The Neurotic Paradox, Volume 1: Progress in Understanding and Treating Anxiety and Related Disorders (World Library of Mental Health) [Volume 1, 1 ed.] 1138850799, 9781138850798

Citation preview

“What a wonderful overview of a program of research that carries with it much of the history of behavior therapy together with the development of an outstanding career. This volume makes for interesting reading for anyone at any level with an interest in clinical psychology and the development of behavior therapy. Moreover, there is a wide range of subject matter and of older papers that still merit reading.” —W. Stewart Agras, MD, Professor of Psychiatry Emeritus, Stanford University “Assembled in one place, here, is a collection of papers by Barlow and colleagues outlining the next paradigm in the psychological science of anxiety disorders. The Neurotic Paradox reflects one important way out of the proliferation of treatment manuals that plagues the dissemination of evidence based practice in clinical psychology. The impact of this work will be as profound as it will be broad based.” —Terry M. Keane, PhD, Professor of Psychiatry & Psychology, Boston University School of Medicine

Taylor & Francis Taylor & Francis Group

http://taylorandfrancis.com

The Neurotic Paradox

This collection of David H. Barlow’s key papers are a testimony to the collaborative research that he engendered and directed with associates who now stand with him at the forefront of experimental psychopathology research and in the treatment of anxiety and related disorders. His research on the nature of anxiety and mood disorders resulted in new conceptualizations of etiology and classification. This research led to new treatments for anxiety and related emotional disorders, most notably a new transdiagnostic psychological approach that has been positively evaluated and widely accepted. Clinical psychology will benefit from this collection of papers with its connecting commentary. David H. Barlow, PhD, is Professor and Founder of the Center for Anxiety and Related Disorders at Boston University. He has received numerous awards and has published over 600 articles and chapters and over 75 books; his research has been continuously funded by the National Institutes of Health for over 40 years.

Routledge World Library of Mental Health

The World Library of Mental Health celebrates the important contributions to mental health made by leading experts in their individual fields. Each author has compiled a career-long collection of what they consider to be their finest pieces: extracts from books, journals, articles, major theoretical and practical contributions, and salient research findings. For the first time ever the work of each contributor is presented in a single volume so readers can follow the themes and progress of their work and identify the contributions made to, and the development of, the fields themselves. Each book in the series features a specially written introduction by the contributor giving an overview of his career, contextualizing his selection within the development of the field, and showing how his own thinking developed over time.

Rationality and Pluralism—The selected works of Windy Dryden By Windy Dryden The Price of Love—The selected works of Colin Murray Parkes By Colin Murray Parkes Attachments: Psychiatry, Psychotherapy, Psychoanalysis—The selected works of Jeremy Holmes By Jeremy Holmes Passions, Persons, Psychotherapy, Politics—The selected works of Andrew Samuels By Andrew Samuels Towards a Radical Redefinition of Psychology—The selected works of Miller Mair Edited by David Winter and Nick Reed Living Archetypes—The selected works of Anthony Stevens By Anthony Stevens Soul: Treatment and Recovery—The selected works of Murray Stein By Murray Stein A Developmentalist’s Approach to Research, Theory, and Therapy—The Selected Works of Joseph Lichtenberg By Joseph D. Lichtenberg

The Neurotic Paradox

Progress in Understanding and Treating Anxiety and Related Disorders Volume 1

Edited by David H. Barlow

First published 2016 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN and by Routledge 52 Vanderbilt Avenue, New York, NY 10017, USA Routledge is an imprint of the Taylor & Francis Group, an informa business © 2016 Taylor & Francis The right of the editor to be identified as the author of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested. ISBN: 978-1-138-85079-8 (hbk) Typeset in Minion by Apex CoVantage, LLC

Epigram

In 1950, O. Hobart Mowrer described a mystery: [It is] the absolutely central problem in neurosis and therapy. Most simply formulated, it is a paradox—the paradox of behavior which is at one and the same time self-perpetuating and self-defeating! . . . Common sense holds that a normal, sensible man or even a beast to the limits of his intelligence, will weigh and balance the consequences of his acts: if the net effect is favorable, the action producing it will be perpetuated; and if the net effect is unfavorable, the action producing it will be inhibited, abandoned. In neurosis, however, one sees actions which have predominantly unfavorable consequences; yet they persist over a period of months, years, or a lifetime. (p. 486)

Taylor & Francis Taylor & Francis Group

http://taylorandfrancis.com

Contents

Permissions Acknowledgmentsxi

Introduction: A Career in Psychology

1

Methodology and Clinical Research

29

  1 Behavior Therapy: The Next Decade

31

DAVID H. BARLOW

  2 Why Can’t We Be More Idiographic in Our Research?

44

DAVID H. BARLOW AND MATTHEW K. NOCK

Treatment of Anxiety and Related Disorders

49

 

 

  5 Behavioral Conception and Treatment of Panic

68

DAVID H. BARLOW

  6 The Process of Fear and Anxiety Reduction: Affective Therapy DAVID H. BARLOW

75

x Contents

  7 A Comparison of Alprazolam and Behavior Therapy in Treatment of Panic Disorder

105

JANET S. KLOSKO, DAVID H. BARLOW, ROBIN TASSINARI, AND JEROME A. CERNY

  8 The Development of an Anxiety Research Clinic

122

DAVID H. BARLOW

  9 Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder: A Randomized Controlled Trial

125

DAVID H. BARLOW, JACK M. GORMAN, M. KATHERINE SHEAR, AND SCOTT W. WOODS

10 Toward a Unified Treatment for Emotional Disorders

141

DAVID H. BARLOW, LAURA B. ALLEN, AND MOLLY L. CHOATE

Nature, Diagnosis, and Etiology of Anxiety and Related Disorders 11 The Phenomenon of Panic

167 169

DAVID H. BARLOW, JAMES VERMILYEA, EDWARD B. BLANCHARD, BONNIE B. VERMILYEA, PETER A. DI NARDO, AND JEROME A. CERNY

12 Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference183 DAVID H. BARLOW

Indexi1 Articles 13–21 appear in Volume 2

Permissions Acknowledgments

1 Barlow, D. H. (1980). Behavior therapy: The next decade. Behavior Therapy, 11(3), 315–328. doi:10.1016/S0005-7894(80)80049-9 2 Barlow, D. H., & Nock, M. K. (2009). Why can’t we be more idiographic in our research? Perspectives on Psychological Science, 4(1), 19–21. doi:10.1111/j.1745-6924.2009.01088.x 3 Agras, S., Leitenberg, H., & Barlow, D. H. (1968). Social reinforcement in the modification of agoraphobia. Archives of General Psychiatry, 19(4), 423–427. doi:10.1001/archpsyc.1968.01740100039006 4 Barlow, D. H., & Wolfe, B. E. (1981). Behavioral approaches to anxiety disorders: A report on the NIMH-SUNY, Albany, research conference. Journal of Consulting and Clinical Psychology, 49(3), 448–454. doi:10. 1037/0022-006X.49.3.448 5 Barlow, D. H. (1986a). Behavioral conception and treatment of panic. Psychopharmacology Bulletin, 22(3), 802–806. 6 Barlow, D. H. (1988). The process of fear and anxiety reduction: Affective therapy. In D. H. Barlow, Anxiety and its disorders (pp. 285–318). New York: Guilford Press. 7 Klosko, J. S., Barlow, D. H., Tassinari, R., & Cerny, J. A. (1990). A comparison of alprazolam and behavior therapy in treatment of panic disorder. Journal of Consulting and Clinical Psychology, 58(1), 77–84. doi:10.1037/0022-006X.58.1.77 8 Barlow, D. H. (1992). The development of an anxiety research clinic. In D. K. Freedheim (Ed.), History of psychotherapy: A century of change (pp. 429–431). Washington D.C.: American Psychological Association.  9 Barlow, D. H., Gorman, J. M., Shear, M. K., & Woods, S. W. (2000). Cognitive-behavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled trial. JAMA, 283(19), 2529–2536. doi:10.1001/jama.283.19.2529 10 Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35(2), 205–230. doi:10.1016/S0005-7894(04)80036-4

xii  Permissions Acknowledgments

11 Barlow, D. H., Vermilyea, J., Blanchard, E. B., Vermilyea, B. B., Di Nardo, P. A., & Cerny, J.A. (1985). The phenomenon of panic. Journal of Abnormal Psychology, 94(3), 320–328. doi:10.1037/0021-843X.94.3.320 12 Barlow, D. H. (1986b). Causes of sexual dysfunction: The role of anxiety and cognitive interference. Journal of Consulting and Clinical Psychology, 54(2), 140–148. doi:10.1037/0022-006X.54.2.140

Introduction: A Career in Psychology

I began graduate school in clinical psychology in 1964. Now, as I write this introductory chapter, I am 50 years into a stimulating and rewarding career. Thus, when presented with an opportunity by my publisher to recap the highlights of this career in the form of a strategic selection of publications, I can only think that it is a rare privilege and pleasure indeed for several reasons. First is the fact that I am still here with my cognitive processes relatively intact (an observation that some would dispute). Second, this opportunity affords me the chance to temporarily pull out of the day-today demands of a busy ongoing program of clinical research, as well as the usual and customary academic duties, in order to reflect on the major trends and developments that have occurred during the last 50 years. In addition to reflecting on these trends and any small part I may have played in developments over this period, the hope is that close observation of the path and process of progress in various areas in which I have worked will be useful in providing some perspective to current and future generations of clinicians and clinical scientists on the nature of scientific advancement in our field. Accordingly, I have organized this retrospective by choosing publications from four major areas. The first area focuses on fundamental methodology in clinical research, particularly in the area of idiographic, single case experimental methods in the development of psychological treatments, as reflected in two papers written almost 30 years apart. The second section carries on with another major theme in my research ongoing for the full 50 years, the development of new treatments for anxiety and related disorders. The third section covers research delving more deeply into the nature of anxiety and negative affect, with implications for the study of the etiology, diagnosis, and classification of anxiety and related disorders. In the final section, I turn more toward policy, with a description of the trials and tribulations of promoting and disseminating evidence-based psychological treatments, a process in which I have been heavily involved for the past 25 years. But before describing the publications chosen to represent developments in each of these four areas, it would seem useful to say something about my own odyssey and the intellectual and experiential influences that provided the context and that, at least in part, determined the substance of my contributions over the decades.

2  Introduction

Boston: The Early Years

My undergraduate years were spent in an idyllic fashion at the University of Notre Dame, where I was deeply immersed in the academic and athletic culture of what was then an all-male university. But due to its strong Catholic heritage, a gaping hole in Notre Dame’s academic portfolio was the absence of a department of psychology. As with many Catholic universities, those responsible for the curriculum at Notre Dame assumed that the principles of psychology were adequately covered in the myriad philosophy courses available. As it turned out, Notre Dame was one of the last Catholic universities to correct this omission. This was not terribly concerning to me at first since I loved to read and became deeply immersed in literature, including fiction but also biography. And I soon realized that the most fascinating books I encountered involved the psychoanalytic study of literary characters! With my long-standing interest in how people behave (one of my less than endearing traits was setting up practical jokes to see how people reacted to improbable situations) and a perception by my friends that I was a good listener, I realized that I was very strongly interested in psychology. Knowing that I wanted to major in psychology but also that I did not want to leave Notre Dame, I began taking advantage of the psychology courses that were offered, mostly in the sociology department (my official major), and I accumulated sufficient credits to apply to graduate school with one exception—a laboratory course in experimental psychology. To meet this requirement, I enrolled at Boston College for an intensive course in experimental psychology with accompanying laboratory experiences during the summer of 1963 prior to my senior year. That course happened to have been taught by Joseph R. Cautela. During that intensive six-week experience, immersed as we were every day in the classroom and lab, Cautela made an indelible impression, persuading me that only through a reliance on the slow but inexorable process of science could the applications of psychological principals to human problems truly advance. This was a very different take on psychology from my initial interests in the in-depth exploration of personality and also a very different message (I was to learn) from that delivered by other clinical faculty at Boston College, who believed that the scientific method was incapable of unraveling the complexities of the human spirit, an activity that required a different, more introspective way of knowing. I began graduate school in the fall of 1964 at Boston College in the MA program in clinical psychology. For me, the program at Boston College afforded me a chance not only to return to my home town of Boston but also to continue working with Cautela. Cautela was an experimental psychologist, as were most of the early adherents to the fledgling movement of behavior therapy or behavior modification, but he maintained a clinical practice mostly in the evenings. This was not uncommon in those days since professional and national guidelines articulating very clear boundaries between clinical psychology and nonclinical psychological training had not yet been developed; indeed, it was common for most experimental psychologists to have some exposure to applied issues during training and to engage in some sort of consulting experience in addition to academic and research duties. Cautela was very interested

