are numerous opportunities for modeling and simulation to facilitate the drug development process and regulatory decision-making, including these61: 1. “Off-label use”: A great example of such cases is when pediatric patients are treated with drugs that were approved for adults but adequate dosing information is not available for the pediatric population. In these cases, there is a great concern about using suboptimal doses as well as the safety of such dosing. These cases provide an excellent opportunity for utilization of modeling and simulation techniques. The approval of etanercept for pediatric patients with rheumatoid arthritis (RA) is an example. It was based on modeling and simulation studies that utilized known patterns of effectiveness and safety as well as established dosing regiments for the adult population suffering from RA.62 2. Diseases that have become resistant to current therapies, or for diseases for which an effective treatment still is not available (populations with unmet needs). Population-based modeling methods are particularly useful to identify trends in the data; for example, these methods identify sources of variability in a given

IV. CLINICAL RESEARCH INFRASTRUCTURE

654

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

population that could influence drug PK, estimate the magnitude of intersubject variability, and find reasons for higher exposure in special populations. Finding a population model that most adequately describes the data is of great importance in a setting where there is a need to individualize dose regimens for specific patient(s) or a patient population; this is important particularly for drugs with a narrow therapeutic range.63 Model-based drug development programs integrate knowledge of the pathophysiology of the disease, drug attributes, and patient characteristics to predict the range of possible outcomes, and these are proving to provide better time and cost-effective use of knowledge in the drug development process and regulatory decision-making.64,65 Pharmacometrics develops and applies mathematical and statistical methods to “characterize, understand, and predict a drug’s PK, PD, and biomarker outcomes behavior.” 66 The process of modern drug development and regulatory decision-making often involves development or estimation of the PK/PD, PD-outcome links, disease progress models, and clinical trials simulations, spanning the spectrum from basic research into disease and mechanisms of drug action to the rational use of medicines in patient care.67

Advent of Pharmacogenetics and Pharmacogenomics Pharmacogenomics has recently become an integral part of the drug development process. More than two decades of pharmacogenetic studies have described the genetic polymorphisms found in the major metabolizing enzymes, and as a result, clearly established the genetic traits responsible for interindividual differences in patients’ drug metabolism. These monogenetic traits have a predictable influence on the PK and the pharmacological effects of a large number of commonly prescribed drugs. This knowledge has been used to develop clinical genotyping methods that can be used by pharmaceutical companies to screen patients prior to initiating drug therapy. Prospective screening of Phase I volunteers for drugmetabolizing enzymes’ polymorphisms is conducted routinely at a number of pharmaceutical companies. With the advent of the pharmacogenomics, more patients enrolled in Phases II and III clinical trials are being genotyped in an effort to correlate drug efficacy with the genetic markers that are predictors of the PD. Currently there are a numerous pharmacogenomics markers that provide useful diagnostic tools to evaluate prospectively dosing regimens, given that multiple genetic polymorphisms may lead to altered drug absorption, distribution, and elimination as well as gene mutations that influence target drug receptors.68

The incorporation of pharmacogenomics into clinical drug development offers the opportunity for more informed drug evaluation based on the knowledge of the effect that specific genetic variants would have on drug response. Further, prospective genetic testing will optimize patient stratification and ensure inclusion of important representative phenotypic subgroups, thus impacting the efficiency of drug development.47,68 Pharmacogenomics assessment in early-phase clinical studies may focus on three aspects: first, identifying the populations that should receive lower or higher doses of a drug, considering genetic effects on drug exposure, doseeresponse, early effectiveness, and/or common adverse reactions; second, identifying the responder populations according to phenotypic, receptor, or genetic characteristics; last, identifying high-risk groups, in which drugs generally cause serious adverse effects. Pharmacogenomics studies can support mechanistic understanding of the drug, which can be used in the design or analysis of later trials, potentially improving their efficiency or likelihood of success if the genetic factor can help predict the likelihood and magnitude of response.69,70 The pharmacogenomic data and prescribing recommendations in product labeling derived from adult studies could be interpreted as applying directly to pediatric patients, but may not always be appropriate. It may be necessary to conduct pharmacogenomic studies directly in children in certain cases.71 Epigenetics is the study of heritable changes in gene expression (active vs. inactive genes) that do not involve changes to the underlying DNA sequence, a change in phenotype without a change in genotype, which in turn affects how cells read the genes. Epigenetic regulation is of importance for drug metabolism, the occurrence of drug resistance in anticancer treatment, and monitoring drug treatment efficacy. Pharmacogenomic and epigenomic biomarkers contribute differentially to interindividual variation in drug response.72

THE ROLE OF THE REGULATORY AGENCY As a science-based regulatory agency whose primary mission is to protect the public and ensure drug safety and efficacy, the FDA is a critical contributor in meeting the public health challenges of the 21st century [see Chapter 6]. Specifically, FDA has pioneered advances in regulatory sciences that facilitate drug development while adhering to rigorous scientific methodologies, and FDA scientists are active contributors to the field of drug development. Over the past decade, in an effort to facilitate much needed modernization of its drug evaluation and

IV. CLINICAL RESEARCH INFRASTRUCTURE

THE ROLE OF THE REGULATORY AGENCY

regulatory process, the agency has identified several key issues that have contributed to the stagnation in drug development in the CPI, (March 2004). Next was published the Critical Path Opportunity List (March, 2006) in which six crucial areas in need of novel tools and approaches were identified in an effort to stimulate the drug development process. These are better evaluation tools, streamlining clinical trials, harnessing bioinformatics, moving manufacturing into the 21st century, developing products to address urgent public health needs, and specific at-risk populations.73 These documents provided a great opportunity for the science of clinical pharmacology to vigorously address the challenge of “getting the dose right.”74 Specifically, there is great potential for the application of clinical pharmacology tools such as quantitative pharmacology and PK and PD modeling as well as developing disease state models, design and simulation of phase III clinical trials, and estimation of benefiterisk probabilities. Furthermore, clinical pharmacologists are eminently equipped for defining optimal dosing schedules, using therapeutic drug monitoring for individualized dosing regimens, and identifying biomarkers as potential predictive tests of drug efficacy and toxicity.75 In line with these possibilities, FDA’s regulatory actions are increasingly reliant on the principles of clinical pharmacology as reflected in the publication of multiple clinical pharmacology guidance documents. Clinical Pharmacology Guidance Documents,76 as with all FDA guidance documents, are nonbinding written documents published by the FDA to communicate the scientific and regulatory requirements to drug manufacturers, as well as to set the much needed standards for the drug development process (Table 37.3).

TABLE 37.3

655

FDA Clinical Pharmacology Guidance Documents

Guidance DocumentdYear Publisheda Drug Metabolism/Drug Interaction Studies in the Drug Development Process: Studies In Vitrod1997 General Considerations for Pediatric Pharmacokinetic Studies for Drugs and Biological Productsd1998 Pharmacokinetics in Patients with Impaired Renal Functiond1998 Population Pharmacokineticsd1999 In Vivo Drug Metabolism/Drug Interaction StudiesdStudy Design, Data Analysis, and Recommendations for Dosing and Labelingd1999 Pharmacokinetics in Patients with Impaired Hepatic Function: Study Design, Data Analysis, and Impact on Dosing and Labelingd2003 ExposureeResponse RelationshipsdStudy Design, Data Analysis, and Regulatory Applicationsd2003 Pharmacokinetics in PregnancydStudy Design, Data Analysis, and Impact on Dosing and Labelingd2004 Clinical Lactation StudiesdStudy Design, Data Analysis, and Recommendations for Labelingd2005 Drug Interaction StudiesdStudy Design, Data Analysis, and Implications for Dosing and Labelingd2006 Pharmacokinetics in Patients with Impaired Renal FunctiondStudy Design, Data Analysis, and Impact on Dosing and Labelingd2010 Drug Interaction StudiesdStudy Design, Data Analysis, Implications for Dosing, and Labeling Recommendationsd2012 Clinical Pharmacogenomics: Premarket Evaluation in Early-Phase Clinical Studies and Recommendations for Labelingd2013 General Clinical Pharmacology Considerations for Pediatric Studies for Drugs and Biological Productsd2014 Physiologically Based Pharmacokinetic AnalysesdFormat and Contentd2016 a

http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm064982.htm.

FDA and Clinical Pharmacology The role of clinical pharmacology in regulatory decision-making has developed markedly since its early stage at the FDA. To date, clinical pharmacology has significantly impacted on risk/benefit and labeling decisions at the FDA mainly through a focus on dose optimization via PK/PD evaluation. The FDA has engaged regulatory processes and created initiatives to expedite patient access to promising therapies. There is a flexible regulatory procedure to accelerate regulatory review and approval of new drugs under certain conditions. FDA established a new classification of drugs designated as “breakthrough” therapies in 2012. A breakthrough therapy is defined as a drug “intended to treat a serious or life-threatening disease or condition and preliminary clinical evidence indicates that the drug may demonstrate substantial improvement over existing therapies on one or more clinically significant endpoints, such as substantial treatment effects observed

early in clinical development.” Clinical pharmacology is essential to the conceptualization of experimental designs and generation of evidence sufficient to update regulatory decisions. The regulation has played an important role in minimizing uncertainty about the drug’s benefiterisk profile.77

FDA and Drug Safety Modernization ActdFood and Drug Administration Amendments Act of 2007 was motivated by the market withdrawals of several high-profile drugs (Table 37.4) due to unexpected safety concerns; this has provided FDA with new and increased authorities, among which are the requirements for postapproval trials and surveillance and safety labeling.78 Since there has been ever increasing attention to safety, identifying approaches to maintain the fine

IV. CLINICAL RESEARCH INFRASTRUCTURE

656 TABLE 37.4

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

Examples of High-Profile Drugs Withdrawn From the Market

Drug

Reason for Withdrawal

Troglitazone (Rezulin) 2000

Risk of hepatotoxicity (superseded by pioglitazone and rosiglitazone)

Trovafloxacin (Trovan) 2001

Risk of liver failure

Rofecoxib (Vioxx) 2004

Risk of myocardial infarction

Lumiracoxib (Prexige) 2007e2008

Progressively withdrawn around the world because of serious side effects, mainly liver damage

Efalizumab (Raptiva) 2009

Increased risk of progressive multifocal leukoencephalopathy

Gemtuzumab ozogamicin (Mylotarg) 2010

Withdrawn in the United States due to increased risks of venoocclusive disease and based on results of a clinical trial in which it showed no benefit in acute myeloid leukemia

Rosiglitazone (Avandia) 2010

Withdrawn in Europe because of increased risk of heart attacks and death. This drug continues to be available in the United States.

Cisapride (Propulsid) 2000s

Withdrawn in many countries because of risk of cardiac arrhythmias

Sibutramine (Reductil/Meridia) 2010

Withdrawn in Europe, Australasia, Canada, and the United States because of increased cardiovascular risk

Propoxyphene (Darvocet/ Darvon) 2010

Withdrawn from worldwide market because of increased risk of heart attacks and stroke

Drotrecoginalfa (Xigris) 2011

Failed to show a survival benefit for patients with severe sepsis and septic shock

balance between safety and effectiveness has become a high priority for the agency over the past several years. This has resulted in publishing several guidance documents through which the agency exclusively addresses safety issues and provides the drug developers with means to adequately implement rigorous safety studies and adverse reactions monitoring in both pre- and postmarketing phases of the drug development process.79,80

FDA and the Special Populations Pediatric PopulationdOver the past decade, pediatric research was facilitated through the Best Pharmaceuticals for Children Act and the Pediatric Research Equity Act81 that resulted in many studies conducted

to assess how drugs behave in children versus adults. These studies provide evidence of important differences between adults and children in the clearance and metabolism of drugs, which resulted in labeling changes for more than 350 marketed medical products, addressing the unique PK, dosing adjustments, and enhanced safety information for children82 (Table 37.3). Pregnant and Lactating WomendPregnancy and Lactation labeling resulted from the needs to develop a more comprehensive and clinically meaningful approach to therapy in pregnant and lactating women, due to the risks of drug exposure to the fetus as well as to determine safe dosing regiments (Table 37.3). Patients with Impaired Kidney and Liver Function and the Elderly PopulationdThese groups of patients have unique PK and PD attributes and these are addressed in detail in the new FDA guidance documents issued in recent years (Table 37.3).

SUMMARY QUESTIONS Question 1. Microdosing studies are conducted during early Phase I clinical trials. The goal of these studies is to: A. B. C. D. E.

Demonstrate safety and efficacy of the compound Determine the maximum tolerable dose limits Assess the PK in special populations Collect receptor-binding imaging data Select the dose for Phase II studies

Question 2. The following objectives are part of the “learning” phase of the “Learn-and-Confirm” paradigm: A. Define the target clinical population and the dose range to be used in Phase III. B. Demonstrate safety and efficacy of the compound for clinical use in a RCT C. Conduct postmarketing surveillance D. Conduct exploratory, FIH studies E. A, C, and D Question 3. The predictive value of new markers of organ toxicity is best evaluated through: A. B. C. D. E.

The animal-to-human test sequence In vitro studies Improved clinical trial design Modeling and simulation None of the above

References 1. Shelley JH, Baur MP. Paul Martini: the first clinical pharmacologist? Lancet 1999;353:1870e3. 2. Honing P. The value and future of clinical pharmacology. Clin Pharmacol Ther 2007;81:17e8.

IV. CLINICAL RESEARCH INFRASTRUCTURE

REFERENCES

3. American Society for Clinical Pharmacology and Therapeutics (ASCPT.org) Knowledge Center. http://www.ascpt.org/Knowledge Center/WhatisClinicalPharmacology/tabid/6665/Default.aspx. 4. American College of Clinical Pharmacology (ACCP1.org) e Public Policy and Position Papers. http://www.accp1.org/documents/ 1999ClinicalPharmacologyRetro.pdf. 5. Breckenridge A, Collery C, Rawlins M, Walport M. The future of clinical pharmacology in the UK. Lancet 2006;367:1051. 6. Aronson JK. Clinical pharmacology: past, present and future. Br J Clin Pharmacol 2006;61:647e9. 7. Decker S, Sausville EA. Drug discovery. In: Atkinson AJ, Abernethy DR, Daniels CE, Dedrick RL, Markey SP, editors. Principles of clinical pharmacology. 2nd ed. San Diego: Academic Press; 2007. p. 1e7. 8. Peck CC, Barr WH, Benet LZ, Collins J, Desjardins RE, Furst DE, Harter JG, Levy G, Ludden T, Rodman JH, Sanathanan L, Schentag JJ, Shah VP, Sheiner LB, Skelly JP, Stanski DR, Temple RJ, Viswanathan CT, Weissinger J, Yacobi A. Opportunities for integration of pharmacokinetics, pharmacodynamics and toxicokinetics in rational drug development. Pharm Res 1992;9:826e33. 9. Committee for Medicinal Products for Human (CHMP). http:// www.ema.europa.eu/ema/index.jsp?curl¼pages/about_us/ general/general_content_000094.jsp. 10. Food, Drug Administration Guidance Documents: Exploratory IND Studies. January 2006. http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ucm 078933.pdf. 11. Madan A, O’Brien Z, Wen J, O’Brien C, Farber RH, Beaton G, Crowe P, Oosterhuis B, Garner RC, Lappin G, Bozigian HP. A pharmacokinetic evaluation of five H (1) antagonists after an oral and intravenous microdose to human subjects. Br J Clin Pharmacol 2009;67:288e98. 12. Pauwels EK, Bergstrom K, Mariani G, Kairemo K. Microdosing, imaging biomarkers and SPECT: a multi-sided tripod to accelerate drug development. Curr Pharm Des 2009;15:928e34. 13. Lappin G, Garner RC. The utility of microdosing over the past 5 years. Expert Opin Drug Metab Toxicol 2008;4:1499e506. 14. Lertora JJL, Vanevski KM. Introduction to pharmacokinetics and pharmacodynamics. In: Thoene JG, editor. Small molecule therapy for genetic disease. 1st ed. New York: Cambridge University Press; 2010. p. 35e55. 15. Sheiner LB. Learning versus confirming in clinical drug development. Clin Pharmacol Ther 1997;61:275e91. 16. Chien JY, Friedrich S, Heathman MA, de Alwis DP, Sinha V. Pharmacokinetics/pharmacodynamics and the stages of drug development: role of modeling and simulation. AAPS J 2005;7: E544e59. 17. DiMasi JA, Hansen RW, Grabowski HG. The price of innovation: new estimates of drug development costs. J Health Econ 2003;22: 151e85. 18. Kola I, Landis J. Can the pharmaceutical industry reduce attrition rates? Nat Rev Drug Discov 2004;3:711e5. 19. Mervis J. Productivity counts e but the definition is key. Science 2005;309:726. 20. US Food, Drug Administration The critical path to new medical products. March 2004. http://www.fda.gov/oc/initiatives/criticalpath/ whitepaper.html. 21. US Food, Drug Administration Critical Path Opportunities Report. March 2006. http://www.fda.gov/oc/initiatives/criticalpath/ whitepaper.html. http://www.fda.gov/downloads/Science Research/SpecialTopics/CriticalPathInitiative/ CriticalPathOpportunitiesReports/UCM077254.pdf. 22. Honig P, Lalonde R. The economics of drug development: a grim reality and a role for clinical pharmacology. Clin Pharmacol Ther 2010;87:247e51. Erratum in: Clin Pharmacol Ther 2010;88:138.

657

23. The critical path initiative report on key Achievements in 2009. http://www.fda.gov/downloads/ScienceResearch/SpecialTopics/ CriticalPathInitiative/CriticalPathOpportunitiesReports/UCM 219432.pdf. 24. Ball R, Robb M, Anderson S, Dal Pan G. The FDA’s sentinel Initiativeda comprehensive approach to medical product surveillance. Clin Pharmacol Ther 2016;99:265e8. 25. Gagne JJ, Han X, Hennessy S, Leonard CE, Chrischilles EA, Carnahan RM, Wang SV, Fuller C, Iyer A, Katcoff H, Woodworth TS, Archdeacon P, Meyer TE, Schneeweiss S, Toh S. Successful comparison of US Food and drug administration sentinel analysis tools to traditional approaches in quantifying a known drug-adverse event association. Clin Pharmacol Ther 2016; 100:558e64. 26. Anger GJ, Piquette-Miller M. Translational pharmacology: harnessing increased specialization of research within the basic biological sciences. Clin Pharmacol Ther 2008;83:797e801. 27. FitzGerald AG. Anticipating change in drug development: the emerging era of translational medicine and therapeutics. Nat Rev Drug Discov 2005;4:815e8. 28. Food, Drug Administration Guidance Documents: Safety Testing of Drug Metabolites. 2016. http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ucm 079266.pdf. 29. Food, Drug Administration Guidance Documents: Estimating the Maximum Safe Starting Dose in Initial Clinical Trials for Therapeutics in Adult Healthy Volunteers. 2005. http://www.fda.gov/ downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm078932.pdf. 30. EMA Guideline on strategies to identify and Mitigate risks for first- inhuman clinical trials with investigational medicinal products. 2007. http://www.ema.europa.eu/docs/en_GB/document_library/ Scientific_guideline/2009/09/WC500002988.pdf. 31. Biomarker Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther 2001;69:89e95. 32. Framework for defining Evidentiary Criteria for biomarker qualification final version Evidentiary Criteria writing group Revised 09/23/16. 2016. http://fnih.org/sites/default/files/final/pdf/Evidentiary %20Criteria%20Framework%20Final%20Version%20Oct%2020% 202016.pdf. 33. US Food, Drug Administration Guidance document: Qualification Process for Drug Development Tools. 2010. https://www.cpath.org/ pdf/FDADraftDDTools.pdf. 34. Guidance for Industry, FDA Staff: Qualification Process for Drug Development Tools. 2014. http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ UCM230597.pdf. 35. Woodcock J. Chutes and ladders on the critical path: comparative effectiveness, product value, and the use of biomarkers in drug development. Clin Pharmacol Ther 2009;86:12e4. 36. Amur S, LaVange L, Zineh I, Buckman-Garner S, Woodcock J. Biomarker qualification: toward a multiple stakeholder framework for biomarker development, regulatory acceptance, and utilization. Clin Pharmacol Ther 2015;98:34e46. 37. Khan E, Batuman V, Lertora JJL. Emergence of biomarkers in nephropharmacology. Biomarkers Med 2010;4:805e14. 38. Biomarker Qualification Management Team: Goodsaid F, Walton M, Zineh I. The principles and process of biomarkers qualification at the FDA. http://www2.rsna.org/re/QIBA_Annual_Meeting_2010/ IndexFiles/Welcome_Intros_PDFs/GOODSAID-FDAPerspective_ 20100525.ppt%20%5BCompatibility%20M.pdf. 39. Joint FDA-EMEA Statement document. 2009. In: http://www.ema. europa.eu/docs/en_GB/document_library/Regulatory_and_procedural_ guideline/2009/10/WC500004205.pdf.

IV. CLINICAL RESEARCH INFRASTRUCTURE

658

37. CLINICAL PHARMACOLOGY AND ITS ROLE IN PHARMACEUTICAL DEVELOPMENT

40. Dieterle F, Sistare F, Goodsaid F, Papaluca M, Ozer JS, Webb CP, Baer W, Senagore A, Schipper MJ, Vonderscher J, Sultana S, Gerhold DL, Phillips JA, Maurer G, Carl K, Laurie D, Harpur E, Sonee M, Ennulat D, Holder D, Andrews-Cleavenger D, Gu YZ, Thompson KL, Goering PL, Vidal JM, Abadie E, Maciulaitis R, Jacobson-Kram D, Defelice AF, Hausner EA, Blank M, Thompson A, Harlow P, Throckmorton D, Xiao S, Xu N, Taylor W, Vamvakas S, Flamion B, Lima BS, Kasper P, Pasanen M, Prasad K, Troth S, Bounous D, Robinson-Gravatt D, Betton G, Davis MA, Akunda J, McDuffie JE, Suter L, Obert L, Guffroy M, Pinches M, Jayadev S, Blomme EA, Beushausen SA, Barlow VG, Collins N, Waring J, Honor D, Snook S, Lee J, Rossi P, Walker E, Mattes W. Renal biomarker qualification submission: a dialog between the FDA-EMEA and predictive safety testing consortium. Nat Biotechnol 2010;28:455e62. 41. Frederick CB, Obach RS. Metabolites in safety testing: “MIST” for the clinical pharmacologist. Clin Pharmacol Ther 2010;87:345e50. 42. Lathia CD, Amakye D, Dai W, Girman C, Madani S, Mayne J, MacCarthy P, Pertel P, Seman L, Stoch A, Tarantino P, Webster C, Williams S, Wagner JA. The value, qualification, and regulatory use of surrogate end points in drug development. Clin Pharmacol Ther 2009;86:32e43. 43. Drug Development Tools Qualification Programs. 2015. In: http://www. fda.gov/Drugs/DevelopmentApprovalProcess/DrugDevelopmentTools QualificationProgram/ucm483052.htm. 44. Heargraves RJ. The role of molecular imaging in drug discovery and development. Clin Pharmacol Ther 2007;83:349e53. 45. Cole PE, Schwartz AJ, Schmidt ME. Application of imaging biomarkers in the early clinical development of central nervous system therapeutic agents. Clin Pharmacol Ther 2012;91:315e20. 46. Parkinson DR, Ziegler J. Educating for personalized medicine: a perspective from oncology. Clin Pharmacol Ther 2009;86:23e5. 47. Flockhart DA, Skaar T, Berlin DS, Klein TE, Nguyen AT. Clinically available pharmacogenomics tests. Clin Pharmacol Ther 2009;86: 109e13. 48. Buyse M, Sargent DJ, Grothey A, Matheson A, de Gramont A. Biomarkers and surrogate end points e the challenge of statistical validation. Nat Rev Clin Oncol 2010;7:309e17. 49. US Food and Drug Administration: Table of pharmacogenomic biomarkers in drug labeling. http://www.fda.gov/drugs/ scienceresearch/researchareas/pharmacogenetics/ucm083378.htm. 50. Paving the Way for Personalized Medicine: FDA’s Role in a New Era of Medical Product Development. 2013. http://www.fda.gov/ downloads/ScienceResearch/SpecialTopics/PrecisionMedicine/ UCM372421.pdf. 51. Dollery CT. Clinical pharmacology e the first 75 years and a view of the future. Br J Clin Pharmacol 2006;61:650e65. 52. Barker AD, Sigman CC, Kelloff GJ, Hylton NM, Berry DA, Esserman LJ. I-SPY 2: an adaptive breast cancer trial design in the setting of neoadjuvant chemotherapy. Clin Pharmacol Ther 2009;86:97e100. 53. Biomarker Consortium. http://www.biomarkersconsortium.org and http://www.biomarkersconsortium.org/images/stories/ docs/ispypress031710.pdf. 54. Kraiczi H, Jang T, Ludden T, Peck CC. Randomized concentrationcontrolled trials: motivations, use, and limitations. Clin Pharmacol Ther 2003;74:203e14. 55. Lledo´-Garcı´a R, Hennig S, Karlsson MO. Comparison of dosefinding designs for narrow-therapeutic-index drugs: concentration-controlled vs. dose-controlled trials. Clin Pharmacol Ther 2009;86:62e9. 56. Gobburu, JVS. Learn-apply paradigm: Re-configuring drug development goals. The 4th Annual FDA/DIA Statistics Forum: “integrating knowledge in clinical development: Meta-analysis, Non-Inferiority, Related Topics.” http://www.fda.gov/downloads/ Drugs/NewsEvents/UCM209136.pdf.

57. US Food and Drug Administration Guidance document: Exposureresponse relationships d study design, data analysis, and regulatory applications. http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ucm 072109.pdf. 58. Ravva P, Gastonguay MR, French JL, Tensfeldt TG, Faessel HM. Quantitative assessment of exposure e response relationships for the efficacy and tolerability of varenicline for smoking cessation. Clin PharmacoTher 2010;87:336e44. 59. Shiner LB, Steimer J-L. Pharmacokinetic/Pharmacodynamic modeling in drug development. Annu Rev Pharmacol Toxicol 2000; 40:67e95. 60. Wang Y, Bhattaram AV, Jadhav PR, Lesko LJ, Madabushi R, Powell JR, Qiu W, Sun H, Yim DS, Zheng JJ, Gobburu JV. Leveraging prior quantitative knowledge to guide drug development decisions and regulatory science recommendations: impact of FDA pharmacometrics during 2004e2006. J Clin Pharmacol 2008;48:146e56. 61. Mould DR, Upton RN. Basic concepts in population modeling, simulation, and model-based drug development. CPT Pharmacometrics Syst Pharmacol September 26 2012;1. e6, http://dx.doi. org/10.1038/psp.2012.4. 62. Lee H, Yim D-S, Zhou H, Peck CC. Evidence of effectiveness: how much can we extrapolate from existing studies? AAPS J 2005;07: E467e74. 63. Miller R. Population pharmacokinetics. In: Atkinson AJ, Abernethy DR, Daniels CE, Dedrick RL, Markey SP, editors. Principles of clinical pharmacology. 2nd ed. San Diego: Academic Press; 2007. p. 1e7. 64. Sheiner LB. Learning vs confirming in clinical drug development. Clin Pharmacol Ther 1997;61:275e91. 65. Gobburu JVS, Lesko LJ. Quantitative disease, drug, and trial models. Annu Rev Pharmacol Toxicol 2009;49:291e301. 66. Williams PJ, Ette EI. Pharmacometrics: impacting drug development and pharmacotherapy. In: Williams PJ, Ette EI, editors. Pharmacometrics: the Science of Quantitative Pharmacology. 1st ed. Hoboken (New Jersey): John Wiley & Sons, Inc.; 2007. p. 1e21. 67. Pfister M, D’Argenio DZ. The emerging scientific discipline of pharmacometrics. J Clin Pharmacol 2010;50:6S. 68. Flockhart DA, Bertilsson L. Clinical pharmacogenetics. In: Atkinson AJ, Abernethy DR, Daniels CE, Dedrick RL, Markey SP, editors. Principles of clinical pharmacology. 2nd ed. San Diego: Academic Press; 2007. p. 1e7. 69. Murphy MP. Current pharmacogenomic approaches to clinical drug development. Pharmacogenomics 2000;1:115e23. 70. Food, Drug Administration Guidance Documents: Clinical pharmacogenomics: Premarket evaluation in early-phase clinical studies and recommendations for labeling. 2013. http://www.fda.gov/downloads/ drugs/guidancecomplianceregulatoryinformation/guidances/ucm 337169.pdf. 71. Green DJ, Mummaneni P, Oh JM, Pacanowski M, Burckart GJ. Pharmacogenomic information in FDA-approved drug labels: application to pediatric patients. Clin Pharmacol Ther 2016;99:622e32. 72. Ingelman-Sundberg M, Cascorbi I. Pharmacogenomic or -epigenomic biomarkers in drug treatment: two sides of the same medal? Clin Pharmacol Ther 2016;99:478e80. 73. US Food and Drug Administration Innovation or stagnation: challenges and opportunity on the critical path to new medical products. March 2004. http://www.fda.gov/oc/initiatives/criticalpath/whitepaper.html. 74. Peck CC, Cross JT. Getting the dose right: facts, a blueprint, and encouragements. Clin Pharmacol Ther 2007;82:12e4. 75. Lesko LJ. Paving the critical path: how can clinical pharmacology help achieve the vision? Clin Pharmacol Ther 2007;81:170e7. 76. US Food and Drug Administration. Clinical pharmacology guidance documents. http://www.fda.gov/drugs/guidancecompliancere gulatoryinformation/guidances/default.htm.

IV. CLINICAL RESEARCH INFRASTRUCTURE

REFERENCES

77. Zineh I, Woodcock J. Clinical pharmacology and the catalysis of regulatory science: opportunities for the advancement of drug development and evaluation. Clin Pharmacol Ther 2013;93:515e25. 78. US Food and Drug Administration Amendments Act (FDAAA) of 2007. http://www.fda.gov/RegulatoryInformation/Legislation/ SignificantAmendmentstotheFDCAct/FoodandDrugAdministration AmendmentsActof2007/default.htm. 79. US Food and Drug Administration. Drug safety. http://www.fda. gov/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm064993.htm.

659

80. US Food, Drug Administration Safety Assessment for IND Safety Reporting. 2015. http://www.fda.gov/downloads/drugs/guidance complianceregulatoryinformation/guidances/ucm477584.pdf. 81. US Food and Drug Administration. Pediatric drug development. http://www.fda.gov/Drugs/DevelopmentApprovalProcess/ DevelopmentResources/ucm049867.htm. 82. Advancing Regulatory Science for Public Health. www.fda.gov/ ScienceResearch/SpecialTopics/RegulatoryScience/ucm228131. htm.

IV. CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

38 Career Paths in Clinical Research Frederick P. Ognibene National Institutes of Health, Bethesda, MD, United States

O U T L I N E Background

661

Student and Resident Training in Clinical Research 662

Additional Educational Approaches and Support for Training

668

PhysicianeScientist Workforce

664

Conclusions

669

Clinical Research Curriculum and Training

665

Summary/Discussion Questions

669

NIH Clinical Center Core Curriculum

666

References

669

This edition of Principles and Practice of Clinical Research provides an updated, comprehensive overview about how to conduct clinical and translational research in the most rigorous manner within the ever-expanding clinical research enterprise. This enterprise has many infrastructure elements, one of which is having well-trained individuals to conduct clinical research; in many cases their careers are devoted to that effort. Individuals who engage in clinical research can have primary careers in medicine, dentistry, nursing, pharmacy, epidemiology, bioinformatics, bioethics, and others. However, one still has to ask certain questions. Who is going to conduct clinical research in the future? How will those various investigators be trained and what are the resource requirements to do so? This chapter’s primary focus will be on careers for physicianescientists in clinical research and will review what has been done, as well as what might need to be done, to address the shortage of physicians as clinical researchers and importantly provide some insights on how to invigorate and sustain that pipeline. Details of the successful efforts by the National Institutes of Health (NIH) and others to answer these questions are included as well as efforts by different professional specialties to help develop those careers. Once training has been provided there are additional challenges related to

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00038-1

sustaining successful careers of physicianescientists in clinical research. More focused information and details about the new practice domain of clinical research nursing are presented in Chapter 39.

BACKGROUND There is a body of literature that emerged starting in the late 1970s, which focused on the shrinking numbers of physicians in clinical research careers.1e4 Physicians engaged with and interested in clinical research were part of the rapid growth in the biomedical basic sciences in the last quarter of the 20th century. There were articles that posited that the decline in the number of physicians engaged in clinical research as a career was possibly related to increased competition for resources that were going to investigators working in the “booming” basic sciences. This was coupled with a sense of low esteem for those in clinical research careers, since there were data as well as perceptions that both academic medical centers and NIH study sections seemed to favor research conducted by basic scientists.2,5 In 1995, an NIH Director’s Panel on Clinical Research was convened by then director Harold Varmus,

661

Copyright © 2018. Published by Elsevier Inc.

662

38. CAREER PATHS IN CLINICAL RESEARCH

MD, with David G. Nathan, MD, as chair to “make recommendations that might guide the NIH toward policy changes that could alleviate the distress in the clinical research community.”6 One important outcome of that panel was to broaden the definition of clinical research to include patient-oriented research, epidemiologic and behavioral studies, outcomes research, and health services research. In addition to the expanded definition of clinical research, recommendations were made related to federal grants funding for clinical research, including special emphasis study sections to review clinical research grant applications; didactic clinical research training grants for support of fellows and junior faculty in academic medical centers; grants for early career clinical investigators and their mentors; and enhancement of the budgets for the (then) General Clinical Research Centers.5e7 Another recommendation from the NIH Director’s Panel on Clinical Research stated that the NIH should initiate an intramural training program to enhance the attractiveness of careers in clinical research to medical students. This led to the formation of the NIH Clinical Research Training Program (CRTP), which was established in 1997 and flourished for 15 years (see additional program details later in this chapter) with support from both the NIH and grant support from Pfizer Inc. and other private entities to the Foundation for the National Institutes of Health.8 Collectively, these recommendations by Dr. Nathan and colleagues were an important milestone and the groundwork for a number of developments contributing to the recognition and enhancement of clinical research as a viable and productive career path for physicians and other clinicianescientists.

STUDENT AND RESIDENT TRAINING IN CLINICAL RESEARCH Waiting until completion of medical school, a primary residency, and possibly subspecialty training in a clinical fellowship to learn how to conduct clinical research is likely too late to foster most careers in clinical research. Historically, there had been very little formal training in clinical research in medical schools. The need for inclusion of clinical research education in medical schools to heighten the sensitivity of all physicians caring for patients about the importance of clinical research as it relates to direct patient care has been raised by thought leaders in the field in the past.9 In 2006, the Association of American Medical Colleges, which provides oversight to the curriculum in medical schools, focused on this deficiency and advocated for the mandatory inclusion of education about clinical and translational research into the curriculum of medical schools.10e12 Although progress has been made adding clinical

research to the curriculum in some medical schools, it is not a universal principle, and this topic is still the subject of discussion. One recent commentary listed 12 recommendations to enhance clinical research in academic centers, including three addressing the need for training in “clinical research methodology” during early career development, specifically addressing medical school, residency training, and Accreditation Council for Medical Education fellowship programs.13 To further describe the challenges, there has been a steady decline in the number of physicians in the United States reporting research as their primary professional activity from 3.6% in 1982 to only 1.6% by 2011.14 However, the same authors also pointed out that self-reported strong interest in research careers for medical school graduates reached 18.8% in 2009, when the American Recovery and Reinvestment Act was enacted and provided substantial increases in research funding for the NIH.14 Despite this conveyed interest in research, there is a sharp drop off in those conducting research, from their reported aspirations at the time of graduation to actual metrics of reported professional activities. In addition to having an introductory clinical research curriculum while in medical school, actual clinical, translational, and basic research experiences to augment the didactic content learned have been determined to be critical for the successful transitioning to research careers. The NIH-funded Medical Scientist Training Program (MSTP) and non-MSTP programs at other medical schools have been major contributors and a model for training physicianescientists who have received both MD and PhD degrees. These dual degree programs have reported successful outcomes in terms of their graduates securing academic research positions.15,16 Historically, research interests and careers for graduates of those dual degree programs have been directed more toward basic science compared to those students with only MD degrees.17 There have been a number of recent publications looking more closely at outcomes and metrics of dual degree programs. A report of career outcomes of graduates from 24 MDePhD programs indicated nearly twothirds of graduates spent at least 50% of their time in research activities, with 39% of MDePhDs from those programs reporting at least 75% of their time conducting research.18 Based on overall percentages, the top three clinical departmental affiliations in academic institutions for these graduates were internal medicine, pediatrics, and pathology. However, the authors noted that numbers of MDePhD graduates entering those disciplines have declined somewhat, while there have been increases in numbers of graduates in dermatology, ophthalmology, radiation oncology, and surgery, which is similar to trends noted for residencies by medical school graduates in general.19 A more detailed analysis

IV. CLINICAL RESEARCH INFRASTRUCTURE

STUDENT AND RESIDENT TRAINING IN CLINICAL RESEARCH

of students who matriculated into combined MDePhD programs (both MSTP and non-MSTP), demonstrated that about 75% of students ultimately received dual degrees, most of the remainder completed their MD degree requirements, with only a small percentage either withdrawing or dismissed from medical schools.20 The key point, however, is despite the academic success and research productivity of these individuals their overall numbers are small, and a portion ultimately has more active clinical roles and is not conducting clinical, translational, or basic research as envisioned. Since the mid-1980s, there have been a number of year-long research enrichment and training opportunities for medical students, which have been developed by both public and private entities. Unlike the MSTP and non-MSTP MDePhD programs, these year-long opportunities do not lead to an additional degree. These programs originally included the Howard Hughes Medical Institute (HHMI) Research Scholars Program and the HHMIeNIH Medical Fellows Program, the Doris Duke Clinical Research Fellowship Program, the Sarnoff Cardiovascular Research Foundation Research Fellowship Program, and the NIH CRTP.21e25 However, since 2012 three of the research programs have been discontinued: the HHMIeNIH Research Scholars Program, the Doris Duke Charitable Foundation Clinical Research Fellowship Program, and the NIH CRTP.8 In 2012, the NIH transformed the CRTP into a new entity for student trainees, the NIH Medical Research Scholars Program (MRSP).26 This amalgamated program combined features of the former HHMIeNIH Research Scholars Program and the NIH CRTP. The MRSP provides yearlong mentored research experiences, ranging from basic sciences to translational and more clinically oriented research for medical, dental, and veterinary students. The NIH MRSP had its first class in 2012e13, and this publiceprivate partnership has continued to support between 40 and 50 students per year. The programs vary in the types of research experiences and cover the full range from basic to translational to clinical research. In addition, and in varying amounts, the programs provide students with didactic content (clinical research curriculum, journal clubs, lectures, etc.) as well as the practical or “hands-on” mentored research experiences. Outcome data from these programs have been published. The HHMIeNIH Research Scholars Program and the HHMI Medical Fellows Program had been in existence since 1985 and 1989, respectively. Both of these programs provided intensive mentored research experiences; and in 2003, Fang and Meyer analyzed early career outcomes in both programs. They determined that participation in the HHMI Medical Fellows and HHMIeNIH Research Scholars programs increased the likelihood of program alumni receiving NIH postdoctoral awards, with award rates slightly less when

663

compared to participants in MSTP programs but very comparable to awards for those in non-MSTP MDePhD programs.27 In addition, outcome data were published for the Doris Duke Clinical Research Fellowship Program, based on follow-up after the first 3.5 years of that program.28 However, since those surveyed were still in residency training there were no data comparable to the HHMI data in terms of success rates for funded postdoctoral awards. Rather, the Doris Duke data focused on satisfaction with the 1-year clinical research training experience, which was almost uniformly positive. In addition, nearly half of those Doris Duke Clinical Research Fellowship Program alumni surveyed indicated an increased commitment to a research career after their year of mentored research. More comprehensive outcome data from these yearlong programs had not been available. There was a “sense” that these research opportunities for medical students provided an introductory framework through which firsthand exposure to clinical and translational research, in a structured and mentored environment, was a key component to establish and hopefully maintain interest in clinical research careers. In 2016 outcome data from the NIH CRTP were published.8 Data from CRTP alumni who also had completed primary residency and subspecialty training were analyzed. Of the cohort of possible respondents, approximately 65% of program alumni, who had completed primary residency and subspecialty training, self-reported participation in research as part of their career portfolios. Most of the group conducting research (88%) were affiliated with academic medical centers, and of that subset about 55% indicated that at least 25% of their reported time commitments involved research. Of the small subset working in industry, all 3 of 84 (3.6%), reported spending over 50% of their time conducting research.8 Ongoing tracking is in place to continue to monitor the career metrics of those students, including career progression, publications, and research funding (including sources). The aforementioned programs all varied somewhat in opportunities, experiences, and instructional methodologies; however, they all have the overarching goal of harnessing early interest in research and then sustaining it as future clinicianescientists progress in their careers. Table 38.1 provides a composite list of the ideal key programmatic components of these medical student training opportunities, which seem critical to the development of students’ interest in careers in clinical research. Although long-term and confirmatory data have not been yet compiled, the impression is that these trainees should have experiences in research career enrichment programs that are expertly mentored, with the “role models” being established by clinical and translational researchers. Additional educational

IV. CLINICAL RESEARCH INFRASTRUCTURE

664 TABLE 38.1

38. CAREER PATHS IN CLINICAL RESEARCH

Key Core Components of Student Research Programs

Dedicated research mentorship Core curriculum of research principles Journal clubs, seminars, and small group discussions Access to research support servicesdresearch nurses, statisticians, bioethicists, and others Clinical experiences demonstrating primary source of research questions as well as benefits of successful research outcomes Teaching of best practices in oral presentations, poster preparation, and writing skills Introduction to grant writing Leadership skills Administrative oversight

experiences should focus on didactics such as a core clinical research curriculum, interactive sessions including seminars and journal clubs, and other experiences to augment and reinforce the benefits of clinical and translational research careers. The students in these year-long research enrichment programs ultimately may have different clinical and career interests, and while in these programs are participating in different types of research, ranging from basic to translational or clinical. However, they all are doing so as part of an augmentation of their established professional school curricula to include a formal research experience, and tracking the career outcomes of these individuals should be a goal. Unfortunately, participation in the year-long research enrichment programs for medical students is limited in capacity due to the small number and size of those programs, so the overwhelming majority of medical students do not have those opportunities. However, enhanced medical school education about the importance and relevance of clinical research, especially as it relates to established as well as innovative patient care, has been advocated for years.9e12,29 There is a challenge that occurs after medical school when young physicians are pursuing clinical training as residents in primary disciplines prior to additional subspecialty training as fellows. Most individuals consider residency training as particularly strenuous since the majority of their time is spent on clinical training and clinical obligations, and this is further complicated by duty hour limitations that are imposed and strictly regulated by the Accreditation Council for Graduate Medical Education.30 However, once there is a transition to subspecialty residency and fellowship training, not only is there an emphasis on advanced clinical training but also on opportunities

to conduct and learn the principles of research (clinical and basic). There are few formal programs targeted at instructing clinical research for residents; a program at Massachusetts General Hospital entitled “Tools of Human Investigation” was designed in 2003 as an opportunity to increase residents’ exposure to clinical research and also to assist in their career development planning.31 This program consisted of 2-week rotations held four times a year, which were required of all second-year internal medicine residents, and had three stated aims: to provide an overview of hypothesis generation and testing using different study designs and research methodologies; to provide skills to critique the medical literature; and to learn about career development firsthand from research faculty at different points in their careers, i.e., role models.31 The authors indicated this experience was aimed to potentially sustain interest in research that might have developed in medical school but ran the risk of being lost due to the clinical demands of primary residencies. As such, this 2-week program was one approach to harness and reinforce interest in research and academic careers among residents.31

PHYSICIANeSCIENTIST WORKFORCE Ensuring that there is an adequate and diverse biomedical workforce of physicianescientists has been a major focus of the NIH. A PhysicianeScientist Workforce Working Group was established and charged to provide recommendations to help sustain and strengthen such a workforce.32 The report indicated that despite stability in workforce size, the physiciane scientist workforce is still relatively small, which is problematic. In addition, there were important demographic issues and challenges, including a rising average age of that workforce, insufficient numbers of younger investigators, and a lack of diversity based on underrepresentation by women and minorities. The report provided a number of recommendations to apply to all clinical investigators including modifications in types of grants (addressing both training and career transitions), better tracking of outcomes, enhanced efforts to increase diversity, and expansion of programs to help address debt burden of physicianescientists. In addition to the NIH recommendations, which will take additional time to fully materialize, others have responded to the PhysicianeScientist Workforce Working Group report with additional actions primarily aimed at trainees to address the workforce deficiencies.33 The authors also strongly advocated for the entire biomedical community including universities, academic medical centers, the

IV. CLINICAL RESEARCH INFRASTRUCTURE

CLINICAL RESEARCH CURRICULUM AND TRAINING

NIH, and professional societies to partner proactively to address these identified challenges to ultimately enhance the workforce. As already noted, a major impediment to actually pursuing careers in clinical research, thus impacting workforce, has been and remains the large amount of debt incurred by medical students and other professionals considering careers in clinical research. The level of indebtedness was recognized as an impediment for careers in clinical research, and during President William Clinton’s administration four loan repayment programs were created by the NIH for (1) clinical research, (2) pediatric research, (3) health disparities research, and (4) clinical research for individuals in disadvantaged backgrounds.34 These were in addition to an older, existing loan repayment program focused on contraception and infertility research. All of those loan repayment programs still exist and, depending on the type of loan repayment program, support both intramural and extramural investigators.35 That cited website has details about the specific requirements and potential payments for each of the loan repayment programs. Overall success rates for applicants to the loan repayment programs have hovered around 50% from 2013 to 2016,35 so having loan repayment as an option, albeit a competitive one, is at least an attempt to limit the impact of professional school debt on an individual’s consideration of clinical research as a viable professional career. Unfortunately, medical school debt has continued to rise annually. In data from 2015 medical school graduates, the Association of American Medical Colleges reported that the average indebtedness for medical school graduates exceeded $180,000.00.36 These loan repayment programs which, if an applicant is successful lead to the receipt of at least partial loan repayment, may help to limit the impact of debt on one’s decision to pursue a career in clinical research. Although, there are many other impediments to physicianescientist workforce and research careers, limiting the impact of indebtedness has been considered to have been a positive impact on encouraging careers in clinical research.

CLINICAL RESEARCH CURRICULUM AND TRAINING As already noted, for physicianescientists and other clinicianescientists to ultimately have successful careers in clinical research, both direct (“hands-on”) experiences and more formal, didactic instruction and training are critical. Those wishing to pursue clinical research careers must first have a good understanding of clinical and scientific issues that is then augmented by formal training in clinical research principles. This can be

665

accomplished via education for medical students, residents, fellows, and others who are part of the clinical research enterprise. One of the first programs was the Harvard Program in Clinical Effectiveness that was initiated in 1986 as an effort to augment knowledge of clinical trials, clinical epidemiology, and health services research.37 This was a pilot program with only three students in its initial year, and it expanded to enroll over 150 students annually, with nearly 1500 students enrolled in 20 years. This 6- to 7-week intensive summer program was geared toward physicians in fellowships and had core content in epidemiology and biostatistics, as well as elective content and the preparation of a grant proposal.37 Outcome data indicated that a majority of the Harvard program participants continued in research careers; about onethird have pursued and obtained advanced degrees and many also received federal and nonfederal funding to support their research. In 1998, the NIH introduced the Clinical Research Curriculum Award (CRCA) to “improve the quality of training in clinical research.”38 These awards, also known as K30 awards, “.were intended to stimulate the inclusion of high-quality, multidisciplinary didactic training as part of the career development of clinical investigators.” The CRCA program ultimately supported the development or improvement of core courses designed as instruction in the fundamental skills, methodology, theories, and conceptualizations necessary for the well-trained, independent, clinical researcher. Some funded institutions also developed programs that granted advanced degrees in clinical research. The NIH K30 program was terminated; however, the curriculum development concept has been sustained in the NIH Clinical and Translational Science Award (CTSA) program.39 The CTSA program was launched in 2006 by the former National Center for Research Resources. Currently, this program is supported by the National Center for Advancing Translational Sciences (NCATS), and there are more than 50 medical research institutions (hubs), in this network.39 The CTSA Program has five key goals with one being to “train and cultivate the translational science workforce.” There are two types of formal clinical research training awards at the hubs, and both combine course work with research experiences. All CTSA hubs have a KL2 program, which offers formal research training to individuals with MD, PhD, or equivalent doctoral degrees. The KL2 Mentored Clinical Research Scholar Program focuses on career development training individuals from a variety of fields (e.g., medicine, dentistry, nursing, the behavioral sciences, biostatistics, and epidemiology). Those individuals supported with KL2 awards have a range of career development experiences for up to 5 years and

IV. CLINICAL RESEARCH INFRASTRUCTURE

666

38. CAREER PATHS IN CLINICAL RESEARCH

access to the CTSA programs’ facilities.40 In addition to the KL2 program, many hubs also have programs that provide predoctoral trainees with an introduction to clinical and translational research through the TL1 program.40 The goal of the TL1 program is to increase the number of well-trained clinicianescientists who can ultimately have leadership roles leading to the design and oversight of future clinical investigations critical to the overall mission of NCATS and NIH.40 Institutional support may be provided for predoctoral candidates and candidates seeking combined health professional doctorateemaster’s degrees as well as postdoctoral fellows seeking additional training in clinical research. For both programs, access to research facilities including courses, specialized equipment, and databases and access to mentors, administrative research protocol support, and funding of pilot projects are available.40 Collectively, at CTSA sites, formalized instruction in clinical research, in many cases leading to advanced degrees, and access to institutional resources have been supported with public funds as part of the broad NIH commitment to enhance the career development of well-trained clinical researchers. These programs have created infrastructure and mechanisms to identify individuals hoping for a career in clinical research and who would most likely benefit from these training and curricular opportunities. However, many established leaders in clinical research also have realized that despite these efforts, challenges remain. With that background, and utilizing information collected from a variety of sources, the former CTSA Education and Career Development key function committee developed six “consensus statements” for further improvement of career development of clinical and translational scholars.41 Included in the strategic recommendations as part of a future agenda were (1) clinical and translational research requires the career development of a qualitatively different investigator, (2) promotion and tenure requirements should reflect the emerging value on team science and mentoring, (3) the trajectory of training includes a long-term commitment by institutions, (4) discipline-specific training is still required but curricula designed to promote teamwork and interdisciplinary training will promote innovation, (5) PhD trainees can take advantage of multiple pathways to have a successful and satisfying career, and (6) mentoring requires a centralized infrastructure and rewards will promote excellence.41 These strategic points were conveyed by the authors as part of a proposed agenda for next steps and taken together with insights and recommendations provided by the 2014 NIH PhysicianeScientist Workforce Working Group Report,32 they do provide some collective guidance on ways to enhance the career development of the clinicianescientist workforce.

However, as with all career development programs, outcome metrics are important, to help gauge success or failure and to help hone best practices. As already noted8 there are limited career outcome data on programs aimed at early capture of individuals still in medical, dental, or veterinary schools who are interested in pursuing biomedical research careers. The emphasis on, and growth of, formal training in clinical and translational research has been discussed earlier. The impact such training has had on successful career development as well as on research productivity of those trained needs to be evaluated in an ongoing manner. A comprehensive model has been described looking at domains of career success assessing extrinsic and intrinsic factors as possible metrics as well as “higher-order contextual factors” (personal and organizational) that can affect career success.42 In terms of CTSA programs, the goal of this model was to allow each funded and participating institution to specify the elements of that larger model to use as a framework from which to develop metrics able to evaluate outcomes.43 Many believe that the model cannot only be used to determine factors that contribute to career success as well as those that are also barriers to success. Since the data and metrics critical to gauging career outcomes have been so varied, it has been proposed that utilizing such a model will encourage more consistent data collection, within and across training sites, and thus lead to more rigorous, systematic, and generalizable assessments of factors associated with successful clinical research careers.43

NIH CLINICAL CENTER CORE CURRICULUM In 1995, the NIH Clinical Center recognized that a formal curriculum in clinical research was critical and essential to augment the practical experiences’ investigators in training received about how to conduct safe, well-regulated, and ethically sound clinical research. A series of three courses have been developed, and they now comprise the core curriculum in clinical research at the NIH Clinical Center and elsewhere.44,45 For 20 years, these courses were taught at the NIH Clinical Center and available in real time via long-distance learning to academic centers around the world. In addition, archived lectures were made available on the Internet for access by those in time zones where real-time connections were not possible. In addition, somewhat abbreviated versions of the three courses have been taught in live formats internationally. These distance learning and academic collaborations with centers around the world have allowed access of this curriculum to both institutions and individuals who do not have well-developed curricula in clinical

IV. CLINICAL RESEARCH INFRASTRUCTURE

667

NIH CLINICAL CENTER CORE CURRICULUM

research. More recently, the presentation format has been modified to web-based content only for two of the three courses. Lectures are videotaped at the NIH Clinical Center, converted to a YouTube format to easily be downloaded, and are accessible either as single lectures or as a “fully packaged” course. Thus, this core content, which is critical to the training of those interested in clinical research careers plus those conducting clinical research, is readily accessible, thus making it readily available to the adult learner. The first course in the core curriculum was established in 1995 and is entitled, “Introduction to the Principles and Practice of Clinical Research (IPPCR).” It provides formal training on how to effectively conduct clinical research with content ranging from trial design to ethical and regulatory issues and beyond. The course provides participants with information on how to design a successful clinical trial by focusing on biostatistics methods, study design, protocol preparation, data management, patient monitoring, quality assurance, ethical issues in the conduct of clinical research focusing on the protection of human subjects, and Food and Drug Administration regulatory issues. Since its inception through 2017, the course has been offered to nearly 38,000 students, over 50% of whom have been at 768 different long-distance sites on five continents. Table 38.2 represents composite course demographics for this and the other courses in the core curriculum. This textbook, Principles and Practice of Clinical Research, originally published in 200246 and now in its fourth edition,47 is used for the course. Seven modified live international versions of the course have been taught: twice in China48 and once in Nigeria, Russia, India, Brazil, and South Africa. The course participants at these international sites were identified as individuals who would not only use the knowledge gained to enhance their skills in their clinical research careers but also an additional overarching goal of these courses was to have the course participants convey the knowledge they gained to their hospital and medical center colleagues. In effect, it was a model of education to not only convey information to enrollees but also designed to “train the trainer.” For courses in TABLE 38.2

countries where English was not the primary native language, the course leaders also ensured that a local teaching assistant was assigned to work with each instructor during the course, with the goal that these teaching assistants also would become the future faculty for subsequent courses in their respective countries. In China and Russia, these teaching assistants also provided assistance with translation of lectures to ensure that the course content was conveyed accurately and completely. They also assisted the faculty as skilled interpreters in answering questions posed by the attendees. It is important to note that in all of the international venues, the percentage of students who passed a final examination was on a par with the students in the United States who take the course and a nearly identical examination. The textbook Principles and Practice of Clinical Research has been translated into Japanese,49,50 Chinese,51,52 and Russian.53 The second course, which is part of the core curriculum in clinical research and which complements the “IPPCR” course is entitled “Principles of Clinical Pharmacology” and was first offered in 1998. It provides an introductory review of pharmacokinetics, drug metabolism and transport, assessment of drug effects, drug therapy in special populations, contemporary drug development, and other topics. A textbook entitled Principles of Clinical Pharmacology first published in 200154 and now in its third edition55 is used for the course, which had been offered annually at the NIH Clinical Center, and similar to the clinical research course also was broadcast live to long-distance sites. However, similar to the IPPCR course lectures are videotaped at the NIH Clinical Center, converted to a YouTube format to be downloaded easily, and are accessible either as single lectures or as a “fully packaged” course. This modified approach is an attempt to reach as broad an audience as possible and to make is more convenient for the adult learn. Since its inception it has been offered to over 12,000 students worldwide (Table 38.2). In 2009 the course was taught live in Beijing, China, and similar to the principles of clinical research course, Chinese teaching assistants were used to assist with translation of the live lectures and to answer questions from the

Scope and Outreach of National Institutes of Health Core Curriculum in Clinical Research Students Enrolled

All Participating Sites

United States Sites

International Sites

Introduction to the Principles and Practice of Clinical Research

37,634

768

404

364

Principles of Clinical Pharmacology

12,344

80

56

24

Ethical and Regulatory Aspects of Clinical Research

7,203

105

68

37

Totals

57,181

953

528

425

IV. CLINICAL RESEARCH INFRASTRUCTURE

668

38. CAREER PATHS IN CLINICAL RESEARCH

audience. The textbook Principles and Practice of Clinical Pharmacology was translated into Chinese for that course and for subsequent use.56 The third course in the NIH core clinical research curriculum, Ethical and Regulatory Aspects of Clinical Research, was implemented in 1999 and offers formal content in the principles of research ethics. Information about the ethical conduct of clinical research is taught and content includes how to utilize a systematic framework for ethically evaluating a clinical research protocol, review of the regulations and ethical standards governing the conduct of human subject research, challenges related to early phase clinical trials, randomization, vulnerable populations, international research, etc. Since inception over 7000 students worldwide have enrolled (Table 38.2) and utilized the textbook Ethical and Regulatory Aspects of Clinical Research.57 The NIH Clinical Center Bioethics Department, the National Institute of Allergy and Infectious Diseases, and international cosponsors have conducted a number of short, targeted training workshops in bioethics at a variety of international locations with the audience including biomedical researchers, research ethics committee members, Ministry of Health officials, and other policymakers in countries where there is significant NIH funding for clinical research. The successful use of long-distance learning, contemporary educational tools and the translation of textbooks into foreign languages demonstrate that the scope and growth of those with careers in clinical research extends well beyond the United States. As of 2017, over 57,000 individuals from 953 sites (Table 38.2) around the world have participated in the NIH Clinical Center’s core curriculum in clinical research. Thus, expanding the pool of well-trained clinical researchers globally and broadly not only enhances the numbers in clinical research careers, but it should also ultimately enhance the overall quality of clinical research conducted internationally and lead to outcomes, which benefit the public health of humankind.

ADDITIONAL EDUCATIONAL APPROACHES AND SUPPORT FOR TRAINING A review of the literature also indicates that the overarching goal of enhancing the pipeline of well-trained, academic clinicianescientists who conduct research has led to strategies and plans in a number of specialty areas. The Long Range Planning Committee of the American Neurological Association advocated for increased flexibility in the training of neurologists to allow these trainees opportunities to generate preliminary data to assist with submissions of early career grant applications, and this flexibility has been approved by

the neurology Residency Review Committee of the Accreditation Council for Graduate Medical Education.58 In addition, the American Neurological Association also proposed an annual 4- to 6-week summer course in clinical research covering principles of clinical research along with core topics in translational neurosciences. The need for enhanced research training in most specialty and subspecialty areas of medicine has been widely communicated in the literature. Perspectives from the ophthalmology and pediatrics literature provide additional general and specific insights.59,60 In 2000, the Veteran’s Affairs Office of Academic Affiliations developed a consortium of Veteran Affairs sites, entitled the “Special Fellowship Program in Advanced Psychiatry and Psychology.61 Ten sites were selected based on a request for applications, with a key requirement that the participating institution had to have a robust mentoring environment. Those sites also developed a core curriculum that included research methodology, statistics, epidemiology, financing, legal and ethical issues, as well as other topics. In addition to the local instruction and mentoring at the 10 institutions, the core curriculum was broadcast to all participants using an interactive audio and video long-distance format that was felt to enhance the learning experiences of the trainees based on their feedback. This approach not only highlights an additional broader, systems’ approach to training, but it also demonstrates the benefits of using contemporary educational tools, such as web-based audio and visual long-distance learning, to reach the entire cohort of program participants and to create a sense of community of clinical researchers. Initial data also indicated that over 70% of fellows who completed this program were either in academic or clinical research positions. Although their support is relatively small, compared to both public funds and support from industry, private philanthropy also has been instrumental in helping to enhance the pipeline of those pursuing careers in clinical research by investing in both research training and career development.62 Additional details about philanthropic support for both basic and translational biomedical research as well as support for training and clinical research careers are available in Chapter 35. Clinical research careers also exist in private industry, including pharmaceutical, device, information technology, and biotechnology companies. Many experienced clinical and translational investigators with successful careers and outstanding research portfolios in either academia or the public sector may choose to migrate to private industry as part of their personal career trajectories. Reasons for such career movement may range from scientific opportunities that such companies may offer to more personal reasons.

IV. CLINICAL RESEARCH INFRASTRUCTURE

REFERENCES

Finally, there is an emergence of the importance of empowerment via leadership training as an additional tool to enhance the careers of clinicianescientists.63,64 Leadership training should enhance personal skills related to emotional intelligence, negotiation and conflict management, teamwork, time management, etc., which are felt to be beneficial to initiating successful careers and career progression.64 However, like most issues related to career development, it is not entirely clear which programs or skills have the most impact, so further research and assessment are critical.63

CONCLUSIONS Despite the challenges of establishing career paths for those aspiring to be part of the clinical research workforce, the pipelines created by early exposure in medical and dental school followed by additional clinical research training and reinforcing experiences in residency, fellowship, and beyond provide hope that the pool of individuals engaged in clinical research careers will continue to grow. To enhance these careers, it is important that there are plenty of opportunities and not just “one time” experiences. Individuals may become interested in clinical research at different times; not all make that determination at the earliest points of their education. The concept of “early capture” and sustaining interest in clinical research careers is essential. It is critical to ensure that these interests are reinforced by comprehensive educational experiences, utilizing some of the curricular content mentioned, coupled with exposure to role models who are established and successful clinical researchers. With the emphasis on the benefits of these careers, as well as attempts to limit impediments to entering and sustaining such careers, it is hoped that a career in clinical research will be viewed favorably by many for the clinical research career pipeline to grow.

SUMMARY/DISCUSSION QUESTIONS 1. Which of the following is not a critical component in a core clinical research curriculum? a. Study design b. Ethical issues c. Genomic analysis of tissue specimens d. Data and safety monitoring 2. Developing an interest in a career in clinical research occurs at which point in a physician’s training? a. Medical school b. Residency c. Fellowship d. All of the above

669

References 1. Wyngaarden JB. The clinical investigator as an endangered species. N Engl J Med 1979;301:1254e9. 2. Gill GN. The end of the physician scientist. Am Scholar 1984;53: 353e68. 3. Kelley WN, Randolph MA, editors. Careers in clinical research: obstacles and opportunities. Washington (DC): National Academy Press; 1994. 4. Goldstein JL, Brown MS. The clinical investigator: bewitched, bothered, and bewildereddbut still beloved. J Clin Invest 1997;99: 2803e12. 5. Nathan DG. Careers in translational clinical researchdhistorical perspectives, future challenges. JAMA 2002;287:2424e7. 6. Nathan DG. Clinical research: perceptions, reality, and proposed solutions. JAMA 1998;280:1427e31. 7. Nathan DG, Varmus HE. The National Institutes of Health and clinical research: a progress report. Nat Med 2000;6:1201e4. 8. Ognibene FP, Gallin JI, Baum BJ, Wyatt RG, Gottesman MM. Outcomes from the NIH clinical research training program: a mentored research experience to enhance career development of clinicianescientists. Acad Med 2016;91:1684e90. 9. Gallin JI. The need for clinical research education in the medical school curriculum. Proc Assoc Am Phys 1998;110:93e7. 10. Association of American Medical Colleges. Promoting translational and clinical science: the critical role of medical schools and teaching hospitals. Report of the AAMC Task Force II on Clinical Research. Washington (DC): Association of American Colleges; 2006. 11. Dickler HR, Korn D, Gabbe SG. Promoting translational and clinical science: the critical role of medical schools and teaching hospitals. PLoS Med 2006;3:1492e5. 12. Teo AR. The development of clinical research training: past history and current trends in the United States. Acad Med 2009; 84:433e8. 13. Meador KJ. Decline of clinical research in academic medical centers. Neurology 2015;85:1171e6. 14. Garrison HH, Deschamps AM. NIH research funding and early career physician scientists: continuing challenges in the 21st century. FASEB J 2014;28:1049e58. 15. Bradford WD, Pizzo S, Christakos AO. Careers and professional activities of graduates of a medical scientist training program. J Med Educ 1986;61:915e8. 16. Bradford WD, Anthony D, Chu CT, Pizzo SV. Career characteristics of graduates of a medical scientist training programs. 1970e1990. Acad Med 1996;71:484e7. 17. Sutton J, Killian CD. The MDePhD researcher: what species of investigator? Acad Med 1996;71:454e9. 18. Brass LF, Akabas MH, Burnley LD, et al. Are MD-PhD programs meeting their goals? An analysis of career choices made by graduates of 24 MD-PhD programs. Acad Med 2010;85:692e701. 19. Andriole DA, Whelan AJ, Jeffe DB. Characteristics and career intentions of the emerging MD/PhD workforce. JAMA 2008;300: 1165e73. 20. Jeffe DB, Andriole DA, Wathington HD, Tai RH. Educational outcomes for students enrolled in MD-PhD programs at medical school matriculation, 1995-2000: a national cohort study. Acad Med 2014;89:84e93. 21. Howard Hughes Medical Institute (website): http://www.hhmi. org/developing-scientists/program-archive. 22. Howard Hughes Medical Institute (website): https://www.hhmi. org/developing-scientists/medical-research-fellows-program. 23. Doris Duke Charitable Foundation (website): http://www.ddcf. org/grants/grant-recipients/2010-clinical-research-fellowship-formedical-students. 24. Sarnoff Cardiovascular Research Foundation (website): http:// www.sarnofffoundation.org/fellowship_info.

IV. CLINICAL RESEARCH INFRASTRUCTURE

670

38. CAREER PATHS IN CLINICAL RESEARCH

25. National Institutes of Health Clinical Research Training Program (Wikipedia): http://en.wikipedia.org/wiki/National_Institues_ of_Health_Clinical_Research_Training_Program. 26. Medical Research Scholar Program (website): http://www. clinicalcenter.nih.gov/training/mrsp. 27. Fang D, Meyer RE. Effect of two Howard Hughes Medical Institute research training programs for medical students on the likelihood of pursuing research careers. Acad Med 2003;78:1271e80. 28. Gallin EK, LeBlancq SM. Launching a new fellowship for medical students: the first years of the Doris Duke Clinical Research Fellowship program. J Invest Med 2005;53:73e81. 29. Laskowitz DT, Drucker RP, Parsonnet J, Cross PC, Gesundheit N. Engaging students in dedicated research and scholarship during medical school: the long-term experiences at Duke and Stanford. Acad Med 2010;85:419e28. 30. Accreditation Council on Graduate Medical Education common program requirements the learning and working environment (duty hours). Available from: http://www.acgmecommon.org/. 31. Oxnard GR, Zinkus TM, Bazari H, Wolf M. Integrating research training into residency: tools of human investigation. Acad Med 2009;84:1295e300. 32. National Institutes of Health. Physician-scientist workforce working group (chaired by David Ginsburg) report. 2014. Available from: http://acd.od.nih.gov/reports/PSW_Report_ACD_06042014.pdf. 33. Milewicz DM, Lorenz RG, Dermody TS, Brass LF. Rescuing the physician-scientist workforce: the time for action is now. J Clin Invest 2015;125:3742e7. 34. Ley TJ, Rosenberg LE. The physician-scientist career pipeline in 2005. Build it and they will come. JAMA 2005;294:1343e51. 35. National Institutes of Health Loan Repayment Program (website): http://www.lrp.nih.gov/. 36. Association of American Medical Colleges. Medical student education: debt, costs, and loan repayment fact card. Available from: https://members.aamc.org/eweb/upload/2015%20Debt%20Fact %20Card.pdf. 37. Goldhamer MEJ, Cohen AP, Bates DW, Cook EFC, Davis RB, Singer DE, Simon SR. Protecting an endangered species: training physicians to conduct clinical research. Acad Med 2009;84:439e45. 38. National Institutes of Health. K30 clinical research curriculum award. Available from: https://archives.nih.gov/asites/grants/05-212015/training/k30.htm. 39. National Institutes of Health Clinical and Translational Science Awards (website): https://ncats.nih.gov/ctsa/about. 40. National Center for Advancing Translational Sciences CTSA Scholar and Research Programs (website): https://ncats.nih.gov/ ctsa/training/programs. 41. Meyers FJ, Begg MD, Fleming M, Merchant C. Strengthening the career development of clinical translational scientist trainees: a consensus statement of the clinical translational science award (CTSA) research education and career development committees. Clin Trans Sci 2012;5:132e7. 42. Rubio DM, Primack BA, Switzer GE, Bryce CL, Seltzer DL, Kapoor WN. A comprehensive career-success model for physician-scientists. Acad Med 2011;86(12):1571e6. 43. Lee LS, Pusek SN, McCormack WT, et al. Clinical and translational scientist career success: metrics for evaluation. Clin Trans Sci 2012;5:400e7. 44. Gallin JI. The NIH Clinical Center and the future of clinical research. Nat Med 2011;17:1221e3.

