Annual Review of Health Care Management : Revisiting the Evolution of Health Systems Organization 9781783507160, 9781783507153

Citation preview

ANNUAL REVIEW OF HEALTH CARE MANAGEMENT: REVISITING THE EVOLUTION OF HEALTH SYSTEMS ORGANIZATION

ADVANCES IN HEALTH CARE MANAGEMENT Series Editors: Leonard Friedman, Grant T. Savage and Jim Goes Recent Volumes: Volume 4:

Bioterrorism, Preparedness, Attack and Response Edited by John D. Blair, Myron D. Fottler, and Albert C. Zapanta

Volume 5:

International Healthcare Management Edited by Grant T. Savage, Jon A. Chilingerian, and Michael Powell

Volume 6:

Strategic Thinking and Entrepreneurial Action in the Health Care Industry Edited by John D. Blair, Myron D. Fottler, Eric W. Ford, and G. Tyge Payne

Volume 7:

Patient Safety and Health Care Management Grant T. Savage and Eric. W. Ford

Volume 8:

Biennial Review of Health Care Management: Meso Perspectives Edited by Grant T. Savage and Myron D. Fottler

Volume 9:

Strategic Human Resource Management in Health Care Edited by Grant T. Savage, Myron D. Fottler, and Naresh Khatri

Volume 10:

Organizational Development in Health Care: Conversations on Research and Strategies Edited by Jason A. Wolf, Heather Hanson, Mark J. Moir, Len Friedman, and Grant T. Savage

Volume 11:

Biennial Review of Health Care Management Edited by John D. Blair and Myron D. Fottler, with assistance from Grant T. Savage

Volume 12:

Health Information Technology in the International Context Edited by Nir Menachemi and Sanjay Singh, with assistance from Valerie Yeager

Volume 13:

Annual Review of Health Care Management: Strategy and Policy Perspectives on Reforming Health Systems Edited by Leonard H. Friedman, Grant T. Savage, and Jim Goes

Volume 14:

Leading in Health Care Organizations: Improving Safety, Satisfaction and Financial Performance Edited by Hannes Leroy, Grant T. Savage, and Tony Simons

Edited by

ADVANCES IN HEALTH CARE MANAGEMENT VOLUME 15

ANNUAL REVIEW OF HEALTH CARE MANAGEMENT: REVISITING THE EVOLUTION OF HEALTH SYSTEMS ORGANIZATION EDITED BY

JIM GOES Cybernos, LLC and Walden University, Minneapolis, MN, USA

GRANT T. SAVAGE University of Alabama at Birmingham, Birmingham, AL, USA

LEONARD FRIEDMAN The George Washington University, Washington, DC, USA

United Kingdom North America India Malaysia China

Japan

Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2013 Copyright r 2013 Emerald Group Publishing Limited Reprints and permission service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters’ suitability and application and disclaims any warranties, express or implied, to their use. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-78350-715-3 ISSN: 1474-8231 (Series)

ISOQAR certified Management System, awarded to Emerald for adherence to Environmental standard ISO 14001:2004. Certificate Number 1985 ISO 14001

CONTENTS LIST OF CONTRIBUTORS

vii

LIST OF REVIEWERS

ix

EDITORIAL ADVISORY BOARD

xi

PREFACE

xiii

SECTION I: REVISITING COMPLEXITY IN HEALTH CARE ORGANIZATION HEALTH CARE ORGANIZATIONS AS COMPLEX SYSTEMS: NEW PERSPECTIVES ON DESIGN AND MANAGEMENT Reuben R. McDaniel, Jr., Dean J. Driebe and Holly Jordan Lanham 3 COMMENTARY ON “HEALTH CARE ORGANIZATIONS AS COMPLEX SYSTEMS: NEW PERSPECTIVES ON DESIGN AND MANAGEMENT” BY REUBEN R. McDANIEL, DEAN J. DRIEBE, AND HOLLY JORDAN LANHAM Ruth A. Anderson 27

SECTION II: REVISITING HEALTH SYSTEMS INTEGRATION HORIZONTAL AND VERTICAL INTEGRATION OF PHYSICIANS: A TALE OF TWO TAILS Lawton Robert Burns, Jeff C. Goldsmith and Aditi Sen

v

39

vi

CONTENTS

COMMENTARY ON “HORIZONTAL AND VERTICAL INTEGRATION OF PHYSICIANS: A TALE OF TWO TAILS” BY LAWTON ROBERT BURNS, JEFF C. GOLDSMITH, AND ADITI SEN Stephen M. Shortell

119

THE EVOLUTION OF INTEGRATED HEALTH CARE STRATEGIES Jenna M. Evans, G. Ross Baker, Whitney Berta and Jan Barnsley

125

SECTION III: INTEGRATION AND PHYSICIAN RELATIONSHIPS HOSPITAL PHYSICIAN RELATIONSHIPS: IMPLICATIONS FROM THE PROFESSIONAL SERVICE FIRMS LITERATURE Mona Al-Amin, Robert Weech-Maldonado and Rohit Pradhan 165 ENGAGING EMPLOYED PHYSICIANS: RECONCEPTUALIZING THE ROLE OF COLLECTIVE IDENTIFICATION Gregory W. Stevens

185

SECTION IV: FUTURE DIRECTIONS CONCLUDING REMARKS Jim Goes

213

LIST OF CONTRIBUTORS Mona Al-Amin

Department of Healthcare Administration, Sawyer School of Business, Suffolk University, Boston, MA, USA

Ruth A. Anderson

School of Nursing, Duke University Medical Center, Duke University, Durham, NC, USA

G. Ross Baker

Institute of Health Policy, Management & Evaluation, University of Toronto, Toronto, ON, Canada

Jan Barnsley

Institute of Health Policy, Management & Evaluation, University of Toronto, Toronto, ON, Canada

Whitney Berta

Institute of Health Policy, Management & Evaluation, University of Toronto, Toronto, ON, Canada

Lawton Robert Burns

Department of Health Care Management, The Wharton School, Philadelphia, PA, USA

Dean J. Driebe

Independent Researcher, Brussels, Belgium

Jenna M. Evans

Institute of Health Policy, Management & Evaluation, University of Toronto, Toronto, ON, Canada

Jim Goes

Cybernos, LLC and Walden University, Minneapolis, MN, USA

Jeff C. Goldsmith

Department of Public Health Sciences, University of Virginia, Virginia, VA, USA

Holly Jordan Lanham

Department of Medicine, Division of Hospital Medicine, The University of Texas, San Antonio, TX, USA vii

viii

LIST OF CONTRIBUTORS

Reuben R. McDaniel, Jr.

Department of Information, Risk and Operations Management, McCombs School of Business, University of Texas at Austin, Austin, TX, USA

Rohit Pradhan

Department of Health Policy and Management, Fay W. Boozman College of Public Health, University of Arkansas, Little Rock, AR, USA

Aditi Sen

Department of Health Care Management, The Wharton School, Philadelphia, PA, USA

Stephen M. Shortell

School of Public Health, Division of Health Policy and Management, University of California, Berkeley, CA, USA

Gregory W. Stevens

Psychology Department, Auburn University, Auburn, AL, USA

Robert WeechMaldonado

Department of Health Services Administration, University of Alabama at Birmingham, Birmingham, AL, USA

LIST OF REVIEWERS Ann F. Chou University of Oklahoma, Oklahoma City, USA

Ann Scheck McAlearney The Ohio State University, Columbus, USA

Fred van Eenennaam George Washington University, Washington, DC, USA; Erasmus University Rotterdam, Rotterdam, The Netherlands

Nir Menachemi University of Alabama at Birmingham, Birmingham, USA Kathleen Montgomery University of California, Riverside, USA

Leonard Friedman George Washington University, Washington, DC, USA

Peter E. Rivard Suffolk University, Boston, USA

Timothy Huerta The Ohio State University, Columbus, USA

Grant T. Savage University of Alabama at Birmingham, Birmingham, USA

Katharina Janus Ulm University, Ulm, Germany; Columbia University, City of New York, USA

Valerie Yeager Tulane University, New Orleans, USA

ix

EDITORIAL ADVISORY BOARD Jon Chilingerian Brandeis University, Massachusetts, USA

Katharina Janus Ulm University, Germany

Ann Chou University of Oklahoma Health, Sciences Center, Oklahoma, USA

Ann Scheck McAlearney Ohio State University, Ohio, USA Nir Menachemi University of Alabama at Birmingham, Alabama, USA

Alison Cuellar George Mason University, Virginia, USA

Kathleen Montgomery University of California, Riverside, California, USA

Jami DelliFraine University of Texas School of Public, Health, Texas, USA Fred van Eenennaam ESAE Erasmus University, The Netherlands; George Washington University, Washington, DC, USA

Ingrid Nembhard Yale School of Public Health, Connecticut, USA Lori T. Peterson Cleveland State University, Ohio, USA

Mattia J. Gilmartin New York University College of Nursing, New York, USA

Peter E. Rivard Suffolk University-Sawyer Business School, Massachusetts, USA

Jonathon R. B. Halbesleben University of Alabama, Alabama, USA

Valerie Yeager Tulane University, New Orleans, USA

Tim Hoff University at Albany, State University of New York, New York, USA

xi

PREFACE After a century of striving, after a year of debate, after a historic vote, health care reform is no longer an unmet promise. It is the law of the land. (President Barack Obama, March 23, 2010)

It may be a law, but is it a reality? Depending on who measures it, health care in the United States consumes around 18 percent of the gross domestic product of the nation. Few who pay attention to this statistic would regard that expenditure as sustainable. Not surprisingly, the growing cost of health care in the United States has stimulated many efforts at reform. However, success at these efforts has been at best mixed. Passage of the Affordable Care Act of 2010 and subsequent success of the legislation in the courts have raised the stakes and promise a new era of health reform. Yet the act of reforming health organizations has long been underway, and the lessons we have learned along the way can help inform and structure the organization and execution of reform efforts yet to come. Since the late 1980s, health systems and health organizations have undergone dramatic changes in organizational structure, systems, and processes. These changes have included development of large, integrated health organizations, designed to provide scale and scope advantages, improve the quality of care and health outcomes, and provide greater bargaining power relative to payers and large employers. One result of this effort has been consolidation of health systems and physician-health organizations, and dramatic changes in relationships and practices between hospitals and physicians in an effort to improve financial and clinical outcomes. The scope and nature of this transition has created greater complexity and size in consolidated health systems, leading to greater interest in understanding hospitals and health systems as complex systems. Health care is a “target rich” environment for students and researchers of organizational and system change. The changes underway at the organization and system levels are invariably complex, often confusing, and at times even chaotic. Over the last 20 years, researchers of health systems and organizations have devoted considerable effort to understanding the xiii

xiv

PREFACE

processes and outcomes of these organizational and system changes. Well established organizational theories have been applied by researchers and health care executives to help understand, design, and navigate health care reform, and often these approaches have been found wanting. The evolution and revolution of health systems and organizations is still very much a work in progress. In Volume 15 of Advances in Health Care Management, we revisit two influential themes that have emerged from this effort: • Viewing health care organizations as complex systems, leading to new perspectives on design and management of these organizations. • Creation and organization of physician-health organizations, and alignment of these organizations with hospital or system structure and governance. To anchor the critical assessment of research around these two themes, we invited authors of the two most frequently cited articles in the 15 year history of Advances to revisit and extend their original work, and judge the relevance of their perspectives the to the systemic reform underway at the national level. In the first section of this volume, Reuben R. McDaniel, Jr. and colleagues Dean J. Driebe and Holly Lanham revisit the influence of complexity science on the design and management of health care organizations. They argue that “people in medicine want to be sure, but complexity raises doubts,” and that avoiding serious consideration of complexity in the design of reform efforts makes it less likely that these efforts will result in meaningful reform. In an industry increasingly focused on outcomes, McDaniel, Jr. and colleagues argue that quick fixes and simplistic thinking may cause more harm than good given the complexity of health systems. As noted by Keller (1979, p. 718), whom they reference, “what is being evaded is the need for a cognitive structure radically different from the prior existing structure.” This is a provocative idea, and worth considering given the mixed results of most reform efforts thus far. In an invited commentary on this chapter, Ruth A. Anderson suggests mixing complexity theory with a focus on outcomes as means of achieving substantive reform. In the second section of the volume, we revisit health systems integration, a theme that has attracted considerable interest among both researchers and system designers. For many years, health care leaders have struggled to engage physicians organizationally and operationally in structural changes and practice reform reform designed to reduce costs and improve health outcomes. In the second anchor chapter in this volume, Lawton Robert Burns and his colleagues Jeff C. Goldsmith and Aditi Sen

Preface

xv

review two decades of evidence on horizontal and vertical integration of physicians, and provide a critical assessment of this widely used strategy. In the first comprehensive review of scale and scope economies of physician practice in nearly 20 years, Burns and his colleagues find little evidence that physician integration actually generates greater efficiency in care delivery. This is a major finding that calls into question core thinking about health reform efforts, and implies the need for much broader thinking about the best ways to organize for efficient delivery and effective outcomes. In the second invited commentary in this volume, Stephen M. Shortell, an eminent researcher on organizational and system change in health care, discusses conceptual and methodological implications of the research by Burns and colleagues for health reform, and moving from a “culture of care” to a “culture of health.” Along with the two invited chapters and commentaries, we also solicited competitive chapter submissions for this volume around the themes of complex systems and health systems integration. In the first of these refereed chapters, Jenna M. Evans, G. Ross Baker, Whitney Berta, and Jan Barnsley explore the evolution of integrated health care strategies with an extensive analysis and integration of published literature on the subject over the last 25 years. They find six major shifts in integration strategies over this period, moving through horizontal and vertical integration, community-based health, quality of care, patient centered outcomes, and care for specific populations. They offer this as evidence of change and complex adaptation over time, and note a migration away from mechanistic models of health care delivery toward an emergent, self-organizing orientation. This chapter not only provides a wealth of source material for students and researchers on the topic, but also points to the emergence of new trends and conceptual models of system integration. In the third section of this volume, authors of two refereed chapters focus more directly on hospital physician relationships, a subject of considerable effort and research in recent years. Mona Al-Amin, Robert Weech-Moldonado, and Rohit Prahan borrow from literature on professional service firms to argue for a “new governance archetype” with potential to align incentives between physicians and hospitals. This non-equity partnership model offers an intriguing conceptual extension to more established approaches for aligning and integrating physicians in achieving efficiency and outcomes that meet hospital goals. Gregory W. Stevens also explores physician integration by considering professional relationships and collective identity of physicians and the role this plays in achieving physician engagement and coordination of care. Engagement, Stevens

xvi

PREFACE

argues, is best achieved by “understanding a physician’s perception of his or her relationships to the profession and the organization.” He poses a series of propositions around the dynamics of physician identity that beg empirical further research. Taken as a whole, this volume of Advances offers a detailed, critical look at the evolution of health systems organization through two lenses: HCOs as means of designing and managing complexity, and HCOs as mechanisms to align incentives and organization to achieve greater efficiency and superior outcomes, both for the system and for patients. In doing so, the chapters in this volume bring the latest evidence and thinking about these issues to the forefront of health care reform. Jim Goes Editor

REFERENCE Keller, E. F. (1979). Cognitive repression in contemporary physics. American Journal of Physics, 47(8), 718 721.

SECTION I REVISITING COMPLEXITY IN HEALTH CARE ORGANIZATION

HEALTH CARE ORGANIZATIONS AS COMPLEX SYSTEMS: NEW PERSPECTIVES ON DESIGN AND MANAGEMENT Reuben R. McDaniel, Jr., Dean J. Driebe and Holly Jordan Lanham ABSTRACT Purpose We discuss the impact of complexity science on the design and management of health care organizations over the past decade. We provide an overview of complexity science issues and their impact on thinking about health care systems, particularly with the rising importance of information systems. We also present a complexity science perspective on current issues in today’s health care organizations and suggest ways that this perspective might help in approaching these issues. Approach We review selected research, focusing on work in which we participated, to identify specific examples of applications of complexity science. We then take a look at information systems in health care organizations from a complexity viewpoint.

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 3 26 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015007

3

4

REUBEN R. McDANIEL ET AL.

Findings Complexity science is a fundamentally different way of understanding nature and has influenced the thinking of scholars and practitioners as they have attempted to understand health care organizations. Many scholars study health care organizations as complex adaptive systems and through this perspective develop new management strategies. Most important, perhaps, is the understanding that attention to relationships and interdependencies is critical for developing effective management strategies. Research and practice implications Increased understanding of complexity science can enhance the ability of researchers and practitioners to develop new ways of understanding and improving health care organizations. Originality/value This analysis opens new vistas for scholars and practitioners attempting to understand health care organizations as complex adaptive systems. The analysis holds value for those already familiar with this approach as well as those who may not be as familiar. Keywords: Complexity science; complex adaptive system; HCO design; HCO management; HCO interdependencies; HCO information systems

A REPRISE As we have looked at the original chapter from our present vantage point, the following seem to be the most important contributions (McDaniel & Driebe, 2001). Thinking about health care organizations as complex adaptive systems (CAS) is a powerful tool. This is because health care organizations are made up of diverse agents who have nonlinear interdependencies and who learn as they encounter the world. This leads health care organizations to have emergent properties, to self-organize and to coevolve with their environments. This perspective suggests that managerial strategies such as sensemaking, learning, and improvisation are useful in attempting to design and manage health care organizations. Recognition of the importance of these strategies suggests that organizational improvements are likely to result from attention to relationships and interactions among members of an organization; in particular, attention to mindfulness and heedfulness. Perspectives that pay attention to the importance of

Health Care Organizations as Complex Systems

5

relationships contrast with less effective traditional strategies of command and control. They also contrast with traditional strategies of attempting to improve health care organizations by focusing attention on improving the skills and knowledge bases of individual agents in the system. We need the system to get better, not just individuals in the system. Our original effort put some fairly new ideas together to get people to look at health care organizations in a different way and to stimulate new lines of research. We believed that continued refinement of traditional models for health care organizations had limited potential to lead to breakthroughs in the delivery of health care. On the other hand, looking at health care organizations as complex adaptive systems provides a promising frame for identifying significant research efforts to improve health care organizations. In the decade following our original contribution, considerable research has been done that shows how ideas from complexity science both work in practice and generate new research streams. The outline of our chapter is as follows: first we will cover some foundational issues to put our ideas in a broader context. Then we will focus on their impact on thinking about health care systems in particular. Over the course of our work, we have found that attention to relationships and interdependencies is critical for developing management strategies and we share some of these insights. The rising importance of information systems for the management of health care systems has led us to study information systems in health care, and we will share some of our understandings in this area. Last, we will frame our perception of the state of health care organizations and discuss the considerable contributions that a complexity science approach can make.

THE SCIENCE OF COMPLEXITY SCIENCE A review of the complexity literature since our original chapter is far beyond the scope of this chapter. Rather, we identify just a very few ideas that we feel might have particular relevance for efforts to understand and improve health care organizations. These choices certainly reflect our biases and also the streams of research we have been involved in over the past decade. We offer these as ideas grounded in the study of complexity science and ideas that we believe have promise for future study of health care organizations. Nature presents a wide array of both regular and irregular phenomena. The impressive progress of science in the Enlightenment was largely due to

6

REUBEN R. McDANIEL ET AL.

isolating, both experimentally and theoretically, the regular aspects of nature. These regularities lend themselves to mathematization and are often seen to arise from rules that are relatively simple to state. This approach is, perhaps, most forcefully represented by Newton’s three laws of nature, especially the second law, which explains and allows for prediction of behavior of mechanical phenomena on both terrestrial and celestial levels. The enormous success of the Newtonian program inspired people working on aspects of human behavior and social systems to think in Newtonian terms and this viewpoint stubbornly persists in many circles up to the present. The rise of complexity science in the latter half of the twentieth century provided new perspectives to reassess both natural and human systems. Concepts such as chaos, self-organization, dissipative structures, and complex adaptive systems began to be used in order to understand both irregular aspects nature displays as well as providing frameworks to explore regularities that arise in complex systems. A hallmark of these ideas is the existence of nonlinearities in systems addressed by complexity science. Generally speaking, this means that the superposition principle is not valid for such systems; the behavior of the system is more than the sum of the behaviors of individual components. Even though many fundamental aspects of nonlinear dynamics were grasped late in the nineteenth century by the French polymath Henri Poincare´, it is undoubtedly due to the widespread accessibility of computers that the field blossomed beginning in the 1960s. The study of chaotic systems revealed that randomness and determinism are not disjoint concepts. Random behavior can arise in quite simple deterministic systems. This tells us that when we observe a phenomenon that appears irregular it may not be due to a complicated system or to our limited knowledge or limited ability to observe. Another feature of chaotic systems is sensitivity to initial conditions, which means that the evolution of two slightly different initial states of the system widely diverge from each other in time. Research into complex systems engenders approaches that employ techniques and analyses that lean more toward qualitative methods. For example, nonlinear dynamical systems described by differential equations that are nonintegrable, basically meaning unsolvable, can nevertheless be investigated by considering the geometric properties of their solutions in phase space. Even though the solutions are not available, some of their key properties can be determined. This is the approach of the so-called qualitative theory of differential equations that was pioneered by Poincare´, among others. There is also available for dynamical systems an approach that

Health Care Organizations as Complex Systems

7

considers the probabilistic properties of ensembles of trajectories instead of individual trajectories (Lasota & Mackey, 1985). This approach leads to results in chaotic systems that are explicitly irreducible to the level of trajectory analysis and allows for an understanding of broken time symmetry (Driebe, 1999). For inanimate systems, complexity entails approaches that go beyond strictly reductionist techniques. When human beings are part of the system, this becomes even clearer. As noted by Schelling (1978), “With people, in contrast to light beams and water, we usually believe we are dealing with conscious decisions or adaptations in the pursuit of goals, immediate or remote, within the limits of their information and their comprehension of how to navigate through their environment toward whatever their objectives are” (p. 18). An approach to understanding of such situations has been the primary objective of the study of complex adaptive systems. These systems are made up of many interacting elements (agents) that coevolve with their local environments and they typically display the features of selforganization and emergent properties. Many economic systems, such as the stock market, a sector of an economy, a nation’s economy, or the global economy can be considered as a complex adaptive system. This way of thinking provides insights that are not present in a classical approach. Selforganization refers to the spontaneous emergence of order, structure, or stability in a system of many elements interacting only locally. Even though the interactions are local, long-range structures can arise in self-organizing systems without externally imposed controls or command processes. Such structures have sometimes been identified as “dissipative structures” because they often feed off of an energy and entropy flow from an environment in open systems that are far from equilibrium (Nicolis & Prigogine, 1977). For example, cyclonic structures in our atmosphere such as tornadoes and hurricanes. One of the most difficult topics to understand and yet one of the most fundamental in the study of complex systems is the notion of emergence. Maguire, McKelvey, Mirabeau, and O¨ztas (2006) called for more theorybuilding and model-building research in this area. A common belief in organizational management is that if things do not happen as expected as the result of an intervention, then there must have been an error in implementation. Complexity science suggests that at least some of the time the observer may be looking at an emergent phenomenon rather than an error in implementation. Telling the difference between these is not easy. “Emergence refers to the arising of novel and coherent structures, patterns, and properties during the process of self-organization in complex systems”

8

REUBEN R. McDANIEL ET AL.

(Goldstein, 1999, p. 49). We cannot predict the nature of the new phenomena from knowledge of the interacting elements alone. A holistic approach is needed for understanding emergence. Complexity science teaches us that despite our best efforts, history matters. Many scholars have been concerned about the history or path dependence of complex adaptive systems. One of the best known is Peter Allen whose work on the evolutionary complexity of economic systems has extended over many years (Allen, 1976, 1993, 2001). Allen has helped us to understand that organizations are limited in their future behavior by their past behavior. Institutional theorists have also helped us understand that organizational behavior is constrained by the history of the organizations as well as the history of the industry (Scott, 2001). Orlikowski (1996) developed a situated change perspective to help in understanding temporal aspects of change. “A situated change perspective may offer a particularly useful strategy for analyzing change in organizations turning increasingly away from patterns of stability, bureaucracy, and control to those of flexibility, self-organizing, and learning” (Orlikowski, 1996, p. 63). Network theory has developed rapidly as a major branch of the study of complexity, offering a way to examine the structures and dynamics of complex systems (Watts, 2003). In particular, network theory provides a strategy for understanding relationships in systems in a more comprehensive way. Early work in the field was concerned with a particular set of relationships and how ideas from network theory (centrality of a node for example) might help us to better understand a system in question. Now we are becoming more concerned about general characteristics of classes of networks and how these might lead to a new theory. This is particularly important in the study of social networks as new complex systems forms such as Facebook, Twitter, and Instagram have emerged from unexpected advances in information and communications technologies. Table 1 lists contrasting notions of recent complexity science approaches with classical approaches to systems thinking. Note that nonlinearity one element that is often used to define complex systems is not included in Table 1. A case could be made that nonlinearity is a necessary but insufficient condition for complexity. However, nonlinearity does not distinguish a complex system from a classical system and our effort in Table 1 is to identify characteristics that clearly distinguish one paradigm from the other. Classical systems may have nonlinear elements, and linear theories, such as quantum mechanics, may be very useful in understanding complex systems. Therefore, we do not use a linear-nonlinear distinction to contrast these two paradigms.

9

Health Care Organizations as Complex Systems

Table 1. Aspects of the Classical Cognitive Reference Frame for Systems Thinking Contrasted with a Contemporary Complexity Science Approach. Classical Paradigm • Equilibrium • Deterministic description based on trajectories • Stability • Reversibility • Euclidean geometry • Local behavior with short-range correlations • Average provides adequate description • Omniscient managing agent • Reductionist • Inevitable approach to equilibrium characterized by increasing disorder

Complexity Paradigm • Nonequilibrium • Probabilistic description based on ensembles • Order through fluctuations • Irreversibility • Fractal geometry • Coherent behavior with longrange correlations • Fluctuations drive the average • Self-organization • Holistic • Continuous diversification and evolution toward complexity

COMPLEXITY SCIENCE AND HEALTH CARE ISSUES Understanding how design and managerial properties emerge from interactions among agents in health care organizations has been a fruitful area for research. For example, we believe that quality performance of a health care organization may well be an emergent property of the organization and, therefore, quality improvement efforts may not succeed (or they may succeed) because of the nonlinear interdependencies among the agents rather than because of the external pressures for improvement (Lanham et al., 2009). We also agree with Bettis and Prahalad (1995) that the dominant logic of an organization is likely an emergent property, and because emergent properties are the result of relationships, attention to relationships in an organization may be the key to understanding its dominant logic. Cohen et al. (2004) developed a practice change model for quality improvement in primary care practice using concepts from complexity science to guide their thinking. Cohen sees the importance of viewing change as an emergent phenomenon if the manager is to have any ability to manage it. Although we are in the midst of a period of rapid and radical change in health care, we may not fully understand how the history and institutional structure of our health care organizations may severely limit the scope and direction of change. Sometimes managers think that change can be dictated independent of the path dependency of a system, but this is

10

REUBEN R. McDANIEL ET AL.

seldom true. Anderson, Corazzini, and McDaniel (2004) used complexity science thinking to develop an empirical study of nursing home turnover thus looking directly at an important dynamic phenomenon in the health care space. She sees turnover as emerging from the patterns of interactions among workers in the nursing home rather than as the result of economic and labor force issues alone. Leykum et al. (2007, 2010) provided key insights on the importance of the consideration of complex adaptive systems’ characteristics in designing interventions for improving care. In a systematic examination of improvement interventions, they found that interventions designed in ways that were consistent with characteristics of complex adaptive systems were more likely to be effective, and this was true for both diabetes care (2007) and congestive heart failure care (2010). The findings for the diabetes study were that the number of complex adaptive system characteristics present was positively associated with the effectiveness of an intervention and that attention to co-evolution and interconnections between participants were important. When the study was repeated examining interventions designed to improve care of congestive heart failure patients, the number of complex adaptive systems characteristics present in the intervention was also positively associated with effectiveness. However, the specific characteristics that were important were learning, self-organization, and co-evolution, a somewhat different set of characteristics than in the diabetes study. These studies show that health care improvement interventions are more likely to be effective when they are consistent with the complex adaptive systems nature of clinical systems, but they must also take into account attributes of the disease. Complexity science offers valuable insights about specific intervention design efforts. It is not enough to know the etiology of a specific disease, although this is important, one must also understand the complex adaptive nature of clinical settings. Social networks are not independent from each other nor are they independent of technological networks (Miguel et al., 2012), and these interconnections may have serious implications for the success of structural organization efforts such as accountable care organizations and health care exchanges. As we envision powerful electronic health records systems and, perhaps, expect them to create high-functioning networks of providers, we need to gain a much better understanding of the human-technology interface. Lanham et al. demonstrated that physicians’ perceptions of the role of information in providing patient care (2013) and their relationships with practice members (2011) affect their willingness and ability to participate in the virtual networks created via electronic health record use. One might ask

Health Care Organizations as Complex Systems

11

questions such as, “how will relationships among providers change as patients and providers are connected in virtual networks of care?” Shaw et al. (2013) use randomized controlled trials to study “facilitated team meetings” and “learning collaboratives” as interventions to improve screening for colorectal cancer and discovered that these networks have mixed results in improving care. More research is needed to improve our understanding of the role of networks in the management of health care organizations. We have continued to collaborate with research groups utilizing complexity science concepts to enhance our understanding of care delivery in a wide variety of health care settings. A research team of faculty at Duke University, led by Ruth Anderson, PhD, RN, has completed several studies of nursing homes.1 Included among her team’s work are studies of quality relationships among caregivers as a precursor to quality care for residents, turnover in nursing home staffs, and patterns of management organization for effective resident care. Her studies are grounded in the notion of nursing homes as complex adaptive systems and the implication of this for managerial practices in homes. Another research group led by Benjamin Crabtree, PhD at the Robert Woods Johnson Medical School, has been studying primary care organizations using a complexity science framework.2 Crabtree et al. have studied the delivery of preventive services (Crabtree et al., 2009) relationships (Lanham et al., 2009) and conversations (Jordan et al., 2009) in primary care organizations and new forms of primary care delivery such as patient-centered medical homes. A health services research group at The University of Texas Health Science Center in San Antonio led by Jackie Pugh, MD, has done significant work in the use of complexity science to better understand both outpatient clinics and hospitals. One interesting insight from this overall set of research projects is that a wide variety of delivery settings may be usefully studied as complex adaptive systems in efforts to enhance management practices.

RELATIONSHIPS AND INTERDEPENDENCIES Our original chapter argued the importance of relationships among individuals in health care organizations in carrying out both routine and nonroutine aspects of care delivery in ways that lead to high-quality patient outcomes. More specifically, we introduced the notion of developing mindfulness as a strategy for managing uncertainty in health care delivery.

12

REUBEN R. McDANIEL ET AL.

In subsequent years, we extended this model to include a total of seven attributes of relationships: trust, diversity, mindfulness, heedful interrelation, respectful interaction, rich and lean communication, and social and task relatedness (Lanham et al., 2009). Analysis of over 200 primary care practices in the United States (US) found that practices exhibiting all of these seven attributes of work relationships were more likely to consistently provide preventive care, such as colorectal cancer screening, depression screening, and smoking cessation counseling, to their patients than practices lacking these attributes. Research currently underway using a different national sample of US primary care practices is working to better understand how these seven attributes of relationships are structured in practice (Lanham & Palmer, 2008). For instance, the concepts of diversity and mindfulness are very different from each other, and the ways in which they express themselves in organizations are distinct. Our research in unpacking this model of work relationships provides new approaches to improving relationships in health care organizations. The key theoretical contribution of the Lanham et al. model of seven relationship attributes is that quality of care is an emergent property of health care delivery systems that arises from the local interactions and relationships among the many diverse and learning individuals in these systems (Lanham et al., 2009). Thus, efforts to improve health care quality in a given system should not only focus on improving individual components of the system but should also consider the relationships among individuals as an important source of improvement. Other important research examines the role of teaming in health care organizations (Edmondson, 2012). Teaming is important in uncertain and dynamic work environments where stable plans to carry out known tasks are becoming less relevant, and in health care, dangerous. Edmondson explains teaming as the fluid coordination and collaboration that allows team members to connect with each other, become familiar with the task at hand and the work environment, and develop knowledge of what each individual brings to the team. A main tenet of the idea of teaming is that the task is often unknown and changing, thus traditional static plans can be not only useless but also distracting. In teaming, learning and execution occur in tandem, and teaming is considered a mechanism for organizational learning. Teaming theory considers the uncertainty and unpredictable dynamics of organizations, ideas that are consistent with a complex adaptive systems view of organizations. Relationships among individuals in health care delivery systems are important, not simply because they engender a positive work environment, but because they enable important organizational behaviors, such as

Health Care Organizations as Complex Systems

13

sensemaking, learning, and improvisation (McDaniel, 2007). One of the most important concepts that affect our work on health care organizations as complex systems is the concept of sensemaking. Weick (1995) has utilized this concept to assist understanding in many environments, including health care (Weick, 2005; Weick & Sutcliffe, 2003). In our original chapter we made what we believed was a strong case for the critical role of sensemaking as a managerial strategy for complex adaptive systems. Since then, Anderson et al. (2005) have examined the role of sensemaking in providing quality care for nursing home residents. Ghaferi, Birkmeyer, and Dimick (2009) have examined the rates of unexpected complications in hospitals with high and low mortality. They found that hospitals with low mortality were better at recognizing when a patient was declining and so acted to address issues more quickly than hospitals with high mortality. This “failure to rescue” can be conceptualized as a failure to effectively make sense of changes in patients’ status. Finally, improving sensemaking capacity in health care delivery settings can help the people working in these settings to effectively make sense of change interventions (Lanham et al., 2013). Because health care organizations are complex adaptive systems, no two interventions will have the same effects, even if they are designed with strict replication in mind. The unique relationship infrastructures and the selforganization processes occurring at local levels of change interventions contribute to the meaning that people assign to the change effort itself and to their ability to implement change. Relationships also affect learning in health care delivery settings. Learning is continually occurring in complex adaptive systems. Since our original chapter, Edmondson’s work on psychological safety (2003) and voice (2012) has contributed new conceptualizations of learning that can be applied to health care organizations as complex adaptive systems. Additionally, Leykum et al. (2010) have contributed a new scale to assess the concept of reciprocal learning in primary care practice settings. While progress is being made on the underlying mechanisms that facilitate learning, the foundation for our understandings of learning in complex adaptive systems remains grounded in the idea of learning from samples of one, or fewer (March, Sproull, & Tamuz, 1991). March et al. (1991) recognized that many important events in organizations do not occur often and so they suggested strategies for organizational learning in these circumstances, primarily experiencing events more richly and developing ways to experience more fully near events (close airline accidents for example). These ideas seem to provide ways for complex adaptive systems to learn even though the dynamics are unpredictable. Anderson et al. (2005) developed

14

REUBEN R. McDANIEL ET AL.

recommendations for doing case research from a complexity perspective that acknowledges the variable nature of health care organizations as complex adaptive systems. Improvisation may turn out to be one of the next areas of important work in health care delivery settings as complex adaptive systems. McKenna, Leykum, and McDaniel (2013) developed a complex adaptive systems based model of improvising in health care delivery using data from inpatient and outpatient care teams. McDaniel (2007) contributed additional work on the role of improvisation in health care organizations. A major challenge in furthering the ideas of improvisation in health care will be to overcome the somewhat pervasive negative perceptions that people tend to have about the role of improvising in medicine. The complex, nonlinear interdependencies in a health care organization, and the dynamic ways in which people’s health and illness unfold over time, lead to irreducible uncertainty. Both clinicians and patients need to be able to make sense, learn, and improvise as events in care unfold. Additionally, it is through relationships that an organization is able to make sense, learn, and improvise to manage the unpredictable trajectories of health.

INFORMATION SYSTEMS, COMPLEXITY, AND HEALTH CARE The potential for complexity science to inform the large-scale design, dissemination, and implementation of health information technologies underway in health care is significant. Health information technologies have been traditionally viewed as mechanisms to increase efficiency and reduce uncertainty in health care delivery. These are both laudable goals; however, are we placing too much emphasis on achieving these goals at the expense of other, perhaps, equally important goals? Traditional views of health information systems and health information technology (IT) are important for our health care system and have led to considerable advances in health care delivery. We must, however, begin to acknowledge and address the limitations in a health care system that overemphasizes the roles of increasing efficiency and reducing uncertainty in health care delivery. For example, electronic health records (EHR) are seen by many physicians simply as the system through which they are reimbursed for services provided. For these physicians, little or no connection exists between the EHR and the task of providing patient care. Under this condition, an EHR

Health Care Organizations as Complex Systems

15

may be misused and therefore introduce unintentional harm. Similarly, some physicians see the EHR as a barrier rather than a facilitator of patient care. Medical records are both a record of transactions, a series of events, and a record of an investigation, an evolutionary process. Paper medical records are often seen as a record of care processes with transactions embedded in the record. The electronic record, on the other hand, is often seen only as a record of transactions. Perhaps more problematic, EHRs tend to be viewed as a mechanism to homogenize health care delivery. Health IT efforts based in this conceptualization, however, are likely to be dysfunctional. Recent research suggests that we must move to understanding electronic health records as artifacts enabling heterogeneity and as platforms for learning (Lanham, 2009, 2010). If taken seriously, this new conceptualization could lead to radically different health IT designs and management strategies. A problem with assuming that health IT is an effective strategy for increasing efficiency in health care delivery is that agents in the health care system are not clear about what is meant by increased efficiency. The health IT community often argues that implementing health information technologies, such as EHRs, clinical decision support, and health information exchanges (HIE), will increase health care system efficiency by, in large part, reducing the number of redundant tests, labs, and procedures performed. The assumption is that these types of information systems allow physicians to view and act based on results from previously performed tests, labs, and procedures, thereby eliminating the need to unnecessarily repeat them. These statements, and others like it, generally have high face validity. That is, until one actually spends time observing what happens when physicians use a health IT in providing patient care. From a number of large field studies of EHR use that we have conducted or participated in over the last decade, the overwhelming consensus from all provider types (medical doctor (MD), nurse practitioner (NP), physician’s assistant (PA)) is that EHRs reduce clinical efficiency. They introduce new and unexpected work, which most often does not subside over time as some suggest. This is true for a variety of reasons, most often cited, however, is simply the time required to use the EHR system (i.e., inputting data and getting data out). This is less of a criticism of EHRs, as this health IT has a momentous place in health care improvement, but rather a call for health IT stakeholders to rethink their fundamental assumptions of health IT and revise estimates of how health IT can and will support the much needed transformations in health care delivery. With a sharp focus on enhancing efficiency, we may be unintentionally

16

REUBEN R. McDANIEL ET AL.

overlooking opportunities for health IT to fundamentally transform the way care is delivered. One other particularly problematic issue with health IT and the increased efficiency assumption is the potential for inadvertently supporting dysfunctional improvements in efficiency. Managers often believe that IT systems by their very nature will increase the efficiency of the system in which they are implemented; this despite the belief by many of the agents in the organization that IT systems often reduce efficiency. What we are facing in health care today is the challenge of a broken system. Inserting IT may only increase the efficiency of this broken system. For example, if providers have billing systems that are set up to overcharge for services, IT will make that process of overcharging for services more efficient. Uncertainty in health care has been attributed to three sources: the complexity of the system itself, unpredictable trajectories of illness, and limits in scientific knowledge (Han, Klein, & Arora, 2011). Clinicians differ in their views of uncertainty (Gerrity, DeVellis, & Dittus, 1990; Gerrity, White, DeVellis, & Dittus, 1995) and their strategies for managing uncertainty (Brashers, 2001, 2007). One way differences in clinicians’ views of uncertainty are manifest is by differences in how clinicians use diagnostic tests and treatments for the same condition or set of presenting symptoms (Roos & Roos, 1981). Another way these differences are manifest is through differences in information-seeking behavior. For example, differences in the amount of information clinicians seek and the sources of information they use in clinical decision-making (Mishel, 1988). Clinicians who view uncertainty as reducible with information will likely display different information-seeking behaviors than clinicians who view uncertainty irreducible (Lanham, 2010). Thus, differences in how clinicians view uncertainty and its management are likely to be associated with differences in how they incorporate EHRs, and other health IT, into patient care activities, resulting in difficult-to-understand heterogeneous EHR use patterns (Lanham, Leykum, & McDaniel, 2011). If irreducible uncertainty did not exist in health care delivery systems, our current health IT designs and strategies for dissemination and implementation might be adequate. Unfortunately, irreducible uncertainty does exist in the systems of care that so desperately need improvement; and, most current health IT designs and implementation strategies are poorly equipped to deal with this type of uncertainty. Current health IT solutions help physicians decrease reducible uncertainty; these health IT enable improvements in information gathering or information processing, which in turn allows updates in the perceived risks involved in patient care. For

Health Care Organizations as Complex Systems

17

reducible uncertainty, information gathering and processing are generally effective. For irreducible uncertainty, these same activities are less, or not at all, effective. When irreducible uncertainty is mistaken for reducible uncertainty, the perceived risk calculations arrived at via information gathering and processing will be distorted, and the actions taken based on those distortions will likely result in poor, unsafe care for patients. A different set of activities based on relationships and uncertainty absorption are required for managing irreducible uncertainty in health care delivery. For health IT to be consistently linked with the kinds of dramatic improvements needed in health care today, these activities, and others that help health care delivery systems address irreducible uncertainty, must be considered in health IT design and implementation efforts.

A PERSPECTIVE OF TODAY’S HEALTH CARE ORGANIZATIONS A complexity science approach helps improve organizational performance. There is a growing interest in performance improvement in health care on a number of fronts, often characterized as improving access, improving quality, and reducing costs. In this section, we discuss specific ways that complexity thinking might help in developing and implementing ideas to meet these objectives. Lanham et al. (2013) articulate the need for health care intervention designs to recognize and leverage self-organization in large-scale dissemination and implementation initiatives. Anderson and McDaniel (2008) discuss the importance of taking complexity science seriously in the development of new research methods and of interpreting research findings. Jordan, Lanham, Anderson, and McDaniel (2010) deal with issues of complexity science research in health care on a meta-level and suggest that interpretation of research results in health care organizations must recognize that they are complex adaptive systems; otherwise, results will likely be overgeneralized. Leykum et al. (2012) use computer simulations, specifically agent-based modeling, to understand the behavior of clinical care teams in academic hospital settings. They found that the nature of the relationships as well as sensemaking and improvisation in these teams were important in understanding the care that was delivered. This work illustrates the potential in health care research for computer simulations that are grounded in complexity science.

18

REUBEN R. McDANIEL ET AL.

Improving health care organizations may be much more difficult than it seems and some of that difficulty may be because improvement efforts are often managed as though organizations are Newtonian in nature. In fact, they are complex adaptive systems. Sometimes those of us working on health care delivery organization issues may be working on the wrong problem because we are using the wrong model(s) to frame our thinking (Lanham et al., 2013). The evidence for this is that our beliefs about health care organizations are often inconsistent with our direct observations of the world. Table 2 identifies possible errors that can result from ignoring complex adaptive systems characteristics as organizations engage in performance improvement efforts. We can look at the way we treat agents in organizations to see dysfunctions in our managerial strategies that may be related to a misspecification of the system’s nature. Managers sometimes act as though they think doctors are dumb even though most doctors are smart. Doctors are given guidelines and incentives in order to produce outcomes desired, and while sometimes these specific efforts work quite well, the overall incidence of medical errors has hardly budged. Nurses are often treated as commodities despite considerable evidence that this is not a good idea, and sometimes it seems as though even as we train more nurses the nursing shortage does not seem to abate. Patients want to live forever without any troubles and die painlessly; therefore, they do not have a calculus of health and neither providers nor patients think like a rational homo economicus when engaging with each other around an illness. Patients’ roles in what happens in health care are very poorly understood. Health care is a process function, and a very diverse set of agents are a critical part of the process. We need to rethink how these agents participate in the process of health care at every juncture of the organization, and a complexity science perspective may help us to do this. Agents in health care organizations have relationships that are difficult to understand and these affect the unfolding of organizational dynamics. Complexity science perspectives help us to develop new ways of thinking about the roles of agents and their influence on processes on health care delivery. These new ways may help us to avoid the kinds of errors noted above. It is difficult in a health care organization to isolate critical variables and know how to manipulate them to get desired effects. The importance of variables may change from patient to patient and within a patient over time. Patient satisfaction, preventive care, and end of life care can vary as a function of environmental influences such as religious beliefs and life

19

Health Care Organizations as Complex Systems

Table 2. A Sample of Possible Errors That Can Result from Ignoring Complex Adaptive Systems Characteristics. Characteristic Diverse agents that learn

Nonlinear interdependences

Self-organization

Emergence

Co-evolution

Typical Errors • Failing to account for employees’ ability to learn unsafe machine operation methods by experimenting on their own with ways to speed up production and thereby reduce the effort they are required to use. • Putting all employees through the same orientation program regardless of differences in cultural interpretations of organizational hierarchies and thereby missing the potential for conflict among employees and between employees and management caused by different expectations of roles. • Disciplining one employee, expecting a modest change in that employee’s behavior but getting a massive union response. • Changing a work process without considering the role of communities of practice on work performance. • Assuming that employees will not reallocate work assignments based on their perception of the best arrangement even after receiving work allocation assignments from management. • Ignoring the speed with which the informal organization can transmit messages and, therefore, failing to control rumors in a productive manner. • Failing to treat the organization’s dominant logic as an emergent property of the system and, instead, treating it as something that can be imposed on the system. • Looking for one bad apple in a group as a strategy for improving work quality, when the quality of output may be an emergent property of the group. • Conducting careful market analysis to determine whether or not to release a new product but ignoring the fact that the release of the new product may change the market in such a way as to make the market analysis incorrect. • Offering a premium to internal workers for extra production without expecting the change in reward structure to affect relationships with suppliers.

Source: Characteristics and Typical Errors of CASs. Reprinted from “Management Strategies for Complex Adaptive Learning Systems: Sensemaking, Learning, and Improvisation,” by Reuben R. McDaniel, Jr., 2007, Performance Improvement Quarterly, 20, 26.

history. Patient responsibility for payment may, in fact, reduce the demand for care, but this is closely related to disposable income. In some cases, the need for care might be a better driver of demand. When is it true that more care is less? When is more expensive care better? Under what conditions

20

REUBEN R. McDANIEL ET AL.

can I let my parents die? Often factors in the environment affect performance and these factors may be unknown to either the patient or the provider. There are many difficult variables and interdependencies to account for in the management of a health care organization. A complexity science perspective may be required if we are to make serious progress in performance improvement (McDaniel, 2007). The magnitude and unpredictability of the changes that seem to be required in all aspects of managing a health care organization may make taking action to fix any single thing seem futile. There may be incremental solutions that seem to work in one part of an organization but any one change or even any set of changes may, in fact, hurt the overall organization, primarily through unintended consequences of the change. Nothing makes a predictable difference anyway, so some managers may think why should I or my organization change? For example, despite many calls to move from a fee for service system that rewards activity rather than results, this is proving very hard for any health care organization. Complexity science calls our attention to the history dependence of organizational change (Allen, Strathern, & Baldwin, 2005), and this may enable us to develop more imaginative and creative approaches to some of the more intractable problems faced by health care organizations. Some people naively believe that markets work in health care. However, there is no market because everybody has to get it, irrespective of their ability to pay. Some observers believe there is a market solution when, in fact, the situation may be more like a utility. It may be that we need two separate and unequal systems within the same health care organization, one run on a market basis and one on a social welfare basis. Everybody’s expense is somebody else’s income and therefore there is no decision platform for making trade-offs and making decisions based on these tradeoffs. Cost-benefit issues at both individual and population levels are impossible to fully calculate because of a lack of agreement about desirability of outcomes and about cause-effect relationships. Benefits may be highly variable across patients as well as within patients, in part because quality of life issues change over stages of life. Costs and benefits may really be comparative issues rather than absolute issues. There had been considerable work in economics studying complex adaptive systems (Arthur, 1999), and this work may provide new ways for conceptualizing economic issues in health care and get us away from the Newtonian cost-benefit analysis that tends to dominate present thinking. Expertise in a health care organization is broad rather than local and extends beyond a specific local situation, and this increases the difficulty

Health Care Organizations as Complex Systems

21

of collaborations and coordination. Nurses seem to move effortlessly from hospital to hospital and doctors often practice in multiple settings. Variable mental models across professionals and knowledge workers make communication among them difficult. With increased specialization and the increased knowledge that this implies coordination of care is more difficult for patient and provider. Professionals desire autonomy from organizational management; therefore, management finds it difficult to impose coordination and collaboration on them. New research in understanding complex networks and in the development of network theory offers a promising approach for examining health care organizations. Likewise, work in institutional theory (Scott, 2001), coupled with work on high-reliability organizations (Weick & Sutcliffe, 2003), may provide new insights into the behavior of networks in complex professional organizations. The traditional approaches to thinking about health care organizations often miss a critical point, namely, the fundamental uncertainties present. It seems as though some of those engaged in thinking about health care organizations are determined to impose an unrealistic rationality on the system. At the very minimum, any efforts to deal with issues of health care delivery must recognize that much of health care is now and will remain uncertain. Health care is a process of delivering services through complex adaptive systems with professionals and other knowledge workers as key participants in the direct delivery of care. Like all organizations, a health care delivery organization is concerned with acquiring resources and maintaining legitimacy, among other things, but its major goal should be helping people so that they can deal with issues of their health and illness. A health care organization is distinct from other complex adaptive systems because of several critical factors. These factors include patient influence over outcomes, difficulty of cost-benefit analysis, highly variable time horizons, and indeterminate technologies. It is the presence of these factors and their interactions that make the health care organization distinct from other CAS. The impact of complexity science has not had as widespread an effect as some have hoped. There are various reasons for this, including the relatively recent origins of this new field of study. Original research on the foundations of complexity science occurred in the 1960s and 1970s but it was only in the 1980s that it became a recognized field with institutes devoted to its study. The Plexus Institute, the primary organization devoted to developing complexity applications in health care was founded in 2000. The recognition of complexity leads to novel approaches and

22

REUBEN R. McDANIEL ET AL.

insights but the apparent lack heretofore of general laws for complexity (Goldenfeld & Kadanoff, 1999) makes it a difficult approach for many to adopt. This may be particularly true in health care where traditional approaches to science have been the mainstay of the discipline. However, it is entirely possible that the avoidance of complexity science thinking goes deeper. In discussing the great difficulties of fully incorporating the basic tenets of a classical worldview into quantum mechanics, Keller (1979) suggested that physicists experience a cognitive repression leading to difficulties in interpreting quantum mechanics. She states that this has been “A result of the failure of physicists to formulate a cognitive paradigm adequate to their theory…. In particular, what is being evaded is the need for a cognitive structure radically different from the prior existing structure” (p. 718). It seems to us to be entirely possible that the avoidance of complexity thinking by health care researchers and practitioners may be driven by a similar phenomenon. People in medicine want to be sure, but complexity raises doubts. Health care providers want to be in control, but interdependencies create new problems. The fact that the resolution of problems in health care may rest in complexity and uncertainty rather than in standard traditional science is frustrating. In this chapter we have tried to address the challenges of dealing with health care as a complex system forthrightly. Despite the difficulties, we are optimistic about the ongoing contribution of complexity science to improvements in health care organizations. The Institute of Medicine’s commentary on the complex nature of health care organizations and the need to attend to the characteristics of complex adaptive systems in thinking about health care organization management is one indicator (IOM, 2001). The chapters in this present volume are another. The management and design of health care organizations is a critical problem in society and we look forward to continuing our work and to encouraging the work of others as we seek insights from complexity science in addressing this problem.

NOTES 1. For more information on Anderson’s work, see http://nursing.duke.edu/ directories/faculty/ruth.anderson. 2. For more information on Crabtree’s work, see http://rwjms.umdnj.edu/ departments_institutes/family_medicine/faculty/full_time/BenjaminCrabtree.html.

Health Care Organizations as Complex Systems

23

ACKNOWLEDGMENTS It was extremely exciting and gratifying to learn that the 2001 chapter, “Complexity science and health care management,” by Reuben R. McDaniel, Jr. and Dean J. Driebe, was one of the most highly cited in the Advances in Health Care Management series. The editor’s invitation to write an extension of that original chapter was enthusiastically received by McDaniel and Driebe, and Holly Jordan Lanham, a colleague who has been an active researcher in the area since the publication of the original chapter, has graciously agreed to join us in this endeavor. The authors would also like to acknowledge the editors and reviewers whose comments have contributed greatly to the quality of this chapter.

REFERENCES Allen, P. M. (1976). Evolution, population dynamics and stability. Proceedings of the National Academy of Sciences of the United States of America, 73(3), 665 668. Allen, P. M. (1993). Evolution: Persistent ignorance from continual learning. In R. H. Day & P. Chen (Eds.), Nonlinear dynamics and evolutionary economics. New York, NY: Oxford University Press. Allen, P. M. (2001). Knowledge, ignorance, and the evolution of complex systems. In J. Foster & J. S. Metcalfe (Eds.), Frontiers of evolutionary economics: Competition, self-organization, and innovation policy (pp. 313 350). Cheltenham: Edward Elgar. Allen, P. W., Strathern, W., & Baldwin, J. S. (2005). The evolutionary complexity of social economic systems: The inevitability of uncertainty and surprise. In R. R. McDaniel, Jr. & D. J. Driebe (Eds.), Uncertainty and surprise in complex systems: Questions on working with the unexpected (pp. 31 50). Berlin: Springer-Verlag. Anderson, R. A., Ammarell, N., Bailey, D., Jr., Colo´n-Emeric, C., Corazzini, K. N., Lillie, M., & McDaniel, R. R. Jr. (2005). Nurse assistant mental models, sensemaking, care actions, and consequences for nursing home residents. Qualitative Health Research, 15(8), 1006 1021. Anderson, R. A., Corazzini, K. N., & McDaniel, R. R. Jr. (2004). Complexity science and the dynamics of climate and communication: Reducing nursing home turnover. Gerontologist, 44(3), 378 388. Anderson, R. A., & McDaniel, R. R. Jr. (2008). Taking complexity science seriously: New research, new methods. In C. Lindberg, S. Nash & C. Lindberg (Eds.), On the edge: Nursing in the age of complexity (pp. 73 95). Bordentown, NJ: PlexusPress. Arthur, W. B. (1999). Complexity and the economy. Science Magazine, 284, 107 109. Bettis, R. A., & Prahalad, C. K. (1995). The dominant logic: Retrospective and extension. Strategic Management Journal, 16(1), 5 14. Brashers, D. E. (2001). Communication and uncertainty management. Journal of Communication, 51(3), 477 497.

24

REUBEN R. McDANIEL ET AL.