Introduction  3

in what was then called “modern learning theory,” comprised mostly of principles of classical and operant conditioning. The previous summer, pursuing this interest in a more clinical context, he had attended a two-week summer training institute, with the South African psychiatrist Joseph Wolpe, in Philadelphia. Wolpe was already recognized as one of the fathers of behavior therapy based on his 1958 book, Psychotherapy by Reciprocal Inhibition. He developed this point of view early in his training after becoming disillusioned with prevailing psychoanalytic views and, looking for alternatives, becoming deeply interested in the work of Pavlov. At that point, Wolpe decided to do a doctoral thesis as part of his psychiatric residency, an option then available in the British system of medical training in South Africa. In a very strange twist of fate, he then came under the influence of an American psychologist, Leo Reyna, who had recently received his PhD from the University of Iowa working with Kenneth Spence and who was spending a few years in South Africa, the only place he could find employment. Reyna went on to direct Wolpe’s thesis on reducing fear (in cats) through a counterconditioning process. After leaving South Africa, Reyna ended up at Boston University, where he soon began mentoring another doctoral student, Joe Cautela! After escaping South Africa, where he was in danger of arrest due to his strong and activist anti-apartheid views, Wolpe spent some time in England and in the United States before moving to Temple University Medical School and the Eastern Pennsylvania Psychiatric Institute in the early 1960s, where he began conducting his summer institutes. Cautela, having been introduced to Wolpe by his mentor, Leo Reyna, came back from the Institute inspired to begin administering more “learning theory”–based approaches with his patients. Cautela had observed Wolpe’s technique of Systematic Desensitization (SD), in which patients with phobias and anxieties would imagine their phobic situations arranged on a hierarchy based on patient ratings of intensity of fear while deeply relaxed under the therapist’s direction. Patients would slowly move up a hierarchy of these imaginal situations from the least fear provoking (e.g., looking at a spider across the room) to the most fear provoking (holding the spider), contingent on their fear diminishing, and, as the theory went, the anxiety would be “reciprocally inhibited.” In Wolpe’s view, this process occurred at a very basic peripheral neurological level, but it wasn’t long before psychologists pointed out that the physiological components of the theory were not tenable scientifically and that the process almost certainly was best described as counterconditioning. This was an important early example to me that the scientific process could, in fact, direct and guide the development of knowledge in a clinical context. In any case, following this institute, Cautela became very intrigued not only with the application of principles of learning to the clinic but also with the process of utilizing images in therapy. With his background in both operant and classical conditioning (which comprised at the time most of what we knew about learning), he began incorporating behavioral principals into imagination, utilizing terms such as “covert” reinforcement (imagine engaging in a desired behavior and being rewarded, etc.) and “covert” sensitization (for someone with alcohol abuse, imagine engaging in a pleasurable but

4  Introduction

undesirable behavior such as reaching for a bottle and suddenly beginning to feel very nauseous, with the nausea intensifying as the bottle comes closer (e.g., Miller & Barlow, 1973). He would discuss his cases during class, as well as during casual meetings outside of class, and it would seem, anecdotally of course, that he was achieving some considerable success with his patients. During informal conversations and private meetings, Cautela would actually demonstrate some of these procedures to me. It is hard to appreciate how extraordinarily radical this was at the time, since the ubiquitous prevailing approach was psychoanalytic, and deviations from this approach were viewed as heretical and even unethical. In 1972, Cautela went on to become one of the early presidents of the Association for the Advancement of Behavior Therapy (AABT), now the Association for Behavioral and Cognitive Therapy (ABCT). During the summer of 1966, after finishing my master’s degree, Cautela arranged for me to spend several months with Joe Wolpe in Philadelphia, whose offices were at the Eastern Pennsylvania Psychiatric Institute. Wolpe ran a very active clinical unit, and since he had something new and different to offer, many patients were referred who had already failed to benefit from long-term psychodynamic psychotherapy or the medications available at that time. As a graduate student, this was a very exciting time since, under Wolpe’s supervision, we would utilize many behavioral procedures, including systematic desensitization, assertive training (which was a specialized form of helping people function better in demanding interpersonal situations), as well as deep muscle relaxation techniques, then one of the bedrock strategies in behavior therapy used in SD and other applications. Visitors from around the world came to see what this new behavior therapy was all about, and we would engage in stimulating and interesting discussions of these sometimes difficult cases. Once a week I would accompany Wolpe over to the main Department of Psychiatry at Temple where he would treat a patient using behavioral techniques such as SD in an ongoing seminar format in front of 15 or 20 of assorted psychiatric residents and faculty. Wolpe simply sat in the middle of a room with the patient with everyone seated in a circle around him. As the weeks went by, the patient would often improve dramatically, which would be surprising and even shocking to the residents and faculty who were deeply steeped in psychoanalytic thought and who would then spend the rest of the session trying to interpret what Wolpe was doing in psychoanalytic terms. At one point after a particularly skillful session, a resident approached me saying, “Isn’t it just the case that Dr. Wolpe is wonderful at forming alliances and that this patient’s progress is simply due to the effects of the alliance and resulting transference?” Already a “true believer,” I found myself replying no, that any success could not be due to Wolpe’s skill as a clinician because we often interchanged therapists with the same patient during the course of treatment (which was true) and that it really was the new behavior therapy techniques that were important! In later years I would recount this anecdote to Wolpe and how I found myself in the position of renouncing to some extent his therapeutic skills (feeling something like a Judas) because, in fact, Wolpe was a wonderful, warm, and supportive therapist.

Introduction  5

The Vermont Years

After an exhilarating summer studying and working with Wolpe in 1966 and newly married to Beverly Colby, I began my doctoral studies at the University of Vermont. At that point in time, it was very clear to me that I wanted to pursue a more scientific approach to clinical training and that the nascent field of behavior therapy seemed to provide the clearest path. The University of Vermont seemed an ideal setting in many ways because of the presence of a young assistant professor, Harold Leitenberg, who himself had only begun his career several years earlier, as well as a young associate professor of psychiatry, Stewart Agras, who had recently teamed up with Harold to pursue clinical research. Harold Leitenberg trained at Indiana University with James Dinsmoor, who focused on operant conditioning in the animal laboratories very much in the tradition of B. F. Skinner. But Leitenberg at the time had little or no clinical experience. Stewart Agras, on the other hand, who was trained in London with a residency at McGill University in Montreal, brought a British empirical approach to psychiatry with an emphasis on careful observation and measurement and diagnostic precision that was very unusual in those days in North America. What I brought to the table was direct training and experience in the practice of behavior therapy, such as it was in 1966, after training with Cautela and Wolpe. Working within an operant conditioning paradigm and under Leitenberg’s guidance, we began pursuing what at the time was an innovative to approach to clinical research. This approach emphasized the repeated measurement and functional analysis in individual participants that came to be called “single case experimental designs.” Our particular contribution as a new research team was translating these designs, as utilized in the operant animal laboratories, to the clinic, resulting in a series of early research studies such as the one reprinted as Article 3 in this volume (Agras, Leitenberg, & Barlow, 1968). These studies, in turn led to a paper describing this methodology that was published in the Archives of General Psychiatry (Barlow & Hersen, 1973) and that ultimately became the first edition of a book on the same topic (Hersen & Barlow, 1976). What made this research possible was the collaboration among Leitenberg and Agras (and myself), since Agras, from his position in the Psychiatry Department, had a direct access to clinical populations and clinical facilities that was very rare those days in psychology departments. Since we were pursuing a systematic program of clinical research, we managed to secure several inpatient beds in a dedicated federally funded clinical research unit that was open to the whole medical school but was relatively underutilized by other departments such as medicine and surgery. Thus, for example, the severely house-bound patients described in Article 3 (Agras, Leitenberg, & Barlow, 1968) spent several weeks in the hospital participating in very intensive daily treatment, along with careful frequent measurement of their progress and the beginnings of some experimental analysis of treatment components. To accomplish these goals, our research team, by then consisting of several additional young psychiatrists as well as additional doctoral students in psychology, would meet daily for two hours going over new data as they came in. It was during the in-depth discussions in these meetings that the translation of

6  Introduction

basic experimental methodologies to the clinic were further developed. During this period, research focus was not limited to what would now be called “anxiety disorders,” such as agoraphobia (we had not yet come to recognize panic attacks or panic disorder at that point in time), but rather extended to any type of very severe psychopathology, mostly treatment-resistant patients who had not benefited from other treatment approaches and who were looking for something that might be effective. Thus, in these years we published papers on tightly controlled analyses of treatments for conversion disorder, claustrophobia (Agras, Leitenberg, Barlow, & Thomson, 1969), anorexia nervosa (Agras, Barlow, Chapin, Abel, & Leitenberg, 1974), and other presentations, in addition to agoraphobia. It was also the case in that period that diagnosis was an imprecise and very much undervalued activity in psychiatry and clinical psychology. The second edition of the Diagnostic and Statistical Manual (DSM-II) or the seventh edition of the International Classification of Diseases (ICD-7), the contemporary nosological systems, offered only very loose global definitions of disorders focused on presenting symptoms such as paranoid delusions or agoraphobia. The prevailing catch-all category for most nonpsychotic disorders, particularly disorders presenting with strong negative affect, was neurosis, which presumed a specific psychoanalytically based etiological process with the resulting symptoms considered only superficial and relatively trivial manifestations of that process. Because of this, patients were typically identified by sometimes unreliable descriptions of their most prominent presenting symptoms or personality features, and formal diagnostic categories were widely ignored. Although our intensive idiographic analyses ranged across much of psychopathology, for my dissertation I returned to my focus on anxiety by examining some of the mechanisms of fear reduction in procedures we were using at that time. Many of us in those years found a convenient analog of phobia in the very common presentation of fear of snakes among college women. Recruiting participants for the studies was very easy, and so many of us did it that it came to be said that the great snake phobic epidemic of the 1960s was all but eliminated by energetic doctoral students attempting to finish their dissertations! In my case, I focused on the slippage in fear reduction during the process of desensitization that occurred when the image of the snake was presented in imagination versus in reality despite the fact that participants worked up the same fear hierarchy at approximately the same speed (Barlow, Leitenberg, Agras, & Wincze, 1969). Another study focused on the effects of actively shaping approach behavior to the snake as opposed to passively moving the snake closer to the participant who sat in a chair watching (Barlow, Agras, Leitenberg, & Wincze, 1970). But it was also during this time that I undertook what has come to be, from my own personal point of view, the most regrettable initiative in my clinical research career. Specifically, with my expertise acquired from Joe Cautela in the administration of covert sensitization, I began treating and evaluating the effects of treatment in individuals with what came to be called “paraphilias” but what was then called “sexual deviation” (Barlow, 1974a). While our focus was mostly on pedophilia (e.g., Barlow, Leitenberg, & Agras, 1969), the

Introduction  7

aggressive behavior of rapists (e.g., Abel, Barlow, Blanchard, & Guild, 1977), and other paraphilias (e.g., Hayes, Brownell, & Barlow, 1978), included in this series of studies were participants presenting with same-sex arousal patterns with consenting adults (e.g., Barlow, Leitenberg, & Agras, 1969). At that time, homosexuality was considered a disorder in all systems of nosology, and, under extreme pressures from society and the associated stigma, these individuals sought out treatment; so very few clinicians even gave the assumption a second thought. But by the mid-1970s, several individuals began questioning these treatment goals. One of the first mental health clinicians to do so, in a shining example of groundbreaking ethical thinking, was the then president of AABT, Jerry Davison, in his presidential address in 1974. Later in the decade, the American Psychiatric Association “declassified” homosexuality as a disorder, and these events began what most observers of culture regard as the most rapid shift in cultural attitudes and behavior ever to occur. Looking back on that period from today’s vantage point, it is very hard to even conceive how we could not have realized the inherent conflicts in attempting to treat harmless consenting adult behavior involving love and affection. But the lesson learned by most of us is that the definitions and classification of psychopathology do not represent qualitatively different entities but rather are embedded in the continually shifting landscape of cultural values and mores and that these ethical and moral issues must be transparent and debated, and they must occupy a central role in all of our endeavors (Barlow & Durand, 2014). After three years at the University of Vermont, in 1969, I had my PhD in hand. It was at this time that Stewart Agras had decided it was the time in his career to seek a position as chair of a psychiatry department. He invited me to join him in the search for the appropriate place with a view towards continuing our collaboration. This search was a much more difficult task than might be imagined. Although Agras had clearly been very productive in clinical research and, with his CV, would today be highly valued as a chair at any psychiatry department in the country, in those days, in psychiatry departments and clinical psychology training programs, empirical research was at best a secondary activity and at worst discouraged. Furthermore, actually influencing a mature, established department in any meaningful way—deeply immersed, as they were, in intensive (psychoanalytic) training and clinical work—would have been very difficult if not impossible. But Stewart, in the process of his search, visited a small department in Jackson, Mississippi, that had few established programs and therefore was in a position to hire a relatively large number of people in a short period of time. Surprised though I was as a New Englander at the location, a visit persuaded me of the excellent opportunity that existed. Thus, in the fall of 1969, we both settled a very long way from New England in the Department of Psychiatry and Human Behavior at the University of Mississippi Medical Center. The Mississippi Years