45. Honey K. True dedication to clinical research: the clinical center of the National Institutes of Health receives the 2011 Mary Woodard Lasker award for public service. J Clin Invest 2011;121:3778e81. 46. Gallin JI. Principles and practice of clinical research. 1st ed. San Diego (California): Academic Press; 2002. 47. Gallin JI, Ognibene FP, Johnson LL. Principles and practice of clinical research. 4th ed. Oxford (UK): Elsevier; 2018. 48. Ognibene FP, Shi TZ, Gallin JI. Global outreach in clinical research by the NIH Clinical Center: building training partnerships with China. In: Sanders S, editor. Selected presentations from the 2011 Sino-American symposium on clinical and translational medicine, a sponsored supplement to science. Washington (DC): Sponsored by Global MD Organization; Science/AAAS; 2011. p. 8e9. Available from: http://science.imirus.com/Mpowered/book/vscim11/i3/p1. 49. Gallin JI. Principles and practice of clinical research. 1st ed. Tokyo: Maruzen Co. Ltd.; USA: Elsevier Science; 2004. 50. Gallin JI, Ognibene FP. Principles and practice of clinical research. 3rd ed. Tokyo: Maruzen Publishing Co. Ltd.; USA: Elsevier Science; 2016. 51. Gallin JI, Ognibene FP. Principles and practice of clinical research. 2nd ed. Burlington (MA): Elsevier; China Science Publisher; 2007. 52. Gallin JI, Ognibene FP. Principles and practice of clinical research. 3rd ed. Burlington (MA): Elsevier (Singapore); China Science Publishing & Media Ltd.; 2013. 53. Gallin JI, Ognibene FP. Principles and practice of clinical research. 2nd ed. Russia: Elsevier; Medprint Russia; 2013. 54. Atkinson AJ, Daniels CE, Dedrick RL, Grudzinskas CV, Markey SP. Principles of clinical pharmacology. 1st ed. San Diego (California): Academic Press; 2001. 55. Atkinson AJ, Huang S-M, Lertora JJL, Markey SP. Principles of clinical pharmacology. 3rd ed. Oxford (UK): Elsevier; 2012. 56. Atkinson AJ, Abernethy DR, Daniels CE, Dedrick RL, Markey SP. Principles of clinical pharmacology. 2nd ed. Burlington (MA): Elsevier; China Science Publisher; 2007. 57. Emanuel EJ, Crouch RA, Arras JD, Moreno JD, Grady C. Ethical and regulatory aspects of clinical research. Baltimore (MD): Johns Hopkins University Press; 2003. 58. ANA Long Range Planning Committee Co-chairs, Hauser SL, McArthur JC. Saving the clinician-scientist: report of the ANA long range planning committee. Ann Neurol 2006;60:278e85. 59. Ambati BK, Cahoon J. Rejuvenating clinician-scientist training. Invest Ophthalmol Vis Sci 2014;55:1853e5. 60. Burns AM, Kushner JA, Ward MA, Turner TL, Kline MW, Orange JS. Strengthening the pipeline for clinician-scientists: the pediatrician-scientist training and development program at Texas Children’s Hospital. J Peds 2016;172:5e6.e5. 61. O’Hara R, Cassidy-Eagle EL, Beaudreau SA, Eyler LT, Gray HL, Giese-Davis J, Hubbard J, Yesavage JA. Increasing the ranks of academic researchers in mental health: a multisite approach to postdoctoral fellowship training. Acad Med 2010;85:41e7. 62. Sung NS, Crowley WF, Genel M, Salber P, Sandy L, Sherwood LM, et al. Central challenges facing the national clinical research enterprise. JAMA 2003;289:1278e87. 63. Straus SE, Soobiah C, Levinson W. The impact of leadership training programs on physicians. Acad Med 2013;88:710e23. 64. Yin HL, Gabrilove J, Jackson R, Sweeney C, Fair AM, Toto R. Sustaining the clinical and translational workforce: training and empowering the next generation of investigators. Acad Med 2015; 90:861e5.

IV. CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

39 Clinical Research Nursing: A New Domain of Practice Gwenyth R. Wallen, Cheryl A. Fisher National Institutes of Health, Bethesda, MD, United States

O U T L I N E Introduction

671

Clinical Research Nursing: An Evolving Practice Specialty

672

Defining and Documenting the Specialty of Clinical Research Nursing Conceptual Framework: The Domain of Practice Practice Standards for Clinical Research Nursing Standards of Care Standards of Practice Job Descriptions Competency Assessment Defining a Core Curriculum What About Certification?

674 675 677 678 678 678 678 679 679

Legal Scope of Practice Issues What Regulations Govern Practice and Liability in Clinical Research Settings?

679 679

INTRODUCTION Throughout the modern history of clinical trials research, nurses have been associated with the implementation and coordination of research involving human subjects, and with the care, education, and advocacy of clinical research participants. Nurses were essential to the creation of designated centers for clinical research such as the National Institutes of Health (NIH) Clinical Center1,2 (Fig. 39.1) and the national network of General Clinical Research Centers (GCRCs) located at academic medical centers throughout the

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00039-3

Tools to Assist a Principal Investigator in Staffing a Study 680 Planning a Study in the Clinical Setting 680 Assessing the Need for Nursing Support 681 Creating the Staffing Plan 681 The Concept of “Research Intensity” 682 Future Considerations Career Potential for Nurses in Clinical Research Meeting the Need for Nurses to Fill Clinical Research Roles Nursing Role in Community-Based Research Supporting the Transition of Nurses Into Clinical Research From Clinical Practice

682 682

Summary/Discussion Questions

684

Acknowledgment

684

References

684

683 683 684

United States.3e5 Nurses within these federally funded centers provided a critical mass of expertise as the research enterprise was transformed, including shifting of fixed resources of the former GCRCs into the much more fluid consortium of Clinical and Translational Science Award (CTSA) centers with their focus on team science, network-based resource sharing, and extension into the community6e10 (Fig. 39.2). Although their role often has not been documented, nurses have been active participants in the US drug development industry, both as clinical trial coordinators and as clinical staff working in industry-sponsored

671

Copyright © 2018. Published by Elsevier Inc.

672

39. CLINICAL RESEARCH NURSING

nurses in community clinical practice are increasingly likely to find themselves caring for a patient who is enrolled in a clinical trial.26e34 The current number of nurses working in clinical research in various capacities is unknown. Efforts to define nursing work supporting clinical research as an evolving area of clinical specialization and leadership are currently underway by the International Association of Clinical Research Nurses (IACRN). This work is building on previous work by Castro6 and colleagues from the NIH Clinical Center Nursing Department. Defining clinical research nursing as a specialization includes formally documenting the scope of practice of nurses practicing in clinical research in contrast to other research team members and other clinical staff working in health-care settings.

CLINICAL RESEARCH NURSING: AN EVOLVING PRACTICE SPECIALTY FIGURE 39.1

Admitting the first patient to the NIH Clinical Center

in 1953.

FIGURE 39.2

Nurse interacting with a research participant in 2010.

clinical pharmacology units.11e16 Nurses also have been essential to the growth of the international clinical research infrastructure,17e22 with their scope of practice varying according to the professional environment encountered in different countries around the world.23e25 In addition to increased visibility of nurses working in designated research centers and programs,

Clinical research nursing is defined as clinical nursing practice with a specialty focus on research implementation and the care of subjects participating in clinical research. In addition to providing and coordinating clinical care, clinical research nurses have a central role in ensuring participant safety, ongoing maintenance of informed consent, integrity of protocol implementation, accuracy of data collection, and data recording, and follow-up. Care received by research participants is driven by study requirements and the collection of research data, as well as by clinical indications. Study procedures may include administration of investigational drugs, performance of an experimental or investigational surgical or radiologic procedure, detailed clinical assessment or phenotyping to characterize the natural history and origin of a disease, or delivery of a psychosocial intervention. Additional care requirements may be imposed by the response of the participant to the investigational intervention and may range from education strategies regarding self-monitoring to comprehensive physiologic support provided in an intensive care unit. Clinical research nurses must continually balance the individual clinical needs of research participants and protocol requirements as they practice, while assuring consistency of baseline treatment and monitoring across participants within a study. As a clinical specialty, clinical research nursing incorporates the two main roles assumed by nurses practicing in clinical research settings: clinical research nurses and research nurse coordinators. Clinical research nurses or clinical research care nurses are clinical research staff nurses with a central focus on care of research participants. They support study implementation within the context of the care delivery setting and are located primarily

IV. CLINICAL RESEARCH INFRASTRUCTURE

CLINICAL RESEARCH NURSING: AN EVOLVING PRACTICE SPECIALTY

in dedicated clinical research settings, such as the NIH Clinical Center, in clinical research units located in academic medical centers, and in private sector research units across the country. These clinical research care nurses are part of the permanent infrastructure of the research organization and are available to investigators accessing the facility. In the second role, research nurse coordinators are primarily responsible for study coordination and data management, with a central focus on managing subject recruitment and enrollment, consistency of study implementation, data management and integrity, and compliance with regulatory requirements and reporting. Research nurse coordinators often are hired by and report to a principal investigator or to the contract research organization (CRO) for support of a specific study or group of studies. They may rely on clinical staff to deliver “hands on” care, including administration of investigational drugs or interventions. Fig. 39.3 shows the continuum of focus in clinical research nursing from clinical care to study management. Nurses in clinical research practice must continually balance the requirements for protocol integrity and data quality with the clinical needs, comfort, and safety of research participants.26,35e40 This “balancing act” is one of the hallmarks of the specialty and requires clinical expertise combined with a strong sense of timing. A classic example of the conflict that presents itself to staff working in clinical research is that of a postoperative patient receiving an investigational drug that is being studied for pharmacokinetics and requires precisely timed blood draws. The nurse may be faced with the situation wherein pain management for the patient is in competition with accurate blood draws and possibly the entire data set for that participant. Expert clinical research nurses know these intersections of care and research very well and plan ahead to avoid conflicts. In cases where there is a conflict between research goals and patient comfort and safety, it is a key standard of care for

673

clinical research nursing that patient safety and comfort always come first. See Fig. 39.4 for a depiction of the ways that research and care are intertwined in the process of clinical research. A role for nurses in clinical and translational science that continues to cause confusion and discourse in the domain of clinical research nursing is that of the nurse scientist, who assumes the role of principal or associate investigator for a given study or program of research. The nurse scientist role is a critical one in the development of portfolios of clinical research and in ensuring balance among perspectives in team science and interdisciplinary studies. However, the role of the nurse scientist should be distinguished from the clinical research nursing roles of the clinical research care nurse and the research nurse coordinator. Nurse scientists may have clinical responsibilities for patient management just as a physician scientist could, but the focus of the nurse scientist’s role is in the development and leadership of research, not clinical practice. Conversely, work in developing the new clinical practice specialty of clinical research nursing is directed at roles of nurses in clinical practice, either directly caring for research participants or coordinating studies on behalf of a principal investigator (See later section on career paths for nurses in clinical research). The roles of nurses in clinical research and clinical trials have been described by numerous authors over several decades.41 Content for these descriptions was drawn primarily from settings where there is a high concentration of clinical trials, such as oncology practices in large urban and academic centers, and the national network of GCRCs and later CTSAs funded by NIH to support the development of an infrastructure for clinical research within academic medical centers. Additional information about clinical research roles for nurses and others managing and coordinating studies comes from the pharmaceutical development industry within the United States and globally.

FIGURE 39.3 The continuum of clinical research nursing: clinical care to study management.

IV. CLINICAL RESEARCH INFRASTRUCTURE

674

39. CLINICAL RESEARCH NURSING

FIGURE 39.4 The process of integrating research and clinical care.

Practice settings for clinical research nurses may be found along the continuum of clinical translation from bench science to phase IV clinical trials and clinical effectiveness studies (Fig. 39.5). Table 39.1 displays the steps in translation of health-care discoveries and innovations into general practice and describes likely practice settings in which these types of research would occur and potential roles for nurses at each step.

DEFINING AND DOCUMENTING THE SPECIALTY OF CLINICAL RESEARCH NURSING It has been a challenge to formally document and support the roles of nurses in clinical research, and there has been no standardization of job titles, scope of practice, or specialty identification. For many years, the most consistent group of nurse leaders focused on developing and defining the roles of clinical staff nurses in support of the care of research participants in dedicated research units across the country was the GCRC Nurse Manager group. This group met once per year, had a regional structure and a website supported by the former NIH National Center for Research Resources (NCRR), which funded the GCRCs. Nurses have emerged as leaders over the past 20 years in industry-

focused clinical research organizations such as the Drug Information Association (http://www.diahome. org/DIAHome/Home.aspx), the Association of Clinical Research Professionals (ACRP) (http://www.acrpnet. org/), the Society of Clinical Research Associates (SoCRA) (http://www.socra.org/), and the Association of Clinical Pharmacology Units (http://www.acpu.net), although their voices has often been hidden in the interdisciplinary category of “clinical research coordinator.” In 2009, a new professional association, the IACRN (http://iacrn.memberlodge.org/) was created to focus on developing and communicating the role and impact of clinical research nursing practice globally. Since its inception, the organization has grown to more than 300 nurses with representatives from over 10 countries. Beginning in 2006, nurses at the NIH Clinical Center capitalized on its critical mass of experience and resources and its diversity of research practice settings to formally undertake national leadership in defining the roles and contributions of nurses practicing in clinical research, as well as the needed clinical practice tools to support development of the specialty. This agenda, coined Clinical Research Nursing2010 (CRN2010), has driven the development of a conceptual framework to support nursing practice in clinical research, as well as consensus around standards and documentation of tools to support clinical research practice across the national and global clinical research enterprise.42 In 2007, this work was linked

FIGURE 39.5 The continuum of clinical translation science.

IV. CLINICAL RESEARCH INFRASTRUCTURE

DEFINING AND DOCUMENTING THE SPECIALTY OF CLINICAL RESEARCH NURSING

TABLE 39.1

Nursing Roles along the Continuum of Clinical Discovery

Translation Step

Clinical Research Nursing (CRN) Role

Disease characterization/ phenotyping (natural history studies)

Full role supporting descriptive clinical cohort studies of populations at risk for or diagnosed with specific diseases or conditions. Role may include activities from all CRN dimensions, as well as clinical care, continuity management, and community-based care coordination

Basic science to identify targets, develop assays, and invent or discover compounds or novel treatment approaches for testing

Limited role focused on the interface with human subjects who may be providing samples and demographic or clinical information, including subject recruitment, informed consent, education, and care related to sample collection, processing, and analysis

Preclinical studies

Limited clinical roles for nurses, although knowledge of these studies and their potential clinical application is important for expert clinical research practice

Phase 0/I studies (“first in humans”)

Full role crossing all dimensions, including subject recruitment, study management and regulatory compliance, administration of study intervention and monitoring, assessment of adverse effects, sample collection and processing, support of clinical and diagnostic procedures per study, management of research records and data, subject education and return to community care, participation in team discussion of results, and next steps

Phase II and III studies (clinical trials)

As in Phase 0/I with more of a focus on recruitment and subject retention, management of adherence and treatment fidelity, management of larger datasets, coordination of multiple study procedures and possibly multiple sites, observation for emerging evidence of positive intended clinical effects or identification of adverse effects

Phase IV studies (postmarketing studies, dissemination research, clinical effectiveness research)

Participation in clinical effectiveness research and postmarketing evaluation studies as research coordinator or participating clinical caregiver, application of evidence to practice and incorporation into standards, evaluation of current practice to assess relevance of new evidence

675

with efforts of the GCRC Nurse Manager group, as nurses in the GCRCs grappled with the transition to the funding framework for the CTSA consortium, and the National Clinical Research Nursing Consortium was formed. Leaders from the Consortium met as a “think tank” three times at Rockefeller University in 2006e2007 to lay the groundwork for defining clinical research nursing as a specialty.7,8,10 In 2009, the Consortium was formally transitioned into the new association of IACRN. Collaboration of the IACRN and the NIH Clinical Center continued through the first three national meetings, held 2009e2011 in Boston and Bethesda.

Conceptual Framework: The Domain of Practice Development of a specialty identity that can lead to certification begins with clarification of the domain of practice for the specialty. This conceptual statement describes the overall areas of work, unique contributions, and expected work activities for the specialty. Therefore, defining the unique domain of specialty practice that makes up clinical research nursing was a key first step in documenting that a specialty practice exists. The domain framework creates a conceptual basis for developing specialty practice tools such as job descriptions, practice standards, competency assessment, educational content, and, ultimately, certification. This is similar to the steps taken by various specialty groups ranging from oncology nursing to informatics nursing. The clinical research nursing domain of practice includes 5 dimensions and 53 individual activities, which make up the full range of practice of both clinical nurses providing research-based patient care and study coordinators managing studies (Fig. 39.6 and Table 39.2). The

FIGURE 39.6 The clinical research nursing domain of practice.

IV. CLINICAL RESEARCH INFRASTRUCTURE

676 TABLE 39.2

39. CLINICAL RESEARCH NURSING

Dimensions and Activities Within the Domain of Clinical Research Nursing Practice

CLINICAL PRACTICE DIMENSION (CC) Provision of direct nursing care and support, using the nursing process, to participants in clinical research, their families and significant others. Care requirements are determined by the scope of study participation, the clinical condition of the patient, and the requirements and clinical effects of research procedures. CP 1

Provide direct nursing care to research participants (e.g., interaction with research participants to provide nursing care, administration of research interventions, specimen collection)

CP 2

Provide teaching to research participants and family regarding study participation, participant’s current clinical condition, and/ or disease process

CP 3

Monitor the research participant and report potential adverse events to a member of the research team

CP 4

Record research data (e.g., documenting vital signs, administration of a research compound, participant responses) in approved source document (e.g., the medical record, data collection sheet)

STUDY MANAGEMENT DIMENSION (SM) Management of clinical and research support activities to ensure patient safety, address clinical needs, and ensure protocol integrity and accurate data collection. SM 1

Participate in study development

SM 2

Participate in research participant recruitment

SM 3

Participate in screening potential research participants for eligibility

SM 4

Coordinate and facilitate the collection of research specimens

SM 5

Develop study-specific materials for research participant education

SM 6

Perform quality assurance activities to assure data integrity

SM 7

Participate in the preparation of reports for appropriate regulatory and monitoring bodies or boards

SM 8

Facilitate accurate communication among research sites

SM 9

Facilitate communication within the research team

SM 10

Contribute to the development of case report forms

SM 11

Participate in the set-up of a study-specific database

SM 12

Comply with International Conference on Harmonization Guideline for Good Clinical Practice

SM 13

Collect data on research participant based on study end points

SM 14

Facilitate scheduling and coordination of study procedures

SM 15

Provide nursing expertise to the research team during study development and implementation

SM 16

Protect research participant data in accordance with regulatory requirements

SM 17

Participate in site visits and/or audits

SM 18

Support study grant and budget development

SM 19

Oversee human resources (people) related to research process

SM 20

Record data on approved study documents (e.g., case report forms, research or study database)

SM 21

Facilitate processing and handling (storage and shipping) of research specimens

SM 22

Identify clinical care implications during study development (e.g., staff competencies and resources, equipment)

SM 23

Participate in the identification and reporting of research trends

CARE COORDINATION AND CONTINUITY DIMENSION (CCC) Coordination of research and clinical activities to meet clinical needs, complete study requirements, and manage linkage with referring and primary care providers. CCC 1

Facilitate the education of the interdisciplinary team on study requirements

CCC 2

Collaborate with the interdisciplinary team to create and communicate a plan of care that allows for safe and effective collection of clinical research data

IV. CLINICAL RESEARCH INFRASTRUCTURE

DEFINING AND DOCUMENTING THE SPECIALTY OF CLINICAL RESEARCH NURSING

677

TABLE 39.2

Dimensions and Activities Within the Domain of Clinical Research Nursing Practicedcont’d

CCC 3

Coordinate research participant study visits

CCC 4

Provide nursing leadership within the interdisciplinary team

CCC 5

Coordinate interdisciplinary meetings and activities in the context of a study

CCC 6

Coordinate referrals to appropriate interdisciplinary services outside the immediate research team

CCC 7

Communicate the impact of study procedures on the research participant

CCC 8

Provide nursing expertise to community-based health care personnel related to study participation

CCC 9

Facilitate research participant inquiries and concerns

CCC 10

Provide indirect nursing care (e.g., participation in clinical, unit, and/or protocol rounds; scheduling study-related test) in the context of research participation

HUMAN SUBJECTS PROTECTION DIMENSION (HSP) Facilitation of informed participation by diverse participants in clinical research. HSP 1

Facilitate the initial and ongoing informed consent or assent process

HSP 2

Support research participant in defining his or her reasons and goals for participating in a study

HSP 3

Collaborate with the interdisciplinary team to address ethical conflicts

HSP 4

Coordinate research activities to minimize subject risk

HSP 5

Serve as institutional review board member

HSP 6

Manage potential ethical and financial conflicts of interest for self

CONTRIBUTING TO WHOM IT MAY CONCERN: THE SCIENCE (CS) Contribution as a research team member to the development of new ideas for study and explorations of innovations arising from clinical research findings to practice. CS 1

Disseminate clinical expertise and best practices related to clinical research through presentations, publications, and interactions with nursing colleagues

CS 2

Serve as an expert in a specialty area (e.g., grant reviewer, editorial board, presenter)

CS 3

Participate in the query and analysis of research data

CS 4

Generate practice questions as a result of a new study procedure or intervention

CS 5

Collaborate with the interdisciplinary team to develop innovations in care delivery that have the potential to improve patient outcomes and accuracy of data collection

CS 6

Identify questions appropriate for clinical nursing research as a result of study team participation

CS 7

Mentor junior staff and students participating as members of the research team

CS 8

Perform secondary data analysis to contribute to the development of new ideas

CS 9

Serve as a resource to new investigators

Clinical Research Nursing Domain of Practice was the first major component of the NIH Clinical Center CRN2010 agenda to come to completion and to be offered for validation and internal and external use. Its five dimensions and associated nursing activities were conceptually validated in a national Delphi study of expert nurses with experience managing or implementing clinical research.6 The domain model was further developed into a survey instrument, which was used in a role delineation study at the NIH Clinical Center designed to profile and compare the roles of clinical research care nurses and research nurse coordinators.43

Practice Standards for Clinical Research Nursing Practice standards take the domain of practice for a professional discipline, or a specialty within the discipline, and make them operational by (1) defining outcomes that recipients of services can expect (standards of care) and (2) setting expectations for key process characteristics that are standardized across practitioners to ensure that outcomes are met (standards of practice). Practice standards can be defined by a licensing, regulatory, or accrediting body, such as a State Board of

IV. CLINICAL RESEARCH INFRASTRUCTURE

678

39. CLINICAL RESEARCH NURSING

Nursing, the US Food and Drug Administration (FDA), or The Joint Commission. They are further defined by expert consensus within a specialty organization and made operational by application at the institutional level (hospital policy, clinical research site policy). Standards of Care Standards of care are overall statements of the outcomes that can be expected from nursing. In a clinical research setting, these standards are stated from the perspective of the research participant. These standards are unique to the research environment and serve as the basis for evaluating the system of research implementation and the process of care delivery. They are global statements intended to be applied across clinical specialty areas and levels of care such as inpatient and outpatient settings (see Table 39.3 for Standards of Care developed at the NIH Clinical Center). Standards of Practice Standards of practice are the “how-to” of the discipline or clinical specialty. They include clinical policy statements, standards of practice, standard operating TABLE 39.3

Clinical Research Nursing Standards of Care

Every research participant can expect the following from nursing at the Clinical Center: 1. Research participants can expect to receive evidence-based nursing care consistent with the accepted professional standard related to their particular condition or therapy. 2. Research participants can expect that their care and treatment are consistent with the research protocol guiding their participation, and that valid data are being collected by the nursing staff. 3. Research participants can expect that treatment and monitoring will be individualized to accommodate individual needs, to the extent allowed by the protocol, and that in all cases, participant safety, comfort, and well-being will be placed above research requirements. 4. Research participants can expect prompt assessment and appropriate response to changes in condition or untoward responses to research procedures. 5. Research participants can expect to know which nurse is accountable for their care and how to contact that person. 6. Research participants can expect that nurses will communicate and collaborate effectively with members of the clinical research team to ensure coordinated, high-quality care. 7. Research participants can expect that information about their care and condition is discussed and communicated with confidentiality, and that care is being appropriately documented. 8. Research participants can expect that while in the Clinical Center they will have a sense of being cared for as an individual, and that they will receive prompt, courteous and individualized services from nurses and patient care staff. 9. Research participants can expect to develop an understanding of their condition and research participation and treatment and will be able to manage self-care as appropriate after discharge. 10. Research participants can expect to be involved in discussions and decisions about their plan of care and research participation.

procedures, clinical practice protocols, and clinical procedures. Policy statements clarify scope and authority related to a specific activity by defining who can do what activity, with what level of supervision, and when. Policy may be stated at multiple levels. For instance, a state board of nursing may define the role of nurses in medication administration by saying which practitioners can dispense medications and what aspects of the medication process can be delegated. This statement may be refined and clarified at the institutional level in a policy on medication administration. Most other practice standards include documents that describe how a process is done, including the principles governing performance and specific psychomotor steps to be taken. Although they may be developed in a general form at the level of a specialty organization, practice standards usually are tailored to the detailed requirements of an individual organization. Practice standards for clinical research include not only the processes for clinical care that may be provided but also the processes for activities related to the other four dimensions in the practice (human subjects’ protection, research coordination and continuity management, study management, and contributing to the science6). Good Clinical Practices (GCP) are another good example of practice standards that apply specifically to clinical research.

Job Descriptions A job description for a specific role is the method of applying the domain of practice and standards accepted by a specific setting to a particular job with a particular location and function. The purposes of a job description are to (1) define key activities of the job, (2) specify knowledge, skills, and abilities that will be expected from the incumbent, and (3) provide a basis for evaluating the qualifications of a given individual to predict success in a given role without extensive training and supervision. The basic structure for a job description consists of an overview of the role and where it fits, as well as descriptions of specific job duties that are expected. For clinical research nurses, specific job duties include which activities from the domain of practice are needed for a particular role. Expected education, licensure, and experience and/or certification may be attached to the job description, or they may be defined when the job is posted.

Competency Assessment Competency assessment is the process of verifying that a specific individual can demonstrate the knowledge, skills, and techniques needed to fulfill a specific role. Competency means that a person is able to do a given

IV. CLINICAL RESEARCH INFRASTRUCTURE

LEGAL SCOPE OF PRACTICE ISSUES

job. Competency validation is an expected prerequisite before an individual is assigned to perform a given function independently. The higher the potential risk of the activity, the more rigorous is the assessment of competency before independent assignment. For instance, if a clinical research care nurse will be responsible for administering a phase I agent in a double-blind study, and there is a potential for anaphylaxis if the participant receives active drug, then the nurse not only must be competent in all activities related to control and documentation of a study drug, informed consent, and participant assessment but also must demonstrate knowledge of what is known about the possible mechanism of action and information from current or previous studies about severe reactions. Additional knowledge required would include the process for responding to anaphylaxis, initiating emergency care, performing resuscitation techniques, and ensuring follow-up assessment. Clinical research nurses must be competent to fulfill their role in the research process, as well as their role in clinical care, and must demonstrate critical thinking skills to determine priorities for action if these two roles conflict. At the NIH Clinical Center, five core competencies were added to the core clinical competencies already in place, reflecting the disease-specific clinical care of research participants. The purpose of adding clinical research competencies was to specifically highlight role competencies related to the research process. The framework for these competencies mapped specifically to dimensions within the domain of practice and included clinical practice, human subject protection, contributing to the science, care coordination and continuity, and study management.

Defining a Core Curriculum A core curriculum is an essential step in clinical specialty development because it defines educational content for the specialized area of practice in such a way that it can be delivered to practitioners new to the specialty. This may include practitioners new to their discipline, such as newly graduated nurses, as well as experienced clinicians who want to move into a new specialty. A core curriculum includes the general content areas to be covered and specifies the level of mastery needed; this lays the groundwork for competency assessment. At the NIH Clinical Center, a course titled Fundamentals for the Clinical Research Nurse was developed as competency based education based on the clinical research core competencies. This approach to the educational program development ensured that the learning was directed toward the knowledge and skills required to meet the competencies and that the appropriate learning

679

resources were provided. The core curricula were designed by using a combination of existing resources, including onsite experts in the field.

What About Certification? Certification is a voluntary activity beyond professional licensure, designed to document and substantiate an individual’s expertise and competence in a given area. In clinical specialty development, it is seen as the final step in documenting specialty expertise. However, certification may be identified as a prerequisite for certain job titles or as a preferred qualification in hiring guidelines. For advanced practice, certification may be required by a state licensing authority as a prerequisite for using a specific title with defined scope that is regulated by that authority (i.e., nurse practitioners in many states). Certification for nurses in clinical research is currently available to them as experts in their clinical specialty (e.g., oncology, critical care, pediatrics) or as Clinical Research Associates (CRAs) or Clinical Research Coordinators (CRCs). These research certifications are offered through the ACRPs and the SoCRA, as interdisciplinary organizations representing all clinical research professionals although not specific to nursing. Both organizations have delineated experience requirements to sit for the certification examination, but no requirement for specific professional licensure. Certification is valued by the field, given the facts that certification may be identified as a preference in hiring and that many people become newly certified every year. Certification of clinical research support staff, in particular CRCs who may or may not be nurses, is seen as a surrogate for high quality when a sponsor evaluates potential clinical research sites.44 One of the goals of IACRN is to develop and offer certification specifically for nurses in clinical research. A certification program through the American Nurses Credentialing Center a subsidiary of the American Nurses Association will enable certified clinical research nurses to demonstrate their specialty expertise and validate their knowledge to their organization and to the patients and families. Internationally renowned credentialing programs provide support with educational materials to promote nursing excellence and quality patient outcomes while providing safe, positive work environments.

LEGAL SCOPE OF PRACTICE ISSUES What Regulations Govern Practice and Liability in Clinical Research Settings? Questions about the legal scope of various roles in clinical research are often raised. This can occur in response

IV. CLINICAL RESEARCH INFRASTRUCTURE

680

39. CLINICAL RESEARCH NURSING

to the fact that licensed staff such as registered nurses may be more expensive than unlicensed research support staff, and substitution is considered as a possible response to constrained budgets. Unfortunately, no overarching clear ruling on the scope of nursing practice in clinical research has been put forth. The highest authority is statutory regulation by state boards of nursing, which govern the scope, authority, and role boundaries, in varying levels of detail, for professionals licensed or regulated by the state such as nurses and physicians. In looking to state boards for guidance, two things must be considered: (1) the independent authority of an individual practitioner and (2) the ability of that practitioner to delegate certain functions, including resulting definitions and requirements of supervision. Regulated functions can include visible processes that are easy to identify, such as prescriptive authority or medication administration, as well as aspects of professional judgment that may be more difficult to identify and articulate. It is easier to investigate the regulatory requirements for medication administration than to assess the function of an unlicensed staff member who may be giving incorrect advice to a participant about an emerging symptom, current medication management at home, or the need to come into the clinical facility to be evaluated. Sometimes other regulatory information on the level of accreditation (Joint Commission) or FDA regulation (GCP, Investigator Responsibilities) may provide guidance. Because of the visibility and risk related to missing a developing complication or untoward effect, it is important that the institution sponsoring research be explicit in its understanding of regulatory guidance and in its assignment of staff to various functions that lie within their legal scope of practice. It also is important that job descriptions be specific enough to permit adequate assessment of competence.

TOOLS TO ASSIST A PRINCIPAL INVESTIGATOR IN STAFFING A STUDY Planning a Study in the Clinical Setting Planning a study that will take place in a clinical setting is a major driver of success in participant accrual, subject retention, accuracy and completion of data collected, and prevention or early detection of adverse events. This is especially true in a clinical setting that is used primarily for community clinical care, as the staff may not be expert at implementing research and may make trade-off decisions that compromise the study when pressed for time, or if a patient becomes clinically unstable. Four major areas should be considered in planning, all driven by the design of the research study (Fig. 39.7).

FIGURE 39.7

Implementing a clinical study.

A clear understanding of the types of participants that will be expected (including special clinical monitoring or concomitant care requirements) is the first step in planning, as it dictates the type of clinical site that would be most appropriate. For instance, patients with acute trauma are most likely to be found in emergency departments, trauma centers, or acute care/critical care settings in hospitals. It would be challenging to design a study that could recruit such patients to an alternate site such as a free-standing clinical research unit. The second factor that needs to be considered includes the procedural research requirements for the study. First and foremost, does it require clinical intervention? Even an intervention as simple as administering a study drug to a healthy volunteer, or a novel contrast agent for clinical imaging to a stable individual, requires consideration of possible side effects, adverse effects, and monitoring needed from the clinical staff. If clinical monitoring or data collection is required, this needs to be stated explicitly, especially the requirements for support from staff members who are part of the clinical care staff on the unit. Consideration of intervention and monitoring requirements in the study design will reveal the clinical competencies needed; this also should help to clarify the types of clinical support needed. This constitutes the third area for planning. Does the work require intervention by a registered nurse? What level of critical thinking and decision-making will be assumed by the clinical staff versus the research team? If the study takes place in a health-care setting, what role will study team members have on the clinical unit, and do their

IV. CLINICAL RESEARCH INFRASTRUCTURE

TOOLS TO ASSIST A PRINCIPAL INVESTIGATOR IN STAFFING A STUDY

duties require institutional credentialing? This step clarifies the investigator’s understanding of the types of participants expected, the specific research activities required, and the new or existing competencies needed to support implementation. The first three steps lead to the fourth and final step, which is detailed planning for protocol set-up and implementation. This step proceeds most quickly if there is direct communication between the principal investigator or delegated research manager and the clinical leadership in the facility, whether a community-based health-care setting or a dedicated research unit. This final step should occur before recruitment and admission of the first participant. Completion of the protocol planning process also can include sharing plans for progress updates and final discussion with the staff of any findings that can be shared. This enhances receptivity to supporting clinical research when future investigators approach the site.

Assessing the Need for Nursing Support A major challenge in allocating resources to support clinical research is the fact that having a clinical research subject in a health-care setting may be totally independent of any clinical need for health-care services but instead may be driven by the need to have the subject available for frequent monitoring for data collection or for a tightly timed series of diagnostic procedures and tests, each requiring preparation and data collection. For instance, a healthy volunteer may be admitted overnight to a clinical research unit, have an intravenous line (IV) inserted, and require every-15-min blood sampling for 24 h. Timing and handling of specimens must be accurate and precise for the data to be included in the study analysis. Blood sampling may have to be interrupted at a precise point, then an investigational drug given, and then the person sent for imaging. Historically, resource allocation at clinical research sites has been generous, to permit exact timing and sequencing of research and clinical activities, so as not to jeopardize the time-sensitive relationships between treatment and response that are often being studied. If the study design requires specific timing and sequencing of treatments and/or sampling and data collection as a component of the research design, a delay in treatment, blood draws, or getting a patient to imaging services may prevent the use of data from that patient in analysis, thus wasting the efforts of all concerned. This translates into a requirement for nursing staffing that ensures adequate staffing not to delay administration of scheduled medications, the start of treatment, assessment of response, or management of side effects. Several recent studies identify activities that should only be performed by a nurse or under the direct supervision of a nurse.

681

These activities cross the spectrum from clinical care when patients are hospitalized for complex trials, to managing adverse events, administering medication, assessing for inclusion/exclusion criteria, case management and education.46,47 These findings led to the development of institutional policies in one facility which stated what activities could be delegated and which were not to be delegated to nonlicensed staff. Another factor often overlooked is that of the need for patient education related to the complexities of participation in clinical research. Clinical research nurses are best prepared to determine participant readiness and the patient’s ability to understand information about research using assessment techniques. Nurses are best suited to determine participant understanding of elements of risk, confidentiality of data, alternatives to participation, injuries associated with greater than minimal risk and volunteerism. As the complexity of informed consent increases, research participants with inadequate health literacy may not be able to fully comprehend information being disclosed.48 In studies involving severely or terminally ill patients, or patients with chronic conditions who must manage their condition, on study, at home between visits to the clinical research site, the relationship between care required to implement study procedures and care needed by a person with acute health disruptions may be blurred. When protocol planning occurs, an estimate is often made of overt impact in terms of planned visits and procedures, but not of intermediate services required to maintain the person as a study participant during disease progression or during potential side effects of study treatment. As resources have been constrained, there has been much discussion of the boundary between standard of care treatment and protocol-required treatment. It is a strong value, deeply held by many clinical research staff, that the research team has a commitment to study subjects to support them as much as possible through treatment, disease sequelae, and follow-up care, all of which can be resource intensive for subjects with advanced clinical disease. The linkage between research activities and clinical activities is embedded into the care delivery and research support process.

Creating the Staffing Plan Staffing decisions for clinical research teams should not only be driven by the scientific and clinical goals of the study but must also balance cost, state-legislated scope of practice, efficiency, potential risks to participants, and the challenges of subject recruitment and retention. The complexity of study coordination and task delegation should be considered across healthcare environments and for the duration of subject

IV. CLINICAL RESEARCH INFRASTRUCTURE

682

39. CLINICAL RESEARCH NURSING

participation. In any clinical research situation, the knowledge base and scope of practice of the staff involved must match the task to be completed. Appropriate use of clinical research nurses has the potential to contribute to the efficiency of study implementation, reduce overall time required to complete a study, improve subject recruitment and retention, and provide safeguards for patient safety. These are important considerations in building a clinical research team.

developing drivers for resource use, in devising an approach to unit-level staffing, and in setting staffing targets in terms of hours of care to be provided per patient day or outpatient visit. Clinical research intensity definitions also can be used to create a model of cost allocation at the location (research site or research unit) level, as at the NIH Clinical Center (Fig. 39.8).

FUTURE CONSIDERATIONS The Concept of “Research Intensity”

Career Potential for Nurses in Clinical Research

An understanding of the variation in intensity of clinical needs and research support requirements is clearly needed for a valid explanation of variations in resource consumption by patient groups, protocols, or institute programs. This requires an extension of the usual concept of intensity or acuity that is captured in patient classification systems. Most patient classification tools assess the clinical intensity and nursing care requirements driven by the presenting condition and needs of the patient. To account for intensity in a clinical research environment, it is important to clearly differentiate research-based clinical intensity and research intensity that arises from the interaction of the research participant with the research process, and with direct care activities within the dimensions of clinical research nursing that is not captured by the current patient classification tool. One example of a set of clinical trial workload-related determinants described by Good and colleagues in a community oncology clinical trial program describes the use of six criteria: (1) treatment complexity; (2) trial-specific laboratory and testing requirements; (3) potential for treatment toxicity; (4) number and complexity of required data forms; (5) degree of coordination required with other departments and disciplines; and (6) number of random assignments or steps in the trial.45 Definitions developed to guide the process of research participation intensity classification are shown in Table 39.4. These definitions have been used in

Nurses can become involved with clinical research at multiple points in a clinical career, beginning with student experiences during early training. Clinically focused nurses can spend an entire career as clinical practitioners contributing at many levels to the support and coordination of clinical research. This work defines the specialty focus of clinical research nursing. Some nurses, however, may want to move beyond the role of clinical research nursing into roles focused on the actual production of clinical research as a principal investigator and nurse scientist. Because of the size and scope of the clinical research programs at the NIH Clinical Center, nurses coming to practice at the facility have opportunities for contribution at many different levels. The nursing career progression model at the Clinical Center defines two distinct career development trajectories along an eight-level continuum, with increasing involvement in the research process and increasing accountability for the independent conduct of clinical research studies (Fig. 39.9). The career path for clinical research nursing that focuses on the care of clinical trial participants spans from undergraduate nursing student through advanced practice nurses. In some academic medical centers, including the NIH Clinical Center, advanced practice nurses may choose to continue their academic training as they pursue a research path that includes independent research as a Clinical Nurse Scientist. A limitation of the model described in Fig. 39.9 continues to be the many evolving roles in clinical research nursing including clinical research nursing education and clinical research nursing leadership roles. The level of responsibility within the clinical research program increases as the nurse moves from a role focused on study implementation and care of research subjects through active involvement in subject recruitment and data management to study coordination, and then potentially to involvement in study design and analysis, and to the responsibilities for study quality and outcomes. The practice specialty of clinical research nursing offers a variety of career advancement opportunities and can easily accommodate a full career of increasing contribution.

TABLE 39.4

Definitions of Nursing Intensity in Clinical Research

Clinical intensitydNursing care requirements driven by the presenting condition and needs of the patient, including physical, psychiatric, psychosocial, educational, and coordination of care needs. Research-based clinical intensitydNursing care requirements that arise from the interaction of the patient with the research process as identified by the dimension of clinical practice. Research intensitydNursing care requirements that are driven by the research protocol as identified by the dimensions of study management, care coordination and continuity, human subjects protection, and contribution to the science.

IV. CLINICAL RESEARCH INFRASTRUCTURE

FUTURE CONSIDERATIONS

683

FIGURE 39.8 Measurement model: nursing intensity in clinical research.

clinical trials become more complex and require more advanced coordination skills. We do know that aside from the content provided on nursing research and evidence-based practice, there is little focus in undergraduate nursing curricula on the role of clinical research in advancing health, or on the potential career trajectories available in clinical research.50 Filling the gap in educational content, as well as increasing awareness of job opportunities in clinical research among students and faculty, will be a top priority over the next decade. FIGURE 39.9 Nursing career path in clinical science.

The career development model should not be conceptualized as a rigid framework that requires the nurse to proceed sequentially through each step. Observations in the clinical setting show that nurses may move into a research-oriented career at essentially any point along the development of a clinical career, and that an individual career trajectory may include multiple moves into and out of research involvement.

Meeting the Need for Nurses to Fill Clinical Research Roles Because there is no centralized data source on the available nursing workforce in clinical research, it is difficult to assess current capability or to project whether there will be problems with supply and demand. One can project that with the likely significant shortage of nurses over the next decade,49 fewer nurses will be attracted to research. Even now, there are anecdotal examples of challenges faced by research sites in filling positions for research nurse coordinators, especially as

Nursing Role in Community-Based Research As the research agenda moves to community practice settings, and as the focus on patient-oriented clinical effectiveness research increases, more and more nurses in community practice settings will find themselves faced with caring for a patient who is participating in a clinical trial, or actually providing a research intervention as part of a trial being conducted with patients in their practice setting. Whether in a critical care setting26,32 or a primary care setting,27 nurses find themselves juggling time requirements for administering study treatments and assessments with the expected regular care to be provided to the participant. Staff nurses caring for research patients may or may not be familiar with the process of clinical research and the regulations that govern clinical research. For instance, in a critical care setting, if a patient is participating in a device trial, the assigned bedside nurse will need information about the proposed mechanism of action and any potential adverse effects to more effectively assess patient clinical response32 and to provide patient and family education. Nurses may be required to handle questions and concerns from family members when

IV. CLINICAL RESEARCH INFRASTRUCTURE

684

39. CLINICAL RESEARCH NURSING

other members of the investigative team are not available to assist with responding. These situations require significant understanding of the research process and ethical principles governing human subjects’ research; both of these content areas are not normally covered in practical depth in traditional nursing clinical education.

Supporting the Transition of Nurses Into Clinical Research From Clinical Practice A frequent topic of conversation among nurses managing clinical research programs is the best way to recruit nurses into clinical research. Although this need has not been formally evaluated, a frequent statement reflects the view that recruiting a nurse with good clinical experience and critical thinking skills developed in practice settings is preferable to recruiting a nurse whose career experience has been exclusively in research. Managers will say, “I need the clinical competence and critical thinking skillsdwe can teach the research part.”

SUMMARY/DISCUSSION QUESTIONS 1. Which of the following are included as dimensions within the clinical research nursing practice domain? a. Clinical practice, human subjects’ protection, and medication administration b. Clinical practice, study management, and human subjects’ protection c. Patient financial services, patient counseling, and human subjects’ protection d. Clinical practice, human subjects’ protection, and data analysis 2. The focus of nurses practicing as Research Nurse Coordinators is to a. Coordinate studies and manage data b. Manage subject recruitment and enrollment c. Assure compliance with regulatory requirements and reporting d. All of the above 3. One of the major challenges facing nurses practicing in clinical research is a. Managing the balance of clinical and research requirements for research participants b. Maintaining accurate clinical data c. Managing the complexity of working with multiple disciplines d. Keeping up with required clinical competencies 4. The highest authority regulating nursing scope of practice in clinical research nursing is a. Decisions made by the Institutional Review Board b. Orders and instructions from the Principal Investigator

c. The State Nurse Practice Act d. Institutional policies 5. The need for nursing support for a clinical study is driven by a. The clinical condition of the participants to be recruited b. The complexity of clinical procedures required by the protocol c. The need for clinical judgment in following and assessing participants d. All of the above

Acknowledgment The authors gratefully acknowledge Dr. Clare Hastings for her contributions to the earlier edition of this chapter and for leadership in the development of clinical research nursing as a new domain of practice.

References 1. Gallin JI, Varmus H. Revitalization of the Warren G. Magnuson clinical center at the national institutes of health. Acad Med 1998; 73:460e6. 2. Masur J, Thompson N. The national clinical center for chronic disease research. Hospitals 1949;23:45e56. 3. Luft FC. The role of the general clinical research center in promoting patient-oriented research into the mechanisms of disease. J Mol Med 1997;75:545e50. 4. Mori C, Mullen N, Hill EE. Describing the role of the clinical research nurse. Res Pract 2007;8:220e8. 5. Robertson D, Tung CS. Linking molecular and bedside research: designing a clinical research infrastructure. J Mol Med 2001;79: 686e94. 6. Castro K, Bevans M, Miller-Davis C, Cusack G, Loscalzo F, Matlock AM, et al. Validating the clinical research nursing domain of practice. Oncol Nurs Forum 2011;38:E72e80. 7. Coller BS. Thinking globally, acting locally: Rockefeller University’s enterprising CTSA work. Clin Transl Sci 2008;1:190e1. 8. Cunningham J, Miller ST, Joosten Y, Elzey JD, Israel T, King C, Luther P, Vaughn Y, Wilkins CH. Community-engaged strategies to promote relevance of research capacity-building efforts targeting community organizations. Clin Transl Sci 2015;8:513e7. 9. Hastings CE, Fisher CA, McCabe MA, The National Clinical Research Nursing Consortium. Clinical research nursing: a critical resource in the national research enterprise. Nurs Outlook 2012;60: 149e56. 10. Offenhartz M, McClary K, Hastings C. Nursing and realities of clinical research. Nurs Manage 2008;39:34e9. 11. Muller S, Weigelt J. Open-access publiceprivate partnerships to enable drug discoverydnew approaches. IDrugs 2010;13:175e80. 12. Pitler L, Pompei RA, Malarick C, Sutherlin RG, Leamy JH, Larkin ME. Demystifying clinical research nursing. The Monitor 2009;23:31e4. 13. Rico-Villademoros F, Hernando T, Sanz JL, Lopez-Alonso A, Salamanca O, Camps C, et al. The role of the clinical research coordinatorddata managerdin oncology clinical trials. BMC Med Res Methodol 2004;4:6. 14. Smith W, Salenius S, Cobb C, Marzan R, Sabina S, Beccaria L, et al. A survey of clinical research coordinators in the Cooperative group setting of the American College of Radiology imaging network (ACRIN). Acad Radiol 2010;17:1449e54.

IV. CLINICAL RESEARCH INFRASTRUCTURE

REFERENCES

15. Vitiello B, Heiligenstein JH, Riddle MA, Greenhill LL, Fegert JM. The interface between publicly-funded and industry-funded research in pediatric psychopharmacology: opportunities for integration and collaboration. Biol Psychiatry 2004;56:3e9. 16. Yin CX. Issues in clinical trials management. Improving the quality of clinical research: recognizing issues in training. Res Pract 2008;9:20e3. 17. Gordon C. Exploring the new specialty of clinical research nursing. Nurs Times 2008;104:34e5. 18. Groft SC, de la Paz MP. Rare diseasesdavoiding misperceptions and establishing realities: the need for reliable epidemiological data. Adv Exp Med Biol 2010;686:3e14. 19. Merritt Jr MG, Railey CJ, Levin SA, Crone RK. Involvement abroad of U.S. academic health centers and major teaching hospitals: the developing landscape. Acad Med 2008;83:541e9. 20. Rath L, Hitchcock A, Oakley M, Graham J. Clinical drug trials and the role of the nurse. Aust Nurs J 2003;10:19e21. 21. Spilsbury K, Petherick E, Cullum N, Nelson A, Nixon J, Mason S. The role and potential contribution of clinical research nurses to clinical trials. J Clin Nurs 2008;17:549e57. 22. Velasquez GS, Stahl H-D. Multinational Clinical Research Centers: a paradigm shift in recruitment for global clinical trials. The Monitor 2008;22:33e6. 23. Bell J. Towards clarification of the role of research nurses in New Zealand: a literature review. Nurs Prax NZ 2009;25:4e16. 24. Cairns JA, Yusuf S, Cook RJ, Cox J, Dagenais GR, Devereaux PJ, et al. Canadian Network and Centre for Trials Internationally (CANNeCTIN): a national network for Canadian-led trials in cardiovascular diseases and diabetes mellitus. Can J Cardiol 2010;26:353e8. 25. Catania G, Poire I, Bernardi M, Bono L, Cardinale F, Dozin B. The role of the clinical trial nurse in Italy. Eur J Oncol Nurs 2011. http:// dx.doi.org/10.1016/j.ejon.2011.04.001. 26. Alt-White AC, Pranulis MF. Addressing nurses’ ethical concerns about research in critical care settings. Nurs Adm Q 2006;30:67e75. 27. Boase S, Kim Y, Craven A, Cohn S. Involving practice nurses in primary care research: the experience of multiple and competing demands. J Adv Nurs 2011. http://dx.doi.org/10.1111/j.1365e2648. 2011.05764.x. 28. Cohen HL, Davis L, Hunter J, Carp D, Geromanos K, Sunkle S. Coordinating a large multicentered HIV research project. J Assoc Nurses AIDS Care 2007;8:41e50. 29. Fagnan LJ, Davis M, Deyo RA, Werner JJ, Stange KC. Linking practice-based research networks and Clinical and Translational Science Awards: new opportunities for community engagement by academic health centers. Acad Med 2010;85:476e83. 30. Green LA, White LL, Barry HC, Nease Jr DE, Hudson BL. Infrastructure requirements for practice-based research networks. Ann Fam Med 2005;3(Suppl 1):S5e11. 31. Holaday B, Mills DM. Clinical research and the development of new drugs: issues for nurses. Dimens Crit Care Nurs 2004;23:179e86. 32. Lanter J. Clinical research and the development of new devices: considerations for nurses. Dimens Crit Care Nurs 2007;26:117e20.

685

33. Miller KP, Dickerson J. Should critical care nurses participate in pharmacologic research? Dimens Crit Care Nurs 1984;3:162e70. 34. Rosal MC, White MJ, Borg A, Scavron J, Candib L, Ockene I, et al. Translational research at community health centers: challenges and successes in recruiting and retaining low-income Latino patients with type 2 diabetes into a randomized clinical trial. Diabetes Educ 2010;36:733e49. 35. Easter MM, Henderson GE, Davis AM, Churchill LR, King NM. The many meanings of care in clinical research. Sociol Health Illness 2006;28:695e712. 36. Fedor CA, Cola PA, Pierre C. Responsible research: a guide for coordinators. London; Chicago, IL: Remedica; 2006. 37. Grady C, Edgerly M. Science, technology, and innovation: nursing responsibilities in clinical research. Nurs Clin North Am 2009;44: 471e81. 38. Oberle K, Allen M. Ethical considerations for nurses in clinical trials. Nurs Ethics 2006;13:180e6. 39. Stephens-Lloyd A. The extended role of the clinical research nurse: building an evidence base for practice. NT Res 2004;9:18e27. 40. Trocky NM. The journey to becoming a research nurse. Clin J Oncol Nurs 2001;5:157e8. 160. 41. Bird J, Kirshbaum M. Towards a framework of advanced nursing practice for the clinical research nurse in cancer care. Clin Eff Nurs 2005;9:161e71. 42. Hastings C. Clinical research nursing 2010: Background and overview. NIH Clin Cent Nurs Patient Care Serv 2009. 43. Bevans M, Hastings C, Wehrlen L, Cusack G, Matlock AM, MillerDavis C, Tondreau L, Walsh D, Wallen GR. Defining clinical research nursing practice: results of a role delineation study. Clin Transl Sci 2011;4:421e7. 44. Lightfoot GD, Sanford SM, Shefrin A. Can investigator certification improve the quality of clinical research? Qual Manag Health Care 1999;7:31e6. 45. Good MJ, Lubejko B, Humphries K, Medders A. Measuring clinical trialeassociated workload in a community clinical oncology program. J Oncol Pract 2013;9:211e5. 46. Thomas Jones C, Hastings C, Wilson L. Research nurse manager perceptions about research activities performed by non-nurse clinical research coordinators. Nurs Outlook 2015;63:474e83. 47. Olson R, Boisjolie C, Goldman J, Pink V, Williams KA. Scope of practice in clinical research: a cardiovascular research institute’s approach. The Monitor 2012;26(6):23e30. 48. Tamariz L, Palacio A, Robert M, Marcus E. Improving the informed consent process for research subjects with low literacy: a systematic review. J Gen Intern Med 2013;28(1):121e6. 49. Institute of Medicine. Future of nursing: Leading Change, advancing health. Washington, DC: National Academies Press; 2010. 50. Galassi AL, Grady MA, O’Mara AM, Ness EA, Parreco LK, Belcher AE, Hastings CE. Clinical research education: perspectives of nurses, employers, and educators. J Nurs Educ 2014;53: 466e72.

IV. CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

40 The Importance and Use of Electronic Health Records in Clinical Research Jon W. McKeeby, Patricia S. Coffey National Institutes of Health, Bethesda, MD, United States

O U T L I N E Electronic Medical Record

687

Electronic Health Record

688

Electronic Health Record Architecture 688 Example of an Electronic Health Record Architectural Diagram 688 Electronic Health Record System Connectivity at the National Institutes of Health Clinical Center 688 Clinical Research Information Systems Using an Electronic Health Record in Clinical Research Data Characteristics Clinical Decision Support Within Electronic Health Record Protocol Order Sets Within the Electronic Health Record Sample Protocol Map/Research Grid

688 692 694 694 694 696

ELECTRONIC MEDICAL RECORD The electronic medical record (EMR) provides a platform for the collection, review, and storage of patient information as part of the patient care process. The EMR enables patient management of demographics, allergies, advance directives, informed consent documents; vital sign management; order management including computerized provider order entry (CPOE); result management; clinical decision support (CDS) for health information through clinical documentation and orders; result reporting; medication management; electronic

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00040-X

Secondary Use of the Electronic Health Record for Clinical Research 698 Legislation and the Electronic Health Record Health Information Technology for Economic and Clinical Health Act Medicare Access and Children’s Health Insurance Program Reauthorization Act of 2015 U.S. Food and Drug Administration Guidance for Electronic Health Record in Clinical Research

698

Summary

699

Summary Questions

699

Terms

701

References

701

Further Reading

702

698 698 699

communication to hospital systems through interface engines, typically using Health Level-7 (HL7) standards, and electronic communication to monitors and instruments such as IV pumps; and administrative processes and reporting. The EMR enables the care team to monitor the patient’s current treatment, review data over time, and track wellness events, such as preventative visits and vaccines. The EMR must store, transfer, and allow access to patient information in a secure manner to ensure that the patient’s privacy is protected at all times. The EMR is typically practice or site specific and limited to a certain patient population.1,2

687

Copyright © 2018. Published by Elsevier Inc.

688

40. USE OF EHR IN CLINICAL RESEARCH

ELECTRONIC HEALTH RECORD A key difference between an electronic health record (EHR) and an EMR is interoperability. The EHR extends the EMR by providing access to individuals, based on their roles, outside of the organization, including the patient; interfacing with other organizations; and allowing usage beyond patient care, such as population health research and clinical research.1,2 Examples of an EHR include: • Sending data to a regional health information exchange (HIE) allowing allergies collected by a primary care provider at one organization to be utilized by an emergency department at a separate organization when the patient is unable to communicate. • Enabling a patient to review his/her laboratory results and medication administrations over time. • Allowing a referring physician or other specialist to access a patient’s data from another organization. • Allowing the electronic transfer of laboratory results, medication lists, and progress notes or discharge documentation to facilitate the transfer of care of a patient to a new facility.

ELECTRONIC HEALTH RECORD ARCHITECTURE The goal of the EHR is to perform all activities that support the care of the patient. The EHR is where orders are entered, results are reviewed, clinical documentation is entered, and vital signs are documented and managed. There are two models of EHR architecture. The first model supports clinical activities within departments by managing workflow through a specific module within the EHR. In the second model, the EHR and ancillary systems have different vendors and require data to be transferred between the systems to support clinical processes. Table 40.1 reviews an example outlining the people, modules, and processes involved in the use of a medication management module within an EHR. Table 40.2 reviews an example outlining the people, module, and processes involved in the use of a Laboratory Information Management System (LIMS/LIS), which is a separate system and different software vendor from the EHR. Table 40.3 reviews systems that are components of an EHR.

Example of an Electronic Health Record Architectural Diagram Fig. 40.1 presents the EHR used at the National Institutes of Health Clinical Center (NIH CC), based on the model introduced by Stephen Rosenfeld, MD.3 This is a hybrid model where the EHR has multiple modules, but it interfaces to other software vendor systems as well. The items in blue are components and modules within the EHR. Order entry, clinical decision support, patient information, and clinical documentation are core components of most EHRs. The EHR used by the NIH CC includes a medication module and enterprise scheduling. The green circles are ancillary systems (i.e., Laboratory Information System [LIS] and Radiology Information System) that are interfaced to the EHR. Data from the EHR are provided to the clinical research data warehouse (Biomedical Translational Research Information System) which then feeds NIH institute systems. The yellow circle shows a new system currently in the pilot phase.

Electronic Health Record System Connectivity at the National Institutes of Health Clinical Center Fig. 40.2 shows all of the systems and their connectivity to the EHR. HL7 messages are shown between some ancillary systems and the EHR and include the message types of patient demographics (HL7-ADT), order (HL7ORM), results (HL7-ORU), and documentation (MDM).

CLINICAL RESEARCH INFORMATION SYSTEMS Clinical research information system (CRIS) is a generic term for systems used to manage the full spectrum of a clinical research protocol from study participation, recruitment and accrual milestones, study participant randomization, data collection, data analysis, and data reporting. System types include electronic data capture (EDC), clinical research data warehouse (CRDW), and clinical trials data management system (CTDMS). An EDC is used to capture data as part of the research process. Data from the EHR, or from multiple EHRs for multisite research projects, ancillary systems, and any EDC specific to collection of research data are interfaced to the CRDW. The CRDW allows laboratory, medication, vital sign, radiology, and other phenomic data to be reviewed in an integrated fashion. The CRDW stores

IV. CLINICAL RESEARCH INFRASTRUCTURE

TABLE 40.1

Medication Management Module within an Electronic Health Record (EHR) 1. A physician selects the patient (pediatric) to initiate computerized provider order entry. 2. The physician reviews existing orders and results in the patient’s record to determine which medication to prescribe including its dosage, frequency, and route. 3. During the order entry process, the system executes clinical decision support to check drug to drug and drug to allergy interactions by comparing order information to documented allergies and the medication history for the patient. The system displays appropriate alerts to the prescriber.

EHR Medication Management Module

4. The order is made available to the Pharmacy Department to verify the medication in the medication module. 5. A pharmacist logs into the medication module of the EHR and reviews and verifies the medication order. The pharmacist performing order verification in the EHR is responsible for the dose calculation. Acknowledging alerts indicates the calculation was completed and the dose is appropriate for a pediatric patient. The pharmacist’s review also includes evaluating medication orders for the following: a. appropriateness of drug, dose, frequency, and route of administration b. weight-based, BSA-based, or other age-appropriate dose check for patients under the age of 18 c. renal function, if indicated, for drugs requiring dose adjustment d. accuracy of order entry: all required elements of the order are present e. correct start date and time f. duplication of order or therapy with other current medications g. indication for PRN order h. only a range in dose is permitted not a range in frequency i. IV admixture incompatibilities j. potential allergies or sensitivity alerts and other contraindications, including pregnancy/lactation status, if applicable k. potential interactions l. compliance with protocol or other organizational guidelines m. clarity and completeness of take-home prescriptions

EHR

6. The EHR prompts the nurse to administer medications to a patient through review of the medical care plan or the electronic medication administration record (eMAR), which includes the history of medications administered and medications to be administered.

Interface engine (IE)

7. The order is electronically sent from the EHR to an automated dispensing management system and to the automated dispensing cabinets (ADCs) via an IE.

ADC

8. The nurse reviews the patient’s medication profile at the ADC and removes the medications. The nurse is required to enter credentials and then select the proper patient’s profile to remove medications.

EHR and barcode

9. The nurse goes to the patient’s room, validates patient identification, scans the patient’s wristband with a barcode scanner and scans the medication, while in the EHR. If the medication, patient, dosage, strength, and administration route match, the nurse can administer the medication to the patient. If there is a discrepancy, the nurse is alerted to evaluate and resolve the issue. Resolution may involve calling the Pharmacy Department or the provider and may require orders to be discontinued or new orders to be entered.

EHR eMAR

10. The administration of the medication is documented on the eMAR enabling it to be reviewed by the care team.

Clinical research data warehouse (CRDW)

11. All ordering information, as well as medication administrations, is provided to the CRDW to support analysis as part of clinical research.

CLINICAL RESEARCH INFORMATION SYSTEMS

IV. CLINICAL RESEARCH INFRASTRUCTURE

EHR

689

690

TABLE 40.2 Electronic Health Record (EHR) that Interfaces with a Laboratory Information Management System (LIMS/LIS) External to the EHR 1. Admissions, nurses, and other care team members update patient demographics, allergies, and other patient information.

Interface engine (IE)

2. These demographic changes are converted to an Admissions, Discharge, Transfer (ADT) Health Level 7 (HL7) message and transmitted to the Laboratory Information System (LIS) via the interface engine.

LIMS/LIS

3. LIMS/LIS processes the HL7 message and stores the ADT information.

EHR

4. A physician selects the patient to initiate computerized provider order entry. 5. The physician reviews the patient record and existing orders and determines that a complete blood count (CBC) is to be ordered. 6. A CBC order is entered in the EHR by the physician for the selected patient

IE

7. The interface converts the order to an HL7 message and transmits it to the LIS via the interface engine.

LIMS/LIS

8. LIMS/LIS processes the HL7 message. 9. LIS notifies phlebotomy about the order and the phlebotomist collects the specimen from the patient.

LIMS/LIS barcode

10. The phlebotomy technician verifies identification of the patient, scans the patient’s wristband, prints the order specimen label, confirms the order, collects the specimen, and scans the order specimen label to update the order to a collected status.

IE

11. Upon collection of the specimen an HL7 message is sent to the EHR via the interface engine.

EHR

12. Users are able to see that the specimen was collected. Other statuses that can be sent between the systems include order modifications and order cancellations.

Lab instruments

13. The specimen is delivered to the laboratory medicine department for processing. This process is sometimes referred to as accessioning. 14. The specimen is loaded into the appropriate laboratory instrument and the ordered tests are performed. The laboratory instrument transfers the results to the LIMS/LIS through an instrument interface.

LIS

15. A lab technician verifies the CBC results within the LIS.

IE

16. Results are sent to the EHR as a final result as an HL7 message via the interface engine. Correct and supplemental results can also be sent from the LIMS/LIS to the EHR as an HL7 message via the interface engine.

EHR

17. The results are available in the EHR for review by the care team.

Clinical research data warehouse (CRDW)

18. The order, collection data (including specimen collection date and time), and results are provided to the clinical research data warehouse for analysis as part of clinical research.

40. USE OF EHR IN CLINICAL RESEARCH

IV. CLINICAL RESEARCH INFRASTRUCTURE

EHR

TABLE 40.3

Systems that Comprise the Electronic Health Record (EHR) Main component used to support computerized provider order entry by managing patients, patient information, order entry, vital signs, nursing documentation, clinical documentation, electronic medication administration record, and display of results.

Interface engine

Used to transport data between systems, typically the electronic medical record/EHR and ancillary systems. The data is usually transferred using Health Level 7.

Enterprise scheduling

Manages the scheduling of a patient from an initiated order and/or appointment, including resources such as time, staff, instruments, scanners, and specific rooms.

Laboratory Information Management System (LIMS/LIS)

Manages laboratory-related information from the order, specimen collection, specimen accessioning, and connection to instruments, resulting, verification of results, and transfer to the EHR. Types of tests may include laboratory, anatomic pathology, microbiology, and transfusion medicine.

Transfusion Medicine Systems

Manages the transfusion medicine systems from order entry, specimen collection, blood component management, and result management. Systems may support blood transfusion management, cell processing, human leukocyte antigen testing, and transfusion medicine testing.

Radiology Information System

Manages radiology information from the order, appointment, appointment schedule, scanner/resource management, exam/scan, and result. Modules include diagnostic radiology, ultrasound, computed tomography and computerized axial tomography scan, magnetic resonance imaging, positron emission tomography, nuclear medicine, and interventional radiology.

Picture Archiving and Communication System

Manages storage, retrieval, management, distribution, rendering, manipulation, and presentation of medical images.

Surgical Information System

Manages surgical information from the order, scheduling, case management, anesthesia documentation, surgical information, as well as inventory control within the surgery unit.

Inpatient Pharmacy Management System

Manages the medication administration process from order and verification of order for processing within the Pharmacy Department.

Outpatient Pharmacy System

Manages the outpatient medication management process from order, verification to dispensing.

Nutrition System

Manages nutrition information at the patient level for managing diet orders, meal selection, and inventory within a Nutrition Department.

Health Information Management Systems

Manages health information within an organization. Systems may include coding, dictation and transcription, medical record location, medicolegal tracking.

Patient Engagement Systems

Systems used to provide information to patients. Two main external systemsdthe first is a Patient Portal that allows internet access to one’s patient record at a medical institution typically provided through an EHR and the second is a Personal Health Record that enables a patient/patient advocate to manage his/her health record across organizations.

Physician Engagement Systems

Physician engagement systems enable the communication of patient information in a secure and private manner to patients’ referring or primary care physicians. This access may be through a physician or referring physician portal.