Brashers, D. E. (2007). A theory of communication and uncertainty management. In B. B. Whaley & W. Santer (Eds.), Explaining communication: contemporary theories and exemplars (pp. 201 218). Mahwah, NJ: Lawrence Erlbaum Associates. Cohen, D., McDaniel, R. R., Jr., Crabtree, B. F., Ruhe, M. C., Weyer, S. M., Tallia, A., & Stange, K. C. (2004). A practice change model for quality improvement in primary care practice. Journal of Healthcare Management, 49(3), 155 168. Crabtree, B. F., Miller, W. L., McDaniel, R. R., Stange, K. C., Nuttingx, P. A., & Carlosx, R. J. (2009). A survivor’s guide for primary care physicians. The Journal of Family Practice, 58(8), E1–E7. Driebe, D. (1999). Fully chaotic maps and broken time symmetry. Dordrecht: Kluwer Academic. Edmondson, A. C. (2003). Managing the risk of learning: Psychological safety in work teams. In M. A. West, D. Tjosvold & K. G. Smith (Eds.), International handbook of organizational teamwork and cooperative working (pp. 255 275). Chichester, West Sussex: John Wiley & Sons, Ltd. Edmondson, A. C. (2012). Teaming: How organizations learn, innovate, and compete in the knowledge economy. San Francisco, CA: Jossey-Bass. Gerrity, M. S., DeVellis, R. F., & Dittus, R. S. (1990). Physicians’ reactions to uncertainty in patient care: A new measure and new insights. Medical Care, 28, 724 735. Gerrity, M. S., White, K. P., DeVellis, R. F., & Dittus, R. S. (1995). Physicians’ reactions to uncertainty: Refining the constructs and scales. Motivation and Emotion, 19, 175 191. Ghaferi, A. A., Birkmeyer, J. D., & Dimick, J. B. (2009). Variation in hospital mortality associated with inpatient surgery. New England Journal of Medicine, 361(14), 1368 1375. Goldenfeld, N., & Kadanoff, L. P. (1999). Simple lessons from complexity. Science, 284, 87 89. Goldstein, J. (1999). Emergence as a construct: History and issues. Emergence, I(1), 49 72. Han, P. K. J., Klein, W. M. P., & Arora, N. K. (2011). Varieties of uncertainty in health care: A conceptual taxonomy. Medical Decision Making, 40(20), 1 11. IOM (Institute of Medicine). (2001). Crossing the quality chasm: A new health system for the 21st century. Washington, DC: The National Academies Press. Jordan, M., Lanham, H. J., Anderson, R. A., & McDaniel, R. R. Jr. (2010). Implications of complex adaptive systems theory for interpreting research about health care organizations. Journal of Evaluation in Clinical Practice, 16, 228 231. Jordan, M. E., Lanham, H. J., Crabtree, B. F., Nutting, P. A., Miller, W. L., Stange, K. C., & McDaniel, R. R. Jr. (2009). The role of conversation in health care interventions: Enabling sensemaking and learning. Implementation Science, 4(15). Retrieved from http://www. implementationscience.com/content/4/1/15 Keller, E. F. (1979). Cognitive repression in contemporary physics. American Journal of Physics, 47(8), 718 721. Lanham, H. J. (2010). Investigating heterogeneity in physician use of electronic medical records: The role of professional values and perspectives of uncertainty. Doctoral dissertation, The University of Texas at Austin, Austin. Lanham, H. J., Leykum, L. K., & McDaniel, R. R. Jr. (2012). Same organization, same electronic health records (EHRs) system, different use: Exploring the linkage between practice member communication patterns and EHR use patterns in an ambulatory care setting. Journal of the American Medical Information Association, 19(3), 382 391.

Health Care Organizations as Complex Systems

25

Lanham, H. J., Leykum, L. K., Taylor, B., McCannon, C. J., Lindberg, C., & Lester, R. (2013). How complexity science can inform scale-up and spread: Understanding the role of self-organization in variation across local contexts. Social Science & Medicine, 93, 194 202. Lanham, H. J., McDaniel, R. R., Jr., Crabtree, B. F., Miller, W. L., Stange, K. C., Tallia, A. F., & Nutting, P. A. (2009). Organizational change and learning: How improving practice relationships among clinicians and nonclinicians can improve quality in primary care. Joint Commission on Accreditation of Healthcare Organizations, 35(9), 457 466. Lanham, H. J., & Palmer, R. F. (2008, November). Organizational and interpersonal attributes of primary care health practices: Testing a theoretical path model. Poster Presentation presented at the North American Primary Care Research Group Annual Conference, San Juan, PR. Lanham, H. J., Sitting, D. F., Leykum, L. K., Parchman, M. L., Pugh, J. A., & McDaniel, R. R. Jr. (2013). Understanding differences in electronic health records (EHR) use: Linking individual physicians’ perceptions of uncertainty and EHRuse patterns in ambulatory care. Journal of the American Medical Association. doi:10.1136/amiajnl-2012-001377. Lasota, A., & Mackey, M. C. (1985). Probabilistic properties of deterministic systems. Cambridge: Cambridge University Press. Leykum, L. K., Kumar, P., Parchman, M., McDaniel, R. R., Jr., Lanham, H. J., & Agar, M. (2012). Use of an agent-based model to understand clinical systems. Journal of Artificial Societies and Social Simulation, 15(3), 1 25. Leykum, L. K., Parchman, M., Pugh, J., Lawrence, V., Noe¨l, P. H., & McDaniel, R. R. Jr. (2010). The importance of organizational characteristics for improving outcomes in patients with chronic disease: a systematic review of congestive heart failure. Implementation Science, 5(66), 1 10. Leykum, L. K., Pugh, J., Lawrence, V., Parchman, M., Noe¨l, P. H., Cornell, J., & McDaniel, R. R. Jr. (2007). Organizational interventions employing principles of complexity science have improved outcomes for patients with Type II diabetes. Implementation Science, 2(28), 1 8. Maguire, S., McKelvey, B., Mirabeau, L., & O¨ztas, N. (2006). Complexity science and organization studies. In S. Clegg, C. Hardy, T. B. Lawrence & W. R. Nord (Eds.), The Sage handbook of organization studies (2nd ed., pp. 165 214). London: Sage Publications. March, J. G., Sproull, L. S., & Tamuz, M. (1991). Learning from samples of one or fewer. Organization Science, 2(1), 1 13. McDaniel, R. R. Jr. (2007). Management strategies for complex adaptive systems: Sensemaking, learning, and improvisation. Performance Improvement Quarterly, 20(2), 21 42. McDaniel, R. R., Jr., & Driebe, D. J. (2001). Complexity science and health care management. In J. D. Blair, M. D. Fottler, & G. T. Savage (Eds.), Advances in health care management (Vol. 2, pp. 11 36). Stamford, CN: JAI Press. McKenna, K., Leykum, L. K., & McDaniel, R. R. Jr. (2013). The role of improvising in patient care. Health Care Management Review, 38(1), 1 8. Miguel, M. S., Johnson, J. H., Kaski, J. K. K., Dı´ az-Guilera, A., MacKay, R. S., & Helbing, D. (2012). Challenges in complex systems science. The European Physical Journal Special Topics, 214, 245 271. Mishel, M. H. (1988). Uncertainty in illness. Journal of Nursing Scholarship, 20(4), 225 232. Nicolis, G., & Prigogine, I. (1977). Self-organization in nonequilibrium systems: From dissipative structures to order through fluctuations. New York, NY: John Wiley & Sons, Ltd.

26

REUBEN R. McDANIEL ET AL.

Orlikowski, W. J. (1996). Improvising organizational transformation over time: A situated change perspective. Information Systems Research, 7(1), 63 92. Roos, N. P., & Roos, L. L. (1981). High and low surgical rates: Risk factors for area residents. American Journal of Public Health, 71(6), 591 600. Schelling, T. (1978). Micromotives and macrobehavior. New York, NY: Norton. Scott, W. R. (2001). Institutions and organizations. Thousand Oaks, CA: Sage Publications. Shaw, E. K., Ohman-Strickland, P. A., Piasecki, A., Hudson, S. V., Ferrante, J. M., McDaniel, R. R., Jr. & Crabtree, B. F. (2013). Effects of facilitated team meetings and learning collaboratives on colorectal cancer screening rates in primary care practices: A cluster randomized trial. Annals of Family Medicine, 11(3), 220 228. Watts, D. J. (2003). Six degrees: The science of a connected age. New York, NY: W.W. Norton and Company. Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage Publications. Weick, K. E. (2005). Managing the unexpected: Complexity as distributed sensemaking. In R. R. McDaniel, Jr. & D. J. Driebe (Eds.), Uncertainty and surprise in complex systems: Questions on working with the unexpected (pp. 51 65). Berlin: Springer-Verlag. Weick, K. E., & Sutcliffe, K. M. (2003). Hospitals as cultures of entrapment: A re-analysis of the Bristol Royal Infirmary. California Management Review, 45(2), 73 84.

COMMENTARY ON “HEALTH CARE ORGANIZATIONS AS COMPLEX SYSTEMS: NEW PERSPECTIVES ON DESIGN AND MANAGEMENT” BY REUBEN R. MCDANIEL, DEAN J. DRIEBE, AND HOLLY JORDAN LANHAM Ruth A. Anderson ABSTRACT In this commentary, I highlight a few of the assertions made by McDaniel et al. (2013) about the importance of complexity science guided management practices, and extend these ideas specifically to how we might think about reducing seemingly intractable problems in health care such as patient safety, patient falls, hospital acquired infection, and the rise of chronic illness and obesity. I suggest that such changes will require managers and providers to view health care organizations

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 27 36 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015008

27

28

RUTH A. ANDERSON

and patients as complex adaptive systems and include patients as full participants in co-producing their health care. Keywords: Complexity science guided management practices; complex adaptive systems; improving patient outcomes; adaptive leadership

In this commentary, I attempt to push McDaniel, Driebe, and Lanham’s argument for management practices based on complexity science even further in terms of its application in the health care context, particularly for improving patient care outcomes. Complexity science has much to contribute to our understandings of how to use management practices as levers to improve care outcomes for patients, particularly when addressing seemingly intractable problems of inpatients such as patient safety, patient falls, infection or intractable problems of outpatients such as continuing rise in chronic diseases such as diabetes, obesity, and their related co-morbidities. Applying insights that complexity science provides might help us begin reducing or eliminating seemingly intractable problems and change health outcomes in the United States and globally. These types of intractable problems all have a knowledge base of evidence available to guide practice. Despite widespread adoption of bestpractice guidelines and quality improvement models, such as total quality management, business process reengineering, rapid cycle change, Lean thinking, Six Sigma (Powell, Rushmer, & Davies, 2009), payment incentives, and quality indicator reporting (De Vos et al., 2009), we have seen only incremental improvements in hospital-related problems (Dixon-Woods, McNicol, & Martin, 2012) and almost no improvements in population health parameters for chronic illness (Nolte & McKee, 2012). In fact, rises in damaging chronic illnesses (Pendo, 2009) are alarming. By their very nature, intractable problems in health care, because of their complexity, require interprofessional teams to bring various disciplinary perspectives to problem solving, followed by highly coordinated work efforts and effective interdependence. A reason that health care has not been as successful as hoped at implementing scientific knowledge for reducing intractable problems is that managers and management systems tend to only superficially address coordination and interdependencies and rely heavily on technical solutions, such as using information technology to improve communication among team members or inventing faster ways to report system problems. The nature of these seemingly intractable health

Commentary on Reuben R. McDaniel, Dean J. Driebe, and Holly Jordan Lanham 29

problems, however, is such that technical solutions will not adequately address development of relationships needed to support coordination and interdependency in the work (Powell et al., 2009). As McDaniel et al. assert, technical approaches in isolation might actually make outcomes worse by making the system more efficient at producing its poor outcomes and also by interrupting existing interaction patterns among interprofessional teams. Complexity science provides useful guidance about the types of complexity inspired management practices that are needed to complement technical solutions, particularly for problems that require team approaches and effective interdependence in accomplishing the work. In my own research program, for example, my team has been addressing intractable problems in nursing homes using the important levers that complexity science inspired management practice provides. We have tested several complexity inspired management practices and complexity science derived hypotheses and found that relationship-oriented management practices, such as communication, participation in decision making, relationship-focused leadership, and less reliance on rules, resulted in better resident outcomes (Anderson & McDaniel, 1998, 1999; Anderson, Issel, & McDaniel, 2003) and lower nursing staff turnover (Anderson, Corazzini, & McDaniel, 2004). A major contribution of these studies is that they not only supported the significance of complexity science derived management practice for better care outcomes and staff psychosocial well-being, but they also showed that better patient outcomes require attention to multiple system parameters (Stacey, 1996). Thus, in subsequent research, we conducted mixed method case studies in nursing homes to explore staff and manager’s descriptions and explanations about what management practices were used and how and why they worked for better care outcomes. Using complexity science, we were guided in a whole system approach to studying management practices and paid particular attention to relationships (Anderson, Crabtree, Steele, & McDaniel, 2005); we included participants from all organizational roles. Through these case studies and long engagement in the field, our interdisciplinary team learned that the capacity for staff in all job categories to deliver better patient care emerged from their local interactions. Data from workers at all levels (frontline to top administration) and roles (e.g., dietary, nursing, activities, housekeeping, and maintenance) suggested that many important outcomes accrued during interactions as they created networks and communication channels for learning together, exchanging care information, and problem solving; we reported findings in several

30

RUTH A. ANDERSON

publications (see, for example, Anderson, Ammarell, & Bailey, 2005; Colo´n-Emeric et al., 2006, 2010; Utley-Smith et al., 2009). Using these findings and complexity science we posited that organizational capacity to improve is embedded in the interactions among the interdisciplinary team (broadly defined to include frontline), particularly when they use real-time information and explore multiple perspectives; these interactions are used to quickly create new knowledge and facilitate the staff’s ability to create solutions to emerging problems. The case study data provided rich knowledge, revealing areas in which nursing homes possess capacity for better care that is currently untapped because of existing relationship patterns (poor information flow, thin connections, and little cognitive diversity in care planning and problem-solving) (Anderson & McDaniel, 2008) and the new management practices that might be used to alter these relationship patterns to improve the nursing home’s ability to perform. Our team used this knowledge to develop an intervention, CONNECT, to help nursing home staff to use local interaction strategies systematically to improve care. CONNECT is a multicomponent intervention that helps staff to: (1) learn new strategies to improve day-to-day interactions; (2) establish relationship networks for creative problem solving; and (3) sustain newly acquired interaction behaviors through mentoring. A preliminary study of eight nursing homes showed promising results for improving staff interactions and reducing falls (Colo´n-Emeric et al., in press; 2013, ahead of print). We are now conducting a full test of the intervention to reduce falls in nursing homes (Anderson et al., 2012). Although we are addressing only one type of intractable problem, that of patient falls, the complexity science guided approaches create system level changes through new patterns of local interaction, which we suspect will impact other problems as well, because as complexity science posits, global patterns, such as work environments, and care quality, emerge from local interactions (Stacey, 2005). Thus, the first point I would strongly like to emphasize from McDaniel et al.’s (2013) paper is that we must better attend to relationships among staff and between staff and patients in health care to create real change in how care is delivered to truly be able to improve health. As an industry, health care organizations have made some strides in this area, particularly since the Institute of Medicine report (2001) brought attention to changing relationships as a way to improve health care for patients. We can now see this complexity science inspired recommendation in action in some important arenas.

Commentary on Reuben R. McDaniel, Dean J. Driebe, and Holly Jordan Lanham 31

One sweeping national change in this regard is the adoption of the Patient Center Medical Home model that combines new payment strategies with a team-based model of care in which the team provides continuous and coordinated care, hopefully throughout a patient’s lifetime to maximize health outcomes (Rittenhouse & Shortell, 2009). Many technical solutions have been used in creating medical homes, such as extended office hours, increased communication between providers and patients via email and telephone, and increased care coordination. Medical home models also call for other types of changes, however, such as moving from a physiciancentric to a team-centric model, developing highly effective interdependence among team members, and patient-centeredness. These aspects of the medical home model require personal change by physicians and other team members (Nutting et al., 2009), including the patient. These are “adaptive challenges” (Heifetz, Grashow, & Linsky, 2009) that require learning and change. Development of adaptive approaches to facilitate adaptive change have not kept pace with technical approaches and questions remain such as: How do we address “patient-centeredness,” which in this framework is more than patient engagement, which is often defined as involving patients in decisions about their care. Once decisions are made then what? How do patients carry it out after they leave the office? How can physician shift from the engrained physician-centric model to a team-centric model? Such challenges require significant change in attitudes and skills. Patient education, delivered by nurses or medical aides, is also recommended but it is well known that education, while necessary, is not sufficient to create adaptive change in patients’ lifestyles and health behaviors. This brings me to the second point in McDaniel et al.’s article that I wish to emphasize, “Patients’ roles in what happens in health care are very poorly understood” (p. 18). Complexity science, again, has much insight to offer as we seek to understand the exiting nature of patient roles and to develop new strategies for what they could be to finally achieve better outcomes for seemingly intractable health conditions. Thinking with a complexity perspective requires envisioning the patient as a full member of the health care team to co-produce care. For example, using the complexity science guided research of Leykum et al. (2011), about reciprocal learning among care team members in primary care, and Nutting et al. (2010), about adaptive reserve, it would be an incremental step to extend these ideas to include the patient in the learning team. We could ask questions such as how care team members learn more effectively or adapt more efficiently by including the patient in the learning team. Having the patient’s impressions and assessments about the actual

32

RUTH A. ANDERSON

challenges they face, rather than just what the care team thinks they face, will change the way the group makes sense and takes action. Similarly, what would be learned by extending the well-studied concept of relational coordination (Gittell, 2002; Gittell et al., 2000; Gittell, Weinberg, Pfefferle, & Bishop, 2008) to include patients as one of the work groups? Assessing relationships among patients and care teams on parameters such as frequency, timeliness, and accuracy of communication, problem solving communication, shared goals, share knowledge, and mutual respect (Gittell, 2011) would reveal areas for relationship development and new insights about how well a particular team is doing relative to patient engagement with the care team. To really move forward in co-producing care, however, will require an even greater shift in thinking beyond including patients in team learning, care decisions, and care planning. We need a shift that recognizes the limits of medical practice and places more focus on the work that only the patient can do to intervene in their health. The Adaptive Leadership framework, originally developed by Heiftz et al. for business and management (Heifetz et al., 2009), can guide us in this shift. Marcus Thygeson and colleagues wrote the first paper to explicate the adaptive leadership framework for the health care context, describing its roots in complexity science (Thygeson, Morrissey, & Ulstad, 2010). If used, this framework could guide health care managers and professionals to make real inroads into relationship changes that also involve a shift to truly co-producing care with patients. The framework calls for adaptive leaders who can arise at any level of the system, akin to theories of complexity leadership (Lichtenstein et al., 2006; Uhl-Bien, Marion, & McKelvey, 2007), but which I suggest must be present at the point of care where providers and patients interact to ensure that relationships with patients change and providers begin to address the adaptive changes that patient can achieve (Bailey et al., 2012). Only then will we finally begin improving population health parameters, particularly in chronic illness. In a follow-up paper, Thygeson (2013) describes how to create an adaptive-health practice and offers several adaptive strategies for providers to use in changing their own behavior and their approaches to patients as well as how to address adaptive work with patients. Viewing the patient as a complex adaptive system, the adaptive leadership theory suggests that providers engage in sensemaking with patients to understand not just symptoms but the challenges the symptoms create. For example, one patient experiencing “mild” pain might have significant adaptive challenges due to work requirements while another might have moderate or severe

Commentary on Reuben R. McDaniel, Dean J. Driebe, and Holly Jordan Lanham 33

pain but is able to cope with this level of pain because they are retired. Approaches for these two patients would be very different. Thygeson (2013) gives several examples of how technical approaches, particularly overuse of medications, stalls patients from doing the adaptive work needed to improve their health and health might actually decline. Thygeson suggests that we use technical interventions sparingly and only long enough for patients to do adaptive work to heal or improve. Primary care practitioners, attending a recent workshop in which Thygeson and I joined other team members to present the concepts of adaptive leadership, raised concerns about asking patients to do adaptive work, expressing worries that patients would just leave the provider and go to someone who didn’t ask them to work hard on their health. This concern brings us full circle to the need for organization and system level management practices that would ensure that changes such as these would be system-wide; thus patients have nowhere to run. For that matter, neither would be providers! With the aim of increasing research capacity to advance science for symptoms of chronic illness, Duke University School of Nursing was funded by the National Institutes of Health for a Center of Excellence in “Adaptive Leadership for Cognitive/Affective Symptom Science.” In this Center, we will be promoting research that uses the adaptive leadership framework to better understand how patients manage symptoms of chronic illness, and develop interventions that address increasing patients’ capacity for the adaptive work they must do to address the numerous adaptive challenges that accompany complex chronic illness. Important to these approaches will be studies that are at the intersection of chronic illness and care systems to ensure that we also address necessary system changes to support and facilitate adaptive work; here again, we are drawing from complexity science to guide this work (Bailey et al., 2012). In closing, McDaniel et al.’s paper is replete with examples of how and why complexity science thinking would help managers improve health care. I challenge researchers and practitioners to apply these ideas to address the seemingly intractable problems in health care. The field is wide open with possibilities for developing new knowledge and theories derived from complexity science and these could provide new levers through which management practices can influence health outcomes. However, as McDaniel et al. describe, people in health care might have difficulty adopting complexity thinking. They state, “People in medicine want to be sure, but complexity raises doubts. Health care providers want to be in control, but interdependencies create new problems (p. 22).” To me, this suggests that we need adaptive leaders (Heifetz et al., 2009) who can help managers

34

RUTH A. ANDERSON

and practitioners understand when problems have both technical and adaptive challenges. The adaptive leader then helps people develop new skills and behaviors to enable them to do the adaptive work such as shifting their understanding of organizations and relationships as complex adaptive systems.

ACKNOWLEDGMENTS The work of the Center of Excellence: Adaptive Leadership for Cognitive/ Affective Symptom Science is funded by the National Institute of Nursing Research, National Institutes of Health (Anderson and Docherty, MPIs; P30NR014139).

REFERENCES Anderson, R. A., Ammarell, N., & Bailey, D. (2005). Nurse assistant mental models, sensemaking, care actions, and consequences for nursing home residents. Qualitative Health Research, 15(8), 1006 1021. Anderson, R. A., Corazzini, K., & McDaniel, R. R. (2004). Complexity science and the dynamics of climate and communication: Reducing nursing home turnover. Gerontologist, 44(3), 378 388. Anderson, R. A., Corazzini, K., Porter, K., Daily, K., McDaniel, R. R., & Colon-Emeric, C. (2012). CONNECT for quality: Protocol of a cluster randomized controlled trial to improve fall prevention in nursing homes. Implementation Science, 7(1), 11. Retrieved from http:// www.implementationscience.com/content/7/1/11 Anderson, R. A., Crabtree, B. F., Steele, D. J., & McDaniel, R. R. (2005). Case study research: The view from complexity science. Qualitative Health Research, 15(5), 669 685. Anderson, R. A., Issel, L. M., & McDaniel, R. R. (2003). Nursing homes as complex adaptive systems Relationship between management practice and resident outcomes. Nurs Res, 52, 12 21. Anderson, R. A., & McDaniel, R. R. (1999). RN participation in organizational decision making and improvements in resident outcomes. Health Care Manage Rev, 24, 7 16. Anderson, R. A., & McDaniel, R. R. Jr. (1998). Intensity of registered nurse participation in nursing home decision making. Gerontologist, 38(1), 90 100. Anderson, R. A., & McDaniel, R. R. (2008). Taking complexity science seriously: New research, new methods. In C. Lindberg, S. Nash & C. Lindberg (Eds.), On the edge: Nursing in the age of complexity (pp. 73 95). Allentown, NJ: Plexus Institute. Bailey, D. E., Docherty, S. L., Adams, J. A., Carthron, D. L., Corazzini, K., Day, J. R., … Anderson, R. A. (2012). Studying the clinical encounter with the adaptive leadership framework. Journal of Healthcare Leadership, 4(1), 83 91.

Commentary on Reuben R. McDaniel, Dean J. Driebe, and Holly Jordan Lanham 35 Colo´n-Emeric, C., Bailey, D., Corazzini, K. N., Lillie, M., Piven, M. L., Utley-Smith, Q., & Anderson, R. A. (2006). Patterns of medical and nursing staff communication in nursing homes: Implications and insights from complexity science. Qualitative Health Research, 16(2), 173 188. Colo´n-Emeric, C. S., Lekan-Rutledge, D., Utley-Smith, Q., Ammarell, N., Bailey, D., Piven, M. L., … Anderson, R. A. (2006). Connection, regulation, and care plan innovation: A case study of four nursing homes. Health Care Manage Rev, 31(4), 337 346. Colo´n-Emeric, C., McConnell, E., Pinheiro, S., Corazzini, K., Porter, K., Simpson, K., … Andersonx, R. A. (in press). CONNECT for better fall prevention in nursing homes: Results from a pilot intervention study. Journal of the American Geriatrics Society. Colo´n-Emeric, C. S., Pinheiro, S. O., Anderson, R. A., Porter, K., McConnell, E., Corazzini, K., … Simpson, K. M. (2013, ahead of print). Connecting the learners: Improving uptake of a nursing home educational program by focusing on staff interactions. The Gerontologist. doi: 10.1093/geront/gnt043 Colo´n-Emeric, C., Plowman, D. A., Bailey, D., Corazzini, K. N., Utley-Smith, Q., Ammarell, N., … Anderson, R. A. (2010). Regulation and mindful resident care in nursing homes. Qualitative Health Research, 20(9), 1283 1294. First published on May 17, 2010, as doi:10.1177/1049732310369337 De Vos, M., Graafmans, W., Kooistra, M., Meijboom, B., Van Der Voort, P., & Westert, G. (2009). Using quality indicators to improve hospital care: A review of the literature. International Journal for Quality in Health Care, 21(2), 119 129. doi: 10.1093/intqhc/mzn059 Dixon-Woods, M., McNicol, S., & Martin, G. (2012). Ten challenges in improving quality in healthcare: Lessons from the health foundation’s programme evaluations and relevant literature. BMJ Quality & Safety, 21(10), 876 884. doi: 10.1136/bmjqs-2011-000760 Gittell, J. H. (2002). Coordinating mechanisms in care provider groups: Relational coordination as a mediator and input uncertainty as a moderator of performance effects. Management Science, 48(11), 1408 1426. Gittell, J. H. (2011). Relational coordination: Guidelines for theory, measurement and analysis. Retrieved from http://rcrc.brandeis.edu/downloads/Relational_Coordination_Guidelines_ 8-25-11.pdf Gittell, J. H., Fairfield, K. M., Bierbaum, B., Head, W., Jackson, R., Kelly, M., … Zuckerman, J. (2000). Impact of relational coordination on quality of care, postoperative pain and functioning, and length of stay: A nine-hospital study of surgical patients. Medical Care, 38(8), 807 819. Gittell, J. H., Weinberg, D., Pfefferle, S., & Bishop, C. (2008). Impact of relational coordination on job satisfaction and quality outcomes: a study of nursing homes. Human Resource Management Journal, 18(2), 154 170. Heifetz, R., Grashow, A., & Linsky, M. (2009). The practice of adaptive leadership: Tools and tactics for change your organization and the world. Boston, MA: Harvard Business Review Press. Institute of Medicine. (2001). Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academy Press. Leykum, L., Palmer, R., Lanham, H., Jordan, M., McDaniel, R., Noel, P., & Parchman, M. (2011). Reciprocal learning and chronic care model implementation in primary care: Results from a new scale of learning in primary care. BMC Health Services Research, 11(1), 44. Lichtenstein, B. B., Ul Bien, M., Marion, R., Seers, A., Orton, D. J., & Schreiber, C. (2006). Complexity leadership theory: An interactive perspective on leading in complex adaptive systems. E:CO, 8(4), 2 12.

36

RUTH A. ANDERSON

McDaniel, R. R., Jr., Driebe, D. J., & Lanham, H. (2013). Health care organization as complex systems: New perspectives on design and management. Advances in Health Care Management, 15, 3–26. Nolte, E., & McKee, M. (2012). Variations in amenable mortality: A comparison of sixteen high-income nations. Public Health Forum, 20(3), 14.e11 14.e13. doi:http://dx.doi.org/ 10.1016/j.phf.2012.06.011 Nutting, P. A., Crabtree, B. F., Stewart, E. E., Miller, W. L., Palmer, R. F., Stange, K. C., & Jae´n, C. R. (2010). Effect of facilitation on practice outcomes in the national demonstration project model of the patient-centered medical home. The Annals of Family Medicine, 8(Suppl 1), S33 S44. doi:10.1370/afm.1119 Nutting, P. A., Miller, W. L., Crabtree, B. F., Jaen, C. R., Stewart, E. E., & Stange, K. C. (2009). Initial lessons from the first national demonstration project on practice transformation to a patient-centered medical home. The Annals of Family Medicine, 7(3), 254 260. doi:10.1370/afm.1002 Pendo, E. (2009). Working sick: Lessons of chronic illness for health care reform. Yale Journal of Health Policy, Law, and Ethics, 9(2), Article 4. Powell, A. E., Rushmer, R. K., & Davies, H. T. O. (2009). A systematic narrative review of quality improvement models in health care. Retrieved from http://www.healthcareimprovement scotland.org/previous_resources/hta_report/a_systematic_narrative_review.aspx. Accessed on July 8, 2013. Rittenhouse, D. R., & Shortell, S. M. (2009). The patient-centered medical home: Will it stand the test of health reform? JAMA, 301(19), 2038 2040. doi:10.1001/jama.2009.691 Stacey, R. (2005). Introduction: Emergence and organization. In R. Stacey, D. Griffin & P. Shaw (Eds.), Experiencing emergence in organizations: Local interaction and the emergence of global pattern. New York, NY: Routledge. Stacey, R. D. (1996). Complexity and creativity in organizations. San Francisco, CA: BerrettKoehler Publishers. Thygeson, N. M. (2013). Implementing adaptive health practice: A complexity-based philosophy of health care. In J. P. Sturmberg & C. M. Martin (Eds.), Handbook of systems and complexity in health. New York, NY: Springer Science + Business Media. Thygeson, M., Morrissey, L., & Ulstad, V. (2010). Adaptive leadership and the practice of medicine: a complexity-based approach to reframing the doctor patient relationship. Journal of Evaluation in Clinical Practice, 16(5), 1009 1015. doi:10.1111/j.1365-2753. 2010.01533.x Uhl-Bien, M., Marion, R., & McKelvey, B. (2007). Complexity leadership theory: Shifting leadership from the industrial age to the knowledge era. The Leadership Quarterly, 18(4), 298 318. doi:10.1016/j.leaqua.2007.04.002 Utley-Smith, Q., Colo´n-Emeric, C. S., Lekan-Rutledge, D., Ammarell, N., Bailey, D., Corazzini, K., … Anderson, R. A. (2009). The nature of staff-family interactions in nursing homes. Journal of Aging Studies, 23(3), 168 177.

SECTION II REVISITING HEALTH SYSTEMS INTEGRATION

HORIZONTAL AND VERTICAL INTEGRATION OF PHYSICIANS: A TALE OF TWO TAILS Lawton Robert Burns, Jeff C. Goldsmith and Aditi Sen ABSTRACT Purpose Researchers recommend a reorganization of the medical profession into larger groups with a multispecialty mix. We analyze whether there is evidence for the superiority of these models and if this organizational transformation is underway. Design/methodology approach We summarize the evidence on scale and scope economies in physician group practice, and then review the trends in physician group size and specialty mix to conduct survivorship tests of the most efficient models. Findings The distribution of physician groups exhibits two interesting tails. In the lower tail, a large percentage of physicians continue to practice in small, physician-owned practices. In the upper tail, there is a small but rapidly growing percentage of large groups that have been organized primarily by non-physician owners.

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 39 117 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015009

39

40

LAWTON ROBERT BURNS ET AL.

Research limitations While our analysis includes no original data, it does collate all known surveys of physician practice characteristics and group practice formation to provide a consistent picture of physician organization. Research implications Our review suggests that scale and scope economies in physician practice are limited. This may explain why most physicians have retained their small practices. Practical implications Larger, multispecialty groups have been primarily organized by non-physician owners in vertically integrated arrangements. There is little evidence supporting the efficiencies of such models and some concern they may pose anticompetitive threats. Originality/value This is the first comprehensive review of the scale and scope economies of physician practice in nearly two decades. The research results do not appear to have changed much; nor has much changed in physician practice organization. Keywords: Physicians; group practice; horizontal integration; vertical integration

INTRODUCTION As the United States (US) moves to implement health care reform, providers, policy-makers, and researchers are focusing on achieving structural change in health services, both to moderate cost increases and improve quality. Federal policymakers hope that the cost of extending insurance coverage under the Patient Protection and Affordable Care Act (PPACA, 2010) might be partially offset by departing from traditional forms of physician organization and payment. Organizational models that might achieve these changes include horizontal integration of physicians into group practices (multispecialty, single specialty), vertical integration of physicians into hospitals or health plans via salaried employment, and “virtual” integration via contractually based physician-hospital organizations (PHOs), independent practitioner associations (IPAs), and quasi-risk arrangements such as accountable care organizations (ACOs). The context for discussing integrated care has changed markedly since 2000. While discussions of integration in the 1990s focused on hospital

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

41

systems (e.g., Shortell, Gillies, Anderson, Erickson, & Mitchell, 1996), more recent discussion has been extended to include the role of physician organizations (Solberg et al., 2009). There has been a groundswell of interest in large, integrated multispecialty medical groups particularly those sponsoring health plans or engaged in risk contracting as solutions to the problems of low-quality, high-cost, and uncoordinated care (Crosson, 2005; Miller & Bovbjerg, 2002; Shortell & Schmittdiel, 2004). There is also growing interest in the medical community regarding the future organization of medical practice. Over the past 2 years, researchers and consultants have issued several reports on the practice options for the medical profession (Goldsmith, 2012; Isaacs & Jellinek, 2012; Kirchhoff, 2013). All of the reports discuss the increased financial pressures on private physician practice and the relative merits of various integration strategies. Some portend the accelerating collapse of private practice (Isaacs & Jellinek, 2012). Those advocating change in physician organization need to confront several inconvenient truths. First, despite repeated calls to do so, hundreds of thousands of physicians have so far declined to organize themselves into larger economic units that may be better able to achieve coordinated, highvalue care. Surveys indicate that roughly two-thirds of US office-based physicians continue to practice in solo settings, two-person partnerships, and small (mostly single specialty) groups with five or fewer physicians, and that nearly four-fifths of all organized groups have fewer than 10 physicians. Such observations fly in the face of repeated forecasts of physician consolidation into larger scale, multispecialty entities (e.g., Bailey, 1968; Bellows, McMenamin, & Halpin, 2010; Shih et al., 2008). In addition to the mass of physicians in very small settings, however, there is a growing percentage of office-based physicians practicing in groups of 11 or more doctors, including a rapidly growing number of very large practices. These practices are noteworthy in at least two respects: they have been assembled by hospitals and other (e.g., corporate) owners, not physicians themselves, and they still contain a little more than one-third of all group physicians (and somewhat less than 20% of all practicing physicians). This suggests that any analysis of physician practice organization needs to explain this strange two-tailed distribution of physician practices: a mass of small group practices, primarily in single specialties, and large groups of physicians, often in multispecialty practices, that are inside or tightly affiliated with other health care organizations such as hospitals or health plans. There is also a growing hybrid form: large single-specialty groups of hospital-based practitioners consolidated by equity capital firms.

42

LAWTON ROBERT BURNS ET AL.

Second, the presumed inevitability of large multispecialty groups flies in face of an economic record devoid of evidence of measurably superior performance (Tollen, 2008). Part of the problem is that the evidence cited in favor of their performance advantage comes from studies of a handful of mature prepaid multispecialty group practices, such as Kaiser’s Permanente Medical Groups. Such research suffers from what evaluation researchers call “multiple treatment interference”: they combine multiple financial and organizational models (group practice, multispecialty mix, scale, and prepayment) whose effects are virtually impossible to disentangle. As a result, studies documenting their effects may reflect health maintenance organization (HMO) (e.g., financing model) performance rather than group and scale (care delivery) performance. Indeed, the evidence of scale economies (defined as the ability of groups to take advantage of technical indivisibilities such as investments in technology and staffing to achieve production-based efficiencies), scope economies (defined as the ability of groups to leverage the presence of diverse specialists or shared resources to efficiently produce multiple patient services), or quality performance advantages has been strikingly thin, in some cases for decades. These findings should not be surprising. The initial volume of this publication included a review of the options for physicians as they confronted a consolidating landscape (Burns & Wholey, 2000). These options included partnering with other physicians (e.g., group practice, virtual networks using IPAs, national federations), partnering with Wall Street (e.g., physician practice management firms), partnering with hospitals and health systems, and partnering with organized labor (e.g., efforts to legalize unions of community-based physicians). The available evidence then, as well as the subsequent decade of experience, has suggested that there is no one best option with a demonstrated competitive advantage. Rather, the different models presented a menu from which physicians could choose depending, for example, on their desire for professional autonomy versus economic security. We have been asked to revisit this topic, assess what has changed, and discuss how the conclusions reached in the earlier review might be altered. The chapter is organized into the following sections. We first discuss the history of physician practice organization and then analyze the size and specialty distribution of group practices over time to see what has changed. We then summarize the existing evidence on the effects of physician group size and type of integration (horizontal, vertical, and virtual) on quality, cost, efficiency, and productivity/profitability. We also summarize

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

43

the findings from three sets of field investigations of organized physician groups to distill additional insights on their functioning and performance. We then turn our attention to two areas where practice consolidation is taking place on a massive scale: employment of physicians by hospitals and other firms and the formation of (largely equity-based) single specialty networks. We conclude by discussing the future trajectory of physician organizations and their ability to deliver on the policy aims sought in the PPACA.

EARLY HISTORY OF GROUP PRACTICE IN THE US Historians commonly point to the establishment of the Mayo Clinic in 1892 by a family of surgeons (father and two sons) along with an invited fourth partner as the beginning of group practice and the rise of private medical clinics. In actuality, groupings of collaborating physicians had existed for decades in a variety of part-time or short-lived arrangements, such as military medicine, industrial medical worksites, public dispensaries, hospital outpatient departments, and hospital medical staffs (combining medical faculty and residents) (Madison, 1990; Madison & Konrad, 1988). As a result, the storyline is somewhat more complex. Historians of group practice also point to the important role played by military medicine (both prior to, but especially during World War I) in fostering the initial growth of this organizational form (Madison & Konrad, 1988). Wartime practice exposed enlisted physicians to structured practice arrangements, cooperative relationships with specialists, ease of consultation, and efficiency of care. Reluctant to forego the advantages afforded by army field hospitals, many surgeons returned home to establish structured groups with the same benefits of collaboration and standardization they enjoyed during wartime practice. Such was the origin of the Cleveland Clinic, founded in 1921 by three wartime surgeons (who invited in a fourth physician, an internist, to form a multispecialty practice). After World War I, the number of such clinics grew rapidly. There were 125 groups by 1926, and 220 by 1932 (Shouldice & Shouldice, 1978). The first national survey of physician groups (1930) uncovered 13 such clinics with more than 15 doctors (though only 4 had more than 25 doctors). The Mayo Clinic was the outlier, with a staff of 200 physicians; the next largest clinic had only 40. Two surveys conducted by the American Medical

44

LAWTON ROBERT BURNS ET AL.

Association (AMA, 1933, 1940) downplayed the growth of group practices in part because the AMA’s own published principles espoused solo practice, fee-for-service medicine. These principles were directly challenged by the ongoing rise of private clinics, contract medicine (sponsored by mutual benefit societies, lodges, and employers), corporate practice of medicine (e.g., employment by mining companies, hospitals), the growth of medical school faculty practice plans, and the emergence of group or staff model prepaid insurance plans (Havlicek, 1990; Starr, 1982). They also gave rise to state laws explicitly prohibiting the corporate practice of medicine. The growth of the medical profession and the rise of specialty boards during the 1930s provided further impetus to group formation (Stevens, 1971). Physicians sought out settings where they could practice with their professional colleagues (often in the same specialty) and make use of specialized equipment and diagnostic technologies. Some sought settings where they could access colleagues in other specialties. The advent of World War II provided additional stimulus to the growth of private clinics and larger-sized group practices. While the number of clinics with more than 15 physicians remained in the teens up to World War II, the number increased after the war to 31 (1946) and then to 128 (1959), when the AMA softened its opposition to group practice. The number of groups with 16 + doctors then mushroomed to 180 (1965) and then to 301 (1969); the number of groups with 100 + physicians (like the Mayo Clinic) remained in the single digits. Overall, the number of groups climbed from 1,546 in 1959 to 6,371 by 1969. Nevertheless, as a percentage of all physicians, group practice accounted for only 10% of US physicians in 1965 (Smart, 2004).

LONGITUDINAL ANALYSES OF THE SIZE DISTRIBUTION OF PHYSICIAN PRACTICES Researchers commonly argue there has been an ongoing trend toward larger physician practices. Such a postulated trend is consistent with hypothesized scale economies in medical practice. Long ago, Stigler (1958) observed that a good measure of effective scale economies in a given industry can be found by examining the change in the size distribution of firms in that industry and the share of the market each size category controls. This observation has been formalized as the “survivorship principle”: over time, market forces lead firms to change their scale of operations, with the

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

45

most efficient sized firm emerging from industry market dynamics. Is it really true there is a trend toward larger-scale practice consistent with this scale hypothesis? Given there is no data gold standard on physician organizations over time, we collated the results from multiple surveys of physician practices to answer this question. To test the survivorship principle, we examined four different measures utilized in the various surveys: the percentage of physician groups falling into defined size categories, the percentage of group practice positions falling into these size categories, the percentage of physicians practicing in settings of different sizes, and the percentage of office visits to practices of different sizes. These surveys are reviewed below and compared for consistency in the trend data they report.

Survey Data from the AMA The AMA defines group practice as “the application of medical service by three or more physicians formally organized to provide medical care, consultation, diagnosis, and/or treatment through the joint use of equipment, records, and personnel, and with income from medical practice distributed according to some prearranged plan” (Havlicek, 1990). This definition has been consistently applied in the AMA’s census surveys of physician groups since 1964. It includes groups of different ownership models (physician, hospital, staff model health plans) as well as groups that contract with hospitals to provide specific services (e.g., radiology, emergency care), but it does not include the hospitals that directly employ physicians. The AMA has provided two independent series of reports on the size distribution of physician practices. The first is a survey of physician groups conducted periodically over time (published under the name Medical Group Practices in the US); the second is a periodic survey of a sample of physicians in the AMA Masterfile regarding the practice arrangements of patient care physicians (published under the title Physician Marketplace Report). The Medical Group Practices in the US reports provide survey data on the prevalence of physician groups over the longest time frame (e.g., Havlicek, 1990, 1993, 1996, 1999; Smart, 2004, 2005, 2006). Most recently, these data have been incorporated into another AMA publication (Smart, 2013). Since 1969, the number of groups has grown from 6,371 to over 29,000 in 2011. This growth has come in spurts: steady growth to 15,485

46

LAWTON ROBERT BURNS ET AL.

groups from 1969 to 1984, low growth to 16,576 groups from 1984 to 1991, followed by a rise to 19,787 groups by 1995, where the number remained stable over the next 10 years, and finally rising substantially to 29,612 groups between 2005 and 2011. Despite the growth in the number of groups, the size distribution has changed only modestly. Between 1988 and 2011, the percentage of groups smaller than five physicians declined from 50.3% to 41.9%, while the percentage of groups with 5 9 physicians changed from 33.4% to 37.3%. The data also reveal an increase in the percentages of groups with 10 49 physicians (from 14.4% to 18.3%) and groups with 50 + doctors (from 1.9% to 2.4%). The fastest rate of growth occurred among groups in the upper tail of the size distribution (100 + physicians), although their prevalence is low (see Fig. 1). As an alternative test of the survivorship principle, Table 1 shows the distribution of group positions across practices of different sizes. Similar to the distribution of groups, the distribution of group positions shows a decline in the percentage of positions among groups with 3 4 physicians (from 18.1% to 12.5%). Unlike the distribution of groups, the data in Table 2 show little change in the percentage of positions among groups of 5 9 physicians (from 21.5% to 20.7%), groups of 10 49 physicians (from 27.0% to 27.6%), and groups of 50 99 physicians (from 8.5% to 7.7%). 3 16 to 25

4 26 to 49

5 to 6 50 to 75

7 to 9 76 to 99

10 to 15 100+

30% 25% 20% 15% 10% 5% 0% 1988

1991

1994

1996

2003

2004

2005

2011

Fig. 1. Physician Group Practice Size Distribution. Source: American Medical Association (Havlicek, 1990, 1993, 1996, 1999; Smart, 2004, 2005, 2006, 2013).

47

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

Table 1.

Physician Group Practice Positions: Percent by Size Category.

Size

1988

1991

1994

1996

2003

2004

2005

2011

3 4 5 6 7 9 10 15 16 25 26 49 50 75 76 99 100 +

8.8 9.3 12.3 9.2 9.7 8.4 8.9 5.7 2.8 24.8

6.0 7.7 11.0 8.5 8.9 8.0 8.7 4.7 3.1 33.3

6.8 7.9 11.4 9.0 9.6 8.8 8.8 4.4 2.8 30.7

6.9 8.1 11.7 9.2 9.9 8.9 9.1 4.6 2.9 28.7

6.4 6.4 10.4 9.5 10.2 9.1 10.2 5.2 3.0 29.5

5.5 5.8 10.1 9.9 10.3 9.1 9.8 5.3 2.7 31.6

5.6 5.8 10.1 9.9 10.3 9.0 9.9 5.4 2.7 31.4

6.2 6.3 10.5 10.2 11.2 7.9 8.5 5.0 2.7 31.5

Source: American Medical Association (Havlicek, 1990, 1993, 1996, 1999; Smart, 2004, 2005, 2006, 2013).

The most remarkable change occurred in the percentage of positions in large-sized groups (100 + physicians) from 24.8% to 31.8% (percentage stable since 2004).1 Fig. 2 indicates that the percentage of US physicians practicing in groups of three or more (as opposed to solo and two-partner offices) has remained fairly stable since 1995 at roughly one-third. The stability in the percentage of group physicians in the face of a growing number of groups is explained by the increase in the physician population. The latter rose from 467,679 in 1980 to 1,004,635 by 2011, while the number of patient care physicians rose from 376,512 to 767,782. This stability may come to an end with the retirement of baby boomer physicians, however. The Physician Marketplace Report provides somewhat comparable data on the percentage of nonfederal physicians in practices of different sizes; solo practice, practices of 2 4 physicians (which include 2-physician partnerships), groups of 5 9, 10 49, and 50 + physicians, and physicians working in other settings (Kane, 2004a, 2004b, 2009). Trend data between 1999 and 2007 2008 reveal a slight drop in the percentage of physicians in solo practice (from 26.6% to 24.6%) and small increases in the percentage of physicians in group practices of 2 4 physicians (from 18.0% to 21.4%), 5 9 physicians (from 10.9% to 12.9%), and 10 49 physicians (from 10.7% to 12.1%). The only sizeable increase is the percentage of physicians working in hospital settings (more than double from 7.7% to 16.3%) a topic covered more fully later on. The percentage of physicians practicing in groups of five or more doctors increased slightly between 1999 and 2001 but has remained constant at 29 30% from 2001 to 2008.

48

LAWTON ROBERT BURNS ET AL. 40% 35% 30% 25% 20% 15% 10% 5%

Fig. 2.

11 20

05

04

20

03

20

96

20

95

19

91

19

88

19

19

84

80

19

75

19

69

19

19

19

65

0%

Percentage of Physicians in Groups. Source: American Medical Association (Havlicek, 1990, 1993, 1996, 1999; Smart, 2004, 2005, 2006, 2013).

Data from the Center for Studying Health System Change/Community Tracking Study Casalino, Devers, Lake, Reed, & Stoddard (2003) present data from a survey of 6,000 practicing physicians, conducted as part of the Community Tracking Study (CTS) by the Center for Studying Health System Change (CSHSC). CSHSC has tracked the changing market conditions in 12 US metropolitan health care markets over time. The survey depicts trends in the practice size distribution of office-based physicians between 1997 and 2001. This survey shows a drop between 1997 and 2001 in the percentage of physicians in very small practices of 1 2 doctors (from 54% to 47%) and small increases in the percentage of physicians in practices of 3 9 doctors (from 30.5% to 34.9%), 10 19 doctors (from 6.3% to 8.5%), and 20 49 doctors (from 4.3% to 4.7%). There is no change evident at that time in the percentage of physicians in practices of 50 + doctors. CSHSC researchers provide additional data on the distribution of physicians in a wider variety of practice settings (including hospitals) over a longer time period 1996 2008 (Boukus, Cassil, & O’Malley, 2009; Liebhaber & Grossman, 2007).2 These data show a roughly 20% decline

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

49

between 1996 1997 and 2008 in physicians in practices of 1 2 doctors (from 40.7% to 32%), small growth among groups of 3 5 doctors (from 12.2% to 15.0%), but larger growth among groups of 6 50 practitioners (from 13.1% to 19%). There is also a sizeable increase in the prevalence of groups of 50 + doctors, which doubled between 1996 1997 and 2008 (from 2.9% to 6.0%). In 2008, group practices accounted for roughly 40% of physicians, followed by solo practitioners and two-person partnerships (32%), and other settings (28%).

Panel Data from the National Ambulatory Medical Care Survey More recent trend data on office-based physician practices come from the National Ambulatory Medical Care Survey (NAMCS). The survey, conducted by the Centers for Disease Control and Prevention and the US Bureau of the Census, includes a representative sample of ambulatory care visits to physician offices and gathers statistics on the physician’s practice (CDC, 2007a, 2007b, 2008a, 2008b, 2010a, 2010b). This source yields data on the distribution of physicians across practices of different sizes from 2007 to 2011.3 According to Fig. 3, there have been small decreases in doctors in solo practice (from 30.7% to 28.4%), two-physician partnerships Solo

2

3 to 5

6 to 10

11+

35% 30% 25% 20% 15% 10% 5% 0% 2007

2008

2009

2010

2011

Fig. 3. Distribution of Office-Based Physicians by Practice Size. Source: National Ambulatory Medical Care Survey (CDC, 2007a, 2007b, 2008a, 2008b, 2010a, 2010b).

50

LAWTON ROBERT BURNS ET AL.

(from 12.3% to 10.2%), and in practices of three to five physicians (from 29.9% to 26.4%). There has been an increase in physicians practicing in groups of 6 10 doctors (from 16.4% to 18.2%) and an even bigger increase in physicians practicing in groups of 11 + doctors (from 10.6% to 16.8%). Fig. 4 shows the distribution of office visits across these different sized practice settings over a slightly different period (2005 2010). The data suggest that solo practitioners still command the largest market share (nearly one-third) of patient visits, although their share has declined a bit, while visits to small partnerships and groups of three to 5 physicians have remained steady. By contrast, a growing percentage of patient visits occurred in the larger group practice settings of 6 10 physicians (from 14.1% in 2005 to 17.9% in 2010) and a sizeable increase in visits to groups of 11 + doctors (from 9.5% to 14.1%). These data mirror the distribution of groups across size categories above. The smaller practice settings (solo practitioners, two-person partnerships) and smaller groups (three to five physicians) account for two-thirds of physician office visits. While their share of groups and office visits has declined from nearly three-quarters to two-thirds, the bottom tail of the distribution has shown remarkable persistence. If there is a survivorship advantage to scale, it is not apparent in the recent past. The analysis also shows rapid growth in visits to doctors in the upper tail.

Solo

2

3 to 5

6 to 10

11+

45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 2005

2006

2007

2008

2009

2010

Fig. 4. Distribution of Physician Office Visits by Practice Size. Source: National Ambulatory Medical Care Survey (CDC, 2007a, 2007b, 2008a, 2008b, 2010a, 2010b).

51

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

Industry Data from the Medical Group Management Association A prominent industry trade group, the Medical Group Management Association (MGMA), surveys the group size of their members. Fig. 5 shows the change in the size distribution of MGMA member groups between 2004 and 2012. The majority of groups have 10 or fewer physicians. Nevertheless, the data reveal changes at the extremes: a drop in groups of 10 or fewer physicians from 63% to 54%, a 1% increase in groups of 11 25, 26 50, 51 75, and 76 150 physicians, and a 4% increase in the number of large groups of 151 + physicians.4 The MGMA membership is quite diverse, and includes groups that are owned by physicians, hospitals, universities/medical schools, payers, and “others” (foundations, integrated delivery networks (IDNs), equity investors). All need to be dues paying members to be counted in the MGMA database, however. The latter ownership categories have groups of much larger size. MGMA data suggest that the biggest change in the size distribution of its member groups has occurred not among the physician-owned practices but among practices owned by outside organizations. We will revisit this topic later on in our discussion of vertical integration.

1 to 10 FTEs

11 to 25 FTEs

26 to 50 FTEs

51 to 75 FTEs

76 to 150 FTEs

151 or more FTEs

70% 60% 50% 40% 30% 20% 10% 0% 2004

Fig. 5.

2005

2006

2007

2008

2009

2010

2011

2012

MGMA Medical Groups by Size. Source: David Gans, Medical Group Management Association.

52

LAWTON ROBERT BURNS ET AL.

Industry Data from Pharmaceutical Marketing Firms The most recent data has been collected by a pharmaceutical services firm Cegedim Dendrite. Their data suggest there may be as many as 53,000 medical groups in the US as of 2012. We analyzed the trend in the size distribution of groups between 2010 and 2012 using different categorizations of group size to facilitate comparison with the prior surveys presented above. Regardless of which categorization is analyzed, the data suggest remarkably little change in the distribution. Similar to the AMA data, the Cegedim Dendrite statistics suggest that the vast majority of groups are small: 51.2% have 3 4 physicians, another 31.3% have 5 9 physicians, 15.7% have 10 49 physicians, and 1.7% have 50 + physicians.

Summary of Trend Data on Physician Practice Size Distribution The various data sources reviewed above are fairly consistent in the historical trend they depict. Contrary to the assumptions of many researchers, there has been no seismic shift to larger-sized physician practices. The majority of physicians have continued to practice in small settings, including solo practice. Much of the shift in practice organization has been from really small practices (from one, to three to four doctors) to slightly larger practices (from five to nine or more doctors). Consistent with Newhouse’s (1973) observation, many physicians still favor a cottage industry model of practice organization. As some recent observers have noticed, the death of private and small practice is exaggerated (Joszt, 2013). Large groups with 50 or more doctors are still a small, though rapidly growing, proportion of the group population.5 Despite relative stability in the distribution of practices of various sizes, the upper tail accounts for a larger percentage of practicing physicians and the most rapid growth in total physicians and physician visit volumes. Moreover, the average size of these large groups (mostly hospital sponsored) appears to have increased much more than the smaller, physician-owned groups, both in absolute and percentage terms. Such changes within the largest sized groups are masked by the size distributions reported in the figures above. These differences have enormous implications that we shall consider below in our discussion of vertical integration.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

53

LONGITUDINAL ANALYSES OF THE SPECIALTY MIX AMONG PHYSICIAN PRACTICES While there are several surveys that speak to the size distribution of physician groups, only a handful describe the changes in their specialty distribution. Using such trend data we attempt a second survivor analysis regarding specialty mix that may speak to the issue of scope economies. Here too, the story is one of persistence, not of massive change.