From lowly graduate student in the previous months, I was suddenly Chief of Psychology in the Department of Psychiatry and Human Behavior and Director of the Psychology Internship Program (now called the Psychology

8  Introduction

Residency Program). There were very few psychologists on site, and psychology interns had yet to be admitted. In addition, when the Chief of Psychology at the adjacent Veterans Administration hospital, closely affiliated with the medical center, moved on several months after I arrived, I found myself as acting Chief of Psychology at the VA as well. With this surfeit of titles but little or no substance to the programs under my direction, my first task was to continue to organize our clinical research efforts along with Stewart Agras who, although devoting a fair amount of time to administration, retained his active role in research. My second task was to initiate a clinical psychology internship program, and my third was to begin to fill many of the open positions with like-minded psychologists with a strong research and scientific approach to clinical work, something of a rarity in those days. Fortunately in those very early years of behavior therapy and behavior modification, several outstanding opportunities presented themselves. First, Ed Blanchard, then at the University of Georgia after having completed his PhD at Stanford with Al Bandura, showed some interest in joining us to help build new programs. We were fortunate to successfully recruit him with his strong expertise in psychophysiological methods and in the emerging field of biofeedback. Second, after consultation with the central office of the VA, we identified Michel Hersen who at the time was looking to relocate from a state hospital in Connecticut. With Michel’s experience and a recommendation from VA headquarters, he was quickly appointed as Chief of Psychology at the VA. Michel also became Associate Director of the Psychology Internship Program, since the plan was to integrate the internship with stipends provided by both the VA and the Medical Center. Other strong psychologists recruited over the next several years into the program included Peter Miller with expertise in addictions and Dick Eisler who had been a colleague of Michel Hersen’s during graduate school days at SUNY Buffalo. And in the fall of 1970 we welcomed the first group of three interns. In addition to opportunities for research, one of the very attractive features of the department under the extraordinarily forward-looking leadership of Stewart Agras was the notion that psychology was fully an equal partner with psychiatry in both clinical and research endeavors. Thus, psychologists were afforded principal responsibility for patients in both inpatient and outpatient units and were on call to the emergency room for psychiatric consultation nights and weekends. As the training program began, this opportunity was also afforded to psychology interns, who would alternate this duty with psychiatry residents (Barlow, 1974b). It was also noteworthy that the clinical psychology internship was the first in the country to orientate itself as an empirically based training program. An ad that appeared in the journal Behavior Therapy stated the philosophy very clearly: The psychiatry department of the University of Mississippi Medical Center offers an approved pre-doctoral internship in clinical psychology. The focus is on a behavioral analysis–behavior modification approach to the variety of clinical problems found in community settings, outpatient and inpatient settings, experimental preschools, and a general hospital emergency room. Diagnostic testing requirements are minimal. Applied

Introduction  9

clinical research using single case experimental designs is encouraged and taught in most clinical settings. (Vol. 4, p. 627, 1973) One of the key sentences in this advertisement was, “Diagnostic testing requirements are minimal.” Thus, the program moved very far away from the prevailing mode of training, focusing on detailed projective and related personality and cognitive testing of patients with results in the form of a psychological report forwarded to the psychiatrist, who would then incorporate the report into a treatment plan. In this way, clinical psychology programs within psychiatry departments operated very much as a laboratory similar to clinical laboratories analyzing blood samples. Of course, objective psychological testing was conducted if indicated, but the elimination of projective testing, as well as the routine psychological battery administered to every patient, was a relatively radical move at the time that prompted the predictable backlash in the community and to some extent nationally. Early in this phase of program development, I received a notice from the VA that, based on a report from a committee of regional directors of training, they wished to come and investigate developments at the VA hospital in Jackson. Of course, it turned out that politics was behind this notification since one of the members of the committee was friends with the previous Chief of Psychology who had, in fact, been “pushed out” by the Chief of Staff of the VA hospital who was not impressed with the way he conducted his small department. The Chief of Staff cleverly used our arrival to ask whether we were going to retain this individual, and, after a few brief conversations, we reported back that he probably would not fit in with our developing plans. The Chief of Staff was obviously delighted and moved him out at the first possible moment, but the blame fell on us, on our new behavior modification approach, and on the customary implication in those days that what we were doing was radically different and probably unethical. A visit from Central Office of the VA quickly dispelled any misconceptions, and the recommendation was to hire Hersen as the new Chief of Psychology. Fortunately there was enough support in the psychological community that, with the strong backing of the chair Stewart Agras, the training program not only survived this and other early attacks but prevailed and flourished. With strong and effective leadership over the ensuing years, it is gratifying to note that the core concepts that we established back in 1969 have been retained in the form of a broad and deep scientist practitioner model of training, and the Psychology Residency Program at Mississippi continues to be one of the strongest research-oriented internships in the country. But our principal focus remained on clinical research. As Stewart Agras has observed, it is noteworthy that the productivity of the whole department of psychiatry in Mississippi prior to 1969 averaged one published paper per year, whereas four years later, close to 100 papers per year were published (Agras, 2012). As in internship training, the focus was on behavioral analytic single case approaches to clinical problems, although not to the exclusion of other more nomothetic approaches. While some of this idiographic research focused on anxiety disorders, including phobias and obsessive-compulsive disorders, the emphasis continued to be on idiographic approaches rather than on any

10  Introduction

one content area. For example, selected publications from this era included work on severe speech disorders (Pineda, Barlow, & Turner, 1971), spasmodic torticollis (Bernhardt, Hersen, & Barlow, 1972), and the measurement and modification of incestuous behavior (Harbert, Barlow, Hersen, & Austin, 1974). Research meetings continued every single day, and in 1971 I was fortunate to obtain my first National Institute of Mental Health (NIMH) grant to study the psychological aspects of sexual dysfunctions and paraphilia. In 1974, I was promoted to full professor of psychiatry, but some of the very radical innovations in the department came under attack. A political fight over the control of a local community mental health center, waged between the Medical Center and a community hospital, provided an opportunity for some in the mental health community to rail against the evils of behavior modification, as propagated by Stewart Agras and the faculty. During this period of somewhat nasty politics, Stewart was offered a much more felicitous professorship at Stanford University and relocated in 1974, where he remains to this day (early 2015) as productive as ever at the age of 85! Approximately six months later and missing New England, I accepted a position as Professor of Psychiatry and Psychology at Brown University and Director of Education and Training and Psychology at Butler Hospital in Providence Rhode Island where Beverly and I, now with two children, relocated in January of 1975. The Brown Years

The new chairman of the Department of Psychiatry at Brown, Ben Feather, had been trained as a psychoanalyst, as were almost all psychiatrists in those days, but he also had received some training in behavior therapy and biofeedback and was interested in the variety of different clinical approaches and how they could be integrated for training purposes. As the founding chair of the new Department of Psychiatry at Brown, Ben was creative and bright but faced a number of administrative hurdles as he attempted to build the fledgling department. First, like many New England medical schools, Brown University did not “own” its own teaching hospital. Rather, it affiliated with a number of existing community hospitals. The freestanding private psychiatric hospital in Rhode Island, Butler Hospital, became the principal seat of the Department of Psychiatry. On the basis of an administrative arrangement worked out between the university and the various hospitals, the Medical Director of Butler Hospital would also become the Chair of Psychiatry at Brown. But other hospitals also provided some mental health services, and they had to be integrated into the larger department—no small undertaking. The Bradley Hospital was the child psychiatric hospital with its own Medical Director. Some of the other general hospitals, such as the Rhode Island Hospital, also had psychiatric components and potential slots for hiring faculty. But the Chair of Psychiatry did not actually control the budgets or salary lines in these other hospitals, with the exception of Butler Hospital, requiring that any negotiations to build up the Department of Psychiatry depended on the personal persuasiveness of the chair and an overarching sense on the part of the hospitals that being affiliated with Brown University was a good thing. Nevertheless, conflicts soon arose over expending hospital monies on training

Introduction  11

activities since this did not generate revenues for the hospitals to pay the bills. On the contrary, it cost the hospitals money, and any reimbursement from Brown University did not begin to cover the time lost. Even though Ben Feather was very interested in beginning to build a clinical research program with external support, he was more immediately interested in setting up exemplary training programs. He saw the internship program in clinical psychology at the Medical Center in Mississippi, which he had visited the previous year, as an exemplar of what could be set up at Brown, and creating this internship became my first task. Having arrived in January of 1975 and with internship acceptances due to be mailed out in early February, there was little time to be lost, and all recruiting had to be done by phone. In those days with few formal procedures in place, this was somewhat easier to do (it would not be possible today), and I was able to recruit five students mostly through contacts with directors of clinical training at universities with whom I had built a relationship while in Mississippi. Nevertheless, these five students were taking a large risk accepting admission to a program that did not yet exist. The five students who comprised the first internship class at Brown in the fall of 1975 were Kelly Brownell, formerly Professor of Psychology and Epidemiology and Public Health at Yale and currently Dean of the Sanford School of Public Policy at Duke University; Steve Hayes, Foundation Professor at the University of Nevada, Reno, and well-known for his creative accomplishments; Toy Caldwell-Colbert, formerly Provost and Chief Academic Officer at Howard University before her untimely passing; Peter Monti, who went on to become Professor of Psychiatry and Human Behavior and Medical Sciences and the Donald G. Miller Distinguished Professor of Alcohol and Addiction Studies at Brown and who ran the internship program himself for a number of years; and Carol Landau, who became Clinical Professor of Psychiatry and Human Behavior and Clinical Professor of Medicine at Brown. I was also fortunate to be able to recruit my colleague from graduate school days at Boston College and the University of Vermont, John Wincze, then at Dalhousie University, to become Chief of the Providence VA Hospital and Associate Director of the Psychology Internship. John remained at Brown for the rest of his career. His son Jeff would later become one of my doctoral students. With a heavy emphasis on training, a number of publications from that period of time focused on the integration of science and practice, still something of a new endeavor from a more practical point of view (e.g., Barlow, 1981; Barlow, Hayes, & Nelson, 1984). Research on sexuality (e.g., Abel, Barlow, Blanchard, & Guild, 1977) and anxiety (e.g., Barlow, Mavissakalian, & Schofield, 1980) continued at Brown as well, but research productivity diminished somewhat due to heavy administrative responsibilities. The Chair of Psychiatry, Ben Feather, unexpectantly resigned in the summer of 1975. Thus, for a period of four years the several of us who were department heads at the hospital were responsible for both the administration of the hospital while another group of us ran the Department of Psychiatry. But it was the hospital, with its large budget and the myriad of issues attendant with its mission that consumed most of my administrative time and effort. This left less time for research than was desirable from my point of view, and in 1979, with the internship program firmly established but the politics in the department

12  Introduction

continuing to be unsettled, it seemed time to relocate to a setting that allowed a more complete focus on clinical research. This decision was solidified by the arrival of a new Chair of Psychiatry from Canada, about whom those of us on the search committee knew less than we thought we did. He proved to be a difficult, autocratic chair who, in addition, was not supportive of psychology. Within the year, all of the full professors and chiefs of psychiatry recruited by Ben Feather at the various hospitals had left, and it was time for me to move on also. Thus, in the fall of 1979 I accepted a position as Professor of Psychology at the State University of New York (SUNY) at Albany, joining my old friend from Mississippi days, Ed Blanchard, who had been recruited to rejuvenate the Clinical Psychology Program at that university two years previously. The Albany Years