CLINICAL RESEARCH INFORMATION SYSTEMS

IV. CLINICAL RESEARCH INFRASTRUCTURE

Core EHR

691

692

40. USE OF EHR IN CLINICAL RESEARCH

FIGURE 40.1

The logical review of an electronic health record (EHR).

links to genomic data warehouses and allows the phenomic data to be visualized with genomic data.4 The CRDW provides data to systems to allow complex analysis requiring “interfaces for queries, honest broker services, data extraction, chart review, and IRB approval”5; pp. 113e114. The CTDMS manages the clinical trial from administrative, financial planning, participant recruitment, participant scheduling, data management and analysis, and adverse reporting perspectives. Data are provided to the CTDMS either through a feed from the CRDW, a request to the CRDW broker or a link to the CRDW. A CTDMS must be utilized for clinical trials with Food and Drug Administration (FDA)-reportable products. For natural history studies, a CRDW or a variant of a CTDMS is an alternative to a CTDMS used for

regulatory reporting. A review of CTDMS systems can be found in Chapter 31.6 Table 40.4 reviews system characteristics of an EHR and CTDMS or CTDMS variant.

USING AN ELECTRONIC HEALTH RECORD IN CLINICAL RESEARCH The primary role of the EHR is to support clinical care of patients. At each step of the clinical care process data are either collected or generated. To fully understand clinical research and how the EHR is utilized, it is important to review the clinical care and clinical research processes and how they map to each other. Table 40.5

IV. CLINICAL RESEARCH INFRASTRUCTURE

NIH/CC Systems ADT Only

eBot/FileBot

Orders to Ancillary, Results to SCM SCMProd

CRIS Datamart

CDR/ CDRnew

CLLAB

HSS

Common

DTI

USING AN ELECTRONIC HEALTH RECORD IN CLINICAL RESEARCH

IV. CLINICAL RESEARCH INFRASTRUCTURE

Documents to SCM

ADT, Orders to Ancillary, Results and Documents to SCM

FIGURE 40.2 A view of the physical systems of the electronic health record (EHR) used at the National Institutes of Health Clinical Center (NIH CC).

693

694 TABLE 40.4

40. USE OF EHR IN CLINICAL RESEARCH

System Characteristics of Electronic Health Record (EHR) Versus Clinical Research Information System (CRIS)

EHR System Characteristics

CRIS MS Variant System Characteristics

Data viewing at the patient level.

Data analysis across patients typically within a protocol or related protocols.

Majority of data has clinical relevance.

There may be multiple systems used to store data that has no clinical relevance, and it may be stored in a separate system.

Majority of data is unstructured.

Data is converted to coded data either through the use of terminology mechanisms or metaknowledge.

EHR is system of record for clinical care.

There may be multiple data sources, multiple organizations, multiple EHRs, multiple laboratory systems providing data to the clinical research data warehouse.

EHR represents data from one organization.

There may be multiple EHRs that use multiple ancillary systems across multiple organizations.

Data analysis is through reports and components within the EHR.

Data is analyzed in the CRIS separate from the EHR.

All data managed for individual patients in an identified format.

Data with limited data sets (deidentified data) may be part of CRIS.

reviews the processes in both clinical care and clinical research,7 with the addition of systems/tools utilized for each process.

Data Characteristics In clinical care, the focus is the care of the patient and understanding the status of the patient. All of the data collected or generated as part of the clinical care process can be utilized as part of clinical research. Some examples of data types from clinical care include orders, laboratory results, radiology reports, procedure reports, medication administrations, vital signs, progress notes, and structured notes. Certain data elements are collected as part of the clinical trial within the EHR (see Chapter 30).8 The focus for clinical research is the analysis of data across patient populations. Desired characteristics for clinical care research data analysis include structured, discrete, single, tests or events coded with one known

identified code, and the ability to compare data across data types (i.e., review changes of laboratory results in respect to medication administration). However, in clinical research, data is typically textual, unstructured data, and originates from multiple sources. Table 40.6 reviews other differences in desired data characteristics for clinical care and clinical research.9

Clinical Decision Support Within Electronic Health Record CDS provides rules-based management within orders, order sets, and clinical documentation templates to augment standardization, compliance, and enhance patient safety in clinical care. CDS enables EHR administrators to add logic within ordering and clinical documentation processes. It supports mechanisms that provide the right information, to the right person (physician, nurse, technician), in the right intervention format (alert, order set, providing relevant clinical information), in the right channel (user screen, email message, text to a mobile device), at the right time in the workflow to provide guidance to make the best healthcare decision.10e12 For clinical research, CDS can enhance compliance with research study requirements, ensure completion and accuracy of fields needed for research analysis, and ensure the integrity of the research study. Based on defined conditions of data entered into fields on order forms, system builders can use calculations for other fields and require ordering of other tests or support complex ordering such as adding pharmacogenetics/ pharmacogenomics information to the ordering process.10 The system also can restrict certain conditions such as not allowing the ordering of a specific medication13 while a patient is on a specific research protocol or require a specific test when a medication is ordered.10 CDS can be added to clinical documentation to ensure the completeness of the document by adding calculations, requiring fields, and alerting the user on conditions to address during documentation.14

Protocol Order Sets Within the Electronic Health Record As part of the Institutional Review Board (IRB) process, many organizations require the development of a map or research grid of the study as part of the protocol submission. The protocol map/research grid, at a minimum, typically shows the time points, tests, medication administrations, exams, and procedures. Variations of a protocol map/research grid can be used to identify items that are billable versus nonbillable or research

IV. CLINICAL RESEARCH INFRASTRUCTURE

TABLE 40.5

Review of Clinical Research and Clinical Care Processes and Associated Systems

Clinical Care Process

Clinical Research Information System(s)

Clinical Information System(s)

Protocol approval

Protocol order set creation tool Electronic health record (EHR)

Protocol management system

Admission

Patient recruitment

Patient registration system

Patient recruitment system Clinical trials data management system (CTDMS)

Admission

Initial informed consent

EHR

CTDMS

Initial assessment

Protocol consent Randomization

EHR

CTDMS

Multidisciplinary plan of care

Research plan

EHR

CTDMS

Treatment

Research intervention/data collection

EHR

Electronic data capture (EDC), Clinical research data warehouse (CRDW), CTDMS

Patient monitoring/evaluation

Endpoint data/data safety monitoring

EHR

EDC, CRDW, CTDMS

Plan for next level of care

Follow-up data collection plan

EHR

EDC, CRDW, CTDMS

Discharge/transfer

EHR

EDC, CRDW, CTDMS

Follow-up plan communicated

EHR Patient portal Personal health record Referring physician portal

EDC, CRDW, CTDMS

Data analysis External review Dissemination of results

USING AN ELECTRONIC HEALTH RECORD IN CLINICAL RESEARCH

IV. CLINICAL RESEARCH INFRASTRUCTURE

Clinical Research Process

CTDMS, CRDW EHR

CTDMS, CRDW Journals, conferences

695

696 TABLE 40.6

40. USE OF EHR IN CLINICAL RESEARCH

Desired Characteristics of Data in Electronic Health Record (EHR) for Clinical Care and Clinical Research9

Clinical Care Data Characteristics Focus on Clinical Care

Clinical Research Data Characteristics Focus on Research

Sample Protocol Map/Research Grid

Structured data: Unstructured data: 1. Allow prose or free-text entry 1. Discrete values to allow data analysis. of clinical documentation. 2. Store PDF or images of reports 2. Provide metaknowledge and/ to ensure data is within EHR. or summaries of all data where 3. Allow nonrequired fields. source is PDF or scanned 4. Allow suggested dictionaries. image. 3. Present data for required fields with limited null values across fields and result values. 4. Use of restricted dictionaries. Focus on standard of care/billing. Rules enforced on order entry for frequency of a given test (i.e., complete blood count (CBC) with Differential ordered every day vs. the standard being every 3 days).

Frequency of tests is in relation to analysis supporting the research question (i.e., CBC with Differential ordered daily based on protocol).

All tests and medications have Potential for masked medication names that are presented to users orders to allow various arms of a of the EHR. research protocol, one which may be a placebo, and the need to prevent the research team from knowing which subject is on each arm of the study. Results for all ordered tests.

Potential for masked results or results not shown in the EHR to prevent the research team from knowing which subject is on each arm of the research protocol. Potential to collect specimens for tests that are not Clinical Laboratory Improvement Amendments certified and do not show clinical relevance and, thus, are not part of the EHR.

Order sets exist for standard of care for multiple diagnoses but not all.

Research protocol order sets desired for all research studies.

Required coding for billing purposes.

Data coded for research and analysis purposes.

Test CT Adrenals DX Spine-Entire for Vertebral Count CBC PTT PT Cortisol, 24 h Urine Aspirin – 325 mg tablet

Baseline Evaluation (M0) X

versus standard of care. Such variations can be used to estimate the cost of the research study as well as support the billing process.

This variation shows the name of the tests and medications as rows and time points as columns. An X indicates when a specific event is to occur. Complex protocols may have multiple arms as well as branches that need to be represented in the protocol map/ research grid (Fig. 40.3). From the protocol, protocol map, and the research grid, organizations build research protocol order sets within the EHR. Research protocol order sets are a form of CDS that enable clinicians to define events including tests, exams and procedures; time points for vital signs collection and medication ordering; and to group items together by logical grouping and label the grouping. The research protocol order set can preserve the integrity of not only the data points of tests and medication administrations but also data fields. The research protocol order set can be configured to specify the time and order of tests, exams, and medication administrations, as well as default values for fields on the order forms such as dosage, frequency, administration route, and special instructions for medication orders; tube type, special instructions, and specimen type for laboratory tests and microbiology tests; and contrast requirements and special instructions for radiology exams. Research protocol order sets also facilitate standardization of the ordering of tests, exams, and procedures to ensure integrity of the study. Requiring all ordering through research order sets, though desired, is not always possible. At the NIH Clinical Center, the goal is that all research protocols have at least one order set with 90% of all ordering via order sets. The organization is currently in the 75% range with 25% of items being ordered as reorders or ad-hoc orders (Figs. 40.4 and 40.5). This grid format can be used for standard of care or research protocols.

3 Monthly Evaluation (M3)

3 Monthly Evaluation (M6, M9) X

X X X X X X

X X X

6 Monthly Evaluation (M12, M18) X

X X X X X

X X X

FIGURE 40.3 Example of a simple protocol map/research grid.

IV. CLINICAL RESEARCH INFRASTRUCTURE

USING AN ELECTRONIC HEALTH RECORD IN CLINICAL RESEARCH

697

FIGURE 40.4 Example of a simplistic, more linear-style order set from the Allscripts Sunrise Clinical Manager electronic health record (EHR).

FIGURE 40.5 Example of an order set configured in the Allscripts Sunrise Clinical Manager electronic health record used for pharmacogenetics.

IV. CLINICAL RESEARCH INFRASTRUCTURE

698

40. USE OF EHR IN CLINICAL RESEARCH

SECONDARY USE OF THE ELECTRONIC HEALTH RECORD FOR CLINICAL RESEARCH Most organizations are not focused on clinical research. In organizations where clinical research is not the focal point, the EHR is used primarily for billing and clinical care. The frequency of data points for collecting results and documenting vital signs, completeness of the data, and the quality of the data as well as the data characteristics identified in Table 40.5 make it difficult to reuse data collected in a nonresearch setting for clinical research.4,9 By focusing on the patient and billing, an EHR not specifically used for research is not inherently structured to meet the needs of data capture for a research protocol.4 A research protocol is structured based on research participants; assignment of research participants to study arms; a detailed schedule of treatments, procedures, devices, blood products, and medications to administer; and tests to use to analyze the outcome of the research study.4 In an EHR not specifically designed for research the data may be incomplete, transformed in a way which does not preserve the raw data, and may not include all the data needed or data in the necessary format to perform analysis. Reusing data from EHRs not specific to a research setting requires the analysis of the problem being studied, the patient care process in using the EHR and collecting data and the quality of the data.15 Processes, tools, and analysis can be used to assess the data, evaluate their validity and perform the migration of the data to a CRDW.15 The CRDW then interfaces to a CTDMS or the CTDMS links to the CRDW. Clinical informatics literature reviews different architectures and uses of EHRs within clinical research.5,6,15e18

LEGISLATION AND THE ELECTRONIC HEALTH RECORD Legislation has led to a more complete EHR, allowing facilities to consider the use of the EHR for clinical research. The Health Insurance Portability and Accountability Act of 1996 (HIPAA) is reviewed elsewhere (see Chapter 12)19 and 21 Code of Federal Regulations (CFR): Food and Drug Administration (FDA) Regulations for Electronic Records (see Chapter 30).8 In this section, we will review the Health Information Technology for Economic and Clinical Health Act (HITECH Act),20 CFR 45 Part 170 Health Information Technology Standards, Implementation Specifications, and Certification Criteria and Certification Programs for Health Information Technology,21e23 the Medicare Access and Children’s Health Insurance Program (CHIP)

Reauthorization Act of 2015 (MACRA),24,25 as well as the FDA Guidance for Industry Computerized Systems Used in Clinical Investigations that was provided for comment as Draft.26,27

Health Information Technology for Economic and Clinical Health Act The HITECH Act was enacted under Title XIII of the American Recovery and Reinvestment Act of 2009.20 With a goal of improving health system performance, the HITECH Act program focuses on attaining meaningful use of EHRs and in the process defined the meaningful use of EHRs, thus defining specific criteria for an EHR. In conjunction with the HITECH Act, the United States Department of Health and Human Services released 45 CFR Part 170 on June 18, 2010 to establish a certification program for Health Information Technology.21 As part of 45 CFR Part 170, the Centers for Medicare and Medicaid Services (CMS) and the Office of the National Coordinator for Health Information Technology (ONC) established standards and meaningful use criteria for EHRs to be a part of the EHR incentive program. These rules identify the criteria for a certified EHR and the meaningful use of EHRs. Table 40.7 reviews the meaningful use stages, with the third stage focused on improved outcomes, which includes use for population health and clinical research. Table 40.8 reviews an example of meaningful use criteria showing movement to focus on the use of EHRs, the desire to have a complete picture of each patient’s care, provide the data to the patients, and share data across organizations.23 These criteria enable a more complete EHR and provide the ability to collect data across EHRs, which in turn enables the ability to use data from the EHR for clinical research.

Medicare Access and Children’s Health Insurance Program Reauthorization Act of 2015 MACRA will replace the incentive program provided by HITECH through the addition a new Medicare payment model for clinicians that will start in 2019.24,25 While still in the comment period, this legislation considers quality, cost, and clinical practice improvement activities in calculating how Medicare physician payments are determined. While MACRA still requires that clinicians be measured on their meaningful use of certified EHR technology for purposes of determining Medicare payments, it also provides opportunities to transition the current incentive program even further. MACRA supports the need for a strong interoperable EHR system with the goal of providing optimal care for patients.24,25

IV. CLINICAL RESEARCH INFRASTRUCTURE

699

SUMMARY QUESTIONS

TABLE 40.7 Review of Meaningful Use Stages Provided by Office of the National Coordinator for Health Information Technology22 Data Capture and Sharing 2011e2012 Stage 1: Meaningful Use Criteria Focus On:

Advance Clinical Processes 2014 Stage 2: Meaningful Use Criteria Focus On:

Improved Outcomes 2016 Stage 3: Meaningful Use Criteria Focus On:

Using that information to track key clinical conditions

Increased requirements for e-prescribing and incorporating lab results

Decision support for national highpriority conditions

Communicating that information for care coordination processes

Electronic transmission of patient care summaries across multiple settings

Patient access to self-management tools

Initiating the reporting of clinical quality measures and public health information

More patientcontrolled data

Access to comprehensive patient data through patientcentered health information exchange

Using information to engage patients and their families in their care

Improving population health

U.S. Food and Drug Administration Guidance for Electronic Health Record in Clinical Research Though the FDA does not regulate clinical trials, they have provided guidance in multiple documents on the use of EHRs, namely the 2013 document titled Clinical Research Guidance for Industry Computerized Systems Used in Clinical Investigations and the document titled Draft: Use of Electronic Health Record Data in Clinical Investigations Guidance for Industry.26,27 The recommendations center on the use of an EHR and the electronic integration to an EDC or CRDW. By using the EHR as the source system, researchers are able to ensure that the data is attributable, legible, contemporaneous, original, and accurate (ALCOA).27 There is a strong recommendation to use an ONCapproved EHR when possible, but there is also a need to document how the EHR will be used for research regardless of it being ONC-approved. Additional recommendations relate to the use of audit trails to track the addition of the data to the EHR and all subsequent

modifications; informed consent by the patient to understand who and how the data will be used; and use of privacy and security controls.26,27

SUMMARY The EMR and EHR can be used in clinical research environments to support clinical research processes. In environments that are solely focused on research, the EHR is configured to support clinical research through the use of specific form design, documentation, clinical decision support, and research protocol order sets. Federal laws and documents that support the use of EHRs have provided opportunities for consideration to reuse data for clinical research. When reusing data from EHRs collected specifically for clinical care, organizations need to assess the processes used to collect the data, as well as the completeness of the data, and evaluate their use in respect to the research being performed.

SUMMARY QUESTIONS 1. The difference between an EMR and an EHR is: a. EMR includes patient engagement modules. b. The EHR extends the EMR by providing access to individuals, based on their roles, outside of the organization. c. EHR is limited to one organization. d. Only the EHR links to ancillary systems. 2. A hybrid EHR architecture can be best described as: a. All modules are from the same vendor. b. All modules are from different vendors. c. Modules follow best of breed and can include products from the vendor and other vendors. d. Core modules are from the EHR vendor. 3. The following is a true statement in relationship to clinical processes and how they map to clinical research processes: a. There is a mapping between processes such as admission to patient recruitment; admission to initial informed consent; initial assessment to protocol consent and randomization, and treatment to research intervention/data collection. b. There is a mapping between processes such as multidisciplinary plan of care to the research plan, treatment to research intervention/data collection, treatment plan to data analysis. c. All processes within clinical care map to a clinical research process. d. There are no processes that relate between clinical care and clinical research.

IV. CLINICAL RESEARCH INFRASTRUCTURE

700

TABLE 40.8

Review of Selected Meaningful Use Metrics Stage 223

Measure Title

Measure Description

Computerized provider order entry (CPOE)

Use CPOE for medication, laboratory, and radiology orders directly entered by any licensed healthcare professional who can enter orders into the medical record per state, local, and professional guidelines. electronic health record (EHR) reporting period are recorded using CPOE

Clinical lab-test results

Incorporate clinical lab-test results into Certified EHR Technology as structured data More than 55% of all clinical lab tests results ordered by the EP during the EHR reporting period whose results are either in a positive/negative or numerical format are incorporated in Certified EHR Technology as structured data

Summary Of care

The EP who transitions his or her patient to another setting of care or provider of care or refers his or her patient to another provider of care should provide summary care record for each transition of care or referral. There are three measures. Measure 3 is below: An EP must satisfy one of the following criteria: • Conducts one or more successful electronic exchanges of a summary of care document, as part of which is counted in “measure 2” (for EPs the measure atx495.6(j) (14) (ii) (B) with a recipient who has EHR technology that was developed designed by a different EHR technology developer than the sender’s EHR technology certified to 45 CFR 170.314(b) (2). • Conducts one or more successful tests with the CMS designated test EHR during the EHR reporting period.

Patient ability to electronically view, download, and transmit (VDT) health information

Provide patients the ability to view online, download, and transmit their health information within four business days of the information being available to the EP. Measure 1dMore than 50% of all unique patients seen by the EP during the EHR reporting period are provided timely (available to the patient within 4 business days after the information is available to the EP) online access to their health information. Measure 2dMore than 5% of all unique patients seen by the EP during the EHR reporting period (or their authorized representatives) view, download, or transmit their health information to a third party.

40. USE OF EHR IN CLINICAL RESEARCH

IV. CLINICAL RESEARCH INFRASTRUCTURE

More than 60% of medication, 30% of laboratory, and 30% of radiology orders created by the eligible professionals (EPs) during the

REFERENCES

4. Select the statement that describes data within an EHR for use within clinical research: a. The quality of data from an EHR does not need to be assessed before feeding a CRDW or a CTDMS. b. Data from an EHR cannot be reused for clinical research. c. Data within an EHR is structured, discrete, and has no need for additional coding to be used within an EHR. d. In an EHR not specifically designed for research the data may be incomplete, transformed in a way which does not preserve the raw data, and may not include all the data needed or data in the necessary format to perform analysis. 5. Federal legislation recognizes the importance for the use of the EHR in clinical research. a. True b. False

TERMS Ancillary systems Systems separate from the EHR used to support workflow within a clinical department. Examples are Laboratory Information Management Systems (LIMS/LIS), Transfusion Medicine Systems, Picture Archiving and Communication System, Radiology Information System, Surgery Information System, Outpatient Pharmacy System, Medication Management System, Nutrition Systems, and Health Information Management Systems. Clinical decision support (CDS) Mechanisms that provide the right information, to the right person (physician, nurse, technician), in the right intervention format (alert, order set, providing relevant clinical information), in the right channel (user screen, email message, text to a mobile device), at the right time in the workflow to provide guidance to make the best healthcare decision.10e12 Clinical research data repository/warehouse (CRDW) System used to manage data from multiple clinical information systems and multiple clinical research system. Clinical research information system (CRIS) Generic term for systems used to manage the research protocol from study participation recruitment and accrual, study participant randomization, data collection, data analysis, and data reporting. Systems include electronic data capture (EDC), clinical research data warehouse (CRDW), and clinical trial data management system. Computerized provider order entry (CPOE) Physician has sometimes been interchanged with the provider. CPOE is the practice of a licensed healthcare professional entering orders into the EMR or EHR. Electronic medication administration record (eMAR) The electronic recording of the administration of a medication to a patient in the patient’s record within the EHR/EMR. Data elements stored are medication, dose, route, frequency, dosage, person who administered the medication, witness to the administration when required. Electronic medical record (EMR) and electronic health record (EHR) The electronic medical record supports the clinical care process within a medical organization allowing the storage, retrieval, and use of clinical data. An EMR is typically only used within an organization and only for clinical care. The EHR is an extension of the EMR outside the organization, such as other medical facilities, to support continuation of care and for others purposes such as clinical research and population health. Health information exchange (HIE) The ability to provide access to clinical data to other organizations through a regional system. It

701

enables multiple medical facilities using EHRs from different vendors to access a central system that can be used by all organizations that participate in the HIE. Health Level 7 (HL7) The coding standard used to transport information between clinical information systems. Interface engine A system used to transport data between systems typically the EMR/EHR and ancillary systems. Typically the data is transferred using HL7. Pharmocogenomics/pharmacogenetics CDS Incorporates genetic information to provide information to guide the efficacy, dosage, or toxicity of the medication to the ordering process within an EHR.10 Protocol map/research grid A CDS mechanism to facilitate the inclusion of multiple orders on a protocol study participant. A research protocol order set allows tests, exams, procedures to be grouped, and fields on the order forms defaulted to allow standard and efficient ordering.

References 1. Garrett P, Seldman J. EMR vs EHR e what is the difference? U.S. Department of Health and Human Services: Office of the National Coordinator for health information technology (ONC). 2011. Retrieved from: http://www.healthit.gov/buzz-blog/electronic-health-andmedical-records/emr-vs-ehr-difference/. 2. McMullen PC, Howie WO, Philipsen N, Bryant VC, Setlow PD, Calhoun M, Green ZD. Electronic medical records and electronic health records: overview for nurse practitioners. J Nurse Pract 2014;10(9):660e5. http://dx.doi.org/10.1016/j.nurpra.2014.07.013. 3. Garnett C. Aged pioneer to retireegradually: new clinical research information system planned to replace MIS. NIH Rec 2001;LIII(21). 4. Hersh W, Weiner M, Embi P, Logan J, Payne P, Bernstam E, et al. Caveats for the use of operational electronic health record data in comparative effectiveness research. Med Care 2013;51(8 Suppl. 3): S30e7. http://dx.doi.org/10.1097/MLR.0b013e31829b1dbd. PMID: 23774517. 5. Shin S, Kim W, Lee J. Characteristics desired in clinical data warehouse for biomedical research. Healthc Inform Res 2014;20(2): 109e16. http://dx.doi.org/10.4258/hir.2014.20.2.109. 6. Cimino J. Chapter 31: Clinical research data: characteristics, representation, storage and retrieval. In: Gallin J, Ognibene F, Johnson L, editors. Principles and practice of clinical research. 4th ed. London: Academic Press; 2013. 7. Wallen G, Fisher C. Chapter 5: Clinical research and clinical processes. In: Gallin J, Ognibene F, Johnson L, editors. Principles and practice of clinical research. 4th ed. London: Academic Press; 2016. 8. St Germain D, Good M. Chapter 30: data management in clinical trials. In: Gallin J, Ognibene F, Johnson L, editors. Principles and practice of clinical research. 4th ed. London: Academic Press; 2014. 9. Terry A, Chevendra V, Thind A, Stewart M, Marshall J, Cejic S. Using your electronic medical record for research: a primer for avoiding pitfalls. Fam Pract 2010;27(1):121e6. http://dx.doi.org/ 10.1093/fampra/cmp068. 10. Goldspiel B, Flegel W, DiPatrizio GT, Penzak S, Biesecker LT, Patel J, McKeeby J. Integrating pharmacogenetic information and clinical decision support into the electronic health record. J Am Med Inform Assoc (Jamia) 2013:522e8. http://dx.doi.org/10.1136/ amiajnl-2013-001873. 11. Osheroff J. Approaching clinical decision support in medication management. In: Improving medication use and outcomes with clinical decision support: a step-by-step guide. HIMSS clinical decision support guidebook series. Chicago, IL: Healthcare Information and Management Systems Society Mission; 2009. p. 1e10. 12. Osheroff J, Teich J, Middleton B, Steen E, Wright A, Detmer D. A roadmap for national action on clinical decision support. J Am Med Inform Assoc (Jamia) 2007;14(2):141e5.

IV. CLINICAL RESEARCH INFRASTRUCTURE

702

40. USE OF EHR IN CLINICAL RESEARCH

13. Cimino J, Farnum L, DiPatrizio G, Goldspiel B. Improving adherence to research protocol drug exclusions using a clinical alerting system. AMIA Annu Symp Proc 2011:257e66. 14. Haerian K, McKeeby J, DiPatrizio G, Cimino J. Use of clinical alerting to improve the collection of clinical research data. In: AMIA Fall Symposium. San Francisco, CA; 2009. 15. Hersh W, Cimino J, Payne P, Embi P, Logan J, Weiner M, Saltz J. Recommendations for the use of operational electronic health record data in comparative effectiveness research. Generating Evid Methods Improve Patient Outcomes (EGEMS) 2013;1(1). http:// dx.doi.org/10.13063/2327-9214.1018. PMID: 25848563. 16. Kahn M, Weng C. Clinical research informatics: a conceptual perspective. J Am Med Inform Association JAMIA 2012;19(e1): e36e42. http://dx.doi.org/10.1136/amiajnl-2012-000968. 17. Vawdrey D, Weng C, Herion D, Cimino J. Enhancing electronic health records to support clinical research. AMIA Summits Translational Sci Proc 2014:102e8. 18. Cole A, Stephens K, Keppel G, Estiri H, Baldwin L. Extracting electronic health record data in a practice-based research network: processes to support translational research across diverse practice organization. Generating Evid Methods Improve Patient Outcomes (EGEMS) 2016;4(2). http://dx.doi.org/ 10.13063/2327-9214. 19. Bonham V. Chapter 12: Legal issues in clinical research. In: Gallin J, Ognibene F, Johnson L, editors. Principles and practice of clinical research. 4th ed. London: Academic Press; 2017. 20. One Hundred Eleventh U.S. Congress, United States of America. Health information technology for Economic and clinical health Act (HITECH Act), Title XIII of the American Recovery and Reinvestment Act of 2009. Washington, DC. 2009. Retrieved from: https://www. healthit.gov/sites/default/files/hitech_act_excerpt_from_arra_ with_index.pdf. 21. U.S. Department of Health, Human Services, United States of America. Code of federal regulation 45 Part 170 health information technology standards, Implementation Specifications, and certification criteria and certification programs for health information technology. Federal Register, vol. 75 (8). Washington, DC: United States Government Publishing; 2010. Retrieved from: https://www.gpo.gov/ fdsys/pkg/FR-2010-01-13/pdf/E9-31216.pdf.

22. U.S. Department of Health, Human Services Office of the National Coordinator for Health Information Technology, U.S. Department of Health and Human Services. EHR incentives & certification: how to attain meaningful use. 2013. Retrieved from: https://www. healthit.gov/providers-professionals/step-5-achieve-meaningfuluse-stage-2. 23. U.S. Department of Health, Human Services Office of the National Coordinator for Health Information Technology, U.S. Department of health and Human services. EHR incentives & certification: step 5 Achieve meaningful use. 2013. Retrieved from: https://www.healthit.gov/providers-professionals/how-attainmeaningful-use. 24. U.S. Department of Health, Human Services, Centers for Medicare & Medicaid Services. CMS quality measure development plan: Supporting the transition to the Merit-based incentive payment system (MIPS) and alternative payment models (APMs). 2015. Baltimore, MD. Retrieved from: https://www.cms.gov/Medicare/Quality-Initiatives-PatientAssessment-Instruments/Value-Based-Programs/MACRA-MIPSand-APMs/MACRA-MIPS-and-APMs.html. 25. U.S. Department of Health, Human Services, Office of the National Coordinator for Health Information Technology. EHR incentive programs: where We Go Next. 2016. Retrieved from: https://www. healthit.gov/buzz-blog/meaningful-use/8791/. 26. U.S. Department of Health, Human Services, Food and Drug Administration (FDA),Center for Drug Evaluation and Research (CDER), Center for Biologics Evaluation and Research (CBER) and Center for Devices and Radiological Health (CDRH). Draft: use of electronic health record data in clinical Investigations guidance for Industry. 2016. 27. U.S. Department of Health, Human Services, Food and Drug Administration (FDA), Office of the Commissioner (OC). Guidance for industry computerized systems used in clinical Investigations. 2007.

Further Reading 1. Chapter 4: Institutional review boards. In: Gallin J, Ognibene F, Johnson L, editors. Principles and practice of clinical research. 4th ed. London: Academic Press; 2006.

IV. CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

41 The Clinical Researcher and the Media John T. Burklow National Institutes of Health, Bethesda, MD, United States

O U T L I N E What Makes News in Science and Medicine?

704

The Interview

708

Published SciencedThe Media’s Bread and Butter

704

What if You Are Misquoted?

710

Novelty

704

What the Public Does Not Know About Science

710

The Unexpected

705

Unexpected Questions

710

Celebrity

705

When the News Is Not Good

Controversy

705

710

Impact

706

A Word About Investigative Reporters

710

Why Talk to Reporters?

706

The Freedom of Information Act

711

Why Reporters Want to Talk to You

706

Why You Should Talk to Reporters

706

Embargoes The Ingelfinger Rule

711 711

Social Media: What to Keep in Mind

707

When to Contact Your Communications Office

712

Engaging the MediadThe Process

707

Conclusion

712

A Word About Email, the Web, and Social Media

707

Summary Questions

712

The chapter discusses how National Institutes of Health (NIH) scientists can deal effectively with the media. It provides an overview of media relations, what drives science/health news, reasons to speak with reporters, and common pitfalls of media interviews. The author describes specific techniques that can be used to ensure that an interview goes well and the scientist gets his or her message across well, which helps to improve the chances of a fair, accurate story. Also discussed is the rapidly changing landscape of social media and its impact on news coverage and dissemination. If you are a scientist conducting clinical research, chances are that you will interact with the media at some point in your career. Dealing effectively with Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00041-1

reporters is like any other skilldto succeed, you need to learn the basics, be receptive to counsel; and practice. And like any other skill, it may look easy; however, actually doing it can be another matter. This chapter addresses the changes in media relations due to the explosion of social media. It also reviews what makes news and why, discusses how to handle media inquiries, and offers tips on how to help raise the odds that your message will come across clearly. Your goal is to be accurate, empathetic, and effective, while avoiding some common pitfalls. Health and medicine, covered daily by the media, are two of the most popular topics among the public. Consequently, reporters are searching constantly for new stories and new angles. Whether it is good news or

703

Copyright © 2018. Published by Elsevier Inc.

704

41. THE CLINICAL RESEARCHER AND THE MEDIA

bad news in medical research, the media want your story. If your research shows results that could lead to a promising treatment, people want to know about it. The more impact a disease has on society, the more the public wants to know. And, they might want to know before you have all the answers to provide them. Conversely, if something bad happens during your research, the media will soon be on your doorstep. For example, if a patient dies because of an adverse reaction to an investigational therapy, or a patient suffers because a research protocol did not comply with regulations or guidelines, or an investigative reporter gets wind of an allegation of a conflict of interest, you can count on a calldor many callsdfrom the media. This applies not only if the situation directly involves you. You may get calls just because you work in that particular field of research. Also, there is a good chance that you will be called even when an eventdpositive or negatived happens at another research institution. Reporters will want to get your reaction or perspective. If it was an NIH-sponsored clinical trial, reporters assume NIH staff may have been involved in some way. As a clinical investigator using public funds, such as from the NIH, your work is supported by public dollars. The public has a right to know the good news and the bad. That means that you have to be ready to deal with a full range of possible media inquiries. The media disseminate information that often brings about improvements in clinical research and patient protections. Many of the changes in the rules and regulations that govern modern clinical practice have occurred as a result of missteps and abuses in history. Beginning in 1932 with Tuskegee, to 1999 and the tragic death of a young man who participated in a gene therapy clinical trial, to the revelation in 2010 of unethical studies in Guatemala in the 1940s, the media have played an important role in shaping society’s perception of the ethics of clinical research and the necessity of continual improvement. It is truedyour clinical research could end up on the front page of a major newspaper, or on the evening news, or go viral on social media. A good example is the outbreak of the Ebola virus epidemic in the summer and fall of 2014. Researchers at the NIH and NIHsupported institutions around the country went to work immediately to develop an effective vaccine. The media coverage was extensivedprint, radio, television, and Web. Dr. Anthony Fauci, director of NIH’s National Institute of Allergy and Infectious Diseases, did countless interviews and appeared numerous times on television, radio, and in newspaper articles. In a matter of weeks, clinical trials began to test candidate vaccines. By fall, the NIH Clinical Center admitted several patients who had been exposed to the virus, including a nurse from Texas who contracted the virus while treating an Ebola

patient. Her successful treatment and release from the NIH Clinical Center garnered nation-wide attention, including two major press conferences, front page headline news coverage, and a visit to the White House to meet the President. For NIH clinical research, it was an amazing story of a “turning-on-a-dime” research response to a public health crisis. Dr. Fauci understands the importance of media and made himself available to reporters, providing crisp, cogent answers to complex scientific questions in ways the public could understand. It was no wonder that all five Sunday talk shows invited him on their shows simultaneously, several times, to discuss the Ebola crisis.

WHAT MAKES NEWS IN SCIENCE AND MEDICINE? Most of the time, media coverage of your clinical trials is desirable, and you may wonder why a particular study attracts media attention while others do not. The following categories describe what draws reporters to cover science and medicine. Keep in mind, however, that large clinical studies will get more attention than basic laboratory findings and more attention than phase I and II clinical studies.

PUBLISHED SCIENCEdTHE MEDIA’S BREAD AND BUTTER Scientific studies and research advances that have been published in peer-reviewed journals get the most newsprint, radio, television air time, and Web coverage, by a huge margin. This constitutes the major source of news in science and medicine. By and large, medical research results reported in journals such as Science, Nature, New England Journal of Medicine, and the Journal of the American Medical Association help dictate what is covered by science and health reporters from week to week. Once you have a manuscript that has been accepted for publication by a major journal, begin thinking early about how you will handle media inquiries and what your “core” message should be, including a brief, topline message, with the public in mind as your main audience.

NOVELTY As with all news stories, the “unusual” in science and medicine gets the attention of the general public. Even people with little knowledge or interest in science want to hear about stem cell clinical trials, people with spinal cord injuries regaining movement, or new

IV. CLINICAL RESEARCH INFRASTRUCTURE

705

CONTROVERSY

therapies for weight loss. Since NIH studies hundreds of common and rare conditions and diseases, there is a reasonable chance that NIH-funded areas of research will generate media attention as hard news, or perhaps a human interest story.

THE UNEXPECTED In September 2015, NIH announced the results of a large, nation-wide clinical trial conducted with a group of adults 50 years and older with high blood pressure. Results indicated that more intensive management of high blood pressure below a commonly recommended blood pressure target significantly reduces rates of cardiovascular disease, and lowers risk of death. The landmark clinical trial called the Systolic Blood Pressure Intervention Trial, found that achieving a target systolic pressure of 120 mm of mercury reduced rates of cardiovascular events such as heart attack and heart failure, as well as stroke, by almost a third, and the risk of death by almost a quarter as compared to the target systolic pressure of 140 mm of mercury. In fact, the results were so evident that the Data and Safety Monitoring Board (see Chapter 10) stopped the trial early, even before the data were published in a peer-reviewed journal. NIH went public with the results immediately, and some in the field clamored for the full set of data, which was not yet available. There was heavy media coverage of this news, and heavy traffic on social media, particularly via Twitter, from patients, health care providers, and researchers. This is a good example of the tension between getting important health information out to the public as quickly as possible, and having a scientific article available. Another example is when the National Cancer Institute announced in October 2010 that it was discontinuing a lung cancer detection trial because the results were overwhelmingly positiveda 20% reduction in lung cancer mortality among those receiving a low-dose helical computed tomography (CT) scan. Since lung cancer is the leading cause of cancer death and is difficult to treat, there was considerable debate prior to the study about the efficacy of using CT for lowering lung cancer mortality; this generated a great deal of news coverage. When the public hears unexpected news, sometimes it is when a therapy previously considered safe poses risks. It is likely that the public reaction will be negative, and a strong demand for clear guidance. Most people will want to learn how to assess their own risk, and about safe alternatives. It will not be sufficient to announce research results and merely recommend that the public talk to their health care provider. With immediate access to health information through the internet, many patients have the same information about a certain topic as does their provider. Conversely, they

could have very different information than that of their provider, and if there is an information vacuum, both the patient and provider are frustrated. Sometimes there is no safe alternative. In that case, you need to convey through the media that you understand the difficult situation and that you feel compassion for patients and the public. You also will want to communicate that researchers will continue to seek answers.

CELEBRITY Think about the impact of Angelina Jolie’s public announcement that she had a preventive double mastectomy because she tested positive for the BRCA-1 gene mutation. It was a major news event. When Michael J. Fox announced that he had been diagnosed with Parkinson’s disease, it raised public awareness, interest, and concern greatly. In 1995, Christopher Reeve’s paralysis focused national attention on spinal cord injury and research, and later on the debate over human embryonic stem cell (hESC) research. The public wants to know about celebrities and their medical problems, and their stories often raise national public discussion about their disease or condition. What if you are called on to treat a celebrity in your clinic, or a celebrity has developed a disease that happens to be in your area of expertise? You may be called by a member of the media to comment on the celebrity’s condition. It is a general policy that clinicians do not comment on their patients, or even acknowledge that an individual is under their care. Furthermore, it is best not to comment to the media about a certain disease, such as the acquired immune deficiency syndrome or stroke, if they are asking because a famous person has it. One’s academic institution, however, may provide general information about research being conducted in that particular area. If you comment, even if only about the disease, there is a good chance that you will be perceived as commenting on the celebrity, and you may well find yourself woven into the story.

CONTROVERSY For controversial biomedical research topics, there is often interplay among NIH, Congress, the Administration, and the media. A prime example was the debate over federal funding of hESC research. Between the years 2000 and 2010, no other science and medicine story dominated the news like stem cell research. This involved researchers at NIH and from institutions around the world. Since it was such a politically and emotionally charged issue, it required a great deal of judgment, tact, and preparation on the part of any researcher who spoke

IV. CLINICAL RESEARCH INFRASTRUCTURE

706

41. THE CLINICAL RESEARCHER AND THE MEDIA

to the press. When talking about an issue that people feel strongly about, it is especially important to keep in mind the various audiences who are reading or listening to your words. What you may consider innocuous, or a straightforward fact, others may find explosive, or consider it part of a larger ideological agenda.

IMPACT Research that has a significant and immediate impact on people gets a great deal of media coverage. For example, findings in the late 1990s that tamoxifen reduced the risk of developing breast cancer by 50% drew more than 100 reporters to a press conference that took place in the Health and Human Services building in Washington, DC. The hormone therapy results in 2002 had direct implications for more than 14 million women and extensive media coverage reflected this impact. More recently, any news or research findings regarding the Zika virus generates tremendous media and public interest. Rare diseases, although they do not affect large numbers of people, can have a devastating impact on people and their families and also may draw media attention.

WHY TALK TO REPORTERS? Although some scientists would not mind if they went through their entire career without speaking to a reporter, there are benefits to interacting with the press. The media help the public understand the importance of research, that research is moving forward, and that it is essential making to making advances in human health. No other topic, except perhaps sports or a natural or man-made disaster, generates this kind of automatic news interest. The NIH conducts and supports a great deal of the clinical research in this country, and it is part of NIH’s mission to let people know about medical research progress. The majority of the public gets much of their health and medical information from the media; therefore, it is a logical extension that scientists be adept in dealing with reporters and work at becoming effective public communicators.

example, in reports on the Precision Medicine Initiative, announced by President Obama in January 2015 (now called the All of Us Research Program), reporters always try to include a quote from the NIH Director, or another well-known NIH researcher. If they are covering a clinical trial result, reporters will want to talk to the lead scientistsdand their patients. NIH and most institutions go to great lengths to protect patient privacy and that includes shielding them from the media. Sometimes patients want to talk to reporters and that is their prerogative. Their clinician, however, must ensure that the patients do not feel obligated or that there is an implicit expectation that they speak to the press. • Clarity and lively flavor. Quotes are what make a news story different from an editorial or an essay. A good quote is usually more interesting than the same information written in a reporter’s words, and quotes often make the story easier to read. • Tension. Reporters want interviews because often they find hints of controversy in what you say, and controversy heightens the interest of the public. This fact exemplifies the need for individuals to think through their words carefully, get advice when needed, and be well prepared before speaking to a reporter. If a reporter asks you to critique or contradict another scientist’s comments, it is likely best to defer and transit to your core message. • Limited time. News reporters are in a hurry. In contrast, scientists are typically deliberate and meticulous in their process. Sometimes these worlds clash. Scientists will complain that the reporter waited until the last minute to request an interview and did not allow enough preparation time. The reality is that the news cycle is short and quickdnow 24 h a day, 7 days a weekdand reporters often only have a few hours to put together a story. A quote saves a reporter the time it takes to uncover the facts from other sources. This does not mean that you should feel rushed into an interview; normally there is at least a small window of time between the reporter’s first call and her/his deadline. Allow yourself enough time to gather your thoughts and prepare adequately for the interview.

WHY YOU SHOULD TALK TO REPORTERS WHY REPORTERS WANT TO TALK TO YOU Reporters tenaciously seek out quotes from experts for many reasons. For example: • Credibility. Quotes from experts and the people directly involved make the story more credible. For

Occasionally, a Principal Investigator does not want to talk to reporters. He or she would rather have the official institutional spokesperson talk to reporters about the research, or send a written statement through email. This is almost never satisfying to a reporter (especially from radio or television) because an agency spokesperson is usually not the subjectematter expert, and

IV. CLINICAL RESEARCH INFRASTRUCTURE

A WORD ABOUT EMAIL, THE WEB, AND SOCIAL MEDIA

not viewed to be as credible as the scientists involved in a study. You, the clinical researcher, are the expert. Institutional spokespeople do speak for their institutions when a brief, factual answer is needed quickly and a scientific expert is not necessary. They also can help you prepare to talk with reporters. At the NIH we believe firmly that you should talk to reporters, unless you feel that you are not the appropriate person to be interviewed, or you have evidence that the reporter is not operating in good faith. There are clear benefits to speaking to reporters, and they usually outweigh the risks of being misquoted or your comments being taken out of context. • You can improve the accuracy of the story. Many science reporters are knowledgeable and very experienced, but they do not know everything about every scientific or medical subject. Even the best science reporters need guidance on emphasis, nuance, or help in understanding methodology. • You help create a favorable climate for your institution. Your input improves public understanding of the importance of medical research and its relevance to people’s everyday lives. In turn, it helps maintain public support for your institution and medical research in general. • You owe it to the American taxpayers. Since NIH is supported by the taxpayers, NIH-funded researchers owe it to them to explain their work. The best way to do this is through the mass media. Clinical investigators especially need to help people understand what an advance in medical research could mean or not mean to their lives. Your participation in the story provides an appreciated context for the American people.

SOCIAL MEDIA: WHAT TO KEEP IN MIND In recent years, social media have become a standard part of media and the public communication landscape. How does this affect the way you interact with reporters or manage a news story? First, social media make everything instantaneousdeven more than email did back in the day. As soon as you make an announcement, your words fly virtually around the globedon Twitter, Facebook, etc. If a reporter quotes you from an interview, or as you present in a public forum, they may tweet about it immediatelydwith photosdand likely will not wait for their story to run in the print version the next day, or on the evening news. Just keep in mind that your “rollout” time frame is compressed dramatically. In other words, if you are trying to stage your announcement, keep in mind that news travels fast and in an increasingly expansive way. If you need to

707

give anyone a heads up before your public announcement, you may want to do it all within 24 h, and use “old-fashioned” methods, such as telephone calls and meetings. Social media, in addition to stepping up the pace of news, provide additional opportunities for people to voice their opinions, and it is important to monitor the reaction to your news in social media, as well as in the more traditional outlets, newspapers, radio, TV, etc. If there is an overwhelmingly positive response, this may indicate that your news “shelf life” could be longer than expected. Likewise, if the reaction is negative, your story may be in the news longer, and you may need to address the issues raised in social media. Before, however, engaging in a debate, on Twitter for example, please consult your communications office, whose staff have experience with social media. They can advise you on whether to engage and what to say to avoid common pitfalls or unintended consequences.

ENGAGING THE MEDIAdTHE PROCESS What if a reporter calls you directly, without going through your institution’s communications office? This does happen, and there are specific steps you can take to handle such a call. Even if the reporter is friendly and wants to do a positive story, you could find yourself in an unfortunate position if you forget to ask a few key questions, especially if you work for the federal government. It is best to get approval and advice from your communications office before you agree to an interview. This is especially important when contacted by major media outlets such as the Washington Post, New York Times, Wall Street Journal, any of the radio or television networks, or Web outlets. Also, your laboratory, NIH institute, or academic institution might have a rule requiring clearance, or approval. Make sure you know your local policy before you talk to the media.

A WORD ABOUT EMAIL, THE WEB, AND SOCIAL MEDIA Traditionally, the major media outlets generate the most attention. Although that is still the case, generally speaking, email, the Web, and social media have leveled the playing field considerably. It is often difficult for the casual reader to distinguish between a major news story and an obscure one. Therefore, keep in mind that your interview with a relatively unknown Web-based news outlet may reach millions. In addition, email enables broad redistribution of interviews with all types of

IV. CLINICAL RESEARCH INFRASTRUCTURE

708

41. THE CLINICAL RESEARCHER AND THE MEDIA

media, including industry and advocacy organization newsletters. Also, if you are in a newswire service story, such as the Associated Press (AP) or Reuters, your name will appear in Google News and other search engines each time the story appears, which could be hundreds of citations. Therefore, it is wise to be as careful about what you say to a newsletter reporter, as what you would say to the Washington Post or an AP reporter. As the number of print news outlets shrinks, and as the power of Facebook, Twitter, YouTube, blogging, and the number of new social media platforms grows, the world of media is constantly evolving. Increasingly, many people get their news from a social media source rather than from traditional sources such as newspapers, magazines, radio, and TV. The NIH, for example, now has more than 200 social media sites available to the public. These seismic changes in the media landscape have important implications for you, the clinical researcher, as you tell your story. For example, keep in mind when you speak or show slides that someone in the audience is tweeting or may be blogging about your presentation. They also may be video- or audiotaping it on their mobile phone. Your comments may appear around the world within seconds. If you plan to blog about your research, talk to your communications office about the pros and cons, and read your institution’s policy on social media. Often, individuals underestimate the amount of time and attention that blogs require. Also, for government employees, keep in mind that some social network platforms are not accessible on government-owned computers because of cybersecurity concerns. This may make it difficult to incorporate them into your standard outreach practices. If you get a call from a reporter requesting an interview, there are several things you should do. First, take the reins confidently, and ask a few questions of your own. You need the answers to the following questions from a reporter before you agree to do anythingd and be sure to write down the answers. 1. What is your name and phone number? 2. What publication/network/station are you with? 3. What is your deadline? (This gives you an idea of how much time you have to think about your answer). 4. What is the angle or story line? 5. Who else are you talking to? 6. What information are you looking for from me? (For example, do they just want a background discussion about T cells? Do they want to feature you or just get a quick quote on someone else’s work?) Alternatively, you can ask the press officer in your communications office to gather this information and report back to you. To allow yourself some time to gather your thoughts, you can say, “I would be happy to talk with you. Could you coordinate this with our

communications office? Here is who you should call, and this is the number.” You always should have the name and phone number of your press officer within reachdat your desk and in your wallet or pocketbook. Your next step should be to consult the communications office of your institution. Your press officer can help you with the following types of questions: • Are you the right person to talk to the reporter or is this a sensitive issue that should be handled by the communications office, the institution spokesperson, or someone else? • Should someone outside your institution field the questions? • Is there an institutional position on the subject in question? • What information about the reporter would be helpful to you, such as the line of questioning they might take? • What experience has your press officer had with this reporter, if any? • How do you decide to say “No” gracefully, if that is what you decide? If you decide not to do the interview, decline truthfully and firmly. Below are three common answers, however, the first one is not recommended: 1. I would like to talk to you, but I have been told not to. This is a fairly common, but inappropriate answer. It will only entice the reporter to pursue you and it will become its own story. 2. I am not the best person to talk to you about this. Why don’t you call_______? This is an appropriate answer if it is true. Be sure to notify the person you named, however, before making the referral. 3. It’s really too early in the research to have anything firm to say about it. Again, this is a good answer if it is true, although it probably will not satisfy the reporter’s questions. Perhaps give the reporter a projected date that he or she could call back for better information.

THE INTERVIEW If you decide to do the interview, keep in mind that reporters may be friendly, but they are not your “pals.” They are not necessarily cheerleaders for science or your point of view. They are in the business of reporting what they think is a news story. The overwhelming majority of science and health reporters, however, strive for accuracy and context, and are very knowledgeable. You can help them achieve their goals while you convey important health and science information to their readers, listeners, or viewers. If you have not had any experience speaking with reporters, it is a good idea to

IV. CLINICAL RESEARCH INFRASTRUCTURE

THE INTERVIEW

get some training. It could be given by your communications office or by a specialist who conducts media training for a living. Usually, the sessions last a half day, and include mock on-camera interviews, as well as role-playing. You will feel more confident, and the reporter will be pleased to get clear, concise answers. Even if you have media experience, it is helpful to take a brief training session to sharpen your skills. If you do not have time to get training before you are faced with a media interview, here are a few tips to keep in mind: 1. Although your training and orientation may lead you to provide detailed, comprehensive answers to questions, with many qualifiers and caveats, your answers to reporters should be concise, directed to the question asked, and void of any “filler.” You should prepare your core message in your mind, and on paper. On an index card, write the message, the exact words that you hope will appear in the story. This is the message that you will keep coming back todor “bridge to”din all of your answers. All of your answers should support your core message. Do not add many qualifiers or caveats; they will increase your chances of not getting your message across. When you are speaking, picture holding a tape measure being pulled out. As you talk, think of the inches as words. You hope the reporter will use every inch. Instead, the reporter uses 1e8 and 16e23 and it may be out of context. One of the most common complaints from scientists about news stories is that the story did not include what they told the reporter. Say only the words you would like to see in print, and keep it focused. Everything you say is potentially fair game to the reporter. An interview should not be a free-flowing discussion, however, it should be conversational and not just a series of talking points. 2. Your core message must encapsulate not only the facts, but also your perspective, the NIH perspective (if applicable), and a human dimension. It is important to take time to think about your message, write it, ask others to review it, and edit it. It should be brief, direct, compelling, and interesting. Always speak in plain language, avoid technical jargon, and never use acronyms, even if you think everyone knows what they mean. 3. Beware of a few terms when you are talking with a reporter. Terms of the journalistic trade may not mean what you think they mean. Keep these definitions in mind, in case you hear the reporter use them: a. On the record. This means a reporter can quote you directly, using your name and title. b. Not for attribution and on background. This means that the information you give, including direct quotes, can be used by the reporter, but you are not to be named. You may be identified as an

709

NIH official, an institutional official, or source. It is rare to have science and medical sources speak under a condition of anonymity. We recommend that you stay on the record at all times. c. Off the record. This means the reporter cannot use your information in a story as coming from you; however, he or she can use it in other waysdto get another source to respond to your comment, for example. Before your interview begins, work out the ground rules with the reporter regarding how he or she plans to use your comments. You cannot take it back after you have said it. Despite these informal rules, the strong recommendation is always to speak on the record. 1. Keep the interview relatively short. Make it clear in the beginning that you have 10e15 min to talk. Ideally, your press officer has called the reporter and relayed to you the types of questions he or she intends to ask. That will help you prepare for the interview. If you allow the interview to go on too long, you become fatigued, or the reporter moves on to other topics, the chances for unintended results increase. 2. Be cordial, but not too casual, cavalier, or humorous. Humor is important, but in a media interview it can easily be taken out of context or misinterpreted by the reader. It is best kept at a minimum. 3. Decide ahead of time whether you prefer to do the interview over the telephone or in person. You have more control over your time if it is over the phone. You may, however, want to invite a reporter to your office if you are beginning to establish a long-term working relationship with them. 4. If you are doing a television interview, try to do it in person as opposed to a remote location. It is difficult to concentrate on a camera lens and easy for your eyes to shift around in a remote format, which will make you look less credible. 5. Have someone (preferably a communications expert) in the room with you when you do the interview. If it is by phone, at the outset of the conversation tell the reporter who else is in the room, and if you are tape-recording the interview. 6. Keep focused on the outcome. Some of the reporter’s questions may irritate you or surprise you. Sometimes that is a technique used to evoke an emotional response, which makes for an interesting quote, at your expense. Remain calm and think about what you want to read or hear in the story. If the reporter asks a negative question, do not repeat their words in your answer. For example, if the reporter asks, “Were you involved in the scandal?” and your answer is, “I had nothing to do with that scandal,” the reporter now has a quote from you mentioning a scandal. That is what readers will remember. Do not

IV. CLINICAL RESEARCH INFRASTRUCTURE

710

41. THE CLINICAL RESEARCHER AND THE MEDIA

take the bait. Rather, you answer with a “No” and move on to your core message. 7. Practice. Like everything else, practice improves your chances for success. Rehearse with your press officer or colleague before the interview, particularly if it is about a complex or controversial issue.

WHAT IF YOU ARE MISQUOTED? Misquotes happen. Even if you follow all of this advice and more, there is a chance your information will be reported in a different light than you anticipated. If the health message is incorrect and may have an effect on patients or the public, it is important to get the mistake corrected. Call the reporter immediately with the correct information and request a correction. If the health message is accurate, but you feel misrepresented, you can call the reporter or the editor, or write a letter to the editor. Talk to your communications office and they will help you decide on a case-by-case basis. As counterintuitive as it may seem, sometimes the best thing to do is just let it be. If you pursue a story correction, you may inadvertently keep the story in the news longer than it would have been if you had done nothing about it. If, however, you have been misquoted, or your comments have been represented inaccurately, you may consider seeking a clarification or a correction.

WHAT THE PUBLIC DOES NOT KNOW ABOUT SCIENCE Surveys show that about 70% of Americans say they get their health information from the mass media. That means your words have a great deal of impact. There are some good basic guidelines to keep in mind when you do an interview with the mass media. The first is that the publicdyour audiencedmay not know much about how science works. What you say and what they hear might not be the same thing. What You Know

What the Public Perceives

Research yields new knowledge and raises new questions.

A piece of published science is “The Truth.” For example, you might see a study on the risk of cancer and cell phones as raising more questions than it answers. The public might see it as a definite causal link.

Legitimate scientific differences of opinion exist.

They view differences as confusion. They want the final answer. The challenge is not to overstep what is known, and at the same time avoid waffling on the issue.

UNEXPECTED QUESTIONS Another situation that you should be prepared for is when you are asked unexpected questions. These questions will not necessarily be about your research, or even science. The following are examples of unexpected questions about a new therapeutic: • Who holds the patent or license? • What will the test cost patients? Will it save money in the long run? • Will insurance companies cover the cost of the test once it is developed? • Did you have a financial interest in this discovery? People care about these issues. Even if you do not believe you are the person to answer them, anticipate that they will arise. Work with your communications office to prepare for the toughest questions. The rule of thumb is to be responsive to the question, but move to the main purpose of the interview or press conference.

WHEN THE NEWS IS NOT GOOD Clinical research is not without risk. Some types of issues are guaranteed to attract media attention. Extreme cases, such as an unexplained death in a study, or deaths, will draw immediate media attention. Other bad news includes scientific misconduct or allegations of conflict of interest, both of which can turn into news stories that do not soon go away. In these cases, it is especially important for government scientists, such as those at NIH, who are involved, and even those not involved, to think through what they are going to say to reporters, if appropriate, and to seek advice from their communications colleagues on what, when, and how to engage reporters.

A WORD ABOUT INVESTIGATIVE REPORTERS As you consider how you are going to engage the media, take a moment to learn about a special type of reporter: the investigative journalist. They are likely to be the reporters who uncover bad news. Keep the following in mind: 1. They have more time than the average reporter who has a daily deadline to file. 2. They are most interested in irregularities, violations and/or misconduct, and seek to confirm or not confirm allegations. 3. They try to cultivate “whistle blowers” or unconventional sources.

IV. CLINICAL RESEARCH INFRASTRUCTURE

711

EMBARGOES

4. They will use the Freedom of Information Act (FOIA) to obtain documents, perhaps more often than a daily reporter. They may collect thousands of pages over many years. 5. From their painstaking research, they become extremely knowledgeable about your organization and gather a great deal of information, sometimes from people with an axe to grind. 6. Keep all of the above in mind when considering how to engage them if they ask to interview you. If you determine that a reporter interested in interviewing you is conducting an investigation, contact your institution’s communications office. They, in turn, will call the reporter, get additional information, and help you decide how to proceed.

THE FREEDOM OF INFORMATION ACT The FOIA applies to government materials and makes documents available to anyone, whether or not they are a citizen and whether or not we think they have a need to know. You cannot withhold documents because they make you look bad or because they could be misinterpreted by the public. There are nine exemptions in the FOIA that permit the government to withhold documents. The two exemptions that are most often used are the following: 1. Invasion of personal privacy, such as release of medical records 2. Commercial or financial information The following are documents that are available to anyone at any time under FOIA: • Approved research protocols • Minutes of NIH Institution Review Boards, with some possible deletions • Your email messages • Your computer files • Document drafts Under FOIA, it does not matter whether or not you stamp a document “Confidential.” Also, the term “predecisional” does not automatically exempt a document from release. Each request is considered individually. Each NIH institute and center has a FOIA officer to help you with requests. You will be involved in the process, however, you will not have any say in what is or is not released through FOIA. In the instance of NIH, only, one persondan attorney in the Office of Communicationsdhas the authority to deny documents under the FOIA.

EMBARGOES An embargo is an agreement between a scientific journal and reporters. It designates the time frame in which a story may be released. In other words, embargoes are dates established by scientific journals that prevent the premature release of a news story. For example, the December issue of the New England Journal of Medicine (NEnglJMed) hits the newsstands on December 14th. Copies are sent to the press and your institution several days ahead of that, but information on the stories cannot be released until the evening of December 13th. Likewise, if you are an author of an embargoed story, you must remind reporters that you are speaking under an embargo. In addition, refrain from talking to nonjournalists about an embargoed article because they are not held to the same checks as reporters. In science and medicine, there is a danger in talking about a pending study result that could potentially lead to “insider trading,” and stimulate an investigation by the Securities and Exchange Commissiondfor example, if the stock market showed unusual movement. In fact, if a research result is considered by NIH to be market sensitive, the press release is not made available until the stock market closes that day. Sometimes an ambitious reporter will jump the gun and break a story before the embargo. What happens then? • The journal may lift the embargo. • News stories may run ahead of schedule, and you may be permitted to proceed immediately with interviews. Whatever happens, you should contact your communications office for instructions. It is important to note that the future of embargoes is uncertain. However, despite more clinical papers having many authors, with fierce competition from reporters, with people posting their data on the Internet, and with the economic importance of clinical advances, embargoes are still honored by credible journalists.

The Ingelfinger Rule The Ingelfinger Rule, named after a former editor of the NEnglJMed was levied in the 1960s to control early release of NEnglJMedarticle information. The Ingelfinger Rule has succeeded in intimidating some scientists to the point that they feel uncomfortable giving media interviewsdeven delivering abstracts at a medical meetingdfor fear they will not get published.

IV. CLINICAL RESEARCH INFRASTRUCTURE

712

41. THE CLINICAL RESEARCHER AND THE MEDIA

The rules since have been clarified by NEnglJMed and JAMA and are more liberal. You can talk freely at meetings and still have that information published. You can talk to reporters about what you presented at meetings, but it is probably not a good idea to go beyond what you presented in public sessions, or to hand out the details of your data before publication. Also, never give out your manuscript to a reporter. Infrequently, events do occur that make both embargoes and the Ingelfinger Rule moot.

WHEN TO CONTACT YOUR COMMUNICATIONS OFFICE Throughout this chapter, several circumstances have been discussed that call for you to seek assistance from your institution’s communications office. The following is a condensed list of occasions in which this would be appropriate: • When you get a request from a reporter for an interview. It is the policy for most institutions, and a good idea for your own comfort level. • Any time you are doing an interview with a major newspaper, magazine, radio outlet, TV network, or Web-based news outlet. • To get help on how to phrase answers for the public. • To do a dry run for TV or radio. • To learn what your institutional policy is on a matter. • When you are concerned about an investigative reporter who wants to talk to you. • If you have a question about embargoes.

CONCLUSION Although speaking to the media comes with its challenges, there are clear guidelines and help available to get the message out about your research, the benefits of science, and how we all are working to make discoveries that will improve public health. As a clinical researcher, you may find yourself in demand for media interviews. Remembering these key points will help you engage successfully with reporters: • People want to hear science and medical news. • For those who work at a government institution, we are obligated to inform the public about our work. • Understand that bad news or an ethically questionable problem draws media attention.

• Most institutions, but especially the NIH, encourage you to talk with reporters. It adds credibility and reflects well on the institution and the biomedical research community. • If you get a call from a reporter, get the information you need before agreeing to the interview. • Use plain language in explaining your work for the general public. • Develop a core message and various ways to convey it. • Assume everything is on the record and only say what you would want to read in the story. • Be aware of media issues, such as embargoes, that are unique to scientists. • Contact your institute’s communications officer for assistance. • Keep in mind the ever-changing landscape of social media.

SUMMARY QUESTIONS 1. When you agree to an interview with a reporter with the explicit understanding that they can use any or all that you say to them in their story, it is called speaking: a. off the record b. on the record c. on background, and not for attribution 2. The law that makes government documents available to the public is called: a. Freedom of Information Act b. Ingelfinger Rule c. The Embargo Act 3. What makes news in science and medicine? a. Published science b. Novelty c. The unexpected d. Celebrity e. All of the above 4. Why would most reporters want to interview you? a. To catch you in a contradiction b. To create controversy c. To misquote you to advance the story d. To include your scientific perspective 5. What tools and resources are available to you in speaking with reporters? a. The press officer in your communications office b. Media trainer c. Colleagues with media experience d. All of the above

IV. CLINICAL RESEARCH INFRASTRUCTURE

C H A P T E R

42 Information Resources for the Clinical Researcher 1

Josh A. Duberman1, Pamela C. Sieving2

National Institutes of Health, Bethesda, MD, United States; 2Sieving Information Solutions, Bethesda, MD, United States

O U T L I N E Introduction

714

Organization and Features of Information Resources 714 Origin

715

Content and Structure

715

Search Capabilities

717

Citation Searching

721

Access and Business Models

721

Familiarity and Currency

723

Biomedical Databases

724

Biomedical Databases (PubMed, Embase)

724

Cited Reference Databases (Web of Science, Scopus)729 Clinical Decision Resources (Micromedex, Lexicomp Online, Facts & Comparisons eAnswers, UpToDate)732 Drug Development Information (Pharmaprojects, other commercial databases, FDA information)

733

Clinical Trials Resources (Cochrane Reviews, clinicaltrials.gov)

733

Evidence-based Medicine and Systematic Reviews

735

Textbooks and reference works (NCBI Bookshelf)

737

Fulltext Searching (PubMed Central)

737

Direct Searching of Journal Articles (ScienceDirect, Google Scholar, Microsoft Academic) 738

Principles and Practice of Clinical Research http://dx.doi.org/10.1016/B978-0-12-849905-4.00042-3

Specialized Clinical Resources (ToxNet, PsycINFO, CINAHL, RePORTER, grey literature, dissertations and theses, FDA medical devices databases, EI Compendex, Inspec, patents 738 Internet Search Engines (Google/Google Scholar/ Google Books/Google Patents/Google News/Google Translate, Bing/Microsoft Academic, Startpage/ IxQuick, DuckDuckGo, other Internet search engines) 741 Bioinformatics Resources Major Bioinformatics Organizations Bioinformatics Directories Browsers Commercial Software

744 745 746 746 746

Data Management

746

Data Integration and Precision Medicine

746

Bibliometrics

747

Bibliographic Managers

748

Resource Selection and Search Strategy

748

Educational Resources

749

Final Notes

749

Acknowledgments

750

References

750

713

Copyright © 2018. Published by Elsevier Inc.

714

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

INTRODUCTION An incomplete literature search can have disastrous consequences in clinical research. In 2001, a healthy 24-year-old technician named Ellen Roche volunteered as a normal experimental subject in a Johns Hopkins University study. She was given hexamethonium by inhalation to produce ganglionic neural blockade and to simulate asthmadbut she developed increasingly severe lung problems and died. Hexamethonium’s potential toxicity was not uncovered by the principal investigator’s literature search, despite a “standard PubMed search” and consulting “standard current editions of textbooks of pharmacology and medicine,” according to a Johns Hopkins internal investigation.1,2 One issue with a PubMed-based literature search in this subject area was that many of the relevant articles that mentioned hexamethonium-related pulmonary toxicity were published in the 1950s. In 2001, MEDLINE, on which PubMed is based, only included citations from 1966 forward. For a comprehensive literature search, it would have been necessary to search additional resources and to use search strategies appropriate to the content, structure, and search features in each of those resources. The Johns Hopkins internal investigators referred to “the vagaries of performing such a search” and gave various examples.1 Some of the issues and search strategies involved were detailed in other articles published at the time.3,4 After this tragedy, research oversight at Johns Hopkins was strengthened in a number of ways. One of these was that investigators were required to work with a librarian and a pharmacist to help search appropriate databases for potential side effects.5 Guidelines for comprehensive literature searches were established at Johns Hopkins and emulated elsewhere, including Mount Sinai School of Medicine.6 The focus of this chapter is on information for a variety of clinical research tasks, including searching for adverse effects, evidence-based practices, and research protocol development. These resources and strategies are broadly applicable to many of the topics discussed in other chapters in this text. This chapter presents an overview of information resources for clinical research, their scope and capabilities, in an attempt to demystify the process and results of searching. Features and characteristics of different types of information resources are covered, including MEDLINE/PubMed, Embase, other free and subscription-based databases, and Internet search engines. Summaries of search strategies, tips, associated

topics, and emerging issues in information resources are also included, with references to tutorials and supporting material. The purpose of this chapter is to provide a background framework on which to build and learn more. Major resources important for general clinical research are included, as well as representative resources covering a variety of associated subjects and areas. However, presenting a comprehensive list of resources in this chapter is not possible due to space limitations. There are many more information resources available, and interested readers are urged to check the references provided, in particular Huber and Srodin.7,8 Please note that information resources are in a state of constant transformation due to economic and technical conditions as well as the rapid growth of medical knowledge. Illustrations are included in this chapter to demonstrate underlying principles and examples; however, specific examples may not be available in the future. Links to Web pages were current when the text went to press. If a specific referenced web page is not available, the historic page may possibly be found by searching for the link in the Internet Archive’s “WayBack Machine” (www.archive.org).