Industry Data from the AMA The AMA reports on Medical Group Practices in the US indicate growth in the number of both single-specialty and multispecialty group practices between 1984 and 1997, with only slight change in the percentage mix between the two. Single-specialty practices comprised 70 71% of all groups throughout the period; multispecialty groups, on the other hand, rose from 18% of groups to 22% by 1997. The residual consisted of small groups of family practitioners, which decreased slightly. There was little change in the average size of the single-specialty groups (5.8 6.4 doctors) and multispecialty groups (23.4 26.6). These statistics do not suggest any widespread shift to multispecialty practice as the prior millennium came to a close.

Data from the Center for Studying Health System Change/Community Tracking Study CSHSC researchers suggest that the prominence of single-specialty groups may have grown in the new millennium. Between Round 2 (1998 1999) and Round 3 (2000 2001) of the CTS, the proportion of physicians in large multispecialty groups fell markedly: the percentage in groups of 50 + doctors dropped from 46.4% to 27.4%, while the proportion in groups of 20 49 decreased from 25.2% to 18.8%. Smaller multispecialty groups increased, on the other hand. The researchers note that not a single large multispecialty group was created in any of the 12 CTS sites during the interval, and that several actually disbanded (Casalino, Pham, & Bazzoli, 2004). In California, for example, nearly 150 physician organizations closed or went bankrupt between 1998 and 2002 (Kirchhoff, 2013), possibly reflecting collateral damage from the managed care backlash and the

54

LAWTON ROBERT BURNS ET AL.

triumph of health plans based on broad physician networks over closed panel models. The shrinkage of large multispecialty groups could also have resulted from the collapse of the large physician practice management firms (PPMs) like Phycor and MedPartners, which concentrated their development activities on large groups. Subsequent CTS reports suggest that physicians were increasingly gravitating to mid-sized single-specialty groups (6 50 doctors), with no parallel movement toward multispecialty practice (Liebhaber & Grossman, 2007). This may have reflected an effort by specialists to achieve scale for bargaining leverage with health plans in their communities. The percentage of physicians in multispecialty groups dropped from 30.9% to 27.5% between 1998 1999 and 2004 2005, again perhaps due to the collapse of PPMs built on the roll-up of multispecialty practices. Physician interviews suggest several advantages of single-specialty groups that may explain their growth: simpler governance and operational issues (due to the absence of cross-subsidies to primary care practitioners in the group), ability to gain bargaining leverage with payers, greater profitability or higher earnings per physician due to the absence of money-losing primary care physicans (PCPs), and, crucially, the ability to bill for technical fees through captive imaging and surgical facilities.

Summarizing the Historical Trend Evidence The analyses above cast doubt on the presumed production efficiency benefits of larger scale and multispecialty practice. Physician-owned groups have displayed no massive long-term movement toward such settings. Instead, they have persisted in traditional arrangements of solo/small groups and single-specialty practices, where they still account for a majority (albeit falling) of physicians and patient visits. Groups owned by outside organizations (hospitals, health plans, equity investors) have developed much larger scale, however, and may also have more specialties represented. This suggests there may be two stories at work in physician organization: one characterizing a large number of groups at the bottom tail of the size distribution featuring a traditional, highly fragmented small scale of practice, and one characterizing a much smaller number of very large and growing groups with non-physician sponsorship at the upper end of the size distribution. These two stories are summarized here as a “tale of two tails.” One possible explanation for the first tail is that a sizeable fraction of physicians believe there are limited advantages to scale and scope, and

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

55

therefore have voted with their feet to retain small, single-specialty practices.6 Patients may likewise prefer doctors in smaller settings, which may have helped to sustain them. To test this explanation, the next section analyzes the academic argument for horizontal integration among physicians and the research evidence for scale and scope economies. One possible explanation for the second tail is that, like the IDNs and the hospital system formations of the 1990s, groups of large scale have been assembled (primarily by outside parties in vertically integrated arrangements) to develop market power, contracting leverage over payers, influence referral patterns for lucrative inpatient and outpatient hospital services, and diversify revenue streams. To test this explanation, the following sections analyze the academic arguments for vertical (and virtual) integration and the research evidence.

HORIZONTAL INTEGRATION: ORGANIZATIONAL SCALE AND SCOPE IN MEDICAL PRACTICE Rationale for Horizontal Integration Physician groups can be formed either organically as physicians join one another in same-site and multi-site group practices or via mergers with other groups. Both represent forms of horizontal integration. As noted above, physician groups can also be assembled by hospitals, universities/ medical schools, foundations, and health plans; such instances of vertical integration are analyzed in subsequent sections. Before examining the evidence on horizontal integration among physicians, it is important to discuss the theoretical rationales for such integration. Such a discussion not only serves to highlight the hypotheses that are tested in group practice analysis but also discerns whether rationales for horizontal integration among providers are similar to those posited in the economics and strategy literature in other economic sectors. Academic Theory Academic theory posits several potential motivations for horizontal integration. These include scale economies, synergies, enhanced market power, expedited market entry, market entry without adding a new competitor, capital raising potential, and better use of a target firm’s assets. Some sources of scale economies are based in production efficiencies; others are based in marketing efficiencies and bargaining leverage with suppliers or

56

LAWTON ROBERT BURNS ET AL.

financing sources. For publicly traded firms, there can be the additional benefits of expected stock price increases from an accretive merger, using stock rather than cash in a market upswing to finance the merger, and the opportunity to discipline the managers of the target firm (Besanko, Dranove, & Shandley, 2000). Provider Rationales Over time, health care providers have enunciated their own rationales for horizontal integration. Some of these coincide with industrial organization theory: scale economies, reductions in excess capacity, market power (e.g., increased negotiating leverage with suppliers or purchasers), increased access to capital, and geographic expansion of their product distribution networks. Other rationales for horizontal integration in health care, however, seem unrelated to conventional academic theory: preparation for managing capitated risk contracts, and alignment of strategic purposes among operating units (Burns & Pauly, 2002). The enunciated rationales for physician group development include creating modern practice infrastructure such as information technology (IT) and revenue cycle enhancement, enhancing operating efficiency, creating negotiating leverage, relieving physicians of administrative duties, income preservation, improving quality, increasing scale to manage risk contracts, improving the ability to coordinate care and referrals, positioning to serve as an ACO under health reform, fostering physician leadership, supporting population health, and improved ability to manage an uncertain and turbulent environment (Crosson, 2005; Goldstein, 1996; Krohn, 1998; Shortell & Schmittdiel, 2004). There is thus only partial overlap in the rationales for horizontal integration in industry and in physician practice. This suggests that group practice formation may achieve some benefits of larger scale, but might also serve broader strategic aims of a sponsoring organization. It also suggests that the logistical complexities of medical practice may vitiate scale benefits that might otherwise accrue from horizontal integration in a less complex economic activity. Economies of Scale in Physician Practices Newhouse (1973, p. 51) long ago observed that economies of scale in physician practice are often just assumed to be true. The same can be said for scale economies in other health care organizations. Potential scale economies can include shared fixed costs, specialization of labor inputs (e.g., use

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

57

of non-physician personnel), internalization of referrals, exploitation of reputational economies, bulk purchasing, use of internal quality monitoring, and extended patient coverage (Pope & Burge, 1992). At the same time, physician practices can suffer from several types of inefficiencies including inefficient scale (number of physicians, use of non-physicians, and ancillary services), scale diseconomies due to free riding and higher patient travel costs, excessive use of inputs, excessive administrative costs, and failure to use a cost-minimizing mix of inputs and outputs (Pope & Burge, 1992). What does the evidence say? There are two main lessons from the literature on economies of scale in physician practices. First, group practices appear to be more productive and efficient than solo practices in terms of number of patient visits or gross revenues per physician (Boan, 1966; Bradford, 1995; Bradford & Martin, 2000; Brown, 1988; Frech & Ginsburg, 1974; Gaynor, 1989; Gaynor & Pauly, 1990; Kimball & Lorant, 1977; Lee, 1990; MGMA, 1998; Newhouse, 1973; Reinhardt, 1972; Sarma, Devlin, & Hogg, 2010). Nevertheless, the evidence is far from absolute. Defelice and Bradford (1997), for example, find no consistent differences between PCPs in solo versus group practices in terms of their utilization of physician hours, clinical and clerical staff time, and laboratory and diagnostic equipment. Group practice has the potential to offer a number of benefits, relative to solo practice, including leverage with health plans and hospitals (Casalino, 2003; Haas-Wilson & Gaynor, 1998a; Robinson & Casalino, 1995), profit from ancillary services (e.g., imaging, diagnostic testing), and improved lifestyle for physicians (Casalino et al., 2003). In addition, group practice has the potential to improve quality as well as contain costs through centralized administration, purchasing, and investments in IT (Greenfield et al., 1992; Ketcham, Baker, & MacIsaac, 2007; Kralewski, Wingert, Knutson, & Johnson, 1999; Kralewski et al., 2000; Weeks et al., 2010). Second, however, contrary to the conventional wisdom about scale, small groups tend to be more productive than large groups, with much of the limited evidence suggesting an optimal size of no more than 7 10 physicians (Bailey, 1968; Getzen, 1984; Ketcham, 2001; Weil, 2002; Wheelan, 2009). Ketcham (2001) found a decrease in the ratio of operating costs to relative value units (RVUs, a measure of productivity) as group size increased. However, this effect plateaued once group size reached 4 16 physicians (49,000 63,000 RVUs), after which costs began to increase relative to unit production. Weil (2002) found that a practice size of 10 provided the maximum scale economies based on RVUs; multispecialty

58

LAWTON ROBERT BURNS ET AL.

practices with 50 + physicians actually exhibited diseconomies of scale, potentially due to care delivery at multiple sites, a higher percentage of managed care patients, and less effective control of time and resources. A practice size of 5 10 may be optimal in terms of taking advantage of scale economies and decision-making without having to delegate to semi-corporate structures such as a board (Hough, 2002). Two studies suggest that the relationship between group size and cost has an inverted U-shape, with smaller groups and extremely large groups being more efficient (Frech & Ginsburg, 1974; Marder & Zuckerman, 1985). Efficiencies in large groups were confined to multispecialty practices, and did not extend to single-specialty or family practices; large concentrations of physicians in a single specialty likely require longer patient travel times that, in turn, may dampen patient volume. Large group efficiencies were also more likely among prepaid than fee-for-service practices, suggesting a confounding of practice size and payment method. Large group efficiency stems from using different forms of physician remuneration (e.g., salaries) and spreading the costs of management investments. This conclusion about the efficiency of large-sized groups is based on survivor analyses of AMA groups between 1965 1969 and 1969 1980. Data presented in Table 1 suggests that such efficiencies in large groups are no longer evident after 1991, when the percentage of physicians practicing in large groups plateaued. What explains the apparent limited scale economies in physician practice? As the size of the practice grows, it is possible that monitoring of physician productivity, cost discipline, and coordination of practitioners become increasingly challenging, thus threatening the quality of patient care (although this latter effect has not been rigorously analyzed). Some research suggests that as group size increases, group culture can become less collegial and cohesive, resulting in less organizational trust and focus on quality (Curoe, Kralewski, & Kaissi, 2003). According to Wheelan (2009), increased group size is associated with decreases in intimacy, cohesion, communication, participation, trust, and satisfaction, and with increases in conflict, argumentation, and competition among physicians. This suggests that the associated adverse consequences may offset benefits of larger group size. Newhouse (1973) argued that the combination of revenue sharing and the diffusion of accountability for managing costs led to reductions in hours worked and inefficiency in resource utilization, whose joint effects overwhelmed any potential scale economies. Newhouse concluded that “the cottage industry [of physicians] may not be so bad after all” (1973, p. 39). Subsequently, researchers concluded that there may be scale

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

59

economies that are reached pretty early with increased group size, after which there may be diseconomies (Pope & Burge, 1996). There is some recent evidence that scale economies disappear entirely when data on physician groups are disaggregated. Hough, Liu, and Gans (2013) found increasing returns to scale (measured by total gross charges per full-time equivalent (FTE) physician) in a sample that collapsed groups of different single specialties; however, returns to scale were constant when each specialty was considered separately. Returns to scale were also constant for multispecialty groups. The authors conclude there are few advantages to practice size. Similarly, Escarce and Pauly (1998) find no significant impact of practice size on per-unit practice costs. More generally, Hough et al. (2013) show that the structure and production functions of physician practices differ substantially across specialties. Not only are multispecialty and single-specialty practices different, but also does one single-specialty practice differ from another. Size has different implications for different types of practices in terms of how they generate efficiencies, and may be constrained by local market or cultural factors. They concluded that either physicians have personal preferences for smaller practices (that may/may not sacrifice some efficiencies) or larger size is not rewarded in the market but requires external subsidy by a sponsoring organization. We return to the latter point in the section on vertical integration. In the same vein, Kimball and Lorant (1977) considered two possible sources of higher group productivity: scale economies resulting from the technical combination of resources, and a mix of organizational and environmental factors (e.g., incentives, administration, physician and patient characteristics). For example, physician age exhibited an inverse U-shaped relationship with productivity. The age of the group was also negatively associated with productivity. The presence of a full-time group manager, the current value of the practice’s assets, and the percentage of patients referred inside the group were positively associated with productivity; the percentage of physician visits to see hospitalized patients, the percentage of patient visits that were initial visits, the percentage of patients referred outside, and the age of the office and its equipment were all negatively associated with productivity. Their research highlights an important point made long ago by Bailey (1968) and most recently by Hough et al. (2013) that there may be no one optimal size group, since different groups can utilize different resources, inputs, and product mixes to their advantage. This suggests the important role of management effectiveness, rather than scale, in group performance (Goldsmith, 2012).

60

LAWTON ROBERT BURNS ET AL.

Economies of Scope in Physician Practice Another potential dimension that affects physician group performance is the practice’s scope of services for example, multispecialty or single specialty. The potential benefits of multispecialty group practice are mainly derived from opportunities to improve coordination and quality of patient care, keep referrals in-house, and capture high-revenue services such as outpatient surgeries and imaging services. The empirical evidence on economies of scope, however, is very limited and mixed (Pauly, 1996). Some researchers have found large multispecialty groups to be the most efficient type of practice (Frech & Ginsburg, 1974; Lee, 1990; Marder & Zuckerman, 1985). However, the researchers note that large groups may suffer cost disadvantages if there are multiple sites of care; for example, overhead costs and the challenges of monitoring physician performance increase with practice size and geographic separation. Two other studies report positive evidence for scope economies. Hillson, Feldman, and Wingert (1990) reported strong scope economies (in terms of the amount of physician-reported work and time, similar to RVUs) in the provision of different professional services during a patient office visit within the same group, although their results are based on patient vignettes examined by physicians in two large multispecialty practices in one city. Weeks et al. (2010) found that patients treated in large multispecialty group practices received higher-quality, lower-cost care compared to Medicare beneficiaries treated in other types of practices. However, their results are limited to fee-for-service beneficiaries and a sample of 20 multispecialty groups that self-selected to participate in a council focused on accountability for costs and quality of care. Their study also notes the importance of favorable selection of patients into the multispecialty groups. There are also three negative studies. Kimball and Lorant (1977) found no evidence for scale economies among multispecialty groups: physician productivity (measured in terms of net income per doctor or annual visits per doctor) declined monotonically with size, or exhibited a shallow U-shaped association with productivity (gross revenue per doctor). Rosenman and Friesner (2004) found that single-specialty groups were more efficient than multispecialty practices. Their evidence also suggested that single-specialty primary care groups had the highest levels of allocative efficiency (use of input resources) and scale efficiency. The researchers conclude that large multispecialty practices could improve their efficiency by dissolving their combination of specialists and family/general practitioners and becoming single-specialty practices. In this way, the dissolution of

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

61

multispecialty groups may contribute to the growth in single-specialty groups. Finally, Sarma et al. (2010) find that combining family practitioners with specialists fails to improve office productivity. Indeed, as noted earlier, physician efforts to develop large multispecialty groups had ceased by 2001 in all 12 CTS sites (Casalino et al., 2003). Casalino et al. (2004) suggest that the formation of large multispecialty groups has been retarded by the decline in HMOs, in whose panels multispecialty groups played a pivotal role, as well as the decline of capitated contracting generally. Conversely, other factors have fueled the increase in single-specialty groups: capital and scale economies to invest in equipment and facilities to provide imaging and surgical services, negotiating leverage with payers, reputation as a high-quality group, professional management to deal with the regulatory environment, and lifestyle benefits due to the presence of colleagues and shared call (Casalino et al., 2004).7 Given limited evidence of their economic efficiency, Pauly (1996) suggests that the primary competitive advantage of multispecialty group practices may be based in coordinating processes of care (e.g., to execute managed care risk contracts) as opposed to any inherent economies of scale or scope.

Additional Evidence on Scale and Scope: Results from Three Field Investigations of Physician Group Practices Three field investigations of group practices provide additional evidence regarding the effects of group size and specialty mix, although such effects were not the major questions of interest. We briefly summarize these studies and cull their findings relating to size and specialty mix; a detailed summary of the three investigations is available from the lead author. Center for Organized Delivery Systems/Center for Health Management Research Between 1996 and 1999, a large research team from the Center for Organized Delivery Systems (CODS) and the Center for Health Management Research (CHMR) combined forces to study 61 physician organizations associated with 14 organized delivery systems. The study sought to identify the factors associated with physician alignment with hospital systems and the implications of such alignment for implementing evidence-based care management practices (CMPs). The study also analyzed the facilitators and barriers to achieving this alignment.

62

LAWTON ROBERT BURNS ET AL.

Shortell et al. (2001) found that the size and specialty mix of the group exhibited no association with an index of care management protocol use. Waters et al. (2001) likewise reported no impact of size on receptiveness to CMPs and participation in care management activities. Primary care groups were more comfortable with CMPs than were either single- or multispecialty groups, likely because they were perceived as less disruptive to their practices. Group size and structure bore little relationship with doctor attitudes toward CMPs. National Study of Physician Organizations Researchers at the University of California Berkeley (and their colleagues) have conducted three surveys over time of large physician organizations: groups of 20 + doctors and IPAs. The surveys were conducted in 2000 2001, 2006 2007, and (most recently) 2011 2012. The surveys have addressed a host of research issues and uncovered many important findings. Many of these deal with the implementation of CMPs, as studied in the CODS/CHMR project, and extend the earlier research considerably. Other findings deal with the impact of size and specialty mix and are thus quite germane to this review. At present, only data from the first two survey waves have been analyzed and published. Below, we consider the results that pertain to CMPs and group structure emanating from the two waves of data. In the first wave of the National Survey of Physician Organizations (NSPO), researchers found that CMP use was associated with larger group size but not with the group’s specialty mix (Casalino et al., 2003). An additional set of studies found positive effects of group size on a host of measures (CMP use, chronic care model use, electronic medical record (EMR) adoption, use of disease registries and health risk appraisals, etc.), but the effects were small in magnitude; again, multispecialty practice did not exhibit any consistent relationship. Similarly, Shortell et al. (2005) found that high-performing groups on an overall dimension of quality were likely to be larger in size, although the effect size was weak and the relationship did not hold up for two of the quality measures that comprised the overall dimension. Larger size also distinguished groups with higher profitability. Multispecialty groups, on the other hand, were not distinguished as high performing groups in terms of clinical quality or financial performance. Rittenhouse and Robinson (2006) found only inconsistent evidence on the relationship between group size and quality. Finally, CMP use was no more pronounced in nine of the largest physician groups in the US compared to the wider sample of groups studied (Rundall et al., 2002).

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

63

Results from the second wave of NSPO parallel those of the first. CMP use was associated with group size, but only in really large practices (threshold effects evident only in groups with more than 440 doctors); CMP use was not associated with multispecialty practice (Rittenhouse et al., 2010). CMP use was much lower in practices with fewer than 20 physicians (Alexander, Maeng, Casalino, & Rittenhouse, 2012). High threshold effects of group size were also evident in patterns of clinical IT capabilities such as computerized physician order entry (CPOE) and electronic registries (Robinson et al., 2009) and patient-centered medical home and care coordination activity (Rittenhouse, Casalino, Gillies, Shortell, & Lau, 2008). Longitudinal analyses revealed, however, that changes in size and CMP usage were not associated (Shortell et al., 2009). Finally, size was associated with the percentage of time recommended care was delivered or outcomes were achieved (Damberg et al., 2010). University of Minnesota Studies of physician group practices by Kralewski and colleagues differ from the above field investigations in several important respects. First, they analyzed groups for far longer, beginning in the mid-1980s. Second, each analysis encompassed different numbers and types of groups, leading to greater diversity in the populations studied. Third, they analyzed not only CMPs and quality programs but also group structure (bureaucracy, staffing mix) and culture, as well as outcomes such as cost, efficiency, and quality. Fourth, their studies usually focused on groups practicing in Minnesota and the Upper Midwest. Kralewski, Pitt, and Shatin (1985) documented that multispecialty groups are the most bureaucratically complex form of group practice, followed by family/general practice and then single-specialty groups. Two studies (Curoe et al., 2003; Kralewski, Dowd, Kaissi, Curoe, & Rockwood, 2005) found that larger-sized and multispecialty groups exhibited lower scores on most dimensions of group practice culture (e.g., quality emphasis via peer control). Subsequent studies found that group size was associated with employment of physician extenders (Kaissi, Kralewski, & Dowd, 2003), adoption of electronic health records, and the percentage of prescriptions sent electronically (Kralewski et al., 2008). Multispecialty practice was associated with e-prescriptions but not with use of extenders. With regard to quality and cost, the researchers found no evidence that group size or specialty mix was associated with prescribing errors

64

LAWTON ROBERT BURNS ET AL.

(Kralewski, Dowd, Heaton, & Kaissi, 2005). Larger groups had higher costs per episode in one study (Kralewski et al., 1999), but equivalent levels of resource use in two others (Kralewski et al., 2000; Kralewski, Dowd, Xu, & Knutson, 2011). Multispecialty practice was associated with higher cost. Summary of Findings from Field Investigations Overall, the three investigations point to only limited benefits of group size. Any size effects that exist may occur only at high threshold levels of group size and may therefore be weak or non-existent in the vast majority of smaller groups. There also do not appear to be any consistent benefits to multispecialty practice. Indeed, the investigations support earlier conjectures that large multispecialty practices may incur bureaucratic and coordination costs that vitiate any benefits of scope. These findings are generally consistent with other research evidence that finds no effect of group size (Pham, Schrag, Hargraves, & Bach, 2005) and that structural characteristics of physician practices explain little variation in quality of care (Keating et al., 2004) and exert few consistent relationships (Greenfield, Rogers, Mangotich, Carney, & Tarlov, 1995). Summary: Scale and Scope Economies in Physician Practice The literature reviewed finds limited evidence of scale and scope economies in physician practice. First, scale economies appear to be quickly reached by groups of 10 or so physicians; second, scope economies do not appear to exist or are weak at best. These findings suggest why there is a thick tail at the lower end of the size distribution of physician groups and why there has been no change in group specialty mix over time. This reinforces the survivorship hypothesis: if there were measureable advantages to scale and scope in physician practice, the physician practice landscape would have consolidated at the large end, and be dominated by large multispecialty physician groups. Rather, it appears that smaller and single-specialty practices dominate the medical group landscape because they are equally or more efficient than larger multispecialty groups. Moreover, like the distributions of practice size reviewed above, the results pertaining to physician scale and scope efficiencies appear fairly stable over time.8 The next section explores the benefits of vertical integration and the consequences of physician alignment with outside partners (e.g., hospitals, universities/medical schools, foundations, health plans). The analysis here shifts to groups occupying the upper tail of the size distribution.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

65

VERTICAL INTEGRATION Types of Vertical Integration Physicians occupy a central position in the health care value chain that links payers, providers, and producers (Burns & Wholey, 2000, Figure 1). Several researchers have noted that physicians control (directly or indirectly) 80% or more of all health care spending via their decisions regarding inpatient admissions, specialist referrals, surgical procedures, diagnostic testing, and drug prescribing (Eisenberg, 2002; Sager & Socolar, 2005; Sirovich, Gallagher, Wennberg, & Fisher, 2008). They are also central because physicians are targets of marketing efforts undertaken by many players in the value chain (e.g., pharmaceutical, medical device, capital equipment firms). Finally, they are central to the provision of primary care, and are for most patients the initial point of contact with the health care system. Physicians thus serve as the air traffic controller directing the patient’s flight plan across various delivery settings. (Though, to extend the metaphor, sometimes they monitor and guide the takeoff but delegate the landing to the family in the post-discharge, post-acute realm). This makes them ideal partners for employers, hospitals, and insurers in developing more economic care trajectories. It is thus not surprising that many of these value chain participants might be tempted formally to incorporate physicians into their organizations to guide the flow of clinical resources. Physicians can be salaried and/or employed by hospitals, insurers, and employers, or they can be enmeshed in a rich matrix of subsidies, such as call pay, directorships, service contracts, incentive payments, gain sharing and the like that one might term “partial integration.” The sections below explore some of the rationales for these various relationships.

History of Physician Hospital Integration The physician hospital relationship has had a fascinating and turbulent history (Burns, Goldsmith, & Muller, 2010). Yet federal health payment reforms over the past three decades have placed an increasing premium upon their closer collaboration. During the early 1980s, passage of the Medicare Prospective Payment System shifted hospital inpatient payment away from an a’la carte toward a prix fixe model, but left physician fee-forservice payment intact. This divergence in payment methods (now referred

66

LAWTON ROBERT BURNS ET AL.

to as a lack of financial alignment) led to increased hospital interest in partnering with their medical staffs. Nevertheless, such interest was typically confined to joint ventures and PHOs (Shortell, 1991); the 1980s saw little hospital movement toward physician employment. According to the AMA’s Socioeconomic Monitoring System (SMS) Survey of a physician random sample, as of 1988, hospital employment of physicians was minimal (3.4% of non-federal patient care physicians), as was employment by universities/medical schools (4.7%). The most typical employer of a physician was another physician (Marder, Emmons, Kletke, & Willke, 1988). During the early 1990s, the proposed Clinton health reforms called for the development of “accountable health plans” that were composed of providers (both hospitals and physicians) that were to contract on an at-risk basis with proposed new regional health alliances (formerly known as health insurance purchasing cooperatives) (Enthoven, 1993). The intention was to foster the creation of Kaiser-like group or staff model integrated health care enterprises that accepted what is now called global risk contracts. The prospect of this contracting model, which never came to fruition, stimulated a market panic in provider communities, and led to a proliferation of hospital physician economic relationships: PHOs, management services organizations (MSOs) to provide services to independent physicians, hospital-sponsored or hospital-affiliated IPAs, group practices without walls, and direct hospital employment.9 With the exception of the salaried model, the other models of physician hospital collaboration rose and then declined during that decade (Fig. 6). As a consequence of the PPACA 2010, there has been renewed focus on vertical integration between physicians and hospitals. All of the economic models initially developed during the 1990s can support the physician infrastructure for ACOs: PHOs, MSOs, IPAs, salaried employment, as well as so-called “clinically integrated networks” of affiliated primary care and multispecialty group practices. Thus, there has been reversal of the downward trends observed in Fig. 6 going forward.

Rationale for Vertical Integration Academic Theory Academic theory suggests several rationales for vertical integration in industry. These include minimizing the sum of production and transactions costs, for example, by fostering closer collaboration between adjacent stages in the value chain when the gains from their coordination under a

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

67

35% 30% 25% 20% 15% 10% 5%

19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11

0%

Fig. 6.

Independent Practice Association

Group Practice Without Walls

Open Physician Hospital Organization

Closed Physician Hospital Organization

Management Service Organization

Integrated Salary Model

Equity model

Foundation

Physician-Hospital Arrangements 1993 2011. Source: Peter Kralovec, Health Forum, American Hospital Association.

common hierarchy exceed the loss of scale economies from market-based transactions. Other rationales include reducing the threat of opportunistic behavior by trading partners, securing stable distribution systems for finished products, pooling of complementary assets, ensuring access to needed inputs (and blocking competitor access to those same inputs), and creating market power over buyers and suppliers (Besanko et al., 2000). In the health care instance, hospital physician integration can theoretically lead to efficiency gains by lowering transaction costs and improving efforts to monitor, manage, and coordinate patient care. Provider Rationales Rationales for vertical integration in health services diverge somewhat from those offered by academic theory. During the 1990s, a period of frenzied vertical integration activity, the rationales included preparing for global risk contracting or capitation (e.g., by incorporating PCPs into hospital networks), increasing network size and geographic coverage to handle risk contracting, taking responsibility for the health status of the local population, offering a seamless continuum of care, responding to federal and state health reform legislation, and protecting and expanding the supply of physicians (Burns & Pauly, 2002). During the 2000s, some additional rationales

68

LAWTON ROBERT BURNS ET AL.

were added: mitigating competition between hospitals and their medical staffs, sharing the cost of clinical IT with physicians, helping physicians stabilize their incomes and supporting malpractice expenses, increasing the predictability of the physician’s caseload with a desire to improve care, developing regional service lines, creating entry barriers to key clinical services, helping hospitals deal with physician shortages and recruitment needs, developing a branding and differentiation strategy, enhancing clinical quality, leveraging payers, and preparing for ACOs and the “triple aim” (Goldstein, 2005). Hospital versus Physician Perspectives Moreover, the vertical integration rationales offered by hospitals and by physicians have often differed from one another (Burns & Muller, 2008). Hospital goals have centered on capturing outpatient market share, increasing hospital revenues and margins, increasing hospital leverage over pricing, improving care processes and outcomes, addressing pathologies in the traditional medical staff by “aligning incentives,” increasing physician loyalty, and increasing physicians’ incomes. For their part, physicians have sought increased access to capital and technology, greater physician influence (vis-a`-vis payers), greater physician satisfaction, increased patient service quality, and increased incomes with reduced business risk. The overlap in the two sets of rationales seems to be limited to enhanced quality and increasing physician incomes. Benefits to Hospitals From the hospital’s perspective, physician alliances can generate increased inpatient admissions and outpatient visits, as well as consultations with hospital specialists through referrals. Many hospitals have utilized acquired physicians to build up outpatient volumes and revenues in the face of flat inpatient business. Medicare has reimbursed hospital-based outpatient care at much higher rates than similar care provided in the community: evaluation and management (E&M) visits were priced 80% higher, ambulatory surgical services were priced 74% higher, and outpatient imaging tests were priced 141% higher (MedPAC, 2012; Regents Health Resources, 2011). Commercial insurers may also face higher rates charged by hospitalacquired practices (Advisory Board, 2012). The result has been faster rates of growth in Medicare fee-for-service outpatient spending (37.0% during 2005 2010) compared to inpatient spending (9.4%), where admission rates have been flat.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

69

Vertical integration with physicians can also theoretically improve hospitals’ bargaining positions in the face of increasing HMO/managed care presence. Research from the 1990s showed that the presence of any hospital physician contractual model, and especially tighter linkages such as employment, was associated with a higher percentage of hospital revenue from MCOs (Morrisey, Alexander, Burns, & Johnson 1996). Integration can also potentially help hospitals to succeed under pay-for-performance and shared savings models by improving cost control and quality of care; for example, physician employment might allow for greater collaboration, improved monitoring of quality of both physician and hospital treatment, and improved CMPs (Budetti et al., 2002; Madison, 2004; MedPAC, 2008). In the absence of affiliation with local physicians, a hospital may be concerned that these physicians will affiliate with another hospital, develop their own risk contracting capacity through IPAs, or invest in physicianowned competitors such as ambulatory surgery centers or specialty hospitals which could divert patients and revenues, especially in competitive markets (Casalino & Robinson, 2003; MedPAC, 2008). Physician acquisitions can help hospitals deal with shortages of both primary care and (some) specialist physicians, as well as with decreased availability of physicians due to the historical trend among younger doctors to work fewer hours and decreased willingness to take call and work in the emergency room (Kirchhoff, 2013; Staiger, Auerbach, & Buerhaus, 2010). Acquisitions can thus facilitate the hospital’s physician recruitment and medical staff planning strategies. Similarly, physician employment may allow hospitals flexibility to deal with either a continued fee-for-service environment or a possible shift to more risk-based contracting (Kocher & Sahni, 2011).

Benefits to Physicians From the physician’s perspective, vertical integration with a hospital may lead to additional income (e.g., through ancillary services, higher professional fees in hospital outpatient settings, or billable technical fees for hospital-based providers under Medicare), better access to insurers’ networks and better payment rates, and improved lifestyle (MedPAC, 2008). The potential for gains from hospital physician integration, however, is counterbalanced by the challenges of aligning the two parties’ interests and incentives (Budetti et al., 2002). Many physicians may seek employment to limit their business risk, pay down their debts, or bridge to a planned nearterm retirement on the hospital’s salary guarantee. Many hospital CEOs report a marked fall-off in physicians’ productivity as they transition from

70

LAWTON ROBERT BURNS ET AL.

private practice to hospital employment. They report some younger physicians choosing hospital employment solely for lifestyle reasons. Physicians might also seek employment by hospitals to avoid bearing the costs of compliance with the HITECH Act (2009) requirement to demonstrate meaningful use of electronic medical records. In addition, employment may enable physicians to expand their group to afford new equipment and services, prepare for risk contracting, and seek safety from the impact of reform.

Potential for Anticompetitive Effects Hospitals formed IDNs with physicians during the 1990s to garner risk contracts and leverage insurers for higher reimbursement. The former strategy met with limited success; the latter strategy met with little in that decade. There is concern today that the main motivation behind hospitals’ development of ACOs and acquisition of physicians isn’t saving Medicare money, but generating more fee-for-service income. Academic researchers recognize that vertical integration can have significant anticompetitive effects (Gal-Or, 1999; Gaynor, 2006; Haas-Wilson & Gaynor, 1998a; Simpson & Coate, 1998). For example, if a hospital controls a large number of physicians in a region, this could limit the access of both health insurers and other hospitals to physician services and/or reduce acquisition opportunities for competitors or outside firms in the physician services market. Such vertical integration and resultant market leverage might well be used by the hospital to increase the prices paid to its integrated physicians and hospital services by threatening not to contract with one or more local health plans; here, a competitive physician market is rendered less competitive through integration. In these cases, integration might lead to higher prices of hospital/physicians services and potential price discrimination (Simpson & Coate, 1998). It is also possible that vertical integration does not have anti-competitive effects, if health plans and consumers have multiple choices of competing integrated networks in a given market. Vertical integration might promote competition when competing hospital physician networks coordinate patient care to lower costs and improve quality, reduce their transactions costs, and pass along the savings to payers, forcing their competitors to match their efforts. This was the core thesis behind the “managed competition” model advocated by Enthoven (1993). The extent to which this happens is a vitally important research issue going forward.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

71

We suspect that the current flurry of vertical integration activity will have a net anti-competitive and thus cost-increasing effect. First, as reviewed below, vertical integration in health services has been shown to lead to higher prices rather than lower prices. Second, there is a chorus of allegations about the anticompetitive nature of hospitals’ efforts to acquire and employ physicians. These complaints are voiced by competitor hospitals, independent physicians, and commercial health plans (Advisory Board, 2013; Creswell & Abelson, 2012; Indest, n.d.). Such complaints (typically by insurers) often prompt Federal Trade Commission (FTC) or Justice Department investigations of anticompetitive conduct by physician hospital combinations. Prevalence of the Physician Employment Model in Hospitals There are multiple data sources on the level and diffusion of physician hospital employment as well as non-employment integration strategies. Some of these data sources describe the models utilized by hospitals to align with their physicians. Fig. 6 depicts the rise and fall of most contracting vehicles (e.g., PHOs, MSOs, IPAs) during the 1990s and 2000s; the only model that has increased in prevalence is salaried employment by hospitals. This model has thus become the focus of interest to both providers and researchers.10 In addition to salaried employment of physicians, the AHA also tracks the prevalence of hospital systems that have integrated both physician and insurance components (cf. Bazzoli, Shortell, Ciliberto, Kralovec, & Dubbs, 2001; Bazzoli, Shortell, Dubbs, Chan, & Kralovec, 1999). The number of such systems nationally has consistently hovered around 40 despite the increase in systems from 325 in 2000 to 427 in 2010 (Burns, Wholey, McCullough, Kralovec, & Muller, 2012). We can find no research on the physician groups that comprise these systems; they are likely captured by the “other” ownership category in the MGMA database. Thus, even as systems have grown as a share of total hospital owners, the proportion of them participating directly in health insurance markets is falling. Prevalence of Employed Physicians These surveys describe the percentage of US hospitals with the employed model, not how many physicians are involved. Estimates of the number and percentage of employed physicians vary widely. According to the AHA, the 2000 2011 period saw a 57% rise in physician employment. The

72

LAWTON ROBERT BURNS ET AL.

number of employed doctors (FTEs, part-time equivalents, and a small number of dentists) rose from 79,330 to 125,087. If one weights part-time physicians as half of full-time physicians, the number of employed doctors rose from 70,987 to 111,741. As a percentage of total US patient care physicians, the ranks of physicians employed by hospitals grew from 11.0% to 14.6%. The trend data on hospital employment of physicians are presented in Table 2. Other trend data suggest similar sharp increases in the percentages of physicians employed by hospitals. Data from the AMA’s Physician Marketplace Report reveals that the percentage of all physicians employed by hospitals rose from 7.7% in 1999 to 16.4% by 2007 2008 (Kane, 2004a, 2009). Based on periodic community surveys, CSHSC researchers reported a rise in physicians working as hospital employees in the 12 markets they track from 10.7% (1996 1997) to 12.0% (2004 2005) (Liebhaber & Grossman, 2007). A portion of these employed physicians are hospitalists. There are no definitive data on the total number of hospitalists or the percentage employed by hospitals; some are employed by physician groups. A 2009 survey conducted by the Association of American Medical Colleges (AAMC) suggests that hospitalists comprise 9% (21,100 22,900) of PCPs Table 2.

Physicians Employed by Hospitals (1998 2011).

Year

Full-Time MDs and DDs

Part-Time MDs and DDs

Total Part- and Full-Time

Total FTEa MDs and DDs

Total FTE Residents

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

62,152 62,570 62,697 61,972 63,845 61,956 64,392 67,792 71,277 76,785 83,495 85,634 91,282 98,323

15,837 17,484 16,633 16,734 17,939 18,752 19,514 20,592 21,943 23,875 23,850 26,412 24,139 26,734

77,989 80,054 79,330 78,706 81,784 80,708 83,906 88,384 93,220 100,660 107,345 112,046 115,421 125,057

70,074 71,302 70,987 70,324 72,823 71,335 74,148 78,096 82,249 88,681 95,400 98,840 103,332 111,741

78,345 77,796 77,411 77,731 78,715 77,813 84,628 83,823 85,320 92,311 90,543 94,729 95,270 99,458

a

Computed by adding Full-time MDs & DDs to one-half of the part-time MDs and DDs at hospital level. M.D.=Doctor of Medicine; D.D.=Doctor of Dentistry. Source: Peter Kralovec, Health Forum, American Hospital Association.

73

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

and 4% (27,600 29,700) of physicians overall (Harbuck, Follmer, Dill, & Erikson, 2012). This total figure corresponds with the reported 2009 membership (28,177) of the Society of Hospital Medicine (SHM); by 2011, membership had reached 34,411 physicians.11 Additional SHM data for 2000 2006 suggest that hospitals employed roughly one-third (33 34%) of all hospitalists, while physician groups employed another 14 24%. (The HealthLeaders survey similarly reported that 38% of hospitalists were hospital employees.) Applying this one-third statistic to the 2009 AAMC survey figures provides a point estimate of 9,200 9,900 hospitalists employed by hospitals in 2009. Between 2003 and 2009, the percentage of hospitals with a hospitalist group rose from 29% to 58%; the SHM now estimates that as many as 57% of hospitalists are now employed by IDNs.12

Prevalence of Employed/Owned Physician Groups Trend data (2003 2012) provided by MGMA, presented in Fig. 7, indicate a decline in the percentage of groups owned by physicians (from 83.2% to 72.5%), and a rise in the percentage owned by hospitals (from 8.3% to 13.6%), universities/medical schools (from 2.5% to 4.7%), and other

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Physician owned

Fig. 7.

Hospital owned

University or Insurance Co. Government Med.School or HMO owned

All other

MGMA-ACMPE Medical Groups by Ownership. Source: David Gans, Medical Group Management Association.

74

LAWTON ROBERT BURNS ET AL.

owners (from 4.1% to 7.4%). Trend data from the NAMCS, depicted in Fig. 8, show the changes in the distribution of practice ownership over the recent period (2008 2011). The major change has been the marked decline in physician ownership and the rise of hospital ownership (approaching 30% of practices in 2011).13 While the share of medical groups employed by non-physician firms (e.g., hospitals, payers, PPMs) in the MGMA database is small, their share of the MGMA member physicians has grown much more. As noted above, between 2003 and 2012 the percentage of hospital-owned groups increased from 8.3% to 13.6%; their share of MGMA physicians increased more substantially from 17.0% to 33.8%. This is because hospital-owned groups are much larger than the physician-owned groups. In fact, while the average sized physician-owned group in the MGMA database increased from 16.4 to 21.3 doctors (2003 2012), the average sized hospital-owned group nearly doubled from 64.3 to 120.6 physicians.14 The average size of university/medical school-owned groups rose from 258.0 to 406.4, although their share of MGMA doctors remained flat (20.7% in 2003, 19.3% in 2012). Physician groups owned by “other” firms increased from 6.1% to 19.3% of the total; their average size skyrocketed from 46.1 to 143.8 physicians. MD/MDGroup

HMO

Community Health Center

Hospital/Other

90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2008

2009

2010

2011

Fig. 8. Distribution of Office-Based Physicians by Practice Ownership. Source: National Ambulatory Medical Care Survey (CDC, 2007a, 2007b, 2008a, 2008b, 2010a, 2010b).

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

75

MGMA data further reveal that the increase in the right-hand tail of the size distribution of medical groups (groups with greater than 150 physicians) comes primarily from those formed by hospitals, universities/medical schools, and other firms. Between 2005 and 2012, there was a net gain of 236 groups with more than 150 physicians (from 199 to 435). As a percentage of all MGMA member groups, this represented an increase from 3% to 7% (hence, the use of “tail”). Of these 236 groups, 89 were hospital-owned, 95 were university/medical school-owned, 32 were other-owned, and only 31 were physician-owned. While some physician-owned groups experienced organic growth, they lost market share, particularly in the largest-size category.

Performance of Vertically Integrated Arrangements Between Hospitals and Physicians Empirical Evidence: Impact of Vertical Integration on Prices The effects of vertical integration on hospital prices are inconsistent. Cuellar and Gertler (2006) found that integration (via PHOs) was associated with an increase in prices for both indemnity and managed care patients; conversely, integration via IPAs and salaried models did not impact price. Ciliberto and Dranove (2006) found that integration of hospital physician arrangements was not associated with significant changes in hospital prices. These contradictory results may be due to their different settings; Cuellar and Gertler studied integration in Arizona, Florida, and Wisconsin between 1994 and 1998, while Ciliberto and Dranove used data from California between 1994 and 2001 (Gaynor, 2006). Berenson, Ginsburg, and Kemper (2010) note that physician hospital alliances fostered joint negotiations with payers and higher bargaining power in California. Recent antitrust actions undertaken by the FTC highlight some of the possible issues related to practice acquisition. Along with the Attorney General of Idaho, the FTC has challenged the acquisition of a large primary care group by St. Luke’s Health System in Boise (Federal Trade Commission, 2013). According to the FTC complaint, the acquisition would give the hospital a 60% share of the adult primary care market and lead to several negative downstream consequences. St. Luke’s Health System could induce the newly acquired physicians to shift their admissions and specialty referrals to the hospital system and away from competitors. It could, in turn, charge higher prices to commercial payers who, facing such a dominant provider, would be unable to find alternative providers

76

LAWTON ROBERT BURNS ET AL.

and thus be forced to pay quasi-monopoly prices. Such higher prices would ultimately be passed onto employers (in the form of higher premiums) and patients (in the form of higher co-pays for patient visits and ancillary tests), and might lead employers to reduce health insurance benefits. Conversely, the higher rates charged along with the higher payments the hospital would receive from Medicare for hospital-based ambulatory care would be passed on to the acquired physicians in the form of higher compensation. The FTC supported these allegations by citing the hospital’s history of multiple group acquisitions and the higher prices charged payers in the local market. The FTC also alleged that St. Luke’s documents admitted the employment strategy was designed to raise profits, not achieve lower costs. Moreover, by aligning, St. Luke’s Health System and the physicians could each seek significant rate increases from payers, knowing the latter must contract with one or the other. We should note that employment models are not alone in their anticompetitive conduct. Prior FTC investigations of PHO and IPA models also found market foreclosure and resulting higher prices charged to commercial payers (cf. FTC vs. Piedmont Health Alliance, FTC vs. Evanston Northwestern Healthcare Medical Group). Empirical Evidence: Impact on Hospital Cost and Quality Evidence regarding the impact of hospital physician integration on cost and quality likewise remains scattered and ambiguous (cf. Burns & Muller, 2008). Cuellar and Gertler found that IPA, PHO, and salaried employment models of integration all failed to lower hospital costs. Madison (2004) found that patients treated in hospitals with salaried employment of physicians received more intense treatment (i.e., higher procedure rates), resulting in higher expenditures. Nevertheless, the effects are small, the results are limited to heart attack patients, and the findings come from the early years of hospital physician integration. The Minnesota field investigations likewise found higher resource use and higher costs among vertically integrated groups (Kralewski et al., 2000; Kralewski et al., 2011). To the extent that physician employment/acquisition is driven by the desire to bill at higher rates in hospital outpatient settings, costs to Medicare are bound to rise. Given the 20% cost-sharing that Medicare patients bear, such arrangements also increase the prices to patients (Matthews, 2012). With regard to quality, Madison found that unlike IPA and PHO models, salaried employment models were associated with lower hospital mortality rates. The NSPO field investigation found that ownership by hospitals and plans was one feature that distinguished medical groups in the

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

77

top versus the bottom quartile of performance in terms of care management and health promotion. A series of NSPO studies found that group ownership was associated with activities in health promotion, chronic care management, EMR adoption, use of chronic disease registries and CMPs, and use of patient and physician reminders; the results were oftentimes statistically insignificant, however. Ownership was also associated with an increase in CMP use over time. On the other hand, some NSPO studies found no effect of vertical integration on clinical IT capabilities in the group or the presence of patient-centered medical home processes in large groups. The Minnesota field investigations found higher rates of inappropriate emergency department (ED) visits and avoidable hospitalizations in the acquired groups. Empirical Evidence: Impact on Productivity and Profitability Research on the effect of integration on physician productivity and hospital profitability has produced mixed results. On the positive side, Wan, Lin, and Ma (2002) found that the presence of a PHO, MSO, or IPA model was associated with greater hospital efficiency (cost per admission, occupancy). Similarly, Goes and Zhan (1995) found that greater levels of financial integration between hospitals and physicians were associated with lower hospital costs and higher occupancy, but might result in lower operating margins. Integration with a multihospital system was not associated with hospital cost or occupancy, but was positively related to operating margin. By contrast, the NSPO field investigation found that hospital ownership of physician groups was not associated with better group financial performance. On the negative side, Stensland and Stinson (2002) found that tighter forms of integration (i.e., where the hospital owns or manages the physician practice, similar to the MSO and employment models) in competitive markets resulted in decreased length of stay, inpatient admissions, and net income. In geographically isolated markets, integration decreased length of stay and increased admissions, with no impact on hospital profitability. In a similar vein, Burns, Gimm, and Nicholson (2005) found that hospitals that employed more than 140 physicians in the 1990s experienced significantly lower returns on total assets. Hospitals that invested $40 million or more in physician integration experienced significantly lower operating margins and return on assets. Hospitals that elected to acquire and salary physicians (as opposed to acquisition only, salary only, or no acquisition/salary) experienced lower total and operating margins and lower returns on assets between 1995 and 1999. Overall investments in integration (physician, hospital, health plan) and investments as a percentage of capital expenditures

78

LAWTON ROBERT BURNS ET AL.

were likewise significantly associated with declines in operating margins and returns on assets. During the 1990s, analysts commonly observed that hospitals lost an average of $100,000 per acquired physician per year (Advisory Board, 1999). Most of these physicians were PCPs who were offered guaranteed salaried contracts with no performance incentives. Adding some validity to these figures, the bankruptcy proceedings of a large hospital system in the late 1990s revealed that its physician division of 500 + employed doctors lost $50 million annually. More recent data suggest that hospitals’ operating losses on physician practices have widened well beyond inflation. In a survey of 189 hospital-sponsored multispecialty group practices, MGMA researchers found annual average losses of $189,910 per FTE physician in 2010 (Gans, 2012a); only 13 of the 189 reported that median total net income (excluding financial support per physician) was zero or positive. Conversely, physician-owned groups reported a total net income of $3,376 per FTE physician, after distributing most of the profit back to physicians as additional compensation (Gans, 2012a, 2012b). MGMA data from 2011 indicate median losses among hospital-owned groups of $174,430 per FTE physician, compared to a positive net income of $4,179 among physicianowned groups (Gans, 2012c). Why have hospital-sponsored groups incurred such large losses compared to physician-owned groups? The answer does not appear to be the different financial environments they face. Recent MGMA survey data indicate that physician-owned and hospital-owned groups identify (a) financial management as their greatest challenge and (b) the same financial issues as most challenging.15 These include dealing with rising operating costs, preparing for reimbursement models that place a greater share of financial risk on the practice, and managing finances with uncertain Medicare reimbursement. Different sized groups also cite similar financial challenges. Instead, the answer may lie in the structure of the two sets of group practices and their management. Published 2010 data from the MGMA Cost Survey (2011 Report) reveal the hospital-sponsored groups include a higher percentage of PCPs (58% vs. 51% in physician-owned groups) who generate less revenue than specialists. Such groups also have more Medicaid patients (12.2% vs. 7.4%), slightly more charity care (1.3% vs. 0.3%) and self-pay (4.1% vs. 3.2%) patients, and fewer commercial patients (50.6% vs. 55.5%). Crucially, physicians salaried by hospitals exhibit 16 29% lower productivity (RVUs per physician). Collection rates are negligibly lower for employed physicians. This suggests that the reduced

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

79

income incentives associated with salary guarantees as well as adverse selection of less productive physicians from the broader physician community and larger than normal numbers of young physicians just building their practices may be the culprits. Finally, the hospital-owned physician groups have much lower revenues from ancillary tests (lab, x-ray) and nonprocedural items (e.g., infused drugs, durable medical equipment) (Gans, 2012a, 2012c). This is because the hospital accounts for those revenues separately. Much of the economic impetus for salaried employment of physicians by hospitals has been to increase these so-called “ancillary revenues,” which generate a significant fraction of hospital profits. Unpublished 2011 data from the latest MGMA Cost Survey (2012 Report) provides additional insight into the performance differences between physician-owned and hospital-owned groups (Gans & Wolper, 2013). Researchers report that the top quartile of hospital-owned groups (in terms of overall performance) resemble the physician-owned groups in many aspects of their staffing and operations; both differ substantially with the bottom three quartiles of the hospital-owned groups. For example, the lower-performing groups employ fewer support staff (e.g., nurses) per FTE physician, provide less square footage of space per FTE physician, and have more branch clinics (with fewer physicians). Such differences in management and operational efficiency may explain the financial performance differences between hospital-owned and physician-owned groups. Many hospitals actually operate their “groups” as dispersed collections of solo and partnership practices where the only things that really change postacquisition are the nameplates on the door and the source of the physicians’ and office staff’s W-2s. Another explanation is the possible lack of due diligence by hospitals undertaking rapid practice acquisitions and information asymmetry between buyers and sellers. Some analysts argue that hospitals may lack transparency into the acquired group’s clinical and financial performance due to the lack of robust IT inside the practice. The hospital thus lacks information on physician productivity and billing, and thus cannot accurately forecast cash flows (Baldwin, 2012). Beyond due diligence, hospitals may overpay physicians to avoid losing them to competing hospitals. To secure the transaction, hospitals may also acquiesce to the physicians’ desire to remain in their current locations and retain their current staffing and systems. This is not meant to imply that all hospitals with employed physicians sustain this level of losses. There are several illustrations (that we are aware

80

LAWTON ROBERT BURNS ET AL.

of) of hospitals that have eliminated their losses from their PCP divisions. Intermountain Healthcare (IHC), for example, focused on process improvement to solve concrete business problems within its employed practices. It benchmarked accounts receivable, staffing levels, and physician productivity across all medical sites. It thereby increased collected co-pays and appointments kept, and reduced days in accounts receivable (from 80 to 34; Phil White, personal communication). IHC also ensured that hospital overhead was not transferred to the physician division, that hospital pay scales were not imposed on physician practices, and that the division could develop its own strong clinical leadership and business infrastructure to manage operations. Empirical Evidence: Impact on Group Culture and Hospital Alignment The University of Minnesota studies found that hospital/health plan ownership negatively impacted most dimensions of group culture (e.g., collegiality, organizational identity, trust, autonomy). Hospital ownership is correlated with the number of specialties in the groups studied, leading to some multicollinearity between vertical integration and multispecialty mix. Early research on physician alignment (e.g., trust, commitment) suggested limited and mixed benefits of vertical integration. On the one hand, Dukerich, Golden, and Shortell (2002) and Burns, Shortell, and Andersen (1998) reported that employment exerted an indirect effect on alignment, mediated by the perceived quality of the work relationship. On the other hand, researchers found that salaried roles failed to improve physician satisfaction or reduce physician hospital conflict (Burns, Andersen, & Shortell, 1990). A national study of physicians in eight integrated systems in the early 1990s compared the degree of physician hospital alignment between doctors on the voluntary medical staff, doctors involved in contracting alliances (PHOs, IPAs, MSOs), and doctors in salaried models. The alignment of salaried physicians was significantly higher, due to the large sample size of the study; differences between the three groups of physicians were small, however. Most physicians expressed tepid relationships with their hospitals, regardless of the arrangement (Burns, Alexander, Zuckerman, Andersen, & Torrens, 1995). These data suggest that the different governance models of vertical integration (e.g., salaried models), virtual integration (alliance models such as PHOs and IPAs), and reliance on the traditional medical staff exerted little impact on relationships. These results were subsequently validated in a study of 14 hospital systems and 61 hospital-sponsored medical groups. Burns et al. (2001) found

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

81

that membership in a PHO or IPA failed to increase the physician’s commitment to the hospital system. By contrast, the salaried model increased the physician’s commitment and identification with the system, as well as promoted greater citizenship behaviors. The impacts were statistically significant but small in magnitude, however. Alignment was not affected by the group’s size or specialty mix (Alexander et al., 2001a, 2001b). What explains the historical lack of alignment between physicians and hospitals repeatedly observed? One global reason is the historical tension that has existed between the hospitals and medical communities for nearly a century (Burns et al., 2010). Managers and professionals have profoundly different cultures and norms (Laufer, 2011; Shortell, 1991). Another explanation is the hospital’s typical focus on structures to integrate doctors while sometimes ignoring the hospital processes that physicians find most dissatisfying. As one example, a survey of physicians found that doctors perceive huge customer service gaps in three main areas: variable quality of physicians on the medical staff, adequacy of nursing staff, and efficient and timely scheduling of patients (VHA, 2003). As another example, Shortell (1991) reported the difficulties of developing trust to promote physician hospital relationships. A third reason is the presence of third parties (e.g., PPMs, medical device firms) that compete with the hospital for the physician’s attention, much in the manner of Georg Simmel’s notion of tertius gaudens: “the enjoying third” (Burns, Nash, & Wholey, 2007). Finally, some research suggests that past hospital efforts to control physicians have spawned a union mentality among the medical staff, whose elected leadership saw its primary role as collectively representing the interests of the medical staff to hospital administration (VHA, 2002). Most recently, Deloitte surveyed a random sample of 613 primary care and specialist physicians (Keckley, Coughlin, & Stanley, 2013). Regardless of their hospital relationships, physicians were pessimistic about the future of their profession, with 6 in 10 reporting that many of their colleagues will retire earlier than planned in the next 1 3 years. The three most satisfying attributes of their practices patient relationships, protecting the health of individuals, and intellectual stimulation do not seem to be emphasized in integrated models. Two attributes that might accompany such relationships leading a team of health professionals and administering a complex health care organization are ranked as the least satisfying. It is hard to imagine how hospitals might develop “alignment” among such a disillusioned group, given the historic lack of trust between the two parties (Burns et al., 2010; Laufer, 2011).