Shortly after arriving in Albany, and in a department of psychology rather than a department of psychiatry and associated hospital for the first time, it seemed necessary to create a venue to carry on clinically meaningful research since there was no ready supply of patients. While some psychology departments in those years had established training clinics, usually called “psychological service centers,” as had Albany, the setup of these training clinics along with their mission made the conduct of clinical research difficult if not impossible beyond simply collecting a few questionnaires and the like. Indeed much of the ongoing research in clinical psychology programs in the 1970s could be characterized as “analog” research, focusing on personality traits or fears mostly in college students. Research from my own dissertation described earlier, focusing as it did on modifying relatively normal fears of snakes in college sophomores, remained the prototype for clinical research approaches in psychology departments despite the fact that findings from these research paradigms produced few results that were generalized to more severe psychopathology. Thus, Ed Blanchard and I together initiated the Center for Stress and Anxiety Disorders, which grew to become a large federally funded research clinic. This Center was, in turn, organized around the Phobia and Anxiety Disorders Clinic housing my research and the Stress Disorders Clinic for Ed’s research on biofeedback for such conditions as hypertension and migraine headaches. We shared administrative space and a reception area, and over the years flexibly allocated space based on grant portfolios. Here my knowledge of organizing and supervising large outpatient clinics in departments of psychiatry and hospitals came in handy, and we purposely set up the Phobia and Anxiety Clinic to serve the general public rather than college students. We also marketed the clinic as offering specialized effective brief psychological treatments for people with anxiety and related disorders, and patients paid for these services. This communicated the fact that the clinic was squarely in the realm of mental health care delivery with a focus on accurate diagnosis and effective treatment (in other words getting people better) rather than a psychology lab where patients paid nothing in return for being guinea pigs in a research project. We also saw that the waiting area and offices were comfortable and well appointed and that staff and receptionists dressed and acted professionally. The functioning of this clinic, as it evolved, is described in Barlow (1992,

Introduction  13

Article 8). Treatment was provided by me and other young faculty members collaborating on setting up clinical research programs, such as Rick Heimberg who began a program for social phobia and Gerry O’Brien working in the area of agoraphobia and the newly created disorder (in DSM-III) of panic disorder. Services were also delivered by doctoral students working on the team who were supported by one or more NIH grants focused on developing psychological treatments for anxiety disorders. Despite marketing ourselves as a fully functioning clinic serving the public in order to attract sufficient numbers of patients, we did not lose sight of our principal mission as a center for clinical research. With this structure in place and taking advantage of every media invitation to make the public aware of our programs, the clinic developed a reputation of providing effective clinical services in a professional manner and positive word of mouth assured steady growth in the number of referrals. This growth necessitated several moves in those first few years to accommodate expansion, and by 1983, we finally settled in our own building leased by the university on a busy road just off campus. That year we recruited Wendy Silverman to the faculty, and she organized a Child and Adolescent program. When Wendy left for Florida in 1990, a new post-doc, Anne Marie Albano took over. In 1987, Bonnie Brown, who had been with us in more junior positions, became administrator of the anxiety clinic and, when she finished her nursing degree, became Nurse-Administrator. Our research focus on anxiety in this center was effectively launched with the awarding of an NIMH contract for a conference on behavioral approaches to anxiety disorders that was held in Albany on April 28–30, 1980. This conference, attended by 20 of the leading clinical researchers in the world at that time, noted that psychologically based treatments, particularly exposure therapies for phobia, were proving to be an important advance in effective treatment, which had been dominated up to that time by pharmacological treatments. But there were sizeable disagreements among clinical investigators on the appropriate directions for research in this area. A report of the consensus conclusions of this conference, published in 1981 (Barlow & Wolfe, 1981, Article 4), recommended general research strategies, new approaches to assessment and classification, ideas for fruitful process and outcome research, as well as the beginnings of dissemination strategies. Among the interesting recommendations from the conference participants at that time was that there was no further need for outcome studies utilizing exposure therapies for phobias since these procedures were well established. However, important outcome research should be initiated in the newly conceptualized areas of generalized anxiety disorder, social phobia, and panic disorder. Furthermore, all funded research should be carried out with clinical populations “whether the goal was to uncover basic mechanisms of action of psychological treatment or to pursue future outcome goals” (p. 449). Thus the committee suggested that analogue research had probably run its course and that priority should be given to studies in the context of the more commonly encountered clinical disorders. Recommendations were also made to study the generalization and maintenance of changes as a function of psychological treatments in order to study the epidemiological and natural histories of the anxiety disorders; further, not surprisingly, it was recommended that a greater use of single case experimental designs could prove fruitful in

14  Introduction

teasing out mechanisms. It is interesting to contrast knowledge of the anxiety disorders as it existed in 1980 and as represented by the conference recommendations with what we have learned in the past 35 years. In 1981, Peter DiNardo, who was on the faculty of nearby SUNY Oneonta, joined us for a year-long sabbatical, and together we collaborated on writing a much needed, semistructured diagnostic interview focusing exclusively on anxiety and related emotional disorders. Modeled after the Schedule for Affective Disorders and Schizophrenia (SADS), this instrument became the Anxiety Disorders Interview Schedule (ADIS), which proved to be an extremely useful instrument and the focus of our new project on classifying anxiety disorders. With a venue in place and the focus squarely on anxiety and related disorders, the major themes of my research became more fully elaborated. Early research on the nosology of anxiety and mood disorders resulted in explorations of the reliability of these new DSM-III categories, as well as new conceptualizations of generalized anxiety disorder and panic disorder (e.g., Barlow, 1985; Barlow, 1987; Barlow, Blanchard, Vermilyea, Vermilyea, & Di Nardo, 1986). These developments were communicated through my membership on the anxiety disorders work group for DSM-III-R and, later, on the Task Force for DSM-IV. Our long-running NIMH grant on the classification of anxiety disorders, first funded in 1984 and taken over by Tim Brown as Principal Investigator in 2000, continues to this day. At the same time, experimental psychopathology research focused on the nature of anxiety as evidenced in men and, later, in women presenting with sexual dysfunction. First funded by NIMH in 1979, the major paradigms used in this grant afforded an easily quantifiable output of the influence of the cognitive and affective components of anxiety in the form of psychophysiological measures of sexual arousal, which could be manipulated, sometimes without patients’ awareness. This program of research resulted in a model of the process of anxiety manifested in sexual dysfunction (Barlow, 1986a, Article 12). In 1985, Michelle Craske arrived to spend a postdoctoral year after finishing up her PhD at the University of British Columbia with Jack Rachman. During this period, my colleagues and I, particularly Michelle and Ron Rapee, also spending a postdoctoral year after getting his degree at the University New South Wales in Sydney, developed new treatments for anxiety and related disorders, most notably a new psychological approach to treating panic disorder that was positively evaluated and widely accepted (Barlow, 1986b, Article 5; Barlow, Gorman, Shear, & Woods, 2000; Klosko, Barlow, Tassinari, & Cerny, 1990; Barlow & Cerny, 1988). We also extended treatment development efforts to generalized anxiety (e.g., Rapee & Barlow, 1991). Fortunately for me, Michelle and Ron ended up staying approximately five years, contributing substantially to a very productive period of clinical research. Programmatic research on the nature, classification, and treatment of anxiety and its disorders during the 1980s resulted in a book integrating basic and applied research on anxiety from a variety of different perspectives (Barlow, 1988). The several years it took to write this book proved to be a crucial step in my research inasmuch as it deepened and broadened my knowledge of anxiety, and I became acquainted with the literature, such as it was, in emotion science. This, in turn, led to the first statement of a transdiagnostic approach

Introduction  15

to treatment in the emotional disorders based on effective principles of change as discussed later (see p. 19). By 1990, evidence-based psychological treatments had been evaluated positively from emerging clinical trials such that sufficient evidence for efficacy was available for at least some disorders, and around this time we turned our attention to dissemination and implementation. The first step was made possible by the necessity in well conducted clinical trials of writing treatment manuals describing in some detail the administration of treatments undergoing evaluation. This allowed multiple therapists to administer these treatments in a reliable manner, thus ensuring the existence of an independent variable. It became apparent to us that these manuals, which had been restricted to internal use among research teams until that point, would be valuable to clinicians more generally wishing to incorporate these treatments into their practice. Thus, in the late 1980s, Michelle Craske and I wrote up our treatment for panic disorder and self-published it through my own publishing company set up for this purpose, Graywind Publications, Inc. (Barlow & Craske, 1988). When it quickly became apparent that there was considerable demand for this product, we published additional materials such as a therapist’s guide (Craske & Barlow, 1990), as well as programs for treating generalized anxiety disorder (Craske, Barlow, & O’Leary, 1992; Zinbarg, Craske, & Barlow, 1993) and began publishing manuals used in other clinical trials by other investigators for a variety of other disorders as well. This was a new concept at the time that did not fit comfortably with traditional publishing companies since we were making available an integrated treatment program to address specific disorders, and it seemed that a new and different marketing and dissemination strategy was needed. Although this dissemination strategy was effective to a point, it became clear that broader efforts were needed to make these treatments better known and available in order to provide some semblance of choice to people suffering from these disorders, most of whom were restricted to available medications from their providers. In 1993, when I was elected President of the Division of Clinical Psychology (now the Society of Clinical Psychology) of the American Psychological Association, I organized a task force (one of the prerogatives of the president each year) and titled it “Promotion and Dissemination of Psychological Interventions.” I was fortunate to have my first choice for chair of this task force, Diane Chambliss, accept this responsibility, and, working together, we recruited diverse members from research and practice settings. In order to implement promotion and dissemination initiatives, it became necessary to first survey the state of the evidence for psychological treatments in order to ascertain which interventions could be judged to be empirically supported or evidence based, but no a priori criteria existed for making these determinations. The Task Force settled on a criterion consisting of a minimum number of studies showing efficacy, which was admittedly arbitrary but formed a starting point for what proved to be a very important debate on this topic. This was followed by the listing of psychological interventions that met this new criterion, and so began the active process of communicating the empirical support of psychological treatments to psychologists, other mental health professionals, and the public at large. These ideas merged with the growing

16  Introduction

mandate emanating from medicine for evidence-based practice, and it became clear that this was an idea whose time had come (Barlow, 1996, 2004). Boston: The Later Years

In 1996, after 17 productive and fulfilling years in Albany, I was presented with an exciting opportunity to rebuild a clinical psychology program in my hometown, and Beverly and I, after an absence of 30 years, returned to Boston, where I became Professor of Psychology and Psychiatry, Director of Clinical Training, and Director of the Center for Anxiety and Related Disorders (CARD) at Boston University. CARD was actually the Phobia and Anxiety Disorders Clinic of the Center for Stress and Anxiety Disorders at Albany, which, by that time, had grown to approximately 30 people including staff and students, almost all of whom made the trek down the Mass Pike from Albany to Boston. Notable among them were Tim Brown and Stefan Hofmann, who had taken over the Social Phobia Program from Rick Heimberg, David Spiegel, the clinic psychiatrist, and, of course, Bonnie Brown, our Nurse-Administrator. Construction of our new quarters in Kenmore Square, very close to Fenway Park, proceeded quickly, and we were open for business in the fall of 1996. The design of the new clinic was greatly assisted by David Spiegel, who, prior to attending medical school, had completed a degree in engineering, skills he put to good use in designing and adapting space to suit our needs. Shortly after arriving, Donna Pincus joined us at CARD and took over the Child and Adolescent program. In the years to follow, Pincus, Brown, and Hofmann would all cross over from soft research, funded positions to tenured professorships at BU. CARD was principally a clinical research operation supported by five different NIMH grants in 1996. The complexity of the transfer prompted the Provost to equate efforts to recruit us with the recruitment of a large physics lab. But once again, as had happened at Brown University, one of the principal objectives of my recruitment by the administration at BU, along with research initiatives, was to revitalize the clinical psychology training program. This program, one of the original 12 programs approved by the American Psychological Association (APA) in 1948, retained many existing elements from those early years, including a very heavy focus on practice and theory. As with most of the early programs, research was not emphasized, and research methods were introduced only late in the training sequence. There were no existing externally funded research grants in the clinical program, and the university made it clear that, unless the program became more scholarly and research based, they would consider closing it down. Initially, we focused largely on revamping the curriculum, hiring or promoting more research-oriented clinical faculty, and making research training an integral part of the clinical program experience beginning in the first year. Very ably assisted by a new graduate of the clinical program, Lynn Bufka, who was intimately familiar with the inner workings of the training program, the department, and the university, as well as with the cooperation of the faculty who were, for the most part, on board with the necessity of change if not necessarily the specifics of what needed to change, the transition went relatively smoothly. During this transition, Lynn performed superbly as Associate Director of Clinical Training