ORGANIZATION AND FEATURES OF INFORMATION RESOURCES A wide spectrum of types of information can be relevant to clinical research including but not limited to • biomedical and technical journals; • bibliographic databases useful in discovering relevant information; • laboratory and patient records, clinical and drug data; • chemical properties data and other chemical information; • genomic and biochemical information; • conference reports and abstracts; • scientific and technical books; • standards and specifications; • patents and engineering diagrams and data; • statistics, policy documents, and white papers; • economic and legal literature, and that of the social and behavioral sciences; • theses and dissertations; • Web pages, blogs, tweets, e-mails; and • manufacturer-provided drug and device information. There are a number of useful ways to categorize these information resources to differentiate between them and understand how to use them more effectively. We have

IV. CLINICAL RESEARCH INFRASTRUCTURE

CONTENT AND STRUCTURE

chosen five categoriesdorigin, content, capabilities, access, and familiaritydwhich can help in deciding which resources to search and when. However, be aware that information resources are in flux, with frequent updates and changes; some of these categories can overlap, and divisions between them blur.

ORIGIN Proximity to source material is a classic method for differentiating information. In this system, information is organized according to its distance from primary literature, which is defined as “reports of original research” of “the initial discovery of health knowledge usually with original data”.9 The primary literature contains methodology, details, and results that enable assessment of the research conclusions.10 Examples of primary literature include journal articles (published in refereed or nonrefereed journals, online, or in print), conference abstracts and proceedings, patent applications and dissertations or theses.11 Secondary literature resources are indexing and abstracting services, or bibliographic databases used to identify other information resources, such as primary or tertiary literature. MEDLINE/PubMed, Embase, Web of Science, Scopus, and CINAHL are examples of commonly available secondary literature resources. Tertiary literature resources analyze and synthesize information from the primary literature and provide a synopsis or guided overview; they are generally useful for identifying primary literature. Tertiary literature can include reviews, textbooks, compendia, encyclopedia entries, reference works, books, evidenced-based resources, and some databases. Micromedex and UpToDate are both tertiary literature resources. These categories can provide a good basic approach to understanding how these types of literature can be used, though there can be additional refinements and ambiguities. Research data sets, electronic medical records, and laboratory data, all “primary data,” may be available as stand-alone publications or in databases, though there should be links to the methodology used to derive the data. E-mails, blog entries, tweets, press releases, and newspaper articles could all function as leads to primary or tertiary literature. The Publication Cycle is a concept that is analogous to proximity to source material, and is important to consider when searching for information about new research. This concept describes the researchpublication cycle from the inception of the idea for research, to the publication of the results in various

715

forms. One example might include idea, preliminary research, research grant application, receipt of research grant, conducting of research, patent application submission (to be published in 18 months from submission), conference presentation with publication of abstract, primary journal submission for peer review, primary literature publication with simultaneous data set availability, news releases and/or blog mentions and tweets, and patent application publication. The cycle can be repeated indefinitely over the course of a research project or career. It also can be cut off before the cycle is complete: a clinical trial may be canceled because of early conclusive unexpected positive or negative results; a project may be reported at a conference but not published in the peer-reviewed journal literature.

CONTENT AND STRUCTURE The content and scope of information resources are obvious practical distinctions. Is the desired information likely to be in the resource? What subjects are covered, in what degree of comprehensiveness, in a given information resource? And, what is searchable and how? Does the resource include only bibliographic information, such as titles and authors? Are abstracts available? Can the full text of the article or book be searched electronically? Have keywords been added to a bibliographic record and, if so, have they been selected from a structured thesaurus with broader and narrower descriptors? Are there other searchable elements: year of publication, information about funding agencies, language of the text, cited and citing references, links to other information resources, availability of the full text online? Electronic or online databases often comprise structured records representing each individual item referred to in the database. In a bibliographic database such as PubMed/MEDLINE, these records consist of “fields” such as article title, author, journal title, year, etc. Fig. 42.1 shows the PubMed/MEDLINE record for an article by J. Savulescu and M. Spriggs, “The hexamethonium asthma study and the death of a normal volunteer in research”.2 The MEDLINE database record is available at http://www.ncbi.nlm.nih.gov/PubMed/?term¼118 34748&report¼medline&format¼text; the descriptions of the fields are available at https://www.nlm.nih.gov/ bsd/mms/medlineelements.html. Some of the fields are obvious and expected, such as AU (author), TI (title), and JT (journal title); others such as RN, OTO, and OID are cryptic. In general, bibliographic databases support inquiries that “ask” the search engine to query any or several of the fields in the individual database records.

IV. CLINICAL RESEARCH INFRASTRUCTURE

716

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

PMIDOWN STATDA DCOMLR IS IS VI IP DP TI PG FAU AU AD -

FAU AU LA PT PT PL TA JT JID RN RN SB SB CIN MH MH MH MH MH MH MH MH MH MH MH MH MH MH MH PMC OID OID OTO OT OT GN GN EDATMHDACRDTPST SO -

11834748 NLM MEDLINE 20020208 20020314 20140612 0306-6800 (Print) 0306-6800 (Linking) 28 1 2002 Feb The hexamethonium asthma study and the death of a normal volunteer in research. 3-4 Savulescu, J Savulescu J Ethics Program, The Murdoch Childrens Research Institute, Royal Children's Hospital, Flemington Road, Parkville, Melbourne, Victoria 3052, Australia. [email protected] Spriggs, M Spriggs M eng Case Reports Journal Article England J Med Ethics Journal of medical ethics 7513619 0 (Ganglionic Blockers) 3C9PSP36Z2 (Hexamethonium) E IM J Med Ethics. 2002 Feb;28(1):1-2. PMID: 11834747 Administration, Inhalation Asthma/*drug therapy Baltimore Controlled Clinical Trials as Topic *Ethics Committees, Research Fatal Outcome Female Ganglionic Blockers/administration & dosage/*adverse effects Hexamethonium/administration & dosage/*adverse effects *Human Experimentation Humans *Professional Misconduct Professional Staff Committees United States Universities PMC1733509 KIE: 101672 NLM: PMC1733509 KIE Biomedical and Behavioral Research Johns Hopkins University KIE: 24 refs. KIE: KIE Bib: human experimentation/ethics committees 2002/02/09 10:00 2002/03/15 10:01 2002/02/09 10:00 ppublish J Med Ethics. 2002 Feb;28(1):3-4.

FIGURE 42.1 MEDLINE record 11834748. Record available at http://www.ncbi.nlm.nih.gov/pubmed/?term¼11834748&report¼medline&format¼text. Descriptions of fields or data elements are available at https://www.nlm.nih.gov/bsd/mms/medlineelements.html.

Some systems default to a small set of these fields, such as title, abstract, and keyword; some present choices with pull-down menus; others require the user to specify the fields, such as this query, which asks PubMed to find

citations which are case reports (a publication type) with the word “hexamethonium” in the title of the article: hexamethonium[ti] AND case reports[pt]

IV. CLINICAL RESEARCH INFRASTRUCTURE

717

SEARCH CAPABILITIES

Database producers frequently add subject indexing terms to the individual records in an attempt to better describe the information in the item and to improve the probability that the record will be retrieved by a well-planned search. These index terms may be chosen from a selected list, or a controlled vocabulary thesaurus, to help solve the problems of how to express different concepts, and how to deal with synonyms. In both MEDLINE/PubMed and Embase, added index terms, which describe the subject of the article, are chosen from a taxonomy, a hierarchically structured set of broader, narrower, and related terms called, respectively, MeSH (Medical Subject Headings) and EMTREE. In Fig. 42.1 the MeSH term fields are abbreviated “MH,” the terms following the slashes are subheadings, and the asterisks indicate that the term is a major concept for this item. These features of MEDLINE indexing can be used to retrieve records for which a concept is not well described in title or abstract words, limit the retrieval to a specific aspect (such as adverse effects of hexamethonium), and retrieve articles in which the specified subject is the major focus rather than one of many. The MeSH record for hexamethonium is shown in Fig. 42.2 and is also available at http://www.ncbi.nlm. nih.gov/mesh/68018738. The record shows various synonyms or “Entry Terms,” which are generally interchangeable with the query term, so that searching for either the MeSH term or an entry term will retrieve the same item (see https://www.nlm.nih.gov/mesh/intro_entry. html). The record indicates which subheadings can be used to qualify this MeSH term; all the subheadings used in PubMed/MEDLINE are listed at http://www. ncbi.nlm.nih.gov/books/NBK3827/table/PubMedhelp.T. mesh_subheadings/. The record also shows that the MeSH term “Hexamethonium” was introduced in 1995; previous to that date, from 1980 to 1994, the MeSH term used was “Hexamethonium compounds.” MeSH terms are updated annually, and for comprehensive retrieval it is often necessary to verify previous indexing terms.4 To conduct a thorough search in 2001, it may have been necessary to use both of the MeSH terms. More recently, some MeSH terms, including “Hexamethonium,” have been applied to older literature. The PubMed tutorial at https://www.nlm.nih.gov/bsd/ disted/pubmedtutorial/020_490.html, as well as some of the links in the PubMed section below, provide more information about MeSH.

SEARCH CAPABILITIES The search interface and capabilities of the information resource can also make a great deal of difference in creating effective search strategies. If data in the resource are organized in fields, can particular fields be specified as part of a search? Can Boolean logic, wildcards, or truncation be used in searches to widen retrieval efficiently? Is the search history available? Can previous result sets be combined with future sets, or must each iteration of the search strategy be created anew? When formulating a literature search strategy, it is often useful to break the search question into its essential concepts. For clinical care questions, a useful mnemonic is PICO: P for patient/condition/problem; I for intervention being considered; C for the comparison intervention; O for the outcome desired or to be measured. Many clinical questions do not have enough of an evidence base for PICO to work as hoped; one tactic is to start with the broadest aspect of the question, on which the researcher expects to find the most information, then add in each of the other aspects in probable order. A Boolean AND will always limit search retrieval further; “ANDing” a query to retrieval that is already a small set of references may be counterproductive. Each aspect of the question can be addressed separately, and the results combined, or in some cases a single search query can be constructed. To ensure the most comprehensive retrieval, researchers should consider possible alternate terms, plurals, synonyms, and alternate spellings for each concept. Examining the MeSH headings in Fig. 42.1 identifies three possible MeSH terms for the concept “hexamethonium.” “Hexamethonium” and the previous MeSH term “Hexamethonium Compounds” were mentioned above. The term “Ganglionic Blockers” is also used to index the Savulescu and Spriggs article. The concept of “adverse effects” as a subheading can be seen applied in Fig. 42.1, and as an allowed subheading in Fig. 42.2; the “poisoning” and “toxicity” subheadings shown in Fig. 42.2 could also be considered for more complete retrieval. As with the MeSH term definitions supplied by the online taxonomy for PubMed/MEDLINE, it is critical to know how a database defines such subheadings. The following might all seem to be useful in a search for possible negative reactions to a drug: Administration and dosage Adverse effects Complications Drug effects Pharmacology

IV. CLINICAL RESEARCH INFRASTRUCTURE

718

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

FIGURE 42.2

Hexamethonium MeSH record. Record available at http://www.ncbi.nlm.nih.gov/mesh/68018738.

IV. CLINICAL RESEARCH INFRASTRUCTURE

SEARCH CAPABILITIES

FIGURE 42.2

Poisoning Toxicity However, the term “adverse effects” is used for procedures as well as chemicals, drugs, and biological agents; the term “complications” is used only with diseases, and thus is not applicable here; “poisoning” is used for accidental or medication error as well as environmental exposure. Informed use of subheadings can be powerful in focusing retrieval; for a longer explication see Golder et al.12 An important MeSH feature is its hierarchical structure. MeSH is constructed with broad terms at the top and increasingly narrow terms below it. For example, “eye diseases” can be used in a broad-stroke search for most eye diseases; under that one finds “corneal diseases” with further subdivisions for “keratitis” and then “keratitis, herpetic.” Using the pull-down box to the left of the PubMed search box lets the searcher browse the MeSH headings; other systems provide various ways to examine indexing taxonomies. Another important task in constructing a search strategy is a form of divergent thinking. In addition to focusing on a disease, consider tests for that condition, therapeutic options and drugs unique to it, outcome measures, etc. For cataract surgery, for instance, consider: cataract/ su; cataract extraction; phacoemulsification; lens, crystalline/su.

719

cont’d.

Starting with a straightforward search statement, then reviewing retrieval for additional terms (including looking carefully at the subject indexing of citations retrieved), is a powerful way to build a comprehensive search. If the information resource has the capability, Boolean logic can be used to group similar concepts together and to combine search terms. The Boolean “OR” operator specifies that at least one of the search terms must be in the results, while the “AND” operator specifies that all the terms must be in each results. Every bibliographic database has its own rule for the default operator; PubMed’s is AND. Thus hexamethonium [ti] rats[ti] searches for all articles with both “hexamethonium” and “rats” in the titles. hexamethonium[ti] OR nicotine[ti] searches for either word; (hexamethonium[ti] OR nicotine[ti]) AND rats[ti] looks for either of the substances, plus “rats”, in the titles. Using parentheses is a help to the searcher in grouping concepts together, and in specifying how sets of results are constructed by the search system, avoiding, for instance, retrieval of all articles about either

IV. CLINICAL RESEARCH INFRASTRUCTURE

720

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

FIGURE 42.3 PubMed Search DetailsdThe Search Details box at lower right shows exactly how the search was interpreted (if it is a long query, click “see more” to see how the entire query was interpreted).

hexamethonium OR about (nicotine and rats), a quite different result. A broad search query using “OR” logic in PubMed with all 9 variations mentioned above would be: (“hexamethonium/adverse effects” [MeSH] OR “hexamethonium/poisoning” [MeSH] OR “hexamethonium/toxicity” [MeSH]) OR (“hexamethonium compounds/adverse effects” [MeSH] OR “hexamethonium compounds/ poisoning” [MeSH] OR “hexamethonium compounds/toxicity” [MeSH]) OR (“ganglionic blockers/adverse effects” [MeSH] OR “ganglionic blockers/poisoning” [MeSH] OR “ganglionic blockers/toxicity” [MeSH]) Note the double quotes in the above search statement, which specify an exact phrase or “string” search, as well

as the postterm field qualifier “[MeSH]”. These search features are available in PubMed and in many other information resources, though the exact grammar may vary. However, double quotes in PubMed produce the desired phrase retrieval only if indexing has created that phrase as a searchable term. The “highlighting” feature (see https://www.ncbi.nlm.nih.gov/books/ NBK53593/ for information on setting preferences) allows the searcher to quickly see, for each record, whether the words in the phrase are adjacent to each other in the title or abstract; PubMed also displays Search Details in a box, displaying exactly how the search was interpreted. For an example, see the search for “hexamethonium death” shown in Fig. 42.3. This is not an indexed phrase in MEDLINE, so the search function defaults to the Boolean AND and retrieves over 60 citations, including one article titled “Death associated with hexamethonium” from 1953; however none of the 62 have the phrase “hexamethonium death,” and the

IV. CLINICAL RESEARCH INFRASTRUCTURE

721

ACCESS AND BUSINESS MODELS

Search Details box at the lower right of Fig. 42.3 explains how the search query was interpreted: (“hexamethonium”[MeSH Terms] OR “hexamethonium”[All Fields]) AND (“death”[MeSH Terms] OR “death”[All Fields]) Most files allow the Boolean NOT command in addition to AND and OR. Use this cautiously, as it may eliminate articles that discuss both a desired aspect as well as the one specified to be eliminated. A number of other useful search features may be available depending on the information resource. Wild card and truncation operators can indicate one or multiple characters in search terms and so can help cover many term variations; they may function at the right- or left-hand of a term, or in place of characters not at either end: *vision might retrieve “envision” as well as “vision”; read* retrieves “readers,” “reading,” and “readjustment”; an*esthesia retrieves the British form of the term as well as the American. Automatic British/American spelling and term conversion also can be useful (color/colour, lorry/truck). Watch for “help” or tutorial information for a database that notes “stemming” or “lemmatization” in the search function; this indicates that the search will automatically include normal variations on a term. For American searchers, it can be very helpful to not be required to remember that edema and esophagus sometimes begin with an O. A search history feature that displays previous searches and enables their use in creating new searches can be extremely useful in working out possible search strategies; usually these features assign a number to each of the previous search statements and allow combinations using Boolean commands: (#1 OR #2 OR #3) AND (#5 NOT #7). When performing a comprehensive search, it is important that the search is transparent. The mechanism of how the terms are actually searched, in which fields, should be readily apparentdas in PubMed’s Search Details boxdso that the search strategy can be improved if necessary. Of course, it is also important that the search be reproducible; that repeating the search will give the same results. Finally, many journals now require comprehensive reporting of the literature searches performed, not simply a statement that X concept was “searched in PubMed.” Some research, such as systematic reviews, requires detailed reporting of search terms and limits, databases searched, years and languages included in the retrieval, and other specifics that would enable other researchers to verify and replicate the search. Good record keeping is, thus, essential. Many bibliographic databases allow some form of user registration. All searches done after logins then can be saved, for future reference, for modification, and for updating. The “My NCBI” (National Center for Biotechnology Information) feature in PubMed, for example, does this, and allows the searcher to request daily, weekly, or monthly automated updates.

CITATION SEARCHING Citation searching takes an idea (or an author or a funding line) both forward and backward in time, and illustrates and scientific progress as reflected in the published literature. An author builds on work already completed; references cited by a paper incorporate that knowledge, and the paper continues the work represented in those references in a direct or indirect manner. It is a simple matter to examine the references in a useful paper. Finding the papers which cite that paper is the function of databases, which support citation searching. The two most comprehensive files, Web of Science and Scopus, as well as several other less general databases, both feature cited-reference searching capabilities: for a given article, the database identifies other articles and books, published later, which cite the original article. This capability can identify subject-related articles without searching for them with subject indexing terms or terms from the title or abstract of other articles. In fact, Eugene Garfield’s original concept for a science citation index13 was focused on the “association of ideas” to move science forward quickly in the absence of adequate subject indexing of the literature. One also can track research on a particular subject forward in time. The Johns Hopkins University internal investigators found a review article on drug-induced pulmonary disease with 224 references14 and noted that it included six articles on hexamethonium toxicity published between 1954 and 1962.1 Hexamethonium did not appear in the indexing for this review article in MEDLINE or Embase; however, it is possible that this article, and the six articles it referenced, could have been found through cited reference searching. Limited cited reference searching is also available in the following files: PubMed, CINAHL, PsycINFO, Google Scholar, and Microsoft Academic.

ACCESS AND BUSINESS MODELS There are many pragmatic questions regarding choice of and access to information resources for clinical research. Is the resource readily available to youddo you subscribe to it, or is it in the collection of a library to which you have access? Is the resource available online, and can you access it remotelydor is it only available in paper? And, is it available free, or is there a fee, a subscription or cost required? Clinical researchers often start with PubMed, Internet search engines, and full text resources such as PubMed Central (PMC), because these do not require subscriptions and are easy to access. However, these resources by themselves are often inadequate for comprehensive

IV. CLINICAL RESEARCH INFRASTRUCTURE

722

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

searches. As a preliminary to initiating a research project, it is important to check for access to other information resources through a relationship with a school, an employer or professional affiliations. Other places to investigate include local universities and medical schools (particularly those supported by public funds), public libraries, professional societies, and alumni associations. Collaborators on a project may have individual access to resources not available to all. In the United States, biomedical information resources are available via the National Network of Libraries of Medicine (NN/LM, https://nnlm.gov/), which includes local centers, training opportunities (https://nnlm.gov/ training, http://nnlm.gov/moodle/), and links to Internet health information (https://nnlm.gov/hip). A number of programs provide assistance with literature resources in other countries. The Health InterNetwork Access to Research Initiative (HINARI), a partnership of the World Health Organization (WHO) and scientific publishers, enables biomedical literature access for researchers in low and middle-income countries (http://www.who.int/hinari/en/). HINARI is part of the Research4Life Initiative, http://www. research4life.org/, which also includes resources for agriculture, the environment, and development and innovation. Researchers in developing and transition countries may also obtain information services from the International Network for the Availability of Scientific Publications (INASP, http://www.inasp.info/en/ and http://www.inasp.info/training-resources/). WHO provides literature access via the database Global Index Medicus http://www.globalhealthlibrary.net, which has links to regional WHO libraries: Africa, Americas, Eastern Mediterranean, Europe, South-East Asia, and Western Pacific. These databases provide access to journals not indexed in MEDLINE, often providing important information on diseases, epidemiology, public health, and research in developing countries. Additional WHO Health information and resources are available at http://www.who.int/ healthsystems/topics/information/en/. More information about international access to biomedical information is available from the National Library of Medicine (NLM)dsee details at https:// www.nlm.nih.gov/pubs/factsheets/intlmedlars.html and https://www.nlm.nih.gov/services/international. html. Additional information may also be available at various medical schools and universities; for example, see Global Access to Journal Literature, Rudolph Matas Library of the Health Sciences, Tulane University http://libguides.tulane.edu/openaccess. Scientific papers have been made available free online starting from the early days of the Internet. The repository arXiv, with e-print content in physics and related fields,

opened in 1991.15 Additional archives following the Open Archives Initiative (http://www.openarchives. org) are now online, including Cold Spring Harbor Laboratory’s bioRxiv (pronounced “bioarchive”, http:// biorxiv.org/) for life sciences preprints. “Green” open access refers to such programs in which papers, data, etc. are deposited and made freely available in a repository. Scientific journals have also been published online since the 1990s on the Internet.16 Generally, access to journal content continues to be dependent on payments, either personal or institutional subscriptions or perarticle payments. Some journals have evolved or begun as freely available without cost to the reader or the reader’s institution: the scientific publisher Public Library of Science (PLoS) was founded in 2000, and others have followed. Another model is “gold” open access, requiring a payment by the author, the author’s institution or funding agency, to the publisher to provide access to any interested reader. The term “Open Access” for online free-to-access literature came into general use after the 2002 Budapest Open Access Initiative (http://www.budapestopenaccessinitiative.org/); other documents describing the principles of this initiative have been published after meetings in Bethesda, 2003 (http://legacy.earlham.edu/wpeters/fos/bethesda.htm); Berlin, 2003 (https://openaccess.mpg.de/BerlinDeclaration); and Brasilia, 2013 (https://openaccess. mpg.de/1526935/Session-3-Costa.pdf). The Directory of Open Access Journals (DOAJ, http://www.doaj.org/) currently lists more than 9000 journals that meet the DOAJ’s standards for publishing quality and access standards. There have been a number of recent moves toward making more research literature freely available to the public. Starting in April 2008, peer-reviewed articles arising from NIH-funded research must be made available in PubMed Central within 1 year of publication according to the NIH Public Access Policy (http:// publicaccess.nih.gov/). In 2013 the US Obama presidential administration directed that published results of federally funded research be made available for free within 1 year of publication.17 It is expected that there will be increased access to research data in the future as well (for more information, see the “Data Management” section). There are costs involved in acquiring, editing, managing, distributing, and archiving the content of scholarly journals. Those costs, plus profit returned to commercial publishers of scientific journals, or surplus revenue that scholarly societies have used to provide services to their members, have traditionally been funded with subscriptions. The shift to free access at the reader’s level currently involves several models: author page charges, funding by the author’s

IV. CLINICAL RESEARCH INFRASTRUCTURE

FAMILIARITY AND CURRENCY

institution, the author’s grant funds, or the author’s personal resources; direct funding by some funders such as the Wellcome Trust and the Howard Hughes Medical Institute; increased resources from scholarly societies devoted to the costs of their publications; and creative approaches at the journal level with some relying on temporary support from their home institutions or actively building endowments to support their costs in the long term. The business model of an information resource can greatly affect questions of access, user interface, and efficacy of use. It can also affect the quality of the publication. The ease of establishing online journals and the relatively low cost of online publishing has led to many new online journals, some of which may take publishing fees but provide no peer review, little editorial input, and no program of distribution and archiving of content; in some cases the author fees are collected without publishing articles. Scientists have been cautioned to avoid these fraudulent “predatory publishers”18; the list of questionable journals and publishers mentioned in the article is not available at this writing, but a recent historic list is available at http://web. archive.org/web/20170111172309/https://scholarlyoa. com/individual-journals/. For more information, see “Identifying Predatory or Pseudo-Journals” from the World Association of Medical Editors, February 18, 2017, at http://www.wame.org/about/identifying-predatoryor-pseudo-journals. MEDLINE/PubMed and other NLM and US government information resources are supported by US tax dollars. Many of the other indexing and abstracts services below are available via commercial subscriptions from the information producers. However, some resources may also be available through aggregators, such as STN (from FIZ Karlsruhe and the CAS division of the American Chemical Society), ProQuest Dialog, and Lexis-Nexis. Since some aggregators have a “payas-you-go access” business model, researchers may be able to request funding for access to these resources as part of their grant applications. Note that many different versions of a specific resource, with differing date ranges, features, and capabilities, may be available from various vendors at different prices. “Free” information resources still cost money to produce, and users should be aware of the business model of the information provider, and of any affiliations, sponsorships, or biases. For example, Google and many other Internet search engines are supported by selling advertising space to companies, which wish to market to people who use the search engines. If Google increases its market share, it can charge advertisers moredso Google’s optimal strategy would be to attract more users, and to keep current users coming back. To do that, Google searches should produce what most

723

users are looking for: accurate and appropriate answers. However, answers that satisfy most users are not necessarily the complete, thorough, unbiased, and scientifically valid answers researchers may need. Google does not reveal exactly how it prioritizes the output of its searches, since “Search Engine Optimization” companies may try to game the system and help companies achieve higher rankings and thus get more business.19 Google also may tailor its search results depending on the user’s location and history of searches. Google searches are not transparent and may be difficult to replicate; see more information in the Internet Search Engines section.

FAMILIARITY AND CURRENCY Many researchers turn first to information resources that they are familiar withdsources that they have used before or that they have heard about. However, the information field is rapidly changing and evolving, and there may have been important shifts in coverage or capabilities since the last time the researcher used the resource. Resource-specific learning opportunities and support are often available from the producer or vendor, and links are listed below. New information resources become available frequently. Researchers should adopt a regular scan to keep aware of any new information developments: good places to check include conferences, networking with colleagues, and libraries. Many resources are updated regularly with new content, as new material is published; this ranges from daily to monthly or even annually varying with the resource and the vendor. In some resources, researchers can set up an “alert” to rerun a search automatically and receive new information on their topic. If this capability is not available, researchers should consider manually repeating their searches at regular intervals. In any case it is also vital to monitor the information source for changes over time: subject indexing terms can be added or changed in significant ways; features that allow the searcher to limit retrieval by such conditions as date of publication, publication types, and funding agencies may also be changed, added, or deleted; bibliographic databases may deselect journals previously indexed, or add others, sometimes indexing the back volumes, other times not doing so. Thus the same search run in a database even a few days apart may retrieve markedly different citations. For general information about searching, researchers should consult the “Educational Resources” section as well as the references in this chapter.

IV. CLINICAL RESEARCH INFRASTRUCTURE

724

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

BIOMEDICAL DATABASES

PubMed Name/producer/cost/type: PubMed http://www.ncbi.nlm.nih.gov/PubMed/; National Center for Biotechnology Information (NCBI), National Library of Medicine (NLM), National Institutes of Health (NIH); free; bibliographic citations with added indexing, many with author-provided abstracts. Sources/subjects: Citations to biomedical research articles from more than 5600 journals in 39 languages from the 1940s to the present with some older articles; broad subject areas include medicine, nursing, dentistry, public health, veterinary medicine, allied health, preclinical sciences. PubMed included nearly 27 million records at the time this chapter was written; to view the current number, enter all [sb] in the PubMed search box. MEDLINE is the underlying database; PubMed is the web-based interface, plus additional content, including in-progress citations, nonindexed content of some journals indexed in MEDLINE, and articles included in PubMed Central that are not indexed for MEDLINE.20 MeSH (Medical Subject Headings) is the NLM’s hierarchically structured controlled vocabulary thesaurus used for indexing. Special features: “Advanced” search (https://www.ncbi.nlm.nih. gov/PubMed/advanced) facilitates building searches with Boolean logic operators; use My NCBI for saved searches, automatic updates, and customizing the interface, see http://www.ncbi.nlm.nih.gov/books/ NBK53592/; consider using PubMed Clinical Queries for noncomprehensive searches limited to specific clinical research areas; see https://www.ncbi.nlm.nih. gov/books/NBK3827/#pubmedhelp.Clinical_ Queries_Filters for complete information on the filters used to limit these searches to evidence-based medicine, and https://www.nlm.nih.gov/bsd/ pubmed_subsets/sysreviews_strategy.html for systematic reviews. Limited cited reference searching is available based on the citations in full text articles in PubMed Central. PubMed uses NCBI’s Entrez search and retrieval system https://www.ncbi.nlm.nih.gov/ books/NBK184582/ and can be accessed using a variety of automated techniques: APIs https://www. ncbi.nlm.nih.gov/home/develop/api.shtml; Entrez

Programming Utilities (eUtils) http://www.ncbi.nlm. nih.gov/books/NBK1058/; and Entrez Direct (EDirect) http://www.ncbi.nlm.nih.gov/books/ NBK179288/ Information/support: Overviews https://www.nlm.nih.gov/bsd/ pmresources.html#overviews; FAQ (answers to Frequently Asked Questions) https://www.nlm.nih.gov/services/PubMed.html; PubMed Help with QuickStart http://www.ncbi. nlm.nih.gov/books/NBK3827/; Online training https://learn.nlm.nih.gov/rest/ training-packet/T0042010P.html; Tutorials https://www.youtube.com/playlist? list¼PLBD13A2628C7A9965; MEDLINE/PubMed and other databases FAQ https://www.nlm.nih.gov/services/faqmed.html; MeSHddescription and links https://www.nlm. nih.gov/mesh/; MeSH FAQ https://www.nlm.nih.gov/pubs/ factsheets/mesh.html It is recommended that researchers check with the libraries at their institutions for access to PubMed, as some institutions may provide links to the full text of subscribed publications. The MEDLINE database is also available with additional search functionality for a fee from a number of vendors. To achieve comprehensive results using PubMed, it is vital to understand how to search it. There are search aids available online: clicking “Help” will bring you to an online book, and clicking “PubMed Tutorials” will bring you to web-based interactive tutorials. Additional PubMed educational materials are available as shown in the PubMed information capsule, and as indicated in Fig. 42.4, the PubMed home page. Also in Fig. 42.4, note the link to the MeSH database discussed earlier, and the “Advanced” link to the PubMed Advanced Search Builder, which facilitates combining previous PubMed searches using Boolean operators. Fig. 42.5 shows the PubMed Advanced Search Builder used to create the search of Hexamethonium OR Hexamethonium compounds OR ganglionic blockers with appropriate subheadings described earlier. This search yielded 151 items, including several of the articles cited as relevant by the JHU internal investigators.1 Fig. 42.6 shows these search results with some of the earlier published articles.

IV. CLINICAL RESEARCH INFRASTRUCTURE

BIOMEDICAL DATABASES

FIGURE 42.4

725

PubMed home page https://www.ncbi.nlm.nih.gov/pubmed/. Note useful links including (at top) Sign in to NCBI, Advanced; (middle) PubMed Quick Start Guide, PubMed Tutorials, MeSH Database; (bottom) Getting Started. IV. CLINICAL RESEARCH INFRASTRUCTURE

726

FIGURE 42.5

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

PubMed Advanced Search Builderdfacilitates combining previous PubMed searches using the Boolean logic operators: AND,

OR, or NOT.

One relevant article uncovered by this search is by Cockersole, “Hexamethonium lung: report of a case associated with pregnancy.”21 The PubMed record for this article is shown in Fig. 42.7 and is also available at https://www. ncbi.nlm.nih.gov/pubmed/13367924. The PubMed record for this 1956 article lacks an abstract, but includes the MeSH term “Hexamethonium Compounds/adverse effects*”; the asterisk indicates that this MeSH term is one of the main topics discussed in the article. However, this search did not find another article referred to by the JHU internal investigators, a 1954 article by Doniach, “Lung changes during hexamethonium therapy for hypertension.”.22 Although PubMed includes this article, it would be difficult to find in a toxicity search, since its record does not include an abstract or any indexing terms regarding adverse effects of hexamethonium (see Fig. 42.8). This article may be more easily found using other search engines, such as Embase.

Embase Name/producer/cost/type: Embase; Elsevier; subscription; bibliographic Sources/subjects: Covers pharmacological and biomedical research articles from more than 8500 journals in 40 languages from 1947 to the present; subject areas include medicine, nursing, drug side effects/interactions, toxicology, biotechnology, medical devices, health policy/management, alternative/complementary medicine; and public/occupational/environmental health; includes conferences records from 2009 to the present; contains more than 32.6 million records as of December, 2016 (to find the current number of records, search using a question mark); includes all articles covered by MEDLINE but with Embase indexing; The Embase Classic subfile provides coverage of articles from 1947 to 1973 and may be included in some Embase

IV. CLINICAL RESEARCH INFRASTRUCTURE

BIOMEDICAL DATABASES

727

FIGURE 42.6 PubMed search results of the search: (“hexamethonium/adverse effects” [Mesh] OR “hexamethonium/poisoning” [Mesh] OR “hexamethonium/toxicity” [Mesh]) OR (“hexamethonium compounds/adverse effects” [Mesh] OR “hexamethonium compounds/poisoning” [Mesh] OR “hexamethonium compounds/toxicity” [Mesh]) OR (“ganglionic blockers/adverse effects” [Mesh] OR “ganglionic blockers/poisoning” [Mesh] OR “ganglionic blockers/toxicity” [Mesh]) Note the possible filters or limits available listed at the left. Items number 142, “Fibrous healing of iatrogenic pulmonary edema (Hexamethonium lung) and 145 “Pulmonary changes associated with Hexamethonium therapy” were cited by Johns Hopkins University internal investigators (Becker et al.1).

subscriptions; Emtree is the hierarchically structured controlled vocabulary thesaurus used for indexing articles. Special features: Modified menu-based interface provides access to many specialized searches with selectable subjectappropriate limits or filters, including disease and drug (see drop-down menus in Fig. 42.9); the phrase search operator is single quotation marks before and after the phrase (Fig. 42.9); register to use saved searches and alerts features.

Information/support: Embase Quick User Guide http://supportcontent. elsevier.com/Support%20Hub/Embase/Files%20& %20Attachements/4559-EMBASE%20QUG.pdf Searching in Embase http://supportcontent.elsevier. com/Support%20Hub/Embase/Files%20&% 20Attachements/5414-Embase%20E102%20-% 20Searching%20in%20Embase%20-%20March% 202015.pdf Searching using Emtree http://supportcontent. elsevier.com/Support%20Hub/Embase/Files%20& Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

728

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

FIGURE 42.7 Cockersole, Hexamethonium lung: report of a case associated with pregnancydPubMed record. Record is also available at https:// www.ncbi.nlm.nih.gov/pubmed/13367924.

%20Attachements/5415-Embase%20E103%20-% 20Searching%20with%20Emtree%20-%20March% 202015.pdf Emtree comparison with MeSH http:// supportcontent.elsevier.com/Support%20Hub/ Embase/Files%20&%20Attachements/4685Embase_White%20Paper_Comparison%20of% 20Emtree%20and%20MeSH_July%202015.pdf More information about Embase https://www. elsevier.com/solutions/embase-biomedicalresearch/learn-and-support# Embase is a key clinical biomedical subscription database, produced by Elsevier, covering the literature from 1947 to the present. It includes all of the articles in MEDLINE but uses its own indexing system, Emtree; see the information capsule above for a link to a comparison with MeSH. Embase also includes more than 2700 journals

that MEDLINE/PubMed does not cover, as well as many conference materials (https://www.elsevier.com/ solutions/embase-biomedical-research/embasecoverage-and-content). Elsevier is headquartered in Europe, and Embase has more international material than MEDLINE. The Embase search interface is shown in Fig. 42.9. Users can browse the Emtree thesaurus, journals or authors, or use menu-based searches. Each of seven search types has relevant features or limits: quick, advanced, drug, disease, device, article, and PICO. An autocomplete feature suggests Emtree terms when you begin to type; if you click on a desired Emtree term you can choose to search broadly, narrowly or limit the articles retrieved to those in which the term is a major focus. PICO is an acronym for a method used to divide a research question into its components to assist in constructing search strategies. It stands for Population/

IV. CLINICAL RESEARCH INFRASTRUCTURE

729

BIOMEDICAL DATABASES

FIGURE 42.8 Doniach, Lung changes during hexamethonium therapy for hypertensiondPubMed record. Note that this record contains no indications of the adverse effects of hexamethonium. Also note that the full text of this article is available free from several locations, as indicated in the upper right of the record. This record is also available at https://www.ncbi.nlm.nih.gov/pubmed/13126364.

problem/patient; Intervention/indicator; Comparison; Outcome; Embase adds an additional field for study design, although this is not included in most PICO formats. The Embase PICO Search screen is shown in Fig. 42.10. In contrast with the PubMed record, the Embase record for Cockersole’s 1956 article “Hexamethonium lung: report of a case associated with pregnancy”21 does not include “adverse effects” but does include indexing for “edema” and “autopsy” as well as an abstract which includes “oedema”. The abstract is revealing: “.a woman of 37, treated with hexamethonium compounds.died after the birth of a live infant.Treatment with hexamethonium drugs.from the 17th week. Severe, ultimately fatal, respiratory disease developed during the 35th week.” The Embase record for Doniach’s 1954 article, “Lung changes during hexamethonium therapy for hypertension”22

does not include “toxicity” or “adverse effects” but does include an abstract and the terms “dyspnea” and “edema”, as well as their British spellings.

Cited Reference Databases Web of Science Name/producer/cost/type: Web of Science; Clarivate Analytics (formerly the IP & Science business of Thomson Reuters) Sources/subjects: Citations to journal articles, reviews, and editorials from more than 12,000 multidisciplinary journals, books and conference proceedings; covers the sciences, agriculture, social sciences, arts, and humanities from 1900 to the present. Institutional subscriptions may not include all subjects and all years. See the Web of Science Quick Reference Guide for full information.23 Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

730

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

FIGURE 42.9

Embase search interface. Note the search tips on the right; the “heart attack” phrase search example below is indicated with single quotation marks not with double quotation marks; “Help,” “Webinars,” and “Guides and videos” links are available at the bottom of the search screen; and drop-down menus are shown at the bottom of this figure. Note: Please note the first two images are the top and the bottom of the Embase search screen, and below that are the drop-down menus expanded from the top of the search screen.

IV. CLINICAL RESEARCH INFRASTRUCTURE

BIOMEDICAL DATABASES

FIGURE 42.10

Special features: Bibliographic citation database with complete citation counts for all records; finds cocitations, articles which cite a common article, and so finds articles which may be related by subject matter via citation analysis without searching for subject terms; includes institutional affiliations for all authors, grant numbers, and funding agencies. Information/support: http://ipscience.thomsonreuters.com/product/ web-of-science/; Web of Science Quick Reference Guide http://wokinfo.com/media/pdf/qrc/webofscience_ qrc_en.pdf; recorded training http://wokinfo.com/ training_support/training/web-of-science/

Scopus Name/producer/cost/type: Scopus; Elsevier; subscription; bibliographic Sources/subjects: Citations to journal articles from more than 21,500 multidisciplinary journals; trade publications, books, and patents; covers the sciences, technology, social sciences, arts, and humanities from 1823 to the present; see the Scopus content information for complete details.24

731

Embase PICO search.

Special features: Citation database with complete citation counts for all records from 1996 to the present; finds cocitations, articles which cite a common article, and so finds articles which may be related by subject matter via citation analysis without searching for subject terms; includes authors’ affiliations and conference proceedings; covers all of the journals indexed by MEDLINE and Embase25 as well as many of the sources covered by Ei Compendex, but may not include the specialized indexing or keywords from those files. Scopus includes Emtree and MeSH indexing in its display of the record for each citation; one can open the “advanced” search function and choose the field code for “Index” to retrieve citations using MeSH or “Index Terms” for all subject indexing assigned by MEDLINE, Embase and Ei Compendex. Information/support: https://www.elsevier.com/solutions/scopus; http://help.scopus.com and http://help.elsevier. com/app/answers/list/p/8150/c/0 Both Web of Science and Scopus are useful for their coverage of the conference literature and for obtaining citation rankings of authors’ publications. Though there Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

732

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

is overlap between these two databases, searches of both Web of Science and Scopus retrieve unique items. Both are fee-based subscription databases. Scopus is generally broader in scope and has more abstracts but only has complete citation counts from 1996 to the present. “Complete citation counts” means that Scopus only includes all the citations that were published after 1995 to a given earlier article. Web of Science has complete citation counts from 1900 to the present but covers fewer journals.

Clinical Decision Resources Micromedex Name/producer/cost/type: Micromedex; Truven Health Analytics; subscription; drug monographs, drug information, specialty databases, and reference books Sources/subjects: Drug, toxicology, and environmental health information; summaries and detailed monographs for drugs marketed worldwide; names, actions, indications, dosing, toxicology, drug interactions, comparisons, product lookup, and adverse effects. Special features: Available resources include Martindale: The Complete Drug Reference, Index Nominum, Physician’s Desk Reference (PDR), Detailed Drug Information for the Consumer, POISINDEX, DRUGDEX, AltMedDex, DISEASEDEX, REPRORISK; resources may be purchased in various combinations; the information is integrated and may not be available in separate databases; mobile access is available (Huber and Swogger7, pp. 239e240). Information/support: https://micromedex.com/about-micromedex; “Help” in upper right cornerdlinks to Training, Search Tips, Results, Tools, etc. Micromedex “had lung toxicity as the first adverse effect of hexamethonium” according to the Johns Hopkins internal investigation.1 Micromedex’s drug profiles can be mined for comparative drug information, i.e., hypoglycemic data for antidiabetes drugs as compared with Metformin. POISINDEX is an essential resource for US Poison Control Centers; Index Nominum is an international drug directory; and PDR is an annual compilation of FDAapproved drug package inserts.

Lexicomp Online Name/producer/cost/type: Lexicomp Online/ American Hospital Formulary System (AHFS)/Facts & Comparisons information;

Wolters Kluwer N.V.; subscription; drug information resources, databases, decision support modules, monographs, specialty databases, and reference books. Sources/subjects: Point-of-care drug information; summaries and details including dosing, administration, warnings, precautions, clinical guidelines, IV compatibility, toxicology, and comparative information. Special features: Includes detailed drug monographs from the AHFS, produced by the American Society of Health-System Pharmacists (ASHP), drug information from Facts & Comparisons, and natural products, pregnancy, neonatal, and lactation drug information. Information/support: http://www.wolterskluwercdi.com/lexicomponline; “Help” in upper right cornerdlinks to Training, Search Tips, Results, and Tools; Details at http://www. wolterskluwercdi.com/lexicomp-online/databases/; User Guide at http://online.lexi.com/lco/help/lcoug.pdf American Hospital Formulary System (AHFS) is an authoritative source of drug information and evidencesupported recommendations; other authoritative sources are also included in Lexicomp Online.

Facts & Comparisons eAnswers Name/producer/cost/type: Facts & Comparisons eAnswers; Wolters Kluwer N.V.; subscription; drug information resources, databases, and monographs. Sources/subjects: Drug information, pharmacology, administration and dosage, interactions, adverse reactions, off-label uses, drug identifier, and natural products information. Special features: A to Z Drug Facts includes drug monographs and information for combination drugs; also information for patient education, dosing and clinical calculators, and comparative data tables. Information/support: http://www.wolterskluwercdi.com/factscomparisons-online/; quick start guide, tutorials available

UpToDate Name/producer/cost/type: UpToDate; Wolters Kluwer N.V.; subscription; review articles on many clinical topics and drug information and databases.

733

BIOMEDICAL DATABASES

Sources/subjects: Evidence-based point-of-care clinical decision support resources; etiology, diagnosis, and treatment questions; peer-reviewed text includes drug information with medical calculators, drug interaction, and patient information; shares some resources with Lexicomp. Special features: References can link to an institution’s full-text subscriptions. Information/support: http://www.uptodate.com/home/help; quick start and reference guide, user manual, and FAQ available

Drug Development Information Pharmaprojects Name/producer/cost/type: Pharmaprojects; Citeline/Pharma intelligence/ Informa PLC; subscription; detailed drug profiles. Sources/subjects: Searchable profiles of drugs in development worldwide since 1980 include disease, therapy, mechanism, biological targets, delivery routes, development and marketing histories, and clinical trial information; sources include clinicaltrials.gov, technical meetings and conferences, and business/ investment sources. Special features: Profiles include extensive lists of drug names, laboratory compound names, and licensing information; help from analysts is readily available. Pharmaprojects is useful for competitive intelligence and technology landscape research. Its drug profiles can be mined for information about clinical trials beyond those included in clinicaltrials.gov. Information/support: FAQ Quick Start https://nih-pipeline.citeline.com/ Documents/Quick%20Start.pdf Many other subscription databases cover drug development and regulatory information, including Adis Insight http://adis.springer.com/, Biomedtracker https://www. biomedtracker.com/, BioPharm Insight http://www. biopharminsight.com/, Medtrack https://pharma intelligence.informa.com/products-and-services/data-andanalysis/medtrack, and Thomson Reuters’ Cortellis and Integrity (both now from Clarivate Analytics) http:// lifesciences.thomsonreuters.com/training/integrity. A wide variety of drug regulatory information is available free from the US Food and Drug Administration (FDA). Information about the drug information section of the FDA website is at https://www.fda.gov/Drugs/ default.htm with links to approval, safety, development,

and regulatory information. A list of the FDA Drug Approvals and Databases is available at https://www.fda. gov/Drugs/InformationOnDrugs/default.htm, including the following: • Orange Book: Approved Drug Products with Therapeutic Equivalence Evaluations, http://www. accessdata.fda.gov/scripts/cder/ob/dfind approved drugs by name, active ingredient, applicant company, dosage form, route of administration, or patent number; additional information at https://www.fda. gov/Drugs/InformationOnDrugs/ucm129662.htm. • Drugs@FDA: FDA Approved Drug Products www.fda. gov/drugsatfdadFDA-approved drugs from 1939 to the present, with full information available for drugs approved since 1998; search by drug name or active ingredient or browse by drug name; additional information available at https://www.fda.gov/Drugs/ InformationOnDrugs/ucm135821.htm, and Drugs@ FDA Data Files are available at http://www.fda.gov/ Drugs/InformationOnDrugs/ucm079750.htm. • MedWatch: The FDA Safety Information and Adverse Event Reporting Program, www.fda.gov/medwatch/ index.html, provides postmarketing surveillance from pharmaceutical companies as required by FDA; emailed Medwatch Safety Alerts are available. Other regulatory agencies may also have useful information including the European Medicines Agency (EMA), and the National Institute for Health and Care Excellence (NICE, https://www.nice.org.uk/) of the United Kingdom, which permits access to journals and for NHS (National Health Service) users https://www. nice.org.uk/about/what-we-do/evidence-services/ journals-and-databases. Researchers interested in searching for drug information should remember to verify drug names and synonyms, possibly using ChemIDplus available at ToxNet (see below), and to also search other databases mentioned elsewhere in this chapter, including Embase. The National Library of Medicine has a directory of drug information resources at https://www.nlm.nih.gov/services/Subject_Guides/ druginformation/ with lists of United States and international drug information resources at https://www.nlm. nih.gov/services/Subject_Guides/druginformation/ unitedstatesdrugresources/ and https://www.nlm.nih. gov/services/Subject_Guides/druginformation/ internationaldrugresources/.

Clinical Trials Resources Cochrane Reviews Name/producer/cost/type: Cochrane reviews (Cochrane Database of Systematic Reviews, CDSR) www.thecochranelibrary.com/ Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

734

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

Cochrane Collaboration; subscription; reviews, bibliographic data, extensive abstracts, other information. Sources/subjects: Evidence-based systematic reviews of research in medicine and health policy covering the effectiveness of health interventions and the accuracy of diagnostic tests; the Cochrane Library also offers other databases: Cochrane Central Register of Controlled Trials (CENTRAL) covering published and unpublished clinical trials, Database of Abstracts of Reviews of Effects (DARE), Cochrane Methodology Register (CMR), Health Technology Assessment Database (HTA), National Health Service (NHS), and Economic Evaluation Database (EED) (Huber and Swogger7, pp. 106, 171). Special features: Cochrane reviews are comprehensive, updated regularly, and include extensive abstracts, which are available free in MEDLINE/PubMed. Some reviews are available as open access immediately on publication, and all Cochrane reviews published after February, 2013, are available freely 12 months after publication (http://www.cochranelibrary.com/help/open-accessoptions-for-the-cochrane-library.html). Some countries have free or low-cost access to the Cochrane Library through HINARI or national programs http://www.cochranelibrary.com/help/ access-options-for-cochrane-library.html. The Cochrane Collaboration is a unique network of independent contributors organized by Cochrane groups http://onlinelibrary.wiley.com/o/cochrane/ cochrane_clabout_contents_fs.html. Cochrane reviews are seen as achieving the highest standards in evidence-based medicine. Information/support: Cochrane information http://www.cochrane.org/ about-us; Cochrane reviews information http://www. cochranelibrary.com/about/about-cochranesystematic-reviews.html; CENTRAL information http://www.cochranelibrary.com/about/centrallanding-page.html; Training in how to use the Cochrane Library http://training.cochrane.org/ resource/how-use-cochrane-library; “Navigating the healthcare information jungle,” a general introduction http://training.cochrane.org/resource/navigatinghealthcare-information-jungle; Cochrane also has extensive additional training available in many aspects of the Cochrane Collaboration and evidence-based medicine, including learning resources, pathways, courses, and handbooks http://training.cochrane.org/

An article by Katie Thomas in the June 29, 2013, issue of the New York Times26 reports on research done by Dr. Peter Doshi, then a postdoctoral student at the Johns Hopkins University and Dr. Tom Jefferson, a British epidemiologist based in Rome, who used the Cochrane Library in 2009 to update earlier reviews on the effectiveness of Tamiflu, a Roche product. Governments and businesses around the world were stockpiling the drug on the assumption it could be used to ameliorate the effects of an influenza pandemic; as much as 60% of the drug’s $3 billion in 2009 sales were to these stockpiles. A third researcher, Dr. Keiji Hayashi, had contacted Dr. Jefferson, concerned that Tamiflu was not as effective as reported in his pediatric practice. That contact led to a finding that several of the trials on which the earlier review and recommendations had been based had not actually been published in full. A cascade followed: The 2009 Cochrane review could not be supported by the new evidence; the BMJ published its own investigation into the results of those “missing” trials27; the European Medicines Agency became involved. The Cochrane Library published a new review in 2014 on neuraminidase inhibitor drugs28 for preventing and treating influenza in adults and children, based on 107 clinical study reports of nearly 25,000 patients; its conclusions highlighted a finding that inadequate reporting of trials did not disclose that most of the Relenza and Tamiflu studies had a high risk of selection bias. The Cochrane Library is produced by the Cochrane Collaboration, an international organization of both paid and volunteer researchers who investigate the effectiveness of medical interventions on the basis of exhaustive systematic searches and appraisals of the published medical literature, identifying human clinical trials, evaluating their design and results, combining results in systematic reviews and metaanalyses. Cochrane reviews are indexed in MEDLINE and other databases; the reviews are available through the Cochrane Library, offered by most medical libraries as part of their subscription databases, and in some countries via special funding as a “national provision;” many low and moderate-income countries have free access through the HINARI program, see table at http://www.cochranelibrary.com/help/access-optionsfor-cochrane-library.html. In addition to the access described above, authors can choose the “gold” open access model to pay for universal free access to their reviews immediately on publication; all reviews published on or after February 1, 2013, are available freely 1 year after publication. The reviews are identified in the database with the symbols

for gold open access, and

available 1 year after publication.

IV. CLINICAL RESEARCH INFRASTRUCTURE

for those

735

BIOMEDICAL DATABASES

Cochrane encourages its authors to publish modified versions of their reviews in the peer-reviewed biomedical literature.29 These copublications are cited an average of 3.5 times as frequently as the complete Cochrane reviews. Why search for and use Cochrane reviews? Some clinical trialists recommend that preparation of a protocol for a human trial should involve the same attention to the literature search and review that is standard for Cochrane reviews30: identification of a research question, careful consideration of information resources including multiple bibliographic databases, unpublished or “grey” literature and conference abstracts, and citation lists in publications reviewed; the searches themselves must be comprehensive, constructed with attention to the strengths and limitations of each resource using both natural language terms and structured taxonomies. See Fig. 42.11 for an example of a recent systematic review’s data sources and search methodology. The Cochrane Collaboration’s origin, based in a collegial response to Archie Cochrane’s frustration with the difficulty in identifying and evaluating the most appropriate clinical trials relevant to a particular patient care question, has led to an international organization of both paid staff and volunteer contributors who research and publish not only systematic reviews but also investigate methods of clinical research, dissemination and uptake of clinical research findings, and literature search methodologies. What else does the Cochrane Collaboration offer? Central: a database of trial reports including not only those published in the peer-reviewed literature but also conference abstracts, notices of trial recruitment, and personal communications. The latter categories are identified in several ways, including “hand searching” by specially trained Cochrane staff and volunteers. These reports can be important in understanding findings not in the peerreviewed literature, sometimes because the results were negative,31 or because a trial was not completed or analyzed as originally planned. Evaluating clinical trial reports and systematic reviews: Standards for trials continue to be developed. Of interest are several: CONSORT http://www.consort-statement. org includes a graphic representation accounting for all recruited trial subjects; and PRISMA standards (http:// www.prisma-statement.org) describe best practices for reporting systematic review research. Registries of clinical trials include: • clinicaltrials.gov, www.clinicaltrials.govdworldwide clinical trials registry and results database maintained by NIH, but not limited to NIH-funded research; for information about searching see https://clinicaltrials. gov/ct2/search/index • International Clinical Trials Registry Platform (ICTRP), http://www.who.int/ictrp/en/dapproximately 20 clinical trials registries maintained by the WHO; FAQ available at http://www.who.int/ictrp/faq/en/.

Evidence-Based Medicine and Systematic Reviews In 1996, David Sackett and colleagues, noting the “ancient” philosophical origins of evidence-based medicine in 19th century Paris, proposed a simple definition: “Evidence based medicine is the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients”.32 That definition explicitly addresses the elements of clinical expertise, medical research, and the situation and desires of the patient. Clinical research both incorporates and contributes to the evidence base. This section addresses identification of commonly available sources of medical research pertinent to an “evidence-based” clinical research protocol, as well as the role of such research in clinical expertise. The “best evidence” is a multipart composite. Consideration of the PICO question format (Patient/condition/ Problem; Intervention being considered or studied; Comparison intervention; Outcome) reminds us that there are several moving parts involved in a clinical decision point; optimal retrieval would find the exact set of those four elements addressed in one or more well-designed randomized controlled trials, or even better, a systematic review and meta-analysis examining those trials. Finding the best evidence requires consideration of all the elements with the realistic understanding that there may be a possible compromise on some. Looking for the “best evidence” might include sifting upward on the evidence pyramid from studies in vitro and in animal models, to case reports and case series, to cross sectional or case-control studies, and finally to randomized controlled clinical trials. Depending on the researcher’s familiarity with the clinical question, the search might start with a textbook or clinical decision support tool for grounding. The latter category includes such tools as DynaMed, ClinicalKey and UpToDate (see information on the latter elsewhere in this chapter). All of them incorporate reviews by physicians of the current literature on clinical questions. They provide added value by incorporating such features as selective full-text access to the journal literature on which the reviewers base their analysis, referral to guidelines of medical societies, patient education aids, or graphic material with explicit permissions to reproduce in slide presentations. Watch for currency of the review or last update of the search on which the recommendations are based and for explicit use of a grading system (“good evidence, discretionary recommendation,” or “level 1 [likely reliable] evidence” or “good evidence, strong recommendation”dall from DynaMed; “Grade 1A” or “Grade 2B” dall from UpToDate, using the GRADE criteria: http://www. gradeworkinggroup.org). These tools are online resources, requiring subscriptions for access to the content, supported Continued

736

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

Databases searched: “We conducted a systematic search of bibliographic databases: MEDLINE, Embase, CINAHL, SCOPUS, LILACS, ProQuest Dissertations and Theses Global and evidence-based medicine (EBM) reviews sources. These included: Cochrane Database of Systematic Reviews (2005 to July 2015), ACP Journal Club (1991 to July 2015), Database of Abstracts of Reviews of Effects (DARE) (second quarter 2015), Cochrane Central Register of Controlled Trials (CENTRAL) (June 2015), Cochrane Methodology Register (third quarter 2012), Health Technology Assessment (second quarter 2015), and NHS Economic Evaluation Database (second quarter 2015). This search spanned from 1946 to July 2015. We also searched the Cochrane Airways Group register of trials which was most recently conducted on July 2016.”

MEDLINE search strategy [strategies for all databases searched is available in the review]: Database: Ovid MEDLINE In-Process & Other Non-Indexed Citations, Ovid MEDLINE Daily and Ovid MEDLINE, 1946 to July 17, 2015 Search strategy: -------------------------------------------------------------------------------1. exp asthma/ 2. asthma*.mp. 3. 1 or 2 4. exp Emergency Service, Hospital/ or (acute or relaps* or exacerbat*).ti,ab. 5. (emergency adj3 (room* or ward or wards or department* or doctor* or nurse* or clincian* or practitioner*)).mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier] 6. ("critical care" or "acute care").mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier] 7. 4 or 5 or 6 8. 3 and 7 9. anticholinergic*.mp. 10. (ipratropium or atrovent or oxitropium or oxivent).mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier] 11. exp Ipratropium/ 12. cholinergic.mp. or exp Cholinergic Agents/ 13. PARASYMPATHOMIMETICS.mp. or exp Parasympathomimetics/ 14. limit 13 to yr="1975 - 1994" 15. 9 or 10 or 11 or 12 or 14 16. 8 and 15 17. salbutamol.mp. or exp Albuterol/ 18. ("levalbuterol hydrochloride" or sultanol or albuterol or "2-t-butylamino-1-(4-hydroxy-3-hydroxy-3hydroxymethyl)phenylethanol" or ventolin or "levosalbutamol hydrochloride" or proventil or "hydrochloride levalbuterol" or "xopenex levalbuterol").mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier] 19. exp Adrenergic beta-2 Receptor Agonists/ 20. 17 or 18 or 19 21. 16 and 20 (336) 22. (combivent or berodual).mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier] 23. 16 and 22 24. 21 or 23 25. limit 24 to "all child (0 to 18 years)" 26. limit 25 to "all adult (19 plus years)" 27. 24 not 25 28. 26 or 27

FIGURE 42.11

Regarding Cochrane methodologies, with added explanatory notes enclosed in square brackets. From Kirkland SW, Vandenberghe C, Voaklander B, Nikel T, Campbell S, Rowe BH. Combined inhaled beta-agonist and anticholinergic agents for emergency management in adults with asthma. Cochrane Database of Systematic Reviews January 11, 2017;1:CD001284. http://dx.doi.org/10.1002/14651858.CD001284.pub2.

737

BIOMEDICAL DATABASES

in most circumstances by institutional subscriptions provided by a library or hospital. Similarly, textbooks should be evaluated not only by the edition number or date of publication, but a glance at the references at the end of chapters allows a basic evaluation of the currency of the material. Access via PubMed to journal literature describing the results of clinical trials can be structured to take advantage of multiple aides built into the system. As noted elsewhere in this chapter, the “Clinical Queries” search, clickable on the main page of PubMed, takes the researcher into an environment in which searches on a subject, using title or abstract terms, or MeSH, are retrieved using a background search for clinical studies and further filtered for etiology/diagnosis/therapy/prognosis/clinical prediction. Each of these may be modified for broad or narrow (sensitivity or specificity) retrieval; the most recent studies appear first, and searches within the clinical studies subset may also be done using an author name (e.g., Fauci AS [an]) or even a research group or institution (e.g., National Cancer Institute[ad]). The standard PubMed search can be narrowed by the filters which appear on the left side of the screen after the first terms are entered in the search box; appropriate article types include “Clinical Trial, Phase I/II/III/IV,” “Pragmatic Clinical Trial,” and “Randomized Controlled Trial.” The search can then be refined for patient population using filters for sex and ages. Systematic reviews provide a shortcut to the task of identifying evidence-based research and add the expertise of the review authors in appraising the available research. While there are widely supported standards for conducting and reporting clinical trial research, such as the CONSORT, PRISMA, and standards for other study designs, such as MOOSE, for observational studies (https://www.editorialmanager.com/jognn/account/ MOOSE.pdf), many publications with “systematic review” in the title or abstract do not report the result of systematic searches and appraisal. Cochrane systematic reviews are described elsewhere in this chapter and continue to fill an important niche in supporting, producing, and disseminating high-quality systematic reviews. In the social sciences, the Campbell Collaboration (campbellcollaboration.org) has a similar mission and will be worth searching for diverse research projects, including drug abuse, education, and nutrition. Resources for evidence-based medicine include the following: • Guidelines.gov, www.guidelines.govdfrequently updated evidence-based guideline clearinghouse from the Agency for Healthcare Research & Quality (AHRQ), part of the US Department of Health and Human Services; it includes extensive filtering capabilities and the ability to compare several guidelines side-by-side.

• PubMed Health, http://www.ncbi.nlm.nih.gov/ PubMedhealth/dclinical effectiveness research; includes systematic reviews and research on effectiveness methods, and information from AHRQ, Cochrane, and other sources; more information is available at http://www.ncbi.nlm.nih.gov/ PubMedhealth/about/.

Textbooks and Reference Works Both current and older editions of textbooks and reference works can be important resources in clinical research. The Johns Hopkins internal investigation report1 stated that although hexamethonium toxicity was not mentioned in the last four editions of Goodman and Gilman’s key textbook (The Pharmacological Basis of Therapeutics, current edition edited by Brunton, Chabner, and Knollman33) or in the current edition of Fishman’s text on pulmonary medicine (Fishman’s Pulmonary Diseases and Disorders, current edition by Grippi et al.34), it was included in a previous edition of Fishman’s text. NCBI Bookshelf described below is an important free full-text resource for texts and reference works. Another resource is Google Books https://books.google.com/ described in the “Internet Search Engines” section.

NCBI Bookshelf Name/producer/cost/type: NCBI Bookshelf http://www.ncbi.nlm.nih.gov/ books; National Center for Biotechnology Information (NCBI), National Library of Medicine (NLM), National Institutes of Health (NIH); free full-text books and some booklike material. Sources/subjects: Full-text searchable books including biomedical textbooks, scientific titles, and NCBI help manuals. Special features: Many books are browsable via hyperlinked tables of contents. Information/support: Help http://www.ncbi.nlm.nih.gov/books/ NBK3833/; quick start http://www.ncbi.nlm.nih.gov/ books/NBK45613/; FAQ http://www.ncbi.nlm.nih. gov/books/NBK45610/; tutorials http://www.ncbi. nlm.nih.gov/books/NBK45612/

Full-text Searching Researchers can search the full text of journal articles to find information that is not available through any other means, such as details about experimental techniques, methods, instrumentation, and materials, and subjects covered in review articles. These details would usually not be included in abstracts, keywords, or added indexing terms. Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

738

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

For example, recall the previously mentioned review article on drug-induced pulmonary disease14 found by Johns Hopkins University internal investigators. Hexamethonium was only mentioned in two paragraphs out of 14 total pages, and no indexing of this article would have included all of the substances covered. However, it is possible that this review article could have been discovered if its full text was searched.

PubMed Central Name/producer/cost/type: PubMed Central (PMC) http://www.ncbi.nlm.nih. gov/pmc/; National Center for Biotechnology Information (NCBI), National Library of Medicine (NLM), National Institutes of Health (NIH); free fulltext journal articles. Sources/subjects: Repository of more than 4 million full-text searchable biomedical and life sciences journal articles, deposited under several parallel programs: deposit of research papers mandated by several funders’ public access policies, including the National Institutes of Health (see https://publicaccess.nih.gov), Wellcome Trust, and the Bill and Melinda Gates Foundation; archives of 300þ journals depositing all of their NIHfunded papers from a date preceding the NIH mandate; nearly 2000 additional journals archive complete or partial content in PMC to provide permanent access. Journal list and deposit details at https://www.ncbi.nlm.nih.gov/pmc/journals/. Special features: Provides data for PubMed cited reference searching; find related data in other NCBI files; some articles in the Open Access subset are available for bulk downloads http://www.ncbi.nlm.nih.gov/pmc/ tools/openftlist/; use My NCBI for saved searches and automatic updates http://www.ncbi.nlm.nih.gov/ books/NBK53592/; search can be filtered to limit to articles published under a specific funding mandate. Information/support: Help http://www.ncbi.nlm.nih.gov/books/ NBK3825/; FAQ http://www.ncbi.nlm.nih.gov/pmc/ about/faq/; public access http://www.ncbi.nlm.nih. gov/pmc/about/public-access/; PMC International http://www.ncbi.nlm.nih.gov/pmc/about/pmci/ describes Europe PMC and Canada PMC

Direct Searching of Journal Articles Individual journals and publishers’ collections of journals may allow full-text searching; access to the content through open access, subscription, or pay-per-view programs. The Directory of Open Access Journals, http://

www.doaj.org, evaluates journals for quality standards and provides full-text searching of nearly 3 million articles available freely. Other full text journal article resources include the following: • ScienceDirect, www.sciencedirect.com, a full text resource from Elsevier, with journal articles, book chapters, and open access content; • Google Scholar http://scholar.google.comdsee details in Internet Search Engines section; and • Microsoft Academic https://academic.microsoft.com/ dsee details in Internet Search Engines section.

Specialized Clinical Resources ToxNet Name/producer/cost/type: TOXNET (Toxicology Data Network) http://toxnet. nlm.nih.gov; National Library of Medicine (NLM), National Institutes of Health (NIH), US Department of Health and Human Services (HHS); free; data, chemical, and bibliographic information. Sources/subjects: Fourteen databases covering chemistry, toxicology data, toxicology literature, and toxics release information can be searched separately or together. Special features: Two TOXNET databases of particular interest are TOXLINE (Toxicology Literature Online) https:// toxnet.nlm.nih.gov/newtoxnet/toxline.htm, which includes toxicology material from MEDLINE/PubMed and other sources; and ChemIDplus https://chem.sis. nlm.nih.gov/chemidplus/chemidlite.jsp, a chemical dictionary with more than 400,000 records, including synonyms, CAS Registry Number, molecular formulae, and some physical properties. More than 3000,000 records are structure-searchable. Information/support: TOXNET FAQ http://www.nlm.nih.gov/pubs/ factsheets/toxnetfs.html. TOXNET manual https://sis.nlm.nih.gov/enviro/ toxnet_manual.pdf. TOXLINE Fact Sheet https://www.nlm.nih.gov/ pubs/factsheets/toxlinfs.html; TOXLINE help https://toxnet.nlm.nih.gov/ newtoxnet/dart-toxlinehelp.html; ChemIDplus Fact Sheet https://www.nlm.nih.gov/ pubs/factsheets/chemidplusfs.html; ChemIDplus Help https://chem.sis.nlm.nih.gov/ chemidplus/jsp/chemidheavy/help. jsp#LiteSearch

IV. CLINICAL RESEARCH INFRASTRUCTURE

BIOMEDICAL DATABASES

When starting a drug search, verifying drug names and synonyms in ChemIDplus can save a lot of time and effort.