82

LAWTON ROBERT BURNS ET AL.

Three reports address the issue of changes in physician alignment that might follow from hospital integration efforts. Earlier surveys of Arizona physicians revealed an increase in physician hospital conflicts in most areas studied, particularly those concerning nursing, ancillary services, and equipment requests (Burns, Andersen, & Shortell, 1993). In a second study, researchers tracked the alignment of employed physicians with the Allina Health Care System between 1995 and 1997 (Bunderson, Lofstrom, & Van de Ven, 2000; Van de Ven, Rogers, Bechara, & Sun, 2008). The physicians’ commitment to the medical profession increased while their commitment to the hospital system decreased.16 In a third unpublished study, researchers from the Center for Organized Delivery Systems (CODS) and Center for Health Management Research (CHMR) compared their surveys of physicians at three IDNs conducted at two points in time (1995, 1998) using comparable instruments. The level of alignment fell at each IDN. The researchers repeated the analyses for those physicians who responded to both surveys; the results were identical.

Empirical Evidence: Impact on Clinical Integration Clinical integration requires structures and systems to coordinate patient care across people, functions, activities, and sites over time. Common activities include population health management, disease and demand management, electronic patient records, common patient identifiers and patient registries, CMPs, clinical service lines, continuous quality improvement, and information systems to track utilization by patient and provider. According to one review, economic integration between physicians and hospitals does not automatically lead to functioning clinical integration (Burns & Muller, 2008). One explanation may be that clinical integration infrastructure requires (a) substantial time and resource investments to develop and (b) a long time to realize positive gains from these investments. According to the NSPO study, between 2000 and 2006 larger physician groups (20 + doctors) increased their overall use of 17 different CMPs only slightly from 6.25 to 7.67 (Shortell et al., 2009). Ownership by a hospital or HMO was not associated with a scale of 19 clinical IT functions such as electronic registries and CPOE (Robinson et al., 2009). Ownership was associated with the use of CMPs (Shortell et al., 2009) as well as the presence of care coordination and quality/safety processes (Rittenhouse et al., 2011).

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

83

Summary: Physician Hospital Integration and Performance The evidence base raises major questions about the effectiveness of hospitals’ vertical integration strategies, particularly the employment model. Many hospitals have pursued this strategy, and incurred losses on the practices themselves, in order to grow their inpatient and outpatient volumes and revenues (e.g., through higher prices). Past antitrust actions undertaken by the FTC suggest hospitals may have had some success with this strategy. However, there is little evidence that integration actually improves the value of the health system’s “product,” for example, that these arrangements improve the quality or lower the cost of services jointly delivered. There is little evidence to date that integration satisfies another hospital objective: improved alignment between the two parties. Employed physicians express only slightly higher levels of alignment compared to those in strategic alliances and on the medical staff. In hospital administratorspeak, “alignment” is often a code word for control exerted through the employment relationship, not actual improvement in clinical service relationships. Many hospitals have used employment as a quick and easy way to develop closer working relationships with physicians which, in turn, might assist in care coordination efforts that, in turn, might help hospitals to achieve pay-for-performance and shared savings targets. This hypothesized causal chain of aspiration has too many untested and likely weak linkages to bear real fruit. From their perspective, physicians may have sought employment relationships as an escape from the business risk of independent practice: for example, to avoid capital expenses needed to comply with the HITECH Act of 2009 (and its meaningful use requirements), to surmount the difficulty in renewing their practices with younger colleagues, or to cope with continuing payment reductions or the broader uncertainties posed by health care reform. Integration via employment may have helped physicians gain economic security, better payer contracts, and perceived safety, but at the cost of their clinical autonomy and perhaps some practice attributes they find most satisfying (Keckley et al., 2013). By seeking salaried employment, tens of thousands of physicians have successfully shifted most of their business risk to hospitals. It remains to be seen whether hospitals will generate a sustainable return on their investment in physician practice. Over the longer term, both hospital and physician partners may find these relationships further strained by a double-whammy: continuing cuts to Medicare and Medicaid payments to providers (price-side effect) but also growing share of admissions accounted for by these public programs

84

LAWTON ROBERT BURNS ET AL.

(volume-side effect), which generate much lower operating margins. Indeed, between 2001 and 2011, the percentage of the non-elderly population covered by public programs rose from 15% to 22%, while the percentage covered by employers fell from 68% to 58% (Fronstin, 2012). Nationally, the proportion of national health expenditures paid by the two big public programs Medicare and Medicaid (state and federal combined) accounted for 35.6% by 2011. Given the declining real dollar payment levels from the two public programs and their lower payment-tocost ratios (compared to private insurance), hospitals may have difficulty financing their investments in physician practices. Hospitals may be faced with the prospect of unwinding some of these acquisitions, as they did in the late 1990s and early 2000s, or of dramatically restructuring the employment contracts with physicians to reduce operating losses. In these contract renewals, hospitals could be pressuring their employed physicians both to do more (see more patients per hour, admit more patients from the ED, refer more patients in-network) and to do less (reduce lengths of stays, reduce inpatient costs in bundled payment, and gain-sharing programs). Managing these often conflicting incentives and economic pressures will create yet new strains in the relationship. Under the pressures of health reform, many hospitals may transition from their former role as “physician’s workshop” to a new role as “physician’s sweatshop” (Burns et al., 2010; Creswell & Abelson, 2012; Laufer, 2011).

Integration by Equity Capital: Single-Specialty Networks Hospitals are not the only economic actor that has sought to acquire and consolidate physician practices. During much of the 1990s, hospitals competed with the investor-owned physician practice management (PPM) firms for such acquisitions. After the collapse of the PPM industry at the end of the 1990s, equity-based physician enterprises fell off the radar screen. However, formidable clinical enterprises, some publicly traded and others privately held, have since arisen to dominate certain singlespecialty markets. These companies have created administrative support and contracting infrastructure around large single-specialty groups with franchises in local markets, and contract exclusively with hospitals to provide specialty coverage in their respective disciplines. Consultants suggest that acquisitions of specialists and private equity investments likely doubled between 2008 and 2012 (Kitchell & Hurst, 2011).

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

85

Anesthesiology Such investments feed on the growing supply challenges and discontentment among several specialties. Anesthesiologists, for example, have suffered declining payment from Medicare, competition from substitutes (certified registered nurse anesthetists, or CRNAs), lack of evidence regarding superior outcomes of anesthesiologist-directed care, the threat of commodification of anesthesia services, pushback from hospitals on their requests for practice support, targeting of anesthesiologists as the deep pockets in malpractice cases and increased liability costs, medication shortages, and difficulties in demonstrating the value-added of their services. In response, anesthesiologists have joined one of several equity-backed firms consolidating the specialty. Nearly all of them rely on the scale economies argument to justify their strategy. One of the most successful is Sheridan Healthcare, backed by the private equity firm Hellman and Friedman, which generated $140 million in operating earnings in 2012. Starting with anesthesiology, Sheridan has diversified into three other hospital-based specialties (emergency medicine, neonatology, radiology) and now employs 1,600 + physicians at 130 sites in 20 states. Another successful anesthesiology company is Pinnacle Partners in Medicine. Pinnacle offers a network of 770 anesthesia providers and provides coverage at 130 hospitals, ambulatory surgery centers, and medical centers. Pinnacle aspires to be a national physician-owned organization of hospital-based practitioners. Physician members retain local management and control, and enjoy an equity stake in the firm. Pinnacle provides its physicians with back-office functions, large scale to help with contract negotiations with payers, and presumed scale economies. Another major hospital-based physician company is publicly traded Mednax/Pediatrix. The Pediatrix Medical Group was founded in 1979 as a single entrepreneurial Florida-based neonatology practice. It has employed a similar, exclusive contracting model to operate the neonatal ICUs (NICUs) in over 300 hospitals nationally, and diversified into maternal and fetal medicine by creating a multi-site obstetrical group. In 2009, it merged with a large national anesthesia group similar in structure to Sheridan, and is publicly traded as MedNax with a market capitalization of $4.4 billion. Mednax employs over 1,675 specialists in neonatology (968 physicians), anesthesiology (308 physicians), maternal/fetal medicine (172 physicians), and pediatric cardiology (104 physicians), as well as 600 nurse practitioners. MedNax hopes to differentiate itself from other PPMs by participating in clinical trials, developing its own EMR and preparing publications from its own clinical information system, developing into a “patient safety

86

LAWTON ROBERT BURNS ET AL.

organization,” and developing relationships with both hospital and surgeon customers. Oncology The field of medical oncology also contains a dominant, investor-financed consolidator. US Oncology was formed through the 1999 merger of two large PPMs with multi-site oncology practices: American Oncology Resources (AOR) and Physicians Reliance Network. US Oncology functions as a classic PPM, operating and providing management services for 83 comprehensive cancer centers, and a national network of independent oncology practices composed of 1,300 physicians. The principal source of profits is physician-directed chemotherapy for cancer care, which accounts for roughly two-thirds of its revenues. US Oncology was taken private by Welsh Carson, a private equity firm in 2006 for $1.6 billion. In 2012, having grown to $3.5 billion in revenues, it was acquired by the diversified drug distribution and IT firm McKesson for $2.2 billion. Hospital Medicine The fastest growing clinical specialty of the past 15 years has been hospital medicine. This discipline has also seen the emergence of large corporate actors consolidating the field. Nationally, there are now over 34,000 hospitalists, compared to less than 10,000 in 1995, at least one-third of whom are actually employed by hospitals. The largest consolidator in the hospitalist specialty is IPC The Hospitalist Company, which employs over 1,250 full-time hospitalists. IPC was founded in 1995, and operates 180 medical groups which staff 350 hospitals around the clock, as well as 550 post-acute facilities. IPC is publicly traded on the NASDAQ and has a market capitalization of $803 million (as of May 13, 2013); its share price has risen 75% since 2010 ($24.69 on July 26, 2010; $43.28 on February 21, 2013). Other major actors in this space include Cogent, a privately held company that provides both hospitalist and intensivist staffing services through both Cogent-employed and hospital-employed physicians. Cogent claims to have 1,000 clinicians practicing in over 100 health facilities. These large entities capitalize on the difficulties that hospitals and small hospital systems have in recruiting and retaining physicians to staff their 24/7 operations. While the groups are independent of hospitals corporately, they could not exist without hospital contracts. Whether through salaried physician groups or contracted physicians, the investor-owned companies are able to recruit from national specialty markets and offer attractive pay and benefits packages. They offer hospitals proprietary care management

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

87

tools and stability and consistency in staffing, in exchange for significant mark-ups on the professional time and services of their physicians.

Vertical Integration by Health Insurers History Another form of vertical integration occurs between health insurers and physicians. One of the oldest and largest insurers in the US is the $50 billion group Kaiser Health Plan, which integrates physician care with hospital and other medical services for its 9 million members inside an insurance product envelope. Physician care is provided by the enormous multispecialty Permanente Medical Groups, whose leaders effectively control the Plan in each of its regions. Despite this 70 year long success story, however, most competing health plans neither employ nor are tightly linked to physician practices in the same way as Kaiser. During the 1980s, several large insurers (e.g., PruCare, Aetna, Cigna, FHP) pursued a staff model physician employment strategy within their HMOs. Yet another health insurer, Humana, sprang fully blown from a successful hospital chain. Neither of these strategies worked well, as these insurers divested their physician and hospital assets during the 1990s; Cigna was an exception, retaining its Cigna Medical Group in the Phoenix market. Insurers encountered problems in matching enrollees with their staff model physicians geographically. Former indemnity insurers also had limited experience in managing physician practices and found that enrollees were more loyal to the doctors than the health plans. During the 1990s, HMO-style plans shifted their physician strategy away from employment (staff model) toward more arms-length, contractual relationships (IPA and group models). Several staff model plans divested those groups during the past 20 years. Group Health Co-operative of Puget Sound spun off its staff model physicians into a Permanente-like medical group in the late 1990s, while Harvard Community Health Plan spun off its staff model physicians into Harvard Vanguard Medical Associates, now part of the Atrius Medical Group, the largest multispecialty medical group in New England. In both cases, the stated purpose of the separation was to encourage the physicians to become a self-governing enterprise with its own P + L and significant operating autonomy. Many physician clinics divested by other insurers, such as PruCare, Cigna, and Aetna, were subsequently acquired by the PPM firms that were ascendant in the early 1990s and went bankrupt by the end of the decade.

88

LAWTON ROBERT BURNS ET AL.

The HMO backlash of the late 1990s doomed many provider and insurer efforts to diversify into one another’s domain. Since 1996, the managed care marketplace transitioned from primarily HMO models to PPO and point-of-service (POS) plans that relied on broad provider networks. To counter consumer resistance to the narrow networks found in HMOs, some integrated health plans focused on quality competition and improving patient outcomes, with a particular focus on preventive care. Ho (2009) provides evidence that Kaiser plans have a quality advantage (based on HEDIS scores) compared to local non-integrated plans in areas where Kaiser has expanded successfully. In addition, local competitor plans increased their quality as a result of Kaiser entry. Other physician group and IPAsponsored health plans have taken a similar approach, and a disproportionate number of the health plans with 4.5 or 5.0 Star Medicare ratings are indeed physician sponsored or controlled (IPA or group models). Even so, successful integrated plans like Kaiser have had difficulty expanding beyond their core markets. Kaiser attempted to enter seven new markets since 1980; by 2001, it had exited four of these (Ho, 2009). Physician insurer integration was hindered by physician hostility to prepaid group practice, the absence of two of the three components of the traditional Kaiser model (dedicated medical group and owned hospital) in new markets, the difficulty in ramping up patient enrollment in the short term in order to compete, and, perhaps most critically, employers’ preference to contract with a single insurer offering a menu of health plans (Gitterman, Weiner, Domino, McKethan, & Enthoven, 2003; Ho, 2009). Kaiser’s plans were also not demonstrably cheaper than the less integrated plans against which they competed, providing no leverage to grow enrollment. In addition, requiring enrollees to switch to a restricted network of physicians proved a major constraint (Ho, 2009). Current Payer-Led Integration Efforts In the past few years, in response to health reform, insurers such as Humana, WellPoint, and United Healthcare/Optum have purchased medical groups in efforts to cut costs by managing patient care and physician networks (especially specialist care) more tightly (Weaver, 2011). In December 2010, Humana paid $790 million to acquire Concentra, a chain of 300 urgent care and occupational medicine clinics in 42 states; in 2011, Humana followed up with additional acquisitions of primary care and occupational medicine clinics (Vesely, 2012). Humana also acquired a home health provider (SeniorBridge) and its 1,500 care managers for $72 million in 2011 as well as a chain of urgent care centers (NextCare) in 2011.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

89

In August 2011, WellPoint paid $800 million to acquire the 26 clinics of CareMore in California, Nevada, and Arizona providing care for 54,000 patients, as well as its Medicare Advantage health plan. In Pennsylvania, Highmark (the state’s largest Blue Cross/Blue Shield plan) acquired not only the West Penn Allegheny Health System but also several large physician practices (Premier Medical Associates, Triangle Urological Group). Unlike the approach taken in hospital acquisitions of physician practices, Highmark reportedly will allow its employed physicians to have control over ancillary care and patient referrals. These acquisitions were intended to protect Highmark members from the possible loss of a contract with the University of Pittsburgh Medical Center, which fields a competing health plan. In August 2011, United Healthcare’s OptumHealth subsidiary acquired Monarch HealthCare (purchase price not disclosed), a multispecialty group practice comprising 2,300 physicians. The Monarch’s acquisition precipitated an angry reaction from some of Monarch’s other health insurance partners. Following the Monarch deal, Blue Shield of California (BSC) asked arbitrators to award it millions of dollars from Monarch in damages, claiming the group induced its patients to disenroll from BSC and join another insurance plan; Anthem Blue Cross also withdrew from a pilot ACO with Monarch. It was lost on many people that Optum and United Healthcare’s insurance businesses are separately managed. The Collaborative Care division of OptumHealth (which is the controlling owner of the practices) has established multiple integrated delivery models to contain hospital re-admissions, reduce unnecessary admissions, and substitute outpatient for inpatient care.17 Collaborative Care has developed physician networks beyond those it operates directly to work with multiple payers, particularly Medicare Advantage plans. One network, Lifeprint, serves Medicare Advantage private plans in Arizona. Two other networks, Evercare and Inspiris, take risk for the continuum of post-acute care by focusing on nursing home and home health patients, respectively. Finally, in early 2013, Optum developed a partnership with Cornerstone Health Care, a large multispecialty group with 360 physicians in North Carolina, to create an ACO. As of 2013, Optum had developed a network of 425 “affiliated” (e.g., employed) physicians (up from 350 in 2011) and 300 nurse practitioners and physician assistants in 90 primary care and urgent care clinics. Optum’s network also includes an additional 4,500 “contracted” physicians, up from 1,500 in 2011, which Optum developed following its $1 billion acquisition of WellMed Medical Management Company which provides management

90

LAWTON ROBERT BURNS ET AL.

services to 4,000 physicians in Texas and Florida (Pricco, 2013). This twofold network model resembles a staff model organization with a wraparound IPA. The network engages in risk contracts with payers, provides data to the physicians to help them with population health management, and offers higher physician compensation to handle care coordination tasks (Stapleton, 2012). One exception to the growth trend has been Cigna. By 2011, Cigna had expanded its Phoenix Medical Group operations to 32 locations in the greater metropolitan area and employed over 131 physicians. In contrast, by early 2012, it had cut staffing by 100, including 15 of its physicians, and closed down several of the clinics and specialty services. The insurer reportedly retrenched, offering primary and preventive care services through its captive group, and contracting out for specialty care, in an effort to cut its overhead costs.

Rationale for Payer-Led Integration There are several rationales behind these acquisitions. First, insurers are positioning themselves for increased Medicare Advantage enrollment, which has been surging and will increase substantially with the retirement of the baby boomers, as well as for increased Medicaid enrollment following PPACA implementation in 2014. These two public programs represent the two biggest growth markets for insurers. The increased number of enrollees may be challenged to find PCPs to treat them, given the nationwide shortage and falling physician participation in both public programs but particularly Medicaid. Physician practice acquisitions may give the insurers an edge in attracting and retaining enrollees. Second, they are developing networks to help manage the care of the sickest patients such as the chronically ill, the dual eligibles, and those with pre-existing conditions which are the target of several initiatives in the PPACA. Several of these organizations (CareMore, Monarch) have extensive care management expertise geared to high-risk populations (Main & Slywotzky, 2011), and some (e.g., Monarch) have already been selected by the Centers for Medicare and Medicaid Services (CMS) to participate in the Pioneer ACO program. WellPoint may believe it can learn from and export the best practices of the CareMore model to its other markets; other ACOs in the US have tapped the CareMore model. Insurers’ seeking out the sickest patients is a profound departure from the traditional risk-underwriting/avoidance strategies pursued prior to the enactment of health reform.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

91

Third, some insurers believe that the only way to manage risk contracts and satisfy the dictates of value-based contracting is by owning the front end of (ambulatory) care and incentivizing their employed physicians to treat enrollees cost-effectively (Weaver, 2011). Insurers who have been engaged in data analytics, population health management, and disease management for several years may feel that they have an edge over hospital-based networks in delivering on these goals. Fourth, some insurance executives believe their physician networks will better coordinate information and improve efficiency (Kirk Stapleton, personal communication). Physician employment inside risk-contracting networks can counteract the fragmentation inherent in fee-for-service and improve compliance with clinical protocols. Health plans intend to aggregate data tied to specific patients from their claims systems and present it to the PCP to help them manage the patient end to end. The system also improves measurable performance on quality measures and reduces practice variability, which not only increases efficiency but also positions the provider network to secure bonus payments based on quality metrics or potential shared savings. Finally, insurers may be threatened by hospital efforts to develop captive physician networks and ACOs which might have as their real goal limiting insurer contracting options and increasing the prices charged them. Insurers may be vertically integrating back into the physician market to develop countervailing power and/or avoid being locked out (Terry, 2012).

Summary The above review suggests that vertical integration potentially benefits both hospitals and physicians. Physicians gain income security and sometimes increased income and a more satisfying lifestyle as well by shifting their economic risk onto hospital employers. Hospital employment of physicians may provide them market leverage over payers, which in turn generates higher payment rates and possibly increased profits. Recent evidence suggests hospitals invest these profits in technology to attract patients and physicians and to pursue more physician acquisitions. There is thus a positive relationship between integration, market power, and technology (Lake et al., 2003). What the data do not tell us at this point is the level of return on the hospital investment in vertical integration. On the other hand, there is limited evidence for societal benefits of physician hospital integration. Studies conducted over the past two

92

LAWTON ROBERT BURNS ET AL.

decades have investigated multiple outcomes of integration and found few positive results. There are concerns that such integration may lead to higher health costs (Berenson et al., 2010). Evidence on the effects of physician practice acquisition by equity investors and payers is limited; prior efforts in the 1990s failed, while recent efforts are too new to evaluate. What is clear, however, is that these financial partners have continued to invest heavily and built large provider networks.

FAVORABLE RESULTS FROM HORIZONTAL AND VERTICAL INTEGRATION: SOME EXCEPTIONS Group Employment Models This is not to say that large horizontally and vertically integrated physician groups are never successful, however. Prominent multispecialty physician groups such as Kaiser Permanente, Geisinger Clinic, Mayo Clinic, Cleveland Clinic, and others have created high performing organizations and high levels of integration among their physician and hospital administrative personnel. In surveys of two of these institutions, the lead author found unified clinical and administrative cultures. Using an instrument developed earlier (Beach, 1992), physician and administrative personnel were asked to allocate 100 points to a series of enunciated corporate values posted on the group’s website in order to describe the group’s culture. The value rankings of the medical and administrative hierarchies within each institution were nearly identical. What explains why these systems are so tightly aligned and apparently effective in their markets? One obvious reason is that they have enjoyed a long history and sufficient time to develop such cultures. Table 3 lists the founding dates of many of these systems. They were typically founded as large multispecialty groups that became the core of their current IDNs, which in turn became physician centric, physician led, and physician dominated. This suggests that searching for empirical evidence of the positive effects of integration in the more recently formed IDNs may be premature. It may take decades for the potential organizational and societal benefits to be realized. It is important, however, to point out that in all these mature integrated organizations, the physician group was the core. Hospitals were not (and are not today) major stakeholder and decision-makers in these groups;

93

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

Table 3.

Group Employment Models.

Group Employment Model Organization Bassett Healthcare Billings Clinic Cleveland Clinic Geisinger Clinic Gunderson Clinic Guthrie Clinic Henry Ford Kaiser Lahey Clinic Marshfield Clinic Mayo Clinic Palo Alto Medical Clinic Permanente Medical Group Scott & White Clinic

Location Cooperstown (NY) Billings (MT) Cleveland (OH) Danville (PA) La Crosse (WI) Sayre (PA) Detroit (MI) Mohave Desert (CA) Burlington (MA) Marshfield (WI) Rochester (MN) Palo Alto (CA) Northern California Temple (TX)

Startup 1927 1911 1921 1915 1891 1910 1915 1932 1939 1916 1880 1930 1945 1897

Source: Minott, Helms, Luft, Guterman, and Weil (2010).

indeed, some like Marshfield do not even own a hospital. Instead, the hospital is a subsidiary of the physician enterprise and, in fee-for-service models and environments, functions as a capital accumulation device that retains earnings (for future investment) and protects surplus capital from being distributed back to physicians at year’s end.18 The key strategic question is whether in the newer integrated delivery enterprises, the sponsoring hospitals will devolve authority and legitimacy to the physician groups they developed, and subordinate the hospitals’ role to the professional values and objectives of their physicians. It is not appropriate simply to assume that these hospital founded IDNs will evolve seamlessly over time into Kaiser or Geisinger type enterprises, which are truly physician directed. Many of these systems also developed in rural areas of Wisconsin, Illinois, Pennsylvania, and Texas, where they established dominant market shares and created entry barriers to other clinics as well as hospitals. Many also formed their own health plans, especially in the 1970s and 1980s. Historically, these systems have enjoyed stable and homogeneous leadership. Finally, there is strong economic interdependence among the three arms of their IDNs: physician group, hospital, and health plan. The dominant physician group plays a major role in the hospital’s admissions; the health plan’s enrollees serve as the central supply of business to both physicians and the hospital. These economic ties are also often supplemented by (a) a

94

LAWTON ROBERT BURNS ET AL.

high degree of shared risk, reward, and ownership among the three entities, (b) overlapping boards among the three entities, (c) overlapping medical leadership among the three entities, (d) a common culture (e.g., academic, faith mission), and (e) presentation of a uniform face to the customer and community (Burns, 1999). There have been at least four studies of employment models that highlight their cost and/or quality advantages. One analysis compared the levels of resource utilization across five “systems of care” combining practice setting, specialty mix, and payment method: solo/single-specialty groups (fee for service), multispecialty groups (fee for service), solo/single-specialty groups (prepaid), multispecialty groups (prepaid), and HMO (Greenfield et al., 1992). The results indicated that prepaid multispecialty groups achieved the lowest hospitalization rates and diagnostic costs per visit; conversely, they had the highest rates of office visits per patient day. The researchers attributed the advantages of such groups to their efficiency in out-of-hospital patient management, including off-hours physician call coverage. However, such groups typically received the lowest patient ratings of the quality of outpatient visits: overall care, technical care, personal care, and office waiting time (Rubin et al., 1993). Newhouse (1973) made the same observation four decades ago. A group of Harvard researchers compared the quality of primary care rendered by physicians in integrated medical groups versus those in IPAs and in hybrids of the two (Mehrotra, Epstein, & Rosenthal, 2006). Quality measures reflected the quality improvement strategies utilized (e.g., data collection, use of practice guidelines, contact of patients who missed screening), as well as the percentage of patients receiving appropriate screening exams and preventive care. They found a gradient in quality that was highest in the medical groups and lowest in the IPAs. The differences in screening and receipt of preventive care persisted even after controlling for the presence of an electronic medical record and the use of quality improvement strategies. The researchers concluded that physician group models influence quality. Researchers at the University of California Berkeley similarly compared the level of adoption of clinical IT in organized medical groups with IPAs as part of the NSPO (Robinson et al., 2009). They found significantly higher adoption levels on most measures in groups compared to IPAs. As part of the same project, researchers also found higher levels of adoption of other CMPs. Finally, the Medicare Physician Group Practice Demonstration analyzed the ability of 10 physician groups (many of them group employment

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

95

models) to simultaneously achieve quality targets and lower cost growth in order to earn bonus payments from CMS. While most groups achieved their quality targets, few earned the bonuses associated with lower cost growth. The two best performing entities were the Marshfield Clinic (which owns no hospital) and the University of Michigan’s faculty practice plan (where one could plausibly argue that the hospital works for them, rather than the other way around) both operating in high-utilization/high-cost Medicare markets. Annual savings per Medicare beneficiary were modest at best, and there was great variation in the savings (and losses) across the 10 groups. Most savings occurred in the treatment of dual-eligible patients: high-risk/high-cost Medicare patients too poor to pay their own deductibles and hence enrolled in Medicaid as well (Haywood & Kosel, 2011; Colla et al., 2012).

VIRTUAL INTEGRATION INTO PHYSICIAN NETWORKS The preceding section finds that the integrated groups often have higher levels of performance on utilization metrics or quality scores compared to IPAs. This is not meant to diminish the accomplishments of the latter, however. IPAs offer many advantages over the freestanding solo practices that they aggregate into physician networks. Historical Development In addition to forming actual groups of co-located practitioners, physicians have also formed virtual groups among solo and small practices. Independent practitioner associations (IPAs) are collectivities of doctors that serve as managed care contracting vehicles.19 IPAs developed in the mid-twentieth century as a competitive reaction of mainstream medicine to the growing threat posed by the nascent prepaid medical groups. One of the very first prepaid medical plans in the US developed in Clackamas County in the State of Oregon in the 1920s around an independent practice association. The California Medical Association encouraged its members to join IPA models of prepaid practice to compete with the Kaiser Permanente Medical Group model (Shouldice & Shouldice, 1978). IPAs were also often established by county medical societies and called Foundations for Medical Care (Starr, 1982). Research documents significant new development of IPAs beginning during the 1970s, reaching

96

LAWTON ROBERT BURNS ET AL.

1,500 in 1990, and as many as 4,000 in 1996, with an average of 300 physicians each (Haas-Wilson & Gaynor, 1998b). Two publications from the NSPO project analyze a sample of 347 366 IPAs in 2003. Roughly one-fifth of these IPAs were owned by a hospital and/or an HMO, with higher levels of such ownership in California (Casalino et al., 2003; Gillies et al., 2003). Based on the CTS, the number of IPAs may actually have declined, evidenced by a drop in the prevalence of hospital-affiliated IPAs from 33% to 20% between 1994 and 2000 (Lake et al., 2003). IPA Benefits IPAs are loosely integrated networks of independent physicians and physician groups that organize to accept some form of insurance risk from health plans either professional capitation or full capitation. They provide two potential advantages to physicians: the opportunity to collectively contract with health plans, and an opportunity to gain some of the advantages of group practice shared services while maintaining physician autonomy. This form of “virtual” (contractual) integration has the potential to benefit from centralized administration, risk spreading, and leverage with health plans, while being easier to establish than integrated systems or large multispecialty groups due to lower start-up costs (e.g., physicians remain in their own offices).20 IPAs can also achieve cost and quality improvements in quality through enhanced monitoring, utilization review, and case management, as well as physician incentives to remain productive (Casalino & Robinson, 2003; Haas-Wilson & Gaynor, 1998b; Penner, 1997). The popularity of IPAs reflects their accommodation to the revealed preferences of employers for broad networks and to physicians’ desire for independence. Because physicians retain their independence, however, IPAs are likely to lack the culture and organizational loyalty that characterize successful integrated systems such as Kaiser or Group Health Co-operative, which, could, in turn, limit their effectiveness (Center for Studying Health System Change, 1999). IPA Structure and Performance Physician Selection and Practice Organization Differences in the organization of medical practice found in IPA, prepaid medical group, and staff models have been documented by many researchers

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

97

over time (Greenfield et al., 1992; Rosenthal, Frank, Buchanan, & Epstein, 2001; Wholey & Burns, 1993; Wolinsky & Marder, 1985). Wolinsky and Marder used the AMA’s Periodic Survey of Physicians to study active physicians in different office settings (e.g., prepaid group, fee-for-service group, IPA, solo practice). They found that physicians self-selected into different settings based on a host of criteria. Compared to IPA and (especially) solo doctors, prepaid group physicians (especially those from Kaiser) were more likely to report involvement in “the business side of medical practice” as important in their practice choice; prepaid group and fee-for-service group physicians were also more likely to mention “predictability of practice schedule” in their choice of settings, compared to IPA and solo doctors. Conversely, IPA and solo doctors were more likely to mention “personal autonomy” as important in their choice. They also found that such preferences shaped their practice patterns. For example, doctors who preferred predictable schedules worked fewer hours per week and devoted more time to scheduling routine visits; those who preferred personal autonomy worked more hours per week and devoted less time to scheduling routine visits. Wolinsky and Marder also investigated the direct effect of the physician’s practice setting on practice patterns. Patients of physicians in prepaid group settings spent more time waiting to schedule routine office visits, compared with patients seen by fee-for-service solo and group physicians; patients of IPA physicians were in between. The opposite pattern held for the length of patient office visits: fee-for-service solo and group doctors met the longest, while prepaid group doctors met the shortest, with IPA physicians in between. Despite these differences, the various practice settings were equivalent in the number of hours worked weekly by their physicians. There were, however, significant differences in utilization. Fee-for-service solo and group doctors had the highest professional expenses, closely followed by the IPA physicians; prepaid group physicians had much lower expenses. Prepaid group physicians also substituted more office visits for hospital visits, compared to other doctors. IPA physicians resembled fee-for-service group physicians in their utilization patterns. In a similar vein, researchers from the Medical Outcomes Study (MOS) compared five systems of care in the late 1980s (described earlier) (Tarlov et al., 1989). One analysis found that IPAs had higher hospitalization rates than prepaid medical groups, but rates lower than other practice settings (Greenfield et al., 1992). IPAs also exhibited lower office visits per patient day, suggesting they were less capable of substituting outpatient for

98

LAWTON ROBERT BURNS ET AL.

inpatient care relative to medical groups. Alternatively, they may have been less concerned about resource consumption and more concerned about their lifestyles. IPAs also exhibited the lowest prescribing rates per patient, and incurred higher costs of tests per visit than the prepaid groups. On the other hand, compared to medical groups, IPAs typically received higher ratings by patients in terms of the quality of outpatient visits (Rubin et al., 1993). Finally, Safran, Tarlov, and Rogers (1994) showed that IPAs outperformed prepaid groups in terms of continuity and comprehensiveness of care (but not in terms of coordination). They found few differences between IPAs and prepaid medical groups in terms of provider’s interpersonal accountability, treatment manner, or technical skill. Wholey and Burns (1993) utilized a different survey database on the four types of prepaid HMO settings (group, staff, network, and IPA) and the physicians within them. Consistent with Wolinsky and Marder, they found that prepaid group models had lower levels of utilization (admissions, inpatient days) than did the IPA models. Prepaid group physicians also had the highest professional satisfaction and the lowest rate of patient complaints. Physician survey data revealed that the IPA doctors had the least interaction with peers and the medical director, and the lowest involvement in decision-making. Cost and Quality Outcomes The evidence on IPA performance is mixed. Friedberg et al. (2007) find that networks of physician groups produce higher quality care than nonaffiliated practices. Similarly, Mehrotra et al. (2006) show that integrated medical groups provide higher quality care than IPAs, while the NSPO studies found that IPAs had lower levels of CMP use and less (more) developed IT capabilities for clinical (administrative) tasks. IPAs also scored lower than groups on activities dealing with health promotion, chronic care model implementation, use of health risk appraisals, use of chronic disease registries, and use of patient and physician reminders; many of these differences were not statistically significant. Finally, the NSPO studies found lower levels of clinical performance among the IPAs. With regard to cost containment, IPAs may be disadvantaged relative to organized groups in managing capitated risk, particularly on the inpatient side where the single largest expenditures are incurred. To be fair, some IPAs may be unable to contract with hospitals on a capitated basis, and thus are restricted to capitation of professional services. Rosenthal et al. (2001) found that organized medical groups are more likely to have affiliations with hospitals or medical foundations (44% vs. 10%), to have a

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

99

primary or preferred hospital (96% vs. 85%), and to have a smaller number of preferred hospitals (two vs. three) where patients are channeled.

From “Either/Or” to “Both/And” As Collins and Poras (1994) long ago pointed out, competitive advantage often comes from combining features that are seemingly in conflict, requiring executives to shift mindsets from either/or to both/and. Despite their historical competition, IPAs are not antithetical to medical groups. During the past few decades, medical groups have often used IPA networks as “wraparounds” for several reasons: to provide additional scale (i.e., more physicians), to provide increased geographic coverage (i.e., more delivery sites), to provide greater flexibility to physician coverage, to channel increased patient volume through a limited number of specialists (and thereby achieve greater coordination, lower transaction costs, and increased group bargaining), and to open new referral channels to the core medical group (Rosenthal et al., 2001). IPA wraparounds may also enable medical groups to keep more patient referrals “in network” and thus promote cost containment in a likely shift to capitated (risk) contracting. In their study of California, Rosenthal et al. (2001) found that wraparound models devote only 10% of professional spending to out-of-group referrals, compared to 18% in stand-alone medical groups. The IPA component absorbed 43% of the total revenues flowing into the core medical group. Large integrated group practices like HealthCare Partners have successfully employed wraparound IPAs to strengthen their bargaining position with payers, in both established and new markets, suggesting they have scale economies based on leverage. This hybrid model makes sense given the preponderance of IPAs in some large markets, and thus their availability as contracting partners to (relatively) smaller organized groups. Two studies have found that IPAs outnumber medical groups in the State of California (Gillies et al., 2003; Rosenthal et al., 2001). Of the 153 physician organizations surveyed, Rosenthal et al. found that 53% were IPAs, 16% were medical groups, and 31% were hybrids. The IPA and hybrid model they studied were also much larger in size than the core medical groups in mean number of physicians (364 vs. 379 vs. 209), median size (236 vs. 203 vs. 93), and percentage of PCPs (30% vs. 20% vs. 27%). The wraparound model is also consistent with the dominance of IPA and network model HMOs, while staff models have decreased in number.

100

LAWTON ROBERT BURNS ET AL.

Illustration of the IPA’s Advantage The nation’s largest IPA, Hill Physicians Medical Group (HPMG), has been successful in improving patient outcomes and practice efficiency while creating market leverage for their physician membership. The IPA includes over 2,000 physicians in both the San Francisco and Sacramento metropolitan areas, encompassing practices of varying sizes, but predominantly solo and partnership practices. An active physician board oversees HPMG in partnership with a management company that provides administrative coordination, management services, technology infrastructure, claims processing, and negotiation and utilization management for payer contracts. Hill’s physician board has led efforts to improve electronic health records and other IT and put in place innovative care management programs incorporating predictive modeling to identify at-risk patients. Hill also coordinates post-hospital follow-up by nursing case managers to prevent repeat emergency room visits. HPMG uses financial incentives (pay-for-performance) to reward physicians for resource utilization, clinical performance, and participation in the IPA’s initiatives. HPMG’s experience suggests that key factors in making the IPA successful included gaining physician trust by demonstrating concern about physician satisfaction and promoting communication and offering upfront monetary support to establish systems such as IT (Emswiler & Nichols, 2009). HPMG has also succeeded in obtaining higher payment rates from payers, which it has used to reward its physicians and maintain their loyalty.

CONCLUSION: A TALE OF TWO TAILS Looking back to the prior review of physician organizations in this volume (Burns & Wholey, 2000), we can now assess what has remained the same, what has changed, and how the earlier conclusions need to be updated. Three similarities exist between now and then. First, the physician sector still remains the least consolidated portion of the health care value chain; the vast majority of doctors continue to practice in small groups and solo settings. Second, physicians still confront a variety of options for consolidation, including horizontal consolidation with one another and vertical integration with other partners. Third, there continues to be an extremely thin evidentiary basis for recommending any particular approach.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

101

Some major changes are also evident. First, amid signs of stability among the mass of physicians, physician markets are in flux. At the lower end of the distribution of group practice size, there is a long fat tail (composed of a large number of very small groups) that has undergone only modest change. At the other end of the distribution, there is a short skinny tail (composed of a small number of very large groups) that has undergone significant recent growth. The percentage distributions of physician practice size mask four important features of this tail. First, there are a small number of very large groups. Second, the size of these groups is growing more rapidly than other groups, due primarily to hospital sponsorship. Third, there is an increasing number of these large groups, reflecting the growth in the total number of group practices in the last few years. Fourth, these groups have grown through the investment by non-physician owners (including private equity firms, health insurers, and others), in contrast to the slower organic growth of the physician owned groups via horizontal integration. This implies that the future shape of this tail may remain dependent on external (e.g., nonphysician services) subsidy flows both for growth and continuing operation. What can we conclude from these two tails? The persistence of the lower tail may be due to the lack of demonstrable scale and scope economies in physician practice. Or it may simply be that this tail has persisted as the dominant pattern of physician organization in the US, subject to gradual loss of market share and erosion of its economic position. Whether practices in this tail can find the economic resources to renew themselves (e.g., recruit and sustain younger colleagues) remains to be seen. IPAs might have assisted the physicians practicing in this tail in remaining viable in the face of managed care growth, but not given them the resources or leverage to prosper long term. It is entirely possible this tail erodes sharply as the baby boomer doctors who comprise perhaps two-thirds of these smaller practices retire in the next decade. Likewise, the growth of the upper tail does not appear to be based on measurable economic returns generated within the physician marketplace. There is no evidence these large groups have been able to achieve scale and scope economies from their clinical operations (production-side economies). Rather, we believe they have been formed by hospitals and IDNs for reasons extrinsic to the physician market effects: to build up outpatient hospital utilization or inpatient market share, to leverage insurers for more favorable contracts and rates, or to position the owner for a transition away from fee-for-service payment by Medicare or commercial insurers. Indeed, the Department of Justice has already investigated several of the

102

LAWTON ROBERT BURNS ET AL.

hospital-formed groups as well as some large single-specialty networks that contain 50 70% of all doctors in that specialty in the local/regional market for possible anti-competitive effects. As our economist colleagues would say, we may have an equilibrium with two very different configurations of firms (and likely of products too) that reflects both demand and supply side differences. On the demand side, perhaps the answer lies in differences in the groups’ customers. Some patients may want to coordinate their own care by dealing with separate and smallish practices, while some others want to deal with a large multispecialty practice or health insurer, for that matter, that coordinates care for them (whether they want it or not). There is some evidence of favorable patient selection into multispecialty groups (Weeks et al., 2010). Alternatively, on the supply side, there is variation in firm size and managers’ ability to run firms of different sizes. This may help to explain why technical economies of scale (where they exist) are not the full story. Or perhaps physicians have preferences regarding group size and what proportion of their relationships with other physicians they want coordinated within the firm or by the market. Or, perhaps doctors respond differently to within-firm incentives: some cease being productive if they get a salary versus productivity-based rewards, whereas others work just as hard and prefer the security of a predetermined salary. Or there may be generational differences in physician needs from various practice settings (need for economic security or work life balance) that predispose physicians to make different choices of practice settings. We need a theory of variation in consumer or physician preferences across different markets and practice settings to test these alternative explanations. Interestingly, each tail could be (and has been) characterized as the more agile and adaptive to changes in the health care system: the lower end due to small practice sizes and low bureaucratic mass of their practices, the upper end due to the enhanced ability of their practices to organize new service offerings and manage risk. It is not clear at this point how each tail will fare competitively going forward. Practices in the upper tail appear to incur higher operating costs that require some compensating benefits in order to survive and thrive. Such benefits as economies of scale, which are often presumed to exist, do not appear to have been achieved thus far on a consistent basis, even in the medium or long term. If they are achieved, it may be due to management efforts and execution in spite of potential diseconomies of practice scale. Regardless of whether groups in the upper tail enjoy any scale or scope economies, it is not clear whether they will be able to sustain their recent

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

103

growth in the absence of external capital. Such capital which is needed to aggregate, integrate, and wire the practices has been supplied by hospitals, insurers, and equity firms. The lesson of the 1990s is that such external capital sources can quickly dry up. Should history repeat itself during this decade, the survivorship principle exhibited by the lower tail of physician groups will be further supported. Enhanced market power may be another benefit to the vertically integrated physician enterprises that regulators may have trouble confronting across a wide provider landscape. Further research on this question is needed. As long as vertically integrated arrangements are able to charge and receive higher prices, they may be able to continue making the subsidy investments that enabled them to grow in the first place. However, as the insurance landscape tilts more to public payers and administered pricing, such arrangements will be increasingly challenged to deliver on quality and cost containment metrics that have proved elusive to them in the past. While integrated systems such as Kaiser, Mayo, and Geisinger remain a “counter-culture” in delivery of health care services (Goldsmith, 2002), huge new physician enterprises are emerging at the upper tail of group size that are major players in rapidly tightening health services markets. Some are backed by powerful, regional hospital systems. Others are backed by private equity investors and are even traded on public stock exchanges. The deeply ingrained history, mission, organization, and culture of the classic integrated care systems make them difficult to replicate in the current setting of small-group practices and fragmented care (Gitterman et al., 2003; Mechanic, 2010). However, the newer organizations appear to be more suited to a changing health care landscape that is consolidating on a large scale and is capable of managing risk. The implementation of the Affordable Care Act is ushering in a new era that focuses on coordinating care and reducing overutilization. Whether this care coordination will result in new payment paradigms for Medicare or commercial insurance, like ACOs or bundled payment, is unclear. Past experience and the lack of consistent evidence on the impact of various organizational strategies (e.g., horizontal integration, vertical integration) suggest that widespread success in cost containment and quality improvement will be challenging. This point is further emphasized by evidence that the vast majority of the variation in patient costs is due to patient-specific factors, rather than to the delivery system and its organization (Reschovsky, Hadley, Saiontz-Martinez, & Boukus, 2011). Policymakers and providers may likely require a more systematic approach to organizing provider delivery systems and incentives.

104

LAWTON ROBERT BURNS ET AL.

NOTES 1. According to Marder and Zuckerman (1985, Table 1), positions in groups of 100 + physicians constituted only 8.6% (1965), 9.7% (1969), 14.1% (1975), and 21.4% (1980) of group practice physicians. According to Table 1, that percentage reached 24.8% by 1988. 2. Some of the later data are not entirely comparable with the earlier data, so care must be taken in interpreting the trends. Nevertheless, many of the years parallel those reported in the Physician Marketplace Report. 3. NAMCS data are also available for two earlier years (1999, 2002). The CDC reported that some of the statistics did not meet standards of reliability or precision. They are therefore omitted here. 4. The last statistic suggests a near doubling in the prevalence of large groups. MGMA researchers believe the number of very large groups is overstated by approximately 20% due to double counting of some groups (David Gans, personal communication). 5. Some consultants argue that the small size of many practices provides physicians the agility to more easily adapt to market changes due to their simple governance structures, thus explaining their persistence over time (Isaacs & Jellinek, 2012). An alternative explanation, per Balzac, might be that the small are difficult to crush because they lie so flat beneath the foot. In addition to retaining small-sized practices, some physicians have developed “micropractices” with little or no office staff and patient amenities to reduce overhead costs. Others have shifted to concierge practices to avoid third-party payers (see Isaacs & Jellinek, 2012). 6. Alternatively, physicians may be too stubborn to throw in their cards and admit they are losing the game. Many solo practitioners and two-person partnerships persisted after 2008 because they lacked the financial resources to retire, and have remained trapped in the lower tail of the distribution. 7. Not all of these reasons may be true, however. Leaders of single-specialty groups report they have little or no ability to negotiate higher rates from payers (at least in the early 2000s). Moreover, the retreat from HMOs and capitation at the end of the 1990s may have shifted physicians’ organizing efforts away from multispecialty to single-specialty groups for several reasons: there was less of a need to belong to groups with primary care physicians (PCPs) to coordinate care, specialists generate higher revenues than PCPs and thus do not need to share their fees with them, and there are less complex governance mechanisms in single-specialty groups. 8. Recent research to generalize Reinhardt’s (1972) production function for physician services reaches a similar conclusion. Thurston and Libby (2002) find that the technical relationships that describe the production process for physician services (the impact of capital, physician labor, and non-physician labor inputs on patient visit output) are stable over time (1965 1988). This stability holds despite the fact that the external market for physician services changed drastically (e.g., introduction of Medicare, the Prospective Payment System, managed care) during the time interval. The authors conclude that physician practices continued to treat patients in the same way they always had. 9. While the majority of IPAs are physician-owned, roughly one-fifth are hospital-owned (Casalino et al., 2003; Gillies, Shortell, Casalino, Robinson, & Rundall,

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

105

2003). We discuss them in greater detail in a later section on virtual integration. In addition to these models, physicians and physician groups can also enter into contracts with hospitals to cover specific clinical service areas such as radiology, anesthesiology, or pathology. Such physicians have traditionally been labeled as hospital-based practitioners. Data from 2010 published by the American Hospital Association (AHA) indicates that 7% of US physicians have individual contracts and another 20% have group contracts with hospitals (American Hospital Association, 2012). 10. The AHA data in Fig. 6 suggest that roughly one-third of hospitals operate an employed model. The one-third estimate seems low compared to much higher prevalence rates frequently reported in the trade literature. For example, the Community Tracking Study found that 65% of hospitals in the 12 markets they followed owned physician practices in 2000 2001 (Lake, Devers, Brewster, & Casalino, 2003). According to Peter Kralovec, who maintains the AHA’s database, both the percentage of hospitals with the salaried model and the number of physicians in such models have grown only slowly. Conversely, a recent survey by HealthLeaders of medical staff leaders suggests the employed model is now found in most hospitals alongside the traditional voluntary medical staff: 78% of surveyed hospitals offer full employment, 67% offer the traditional voluntary medical staff, 63% offer paid directorships, and 25% offer clinical co-management (Cantlupe, 2010). These findings diverge from the AHA statistics likely due to the small sample size in the HealthLeaders survey and the tendency to respond based on having an employment model. 11. The authors thank Professor Guy David and Mr. Joseph Miller, Senior VicePresident at the Society of Hospital Medicine for supplying these data. 12. The 58% figure is much higher than the AHA’s data suggesting one-third of hospitals use an employment model. The discrepancy is likely due to (a) the fact that AHA data on hospitalists do not discriminate whether they are hospitalemployed or (b) the possibility that hospitals do not consider them when they report the various physician affiliation models they utilize. 13. The HealthLeaders survey reveals that over 50% of hospitals employ no more than 20% of their physicians, although there is a long tail in the distribution: 6% of hospitals employed no doctors, 46% of hospital medical staffs contained 1 20% employed doctors, 20% of hospital medical staffs contained 21 40% employed doctors, 7% of medical staffs contained 41 60% employed doctors, 6% of medical staffs contained 61 80% employed doctors, and 15% of medical staffs contained 81 10% employed doctors (Cantlupe, 2010). 14. As a percentage of all MGMA members, physician-owned groups dropped from 43.6% to 35.9%. Data collected by the pharmaceutical firm Sanofi (2013) also shows a size advantage for hospital-owned over freestanding medical groups. The former are 50% larger than the latter; the differences are most notable in the largest size category they measure (20 + physicians). Twenty percent of hospital-owned groups have 20 or more doctors; among independent groups, only 9% have 20 + doctors. Hospital-owned groups are more likely to be multispecialty (55%) compared to independent groups (40%). Overall, roughly one-quarter of all medical groups tracked are affiliated with hospital systems. Note: In 2004, Sanofi merged with Aventis, which historically maintained the database.

106

LAWTON ROBERT BURNS ET AL.

15. The authors thank Dave Gans for sharing the survey results. 16. These aggregate trends masked some important differences, however. The Allina physicians who joined the employed group and moved through the “grieving process” of losing their autonomy became important contributors to the group and exhibited an increase in organizational commitment; those who joined the group and bemoaned the loss of their autonomy (and thus delayed the grieving process) stalled the integration process and led to a sharp decrease in organizational commitment. 17. The authors thank Chad Peel for verifying the Optum information presented here. 18. It is possible, of course, that more hospital-centric systems that dominate their local markets such as Advocate and Partners Health Care may accumulate wealth over time (extracted from payers), invest it in their owned and affiliated medical groups, and make them just as successful. 19. IPA-model HMOs are not the same as IPA physician organizations. The former developed rapidly in the mid- and late 1980s along with the managed care revolution, and became the dominant physician staffing model for HMOs. The latter arose primarily during the 1990s as one model of integrated delivery network integration between physicians and hospitals. 20. Of course, the high entry rates of IPAs may be matched by high exit rates, which occurred during the 1990s (Haas-Wilson & Gaynor, 1998b; Kirchhoff, 2013).

ACKNOWLEDGMENTS The authors thank Mark Pauly, Stephen Shortell, and John Kralewski for their comments on this chapter. They also thank Peter Kralovec at Health Forum, David Gans at the Medical Group Management Association, and Esther Hing at the Centers for Disease Control and Prevention for making available much of the data reported here.