Introduction  17

with her strong interpersonal skills and ability to relate to both faculty, with whom she was popular and respected, and students, who were, after all, just a few years younger than she. Lynn was able to smooth out many of the inevitable bumps in the road during this rebuilding process. This allowed a more complete focus on clinical research at CARD then I had anticipated. By 2004, after eight years of running the clinical program, the transitions were complete, and I stepped down. As I write this introduction, it has now been over 18 years since I arrived in Boston, and my program of research and writing has continued to focus on the nature and treatment of anxiety and related disorders of emotion, as well as policy issues involved in dissemination and implementation. In 2004, recognizing the plethora of treatment manuals that had been developed for anxiety and related emotional disorders, each targeting a very narrow slice of psychopathology represented by DSM-IV categories, we returned to the approach first articulated in 1988, identifying a common set of principles of change applicable to all disorders. We referred to this approach as a Unified Transdiagnostic Treatment for Emotional Disorders in the first article on this topic (Barlow, Allen, & Choate, 2004, Article 10). Moving more deeply into this area, I began to realize, based on our ongoing research with my colleague Tim Brown on the classification and nature of emotional disorders, that fundamental temperamental aspects underlying anxiety, mood, and related disorders seemed more central to the nature of these disorders than did the symptom presentations that were their defining feature in DSM-IV and DSM-5. In 2009, Tim Brown and I proposed the conceptual outlines of a new hybrid dimensional-categorical classification system based on these shared features and began to spell out implications for assessment and treatment (Brown & Barlow, 2009, Article 16). This led, in turn, to a greater focus on the underlying temperament of neuroticism, along with related temperaments such as positive affect, and a modification of existing models of etiology to encompass the nature and development of neuroticism itself, with implications for diagnosis and treatment (Barlow, Sauer-Zavala, Carl, Bullis, & Ellard, 2014, Article 17). At the same time, we continued to focus on dissemination and implementation with policy-based articles outlining the status of some of these fledgling efforts (McHugh & Barlow, 2010 Article 19) and a more general update on the status of evidence-based psychological treatments with some predictions of necessary future steps (Barlow, Bullis, Comer, & Ametaj, 2013, Article 21). By 2009, CARD had morphed into something much more than my laboratory, which it had been during the Albany years and the early years in Boston, to a university-wide resource that was used for training for not only doctoral-level psychology students at BU but also psychiatric residents, social work students, as well as students from other clinical psychology programs in and around the city of Boston. Research projects and programs have multiplied, and CARD is currently staffed with over 70 individuals including a number of principal investigators pursuing different aspects of adult and child emotional disorders. In a typical year, over 500 patients are admitted to the clinic (after a telephone screen for appropriateness), with about half of them

18  Introduction

triaged into ongoing clinical research projects. The other half, including those who refuse participation in research for one reason or another, are assigned to staff and students for treatment. In 2009, I gave up all remaining administrative roles in the department and clinic, and a full-time director was appointed to administer CARD. In view of the complexities of running this large research and teaching unit, the decision was made to appoint someone on a nontenure track who was not also pursuing research. Dr. Lisa Smith, the former Clinical Director of CARD, was appointed, and she in turn reports to a steering committee of principal investigators, leaving me time to focus more fully on the already mentioned strands of research. With this description of the external influences and contexts that led to the various strands of my program of research now complete, it is time to return once again to the major themes outlined at the beginning of this chapter and to consider briefly the articles I have chosen to illustrate the development of these themes, beginning with methodology and clinical research. Methodology and Clinical Research

The first article in this section (Barlow, 1980) represents my presidential address to AABT (now (ABCT). The address was actually delivered in November of 1979, at a time when behavioral and cognitive interventions were beginning to be seen in a more favorable light. This followed a decade ranging through the mid- to late 1970s when behavior modification or behavior therapy received very bad press and was often described as nothing more than unethical brainwashing, as famously represented in the popular book and movie of the time, Clockwork Orange. This poor image was best represented by a survey of articles in the New York Times during the mid-1970s demonstrating that approximately 50% of the articles addressing the subject equated behavior modification with such procedures as brainwashing, psychosurgery, sensory deprivation, and Chinese water torture! Nevertheless, by 1980 both the NIMH and the American Psychiatric Association, along with Science magazine and the President’s Commission on Mental Health, strongly endorsed further research into behavioral interventions, noting the preliminary evidence for their efficacy. But what I noted in that address was that our science was becoming shallow. Basically this was due to an overemphasis on broad nomothetic comparisons of a treatment to no treatment with results analyzed in terms of statistical rather than clinical significance: “[I]nstead of asking why does a treatment work and what are the ingredients of a given technique that are truly effective, the question is often how much of a chance is there that this package treatment as it currently exists might work with some broadly defined problem. And any probability at all is usually good enough” (pp. 319). In this report, and following the tenets of applied behavioral analysis, I advocated for “. . . a fine grained analysis of an individual’s behavior or a series of individuals’ behavior, with attention to technique building, repeated measurement, and social rather than statistical significance; an approach to science that in its very nature attends to our failures” (p. 322). In this article, I go on to hazard some predictions about how this focus on a more idiographic analysis might come about.

Introduction  19

It is interesting to fast-forward 30 years to the second paper in this section (Barlow & Nock, 2009). The same plea is made, but it is noted that more idiographic single case experimental analyses had not been as widely adopted as I predicted they might be in 1979. The principal reason for this (in retrospect) seems to be the beginning of the funding of large randomized clinical trials (RCTs) by the NIMH. These trials very clearly became the gold standard for establishing the efficacy of treatments, and the outcomes deeply influenced health care policies and practices. Without questioning the enormous contribution of this methodology, it is interesting to note that the NIMH in its recent policy pronouncements has proposed abandoning large clinical trials that simply compare treatment A to treatment B. Rather, any new trials must undertake a deeper analysis of active mechanisms in these treatments, along with a focus on discerning appropriate targets of any intervention integrating biological and psychological factors that might reflect important causal mechanisms in psychopathology. The NIH, with its recent launch of the Science of Behavior Change (SOBC) initiative, seems to be pursing the very same objectives noted as desirable in these articles. In Barlow and Nock (2009), we hazard a suggestion as to how this technology can be better integrated into current clinical science initiatives. Treatment of Anxiety and Related Disorders

Articles in this section span most of my career and trace an interesting story in the development of our conceptions of treatment, even predating exposure as we know it today. The first article in this section, published in 1968 in the Archives of General Psychiatry but reflecting treatments that occurred a year or two earlier, seems naïve in its conception, as one might expect from an effort over 40 years old. As a doctoral student at the time, I was the therapist for these patients, which meant that I saw them daily for a number of weeks during these intensive interventions. These patients suffered from very severe agoraphobia, such that they could not leave their homes and had to be transported to the clinical research unit in the hospital via ambulance while heavily sedated. Treatment consisted of setting up a walking course of about a mile in length that proceeded into an ever more crowded area in Burlington, Vermont. Patients would venture out as far as they could away from the hospital and would receive enthusiastic praise (from me) for successfully meeting behavioral targets for the day. The choice of selective positive reinforcement was based on strong results from the operant laboratories, as a well established procedures for effecting behavior change. At that time of exposure, procedures themselves were not considered central to any therapeutic approach, including the new behavioral approaches. Wolpe had been using SD, particularly for specific phobia, in which patients would be gradually exposed in imagination to feared situations along a hierarchy while deeply relaxed, based on the idea that relaxation would inhibit anxiety as long as the anxiety was not too intense. So the principal goal was to inhibit or decrease the anxiety. This also reflected a theme present in both psychoanalytic and behavioral theorizing of the day that experiencing intense anxiety could produce a very dangerous outcome up to and including a psychotic break. In psychoanalytic theory, the purpose of defense mechanisms was

20  Introduction

to prevent intense anxiety from happening. In behavioral theorizing based on Pavlov, if anxiety was too intense, the organism could enter a state of paralysis described by Pavlov as “transmarginal inhibition” (Barlow, 1988). So all anxiety reduction procedures of the day were carried out very gradually indeed. That our interventions were successful even with very severe patients was something of a surprise to us but did engender a number of questions on the active mechanisms of change. It was not long before we discovered that the exposure procedures themselves, particularly in vivo exposure as was happening in this study, were the principal and most powerful mechanism of change and that reinforcement simply motivated this new behavior (Barlow, 1988). Other articles in this section detail the accumulation of knowledge over the decades as we developed ever more successful treatments. The aforementioned NIH conference that occurred at SUNY, Albany, in 1980 summarized the state of the art at that time (Article 4). By the mid-1980s, we had conceptualized a new approach to treating panic attacks based on the notion that these attacks represented the conditioning of anxiety to internal somatic (interoceptive) cues and that the required treatment would involve the “twist” of directly exposing patients suffering from panic attacks to these interoceptive cues in a procedure that came to be called “interoceptive exposure” (Barlow, 1988). This early conceptualization is presented in Barlow (1986a, Article 5). In 1988, and as noted, while attempting to integrate everything we knew about anxiety and its disorders at the time, I wrote a chapter “The Process of Fear and Anxiety Reduction: Affective Therapy” (Barlow, 1988, pp. 285–318, Article 6). I called it “affective therapy” since I drew on the principles of emotion science as the basis for new and continuing developments in treating disorders of emotion. Particularly important strategies gleaned from this analysis included changing action tendencies associated with pathological emotions, along with interoceptive exposure and altering attributions and appraisals about the emotional experience. This was actually the first statement of what would develop 15 years later into an integrative transdiagnostic treatment for emotional disorders. In the meantime, clinical trials and had already become the gold standard for evaluating treatment efficacy, and our clinic benefited from a number of NIMH grants supporting trials for one disorder or another. One of my students, Janet Klosko, with support from what was then the Upjohn Pharmaceutical Company, conducted the first direct comparison of our new treatments for panic disorder incorporating interoceptive exposure with the most popular drug of the day Alprazolam (Xanax) a high-potency benzodiazepine (Klosko, Barlow, Tassinari, & Cerny, 1990, Article 7). The somewhat surprising result was that what we then called “panic control treatment” actually outperformed the drug, an outcome quite unexpected for many. “Surprising” because there was little evidence at that time for the efficacy of psychological interventions for what was thought by many, particularly biological psychiatrists, to be a biologically based disorder. But this finding led directly to a large multisite clinical trial finally published in 2000 in JAMA comparing panic control treatment with the tricyclic antidepressant imipramine, as well as their combination (Barlow, Gorman, Shear, & Woods, 2000, Article 9). This study broke new ground at NIMH since they had not yet had an application

Introduction  21

for a study from multiple principal investigators acting independently at four different sites (as opposed to one site subcontracting to another). This forced the NIMH to create a new grant mechanism to accommodate this collaboration, which included two sites with a pharmacological approach to anxiety and panic (Gorman & Woods) and two sites better known for a psychological approach (Barlow & Shear). This innovation, intended to handle the obvious confounding of allegiance effects at any one site, was widely praised and, given the success in obtaining funding, widely imitated in years to come. The ongoing activities in our anxiety disorders clinic in the 1990s (see the discussion in “The Albany Years”) are detailed in a brief article published in 1992 (Barlow, 1992, Article 8). Finally, as a logical extension of the thinking detailed in my 1988 book Anxiety and Its Disorders on “Affective Therapy,” we published in 2004 the first statement describing a Unified Transdiagnostic Treatment for Emotional Disorders, consisting of five core therapeutic procedures thought to be widely applicable to all anxiety, mood, and related disorders (Barlow, Allen, & Choate, 2004, Article 10). In many ways, this was a culmination of work begun back in the 1960s that reflected our deepening understanding of the nature of anxiety and its disorders and the development of effective psychological interventions targeting these underlying mechanisms. It certainly will not be the last statement since clinical science will continue to advance, but it does form the current focus of our efforts in the twilight of my career. The Nature, Diagnosis, and Etiology of Anxiety and Related Disorders

Upon my arrival at SUNY Albany in the fall of 1979, two NIMH grants were simultaneously awarded; the first focused on the treatment of agoraphobia and panic disorder with a particular emphasis on including a significant other, usually the spouse, in the treatment process (see Barlow, O’Brien, & Last, 1984). The second was focused on elucidating the nature of anxiety and arousal in sexual dysfunction (as previously mentioned and to be described further). In 1982, a third NIMH-sponsored clinical trial focused on evaluating new treatments for generalized anxiety disorder. As the principal investigator (PI) on three NIMH awards, we discussed with NIMH staff the desirability of applying for a Center grant that would support the infrastructure of the anxiety clinic, making each of the treatment outcome grants less expensive. Due to fluctuations in policies regarding Center grants at that time, this particular application was not encouraged, but working with NIMH we decided that an additional grant application, focused on the detailed assessment, diagnosis, and classification of all patients coming into the clinic would serve the same purpose. That is, all patients could then be worked up in considerable detail and triaged to the appropriate clinical trial focused on treatment outcomes, thereby saving the clinical trials the expense of this initial screening. The grant, titled “Classification of Anxiety Disorders,” was funded in April of 1984. In September of 2000, after relocating to Boston, the PI status was transferred to Tim Brown, and this grant is now in its 32nd year of continual funding. The first article in this section, “The Phenomenon of Panic” (Barlow, Vermilyea, Blanchard, Vermilyea, DiNardo, & Cerny, 1985, Article 11), emanating