PsycINFO Name/producer/cost/type: PsycINFO; American Psychological Association; subscription; bibliographic records. Sources/subjects: Psychology, behavioral, health and social sciences; covering almost 2500 journals (80% of records), books, book chapters, and dissertations from 1597 to the present with comprehensive coverage from the 1880s to the present. Special features: Controlled vocabulary from the Thesaurus of Psychological Index Terms; detailed publication type, document type, and methodology indexing enables deep searching; citation searching available. Information/support: Overview http://www.apa.org/pubs/databases/ psycinfo/; Details http://www.apa.org/pubs/databases/ psycinfo/?tab¼3; Training http://www.apa.org/pubs/databases/ psycinfo/index.aspx?tab¼4; Tips http://www.apa.org/science/about/psa/ 2013/10/using-psycinfo.aspx Although PsycINFO has a substantial overlap with MEDLINE/PubMed, its additional indexing can be very effective in obtaining relevant items. For example, PsycINFO’s indexed age groups are more detailed than those used in MEDLINE/PubMed, including infancy, preschool age, school age, childhood, adolescence, young adulthood, adulthood, thirties, middle age, aged, and very old. PsycINFO enables comprehensive cited reference searching beginning in 2001 with some partial previous coverage (details at http://www.apa.org/pubs/ databases/psycinfo/cited-references.aspx).

CINAHL Name/producer/cost/type: CINAHL (Cumulative Index to Nursing and Allied Health Literature); EBSCO Information Services; subscription; bibliographic entries and full-text journals. Sources/subjects: Nursing, healthcare and allied health subjects, including physical therapy, nutrition, and athletic training; indexes >4000 journals, includes 70 full-text journals and searchable cited references for more than 1300 journals.

739

Special features: CINAHL subject headings are modeled after MeSH; coverage includes books, audiovisual materials, dissertations, and research instruments; document types include care plans, accreditation, and clinical innovations (Huber and Swogger7, p. 101). Information/support: https://health.ebsco.com/products/the-cinahldatabase; training http://support.epnet.com/cinahl/ training.php CINAHL covers a significant number of peer-reviewed and other sources that are not covered by MEDLINE or EMBASE/Scopus.25 Versions of CINAHL are available, which cover additional journals, years of coverage back to 1937, and access to full-text sources; for details, see https://www.ebscohost.com/promoMaterials/ CINAHL_Comparison_Chart_0816.pdf.

Additional Specialized Resources Many other information resources can be useful in clinical research beyond the strictly clinical information resources. RePORTER, detailed below, lists governmentfunded research and so can yield information about research efforts and programs before any results are published. Dissertations are an often overlooked resource; patents, engineering and scientific information can be useful with regard to medical devices and broad areas of research.

RePORTER Name/producer/cost/type: RePORTER (Research Portfolio Online Reporting Tools Expenditures and Results), projectreporter.nih. gov/reporter.cfm; National Institutes of Health (NIH), US Department of Health and Human Services (HHS); free; information on research funded by several government agencies. Sources/subjects: Reports on research funded by NIH, CDC, AHRQ, VA, and other federal agencies for the most recent 25 fiscal years. Special features: Includes listings of some publications and patents arising from funded research; MY RePORTER (free registration available) facilitates saving query criteria and creation of research portfolios and hit lists. Information/support: Manual https://projectreporter.nih.gov/ RePORTER_Manual_files/RePORTERManual.pdf; FAQ https://report.nih.gov/faq.aspx; Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

740

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

My RePORTER tutorial https://report.nih.gov/ tutorial/my_reporter.aspx RePORTER can be used to search for US government HHS- and VA-funded biomedical research from a rolling window of the 25 previous fiscal years. It replaced the older CRISP (Computer Retrieval of Information on Scientific Projects) system; however, downloadable data from the CRISP and from the ExPORTER systems are available at https://exporter.nih.gov/default.aspx. RePORTER is very useful to find out who is researching what subjects or using which techniques. It has many options for fields and limits; for example, searches can differentiate between types of intramural and extramural grants (https://report.nih.gov/catalog.aspx). Details discovered from RePORTER can be used for further information searches in other resources; e.g., a search in Web of Science for publications resulting from specific NIH grants. Grey literature is information from sources other than commercial or academic publishing, including dissertations, research reports, conference papers, and regulatory information, according to GreyNet, the Grey Literature Network Service http://www.greynet.org/. Some general grey literature resources include the following: • GreySource http://www.greynet.org/ greysourceindex.htmldexamples of grey literature with profiles of producing organizations • Grey Literature Report http://www.greylit.org/ publishers/listdgrey literature publishers list from the New York Academy of Medicine • Grey Matters: a practical tool for searching healthrelated grey literature, https://www.cadth.ca/ resources/finding-evidence/greymattersddownloadable checklist of searchable sites with live links organized by topic from the Canadian Agency for Drugs and Technologies in Health (CADTH); also includes links to search strategies and search filters • OpenGrey http://www.opengrey.eu/dbibliographic citations of multidisciplinary European grey literature • Google Scholar http://scholar.google.comdsee details in Internet Search Engines section Dissertations can be very helpful in research with indepth information often unavailable elsewhere. Some dissertation search resources include the following: • ProQuest Dissertations & Theses Database http://www. proquest.com/products-services/pqdt. htmldsubscription-based resource covers from 1743 to present with some full-text items since 1990

• Networked Digital Library of Theses and Dissertations http://www.ndltd.org/resources/find-etdsdlinks to many searchable sources with new Global ETD (electronic thesis or dissertation) Search http://search. ndltd.org/ • Open Access Theses and Dissertations database https:// oatd.orgdcovers material from 1880 to the present with more than 3 million titles, many linked for open access • OpenDOAR, the Directory of Open Access Repositories http://www.opendoar.org/index.htmldacademic open access repositories • OpenThesis http://www.openthesis.org/drepository of theses and dissertations; see more information at http://www.openthesis.org/faq.html • BASE, Bielefeld Academic Search Engine http://www. base-search.net/dacademic web resources from Bielefeld University in Germany, available in English, German, and Chinese; it is possible to refine search results by document type, theses; more information is available at https://www.base-search.net/about/en/ faq.php Resources useful for medical device research include the following: • US FDA Medical Device Databases list, available at http://www.fda.gov/MedicalDevices/ DeviceRegulationandGuidance/Databases/ • Ei Compendex https://www.elsevier.com/solutions/ engineering-village/content/ compendexdsubscription-based engineering literature database from Elsevier containing more than 20 million bibliographic records with indexing terms from the Engineering Index Thesaurus • Inspec http://www.theiet.org/resources/inspec/ dsubscription-based engineering literature database from The Institution of Engineering and Technology containing more than 16 million bibliographic records with specialized indexing Patents can be a useful source of clinical research information with details about state-of-the-art and forthcoming devices, drugs, formulations, tests, and standards, which are not available elsewhere. A good resource about what is in a US patent record is “How to read a patent,” http://www.lens.org/support/help-resources/thebasics/how-to-read-a-patent/, from free patent search site The Lens https://www.lens.org. Patent documents can include published applications and granted patents depending on the countries involved; a list of the codes used for the patents of different countries is available at the US Patent and Trademark Office (USPTO) site https://www.uspto.gov/patentsapplication-process/applying-online/country-codes-

IV. CLINICAL RESEARCH INFRASTRUCTURE

741

BIOMEDICAL DATABASES

wipo-st3-table. Patents for the same subject material in different countries can be called “patent equivalents” and organized into patent families as explained at the European Patent Organization (EPO) site at http://www. epo.org/searching-for-patents/helpful-resources/firsttime-here/patent-families.html. Patents are categorized by technology according to several classification schemes, including the US Patent Classification (USPC) and Cooperative Patent Classification (CPC), explained at https://www.uspto.gov/ patents-application-process/patent-search/classificationstandards-and-development. The web page “General Searching Best Practices”, from Landon IP, at http://web. archive.org/web/20160515054221/http://www.intellogist. com/wiki/General_Searching_Best_Practices offers an explanation of a general approach to searching patents by subject. More detailed information is available from texts such as Patent Searching: tools & techniques.35 A list of the patent offices of different countries is available at https://en.wikipedia.org/wiki/List_of_patent_ offices. Some free patent search sites include the following: • The Lens http://www.lens.org/lens/ with more information available at http://www.lens.org/about/ what/ and help at http://www.lens.org/support/ • FreePatentsOnline www.freepatentsonline.com with help available at http://research.freepatentsonline. com/help • USPTO http://patft.uspto.gov/dUS only patents and applications only; help available at http://patft.uspto. gov/help/help.htm • Google Patents http://www.google.com/patentsdsee more information in the Internet Search Engines section • Espacenet http://worldwide.espacenet.com/dfrom the EPO, http://www.epo.org/; searchable in English, French, or German. Many commercial vendors of patent information provide useful features to enable better searching. Some of these are briefly summarized here; however, there are frequent changes in this competitive market, so contact the vendors for more information. Derwent World Patents Index (DWPI), from Clarivate Analytics, offers additional indexing, abstracts and titles rewritten for clarity, and global coverage. SciFinder/STN from ACS/American Chemical Society offers worldwide coverage of broadly chemical-related subjects, additional indexing, rewritten abstracts, and chemical structure and substructure searching of compound covered by patents, including prophetic compounds, which may not yet have been synthesized (Markush structures). Orbit from Questel also offers

additional indexing, worldwide coverage, and chemical structure and substructure searching. Reaxys from Elsevier offers worldwide coverage of broadly chemical-related subjects back to the 1800s with additional indexing and chemical structure searching. Patbase from Minesoft/ RWS Group offers some of the broadest worldwide patent coverage available with an enhanced and popular search interface. Identifying and locating a particular book can be important in clinical research. WorldCat, http://www. worldcat.org/, is a searchable database of the collections of more than 10,000 libraries worldwide. Help is available at http://www.oclc.org/support/help/worldcatorg/ ApplicationHelp.htm and an FAQ is available at https://www.oclc.org/support/services/worldcat-org/ faq.en.html.

Internet Search Engines Google Name/producer/cost/type: Google https://www.google.com; Google Inc., Alphabet Inc.; free advertising supported; indexes web pages and many other types of files posted on the Internet including PowerPoint slides, pdf files, and images. Sources/subjects: Web pages and other content crawled by Google, as well as content digitized by Google; see below for specialized Google search engines Google Scholar (fulltext articles), Google Books (full-text books), Google Patents (full-text patents), and Google News. Special features: To see more search results, at Search Settings http:// www.google.co.in/preferences?hl¼en click “Never show Instant results” and change “results per page” from 10 to 100. Search operators include “phrase search”- search for the phrase enclosed within double quotes; exclude “-” which is a hyphen immediately before the excluded word (no spaces); “inurl:”dthe URL (Uniform Resource Locator, the website address) must include the word after the colon, for example, “inurl:pdf” would retrieve pdf filesdfor more information see https://support.google.com/websearch/answer/ 2466433. Information/support: Google search help https://support.google.com/ websearch#topic¼3378866 Advanced search https://www.google.com/ advanced_search Continued

IV. CLINICAL RESEARCH INFRASTRUCTURE

742

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

Advanced search help https://support.google. com/websearch/answer/35890 Search filters https://support.google.com/ websearch/answer/142143 Image search https://support.google.com/ websearch/answer/112511 Reverse Image search https://support.google. com/websearch/answer/1325808 Advanced image search https://www.google. com/advanced_image_search Google Power Searching http://www. powersearchingwithgoogle.com/donline courses from Google Advanced Power Searching Skills http://www. powersearchingwithgoogle.com/course/aps/ skillsdlinks to Google Search Tactics and Research Strategies texts and videos Power Searching Quick Reference http://www. powersearchingwithgoogle.com/assets/misc/ AdvancedPowerSearchingQuickReference.pdf Additional Google specialty resources: Google Scholar https://scholar.google.com/ dindexes journal articles, books, dissertations, and optionally patents and/or case law; many citations have links to full text of source material; easily search-related citations and authors; Advanced Search and Settings are available from drop-down menu at the far right of the search results screen; Settings https://scholar.google. com/scholar_settings allow you to change the results per page from 10 to 20, show links to download citations into a bibliographic citation manager, and show Library Links from an article to libraries for which searcher has access. Use Settings to include citing references to retrieved citations; includes citations from journals not indexed by other citation search resources, including many in non-English language journals and books. Set alerts for updates on searches as content is added to the database using “create alert” link. About https://scholar.google.com/intl/en/ scholar/about.html Search Tips https://scholar.google.com/intl/en/ scholar/help.html Google Scholar, lesson from Power Searching with Google http://www.powersearchingwithgoogle. com/assets/textversions/7-10-Google_Scholar/ 710GoogleScholar.html Google Books https://books.google.com/dfulltext books and magazines scanned by Google, many of which can be viewed and downloaded in full, some are only available with limited views.

About https://books.google.com/googlebooks/ about/ How to use https://support.google.com/ websearch/answer/43729 Advanced https://books.google.com/advanced_ book_search Google Books, from Power Searching with Google http://www.powersearchingwithgoogle.com/ assets/textversions/11-14-Book_Search_for_ research/1114BookSearchforresearch.html Google Patents https://www.google.com/? tbm¼pts search and view patent documents and applications from many countries; many of the documents are available in full text Advanced https://www.google.com/advanced_ patent_search About https://support.google.com/faqs/answer/ 6390996?hl¼endoverview, coverage, searching, and results Google Patents, from Power Searching with Google http://www.powersearchingwithgoogle.com/ assets/textversions/10-13-Patents/1013Patents. html Google News and Google News Archive https:// news.google.comdlinks to news articles in full-text newspapers with customizable sources and topics; after a search create an alert to be notified of new articles on the topic Archives https://news.google.com/ newspapersdsearch and view full-text historic newspapers and magazines Help https://support.google.com/news/answer/ 106259?hl¼en https://support.google.com/news#topic¼2428790 Google News, from Power Searching with Google http://www.powersearchingwithgoogle.com/ assets/textversions/9-12-Google_News_and_ News_Archive/912GoogleNewsandNewsArchive. html Google Translate https://translate.google.com/ dwebsite for machine translation; also available as a downloadable app and a plug-in for the Chrome Web browser. Help https://support.google.com/translate/? hl¼en#topic¼7011755

Bing Name/producer/cost/type: Bing https://www.bing.com/; Microsoft Inc.; free advertising supported; web pages and many other types of files posted on the Internet including PowerPoint slides, pdf files, and images.

IV. CLINICAL RESEARCH INFRASTRUCTURE

BIOMEDICAL DATABASES

Sources/subjects: Web pages and other content crawled by Microsoft, as well as content digitized by Microsoft including academic papers, also available at Microsoft Academic (see below) Special features: To see more search results, at preferences http:// www.bing.com/account under “Results” change “Auto” to 50; other settings are also available. Search operators include “phrase search”- search for the phrase enclosed within double quotes; exclude operator “-” use a hyphen immediately before the word (no spaces) to be excluded from the results; “filetype:” retrieves only the files of the type specified

Internet search engines provide access to web pages as well as articles, databases, books, patents, newspapers, magazines, blogs, dissertations, white papers, lecture videos and notes. Internet searches can be very useful in clinical research since users can obtain information not readily available anywhere else. The first indication of hexamethonium’s toxicity was found by the Johns Hopkins University principal investigator through an Internet search on the day that Ellen Roche was admitted to the hospital with declining pulmonary function. The website found “listed 13 medical journal articles from the 1950s and 1960s that linked hexamethonium to sometimes-fatal lung damage.”36 In this case the website, Pneumotox Online, www. pneumotox.com,1 was reputable, run by a university hospital and had correct information. However, much of the information available online may not be correct or may be presented with an incorrect bias. For example, see the satire site about the dangers of dihydrogen monoxide, http://www.dhmo.org/facts.html, which states “each year, Dihydrogen Monoxide is a known causative component in many thousands of deaths.” Dihydrogen monoxide is, of course, another name for water, and while the information in the website is technically correct, it offers an incomplete and alarming portrayal of this “dangerous” chemical. So, it is important to remember to verify any information found on the Internet, and check it for accuracy, publisher reputation, and currency. One of the best strategies is to use the Internet as a pointer to find either verifiable information or citations, or to find information that can be searched further to obtain verifiable citations or information. It is also important to learn where and how to search for information online, so that relevant information can be rapidly identified. The Johns Hopkins internal

743

after the colondfor example, “filetype:pdf”would retrieve pdf filesdfor more information see http:// help.bing.microsoft.com/#apex/18/en-us/10002/ 0 and keywords http://help.bing.microsoft.com/ #apex/18/en-us/10001/0. Information/support: Help http://help.bing.microsoft.com/#apex/18/ en-US/n1999/-1/en-US Additional Microsoft specialty resources: Microsoft Academic https://academic.microsoft. com/dacademic content from many publications. FAQ https://academic.microsoft.com/FAQ

investigation found that only two out of four Internet search engines were able to locate the Pneumotox Online website.1 Here are some overview strategies for effective Internet searching, some of which are detailed above for Google and Bing: • Change the default preferences from 10 to 100 (Google) or to 50 (Bing) to see more results readily without needing to click on the next page of results. • Learn to use the advanced search forms and the search operators routinely, particularly the “” phrase search double quotation marks; take whatever training is available, such as the Google Power Searching classes. • Take advantage of the automatic spelling correction, synonym, and alternate spelling matching (lorry/ truck, oedema/edema) that is part of Internet searching using Google, Bing, or other major search engines; remember to note any relevant related words, synonyms, concepts, or spellings to use in further searches in systems without this automated assistance. • Think about which kinds of information sources would be most appropriate and authoritative for a particular search; use the specialty engines and resources detailed, particularly the full-text resources Google Scholar and Microsoft Academic. • Realize that Internet search results may not be reproducible, particularly if a user is signed in. The search engine may use previous searches to “personalize” and bias future search retrieval, as described by Eli Pariser in The filter bubble: how the new personalized web is changing what we read and how we think (Penguin Books, 2012).37 • Good practice includes making file copies of web pages used, including the URL and the date accessed.

IV. CLINICAL RESEARCH INFRASTRUCTURE

744

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

• Much of the Internet search interfaces, mechanisms, help documentation, specialty engines and resources are subject to change without notice, to foil Search Engine Optimization efforts as previously noted, or to better succeed in a competitive market; the searcher must be vigilant. • Change search terms regularly and try different Internet search enginesdpossibly adopt a “grazing” strategy and avoid getting bogged down or distracted; keep notes of which search engine was used and what terms were searched, with dates and key retrieved items. Privacy considerations are an important issue when searching the Internet. Search terms are not confidential on most Internet search engines; in fact, users are usually tracked and their search terms are mined to better sell them advertisements. However, these search engines advertise that they maintain users’ privacy: • Startpage, www.startpage.com, provides Google search results with privacy protection; advanced search, settings, and more details are available. • Ixquick, www.ixquick.eu, Startpage’s sister site, provides top results from multiple search engines while complying with European Union privacy standards; advanced search, settings including multiple languages, and more details are available. • DuckDuckGo, https://duckduckgo.com, provides results from many sources including Bing, Yandex, and Wikipedia with privacy protection; advanced search, settings, multiple languages, and details are available. There are many other Internet search engines with different features and capabilities. Leading nonEnglish search engines include Baidu www.baidu.com/ from China and Yandex www.yandex.com from Russia. Baidu is searchable in English using the Google Translate feature in the Google Chrome Web browser www. google.com/chrome/browser/desktop/index.htm; Yandex is also searchable in English. Internet search engines are described, listed, and compared in scholarly journals (see, for example, Lewandowski38 and in Wikipedia articles such as https://en. wikipedia.org/wiki/Outline_of_search_engines, https:// en.wikipedia.org/wiki/List_of_search_engines, and https://en.wikipedia.org/wiki/Comparison_of_web_ search_engines). Caution is indicated when consulting Wikipedia, since it is a crowd-sourced encyclopedia and may contain dated or biased information, or information that has not yet been corrected or verified. A good strategy is to extract key references and citations from Wikipedia articles, while treating the text of the articles as background information to be verified.

There are many specialty Internet search engines, directories, listing services, portals, and hubs. For example, Harvard’s Think Tank Search, http://guides.library. harvard.edu/hks/think_tank_search/US, searches white papers and reports from approximately 600 think tanks and research centers; this resource could be very useful for policy research and analysis. DMOZ (formerly the Open Directory Project), http://www.dmoz.org/, is a human-edited web directory with many useful links, including 18 sites listed under Health > Medicine > Evidence-Based Medicine > Guidelines, http://www. dmoz.org/Health/Medicine/Evidence_Based_Medicine/ Guidelines/. Since these can frequently change or become superseded, one particularly useful service mentioned earlier is the Internet Archive’s Wayback Machine, www. archive.org, where users may find historic web pages. One of the most effective tactics for finding useful information on the Internet is to understand the coverage of large search engines and which special operators to use as mentioned above. An example is how to search for an overview of a developing technologydsay, medical image analysisdto better understand new research. This recent background material can be difficult to find but is often presented in college classes. Many leading universities have made their class notes available online, so searching Google with the terms, medical image analysis inurl:edu, retrieves a relevant course syllabus from Carnegie Mellon University with downloadable lecture PowerPoints and videos (Methods in Medical Image Analysis, John Galeotti, Carnegie Mellon University, Spring 2017 http://www.cs.cmu.edu/wgaleotti/ methods_course/). Another example is if one were interested in finding studies regarding the toxicity of using a 7-day levothyroxine dose for noncompliant patients. Since both Google and Bing have indexed scientific articles and PMC or PubMed entries, the following search strategy was devised using ncbi as a simple search term to include items from both PubMed and PMC: levothyroxine noncompliant weekly ncbi. This search found several very relevant PMC records within the initial sets of retrieved items from both Google and Bing searches.

BIOINFORMATICS RESOURCES Bioinformatics is generally understood to include the development and application of computational methods to research questions in biology and molecular biology often involving genetic and protein data, functions, and interactions. Bioinformatics is interdisciplinary and associated with a wide variety of subjects such as biology, biochemistry, genetics, population modeling,

IV. CLINICAL RESEARCH INFRASTRUCTURE

BIOINFORMATICS RESOURCES

evolution, statistics, computer programming, data analysis, and visualization. Following on the successful sequencing of the human genome, bioinformatics is increasingly relevant in clinical research illuminating the complex relationships between diseases, therapies, environment, and human health. It is a key factor in the development of personalized medicine, treatments customized to be most effective for an individual patient. For example, a recent article researching the mechanism of action of the antidiabetic drug metformin reported on the results of a genome-wide association study (GWAS) of more than 13,000 patients.39 There are many important bioinformatics information resources with new ones appearing regularly. Below is a sampling.

Major Bioinformatics Organizations • NCBI (National Center for Biotechnology Information, National Library of Medicine (NLM), National Institutes of Health (NIH)), https://www. ncbi.nlm.nih.gov/ provides many freely available curated resources, all listed at http://www.ncbi.nlm. nih.gov/guide/all/, along with extensive explanations, tutorials, and FAQs. The NCBI Genetics & Medicine subset is available at http://www.ncbi. nlm.nih.gov/guide/genetics-medicine/. Selected resources include the following: Databases: • Gene http://www.ncbi.nlm.nih.gov/ genedsearchable database of sequenced genomes • GenBank https://www.ncbi.nlm.nih.gov/ genbank/dannotated collection of publicly available DNA sequences, also available from DDBJ (the DNA Data Bank of Japan) and ENA (the European Nucleotide Archive) • ClinVar http://www.ncbi.nlm.nih.gov/ clinvardreported relationships between human variation and health status • Genetic Testing Registry (GTR) http://www.ncbi. nlm.nih.gov/gtr/dvoluntary registry of genetics tests and laboratories • Online Mendelian Inheritance in Man (OMIM) http://www.ncbi.nlm.nih.gov/omim/ddatabase of human genes and genetic disorders • Database of Single Nucleotide Polymorphisms (dbSNP) https://www.ncbi.nlm.nih.gov/ snpdsingle nucleotide variations, microsatellites, and small insertions/deletions • Database of Genotypes and Phenotypes (dbGaP) http://www.ncbi.nlm.nih.gov/gapddata from studies on the interaction of genotype and phenotype • RefSeqGene http://www.ncbi.nlm.nih.gov/refseq/ rsg/dhuman gene-specific reference genomic

745

sequences, a subset of RefSeq http://www.ncbi. nlm.nih.gov/RefSeq/ • Genome https://www.ncbi.nlm.nih.gov/ genomedsequence and map data from the genomes of over 1000 organisms both completely sequenced and in progress Tools: • BLAST (Basic Local Alignment Search Tool) https://blast.ncbi.nlm.nih.gov/Blast.cgidfinds regions of similarity between nucleotide or protein sequences, one of many sequence analysis resources available at http://www.ncbi.nlm.nih. gov/guide/sequence-analysis/ • HomoloGene http://www.ncbi.nlm.nih.gov/ homologene/dgene homology tool • Phenotype-Genotype Integrator (PheGenI) https://www.ncbi.nlm.nih.gov/gap/ phegenidsearch for human phenotype/genotype relationships • 1000 Genomes Browser https://www.ncbi.nlm. nih.gov/variation/tools/1000genomes/ dinteractive graphic viewer enabling exploration of data from the 1000 Genome Project, the largest public catalog of human variation and genotype data, collected between 2008 and 2015 (http:// www.internationalgenome.org/about) Support: • NCBI Learn http://www.ncbi.nlm.nih.gov/ home/learn.shtml webinars, courses, conferences, presentations, tutorials, and documentation • NCBI YouTube channel https://www.youtube. com/user/NCBINLM Note: PubMed records are annotated with links to the NCBI bioinformatics tools as appropriate. From within the PubMed search interface, all of these databases can be searched directly using the pull-down menu to the left of the search box to select an appropriate file. Choosing “all databases” runs a quick search across all files and displays the number of hits in each. • EMBL-EBI, https://www.ebi.ac.uk/, the European Molecular Biology Laboratory and European Bioinformatics Institute, provide many databases with bioinformatics data as well as additional resources. • DDBJ, http://www.ddbj.nig.ac.jp/, the DNA Data Bank of Japan, provides access to databases, software tools, and documentation. • ExPASy, http://www.expasy.org/ (formerly Expert Protein Analysis System), from the Swiss Institute of Bioinformatics (SIB), is a resource portal with databases and software tools.

IV. CLINICAL RESEARCH INFRASTRUCTURE

746

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

Bioinformatics Directories • OBRC, http://www.hsls.pitt.edu/obrc/, the Online Bioinformatics Resources Collection, from the Health Sciences Library, University of Pittsburgh, is an annotated, searchable list of bioinformatics databases and tools. • Bioinformatics Links Directory, https:// bioinformatics.ca/links_directory/, from the Canadian Bioinformatics Workshops, is a searchable list of bioinformatics web resources. • The journal Nucleic Acids Research publishes an annual database issue, http://www.oxfordjournals. org/nar/database/c. • A list of biological databases, https://en.wikipedia. org/wiki/List_of_biological_databases, and a list of open-source bioinformatics software, https://en. wikipedia.org/wiki/List_of_open-source_ bioinformatics_software, are available from Wikipedia.

Browsers • Gene browser Ensembl http://www.ensembl.org/ from EMBL-EBI has tools and other functionalities, as well as other browsers available. • UCSC Genome Browser https://genome.ucsc.edu/ training/index.html from the University of California at Santa Cruz also has other tools available. • MapViewer https://www.ncbi.nlm.nih.gov/ projects/mapview/from NCBI provides special browsing capabilities for a subset of organisms from the NCBI Genome database.

Commercial Software A wide variety of bioinformatics software is available from commercial vendors, as part of a very competitive and rapidly changing marketplace. Researchers might benefit from the significant advantages offered by specialized software packages and suites. Some vendors include Qiagen (Human Gene Mutation Database Professional, Ingenuity Pathways Analysis), DNASTAR (Lasergene), Clarivate Analytics (MetaCore), and Elsevier (Pathway Studio).

DATA MANAGEMENT Data management frequently plays an important part in clinical research (see Chapter 30). The subject of data management can include where and how the researcher gets data, how it is described, organized, manipulated, analyzed, visualized, displayed, archived, shared and reused.

Here is an example of a research data life cycle: Proposal planning and writing; Project start up; Data discovery; Data collection; Data analysis (possibly leading to data repurposing); Data sharing; End of project; Deposit data in data archive; and finally the Reuse of data for data discovery, since this is an iterative process not a linear one. Data management can often involve some programming tools depending on specific subdiscipline and laboratory practices, such as R, Python, and statistical analysis and data visualization specialties. Researchers might consider developing a preproject data management plan, perhaps using open data formats rather than proprietary formats, and possibly collecting meta-data including synonyms and controlled vocabulary terms to use in the data collection phase. There are also considerations regarding sharing research data. Many journals have requirements about depositing research data in repositories, such as the Gene Expression Omnibus (GEO, www.ncbi.nlm.nih. gov/geo/) or GenBank (www.ncbi.nlm.nih.gov/ genbank/). Certain types of genomic data generated in connection with NIH-funded research must be submitted to a designated data depository, according to the NIH’s Genomic Data Sharing Policy, effective January 25, 2015 (https://gds.nih.gov/). With regard to all research data generated by NIHfunded research, the NIH’s Data Sharing Policy currently applies to larger grants of $500,000 or more in direct costs (http://grants.nih.gov/grants/guide/ notice-files/NOT-OD-03-032.html). It is expected that soon the NIH will require all funding proposals to include a data sharing plan (https://grants.nih.gov/ grants/NIH-Public-Access-Plan.pdf). There are some indications that researchers who share data receive increased citations of their work and greater citation impact.40,41

DATA INTEGRATION AND PRECISION MEDICINE Major advances in medicine are expected from the integration of data from genomic and bioinformatics sources, medical records, drug screening, and literature and text resources. Researchers are mining these networks of data and information for the discovery of disease mechanisms and new therapeutics. This concept is illustrated in Fig. 42.12, after an image from a lecture delivered by Kevin White, of the University of Chicago, at the National Institutes of Health, “Genomic Networks in Development and Cancer: Resolving Biomarkers and Therapeutic Targets from a

IV. CLINICAL RESEARCH INFRASTRUCTURE

747

BIBLIOMETRICS

FIGURE 42.12

Defining Networks, Mining Networks. Image from Kevin White. Genomic networks in development and cancer.In: NIH Seminar. Bethesda (MD); February 14, 2012, videocast available at http://videocast.nih.gov/launch.asp?17107, used with permission.

Cloud of Data” (February 14, 2012, available at http:// videocast.nih.gov/launch.asp?17107). Another lecture on a similar concept was delivered by Atul Butte, of the University of California at San Francisco, also at the NIH, “Translational Bioinformatics: Transforming 300 Billion Points of Data into Diagnostics, Therapeutics, and New Insights into Disease” (June 20, 2012, available at http://videocast.nih.gov/launch. asp?17321). Russ Altman of Stanford University delivered a lecture on this topic more recently at the NIH, “Integrating multi-scale data for biomedical discovery and clinical implementation” (October 7, 2015, available at https:// videocast.nih.gov/launch.asp?19221). All of Us (sm), the NIH Precision Medicine Initiative, http://www.nih.gov/precisionmedicine/, intends to extend this concept with plans for a research cohort of more than one million volunteers who will share genetic data, biological samples, and diet/lifestyle information.

BIBLIOMETRICS Bibliometrics is the statistical analysis of publications, books, or articles. It is often used to measure the impact of an author, a publication or publications, or a topic within a portfolio or subject area. Two common bibliometric indicators are the impact factor of a journal and the h-index, which can be applied to an author, group of authors, or other publications. The calculation of a Journal Impact Factor (JIF) for a journal for a given year divides the number of citations received in the given year for articles published in that journal in the preceding 2 years, by the total number of articles published in that journal during the preceding 2 years. Journal Citation Reports, available from Web of Science, publishes the annual JIFs, as well as additional metrics such as a 5-year Impact Factor, Immediacy Index, and the Cited Half Life; see detailed descriptions of the metrics and their calculation on the Web of Science site. A similarly computed metric is also available for Scopus data

IV. CLINICAL RESEARCH INFRASTRUCTURE

748

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

from Scimago; for more information, see http://www. scimagojr.com/journalrank.php. Bear in mind that the JIF only applies to a journal not to an article or a set of articles nor the work of an author. JIF numbers reflect citation patterns in subject disciplines; they also reflect the size of the scholarly community in a discipline. Thus journals in cardiology have, on average, higher JIF numbers than those in ophthalmology. The h-index, the first widely used metric to assess the impact of an individual researcher, is calculated by taking the number of articles published by that author, and the number of times each article has been cited; the h-index is how many of the author’s articles have received at least that many citations. An author with an h-index of five must have published at least five articles each of which has been cited at least five times. The h-index is dependent on the subject area, since different fields have varying citation patterns; this is also true, at the journal level, for the JIF. The h-index also depends on how long the author has been publishing. An author with a few highly cited publications may still have a low h-index: a researcher with 5 papers, each of them cited 100 times, would have an h-index of 5, the same number as a researcher with 50 papers, only 5 of them cited at least 5 times. When calculating an h-index, care should be taken to examine the data used and to avoid incorrect or duplicate publications. H-index values, and all other bibliometric indicators, may also vary depending on the database used as a citation source. Commercial databases Scopus and Web of Science cover different journals and so may produce different h-index values. Google Scholar can also be used for h-index calculations (for details see http:// scholar.google.com/intl/en-US/scholar/metrics.html). The free software Publish or Perish from http://www. harzing.com/pop.htm uses Google Scholar and Microsoft Academic citation data and can automatically calculate h-index values; these four resources rarely have the same data, thus rarely compute identical h-indexes. Several additional free or open access tools are available to assist with bibliometrics: • MyRI, http://myri.conul.ie/, is an open access toolkit for bibliographic training and awareness. • Scholarometer, http://scholarometer.indiana.edu/, from Indiana University, is a social tool to facilitate citation analysis. Additional bibliometric indicators have been proposed, including some generally grouped under the rubric “altmetrics” for alternative or nontraditional metrics (https://www.altmetric.com/about-altmetrics/ what-are-altmetrics/, http://altmetrics.org/manifesto/).

BIBLIOGRAPHIC MANAGERS Bibliographic citation management software is used to organize and manage references for research and writing. Also called reference managers, this software usually permits the format of the references to be changed to conform to the citation styles used by different publishers. Citation managers may be based in a single computer or be web-based and may be used to share references among a work group. Many databases allow search results to be output or exported in a format easily read by citation managers (possibly including filetypes csv or ris). A variety of other citation manager features may include annotation, automatic searching of bibliographic databases, placing formatted references into a manuscript, and organizing pdf files. Examples of popular subscription-based citation managers include Endnote from Clarivate Analytics (formerly Thomson Reuters), Mendeley from Elsevier (with a free, limited version), and Papers from ReadCube. There are also free and open-source citation managers, such as Zotero and BibDesk. A number of articles compare citation managers features, such as the Wikipedia entry, “Comparison of reference management software” (https://en.wikipedia.org/wiki/Comparison_ of_reference_management_software). A slightly different approach is taken by CiteULike, http://www.citeulike.org/, a free crowd-sourced service utilized to store, organize, and share scholarly papers. Citations can be imported from, and exported to, a user’s citation management software. Since citations in CiteULike can be tagged and shared, it can be used as a discovery tool to find relevant literature. Many researchers find citation management software invaluable for efficient and effective research to save time and avoid errors and duplication in using and citing references. A common comment from young researchers and future clinician-scientists at academic institutions is that they wish they had started using citation managers earlier in their research careers. One should plan on regularly updating this software; bibliographic databases and journal requirements change regularly, and work-arounds for an outdated program can be frustrating.

RESOURCE SELECTION AND SEARCH STRATEGY One consideration in selecting information resources is understanding the selection process by which information is added to the resource. MEDLINE has standards for a journal’s viability, quality standards for

IV. CLINICAL RESEARCH INFRASTRUCTURE

749

FINAL NOTES

both the production and the scientific content, editorial input, independence of the editorial process, etc., as well as general requirements for the journal to be in scope for the database; some information about these criteria can be seen at www.nlm.nih.gov/pubs/ factsheets/j_sel_faq.htmlda2. Web of Science includes journals shown to meet quality standards, but the citability of a journal’s articles, used to calculate JIF, is a major consideration. Google Scholar, on the other hand, is available freely but does not select content with the same quality standards. Google Scholar may include multiple citations to the same content and incomplete or inaccurate references; it also includes articles in journals recognized as “predatory.” Some subscription database producers may adjust their inclusion standards to create products that are comprehensive for a discipline; for example, CINAHL includes many types of information material, all with applicability to nursing and allied health. Another consideration is to determine which resources were searched in relevant systematic reviews covering the research topic. Researchers could start with these resources to better ensure comprehensive retrieval. Some of the search strategies used in those systematic reviews could also indicate useful keywords, indexing and controlled vocabulary terms for further searching. Here is a general example of building a search strategy based on this iterative model: 1. Break the query into essential conceptsdpossibly using PICO. 2. Find synonyms/alternate terms for each conceptduse controlled vocabularies if available. 3. Expand, then reduce retrievaldgroup synonyms together for each concept with “OR” logic; then “AND” the different groups of concepts together. 4. Use as many information resources as necessary, including sources searched in relevant systematic reviews. 5. Search for information to find more information: a. find key citations, possibly with Wikipedia (check the footnotes), Google, Bing, etc.; b. search for these key citations in PubMed/Embase/ etc. to find controlled vocabulary terms to expand your synonym lists; c. use these key citations in cited reference searching (Scopus/Web of Science/Google Scholar); d. check the reference lists of pertinent papers, repeating as far back as helpful references continue to be identified; do not limit this process to the journal literature. 6. Repeat as necessary. 7. Keep track of all searches and retrievals. 8. Remove duplicate retrievals.

9. Check for any newly published material before completing research or submitting article; it may also be useful to check whether key articles have been corrected or retracted since first used. PubMed/ MEDLINE citations are tagged with links to corrections and retractions published in the original journals.

EDUCATIONAL RESOURCES There are a number of educational resources that can provide additional information about comprehensive literature searching and the subjects covered in this chapter. One key resource is Charles Wessel’s course, “Responsible Literature Searching for Research: A SelfPaced Interactive Program,” available free from the University of Pittsburgh.11 Guidelines from Mount Sinai School of Medicine6 could be useful as mentioned previously; they include overview, search log, summary and bibliography forms, lists of sources, and search technique suggestions. Yale University also has useful guidelines with a good checklist.42 Certainly a number of the reference works cited in this chapter could be useful. The vendors or producers mentioned above can also a good resource for support material, training seminars, and information about their products. Since information resources frequently change, conscientious researchers can search for updated guides and overviews. These subject guides are often produced by libraries, which may mount them on the web using the libguides platform. So, when searching online, useful terms to include along with the subject of interest are “libguides,” “library,” or “campusguides” (an alternate name for libguide). Searching Google or Bing with the terms “bioinformatics libguides” retrieves many relevant sites including one from MIT Libraries.43

FINAL NOTES What are some good search quality checkpoints? • You have identified the essential concepts involved in the search, as well as any synonyms, plurals, or alternate spellings. • You have identified and searched all relevant resources and understand their scope and coverage. • You understand how each search engine works, what it searches, and what assumptions it makes.

IV. CLINICAL RESEARCH INFRASTRUCTURE

750

42. INFORMATION RESOURCES FOR THE CLINICAL RESEARCHER

• You have repeated your searches regularly during your research, or set up alerts to automatically send you any new results, such as My NCBI in PubMed. • Your results include known relevant articles. • Different searchers/searches yield no new relevant results. Can you trust your search results? When is a search complete? These are difficult questions, best answered by: • experience, • following the guidelines and suggestions mentioned above, • verifying your strategy and results, • collaborating with a reference librarian, and, if applicable, a pharmacist as well.

6.

7. 8. 9. 10. 11.

12.

Acknowledgments

13.

The authors wish to gratefully acknowledge assistance from Kevin White, Brian Haugen, Russ Altman, Chris Fortney, NIH Library and NLM staff members including Mary Ryan (ret.), Brian Brown, Chris Belter, Lisa Federer, Medha Bhagwat, Rex Robison, Lou Knecht (ret.), and Barbara Brandys, and the NIH Office of the Director, Office of Research Services.

14. 15. 16.

17.

Disclosure The authors have no financial interests in any of the information resources discussed in this chapter. Elsevier, the publisher of this book, owns a number of the information resources covered. However, Elsevier has had no input regarding the actual content of this chapter or opinions expressed within it. In the spirit of transparency, the authors wish to disclose backgrounds as Informationists/Research Librarians, with many years of service at the National Institutes of Health (NIH) Library. The NIH Library is separate from the National Library of Medicine (NLM) and supports the information needs of NIH staff. More information about the NIH Library is available at the web site http://nihlibrary.nih.gov and elsewhere.44

18.

19.

20.

21.

22.

References 1. Becker LC, Brower RG, Faden R, Johns RA, Lipsett PA, Petty BG, Pronovost PJ. Report of internal investigation into the death of a volunteer research subject. Baltimore (MD): Johns Hopkins University School of Medicine; 2001. Retrieved from: https://web.archive. org/web/20150311234613/http://www.hopkinsmedicine.org/press/ 2001/JULY/report_of_internal_investigation.htm. 2. Savulescu J, Spriggs M. The hexamethonium asthma study and the death of a normal volunteer in research. J Med Ethics 2002;28:3e4. 3. McLellan F. 1966 and all thatdwhen is a literature search done? Lancet 2001;358:646. 4. Perkins E. Johns Hopkins’ tragedy: could librarians have prevented a death? NewsBreaks, Inf Today August 7, 2001. Retrieved from: http://newsbreaks.infotoday.com/nbreader.asp?ArticleID¼17534. 5. JHMI Office of Communications and Public Affairs. Hopkins response to FDA observations. September 7, 2001. Retrieved from:

23. 24. 25. 26.

27. 28.

https://web.archive.org/web/20070903162020/http://www.hopkinsmedicine.org/press/2001/SEPTEMBER/010907A.htm. Mount Sinai School of Medicine, Institutional Review Board. Guidelines for determining an adequate and comprehensive literature search of drug and device safety for use by investigator and institutional review boards. January 2003. Retrieved from: http://icahn.mssm.edu/ static_files/Old%20Sites/www.mssm.edu/cmca/pdf/ SearchGuidelines.pdf. Huber JT, Swogger S, editors. Introduction to reference sources in the health sciences. 6th ed. Chicago: ALA-Neal Schuman; 2014. Srodin S. Using the pharmaceutical literature. New York: Taylor & Francis; 2006. Hersh WR. Information retrieval: a health and biomedical perspective. 3rd. ed. New York: Springer; 2009. pp. 68, 57. Malone PM. Drug information: a guide for pharmacists. New York: McGraw-Hill; 2001. pp 54, 81. Wessel C.B. (n.d.). Responsible literature searching for research: a self-paced interactive program. University of Pittsburgh. Health Sciences Library System, and Internet-Based Studies in Education and Research. https://cme.hs.pitt.edu/ISER/servlet/IteachController Servlet?actiontotake¼loadmodule&moduleid¼8381. Golder S, McIntosh HM, Duffy S, Glanville J. Developing efficient search strategies to identify reports of adverse effects in Medline and Embase. Health Inf Libr J 2006;23:3e12. Garfield E. Citation indexes for science e new dimension in documentation through association of ideas. Science 1955;122:108e11. Rosenow III EC. The spectrum of drug-induced pulmonary disease. Ann Intern Med 1972;77:977e91. Ginsparg P. ArXiv at 20. Nature 2011;476:145e7. Laakso M, Welling P, Bukvova H, Nyman L, Bjo¨rk BC, Hedlund T. The development of open access journal publishing from 1993 to 2009. PLoS One 2011;6(6):e20961. Holdren JP. Increasing access to the results of federally funded scientific research. Executive Office of the President, Office of Science and Technology Policy; 2013. http://www.whitehouse.gov/sites/ default/files/microsites/ostp/ostp_public_access_memo_2013.pdf. Kolata G. For scientists, an exploding world of pseudo-academia. N Y Times April 7, 2013. http://www.nytimes.com/2013/04/08/ health/for-scientists-an-exploding-world-of-pseudo-academia. html?_r¼0&pagewanted¼all. Segal D. The dirty little secrets of search. N Y Times February 12, 2011. http://www.nytimes.com/2011/02/13/business/13search. html. Medline, PubMed, and PMC (PubMed Central): how are they different. 2016. https://www.nlm.nih.gov/pubs/factsheets/dif_med_pub. html. Cockersole FJ, Park WW. Hexamethonium lung; report of a case associated with pregnancy. J Obstet Gynaecol Br Emp 1956;63: 728e34. Doniach I, Morrison B, Steiner RE. Lung changes during hexamethonium therapy for hypertension. Br Heart J 1954;16:101e8. Thomson Reuters. Web of Science quick reference guide. 2013. http:// wokinfo.com/media/pdf/qrc/webofscience_qrc_en.pdf. Elsevier. Scopus content. 2016. https://www.elsevier.com/ solutions/scopus/content. Hill B. Comparison of journal title coverage between CINAHL and Scopus. J Med Libr Assoc 2009;97:313e4. Thomas K. Breaking the seal on drug research. N Y Times June 29, 2013. http://www.nytimes.com/2013/06/30/business/breakingthe-seal-on-drug-research.html. Kmietowica Z. Study claiming Tamiflu saves lives was based on “flawed” analysis. BMJ 2014;348:g2228. Jefferson T, Jones MA, Doshi P, Del Mar CB, Hama R, Thompson MJ, et al. Neuraminidase inhibitors for preventing and treating influenza in healthy adults and children. Cochrane Database Syst Rev 2014;4. http://dx.doi.org/10.1002/14651858.

IV. CLINICAL RESEARCH INFRASTRUCTURE

REFERENCES

29. Wang S, Hawkins BS, Dickersin K. Cochrane systematic reviews and co-publication: dissemination of evidence on interventions for ophthalmic conditions. Cochrane Database Syst Rev 2015;118. http://dx.doi.org/10.1186/s13643-015-0104-5. 30. Lund H, Brunnhuber K, Juhl C, Robinson K, Leenaars M, Dorch BF, et al. Towards evidence based research. BMJ 2016;355:i5440. 31. Chalmers I, Glasziou P, Godlee F. All trials must be registered and the results published. BMJ 2013;346:f105. 32. Sackett DL, Rosenberg WM, Gray JAM, Haynes RB, Richardson WS. Evidence based medicine: what it is and what it isn’t. BMJ 1996;312:71e2. 33. Brunton LL, Chabner BA, Knollman BC. Goodman and Gilman’s the pharmacological basis of therapeutics. 12th ed. New York: McGrawHill; 2011. 34. Grippi M, Elias J, Fishman J, Pack A, Senior R, Kotloff R. Fishman’s pulmonary diseases and disorders. 5th ed. New York: McGraw-Hill; 2015. 35. Hunt D, Nguyen L, Rodgers M. Patent searching: tools & techniques. Hoboken (NJ): Wiley; 2007. 36. Pelton T. Death in test haunts Hopkins researcher. Baltimore (MD): The Sun; January 27, 20021A. 37. Pariser E. The filter bubble: how the new personalized web is changing what we read and how we think. New York: Penguin Books; 2012. 38. Lewandowski D. Evaluating the retrieval effectiveness of web search engines using a representative query sample. J Assoc Inf Sci Technol 2015;66(9):1763e75.

751

39. Zhou K, Yee SW, Seiser EL, van Leeuwen N, Tavendale R, Bennett AJ. Variation in the glucose transporter gene SLC2A2 is associated with glycemic response to metformin. Nat Genet 2016; 48:1055e9. 40. Piwowar HA, Day RS, Fridsma DB. Sharing detailed research data is associated with increased citation rate. PLoS One 2007;2007(2): e308. 41. Leitner F, Bielza C, Hill SL, Larran˜aga P. Data publications correlate with citation impact. Front Neurosci 2016;10:419. 42. Glover J, Gentry M. Literature searching guidelines checklist for researchers. Yale University, Cushing Whitney Medical Library; 2003. http://library.medicine.yale.edu/sites/default/files/services/ crs/Literature_Searching_Guidelines1.pdf. 43. Bioinformatics: Home (n.d.) MIT libraries. http://libguides.mit. edu/bioinfo. 44. Cogdill K. Government agency and professional association resources and services. In: Huber JT, Tu-Keefner F, editors. Health librarianship: an introduction. Santa Barbara: Libraries Unlimited; 2014. 45. Kirkland SW, Vandenberghe C, Voaklander B, Nikel T, Campbell S, Rowe BH. Combined inhaled beta-agonist and anticholinergic agents for emergency management in adults with asthma. Cochrane Database Syst Rev January 11, 2017;1:CD001284. http://dx.doi.org/ 10.1002/14651858.CD001284.pub2.

IV. CLINICAL RESEARCH INFRASTRUCTURE

Appendix 1 Answer Key to Summary Questions CHAPTER 1

CHAPTER 5

1.1 (j) All of the above 1.2 False. (Correct answer: This statement describes Hippocrates.) 1.3 (a) Semmelweis started his career as a student on an obstetric ward, (c) The second division used midwives and the death rate was only 2%e3%, (d) Semmelweis started his work by studying cadavers, (e) Semmelweis introduced handwashing with chlorinated lime to decrease mortality rates, and (f) Despite convincing data, Semmelweis’ work was condemned by colleagues. 1.4 (d) Benjamin Franklin

5.1 (b) Responsibilities of IRBs, researchers, and organizations 5.2 (a) Three years 5.3 (d) All of the above

CHAPTER 6 6.1 (f) All of the above 6.2 (e) a and c 6.3 (c) has programs to facilitate the development of new drugs and biologics 6.4 (g) a and d 6.5 (f) a, b, and c

CHAPTER 2 2.1 2.2 2.3 2.4

(a) Sample size and study design (b) The Common Rule (c) Coercion or undue influence (b) The risks are minimized and justified by benefits or the value of the study

CHAPTER 3 3.1 (c) Failure to retain research records 3.2 (d) All of the above 3.3 (a) Substantial contributions to conception and design 3.4 (b) Publication of negative data

CHAPTER 4 4.1 (b) Protect the rights and safeguard the welfare of human research subjects 4.2 (e) All of the above 4.3 (e) All of the above

CHAPTER 8 8.1 (d) All of the above 8.2 (c) When the additional risk posed by the use of placebo is minor and withholding the current standard therapy would not lead to serious or permanent harm. 8.3 (d) Cultural values and norms can influence the informed consent process 8.4 (e) All of the above 8.5 (e) b and c

CHAPTER 9 9.1 (d) All of the above 9.2 (c) Summary data displayed in a tabular format for certain registered trials 9.3 (a) Medical journal editors 9.4 (c) Systematically verified against external, objective data sources 9.5 (a) The database is not comprehensive: it does not include all clinical trials.

753

754

APPENDIX 1 ANSWER KEY TO SUMMARY QUESTIONS

CHAPTER 10

CHAPTER 18

10.1 (e) a and b 10.2 (e) All of the above 10.3 (c) They have fewer conflicts of interest (perceived or real) 10.4 (b) False 10.5 (b) False 10.6 (d) Neither a nor b

18.1 (b) Articulating the precise research question to be addressed 18.2 (d) There is not a single correct answer to this question because the most appropriate control group depends on the question and outcomes of interest 18.3 (c) Can be used when randomization is not possible, such as in naturalistic settings

CHAPTER 12

CHAPTER 20

12.1 (d) Voluntary consent and independent ethical review 12.2 (e) Legally authorized representative and/or parent with health care decision- making authority 12.3 (d) All of the above 12.4 (e) All of the above 12.5 (b) No, though individual agreements may inhibit sharing

CHAPTER 13 13.1 (e) Both b and c 13.2 (a) Women and people from minority groups be included unless there is scientific justification for not including them 13.3 (d) All of the above 13.4 (d) Both a and b

CHAPTER 15 15.1 (d) All of the above 15.2 (a) Unintentional bias by the investigator either favoring or disfavoring the intervention being tested can impact study and analytic decisions throughout the conduct of the study. 15.3 (b) Unmasked studies.

CHAPTER 17 17.1 (b) No 17.2 (c) Association study of maternal use of stilbesterol with tumor appearance 17.3 (b) Yes

20.1 (c) Frequency of a specific current procedural terminology code within Medicare outpatient billing claims 20.2 (d) All of the above 20.3 (d) Database linking

CHAPTER 21 (d) The meaningfulness of change True (c) Health status (b) FIM (a) SF36 False True False Function is a term that refers to an individual’s ability to perform needed/desirable activity. 21.7 True 21.7a True

21.1 21.2a 21.2b 21.3a 21.3b 21.4 21.5a 21.5b 21.6

CHAPTER 22 22.1 (d) There are several described approaches to detecting publication bias in studies collected for analysis. 22.2 (c) The Preferred Reporting Items for Systematic reviews and Meta-Analyses Protocols (PRISMA-P) provides a checklist that helps to improve the quality of reporting for metaanalysis. 22.3 (d) All of the above 22.4 (c) A “fixed effect” approach to combining studies for analysis assumes that the treatment effects are comparable across studies.

755

APPENDIX 1 ANSWER KEY TO SUMMARY QUESTIONS

CHAPTER 23 23.1a For simple randomization, each new patient who enrolls should have 50% chance of receiving either treatment, say treatment A or B. One can accomplish this, using either one die or both dice, by assigning half the possible die outcomes to treatment A and the other half to treatment B. For example, using one die, rolls with odd numbers {1,3,5} could be assigned to treatment A and even numbers {2,4,6} to treatment B. The die is then rolled 48 times, and the sequence of rolls and corresponding treatment is recorded. By recording outcomes of each roll that were observed and corresponding treatment assignment, there is at least some documentation provided that the intended algorithm for the randomization list was properly carried out. Today, randomization in clinical trials is generally carried out using computer software, and not physical mechanisms such as rolling dice or flipping a coin, in order better ensure the fairness and reproducibility of the resulting treatment assignment sequence. 23.1b To stratify a simple randomization procedure on gender, one simply needs to repeat the process described in part A twice to create a separate randomization list for each gender. By making the randomization list for each gender have 48 entries, you are guaranteed to have lists long even in the extreme scenario of all male or all female participants. 23.1c Potential block sizes should be somewhat small relative to the total sample size, a multiple of the number of treatments, and a divisor of 48. Some choices are as follows: 4, 6, or 8. Using a block of 6, one could randomly permute 3 As and 3 Bs in 8 blocks of 6. For additional protection against pattern recognition, one could use a computer program to randomly choose the block size from 4, 6, and 8 and then permute equal numbers of As and Bs in each block. Doing it this way you are not guaranteed to have a completed block at the end, but the resulting imbalance should be small in this scenario if the block sizes are kept small relative to 48. 23.2 Randomization naturally balances observed and unobserved covariates across treatment arms. By having more reasons to exclude patients in one arm postrandomization, this treatment could wind up with patients who are no longer comparable with those on the other arm. A second way this could ruin randomization is that by replacing a withdrawn patient with the next available patient, the person who recruits the next patient could easily be unblinded to the treatment assignment the patient would get and unconsciously, or consciously, choose which

patient to recruit based on whether they thought the potential subject would respond well to surgery. 23.3 A scheme similar to that described in 1A could be used to perform simple 1:1 randomization. With this scheme it is possible to wind up with all 10 assignments on one arm; however, this is very unlikely. 23.4 A permuted block algorithm, using 4 blocks of size 6 could have been used to randomly assign 24 subjects in equal numbers to two treatments. Randomization would have naturally balanced patient characteristics. 23.5 (a) Investigators enrolling subjects could be easily be unblinded and subsequently choose to accept or invite the next patient walking into the clinic to participate in the trial based on whether they thought the patient was a good candidate for the next treatment. (b) There is a danger that patients could become aware of the enrolling/treatment assignment process and organize themselves in a manner to get the desired treatment, particularly in trials where potential participants know each other. (c) If at any time one subject is unblinded the entire cohort would be unblinded. Observations by unblinded investigators could be biased and hence bias trial results.

CHAPTER 24 24.1 Let mD represent the mean number of fractures in the vitamin D supplement group and mP represent the mean for the placebo group. Null Hypothesis H0: mD 1=4 mP Alternative Hypothesis HA: mDsmP 24.2a The study found a significant difference in hip density between the two treatment groups. The null hypothesis can be rejected at the 0.05 level. 24.2b In the ITT analysis, women in the vitamin D supplementation group had a 12% decreased risk of hip fracture relative to women in the placebo group, but this difference is not significant. Hence, the null hypothesis of no difference (rate ratio ¼ 1) between the groups cannot be rejected. In the analysis examining compliant women only, the vitamin D supplement group had a 29% decreased risk of hip fractures compared with placebo, and this result is significant. Hence, the null hypothesis of no difference in fracture rate among the compliant women can be rejected; however, this is a nonrandomized comparison, and thus this result must be interpreted with caution.

756

APPENDIX 1 ANSWER KEY TO SUMMARY QUESTIONS

24.3 Let p1 be the true underlying probability of surviving 30 days on treatment 1. Let p2 be the same probability for treatment 2. Null Hypothesis H0 : p1 1=4 p2 Alternative Hypothesis HA: p1sp2 24.4a The P-value is the probability that results as extreme or more extreme than what was observed would occur given the null hypothesis is true. 24.4b A type I error occurs when the null hypothesis is rejected given that it is true. 24.4c A type II error occurs when the null hypothesis is not rejected given that it is false. 24.5 The goal of hypothesis testing is to make statements about the general population. This goal is in perfect alignment with the purpose of a phase III clinical trial, which seeks definitive answers to a scientific question that are generalizable to a broader population than just the cohort in the trial. 24.6 (b) No, failing to reject the null hypothesis means there is insufficient evidence to reject the null hypothesis. 24.7 (a) The P-value is the probability that the null hypothesis is true

CHAPTER 26 26.1 (e) All of the above 26.2 (c) Use a Cox model. You expect the Kaplan Meier estimator will be wrong

CHAPTER 27 27.1 Subjects who were observed to still have bone loss after 12 months of treatment could have been randomized at 12 months to receive additional vitamin D supplement or placebo. The difference in response between these two randomized treatment groups would have been an unbiased measure of the treatment benefit of vitamin D. 27.2 (a) A test with some value as a diagnostic test would have an AUC > ½; a perfect test would have an AUC ¼ 1; a useless test would have AUC ¼ ½. (i) Null hypothesis: AUC ¼ ½ (ii) Alternative hypothesis: AUC > ½ (b) Sensitivity is the probability a test detects disease among truly diseased individuals. Let Se1 be the sensitivity of test one, and Se2 be the sensitivity of test two. Null hypothesis: Se1 1=4 Se2 Alternative hypothesis: Se1sSe2

27.3a This will not introduce bias if you choose the appropriate survival analysis methods, discussed in Chapter 23. 27.3b This could bias trial results if the reasons for loss to follow-up are different on the two arms and related to treatment response, but due to the small percentage of lost data the missing results will likely have a small impact on the results. 27.3c A rate of 15% is high enough to be of concern as a possible and noticeable source of bias if data are analyzed without methods to adjust for or examine the sensitivity of results to different assumptions about the missing data. 27.4 To protect the randomization, and comparability of individuals on the two arms, an ITT analysis should be performed. To do this all participants, regardless of whether they need to be reassigned to an alternative therapy, should be followed up in the same manner and treated as members of their original randomized treatment group in the analysis. 27.5 (b) Noninferiority trial.

CHAPTER 28 28.1 (d) All of the above 28.2 (a) Unintended biological effects are relatively uncommon. 28.3 (d) All of the above 28.4 (c) Data and safety monitoring committee (DSMC) 28.5 (c) Research conducted in emergency contexts

CHAPTER 29 29.1 (e) Arrange to file a PCT in 4 months, and in 22 months, to file a US-371 and National Phase applications in major markets. Wrong answers explained: (a) is wrong because failing to file National Phase in major markets means losing patent rights there, and having those rights likely will make finding an exclusive licensee substantially easier. (b) is wrong because you cannot file National Phase applications if you do not file a PCT application first, and the US-111 application, while equivalent in content to a PCT, is not a substitute. (c) is wrong because your predecessor already filed a Provisional. (d) is wrong simply because prospective licensees may be holding back until the research has progressed enough to warrant taking the risk, which easily might not happen until after National Phase. 29.2 (a) Whether the technology is a Cooperative Research and Development Agreement “subject invention.”

757

APPENDIX 1 ANSWER KEY TO SUMMARY QUESTIONS

Wrong answers explained: all of the other four choices are elements explicitly required by 35 USC x 209 and 37 CFR Part 404. 29.3 (c) The contracting agency has determined that changing the standard patent clauses will better promote the policy and objectives of the Bayh-Dole Act. Wrong answers explained: (a) is wrong because the interests of an intramural research program are not grounds for implementing a DEC; rather, a DEC should only be used to advance the stated goals of the Bayh-Dole Act itself. (b) is wrong because the Bayh-Dole Act only applies to contracts for research and development. (d) is wrong because the prospective bidders do not have a “veto,” just an opportunity to comment (though those comments should be taken seriously). (e) is wrong because the “Greater Rights” refers to a contracting clause that enables the contractor to assert greater rights in its invention over the agency (on a case-bycase basis), not the other way around.

CHAPTER 30 30.1 (d) All of the above 30.2 (b) A telephone call 30.3 (d) An untoward event that may or may not be related to the study agent 30.4 (c) An audit performed to investigate fraud or misconduct

CHAPTER 31 31.1 (a) If the subject did not appear, to the researcher, to be in pain, then he probably was not really having pain 31.2 (a) It is subjective and quantitative 31.3 (b) The name of the researcher who recorded the answer 31.4 (c) A numeric test result (e.g., “4.0”) 31.5 (e) All of the above need to be recorded 31.6 (c) Adhering closely to the Vancouver citation style when submitting a manuscript for publication

CHAPTER 32 32.1 (a) A storage center for biological specimens 32.2 (b) That the type of biospecimen used has been optimized for a particular type of analysis 32.3 (d) Protocols that have been optimized using highquality specimens and other reference standards to ensure that the procedures developed will yield scientifically significant results 32.4 (d) All of the above

32.5 (d) All of the above 32.6 (a) They contain information stored in a vertical as well as a horizontal direction on the label (c) They usually take up a smaller space on a label (d) They may be printed on the bottom of a vial 32.7 (b) An MTA is a Material Transfer Agreement that it is used to document how a biospecimen may be used once it is transferred to its next destination

CHAPTER 35 35.1 The terminology used to describe organizations in the philanthropic or not-for-profit, nongovernment sector can be confusing because terms often lack a precise meaning, and the tax codes and regulations relating to nonprofit organizations may differ between countries. With that caveat, the following terminology and organizations are part of the lexicon of the US philanthropic sector that supports health/biomedical research and related training activities. Foundations • Foundations that support charitable activities by making grants to unrelated organizations or institutions. These include family or independent foundations such as the Bill and Melinda Gates Foundation (the largest US foundation) as well as company-sponsored or corporate foundations. • Operating foundations that run their own programs and services and typically do not provide much grant support to outside organizations. An example is the Kaiser Family Foundation. • Hybrid foundations that serve as a link between government agencies and the private sector. Public Charities: In contrast to foundations, public charities raise their funds from diverse sources, and they are often disease-focused and patient-centered organizations. They include organizations such as the American Cancer Society and the American Heart Association. Many of these organizations are also referred to as voluntary health organizations. The philanthropic sector supports a variety of activities related to biomedical research including providing training fellowships and research grants to investigators at different stages of their careers. Other organizations focus on supporting newly emerging research areas and providing “risk capital” to stimulate innovation or fill important “knowledge gaps.” Other areas supported by philanthropies include establishing product development partnerships,

758

APPENDIX 1 ANSWER KEY TO SUMMARY QUESTIONS

fostering the dissemination of information, data sharing, and patient engagement, and even advocating for resources and policy changes. 35.2 There are three major reasons why the philanthropic sector occupies a vital and distinctive niche in biomedical and health research. 1 Philanthropic sector organizations are often more flexible than government or for-profit organizations. They are freer to invest in highrisk research, investigate unconventional hypotheses, work on difficult and seemingly intractable problems that require long-term investments, and support research in areas with little potential for large financial returns such as research on very rare diseases. In contrast, government support for research is determined primarily by politicians who must consider many competing priorities and the greatest public good, avoid issues their constituencies deem too fractious, and for-profit companies need to consider the potential for timely financial gains. Nonprofit, nongovernment philanthropic organizations are not usually constrained by these pressures. 2 Philanthropic organizations are often less bureaucratic, enabling them to act quickly and flexibly in response to emerging opportunities and needs. This flexibility can facilitate product development. 3 Finally and perhaps most importantly, the breadth and diversity of philanthropic organizations in the United States, as well as in many other developed countries, provides a myriad of opportunities for those seeking funding, making it likely that with perseverance committed investigators and excellent institutions will find support for their work. 35.3 The list of organizations that have contributed to advancing biomedical or health research is long and varied. The following are a few of the many examples cited in this chapter. 1 The Rockefeller Foundation: • Helped implement some of the Flexner Report (which argued that medical education should become a more academic rigorous program grounded in basic science and research recommendations) by providing funds to modernize medical schools and build their faculty. • Established the first school of public health at Johns Hopkins University and launched two public health campaigns (to eradicate hookworm disease in the United States and to tackle yellow fever in the Western Hemisphere).

2 The National Foundation for Infantile Paralysis (now the March of Dimes) • Advocated for and cared for polio patients • Supported research that led to the polio vaccine • Laid the groundwork for launching the world’s biggest public health experimentdtesting the efficacy of the Salk vaccine in more than 1.8 million school children. 3 The Burroughs Wellcome Fund • Piloted and evaluated a bridging award shown to be valuable in facilitating the transition to a career as an independent investigator and providing physicianscientists with much needed protected time for research. After noting the success of the program, other philanthropic funders and the National Institutes of Health established bridging award programs as well. 4 The Whitaker Foundation • Contributed more than $700 million during its 30 year history, primarily to build the interdisciplinary field of biomedical engineering by supporting research, curriculum development, education programs, faculty hires, building construction, textbooks, and conferences, essentially establishing 30 departments of bioengineering at universities across the United States. This investment has been leveraged by many subsequent funders, including the NIH when it established the National Institute of Biomedical Imaging and Bioengineering (NIBIB). 5 The Doris Duke Charitable Foundation and the Howard Hughes Medical Institute • At the height of the HIV/AIDS epidemic in the province of KwaZulu-Natal (UKZN) South Africa, the Doris Duke Charitable Foundation partnered with the University of KwaZulu-Natal (UKZN) to build Doris Duke Medical Research Institute (DDMRI). DDMRI was the first stand-alone biomedical research institute in Durban, South Africa and housed a cadre of outstanding HIV/AIDS investigators. This seminal investment quickly attracted additional NIH support and paved the way for the Howard Hughes Medical Institute’s later investment that built and staffed the KwaZulu-Natal Research Institute for Tuberculosis and HIV (K-RITH) on the same campus. Together with significant investments by the Welcome Trust, KZN has emerged as an important research hub in southern Africa.