REFERENCES Advisory Board. (1999). Stopping the bleed: Executive briefing. Washington, DC: Health Care Advisory Board. Advisory Board. (2012). The power of joining a hospital network: Physician docs raise rates. Washington, DC: Health Care Advisory Board. Advisory Board. (2013). Doctors offer ‘survival tools’ to avoid affiliations with hospitals. Washington, DC: Health Care Advisory Board. Alexander, J. A., Maeng, D., Casalino, L. P., & Rittenhouse, D. (2012). Use of care management practices in small- and medium-sized physician groups: Do public reporting of physician quality and financial incentives matter? Health Services Research, 48(2), 376 397.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

107

Alexander, J. A., Waters, T. M., Boykin, S., Burns, L. R., Shortell, S. M., Gillies, R. R., Budetti, P., & Zuckerman, H. S. (2001a). Risk assumption and physician alignment with health care organizations. Medical Care, 39(7), I46 I61 [Physician-System Alignment Supplement]. Alexander, J. A., Waters, T. M., Burns, L. R., Shortell, S. M., Gillies, R. R., Budetti, P., & Zuckerman, H. S. (2001b). The ties that bind: Interorganizational linkages and physician-system alignment. Medical Care, 39(7), I30 I45 [Physician-System Alignment Supplement]. American Hospital Association. (2012). AHA hospital statistics 2012 edition. Chicago, IL: American Hospital Association. American Medical Association. (1933). Private group practice. JAMA, 100, 1605 1608, 1693 1699, 1773 1778. American Medical Association. (1940). Group medical practice. Chicago, IL: American Medical Association. Bailey, R. M. (1968). A comparison of internists in solo and fee-for-service group practice in the San Francisco Bay Area. Bulletin of the New York Academy of Medicine, 44, 1293 1303. Baldwin, G. (2012). The price of I.T. progress? HealthData Management, May 1. Retrieved from http://www.healthdatamanagement.com/issues/20_5/hospital-physician-i.t.-informationtechnology-ehr-44378-1.html. Accessed on May 25, 2013. Bazzoli, G., Shortell, S., Ciliberto, F., Kralovec, P., & Dubbs, N. (2001). Tracking the changing provider landscape: Implications for health policy and practice. Health Affairs, 20(6), 188 196. Bazzoli, G., Shortell, S., Dubbs, N., Chan, C., & Kralovec, P. (1999). A taxonomy of health networks and systems: Bringing order out of chaos. Health Services Research, 33(6), 1683 1717. Beach, L. R. (1992). Making the right decision: Organizational culture, vision, and planning. Englewood Cliffs, NJ: Prentice-Hall. Bellows, N. M., McMenamin, S. B., & Halpin, H. A. (2010). Adoption of health promotion practices in a cohort of U.S. physician organizations. American Journal of Preventive Medicine, 39(6), 555 558. Berenson, R. A., Ginsburg, P. B., & Kemper, N. (2010). Unchecked provider clout in California foreshadows challenges to health reform. Health Affairs, 29(4), 699 705. Besanko, D., Dranove, D., & Shandley, M. (2000). The economics of strategy (2nd ed.), New York, NY: Wiley. Boan, J. A. (1966). Group practice. Ottawa, Canada: Royal Commission on Health Services. Boukus, E., Cassil, A., & O’Malley, A. S. (2009). A snapshot of U.S. physicians: Key findings from the 2008 Health Tracking Physician Survey. Data Bulletin No. 35. Center for Studying Health System Change, Washington, DC. Bradford, W. D. (1995). Solo versus group practice in the medical profession: The influence of malpractice risk. Health Economics, 4(2), 95 112. Bradford, W. D., & Martin, R. E. (2000). Partnerships, profit sharing, and quality competition in the medical profession. Review of Industrial Organization, 17, 193 208. Brown, D. M. (1988). Do physicians underutilize aides? Journal of Human Resources, 23(3), 342 355. Budetti, P. P., Shortell, S. M., Waters, T., Alexander, J. A., Burns, L. R., Gillies, R. R., & Zuckerman, H. (2002). Physician and health system integration. Health Affairs, 21(1), 203 210.

108

LAWTON ROBERT BURNS ET AL.

Bunderson, J. S., Lofstrom, S. M., & Van de Ven, A. H. (2000). Conceptualizing and measuring professional and administrative models of organizing. Organization Behavior Research, 3(4), 366 391. Burns, L. R. (1999). Polarity management: The key challenge for integrated health systems. Journal of Healthcare Management, 44(1), 14 31. Burns, L. R., Alexander, J. A., Shortell, S. M., Zuckerman, H. S., Budetti, P., Gillies, R. R., & Waters, T. M. (2001). Physician commitment to organized delivery systems. Medical Care, 39(7), I9 I29 [Physician-System Alignment Supplement]. Burns, L. R., Alexander, J. A., Zuckerman, H. S., Andersen, R. M., & Torrens, P. (1995). Make, buy, or ally: Impact of governance mode on trading partner alignment. Paper presented at Annual Meeting of Association of Health Services Research. Burns, L. R., Andersen, R. M., & Shortell, S. M. (1990). The effect of hospital control strategies on physician satisfaction and hospital-physician conflict. Health Services Research, 25(3), 527 560. Burns, L. R., Andersen, R. M., & Shortell, S. M. (1993). Trends in hospital/physician relationships. Health Affairs, 12(3), 213 223. Burns, L. R., Gimm, G., & Nicholson, S. (2005). The financial performance of integrated health organizations (IHOs). Journal of Healthcare Management, 50(3), 191 213. Burns, L. R., Goldsmith, J., & Muller, R. (2010). History of hospital/physician relationships: Obstacles, opportunities, and issues. In J. Crosson & L. Tollen (Eds.), Partners in health (pp. 18 45). San Francisco, CA: Jossey-Bass. Burns, L. R., & Muller, R. W. (2008). Hospital-physician collaboration: Landscape of economic integration and impact on clinical integration. Milbank Quarterly, 86(3), 375 434. Burns, L. R., Nash, D., & Wholey, D. R. (2007). The evolving role of third parties in the hospital-physician relationship. American Journal of Medical Quality, 22(6), 402 409. Burns, L. R., & Pauly, M. V. (2002). Integrated delivery networks: A detour on the road to integrated health care? Health Affairs, 21(4), 128 43. Burns, L. R., Shortell, S. M., & Andersen, R. M. (1998). Does familiarity breed contentment? Attitudes and behaviors of integrated physicians. In J. Kronenfeld (Ed.), Research in the sociology of health care (Vol. 15), Greenwich, CT: JAI Press. Burns, L. R., & Wholey, D. R. (2000). Responding to a consolidating health care system: Options for physician organizations. In J. Blair, M. Fottler, & G. Savage (Eds.), Advances in health care management (Vol. 1, pp. 261 323). Greenwich, CT: JAI Press. Burns, L. R., Wholey, D. R., McCullough, J. S., Kralovec, P., & Muller, R. (2012). The changing configuration of hospital systems: Centralization, federalization, or fragmentation? In L. Friedman, G. Savage & J. Goes (Eds.), Annual review of health care management: Strategy and policy perspectives on reforming hospital systems (Advances in Health Care Management, 13, 189 232). Bingley, UK: Emerald. Cantlupe, J. (2010). Physician alignment in an era of change. Danvers, MA: HealthLeaders Media/Intelligence. Casalino, L. (2003). IPA overview. Testimony to FTC Health Care Competition Hearing. Retrieved from http://www.FTC.Gov/OGC/healthcarehearings/DOCS/030925LAWRENCE CASALINO.pdf Casalino, L. P., Devers, K. J., Lake, T. K., Reed, M., & Stoddard, J. (2003). Benefits of and barriers to large medical group practice in the United States. Archives of Internal Medicine, 163(16), 1958 1964.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

109

Casalino, L., Gillies, R. R., Shortell, S. M., Schmittdiel, J. A., Bodenheimer, T., Robinson, J. C., Rundall, T. G., Oswald, N., Schauffler, H., & Wang, M. C. (2003). External incentives, information technology, and organized processes to improve health care quality for patients with chronic diseases. JAMA, 289(4), 434 441. Casalino, L. P., Pham, H., & Bazzoli, G. (2004). Growth of single-specialty medical groups. Health Affairs, 23(2), 82 90. Casalino, L. P., & Robinson, J. C. (2003). Alternative models of hospital-physician affiliation as the United States moves away from tight managed care. Milbank Quarterly, 81(2), 331 351. Centers for Disease Control and Prevention (CDC). (2007a). National Ambulatory Medical Care Survey: 2005 summary. Advance Data. Number 387. National Center for Health Statistics, Hyattsvile, MD. Centers for Disease Control and Prevention (CDC). (2007b). Characteristics of office-based physicians and their practices: United States, 2003 2004. Series 13, Number 164. National Center for Health Statistics, DHHS, Hyattsvile, MD. Centers for Disease Control and Prevention (CDC). (2008a). National Ambulatory Medical Care Survey: 2008 summary tables. Atlanta, CA: Esther Hing. Centers for Disease Control and Prevention (CDC). (2008b). Characteristics of office-based physicians and their practices: United States, 2005 2006. Series 13, Number 166. National Center for Health Statistics, DHHS, Hyattsville, MD. Centers for Disease Control and Prevention (CDC). (2010a). National Ambulatory Medical Care Survey: 2007 summary. Number 27. National Center for Health Statistics, Hyattsvile, MD. Centers for Disease Control and Prevention (CDC). (2010b). National Ambulatory Medical Care Survey: 2010 summary tables. Atlanta, CA: Esther Hing. Center for Studying Health System Change. (1999). How physician organizations are responding to managed care. Issue Brief No. 20. CSHSC, Washington, DC. Ciliberto, F., & Dranove, D. (2006). The effect of physician hospital affiliations on hospital prices in California. Journal of Health Economics, 25, 29 38. Colla, C. H., Wennberg, D. E., Meara, E., Skinner, J. S., Gottlieb, D., Lewis, V. A., & Fisher, E. S. (2012). Spending differences associated with the Medicare physician group practice demonstration. JAMA, 308(10), 1015 1023. Collins, J., & Poras, J. (1994). Built to last: Successful habits of visionary companies. New York, NY: Harper Business. Creswell, J., & Abelson, R. (2012). A hospital war reflects a bind for doctors in the U.S. New York Times, November 30. Crosson, J. (2005). The delivery system matters. Health Affairs, 24(6), 1543 1548. Cuellar, A. E., & Gertler, P. J. (2006). Strategic integration of hospitals and physicians. Journal of Health Economics, 25, 1 28. Curoe, A., Kralewski, J., & Kaissi, A. (2003). Assessing the cultures of medical group practices. Journal of the American Board of Family Medicine, 16, 394 398. Damberg, C. L., Shortell, S. M., Raube, K., Gillies, R. R., Rittenhouse, D. R., McCurdy, R. K., & Casalino, L. P. (2010). Relationship between quality improvement processes and clinical performance. The American Journal of Managed Care, 16(8), 601 606. Defelice, L. C., & Bradford, W. D. (1997). Relative inefficiencies in production between solo and group practice physicians. Health Economics, 6(5), 455 465.

110

LAWTON ROBERT BURNS ET AL.

Dukerich, J., Golden, B., & Shortell, S. M. (2002). Beauty is in the eye of the beholder: The impact of organizational identification, identity, and image on the cooperative behaviors of physicians. Administrative Science Quarterly, 47, 507 533. Eisenberg, J. (2002). Physician utilization: The state of research about physicians’ practice patterns. Medical Care, 40(11), 1016 1035. Emswiler, T., & Nichols, L. M. (2009). Hill Physicians Medical Group: Independent physicians working to improve quality and reduce costs. Case study on high-performing health care organization (Publication No. 1247, Vol. 11). New York, NY: The Commonwealth Fund. Enthoven, A. (1993). The history and principles of managed competition. Health Affairs, 12(Suppl.), 24 48. Escarce, J., & Pauly, M. V. (1998). Physician opportunity costs in physician practice cost functions. Journal of Health Economics, 17, 129 151. Federal Trade Commission. (2013). Federal Trade Commission and State of Idaho v. St. Luke’s Health System and Saltzer Medical Group. Retrieved from http://www.ftc.gov/os/caselist/ 1210069/130312stlukescmpt.pdf. Accessed on July 10, 2013. Frech, H. E., & Ginsburg, P. B. (1974). Optimal scale in medical practice: A survivor analysis. Journal of Business, 47, 26 41. Friedberg, M. W., Coltin, K. L., Pearson, S. D., Kleinman, K., Zheng, J., Singer, J., & Schneider, E. (2007). Does affiliation of physician groups with one another produce higher quality primary care? Journal of General Internal Medicine, 22(10), 1385 1392. Fronstin, P. (2012). Sources of health insurance and characteristics of the uninsured: Analysis of the March 2012 current population survey. Issue Brief No. 376. Employee Benefits Research Institute, Washington, DC. Gal-Or, E. (1999). The profitability of vertical mergers between hospitals and physician practices. Journal of Health Economics, 18(5), 623 654. Gans, D. (2012a). Why hospital-owned medical groups lose money. MGMA Connexion, 12(4), 19 20. Gans, D. (2012b). Best of breed: What enables some hospital-owned groups to prosper? MGMA Connexion, 12(5), 21 22. Gans, D. (2012c). Perspective is everything: Look beyond reported losses. MGMA Connexion, 12(10), 17 19. Gans, D., & Wolper, L. (2013). Management makes the difference: Optimizing medical group performance. ACMPE Executive View (Spring), 14 18. Gaynor, M. (1989). Competition within the firm: Theory plus some evidence from medical group practice. The RAND Journal of Economics, 20(1), 59 76. Gaynor, M. (2006). Is vertical integration anticompetitive? Definitely maybe (but that’s not final). Journal of Health Economics, 25, 175 180. Gaynor, M., & Pauly, M. V. (1990). Compensation and productive efficiency in partnerships: Evidence from medical group practice. Journal of Political Economy, 98(3), 544 573. Getzen, T. E. (1984). A ‘brand name firm’ theory of medical group practice. Journal of Industrial Economics, 33(2), 199 215. Gillies, R. R., Shortell, S. M., Casalino, L., Robinson, J. C., & Rundall, T. (2003). How different is California? A comparison of U.S. physician organizations. Health Affairs Web Exclusive, W3/493 502. Gitterman, D. P., Weiner, B., Domino, M., McKethan, A., & Enthoven, A. (2003). The rise and fall of a Kaiser Permanente expansion region. Milbank Quarterly, 81(4), 567 601.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

111

Goes, J., & Zhan, C. (1995). The effects of hospital-physician integration strategies on hospital financial performance. Health Services Research, 30(4), 507 530. Goldsmith, J. (2002). Integrating care: A talk with Kaiser Permanente’s David Lawrence. Health Affairs, 21(1), 39 48. Goldsmith, J. (2012). The future of medical practice: Creating options for practicing physicians to control their professional destiny. Bonita Springs, FL: The Physicians Foundation. Goldstein, D. (1996). Building and managing effective physician organizations under capitation. Gaithersburg, MD: Aspen. Goldstein, L. (2005, September). Not-for-profit hospital-physician integration: New strategies or back to the future? Report # 94423. Moody’s Investors Service, New York, NY. Greenfield, S., Nelson, E. C., Zubkoff, M., Manning, W., Rogers, W., Kravitz, R., & Ware, J. E. (1992). Variations in resource utilization among medical specialties and systems of care. JAMA, 267(12), 1624 1630. Greenfield, S., Rogers, W., Mangotich, M., Carney, M., & Tarlov, A. (1995). Outcomes of patients with hypertension and non-insulin dependent diabetes mellitus treated by different systems and specialties: Results from the medical outcomes study. JAMA, 274, 1436 1444. Haas-Wilson, D., & Gaynor, M. (1998a). Increasing consolidation in healthcare markets: What are the antitrust policy operations? Health Services Research, 33(5), 1403 1419. Haas-Wilson, D., & Gaynor, M. (1998b). Physician networks and their implications for competition in health care markets. Health Economics, 7(2), 179 182. Harbuck, S., Follmer, A., Dill, M., & Erikson, C. (2012). Estimating the number and characteristics of hospitalist physicians in the United States and their possible workforce implications. Analysis in Brief, 12(3), 1 2. Havlicek, P. (1990). Medical group practices in the US 1990. Chicago, IL: AMA. Havlicek, P. (1993). Medical group practices in the US 1993. Chicago, IL: AMA. Havlicek, P. (1996). Medical group practices in the US 1996. Chicago, IL: AMA. Havlicek, P. (1999). Medical group practices in the US 1999. Chicago, IL: AMA. Haywood, T., & Kosel, K. (2011). The ACO model a three year financial loss? New England Journal of Medicine, 364, E27. Hillson, S. D., Feldman, R., & Wingert, T. D. (1990). Economies of scope and payment for physician services. Medical Care, 30(9), 822 831. Ho, K. (2009). Barriers to entry of a vertically integrated health insurer: An analysis of welfare and entry costs. Journal of Economics & Management Strategy, 18(2), 487 545. Hough, D. E. (2002). Is bigger always better? The optimal size of a group practice. Journal of Medical Practice Management, 17(6), 312 318. Hough, D. E., Liu, K., & Gans, D. N. (2013). Size matters: The diversity of physician practice production functions. Unpublished manuscript. Indest, G. (n.d.). The association of independent doctors has concerns about acquisitions by hospitals and medical groups. Retrieved from http://www.indest.com/blog/posts/theassociation-of-independent-doctors-has-concerns-about-acquisitions-by-hospitals-and-medicalgroups.html. Accessed on May 25, 2013. Isaacs, S., & Jellinek, P. (2012). Survival of the fittest: A review of promising models for the maintenance of independent medical practice. Bonita Springs, FL: The Physicians Foundation. Joszt, L. (2013). Practice management: Death of private practice ‘exaggerated’. Physician’s Money Digest HCP Live. (June 7th). Retrieved from http://www.hcpLive.com/physiciansmoney-digest/PRACTICE-MANAGEMENT/DEATH-OF-PRIVATE-PRACTICE-EXAG GERATED-LBJ

112

LAWTON ROBERT BURNS ET AL.

Kaissi, A., Kralewski, J. E., & Dowd, B. (2003). Financial and organizational factors affecting the employment of nurse practitioners and physician assistants in medical group practices. Journal of Ambulatory Care Management, 26(3), 209 216. Kane, C. K. (2004a). The practice arrangements of patient care physicians, 1999-revised. Physician Marketplace Report No. 2004-1. AMA, Chicago, IL. Kane, C. K. (2004b). The practice arrangements of patient care physicians, 2001. Physician Marketplace Report No. 2004-2. AMA, Chicago, IL. Kane, C. K. (2009). The practice arrangements of patient care physicians, 2007 2008. Policy Research Perspectives. AMA, Chicago, IL. Keating, N., Landrum, M., Landon, B., Ayanian, J., Borbas, C., & Wolf, R. (2004). The influence of physicians’ practice management strategies and financial arrangements on quality of care among patients with diabetes. Medical Care, 42, 829 839. Keckley, P., Coughlin, S., & Stanley, E. (2013). Deloitte 2013 survey of U.S. physicians. Washington, DC: Deloitte Center for Health Solutions. Ketcham, J. D. (2001). Data microscope: Focus on group size. In Performance and practices of successful medical groups: 2000 report based on 1999 data (pp. 18 21). Englewood, CO: Medical Group Management Association. Ketcham, J. D., Baker, L. C., & MacIsaac, D. (2007). Physician practice size and variations in treatment and outcomes: Evidence from Medicare patients with AMI. Health Affairs, 26(1), 195 205. Kimball, L., & Lorant, J. (1977). Physician productivity and returns to scale. Health Services Research, 12(4), 367 380. Kirchhoff, S. M. (2013). Physician practices: Background, organization, and market consolidation. Cornell University ILR School, Ithaca, NY. Kitchell, W., & Hurst, R. (2011). Specialty physician practice trends. MANow (October 2011). Retrieved from http://www.mossadams.com/mossadmas/media/Documents/Pulbications/ MA%20Now/MA-Now_HCG_Oct2011.pdf. Accessed on May 16, 2013. Kocher, R., & Sahni, N. R. (2011). Hospitals’ race to employ physicians the logic behind a money-losing proposition. New England Journal of Medicine, 364(19), 1790 1793. Kralewski, J. E., Dowd, B., Cole-Adeniyi, T., Gans, D., Malakar, L., & Elson, B. (2008). Factors influencing physician use of clinical electronic information technologies after adoption by their medical group practices. Health Care Management Review, 33(4), 361 367. Kralewski, J. E., Dowd, B., Heaton, A., & Kaissi, A. (2005). The influence of the structure and culture of medical group practices on prescription drug errors. Medical Care, 43(8), 817 825. Kralewski, J. E., Dowd, B., Kaissi, A., Curoe, A., & Rockwood, T. (2005). Measuring the culture of medical group practices. Health Care Management Review, 30(3), 184 193. Kralewski, J. E., Dowd, B. E., Xu, Y., & Knutson, D. (2011). The organizational characteristics of best medical group practices. Unpublished manuscript. Kralewski, J. E., Pitt, L., & Shatin, D. (1985). Structural characteristics of medical group practices. Administrative Science Quarterly, 30(1), 34 45. Kralewski, J. E., Rich, E. C., Feldman, R., Dowd, B., Bernhardt, T., Johnson, C., & Gold, W. (2000). Effects of medical group practice and physician payment methods on costs of care. Health Services Research, 35(3), 591 613. Kralewski, J. E., Wingert, T., Knutson, D., & Johnson, C. (1999). The effects of medical group practice organizational factors on physicians’ use of resources. Journal of Healthcare Management, 44(3), 167 182.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

113

Krohn, R. (1998). Physician networks. Chicago, IL: American College of Healthcare Executives. Lake, T., Devers, K., Brewster, L., & Casalino, L. (2003). Something old, something new: Recent developments in hospital-physician relationships. Health Services Research, 38(1), 471 488. Laufer, N. (2011, November). The employment of doctors by hospitals indentured servitude or practice salvation? Maricopa County Medical Society Online. Retrieved from http://www. mcmsonline.com/president/nathan-laufer-md/employment-doctors-hospitals- -indenturedservitude-or-practice-salvation. Accessed on May 25, 2013. Lee, R. H. (1990). Monitoring physicians: A bargaining model of medical group practice. Journal of Health Economics, 9(4), 463 481. Liebhaber, A., & Grossman, J. M. (2007). Physicians moving to mid-sized, single-specialty practices. Center for Studying Health System Change Tracking Report, No. 18. CSHSC, Washington, DC. Madison, D. L. (1990). Notes on the history of group practice: The tradition of the dispensary. MGM Journal, 37(5), 52 54, 56 60, 86 92. Madison, D. L., & Konrad, T. R. (1988). Large medical group-practice organizations and employed physicians: A relationship in transition. Milbank Quarterly, 66(2), 240 282. Madison, K. (2004). Hospital physician affliations and patient treatments, expenditures, and outcomes. Health Services Research, 39(2), 257 278. Main, T., & Slywotzky, A. (2011, November). The quiet health-care revolution. The Atlantic Monthly. Retrieved from http://www.theatlantic.com/magazine/archive/2011/11/the-quiethealth-care-revolution/308667/. Accessed on May 22, 2013. Marder, W. D., Emmons, D., Kletke, P., & Willke, R. (1988). Physician employment patterns: Challenging conventional wisdom. Health Affairs, 7(5), 137 145. Marder, W. D., & Zuckerman, S. (1985). Competition and medical groups: A survivor analysis. Journal of Health Economics, 4(2), 167 176. Matthews, A. W. (2012). Same doctor visit, double the cost. Wall Street Journal. August 30. Mechanic, D. (2010). Replicating high-quality medical care organizations. JAMA, 303(6), 555 556. MedPAC. (2008). Report to the Congress: Reforming the delivery system. Chapter 3. Washington, DC: Medical Payment Advisory Commission. MedPAC. (2012). Report to the Congress: Medicare payment policy. Washington, DC: Medical Payment Advisory Commission. Mehrotra, A., Epstein, A. M., & Rosenthal, M. B. (2006). Do integrated medical groups provide higher-quality medical care than individual practice associations? Annals of Internal Medicine, 145, 826 833. MGMA. (1998). Performance and practices of successful medical groups. Englewood, CO: Medical Group Management Association. Miller, R. H., & Bovbjerg, R. R. (2002). Efforts to improve patient safety in large, capitated medical groups: Description and conceptual model. Journal of Health Politics, Policy and Law, 27, 401 440. Minott, J., Helms, D., Luft, H., Guterman, S., & Weil, H. (2010, April). The group employed model as a foundation for health care delivery reform. Issue Brief. New York, NY: The Commonwealth Fund. Morrisey, M. A., Alexander, J., Burns, L. R., & Johnson, V. (1996). Managed care and physician/hospital integration. Health Affairs, 15(4), 62 73.

114

LAWTON ROBERT BURNS ET AL.

Newhouse, J. P. (1973). The economics of group practice. Journal of Human Resources, 8(1), 37 56. Pauly, M. V. (1996). Economics of multispecialty practice. Journal of Ambulatory Care Management, 19(3), 26 33. Penner, M. J. (1997). Capitation in California: A study of physician organizations managing risk. Ann Arbor, MI: Health Administration Press. Pham, H., Schrag, D., Hargraves, J., & Bach, P. (2005). Delivery of preventive services to older adults by primary care physicians. JAMA, 294(4), 473 481. Pope, G., & Burge, R. T. (1992). Inefficiencies in physician practices. In L. Rossiter & M. Scheffler (Eds.), Advances in health economics and health services research (Vol. 13, pp. 129 164). Greenwich, CT: JAI Press. Pope, G., & Burge, R. T. (1996). Economies of scale in physician practice. Medical Care Research and Review, 53(4), 417 440. Pricco, C. (2013). Optum: Accountable care solutions. Retrieved from http://www.stthomas. edu/business/centers/healthmedical/events/2013-01-31-UHG-Sleeping.html. Accessed on May 23, 2013. Regents Health Resources. (2011). Cost comparison: Hospital-based versus freestanding outpatient imaging services. RBJ Imaging Market File, 1 2. Reinhardt, U. (1972). A production function for physician services. Review of Economics and Statistics, 54(1), 55 66. Reschovsky, J. D., Hadley, J., Saiontz-Martinez, C. B., & Boukus, E. R. (2011). Following the money: Factors associated with the cost of treating high-cost Medicare beneficiaries. Health Services Research, 46(4), 997 1021. Rittenhouse, D., & Robinson, J. C. (2006). Improving quality in Medicaid: The use of care management processes for chronic illness and preventive care. Medical Care, 44(1), 47 54. Rittenhouse, D. R., Casalino, L. P., Gillies, R. R., Shortell, S. M., & Lau, B. (2008). Measuring the medical home infrastructure in large medical groups. Health Affairs, 27(5), 1246 1258. Rittenhouse, D. R., Casalino, L. P., Shortell, S. M., McClellan, S. R., Gillies, R. R., Alexander, J. A., & Drum, M. L. (2011). Small and medium-size physician practices use few patient-centered medical home processes. Health Affairs, 30(8), 1575 84. Rittenhouse, D. R., Shortell, S. M., Gillies, R. R., Casalino, L. P., Robinson, J. C., McCurdy, R. K., & Siddique, J. (2010). Improving chronic illness care: Findings from a national study of care management processes in large physician practices. Medical Care Research and Review, 67(3), 301 320. Robinson, J. C., & Casalino, L. P. (1995). The growth of medical groups paid through capitation in California. New England Journal of Medicine, 333, 1684 1687. Robinson, J. C., Casalino, L. P., Gillies, R. R., Rittenhouse, D. R., Shortell, S. M., & Fernandes-Taylor, S. (2009). Financial incentives, quality improvement programs, and the adoption of clinical information technology. Medical Care, 47(4), 411 417. Rosenman, R., & Friesner, D. (2004). Scope and scale inefficiencies in physician practice. Health Economics, 13, 1091 1116. Rosenthal, M., Frank, R. G., Buchanan, J. L., & Epstein, A. M. (2001). Scale sand structure of capitated physician organizations in California. Health Affairs, 20(4), 109 119. Rubin, H. R., Gandek, B., Rogers, W. H., Kosinski, M., McHorney, C. A., & Ware, J. E. (1993). Patients’ ratings of outpatient visits in different practice settings: Results from the medical outcomes study. JAMA, 270(7), 835 840.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

115

Rundall, T. G., Shortell, S. M., Wang, M. C., Casalino, L. P., Bodenheimer, T., Gillies, R. R., & Robinson, J. C. (2002). As good as it gets? Chronic care management in nine leading U.S. Physician organizations. British Medical Journal, 325, 958 961. Safran, D., Tarlov, A., & Rogers, W. H. (1994). Primary care performance in fee-for-service and prepaid health care systems: Results from the medical outcomes study. JAMA, 271(20), 1579 1586. Sager, A., & Socolar, D. (2005). Health costs absorb one-quarter of economic growth, 2000 2005. Retrieved from http://dcc2.bumc.bu.edu/hs/health%20costs%20absorb%20onequarter%20of%20economic%20growth%20%202000-05%20%20sager-socolar%207%20 february%202005.pdf. Accessed on May 21, 2013. Sanofi. (2013). Managed care digest series: Hospitals/systems digest 2013. Bridgewater, NJ: Sanofi-Aventis. Sarma, S., Devlin, R. A., & Hogg, W. (2010). Physician’s production of primary care in Ontario, Canada. Health Economics, 19, 14 30. Shih, A., Davis, K., Schoenbaum, S., Gauthier, A., Nuzum, R., & McCarthy, D. (2008). Organizing the U.S. health care delivery system for high performance. New York, NY: The Commonwealth Fund. Shortell, S. M. (1991). Effective hospital-physician relationships. Ann Arbor, MI: Health Administration Press. Shortell, S. M., Gillies, R. R., Anderson, D., Erickson, K., & Mitchell, J. (1996). Remaking health care in America: Building organized delivery systems. San Francisco, CA: Jossey-Bass. Shortell, S. M., Gillies, R. R., Siddique, J., Casalino, L. P., Rittenhouse, D. R., Robinson, J. C., & McCurdy, R. K. (2009). Improving chronic illness care: A longitudinal cohort analysis of large physician organizations. Medical Care, 47(9), 932 939. Shortell, S. M., & Schmittdiel, J. (2004). Prepaid groups and organized delivery systems: Promise, performance, and potential. In A. Enthoven & L. Tollen (Eds.), Toward a 21st century health system: The contributions and promise of prepaid group practice (pp. 1 21). San Francisco, CA: Jossey-Bass. Shortell, S. M., Schmittdiel, J. A., Wang, M. C., Li, R., Gillies, R. R., Casalino, L. P., & Rundall, T. G. (2005). An empirical assessment of high performing physician organizations: Results from a national study. Medical Care Research and Review, 62(4), 407 434. Shortell, S. M., Zazzali, J., Burns, L. R., Alexander, J. A., Gillies, R. R., Budetti, P., & Zuckerman, H. S. (2001). Implementing evidence-based medicine: The role of market pressures, compensation incentives, and culture of physician organizations. Medical Care, 39(7), I62 I78 [Physician-System Alignment Supplement]. Shouldice, R. G., & Shouldice, K. H. (1978). Medical group practice and health maintenance organizations. Washington, DC: Information Resources Press. Simpson, J. D., & Coate, M. B. (1998). Efficiencies or anticompetitive effects: Vertical integration of hospitals into medical practices. Antitrust Bulletin, 43, 376 401. Sirovich, B., Gallagher, P., Wennberg, D., & Fisher, E. (2008). Discretionary decision making by primary care physicians and the cost of U.S. health care. Health Affairs, 27(3), 749 758. Smart, D. (2004). Medical group practices in the US 2004. Chicago, IL: AMA. Smart, D. (2005). Medical group practices in the US 2005. Chicago, IL: AMA. Smart, D. (2006). Medical group practices in the US 2006. Chicago, IL: AMA. Smart, D. (2013). Physician characteristics and distribution in the US. Chicago, IL: AMA.

116

LAWTON ROBERT BURNS ET AL.

Solberg, L. I., Asche, S. E., Shortell, S. M., Gillies, R. R., Taylor, N., Pawlson, L. G., & Young, M. R. (2009). Is integration in large medical groups associated with quality? The American Journal of Managed Care, 15(6), e34 e41. Staiger, D., Auerbach, D., & Buerhaus, P. (2010). Trends in the work hours of physicians in the United States. JAMA, 303(8), 747. Stapleton, K. (2012). OptumHealth collaborative care: Overview and discussion. Presentation to the Wharton School. Philadelphia, PA (April). Starr, P. (1982). The social transformation of American medicine. New York, NY: Basic Books. Stensland, J., & Stinson, T. (2002). Successful physician-hospital integration in rural areas. Medical Care, 40(10), 908 917. Stevens, R. (1971). American medicine and the public interest. New Haven, CT: Yale University Press. Stigler, G. (1958). The economies of scale. Journal of Law and Economics, 1, 54 71. Tarlov, A. R., Ware, J. E., Greenfield, S., Nelson, E. C., Perrin, E., & Zubkoff, M. (1989). The medical outcomes study: An application of methods for monitoring the results of medical care. JAMA, 262(7), 925 930. Terry, K. (2012). Why are insurers buying physician groups? Hospital and Health Networks Magazine. Retrieved from http://www.hhnmag.com/hhnmag/jsp/articledisplay.jsp?dcrpath = HHNMAG/Article/data/01JAN2012/0112HHN_FEA_trendwatching&domain = HHNMAG. Accessed on May 26, 2013. Thurston, N. K., & Libby, A. M. (2002). A production function for physician services revisited. The Review of Economics and Statistics, 84(1), 184 191. Tollen, L. (2008). Physician organization in relation to quality and efficiency of care: A synthesis of recent literature. New York, NY: The Commonwealth Fund. Van de Ven, A., Rogers, R., Bechara, J., & Sun, K. (2008). Organizational diversity, integration and performance. Journal of Organizational Behavior, 29, 335 354. Vesely, R. (2012). Marriage of convenience. Modern Healthcare.com. Retrieved from http://www.modernhealthcare.com/article/20120102/MAGAZINE/301029918. Accessed on July 10, 2013. VHA. (2002). Medical staff organizations: The forgotten covenant. Irving, TX: Voluntary Hospitals of America. VHA. (2003). Physician hospital relationships: The empty clubhouse. Irving, TX: Voluntary Hospitals of America. Wan, T., Lin, B., & Ma, A. (2002). Integration mechanisms and hospital efficiency in integrated health care delivery systems. Journal of Medical Systems, 26(2), 127 143. Waters, T. M., Budetti, P. P., Reynolds, K. S., Gillies, R. R., Zuckerman, H. S., Alexander, J. A., Burns, L. R., & Shortell, S. M. (2001). Factors associated with physician involvement in care management. Medical Care, 39(7), I79 I91 [Physician-System Alignment Supplement]. Weaver, C. (2011). Managed care enters the exam room as insurers buy doctor groups. Kaiser Health News, July 1. Retrieved from http://www.kaiserhealthnews.org/Stories/2011/July/01/ unitedhealth-insurers-buy-doctors-groups.aspx. Accessed on December 15, 2012. Weeks, W. B., Gottlieb, D., Nyweide, D., Sutherland, J., Bynum, J., Casalino, L., Gillies, R. R., Shortell, S. M., & Fisher, E. S. (2010). Higher health care quality and bigger savings found at large multispecialty medical groups. Health Affairs, 29(5), 991 997. Weil, T. P. (2002). Multispecialty physician practices: Fixed and variable costs, and economies of scale. Journal of Ambulatory Care Management, 25(3), 70 77.

Horizontal and Vertical Integration of Physicians: A Tale of Two Tails

117

Wheelan, S. A. (2009). Group size, group development, and group productivity. Small Group Research, 40(2), 247 262. Wholey, D. R., & Burns, L. R. (1993). Organizational transitions: Form changes by health maintenance organizations. In S. Bacharach (Ed.), Research in the sociology of organizations (pp. 257 293). Greenwich, CT: JAI Press. Wolinsky, F. D., & Marder, W. D. (1985). The organization of medical practice and the practice of medicine. Ann Arbor, MI: Health Administration Press.

COMMENTARY ON “HORIZONTAL AND VERTICAL INTEGRATION OF PHYSICIANS: A TALE OF TWO TAILS” BY LAWTON ROBERT BURNS, JEFF C. GOLDSMITH, AND ADITI SEN Stephen M. Shortell Keywords: physician organizations; hospital physician integration; payment incentives; integrated delivery systems

Burns et al. have done a “tour de force” in summarizing and interpreting the body of evidence on the horizontal and vertical integration of physicians over the past 12 years. Of particular note is their historical summary of the origin and growth of group practices over time, synthesizing multiple data sources. They find that, despite conventional wisdom, there has been some but not a widespread shift toward larger size group practices. Further, the review of the evidence leads them to conclude that small groups (e.g., 7 10 physicians) tend to be more productive than large

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 119 124 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015010

119

120

STEPHEN M. SHORTELL

groups. The evidence also suggests that groups enjoy some economies of scale versus solo, independent practices or partnerships, although that evidence is “far from absolute.” They find little evidence that multispecialty practices exhibit economies of scale. The evidence on the vertical integration of physicians and hospitals on cost and quality of care is mixed. Overall, they conclude that there is no “one best way” for physicians to organize. This is a conclusion, of course, that others have arrived at as well, and one that we might best accept and put to rest as we look to the future.

IMPLICATIONS As someone who has conducted research on physician organizations, hospitals, hospital physician relationships, and integrated delivery systems over three decades, I think we have underestimated the importance of external incentives for change on the one hand and the internal capabilities of organizations and their leaders on the other hand to respond to these changes. As noted 12 years ago about the prospects for more integrated health care delivery in the future: “The negative forces lack of aligned economic incentives, lack of a coherent national health policy, lack of a business case for quality or value, and continuing pressures on financial bottom lines are simply too strong” (Shortell, Gillies, Anderson, Morgan-Erickson, & Mitchell, 2000). So what has changed? Could it be that the real test of the performance of more organized physician groups and more integrated delivery systems will be afforded by the Accountable Care Act (ACA)? While primarily expanding financial access to care by providing insurance coverage to approximately 32 million Americans, the Affordable Care Act (ACA) also creates new payment incentives for rewarding lower cost/higher quality (i.e., greater value) care on the part of providers. These new payment models are associated with the development of Accountable Care Organizations (ACOs) and Patient-Centered Medical Homes (PCMHs) in which providers have predetermined risk-adjusted expenditure targets for a given population of patients. Under one-side risk models they share only in any savings generated and are not accountable for losses. But under two-sided risk models they are accountable for both savings and losses. The goal is to move away from fee-for-service payment and toward bundled payments such as paying a lump sum to both hospitals and physicians for given

Commentary on Lawton Robert Burns, Jeff C. Goldsmith, and Aditi Sen

121

conditions such as total hip replacements, total knee replacements, and coronary artery bypass graft surgery. Other models include episode of carebased payment for conditions such as diabetes, capitation based on permember per month payments for all or part of a patient care, and overall risk-adjusted global budgets. In these new payment models, providers are rewarded for keeping people well and the hospitals’ revenue-generating business model changes from maximizing the in-patient margin to maximizing the total margin, involving potentially the entire continuum of care. “Value, not volume” has become the rallying cry for many hospitals and physician organizations across the country. Given the shared economic alignment and interdependencies, hospitals and physicians are working on developing the capabilities to successfully respond to the new incentives. These capabilities include, among others, electronic health record functionality, continuous quality improvement processes for both internal improvement and external performance reporting, care management transition tools, reductions in hospital-acquired infections, reducing readmissions, developing the ability to work in teams, and care for patients with multiple complex chronic illnesses. Tools are needed to advance clinical integration across the continuum of care. The organizational arrangements of physicians and hospitals that have been studied in the past provide the structure and platform for more coordinated patientcentered care to occur. But, until recently, such care has not been rewarded, and there has been little incentive for providers to invest in the very difficult work of developing the capabilities to provide such care. It should be no surprise that the myriad studies conducted to date show equivocal results at best.

FUTURE RESEARCH So where should future research focus? First, there is a need to examine the existing and new organizational arrangements in association with their payment incentives and payment models. We should not attempt to disentangle the financial model from the organizational model because they go hand in hand and need to be examined as an “innovation package” rather than being concerned about “multiple treatment interference.” But we do need fine-grained analysis of the contracts provider organizations such as ACOs have signed with Center for Medicare and Medicaid Services (CMS) and private insurers. We need to know what percentage of revenue is at risk

122

STEPHEN M. SHORTELL

and what specific rewards, sanctions, and performance criteria are involved. These data then need to be linked with the provider organization models. A simple but informative hypothesis is that the greater the percentage of patient revenue or “enrolled/attributed lives” at financial risk, the greater the likelihood that more integrated care systems with more organized groups of physicians will provide higher quality/lower cost care and greater patient-experience than less-integrated care systems with more loosely organized physicians. Second, and related to the first, is to recognize that despite the national legislation there will be great variation in ACA implementation by state. Some states such as California, Massachusetts, Minnesota, Oregon, Vermont, and Wisconsin are further along in their reform efforts than others and can serve as “hotbeds” for studying innovative payment and delivery system arrangements (Scheffler et al., 2013; Song et al., 2012). We are likely to learn more rapidly by studying new provider arrangements in these kinds of states rather than expecting that the health care organizational landscape of America will change rapidly across the country. Third, we need to expand our use of mixed methods by placing greater emphasis on qualitative methods that look “under the hood” of the physician hospital and integrated delivery system arrangements to learn about what they are actually doing to provide higher quality/lower cost and better patient-experience care. We suspect that leadership, culture, information management, quality improvement methods, and innovative care management tools play key roles here. Further, given that ACOs and related arrangements require building effective partnerships among previously largely separate organizations, a premium will be placed on relational coordination issues (Gittell, Godfrey, & Thistlewaite, 2013). These include assessing the extent of shared goals among the parties involved, shared knowledge, and mutual respect, supported by frequent, timely, accurate, problem-solving communication. Particularly needed are longitudinal (3 + years) comparative case studies. Such research is needed to capture the changes that organizations make over time as they attempt to address the challenges of fundamentally changing the way in which care is delivered. A fourth area of research is to examine the potential negative anticompetitive effects of increased consolidation within the industry that Burns, Goldsmith and Sen (2013) and others raise (Berenson et al., 2012; Robinson, 2004). New, innovative approaches to mitigating the potential harmful effects of increased consolidation and those allowing for the potential greater value of such arrangements need to be developed and examined (Scheffler, Shortell, & Wilensky, 2012).

Commentary on Lawton Robert Burns, Jeff C. Goldsmith, and Aditi Sen

123

Finally, we need to pay attention to the unit of analysis and to the importance of subgroup analysis. To return to the title of the Burns, et al. paper, the “tails” of the distribution may not lend themselves to telling a very clear “tale” in the future. For example, many ACOs are made up of small practice units in addition to larger entities. While most include hospitals, others do not. Various combinations of federally qualified health centers (FQHCs), nursing homes, and home health agencies are also included or not. In examining cost, quality, and patient-experience performance, most analyses will be conducted at the level of the ACO overall. This means that there needs to be a fine-grained description of the ACO and its various components. There is need to develop a taxonomy of ACOs based on clusters of similar attributes that can be used in various analyses. Conversely, there will also be a need to examine some of the individual components of the ACO and their performance. For example, a specific ACO-affiliated hospital may have done an outstanding job preventing preventable hospital readmissions but this may not be reflected in the ACO’s overall performance. Further, in addition to examining overall performance across all patients and conditions, studies should also focus on specific patient populations and conditions where improvement is most needed and/or expected. Examples include the dual-eligible population, those with complex multiple chronic illnesses, and those with diabetes, cancer, cardiovascular disease, and depression.

CONCLUSION Much of the above discussion is based on the assumption that the changes set in place by the ACA are permanent rather than transitory. This may not be the case. But even if key elements of the legislation are removed or blunted in their intended effect, ongoing forces in the United States to improve quality, lower the rate of growth in costs, and improve population health remain. Moving toward arrangements that provide greater value will vary greatly across the country and will take time. In the next 10 15 years, we may be witnessing a migration from the past decades of a “culture of sickness” to the current emphasis on a “culture of care,” to a possible future emphasis on creating a “culture of health.” However, this will require payment and organizational arrangements that create markets for health in which the winners will be rewarded for keeping people well. In such a world, we move from the hospital as the “physician workshop” of the past and

124

STEPHEN M. SHORTELL

from the hospital as the “physician sweatshop” of the present to the hospital as the physician “backstop” of the future. In brief, if all else fails, the hospital becomes the last resort. From a research perspective, the opportunity to examine this transition bodes well as we develop even better risk adjustment measures of quality and health status, implement all-payer claim databases to examine total cost of care, and work in interdisciplinary teams of clinicians and social scientists, using increasingly sophisticated quantitative and qualitative mixed methods. One unobtrusive measure of the extent to which this transition may be occurring would be the appearance of a billboard showing an empty hospital bed and the following byline: “Visionary Health System: Working to Keep You Out of Bed!” The empirical question is whether “Visionary Health System” meets the triple aim of higher quality, lower cost, and better population health than others in their community, and whether or not that is in any way related to the degree of clinical integration achieved across the continuum of care associated with the various organizational arrangements within the system.

REFERENCES Berenson, R. A., Ginsburg, P. B., Christianson, J. B., & Yee, T. (2012). The growing power of some providers to win steep payment increases from insurers suggests policy remedies may be needed. Health Affairs (Millwood), 31(5), 973 81. Burns, L. R., Goldsmith, J. C., & Sen, A. (2013). Horizontal and vertical integration of physicians: A tale of two tails. In J. Goes, G. T. Savage & L. Friedman (Eds.), Annual review of health care management: Revisiting the evolution of health systems organization (Vol. 15, pp. 39 117). Advances in Health Care Management. Emerald Group Publishing Limited. Gittell, J. H., Godfrey, M., & Thistlewaite, J. (2013). Interprofessional collaborative practice and relational coordination: improving healthcare through relationships. Journal of Interprofessional Care, 27(3), 210 213. Robinson, J. C. (2004). Consolidation and the transformation of competition in health insurance. Health Affairs (Millwood), 23(6), 11 24. Scheffler, R. M., Bowers, L. G., Fulton, B. D., Connors, C., Shortell, S. M., & Morrison I. (2013, February). A new vision for California’s healthcare system: Integrated care with aligned financial incentives. Berkeley, CA: The Berkeley Forum. Scheffler, R. M., Shortell, S. M., & Wilensky, G. R. (2012). Accountable care organizations and antitrust: Restructuring the health care market. JAMA, 307(14), 1493 1494. Shortell, S. M., Gillies, R. R., Anderson, D. A., Morgan-Erickson, K., & Mitchell, J. B. (2000). Remaking healthcare in America: The evolution of organized delivery systems (2nd ed.). San Francisco, CA: Jossey-Bass. Song, Z., Safran, D. G., Landon, B. E., Landrum, M. B., He, Y., Mechanic, R. E., … Chernew, M. E. (2012). The ‘Alternative Quality Contract,’ based on a global budget, lowered medical spending and improved quality. Health Affairs (Millwood), 31(8), 1885 1894.

THE EVOLUTION OF INTEGRATED HEALTH CARE STRATEGIES Jenna M. Evans, G. Ross Baker, Whitney Berta and Jan Barnsley ABSTRACT Purpose To examine the evolution of health care integration strategies and associated conceptualization and practice through a review and synthesis of over 25 years of international academic research and literature. Methods A search of the health sciences literature was conducted using PubMed and EMBASE. A total of 114 articles were identified for inclusion and thematically analyzed using a strategy content model for systems-level integration. Findings Six major, inter-related shifts in integration strategies were identified: (1) from a focus on horizontal integration to an emphasis on vertical integration; (2) from acute care and institution-centered models of integration to a broader focus on community-based health and social services; (3) from economic arguments for integration to an emphasis on improving quality of care and creating value; (4) from evaluations of integration using an organizational perspective to an emerging interest in

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 125 161 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015011

125

126

JENNA M. EVANS ET AL.

patient-centered measures; (5) from a focus on modifying organizational and environmental structures to an emphasis on changing ways of working and influencing underlying cultural attitudes and norms; and (6) from integration for all patients within defined regions to a strategic focus on integrating care for specific populations. We propose that underlying many of these shifts is a growing recognition of the value of understanding health care delivery and integration as processes situated in Complex-Adaptive Systems (CAS). Originality/value This review builds a descriptive framework against which to assess, compare, and track integration strategies over time. Keywords: Integrated delivery system; integrated care; health system integration; whole systems change; complex adaptive systems

INTRODUCTION For many years health care systems have focused on improving integration between organizations and levels of care. It is widely believed that the greater alignment and synergy achieved through integration enhances quality of care, efficiency, and patient satisfaction. However, there is a general lack of consensus on what integration is, and how it can be achieved. Literature reviews report the use of numerous definitions, models, and strategies for integration (Armitage, Suter, Oelke, & Adair, 2009; Kodner, 2009; Strandberg-Larsen & Krasnik, 2009; Suter, Oelke, Adair, & Armitage, 2009). This lack of conceptual consistency creates confusion among practitioners when selecting integration strategies, and among the research community when comparing research findings and insights. Increasingly, scholars call for the establishment of a common language and framework of integrated care to provide a base for advancing research and practice (Kodner, 2009; Stein & Rieder, 2009). There are over two decades of academic research and practical experiences on integrating elements of health care systems, so this is an apt time to explore how integration conceptions and strategies have evolved. The meanings of concepts such as “health,” “health care,” and “integration” lack rigid boundaries and change in relation to context, time, nature of the problem, and evidence of utility (Rodgers, 2000). As well, strategic change represents an evolutionary and iterative process in which managers learn

The Evolution of Integrated Health Care Strategies

127

from their experiences (Lovas & Ghoshal, 2000; Rajagopalan & Spreitzer, 1997). Much of the early thinking on integration is rooted in mechanistic views of care delivery and systems change (Ackerman, 1992; Charns, 1997; Conrad & Dowling, 1990; Fox, 1989; Shortell, Gillies, & Anderson, 1994). It was believed that integrated health systems could be designed from the top-down through a series of steps that brought various elements of health care delivery together under large, centralized structures. Many of these integration efforts failed, prompting scholars and practitioners to recognize the complexity and nonlinear dynamics inherent in the integration process (Baskin, Goldstein, & Lindberg, 2000; Begun, Zimmerman, & Dooley, 2003; Zimmerman & Dooley, 2001). Seminal papers characterizing health care organizations as Complex-Adaptive Systems (CAS) argue that CAS can self-organize without external control as a function of relationships and interactions among diverse agents; in CAS, the future is unpredictable, so learning and sense-making are more important than structure and control (McDaniel & Driebe, 2001; Plsek & Greenhalgh, 2001; Zimmerman, Lindberg, & Plsek, 1998). These ideas are increasingly being applied to integration efforts and other complex change initiatives to better understand the challenges of integrating and to inform the design and management of new or existing integration initiatives (Dattee & Barlow, 2010; Edgren, 2008; Edgren & Barnard, 2012; Nugus et al., 2010; Tsasis, Evans, & Owen, 2012; Tsasis, Evans, Forrest, & Jones, 2013). In this chapter, we review and synthesize over 25 years of international research to examine the evolution of health care integration concepts and strategies as described in the academic literature. Tracing the development of health care integration strategies can reveal differing viewpoints and applications as well as emerging or outmoded meanings; these insights, in turn, can be used to build a current and common conceptual base for future research and practice (Rodgers, 2000). We identify 14 integration strategies and 6 key shifts in strategy content; we propose that underlying many of these shifts is a growing recognition of the value of understanding health care delivery as a CAS. We conclude with recommendations for researchers, practitioners, and policymakers.

METHODS We situate our review within the literature on strategy content, which aims to understand the subject of strategic decisions (i.e., “what” is decided) by

128

JENNA M. EVANS ET AL.

identifying and examining different types of strategies in relation to strategic organizational issues such as goals, scope, and competition (Fahey & Christensen, 1986). Much of the literature on strategy content relates to organizations functioning in businesses other than health care referring to categories like “new market development,” “product-service refinement,” and “generic/novel” to organize strategic decisions (Andrews, Boyne, Law, & Walker, 2009; Blair & Boal, 1991; Shortell, Morrison, & Robbins, 1985). Our interests, however, are in the content of strategies used to integrate health care services. To guide our review and analysis of the literature on health care integration, we thus developed four strategy content questions that reflect the key decisions that shape integration strategies: (1) Which organizations and services are targeted for integration? (2) What are the desired outcomes of integration? (3) How will integration be achieved? (4) When, where, or for whom does integration add the most value? We felt that these questions represent the main strategic issues likely to generate discussion and disagreement among health system planners and decision-makers; strategic conflicts typically occur over ends (what), means (how), philosophy (why), location (where), timing (when), and impacts on stakeholders (who) (Bryson, 1988). We hypothesized that a model capturing these potential points of conflict would enable us to identify shifts in health care integration strategies and associated conceptualization and practice. The underlying assumption is that if the four questions represent key strategic issues that generate discussion and disagreement, they may also represent the areas where dominant thought and research have shifted over time. The review is based on a search of the health sciences literature indexed in PubMed and EMBASE. Although reports on the concept of integration in health care can be traced back several decades, only in the mid-1980s did substantial numbers of academic articles emerge on this topic. As such, only peer-reviewed papers published in English between 1985 and 2013 were included. Commentaries, editorials, and books or book chapters were excluded. We focused the review on papers addressing integration in the United States (US), Canada, Australia, New Zealand, the United Kingdom (UK), and other Western European nations, as these are jurisdictions with the longest history of integration efforts from which we stand the most to learn. Search terms included integrated/organized delivery system, health care/services/systems integration, and integrated health care/services/system/ delivery. The review aims to identify major shifts in thinking about integration at the systems level. We define “system” as encompassing multiple sectors,

The Evolution of Integrated Health Care Strategies

129

organizations, and professionals involved in the delivery of health care services. This criterion excluded articles focused exclusively on integrating specific programs or services (i.e., cancer care, mental health) and those reporting on integrating care for a specific population group (i.e., children, frail elderly). We also excluded articles focused exclusively on integration enablers (i.e., health informatics, physician integration, governance models) or integration tools (i.e., disease management, joint care plans) as our interests were in pursuing our questions at the strategic level, as opposed to the operational. Because the review is interpretive and emphasizes variations in the conceptualization of integration, criteria for inclusion and exclusion as outlined above are based on relevance not methodological quality; this standard allowed for the inclusion of descriptive or discussion papers and theoretical explorations. To be included, papers had to discuss or examine integration in sufficient depth to clarify or further develop understanding of its perceived meaning and application. Information pertaining to the four strategy content questions was extracted from each of the articles selected for inclusion, and synthesized through an iterative process into 14 categories of strategic approaches to integrating care (Table 1). This process involved three steps: (1) recording integration strategies (i.e., responses to the four strategy content questions), retaining the authors’ own words to the extent possible; (2) grouping similar integration strategies together under preliminary categories; and (3) continuously revising the categories and their definitions to improve clarity and applicability. The categories, thus, emerged from an analysis of the characteristics and foci of the articles in relation to the four strategy content questions we posed to frame the review. The 14 categories identified are intended to be flexible and broad to accommodate the heterogeneity of international approaches to integration, while also providing a meaningful set of concepts for high-level comparisons and tracking of integration strategies over time. The categories associated with each question are not mutually exclusive. Following the development of the strategy content categories, the articles were reviewed a second time and coded based on the extent to which each article: (a) promotes, emphasizes, or reports high prevalence of the strategy, (b) argues that the strategy is insufficient (but is, or may be, necessary or beneficial), (c) discourages or reports a decline in the strategy, or (d) does not address the strategy at all or in any depth. Coding was based on a combination of three factors (in order of increasing importance): number of times the strategy was mentioned, amount of space devoted to discussing the strategy, and the tone and type of language used to describe the strategy

130

JENNA M. EVANS ET AL.