22  Introduction

from early work on this project, provides detailed descriptive data on patients coming into our clinic. In this article, it was demonstrated for the first time that panic attacks were a ubiquitous process occurring across all of the anxiety disorders and depression, and therefore from a nosological point of view, it could not be restricted to what was then called “panic disorder.” We also developed in this article for the first time a scheme for categorizing panic attacks based on whether triggers or cues for the panic attack were recognized by the patient (as in a phobic situation) and whether the attack was expected or unexpected. A panic attack categorized as uncued and unexpected by the patient was meant to replace the prevailing more biologically based conception of a “spontaneous” panic attack, a term that was judged to be unscientific since nothing was truly “spontaneous” in nature. We also made clear that whether or not the attacks were cued and expected was a construction of the patient rather than a biologically based phenomenon since the different types of attacks presented all but identically in terms of symptom clusters and other characteristics. This terminology and the ubiquity of panic attacks then began to make its way into the DSM process, beginning with DSM-III-R, published in 1987. The second article, “Causes of Sexual Dysfunction: The Role of Anxiety in Cognitive Interference” (Barlow, 1986b, Article 12), ultimately provided the theoretical underpinning for a new conception of the nature of anxiety described in more detail in my book Anxiety and its Disorders, published in 1988. The gist of this article was a fundamental refutation of the then widely accepted notion from Masters and Johnson that anxious arousal was the cause of sexual dysfunction. What was demonstrated in this article was that introducing anxiety in the form of shock threat while “normal” males (and later females) without sexual dysfunction were watching erotic content not only did not diminish sexual arousal, as objectively measured in the form of changes in penile circumference, but actually enhanced sexual arousal. Indeed, what seemed to differentiate sexually functional from dysfunctional males in this paradigm was the type and focus of cognitive activity in an erotic context, as well as the extent to which this sexual arousal was perceived to be under the participants’ control. This contribution is also the last article in Volume 1 with remaining articles appearing in Volume 2. A sense of control became a major theoretical underpinning in my conceptions of anxiety (and later neuroticism itself), and the origins of perceptions of uncontrollability and anxiety were traced to early developmental experiences first in 1988 (Barlow, 1988) and then, in considerably more detail, in a Psychological Bulletin article first authored by one of my students at the time, Bruce Chorpita (Chorpita & Barlow, 1998, Article 14). In 1991, anticipating later work attempting to identify common underpinnings of all emotional disorders and building on theoretical work developed in Barlow (1988), I published an article titled “Disorders of Emotion” (Barlow, 1991, Article 13), extending our conceptions of the etiology of panic disorder to other emotional disorders such as depression, stress and anger, and mania (excitement). Ten years later, during a very productive year at the Center for Advanced Study of Behavioral Sciences, my colleagues, Mark Bouton and Sue Mineka, and I wrote a paper updating in some detail the theory of the etiology of panic disorder integrating new findings from cognitive science and neuroscience (Bouton, Mineka, & Barlow, 2001, Article 15).

Introduction  23

By that time, DSM-5 was already in the works as already noted, and Tim Brown and I began speculating on what new findings from the classification of anxiety disorders grant would portend for classification in DSM-5 and beyond. Taking a unified transdiagnostic perspective, we published in 2009 an invited paper proposing a new hybrid dimensional-categorical classification system based on the shared features of anxiety and mood disorders (Brown & Barlow, 2009, Article 16). This work continues to be the major focus of the classification grant under Tim’s direction. Finally, as also previously mentioned briefly, following the substantial honor of being awarded the James McKeen Cattell award from the Association for Psychological Science in 2012, our team fashioned a paper based on portions of my award address, entitled “The Nature, Diagnosis, and Treatment of Neuroticism: Back to the Future,” in which we proposed that earlier ideas from Barlow (1988) on the origins of anxiety required refocusing to higher-order dimensions of temperament, specifically neuroticism itself (Barlow, Sauer-Zavala, Carl, Bullis, & Ellard, 2014, Article 17). We note in that article that neuroticism may be more malleable than previously thought and would ideally be the target of direct therapeutic intervention. This constitutes the very heart of our research approach at the current time. The Ascendance of Evidence-Based Psychological Treatments

As mentioned in the section describing the Albany years, I served as President of the Society for Clinical Psychology (Division 12 of the American Psychological Association) in 1993 and created a task force on the “Promotion and Dissemination of Psychological Interventions.” This initiative grew out of efforts in the late 1980s, previously alluded to, when it had become very clear from a public policy point of view that the prevailing pharmacological approaches to mental disorders were being widely adopted and recommended in emerging health care policy statements. Most of these policies were targeting accountability under the relatively new concept of evidence-based practice, and, at the same time addressing the spiraling cost of health care. These policies often took the form of clinical practice guidelines. As already noted, medications were deemed the treatment of choice for panic disorder despite the fact that our data had already indicated that new psychological interventions were at least as effective if not more so (e.g., Klosko, Barlow, Tassinari, & Cerny, 1990; Barlow & Cerny, 1988). Since these treatments were not readily available to clinicians or the public, and there were no efforts other than the occasional workshop to make them available, we formed Graywind Publications in order to disseminate new treatments. With unbridled optimism and a firmly established illusion of control over how easy this was going to be, we incorporated our company, invested money in printing several hundred copies of our treatment for panic disorder entitled Mastery of Your Anxiety and Panic, and set up shop in the basement of our house in 1988, with my wife Beverly running the show. After obtaining several mailing lists and testing the waters, the initial investment was recouped quickly, and the company became profitable. The business expanded rapidly and was moved into commercial space, and additional employees were hired. This experience convinced us that the demand existed for these programs and

24  Introduction

that the problem was little or no infrastructure for dissemination. After five years, the business required an infusion of substantial cash for expansion, a much advanced computer system for fulfillment, and a CEO to actually run the business—all steps that neither Beverly nor I were willing to take due to other commitments and a bit of fatigue from the very long hours and inevitable set of problems involved in any rapidly growing business. So we sold the business with it ultimately ending up in the hands of Oxford University Press where the series is known as “Treatments that Work.” But these early experiences made apparent the necessity of further promoting the existence of what we then called “empirically validated treatments” and of exploring methods of dissemination and implementation In the first article in this section, published in 1996 (Barlow, 1996, Article 18), I identified emerging policies, the status of research, and the necessity of responding to developing clinical practice guidelines, such as they were at that time. By 2010, dissemination and implementation had become its own field of endeavor with appropriate funding mechanisms and emerging methodologies. Along with my student at the time, Kate McHugh, we detailed the status of those efforts offering judgments on where the field was lacking and what additional research was needed (McHugh & Barlow, 2010, Article 19). This topic was expanded into a book published in 2012 (McHugh & Barlow, 2012) and in 2013 our team once again updated the status of research on evidence based psychological treatments with suggestions for more broad based future efforts (Barlow, Bullis, Comer, & Ametaj, 2013, Article 21). To improve dissemination, one of those recommendations focused on taking advantage of the power of direct-to-consumer marketing of psychological interventions, a strategy that has proved enormously successful for the large pharmaceutical companies. Initial demonstrations of the potential of this strategy for psychological interventions were detailed in a special series (Santucci, McHugh, & Barlow, 2012, Article 20). Notably, another of my students Katlin Gallo, conducted an important dissertation on this topic only recently published (Gallo, Comer, & Barlow, 2013; Gallo, Comer, Barlow, Clarke, & Antony, 2015). Thus, the articles in this last section, representing as they do substantial advances in policy, is perhaps one of the more remarkable developments over the course of my career inasmuch as we literally began with nothing. From that humble beginning, we have now reached a point, detailed in some of the preceding publications, where governments and health care policy makers around the world, including the National Health Care System in the UK and the Veterans Health Administration in the United States, are spending billions of dollars to make evidence-based psychological treatments more readily available. Although we have a very long way to go, I believe we can all be gratified by this progress.

References Abel, G.G., Barlow, D.H., Blanchard, E.B., & Guild, D. (1977). The components of rapists’ sexual arousal. Archives of General Psychiatry, 34, 895–908. doi:10.1001/archpsyc. 1977.01770200033002 Agras, W.S. (2012). The Mississippi years (1969–1974). Behavior Modification, 36, 436–443. doi:10.1177/0145445512443979

Introduction  25 Agras, W.S., Barlow, D.H., Chapin, H.N., Abel, G.G., & Leitenberg, H. (1974). Behavior modification of anorexia nervosa. Archives of General Psychiatry, 30, 279–286. doi:10.1001/ archpsyc.1974.01760090003001 Agras, W.S., Leitenberg, H., & Barlow, D.H. (1968). Social reinforcement in the modification of agoraphobia. Archives of General Psychiatry, 19, 423–427. doi:10.1001/archpsyc. 1968. 01740100039006 Agras, W.S., Leitenberg, H., Barlow, D.H., & Thomson, L.E. (1969). Instruction and reinforcement in the modification of neurotic behavior. American Journal of Psychiatry, 125, 1435–1439. doi:10.1176/ajp.125.10.1435 Barlow, D.H. (1974a). The treatment of sexual deviation: Towards a comprehensive behavioral approach. In K.S. Calhoun, H.E. Adams, & K.M. Mitchell (Eds.), Innovative treatment methods in psychopathology (pp. 121–147). New York: John Wiley & Sons. Barlow, D.H. (1974b). Psychologists in the emergency room: Implications for training. Professional Psychology, 5, 251–256. doi:10.1037/h0037278 Barlow, D. H. (1980). Behavior therapy: The next decade. Behavior Therapy, 11(3), 315–328. doi: 10.1016/S0005-7894(80)80049-9 Barlow, D.H. (1981). On the relation of clinical research to clinical practice: Current issues, new directions. Journal of Consulting and Clinical Psychology, 49, 147–156. doi:10.1016/ S0005-7894(80)80049-9 Barlow, D.H. (1985). The dimensions of anxiety disorders. In A.H. Tuma & J.D. Maser (Eds.), Anxiety and the anxiety disorders. Hillsdale, NJ: Lawrence Erlbaum Associates. Barlow, D. H. (1986a). Causes of sexual dysfunction: The role of anxiety and cognitive interference. Journal of Consulting and Clinical Psychology, 54(2), 140–148. doi:10.103 7/0022-006X.54.2.140 Barlow, D. H. (1986b). Behavioral conception and treatment of panic. Psychopharmacology Bulletin, 22(3), 802–806. Barlow, D.H. (1987). The classification of anxiety disorders: In G.L. Tischler (Ed.), Diagnoses and classification in psychiatry: A critical appraisal of DSM-III (pp. 223–242). Cambridge University Press. Barlow, D.H. (1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: Guilford Press. Barlow, D. H. (1991). Disorders of emotion. Psychological inquiry, 2(1), 58–71. doi:10.1207/ s15327965pli0201_15 Barlow, D.H. (1992). The development of an anxiety research clinic. In D.K. Freedheim (Ed.), History of psychotherapy: A century of change. (pp. 429–431). Washington, DC: American Psychological Association. Barlow, D. H. (1996). Health care policy, psychotherapy research, and the future of psychotherapy. American Psychologist, 51(10), 1050–1058. doi:10.1037/0003-066X.51.10.1050 Barlow, D.H. (2004). Psychological treatments. American Psychologist, 59(9), 869–878. doi: 10.1037/0003-066X.59.9.869 Barlow, D.H., Agras, W.S., Leitenberg, H., & Wincze, J.P. (1970). An experimental analysis of the effectiveness of “shaping” in reducing maladaptive avoidance behavior: An analogue study. Behaviour Research and Therapy, 8, 165–173. doi:10.1016/0005-7967 (70)90086-0 Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35(2), 205–230. doi:10.1016/S0005-7894(04)80036-4 Barlow, D.H., Blanchard, E.B., Vermilyea, J.A., Vermilyea, B.B., & Di Nardo, P.A. (1986). Generalized anxiety and generalized anxiety disorder: Description and reconceptualization. The American Journal of Psychiatry, 143, 40–44. doi:10.1176/ajp.143.1.40