759

APPENDIX 1 ANSWER KEY TO SUMMARY QUESTIONS

CHAPTER 36 36.1 (b) Punitive approach to managing untoward events 36.2 (b) Failure mode and effects analysis 36.3 (c) Outcome measure 36.4 (c) Ernest Codman

CHAPTER 39 39.1 (b) Clinical practice, study management, and human subjects protection 39.2 (d) All of the above 39.3 (a) Managing the balance of clinical and research requirements for research participants 39.4 (c) The State Nurse Practice Act 39.5 (d) All of the above

CHAPTER 37 37.1 (d) Collect receptor-binding imaging data 37.2 (e) A, C, and D 37.3 (a) The animal-to-human test sequence

CHAPTER 38 38.1 (c) Genomic analysis of tissue specimens 38.2 (d) All of the above

CHAPTER 41 41.1 41.2 41.3 41.4 41.5

(b) On the record (a) Freedom of Information Act (e) All of the above (d) To include your scientific perspective (d) All of the above

Appendix 2 Acronyms Chapter 1 (John I. Gallin) A Historical Perspective on Clinical Research NIH

dcont’d

National Institutes of Health

PART IdETHICAL, REGULATORY, AND LEGAL ISSUES Chapter 2 (Christine Grady) Ethical Principles in Clinical Research

CRADA

Collaborative Research and Development Agreement

CSCE

Committee on Scientific Conduct and Ethics

DSMB

Data and Safety Monitoring Board

FDA

US Food and Drug Administration

FDAAA

US Food and Drug Administration Amendments Act

FOIA

Freedom of Information Act

ICMJE

International Committee of Medical Journal Editors

IOM

Institute of Medicine

IRP

Intramural Research Program

MTA

Material Transfer Agreements

NIAMS

National Institute of Arthritis and Musculoskeletal and Skin Diseases

CIOMS

Council for International Organizations of Medical Sciences

DHHS

Department of Health and Human Services

FDA

US Food and Drug Administration

GCP

Good Clinical Practice

NIH

National Institutes of Health

HIV

Human Immunodeficiency Virus

ORI

Office of Research Integrity

ICH

International Conference on Harmonization

PHS

Public Health Service

IRBs

Institutional Review Boards

PRIM&R

Public Responsibility in Medicine and Research

NIH

National Institutes of Health

NPRM

Notice of Proposed Rule Making

RCTs

Randomized Clinical Trials

REC

Research Ethics Committee

USCFR

US Code of Federal Regulations

WHO

World Health Organization

WMA

Chapter 4 (Julia Slutsman, Lynnette Nieman) Institutional Review Boards

World Medical Assembly

Chapter 3 (Melissa Colbert, Robert Nussenblatt, Michael Gottesman) Integrity in Research: Responsibilities and Prevention of Conflicts of Interest AAALAC

Association for Assessment and Accreditation of Laboratory Animal Care International

ACLU

American Civil Liberties Union

ACUC

Animal Care and Use Committee

CITI

Collaborative Institutional Training Initiatives

COI

Conflicts of Interest

AAHRPP

Association for the Accreditation of Human Research Protection Programs

ANPRM

Advance Notice of Propose Rule Making

CC

Clinical Center

CFR

Code of Federal Regulations

DHEW

Department of Health, Education and Welfare

DHHS

Department of Health and Human Services

DSMC

Data Safety and Monitoring Committee

FDA

US Food and Drug Administration

FWA

Federal-wide Assurances of Compliance

GAO

Government Accountability Office

HEW

Department of Health, Education, and Welfare

HIPAA

Health Insurance Portability and Accountability Act

HIV

Human Immunodeficiency Virus

HRPP

Human Research Protection Program Continued

761

762

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

IOM

Institute of Medicine

NDA

New Drug Application

IRB

Institutional Review Board

NIH

National Institutes of Health

IRP

Intramural Research Program

ODA

Orphan Drug Act

NIH

National Institutes of Health

OHRP

Office for Human Research Protections

PDUFA

Prescription Drug User Fee Act

PHI

Protected Health Information

PMC

Postmarket Commitments

PI

Principal Investigator

PMR

Postmarket Requirements

USDA

US Department of Agriculture

Chapter 5 (Elyse Summers and Michelle Feige) Accreditation of Human Research Protection Programs AAHRPP

Chapter 7 (Theresa Mullin) International Regulation of Drugs and Biological Products

Association for the Accreditation of Human Research Protection Programs

CTD

Common Technical Document

AAU

Association of American Universities

DRAs

Drug Regulatory Authorities

EC

Ethics Committee

eCTD

Electronic Standard of the CTD

HRPP

Human Research Protection Program

E

HRPPP

Human Research Participant Protection Programs

Efficacy (a criteria for approving and authorizing new medicinal products; ICH guideline)

IOM

Institute of Medicine

EC

European Commission

IRB

Institutional Review Board

EFPIA

European Federation of Pharmaceutical Industries and Associations

NBAC

National Bioethics Advisory Commission

EWG

Expert Working Group

NIH

National Institutes of Health

FDA

US Food and Drug Administration

GCG

Global Cooperation Group

ICH

International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use

ICH E6

E6 Good Clinical Practices

IFPMA

International Federation of Pharmaceutical Manufacturers and Associations

JPMA

Japanese Pharmaceutical Manufacturers Association

M

Multidisciplinary (a criteria for approving and authorizing new medicinal products; ICH guideline)

Chapter 6 (Molly Flannery, Amy McKee, Diane Maloney, Jonathan Jarow) The Regulation of Drugs and Biological Products by the Food and Drug Administration BLA

Biologics License Application

CBER

Center for Biologics Evaluation and Research

CDER

Center for Drug Evaluation and Research

CFR

Code of Federal Regulations

CMC

Chemistry, Manufacturing, and Control (information)

EOP2

End-of-phase-2

MedDRA

Medical Dictionary for Regulatory Activities

FDA

US Food and Drug Administration

FDAAA

US Food and Drug Administration Amendments Act

MHLW/ PMDA

Japanese Ministry of Health, Labor and Welfare/ Pharmaceutical and Medical Devices Agency

FDAMA

US Food and Drug Administration Modernization Act

MHRA

UK’s Medicines and Healthcare Products Regulatory Agency

FDASIA

US Food and Drug Administration Safety and Innovation Act

MRCTs

Multi-Regional Clinical Trials

FD&C

Federal Food, Drug, and Cosmetic Act

PhRMA

Pharmaceutical Research and Manufacturers of America

GCP

Good Clinical Practice

Q

IB

Investigator’s Brochure

Quality (a criteria for approving and authorizing new medicinal products; ICH guideline)

ICH

International Council on Harmonization

QT Interval

IND

Investigational New Drug

In cardiology, the QT Interval is a measure of the time between the start of the Q wave and the end of the T wave in the heart’s electrical cycle

IRB

Institutional Review Board

RHIs

Regional Health Initiatives

763

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

S

Safety (a criteria for approving and authorizing new medicinal products; ICH guideline)

WHO

World Health Organization

Chapter 8 (Christopher Olopade, Michelle Tagle, and Olufunmilayo Olopade) Clinical Research in International Settings: Opportunities, Challenges, and Recommendations

FDAMA

US Food and Drug Administration Modernization Act

HIV

Human Immunodeficiency Virus

HOPE

Health Omnibus Programs Extension

ICMJE

International Committee of Medical Journal Editors

IPD

Individual Participant Data

IRB

Institutional Review Board

MI

Myocardial Infarction

NCI

National Cancer Institute

NCT

ClinicalTrials.gov Identifying Number

NIH

National Institutes of Health

NLM

National Library of Medicine

NPRM

Notice of Proposed Rule making

OMB

Office of Management and Budget

AHRQ

Agency for Healthcare Research and Quality

CIOMS

Council for International Organizations of Medical Sciences

DSMP

Data Safety and Monitoring Plan

HICs

High-Income Countries

HIV

Human Immunodeficiency Virus

HIV/ AIDS

Human Immunodeficiency Virus/Acquired Immune Deficiency Syndrome

IRB

Institutional Review Board

PCORI

Patient-Centered Outcomes Research Institute

LMIC

Low- and Middle-Income Countries

PET

Positron Emission Tomography

MDG

Millennium Development Goal

PRS

Protocol Registration and Results System

MEPI

Medical Education Partnership Initiative

TRS

Three-level Trial Reporting System

NCD

Noncommunicable Diseases

VA

Department of Veterans Affairs

NHS

National Health Service (in the United Kingdom)

WHO

World Health Organization

NIH

National Institutes of Health

XML

Extensible Markup Language

SDG

Sustainable Development Goals

SEAR

South East Asia Region

SOP

Standard Operating Procedure

AE

Adverse Events

SSA

Sub-Saharan Africa

CDP

Coronary Drug Project

VOIP

Voice Over Internet Protocol

DSMB

Data and Safety Monitoring Board

WHO

World Health Organization

EMA

European Medicines Agency

WMA

World Medical Association

FDA

US Food and Drug Administration

IRB

Institutional Review Board

NIH

National Institutes of Health

RCTs

Randomized Controlled Trials

SAE

Serious Adverse Events

WHO

World Health Organization

Chapter 9 (Deborah Zarin, Rebecca Williams, Tony Tse, Nicholas Ide) The Role and Importance of Clinical Trials Registries AIDS

Acquired Immunodeficiency Syndrome

CDC

Centers for Disease Control and Prevention

CMS

Centers for Medicare and Medicaid Services

CONSORT

Consolidated Standards of Reporting Trials

CTTI

Clinical Trials Transformation Initiative

DHHS

Department of Health and Human Services

EMA

European Medicines Agency

EU

European Union

EudraCT

European Union Drug Regulating Authorities Clinical Trials database

FDA FDAAA

Chapter 10 (Paul Wakim and Pamela Shaw) Data and Safety Monitoring

Chapter 11 (Stephen Straus) Unanticipated Risk in Clinical Research ACTG

AIDS Clinical Trials Group

AIDS

Acquired Immunodeficiency Syndrome

AZT

Zidovudine

CMV

Cytomegalovirus

DNA

Deoxyribonucleic Acid

US Food and Drug Administration

DSMB

Data and Safety Monitoring Board

US Food and Drug Administration Amendments Act of 2007

FDA

US Food and Drug Administration Continued

764

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

FIAC

Fluoroiodoarabinosylcytosine or Fiacitabine

FY

Fiscal Year

FIAU

Fluoroiodoarauracil or Fialuridine

GAO

General Accounting Office

HBV

Hepatitis B Virus

IOM

Institute of Medicine

HIV

Human Immunodeficiency Virus

IRB

Institutional Review Board

IOM

Institute of Medicine

LDL

Low-Density Lipoprotein

IRB

Institutional Review Board

MESA

Multi-Ethnic Study of Atherosclerosis

NIH

National Institutes of Health

MHT

Menopausal Hormone Therapy

ORI

Office of Research Integrity

NIH

National Institutes of Health

OSI

Office of Scientific Integrity

OMB

Office of Management and Budget

OTC

Ornithine Transcarbamylase

ORWH

Office of Research on Women’s Health

PHS

Public Health Service

SBP

Systolic Blood Pressure

STI

Sexually Transmitted Infection

WHI

Women’s Health Initiative

WISE

Women’s Ischemia Syndrome Evaluation

Chapter 12 (Valerie Bonham) Legal Issues in Clinical Research APA

Administrative Procedure Act

CFR

Code of Federal Regulations

DHHS

Department of Health and Human Services

DoD

Department of Defense

DPOA

Durable Power of Attorney

FDA

US Food and Drug Administration

AARP

American Association of Retired People

FDAAA

US Food and Drug Administration Amendments Act

NCI

National Cancer Institute

FOIA

Freedom of Information Act

NIH

National Institutes of Health

hESC

Human Embryonic Stem Cell

HIPAA

Health Insurance Portability and Accountability Act

IRB

Institutional Review Board

LAR

Legally Authorized Representative

LDS

Limited Data Set

NIH

National Institutes of Health

BIA

Bioelectrical Impedance Analysis

OHRP

Office of Human Research Protection

CASS

Coronary Artery Surgery Study

PHI

Protected Health Information

CAST

Cardiac Arrhythmia Suppression Trial

PI

Principal Investigator

CBT

Cognitive Behavior Therapy

SORN

System of Records Notice

CHF

Chronic Heart Failure

CONSORT

Consolidated Standards of Reporting Trials

DSMB

Data Safety Monitoring Board

HRT

Hormone Replacement Therapy

IRB

Institutional Review Board

Chapter 13 (Janine Clayton and Juliana Blome) National Institutes of Health Policy on the Inclusion of Women and Minorities as Subjects in Clinical Research

Chapter 14 (Jerry Sachs) Clinical Research: A Patient Perspective

PART IIdSTUDY DESIGN AND BIOSTATISTICS Chapter 15 (Catherine Stoney and Laura Lee Johnson) Development and Conduct of Studies

CDC

Centers for Disease Control and Prevention

MI

Myocardial Infarction

CFR

Code of Federal Regulations

MOO

Manual of Operating Procedures

DHHS

Department of Health and Human Services

MOP

Manual of Procedures

DPP

Diabetes Prevention Program

MOST

Multiphasic Optimization Strategy

FDA

US Food and Drug Administration

SHEP

Systolic Hypertension in the Elderly Program

765

APPENDIX 2 ACRONYMS

Chapter 16 (Elizabeth Bartrum and Barbara Karp) Writing a Protocol

dcont’d ICH

International Conference on Harmonization

ISIS-4

Fourth International Study of Infarct Survival

JITAI

Just-In-Time-Adaptive Interventions

MDD

Major Depressive Disorder

MI

Myocardial Infarction

MOST

Multiphase Optimization Strategy

MS

Multiple Sclerosis

MTD

Maximum Tolerated Dose

MRI

Magnetic Resonance Imaging

NCI CTCAE

National Cancer Institute Common Toxicity Criteria for Adverse Events

NINDS

National Institute of Neurological Disorders and Stroke

PRECIS

Pragmatic Explanatory Continuum Indicator Summary

Quality Assurance

RRMS

Relapsing-Remitting Multiple Sclerosis

Unanticipated Problems

RCT

Randomized Controlled Trials

SMART

Sequential, Multiple Assignment, Randomized Trial

WHI

Women’s Health Initiative

CRADA

Cooperative Research and Development Agreements

CTA

Clinical Trial Agreements

DoD

Department of Defense

DPA

Durable Power of Attorney

DSMB

Data and Safety Monitoring Board

DSMC

Data and Safety Monitoring Committee

DTA

Data Transfer Agreements

EKG

Electrocardiogram

FDA

US Food and Drug Administration

IRB

Institutional Review Board

MOU

Memo of Understanding

MTA

Material Transfer Agreements

NIH

National Institutes of Health

QA UP

Chapter 17 (Laura Lee Johnson) Design of Observational Studies CHD

Coronary Heart Disease

CONSORT

Consolidated Standards of Reporting Trials

DSMB

Data and Safety Monitoring Board

HIV/AIDS

Human Immunodeficiency Virus/Acquired Immune Deficiency Syndrome

AHQR

Agency for Healthcare Research and Quality

HRT

Hormone Replacement Therapy

ARRA

MMWR

Morbidity and Morality Weekly Report

The American Recovery and Reinvestment Act (“Stimulus Package”)

MRI

Magnetic Resonance Imaging

CDM

Common Data Model

OR

Odds Ratio

CDRNs

Clinical Data Research Networks

OSMB

Observational Study Monitoring Board

CER

Comparative Effectiveness Research

RCT

Randomized Controlled Trial

CMS

Centers for Medicare and Medicaid Services

RR

Relative Risk or Risk Ratio

CT

CAT Scans

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

DEcIDE

Developing Evidence to Inform Decisions about Effectiveness

WHI

Women’s Health Initiative

DHHS

Department of Health and Human Services

EHR

Electronic Health Records

ESA

Erythropoiesis Stimulating Agents

FDA

US Food and Drug Administration

HCV

Hepatitis C

HITECH

Health Information Technology for Economic and Clinical Health

Chapter 18 (Catherine Stoney and Laura Lee Johnson) Design of Clinical Trials and Studies

Chapter 19 (Joe Selby, Evelyn Whitlock, Kelly Sherman, and Jean Slutsky) The Role of Comparative Effectiveness Research

CONSORT

Consolidated Standards of Reporting Trials

CHD

Coronary Heart Disease

DLT

Dose Limiting Toxicity

HTA

Health Technology Assessments

HRT

Hormone Replacement Therapy

IOM

Institute of Medicine Continued

766

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

IPD

Individual Participant Data

FDA

US Food and Drug Administration

IPTW

Inverse Probability of Treatment Weighting

FIM

Functional Impact Measure

IRB

Institutional Review Boards

FLI

Functional Life Index

IVs

Instrumental Variables

GRC

Global Rating of Change

MMA

Medicare Prescription Drug, Improvement and Modernization Act (Medicare Modernization Act)

HRQL

Health-Related Quality of Life

IADL

Instrumental Activity of Daily Living

MS

Multiple Sclerosis

ICF

NIH

National Institutes of Health

International Classification of Functioning, Disability and Health; International Classification of Function

OMOP

Observational Medical Outcomes Partnership

ICIDH

PCORI

Patient-Centered Outcomes Research Institute

International Classification of Impairments, Disabilities and Health

PICOTS

Population/Patients, Interventions, Comparators, Outcomes, Timing, Setting

IRT

Item Response Theory

MOS

Medical Outcomes Study

PPRNs

Patient-Powdered Research Networks

NCHS

National Center for Health Statistics

PROMIS

Patient Reported Outcomes Measurement Information System

NHANES

National Health and Nutrition Examination Survey

NIH

National Institutes of Health

SMART

Sequential Multiple Assignment Randomized Trials

PCORI

Patient-Centered Outcomes Research Institute

USPSTF

United States Preventive Services Task Force

PCORNET

Patient-Centered Clinical Research Network

PRO/SRO

Patient Report Outcomes/Self-Reported Outcomes

PROMIS

Patient Reported Outcomes Measurement System

PRO

Patient Report Outcomes

QALY

Quality Adjusted Life Years

QOL

Quality of Life

QOLI

Quality of Life Index

SIP

Sickness Impact Profile

SR

Systematic Reviews

WHO

World Health Organization

Chapter 20 (Leighton Chan, Patrick McGarey, and Joseph Sclafani) Using Large Data Sets for Population-Based Health Research CPT

Current Procedural Terminology

her

Electronic Health Record

HMO

Health Maintenance Organization

ICD-10

International Classification of Diseases

IRB

Institutional Review Board

JAMA

Journal of the American Medical Association

SEER

Surveillance Epidemiology and End Results

UNOS

United Network for Organ Sharing

Chapter 21 (Lynn Gerber and Jillian Kallman Price) Measures of Function and Health-Related Quality of Life

Chapter 22 (Junfeng Sun, Bradley Freeman, Charles Natanson) Meta-analysis of Clinical Trials CI

Confidence Intervals

OR

Odds Ratio

PICO

Participants, Interventions, Comparators and Outcomes

ADL

Activity of Daily Living

AIMS

Arthritis Impact Measurement Scales

CDC

Centers for Disease Control and Prevention

PRISMA-P 2015

Preferred Reporting Items for Systematic Reviews and Meta-analyses for Protocols 2015

CID

Clinically Important Difference

QUOROM

Quality of Reporting of Meta-analyses

CLD

Chronic Liver Disease

RR

Relative Risks

CLDQ

Chronic Liver Disease Questionnaire

TNF

Tumor Necrosis Factor

EORTC

European Organization for Research and Treatment of Cancer

767

APPENDIX 2 ACRONYMS

Chapter 23 (Pamela Shaw, Laura Lee Johnson, and Craig Borkowf) Issues in Randomization

Chapter 26 (Laura Lee Johnson) An Introduction to Survival Analysis K-M

KaplaneMeier

MI

Myocardial Infarction

BMI

Body Mass Index

CONSORT

Consolidated Standards of Reporting Trials

CPMP

Committee for Proprietary Medicinal Products

CVD

Cardiovascular Disease

DSMB

Data Safety Monitoring Board

Chapter 27 (Pamela Shaw, Laura Lee Johnson, and Michael Prochan) Intermediate Topics in Biostatistics

ECMO

Extracorporeal Membrane Oxygenators

ACTG

AIDS Clinical Trial Group

EMA

European Medicines Agency

ART

Antiretroviral Therapy

FDA

US Food and Drug Administration

AUC

Area Under The Curve

ICH

International Conference on Harmonization

BOCF

Baseline Observation Carried Forward

ITT

Intent-to-Treat

BMD

Bone Mineral Density

MRC

British Medical Research Council

CC

Complete-Case Analysis

UGDP

University Group Diabetes Program

CONSORT

Consolidated Standards of Reporting Trials

WHI

Women’s Health Initiative

CP

Conditional Power

DSMB

Data and Safety Monitoring Board

Egfr

Estimated Glomerular Filtration Rate

FPG

Fasting Plasma Glucose

Chapter 24 (Laura Lee Johnson, Craig Borkowf, and Pamela Shaw) Hypothesis Testing ANOVA

Analysis of Variance

FDA

US Food and Drug Administration

CI

Confidence Interval

FDR

False Discovery Rate

DF

Degrees of Freedom

fMRI

Functional Magnetic Resonance Imaging

FDR

False Discovery Rate

FPF

False Positive Fraction

IACUC

Institutional Animal Care and Use Committee

GEE

Generalized Estimating Equations

IRB

Institutional Review Board

GFR

Glomerular Filtration Rate

ITT

Intention-To-Treat

GWAS

Genome Wide Association Studies

MI

Myocardial Infarction

HIV

Human Immunodeficiency Virus

MRI

Magnetic Resonance Imaging

HRT

Hormone Replacement Therapy

MS

Multiple Sclerosis

ITT

Intent-to-Treat

RRMS

Relapsing-Remitting Multiple Sclerosis

IV

Intravenous

SPIRIT

Standard Protocol Items: Recommendations for Interventional Trials

K-M

KaplaneMeier

LRCF

Last Rank Carried Forward

WHI

Women’s Health Initiative

LOCF

Last Observation Carried Forward

Chapter 25 (Craig Borkowf, Laura Lee Johnson, and Paul Albert) Power and Sample Size Calculations

MAR

Missing at Random

MCAR

Missing Completely at Random

MI

Myocardial Infarction

GLMs

Generalized Linear Models

MNAR

Nonignorable or Missing Not at Random

GRTs

Group Randomized Trials

MRI

Magnetic Resonance Imaging

HCM

Hypertrophic Cardiomyopathy

MTD

Maximally Tolerated Dose

HIV

Human Immunodeficiency Virus

NIH

National Institutes of Health

ICC

Intraclass Correlation

NPV

Negative Predictive Value Continued

768

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

NRC

National Research Council

ICH

International Conference on Harmonization

oSOC

Optimized Standard of Care

IRB

Institutional Review Board

pAUC

Partial Area Under the Curve

ISIS-1

First International Study of Infarct Survival

PLCO

Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial

ISIS-4

Fourth International Study of Infarct Survival

IVRS

Interactive Voice Randomization Service

PPV

Positive Predictive Value

MDIC

Medical Device Innovation Consortium

PREVAIL II

Partnership for Research on Ebola Virus in Liberia II

MID

Minimally Important Clinical Difference

ROC

Receiver Operator Characteristic Curve

NEST

rROC

Relative Receiver Operator Characteristic Curve

National Evaluation System for healthcare Technology

RTM

Regression to the Mean

NIH

National Institutes of Health

SNP

Single-Nucleotide Polymorphism

NNH

Number Needed to Harm

TPF

True Positive Fraction

NNT

Number Needed to Treat

TPR, FPR

True and False Positive Rates

OHRP

Office for Human Research Protections

WHI CT

Women’s Health Initiative Clinical Trial

PCORI

Patient-Centered Outcomes Research Institute

PRAISE

Prospective Randomized Amlodipine Survival Evaluation trial

PRECIS

Pragmatic Explanatory Continuum Indicator Summary

QbD

Quality by Design

RE-LY

Randomized Evaluation of Long-Term Anticoagulation Therapy trial

Chapter 28 (Robert Califf) Large Clinical Trials and Research Institutes

Registries: Clinical

ACC

American College of Cardiology

ACEIs

Angiotensin-Converting Enzyme Inhibitors

AHA

American Heart Association

T2DM

Type 2 Diabetes Mellitus

AIDS

Acquired Immunodeficiency Syndrome

SUPPORT

aPTT

Activated Partial Thromboplastin Time

Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments

ASSENT-2

Assessment of the Safety and Efficacy of a New Thrombolytic study

BARI

Bypass Angioplasty Revascularization Investigation

BEsT

Biomarkers, Endpoints, and other Tools

CFF

Cystic Fibrosis Foundation

COX

Cyclooxygenase

CROs

Contract Research Organizations

CTTI

Clinical Trials Transformation Initiative

ARIPO

African Regional Intellectual Property Organization

DCC

Data Coordinating Center

BrIDGs

Bridging Interventional Development Gaps

DSMC

Data and Safety Monitoring Committee

BW

Burroughs-Welcome

EHR

Electronic Health Records

CAFC

Court of Appeals for the Federal Circuit

FDA

US Food and Drug Administration

CCPA

Court of Customers and Patent Appeals

FTT

Fibrinolytic Therapy Trialists

CDA

Confidential Disclosure Agreement

GCP

Good Clinical Practice

CDC

Centers for Disease Control and Prevention

GUSTO-I trial

Global Utilization of Streptokinase and rt-PA for Occluded Coronary Arteries trial

CDP

Commercial Development Plan

HER-2

Human Epidermal Growth Factor Receptor

cGMP

Current Good Manufacturing Practice

HIV

Human Immunodeficiency Virus

CIP

Continuation-In-Part

CON

Continuation Application

PART IIIdTECHNOLOGY TRANSFER, DATA MANAGEMENT, AND SOURCES OF FUNDING SUPPORT FOR RESEARCH Chapter 29 (Bruce Goldstein) Intellectual Property and Technology Transfer

769

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

CRADA

Cooperative Research and Development Agreement

ORTA

Offices for Research and Technology Applications

CTA

Clinical Trial Agreement

OTT

Office of Technology Transfer

DEC

Determination of Exceptional Circumstances

PCT

Patent Cooperation Treaty

DIV

Divisional

PGR

Post-Grant Review

DJ

Declaratory Judgment

PI

Principal Investigator

DNA

Deoxyribonucleic Acid

PPH

Patent Prosecution Highway

DoE

Department of Energy

PRV

Provisional Application

EAPO

Eurasian Patent Organization

RCE

Request for Continued Examination

EPA

Environmental Protection Agency

RePORT

Research Portfolio Online Reporting Tools

EPC

European Patent Convention

RNA

Ribonucleic Acid

EPO

European Patent Office

RP

Research Plan

FAOM

First Office Action on the Merits

SIR

Statutory Invention Registration

FDA

US Food and Drug Administration

SPC

Supplementary Protection Certificate

FOIA

Freedom of Information Act

TRIPS

FTTA

Federal Technology Transfer Act

Agreement to Trade-Related Aspects of Intellectual Property Rights

GATT

General Agreement on Tariffs and Trade

TTCA

Technology Transfer Commercialization Act

GCC

Gulf Cooperation Council

USDA

US Department of Agriculture

GDP

Gross Domestic Product

UMBTA

Uniform Biological Material Transfer Agreement

IC

Institute or Center

USPTO

US Patent and Trademark Office

IDS

Information Disclosure Statement

WIPO

World Intellectual Property Organization

iEdison

Interagency Edison System

WTO

World Trade Organization

IND

Investigational New Drug

IP

Intellectual Property

IPER

International Preliminary Examination Report

AE

Adverse Event

IPR

Inter Partes Review

AAHRPP

ISR

International Search Report

Association for the Accreditation of Human Research Protection Programs

ITC

International Trade Commission

caDSR

Cancer Data Standards Registry and Repositories

JHU

Johns Hopkins University

CAPA

Corrective and Preventive Action Plan

LOI

Letter of Intent

CBIIT

Center for Bioinformatics and Information Technology

MTA

Material Transfer Agreement

CDE

Common Data Elements

NASA

National Aeronautics and Space Administration

CFR

Code of Federal Regulations

NCATS

National Cybersecurity Assessment and Technical Services

CRA

Clinical Research Associate

NCI

National Cancer Institute

CRF

Case Report Form

NIH

National Institutes of Health

CRO

Contract Research Organizations

NPE

Nonpracticing Entities

CTMS

Clinical Trials Management Systems

NTTAA

National Technology Transfer Advancement Act

CTN

Clinical Trials Nurse

OAMPI

African and Malagasy Patent Rights Authority

CTCAE

Common Terminology Criteria for Adverse Events

OAPI

Organisation Africaine de la Proprie´te´ Intellectuelle

CTEP

Cancer Therapy Evaluation Program

DBA

Database Administrator

Chapter 30 (Diane St. Germain, Marjorie Good) Data Management in Clinical Trials

Continued

770

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

DHHS

Department of Health and Human Services

DMP

Data Management Plan

DSMC

Data Safety and Monitoring Committee

EMR

Electronic Medical Record

FDA

US Food and Drug Administration

GCP

Good Clinical Practices

HIPAA

Health Insurance Portability and Accountability Act

ICH

International Conference on Harmonization

IND

Investigational New Drug

IRB

Institutional Review Board

IT

Information Technology

MedDRA

Medical Dictionary for Regulatory Activities

NCI

National Cancer Institute

NIH

National Institutes of Health

OHRP

Office of Human Subjects Protections

PI

Principal Investigator

1D

One-dimensional Scannable Barcode

PRO-CTCAE

Patient Reported OutcomesdCommon Terminology Criteria for Adverse Events

2D

Two-dimensional Scannable Barcode

SAE

Serious Adverse Events

ASCP

American Society for Clinical Pathology

SOPs

Standard Operating Procedures

BBRB

Biorepositories and Biospecimen Research Branch

UPs

Unanticipated Problems

BRN

Biospecimen Research Network

VPN

Virtual Private Network

CAP

College of American Pathology

CTRNet

Canadian Tissue Repository Network

DHHS

Department of Health and Human Services

DNA

Deoxyribonucleic Acid

IATA

International Air Transportation Association

ID

Identifier

IRBs

Institutional Review Boards

ISBER

International Society for Biological and Environmental Repositories

MTA

Material Transfer Agreement

NCI

National Cancer Institute

NIH

National Institutes of Health

OBBR

Office of Biorepositories and Biospecimen Research

Chapter 31 (James Cimino) Clinical Research Data: Characteristics, Representation, Storage, and Retrieval

ICD-9-CM

International Classification of Diseases, Ninth Edition, with Clinical Modifications

ISO

International Standards Organization

LOINC

Logical Object Identifier Name and Codes

MedRA

Medical Dictionary for Regulatory Activities

NCATS

National Center for Advancing Translational Science

NCBI

National Center for Biotechnology Information

NIH

National Institutes of Health

NLM

National Library of Medicine

REDCap

Research Electronic Data Capture

SHRINE

Shared Health Research Information Network

SNOWMED-CT

Systematized Nomenclature of MedicinedClinical Terms

Chapter 32 (Karen Berliner and Amy Skubitz) Management of Patient Samples

ACT

Advanced Clinical Trials

BRIDG

Biomedical Research Integrated Domain Group

CDA

Clinical Document Architecture

CDISC

Clinical Data Interchange Standards Consortium

CTDMs

Clinical Trials Data Management Systems

CTSA

Clinical and Translational Science Awards

dbGAP

Database of Genome and Phenome

EHRs

Electronic Health Records Systems

FDA

US Food and Drug Administration

GO

Gene Ontology

OCT

Optimal Cutting Temperature

HL7

Health Level Seven

RNA

Ribonucleic Acid

i2b2

Informatics for Integrating Biology and the Bedside

SOPs

Standard Operating Procedures

SPREC

Sample PREanalytic Code

771

APPENDIX 2 ACRONYMS

Chapter 33 (Phyllis Klein) Evaluating a Protocol Budget

dcont’d OHRP

Office for Human Research Protections

PA

Program Announcement

PDF

Portable Document Format

PD/PI

Project Director/Principal Investigator

PHS

Public Health Service

PO

Program Officer

POR

Patient-Oriented Research

RFA

Request for Applications

R&R

Research and Related (application form)

SO

Signing Official

SRG

Scientific Review Group (also called study sections)

SRO

Scientific Review Officer

CAP

College of American Pathologists

CLIA

Clinical Laboratory Improvement Amendments

CPT

Current Procedural Terminology

CRF

Case Report Form

CV

Curriculum Vitae

ECG

Electrocardiogram

ERB

Ethics Review Board

FDA

US Food and Drug Administration

ICF

Informed Consent Form

IEC

Independent Ethics Committee

IND

Investigational New Drug

IRB

Institutional Review Board

MRI

Magnetic Resonance Imaging

MSP

Medicare’s Secondary Payer

NIH

National Institutes of Health

PI

Principal Investigator

AACR

American Association for Cancer Research

PK

Pharmacokinetics

ACS

American Cancer Society

RE

Research Ethics Board

ACT UP

AIDS Coalition to Unleash Power

SOC

Standard of Care

AID

Agency for International Development

AIDS

Acquired Immunodeficiency Syndrome

Chapter 34 (Valerie Prenger) Getting the Funding You Need to Support Your Research: Navigating the National Institutes of Health Peer Review Process

Chapter 35 (Elaine Gallin, Maryrose Franko, and Enriqueta Bond) Philanthropy’s Role in Advancing Biomedical Research

ALS

Amyotrophic Lateral Sclerosis

amfAR

Foundation for AIDS Research

BWF

Burroughs Wellcome Fund

CDC

Centers for Disease Control and Prevention

CF

Cystic Fibrosis

CFF

Cystic Fibrosis Foundation

CFFT

Cystic Fibrosis Foundation Therapeutics, Inc.

CSO

Common Scientific Outline

DDCF

Doris Duke Charitable Foundation

DDMRI

Doris Duke Medical Research Institute

DELTAS

Developing Excellence in Leadership, Training, and Science

ASSIST

Application Submission System and Interface for Submission Tracking

CSR

Center for Scientific Review

DHHS

Department of Health and Human Services

eRA

Electronic Research Administration (Commons)

ESI

Early-Stage Investigator

FOA

Funding Opportunity Announcement

GCP

Good Clinical Practice

IACUC

Institutional Animal Care and Use Committee

IC

Institute/Center

DNDi

Drugs for Neglected Disease Initiative

IRB

Institutional Review Board

EEG

Electroencephalogram

LRP

Loan Repayment Program

EGPAF

Elizabeth Glaser Pediatric AIDS Foundation

NCI

National Cancer Institute

FDA

US Food and Drug Administration

NHLBI

National Heart, Lung, and Blood Institute

FNIH

Foundation for the National Institutes of Health

NIAID

National Institute of Allergy and Infectious Diseases

HCV

Hepatitis C virus

NIH

National Institutes of Health

HHMI

Howard Hughes Medical Institute Continued

772

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

HIV

Human Immunodeficiency Virus

IRBs

Institutional Review Boards

HRA

Health Research Alliance

NIH CC

National Institutes of Health Clinical Center

ICRP

International Cancer Research Partnership

ORS

Occurrence Reporting System

KHN

Kaiser Health News

SARS

Severe Acute Respiratory Syndrome

K-RITH

KwaZulu-Natal Research Institute for Tuberculosis and HIV

LLCs

Limited Liability Corporations

MIT

Massachusetts institute of Technology

MMRF

Multiple Myeloma Research Foundation

MMV

Medicines for Malaria Venture

NBCC

National Breast Cancer Coalition

NIBIB

National Institute of Biomedical Imaging and Bioengineering

NIH

National Institutes of Health

NLM

National Library of Medicine

NSF

National Science Foundation

PDPs

Product Development Partnerships

PLoS

Public Library of Science

PMC

PubMed Central

PPRs

Patient-powered registries

PPRNs

Patient-powered research networks

RFA

Request For Applications

SCDA

Simons Center for Data Analysis

SFARI

Simons Foundation Autism Research Initiative

SU2C

Stand Up to Cancer

TAP

Therapy Acceleration Program

UCLA

University of CaliforniadLos Angeles

UCSF

University of CaliforniadSan Francisco

UK

United Kingdom

UKZN

University of KwaZulu-Natal

VHO

Voluntary Health Organizations

Chapter 37 (Sue Cheng, Konstantina Vanevski, and Juan J.L. Lertora) Clinical Pharmacology and its Role in Pharmaceutical Development ACCP

American College of Clinical Pharmacology

ADME (processes)

Absorption, Distribution, Metabolism, and Excretion

AE

Adverse Events

AMQ

Animal Model Qualification (program)

ARIA

Active Postmarket Risk Identification and Analysis (system)

ASCPT

American Society for Clinical Pharmacology and Therapeutics

ASPE

Assistant Secretary for Planning and Evaluation

AUC

Area Under the Curve

BPCA

Best Pharmaceuticals for Children Act

CHMP

Committee for Medicinal Products for Human Use

CL

Clearance

CMS

Centers for Medicare and Medicaid Services

COA

Clinical Outcome Assessments

CPI

Critical Path Initiative

CTA

Clinical Trial Application

CTTI

Clinical Trials Transformation Initiative

DAPT

Dual Antiplatelet Therapy

DNA

Deoxyribonucleic Acid

EMA

European Medicines Agency

EU

European Union

FDA

US Food and Drug Administration

FDAAA

US Food and Drug Administration Amendments Act

FIH

First In Human

HED

Human Equivalent Dose

IND

Investigational New Drug

Chapter 36 (Laura Lee and David Henderson) Identifying, Understanding, and Managing Patient Safety and Clinical Risks in the Clinical Research Environment

I SPY 2 TRIAL

Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2

MIST

Metabolites in Safety Testing (committee)

FMEA

Failure Mode and Effects Analysis

MRSD

Maximum Recommended Starting Dose

HROs

High Reliability Organizations

NDA

New Drug Application

PART IVdCLINICAL RESEARCH INFRASTRUCTURE

773

APPENDIX 2 ACRONYMS

dcont’d

dcont’d

NOAEL

No Observed Adverse Events Level

CRO

Contract Research Organization

PAD

Pharmacologically Active Dose

CTSA

Clinical and Translational Science Award

PBPK

Physiologically Based Pharmacokinetic Modeling

DIA

Drug Information Association

PET

Positron Emission Tomography

FDA

US Food and Drug Administration

PK/PD

Pharmacokinetic/Pharmacodynamic

GCP

Good Clinical Practices

PREA

Pediatric Research Equity Act

GCRC

General Clinical Research Center

PRISM

Postmarket Rapid Immunization Safety Monitoring (system)

IACRN

International Association of Clinical Research Nurses

IRB

Institutional Review Board

RA

Rheumatoid Arthritis

NCRR

National Center for Research Resources

RCCT

Randomized Concentration Controlled Trial

NIH

National Institutes of Health

RNA

Ribonucleic Acid

SoCRA

Society of Clinical Research Associates

RTCs

Randomized Controlled Trials

TJC

The Joint Commission

SPECT

Single Photon Emission Computed Tomography

Chapter 38 (Frederick P. Ognibene) Career Paths in Clinical Research ANA

American Neurological Association

CRCA

Clinical Research Curriculum Award

CRTP

Chapter 40 (Jon McKeeby and Patricia Coffey) The Importance and Use of Electronic Health Records in Clinical Research ALCOA

(Data that is) Attributable, Legible, Contemporaneous, Original, and Accurate

Clinical Research Training Program

BTRIS

Biomedical Translational Research Information System

CTSA

Clinical and Translational Science Award

CDS

Clinical Decision Support

FDA

US Food and Drug Administration

CFR

Code of Federal Regulations

HHMI

Howard Hughes Medical Institute

CHIP

Children’s Health Insurance Program

IPPCR

Introduction to the Principles and Practice of Clinical Research

CMS

Centers for Medicare and Medicaid Services

CPOE

Computerized Provider Order Entry

MRSP

Medical Research Scholars Program

CRDW

Clinical Research Data Warehouse

MSTP

Medical Scientist Training Program

CRIS

Clinical Research Information Systems

NCRR

National Center for Research Resources

CTDMS

Clinical Trials Data Management System

NIAID

National Institute of Allergy and Infectious Diseases

EDC

Electronic Data Capture

NIH

National Institutes of Health

EMR

Electronic Medical Record

EHR

Electronic Health Record

HIPAA

Health Insurance Portability and Accountability Act

HITECH

Health Information Technology for Economic and Clinical Health Act

Chapter 39 (Gwenyth Wallen and Cheryl Fisher) Clinical Research Nursing: A New Domain of Practice ACPU

Association of Clinical Pharmacology Units

ACRP

Association of Clinical Research Professionals

HL7

Health Level 7

ANA

American Nurses Association

HL7-ADT

Health Level 7dAdmit Discharge Transfer Message

ANCC

American Nurses Credentialing Center

HL7-ORM

Health Level 7dOrder Message

CRA

Clinical Research Associates

HL7-ORU

Health Level 7dResults Message

CRC

Clinical Research Coordinators

HL7 -MDM

Health Level 7dDocumentation Message

CRN2010

Clinical Research Nursing2010

IRB

Institutional Review Board Continued

774

APPENDIX 2 ACRONYMS

dcont’d IV

Intravenous Therapy

LIMS

Laboratory Information Management System

LIS

Laboratory Information System

MACRA

Medicare Access and Children’s Health Insurance Program (CHIP) Reauthorization Act

NIH CC

National Institutes of Health Clinical Center

ONC

Office of the National Coordinator for Health Information Technology

RIS

Radiology Information System

Chapter 41 (John Burklow) The Clinical Research and the Media

Chapter 42 (Josh Duberman and Pamela Sieving) Informational Resources for the Clinical Researcher AHFS

American Society of Health-System Pharmacists

AHRQ

Agency for Healthcare Research and Quality

CINAHL

Cumulative Index to Nursing and Allied Health Literature

DOAJ

Directory of Open Access Journals

FDA

US Food and Drug Administration

HINARI

The Health Internetwork Access to Research Initiative

INASP

International Network for the of Scientific Publications

MeSH

Medical Subject Headings

NCBI

National Center for Biotechnology Information

NIH

National Institutes of Health

NLM

National Library of Medicine

NN/LM

National Network of Libraries of Medicine

PICO

Population/Problem/Patient; Intervention/Indicator; Comparison; Outcome

AIDS

Acquired Immune Deficiency Syndrome

AP

Associated Press

CT

Computed Tomography (CAT or CT scan)

FOIA

Freedom of Information Act

HESC

Human Embryonic Stem Cell

PMC

PubMed Central

IC

Institute/Center

RePORTER

JAMA

Journal of the American Medical Association

Research Portfolio Online Reporting Tools Expenditures and Results

NEJM

New England Journal of Medicine

TOXLINE

Toxicology Literature Online

NIH

National Institutes of Health

TOXNET

Toxicology Data Network

SPRINT

Systolic Blood Pressure Intervention Trial

WHO

World Health Organization

Index

‘Note: Page numbers followed by “f” indicate figures, “t” indicate tables and “b” indicate boxes.’

A AAALAC guidelines. See Association for Assessment and Accreditation of Laboratory Animal Care International guidelines (AAALAC guidelines) AACR. See American Association for Cancer Research (AACR) AAHRPP. See Association for Accreditation of Human Research Protection Programs (AAHRPP); Association for the Accreditation of Human Research Protection Programs (AAHRPP) AARP Inc. See American Association of Retired Persons (AARP Inc.) AAU. See Association of American Universities (AAU) Abacavir, 650, 651t Abbott study, 492 Abciximab, 419 “About Grants” Page, 609 “Absolute novelty” jurisdictions, 461 Absolute risk, 243 Absorption, distribution, metabolism, and excretion (ADME), 649 Abu al-Qasim al-Zahrawi, 3 Abusive tactics, 482e483 ACC/AHA. See American College of Cardiology/American Heart Association (ACC/AHA) Access models, 721e723 Accreditation, 58, 63e64 accreditation-eligible organizations, 63 human research protection programs, 65 principles, 64e65 AAHRPP expecting from organizations, 64 organizations expecting from AAHRPP, 64e65 standards, 66e70 Institutional Review Board or Ethics Committee domain, 67e69 organization domain, 66e67 researcher and research staff domain, 69e70 steps to, 70 training and, 561e562 value of, 70e72 Accuracy measures, 396e398 ACEIs. See Angiotensin-converting enzyme inhibitors (ACEIs)

ACLU. See American Civil Liberties Union (ACLU) Acquired immunodeficiency syndrome (AIDS), 113, 144 ACRP. See Association of Clinical Research Professionals (ACRP) ACS. See American Cancer Society (ACS); American Chemical Society (ACS) ACT. See Advanced Clinical Trials (ACT) ACTG. See AIDS Clinical Trials Group (ACTG) Activated partial thromboplastin time (aPTT), 424e425, 432 Activity of daily living (ADL), 310 ACUC. See Animal Care and Use Committee (ACUC) Acute hemorrhage, 548 ADA. See Americans with Disabilities Act (ADA) Adaptive design clinical trials, 280e281 Adaptive designs, 387e389, 653 Adaptive intervention treatment. See Adaptive treatment Adaptive randomization methods, 336e337 Adaptive treatment, 257 designs, 257 Additional specialized resources, 724be743b Adherence, 214e215 ADL. See Activity of daily living (ADL) ADME. See Absorption, distribution, metabolism, and excretion (ADME) Administrative censoring, 374 Administrative data, 116 Administrative Procedure Act (APA), 167 Advance directives, 163e164 Advance Notice of Propose Rule Making (ANPRM), 57e58 Advanced Clinical Trials (ACT), 555 Adverse effects, 719 Adverse events (AEs), 130, 224e225, 540, 648e649 monitoring and reporting, 540e542 AEC. See Asian Economic Community (AEC) AEs. See Adverse events (AEs) Afatinib, 650, 651t African Americans underrepresented in clinical trials, 192e193

775

African Regional Intellectual Property Organization (ARIPO), 452 Agency for Healthcare Research and Quality (AHRQ), 270, 275, 278, 286, 724be743b Agency for International Development (AID), 624 Aggregation bias, 321 Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), 452e453 Agreements to protect confidentiality, 506e509 to transfer materials, 509e514 AHFS. See American Hospital Formulary System (AHFS) AHRQ. See Agency for Healthcare Research and Quality (AHRQ) AID. See Agency for International Development (AID) AIDS. See Acquired immunodeficiency syndrome (AIDS) AIDS Clinical Trials Group (ACTG), 144e145, 145t Protocol 076 trial, 385 AIDS-TRIALS database, 113 AIMS. See Arthritis Impact Measurement Scales (AIMS) ALCOA. See Attributable, legible, contemporaneous, original, and accurate (ALCOA) Alcohol, 3 Alexandre Louis, Pierre Charles, 7 Algorithms and software, 458e459 Alliances, 614e615 Allopurinol, 144 Ally in scientific research, 198e199 Alpha error. See Type I error Alpha spending, 384e387 efficacy boundaries, 384e386 futility, 386e387 O’BrieneFleming boundary, 387 ALS. See Amyotrophic lateral sclerosis (ALS) Alternative hypothesis (HA), 344e346 Alternative reviews, 306 Alzheimer’s disease, 619e620 therapy, 418e419 Amantadine, 144 American Association for Cancer Research (AACR), 617

776 American Association of Retired Persons (AARP Inc.), 195 American Cancer Society (ACS), 614 American Chemical Society (ACS), 724be743b American Civil Liberties Union (ACLU), 457e458 American College of Cardiology/ American Heart Association (ACC/AHA), 414 American Hospital Formulary System (AHFS), 724be743b American Neurological Association, 668 American Recovery and Reinvestment Act (ARRA), 270e271, 271b American Society of Health-System Pharmacists (ASHP), 724be743b Americans with Disabilities Act (ADA), 492 amfAR. See Foundation for AIDS Research (amfAR) Ammar ibn Ali al-Mawsili, 3 AMP. See Association for Molecular Pathology (AMP) Amyotrophic lateral sclerosis (ALS), 619e620 Analysis of variance (ANOVA), 11e13, 354, 356 Analysis population, issues related to, 120 “Analysis-ready” data set, 121 Ancillary care, 105 Ancillary systems, 688 Ancillary therapy, 421e422 “AND” operator, 719e720 Anesthesia, 8 Angiotensin-converting enzyme inhibitors (ACEIs), 419 Animal carcinogenicity, 80 Animal Care and Use Committee (ACUC), 39 Animal magnetism, 6e7 Animal Rule, 650 Animal subject research, 38e39 Annals of Internal Medicine, 43 ANOVA. See Analysis of variance (ANOVA) ANPRM. See Advance Notice of Propose Rule Making (ANPRM) Anthrax, 9 Anticipation, 519 Anticoagulants, 565 Antidotes, 641 Antiinflammatory agents meta-analysis of clinical trials in sepsis, 321e323 analyzing data, 322e323 data extraction, 322 eligibility criteria, 322 formulating question, 321e322 identifying studies, 322 inflammation in mediating sepsis, 321 Antiretroviral therapy (ART), 384 Antitoxin, 10 AP. See Associated Press (AP)

INDEX

APA. See Administrative Procedure Act (APA) APEC. See Asia-Pacific Economic Cooperation (APEC) Application Submission System and Interface for Submission Tracking (ASSIST), 592 aPTT. See Activated partial thromboplastin time (aPTT) Archiving, 37 Area under the curve (AUC), 398, 649 ARIPO. See African Regional Intellectual Property Organization (ARIPO) “Arm’s length” transaction, 471e472 Army R&D lab, 496 ARRA. See American Recovery and Reinvestment Act (ARRA) ART. See Antiretroviral therapy (ART) Arthritis Impact Measurement Scales (AIMS), 308e309 Ascertainment bias, 399 ASEAN. See Association of Southeast Asian Nations (ASEAN) ASHP. See American Society of HealthSystem Pharmacists (ASHP) Asia-Pacific Economic Cooperation (APEC), 87e88 Asian Economic Community (AEC), 87e88 Aspirin, 419e420 Assertive patient, 198e199 Assessment of Safety and Efficacy of a New Thrombolytic study (ASSENT2 study), 429 Assignment, 468e469, 519 ASSIST. See Application Submission System and Interface for Submission Tracking (ASSIST) Associated Press (AP), 707e708 Association for Accreditation of Human Research Protection Programs (AAHRPP), 538e539 AAHRPP-accreditation model, 63 expecting from organizations, 64 organizations expecting from, 64e65 Association for Assessment and Accreditation of Laboratory Animal Care International guidelines (AAALAC guidelines), 39 Association for Molecular Pathologists v. USPTO and Myriad Genetics, 456e457 Association for Molecular Pathology (AMP), 457e458 Association for the Accreditation of Human Research Protection Programs (AAHRPP), 58, 63 expecting from organizations, 64 importance of achieving, 71f organizations expecting from, 64e65 results, 71f Association of American Medical Colleges, 662 Association of American Universities (AAU), 64

Association of Clinical Research Professionals (ACRP), 674 Association of Southeast Asian Nations (ASEAN), 87e88 Association of University Technology Managers (AUTM), 487e488, 502 Asthma, 590e591 Attributable, legible, contemporaneous, original, and accurate (ALCOA), 699 Attributable risk, 243 AUC. See Area under the curve (AUC) Auditing, 538e540, 539t Authentication of key biological resources, 598 Authorization, 169 Authorship, 42e43, 44t, 212e213 AUTM. See Association of University Technology Managers (AUTM) Avandia. See Rosiglitazone AZT. See Zidovudine (AZT)

B Baidu, 744 Bailly, Jean-Sylvain, 6e7 “Balancing act”, 673 Banting, Frederick, 11 BARI trial. See Bypass Angioplasty Revascularization Investigation trial (BARI trial) BASE. See Bielefeld Academic Search Engine (BASE) Baseline hazard, 379e380 Baseline observation carried forward (BOCF), 404 Basic Local Alignment Search Tool (BLAST), 745 Bayesian adaptive methods, 389 Bayesian approach, 653 Bayh-Dole Act, 41, 487e493 core terms required in BayheDole funding agreementsex 202, 488e489 x 202edetermination of exceptional circumstances, 489e490 duty of US manufactureex 204, 493 funding agreements outside, 493 history and philosophy, 487e488 key conceptsexx 200 and 201, 488 “March-In”ex 203, 490e493 organization of clauses, 488 x 202ereporting obligations, 489 BBRB. See Biorepositories and Biospecimen Research Branch (BBRB) Beecher, Henry, 13, 21, 48 Before American Revolution, 449 Behavioral research, 56 Belmont Report, The, 21e22, 24, 49e50 Beneficence, 49 Benson-Flook-Diehr, 458e459 Best mode, 463e464 Best Pharmaceuticals for Children Act, 656 Best practices for licensing genomic inventions, 501 Best proven intervention, 103

INDEX

BEsT resource. See Biomarkers, Endpoints and other Tools resource (BEsT resource) Beta error. See Type II error Beta-blockers, 432 Beta-interferon effect, 265e266 study, 265 Beta-interferon/magnetic resonance imaging study, hypotheses for, 345, 351e352, 351t Bias(es), 238, 330 biased coin randomization method, 336 publication, 427e428 Bibliographic managers, 748 Bibliometrics, 747e748 Bielefeld Academic Search Engine (BASE), 724be743b Bilski, 458e459 Binary data, 347e348 calculations for, 367e368 Bing, 724be743b Bioengineering, 619 Bioethics research, 23 Biohazards, 597 Bioinformatics directories, 746 organizations, 745 resources, 744e746 bibliographic managers, 748 bibliometrics, 747e748 browsers, 746 commercial software, 746 data integration and precision medicine, 746e747 data management, 746 defining networks, mining networks, 747f educational resources, 749 organizations, 745 selection and search strategy, 748e749 Bioinformatics Links Directory, 746 Biologics Control Act (1902), 73e74 Biologics licensing application (BLA), 90 Biologics licensing application/New Drug Application (BLA/NDA), 82e83 Biomarkers, 425, 426t, 559, 620, 649e652 efficacy biomarkers and surrogate end points, 650e652 qualification, 650 safety biomarkers, 650, 651t Biomarkers, Endpoints and other Tools resource (BEsT resource), 425 BioMedCentral, 41e42 Biomedical imaging, 619 Biomedical information resources, 722 Biomedical research, 56 Biomedical Research Integrated Domain Group (BRIDG), 551 Biopsychosocial model, 303e305 Biorepositories and Biospecimen Research Branch (BBRB), 560 Biorepository, 564 Bioresearch monitoring inspection, 83

Biosketches, senior/key personnel profiles component and, 604 Biospecimen, 559 research “regardless of identifiability”, 172 BLA. See Biologics licensing application (BLA) BLA/NDA. See Biologics licensing application/New Drug Application (BLA/NDA) BLAST. See Basic Local Alignment Search Tool (BLAST) Blinded trials, 427 Blinding, 28, 43e44, 210, 215, 258, 424 Block randomization, 332e333 Blood pressure, 216, 550 Blundell, James, 3e4 BOCF. See Baseline observation carried forward (BOCF) Bolar Amendment, 471 Bonferroni correction, 355, 368 Book of Optics (Ibn al-Haytham), 3 Boolean AND operator, 717 Boolean NOT command, 721 Boolean OR operator, 719e720 “Brain drain” of health care, 104 Breaking patents, 483 Breakthrough therapy designation, 79 Breast cancer, 208 BRIDG. See Biomedical Research Integrated Domain Group (BRIDG) Bridging Interventional Development Gaps (BrIDGs), 490 British Medical Research Council (MRC), 331, 412e413 Browsers, 746 Budget and justification, 604 Budget and period of support, 598 Building knowledge and expanding scientific disciplines, 619e620 Burroughs Wellcome Fund (BWF), 617 BurroughseWellcome Company (BW Company), 144, 467 Business models, 721e723 BW Company. See BurroughseWellcome Company (BW Company) BWF. See Burroughs Wellcome Fund (BWF) By-arm treatment effect, 396 Bypass Angioplasty Revascularization Investigation trial (BARI trial), 429, 432

C C-reactive protein, 185 caDSR. See Cancer Data Standards Registry and Repositories (caDSR) CADTH. See Canadian Agency for Drugs and Technologies in Health (CADTH) CAFC. See Court of Appeals for Federal Circuit (CAFC) Campbell Collaboration, 305e306 Campostar. See Irinotecan Canadian Agency for Drugs and Technologies in Health (CADTH), 724be743b

777 Canadian Tissue Repository Network (CTRNet), 561e562 Cancer Data Standards Registry and Repositories (caDSR), 534e535 Cancer Therapy Evaluation Program (CTEP), 534 CancerNet Website, 113 Canon of Medicine, The, 3 CAP. See College of American Pathology (CAP) CAPA. See Corrective and preventive action plan (CAPA) Carbamazepine, 650, 651t Carcinogenesis, 177 Cardiac Arrhythmia Suppression Trial, 211e212, 425 Cardiovascular mortality, 206 risk factors, 185 Cardiovascular disease (CVD), 330 Career awards, 594 Career development programs, 617, 618t Career paths clinical research curriculum and training, 665e666 educational approaches and support for training, 668e669 key core components of student research programs, 664t NIH, 661 clinical center core curriculum, 666e668, 667t PhysicianeScientist workforce, 664e665 student and resident training in clinical research, 662e664 Career potential for nurses in clinical research, 682e683, 683f Career transition awards K22 career transition awards, 608 K99/R00 pathway to independence award, 607e608 Careful record keeping, 562 Caregiver, 195 Carisoprodol, 650, 651t Case report, 232e233 advantages and disadvantages, 234 objectives and design, 233e234 observations and analysis, 234 Case report forms (CRFs), 535e536, 574e575 Case series, 232e234 Case-control methods, 282 studies, 233, 282e283 advantages and disadvantages, 239 objectives and design, 236e238, 236f observations and data analysis, 238e239, 239t “Cases and controversies”, 472 Cassandra revealed, 147 Causal inference in observational studies, 405e406 CBER. See Center for Biologics Research and Evaluation (CBER) CBIIT. See Center for Bioinformatics and Information Technology (CBIIT)

778 CC. See Clinical Center (CC) CC analysis. See Complete-case analysis (CC analysis) CCPA. See US Court of Customs and Patent Appeals (CCPA) CDA. See Clinical Document Architecture (CDA); Confidential disclosure agreement (CDA) CDC. See Centers for Disease Control and Prevention (CDC) CDER. See Center for Drug Evaluation and Research (CDER) CDEs. See Common data elements (CDEs) CDISC. See Clinical Data Interchange Standards Consortium (CDISC) CDM. See Common data model (CDM) CDP. See Commercial development plan (CDP); Coronary Drug Project (CDP) CDRNs. See Clinical data research networks (CDRNs) CDS. See Clinical decision support (CDS) CDSR. See Cochrane Database of Systematic Reviews (CDSR) Cell, 333 CellPro, Inc., 491 Censoring, 374 Center for Bioinformatics and Information Technology (CBIIT), 534e535 Center for Biologics Research and Evaluation (CBER), 73 Center for Disease Control and Prevention, 633e634 Center for Drug Evaluation and Research (CDER), 73 Center for Scientific Review (CSR), 593 Center for scientific review home page, 609 Centers for Disease Control and Prevention (CDC), 113, 304, 614 Centers for Medicare and Medicaid Services (CMS), 115, 274, 304, 698 CENTRAL. See Cochrane Central Register of Controlled Trials (CENTRAL) Central IRB, 58 Central limit theorem, 348 CER. See Comparative clinical effectiveness research (CER) Certificate of Confidentiality (CoC), 169e171, 226 Certification for nursing, 679 CF. See Cystic fibrosis (CF) CFF. See Cystic Fibrosis Foundation (CFF) 45CFR46. See Code of Federal Regulations in Title 45, USCFR, Part 46 (45CFR46) Chakrabarty to Mayo to Myriad, 456e458 Charles Best’s discovery of insulin, 11 CHD. See Coronary heart disease (CHD) ChemIDplus, 724be743b Chemistry, manufacturing, and control (CMC), 77e78 Chemotherapy, 11 CHF. See Chronic heart failure (CHF)

INDEX

Childbearing potential, women of, 182e183 Childhood exposure to trauma, 205e206 Children, 162, 165e166, 182e183 assent, 163e164 inclusion of, 603 appendix, 605 budget and justification, 604 facilities and resources, 604e605 senior/key personnel profiles component and biosketches, 604 vertebrate animals, 603e604 Children’s Health Insurance Program (CHIP), 698 Choice of control, 28 Cholera, 7 Cholesterol, 185 Chronic active hepatitis study, 378, 378t Chronic concern, 40 Chronic disease studies, 264 Chronic HBV infection, 145 Chronic heart failure (CHF), 206 Chronic hepatitis, 145, 156 Chronic interruptions, 197 Chronic lateness, 197 Chronic liver disease (CLD), 312 Chronic Liver Disease Questionnaire (CLDQ), 312e313 Chronic obstructive pulmonary disease, 207 CI. See Confidence interval (CI) CINAHL. See Cumulative Index to Nursing and Allied Health Literature (CINAHL) CIOMS. See Council of International Organizations of Medical Sciences (CIOMS) CIP. See Continuation-In-Part (CIP) Cirassians, The, 5e6 Cirrhosis, 145 Cisapride, 655, 656t Citation searching, 721 Cited reference databases, 724be743b CiteULike, 748 CITI. See Collaborative Institutional Training Initiatives (CITI) Civil disorders, 101e102 “Civil” wrong, 469 CL. See Clearance (CL) Claims, 473e474 data, 294 CLD. See Chronic liver disease (CLD) CLDQ. See Chronic Liver Disease Questionnaire (CLDQ) Clearance (CL), 649 CLIA. See Clinical Laboratory improvement Amendments (CLIA) Clinical and Translational Science Award (CTSA), 289, 555, 665, 671 Clinical care, 694 Clinical Center (CC), 48 Clinical Data Interchange Standards Consortium (CDISC), 551 Clinical data repositories, 552e553

Clinical data research networks (CDRNs), 288 Clinical decision resources, 724be743b Clinical decision support (CDS), 687 within EHR, 694 Clinical Document Architecture (CDA), 551 Clinical end point, 650e651 Clinical equipoise, 27 Clinical event reporting systems, 640e641 Clinical investigator, 704 Clinical Laboratory improvement Amendments (CLIA), 578 Clinical outcome assessments (COA), 650 Clinical pharmacology, 645 contribution, 649e654 advent of pharmacogenetics and pharmacogenomics, 654 design and conducting of phase I-II studies, 652e653 FIH study, 649 functional imaging tools, utilization of, 649e652 identification, development, and qualification of biomarkers, 649e652 modeling and simulation and modelbased drug development, 653e654 personalized medicine, 652 drug development, 646e647 current state of affairs in, 647e649 FDA and, 655 and drug safety, 655e656 and special populations, 656 regulatory agency role, 654e656 tools, 655 as translational discipline, 645e646 Clinical Pharmacology Guidance Documents, 655 Clinical practice, clinical research from, 19e20 Clinical protocol, 80 Clinical quality measurement, 641e642, 641t performance measurement, 635 Clinical research, 1, 111, 166, 219, 531, 710. See also Comparative effectiveness research; Population-based health research challenges, 100e103 civil disorders, 101e102 deficient research infrastructures, 100e101 economic and seasonal migration, 102 ethical issues, 102e103 inadequate human resources, 100 information gaps, 101 natural disasters, 101e102 physical barriers, 102 political instability, 101e102 study participant characteristics, 102 subpar health-care systems, 101

INDEX

CRIS, 688e692 data, 547 cooperative sharing efforts, 555e556 data as surrogates, 547e550 data capture, storage, and retrieval, 551e553 data standards, 550e551 responsible stewardship of data, 553e555 types of data, 550 earliest, 1e2 EHR, 688, 692e696 architecture, 688 secondary use of EHR for clinical research, 698 eighteenth century, 4e7 EMR, 687 environment, 633e634 applying high reliability principles in, 635, 636te637t electronic surveillance for errors and system failures, 641 identifying and managing clinical risk in, 633e635 leveraging patient safety in clinical research, 635e637 patient safety and clinical quality measurement, 641e642, 641t proactively assessing clinical and operational risk, 638e641, 639f Greek and Roman influence, 2 in international settings, 100 legislation and EHR, 698e699 methodology, 662 middle ages and renaissance, 2e3 nineteenth century, 7e11 operational organization for large-scale, 431f, 433e437 participants perception assessment, 642 protocol, 220 quality improvement techniques in, 635e637 recommendations, 103e106 developing and enhancing local institutional review board capacity, 104e105 DSMP for adverse events, 105 existing infrastructure integration, 105e106 local setting, 103e104 long-term plans, 105 office for sponsored research/office of clinical research development, 105 providing ancillary care, 105 train, mentor, and closely supervise, 104 use technology for effective communication, 105 secondary use of EHR for clinical research, 698 seventeenth century, 3e4 testing puddings and gelatins at Consumers Union, 13f twentieth century and beyond, 11e14

Clinical research associate (CRA), 532e533, 679 Clinical Research Coordinators (CRCs), 679 Clinical research curriculum and training, 665e666 Clinical Research Curriculum Award (CRCA), 665 Clinical research data warehouse (CRDW), 688 Clinical research information system (CRIS), 688e692, 694t Clinical research institutes concepts underlying trial design, 417e420 common qualitative interactions, 419 interactions among therapies, 419e420 long-term effects, 420 treatment effects, 417e418 uncommon qualitative interactions, 418e419 unintended biological targets, 419 controversies and personal perspective, 437e439 composite and surrogate endpoints, 438 governmental regulation vs. professional responsibility, 437e438 randomized trials vs. observational studies, 438e439 sharing of information, 439 critical general concepts, 414e415, 414t generalizability, 415 validity, 415 design considerations, 420e422 ancillary therapy, 421e422 data collection form, 421 entry criteria, 421 multiple randomization, 422 pick the winner approach, 422, 423f pragmatic vs. explanatory, 420 expressing clinical trial results, 415e417, 416fe417f future, 439e440 hypothesis formulation, 427 integration into practice, 437, 437f legal and ethical issues biomarkers and surrogate endpoints, 425 blinding, 424 conflict of interest, 425e426 endpoint adjudication, 424 groups of patients vs. individuals, 422e424 intensity of intervention, 424e425 medical justification, 422 special issues with device trials, 426e427 medical practice, 412 meta-analysis and systematic reviews, 430e431, 430t operational organization for large-scale clinical research, 433e437 phases of evaluation of therapies, 413e414, 413t publication bias, 427e428

779 statistical considerations noninferiority, 429, 429t sample size calculations, 430 type I error and multiple comparisons, 428 type II error and sample size, 428e429 therapeutic truisms, 432e433 understanding covariates and subgroups, 431e432 Clinical Research Nursing (CRN), 672e675 continuum, 673f of clinical translation science, 674f documenting specialty, 674e679 future considerations career potential for nurses in clinical research, 682e683 meeting need for nurses to fill clinical research roles, 683 nursing role in community-based research, 683e684 supporting transition of nurses into clinical research, 684 integrating research process and clinical care, 674f legal scope of practice issues, 680 NIH Clinical Center, 671, 672f nurse interacting with research participant, 672f nursing roles, 675t tools to assist principal investigator in staffing study, 680e682 Clinical Research Training Program (CRTP), 661e662 Clinical researcher(s), 57 and media celebrity, 705 communications office, 712 controversial biomedical research, 705e706 email, web, and social media, 707e708 embargoes, 711e712 engaging media, 707 FOIA, 711 health message, 710 impact, 706 interview, 708e710 investigative reporters, 710e711 news, 710 news in science and medicine, 704 NIH, 703 NIH results, 705 novelty, 704e705 public not knowing science, 710 publishing science, 704 rule of thumb, 710 social media, 707 talking to reporters, 706e707 at stages in careers, NIH grant programs for, 607e609 career transition awards, 607e608 exploratory/development grant (R21) applications, 608e609 independent scientist awards, 608