Table 1. Defining the Contents of Health Care Integration Strategies. Strategy Content

Definition

Which organizations and services are targeted for integration? Same level Organizations that provide similar care services (e.g., hospital hospital relationships and physician group relationships; also called horizontal integration) Different level Organizations that provide different care services (e.g., hospital physician and hospital homecare relationships; also called vertical integration) Institution-centered Organizations that provide full-time in-house monitoring and care to patients; the focus tends to be on addressing, managing, or preventing acute care needs (e.g., acute care and specialty hospitals as well as long-term care homes)a Community-based Organizations that provide health and social services to clients within their homes or communities; the focus tends to be on primary care, the management of chronic diseases or special needs, and total wellness (e.g., homecare agencies, community support service agencies, and mental health agencies) What are the desired outcomes of integration? Economic Minimizing costs or increasing profits through control of resources, efficiencies (economies of scale), marketing, and market share; references to quality are limited or quality is viewed primarily as a means to improve economic performance Value with a focus Cost-effectiveness or health outcomes per dollar spent with an emphasis on quality on the quality of patient care; better quality and better patient outcomes are an end in themselves, but are also inherently less costly to the health care system Organizational level Evaluation and measurement of integration reflects an organizational focus (e.g., structural and process indicators, financial outcomes, and managers’ and clinicians’ perceived levels of integration) Patient level Evaluation and measurement of integration reflects a patient focus (e.g., patient experiences of coordination and patient-centeredness, health outcomes, functional status, and quality of life) How will integration be achieved? Organizational Changes to organizational structures and/or the relationships between change organizations/providers; examples include mergers, acquisitions, partnerships, governance, size, location, geographic distribution, hierarchy, centralization, scope of services, specialization, and staffing patterns Process change Changes to operational activities, including the way health services are delivered; examples include teamwork, communication, incentives, training, use of technology, use of best practices, standards, and protocols, case management, and patient/caregiver engagement Policy change Changes to system-level public policies; examples include laws, regulations, and decisions regarding the funding, management, and organization of the health system

The Evolution of Integrated Health Care Strategies

Table 1.

131

(Continued )

Strategy Content

Definition

Cultural change

Changes to the shared norms, values, attitudes, assumptions, and beliefs that guide the behaviors and interactions of professionals and organizations; examples include leadership and change management initiatives

When, where, or for whom does integration add value? Defined regions All patients within a geographic boundary Targeted Specific groups of patients based on age, gender, socioeconomic status, populations disability, disease, health needs, or other demographic or lifestyle characteristics a

Many early vertical integration initiatives were hospital-driven with relationships between hospitals and other providers, such as physicians, fueled by a desire to control and promote hospital referrals. We classify these relationships as examples of institution-centered or institutiondriven vertical integration.

(i.e., negative vs. positive). The coding scheme was applied to 25% (n = 30) of the papers (random sample) by one of the authors and by a group of five naı¨ ve coders, each of whom coded six papers. Disagreements were resolved by consensus, and minor modifications were made to the coding scheme to improve clarity. The revised coding scheme was applied to all of the papers by the main coder. The results were organized in chronological order based on publication date and analyzed to identify any shifts in relative emphasis over time for each of the four strategy content questions.

RESULTS The search strategy produced over 2,000 bibliographic records after the removal of duplicates. Through a screening process, outlined in Fig. 1, 114 papers were ultimately included in the review; this includes a total of five relevant articles identified through a second search of the databases for the period January March 2013. The majority of papers excluded based on their title and abstract lacked a focus on system-level integration, while the majority of papers excluded based on a full-text review failed to address any one of the strategy content questions in any depth. The final set of papers hailed from several jurisdictions including the US (43%), Europe (excluding the UK) (26%), Canada (18%), the UK (10%),

132

JENNA M. EVANS ET AL.

Databases: PubMed and EMBASE Search terms: integrated/organized delivery system, health care/services/systems integration, integrated health care/services/system/delivery Articles identified after removal of duplicates (n=2,235) Articles excluded based on title and abstract (n=2,093) Articles identified for full review (n=142) Articles excluded based on full text review (n=46)

Articles added through hand search (n=12)

Articles added from second database search for period January to March 2013 (n=5) Included articles (n=114)

Fig. 1.

Article Screening Process.

New Zealand (2%), and Australia (1%). Less than half of the papers were empirical studies (42%). Papers ranged from conceptual or theoretical explorations (15%) and descriptions of integration efforts, characteristics, barriers, and enablers (60%), to evaluation approaches and exercises (25%). The majority of papers were published post-2000 (72%). A summary of all 114 articles is presented in Table 2 in chronological order. Analysis of the selected literature suggests six important shifts in health care integration strategies, and associated conceptualization and practice. Although the shifts overlap and are interrelated, they are separated here for the purpose of clarity. The six shifts are as follows: 1. From a focus on horizontal integration to an emphasis on vertical integration. 2. From acute care and institution-centered models of integration to a broader focus on community-based health and social services. 3. From economic arguments for integration to an emphasis on improving quality of care and creating value.

++

++

++

Brown et al. (1990)

++

++ ++

++

++

+

+

+

++

Leatt et al. (1996)

Miller (1996)

++

+

Shortell et al. (1994)* Goldstein (1995)

Conrad et al. (1996) Griffith (1996)

∼

++ ++

+

∼ ++

++

∼

++

+

Shortell et al. (1993)* Devers et al. (1994)* Goldsmith (1994)

+

∼

++

++

∼

++ ++

++ +

++

+

Gillies et al. (1993)*

Conrad (1993)

Ackerman (1992) Arndt et al. (1992)

∼

+

++

++

+

Fox (1989)

Conrad et al. (1990)

++

++

Shortell (1988)

++

++

++

Brown et al. (1986)

Instit based

Diff level

+

+

+

+

Comm based

Which Organizations and Services are Targeted for Integration? (n=110, 45 Empirical)

++

++ ++

∼

∼

+

++ ++

++

+

++

∼

++

Econ

+

++

++

++

++

+

+

++

+

++

Value (qual)

++

++

++

Org level

How will Integration be Achieved? (n=113, 46 Empirical)

+

∼

++

++ ++

∼

++

++ ++

++

++

++

++ ++

++

++

++

∼

++

+

+

++ ++

++

++

++ +

++

++

++

++

++

+

++

++

+

++

+

++

+

+

+

Pt Org Proc Poli Cult level

What are the Desired Outcomes of Integration? (n=94, 36 Empirical)

++

++

+

+

++

++

++

++

++

++

++

Defined regions

+

Targeted pop’s

When, Where, Or for Whom does Integration Add Value? (n=80, 26 Empirical)

The Evolution of Health Care Integration Strategies in the Academic Literature, 1985 2013 (n = 114).

Same level

Referenceª

Table 2. The Evolution of Integrated Health Care Strategies 133

++ ++

∼

++

∼

++

++

++

Paulus et al. (2000)

+

++ +

+

++ ++

++

+

+

++ +

∼

++

++

Leatt et al. (2000)

++

++

+

Mur-Veeman et al. (1999)* van Raak et al. (1999) Bazzoli et al. (2000)*

++

+

+

+

∼ ∼

+

∼

++

∼

+

∼ ∼

++

++

++

++

+

++ ++

+

+

++

++

++

+

+ ++

+

+

++

∼

++

++

++

++

+

+

++

++

++

++

++

+

+

+

+ +

+

+

++

++ ++

Defined regions

++

++

+

++

++

Targeted pop’s

When, Where, Or for Whom does Integration Add Value? (n=80, 26 Empirical)

134

++

+

++ ++

+

++

++

Leutz (1999) Lin et al. (1999)*

∼

Hardy et al. (1999)

∼

++

++

+

+

++ ++

∼

++

Pt Org Proc Poli Cult level

++

++

++

Org level

++

∼

++

++

Value (qual)

Goldsmith (1998)

+

++

++

++

Econ

Fischer et al. (1998)

Barnsley et al. (1998)

∼

+

+ +

Comm based

How will Integration be Achieved? (n=113, 46 Empirical)

+

∼

+ ∼

Instit based

What are the Desired Outcomes of Integration? (n=94, 36 Empirical)

++

+

++

++

∼

++

++

++

Young (1997)

Diff level

Same level

Which Organizations and Services are Targeted for Integration? (n=110, 45 Empirical)

Leggat et al. (1997)

Charns (1997) Gillies et al. (1997)

Robinson et al. (1996) Walston et al. (1996)

Referenceª

Table 2. (Continued )

.

JENNA M. EVANS ET AL.

+ ++

++

++ ++

+ ++ ++

++

++

++

Plochg et al. (2006)*

++

Fleury (2006)*

++

++

∼

∼

+

+

∼

++

++

+

++

van Raak et al. (2005)* Yu et al. (2005)*

++

++

Triska et al. (2005)*

∼

+

++

++

++

Ahgren et al. (2005)* +

++

++

+

+

∼

+

+ ++ ++

+

Burns et al. (2005)*

+

++

∼

++ ++

++

+

++

+

+

++

+

∼

++

++

∼

++

∼

+

+

++

∼

∼

∼

∼ ++

++

∼

+

++

++ ++

+

++

+

∼

++ ∼

+

++

++

+

∼ +

+

++

∼ ∼

∼

+ +

++

+

+

+

++

∼ ∼

+

++

+

++

+

++

∼

+

++ ++

+

+

Browne et al. (2004)*

Wan et al. (2002)*

Niskanen (2002) Plochg et al. (2002)

Lukas et al. (2002)*

Lee et al. (2002)*

Kodner et al. (2002)

+

++

Delnoij et al. (2002)

++

++

++

Burns et al. (2002)

++

∼

Woods (2001)

++

+

++

+

Parker et al. (2001)*

van der Linden et al. (2001)* Wan et al. (2001)* Wang et al. (2001)*

++

∼

++

Friedman et al. (2001) Grone et al. (2001)

+

++

++

de Jong et al. (2001)

+

++

Burns et al. (2001)*

++

+

+

++

++

++

++

++

++

++ ++

++

+

++

+

++

+

++

++

++

++

++

++

+

∼

+

++

++ ++

+

++

++

++

++

++

+

+

++

+

+

+

+

+

++

+

+

++

++

+

+

++

+

++ +

+

+

++

++

++

++

++

++ ++

++

++

++

++

+

++

+

The Evolution of Integrated Health Care Strategies 135

+

+

+

Rea et al. (2007)

Suter et al. (2007)*

+ ++

+

++

Kodner (2009)

+

+

++

++

++

++

++

Enthoven (2009)

++

++

Thaldorf et al. (2007) Hollander et al. (2008)* Mur-Veeman et al. (2008)* Strandberg-Larsen et al. (2008)* Armitage et al. (2009)* Denis et al. (2009)*

+

++

+

+

++

+

+

Comm based

++

+

∼

++

+

+

++

Lega (2007)

++

++

++

Instit based

Diff level

Browne et al. (2007) Kautz et al. (2007)*

Same level

Which Organizations and Services are Targeted for Integration? (n=110, 45 Empirical)

Van Wijngaarden et al. (2006)* Ahgren et al. (2007)*

Referenceª

∼

∼

Econ

+

++

+

+

+

+

++

Value (qual)

++

++

++

++ ∼

Org level

How will Integration be Achieved? (n=113, 46 Empirical)

+

+

++

++

++

∼

+

+

++

∼

+

∼

++

++ ∼

+

∼

++

++

+

+

++

++

++

++

++

++

++

++ ++

++

++

+

∼

+

++

∼

++

+

∼

+

∼

+

++

++

+

++

++

+

++

++

++

Pt Org Proc Poli Cult level

What are the Desired Outcomes of Integration? (n=94, 36 Empirical)

Table 2. (Continued )

+

++

+

+

++

+

+

Defined regions

++

+

+

+

++

++

+

+

++

++

++

++

++

Targeted pop’s

When, Where, Or for Whom does Integration Add Value? (n=80, 26 Empirical)

136 JENNA M. EVANS ET AL.

++ ++

++ +

++

++

Hudson (2011)

+

++

++

Ham et al. (2011)

Jiwani et al. (2011) Romoren et al. (2011)

++

++

+

++

++

++

++

Glasby et al. (2011)

+

++

Calciolari et al. (2011) Cumming (2011)

++

+

++

+

++

++

+

++

+

∼

+

++

++

++

+

+

+

++

++

++

Bevan et al. (2011)

Rumbold et al. (2010) Strandberg-Larsen et al. (2010)* Ahgren et al. (2011)

Williams et al. (2009)* McCarthy et al. (2010) Pate et al. (2010)*

Strandberg-Larsen et al. (2009)* Suter et al. (2009)*

Stein et al. (2009)

Minkman et al. (2009a)* Minkman et al. (2009b)* Solberg et al. (2009)

+

∼

∼

+

++

+

+

+

+

++

++

+

++

+

+

+

++

∼

++

++

+

+

++ ∼

∼

++

∼

∼

∼

∼

+

∼

+

∼

++

∼

++

+

++ ++

++

++

++

++

++

++

++

++

+

+

++

++

++

++

++

++

++

+ ∼

++

+

+

+

++

++

++

++

++

++ ∼

∼

++

++

∼

+

∼

++

++

+

++

++

++

+

+

++

+

+

++

++

++

+

+

++

++ ++

+

++

+

+

+

+

++

+

++

+

+ ++

+

++

+

+

++

++

++

+

++

+

++

++

The Evolution of Integrated Health Care Strategies 137

+

++ ++

++ ++

++

++

Ye et al. (2012)*

Demers (2013) Harris et al. (2013)*

∼ +

++

Williams (2012)* +

+

∼ ∼

∼

+

++

+

∼

∼ +

∼

+

+

++

++

∼ ++

+

++

++

+ +

++

++ ++

++

++

++

++

++

++

++

++ ++

∼

++

++

++

+

+

++

++

++

∼

+

++

+

++ ++

++

++

++

++

++

++

++

++

++

+

Pt Org Proc Poli Cult level

++

++

∼

+

+

Org level

Tsasis et al. (2012)*

+

++

++

++

+

+

+

++

++

Value (qual)

How will Integration be Achieved? (n=113, 46 Empirical)

Rudkjobing et al. (2012) Sullivan et al. (2012)

++

∼

∼

++

+

+

Econ

What are the Desired Outcomes of Integration? (n=94, 36 Empirical)

++

+

Lillrank (2012) Ling et al. (2012)*

Kreindler et al. (2012)* Pourat et al. (2012)*

++

++

Evans et al. (2012)

Edgren et al. (2012)

++

++

Burns et al. (2012)

++

Comm based

+

Instit based

++

Diff level

Vedel et al. (2011)

Same level

Which Organizations and Services are Targeted for Integration? (n=110, 45 Empirical)

Singer et al. (2011)

Referenceª

Table 2. (Continued )

+

+

++

++

++

Defined regions

+

++

+

++

++

+ ++

+

+

++

++

++

Targeted pop’s

When, Where, Or for Whom does Integration Add Value? (n=80, 26 Empirical)

138 JENNA M. EVANS ET AL.

+

++ ++

++

++

++

++ ++

+ ++

a All references list first author only. Articles published in the same year are ordered alphabetically. * Empirical paper. ++ Promotes, emphasizes, or reports high prevalence of the strategy (strong focus). + Promotes, emphasizes, or reports high prevalence of the strategy (limited focus). ∼ Argues that the strategy is insufficient (but is, or may be, necessary or beneficial). Discourages or reports a decline in the strategy. Blank space Does not address the strategy at all or in any depth.

Valentijn et al. (2013) Walker et al. (2013)*

Tsasis et al. (2013)*

++

++

++

+ ++

++ ++

++

The Evolution of Integrated Health Care Strategies 139

140

JENNA M. EVANS ET AL.

4. From evaluations of integration using an organizational perspective to an emerging interest in patient-centered measures. 5. From a focus on modifying organizational and environmental structures to an emphasis on changing ways of working and influencing underlying cultural attitudes and norms. 6. From integration for all patients within defined regions to a strategic focus on integrating care for specific populations. The six themes above represent shifts in relative emphasis and do not imply that certain conceptualizations are obsolete or that some strategies are no longer implemented. We describe these shifts in strategy content in further detail below followed by an examination of the contributing factors to change.

Which Organizations and Services are Targeted for Integration? From Horizontal to Vertical Integration From the late 1980s to the mid-1990s interest gradually decreased in horizontal integration efforts involving mergers of organizations providing the same level of care, such as among acute-care hospitals or among rehabilitation hospitals (see Table 2). Multihospital systems initially emerged in the US as a means to garner security and protection in a highly uncertain and complex health care environment; achieving economies of scale and increased market power through system participation was a survival mechanism (Conrad & Shortell, 1996; Shortell, 1988). However, multihospital systems largely failed to demonstrate improved profitability, quality, or access compared with independent hospitals (Burns & Pauly, 2002; Shortell, 1988). The lack of improved outcomes has been attributed to a narrow focus on administrative economies with little consideration for the integration of medical care (Shortell et al., 1994). By the early to mid-1990s, the vertical integration of organizations providing different levels of care became the focus of restructuring efforts with hospitals acquiring physician practices, long-term care homes, and medical equipment companies in addition to other hospitals. This integration strategy, which has persisted to date (see Table 2), was driven by environmental conditions, most notably the introduction of managed care in the US which increased interdependencies among organizations at different levels in health care delivery, as well as desires to control increasing costs, respond to changing patient needs, and acquire legitimacy (Ackerman, 1992; Arndt &

The Evolution of Integrated Health Care Strategies

141

Bigelow, 1992; Burns & Pauly, 2002; Conrad & Dowling, 1990; Walston et al., 1996). Yet, like their multihospital predecessors, hospital-led vertically integrated systems experienced few economies of scale due to higher transaction costs, limited coordination of medical care, and slower decisionmaking (Burns & Pauly, 2002; Goldsmith, 1994, 1998). Although evidence of sustained financial or clinical value in horizontally and vertically integrated systems referred to as Integrated Delivery Systems was scarce, belief in the logic of coordinating functions and services across the continuum of care persisted. Alternative strategies began to emerge seemingly in response to the disappointing performance outcomes of these early integration efforts. From Institution-Centered to Community-Based Integrated Care In the last decade there has been an increasing focus on integrating more broadly across various levels of medical care and across public health, social services, and related supports such as education and housing at the community level (Browne et al., 2004; Hollander & Prince, 2008; Kodner & Spreeuwenberg, 2002; Plochg & Klazinga, 2002; Singer et al., 2011). These multi-sectoral approaches recognize the interdependencies between social conditions and health status, and demonstrate deeper consideration for the roles of informal, self-directed, and culturally sensitive care in the delivery of integrated services. This shift may be linked with the emergence of increasing evidence regarding the impact of nonmedical determinants on individual and population health, and associated changes to traditional definitions of what constitutes health and health systems (Arah, Westert, Delnoij, & Klazinga, 2005; Raphael, 2009). There has been significant movement away from conceptualizations in which acute care hospitals are the nucleus of the system (Burns & Pauly, 2002; Hollander & Prince, 2008). An acute care focus initially dominated because hospitals and hospital chains pursued integration strategies (Brown & McCool, 1990; Conrad & Shortell, 1996; van der Linden et al., 2001). As such, system leaders were encouraged to select integration pursuits based on the “needs of the acute care business” and not necessarily population health needs or health system objectives (Fox, 1989, p. 53). However, Table 2 displays an increasing emphasis on linking various home and community-based facilities and services in addition to improving transitions to and from hospitals, with primary care as the hub (Browne et al., 2004; Burns & Pauly, 2012; Cumming, 2011; Niskanen, 2002; Pourat et al., 2012; Valentijn et al., 2013; Williams & Sullivan, 2009).

142

JENNA M. EVANS ET AL.

Expert opinion and empirical evidence suggest that those systems built around primary care demonstrate superior performance as compared to systems built around hospitals (Burns et al., 2005; Goldsmith, 1994; Robinson & Casalino, 1996). Conrad and Shortell (1996) argue that hospitals must become “peripheral organizations” within the system (p. 36), and Burns and Pauly (2002) state that hospitals will have a role, but “by no means the leading role” in linking services (p. 140). Recent reform efforts like Accountable Care Organizations (ACOs) and the Patient-Centered Medical Home (PCMH) emphasize incentives for quality and efficiency, performance-based accountability, and a strong primary care foundation respectively. Experts suggest combining ACOs and PCMHs to create “medical neighborhoods” that support integration and accountability between clinicians and services within PCMHs, and those that fall outside of those boundaries, including nonmedical providers (Fisher, 2008; McClellan, McKethan, Lewis, Roski, & Fisher, 2010; Pham, 2010; Rittenhouse, Shortell, & Fisher, 2009; Walker et al., 2013).

What are the Desired Outcomes of Integration? From Economic to Quality Arguments for Integration Both horizontal and vertical integration strategies emerged in the health care sector for economic reasons, including potential economies of scale, market domination, increased profits, and ultimately better prospects for survival (Thaldorf & Liberman, 2007). The advantages of integration were initially outlined in terms of efficiencies with quality improvement framed as a means to achieve economies of scale or as a potential secondary benefit (Ackerman, 1992; Conrad & Dowling, 1990; Fox, 1989; Goldsmith, 1998; Walston et al., 1996). Some health care views on economic motives may have been rooted in business literature and practices which depicted vertical integration as a rational economic response by organizations to a changing environment (Fox, 1989). Indeed, the dominance of financial performance measures and economic arguments for horizontal and vertical integration is a product of the American policy context in the early to mid-1990s when health service organizations were under growing pressure to perform better financially (Friedman & Goes, 2001). Many health care systems outside of the US also initially adopted integration strategies with the primary aim of minimizing or controlling costs (Ahgren & Axelsson, 2011; Jiwani & Fleury, 2011; Leatt et al., 2000; Niskanen, 2002; Rea et al., 2007). Challenges in identifying and measuring meaningful quality outcomes may

The Evolution of Integrated Health Care Strategies

143

have also contributed to a reliance on financial measures to assess performance (Friedman & Goes, 2001). However, the focus on potential economic benefits of integration in the literature has expanded to a dual emphasis on both efficiency and quality of care, as Table 2 depicts. This shift in thinking may have occurred for two reasons. First, the increasing “corporatization” of health care management practices in the US fueled fear among some observers that quality was at risk, resulting in demands for greater patient protection and public accountability (Shortell, Gillies, Anderson, Erickson, & Mitchell, 2000). Second, by the mid-1990s evidence began to emerge that successfully integrating policy, staff, funding, and clinical processes requires substantial investments that may result in improved quality of care, but not necessarily efficiencies, particularly in the short term (Burns et al., 2005; Leutz, 1999). While some papers reported, or suggested, improved financial performance as a result of integration (e.g., Wan et al., 2002), others illustrated negative, mixed, or no impacts (Bazzoli et al., 2000; Burns et al., 2005; Fischer & Coddington, 1998; Lee & Wan, 2002). The inconsistency in evidence may be caused by implementation difficulties and methodological challenges, such as conceptual ambiguity, contextual differences, and a lack of longterm studies (Stein & Rieder, 2009). Since the early 2000s, there is a growing recognition that efficiency and quality are linked and must be addressed simultaneously to create enduring value for patients, providers, and purchasers. For example, eliminating waste resulting from unnecessary and unsafe care can result in both quality and efficiency improvements (Mazzocato, Savage, Brommels, Aronsoon, & Thor, 2010; Porter & Teisberg, 2006); a quality-centered culture also influences multiple dimensions of performance, including financial outcomes (Shortell et al., 2005). As such, the value of integration is increasingly being defined by arguments for enhancing the quality of care delivered to patients with particular attention to patient-centeredness (see Table 2) (Ahgren & Axelsson, 2011; Kautz et al., 2007; Kodner & Spreeuwenberg, 2002; Kreindler et al., 2012; McCarthy & Zuckerman, 2010; Minkman et al., 2009a; Valentijn et al., 2013). Failure to develop consensus on the purposes of health systems integration among managers, policy-makers, clinicians, and patients can hinder efforts to secure cooperation at all levels (Friedman & Goes, 2001; Stein & Rider, 2009). The shift toward quality arguments for integration may have arisen with the recognition that it is important to align and balance competing motives among various stakeholders in order for integration efforts to succeed.

144

JENNA M. EVANS ET AL.

From Organization-Focused Evaluation of Integration to an Emerging Interest in Patient-Centered Measures Although, there is a reliance on reporting the perceived extent and impacts of integration from the perspectives of managers and clinicians (Strandberg-Larsen & Krasnik, 2009), a growing number of scholars are emphasizing or exploring patient expectations, perceptions, and experiences of integration beyond patient satisfaction (Cumming, 2011; Kautz et al., 2007; Minkman et al., 2009a; Singer et al., 2011; Walker et al., 2013). The definition of integrated care has recently been extended to incorporate the concept of patient-centeredness with measurement focused on the patient experience of integrated care, not underlying organizational structures and processes (Singer et al., 2011; Walker et al., 2013). Table 2 suggests that there is a developing interest in expanding traditional evaluation frameworks to include broader and more complex sets of measures based on the patient perspective. However, 58% of the papers included in the review (n = 66) do not discuss performance measurement or evaluation at all or in any depth, as the blank cells in Table 2 indicate. The limited attention in the literature to measurement and evaluation highlights a gap in knowledge and practice that may be hindering progress toward integrated systems of care. This gap suggests a need for ongoing attention to the development, implementation, and refinement of multilevel performance evaluation and measurement frameworks which incorporate both organizational-focused and patient-centered measures.

How Will Integration be Achieved? From Structural to Operational and Cultural Approaches to Change Initial integration efforts primarily emphasized changes to organizational structures (i.e., mergers, acquisitions, and alliances) with few corresponding modifications to care processes and administrative operations. The reliance on structures to promote integration may stem from the relative ease with which structures can be changed compared with processes embedded in them such as communication, routines, and teamwork. Several studies suggest that structural modifications are inadequate for integrating services and patient care (Burns & Pauly, 2002; Burns et al., 2001; Kautz et al., 2007; Lee & Wan, 2002; Strandberg-Larsen et al., 2010) and many experts attribute integration failures to a narrow focus on bureaucratic mechanisms and a disregard for system functioning (Bevan & Janus, 2011; Demers,

The Evolution of Integrated Health Care Strategies

145

2013; Friedman & Goes, 2001; Hudson, 2011; Shortell, 1988; Tsasis et al., 2012; Williams & Sullivan, 2009; Young, 1997). Similarly, although the broader context (including public policies, laws, and regulations) can positively or negatively impact integration efforts and outcomes (Mur-Veeman et al., 1999; Mur-Veeman et al., 2008), more micro-contextual factors like local management actions including stakeholder engagement, participative leadership, and open communication also play decisive roles in achieving integration goals even if the broader institutional context remains unchanged (Ahgren & Axelsson, 2007; van Raak et al., 1999; Williams & Sullivan, 2009). As shown in Table 2, scholars increasingly suggest that while policy changes to incent and support collaboration may be necessary or beneficial, they are not sufficient for achieving integrated care; in fact, frequent modifications to public policy which result in restructuring or competing incentives can damage existing networks and relationships and focus attention on internal reorganization rather than external relationships (Cumming, 2011; Demers, 2013; Hudson, 2011). In the English National Health Service (NHS), for example, ongoing structural reforms to administrative boundaries and roles paired with policy ambiguity regarding the functions of and relationship between competition and cooperation may have contributed to “increased fragmentation and rivalry across the system” in recent years (Hudson, 2011, p. 4). Similarly, New Zealand implemented a series of reforms over the past decade, including the establishment of District Health Boards and Primary Health Organizations, but reports suggest that these efforts have had limited impact on service delivery (Cumming, 2011). Within the movement away from an exclusive focus on structural interventions lies a sub-shift: an increasing emphasis on more loosely linked relationships like partnerships and alliances rather than mergers and acquisitions. While there was early acknowledgement of the viability of partnership, or “virtual” approaches to integrating services (Brown & McCool, 1986; Conrad & Dowling, 1990; Fox, 1989), ownership models dominated industry through the mid-1990s. Shared ownership was originally assumed to confer the benefits of reducing conflicts of interest among service providers, increasing strategic control and speed in decision-making, and reducing transaction costs (Conrad, 1993; Conrad & Shortell, 1996). However, there is little and often mixed or only indirect evidence to support these assumptions (Bazzoli et al., 2000; Burns et al., 2005; Kautz et al., 2007). In practice, owned integrated systems exhibited more management layers, slower decision-making, higher costs, and wider gaps between system executives and clinicians thus displaying diseconomies of both scale and

146

JENNA M. EVANS ET AL.

coordination (Demers, 2013; Goldsmith, 1994, 1998; Walston et al., 1996; Yu et al., 2005). The political and logistical challenges of shared ownership contributed to struggles for power that diverted attention, resources, and commitment away from the integration vision. The success of earlier established Integrated Delivery Systems, like Kaiser Permanente and the Mayo Clinic, has been attributed to unique advantages that are difficult to imitate or recreate, such as a favorable socioeconomic and political context and early development, allowing for the accumulation of management experience and the development of a coherent culture (Bevan & Janus, 2011; Burns & Pauly, 2002). Evidence also suggests that integration at a corporate level has no relationship with integration at the patient-care level (Kautz et al., 2007; Leatt et al., 2000; Lee & Wan, 2002). In response to the challenges of shared ownership and the prevalence of suboptimal outcomes, there was a shift toward managing, not necessarily eliminating, organizational boundaries (Friedman & Goes, 2001; Glasby et al., 2011; Leatt et al., 2000; Miller, 1996; Robinson & Casalino, 1996; Strandberg-Larsen & Krasnik, 2008; Woods, 2001). Shared information, best practice guidelines, and a common culture can facilitate coordination across boundaries through trust, shared objectives, and recognition of interdependence rather than through control using hierarchy or ownership (Burns & Pauly, 2002; Conrad, 1993; Demers, 2013; Suter et al., 2009). In other words, virtual integration may capture the advantages of ownership models without the associated risks and liabilities. ACOs in the US reflect this shift in thinking. ACOs may adopt diverse structural arrangements and relationships, yet the primary focus of reform efforts is clinical transformation at the process level (Burns & Pauly, 2012; Kreindler et al., 2012; McClellan et al., 2010). Recent studies examine integrative and developmental processes more closely, including teamwork, communication, decision-making, and knowledge exchange (de Jong & Jackson, 2001; Harris et al., 2013; Minkman et al., 2009b; Tsasis et al., 2013; van Wijngaarden et al., 2006; Williams, 2012; Williams & Sullivan, 2009; Yu et al., 2005). Scholars are also increasingly focused on understanding and managing the cultural, cognitive, and power differences between organizations and professional groups as a means to improving integration efforts and outcomes (see Table 2) (Ahgren & Axelsson, 2007; Denis et al., 2009; Evans & Baker, 2012; Kreindler et al., 2012; Pate et al., 2010; Pourat et al., 2012; Sullivan & Williams, 2012; van Wijngaarden et al., 2006; Williams, 2012). For example, Evans and Baker (2012) develop a framework of “mental models of integrated care” which outlines important knowledge and beliefs whose convergence or divergence

The Evolution of Integrated Health Care Strategies

147

across stakeholder groups may influence interorganizational and interprofessional relations. Others echo the need to build common knowledge and perceptions (Ahgren & Axelsson, 2007; Ling et al., 2012; van Wijngaarden et al., 2006; Williams, 2012), in part through training in teamwork and care coordination (Ling et al., 2012; Pourat et al., 2012) as well as through social strategies that engage and unite people working at different levels and across different settings (Denis et al., 2009; Kreindler et al., 2012; Pate et al., 2010). The shift in relative emphasis from context and organizational structures to operational activities and cultural issues does not imply that structural modifications are unnecessary. Modifications to governance, funding, and accountability structures are usually necessary, but not sufficient for achieving integration (Cumming, 2011; Hudson, 2011; Ling et al., 2012; Plochg et al., 2006). Structural changes facilitate the meso- and micro-level work needed to integrate care and encourage widespread implementation, adequate evaluation, and sustainability (Fleury, 2006; Lega, 2007; Minkman et al., 2009b; van der Linden et al., 2001; Valentijn et al., 2013), but Table 2 suggests that operational and cultural changes may be more likely than organizational and policy interventions to directly contribute to the delivery of integrated care.

When, Where, or For Whom Does Integration Add Value? From Integration for All to Integration for Targeted Populations Although integration in some studies is associated with positive outcomes, research evidence as a whole is limited, and in some cases mixed. This ambiguity is due to differences across studies in the integration models and approaches examined, their context, and the evaluation methods used (Armitage et al., 2009; Burns & Pauly, 2002; Stein & Rieder, 2009). While the intuitive potential value of integrating care remains, integration may not be necessary or beneficial in all circumstances and contexts, and may not be desired by some patient groups (Ahgren & Axelsson, 2005; Burns & Pauly, 2002). Studies suggest that successful integration is more likely when provider interdependence is high, situation uncertainty is high, and the services to be integrated have clear boundaries, such as in cardiac care or women’s health (Fleury, 2006; Leutz, 1999). As depicted in Table 2, the emerging method for achieving integration is to adopt a patient-centric approach by integrating care for certain subpopulations or disease groups (Burns & Pauly,

148

JENNA M. EVANS ET AL.

2002; Fleury, 2006; Hollander & Prince, 2008; Kodner, 2009; Kodner & Spreeuwenberg, 2002; Leutz, 1999; Plochg et al., 2006; Rea et al., 2007; Valentijn et al., 2013). Research focused on integrating care within specific sectors or for particular patient populations has a long history. For example, there is a rich and growing literature on integrating care for mental health patients. Papers like these that focused on such specific populations were excluded from inclusion in this review. However, we note that work with a broader emphasis on integration increasingly recognizes the need to identify and engage targeted patient populations as part of system-level integration strategies. There is a growing belief that more focused integration strategies that involve reorganizing care delivery around medical conditions, interventions, or population groups will yield better quality and efficiency outcomes, and offer competitive advantages, when compared with broader, generic approaches to integration (Burns & Pauly, 2002). These benefits may result in part from a reduction in the implementation of costly large-scale integration initiatives and from increasing attention to patient needs and experiences as the drivers of integration. Improving services for complex and vulnerable patients may eventually contribute to system-wide enhancements in integrated care for all patients (Stein & Rieder, 2009). For example, in Sweden, “chains of care” that link health care providers for targeted groups of patients are viewed as prerequisites for the development of broader clinical networks called “local health care” (Ahgren & Axelsson, 2007).

DISCUSSION This review identifies and discusses six shifts over time in the conceptualization and practice of health systems integration as described in the academic literature. Other reviews have focused on defining terms, identifying conceptual distinctions, and outlining various views and approaches, but do not track changes in focus and thinking within the integration literature (Armitage et al., 2009; Kodner, 2009; Strandberg-Larsen & Krasnik, 2009; Suter et al., 2009). Using a strategy content model and a coding scheme, we identified evidence in support of our proposition that conceptualizations and applications of integration have shifted over time (Table 2). The six proposed trends are general themes that provide an overview of health systems integration research over the past 27 years. The trends suggest that many previous connotations of integration are less relevant to current

The Evolution of Integrated Health Care Strategies

149

discussions. Horizontal, acute care and cost-focused, one-size-fits-all, structural approaches no longer represent the recommended methods for integrating services. The academic literature increasingly focuses on integration strategies involving community-based partnerships across the continuum of care, operational and cultural change, quality and patient-centered goals, and targeted patient populations. None of these ideas were completely absent in earlier integration research, but they have become a more central focus in recent years. Three factors are likely contributors to the changes we have observed in integration objectives, desired scope, and proposed methods. First, environmental changes such as aging populations, increasing prevalence of chronic illnesses, and economic pressures have led to increasing demands for integrated care (Grone & Garcia-Barbero, 2001). Although information exchange and interoperability are ongoing challenges, technological advancements in medicine and informatics are providing tools (i.e., video telemonitors and portable patient-friendly biometric devices) for improving the transition to community-based care by making services like telehealth and home care for complex patients more feasible (Darkins et al., 2009), thus reducing the transaction costs of integrated care. Second, there has been growing interest in providing patient-centered care. Researchers, system leaders, and clinicians have been moved to ask “who needs integrated care” and to visualize the patient experience receiving care over time and across providers. The movement from an organizational focus to a patient focus may be influenced in part by increasing use of wellness as opposed to illness models of health (Larson, 1999), a greater emphasis on customer service as a result of the application of business frameworks and tools (Ford & Fottler, 2000), and the growing interest in patient experience as a dimension of quality of care (Berwick, 2009). Finally, research on health systems integration over the past two decades has provided new information regarding the costs, benefits, and challenges of integration. We have a better sense of what integration entails and what works for whom and under what circumstances. Contrary to traditional assumptions, the integration of organizations and their activities to form systems does not necessarily result in the integration of direct patient care (Singer et al., 2011). As such, integration has become less defined by Integrated Delivery Systems, which are one means for promoting integration, and more by the desired output, integrated care. The failure of many early attempts at integration also may have fueled growing recognition of the value of seeing integration through the lens of CAS theory (Baskin et al., 2000; Begun et al., 2003; Dattee & Barlow,

150

JENNA M. EVANS ET AL.

2010; Edgren, 2008; Edgren & Barnard, 2012; Nugus et al., 2010; Tsasis et al., 2012; Zimmerman & Dooley, 2001). Scholars and practitioners are gradually moving away from linear, command-and-control views of system transformation in which change is implemented from the top-down and attention is focused on organizational structures. A CAS lens emphasizes instead the role of relationships and interactions in shaping change efforts; as such, cultural similarities and differences, learning and sense-making processes, and the voices of patients and their caregivers take on more central importance (Edgren & Barnard, 2012; McDaniel & Driebe, 2001). Our findings suggest that there is a growing belief that integration strategies involving community-based partnerships, operational and cultural change, quality and patient-centered goals, and targeted patient populations may be more effective than, or may increase the effectiveness of, more traditional integration methods. CAS theory offers an explanation for why this might be so; replacing top-down control with mechanisms that support dialogue, experimentation, and collaboration from the bottom-up maximize the potential for the system to coevolve and self-organize in ways that positively impact performance (McDaniel & Driebe, 2001; Tsasis et al., 2012). Formal rules, procedures, and structures are typically too rigid and static to effectively address and manage the complex and dynamic nature of change in CAS (Lanham et al., 2013). Agents not only need sufficient structure to allow for information exchange, but also sufficient flexibility to allow for local trial-and-error and mutual adaptation to changing needs and capacities and changing demands and opportunities from the environment (Comfort, 1994). These insights from CAS theory are supported by recent research on whole-systems change in health care, which links success and sustainability to “engaged human agents” and sociocultural and interpersonal factors, rather than hard structures (Best et al., 2012; Greenhalgh, MacFarlane, Barton-Sweeney, & Woodard, 2012, p. 543). Policy-makers, managers, and care providers involved in integrating care may use the four strategy content questions outlined in Tables 1 and 2 to reflect on their current practice and to guide discussions and interactions. Comparing responses to the questions across stakeholder groups may reveal areas of potential conflict that need to be addressed or potential opportunities for securing buy-in or leveraging efforts. This type of dialogue may be particularly important given the increased focus in the literature on the cultural challenges of integration. Given that experiences with integration fail to align with mechanistic views of health care delivery, practitioners may consider applying managerial strategies (Edgren & Barnard, 2012; Lanham et al., 2013; McDaniel & Driebe, 2001; Tsasis et al., 2012)

The Evolution of Integrated Health Care Strategies

151

and scholars may consider developing research designs (McDaniel, Lanham, & Anderson, 2009) rooted in a CAS perspective. Finally, the 14 integration strategies in Table 1 can be used as a descriptive framework of strategy content against which to assess, compare, and track new concepts and approaches emerging in the field. The six proposed shifts are intended to describe and characterize the literature, not prescribe or predict; the evidence base is modest and the shifts are general; therefore, judgments regarding which strategies are best are premature. These trends may also be broken down to examine more specific changes and issues that have emerged over time and within specific contexts. Future studies on integration would benefit from specifying the contextual factors and strategy content of integration initiatives under study, permitting more detailed analyses of how these two factors interact to shape performance outcomes. Describing these details in published reports will enable more objective and accurate comparisons to be made across studies and will help clarify which integration strategies produce the desired outcomes and under what conditions.

LIMITATIONS Changing terms and the interdisciplinary nature of the field suggest that relevant papers may have been missed. Grey literature reports and papers on integration at the program or provider levels were not included, but these may provide further insights for future research. Furthermore, the extent to which the peer-reviewed academic literature accurately captures integration practices as they have unfolded over the years in real world contexts is debatable. For example, although there has been a shift in emphasis in the literature from horizontal, institution-centered, and costdriven integration strategies to vertical, community-based, and qualitydriven integration strategies, in practice horizontal integration schemes, acute care paradigms, and cost-driven policy interventions persist in some contexts (Ahgren & Axelsson, 2011; Hudson, 2011; Jiwani & Fleury, 2011). Professional publications, trade and business magazines, and documents from organizations and systems offer further information on if, how, and to what extent integration strategies and practices have evolved. Trends and changes over time were assessed based on publication year. Considering the lag in published articles and the differences in time to publication from one journal to another, this is an imperfect method. Finally,

152

JENNA M. EVANS ET AL.

many early papers on integrated health systems emerged from the US as a response to the development of managed care; as a result a substantial number of papers in this review are based on the American experience (43%). Nevertheless, the changes in health care integration strategies have transpired along similar trajectories in other national contexts (Delnoij et al., 2002; Grone & Garcia-Barbero, 2001; Hardy et al., 1999; Leatt et al., 2000; Leutz, 1999; Woods, 2001).

CONCLUSION The meanings and focus of health systems integration have shifted over time with the evolution of health care needs and the broader social environment, and with growing experience of the challenges and impact of integration strategies. The six broad shifts that we have described above reflect a broader paradigm shift in health care away from acute inpatient care, the treatment of illness, and responsibility for the individual patient to a focus on integrating services, improving quality of care, and accountability for population health (Shortell & Kaluzny, 2005). Our findings suggest a growing movement away from mechanistic conceptions of health care management and delivery, and an increasing recognition of the value of understanding integration as an agent-based, nonlinear, emergent, selforganizing and coevolving phenomenon (Tsasis et al., 2012). The evolution of integration strategies raises the question, “What new trends will emerge in our quest for integrated care?” In this chapter, we suggest that improving health systems integration requires periodically looking back to understand the development of current strategies and to learn from past experiences and ways of thinking.

ACKNOWLEDGMENTS An earlier and shorter version of this chapter was published in the Academy of Management’s 2013 Best Paper Proceedings. Jenna M. Evans is funded by a Vanier Canada Graduate Scholarship and by the Health System Performance Research Network (HSPRN).

The Evolution of Integrated Health Care Strategies

153

REFERENCES Ackerman, F. (1992). The movement toward vertically integrated regional health systems. Health Care Management Review, 17(3), 81 88. Ahgren, B., & Axelsson, R. (2005). Evaluating integrated health care: A model for measurement. International Journal of Integrated Care, 5. Retrieved from http://www.ijic.org/index. php/ijic/article/view/134/267 Ahgren, B., & Axelsson, R. (2007). Determinants of integrated health care development: Chains of care in Sweden. International Journal of Health Planning and Management, 22(2), 145 157. Ahgren, B., & Axelsson, R. (2011). A decade of integration and collaboration: The development of integrated health care in Sweden 2000 2010. International Journal of Integrated Care, 11. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-101301. Andrews, R., Boyne, G., Law, J., & Walker, R. (2009). Strategy formulation, strategy content and performance: An empirical analysis. Public Management Review, 11(1), 1 22. Armitage, G, Suter, E., Oelke, N., & Adair, C. (2009). Health systems integration: State of the evidence. International Journal of Integrated Care, 9. Retrieved from http://www.ijic.org/ index.php/ijic/article/viewArticle/316/630 Arah, O., Westert, G., Delnoij, D., & Klazinga, N. (2005). Health system outcomes and determinants amenable to public health in industrialized countries: A pooled, cross-sectional time series analysis. BMC Public Health, 5, 81. doi:10.1186/1471-2458-5-81. Arndt, M., & Bigelow, B. (1992). Vertical integration in hospitals: A framework for analysis. Medical Care Review, 49(1), 93 115. Barnsley, J., Lemieux-Charles, L., & McKinney, M. (1998). Integrating learning into integrated delivery systems. Health Care Management Review, 23(1), 18 28. Baskin, K., Goldstein, J., & Lindberg, C. (2000). Merging, de-merging, and emerging at Deaconess Billings Clinic. The Physician Executive, 26(3), 20 25. Bazzoli, G., Chan, B., Shortell, S., & D’Aunno, T. (2000). The financial performance of hospitals belonging to health networks and systems. Inquiry, 37(3), 234 252. Begun, J., Zimmerman, B., & Dooley, K. (2003). Health care organizations as complex adaptive systems. In S. M. Mick & M. Wyttenbach (Eds.), Advances in health care organization theory (pp. 253 88). San Francisco: Jossey-Bass. Berwick, D. (2009). What ‘patient-centered’ should mean: Confessions of an extremist. Health Affairs, 28(4), w555 w565. Best, A., Greenhalgh, T., Lewis, S., Saul, J., Carroll, S., & Bitz, J. (2012). Large-System transformation in health care: A realist review. The Milbank Quarterly, 90(3), 421 456. Bevan, G., & Janus, K. (2011). Why hasn’t integrated health care developed widely in the United States and not at all in England? Journal of Health Politics, Policy and Law, 36(1), 141 164. Blair, J., & Boal, K. (1991). Strategy formation process in health care organizations: A context-specific examination of context-free strategy issues. Journal of Management, 17(2), 305 344. Brown, M., & McCool, B. (1986). Vertical integration: Exploration of a popular strategic concept. Health Care Management Review, 11(4), 7 19. Brown, M., & McCool, B. (1990). Health care systems: Predictions for the future. Health Care Management Review, 15(3), 87 94.

154

JENNA M. EVANS ET AL.

Browne, G., Kingston, D., Grdisa, V., & Markle-Reid, M. (2007). Conceptualization and measurement of integrated human service networks for evaluation. International Journal of Integrated Care, 7. Retrieved from http://www.ijic.org/index.php/ijic/article/view/214/426 Browne, G., Roberts, J., Gafni, A., Byrne, C., Kertyzia, J., & Loney, P. (2004). Conceptualizing and validating the human services integration measure. International Journal of Integrated Care, 4. Retrieved from http://www.ijic.org/index.php/ijic/article/view/ 98/195. Bryson, J. (1988). A strategic planning process for public and non-profit organizations. Long Range Planning, 21(1), 73 81. Burns, L., & Pauly, M. (2002). Integrated delivery networks: A detour on the road to integrated health care? Health Affairs, 21(4), 128 143. Burns, L., & Pauly, M. (2012). Accountable Care Organizations may have difficulty avoiding the failures of Integrated Delivery Networks of the 1990s. Health Affairs, 31(11), 2407 2416. Burns, L., Gimm, G., & Nicholson, S. (2005). The financial performance of integrated health organizations. Journal of Healthcare Management, 50(3), 191 212. Burns, L., Walston, S., Alexander, J., Zuckerman, H., Anderson, R., Torrens, P., & Hilberman, D. (2001). Just how integrated are integrated delivery systems? Results from a national survey. Health Care Management Review, 26(1), 20 39. Calciolari, S., & Ilinca, S. (2011). Comparing (and learning from) integrated care initiatives: An analytical framework. Journal of Integrated Care, 19(6), 4 13. Charns, M. (1997). Organization design of integrated delivery systems. Hospital & Health Services Administration, 42(3), 411 432. Comfort, L. (1994). Self-organization in complex systems. Journal of Public Administration, Research and Theory, 3, 393 410. Conrad, D. (1993). Coordinating patient care services in regional health systems: The challenge of clinical integration. Hospital & Health Services Administration, 38(4), 491 508. Conrad, D., & Dowling, W. (1990). Vertical integration in health services: Theory and managerial implications. Health Care Management Review, 15(4), 9 22. Conrad, D., & Shortell, S. (1996). Integrated health systems: Promise and performance. Frontiers of Health Services Management, 13(1), 3 40. Cumming, J. (2011). Integrated care in New Zealand. International Journal of Integrated Care, 11. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-101655 Darkins, A., Ryan, P., Kobb, R., Foster, L., Edmonson, E., Wakefield, B., & Lancaster, A. (2009). Care coordination/home telehealth: The systematic implementation of health informatics, home telehealth, and disease management to support the care of veteran patients with chronic conditions. Telemedicine and e-Health, 14(10), 1118 1126. Dattee, B., & Barlow, J. (2010). Complexity and whole-system change programmes. Journal of Health Services Research & Policy, 15(Suppl. 2), 19 25. de Jong, I., & Jackson, C. (2001). An evaluation approach for a new paradigm Health care integration. Journal of Evaluation in Clinical Practice, 7(1), 71 79. Delnoij, D., Klazinga, N., & Glasgow, I. (2002). Integrated care in an international perspective. International Journal of Integrated Care, 2. Retrieved from http://www.ijic.org/index. php/ijic/article/view/62/123 Demers, L. (2013). Mergers and integrated care: The Quebec experience. International Journal of Integrated Care, 13. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:101-114229

The Evolution of Integrated Health Care Strategies

155

Denis, J., Lamothe, L., Langley, A., Breton, M., Gervais, J., Trottier, L., … Dubois, C. (2009). The reciprocal dynamics of organizing and sensemaking in the implementation of major public-sector reforms. Canadian Public Administration, 52(2), 225 248. Devers, K., Shortell, S., Gillies, R., Anderson, D., Mitchell, J., & Erickson, K. (1994). Implementing organized delivery systems: an integration scorecard. Health Care Management Revie, 19(3), 7 20. Edgren, L. (2008). The meaning of integrated care: A systems approach. International Journal of Integrated Care, 8. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1100498. Edgren, L., & Barnard, K. (2012). Complex adaptive systems for management of integrated care. Leadership in Health Services, 25(1), 39 51. Enthoven, A. (2009). Integrated delivery systems: The cure for fragmentation. American Journal of Managed Care, 15, S284 S290. Evans, J., & Baker, R. (2012). Shared mental models of integrated care: Aligning multiple stakeholder perspectives. Journal of Health Organization and Management, 26(6), 713 736. Fahey, L., & Christensen, H. K. (1986). Evaluating the research on strategy content. Journal of Management, 12(2), 167 183. Fischer, E., & Coddington, D. (1998). Integrated health care: Passing fad or lasting legacy? Healthcare Financial Management, 52(1), 44 48. Fisher, E. (2008). Building a medical neighbourhood for the medical home. The New England Journal of Medicine, 359(12), 1202 1205. Fleury, M. (2006). Integrated service networks: The Quebec case. Health Services Management Research, 19, 153 165. Ford, R., & Fottler, M. (2000). Creating customer-focused health care organizations. Health Care Management Review, 25(4), 18 33. Friedman, L., & Goes, J. (2001). Why integrated health networks have failed. Frontiers of Health Services Management, 17(4), 3 28. Fox, W. (1989). Vertical integration: More promising than diversification. Health Care Management Review, 14(3), 49 56. Gillies, R., Shortell, S., Anderson, D., Mitchell, J., & Morgan, K. (1993). Conceptualizing and measuring integration: Findings from the health systems integration study. Hospital & Health Services Administration, 38(4), 467 489. Gillies, R., Shortell, S., & Young, G. (1997). Best practices in managing organized delivery systems. Hospital & Health Services Administration, 42(3), 299 321. Glasby, J., Dickinson, H., & Miller, R. (2011). Partnership working in England Where we are now and where we’ve come from. International Journal of Integrated Care, 11. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-101274 Goldsmith, J. (1998). Integration reconsidered: five strategies for improved performance. Healthcare Strategist, 2(11), 1 8. Goldsmith, J. (1994). The illusive logic of integration. Healthcare Forum Journal, 37(5), 26 31. Goldstein, D. (1995). Moving beyond generic integration models. Healthcare Financial Management, 49(4), 56 62. Greenhalgh, T., MacFarlane, F., Barton-Sweeney, C., & Woodard, F. (2012). “If we build it, will it stay?” A case study of the sustainability of whole-system change in London. The Milbank Quarterly, 90(3), 516 547.

156

JENNA M. EVANS ET AL.

Griffith, J. (1996). Managing the transition to integrated health care organizations. Frontiers of Health Services Management, 12(4), 4 50. Grone, O., & Garcia-Barbero, M. (2001). Integrated care: A position paper of the WHO European office for integrated health care services. International Journal of Integrated Care, 1. Retrieved from http://www.ijic.org/index.php/ijic/article/view/28/55 Ham, C., Dixon, J., & Chantler, C. (2011). Clinically integrated systems: The future of NHS reform in England? British Medical Journal, 342, d905. Hardy, B., Mur-Veeman, I., Steenbergen, M., & Wistow, G. (1999). Inter-agency services in England and the Netherlands: A comparative study of integrated care development and delivery. Health Policy, 48, 87 105. Harris, M., Greaves, F., Gunn, L., Patterson, S., Greenfield, G., Car, J., … Pappas, Y. (2013). Multidisciplinary group performance Measuring integration intensity in the context of the North West London Integrated Care Pilot. International Journal of Integrated Care, 13. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-114227 Hollander, M., & Prince, M. (2008). Organizing healthcare delivery systems for persons with ongoing care needs and their families: A best practices framework. Healthcare Quarterly, 11(1), 44 54. Hudson, B. (2011). Ten years of jointly commissioning health and social care in England. International Journal of Integrated Care, 11. Retrieved from http://www.ijic.org. URN: NBN:NL:UI:10-1-101296 Jiwani, I., & Fleury, M. (2011). Divergent models of integration: The Canadian way. International Journal of Integrated Care, 11. Retrieved from http://www.ijic.org. URN: NBN:NL:UI:10-1-101421. Kautz, C., Gittell, J., Weinberg, D., Lusenhop, R., & Wright, J. (2007). Patient benefits from participating in an integrated delivery system: Impact on coordination of care. Health Care Management Review, 32(3), 284 294. Kodner, D. (2009). All together now: A conceptual exploration of integrated care. Healthcare Quarterly, 13, 6 15. Kodner, D., & Spreeuwenberg, C. (2002). Integrated care: Meaning, logic, applications, and implications A discussion paper. International Journal of Integrated Care, 2. Retrieved from http://www.ijic.org/index.php/ijic/article/view/67/133 Kreindler, S., Larson, B., Wu, F., Carluzzo, K., Bbemudu, J., Struthers, A., & Fisher, E. S. (2012). Interpretations of integration in early accountable care organizations. The Milbank Quarterly, 90(3), 457 483. Lanham, H., Leykum, L., Taylor, B., McCannon, C., Lindberg, C., & Lester, R. (2013). How complexity science can inform scale-up and spread in health care: Understanding the role of self-organization in variation across local contexts. Social Science and Medicine, 93, 194–202. Larson, J. (1999). The conceptualization of health. Medical Care Research and Review, 56(2), 123 136. Leatt, P., Pink, G., & Naylor, C. (1996). Integrated delivery systems: Has their time come in Canada? Canadian Medical Association Journal, 154(16), 803 809. Leatt, P., Pink, G., & Guerrier, M. (2000). Towards a Canadian model of integrated healthcare. Healthcare Papers, 1(2), 13 35. Lee, K., & Wan, T. (2002). Effects of hospitals’ structural clinical integration on efficiency and patient outcome. Health Services Management Research, 15(4), 234 244.