26  Introduction Barlow, D. H., Bullis, J. R., Comer, J. S., & Ametaj, A. A. (2013). Evidence-based psychological treatments: An update and a way forward. Annual review of clinical psychology, 9, 1–27. doi:10.1146/annurev-clinpsy-050212-185629 Barlow, D.H., & Cerny, J.A. (1988). Psychological treatment of panic. New York: Guilford Press. Barlow, D.H., & Craske, M.G. (1988). Mastery of your anxiety and panic. Albany, NY: Graywind Publications. Barlow, D. H., & Durand, V. M. (2014). Abnormal psychology: An integrative approach (7th ed.). Belmont, CA: Wadsworth, Cengage Learning. Barlow, D.H., Gorman, J.M., Shear, M.K., & Woods, S.W. (2000). Cognitive-behavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled trial. JAMA, 283, 2529–2536. doi:10.1001/jama.283.19.2529 Barlow, D.H., Hayes, S.C., & Nelson, R.O. (1984). The scientist-practitioner: Research and accountability in clinical and educational settings. New York: Pergamon Press. Barlow, D.H., & Hersen, M. (1973). Single case experimental designs: Uses in applied clinical research. Archives of General Psychiatry, 29, 319–325. doi:10.1001/ archpsyc.1973.04200030017003 Barlow, D.H., Leitenberg, H., & Agras, W.S. (1969). Experimental control of sexual deviation through manipulation of the noxious scene in covert sensitization. Journal of Abnormal Psychology, 74, 596–601. doi:10.1037/h0028087 Barlow, D.H., Leitenberg, H., Agras, W.S., & Wincze, J.P. (1969). The transfer gap in systematic desensitization: An analogue study. Behaviour Research and Therapy, 7, 191–197. doi:10.1016/0005-7967(69)90032-1 Barlow, D.H., Mavissakalian, M.R., & Schofield, L.D. (1980). Patterns of desynchrony in agoraphobia: A preliminary report. Behaviour Research and Therapy, 18, 441–448. doi: 10.1016/0005-7967(80)90009-1 Barlow, D.H., & Nock, M.K. (2009). Why can’t we be more idiographic in our research?. Perspectives on Psychological Science, 4(1), 19–21. doi:10.1111/j.1745-6924.2009.01088.x Barlow, D.H., O’Brien, G.T., & Last, C.G. (1984). Couples treatment of agoraphobia. Behavior Therapy, 15, 41–59. doi:10.1016/S0005-7894(84)80040-4 Barlow, D.H., Sauer-Zavala, S., Carl, J.R., Bullis, J.R., & Ellard, K.K. (2014). The nature, diagnosis, and treatment of neuroticism back to the future. Clinical Psychological Science, 2(3), 344–365. doi:10.1177/2167702613505532 Barlow, D.H., Vermilyea, J., Blanchard, E.B., Vermilyea, B.B., Di Nardo, P.A., & Cerny, J.A. (1985). The phenomenon of panic. Journal of Abnormal Psychology, 94(3), 320–328. doi: 10.1037/0021-843X.94.3.320 Barlow, D.H., & Wolfe, B.E. (1981). Behavioral approaches to anxiety disorders: A report on the NIMH-SUNY, Albany, research conference. Journal of Consulting and Clinical Psychology, 49(3), 448–454. doi:10.1037/0022-006X.49.3.448 Bernhardt, A.J., Hersen, M., & Barlow, D.H. (1972). Measurement and modification of spasmodic torticollis: An experimental analysis. Behavior Therapy, 3, 294–298. doi:10.1016/ S0005-7894(72)80094-7 Bouton, M.E., Mineka, S., & Barlow, D.H. (2001). A modern learning theory perspective on the etiology of panic disorder. Psychological Review, 108(1), 4–32. doi: 10.1037/0033-295X.108.1.4 Brown, T. A., & Barlow, D. H. (2009). A proposal for a dimensional classification system based on the shared features of the DSM-IV anxiety and mood disorders: Implications for assessment and treatment. Psychological assessment, 21(3), 256–271. doi:10.1037/ a0016608

Introduction  27 Chorpita, B.F., & Barlow, D.H. (1998). The development of anxiety: The role of control in the early environment. Psychological bulletin, 124(1), 3. doi:10.1037/0033-2909.124.1.3 Craske, M.G., & Barlow, D.H. (1990). Therapist guide for the mastery of your anxiety and panic program. Albany, NY: Graywind Publications. Craske, M.G., Barlow, D.H., & O’Leary, T. (1992). Mastery of your anxiety and worry: Client workbook. Albany, NY: Graywind Publications. Gallo, K.P., Comer, J.S., & Barlow, D.H. (2013). Direct-to-consumer marketing of psychological treatments for anxiety disorders. Journal of Anxiety Disorders, 27(8), 793–801. doi:10.1016/j.janxdis.2013.03.005 Gallo, K.P., Comer, J.S., Barlow, D.H., Clarke, R.N., & Antony, M.M. (2015). Direct-toconsumer marketing of psychological treatments: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 83(5), 994–998. doi: 10.1037/a0039470 Harbert, T.L., Barlow, D.H., Hersen, M., & Austin, J.B. (1974). Measurement and modification of incestuous behavior: A case study. Psychological Reports, 34, 79–86. doi:10.2466/ pr0.1974.34.1.79 Hayes, S.C., Brownell, K.D., & Barlow, D.H. (1978). The use of self-administered covert sensitization in the treatment of exhibitionism and sadism. Behavior Therapy, 9, 283–290. doi:10.1016/S0005-7894(78)80114-2 Hersen, M., & Barlow, D.H. (1976). Single case experimental designs: Strategies for studying behavior change. New York: Pergamon Press. Klosko, J.S., Barlow, D.H., Tassinari, R., & Cerny, J.A. (1990). A comparison of alprazolam and cognitive-behavior therapy in treatment of panic disorder. Journal of Consulting and Clinical Psychology, 58, 77–84. doi:10.1037/0022-006X.58.1.77 McHugh, R.K., & Barlow, D.H. (2010). The dissemination and implementation of evidence-based psychological treatments: A review of current efforts. American Psychologist, 65(2), 73–84. doi:10.1037/a0018121 McHugh, R.K., & Barlow, D.H. (Eds.). (2012). Dissemination and implementation of evidence-based psychological interventions. New York: Oxford University Press. Miller, P.M., & Barlow, D.H. (1973). Behavioral approaches to the treatment of alcoholism. Journal of Nervous and Mental Disease, 157, 10–20. doi:10.1097/00005053-197 307000-00002 Mowrer, O.H. (1950). Learning theory and the personality dynamics. New York: Arnold Press. Pineda, M.R., Barlow, D.H., & Turner, B.B. (1971). Treatment of a severe speech disorder by behavior modification: A case study. Behavior Therapy and Experimental Psychiatry, 2, 203–207. doi:10.1016/0005-7916(71)90060-7 Rapee, R.M., & Barlow, D.H. (1991). Chronic anxiety, generalized anxiety disorder, and mixed anxiety depression. New York: Guilford Press. Santucci, L.C., McHugh, R.K., & Barlow, D.H. (2012). Direct-to-consumer marketing of evidence-based psychological interventions: Introduction. Behavior Therapy, 43(2), 231–235. doi:10.1016/j.beth.2011.07.003 Zinbarg, R.E., Craske, M.G., & Barlow, D.H. (1993). Mastery of your anxiety and worry: Therapist guide. Albany, NY: Graywind Publications.

Methodology and Clinical Research

Article 1

Behavior Therapy: The Next Decade David H. Barlow State University of New York at Albany

Despite predictions of its death, and poor public relations due to a general disenchantment with science during the last decade, behavior therapy has been proven effective and has prospered. Nevertheless, we are in danger of overselling seemingly effective techniques due to a shallow approach to scientific evaluation which ignores failures and discourages a more thorough analysis. This approach may prevent the development of truly effective therapeutic approaches and will ensure a continued widening of the scientist–practitioner gap. The encouraging of practitioners to collect data as they practice, a development which will almost certainly be required by various government and third-party payers in the next decade, is suggested and examples provided. This approach should ensure close cooperation and communication among scientists and practitioners, broaden our data base, and prevent premature termination of evaluation of behavior therapy techniques after initial successes. Most would agree that the last 10 years have been the most important behavior therapy is ever likely to know. With the coming of a new decade, it seems appropriate to reflect a bit on our immediate past and speculate on some possible directions, particularly in regard to our research and the role of all of our clinicians in this effort. To begin with, I think it is possible now to get some perspective on the growth of our field, and it is particularly interesting to examine some of the predictions about the future of behavior therapy which were made during the last two decades. The first prediction, made at the birth of behavior therapy in the late 50s and early 60s, was that this approach was the wave of the future and that other psychological therapies would soon disappear as our true, scientific approach to solving human behavioral problems gained ascendancy. In retrospect, this was probably a necessary development, in reaction to the relatively nonempirical approaches of the 40s and 50s. The resulting intellectual revolution created considerable excitement, with its host of prophets and true believers. Much has been written about this period (e.g., Kazdin, 1978a). Nevertheless, for all of its seeming historical inevitability, the arrogance and abuses inherent in any revolution came back to haunt behavior therapy toward the end of the 60s and 70s. Many of these arrogances were outlined by Kazdin last year (Kazdin,

32  David H. Barlow

1979), and among them was the exclusive claim to the scientific approach to the study of human behavior problems. And, of course, the inevitable superiority of all behavioral techniques. But, it is interesting to put this arrogance in some perspective. In addition to being a characteristic of most social or intellectual revolutions, it also resembles, to a remarkable degree, the type of arrogance found in other humanistic philosophies and endeavors. For, if nothing else, behavior therapy is a true, humanistic approach to behavioral and emotional problems, as is evident when one examines the humanist manifesto, signed, incidentally, by B. F. Skinner. For despite all the arguments on who is humanistic and who isn’t, at the core of the philosophy of humanism is a supreme faith in human reason and the methods of science to confront and solve the many problems that humans face and to rearrange the world so that human life will prosper. Ehrenfeld (1978), in an interesting book on humanism, has observed that with its lofty goals and ambitions, and beliefs in the power of its technology to engineer all behavior, the behavioral approach epitomizes a humanistic view of dealing with behavioral problems and, in doing so, demonstrates its arrogance. This arrogance contributed in part, I think, to negative views which have characterized behavior therapy in some quarters during a decade of skepticism about science and technology in general—including nuclear technology, genetic engineering, and the like. The major questions that have arisen as the power of science becomes apparent in all areas, are, of course, ethical ones. And I am pleased to say that members of the Association for Advancement of Behavior Therapy (AABT) have been in the forefront of putting our emerging technology of behavior change into a perspective of human values and ethics (e.g., Stolz & Associates, 1978). Nevertheless, with Kazdin last year (Kazdin, 1979), it is safe to say that behavior therapy has not become the wave of the future, at least as outlined by some of the prophets of the 60s, and our technology is not likely to solve soon all of the world’s problems. A second prediction, which is seemingly contradictory, but in fact very related, was made at the beginning of this decade. The essence of this prediction was that behavior therapy would die. These views, put forth in articles by London (1972), Lazarus (1977), Krasner (1976), and others, differed somewhat in the reason behind their conclusions, but each of these futurists basically agreed that the dogma of behaviorism and perhaps the language would fade away, while the basic techniques of behavior therapy would be co-opted into the larger mental health or behavior change professions. Several beliefs were at the heart of this prediction. Foremost among them was the growing effectiveness of behavior therapy techniques, making their widespread adoption, or co-option, desirable. But, also, these prophecies reflected the general poor image of behavior therapy and behaviorism in the public’s eye as manipulative, mechanistic, rigid, and perhaps too scientific or humanistic. The media, of course, has not helped. For example, Turkat and Feuerstein (1978) observed that about 50% of the articles in the New York Times, over a 4-year period, equated behavior modification with such procedures as brainwashing, psychosurgery, sensory deprivation, and Chinese water torture. It is tough to