780 Clinical researcher(s) (Continued ) individual career development awards, 607 loan repayment program, 609 mentored career development awards, 607 midcareer investigator award in patient-oriented research, 608 small research grant (R03) applications, 609 Clinical studies, 232, 250e251 critical issues in clinical study design blinding or masking, 258 choosing comparison group, 258 intervention development, 258 Clinical treatment trial, 193 Clinical trial agreement (CTA), 226, 512e513 Clinical Trial Regulations, 116 Clinical trial(s), 76e79, 220, 232 clinical studies, 250e251 critical issues in clinical study design, 258 control groups, 258e261 design, 250, 255e257 adaptive treatment, 257 crossover, 255e256 enriched enrollment designs, 256 factorial, 256 group-randomized trial, 257 parallel groups, 256e257 sequential trial designs and interim analyses, 257 information, 114 meta-analysis, 321e323 analyzing data, 322e323 data extraction, 322 eligibility criteria, 322 formulating question, 321e322 identifying studies, 322 inflammation in mediating sepsis, 321 mistakes and misconceptions, 262e266 monitoring, 130, 130f patients refuse to participate in, 191e193 African Americans underrepresented in, 192e193 elderly underrepresented in, 193 phases, 220t placebo responses, 261e262 policies affecting clinical trials in United States, 115 purpose, 250e251 registration, 112 ethical and scientific rationale for, 113t rationale for, 112e113 steps in, 112f research, 671 resources, 724be743b results, 415e417, 416fe417f sample approaches to protocol organization, 221t understanding spectrum of research continuum, 251e255

INDEX

Clinical trials data management system (CTDMS), 552, 688, 692 Clinical trials management systems (CTMS), 537e538 Clinical trials nurses (CTNs), 532e533 Clinical Trials Transformation Initiative (CTTI), 121, 434e435 ClinicalKey, 724be743b ClinicalTrials.gov clinical research enterprise, 121 comparison of summary requirements, 115t cumulative number of registered studies, 114f using data, 120e121 intended audience, 120e121 history, 113e115 identifier, 116 registering clinical trials at, 116e118 data standards and minimal data set, 116e117 importance of protocol, 117 interventional vs. observational studies, 117 keeping information up-to-date, 117e118 single clinical trial, 117 reporting results to, 118e120 data preparation, 118 data standards and minimal data set, 118 relation of results reporting to publication, 119 review criteria, 118e119, 119t scientific modules, 118t search tips for, 121 Clinicianescientists, 661e662, 665, 668 ClinVar, 745 ClopperePearson method, 348 Cluster randomization, 280 Cluster randomized trials. See Grouprandomized trials CMC. See Chemistry, manufacturing, and control (CMC) CMR. See Cochrane Methodology Register (CMR) CMS. See Centers for Medicare and Medicaid Services (CMS) CMV. See Cytomegalovirus (CMV) COA. See Clinical outcome assessments (COA) CoC. See Certificate of Confidentiality (CoC) Cochrane Central Register of Controlled Trials (CENTRAL), 724be743b Cochrane Collaborations, 305e306, 724be743b Cochrane Database of Systematic Reviews (CDSR), 724be743b Cochrane Library, 724be743b Cochrane Methodology Register (CMR), 724be743b Cochrane reviews, 724be743b Cochrane Systematic Reviews, 306

Cockersole, Hexamethonium lung, 724be727b, 728f Code of Federal Regulations in Title 45, USCFR, Part 46 (45CFR46), 22e23, 49 Codes of research ethics and regulations, 22e23, 22t Coding system, 536 Coexclusive licensing, 496 Cohort studies advantages and disadvantages, 242e243 objectives and design, 239e241, 240f concurrent or prospective cohort studies, 240e241 nested case-cohort studies, 241 nested case-control studies, 241 nonconcurrent, historical, or retrospective cohort studies, 240 observations and data analysis, 241e242 COI. See Conflict of interest (COI) Coinvestigators. See Subinvestigators Collaboration and inventions, 514e518 Collaborative Institutional Training Initiatives (CITI), 38 Collaborative Research and Development Agreement (CRADA), 39e40, 502 Collaborative science, 39e40 College of American Pathology (CAP), 562 , 578 Collip, James B., 11 Colon cancer, 194 Combined across arms, 388 Commercial development plan (CDP), 500 Commercial exploitation, 465 Commercial IRBs, 572 Commercial software, 746 Commitment, 40e41 Committee for Proprietary Medicinal Products (CPMP), 337 Committee on Scientific Conduct and Ethics (CSCE), 33e34 Common data elements (CDEs), 534 Common data model (CDM), 288e289 Common Rule, 21, 23, 38e39, 49e51, 169e170, 227e229 regulatory requirements, 162 Research, 51 Common Rule regulations, 165 Common Scientific Outline (CSO), 615 Common Technical Document (CTD), 90 Common Terminology Criteria for Adverse Events (CTCAE), 535, 541 Communications office, 712 Community community-engaged participatory research model, 104 involvement in research, 21 nursing role in research, 683e684 participatory research model, 104 Company-sponsored foundation, 614 Comparative clinical effectiveness research (CER), 270e271, 280f articles in PubMed with “comparative effectiveness”, 270f

INDEX

comparative effectiveness research, 272t evidence synthesis in, 285e287 methods, 275e278 choosing clinical outcomes to measurement, 277 choosing study population, 276 engagement role in specifying research questions, 277e278 getting research question, 275e276 selecting appropriate interventions and comparator(s), 276e277 PCORI, 271b role, 273e275 study designs for CER studies, 278e285, 279f Comparative effectiveness research, 254. See also Clinical research; Population-based health research CER, 270e271 evidence synthesis in CER, 285e287 health-care choices, 269 methods of CER, 275e278 national infrastructure for conduct, 287e290, 288f PCORI, 271e273 role of CER, 273e275 study designs for CER studies, 278e285 Comparative effectiveness trial, 258 COMPENDEX, 724be743b Competency, 679 assessment, 679 validation, 679 Competing risks, 402 Competing supplement applications, 597 Complement system, 10 Complete response letter, 83 Complete-case analysis (CC analysis), 404 Complex protocols, 696 Compliance, 84 Composite, 438 Comprehensive Meta-Analysis, 321 Compulsory licensing, 483 Computed tomography scan (CT scan), 705 Computer-based systems, 563 Computerized management of digital information, 440 Computerized provider order entry (CPOE), 687 CON. See Continuation (CON) Conception, 466e467, 519 Concurrent cohort studies, 240e241 Concurrent prospective study, 240, 240f Conditional power (CP), 386 Conduct nontherapeutic research, 167 Conduct of research. See also Ethical principles in clinical research archiving, 37 authorship, 42e43, 44t COI and commitment, 40e41 collaborative and team science, 39e40 data management, 36e37 data sharing, 38 guidelines and principles for, 33e34 peer review, 41e42

publication practices, 42 reproducibility, 43e45, 44t research involving human and animal subjects, 38e39 responsibilities of research supervisors and trainees, 36 scientific integrity and research misconduct, 34e35 Conduct of science, 34 Confidence interval (CI), 343, 347e348 for binomial proportions, 362 estimation, 392 for means of continuous data, 361e362 sample size calculations for precision in, 361e362 Confidential disclosure agreement (CDA), 507 anatomy of, 508e509 Confidentiality, 55e56, 226, 510, 600 agreements to protecting, 506e509 anatomy of confidential disclosure agreement, 508e509 secrets and government, 507e508 trade secrets, 507 Conflict of interest (COI), 40e41, 167e168, 226e227, 425e426, 600 Confounding, 379 Congenital anomaly, 156 Congress, 154, 168, 177 Consensus decision-making model, 638 Consensus statements, 666 Consent process, 227e229, 228t consent form, 227e228 individuals obtaining consent, 227 persons providing consent, 227 Consolidated Standards of Reporting Trials (CONSORT), 119, 216e217, 724be743b guidelines, 233 statement, 262 Container selection, 565 Continuation (CON), 475, 478, 519 Continuation-In-Part (CIP), 475, 478, 519 Continuing review of research, 56 Continuous data calculations for, 362e363 with equal variances and equal sample sizes, 366e367 with unequal variances or unequal sample sizes, 367 Contract research organization (CRO), 434e435, 538e539, 672e673 Contract(s) execution in general, 504e506 management, 436 Contractors, 487e489 Contributory infringement, 469e470 Control groups, 258e261 multiple control groups, 260e261 placebo control group, 259e260 sham control group, 260 time and attention control groups, 259 usual and standard care controls, 260 wait-list control group, 259

781 Controlled clinical trials, 374, 437e438 Controlled trials, 271 Convention on Unification of Certain Points of Substantive Law on Patents for Invention, 451e452 Cooperative agreements, 592 Cooperative Patent Classification (CPC), 724be743b Cooperative Research and Development Agreement (CRADA), 226, 494, 507, 514e518 execution by parties and effective date, 518 negotiating agreement, 516 financial and material contributions, 517 modifications to CRADA language, 516 RP, 516e517 NIH inventions, 514 NIH review of agreement, 518 possibilities, 518 selecting collaborator, 515e516 Cooperative sharing efforts, 555e556 Copyrights, 448e449, 483e485 Core review criteria, 596e597 Coronary angioplasty, 424 Coronary Artery Surgery Study, 206e207 Coronary disease, 207 Coronary Drug Project (CDP), 129 Coronary heart disease (CHD), 208, 233, 250 Corporate foundation, 614 Corrective and preventive action plan (CAPA), 539e540 Correlation, 392e393 Correlational studies. See Ecological studies Cost-effectiveness approaches, 274e275 Council of International Organizations of Medical Sciences (CIOMS), 22 Council on Accreditation, 64 Court of Appeals for Federal Circuit (CAFC), 451, 467 Covariates, 211 Coversheet Provisional, 476 Cox model, 401e402 Cox proportional hazards model, 379e380 Cox regression, 400 COX-2 inhibitors. See Cyclooxygenase-2 inhibitors (COX-2 inhibitors) CP. See Conditional power (CP) CPC. See Cooperative Patent Classification (CPC) CPI report. See Critical Path Initiative report (CPI report) CPMP. See Committee for Proprietary Medicinal Products (CPMP) CPOE. See Computerized provider order entry (CPOE) CPT. See Current Procedural Terminology (CPT) CRA. See Clinical research associate (CRA)

782 CRADA. See Collaborative Research and Development Agreement (CRADA); Cooperative Research and Development Agreement (CRADA) CRCA. See Clinical Research Curriculum Award (CRCA) CRCs. See Clinical Research Coordinators (CRCs) CRDW. See Clinical research data warehouse (CRDW) Credible/Credibility, 128, 360, 460, 706 CRFs. See Case report forms (CRFs) “Criminal” wrongs, 469 CRIS. See Clinical research information system (CRIS) Critical Path Initiative report (CPI report), 646 Critical Path Opportunity List, 654e655 Critical values, 346e347 CRN. See Clinical Research Nursing (CRN) CRO. See Contract research organization (CRO) Cross-Disciplinary Research Awards, 622 Cross-sectional studies, 232e233 advantages and disadvantages, 235e236 objectives and design, 234e235 observations and data analysis, 235 Cross-sectional survey, 232e233 Crossover designs, 255e256 CRTP. See Clinical Research Training Program (CRTP) CSCE. See Committee on Scientific Conduct and Ethics (CSCE) CSO. See Common Scientific Outline (CSO) CSR. See Center for Scientific Review (CSR) CT scan. See Computed tomography scan (CT scan) CTA. See Clinical trial agreement (CTA) CTCAE. See Common Terminology Criteria for Adverse Events (CTCAE) CTD. See Common Technical Document (CTD) CTDMS. See Clinical trials data management system (CTDMS) CTEP. See Cancer Therapy Evaluation Program (CTEP) CTMS. See Clinical trials management systems (CTMS) CTNs. See Clinical trials nurses (CTNs) CTRNet. See Canadian Tissue Repository Network (CTRNet) CTSA. See Clinical and Translational Science Award (CTSA) CTTI. See Clinical Trials Transformation Initiative (CTTI) Cumulative Index to Nursing and Allied Health Literature (CINAHL), 724be743b Cumulative meta-analysis, 430 Curie, Marie, 11, 12f Current Procedural Terminology (CPT), 294 Curriculum vitae (CV), 578

INDEX

CV. See Curriculum vitae (CV) CVD. See Cardiovascular disease (CVD) Cycle(s) NIH review, 594, 594fe595f, 595t of quality, 437 Cyclooxygenase-2 inhibitors (COX-2 inhibitors), 419 Cystic fibrosis (CF), 623e624 Cystic Fibrosis Foundation (CFF), 422, 439, 614 Cytomegalovirus (CMV), 144

D da Foligno, Gentile, 3 da Vinci, Leonardo, 3, 3fe4f Dairymaids, 5 DARE. See Database of Abstracts of Reviews of Effects (DARE) Darvocet. See Propoxyphene Darvon. See Propoxyphene Data analysis, 298 capture, storage, and retrieval, 551e553, 554f clinical data repositories, 552e553 CTDMs, 552 collection, 210e211 and analysis, 531 form, 421 methods, 235 completeness, 133 dependencies, 392e393 elements, 534e535, 534t exploration, 285 extraction, 322 integration, 746e747 interoperability, 288e289 legal and regulatory issues related to data reporting, 542e543 linking data sets, 298e299 management, 36e37, 226, 533e538, 746 adverse event monitoring and reporting, 540e542, 541f, 542t auditing, 538e540 choosing database system, 536e537 CRFs, 535e536 data collection, 537e538 data elements, 534e535, 534t follow-up and analysis, 543 legal and regulatory issues related to data reporting, 542e543 quality control of data, 538 record retention, 543, 543t sources of data, 538 unanticipated problems, 540e542, 541f mining, 298 missing, 297 and model organisms sharing, 498e499 preparation, 118 providers, 118 quality, 133, 297 flags and triggers, 133 repositories, 552e553 responsible stewardship, 553e555

sharing, 38 fostering dissemination of, 624e625 and individual consent, 171e173 plan, 598 standards, 116e118, 550e551 as surrogates, 547e550 indirect nature of clinical research data, 548 metadata, 549e550 objectivity and subjectivity of clinical data, 548e549 transparency, rigor, and reproducibility, 549 types, 550 validation, 552 Data analysis, 392e394 confidence interval estimation, 392 data dependencies, 392e393 essential issues in microarrays, functional MRI, 393e394 study protocol, 211e212 overall analyses, 211 subgroup analysis, 211e212 Data and safety monitoring (DSM), 128, 130e137, 224 DSMB, 128 interim analyses, 133e136 monitoring participant safety, 130e131 monitoring trial conduct, 131e133 Data and Safety Monitoring Board (DSMB), 40e41, 52e53, 128, 156, 212, 224, 384e385, 434, 533 general structure of DSMB meetings, 137 history, 129 masking, 137e138 needs, 129e130 Data coordinating center (DCC), 433, 435 Data management plans (DMPs), 534 Data marts. See Datadrepositories Data Monitoring Committee (DMC), 128 Data safety and monitoring committee (DSMC). See Data and Safety Monitoring Board (DSMB) Data safety and monitoring plan (DSMP), 105, 603 for adverse events, 105 Data Transfer Agreements (DTA), 226 Data warehouses. See Data repositories Database administrator (DBA), 533 Database of Abstracts of Reviews of Effects (DARE), 724be743b Database of Genome and Phenome (dbGAP), 555e556, 745 Database of Single Nucleotide Polymorphisms (dbSNP), 745 Database system, choosing, 536e537 “Daughter” applications, 475 DBA. See Database administrator (DBA) dbGAP. See Database of Genome and Phenome (dbGAP) dbSNP. See Database of Single Nucleotide Polymorphisms (dbSNP) DCC. See Data coordinating center (DCC) DDBJ. See DNA Data Bank of Japan (DDBJ)

INDEX

DDC. See Zalcitabine (DDC) DDCF. See Doris Duke Charitable Foundation (DDCF) DDI. See Didanosine (DDI) DDIR. See Director for Intramural Research (DDIR) DDMRI. See Doris Duke Medical Research Institute (DDMRI) DEC. See Determination of Exceptional Circumstances (DEC) DEcIDE Network. See Developing Evidence to Inform Decisions about Effectiveness Network (DEcIDE Network) Decision-making process, 54e55 Declaration of Helsinki, 13, 22, 102e103, 113 Declaratory Judgement (DJ), 472 actions, 472 Deficient research infrastructures, 100e101 degrees of freedom (df), 346, 350 DELTAS Africa. See Developing Excellence in Leadership, Training, and Science in Africa (DELTAS Africa) Demographic trends in clinical trial participation, 183e184, 184f Denys, Jean-Baptiste, 3e4 Department of Commerce (DoC), 450, 496 Department of Defense (DoD), 163e164, 220, 487, 493, 496 Departments of Energy (DoE), 493, 496 Departments of Homeland Security (DHS), 493 Departments of Interior (DoI), 495 Departments of Transportation (DoT), 493, 495, 497 Dependent censoring, 401e402 Depression, 195 DerwentWorld Patents Index (DWPI), 724be743b Descriptive information, 116 Design patent, 454 Determination of Exceptional Circumstances (DEC), 489e490 x 202eDetermination of exceptional circumstances, 489e490 Developing Evidence to Inform Decisions about Effectiveness Network (DEcIDE Network), 271b Developing Excellence in Leadership, Training, and Science in Africa (DELTAS Africa), 621 Device trials, special issues with, 426e427 df. See degrees of freedom (df) DHHS. See US Department of Health and Human Services (DHHS) DHS. See Departments of Homeland Security (DHS) Diabetes Prevention Program (DPP), 185e186 Diagnostic accuracy studies, 400 criteria, 311e312 likelihood ratio, 397

suspicion bias, 238e239 testing considerations for study design, 398e399 diagnostic accuracy studies, 400 HIV test results, 397t measures of accuracy, 396e398 mistakes and biases, 399e400 Dichotomous data. See Binary data DickeyeWicker Amendment, 166 Didanosine (DDI), 149 Digital information, computerized management of, 440 Diphtheria antitoxin, 10, 74 Direct searching of journal articles, 724be743b “Direct” liability, 469 Director for Intramural Research (DDIR), 518 Directory of Open Access Journals (DOAJ), 722 Discipline, 645 Disease-specific measures, 312e313 Dissemination and implementation studies, 253e254 Dissertations search resources, 724be743b Distributive justice, 24 Division of Biological Standards, 74 Divisional application (DIV application), 475e476, 478, 519 DJ. See Declaratory Judgement (DJ) DLT. See Dose-limiting toxicity (DLT) DMC. See Data Monitoring Committee (DMC) DMOZ, 744 DMPs. See Data management plans (DMPs) DNA, 179, 455e458 DNA Data Bank of Japan (DDBJ), 745 DNDi. See Drugs for Neglected Disease Initiative (DNDi) DOAJ. See Directory of Open Access Journals (DOAJ) DoC. See Department of Commerce (DoC) Documentation of randomization, 337 Documenting specialty for CRN, 674e679 certification, 679 competency assessment, 679 conceptual framework, 675e677, 675f, 676te677t core curriculum, 679 job descriptions, 678 practice standards, 677e678 DoD. See Department of Defense (DoD) DoE. See Departments of Energy (DoE) DoI. See Departments of Interior (DoI) Doniach, Lung changes during Hexamethonium therapy, 724be728b, 729f Doris Duke Charitable Foundation (DDCF), 617 Doris Duke Clinical Research Fellowship Program, 663 Doris Duke Medical Research Institute (DDMRI), 621

783 Dose doseeresponse relationship, 406 dosing/intervention intensity, 209 escalation in FIH study, 649 ranging, 215 response, 653 Dose-limiting toxicity (DLT), 252 Doshi, Dr. Peter, 724be743b DoT. See Departments of Transportation (DoT) Double blinding, 424 Double-dummy technique, 424 Down syndrome, 13 DPOA. See Durable power of attorney (DPOA) DPP. See Diabetes Prevention Program (DPP) DRAs. See Drug regulatory authorities (DRAs) Drotrecoginalfa, 655, 656t Drug Importation Act, 73 Drug regulatory authorities (DRAs), 90e91 Drug(s), 3, 144e145 and biological product life cycle, 76e84 clinical trials, 76e79 development, 155e157, 646e647, 646f clinical drug development, 648t current state of affairs in, 647e649 industryeFDA interactions, 647f information, 724be743b patients in drug trials, 156e157 preclinical testing of new drugs reliable predictor of toxicity, 155e156 discovery/nonclinical investigation, 76 efficacy, 652 FDA and drug safety, 655e656 marketing approval/licensure, 81e83 Postapproval, 83e84 responsibilities and documentation, 79e81 clinical protocol, 80 FDA, 81 IB, 81 Investigational New Drug Safety Reports, 81 investigators, 80 IRB, 80 sponsors, 79e80 Drugs for Neglected Disease Initiative (DNDi), 624 Drugs@FDA, 75e76, 83, 724be743b DSM. See Data and safety monitoring (DSM) DSMB. See Data and Safety Monitoring Board (DSMB) DSMC. See Data safety and monitoring committee (DSMC) DSMP. See Data safety and monitoring plan (DSMP) DTA. See Data Transfer Agreements (DTA) Dual-review system, NIH, 594 DuckDuckGo, 744 Due diligence, 471

784 Durable power of attorney (DPOA), 165, 227 Duty of US manufactureex 204, 493 DWPI. See DerwentWorld Patents Index (DWPI) DynaMed, 724be743b Dynamic allocation algorithms, 336 Dynamic treatment. See Adaptive treatment

E E6 Good Clinical Practices, 89e90 EAC. See East African Community (EAC) EAC-MRH. See East African Medicines Regulatory Harmonization Program (EAC-MRH) EAPO. See Eurasian Patent Organization (EAPO) Earliest clinical research, 1e2 Early phase studies, 391 Early-Stage Investigators (ESIs), 596, 598 Ease of scoring, 308 East African Community (EAC), 87e88 East African Medicines Regulatory Harmonization Program (EAC-MRH), 87e88 Ebola virus Disease Medical Countermeasures Trial, 389 infection, 635 EC. See Ethics committee (EC); European Commission (EC) Ecological bias, 321 fallacy, 233, 298 studies, 232e233 Economic engine, 454 Economic Espionage Act (EEA), 486 Economic Evaluation Database (EED), 724be743b Economic migration, 102 eCRF, 535e538 eCTD. See Electronic Common Technical Document (eCTD) EDC. See Electronic data capture (EDC) EDirect. See Entrez Direct (EDirect) Edit checks, 536 Educational approaches, additional, 668e669 Educational resources, 749 EEA. See Economic Espionage Act (EEA) EED. See Economic Evaluation Database (EED) EEG. See Electroencephalogram (EEG) Efalizumab, 655, 656t Efficacy, 88 biomarkers, 650e652 boundaries, 384e386 of intervention, 28 EFPIA. See European Federation of Pharmaceutical Industries and Associations (EFPIA) eGFR. See estimated glomerular filtration rate (eGFR)

INDEX

EGPAF. See Elizabeth Glaser Pediatric AIDS Foundation (EGPAF) Egyptian medicine, 1e2 EHR. See Electronic health record (EHR) Ehrlich, Paul, 10e11, 11f Ei Compendex, 724be743b Elderly population, 656 Elderly underrepresented in clinical trials, 193 Electroencephalogram (EEG), 623 Electromagnetic radiation, 11 Electron microscopy, 148 Electronic or online databases, 715e716 submission of applications, 593e594 surveillance for errors, 641 systems, 640 Electronic Common Technical Document (eCTD), 90 Electronic data capture (EDC), 537, 537f, 688e692 Electronic health record (EHR), 282, 413, 688 architecture, 688 diagram, 688, 692f medication management module, 689t system connectivity at national institutes, 688, 693f systems comprising, 691t in clinical research, 692e696, 695t CDS, 694 data characteristics, 694, 696t protocol order sets, 694e696 data, 288e289 interfaces with LIMS/LIS, 690t legislation and, 698e699 proliferation, 421 secondary use for clinical research, 698 systems, 552 Electronic medical record (EMR), 687 Electronic medication administration record (eMAR), 689t Electronic research administration (eRA), 592 Eligibility criteria, 318, 322 Elizabeth Glaser Pediatric AIDS Foundation (EGPAF), 626 EMA. See European Medicines Agency (EMA) Email, 102, 707e708 eMAR. See Electronic medication administration record (eMAR) Embargoes, 711e712 Ingelfinger Rule, 711e712 Embase, 717, 724be743b Classic subfile, 724be743b PICO search, 724be730b, 731f search interface, 730f EMBL-EBI. See European Molecular Biology Laboratory and European Bioinformatics Institute (EMBLEBI)

Embryos, 166 EMR. See Electronic medical record (EMR) EMTREE, 717, 724be743b ENA. See European Nucleotide Archive (ENA) Enablement, 463e464, 519 Endpoint adjudication, 424 Engagement rubric, 278 Enhanced usual care groups, 260 Enriched enrollment designs, 256 Ensembl, 746 Entrez Direct (EDirect), 724be743b Entrez Programming Utilities (eUtils), 724be743b Entry Terms, 717 Environment, 597 Environmental Protection Agency (EPA), 489 EORTC. See European Organization for Research and Treatments of Cancer (EORTC) EPA. See Environmental Protection Agency (EPA) EPC. See European Patent Convention (EPC) Epigenetic regulation, 654 Epigenetics, 654 Epilogue, 155e157 clinical research training, 157 drug development, 155e157 personal perspectives, 157 EPO. See European Patent Office (EPO); European Patent Organization (EPO) Equipoise, 27, 213, 329e330 Equivalence trials, 389e390 eRA. See Electronic research administration (eRA) Erythropoiesis-stimulating agents (ESAs), 284 Erythropoietin therapy, 425 ESAs. See Erythropoiesis-stimulating agents (ESAs) ESIs. See Early-Stage Investigators (ESIs) Espacenet, 724be743b Essential Inventions, 492 estimated glomerular filtration rate (eGFR), 395 Etanercept, 653 Ethical foundations, current requirements for research involving human subjects, 49 Ethical framework for clinical research, 23e27, 24t fair subject selection, 24e25 favorable risk/benefit ratio, 25 independent review, 25e26 informed consent, 26e27, 26t respect for enrolled subjects, 27 value and validity, 23e24 Ethical issues, 102e103 Ethical principles in clinical research, 19. See also Conduct of research

INDEX

clinical research from clinical practice, 19e20 codes of research ethics and regulations, 22e23 ethical considerations in randomized clinical trials, 27e29, 28t ethical framework for clinical research, 23e27 ethics and clinical research, 20 history of ethical attention to clinical research, 20e21 benefit to individual, 20 benefit to society, 20e21 community involvement in research, 21 protection of research subjects, 21 research as benefit, 21 research on bioethical questions, 23 Ethical requirements, 20 Ethical review board. See Institutional Review Board (IRB) Ethics committee (EC), 64 domain, 67e69 Etiology, Concept, and Prophylaxis of Childhood Fever, The, 8 EU. See European Union (EU) EudraCT. See European Union Drug Regulating Authorities Clinical Trials database (EudraCT) Eurasian Patent Organization (EAPO), 452 European Commission (EC), 87 European Federation of Pharmaceutical Industries and Associations (EFPIA), 89 European Medicines Agency (EMA), 116, 129, 649, 724be743b European Molecular Biology Laboratory and European Bioinformatics Institute (EMBL-EBI), 745 European Nucleotide Archive (ENA), 745 European Organization for Research and Treatments of Cancer (EORTC), 312 European Patent Convention (EPC), 452 European Patent Office (EPO), 452 national phase application, 479 European Patent Organization (EPO), 724be743b European Patents, 479e480 European Union (EU), 87, 116 European Union Drug Regulating Authorities Clinical Trials database (EudraCT), 116 eUtils. See Entrez Programming Utilities (eUtils) Eve, Christmas, 10 Event reporting systems, 640 Evidence synthesis in CER, 285e287 standards for, 286b Evidence-based medicine, 412e413, 724be743b EWG. See Expert working group (EWG) “Ex Parte” Reexamination, 478 Exact tests, 348 Excisional Atherectomy Trial, 424 Exclusionary criteria, 206e208

Exclusive licensing, 496 Exempt research activities, 51 Exemptions generic drugs, 471 medical practitioner exemption, 471 research-use exemption, 470e471 US Government as Infringer, 471 Existing infrastructure integration, 105e106 ExPASy. See Expert Protein Analysis System (ExPASy) Expedited review procedures, 51 Experimental studies, 232 designs for CER, 279e281 metaanalyses, 318 nonrandomized, 250 Experimental therapies, 20 Expert Protein Analysis System (ExPASy), 745 Expert working group (EWG), 90e91, 93 Explanatory exploratory/development grant (R21) applications, 608e609 research, 270 trials, 254e255, 420e421 Export control, 476 Exposure exposureeresponse relationship, 652e653 suspicion bias, 238e239 Extensible Markup Language (XML), 116 External validity, 251 Eye diseases, 719

F Fabry disease, 492 Factor analysis, 309, 312e313 Factorial designs, 256 using factorial or partial factorial design instead of parallel group design, 265 Factorial trial designs, 422 Facts & Comparisons eAnswers, 724be743b Failure mode and effects analysis (FMEA), 635, 638e639 Fair subject selection, 24e25 Fairness, 453 False-negative result machine, 431 False-positive fraction (FPF), 397 False-positive result machine, 431 Family foundation, 613e614 Family information bias, 238e239 FAOM. See First office action on merits (FAOM) FAQ. See Frequently Asked Questions (FAQ) Fast-track designation, 480e481 program, 79 FasterCures, 615 Fasting plasma glucose (FPG), 396e397 Favorable risk/benefit ratio, 25 FD&C Act. See US Federal Food, Drug, and Cosmetic Act (FD&C Act)

785 FDA. See US Food and Drug Administration (FDA) FDAAA. See Food and Drug Administration Amendments Act (FDAAA) FDAMA. See Food and Drug Administration Modernization Act (FDAMA) FDASIA. See Food and Drug Administration Safety and Innovation Act of 2012 (FDASIA) Federal “intramural” research, 493e495 Federal Aviation Administration, 497 Federal Circuit, 450e451 Federal Constitution, 449e450 Federal employees key statutes relating to, 486e487 senior, 494 Federal funding, 617e619 Federal policy, 35 Federal Privacy Act (1972), 162 Federal regulations, 49, 166, 181 Federal rulings, 39 Federal Technology Transfer Act (FTTA), 13, 493e495 history and philosophy, 494 key concepts and clauses, 494 subsequent supporting acts, 494e495 Federal Trade Secrets Act, 486 Federalwide assurances (FWAs), 50 Felbamate monotherapy trial, 343, 349e350 hypotheses for, 345 hypothesis tests for, 352 Felbamate trial, 343e344 Ferments of microorganisms, 8e9 Fetal tissue, protections for, 166e167 FIAC. See Fluoroiodoarabinosylcytosine (FIAC) Fiacitabine. See Fluoroiodoarabinosylcytosine (FIAC) Fialuridine (FIAU), 145 chemical structures, 144f clinical trials, 145t toxicity, 148e149 FIAU. See Fialuridine (FIAU) Fibrinolytic therapy, 420 Fibrinolytic Therapy Trialists (FTT), 415 analysis, 416 FIH studies. See First-in-human studies (FIH studies) “File wrapper”, 476 FIM. See Functional Impact Measure (FIM) Final product stability, 80 Final Statement on Sharing Research Data, 498 Financial/finance, 436 disclosure, 167e168 incentive, 439 and material contributions, 517 First International Study of Infarct Survival (ISIS-I), 413 hypotheses for ISIS-4 trial, 345e346, 352e353, 353t study, 265, 350

786 “First inventor to file” system, 461, 482 First office action on merits (FAOM), 477 “First to file” system, 461 First US public policy, 48 First-in-human studies (FIH studies), 77, 646e647, 649 First-to-invent, competing claims of, 461e462 Fisher, Sir Ronald Aylmer, 11e13 Fisher’s exact test, 350e351 “Fit for purpose”, 560 Five-step procedure, 94 Fixed-effect model, 319 Flags, 133, 134f Fleming, Alexander, 11 Flexner, Abraham, 11 FLI. See Functional Life Index (FLI) Flowcharting techniques, 635, 637, 642 Fluoroiodoarabinosylcytosine (FIAC), 144, 144f Fluoroiodoarauracil. See Fialuridine (FIAU) FMEA. See Failure mode and effects analysis (FMEA) fMRI. See Functional MRI (fMRI) FNIH. See Foundation for National Institutes of Health (FNIH) FOA. See Funding Opportunity Announcement (FOA) FOIA. See Freedom of Information Act (FOIA) Follow-up and analysis, 543 visits, 133 Food, Drug, and Cosmetic Act (1938), 20 Food and Drug Act, 73e74 Food and Drug Administration Amendments Act (FDAAA), 33e34, 38, 74, 114, 168 Food and Drug Administration Modernization Act (FDAMA), 74, 113 Food And Drug Administration Review, 82e83 Food and Drug Administration Safety and Innovation Act of 2012 (FDASIA), 74 For-profit budgets, 573 type, 571 Foreign donors, 101 Forest plot, 319e320 Formal ICH Procedure, 94e95 Fostering dissemination of information, data sharing, and patient engagement, 624e625 Foundation for AIDS Research (amfAR), 626 Foundation for National Institutes of Health (FNIH), 614 Foundations, 613e614 Founding Regulatory Members, 93

INDEX

Fourth International Study of Infarct Survival trial (ISIS-4 trial), 256, 430e431 FPF. See False-positive fraction (FPF) FPG. See Fasting plasma glucose (FPG) Franklin, Benjamin, 6e7 “Free” information resources, 723 Freedom of Information Act (FOIA), 38, 170, 486, 506e507, 711 FreePatentsOnline, 724be743b Frequently Asked Questions (FAQ), 724be743b Frist-Ganske Amendment, 471 FTT. See Fibrinolytic Therapy Trialists (FTT) FTTA. See Federal Technology Transfer Act (FTTA) Full-text searching, 724be743b “Fully packaged” course, 666e667 Functional imaging tools related to Phase 0 trial, 652 utilization of, 649e652 Functional Impact Measure (FIM), 311 Functional Life Index (FLI), 308e309 Functional measures, 303e313 examples of, 311 features to, 310e311 selecting, 311e312, 311t utility of, 310 Functional MRI (fMRI), 392 essential issues in, 393e394 Functional Status Questionnaire, 311 Funding agreements outside BayheDole Act, 493 Funding Opportunity Announcement (FOA), 592e593 Futility, 386e387 FWAs. See Federalwide assurances (FWAs)

G Galen, Claudius, 2 theory, 3 Ganglionic Blockers, 717 GAO. See US Government Accountability Office (GAO) Gaussian distribution. See Normal distributions GCC. See Gulf Cooperation Council (GCC) GCC-DR. See Gulf Central Committee for Drug Registration (GCC-DR) GCG. See Global Cooperation Group (GCG) GCP. See Good Clinical Practice (GCP) GCRCs. See General Clinical Research Centers (GCRCs) GDP. See Gross domestic product (GDP) GDS. See Genomic Data Sharing Plan (GDS) Gelsinger, Jesse, 63e64 Gemtuzumab ozogamicin, 655, 656t GenBank, 745e746 Gender, 179 Gene Expression Omnibus (GEO), 746 Gene Ontology (GO), 551 Gene(s), 745 patents on, 482 therapy, 142

General Clinical Research Centers (GCRCs), 671 Generalizability, 298, 415 Generalized linear models (GLMs), 369 Generic drugs, 471 Genetic Alliance, 615 Genetic diseases, 432e433 Genetic Testing Registry (GTR), 745 Genetic traits, 284e285 Genome databases, 745 Genome-wide association study (GWAS), 745 Genomic Data Sharing Plan (GDS), 598 Genomics, advent of, 418 Genzyme, 492e493 GEO. See Gene Expression Omnibus (GEO) Geographic variation, 296 Germ basis of fermentation, 8e9 Gilotrif. See Afatinib Glass transition temperature, 566e567 Gleevec. See Imatinib mesylate GLMs. See Generalized linear models (GLMs) Global Consortium for Biomarker Standardization, 620 Global Cooperation Group (GCG), 90e91 Global health field, 99e100 Global Index Medicus, 722 Global Utilization of Streptokinase and rt-PA for Occluded Coronary Arteries (GUSTO-I), 418, 427, 430 Glucosamine/chondroitin, 386 GO. See Gene Ontology (GO) Gold standard, 287, 397 “Gold” open access, 722 Good Clinical Practice (GCP), 23, 76e77, 538e539, 605, 678 guidances, 434e435 principles, 77t Good experimental design, 44e45 Good record keeping, 562 Google, 724be743b Google Books, 724be743b Google News, 724be743b Google News Archive, 724be743b Google Patents, 724be743b Google Scholar, 724be743b, 748e749 Google Translate, 724be743b, 744 optimal strategy, 723 Gottschalk v. Benson, 458 Governmental/government regulation, 437e438 scope of actual authority of government laboratories, 506 secrets and, 507e508 sponsors, 435 Grant(s), 479 applications, 600e605 funding, 608 hints and suggestions for preparing part of application, 601e605 PHS 398 specific human subjects sections, 603 PHS 398 specific research plan component, 601e603

INDEX

SF424 (R&R) project summary/ abstract, 601 instructions closelyesubmit complete and carefully prepared application, 601 makers in Health, 615 NIH, 592 planning application NIH grant application process and forms, 600e601 sufficient time to prepare application, 600 writing, 590 “Greater Rights” clause, 490 Greek dictums of medicine, 3 “Green” open access, 722 Greenberg Report, 129 Greenwood estimator, 376e377 Grey Literature Report, 724be743b Grey Matters, 724be743b GreySource, 724be743b Gross domestic product (GDP), 454 Group randomized trials (GRTs), 257, 369 Group sequential trials, 257, 368e369 Group-based exercise intervention, 203e204 Group-randomized trials, 280, 369 GRTs. See Group randomized trials (GRTs) GTR. See Genetic Testing Registry (GTR) Guide to Avoiding Financial and NonFinancial Conflicts or Perceived Conflicts of Interest in Clinical Research at NIH (Guide), 40e41 Guidelines for the Conduct of Research in the Intramural Research Program at NIH, 33 Guidelines.gov, 724be743b Guiding Principles in Medical Research Involving Humans, 13 Guillotin, Joseph-Ignace, 6e7 Guinea pig, 191 Gulf Central Committee for Drug Registration (GCC-DR), 87e88 Gulf Cooperation Council (GCC), 452 GUSTO-I. See Global Utilization of Streptokinase and rt-PA for Occluded Coronary Arteries (GUSTO-I) GWAS. See Genome-wide association study (GWAS)

H h-index, 747e748 H1N1 influenza pandemic, 182e183 Hansen, Norwegian Gerhard Armauer, 9e10 Harmonized CTD, 90 Harvard Program in Clinical Effectiveness, 665 Harvard’s Think Tank Search, 744 Harvey, William, 3e5 Hata, Sahachiro, 11 Hatch-Waxman Act, 471 Havrix, 503 Hawthorne effects, 259e260

HaybittleePeto boundary, 384e385 method, 384 Hazard function, 379 scores, 638 HBV infection. See Hepatitis B virus infection (HBV infection) HCM. See Hypertrophic cardiomyopathy (HCM) HCV. See Hepatitis C virus (HCV) Health disparities research, 295e296 equity, 99 and medicine, 703e704 Health, Education, and Welfare (HEW), 49 Health and Human Services (HHS), 492, 495 Health Assessment Questionnaire, 311 Health information exchange (HIE), 688 Health Information Technology for Economic and Clinical Health Act (HITECH Act), 287, 698, 699te700t Health Insurance Portability and Accountability Act (HIPAA), 56, 162, 168, 170, 299t, 532, 698 Health InterNetwork Access to Research Initiative (HINARI), 722 Health Level-7 standards (HL7 standards), 551, 687 Health maintenance organizations (HMOs), 293 Health Omnibus Programs Extension Act (HOPE Act), 113 Health research evaluating specific diseases and treatments, 296 geographic variation, 296 health disparities research, 295e296 monitoring secular trends, 294e295 uses of secondary data in, 294e296, 295t Health Research Alliance (HRA), 614e615 Health system data warehouse, 413 Health Technology Assessment Database (HTA). See International Health Technology Assessment (HTA) Health-care environment, 634 patents, 482 performance improvement tools, 635 quality, 640 Health-related quality of life (HRQL), 303e305. See also Patient-reported outcomes characteristics of measures of healthrelated instruments, 308t global measures, 312 measures, 305, 308 “Healthy migrant” effect, 238 “Healthy People”, 295e296 “Healthy worker” effect, 238 HED. See Human equivalent dose (HED) “Hematocrit”, 550 Hemoglobin molecule, 548

787 Heparin, 419 Hepatitis B virus infection (HBV infection), 145, 146f Hepatitis C virus (HCV), 274 Hepatitis drug development, 155 HER2. See Human Epidermal Growth Factor Receptor (HER2) Herceptin. See Trastuzumab hESCs. See Human embryonic stem cell (hESCs) Heterogeneity, 319e320 assessing treatment, 284e285 HEW. See Health, Education, and Welfare (HEW) Hexamethonium, 717 compounds, 717 MeSH record, 717, 718f HHMI. See Howard Hughes Medical Institute (HHMI) HHS. See Health and Human Services (HHS) HICs. See High-income countries (HICs) HIE. See Health information exchange (HIE) High reliability organizations (HROs), 635 High reliability principles in clinical research environment, 635 High-income countries (HICs), 100 HINARI. See Health InterNetwork Access to Research Initiative (HINARI) HIPAA. See Health Insurance Portability and Accountability Act (HIPAA) Hippocrates, 2 Hippocratic oath, 2 Historical cohort studies, 240 Historical foundations, current requirements for research involving human subjects, 47e49 Historical prospective study, 240 HITECH Act. See Health Information Technology for Economic and Clinical Health Act (HITECH Act) HIV. See Human immunodeficiency virus (HIV) HIV/AIDS. See Human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) HL7 standards. See Health Level-7 standards (HL7 standards) HMOs. See Health maintenance organizations (HMOs) Hoechst AG, 11 Holmes, Oliver Wendell, 7 HomoloGene, 745 HOPE Act. See Health Omnibus Programs Extension Act (HOPE Act) Hormone replacement therapy (HRT), 213, 245e246, 250, 405 Hospital discharge abstracts, 294 Hotel Security Checking Co. v. Lorraine Co., 458 Hourly rate, 572e573, 573t Howard Hughes Medical Institute (HHMI), 611, 663

788 Howard Hughes Medical Institute (HHMI) (Continued ) HHMIeNIH Research Scholars Program, 663 Medical Fellows Program, 663 HRA. See Health Research Alliance (HRA) HROs. See High reliability organizations (HROs) HRPP. See Human Research Protection Program (HRPP) HRPPPs. See Human research participant protection programs (HRPPPs) HRQL. See Health-related quality of life (HRQL) HRT. See Hormone replacement therapy (HRT) HTA. See International Health Technology Assessment (HTA) “Hub and spoke” system of financial services, 458e459 Human biospecimens, 38e39, 38t Human capital development, 617e619, 618t Human embryonic stem cell (hESCs), 166e167, 705 Human embryos, 166e167 Human Epidermal Growth Factor Receptor (HER2), 418 Human equivalent dose (HED), 649 Human Fertilisation and Embryology Authority, 167 Human immunodeficiency virus (HIV), 21, 51, 113 Human immunodeficiency virus/ acquired immune deficiency syndrome (HIV/AIDS), 234, 422 Human research participant protection programs (HRPPPs), 65 Human Research Protection Program (HRPP), 58, 63, 65, 65f Human resources, 436 Human subject, 51 current requirements for research involving, 47e50 ethical foundations, 49 historical foundations, 47e49 regulatory foundations, 49e50 protection, 57, 166, 603 research, 38e39 “111” application, 475 Hunter, John, 5 “Hybrid” organizations, 614 Hypertrophic cardiomyopathy (HCM), 363 “HyperX” multicore parallel processor, 496 Hypoalbuminemia, 148 Hypoprothrombinemia, 148 Hypothesis formulation of statistical hypotheses, 345e346 formulation intention to treat, 427 primary hypothesis, 427 secondary and tertiary hypotheses, 427

INDEX

generation, 547 misstatements and misconceptions, 353e354 mistakes in, 353 for motivating examples, 351e353 hypothesis tests for beta-interferon/ magnetic resonance imaging study, 351e352 hypothesis tests for felbamate monotherapy trial, 352 hypothesis tests for ISIS-4 trial, 352e353 and objectives, 222 one-sample hypothesis tests, 346e349 sample size calculations for, 362e366 one sample of data, 362e364 paired data, 364e366 two independent samples, 366e368 special considerations, 354e356 multiple comparisons, 355 multiple linear regression, 354e355 nonparametric vs. parametric tests, 355e356 one-way analysis of variance, 354 simple linear regression, 354e355 statistical inference, 342e343 testing, 341e345 two-sample hypothesis tests, 349e351 Hypothetical scenario, 503e504

I I SPY 2 TRIAL. See Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2 (I SPY 2 TRIAL) i2b2. See Informatics for Integrating Biology and the Bedside (i2b2) IACRN. See International Association of Clinical Research Nurses (IACRN) IACUC. See Institutional Animal Care and Use Committee (IACUC) IB. See Investigator’s brochure (IB) Ibn al-Haytham (Alhazen), 3 Ibrutinib, 650, 651t IC. See Institute or Center (IC) ICC. See Intraclass correlation (ICC) ICD-10-CM. See International Classification of Diseases Tenth Edition with Clinical Modifications (ICD-10-CM) ICD-10. See International Classification of Diseases-10 (ICD-10) ICD-9-CM. See International Classification of Diseases, Ninth Edition, with Clinical Modifications (ICD-9-CM) ICDRA. See International Conference of Drug Regulatory Authorities (ICDRA) ICF. See Informed consent form (ICF) ICF model. See International Classification of Function model (ICF model) ICH. See International Conference on Harmonization (ICH)

ICIDH. See International Classification of Impairments, Disabilities and Health (ICIDH) ICMJE. See International Committee of Medical Journal Editors (ICMJE) ICRP. See International Cancer Research Partnership (ICRP) ICTRP. See International Clinical Trials Registry Platform (ICTRP) Identifier (ID), 562e563 IDS. See Information Disclosure Statement (IDS) IE. See Interface engine (IE) iEdison. See Interagency Edison system (iEdison) IFPMA. See International Federation of Pharmaceutical Manufacturers and Associations (IFPMA) Imatinib mesylate, 650, 651t Imbruvica. See Ibrutinib Immunities generic drugs, 471 medical practitioner exemption, 471 research-use exemption, 470e471 US Government as Infringer, 471 Immunodeficiency, 142 Impaired kidney function, 656 Impairments, 303e304 Importation, 472 Imputation methods, 404 “in rem jurisdiction”, 472 Inadequate human resources, 100 INASP. See International Network for the Availability of Scientific Publications (INASP) Incentives for product development, 453e454 Incidence bias, 238 studies, 234e236 Inclusion changes in inclusion/exclusion criteria, 401e402 of children, 603 inclusionary criteria, 206e208 policy, 178e179 of women and minorities, 603 minorities, and children in research, 597 Incorporation bias, 399e400 IND. See Investigational new drug (IND) Independent censoring, 374 foundation, 613e614 review, 25e26 scientist awards, 608 Independent ethics committee. See Institutional Review Board (IRB) Index Nominum, 724be743b Indirect cost, 572 nature of clinical research data, 548 Indirect infringement. See Contributory infringement

INDEX

Individual career development awards, 607 Individual participant data (IPD), 121, 287 Induced infringement, 469e470 Industrial applicability, 459e460 Industry sponsors, 435 Inequitable conduct, 473, 475, 482e483 Inflammation in mediating sepsis, 321 Inflammatory marker, 185 Informatics for Integrating Biology and the Bedside (i2b2), 553, 555f Information gaps, 101 request, 83 resources for clinical researcher, 714 access and business models, 721e723 bioinformatics resources, 744e746 citation searching, 721 content and structure, 715e717 familiarity and currency, 723 organization and features of, 714e715 origin, 715 search capabilities, 717e721 sharing, 439 systems, 440 Information Disclosure Statement (IDS), 475 Information technology (IT), 436, 533 Informative censoring, 401e402 Informed consent, 26e27, 26t, 54e55, 54t, 102e103, 163e164 informed consenteadults, 164e165 informed consentechildren, 165e171 conflict of interest and financial disclosure, 167e168 protections for fetal tissue, human embryos, and hESCs, 166e167 public transparency, 168 recordkeeping and privacy protection, 168e171 process, 225, 227e229 content of, 164 Informed consent form (ICF), 572 Infringement, 448, 474 Ingelfinger Rule, 711e712 Inherent risks, 25 Innovation, 596 Inotropic therapies, 419, 432 Inspec, 724be743b Institute of Medicine (IOM), 34, 57, 63e64, 152e153, 179, 270e271, 305e306 Institute or Center (IC), 499, 595 Institutional Animal Care and Use Committee (IACUC), 603e604 Institutional procedures, 51 Institutional review boards (IRBs), 21, 37, 47, 50e56, 63, 79e80, 103, 116, 132, 142, 162, 180e181, 212, 220, 280, 299, 434, 532, 562, 571, 603, 694e696 clinical researchers and, 57 Common Rule, 50e51 criteria for IRBs approval of research, 52e56 current requirements for research involving human subjects, 47e50

domain, 67e69 evaluation and evolution of current system of research oversight and, 57e59 critique and proposed changes to IRBs operations, 58e59 proposed changes to current oversight of research with human subjects, 57e58 exempt research activities, 51 fees, 572, 579f membership, 51e52 minimal risk and expedited review procedures, 51 protocol review standards, 53t review of research, 51e56 continuing review of research, 56 criteria for institutional review board approval of research, 52e56 Institutional Review Board Membership, 51e52 review times by type of review, 72f template IRBs assessment of researchrelated risks and benefits, 54t Instrumental variables (IVs), 284 Insulin, 11 Integrated health-care systems, 440 Intellectual leadership, 435 Intellectual property (IP), 448e449 basic elements of patent application process, 473e483 copyrights, 483e485 core concepts of US patent law, 454e473 laws, 448 modern philosophy of patent law, 453e454 patents, 449e453 trade secrets, 485e487 trademarks, 485 Intended audience, 120e121 Intensity of intervention, 424e425 Intent-to-treat principle (ITT principle), 329e330, 334, 405, 427 Inter Partes Review (IPR), 478 “Inter Partes” Reexamination, 478 Interactions among therapies, 419e420 Interagency Edison system (iEdison), 489 Interface engine (IE), 687, 689te691t “Interference”, 461e462, 519 Interim analyses, 130, 133e136, 257 for efficacy, futility, and/or harm, 135e136 sequential designs, 136 stochastic curtailment tests, 136, 136f sample size recalculation, 134e135 Interim monitoring, 384e387 efficacy boundaries, 384e386 futility, 386e387 O’BrieneFleming boundary, 387 Internal validity, 251 International applications and filing procedures, 479e481 combining US and patent cooperation treaty filings, 480e481

789 general strategy notes, 481 patent cooperation treaty applications, 479 regional patent offices, 479e480 International Association of Clinical Research Nurses (IACRN), 672, 679 International Cancer Research Partnership (ICRP), 615 International Classification of Diseases, Ninth Edition, with Clinical Modifications (ICD-9-CM), 551, 552t International Classification of Diseases, Tenth Edition, with Clinical Modifications (ICD-10-CM), 551, 552t International Classification of Diseases-10 (ICD-10), 294 International Classification of Function model (ICF model), 304e307, 304f International Classification of Impairments, Disabilities and Health (ICIDH), 306 International Clinical Trials Registry Platform (ICTRP), 724be743b International Committee of Medical Journal Editors (ICMJE), 43, 43t, 113e114 International Conference of Drug Regulatory Authorities (ICDRA), 87 International Conference on Harmonization (ICH), 23, 76e77, 88, 130e131, 434e435, 542e543 early operations and achievements of, 88e91 efficacy guidelines, 96t guidelines relevanting to clinical research, 95e98 Management Committee, 92 sampling of topic areas addressed by, 89t technical harmonization process, 93e95 Five-Step Harmonization Procedure, 94e95 nomination and selection of topics for harmonization, 93e94 International Ethical Guidelines for Biomedical Research Involving Human Subjects, 22 International Federation of Pharmaceutical Manufacturers and Associations (IFPMA), 89 International harmonization, 481e482 International Health Technology Assessment (HTA), 286, 724be743b International Joint Efficacy Comparison of Thrombolytics, 429 International landscape, 115e116 International Network for the Availability of Scientific Publications (INASP), 722 International Organization for Standards (ISO), 550e551 International phase, 479 International Preliminary Examination Report (IPER), 479

790 International regulation of drugs and biological products future work in regulatory harmonization, 98 International Council on Harmonisation early operations and achievements, 88e91 guidelines relevant to clinical research, 95e98 technical harmonization process, 93e95 New International Council on Harmonisation financing, 92e93 membership in, 91e92 organization of, 92 recent evolution and reforms, 91 International Search Report (ISR), 479 International Society for Biological and Environmental Repositories (ISBER), 559e560 International Society for Stem Cell Research, 167 Internet Search Engines, 724be743b Interstitial lung disease, 505 Interval censoring, 402 estimation, 343 Intervention development, 258 Interventional studies, 117 Intraclass correlation (ICC), 369 Intracranial hemorrhage, 432e433 Intramural Research Program (IRP), 33, 52e56, 265 Introduction to the Principles and Practice of Clinical Research (IPPCR), 667 Inventions collaboration and, 514e518 by National Institutes of Health, 497e503 patent and patent-related policies, 497e499 portfolio size and scope, 499 Inventorship, 468 Inventory management systems, 563 Inverse-variance method, 319 Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2 (I SPY 2 TRIAL), 653 Investigational drug, 77 Investigational new drug (IND), 132, 500 application, 532, 646 exemption, 150 safety reports, 81 submission, 76e78 Investigator(s), 80, 596 bias, 330 initiated applications, 592 Investigator’s brochure (IB), 77e78, 81, 578 Invoiced items, 577e581, 581t, 582f IOM. See Institute of Medicine (IOM) IP. See Intellectual property (IP) IPD. See Individual participant data (IPD)

INDEX

IPER. See International Preliminary Examination Report (IPER) IPPCR. See Introduction to the Principles and Practice of Clinical Research (IPPCR) IPR. See Inter Partes Review (IPR) IRBs. See Institutional review boards (IRBs) Irinotecan, 650, 651t IRP. See Intramural Research Program (IRP) Irresponsible authorship, 43 ISBER. See International Society for Biological and Environmental Repositories (ISBER) ISIS-4 trial. See Fourth International Study of Infarct Survival trial (ISIS-4 trial) ISIS-I. See First International Study of Infarct Survival (ISIS-I) ISO. See International Organization for Standards (ISO) ISR. See International Search Report (ISR) IT. See Information technology (IT) ITC. See US International Trade Commission (ITC) Item response theory, 309 ITT principle. See Intent-to-treat principle (ITT principle) IVs. See Instrumental variables (IVs) Ixquick, 744

J JAMA. See Journal of American Medical Association (JAMA) Japanese Pharmaceutical Manufacturers Association (JPMA), 89 Jefferson Dr. Tom, 724be743b Jenner, Edward, 5 vaccination for smallpox, 7 Jesty, Benjamin, 5e6 JIF. See Journal Impact Factor (JIF) Johns Hopkins internal investigators, 714 Joint authorship, 484 Joint inventorship, 468 Joliot-Curie, Irene, 11 Journal Impact Factor (JIF), 747e748 Journal of American Medical Association (JAMA), 43, 296, 505 JPMA. See Japanese Pharmaceutical Manufacturers Association (JPMA) Jurin, James, 5e6 “Jurisdiction over person”, 472 Justice, 49

K K-M estimator. See KaplaneMeier estimator (K-M estimator) K-RITH. See KwaZulu-Natal Research Institute for Tuberculosis and HIV (K-RITH) K22 career transition awards, 608 K30 awards, 665 K99/R00 pathway to independence award, 607e608

KaplaneMeier estimator (K-M estimator), 375e377, 401 calculation and formula for estimation, 375e376 of variance, 376e377 construction of, 376t Karkinos, 2 Keeping information up-to-date, 117e118 KefauvereHarris amendments, 13, 74 bill, 48 KEI. See Knowledge Ecology International (KEI) Kelsey, Dr. Francis, 74 Kennedy Space Center, researchers at, 497 Kidney injury biomarkers, 650 King, Edmund, 3e4 King Louis XVI of France, 6e7 KL2 program, 665e666 Knowledge Ecology International (KEI), 492 Koch, Robert, 9, 9f KwaZulu-Natal Research Institute for Tuberculosis and HIV (K-RITH), 621

L Lab Compact, 36 Laboratory animals, 39 Laboratory Corporation of America (LabCorp), 456 Laboratory Information Management System (LIMS), 690t Laboratory Information System (LIS), 688 Laboratory methods and measurement error, 216 Lactating women, 656 Language, modifications to CRADA, 516 LAR. See Legally authorized representative (LAR) Last observation carried forward (LOCF), 404 Last rank carried forward (LRCF), 404 Lavoisier, Antoine, 6e7 Law, 154e155, 161 LDS. See Limited data set (LDS) “Learn and Apply” approach, 653 “Learn and confirm” model, 77 Learning health-care systems, 287 Left censoring, 402 Legal issues in clinical research, 161. See also Unanticipated risk in clinical research content of informed consent processes, 164 data sharing and individual consent, 171e173 informed consenteadults, 164e165 informed consentechildren, 165e171 protecting individual participant interests, 162e164

INDEX

independent review and monitoring, 162e163 informed consent, surrogate consent, advance directives, 163e164 Legal issues related to data reporting, 542e543 Legal scope of practice issues, 680 regulations govern practice and liability in clinical research settings, 680 Legally authorized representative (LAR), 164e165 Legislation, 698e699 FDA guidance for EHR in clinical research, 699 HITECH Act, 698, 699te700t MACRA, 698 Lens, The, 724be743b Letter of Intent (LOI), 518 Letters Patent, 449 Lexicomp Online, 724be743b Liability in clinical research settings, 680 License, 468e469, 519 Licensing, 493 by agency exclusive and coexclusive licensing, 496 results, 496e497 scope of licensing authority, 495 various agency missions, 495 Inventions By National Institutes of Health, 497e503 National Institutes of Health licensing program, 499e503 Limited data set (LDS), 170 LIMS. See Laboratory Information Management System (LIMS) Lind, James, 4e5, 6f, 6t Lipase, 148 Lipid-lowering medication effects, 206 LIS. See Laboratory Information System (LIS) Lister, Joseph, 8 Literature searching, 749 Litigation, practical issues of, 472e473 Liver disease, 312 function, 656 Living organisms, 455e458 LMICs. See Low-and middle-income countries (LMICs) Loan Repayment Program (LRP), 607, 609 Local IRBs, 572 capacity, 104e105 Local setting, 103e104 Local stakeholders, 104 Location and contact information, 116 LOCF. See Last observation carried forward (LOCF) Log-rank test, 377e378, 377t, 380 Logical Object Identifiers Names and Codes (LOINC), 551 LOI. See Letter of Intent (LOI) LOINC. See Logical Object Identifiers Names and Codes (LOINC) London Agreement, 480

Long Range Planning Committee of the American Neurological Association, 668 Long-term plans, 105 Longitudinal observational studies, 242 Longitudinal studies, 232 Low-and middle-income countries (LMICs), 100e101 Low-density lipoprotein, 185 Lower, Richard, 3e4 Lower error variance, 206 LRCF. See Last rank carried forward (LRCF) LRP. See Loan Repayment Program (LRP) Lumiracoxib, 655, 656t Lumped across arms, 388 Lung cancer, 705

M “Machine or transformation” test, 459 Macleod, J. J. R., 11 MACRA. See Medicare Access and CHIP Reauthorization Act (MACRA) Macrina, Frank, 40 Madey v. Duke University, 470e471 Magic bullet, 11 Magnesium, 345e346, 352e353 Magnetic resonance imaging (MRI), 252e253, 342e343 Major depressive disorder (MDD), 252 patients with, 262 Mantel-Haenszel method, 319 Manual of operating procedures (MOP), 213e216, 214t adherence, 214e215 dose ranging, 215 laboratory methods and measurement error, 216 masking, 215 recruitment and retention, 214 reporting results, 216e217 treatment fidelity, 216 Manual of procedures (MOO). See Manual of operating procedures (MOP) MapViewer, 746 MAR. See Missing at random (MAR) “March-In”ex 203, 490e493 Abbott and Pfizer study, 492 CellPro, Inc., 491 Genzyme, 492e493 Marcus, Adam, 34e35 Marginal structural models, 283 Marketing approval/licensure, 81e83. See also Postapproval application, 82 Food And Drug Administration Review, 82e83 pre-new drug application/biologics license application submission, 81e82 “Mask Works” rights, 448 Masked protocols, 210 Masking. See Blinding Massive data sets, 393e394

791 Material Transfer Agreements (MTA), 39, 226, 498, 566 anatomy, 509e512 confidentiality, 510 materials, 510 parties, 509e510 “reach-through” rights, 512 rights in materials, 510 termination, 511 uses, 510 warranties and indemnification, 511 key specialized materials in repositories, 513e514 software transfer agreements, 514 routine, 509 Mather, Cotton, 5e6 Maximum likelihood methods, 404 Maximum recommended starting dose (MRSD), 649 Maximum tolerated dose (MTD), 252 Mayo Chakrabarty to Mayo to Myriad, 456e458 Collaborative Svcs. v. Prometheus Labs., Inc., 456e457 McClintock, Barbara, 13e14 McNemar’s test, 365e366 MDD. See Major depressive disorder (MDD) MDIC. See Medical Device Innovation Consortium (MDIC) MDePhD programs, 662e663 Mean, 342 survival, 380 Measures of function. See Functional measures MedDRA. See Medical Dictionary for Regulatory Activities (MedDRA) Media, 153e154, 704 bread and butter, 704 celebrity, 705 engaging, 707 interview, 708e710 investigative reporters, 710e711 social, 707 talking to reporters, 706e707 Median, 342 survival, 380 Medical Device Innovation Consortium (MDIC), 434e435 Medical devices, 274 Medical Dictionary for Regulatory Activities (MedDRA), 90, 92, 541, 551 Medical Education Partnership Initiative (MEPI), 104 Medical errors, 633e634 Medical justification, 422 Medical Outcomes Survey Short Form 36, 312 Medical practice, 412 Medical practitioner exemption, 471 Medical Research Scholars Program (MRSP), 663

792 Medical Scientist Training Program (MSTP), 662 Medical student training opportunities, 663e664 Medical Subject Headings (MeSH), 717, 724be743b Medicare Access and CHIP Reauthorization Act (MACRA), 698 Medicare Modernization Act (MMA), 270, 271b Medicare Prescription Drug, Improvement, and Modernization Act. See Medicare Modernization Act (MMA) Medicine, making news in, 704 Medicines and Healthcare Products Regulatory Agency (MHRA), 90 Medicines for Malaria Venture (MMV), 624 MEDLINE, 714, 724be743b database record, 715e716 MEDLINE/PubMed, 717, 723 record 11834748, 716f, 717 MedWatch, 724be743b Melancholia, 2 Membership, 91 bias, 238 of EWG, 93 IRB, 236e239 in New International Council on Harmonisation, 91e92 review committees, 606 Memoranda of Understanding (MOU), 226 Menopausal hormone therapy (MHT), 185 Mental health, 307 Mentor/mentoring, 36, 104 career development awards, 607 clinical scientist development award (K08), 607 mentoretrainee relationship goal, 36 patient-oriented research career development award (K23), 607 MEPI. See Medical Education Partnership Initiative (MEPI) 6-Mercaptopurine, 144 Merck KGaA v. Integra Life Sciences, Inc., 471 Merck vs. Integra, 471 Mere associations, 455e456 LabCorp v. Metabolite, 456 patents on, 482 Meridia. See Sibutramine MESA. See Multi-Ethnic Study of Atherosclerosis (MESA) MeSH. See Medical Subject Headings (MeSH) Mesmerism, 6e7 Message, 703 core, 704, 709 health, 710 Meta-analysis, 317, 430e431, 430t. See also Survival analysis of clinical trials of antiinflammatory agents in, 321e323 analyzing data, 322e323

INDEX

data extraction, 322 eligibility criteria, 322 formulating question, 321e322 identifying studies, 322 inflammation in mediating sepsis, 321 Funnel plot with horizontal axis, 323f techniques, 318e321 eligibility criteria, 318 formulating question, 318 identifying studies and data extraction, 318 statistical analysis, 318e321 Meta-analysis of Observational Studies in Epidemiology (MOOSE), 724be743b Metadata, 549e550 Metaregression, 321, 430 Metchnikoff, Elie, 10, 10f Methodology Standards, 276 MHLW. See Ministry of Health, Labor and Welfare (MHLW) MHRA. See Medicines and Healthcare Products Regulatory Agency (MHRA) MHT. See Menopausal hormone therapy (MHT) MI. See Myocardial infarction (MI) Mibefradil, 432 Microarrays, essential issues in, 393e394 “Microdosing” trials. See “Phase 0” trials Micromedex, 715, 724be743b Microscope, 4 Microsoft Academic citation data, 748 full text journal article resources, 724be743b Microsteatosis, 146e147 MID. See Minimally important clinical difference (MID) Midcareer investigator award in patientoriented research, 608 Minimal data set, 116e118 Minimal risk, 51e52 Minimally important clinical difference (MID), 430 Minimization, 336 Ministry of Health, Labor and Welfare (MHLW), 89 Minnesota Coronary Survey Dietary Trial, 396 Misconceptions, 244e247 CONSORT statement, 262 large definitive study, 262e263 choosing control group, 263e264 confusing placebo response and regression to mean, 264e265 using factorial or partial factorial design instead of parallel group design, 265 failing to decrease variance, 263 failing to increase treatment effect, 263 placebo treatment in long-term studies, 264

small, open-label, nonrandomized, uncontrolled studies, 265e266 unethical placebo groups, 264 Misinterpretations, 244e247 confusing causation, prediction, association, and confounding, 245 design randomized study, 246e247 implying causation, 244e245, 245t misinterpreting relative measures, 244 observational and randomized studies, 245e246 ORs and relative risks similar magnitude, 244, 244t safe observational study, 247 trusting bivariate associations based on observational study data, 244 Missing at random (MAR), 399, 403 Missing completely at random, 403 Missing data, 403 in clinical research, 403 methods for handling, 404 minimizing, 405 mistakes in, 404e405 types of, 403e404 Mission and organization of NIH, 590e591 Mitochondrial toxicity, 149 Mixed model, 404 MMA. See Medicare Modernization Act (MMA) MMRF. See Multiple Myeloma Research Foundation (MMRF) MMV. See Medicines for Malaria Venture (MMV) MMWR. See Morbidity and Mortality Weekly Report (MMWR) Model-based drug development, 653e654 Modeling and simulation, 653e654 Modernization Act, 655 Monitoring secular trends, 294e295 Montagna, Bartolomeo, 3 Montagu, Lady Mary Worley, 5e6 MOOSE. See Meta-analysis of Observational Studies in Epidemiology (MOOSE) MOP. See Manual of operating procedures (MOP) Morbidity and Mortality Weekly Report (MMWR), 234 “Morella Bill”. See National Technology Transfer Advancement Act of 1995 (NTTAA) Mortality, 548 Morton, William Thomas Green, 8 MOST. See Multiphase Optimization Strategy (MOST) MOU. See Memoranda of Understanding (MOU) MRC. See British Medical Research Council (MRC) MRCTs. See Multiregional Clinical Trials (MRCTs) MRI. See Magnetic resonance imaging (MRI)