The Evolution of Integrated Health Care Strategies

157

Lega, F. (2007). Organisational design for health integrated delivery systems: Theory and practice. Health Policy, 81(2), 258 279. Leggat, S., & Leatt, P. (1997). A framework for assessing the performance of integrated health delivery systems. Healthcare Management Forum, 10(1), 11 26. Leutz, W. (1999). Five laws for integrating medical and social services: Lessons from the United States and the United Kingdom. The Milbank Quarterly, 77(1), 77 110. Lillrank, P. (2012). Integration and coordination in healthcare: An operations management view. Journal of Integrated Care, 20(1), 6 12. Ling, T., Brereton, L., Conklin, A., Newbould, J., & Roland, M. (2012). Barriers and facilitators to integrating care: Experiences from the English integrated care pilots. International Journal of Integrated Care, 12. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:101-113730 Lin, B., & Wan, T. (1999). Analysis of integrated healthcare networks’ performance: A contingency-strategic management perspective. Journal of Medical Systems, 23(6), 467 485. Lovas, B., & Ghoshal, S. (2000). Strategy as guided evolution. Strategic Management Journal, 21(9), 875 896. Lukas, C., Meterko, M., Lowcock, S., Donaldson-Parlier, R., Blakely, M., Davies, M., & Petzel, R. A. (2002). Monitoring the progress of system integration. Quality Management in Health Care, 10(2), 1 11. Mazzocato, P., Savage, C., Brommels, M., Aronsoon, H., & Thor, J. (2010). Lean thinking in healthcare: A realist review of the literature. Quality & Safety in Health Care, 19(5), 376 382. McCarthy, K., & Zuckerman, A. (2010). Realizing the full financial benefits of true integration. Healthcare Financial Management, 64(11), 78 88. McClellan, M., McKethan, A., Lewis, J., Roski, J., & Fisher, E. (2010). A national strategy to put accountable care into practice. Health Affairs, 29(5), 982 990. McDaniel, R. Jr., & Driebe, D. (2001). Complexity science and health care management. In J. D. Blair, M. D. Fottler, & G. T. Savage (Eds.), Advances in Health Care Management (pp. 1 36). Stamford: JAI Press. McDaniel, R. Jr., Lanham, H., & Anderson, R. (2009). Implications of complex adaptive systems theory for the design of research on health care organizations. Health Care Management Review, 34(2), 191 199. Miller, R. (1996). Health systems integration: A means to an end. Health Affairs, 15(2), 92 106. Minkman, M., Ahaus, K., Fabbricotti, I., Nabitz, U., & Huijsman, R. (2009a). A quality management model for integrated care: Results of a Delphi and concept mapping study. International Journal for Quality in Health Care, 21(1), 66 75. Minkman, M., Ahaus, K., & Huijsman, R. (2009b). A four phase development model for integrated care services in the Netherlands. BMC Health Services Research, 9(42). doi:10.1186/1472-6963-9-42. Mur-Veeman, I., van Raak, A., & Paulus, A. (1999). Integrated care: The impact of governmental behaviour on collaborative networks. Health Policy, 49(3), 149 159. Mur-Veeman, I., van Raak, A., & Paulus, A. (2008). Comparing integrated care policy in Europe: Does policy matter? Health Policy, 82(2), 172 183. Niskanen, J. (2002). Finnish care integrated? International Journal of Integrated Care, 2. Retrieved from http://www.ijic.org/index.php/ijic/article/view/57/113

158

JENNA M. EVANS ET AL.

Nugus, P., Carroll, K., Hewett, D., Short, A., Forero, R., & Braithwaite, J. (2010). Integrated care in the emergency department: A complex adaptive systems perspective. Social Science and Medicine, 71(11), 1997 2004. Parker, V., Charns, M., & Young, G. (2001). Clinical service lines in integrated delivery systems: An initial framework and exploration. Journal of Healthcare Management, 46(4), 261 275. Pate, J., Fischbacher, M., & Mackinnon, J. (2010). Health improvement: Countervailing pillars of partnership and profession. Journal of Health Organization and Management, 24(2), 200 217. Paulus, A., van Raak, A., van Merode, F., & Eddy, A. (2000). Integrated health care from an economic point of view. Journal of Economic Studies, 27(3), 200 209. Pham, H. (2010). Good Neighbors: How will the patient-centered medical home relate to the rest of the health-care delivery system? Journal of General Internal Medicine, 25(6), 630 634. Plochg, T., Delnoij, D., Hoogedoorn, N., & Klazinga, N. (2006). Collaborating while competing? The sustainability of community-based integrated care initiatives through a health partnership. BMC Health Services Research, 6(37). doi:10.1186/1472-6963-6-37 Plochg, T., & Klazinga, S. (2002). Community-based integrated care: Myth or must? International Journal for Quality in Health Care, 14(2), 91 101. Plsek, P., & Greenhalgh, T. (2001). The challenge of complexity in health care. British Medical Journal, 323, 625 628. Porter, M., & Teisberg, E. (2006). Redefining health care: Creating value-based competition on results. Boston, MA: Harvard Business Press. Pourat, N., Davis, A., Salce, E., Hilberman, D., Roby, D., & Kominski, G. (2012). In ten California counties, notable progress in system integration within the safety net, although challenges remain. Health Affairs, 31(8), 1717 1727. Rajagopalan, N., & Spreitzer, G. (1997). Toward a theory of strategic change: A multilens perspective and integrative framework. The Academy of Management Review, 22(1), 48 79. Raphael, D. (Ed.). (2009). Social determinants of health: Canadian perspectives (2nd ed.), Toronto, ON: Canadian Scholars’ Press. Rea, H., Kenealy, T., Wellingham, J., Moffitt, A., Sinclari, G., McAuley, S., … Arcus, K. (2007). Chronic care management evolves towards integrated care in Counties Manukau, New Zealand. Journal of the New Zealand Medical Association, 120(1252), U2489. Rittenhouse, D., Shortell, S., & Fisher, E. (2009). Primary care and accountable care Two essential elements of delivery-system reform. The New England Journal of Medicine, 361(24), 2301 2303. Robinson, J., & Casalino, L. (1996). Vertical integration and organizational networks in healthcare. Health Affairs, 15(1), 8 22. Rodgers, B. (2000). Concept analysis: An evolutionary view. In B. L. Rodgers & K. A. Knafl (Eds.), Concept development in nursing (pp. 77 102). Philadelphia, PA: Saunders. Rudkjobing, A., Olejaz, M., Okkels, H., Nielsen, A., Hernandez-Quevedo, C., & Krasnik, A. (2012). Integrated care: A Danish perspective. British Medical Journal, 345, e4451. Rumbold, B., & Shaw, S. (2010). Horizontal and vertical integration in the UK: Lessons from history. Journal of Integrated Care, 18(6), 45 52. Shortell, S. (1988). The evolution of hospital systems: Unfulfilled promises and self-fulfilling prophecies. Medical Care Review, 45(2), 177 214.

The Evolution of Integrated Health Care Strategies

159

Shortell, S., Gillies, R., Anderson, D., Erickson, K., & Mitchell, J. (2000). Remaking health care in America: The evolution of organized delivery system (2nd ed.). San Francisco, CA: Jossey-Bass. Shortell, S., Gillies, R., Anderson, D., Mitchell, J., & Morgan, K. (1993). Creating organized delivery systems: Barriers and facilitators. Hospital & Health Services Administration, 38(4), 447 466. Shortell, S., Gillies, R., & Anderson, D. (1994). The new world of managed care: Creating organized delivery systems. Health Affairs, 13(5), 46 64. Shortell, S., & Kaluzny. (2005). Health care management: Organization design and behaviour (6th ed.). Albany, NY: Delmar Publishing. Shortell, S., Morrison, E., & Robbins, S. (1985). Strategy making in health care organizations: A framework and agenda for future research. Medical Care Review, 42(2), 219 266. Shortell, S., Schmittdiel, J., Wang, M., Li, R., Gillies, R., Casalino, L., … Rundall, T. (2005). An empirical assessment of high-performing medical groups: Results from a national study. Medical Care Research and Review, 62, 407 434. Singer, S., Burgers, J., Friedberg, M., Rosenthal, M., Leape, L., & Schneider, E. (2011). Defining and measuring integrated patient care: Promoting the next frontier in health care delivery. Medical Care Research and Review, 68(1), 112 127. Solberg, L., Asche, S., Shortell, S., Gillies, R., Taylor, N., Pawlson, G., … Young, M. R. (2009). Is integration in large medical groups associated with quality? American Journal of Managed Care, 15(6), e34 e41. Stein, V., & Rieder, A. (2009). Integrated care at the crossroads Defining the way forward. International Journal of Integrated Care, 9, e10. Retrieved from http://www.ijic.org/index. php/ijic/article/view/315/628 Strandberg-Larsen, M., & Krasnik, A. (2008). Does a public single payer system deliver integrated care? A national survey study among professional stakeholders in Denmark. International Journal of Integrated Care, 8. Retrieved from http://www.ijic.org/index.php/ ijic/article/view/248/494 Strandberg-Larsen, M., & Krasnik, A. (2009). Measurement of integrated healthcare delivery: A systematic review of methods and future research directions. International Journal of Integrated Care, 9. Retrieved from http://www.ijic.org/index.php/ijic/article/ view/305/608 Strandberg-Larsen, M., Schiotz, M., Silver, J., Frolich, A., Andersen, J., Graetz, I., … Hsu, J. (2010). Is the Kaiser Permanente model superior in terms of clinical integration? A comparative study of Kaiser Permanente, Northern California and the Danish healthcare system. BMC Health Services Research, 10, 91 104. Sullivan, H., & Williams, P. (2012). Whose kettle? Exploring the role of objects in managing and mediating the boundaries of integration in health and social care. Journal of Health Organization and Management, 26(6), 697 712. Suter, E., Oelke, N., Adair, C., & Armitage, G. (2009). Ten key principles for successful health systems integration. Healthcare Quarterly, 13, 16 23. Suter, E., Hyman, M., & Oelke, N. (2007). Measuring key integration outcomes: A case study of a large urban health center. Health Care Management Review, 32(3), 226 235. Thaldorf, C., & Liberman, A. (2007). Integration of health care organizations: Using the power strategies of horizontal and vertical integration in public and private health systems. The Health Care Manager, 26(2), 116 127.

160

JENNA M. EVANS ET AL.

Triska, O., Church, J., Wilson, D., Roger, R., Johnston, R., Brown, K., & Noseworthy, T. W. (2005). Physicians’ perceptions of integration in three western Canada health regions. Healthcare Management Forum, 18(3), 18 24. Tsasis, P., Evans, J., & Owen, S. (2012). Reframing the challenges to integrated care: A complex adaptive systems perspective. International Journal of Integrated Care, 12. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-113786 Tsasis, P., Evans, J., Forrest, D., & Jones, R. (2013). Outcome mapping for health system integration. Journal of Multidisciplinary Healthcare, 6, 99 107. Valentijn, P., Schepman, S., Opheij, W., & Bruijnzeels, M. (2013). Understanding integrated care: A comprehensive conceptual framework based on the integrative functions of primary care. International Journal of Integrated Care, 13. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-114415 van der Linden, B., Spreeuwenberg, C., & Schrijvers, A. (2001). Integration of care in The Netherlands: The development of transmural care since 1994. Health Policy, 55, 111 120. van Raak, A., Mur-Veeman, I., & Paulus, A. (1999). Understanding the feasibility of integrated care: A rival viewpoint on the influence of actions and the institutional context. International Journal of Health Planning and Management, 14, 235 248. van Raak, A., Meijer, E., Meijer, A., & Paulus, A. (2005). Sustainable partnerships for integrated care: The role of decision-making and its environment. International Journal of Health Planning and Management, 20, 159 180. van Wijngaarden, J., Bont, A., & Huijsman, R. (2006). Learning to cross boundaries: The integration of a health network to deliver seamless care. Health Policy, 79, 203 213. Vedel, I., Monette, M., Beland, F., Monette, J., & Bergman, H. (2011). Ten years of integrated care: Backwards and forwards. The case of the province of Quebec, Canada. International Journal of Integrated Care, 11. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:101-101295 Walker, K., Labat, A., Choi, J., Schmittdiel, J., Stewart, A., & Grumbach, K. (2013). Patient perceptions of integrated care: Confused by the term, clear on the concept. International Journal of Integrated Care, 13. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:101-114281. Walston, S., Kimberly, J., & Burns, L. (1996). Owned vertical integration and health care: Promise and performance. Health Care Management Review, 21(1), 83 92. Wan, T., Ma, A., & Lin, B. (2001). Integration and the performance of healthcare networks: Do integration strategies enhance efficiency, profitability, and image? International Journal of Integrated Care, 1. Retrieved from http://www.ijic.org/index.php/ijic/article/view/31/61 Wan, T., Lin, B., & Ma, A. (2002). Integration mechanisms and hospital efficiency in integrated health care delivery systems. Journal of Medical Systems, 26(2), 127 143. Wang, B., Wan, T., Clement, J., & Begun, J. (2001). Managed care, vertical integration strategies, and hospital performance. Health Care Management Science, 4(3), 181 191. Williams, P. (2012). Integration of health and social care: A case of learning and knowledge management. Health and Social Care in the Community, 20(5), 550 560. Williams, P., & Sullivan, H. (2009). Faces of integration. International Journal of Integrated Care, 9. Retrieved from http://www.ijic.org/index.php/ijic/article/view/509/1016 Woods, K. (2001). The development of integrated health care models in Scotland. International Journal of Integrated Care, 1. Retrieved from http://www.ijic.org/index.php/ ijic/article/view/29/57

The Evolution of Integrated Health Care Strategies

161

Ye, C., Browne, G., Grdisa, V., Beyene, J., & Thabane, L. (2012). Measuring the degree of integration for an integrated service network. International Journal of Integrated Care, 12. Retrieved from http://www.ijic.org. URN:NBN:NL:UI:10-1-113112 Young, D. (1997). Managing clinical integration in integrated delivery systems: A framework for action. Hospital & Health Services Administration, 42(2), 255 274. Yu, J., Engleman, R., & Van de Ven, A. (2005). The integration journey: An attention-based view of the merger and acquisition integration process. Organization Studies, 26(10), 1501 1528. Zimmerman, B., & Dooley, K. (2001). Mergers versus emergers: Structural change in health care systems. Emergence, 3(4), 65 82. Zimmerman, B., Lindberg, C., & Plsek, P. (1998). Edgeware: Complexity resources for healthcare leaders. Irving, TX: VHA Publishing.

SECTION III INTEGRATION AND PHYSICIAN RELATIONSHIPS

HOSPITAL PHYSICIAN RELATIONSHIPS: IMPLICATIONS FROM THE PROFESSIONAL SERVICE FIRMS LITERATURE Mona Al-Amin, Robert Weech-Maldonado and Rohit Pradhan ABSTRACT Purpose The hospital physician relationship (HPR) has been the focus of many scholars given the potential impact of this relationship on hospitals’ ability to achieve socially and organizationally desirable health care outcomes. Hospitals are dominated by professionals and share many commonalities with professional service firms (PSFs). In this chapter, we explore an alternative HPR based on the governance models prevalent in PSFs. Design/methodology approach We summarize the issues presented by current HPRs and discuss the governance models dominant in PSFs. Findings We identify the non-equity partnership model as a governance archetype for hospitals; this model accounts for both the professional

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 165 184 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015012

165

166

MONA AL-AMIN ET AL.

dominance in health care decisions and the increasing demand for higher accountability and efficiency. Research limitations There should be careful consideration of existing regulations such as the Stark law and the antikickback statue before the proposed governance model and the compensation structure for physician partners is adopted. Research implications While our governance archetype is based on a review of the literature on HPRs and PSFs, further research is needed to test our model. Practical implications Given the dominance of not-for-profit (NFP) ownership in the hospital industry, we believe the non-equity partnership model can help align physician incentives with those of the hospital, and strengthen HPRs to meet the demands of the changing health care environment. Originality/value This is the first chapter to explore an alternative hospital physician integration strategy by examining the governance models in PSFs, which similar to hospitals have a high reliance on a predominantly professional staff. Keywords: Physicians; hospitals; hospital physician relationships; governance; non-equity partnership; goal alignment

INTRODUCTION The organization of hospital physician relationships (HPRs) has varied greatly over the last few decades. In broad terms, HPRs refer to collaborative arrangements between hospitals and their physicians, and are also referred to as integration strategies (Burns & Muller, 2008). Initially, HPRs were loose and hospitals were viewed as physician “workshops” (Shortell, Morrisey, & Conrad, 1985), with no formal financial ties (Lake, Devers, Brewster, & Casalino, 2003). However, the advent of the hospital prospective payment system (PPS) in 1983 and the growing dominance of managed care marked the arrival of innovative physician hospital arrangements in the United States, which have included physician hospital organizations (PHOs), and in other cases, direct physician employment by the hospital. Hospitals hoped that closer alignment with physicians would increase

Hospital Physician Relationships

167

their market power, control costs, and help negotiate with managed care (Lake et al., 2003). A report by the California Healthcare Foundation points out that hospital physician integration increases during periods when “patterns of reimbursement align physician and hospital incentives, competition intensifies, or other economic or demographic changes require collaboration” (Witt, Jacobs, & CamdenGroup, 2010, p. 2). Therefore, it is not surprising that with the passage of the Patient Protection and Affordable Care Act (ACA) and its emphasis on “value-based” payment mechanisms improved quality at lower costs there is a renewed impetus for stronger HPRs. The ACA also proposes the accountable care organizations (ACO) model as a potential solution to the fragmented and wasteful delivery of care in the United States (Berwick, 2011). The ACO model is based on three principles: (a) Provider-led organizations are responsible for the entire continuum of care for a defined population, (b) Use of financial incentives for improving quality and controlling costs, and (c) Incorporating reliable performance measures which would address concerns of both the payers as well as consumers (Shortell, 2010). The ACA offers strong financial incentives for greater physician hospital integration through bundled payments which would require physicians and hospitals to collaborate closely to improve both quality and efficiency. The requirements of the ACA, along with pressures to bend the cost curve while improving quality, underline the need for innovative forms of HPRs. These new forms of HPRs would foster a quality driven and cost conscious organization, while recognizing the idiosyncratic nature of health care organizations, which includes professional dominance, asymmetric information between patients and providers, intensive use of technology, shared accountability, and high of level of government regulations. Therefore, the main aim of this chapter is to explore an alternative governance model which can help align physician incentives with those of the hospital, and strengthen HPRs to meet the demands of the changing health care environment. The health care industry has oftentimes adopted organizational models from other industries, such as aviation and manufacturing, with the dual goals of improving quality and increasing efficiency. For example, hospitals turned to the auto industry and specifically the Toyota Production System to improve operations (Bentley, Effros, Palar, & Keeler, 2008), reduce patient wait times in the Emergency Room (Ng, Vail, Thomas, & Schmidt, 2010), increase productivity in laboratory services (Rutledge, Xu, & Simpson, 2010), reduce medication errors (Kim, Spahlinger, Kin, & Billi, 2006), and

168

MONA AL-AMIN ET AL.

improve their operating margins. Similarly, health care organizations have adopted crew resource management (CRM) principles from aviation to medical team training to create a culture of safety and reduce medical errors. While there are obvious differences between aviation and medical services, there are also similarities: (1) both are highly regulated industries, (2) safety is critical and a priority in both industries, and (3) both rely on expert performance (Hunt & Callaghan, 2008). Whether it is Lean Management, Six Sigma, or CRM, there has been initial resistance from staff (Morey et al., 2002) or skepticism from industry leaders, professionals, and scholars, who emphasize the uniqueness of health care (Guo & Hariharan, 2012). However, health care organizations have learned and continue to do so by adopting process improvement methodologies, organizational structures, and strategies from other industries. Similarly in this chapter, we explore alternative hospital physician integration strategies by examining governance structures used in other industries, which similar to hospitals have a high reliance on a predominantly professional staff. We focus on the professional service firms (PSFs), such as law and accounting firms, to understand how their partnership model may be adapted to the hospital industry. In the next section, we discuss the administration professional relationship in PSFs and hospitals, and the shortcomings of current HPRs in minimizing the administration professional conflict. We then describe PSFs including their governance structures and the main attributes hospitals share with these organizations. In the final sections of this chapter, we propose the non-equity partnership model, present in many PSFs, as a viable alternative for hospitals.

HOSPITAL PHYSICIAN RELATIONSHIPS HPRs have varied over time from a symbiotic relationship to a competitive one with the nature and strength of this relationship shaped by health care policy and evolving reimbursement models (Page, Butler, & Bozic, 2012). In the past two decades, deteriorating HPRs coupled with physician dissatisfaction and desire for greater autonomy has pushed physicians toward founding their own health care delivery organizations, such as ambulatory surgery centers and specialty hospitals, and thus competing directly with hospitals. However, with the passage of the ACA, regulatory and environmental forces are forcing hospitals to reconfigure their arrangements with physicians. ACA has institutionalized new payment models such as

Hospital Physician Relationships

169

value-based purchasing and bundled payments. The reform shifts risk from payers and consumers to hospitals, and creates unprecedented incentives for hospitals to transform their current governance models and structures for better physician alignment (Sowers, 2013) Therefore, the main goal of our proposed governance model is to maximize goal alignment between hospitals and physicians, and thereby minimize physician administration conflict. In this section we discuss the various physician hospital integration strategies, and specifically focus our discussion on physician employment since it is an integral part of our proposed model. Burns and Muller (2008) group the variety of HPRs into three integration categories: clinical, noneconomic, and economic integration. Clinical integration involves hospital systems that ensure coordination of patient care between the different entities and are responsible for care delivery such as scheduling, case management, credentialing, total quality improvement, and care standards. Clinical integration is the technical core of hospital services that encompass the activities, individuals, and functions related to the delivery of patient care (Burns & Muller, 2008). Physicians depend on clinical integration to provide medical services to their patients. Noneconomic integration refers mainly to hospitals appealing to physicians through their facilities and operations and includes management information systems, participative decision making, and provision of training opportunities such as leadership development (Burns & Muller, 2008). Noneconomic and economic integration strategies aim at achieving greater clinical integration. For example, one noneconomic integration strategy that has particularly gained ground in the last decade and strengthens clinical integration is integrating community physicians through electronic medical records (EMRs) and related support services. While antikickback regulations and Stark law pose restrictions on the incentives hospitals can deploy with community physicians, the Internal Revenue Services (IRS) in 2007 made it clear that hospitals are allowed to subsidize physician EMR software purchase and support services (Grossman & Cohen, 2008). This integration strategy allows hospitals to increase physician loyalty and facilitates improved quality through better information sharing and coordination of care (Warden, 2009). Physicians on the other hand have an incentive to embrace hospital’s EMR services as they are able to avoid the substantial capital and recurring investments required for a fully functioning EMR system. Economic integration involves linking physicians to the organization through direct monetary compensation. It may include performance bonds, risk and gain sharing, direct employment by the hospital, medical

170

MONA AL-AMIN ET AL.

directorships, and comanagement of service lines and centers of excellence (Burns & Muller, 2008). While in all of these instances physician income is directly affected by the hospital, the level of alignment between the goals of the hospital and that of the physician varies between the various forms of economic integration. Physician employment is gaining grounds as an economic integration strategy. Historically, physician employment was not the norm in the hospital industry; the voluntary medical staff model, whereby physicians use hospitals as their “workshops” through admitting privileges, has been the traditional model (Casalino, November, Berenson, & Pham, 2008). In the traditional model, physicians benefit from all the resources available in the hospital while retaining their professional independence. The hospital in turn benefits through patient referrals, utilization of hospital facilities, and physician’s contribution to the hospital through services such as serving on hospital committees and being on call for the Emergency Department (Berenson, Ginsburg, & May, 2007). However, in response to the changing environment, physician employment by hospitals has been growing at a rapid pace in many markets (O’Mally, Bond, & Berenson, 2011). From a physician’s point of view, hospital employment secures regular work hours along with a steady income, while sheltering the physician from the unstable and increasingly competitive health care environment and rising malpractice insurance premiums (Casalino et al., 2008). Hospitals, on the other hand, may benefit as physician employment can potentially shelter the hospital from physician shortages in high demand specialties, empower the hospital with more control in terms of cost and quality, and provide hospitals with negotiating power with insurance plans (Casalino et al., 2008). However, health care researchers remain cautious of the physician employment model. Some of these concerns stem from the limited success of the first noticeable physician employment wave during the 1990s (Birk, 2010). Hospitals responding to the demands of managed care deployed physician employment strategies to secure referrals and adapt to the capitation reimbursement system. However, it appears that this initial wave of physician employment proved to be more costly than beneficial. Research showed that employed physicians were less responsive to patient demands, were less likely to make changes in order to improve patient perceptions, and had lower overall productivity (Griffith, 1995; Grembowski et al., 2003; Kocher & Sahni, 2011). Birk (2010) attributes failure of the physician employment models in the 1990s to the generous compensations offered to physicians and the limited readiness of hospitals to culturally and

Hospital Physician Relationships

171

structurally integrate physicians within the organization. More recent studies on hospital employment of physicians argue that while stronger physician integration might lead to better quality, physician employment is not necessarily an integration strategy that will guarantee clinical integration (O’Mally et al., 2011). Therefore, it is important that hospitals proceed cautiously with the physician employment model and make the necessary modifications to address its potential shortcomings. One area that deserves particular attention is the physician’s compensation structure in the employment models. It is almost certain that a fixed salary will not strongly align physician interests with the hospital; rather, compensation should be linked to productivity, cost containment, patient outcomes, and organizational goals and hospital financial performance. This to a large extent is similar to the challenges faced by PSFs in designing governance structures, where professionals are not only responsible for the technical/service delivery aspects of their job, but they are also accountable and directly rewarded for the success of the organization as a whole. While it is widely accepted among academic and health care executives that effective cooperation and goal alignment between hospitals and physicians are the cornerstone to hospital performance improvement (Trybou, Gemmel, & Annemans, 2011), there are still significant shortcomings in current hospital physician arrangements. The reasons behind the shortcomings are many, and include the competing logics of business-like health care versus medical professionalism (Reay & Hinings, 2009). The model we propose in this chapter builds on our discussion of hospital employment and attempts to address its potential shortcomings by presenting the partnership model prevalent in most PSFs as a viable option. In the next sections, we further expand on PSFs and their similarities with hospitals, and then discuss a modified partnership/employment model as an alternative hospital governance structure.

HOSPITALS AS PROFESSIONAL SERVICE FIRMS PSFs are knowledge intensive organizations with a highly educated workforce that customize their services to suit clients’ needs (Greenwood, Deephouse, & Li, 2007; Lowendahl, 2000). Von Nordenflycht (2010) argues that there are three attributes that can be used to characterize PSFs: high knowledge intensity, low capital requirement, and a professionalized workforce (Von Nordenflycht, 2010). Firms that are high on the three attributes,

172

MONA AL-AMIN ET AL.

such as accounting and law firms are considered Classic PSFs. In general, PSFs can be thought of as organizations whose highly educated and professional workforce constitutes its main asset, and whose outputs are services that are intangible and customized to meet the needs of the customer (Greenwood, Li, Prakash, & Deephouse, 2005). Admittedly, there are differences between hospitals and classical PSFs for instance, hospitals have high capital requirement. However, hospitals do share significant commonalities with PSFs, especially as it relates to the nature of their workforce. Hospitals rely on physicians, nurses, pharmacists, and other highly trained personnel to provide the majority of services. In most cases, it is difficult for customers (patients) to judge the quality of the services provided by the hospital due to information asymmetry and the fact that outcomes in health care services are hard to quantify. As such, Von Nordenflycht (2010) argues that hospitals belong to the group labeled Professional Campuses. These PSFs are knowledge intensive and have a highly professionalized workforce but differ from Classic PSFs in terms of high capital intensity. The professionalized workforce is one of the key distinguishing features that differentiates hospitals from other organizations and binds it with PSFs. In broad terms, the core of a profession is “prolonged specialized training in a body of abstract knowledge and a collectivity or service orientation”; income, prestige, and power are just derivatives gained from this core (Goode, 1960). Moreover, one of the important distinctive features of a profession is its monopolistic power in setting its own rules, standards, and membership criteria (Star, 1982). Commonalities between all professionals, including physicians, were summarized by Buhai (2012): (1) provision of personal services and not the production of goods, (2) specialized education, licensing, and power to exercise independent judgment, (3) disparities in information between customer and provider, (4) occupational ethos which places customer interest ahead of the other interests, and (5) duties to the public at large. Greenwood et al. (2005, p. 665) explains that the professional “carries and generates the knowledge encoded in the services being offered and he/she develops relationships with clients critical for a substantial flow of work. Professionals within the firm, in other words, are the human capital of the firm and generate its social capital.” Besides physicians, the hospital’s workforce consists of many other professionals. Around 80% of a hospital’s workforce has professional degrees, for example, nursing and pharmacy (Young, Charns, & Heeren, 2004). Therefore, there is a general agreement that hospital services are provided by a professional workforce; however, there is disagreement on how to best control and incentivize these professionals.

Hospital Physician Relationships

173

A system of governance frequently seen in PSFs is the partnership model. It is important to distinguish between “partnership” as a legal terminology and “partnership” in the organizational terminology. In the organizational literature, the professional partnership or simply partnership refers to a “distinctive set of governance characteristics, independent of any legal form” (Von Nordenflycht, 2007, p. 430). Most large PSFs are characterized by private corporate ownership form with a partnership governance structure, where junior professionals start as employees of the firm and contingent on their performance, achievements, and contributions to the overall success of the organization, progress to become partners. The traditional equity partnership model of PSFs consists mainly of partners sharing ownership and authority over the enterprise and making major decisions that affect performance (Pinnington & Morris, 2003), and excludes any outside ownership (Von Nordenflycht, 2008). Partnership is in general the most suitable governance model for PSFs since it limits potentially divergent goals between the different stakeholders: shareholders, professionals, clients, and administrators (Empson & Chapman, 2006). Greenwood, Hinnings, and Brown (1990) argue that the partnership form is very intertwined with the concept of professionalism. Partnership is almost synonymous with professionalism as both of these concepts are centered on three fundamental elements: (1) understanding and appreciating the esoteric knowledge of professionals, (2) giving professionals autonomy over decisions related to their particular domain, and (3) demanding that professionals use their knowledge to serve the clients’ best interests (Greenwood et al., 1990). More importantly, the partnership model might be the best mechanism to align the professionals’ goals with that of the organization, and this alignment would be expected to improve organizational performance. In many instances, it is the lack of goal and incentive alignment between professionals and organization that creates conflict between administration and professionals. Therefore, in the next section, we discuss the nature and causes of the administration and professional conflict evident in hospitals. In our view, given that the partnership model has proven to be successful in other industries in minimizing this conflict, hospitals can learn and adapt this governance model to suit their mission and the nature of their services.

PHYSICIANS AND HOSPITALS GOAL ALIGNMENT Due to the proliferation of large bureaucratic PSFs and the mounting influence of the administrative staff, PSFs in general seem to be facing some of

174

MONA AL-AMIN ET AL.

the same questions health care organizations have been trying to answer in the past 30 years. As Brivot (2011) argues, the struggle between administration and professionals is becoming evident in most industries and countries, “professions are facing intense economic and institutional pressure to deregulate, rationalize, and manageralize their activities” (p. 491). The struggle in hospitals is similar to that of large PSFs: there is pressing need for governance forms which can minimize the tension between the various control systems in an organization and maximize goal alignment between hospitals and physicians. We adopt Brivot’s (2011) and Hopwood’s (1974) typology of control systems to discuss three types of control systems in health care organizations: (1) self-control which is imposed by an individual on her/himself, (2) social controls imposed directly and indirectly on professionals within the same occupation on each other, and (3) administrative controls imposed by managers on subordinates and on professionals providing patient services. Traditionally hospitals like other PSFs have relied on self and social controls (Scott, 1982) with a limited role for administrative controls. However, with the advent of managed care and changes in reimbursement schemes including introduction of pay-for-performance, hospitals are increasingly emphasizing the use of administrative control systems. The traditional reliance on self and social controls has proven insufficient in eliminating waste, ensuring standardization of care, and in contributing to the hospital strategy and financial performance. Thus, the conceptual question we answer in this chapter is how to minimize conflict between the different control systems that coexist in hospitals. Another way of examining the issues associated with hospital governance is by understanding the distinction between the micro- and macrolevel issues faced by hospitals (Alexander, Kaluzny, & Middleton, 1986; Scott, 1982). The micro-level is concerned with the hospital’s technical core and primarily deals with delivering patient care. The macro-level, on the other hand, is concerned with the hospital’s operating environment, such as competition and regulations and its overall economic wellbeing (Alexander, Morrisey, & Shortell, 1986). Therefore, the fundamental challenge hospitals face is how to reconcile the physician focus on micro-level issues with macro-level concerns of the hospital administrative system. Based on the previous discussion, we stress that any optimal HPR needs to ensure two major goals, or what we refer to as the twin goals: (1) physicians embrace and support not only the self and social controls but the administrative control systems in the hospital as well, and that (2) physicians understand the association between micro and macro level

175

Hospital Physician Relationships

issues, and effectively contribute to both levels. Given current reimbursement and regulatory incentives, we argue that physician employment alone is not sufficient to ensure high quality care at lower costs. As an alternative, we suggest that employment coupled with the partnership model as is the case in other PSFs might be the optimal option to align physician and hospital incentives and achieve the two goals discussed previously.

THE NON-EQUITY PARTNERSHIP MODEL We propose the non-equity partnership employment model as a solution toward strengthening physician hospital goal alignment (Fig. 1). Traditionally, the partnership models have included both equity and nonequity based partnerships. In equity-based partnerships, individuals own shares in the organization once they make it to a partner status. While equity partnership used to be the dominant form of professional partnerships, with the increasing dominance of large and sometimes multinational PSFs, alternate models have emerged. In the non-equity partnership model, partners are salaried employees who have earned their partner status but do not own equity in the firm (Galanter & OPalay, 1990; Morris & Pinnington, 1998). Non-equity partnerships have increased significantly

Strength of Physician Hospital Goal Alignment

Physician Hospital Arrangements

Hospital Goals

Voluntary Medical Staff Model

Healing Patients

Sharing EMRs, etc…

Standardization-Evidence Based Care

Traditional Employment Model Salary & Bonuses Partnership Employment Model Partners: permanent tenure, higher salary, gainsharing and other incentives

Cost Containment Competitiveness in the Market Financial Performance Long-term Goals

Fig. 1.

Physician Hospital Goal Alignment.

176

MONA AL-AMIN ET AL.

since the beginning of the 21st century partly to accommodate professionals who have successfully contributed to the organization yet desire a less demanding workload than what is expected from equity partners (Evans, 2007). Non-equity partnership arrangements might vary across firms, but in general the partner receives financial reports on the firm, and incentives are offered to align his/her interests with those of the firm (Evans, 2007). In our view, the non-equity partnership employment model is more suitable for the hospital industry, given that around 52% of hospitals in the United States are not-for-profits (NFPs) (American Hospital Association, 2012). The non-equity partnership model can be adapted to the hospital industry to align physician and hospital interests. In this model, physicians are hired by the hospital as employees, and contingent on their performance and contribution to the overall wellbeing of the organization, they can progress to partner status. There are several advantages to the partnership model. First, achieving partner status is a robust motivator for physicians since it encompasses a wide array of financial and nonfinancial incentives that can influence physicians’ performance, such as (1) prestige, (2) acknowledgement of the physicians’ contributions to the organization, (3) direct input in the development of hospital’s goals and strategies, (4) economic gain such as higher salaries and gainsharing, and (5) the security of permanent tenure. Research indicates that financial and nonfinancial incentives influence physicians’ productivity and quality of care (Conrad & Christianson, 2004; Shortell, Waters, Clarke, & Budetti, 1998). In a recent study, comparing physicians and nurses work motivation; physicians reported the following as significant motivators: achieving goals of the organization, recognition for good performance, opportunity for advancement and promotion, income, and job security (Grujicic, Sipetic, Cvejin, Novakovic, & Bata, 2012). Second, physicians may be more trustful of the monitoring and reward mechanisms adopted by partnership organizations. Professionals in general are skeptical of monitoring systems designed to evaluate their performance (Sharma, 1997), and the lack of trust of such monitoring can result in negative emotions and perceptions, as well as resentment and suspicion (Sharma, 1997). However, in the partnership model, control systems are more likely to be accepted by physicians given that the ultimate power relies on physician partners rather than nonphysician managers. Finally, hospitals can benefit from the partnership model by more direct involvement of physicians in hospital committees and operations, and supervision of more junior physicians. As discussed in the prior section, economic integration is crucial to achieving the twin goals of aligning the interests of physicians and

Hospital Physician Relationships

177

hospitals, as well as minimizing potential physician administration conflict. Therefore, the hospital should provide partner physicians, in addition to competitive salaries, strong economic incentives such as gainsharing, participating bond transactions (PBTs), and salary bonuses, which link productivity and potential cost savings to remuneration. Gainsharing programs rewards physicians who contribute to hospital cost reduction efforts by sharing with them a percentage of the savings (Ketcham & Furukawa, 2008). Gainsharing arrangements can vary but in general include rewarding physicians for their contribution to standardized purchasing, implementation of information management programs, process improvement methodology such as Lean Management or Six Sigma, and other efficiency oriented programs. Prior research shows that hospitals’ gainsharing programs are associated with lower costs (Ketcham & Furukawa, 2008; Leitman et al., 2010; Plumb, Leitman, Factor, Sivaprasad, & Bernard, 2011). Similar to gainsharing programs, PBTs link physician income to hospital performance. PBTs are municipal tax exempt bonds, which carry a high yield because they are issued as subordinate debt (Stevens, 2005). The creator of PBTs, Robert Rosenfield, explains that the interest payment on PBTs varies with hospitals performance, and this can potentially create a strong alignment of the interests of hospitals and physicians (The Governance Institute, 2012). As we move toward value-based purchasing, patient experience and quality measures will directly impact hospital reimbursement and financial performance. As such, PBTs can provide strong incentives for physicians to commit their efforts toward better scores on quality indicators and patient ratings. Moreover, even though physician partners will not hold equity in the hospital (since ACA prohibits physician hospital ownership), PBTs can instill in a physician sense of ownership and control (Cohn, Allyn, Rosenfield, & Schwartz, 2005). Finally, salary bonuses can reward physicians based on their performance on quality indicators, such as readmission rates, and for their contribution to population health programs, such as preventive services and awareness campaigns. Research shows that financial incentives can also have a positive impact on quality of care and health outcomes (Doran et al., 2011; Peabody et al., 2011). While economic incentives can be a key motivator, other intrinsic factors such as personal satisfaction and self-actualization may motivate professionals to work toward the partner status (Burke, 1995). Nonmonetary rewards foster a competitive work environment and confer additional benefits to the employee including social recognition and increased work satisfaction. For instance, in a study of public health physicians, Garcia-Prado

178

MONA AL-AMIN ET AL.

(2005) has shown that the use of non-monetary incentives and public recognition may help improve performance without deleterious and expensive additional monitoring. As Arrow (1963) argued in his seminal paper, physicians are motivated by the concerns of altruism and reputation. This dynamic is further reinforced by their training and social conditioning. The partner status can offer physicians public recognition within the hospital and their community as well as personal satisfaction for their achievement. Promotion to a partner status in the proposed model would be based on productivity assessment which would not only include patient care outcomes but also contributions to the overall functions and the achievement of hospital’s mission, vision, goals, and strategy. As Carlson, Greeley, and Atchison (2010) argue “one strategy to align the interests of hospitals and physicians is to employ physicians using a compensation model focused on quality, service, cost, productivity, profitability, and the values of the integrated system.” In our model, partners will be involved in the selection and monitoring of physician partners. Partners participate in promoting and selecting other partners because the reputation of the hospital will be associated with the qualities of its physician partners in terms of their education, expertise, and achievements (Morris & Pinnington, 1998). Moreover, physician partners have direct interest in the performance of the hospital and therefore “are motivated to monitor their colleagues’ performance thus ensuring the quality of service provided” (Greenwood & Empson, 2003). This discourages physicians from covering up other physicians’ mistakes and creates a culture of shared responsibility. Physician partners will be beneficiaries of the hospital’s performance, and aspiring physicians will be motivated to achieve hospital’s goals by the potential promotion to partner status. Unlike the emphasis on “up-or-out” promotion models traditionally dominant in PSFs, especially consulting and law firms, our proposed model would allow frequent inflows of physicians from the external market, especially when these physicians have scarce specializations or are exceptionally well credentialed. Physician partners are not expected to manage day-to-day operations rather, this function will remain the domain of functional department heads. Partners, however, in collaboration with the administration will ensure physician involvement in major decisions related to operations such as the adoption of a new information system and implementation of best clinical practices. This, in turn, may influence how care is provided and more importantly enhance physician commitment. Physician partners involved in the hospital operations may be referred to as “Managing

Hospital Physician Relationships

179

Partners” and they will act as liaisons between the partners and the management teams (Empson & Chapman, 2006). Senior partners (senior and very well established physicians) will provide oversight on the strategic and financial management of the hospital. In summary, the hospital non-equity partnership model proposed in this chapter will maintain two authorities: the management and the partners’ authority. Major strategic and operational decisions will require the involvement and approval of both authorities. This model integrates physicians with the hospital, and enables those who have contributed to the goals, objectives, and strategy of the organization to be rewarded by a non-equity partnership status; thus it serves the achievement of the twin goals outlined before. Finally, the model proposed here would be particularly suitable to hospitals as the majority of hospitals are NFPs, precluding the use of equity/ownership models.

LIMITATIONS There are ethical considerations and opportunistic behavior that might be associated with the partnership model. However, Greenwood and Empson (2003) and Sharma (1997) argue that partnerships pressure professionals to adhere to professional values because the organization is usually dominated by a professional group that have common values, which might help deter deviant behavior. Another potential limitation of the partnership model is hospital size. The larger the hospital, the harder it is to base all major managerial and strategic decisions on consensus among all physician partners with different specialties. The more complex the organization, the less experience and knowledge partners will have and the less efficient it is to manage the partnership by depending on collegial decision-making (Greenwood & Empson, 2003). However, one potential solution may be limiting full partnership participation to major operational and strategic decisions, whereby the partnership group votes on major and “high symbolic” decisions (Greenwood & Empson, 2003). Finally, it is important to note that as hospitals design the compensation structure for their partnerships, they have to consider existing regulations, such as the Stark law and the antikickback statue. The Stark law prohibits physicians from referring Medicare patients to a health care facility with which the physician or a family member has financial ties, and prohibits providers from presenting claims to Medicare or any other third payer for those referred services, except in certain permitted transactions

180

MONA AL-AMIN ET AL.

(Wales, 2003). While many health care organizations have implemented gainsharing programs in compensating their physicians, these arrangements have been scrutinized by the Office of Inspector General (OIG) of the Department of Health and Human Services. OIG originally viewed gainsharing as a potential violation of current regulations; however, more recently several demonstration projects have been approved (Jacobson, Napiewocki, & Voigt, 2011; Torgerson, 2008). The OIG has viewed favorably arrangements that tie compensation to specific, identifiable cost-saving activities as opposed to generalized cost savings that may potentially have a negative impact on patient care. Other federal agencies, such as the IRS, the Department of Justice, and the Federal Trade Commission have encouraged and embraced the integration between physicians and health care organizations, and the implementation of risk-sharing programs (Jacobson et al., 2011). In its report released in 2005, the Medicare Payment Advisory Commission (MedPac) argued that “properly structured, gainsharing arrangements have the potential to encourage physician and hospital cooperation to lower costs and improve care” (2005).

CONCLUSION In this chapter, we adapt the partnership model which is dominant in PSFs to the hospital industry. We offer an innovative governance model suited for health care organizations, which have traditionally been dominated by professionals. The model has the potential to align hospital and physician goals in terms of patient care, quality improvement, and cost containment. The very nature of professionalism is shaped by the autonomy and social status that professionals enjoy. Organizational scholars have long argued that professionals should not be subject to traditional control mechanisms; the autonomy and self-regulation of professionals are in conflict with bureaucratic organizations (Barley, 2005). Organizations influence the status and satisfaction level of their professionals while at the same time, reliance on professional employees affects organization structure, strategy and behaviors (Von Nordenflycht, 2010). The movement toward ACOs and integrated care delivery mechanisms calls for newer HPR models which may result in better goal alignment of physicians and hospitals, while respecting the professional autonomy of physicians. The non-equity partnership model, with the appropriate financial incentives, may be helpful in securing the highest level of commitment from physicians

Hospital Physician Relationships

181

and in ensuring that physicians are willing to invest time and effort in achieving the optimal outcomes at both the micro- and macro-levels of care. Partnership is an overlooked strategy that might lead to better integration of physician knowledge, expertise, power, and incentives with hospitals’, patients’, and insurers’ goals.

REFERENCES Alexander, J., Kaluzny, A., & Middleton, S. (1986). Organization, growth, survival, and death in the U.S. hospital industry: A population ecology perspective. Social Science and Medicine, 22(3), 303 308. Alexander, J. A., Morrisey, M. A., & Shortell, S. M. (1986). Effects of competition, regulation, and corporatization on hospital-physician relationships. Journal of Health and Social Behavior, 27(3), 220 235. American Hospital Association. (2012). Fast facts on US hospitals. Retrieved from http:// www.aha.org/research/rc/stat-studies/fast-facts.shtml Arrow, K. J. (1963). Uncertainty and the welfare economics of medical care. The American Economic Review, 53(5), 941 973. Barley, S. R. (2005). What we know (and mostly don’t know) about technical work. In S. Ackroyd, R. Batt, P. Thompson, & P. Tolbert (Eds.), The oxford book of work and organization (pp. 377 403). Oxford: Oxford University Press. Bentley, T. G. K., Effros, R. M., Palar, K., & Keeler, E. B. (2008). Waste in the U.S. health care system: A conceptual framework. Milbank Quarterly, 86(4), 629 659. Berenson, R., Ginsburg, P., & May, J. (2007). Hospital-physicians relations: Cooperation, competition, or separation? Health Affairs, 26(1), w31 w43. doi:10.1377/hlthaff.26.1.w31 Berwick, D. M. (2011). Launching accountable care organizations The proposed rule for the medicare shared savings program. New England Journal of Medicine, 364(16), e32. Birk, S. (2010). Physician employment: How it can work this time. Healthcare Executive, 25(4), 8 10. Brivot, M. (2011). Controls of knowledge production, sharing and use in bureaucratized professional service firms. Organization Studies, 32(4), 489 508. Buhai, S. (2012). Profession: A definition. Fordham Urban Law Journal, 40(1), 241 281. Burke, R. (1995). Benefits of promotion within professional services firm. Psychological Report, 76(3), 859 867. Burns, L., & Muller, R. (2008). Hospital-physician collaboration: Landscape of economic integration and impact on clinical integration. Milbank Quarterly, 86(3), 375 434. Carlson, G., Greeley, H., & Atchison, T. (2010). Is the relationship between your hospital and your medical staff sustainable? Practitioner application. Journal of Healthcare Management, 55(3), 158 173. Casalino, L. P., November, E. A., Berenson, R. A., & Pham, H. H. (2008). Hospital-physician relations: Two tracks and the decline of the voluntary medical staff model. Health Affairs, 27(5), 1305 1314. Cohn, K. H., Allyn, T. R., Rosenfield, R. H., & Schwartz, R. (2005). Overview of physicianhospital ventures. The American Journal of Surgery, 189(1), 4 10.

182

MONA AL-AMIN ET AL.

Conrad, D. A., & Christianson, J. B. (2004). Penetrating the “black box”: Financial incentives for enhancing the quality of physician services. Medical Care Research and Review, 61(3 suppl.), 37S 68S. Doran, T., Kontopantelis, E., Valderas, J. M., Campbell, S., Roland, M., Salisbury, C., & Reeves, D. (2011). Effect of financial incentives on incentivised and non-incentivised clinical activities: longitudinal analysis of data from the UK Quality and Outcomes Framework. BMJ: British Medical Journal, 342. Retrieved from http://www.bmj.com/content/342/bmj. d3590?rss=1 Empson, L., & Chapman, C. (2006). Partnership vs corporation: Implications of alternative forms of governance in PSFs. In R. Greenwood & R. Suddaby (Eds.), Professional Service Firms: Research in the Sociology of Organizations (Vol. 24). Bingley, UK: Emerald Group Publishing. Evans, D. (2007). Non-equity partnership: A flawed solution to the disprportionate advancement of women in private law firms. Women’s Rights Law Reporter, 28, 13. Galanter, M., & OPalay, T. (1990). Why big get bigger: The promotion-to-partner tournamemnt and the growth of large law firms. Virginia Law Review, 76, 64. Garcia-Prado, A. (2005). Sweetening the Carrot: Motivating public physicians for better performance. Policy Research Paper. Washington, DC: The World Bank. Goode, W. J. (1960). The profession: Reports and opinion. American Sociological Review, 25(6), 902 965. Greenwood, R., & Empson, L. (2003). The professional partnership: Relic or exemplary form of governance? Organization Studies, 24(6), 909 933. Greenwood, R., Deephouse, D., & Li, S. X. (2007). Ownership and performance of professional service firms. Organization Studies, 28(2), 219 238. Greenwood, R., Hinnings, C., & Brown, J. (1990). “P2-form” strategic management: Corporate practices in professional partnerships. Academy of Management Journal, 33(4), 725 755. Greenwood, R., Li, S. X., Prakash, R., & Deephouse, D. L. (2005). Reputation, diversification, and organizational explanations of performance in professional service firms. Organization Science, 16(6), 661 673. Grembowski, D., Ulrich, C., Paschane, D., Diehr, P., Katon, W., Martin, D., … Velicer, C. (2003). Managed Care and Primary Physician Satisfaction. Journal of the American Board of Family Practice, 16(5), 383 393. Griffith, J. R. (1995). The well managed health care organization. Chicago, IL: Health administration Press. Grossman, J., & Cohen, G. (2008). Despite regulatory changes, hospitals cautious in helping physicians purchase electronic medical records. Center for Studying Health System Change, (123). Grujicic, M., Sipetic, S., Cvejin, M. M., Novakovic, B., & Bata, J. J. (2012). Differences in work motivation and job satisfaction between physicians and nurses. HealthMed, 6(12), 4009 4018. Guo, L., & Hariharan, S. (2012). Patients are not cars and staff are not robots: Impact of differences between manufacturing and clinical operations on process improvement. Knowledge and Process Management, 19(2), 53 68. Hopwood, A. (1974). Accounting and human behavior. London: Haymarket. Hunt, G., & Callaghan, K. (2008). Comparative issues in aviation and surgical crew resource management: (1) Are we too solution focused? ANZ Journal of Surgery, 78(8), 690 693.

Hospital Physician Relationships

183

Jacobson, P., Napiewocki, L., & Voigt, L. (2011). Regulating the US health care system: Failure in motion. Journal of Health Politics, Policy and Law, 36(3), 583 589. Ketcham, J. D., & Furukawa, M. F. (2008). Hospital-physician gainsharing in cardiology. Health Affairs, 27(3), 803 812. Kim, C. S., Spahlinger, D. A., Kin, J. M., & Billi, J. E. (2006). Lean health care: What can hospitals learn from a world-class automaker? Journal of Hospital Medicine, 1(3), 191 199. Kocher, R., & Sahni, N. R. (2011). Hospitals’ race to employ physicians The logic behind a money-losing proposition. New England Journal of Medicine, 364(19), 1790 1793. Lake, T., Devers, K., Brewster, L., & Casalino, L. (2003). Something old, something new: Recent developments in hospital physician relationships. Health Services Research, 38(1), 471 488. Leitman, I. M., Levin, R., Lipp, M. J., Sivaprasad, L., Karalakulasingam, C. J., Bernard, D. S., … Shulkin, D. J. (2010). Quality and financial outcomes from gainsharing for inpatient admissions: A three-year experience. Journal of Hospital Medicine, 5(9), 501 507. Lowendahl, B. (2000). Strategic management of professional firms (2nd ed.), Copenhagen: Handelshojskolens Forlag. MedPac. (2005). Physician owned specialty hospitals. Retrieved from http://www.medpac.gov/ documents/Mar05_SpecHospitals.pdf Morey, J. C., Simon, R., Jay, G. D., Wears, R. L., Salisbury, M., Dukes, K. A., & Berns, S. D. (2002). Error reduction and performance improvement in the emergency department through formal teamwork training: Evaluation results of the medteams project. Health Services Research, 37(6), 1553 1581. Morris, T., & Pinnington, A. (1998). Promotion to partner in professional service firms. Human Relations, 51(1), 3 24. Ng, D., Vail, G., Thomas, S., & Schmidt, N. (2010). Applying the lean principles of the Toyota production system to reduce wait times in the emergency department. CJEM, 12(1), 50 57. O’Mally, A., Bond, A., & Berenson, R. (2011). Rising hospital employment of physicians: Better quality, higher costs? Center for Studying Health System Change, 136. Retrieved from http://www.hschange.org/CONTENT/1230/1230.pdf Page, A. E., Butler, C. A., & Bozic, K. J. (2012). Factors driving physician-hospital alignment in orthopaedic surgery. Clinical Orthopaedics and Related Research, 471(6), 1809 1817. Peabody, J., Shimkhada, R., Quimbo, S., Florentino, J., Bacate, M., McCulloch, C. E., & Solon, O. (2011). Financial incentives and measurement improved physicians’ quality of care in the Philippines. Health Affairs, 30(4), 773 781. Pinnington, A., & Morris, T. (2003). Archetype change in professional organizations: Survey evidence from large law firms. British Journal of Management, 14(1), 85 99. Plumb, P. G., Leitman, I. M., Factor, P. H., Sivaprasad, L., & Bernard, D. S. (2011). Gainsharing in the ICU: Is now the right time? ICU Director, 2(3), 74 77. Reay, T., & Hinings, C. R. (2009). Managing the rivalry of competing institutional logics. Organization Studies, 30(6), 629 652. Rutledge, J., Xu, M., & Simpson, J. (2010). Application of the Toyota production system improves core laboratory operations. American Journal of Clinical Pathology, 133(1), 24 31. Scott, W. R. (1982). Managing professional work: Three models of control for health organizations. Health Services Research, 17(3), 213 240.

184

MONA AL-AMIN ET AL.

Sharma, A. (1997). Professional as agent: Knowledge asymmetry in agency exchange. The Academy of Management Review, 22(3), 758 798. Shortell, S. (2010). Delivery system reform: Accountable care organizations and patient centered medical homes. In The Society for Healthcare Strategy and Market Development (Ed.), Futurescan 2010: Healthcare trends and implications 2010 2015. Chicago, IL: Health Administration Press. Shortell, S., Morrisey, M., & Conrad, D. (1985). Economic regulation and hospital behavior: The effects on medical staff organization and hospital-physician relationships. Health Services Research, 20(5), 597 628. Shortell, S., Waters, T., Clarke, K., & Budetti, P. (1998). Physicians as double agents: Maintaining trust in an era of multiple accountabilities. JAMA, 280(12), 1102 1108. Sowers, K. W. (2013). Evolution of physician-hospital alignment models: A case study of comanagement. Clinical Orthopaedics and Related Research, 471(6), 1818 1823. Star, P. (1982). The social transformation of American medicine: The rise of a sovereign profession and the making of a vast industry. New York, NY: Basic Book. Stevens, T. (2005). Particpating bonds offer new “carrots” to hospitals. Nashville Medical News, May. The Governance Institute. (2012). Retrieved from http://www.governanceinstitute.com/ ResearchPublications/FacultySpeakerDirectory/BiographyRobertHRosenfieldEsq/tabid/49 6/Default.aspx Torgerson, P. M. (2008). Gain-sharing with the hospital: What is possible in the current legal environment. Orthopedic Clinics of North America, 39(1), 33 36. Trybou, J., Gemmel, P., & Annemans, L. (2011). The ties that bind: An integrative framework of physician-hospital alignment. BMC Health Services Research, 11(1), 36. Von Nordenflycht, A. (2007). Is public ownership bad for professional service firms? Ad agency ownership, performance, and creativity. Academy of Management Journal, 50(2), 429 445. Von Nordenflycht, A. (2008). The demise of the professional partnership? The emergence and diffusion of publicly-traded professional service firms. Working Paper. Simon Fraser University, Burnaby, British Columbia. Retrieved from http://fcic-static.law.stanford.edu/ cdn_media/fcic-docs/2006-09-00%20Andrew%20von%20Norden-flycht,%20Draft%20Paper %20on%20The%20Demise%20of%20the%20Professional%20Partnership,%20The%20 Emergence%20and%20Diffusion%20of%20Publicly-Traded%20Professional%20Service %20Firms.pdf Von Nordenflycht, A. (2010). What is a professional service firm? Toward a theory and taxonomy of knowledge-intensive firms. Academy of Management Review, 35(1), 155 174. Wales, S. D. (2003). Stark law: Boon or Boondoggle An analysis of the prohibition on physician self-referrals. The Law and Psychology Review, 27, 1. Warden, J. (2009). Creating sustainable physician-hospital strategies. Chicago, IL: Health Administration Press. Witt, M., Jacobs, L., & Camden Group. (2010). Physician-hospital integration in the era of health reform (p. 2). Prepared for the California Healthcare Foundation. Young, G. J., Charns, M. P., & Heeren, T. C. (2004). Product-line management in professional organizations: An empirical test of competing theoretical perspectives. The Academy of Management Journal, 47(5), 723 734.

ENGAGING EMPLOYED PHYSICIANS: RECONCEPTUALIZING THE ROLE OF COLLECTIVE IDENTIFICATION Gregory W. Stevens ABSTRACT Purpose This chapter proposes a paradigm shift in considering the collective identification of employed physicians and how it influences physician engagement. Design/methodology/approach There are many challenges for organizations employing physicians, particularly in terms of engagement in organizational initiatives. Prior research suggests this conflict stems from how physicians think of themselves as professionals versus employees (as forms of collective identification). Unfortunately, research is limited in addressing these dynamics. Findings This conceptual chapter considers the complex network of relationships that physicians perceive between the collectives to which

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 185 209 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015013

185

186

GREGORY W. STEVENS

they belong. A primary collective identification (i.e., the profession) is proposed to influence subsequent collective identification (i.e., the organization), and that these meanings and relationships along with contextual factors drive engagement. Originality/value Health care organizations increasingly rely on engagement from their physicians to improve upon coordinated care. This proposed conceptualization offers new insight into the dynamics surrounding how and why employed physicians become engaged. Keywords: Physician engagement; employed physicians; professional identification; organizational identification; collective identification

A clear trend within the health care industry is the increasing number of physicians employed within larger organizational settings, such as organized delivery systems (Hoff, 2001; Shortell, Gillies, Anderson, Erickson, & Mitchell, 2000; Zuckerman et al., 1998), but these arrangements have not been without their challenges. At the heart of the issue is the difficulty of engaging employed physicians, who seek to retain their autonomy and professional authority, while administrators of the systems that employ them attempt to improve upon the delivery of care through coordination, efficiency improvements, and the implementation of quality standards (e.g., Dukerich, Golden, & Shortell, 2002; Fiol, Pratt, & O’Connor, 2009; Kreindler, Dowd, Star, & Gottschalk, 2012). In these scenarios, employed physicians must balance a sense of belonging to their profession, referred to as professional identification, in addition to membership within an organization, referred to as organizational identification (Hekman, Steensma, Bigley, & Hereford, 2009; Johnson, Morgeson, Ilgen, Meyer, & Lloyd, 2006). Conflict between the values and goals associated with each form of identification is most often posited as a source of the difficulty of engaging professional employees in organizations (Gouldner, 1957; Hekman, Bigley, Steensma, & Hereford, 2009; Wallace, 1995). Unfortunately, this paradigm is deficient in terms of accurately capturing the complex processes of identification among employed physicians, particularly as a driver of engagement. Rather than continue to consider multiple identifications within silos (e.g., “I am a professional” or “I am an employee”; cf. Hekman, Steensma, et al., 2009; Johnson et al., 2006), this chapter advocates a shift toward thinking about the interactions between identifications from the perspective of the individual physician. Within the context of employed physicians,

Engaging Employed Physicians

187

the argument is made for a primary collective identity (i.e., the professional identity) that influences subsequent identification with other collectives (e.g., “I am a physician working for or with this health care organization”). There are several implications from such a paradigm shift. First, it focuses more closely on the specific content of identifications rather than rely on abstract labels. Second, it moves from prior assumptions that collectives do not interact with each other in the social environment to explicit notions of those relationships (e.g., the physician’s perception of the relationship between doctors and the organization). Finally, it allows for more nuanced prediction of engagement by looking at the different potential configurations of identification rather than assuming that engagement follows from an abstract notion of organizational identification.

REVIEWING THE LITERATURE ON IDENTIFICATION AMONG EMPLOYED PHYSICIANS As noted in the introduction, a common lens through which to interpret dynamics surrounding employed physicians has been collective identification, particularly in terms of their sense of belonging within the larger profession of physicians and within the health care organizations that employ them. Before discussing this specific stream of research, however, it is useful to examine first what is meant by these concepts, and also explore why they are important from an organizational standpoint. What Is Collective Identification and Why Does It Matter? Broadly, collective identification captures an individual’s experience of a sense of belonging or oneness with a group, role, or social category, such that membership within that collective becomes self-defining or selfreferential (Ashforth, Harrison, & Corley, 2008; Ashforth & Mael, 1989; Pratt, 1998). It answers the questions of “Who am I as a member of this group?” or “Who are we as a group?” in a contextually appropriate manner (Brewer & Gardner, 1996). For example, a physician may identify as a clinician, a medical director, an educator, or some combination of those depending on the relevant collective in a given situation: working in a primary care practice, serving on a committee with other medical directors, or teaching residents and medical students with fellow faculty members. Within each of these contexts, collective identification serves an important

188

GREGORY W. STEVENS

function, not only in helping to define the individual as a social actor, but also helping to explain how that individual may interpret and behave in a given situation. Indeed, collective identification has often been referred to as a “root construct” in explaining organizational thought and action (Albert, Ashforth, & Dutton, 2000, p. 13). In general, common correlates and outcomes of identification are buttressed by a common understanding that views identification as the process through which individuals become a “microcosm” of the larger collective, where exerting effort on behalf of a group becomes indistinguishable from doing so on behalf of the self. Illustrative of this connection, collective identification has been shown to relate to cooperation on behalf of the organization (Dukerich et al., 2002), participation and coordination within the collective (Blader & Tyler, 2009; Lipponen, Bardi, & Haapamaki, 2008), demonstration of reciprocity with the collective (Hekman, Bigley et al., 2009), intentions to support organizational action and change (Hekman, Steensma et al., 2009; Jimmieson & White, 2011), and a range of citizenship behaviors and extrarole performance (Bergami & Bagozzi, 2000; Grube & Piliavin, 2000; Van Dick, Grojean, Christ, & Wieseke, 2006). The common thread throughout this disparate range of outcomes is the willingness of the individual to exert some form of effort toward activities that benefit a larger collective. A final note is perhaps warranted regarding the relationship between identification and the expectation of extra-role behaviors, particularly among employed physicians. Despite being salaried and under contract, it has been argued that professional employees’ behavior is usually not as subject to incentives or punishments as would guide typical employee behavior (e.g., Dukerich et al., 2002; Johnson et al., 2006; Scott, 1982). This is particularly true for behaviors for which no formal sanctions exist, such as cooperative and extra-role behavior. Rather, employed professionals’ decisions to engage in such behaviors likely stem from the way in which they identify themselves, or think about themselves in relation to a given context (Kramer, 1993). Taken together, these findings suggest a strong role for identification in explaining why some physicians might engage in pro-organizational behaviors, while others choose not to.

Professional and Organizational Identification The notion of collective identity is a useful lens for investigating the working relationships of employed physicians as they face the challenge of

Engaging Employed Physicians

189

simultaneously balancing a sense of belonging within multiple collectives: the profession and the organization that employs them. This stream of research has its roots in the classical distinction Gouldner (1957, 1958; see also Wallace, 1995) made between cosmopolitans and locals, and the corresponding influence of these orientations on organizational behavior. The cosmopolitan identity captures individuals with an “outer reference group orientation,” those who are more committed to specialized role skills and the larger profession rather than to a specific organization; on the other hand, the local identity captures an “inner reference group orientation,” or those who are more loyal to a specific organization than to a set of specialized role skills or external professional group (Gouldner, 1957, p. 290). Herein lays a potential point of conflict for organizations, which simultaneously depend on the expertise of these professionals and yet also strive to function as an efficient and coherent entity based on the commitment and participation of its members. More recently, this distinction has been recast in the context of employed physicians (and other employed professionals, such as lawyers and accountants) as professional identification and organizational identification (cf. Bamber & Iyer, 2002; Dukerich et al., 2002; Glynn, 2000; Golden, Dukerich, & Fabian, 2000; Hekman, Steensma et al., 2009; Johnson et al., 2006). The argument underlying this research suggests that professional identification, much like the cosmopolitan identity, leads one to perceive the organization and its administrators as outsiders whose values and goals run counter to those espoused by the profession (e.g., profit and cost controls over quality of practice). On the other hand, organizational identification, much like the local orientation described above, contributes to a sense of being on the same team as the administration, sharing the same goals, values, and priorities. Empirical findings tend to support these assertions, to the extent that professional identification relates to fewer pro-organizational behaviors particularly under high levels of perceived monitoring and administrator normative pressure (Hekman, Steensma et al., 2009), and negatively relates to policy adherence and productivity (Hekman, Bigley et al., 2009). Meanwhile, organizational identification has been found to lead to increased cooperative behaviors and participation within the organization (Dukerich et al., 2002). Interestingly, Hekman, Bigley et al. (2009) found that high professional identification tended to obscure the positive effects of organizational identification. This traditional paradigm is illustrated at a high level in the top panel of Fig. 1.

190

GREGORY W. STEVENS

Traditional Paradigm: Professional Identification [–] Individual Physician

Engagement [+] Organizational Identification

Proposed Paradigm: Moderating Influences on relationship between primary and subsequent collectives [P6]

Primary Collective Identification

[P5]

Subsequent Collective Identification

Content [P3] Moderating influences on strength of primary collective identity [P2]

Alignment with engagement, related to congruency [P7] and interdependency [P8] in collectives’ relationship

[P1]

Individual Physician

Alignment of primary identity content and engagement through shared meaning [P4]

Engagement

Fig. 1. A comparison illustrating the traditional paradigm of multiple collective identification among employed physicians (top panel) and the proposed paradigm shift (lower panel). In the traditional paradigm, professional identification negatively influences enagement while organizational identification positively influences engagment; however, these forms of collective identification are considered in isolation from one another. In the proposed paradigm, the focus shifts to the notion of identification with a primary collective, the relationship to subsequent collectives, and the factors that influence their relationships to engagement.

Limitations of Extant Theory Although extant research helps to further an understanding of professionals’ interactions with organizations in terms of different forms of identification, it is nevertheless constrained by several limitations. Perhaps most importantly, the majority of this research has approached the study of

Engaging Employed Physicians

191

identification with multiple collectives as if the collectives to which an individual belongs operate in a vacuum (Hekman, Bigley et al., 2009; Johnson et al., 2006), rather than appreciating that those collectives are all part of the same working environment. This is a serious limitation to the extent that employed physicians may generate complex understandings of the relationships between their profession and the organizations to which they belong (i.e., “this is what the doctors in this organization do vis-a`-vis what the organization does”), particularly given the interdependent nature of providing coordinated and comprehensive care. Relatedly, research has often overlooked the potential influences that identification with one collective may have on other collectives to which the individual also belongs (cf. Ashforth & Johnson, 2001). For example, past research suggests that the only path toward engagement is with a high level of organizational identification and low level of professional identification (Hekman, Bigley et al., 2009), akin to asking physicians to forget their professional socialization and values, and instead become strongly committed members of an organization. This position may not be tenable, however, given the time and investment required to become a physician. More likely, the professional identity acts as a core or primary identity and influences other work-related collectives to which the physician belongs. Once these types of relationships between collectives are acknowledged, the dynamics of those relationships offer fruitful directions for expanding the role of collective identification in the context of driving engagement. This final point, of what it means to be a professional captures a third limitation; namely, collective identification tends to be assessed at a global and abstract level (i.e., assuming the “physician” label means the same thing to all physicians), rather than appreciating the unique meaning and content that individuals may apply to that label. There may be more value in understanding how an employed physician identifies him or herself within the context of the profession and the organization, rather than rely on broad labels that assume all physicians understand professional or organizational membership in the same way. In the following sections, I propose a paradigm shift that rests on the concept of a primary collective identity (i.e., the physician as a professional) that influences subsequent identification with relevant work-related collectives, such as the organization (summarized in the lower panel of Fig. 1). By understanding the relationships between collectives, one is able to reconceptualize a sense of organizational identification that does not require the casting off of a professional identity, but rather incorporates both collectives into a larger understanding of how both collectives fit

192

GREGORY W. STEVENS

together and interact. Finally, leveraging these relationships allows one to be more specific about the forms of engagement that are desired and understanding how the meaning of these complex identifications may contribute to those behaviors.

TOWARD A PRIMARY COLLECTIVE IDENTITY The first step in establishing the profession as a primary collective identity is to more fully understand the nature and meaning of that collective membership to the individual. This process involves not only looking at what it means to be a professional, but also how that meaning is enacted through an individual’s identity. Through this meaning, it becomes clear that professional and organizational identification cannot be treated as similar constructs differing only in terms of the relevant foci; rather, it is important to acknowledge that they may serve an individual in fundamentally different ways.

Establishing the Profession as the Primary Collective Identity Within the context of what it means to be a professional, there are a number of factors that contribute to one’s identity as a physician serving as a primary collective identity. To start with, a profession is typically organized around a systematic and theoretical body of knowledge, as well as a set of cultural values, norms and professional opinions, all which serve to separate and distinguish a given profession from general society (Etzioni, 1969; Hall, 1975; Hickson & Thomas, 1969). Furthermore, a long period of socialization, education, and training are typically required of physicians, the net result of which is a professional attitude and a lifelong commitment to the profession (Etzioni, 1969). These factors help to contribute to a professional identity that is well developed and salient, in terms of: the individual’s membership and “fit” within the collective; the value and status placed on membership both by the individual and society at large; and a sense of interdependence among members of the collective as a result of shared experiences and skillsets (Ashmore, Deaux, & McLaughlin-Volpe, 2004). One must also note that physicians spend a large portion of their days engaged in the tasks that define their profession (i.e., medical decisionmaking, caring for patients). The greater the degree to which an identity is

Engaging Employed Physicians

193

implicated in one’s day-to-day activities, referred to as social embeddedness (Ashmore et al., 2004), the greater importance and salience one would expect that identity to have. Daily experiences also tend to reinforce the professional values and attitudes that are initially developed during the period of training. For example, the high regard in which society holds professionals is likely communicated to physicians through patient encounters and more generally, through everyday interactions with office staff and members of the general public. The net result from these experiences tends to be a professional identity that is quite strong, especially when considered against other work-related identities that are more temporal or require less investment, for example an organizational or workgroup identity (Johnson et al., 2006). Less time, for example, is devoted by physicians to the business side of medicine and interacting with the organization that manages those aspects, than it is to patient care. Furthermore, the medical aspects of practicing medicine are typically held in higher regard than the business aspects that support them, such as ensuring compliance with federal regulations and insurance mandates, ensuring pre-certifications and prior authorizations are in place, and so forth. Finally, although less frequently considered, the content of each identity includes the historical relationships and narratives associated with the tension between the business and medical aspects of practicing medicine from the individual’s perspective (e.g., cycles of divestiture and investiture in primary care clinics or other initiatives, track records of administration and physician relationships) and speaks to the way in those foci of membership may not be assumed to be equally influential. In general, then, one may expect a strong and central identity to pervade across contexts, in the sense that a physician first and foremost identifies him or herself as a professional. Indeed, Golden and colleagues (2000) found that “mixed affiliates,” defined as professionals with substantial organizational roles (e.g., doctors who serve as chief medical officers), tend to view resource allocation issues and decision-making more similarly to their professional rather than management peers. These findings suggest the need to account for a primary collective identity in the context of multiple collectives, against which other identifications are generated and/or held. As a result, a physician employed by an organization is more likely to frame his or her identity as “a physician employed by this organization” rather than as an “employee of this organization.” Whereas the latter represents the simplicity implied by the traditional paradigm, the more complex notion represented by the former is more likely to capture the reality of multiple

194

GREGORY W. STEVENS

identifications as interactive and integrated from the individual’s perspective, reinforcing the need for such a shift in the literature. P1. The professional identity of employed physicians serves as a primary collective identity, in the context of which other work-related identities are generated and/or held. It is important to note at the outset that the strength of this effect is not assumed to be equally influential for all employed physicians in all cases. As Ashforth and colleagues (2008) have noted, it is possible to conceptualize a collective identity as comprising a core set of cognitive, evaluative and affective attributes concerning membership within a collective, surrounded by the content associated with that collective (e.g., values, goals, beliefs, stereotypic traits, knowledge, skills, and abilities). While the core represents the self-descriptive process of identification (e.g., “I am a physician who practices family medicine …”), the content represents the meaning that is associated with that label (e.g. “… so I value personalized relationships with patients and providing care across patients’ life spans …” and so forth). The boundary between the two is permeable to represent the fact that the content may not necessarily follow the core identity if such attributes are “unclear, emergent, in flux, conflicted, tacit, espoused but not enacted, and so on” (Ashforth et al., 2008, p. 330). The extant literature suggests several areas that may contribute to the degree to which this boundary is permeable. First, as noted above, the strength of the professional identity is related to the nature of the profession’s socialization process and barriers to gain membership. For professions where the socialization process is long and/or difficult, such as for physicians, the influence of the professional identity on other identities may be expected to be greater than for professions where the barriers to membership are considerably lower. Within the physician collective, there may similarly exist variability in terms of the socialization process, depending on a number of factors like specialization and training requirements for certain fields, continuing education requirements or certifications, and so forth. Another potential moderating factor is suggested by the extent to which professionals are engaged in daily tasks and activities that reflect their profession (e.g., Johnson et al., 2006). As Wallace (1995) notes, professionals may be employed by professional organizations, where the primary work is that of the profession (e.g., a physician working in a hospital or as a provider within a large health system) and nonprofessional organizations, where the primary work departs from the profession (e.g., a physician working for a pharmaceutical or insurance company). Where the

Engaging Employed Physicians

195

professional spends substantial portions of his or her day engaged in professional work, the strength of the professional identity may be expected to be stronger. Although this is not intended to represent an exhaustive list of all potential moderators, the literature suggests that these domains may be particularly influential. P2. The strength of the primary collective identity is moderated by (a) the investment and time required to join a given profession, and (b) the proportion of time the individual spends engaged in professional versus nonprofessional activities or tasks. Expanding on this last point regarding variability among individuals’ professional identities, it is perhaps useful to take stock of some of the assumptions that have typically followed such conceptualizations. More often than not, treatments of professional identity have failed to acknowledge the potential for different instantiations of that identity, instead assuming that professionals share a group identity that means the same thing to each individual (Hoff, 2001). As Hoff further elaborated, this identity has focused almost solely on professionals as a “unified occupational subculture that is both distinct from and often in conflict with the organization in which it is found,” ignoring differences among individual physicians (2001, p. 56). This view tends to be echoed rather loudly in the organizational sciences (e.g., Glynn, 2000; Gouldner, 1957; Hekman, Steensma et al., 2009; Johnson et al., 2006; for an exception, see Wallace, 1995). In a majority of these treatments, a common dynamic plays out: the professionals’ desire for autonomy versus the organization’s desire for control, resulting in resistance on the part of physicians and frustration on the part of management. There is an inherent limitation to such an assumption. As stated earlier, it ignores the variability among individuals’ professional identities, some aspects of which may be invested in a need for autonomy, but other aspects that are not. The focus on a single aspect of one’s professional identity comes at the expense of other aspects of that identity, such as perceiving one’s role as a caregiver, economic decision-maker, and/or organizational member (cf. Hoff, 2001). In the dominant paradigm, these other aspects are largely ignored, yet each may contribute important dynamics to the relationship between professional and organization, beyond a desire for autonomy and beyond inherent conflict. These arguments illustrate the need to appreciate the multidimensional and multifaceted identities of individual physicians. For example, while certain physicians may primarily orient themselves toward the earnings and status related aspects of their profession, others may identify more with the

196

GREGORY W. STEVENS

socially beneficial aspects of using professional knowledge to help others, or see their profession as a calling. As Postmes and Jetten (2006, p. 259) argue, such content serves as a “rudder” that specifies exactly what the effects and outcomes of identification will be. Not only may these differences in content influence the extent to which relationships are formed between the primary collective and the subsequent collective (e.g., seeing one’s profession as a calling and employment in a nonprofit vs. for-profit organization), but may also influence the nature of engagement. In terms of predicting engagement, for example, earnings-focused professionals may exhibit different patterns of participation than their quality-focused counterparts, particularly when the nature of engagement entails aspects that touch upon compensation versus quality improvement domains. Therefore, the concept of content as a rudder is particularly important given the focus here on the relationships between a primary collective and other relevant collectives in the social environment, both in terms of the impact on an individual’s identification and the outcomes of that identification in the form of engagement. P3. In considering the influence of a primary collective identity on subsequent collective identification, it is important to capture the specific content of that identity rather than rely on abstract labels or categories. P4. The content of one’s primary collective identity will influence patterns of engagement to the extent that such content is aligned with the underlying purpose of engagement.

THE NATURE OF SUBSEQUENT IDENTIFICATION IN CONTEXT Once the existence of a primary collective identity is acknowledged, it becomes possible to discuss how that identity influences subsequent identification. This shifts away from overly simplistic notions of separate multiple identifications to a more realistic consideration of how an individual’s memberships within multiple collectives relate to one another. The implication is that collectives do not exist in a vacuum, but rather have relationships to other collectives in the larger social context; these relationships ultimately influence the way in which the individual identifies. To date, the literature on the management of multiple identifications has focused on the individual and his or her relationships to multiple collectives,

Engaging Employed Physicians

197

such that a distinct individual collective relationship exists for each collective. The implicit assumption within this paradigm is that the individual perceives him or herself primarily as a member within each collective, and consequently must manage the different aspects of those collectives as a part of the self, either in a pluralistic or holistic sense (Ashforth & Johnson, 2001; Thoits & Virshup, 1997). Whereas pluralistic strategies entail maintaining distinctions between one’s membership in different collectives (e.g., one’s identity as a physician is compartmentalized from that of belonging to other collectives), holistic approaches involve combining elements of multiple identities to form a new amalgamation (e.g., a superordinate identity captures membership within both collectives, such that there are no longer any distinctions between them). Unfortunately, few if any of the extant theories on multiple identifications (cf. Foreman & Whetten, 2002; Pratt & Foreman, 2000; Roccas & Brewer, 2002) account for the presence of a primary collective identity, in the context of which other collectives are formed and/or held. For example, the presence of a primary collective identity that is dominant and central, and yet remains distinct from subsequent collectives such as the organization, precludes the potential for a superordinate collective to be instantiated; a superordinate collective is unlikely given the strength and salience of the professional identity and its distinctiveness from the organizational identity. On the other hand, the interdependent and complex nature of providing medical care suggests that a compartmentalized strategy is an unlikely outcome as well, as physicians rely on the support of practice staffs to manage patient flow and corporate structures to handle financial and regulatory concerns; despite the distinctiveness of each collective’s roles, there are important interdependencies to accomplishing the work of providing health care to patients. All of this suggests a middle ground, where the collectives are simultaneously differentiated, yet interdependent, while still acknowledging the subjective and relative importance of one collective in comparison to the other. Thus, the argument is made that employed physician’s experience of and identification with the organization is influenced by the nature of the professional organizational relationship, rather than the individual organizational relationship as past theory would suggest. The nature of this relationship is defined less by the degree of perceived overlap per se, and more by the particular relationships that occur between the organization and its professionals. These relationships allow for each domain to remain distinct, yet interconnected, mirroring more closely the way in which physicians think of themselves in the context of the organizations that employ

198

GREGORY W. STEVENS

them: less as employees of the organization, and more as professionals engaged in an interdependent relationship with the organization. This notion essentially recasts organizational identification, departing from prior work that conceptualizes it as a function of the relationship between the individual and the organization (cf. Ashforth & Mael, 1989; Bergami & Bagozzi, 2000). Rather, I argue that for employed physicians (and potentially for other individuals with salient primary collective identities), organizational identification is better understood as a function of the individual and the perceived relationship between his or her profession and the organization. P5. The nature of relationships and other interdependencies between the primary collective and the subsequent collective drives the nature of subsequent identification, such that an individual identifies with the role and relationships of the profession within the context of the organization rather than the organization itself. There are several ways this shift advances theory. First, one may differentiate between cases where an individual identifies with the organization itself versus when that individual identifies with the role of the profession within the organization, based on the degree of perceived congruity in the relationship between the profession and organization. Congruity may be defined as the extent to which the attributes of two collectives (i.e., values, goals, attitudes, tasks, and so forth) are similar in nature or function. Two relevant arrangements are possible that speak to the level of potential congruity of professionals within organizations (Hall, 1975; Scott, 1965, 1982; Wallace, 1995): employment in an autonomous organization, or employment in a heteronomous organization. The autonomous organization is one in which the structures, norms, goals, and policies are established by one or more professionals, such that the professional aspects of the organization are controlled by professional rather than nonprofessional managers. In contrast, heteronomous organizations are defined as ones in which the work of professionals falls under the control of nonprofessional managers and administrators (Scott, 1982). This form of organization is perhaps more common, particularly in terms of supporting some of the assumptions put forth in the previous sections regarding physician administration conflict. A further distinction among these categories that may be of relevance is the potential difference between organizations that are primarily professional versus those that are primarily nonprofessional (Wallace, 1995; see

Engaging Employed Physicians

199

also Johnson et al., 2006). Professional organizations are defined as those where the core services being provided are congruent with the direct practice of the profession. Examples might include private medical clinics or hospitals. On the other hand, nonprofessional organizations are defined as those were the core work departs from that which the profession typically includes, yet where professionals are still employed by the organization. An example might be where a doctor is employed by a pharmaceutical or insurance company in a medical consulting role. Traditionally, these differences have been approached from the perspective of being rooted in formal organizational structures. However, it may also be possible that such distinctions may be formed by individuals’ perceptions, particularly in organizations that are seen as complex and having multiple identities (Pratt & Foreman, 2000). For example, physicians working for a large health system may perceive the goal of the organization to be the provision of health care services (i.e., as a professional organization) or as a for-profit business which happens to be in the health care industry (i.e., as a nonprofessional organization). So, although these distinctions between different types of organizations are frequently presented as discrete categories, it may be more appropriate to view them as existing on a continuum, differentiated by the degree to which the domains of the professional and the administrative staff are positioned in relationship to one another, as well as the degree to which those features correspond to individuals’ perceptions. To summarize, whereas a sense of congruity between the primary collective and the subsequent collective exists on some level, either in the form of an autonomous organization or in the form of primarily professional work, one may expect the physician to identify with the organization itself, the role of professionals being less separable from that organization. On the other hand, where there is less congruity, either as a result of a heteronomous setting or primarily nonprofessional work, one would expect identification to flow more from the specific role of professionals within the organization. P6. Organizational identification of professionals varies according to the nature of the relationship between the professional collective and the organizational collective, where more congruency in terms of (a) professional leadership and/or (b) professional work is related to identification with the organization itself and less congruency is related to identification with the primary collective’s role in the context of the organization.

200

GREGORY W. STEVENS

EXPANDING UPON THE IDENTIFICATION ENGAGEMENT RELATIONSHIP The preceding distinction is important in the sense that different profiles of engagement may emerge from each form of identification with the organization, based in the understanding that identification predicts engagement as a result of the shared meanings that underlie perceptions of the self and subsequent behavior (Stryker & Burke, 2000). Thus, where prior research has looked rather abstractly at the identification engagement relationship (e.g., broad measures of organizational identification predicting a wide range of participatory outcomes; Dukerich et al., 2002), the position advocated here is that considering more specific forms of identification and their associated meanings will lead to a better understanding of the drivers of engagement. First, it is important to examine the relationship between identification and engagement, which is presently conceptualized as the willingness to engage in extra-role behaviors. These behaviors are not formally included as a part of a specific job, yet contribute to the overall functioning of the organization (Van Scotter, Motowidlo, & Cross, 2000). Broader than organizational citizenship behaviors, extra-role performance may include any number of “volitional” behaviors, such as participation in organizational change efforts (e.g., Hekman, Steensma et al., 2009) or cooperation as defined as desirable referral patterns, voluntary service on committees and task forces, or recommending the organization to colleagues over local competitors (e.g., Dukerich et al., 2002). An interesting dynamic emerges for employed physicians, in that behaviors that are typically required of other employees (e.g., supporting change) may be perceived as entirely voluntary despite administrators’ “mandates.” In part, this dynamic reflects the difficulty in incentivizing such behavior appropriately, as well as the general orientation of seeing professionals in organizations as “rock stars” that are often afforded substantial autonomy and discretion. As a result, extra-role behaviors are conceptualized as those which are largely voluntary from the perspective of the employed professionals. This distinction helps to form a strong case for the role of identification in such behaviors, since formal organizational incentives and/or punishments are largely expected to be of little consequence. Although studying the relationships between identification and these behaviors is not new, the current paradigm shift argues for a richer sense of connecting the meaning of identity (through its content, interrelationships, and context) to the meaning of behaviors (Stryker & Burke, 2000). Prior

Engaging Employed Physicians

201

research has tended to focus at times on aggregated scores of cooperation (Dukerich et al., 2002) or on single behaviors in terms of supporting change (e.g., Hekman, Steensma et al., 2009). These findings have often failed to address the underlying complexity and experience of what it means to actually engage in these behaviors. For example, the beneficiaries of such behaviors (e.g., patient, organization, and physician), the costs of such behaviors, and the effort required have all gone relatively unexamined, thus contributing to the ambiguity surrounding the link with employed physicians’ identification. For example, two physicians who are equally identified with their profession and with their organization under the traditional paradigm may see change in very different ways. Where the first might see the change as improving patient care and thus support the change, the second may see the same change as hurting patient care and thus resist the change. Much of this argument suggests the need for a more person-centric approach to this phenomenon, particularly in placing the emphasis on the meaning of action rather than on action in an abstract sense. One pathway that emerges in terms of providing a clearer link between the identification and the engagement of employed physicians is the extent to which both concepts share an underlying meaning. Expanding upon the prior discussion of autonomous and heteronomous organizations, one may examine the extent to which the identities of the professional and organizational collectives align with one another. In the autonomous case, the organization is led by professionals, who share the values and goals of the profession with that organization’s physician employees. As a result of this shared meaning, physician employees may be more likely to engage in extra-role behaviors that support those values and goals that are espoused by the organization’s leadership. Support for this assertion stems from the research that finds professionals are more engaged and involved in organizations where management is comprised of professionals versus when it is not (Moore & Coddington, 2009). In the case of the heteronomous organization, physicians may perceive much less congruency between their own values and goals and those of the administrators of the organization (even despite the appearance of similarity between espoused goals, like providing high quality care), and consequently, may be less likely to engage in behaviors that support the administration. This distinction may help to explain why employed physicians in Hekman, Bigley, and colleagues’ (2009) sample, working within a heteronomous organization, were less engaged in supporting organizational change despite reporting high levels of organizational and professional identification. Although more research

202

GREGORY W. STEVENS

is required to confirm this notion, accounting for different forms of organizational identification offers a lens through which to interpret such research. P7. Identification with the subsequent collective under higher levels of congruency between collectives (i.e., extent of shared underlying meaning) is related to higher levels of engagement than identification under lower levels of congruency. Building from the notion of incongruency between collectives and the subsequent tendency to identify with the role of the profession within the organization, a further contribution of this theory is the incorporation of different patterns of relationships between the primary collective and organization. Once again, it is important to move beyond the extant paradigms that examine how individuals manage circumstances in which multiple collectives are salient. Most commonly, social identity theorists (cf. Hogg & Terry, 2000) have advocated the importance of the in-group versus outgroup distinction and its associated effects on behavior, namely conflict and lack of trust. Within this theory, individuals undergo a depersonalization process and become microcosms of a social collective or group, forming a salient in-group. Those that occupy out-groups are subsequently viewed with suspicion and a lack of trust, which has helped to explain the traditional conflict model between physicians (as the in-group) and administrators (as an untrusted out-group). Unfortunately, this model fails to account for an important middle ground, in which interdependencies exist between distinct collectives and those relationships that are more based in terms of differentiated roles than competition for resources. More recently, Fiol and colleagues (2009) have advocated a pathway to resolve these types of relationships in which conflict arises out of identification processes. They argue for the establishment of a superordinate collective, framed by the interdependencies that exist between subgroups. Unfortunately, as noted earlier, this possibility is somewhat constrained by the extent to which the collectives are viewed as distinct by employed physicians. In other words, physicians are unlikely to think of their profession as a subgroup within an organization, united by the common mission of providing medical care. The lack of fit with previous theories suggests that a more nuanced alternative needs to be developed. To that end, it may be more beneficial to focus on how employed professionals specifically perceive the relationship between their primary collective and the organization on functional terms. For example, one possibility is that employed physicians do not perceive themselves as

Engaging Employed Physicians

203

“employees” of an organization, but nevertheless understand the working relationship between what they do as professionals and what the organization does to support them. A physician might see him or herself as something of a “gatekeeper” within the process of providing medical care, yet nevertheless understands how the practice staff maintains an efficient patient flow to support the actual patient encounter, while the organization helps to manage the changing external environment and also manage the business aspects of running the practice (i.e., through staffing, ordering supplies, and so forth). This alternative does not fit neatly into either paradigm noted above, in that the professionals do not perceive themselves as belonging to the organization, removing the possibility for engagement that is rooted in the sense of belonging to the same superordinate collective (Fiol et al., 2009), nor do they necessarily perceive the organization as belonging to an out-group. Rather, professionals possess complex understandings of how their work is intertwined with the work of organizations that employ them, which likely serves as a basis for identification processes and potential engagement at the point of intersection between those collectives. To the extent that the content or domains of these functional relationships overlaps, one may expect higher engagement with organizational initiatives. For example, a physician might have entered into an employment relationship with an organization with the expectations of a clearly defined practice management support function. For the physician who sees those practice management aspects as intertwined with what a physician does (e.g., coordinated approach to improve patient care and experience), he or she is more likely to become engaged in initiatives that touch upon that shared domain. On the other hand, for the physician who perceives the professional and organizational functions as distinct and nonoverlapping, he or she is much less likely to become engaged, instead seeing that same initiative as squarely falling under the responsibility of the organization and not the physician. In this manner, when there is initial incongruence between the professional and the organizational collectives, engagement will be driven by the specific expectations that stem from the perception of role relationships between those collectives. P8. Under conditions of greater incongruity, the professional’s perception of the role relationship between the primary collective and subsequent collective will inform the manner of identification, such that engagement will be higher for interdependent functions or shared domains.

204

GREGORY W. STEVENS

CONCLUSION The trend within health care toward employing physicians in large organized delivery systems appears to be as popular today as it was during the initial rush in the late 1990s and early 2000s. Unfortunately, today’s physicians are unlikely to enjoy the same freedom to move in and out of these types of relationships as their colleagues had in the past; changes to reimbursement structures and federal regulations (e.g., value-based purchasing, accountable care organization models) have moved to support these integrated models and made it more difficult for independent practices to remain profitable (Kocher & Sahni, 2011). Rather, the impetus is on health care management scholars and practitioners to examine these physician-system relationships, both in order to avoid the problems of the past in terms of misalignment and conflict, and to move toward higher levels of integration and coordination of care. For this to happen, the current paradigm needs to shift away from the traditional dichotomy of physicians versus organizations, often captured through the notions of professional and organizational identification, respectively, and its overly restrictive assumptions of how engagement flows or doesn’t from those concepts. Instead, the current proposal for a shift in this paradigm argues that the type of engagement required in integrated care delivery models is instead a function of a form of collective identification that acknowledges: the professional identity as a primary identity against which other work-related identities are developed and/or held; and that the specific content and relationships within and between those collectives (i.e., the profession and the organization) will drive particular patterns of engagement. In providing a better understanding of these dynamics, one moves beyond unrealistic and oversimplified notions where an abstract form of high organizational identification and low professional identification contributes to engagement on the part of physicians. Instead, the argument is made that engagement is best achieved by understanding a physician’s perception of his or her relationships to the profession and the organization in terms of how they interact, finding domains of shared values and interdependence, and finally, leveraging those domains to establish a foundation upon which engagement can be built.

Implications for Theory and Practice Essentially, many of the issues that integrated delivery systems face with respect to employed physicians stem from “people problems,” relationships

Engaging Employed Physicians

205

that are more likely to be characterized by conflict or ambivalence than cooperation. This presents a unique opportunity for researchers and practitioners in health care management fields to attempt to understand these relationships, with an eye toward not only improving business processes, but patient care delivery and employee satisfaction as well. This latter point touches upon the potential significance of a paradigm shift like this one, as these ideas provide insight into the complex context and dynamics of employing physicians. Although a majority of the theoretical implications of this paradigm shift have been addressed, there are several areas that merit further attention. Perhaps the primary goal is to encourage work that acknowledges both the complexity of the working environment within health care organizations, in terms of multiple collectives, as well as the interacting nature of those collectives. Trends in the identity literature have begun to appreciate the role of nested and cross-cutting groups on identification processes (Ashforth & Johnson, 2001); yet, the experiences of employed physicians do not fit neatly into either group. Instead, research needs to examine the potential for distinct collectives to share complex relationships with one another, as professions and organizations often do. The current work is only a single, first step toward capturing such relationships, and future research would be well served to expand upon the myriad possibilities offered by the increasing complexity of health care organization structures and arrangements. For example, one of several avenues that may be particularly valuable involves the introduction of interprofessional or team-based care within patient-centered medical home models, which adds another team-level collective to consider alongside the professional and the organizational collectives. Valuable questions center on how team-level identification may influence or be influenced by both professional and organizational collectives. Such research could also be expanded to address more recent trends, as health care systems move to employ more specialists in addition to primary care physicians (Kocher & Sahni, 2011). Given the factors discussed above, such as the impact of the socialization process and the potential for complex and idiosyncratic identities, the professional collective may require further research to understand the relational dynamics between specialists and primary care in a single employment setting, each of which may bring different values and goals to the table, and would likely require different engagement strategies. Fortunately, the focus in the current paradigm on appreciating variability among forms of collective identification seems well suited to the task. There are also a number of practical considerations and implications that stem from the current work. Foremost, the extant literature has

206

GREGORY W. STEVENS

advocated a strong role for organizational identification in encouraging extra-role behaviors, and for professional identification in hindering those behaviors regardless of the level of organizational identification (Hekman, Bigley et al., 2009). Given prior research on the relative resistance of identity to change (Ashforth et al., 2008) and the centrality of the physician identity as proposed here, attempts to change employed physicians’ identities to align with organizational attributes is unlikely to be a successful route to engagement. Rather, practitioners may be well-served to focus instead on identity mobilization and the influence that context plays in that process (cf. Kreindler et al., 2012). Such recommendations fit well into the proposed conceptualization of collective identity, which specifically models the role that context and relationships between collectives play in forming the basis for engagement. For example, by highlighting and focusing on shared or interdependent roles (e.g., emphasizing the shared values that the organization’s implementation of electronic health records and the physician have in delivering patient care), practitioners in health care management may be successful in tapping into or mobilizing a physician’s collective identity (rather than attempting to change it) to support a given change initiative, the key being the establishment of shared meaning in context. The challenge for practitioners then, is to find a way for this pattern of identification to encourage engagement. The current paradigm shift offers a valuable avenue, in that it provides for an understanding of how professionals come to understand their organization through the lens of their professional identity, as well as account for variability in terms of those patterns of identification with other relevant collectives. From a practical standpoint, this suggests the importance of fully understanding one’s physician workforce, in terms of their motivations as physicians, their perceptions of their role within the organization and the organization’s role within their professional work. By leveraging commonalities in terms of interdependent work and building congruency through shared values and goals that are clear to all parties, engagement is a more likely outcome than it would be through attempts to build a strong sense of organizational membership while minimizing the influence of membership within the profession.

ACKNOWLEDGMENT The author would like to thank Malissa Clark for her valuable feedback on earlier drafts, as well as Daniel Svyantek, Stanley Harris, and Jinyan Fan for their assistance during the early development of this manuscript.

207

Engaging Employed Physicians

REFERENCES Albert, S., Ashforth, B. E., & Dutton, J. E. (2000). Organizational identity and identification: Charting new waters and building new bridges. Academy of Management Review, 25, 13 17. Ashforth, B. E., & Johnson, S. A. (2001). Which hat to wear? The relative salience of multiple identities in organizational contexts. In M. A. Hogg & D. J. Terry (Eds.), Social identity processes in organizational contexts (pp. 31 48). Philadelphia, PA: Psychology Press. Ashforth, B. E., & Mael, F. (1989). Sociality identity theory and the organization. Academy of Management Review, 14, 20 39. Ashforth, B. E., Harrison, S. H., & Corley, K. G. (2008). Identification in organzations: An examination of four fundamental questions. Journal of Management, 34, 325 374. Ashmore, R. D., Deaux, K., & McLaughlin-Volpe, T. (2004). An organizing framework for collective identity: Articulation and significance of multidimensionality. Psychological Bulletin, 130, 80 114. Bamber, E. M., & Iyer, V. M. (2002). Big 5 auditors’ professional and organizational identification: Consistency or conflict? Auditing: A Journal of Practice and Theory, 21, 21 38. Bergami, M., & Bagozzi, R. P. (2000). Self-categorization, affective commitment, and group self-esteem as distinct aspects of social identity in the organization. British Journal of Social Psychology, 39, 555 577. Blader, S. L., & Tyler, T. R. (2009). Testing and extending the group engagement model: Linkages between social identity, procedural justice, economic outcomes and extrarole behavior. Journal of Applied Psychology, 94, 445 464. Brewer, M. B., & Gardner, W. (1996). Who is this “we”? Levels of collective identity and self representation. Journal of Personality and Social Psychology, 71, 83 93. Dukerich, J. M., Golden, B. R., & Shortell, S. M. (2002). Beauty is in the eye of the beholder: The impact of organizational identification, identity, and image on cooperative behaviors of physicians. Administrative Science Quarterly, 47, 507 533. Etzioni, A. (1969). The semi-professions and their organization. New York, NY: The Free Press. Fiol, C. M., Pratt, M. G., & O’Connor, E. J. (2009). Managing intractable identity conflicts. Academy of Management Review, 34, 32 55. Foreman, P., & Whetten, D. A. (2002). Members’ identification with multiple-identity organizations. Organization Science, 13, 618 635. Glynn, M. A. (2000). When cymbals became symbols: Conflict over organizational identity within a symphony orchestra. Organization Science, 11, 285 298. Golden, B. R., Dukerich, J. M., & Fabian, F. H. (2000). The interpretation and resolution of resource allocation issues in professional organizations: A critical examination of the professional-manager dichotomy. Journal of Management Studies, 37, 1157 1187. Gouldner, A. W. (1957). Cosmopolitans and locals: Toward an analysis of latent social roles I. Administrative Science Quarterly, 2, 281 306. Gouldner, A. W. (1958). Cosmopolitans and locals: Toward an analysis of latent social roles II. Administrative Science Quarterly, 2, 444 480. Grube, J. A., & Piliavin, J. A. (2000). Role identity, organizational experiences, and volunteer performance. Personality and Social Psychology Bulletin, 26, 1108 1119. Hall, R. T. (1975). Occupations and the social structure (2nd ed.), Englewood Cliffs, NJ: Prentice Hall.

208

GREGORY W. STEVENS

Hekman, D. R., Bigley, G. A., Steensma, H. K., & Hereford, J. F. (2009). Combined effects of organizational and professional identification on the reciprocity dynamic for professional employees. Academy of Management Journal, 52, 506 526. Hekman, D. R., Steensma, H. K., Bigley, G. A., & Hereford, J. F. (2009). Effects of organizational and professional identification on the relationship between administrators’ social influence and professional employees’ adoption of new work behavior. Journal of Applied Psychology, 94, 1325 1335. Hickson, D. J., & Thomas, M. W. (1969). Professionalization in Britain: A preliminary measurement. Sociology, 3, 37 53. Hoff, T. J. (2001). The physician as worker: What it means and why now? Health Care Management Review, 26(4), 53 70. Hogg, M. A., & Terry, D. J. (2000). Social identity and self-categorization processes in organizational contexts. Academy of Management Review, 25, 121 140. Jimmieson, N. L., & White, K. M. (2011). Predicting employee intentions to support organizational change: An examination of identification processes during a re-brand. British Journal of Social Psychology, 50, 331 341. Johnson, M. D., Morgeson, F. P., Ilgen, D. R., Meyer, C. J., & Lloyd, J. W. (2006). Multiple professional identities: Examining differences in identification across work-related targets. Journal of Applied Psychology, 91, 498 506. Kocher, R., & Sahni, N. R. (2011). Hospitals’ race to employ physicians The logic behind a money-losing proposition. New England Journal of Medicine, 364, 1790 1793. Kramer, R. (1993). Cooperation and organizational identification. In J. K. Murnighan (Ed.), Social psychology in organizations (pp. 244 268). Englewood Cliffs, NJ: Prentice Hall. Kreindler, S. A., Dowd, D. A., Star, N. D., & Gottschalk, T. (2012). Silos and social identity: The social identity approach as a framework for understanding and overcoming divisions in health care. Milbank Quarterly, 90, 347 374. Lipponen, J., Bardi, A., & Haapamaki, J. (2008). The interaction between values and organizational identification in predicting suggestion-making at work. Journal of Occupational and Organizational Psychology, 81, 241 248. Moore, K. D., & Coddington, D. C. (2009). Multiple paths to integrated health care. Healthcare Financial Management, 63(12), 46 54. Postmes, T., & Jetten, J. (2006). Reconciling individuality and the group. In T. Postmes & J. Jetten (Eds.), Individuality and the group: Advances in social identity (pp. 258 269). London: Sage. Pratt, M. G. (1998). To be or not to be? Central questions in organizational identification. In D.A. Whetten & P. C. Godfrey (Eds.), Identity in organizations: Building theory through conversations (pp. 171 207). Thousand Oaks, CA: Sage. Pratt, M. G., & Foreman, P. O. (2000). Classifying managerial responses to multiple organizational identities. Academy of Management Review, 25, 18 42. Roccas, S., & Brewer, M. B. (2002). Social identity complexity. Personality and Social Psychology Review, 6, 88 106. Scott, W. R. (1965). Reactions to supervision in a heteronomous professional organization. Administrative Science Quarterly, 10, 65 81. Scott, W. R. (1982). Managing professional work: Three models of control for health organizations. Health Services Research, 17, 213 240. Shortell, S. M., Gillies, R. R., Anderson, D. A., Erickson, K. M., & Mitchell, J. B. (2000). Remaking health care in America. San Francisco, CA: Jossey-Bass.

Engaging Employed Physicians

209

Stryker, S., & Burke, P. J. (2000). The past, present, and future of an identity theory. Social Psychology Quarterly, 63, 284 297. Thoits, P. A., & Virshup, L. K. (1997). Me’s and we’s: Forms and functions of social identities. In R. D. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental issues (pp. 106 133). New York, NY: Oxford University Press. Van Dick, R., Grojean, M. W., Christ, O., & Wieseke, J. (2006). Identity and the extra-mile: Relationships between organizational identification and organizational citizenship behavior. British Journal of Management, 17, 283 301. Van Scotter, J. R., Motowidlo, S. J., & Cross, T. C. (2000). Effects of task performance and contextual performance on systemic rewards. Journal of Applied Psychology, 85, 526 535. Wallace, J. E. (1995). Organizational and professional commitment in professional and nonprofessional organizations. Administrative Science Quarterly, 40, 228 255. Zuckerman, H. S., Hilberman, D. W., Andersen, M. R., Burns, L. R., Alexander, J. A., & Torrens, P. (1998). Physicians and organizations: Strange bedfellows or a marriage made in heaven? Health Services Management, 14, 3 34.

SECTION IV FUTURE DIRECTIONS

This page intentionally left blank

CONCLUDING REMARKS Jim Goes Viewed from a historical perspective, health care reform has been more of a process a very long process than a specific event, or even a series of specific events. From the passage of Medicare in the 1960s through the Affordable Care Act (ACA) in 2010, the evolution of health systems and organizations has ebbed and flowed, and in the process health care organizations have become more efficient in many ways, less in others, and unquestionably more complex. The dominant forms of health care organization in the 1960s were the solo or small group physician practitioner and the freestanding community hospital. Years of restructuring, consolidation, and integration, both horizontally and vertically, have resulted in the far different configuration of system, hospital, and physician practice designs we see today. Efforts to wring greater efficiency and effectiveness out of the system at multiple levels have worked, to a degree, but the United States still spends a far greater percentage of its GDP on health care, about 18 percent or $8,200 per capita (WHO, 2011), than any other nation. On major indices of health, including infant mortality and life expectancy, the United States consistently underperforms relative to other wealthy countries. For a nation and culture that prides itself on getting its money’s worth, we have a long way to go in reaching that standard for health care. The ACA, like many other initiatives in the long history of health care reform, promises to widen coverage and care, achieve greater efficiency,

Annual Review of Health Care Management: Revisiting the Evolution of Health Systems Organization Advances in Health Care Management, Volume 15, 213 215 Copyright r 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-8231/doi:10.1108/S1474-8231(2013)0000015017

213

214

JIM GOES

and improve the health of US citizens. That is a tall order indeed, but the fresh evidence and ideas presented in the chapters in this volume have substantial potential to inform not only the debate around ACA, but also the successful execution of substantive reform. Taken as a whole, two integrative themes emerge from the chapters presented in this volume. First, the great complexity of modern health care organizations continues to challenge both system designers and leaders. Despite brief and intense interest about a decade ago in complexity theory as a paradigm for health system design, this approach, as McDaniel, Jr. and colleagues note, “has not had as widespread an effect as some have hoped.” They suggest that for many health care leaders, “the fact that the resolution of problems in health care may rest in complexity and uncertainty rather than in standard traditional science is frustrating.” Perhaps that frustration results in part from the lack of success of synoptic approaches to design and management in substantively and consistently reducing costs and improving health outcomes. In her commentary on complexity in this volume, Anderson suggests an alternative, by refocusing inside the organization and tapping the “capacity to improve (that) is embedded in the interactions among the interdisciplinary team.” In short, rather than tradition top-down design, system designers may find it more effective to cut loose innovation inside the organization, enlisting rank and file managers and caregivers in the effort. We talk a lot about system reorganization from the outside, but the key to managing complexity may reside in self-organization from the inside out. The second overarching theme of this volume pertains to the effectiveness of efforts at hospital physician integration. Inducing or controlling physician behavior to achieve greater efficiency and improved clinical outcomes has been a long-standing effort in most health systems, but success in doing so has been more problematic. Many systems have devoted considerable efforts to build large, integrated group practices connected to the hospital or system, but the extensive review of this strategy by Burns and his colleagues in this volume finds “the physician sector still remains the least consolidated portion of the health care value chain; the vast majority of doctors continue to practice in small groups and solo settings.” The distribution of physician group practice has largely bifurcated into what Burns et al. call “the two tails”: a large majority of practices in small groups, and a small minority of very large groups. Yet the authors’ review of 10 years of data finds little compelling evidence that either strategy is superior.

215

Concluding Remarks

The remaining chapters in this volume drill down into the processes of physician integration, considering its evolution over time from mechanistic approaches to self-governance (Evans, et al.), emergence of new governance archetypes (Al-Amin, et al.), and rethinking physician engagement and collective identity (Stevens). Each of these chapters offers original insights and thoughtful ideas around how to redefine physician integration strategies to get closer to the goals of greater efficiency and improved outcomes. Overall, the evidence and ideas presented in this volume suggest two important takeaway lessons. First, that complexity and integration in health care organizations remains a major challenge for system designers and industry leaders, and despite considerable effort to achieve integration for economic or managerial purposes there is much yet to be done. As yet, there has been no “silver bullet” strategy that stands out as superior. True success at integration may require both a deeper and a more thoughtful understanding of complexity how to not only manage it but also harness it by enabling self-organization among professionals within the system. Second, as Scott (2000) observed, health care organizations are highly institutionalized, and these institutional pressures extend to the strategies organizations use. To solve the puzzle of health systems integration, system designers and leaders may need to cultivate both a willingness to experiment with new strategies, and the discipline to abandon those that do not result in the desired outcomes.

REFERENCES Scott, W. R. (2000). Institutional change and healthcare organizations: From professional dominance to managed care. Chicago, IL: University of Chicago Press. World Health Organization (WHO). (2011). World health statistics 2011. Geneva: WHO.