INDEX

MRSD. See Maximum recommended starting dose (MRSD) MRSP. See Medical Research Scholars Program (MRSP) MS. See Multiple sclerosis (MS) MSTP. See Medical Scientist Training Program (MSTP) MTA. See Material Transfer Agreements (MTA) MTD. See Maximum tolerated dose (MTD) Multi-Ethnic Study of Atherosclerosis (MESA), 185 Multiarm clinical trials, 368 Multicenter trials, 434 Multidisciplinary guidelines, 88 Multidomain measures, 304 Multiphase Optimization Strategy (MOST), 210, 258 Multiple comparisons, 355 Multiple control groups, 260e261 Multiple linear regression, 354e355 Multiple Myeloma Research Foundation (MMRF), 439, 624 Multiple principal investigators, 594 Multiple randomization, 422 Multiple regression, 355 Multiple sclerosis (MS), 281 Multiregional Clinical Trials (MRCTs), 98 Multivariate regression, 355 Mycobacterium leprae (M. leprae), 9e10 Mylotarg. See Gemtuzumab ozogamicin Myocardial infarction (MI), 120, 206, 238, 256, 343e344, 373, 401 MyRI, 748 Myriad, 41 Chakrabarty to Mayo to Myriad, 456e458 genetics, 457

N Naloxone, 641 Narcan. See Naloxone NASA. See National Aeronautics and Space Administration (NASA) National Academy of Medicine, 438 National advisory councils and boards, review by, 599e600 National Aeronautics and Space Administration (NASA), 493, 495, 497 National Bioethics Advisory Commission, 63e64 National Breast Cancer Coalition (NBCC), 626e628 National Cancer Institute Common Toxicity Criteria for Adverse Events (NCI CTCAE), 252 National Center for Advancing Translational Sciences (NCATS), 289, 553, 555, 665e666 National Center for Biotechnology Information (NCBI), 555e556, 721, 724be743b, 745 Bookshelf, 724be743b Learn, 745 YouTube channel, 745

National Center for Health Statistics (NCHS), 309e310 National Center for Research Resources (NCRR), 674 National Center for Toxicological Research, 74 National Commission, 49 National Cybersecurity Assessment and Technical Services (NCATS), 490 National documents, 481 National Evaluation System for healthcare Technology (NEST), 414 National Foundation, 616 for Infantile Paralysis, 616 National Health and Nutrition Examination Survey (NHANES), 309e310 National Health Service (NHS), 101, 724be743b National Heart, Lung, and Blood Institute (NHLBI), 590e591 National Institute for Health and Care Excellence (NICE), 724be743b National Institute of Allergy and Infectious Diseases (NIAID), 590e591 National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), 43 National Institute of Biomedical Imaging and Bioengineering (NIBIB), 619 National Institute of Neurological Disorders and Stroke (NINDS), 265 National Institute of Standards and Technology, 496 National Institutes of Health (NIH), 13, 21, 33, 48, 64, 74, 111e112, 129, 143, 152, 161e162, 169, 177, 179t, 226, 271b, 304, 421, 434, 487, 536, 551, 560, 575, 590, 591f, 591t, 617, 661, 703, 705, 724be743b, 745 application procedures for NIHefunded clinical trials, 605 challenges to enrolling volunteers, 181e182 core purpose of technology transfer enterprise, 502 demographic trends in clinical trial participation, 183e184 electronic submission of applications, 593e594 extramural funding mechanisms, 592, 592t FDA-approved veterinary vaccines, 502e503 funding announcements funding opportunity announcements, 592e593 requests for applications and PA in NIH guide, 593 grant programs for clinical researchers at stages in careers, 607e609 hints for preparing grant applications, 600e605 inclusion policy, 183e184 institute/center home pages, 609

793 inventions by, 497e503 patent and patent-related policies, 497e499 portfolio size and scope, 499 IRB, 180e181 K30 program, 665 licensing program, 499e503 mission and organization, 590e591 multiple principal investigators, 594 national institutes of health licensing process, 501e502 peer review process, 590, 594e600 policies, 177e179 best practices for licensing genomic inventions, 501 general, 506 Policy on Dissemination of NIH-Funded Clinical Trial Information, 168 PrecisionMedicine Initiative, 747 responsibilities of NIH staff, 591e592 review of agreement, 518 revising unsuccessful applications, 605e606, 605t procedural error during peer review, 606 revise and resubmit application, 606 revise application, 605e606 study section inappropriate or biased, 606 Revitalization Act (1993), 177e178 codified NIH’s inclusion policy, 184e185 royalty management, monitoring, and enforcement, 502 scientific considerations and peer review, 179e180 scope of licensing authority, 501 stay information about NIH peer review, 609 types and structure of national institutes of health licenses, 501 women of childbearing potential, pregnant women, and children, 182e183 National Institutes of Health Clinical Center (NIH CC), 195e196, 199, 503, 635e637, 642, 671, 672f, 679, 688 core curriculum, 666e668, 667t Patient Advisory Group, 189 National Library of Medicine (NLM), 111e112, 555e556, 722, 724be743b, 745 National Network of Libraries of Medicine (NN/LM), 722 National Patient-Centered Clinical Research Network (PCORnet), 438 “National phase”, 479 National Research Act (1974), 49 National Research Corporation (NRC), 642 National Research Council (NRC), 405 National stage application, 519 National Technology Transfer Advancement Act of 1995 (NTTAA), 494 National Wildlife Research Center, 496

794 Natural and Political Observations Mentioned in a Following Index, and Made Upon the Bills of Mortality, 4 Natural disasters, 101e102 Nature article, 41e42 Nazi human experimentation, 13 NBCC. See National Breast Cancer Coalition (NBCC) NCATS. See National Center for Advancing Translational Sciences (NCATS); National Cybersecurity Assessment and Technical Services (NCATS) NCBI. See National Center for Biotechnology Information (NCBI) NCDs. See Noncommunicable diseases (NCDs) NCHS. See National Center for Health Statistics (NCHS) NCI. See US National Cancer Institute (NCI) NCI CTCAE. See National Cancer Institute Common Toxicity Criteria for Adverse Events (NCI CTCAE) NCRR. See National Center for Research Resources (NCRR) NCT number. See ClinicalTrials.govdidentifier NDA. See New Drug Application (NDA) Negative predictive value (NPV), 397 NEnglJMed. See New England Journal of Medicine (NEnglJMed) Neosalvarsan, 11 NEST. See National Evaluation System for healthcare Technology (NEST) Nested case-cohort studies, 241 Nested case-control designs, 241 studies, 241 Nested within cohort, studies, 239e243 Network, 615 meta-analysis, 286 Networked Digital Library of Theses and Dissertations, 724be743b Neuroinflammation, 619e620 New Drug Application (NDA), 90, 646 New England Journal of Medicine (NEnglJMed), 21, 711 New International Council on Harmonisation financing, 92e93 membership in, 91e92 organization of, 92 New topic proposals, 93 New York Times, 724be743b New/early-stage investigators, research project grant applications from, 598 News, 710 Google, 724be743b making news in science and medicine, 704 NHANES. See National Health and Nutrition Examination Survey (NHANES)

INDEX

NHLBI. See National Heart, Lung, and Blood Institute (NHLBI) NHS. See National Health Service (NHS) NIAID. See National Institute of Allergy and Infectious Diseases (NIAID) NIAMS. See National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) NIBIB. See National Institute of Biomedical Imaging and Bioengineering (NIBIB) NICE. See National Institute for Health and Care Excellence (NICE) Nightingale, Florence, 11, 12f NIH. See National Institutes of Health (NIH) NIH CC. See National Institutes of Health Clinical Center (NIH CC) NIH Office of Technology Transfer (NIH OTT), 40, 499, 502 NIH OTT. See NIH Office of Technology Transfer (NIH OTT) NINDS. See National Institute of Neurological Disorders and Stroke (NINDS) NLM. See National Library of Medicine (NLM) NN/LM. See National Network of Libraries of Medicine (NN/LM) NNT. See Number needed to treat (NNT) No Observed Adverse Events Level (NOAEL), 649 Non-ICH drug regulators, 91 Non-MSTP MDePhD programs, 663 Nonclinical investigation, 76 Noncommunicable diseases (NCDs), 100 Nonconcurrent cohort approach, 240 studies, 240 Nonconcurrent prospective study, 240, 240f Nongovernment sector, 613 Nonindependent censoring, 401e402 Noninferiority, 429, 429t margin, 389e390 trials, 389e390, 429 Nonobviousness, 519 “obvious to try”, 463 principles, 462 “secondary considerations”, 462e463 Nonparametric tests, 355e356 Nonpharmacologic interventions, 259 studies, 203 treatment studies, 252 Nonpracticing entities (NPEs), 482e483 Nonprofit Contractors, 489 Nonrandomized studies, 265e266 Nontherapeutic research, 167 Normal distributions, 361 “Not-for profit” studies, 573 type, 571 Notice of Proposed Rule Making (NPRM), 23, 114e115

Novelty, 519 competing claims of first-to-invent, 461e462 principles, 460e461 Now-abolished “Statutory Invention Registration”, 475 NPEs. See Nonpracticing entities (NPEs) NPRM. See Notice of Proposed Rule Making (NPRM) NPV. See Negative predictive value (NPV) NRC. See National Research Corporation (NRC); National Research Council (NRC) NTTAA. See National Technology Transfer Advancement Act of 1995 (NTTAA) Nuisance parameters, 135 Null hypothesis (H0), 343e346, 353, 378 Number needed to treat (NNT), 415 Nuremberg Code in 1947, 13, 22e23, 47e48, 163e164 Nurse(s), 672e673, 681 certification for nurses, 679 clinical research, 672 career potential for nurses in, 682e683, 683f meeting need for nurses to fill clinical research roles, 683 nursing role in community-based research, 683e684 role, 673e674, 675t scientist, 673 supporting transition of nurses into clinical research from clinical practice, 684

O O’BrieneFleming boundary, 387 OAMPI, 452 OAPI. See Organisation Africaine de la Proprie´te´ Intellectuelle (OAPI) Obamacare, 295 OBBR. See Office of Biorepositories and Biospecimen Research (OBBR) “Object code”, 484 Objectivity of clinical data, 548e549 OBRC. See Online Bioinformatics Resources Collection (OBRC) Observational Medical Outcomes Partnership model (OMOP model), 288e289 Observational studies, 117, 232, 279, 438e439 attributable risk, 243 case reports and case series, 233e234 case-control studies, 236e239 causal inference in, 405e406 CER studies, 283e284 cohort studies, 239e243 design, 232, 232t ecological studies, 233 mistakes, misconceptions, and misinterpretations, 244e247 ORs, 243 RR, 243 single time point studies, 234e236

795

INDEX

Observational study designs for CER, 281e282 Observational Study Monitoring Board (OSMB), 247 Observed treatment effect, 135 Observers, 92 Observership, 91e92 “Obvious to try”, 463 Occurrence reporting systems (ORSs), 640 Oclassen Pharmaceuticals, 147 OCT. See Optimal Cutting Temperature (OCT) Odds ratio (OR), 239, 243, 318 ORs and relative risks similar magnitude, 244, 244t survival, 322f OddzOn application, 466 Off-duty hours, 41 “Off-label use”, 653 Off-target effects, 131 Office for Human Research Protections (OHRP), 50, 161e162, 538e541, 603 mailing lists, 542t website, 50 Office for sponsored research/office of clinical research development, 105 Office of Biorepositories and Biospecimen Research (OBBR), 560 Office of National Coordinator for Health Information Technology (ONC), 698 ONC-approved EHR, 699 Office of Research Integrity (ORI), 34, 151 Office of Scientific Integrity (OSI), 151 Offices for Research and Technology Applications (ORTAs), 494 OHRP. See Office for Human Research Protections (OHRP) OMB. See US Office of Management and Budget (OMB) eomic technologies, 652 OMIM. See Online Mendelian Inheritance in Man (OMIM) OMOP model. See Observational Medical Outcomes Partnership model (OMOP model) On the Contagiousness of Puerperal Fever, 7 On-site computer support and help desk availability, 536 ONC. See Office of National Coordinator for Health Information Technology (ONC) One proportion z-test, 356 One-dimensional barcodes (1D barcodes), 563, 563f One-sample hypothesis tests, 346e349. See also Two-sample hypothesis tests binary data, 347e348 determining statistical significance, 346e347 example, 348e349 tests for normal continuous data, 346 One-sided tests, 344, 346, 351e352 One-way analysis of variance, 354

Online Bioinformatics Resources Collection (OBRC), 746 Online communication, 305 Online Mendelian Inheritance in Man (OMIM), 745 Open Access Theses and Dissertations database, 724be743b Open Access, 615, 722 Open Archives Initiative, 722 Open Directory Project. See DMOZ Open-label studies, 265e266 OpenDOAR, the Directory of Open Access Repositories, 724be743b OpenGrey, 724be743b OpenThesis, 724be743b Operating foundations, 614 Operational organization for large-scale clinical research, 431f, 433e437 coordinating functions, 433f, 435 DCC, 435 intellectual leadership, 435 site management organization, 435 executive functions DSMC, 434 industry or government sponsors, 435 IRB, 434 regulatory authorities, 434e435 steering committee, 433 supporting functions contracts management, 436 finance, 436 human resources, 436 information technology, 436 pharmacy and supplies, 436 project management and regulatory affairs, 436e437 randomization services, 436 Opportunities, 104 Optimal Cutting Temperature (OCT), 564 Optimized standard of care (oSOC), 389 Options “after issuance”, 478 OR. See Odds ratio (OR) Oransky, Ivan, 34e35 Organisation Africaine de la Proprie´te´ Intellectuelle (OAPI), 452 Organization domain, 66e67 ORI. See Office of Research Integrity (ORI) Ornithine transcarbamylase deficiency (OTC deficiency), 142 ORSs. See Occurrence reporting systems (ORSs) ORTAs. See Offices for Research and Technology Applications (ORTAs) OSI. See Office of Scientific Integrity (OSI) Osler, William, 10e11 OSMB. See Observational Study Monitoring Board (OSMB) oSOC. See Optimized standard of care (oSOC) OTC deficiency. See Ornithine transcarbamylase deficiency (OTC deficiency) Ovarian cancer, 397

diagnosis, 398e399 national advocacy organization for, 189 Overall impact/priority score and percentiles, 599 Overcommitment, 41 Overhead, 572 Ownership, transfers of, 468e469

P P-value. See Probability value (P-value) PA. See Program Announcement (PA) Pacific Northwest National Laboratory, 496 PAD. See Pharmacologically active dose (PAD) PAHO. See Pan American Health Organization (PAHO) Paired binary data, calculations for, 365e366 Paired continuous data, calculations for, 364e365 Paired data, 349, 364e366 calculations for paired binary data, 365e366 calculations for paired continuous data, 364e365 Paired t-test, 349, 356 Paired test, 393 Paired z-tests, 349 Palliative care role, 196 Pan American Health Organization (PAHO), 87e88 Pan American Network for Drug Regulatory Harmonization (PANDRH), 87e88 Pancreatitis, 146 Parallel groups designs, 256e257 using factorial or partial factorial design instead of, 265 Parameter, 342 Parametric tests, 355e356 “Parent”, 165 application, 475 Paris Convention for Protection of Industrial Property (1883), 451 Parker v. Flook, 458 Partial area under the curve (pAUC), 398 Participant(s) baseline characteristics, 131 flow, 131 safety, 130 monitoring, 130e131 treatment adherence, 133 Participants’ disposition. See Participant flow Partnership for Research in Ebola Virus in Liberia II (PREVAIL II), 389 Pasteur, Louis, 8e9, 9f Patent Act (1952), 450 Patent and patent-related policies general, 497 Research Tools, 497e498 sharing of data and model organisms, 498e499

796 Patent application process, 473e483 content of patent application, 473e475 claims, 473e474 duty of disclosure and “inequitable conduct”, 475 one invention per application, 474e475 specification, 473 technical items, 474 efforts to altering US patent laws, 481e483 international applications and filing procedures, 479e481 US applications, 475e478 Patent cooperation combining US and patent cooperation treaty filings, 480e481 treaty applications, 479 Patent Cooperation Treaty (PCT), 452, 479, 519 Patent documents, 724be743b Patent equivalents, 724be743b Patent infringement, United States, 469e473 civil liability, 469 contributory and induced infringement, 469e470 “declaratory judgment” actions, 472 importation, 472 International Trade Commission, 472 major defenses, 470 practical issues of litigation, 472e473 remedies, 471e472 specific exemptions and immunities, 470e471 Patent law, modern philosophy of, 453e454 economic engine, 454 fairness, 453 incentives for product development, 453e454 “Quid Pro Quo”, 453 Patent Prosecution Highway (PPH), 453 Patent Reform Act, 451, 461, 464e465, 478, 481 Patent trolls. See Nonpracticing entities (NPEs) Patent(s), 448e449 before American Revolution, 449 critical laws, 487e503 design, 454 federal “intramural” research, 493e495 federal funding of private “extramural” research, 487e493 federally supported research, 487e503 licensing, 469 life, 476e477 national documents, 454 patenting and licensing by federal agencies, 495e503 plant, 454 prosecution, 475 rights conveyed by, 455 treaties, 451e453 United States (1789e1951), 450

INDEX

United States Constitution, 449e450 United States: modern framework, 450e451 utility, 454 Patentability, substantive criteria for, 455e464 best mode, 463e464 enablement, 463e464 nonobviousness, 462e463 novelty, 460e462 patentable subject matter, 455e459 utility, 459e460 written description, 463e464 Patentable subject matter, 450 algorithms and software, 458e459 DNA, 456e458 living organisms, 456e458 mere associations, 456 principles, 455e456 “Patentee”, 455 Patenting by agency exclusive and coexclusive licensing, 496 results, 496e497 scope of licensing authority, 495 various agency missions, 495 inventions by national institutes of health, 497e503 National Institutes of Health licensing program, 499e503 Patient care records, 37 Patient communications, effective, 197e198 Patient engagement, fostering dissemination of, 624e625 Patient experience, 193e195 lay expert, 194e195 new world, 194 worst news, 193e194 Patient perspective, 189 assertive patient, 198e199 effective patient communications, 197e198 managing difficult news, 196e197 patient experience, 193e195 patients refuse to participate in clinical trials, 191e193 patientescientist partnership, 190e191 role of palliative care, 196 role of Patient Representative, 195e196 understanding caregiver, 195 Patient Protection and Affordable Care Act (PPACA), 492 Patient Reported Outcomes-Common Terminology Criteria for Adverse Events (PRO-CTCAE), 535 Patient Representative role, 195e196 Patient response bias, 330 Patient safety and clinical event reporting systems, 640e641 and clinical quality measurement, 641e642, 641t in clinical research, 635e637, 638t

Patient samples management access to, 567 good record keeping, 562 role of preanalytic variables in research using patient specimens, 560e561 specimen collection, 563e565 specimen culling, transfer of collections, and repository closings, 567 specimen handling, 565 specimen storage, 566e567 specimen tracking, 562e563 specimen transit, 565e566 successful research rests on foundation of careful planning, 560 training and accreditation, 561e562 Patient specimens, preanalytic variables in research using, 560e561, 561f Patient-centered approach, 285 CER, 284e285 Clinical Research infrastructure, 287e288 Patient-centered clinical research network (PCORNET), 309 Patient-Centered Outcomes Research Institute (PCORI), 115, 270e273, 271b, 277e278, 289e290, 304, 422 Methodology Committee, 275 Methodology Standards, 275e276, 276t National Research Priorities, 273t perspective, 284e285 role in National Clinical Research, 273f Patient-oriented organizations, 616e617 Patient-oriented research (POR), 607 midcareer investigator award in, 608 Patient-powered research networks (PPRNs), 288, 625 Patient-reported outcomes (PROs), 303e305, 307 disease-specific measures, 312e313 function, 310e312 examples of functional measures, 311 features to functional measure, 310e311 functionemeasurement and use, 310 selecting functional measure, 311e312, 311t utility of functional measures, 310 in functional outcome measures, 306e307 measurement and methodology, 307e310 NHANES, 309e310 psychometric properties, 308e309 Patient-Reported Outcomes Measurement Information System (PROMIS), 277, 307 Patient/intervention/condition/problem (PICO), 717, 724be743b Patientescientist partnership, 190e191 patient motivation for trial participation, 190e191 pAUC. See Partial area under the curve (pAUC)

INDEX

PCORI. See Patient-Centered Outcomes Research Institute (PCORI) PCORnet, 287e289 networks, 289 priority activities, 288 PCORnet. See National Patient-Centered Clinical Research Network (PCORnet) PCORNET. See Patient-centered clinical research network (PCORNET) PCT. See Patent Cooperation Treaty (PCT) PD. See Pharmacodynamics (PD) PD/PIs. See Project directors/principal investigators (PD/PIs) PDF. See Portable Document Format (PDF) PDPs. See Product Development Partnerships (PDPs) PDR. See Physician’s Desk Reference (PDR) PDUFA. See Prescription Drug User Fee Act (PDUFA) Pearson correlation, 392 Pediatric population, 656 Pediatric Research Equity Act, 656 Peer review, 41e42, 179e180 NIH, 594e600, 609 “About Grants” Page, 609 assignment of applications to review group and funding institute, 595 center for scientific review home page, 609 confidentiality and conflict of interest, 600 criteria for research project grant applications, 596e598 determines applications awarded, 600 dual-review system, 594 institute/center home pages, 609 by national advisory councils and boards, 599e600 overall impact/priority score and percentiles, 599 possible scientific review group actions, 598e599 research project grant applications from new/early-stage investigators, 598 review “cycles”, 594 review proceed, 596 reviewers selection, 595e596 reviewers thought about application, 599 procedural error during, 606 Penicillin, 11 “Per patient” budget, 571e577 budget comparison grid, 577t cost breakdown, 574t study budget template, 576t Per-protocol analysis, 334 Performance assessment problems, 57 measurement, 641 Peripheral edema, 420 Permuted block randomization. See Block randomization Persian al-Razi, 3

Persian Ibn Sina-Avicenna, 3 Personal integrity of physician, 40 Personalized medicine, 652 PET. See Positron emission tomography (PET) Pfizer study, 492 PGR. See Post-Grant Review (PGR) Phagocytosis, 10 Pharmaceutical and Medical Devices Agency (PMDA), 89 Pharmaceutical Inspection Cooperation Scheme (PIC/S), 87 Pharmaceutical Product Working Group (PPWG), 87e88 Pharmaceutical Research and Manufacturers of America (PhRMA), 89 Pharmaceutical testing, 454 Pharmacodynamics (PD), 646e647, 650 Pharmacogenetics, 652, 654 Pharmacogenomics, 652, 654 Pharmacokinetics (PK), 646e647 Pharmacologic treatment studies, 252 Pharmacologically active dose (PAD), 649 Pharmacometrics, 654 Pharmacy and supplies, 436 Pharmaprojects, 724be743b Phase 0 trials, 646e647 functional imaging tools related to, 652 Phase 1 trials, 77, 404 Phase 2 trials, 77 Phase 3 trials, 77, 80, 404, 413e414, 647 Phase 4 trials, 79 Phase I studies, 252, 652e653 Phase II studies, 252e253, 652e653 Phase III studies, 253, 652e653 Phase IV studies, 253 Phase IV trials, 413e414 Phenotype-Genotype Integrator (PheGenI), 745 Phenytoin, 650, 651t PHI. See Protected health information (PHI) Philanthrocapitalism, 616 Philanthropic donors, 620e621 Philanthropic sector, 611 areas of contribution, 617 organization of philanthropic sector and terminology, 613e615 alliances and umbrella organizations serving philanthropic sector, 614e615 foundations, 613e614 public charities, 614 private foundations, 615e616 public charities and patient-oriented organizations, 616e617 Philanthropy’s role in advancing biomedical research areas of contribution, 617e628 advocating for resources and policy changes, 626e628 building knowledge and expanding scientific disciplines, 619e620 developing human capital, 617e619

797 fostering dissemination of information, data sharing, 624e625 PDPs, 624 philanthropic sector, 617 stimulating innovation, 621e623 supporting institutions, 620e621 translating discoveries into cures, therapeutics, and preventions of disease, 623e624 continuum of biomedical research, 613f history of philanthropic sector, 615e617 organization of philanthropic sector and terminology, 613e615 philanthropic funds devoted to provision of health services, 611e612, 612f Phlebotomy, 197 PhRMA. See Pharmaceutical Research and Manufacturers of America (PhRMA) PHS. See U.S. Public Health Service (PHS) PHS 398 specific human subjects sections, 603 data safety monitoring plan, 603 inclusion of children, 603 inclusion of women and minorities, 603 protection of human subjects, 603 PHS 398 specific research plan component, 601e603, 602t research strategy, 602e603 specific aims, 602 Physical barriers, 102 Physical health, 307 Physician’s Desk Reference (PDR), 724be743b PhysicianeScientist workforce, 664e665 Physiological characteristics, 179 Physiology-based drug discovery, 646 PI. See Principal investigator (PI) PIC/S. See Pharmaceutical Inspection Cooperation Scheme (PIC/S) Pick the winner approach, 422, 423f PICO. See Patient/intervention/ condition/problem (PICO) Pilot Grants, 622 “Pivotal” phase, 414 PK. See Pharmacokinetics (PK) Placebo confusing placebo response and regression to mean, 264e265 control group, 259e260 control in research, 29 effect, 395 responses, 261e262 clinical research, 261 identifying placebo responders, 261e262 treatment in long-term studies, 264 unethical placebo groups, 264 Plant patent, 454 Plant Variety Protection Act, 448 Platform trials, 281

798 PLCO Cancer Screening Trial. See Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO Cancer Screening Trial) PLoS. See Public Library of Science (PLoS) PMC. See Postmarketing commitments (PMC); PubMed Central (PMC) PMDA. See Pharmaceutical and Medical Devices Agency (PMDA) PMR. See Postmarketing requirements (PMR) Pneumotox Online, 743 PO. See Program officer (PO) Point estimation, 343 POISINDEX, 724be743b Poisoning, 719 Polemian’s growth factor, 503 Policy Board, 129 Policy On Sharing of Model Organisms for Biomedical Research, 498 Political instability, 101e102 Population variance, 363 Population-based health research. See also Comparative effectiveness research; Clinical research ethical considerations, 299, 299t HMOs, 293 limitations and solutions, 297e298 data mining and statistical significance, 298 data quality, 297 generalizability and ecological fallacy, 298 lack of clinical detail, 297e298 missing data, 297 linking data sets, 298e299 original sources for data, 294 strengths, 296e297 surveys, 298 uses of secondary data in health research, 294e296 Population-based modeling methods, 653e654 POR. See Patient-oriented research (POR) Portable Document Format (PDF), 593 Positive predictive value (PPV), 397 Positron emission tomography (PET), 652 Post-Grant Review (PGR), 478 Post-marketing trials, 79 Postapproval, 83e84 Posterior distribution, 389 Postmarketing 15-day “alert reports”, 83e84 Postmarketing commitments (PMC), 83 Postmarketing requirements (PMR), 83 Pott’s pupil, 5 Power, 345, 359e360. See also Sample size calculations PPACA. See Patient Protection and Affordable Care Act (PPACA) PPH. See Patent Prosecution Highway (PPH) PPRNs. See Patient-powered research networks (PPRNs) PPV. See Positive predictive value (PPV)

INDEX

PPWG. See Pharmaceutical Product Working Group (PPWG) Practice standards for CRN, 677e678 standards of care, 678 standards of practice, 678 Pragmatic concerns, 222 Pragmatic Explanatory Continuum Indicator Summary (PRECIS), 254e255, 421 PRECIS-2 tool, 254e255 Pragmatic trials, 254e255, 420e422 “Pragmatic” research, 270 PRAISE trial. See Prospective Randomized Amlodipine Survival Evaluation trial (PRAISE trial) Pre-Biologics License Application (Pre-BLA), 78e79 Pre-IND meeting, 78e79 Pre-New Drug Application (pre-NDA), 78e79 pre-NDA/biologics license application submission, 81e82 Pre-Renaissance medicine, 3 Preanalytic variables in research using patient specimens, 560e561, 561f PRECIS. See Pragmatic Explanatory Continuum Indicator Summary (PRECIS) Precision medicine, 746e747 Predatory journals, 42 Predictive safety testing, 648e649 Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA), 724be743b Preferred Reporting Items for Systematic reviews and Meta-Analyses for Protocols 2015 (PRISMA-P 2015), 321 Pregnant women, 162, 166, 182e183, 656 Preliminary data, 27e28 Premarketing clinical research phases, 77 Prescription Drug User Fee Act (PDUFA), 74 PREVAIL II. See Partnership for Research in Ebola Virus in Liberia II (PREVAIL II) Prevalence, 232e233, 238, 397 rate, 234e235 surveys, 234e236 Prevention and Treatment of Missing Data in Clinical Trials, The, 404 Prevention trials, 250e251 Prexige. See Lumiracoxib PRIM&R. See Public Responsibility in Medicine and Research (PRIM&R) Primary data, 36, 715 Primary hypothesis, 427 Primary literature resources, 715 Primary prevention trials, 250e251 Principal investigator (PI), 52e56, 433, 516, 532, 592 department, 572 in staffing study assessing need for nursing support, 681

“concept of research intensity”, 682 planning study in clinical setting, 680e681, 680f staffing plan creation, 681e682 Principles and Practice of Clinical Research, 667 “Prior art”, 464e466, 519 “Priority date”, 476 Priority review, 79 PRISMA-P 2015. See Preferred Reporting Items for Systematic reviews and Meta-Analyses for Protocols 2015 (PRISMA-P 2015) Privacy, 55e56, 225e226 protection, 168e171 Privacy Act, 56, 169e171 Privacy Board, 169e170 Privacy Rule. See Health Insurance Portability and Accountability Act (HIPAA) Private “extramural” research, federal funding of, 487e493 core terms required in BayheDole funding agreementsex 202, 488e489 x 202edetermination of exceptional circumstances, 489e490 duty of US manufactureex 204, 493 funding agreements outside BayheDole Act, 493 history and philosophy, 487e488 key conceptsexx 200 and 201, 488 “March-In”ex 203, 490e493 organization of clauses, 488 x 202ereporting obligations, 489 Private foundations, 613e616 Prize4Life, 622e623 PRO-CTCAE. See Patient Reported Outcomes-Common Terminology Criteria for Adverse Events (PRO-CTCAE) Probability value (P-value), 345, 349, 353e354 Procedural problems, 57 Product development, incentives for, 453e454 Product Development Partnerships (PDPs), 624 Product-limit estimator. See KaplaneMeier estimator (K-M estimator) Professional responsibility, 437e438 Program Announcement (PA), 592e593 requests for applications and PA in NIH guide, 593, 593t Program officer (PO), 605e606 Project directors/principal investigators (PD/PIs), 592 Project management, 436e437 PROMIS. See Patient-Reported Outcomes Measurement Information System (PROMIS) Proof of concept studies, 252 trials, 77 Propensity scores, 283

799

INDEX

Property-owner’s exclusivity, 448 Prophetic conception, 467e468 Proportional hazards model, 379e380 Propoxyphene, 655, 656t Propulsid. See Cisapride ProQuest Dissertations & Theses Database, 724be743b PROs. See Patient-reported outcomes (PROs) Prosecution of patent application, 477e478 Prospective cohort study, 233, 240e242 Prospective longitudinal cohort studies, 233 Prospective Randomized Amlodipine Survival Evaluation trial (PRAISE trial), 432 Prospective studies, 232, 239e243 Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO Cancer Screening Trial), 398e399 Prostate Cancer Challenge Awards, The, 622 Protected health information (PHI), 56, 169 Protection of human subjects from research risk, 597 Proteins, 565 Protocol budget, 571 areas of concern, 585e586 hourly rate, 572e573 IRB fees, 572 overhead or indirect cost, 572 “per patient” budget, 573e577 start-up cost and invoiced items, 577e581 submitting budget to sponsor for approval, 582 walking away, 586 wrapping up, 586e587 Protocol compliance by research staff, 132 Protocol components, 205, 205t, 221e229 alternatives to participation, 225 compensation, 227 confidentiality, 226 conflict of interest, 226e227 consent process and documents, 227e229 data and safety monitoring, 224 determination overall benefit-to-risk ratio, 224 protocol benefit category, 224 protocol risk category, 223e224 eligibility criteria, 225 vulnerable populations, 225 hypotheses and objectives, 222 legal agreements, 226 management of data and samples, 226 precis, 221 privacy, 225e226 QA monitoring, 224 qualifications of investigators, 226 risks, discomforts, and inconveniences, 223 statistical analysis, 226

study design and methods, 222e223, 222t procedures, 223 recruitment, 222 screening, 223 unanticipated problem, adverse event, and deviation/violation reporting, 224e225 Protocol map/research grid, sample, 696, 696fe697f Protocol Registration and Results System (PRS), 116 Protocol writing, 219e220 clinical trials, 220 phases, 220t sample approaches to protocol organization, 221t elements of protocol, 220e229 protocol components, 221e229 regulatory oversight, 219e220 Protocol(s), 204 importance of, 117 modifications, 212 order sets within EHR, 694e696 sample protocol map/research grid, 696 violations, 152 Protracted nausea, 195 Provisional (PRV), 475e476, 519 Proximity to source material, 715 PRS. See Protocol Registration and Results System (PRS) Pseudorandomization methods, 333e334 Psychometric properties, 304, 308e309 methodology in measurement development, 308e309 patient-reported outcomes in large data sets, 309 requirements for all measures, 308t Psychotherapy, 209 PsycINFO, 724be743b Public and private secondary data, 294e295 charities, 614, 616e617 foundations, 613e614 and science, 710 transparency, 168 use, 465 Public Health Service Act, 76e77, 114 Public Health Service’s Bureau of Radiological Health, 74 Public Library of Science (PLoS), 625, 722 Public Responsibility in Medicine and Research (PRIM&R), 38 Publication, 476 bias, 320e321, 427e428 Cycle, 715 practices, 42 PubMed, 721e722, 724be743b advanced search builder, 726f Health, 724be743b home page, 725f PubMed-based literature search, 714

PubMed/MEDLINE record, 715e716 records, 745 Search Details, 719, 720f search results of search, 727f PubMed Central (PMC), 721e722, 724be743b PubPeer, 34e35 Puerperal fever, 7e8 Puerperal Fever as a Private Pestilence, 7 Pure Food and Drug Act (1906), 20

Q QA. See Quality assurance (QA) QoL. See Quality of life (QoL) QOLI. See Quality of Life Index (QOLI) Qualification Process for Drug Development Tools, 650 Qualifications of investigators, 226 Qualitative data, 550 Qualitative interactions common, 419 uncommon, 418e419 Quality, 88 Quality assurance (QA), 224, 538e539 Quality control of data, 538 Quality improvement techniques in clinical research, 635e637, 638t tools, 642 Quality of life (QoL), 307e308 Quality of Life Index (QOLI), 312 Quantitative data, 550 Quantitative pharmacology, 655 Quasiexperimental studies, 255 “Quid Pro Quo”, 453 of patent philosophy, 467 of patent system, 485

R R &R application. See Research and Related application (R &R application) RA. See Rheumatoid arthritis (RA) Radiation, 197 RAID. See Rapid Access to Intervention Development (RAID) Random assignment, 28 Random effect model, 319 Random errors, 297 Randomization, 11e13, 28, 250, 329e330, 412e413 factor, 191 importance of, 330 issues in implementation, 334e335 monitoring, 335 sound allocation, 334e335 mechanisms of, 335 methods, 332e334 objection to, 192 outcome, 131e132 process, 210 services, 436 special considerations, 335e337 adaptive randomization methods, 336e337

800 Randomization (Continued ) documentation, 337 threats to integrity of randomization, 337 Randomized clinical trials (RCTs), 27, 331 ethical considerations in, 27e29, 28t Randomized concentration controlled trial (RCCT), 653 Randomized controlled trial (RCT), 136, 245e246, 250, 254, 331, 437e438, 647 Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY), 425 Randomized phase I studies, 252 Randomized Phase II clinical trials, 253 Randomized studies, 245e246 design randomized study, 246e247 Randomized trial, 331, 438e439 Rapid Access to Intervention Development (RAID), 490 Rapporteur, 93e94 Raptiva. See Efalizumab RCCT. See Randomized concentration controlled trial (RCCT) RCE. See Request for Continued Examination (RCE) RCT. See Randomized controlled trial (RCT) RCTs. See Randomized clinical trials (RCTs) RE-LY. See Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) “Reach-through” rights, 512 Reactive drug metabolites, 650 “Reasonable volunteer” standard, 26 REC. See Research ethics committee (REC) Recall bias, 238e239 Receiver operator characteristic curve (ROC curve), 397e398, 397f “Receiving Office”, 479 “Recklessness”, 35 Recochrane methodologies, 724be735b, 736f Record retention, 543, 543t Recordkeeping, 168e171 research dataeHIPAA, Privacy Act, and Certificates of Confidentiality, 169e171 Recruitment, 116, 132e133, 132f Recurrent events analysis, 402 Red blood cell transfusion, 548 REDCap system. See Research Electronic Data Capture system (REDCap system) Reductil. See Sibutramine Reduction, refinement, and replacement (three R’s), 39, 39t Reduction to Practice, 466e467, 519 Reed, Walter, 10e11 “Reexamination”, 478 Reference managers. See Bibliographic managers Reference works, 724be743b RefSeqGene, 745

INDEX

“Refuse to file” action, 82 Regional health initiatives (RHIs), 90e91 Regional patent offices, 479e480 Regional Patenting Systems, 452 Registration and results reporting, 168 Regression to mediocrity, 394e395 Regression to the mean (RTM), 394e395 examples, 395e396 change after exceeding threshold, 395 placebo effect, 395 screening period vs. trial event rates, 395e396 medical research, 396 ways to addressing, 396 Regulations FDA and Common Rule, 163 Federal laws and, 162 govern practice, 680 Regulatory affairs, 436e437 Regulatory agency role, 654e656 Regulatory authorities, 434e435 Regulatory Chair, 93e94 Regulatory compliance, 132 Regulatory foundations, current requirements for research involving human subjects, 49e50 Regulatory guidelines, 23 Regulatory harmonization, future work in, 98 Regulatory issues related to data reporting, 542e543 Regulatory Member, 92 “Reissue” application, 478 Rejection region, 346e347 Relapsing/remitting multiple sclerosis (RRMS), 252e253, 342e343 relative false-positive fraction (rFPF), 399 Relative risks (RR), 244, 244t, 318, 379e380 Relative treatment effect, 416 relative true-positive fraction (rTPF), 399 Renaissance, 2e3 Renewal applications, 597 Repeated significance tests. See Sequential designs RePORT. See Research Portfolio Online Reporting Tools (RePORT) RePORTER. See Research Portfolio Online Reporting Tools Expenditures and Results (RePORTER) Reporters investigative, 710e711 talking to, 706e707 Reporting key scientific principles and best practices for, 119e120 issues in reporting outcome measures, 119e120 issues related to analysis population, 120 outcome measures four levels of specification in, 120f issues in, 119e120 x 202eReporting obligations, 489 Repository closings, 567 Reproducibility, 43e45, 44t, 549

Reproductive organs, 179 Request for Continued Examination (RCE), 477e478 Requests for Applications (RFAs), 592, 626be628b Research, 51, 706 as benefit, 21 on bioethical questions, 23 clinical, 703e704 controversial biomedical, 705e706 coordination, 678 dataeHIPAA, 169e171 director/manager, 532e533 ethics guidelines, 48 intensity concept, 682, 682t misconduct, 34e35 at NIH, 40e41 nurse coordinators, 672e673 participants, 63e65, 70e71 project grant applications from new/early-stage investigators, 598 review criteria for, 596e598 protocol, 704 order sets, 696 research-use exemption, 470e471 spectrum of research continuum, 251e255 staff domain, 69e70 subjects protection, 21 “turning-on-a-dime”, 704 Research and Related application (R &R application), 592 Research Electronic Data Capture system (REDCap system), 552, 553f Research ethics board. See Institutional Review Board (IRB) Research ethics committee (REC), 25e26 Research plan (RP), 510, 516e517 Research Portfolio Online Reporting Tools (RePORT), 489 Research Portfolio Online Reporting Tools Expenditures and Results (RePORTER), 724be743b Research team, 531e533 clinical research associate, 533 clinical trials nurse, 533 DBA, 533 principal investigator, 532 research director/manager, 532e533 statistician, 533 subinvestigators, 532 Research Tools policy, 497e498 Research!America, 615 Researcher domain, 69e70 Resident training in clinical research, 662e664 Resource selection, 748e749 Resource Sharing Plans, 598 Resource-limited countries, 99 Resource-specific learning opportunities and support, 723 Resources and policy changes, advocating for, 626e628 Respect for enrolled subjects, 27

INDEX

Respect for persons, 49 Response-adaptive allocation, 336 Responsibilities of NIH staff, 591e592, 592t Responsibilities of research supervisors and trainees, 36 Responsible stewardship of data, 553e555 Resubmission applications, 597 Results database, 111e112, 114, 118, 121 Results reporting, 112 ethical and scientific rationale for, 113t to publication, 119 rationale for, 112e113 steps in, 112f Retention recruitment and, 211, 214 of subjects, 369 Retraction Watch, 34e35 Retrospective cohort studies, 240, 282 Retrospective studies, 232, 236, 239e243 Review by national advisory councils and boards, 599e600 Review clock, 82 Review criteria, 118e119, 119t for research project grant applications, 596e598 additional review considerations, 597e598 additional review criteria, 597 core review criteria, 596e597 Review group and funding institute, assignment of applications to, 595 Review proceed, 596 Reviewers selection, 595e596 Revision applications, 597 Rezulin. See Troglitazone RFAs. See Requests for Applications (RFAs) rFPF. See relative false-positive fraction (rFPF) Rheumatoid arthritis (RA), 307, 653 RHIs. See Regional health initiatives (RHIs) Right censoring, 374 Rights conveyed by patents, 455 Rigor, 549 Risk, 52 assessment methodologies, 642 continually monitoring clinical research environment for, 639e640 risk/benefit assessment, 25 Risk ratio (RR), 243 ROC curve. See Receiver operator characteristic curve (ROC curve) Roche, Ellen, 714, 743 Roche Biomedical Laboratories, 456 Rofecoxib, 655, 656t “Role models”, 663e664 Ro¨ntgen, Wilhelm, 11, 12f Roosevelt, Theodore, 74 Root Cause Analysis, 635, 636te637t Rosiglitazone, 655, 656t Rowley, Janet, 13e14 RP. See Research plan (RP) RR. See Relative risks (RR); Risk ratio (RR)

RRMS. See Relapsing/remitting multiple sclerosis (RRMS) RTM. See Regression to the mean (RTM) rTPF. See relative true-positive fraction (rTPF) Rubenstein, Dr. Arthur, 34 “Rule of three”, 392 Rule of thumb, 710

S SADC. See South African Development Community (SADC) SAE. See Serious adverse events (SAE); Suspected adverse event (SAE) Safety, 88. See also Data and safety monitoring (DSM) biomarkers, 650, 651t FDA and drug, 655e656 patient safety and clinical event reporting systems, 640e641 and clinical quality measurement, 641e642, 641t leveraging patient safety in clinical research, 635e637 Salvarsan, 11 Sample PREanalytic Code (SPREC), 560e561, 561f Sample protocol map/research grid, 696, 696fe697f Sample size, 402e403, 428e429 calculations, 359e360, 430 advanced study designs, 368e369 alternative statistics and methods, 368 clinical and scientific research, 360 normal distributions, 361 notational conventions, 360e361 for precision in confidence interval construction, 361e362 retention of subjects, 369 sample size calculation, 360 statistical computing, 369 t-distributions, 361 for hypothesis tests one sample of data, 362e364 paired data, 364e366 two independent samples, 366e368 recalculation, 134e135 based on nuisance parameters, 135 based on nuisance parameters and observed treatment effect, 135 special considerations, 390e392 early phase studies, 391 unequal sample sizes, 391e392 Sampling, 235 distribution, 343 of topic areas addressed by, 89t Sanarelli, Giuseppe, 10e11 SARS. See Severe acute respiratory syndrome (SARS) Satterthwaite’s formula, 350 SBP. See Systolic blood pressure (SBP) SC. See Steering Committee (SC) Scalability, 536

801 SCDA. See Simons Center for Data Analysis (SCDA) Scholarometer, 748 Science making news in science and medicine, 704 public not knowing science, 710 publishing, 704 science-based regulatory agency, 654 science-driven hypotheses, 183e184 ScienceDirect, 724be743b Scientific collaborators, 39, 40t Scientific considerations, 179e180 Scientific data, 36 Scientific disciplines, building knowledge and expanding, 619e620 biomarkers, 620 biomedical imaging and bioengineering, 619 neuroinflammation, 619e620 stem cell research, 620 Scientific information, 26 Scientific integrity, 34e35 Scientific journals, 722 Scientific misconduct, 35, 35t, 151 Scientific papers, 722 Scientific record keeping, 37, 37t Scientific research, ally in, 198e199 Scientific review groups (SRGs), 594 actions, 598e599 Scientific review officer (SRO), 594 Scientist, 703, 706 SciFinder (STN), 724be743b Scoping reviews, 306 Scopus, 721, 724be743b, 748 Screening, 223 period, 395e396 trials test strategies, 250e251 “Search Engine Optimization” companies, 723 Search strategy, 748e749 Seasonal migration, 102 Secondary data, 293e294 uses of secondary data in health research, 294e296 Secondary hypothesis, 427 Secondary literature resources, 715 Secondary research, 285 Secondary use of EHR for clinical research, 698 “Secret prior art”, 465 Secrets and government, 507e508 Section 801, 114 Secular trends, monitoring, 294e295 SEER cancer database. See Surveillance, Epidemiology, and End Results cancer database (SEER cancer database) Seizure Detection and Prediction Challenge, 623 Select agent research, 597e598 Selection bias, 330 Self-Paced Interactive Program, 749 Semmelweis, Ignaz Philipp, 7e8, 8f

802 Senior/key personnel profiles component and biosketches, 604 Sensitivity, 397, 405 Sepsis, meta-analysis of clinical trials of antiinflammatory agents in, 321e323 Sequential designs, 136 Sequential Multiple Assignment Randomized Trials (SMART), 257, 281 Sequential trial designs, 257 Serious adverse event (SAE), 130e131 Serum albumin, 361 Serum amylase, 148 Severe acute respiratory syndrome (SARS), 637e638 Sex, 179 Sexually transmitted infections (STIs), 181 SF36. See Short Form 36 (SF36) SF424 (R&R) project summary/abstract, 601 Sham control group, 260 Shannon, Dr. James, 49 Shared Health Research Information Networks (SHRINE), 555 Sharing Model Organisms, 598 SHEP. See Systolic Hypertension in Elderly Program (SHEP) Short Form 36 (SF36), 312 SHRINE. See Shared Health Research Information Networks (SHRINE) SIB. See Swiss Institute of Bioinformatics (SIB) Sibutramine, 655, 656t Sickness Impact Profile (SIP), 308e309 Significance, 596 level, 344, 361, 364 P-value to, 385t statistical, 298, 346e347 Signing agreements contract execution in general, 504e506 scope of actual authority of government laboratories, 506 Signing official (SO), 593 Simons Center for Data Analysis (SCDA), 625 Simple linear regression, 354e355 Simple randomization, 332 “Simultaneous conception and reduction to practice”, 467e468 Sinclair, Upton, 74 Single blinding, 424 Single clinical trial, 117 Single photon emission computed tomography (SPECT), 652 Single population mean, 362e363 proportion calculations for binary data, 363 two-stage designs for, 363e364 Single time point studies, 234e236 Single-masked studies, 210 Single-nucleotide polymorphisms (SNPs), 394 SIP. See Sickness Impact Profile (SIP)

INDEX

SIR. See Statutory Invention Registration (SIR) Site management organization, 435 Sloane, Sir Hans, 5e6 Small research grant (R03) applications, 609 Small studies, 265e266 Small-area variation, 296 Smallpox, 5e6 SMART. See Sequential Multiple Assignment Randomized Trials (SMART) Smith, James, 7 SNOMED-CT. See Systematized Nomenclature of Medicine-Clinical Terms (SNOMED-CT) Snow, John, 7 Snow’s studies, 7 SNPs. See Single-nucleotide polymorphisms (SNPs) SO. See Signing official (SO) SOC. See Standard of care (SOC) Social health, 307 Social media, 707 word about email, web, and, 707e708 Society of Clinical Research Associates (SoCRA), 674 SoCRA. See Society of Clinical Research Associates (SoCRA) “Soft data”, 304 Software and algorithms, 455e456 Software transfer agreements, 514 Sollmann, Torald, 11e13 SOMA. See Carisoprodol Somatic cell nuclear transplantation, 167 SOPs. See Standard operating procedures (SOPs) SORN. See System of records notice (SORN) Sound allocation, 334e335 “Source code”, 484 Source documents, 538, 538t Sources of data, 538 South African Development Community (SADC), 87e88 Southern California Health Care Organization, 294e295 SPC. See Supplementary Protection Certificate (SPC) Specialized clinical resources, 724be743b Specialty practice, 675 Specification, 473 Specificity, 397 Specimen barcode, 560e561 collection, 563e565 and processing protocols, 560 culling, 567 handling, 565 inventory, 564 storage, 566e567 tracking, 562e563 transit, 565e566 SPECT. See Single photon emission computed tomography (SPECT)

Spectrum bias, 399 Spectrum of research continuum, 251e255 comparative effectiveness research, 254 dissemination and implementation studies, 253e254 explanatory vs. pragmatic trials, 254e255 phase I studies, 252 phase II studies, 252e253 phase III studies, 253 phase IV studies, 253 quasiexperimental studies, 255 Sponsors, 79e80 SPREC. See Sample PREanalytic Code (SPREC) Spring Harbor Laboratory’s bioRxiv, 722 SRGs. See Scientific review groups (SRGs) SRO. See Scientific review officer (SRO) SRs. See Systematic Reviews (SRs) SSA. See Sub-Saharan Africa (SSA) Staffing plan creation, 681e682 requirements and responsibilities, 532 Staggered entry, 374, 380 Stand Up to Cancer (SU2C), 622 Standard care control groups, 260 Standard care groups. See Usual care control groups Standard criteria, 237e238 Standard deviation, 342 Standard error, 342 Standard of care (SOC), 574, 585 Standard operating procedures (SOPs), 536, 562 Standard(s) of care, 678, 678t for evidence synthesis, 286b of practice, 678 systematic reviews, 305e306 Staphylococcus pyogenes (S. pyogenes), 8e9 Start-up cost, 571, 577e581, 580f, 581t, 582f Starting dose in FIH study, 649 Startpage, 744 State of Maine, 113e114 State Street, 458e459 State Street Bank & Trust Co. v. Signature Financial Group, 458e459 Statistic, 342e343 Statistical computing, 369 genetics, 394 hypotheses formulation, 345e346 hypothesis testing, 344e345 inference, 342e343 penalty, 135 power, 360 significance determination, 346e347 theory, 348 Statistical analysis determining measure of treatment effect, 318e319 fixed vs. random effect, 319 heterogeneity, 319e320 publication bias, 320e321 reporting and interpreting results, 321

803

INDEX

software, 321 subgroup analysis and metaregression, 321 Statistician, 533 Statistics in Medicine, 129 “Statutory bar”, 461, 482, 519 Statutory Invention Registration (SIR), 476 Statutory subject matter, 519 Steering Committee (SC), 88 Stem cell research, 620 Step 1 Application, 70 Step 2 Application, 70 Stepped-care treatment. See Adaptive treatment Stevenson-Wydler Act, 493e495 history and philosophy of, 494 key concepts and major clauses, 494 subsequent supporting acts, 494e495 Stewardship of data, 553e555 Stimulating innovation, 621e623 STIs. See Sexually transmitted infections (STIs) STN. See SciFinder (STN) Stochastic curtailment tests, 136, 136f Strasbourg Patent Convention, 451e452 Stratification approaches, 283 Stratification factor, 131e132 Stratified log-rank test, 379 Stratified randomization, 333 Streamlining patient care, 585 Strengthening the Reporting of Observational Studies in Epidemiology guidelines (STROBE guidelines), 233 Streptococcus pyogenes (S. pyogenes), 8e9 STROBE guidelines. See Strengthening the Reporting of Observational Studies in Epidemiology guidelines (STROBE guidelines) Structural problems, 57 Structured trial reporting, 120 Student training in clinical research, 662e664 Student’s t distribution, 347, 349 Studies nested within cohort, 239e243 Study calendar, 535, 535t Study designs for CER studies, 278e285, 279f adjusting for and avoiding confounding in observational CER studies, 283e284 assessing treatment heterogeneity, 284e285 cohort designs, 282e283 experimental study designs for CER, 279e281 observational study designs for CER, 281e282 Study management, 672e673, 673f Study parameters, 535, 535t Study participant characteristics, 102 Study protocol. See also Protocol writing development and conduct of studies, 203e204

development and importance of, 204e213 authorship, 212e213 data analysis, 211e212 data collection, 210e211 dosing/intervention intensity, 209 equipoise, 213 identifying and defining outcomes of interest, 208e209 inclusionary and exclusionary criteria, 206e208 masking/blinding, 210 protocol modifications, 212 recruitment and retention, 211 statement of design, 205 study sample, 205e206 treatment/intervention development, 209e210 MOP, 213e216 study design, 204 Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT), 425 SU2C. See Stand Up to Cancer (SU2C) Sub-Saharan Africa (SSA), 100 Subgroup analysis, 321 Subinvestigators, 532 Subject Invention, 488 Subjectivity of clinical data, 548e549 Submitting budget to sponsor for approval, 582, 582fe585f Subpar health-care systems, 101 Superiority trials, 389e390 Supplemental Examination, 478 Supplementary Protection Certificate (SPC), 476e477 SUPPORT. See Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) Surrogate consent, 163e164 Surrogate decision-making, 165 Surrogate endpoints, 425, 426t, 650e652 biomarkers and, 425 composite and, 438 Surveillance, Epidemiology, and End Results cancer database (SEER cancer database), 294 Survival, 373e374 Survival analysis, 369, 373e374, 400e403. See also Meta-analysis changes over time in coefficients and covariates, 400e401 time-dependent covariates, 401 time-varying coefficients, 400e401 data from first hypothetical example, 375t dependent or informative censoring, 401e402 features of survival data, 374e375 mistakes, 380 recurrent events analysis, 402 sample size, 402e403 survival function, 375e380

Survival function, 375e380 comparing, 377e379 chronic active hepatitis study, 378, 378t at given time point, 377 stratified log-rank test, 379 using whole curve, 377e378 K-M estimator, 375e377 proportional hazards model, 379e380 Survival time, 374 Sushruta Samhita, 2 Suspected adverse event (SAE), 540e541 Swiss Institute of Bioinformatics (SIB), 745 Symptoms related to function, 312 Synagis, 503 System failures, 641 System of records notice (SORN), 170 System-level interventions, 274, 277 Systematic errors, 297 Systematic Reviews (SRs), 306, 724be743b standard systematic reviews, 305e306 Systematized Nomenclature of MedicineClinical Terms (SNOMED-CT), 551 Systolic blood pressure (SBP), 185 Systolic Blood Pressure Intervention Trial, 705 Systolic Hypertension in Elderly Program (SHEP), 208

T t Tests, 347 t-distributions, 361 T2DM. See Type 2 diabetes mellitus (T2DM) Tails of distribution, 344 TAP. See Therapy Acceleration Program (TAP) Target-based drug discovery, 646 Targeted therapeutic approaches, 652 Taxus, 503 Team(s), 531e532 goals, 531e532 responsibilities, 532 science, 39e40 Technical items, 474 Technology transfer, 448 agreements agreements to protect confidentiality, 506e509 agreements to transfer materials, 509e514 CRADA, 514e518 hypothetical scenario, 503e504 signing agreements, 504e506 critical laws, 487e503 federal “intramural” research, 493e495 federal funding of private “extramural” research, 487e493 federally supported research, 487e503 patenting and licensing by federal agencies, 495e503 Technology Transfer Commercialization Act of 2000 (TTCA), 494 Template, standardized, 220e221 Tension, 706

804 “Terminal disclaimer”, 475e477 Tertiary hypothesis, 427 Tertiary literature resources, 715 Test statistic, 344, 349e351 Tetrabenazine, 650, 651t Textbooks, 724be743b Thalidomide, 177 Therapeutic truisms, 432e433 Therapy Acceleration Program (TAP), 624 Third phase. See “Pivotal” phase 1000 Genomes Browser, 745 “371” application, 475, 480 Thrombolytic therapy, 419 Thyrogen, 503 Time and attention control groups, 259 Time to event, 373e374 Time-dependent confounding, 283 covariates, 401 hazard ratios, 400e401 Time-varying coefficients, 400e401 Time-varying covariates. See Timedependent covariates Titrating therapy, 425 TL1 program, 665e666 TNF-a. See Tumor necrosis factor-a (TNF-a) “Tools of Human Investigation”, 664 Toxicity, 646, 649e650, 652 preclinical testing of new drugs reliable predictor of, 155e156 Toxicology Data Network (TOXNET), 724be743b Toxicology Literature Online (TOXLINE), 724be743b TPF. See True-positive fraction (TPF) Trade secrets, 507 and federal employees, key statutes relating to, 486e487 principles, 485e486 Trademarks, 485 Traditional retrospective case-control studies, 238 Transfer materials, agreements to basic material transfer agreement, 509e514 CTA, 512e513 key specialized material transfer agreements, 513e514 uniform biological material transfer agreement, 512 Transfer of collections, 567 Transfusions of blood, 3e4 Translating discoveries into cures, therapeutics, and preventions of disease, 623e624, 623t “Translation”, 190 Translational and clinical research, 590 Translational discipline, clinical pharmacology as, 645e646 Translocation, 13e14 Transnational approaches, 99 Transparency, 549 Transportation, 192 Trastuzumab, 650, 651t

INDEX

Treatment by covariate interaction analysis, 432 fidelity, 216 and intervention development, 209e210 as usual groups. See Usual care control groups Trial conduct, 130 monitoring, 131e133 data quality, 133 participant flow, 131 participants’ baseline characteristics, 131 randomization outcome, 131e132 regulatory compliance, 132 trial performance, 132e133 Trial design, 384e392 adaptive designs, 387e389 alpha spending, 384e387 equivalence, 389e390 interim monitoring, 384e387 noninferiority, 389e390 special considerations for sample size, 390e392 superiority, 389e390 Trial event rates, 395e396 Trial performance, 132e133 attendance at follow-up visits, 133 data completeness, 133 participants’ treatment adherence, 133 protocol compliance by research staff, 132 recruitment, 132e133, 132f Trial reporting system (TRS), 121, 122f, 122t Triggers, 133, 134f Triple blinding, 424 TRIPS. See Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) Troglitazone, 655, 656t Trovafloxacin, 655, 656t Trovan. See Trovafloxacin TRS. See Trial reporting system (TRS) True-positive fraction (TPF), 397 Trusting bivariate associations based on observational study data, 244 TTCA. See Technology Transfer Commercialization Act of 2000 (TTCA) Tubercle bacillus, 9 Tuberculosis, 7 Tumor necrosis factor-a (TNF-a), 322 Tumor suppressor gene, 142 Two-arm trial, 329e330, 344 Two-dimensional scannable barcode (2D scannable barcode), 563, 563f Two-sample hypothesis tests, 349e351. See also One-sample hypothesis tests tests for comparing means of two normal populations, 349e350 tests for comparing two population proportions, 350e351 Two-sample t-test, 350, 356 Two-sample z-tests, 350 Two-sided test, 344, 346, 351e352

Two-stage designs for single population proportion, 363e364 Type 2 diabetes mellitus (T2DM), 420 Type I error, 344e345, 360, 427e428 Type II error, 344e345, 360, 428e429 Typhoid fever, 7

U U.S. Public Health Service (PHS), 35, 180, 603 UBMTA. See Uniform Biological Material Transfer Agreement (UBMTA) UCLA. See University of California, Los Angeles (UCLA) UCSC Genome Browser, 746 UCSF. See University of California at San Francisco (UCSF) UGDP. See University Group Diabetes Program (UGDP) UKZN. See University of KwaZulu-Natal (UKZN) Umbrella organizations serving philanthropic sector, 614e615 Unanticipated problem (UP), 224e225, 540e542 Unanticipated risk in clinical research, 142. See also Legal issues in clinical research Cassandra revealed, 147 Congress, 154 drug, 144e145 epilogue, 155e157 extended studies, 147 FDA, 151e152 Fialuridine toxicity, 148e149 investigations begin, 150e151 IOM, 152e153 law, 154e155 media, 153e154 NIH, 152 reassessing preclinical studies, 149 research oversight, 149e150 scientific misconduct, 151 target, 145 trials, 145e147 Unashamedly utilitarian phase, 20e21 “Unborn children”, 13 Uncertainty principle, 422 Uncontrolled studies, 265e266 Unequal sample sizes, 391e392 Unethical placebo groups, 264 Uniform Biological Material Transfer Agreement (UBMTA), 512 Uniform biological material transfer agreement, 512 Unitary patent system, 480 United Network for Organ Sharing organ transplant database (UNOS organ transplant database), 294 United States (US) applications, 475e478 basic types of, 475e476 options “after issuance”, 478 prosecution of patent application, 477e478

805

INDEX

combining US and patent cooperation treaty filings, 480e481 Constitution, 449e450 “Federal Register”, 170 as Infringer, 471 modern framework, 450e451 Federal Circuit, 450e451 1952 Patent Act, 450 US Patent Reform of 2011, 451 patent infringement, 469e473 civil liability, 469 contributory and induced infringement, 469e470 “declaratory judgment” actions, 472 importation, 472 International Trade Commission, 472 major defenses, 470 practical issues of litigation, 472e473 remedies, 471e472 specific exemptions and immunities, 470e471 1789e1951, 450 timing considerations export control, 476 patent life, 476e477 publication, 476 US-based clinical research, 184f United States Preventive Services Task Force (USPSTF), 277 Unity, 474e475 University Group Diabetes Program (UGDP), 330 University infringement shield, 470 University of California, Los Angeles (UCLA), 620 University of California at San Francisco (UCSF), 505, 620 University of KwaZulu-Natal (UKZN), 621 University Patents Inc. (UPI), 456 UNOS organ transplant database. See United Network for Organ Sharing organ transplant database (UNOS organ transplant database) Unpaired data, 349e350 UP. See Unanticipated problem (UP) UPI. See University Patents Inc. (UPI) UpTo-Date, 715, 724be743b US Code of Federal Regulations (USCFR), 21, 49, 75, 220, 532, 542e543, 698 US Council on Foundations, 613 US Court of Customs and Patent Appeals (CCPA), 451 US Department of Agriculture (USDA), 74, 496 US Department of Health and Human Services (DHHS), 21, 49, 114e115, 161e162, 182e183, 271b, 542, 562, 590, 724be743b US FDA Medical Device Databases list, 724be743b US Federal Food, Drug, and Cosmetic Act (FD&C Act), 13, 74, 76e77

US Food and Drug Administration (FDA), 22e23, 33e34, 48, 73e74, 81, 87, 113, 131, 143, 151e152, 161e162, 164, 178, 192, 220, 273e274, 304, 337, 413e414, 419, 434e435, 437e438, 453e454, 532, 538e541, 551, 578, 614, 646, 648e649, 654, 677e678, 698, 724be743b and clinical pharmacology, 655, 655t compliance, 84 definitions and terms, 76t drug and biological product life cycle, 76e84 and drug safety, 655e656, 656t FDA-approved veterinary vaccines, 502e503 FDA-regulated clinical research, 162 FDA-reportable products, 692 guidance for EHR in clinical research, 699 mailing lists, 542t mission and terminology, 74e76, 75t principal regulations for drug and biological products, 75t and special populations, 656 statutory authorities, 75t US Government Accountability Office (GAO), 57, 178 US International Trade Commission (ITC), 472 US National Cancer Institute (NCI), 113, 457, 534, 538e539, 560, 590e591 US Office of Management and Budget (OMB), 114e115, 178 US Patent and Trademark Office (USPTO), 451, 455, 460e461, 475, 724be743b US Patent Classification (USPC), 724be743b US patent law(s) core concepts “conception” vs. “reduction to practice”, 466e467 “inventorship” and “joint inventorship”, 468 key terms, 464e469 patent, 454e455 patent infringement, 469e473 “prophetic conception” vs. “simultaneous conception and reduction to practice”, 467e468 “prior art”, 464e466 substantive criteria for patentability, 455e464 transfers of ownership, 468e469 efforts to altering, 481e483 abusive tactics, 482e483 compulsory licensing and breaking patents, 483 international harmonization, 481e482 patents on genes and “mere associations”, 482 US Patent Reform of 2011, 451 USCFR. See US Code of Federal Regulations (USCFR) USDA. See US Department of Agriculture (USDA)

USPC. See US Patent Classification (USPC) USPSTF. See United States Preventive Services Task Force (USPSTF) USPTO. See US Patent and Trademark Office (USPTO) Usual care control groups, 260 “Usual care”, 277 Utility, 459e460, 519 patent, 454

V VA. See Veterans Affairs (VA) Vaccination, 5 “Validation” stage, 479 Validity, 23e24, 415 Value, 23e24 of accreditation, 70e72 “value-added” features, 310e311 van Leeuwenhoek, Antony, 4, 5f Variance, failing to decrease, 263 Variolation, 5e6 Vasodilator therapies, 432 Velcade, 503 Venture philanthropy, 616 Verification bias, 399 Vertebrate animals, 597, 603e604 Vesalius, Andreas, 3 Veteran’s Affairs Office of Academic Affiliations, 668 Veterans Affairs (VA), 115, 495 VHOs. See Voluntary health organizations (VHOs) Vidarabine, 144 Vioxx. See Rofecoxib Virtual private network (VPN), 537 Vitamin K, 641 Vitravene, 503 Voluntary and informed individual consent, 163e164 Voluntary health organizations (VHOs), 611, 614 Volunteers, challenges to enrolling, 181e182, 181t, 184t Vomiting, 195 von Behring, Emil, 10, 10f Voxels, 393e394 VPN. See Virtual private network (VPN) Vulnerable research subjects, 55, 55t

W Wait-list control group, 259 Warfarin, 650, 651t Warranties and indemnification, 511 Warren, Dr. John Collins, 8 Washington, George, 5e6 Washout, 255e256 Waterhouse, Benjamin, 7 “Watershed”, 331 Web, 707e708 Web of Science, 721, 724be743b, 748e749 Welch’s t-test, 352 “Welcome Letter” tool, 36 Well-written protocol, 204e205

806 WHI. See Women Health Initiative (WHI); Women’s Health Initiative (WHI) WHI CT. See Women’s Health Initiative Clinical Trial (WHI CT) White-coat hypertension, 208e209 WHO. See World Health Organization (WHO) Wiley, Dr. Harvey, 74 WIPO. See World Intellectual Property Organization (WIPO) WISE. See Women’s Ischemia Syndrome Evaluation (WISE) WMA. See World Medical Assembly (WMA) Women and minority groups, 177e178 Women Health Initiative (WHI), 213 Women of childbearing potential, 182e183 Women’s Health Initiative (WHI), 185, 245e246, 250, 283e284 Women’s Health Initiative Clinical Trial (WHI CT), 405

INDEX

Women’s Ischemia Syndrome Evaluation (WISE), 185 Work up bias, 399 World Development Indicators 2016, 100 World Health Organization (WHO), 22, 48, 87, 100, 115e116, 129, 196, 303e304, 722 World Intellectual Property Organization (WIPO), 452 World Medical Assembly (WMA), 13, 22 World Trade Organization (WTO), 452e453 WorldCat, 724be743b “Worst case” analysis, 405 Wrapping up, 586e587 Wren, Sir Christopher, 3e4 Writing protocol. See Protocol writing Written description, 463, 519 WTO. See World Trade Organization (WTO)

X Xenazine. See Tetrabenazine Xigris. See Drotrecoginalfa XML. See Extensible Markup Language (XML)

Y Yandex, 744 Yellow fever studies, 10e11

Z Z statistic, 378, 380 z Tests, 347 Zalcitabine (DDC), 149 Ziagen. See Abacavir Zidovudine (AZT), 144, 149