Behavior Therapy: The Next Decade  33

overcome that image, and it has clearly had an effect. Many, I’m sure, saw an article by Woolfolk, Woolfolk, and Wilson (1977) who found that when they described an effective, standard behavioral classroom procedure as humanistic education it was rated much more favorably by undergraduates than when it was called behavior modification, even though the procedure was effective in accomplishing its goal no matter what the label. These and other studies suggest that the culprit is not the procedures in the eye of the public, but the language, and behavior therapy will be more widely accepted by the public if it is presented without laboratory conditioning language, as Bandura (1977) has suggested. These prophecies of death came to mind last spring in New Orleans when I attended the death and burial of the largest branch of AABT, the Southeastern Association of Behavior Therapy, which at the time of its demise had over 800 members. The President of that Association, Marilyn Erickson, along with its Board, noted that the goals of the Association had been achieved and that behavioral techniques were widely used in the Southeast and well represented at the meetings of the Southeastern Psychological Association, and thus, there was no further reason for existence. It is not a large inferential leap to suggest that perhaps the AABT itself would cease to exist during the next decade, since there is no advantage, evidently, to being associated with behavior therapy in the public’s eye and since the mental health professions, particularly clinical psychology and psychiatry, may become roughly behavioral in their own nonbiological approaches to mental illness, making a separate organization superfluous. In fact, our founding President, Cyril Franks, made such a prediction in the late 60s. And yet the facts don’t seem to support entirely these predictions of death, no matter how reasonable they seem. During this decade, the National Institute of Mental Health published a policy paper, to which the then Director of the Institute contributed, which strongly endorsed behavior therapy and encouraged its future development (Brown, Wienckowski, & Stolz, 1975). The American Psychiatric Association studied behavior therapy and concluded that every psychiatric residency should have ample training in this area (Birk, Stolz, Brady, Brady, Lazarus, Lynch, Rosenthal, Skelton, Stevens, & Thomas, 1973). Science magazine said that, “Today the content of clinical psychology has been significantly amplified by an explosion in the application of behavioral psychology.” It continues, “There is a body of knowledge and technique much broader than that available to practitioners even a decade ago. As a result, the number of psychologists in professional settings has grown rapidly” (Editorial comment, 1979, p. 339). In a similar vein, the President’s Commission on Mental Health has noted the effectiveness of behavioral techniques with a number of problems and strongly encouraged further research and development. Finally, in one of the more important developments, as many of you know, the AABT recently published an updated training directory in which the number of programs reporting significant behavioral training more than tripled in 4 years (Barlow, 1978). This seeming explosion of training opportunities reflects the fact that the greatest evidence for the success or growth of behavior therapy is within the

34  David H. Barlow

field itself. Contrary to all predictions on the death of behavior therapy and in contrast to the downplaying of the language of behavior therapy among the public because of its poor image, the demand for and the dissemination of knowledge among professionals is awesome. You’re all aware of the growth of books in the area, but one rather amazing development in the last few years is the tendency for authors and publishers to call their books behavioral when they really are not by any conceivable definition (e.g., see Kazdin, 1978b). When the marketing departments of large publishers recognize the selling power of behavior therapy, then we must be doing something right. Also, the oft cited growth in number of journals really deserves mentioning again. From two journals in 1969, we now have 21 by my own rough count, excluding journals in such related fields as biofeedback and programmed instruction, and there are new proposals coming in every day. All of this reflects the overwhelming evidence that behavioral approaches are effective. With both adult and child populations, with institutionalized and outpatient populations, the success of behavioral approaches to specific problems, when compared to no treatment or other approaches, has been demonstrated time and again, as recently summarized in several excellent books (Agras, Kazdin, & Wilson, 1979; Kazdin & Wilson, 1978). Behavior therapy is not dead; it is not dying; indeed it seems to be undergoing unprecedented growth and success, and it seems to me that our approach can contribute substantially to a technology necessary for effective human behavior change in the next decade. I would also agree with Wilson (1978) that the term “behavior therapy” will remain relevant within professional circles, although the laboratory conditioning language will probably disappear. For what the public is seeking, after all, with all of us, is relief from human suffering, and semantics or basic philosophical debates on Skinnerian freedom are unlikely to interfere with this goal for very long. And yet, I think the very consequences of our success represent the beginnings of some trends which, if unchecked, could result in deterioration of the principles which ensured the progress of behavior therapy to date. Now, many factors have contributed to this success, but it can be reasonably argued that the most important factor is the enthusiasm among those developing and practicing behavior therapy to employ and be guided by the findings from an empirical scientific approach. But, after the unprecedented period of research and development just past, I think our science is becoming shallow. With all of our recent technological and preliminary successes, we are rushing to publicize and use the technology, which is certainly understandable, but we are slowing down, it seems to me, when it comes to doing the necessary clinical or applied research which will make our procedures not only truly effective but also practical and useful to those in the front lines who apply them every day. Instead of asking why does a treatment work, and what are the ingredients of a given technique that are truly effective, the question is often how much of a chance is there that this package treatment, as it currently exists, might work with some broadly defined problem. And any probability at all is usually good enough. This erosion, to the extent that it exists, of our careful, fine-grained analysis and technique-building approach could not come at a worse time, since, as

Behavior Therapy: The Next Decade  35

I mentioned earlier, our techniques and procedures are just beginning to pay off. This is nowhere more evident than in the rash of self-help books on the market. To use one example that Warren Steinman (e.g., see Luiselli, Marholin, Steinman, & Steinman, in press; Marholin, Steinman, Luiselli, Schwartz, & Townsend, 1979) has noted: How many package procedures do we now have, many of which have been published in self-help books, which include relaxation training as a component? From weight reduction to social skills training to insomnia and smoking, and in some cases even parenting books, relaxation is recommended as an important ingredient. And yet very seldom, if ever, is this component tested, nor is the real function of relaxation analyzed in this context. Another example I have mentioned before is the use of social approval or differential attention, particularly when delivered by parents. This technique can be found in almost every package treatment applied by technicians working with family problems, as well as in all of the parenting books found on the market. And yet Herbert and Sajwaj, and their colleagues, demonstrated in an important paper in 1973 that differential attention from mothers was not only ineffective in some cases, but actually made children worse in terms of disruptive behavior in other cases (Herbert, Pinkston, Hayden, Sajwaj, Pinkston, Cordua, & Jackson, 1973). At this time I thought this article demonstrated the strength of our research. It seemed this was an important finding that would generate further research and analysis into what factors were responsible for those failures. But 6 years later, where is the research? Where is the analysis? The package treatments continue to be extolled as some sort of miracle for parents. The point has been very well made by members of AABT that the evidence supporting self-help books is either weak or nonexistent (e.g., Glasgow & Rosen, 1979). Until the data are in, it is possible that these self-help books are doing more harm than good. But the point I am more interested in here is that the package treatments themselves are not being evaluated, except in a very superficial way, in that any differences between a large group receiving the package and one that doesn’t are not due to chance. Some have said that eventually a successful package, tested in this way, will be pulled apart and analyzed (Azrin, 1977), but I don’t see that happening and I don’t think our approach to research encourages it. Another area which illustrates this point is the development of effective fear reduction procedures for the treatment of severe clinical phobias, particularly agoraphobia. In reviewing progress in this area, at the request of several agencies this year, it was of particular pleasure to be able to describe what is, in a microcosm, the success story of behavior therapy. For one of the greatest advances in mental health in the past 15 or 20 years, in my opinion, has been in the development of effective treatments for anxiety-based disorders such as phobia and obsessive-compulsive disorders. The development of these procedures is particularly important since anxiety-based disorders comprise a sizeable fraction of adult clinical problems, and the most usual treatment, minor tranquilizers, is not only ineffective, but actually harmful in light of the high incidence of dependence and addiction. Since the pioneering work of Wolpe (1958), we are all aware of the evidence in the last 10 years indicating that even severe, disabling, chronic agoraphobia can be successfully treated in

36  David H. Barlow

several weeks by prolonged exposure (Mavissakalian & Barlow, in press). This has been hailed as a great breakthrough, and indeed it is, in one sense, for the millions of people who suffer from this problem. But now, specialty clinics for agoraphobia have opened around the country. Many have made network TV. And the inevitable self-help books containing the current popular package of procedures are now being published. And yet, how effective is this procedure? Well, we know for a fact that when compared to no treatment or some good placebos, there are only 5 chances in 100 that it is not significantly better than these. But what does this mean in terms of the number of people who get better? This is a much more difficult question to ask when one reviews the literature, since it is very hard to compare outcomes of treatments tested thus far, due to greatly differing ways of assessing progress and other factors. Some investigators use clinical ratings by the therapist, some by the patients, some take behavioral measures, some take physiological measures, and so on. Follow-up on these studies is even more difficult to ascertain. Many studies use crossover designs which don’t permit follow-up, and other studies simply do not report follow-ups. Nevertheless, in the studies reported, it appears that the dropout rate alone from this procedure ranges between 8% and 40%, with a median of about 22%. Similarly, the improvement rate among those who complete treatment seems to average about 60% to 70%, and this is without specifying the degree of improvement. Thus, the number who are unimproved or perhaps worse off as a result of the procedure is as high as 40% of those who complete treatment (Mavissakalian & Barlow, in press). While these figures are gross estimations, due to the obvious problems mentioned earlier, it suggests that while exposure treatments are more effective than alternative treatments, the clinical significance is, as yet, still limited. For whom do these exposure treatments work? What if measures in the three recognized response systems, physiological, behavioral, and verbal, don’t correlate during or after treatment? What is the correct measure or combination of measures on which to judge a favorable outcome? Why does it work when it does work? Do those patients who are successful benefit from unspecified ingredients of an exposure package which in some way inhibit anxiety, or is the process simply one of habituation or extinction? Or is it a way of increasing judgments of self-efficacy or, as Lang (1977) has proposed more recently, a way of facilitating emotional processing in an information processing sense? We don’t know the answers to these questions yet, and in the midst of what is seemingly one of our biggest success stories, we are most likely propagating procedures that are making a lot of people worse. And yet, I think our research methods encourage us to continue to ignore our failures. Obviously this is a day in, day out problem for the practicing behavior therapist or for any clinician. He or she must make do with the very best that is available to relieve the suffering of those who are seeking help. But as of yet we don’t know why this works, and it seems that the kind of outcome study which tests the general package of in vivo exposure, cognitive restructuring and modeling, along with some relaxation thrown in, will not advance our understanding. And we are certainly not ready to make exaggerated claims in workshops or self-help books on the effectiveness of these packages. It will probably come as no surprise to many of you that I think that one of the major problems propelling us toward this premature advertisement of our

Behavior Therapy: The Next Decade  37

technology is the stipulation that we need only prove beyond the probability of 5 chances out of 100 that a given technique differs from no treatment or from some comparison treatment. This means absolutely nothing to the practicing clinicians confronted with an individual in a treatment setting since it ignores our many failures. I think this is the wrong way to advance behavior therapy, and I think this approach is indirectly related to the shallowness of current clinical research and our premature acceptance of package treatments. Another more important consequence of this approach, in my opinion, is that we are losing or have lost the ability to influence clinicians or practitioners doing the day-to-day work, who are either unwilling or unable to do this kind of research themselves and who are becoming skeptical of the data that are produced by the very few who do engage in this research. I’ve been particularly interested in comments from prominent and senior researchers in recent years on this issue, some from outside the field of behavior therapy. Recently Meehl (1978) said, when talking about Fisher and inferential statistics, “I suggest to you that Sir Ronald has befuddled us, mesmerized us, and led us down the primrose path. I believe that the almost universal reliance on merely refuting the null hypothesis as the standard method for corroborating substantive theories . . . is a terrible mistake, is basically unsound, poor scientific strategy, and one of the worst things that ever happened in the history of psychology” (p. 817). Five or six years ago, Strupp and I were on a panel, along with Hersen, where he discussed the inadequacy of single case designs for psychotherapy research. But more recently, in the Archives of General Psychiatry, he said, “As our work progressed, we became impressed with the fact that group comparisons obscure the very phenomena which must be understood in psychotherapy research and that constitute its essence . . .” (Strupp & Hadley, 1979, p. 1135). Azrin (1977), in a very important paper on outcome research, has noted the dangers of losing sight of the individual, who might be buried in the group average. Now group comparison research doesn’t have to be misleading, and I, for one, certainly don’t wish to imply that it always is. After all, statistics or any particular research methodology are simply tools which can be used for better or for worse. One only has to examine the work of Azrin (1977) or Paul (Paul & Lentz, 1977) to see examples of work that does not lose sight of the individual and that attends to follow-up and clinical significance. But I think that this approach, with its emphasis on factorial designs and multivariate statistics as the zenith of accomplishment, discourages an emphasis on the individual and clinical significance. Thirty years of experience with this approach underscores this concern, despite warnings and cautions to the contrary (e.g., Bergin & Strupp, 1972) to those who should already know better. What I’m advocating, of course, as the more appropriate strategy for our science, is a fine-grained analysis of an individual’s behavior, or a series of individuals’ behavior, with attention to technique building, repeated measurement, and social rather than statistical significance: an approach to science that in its very nature attends to our failures (Barlow & Hersen, 1973; Hersen & Barlow, 1976). Without this approach, I think our research will advance only very slowly. We’ve begun to apply technology in an overenthusiastic manner and in a way that precludes a behavioral analysis. In this respect, we’re no

38  David H. Barlow

different from the psychotherapists in the middle of the century or the social planners of the 60s. I think we need to resist the tendencies to be co-opted into the mainstream of clinical psychology and psychiatry if this means reverting back to the kind of theorizing and premature clinical application and low yield clinical research that characterized the early years in the behavior change field. I think we need a reaffirmation of fine-grained analysis of behavior, greater attention to the individual, and clinical and social significance. This should not be a step backward into a defensive stance where all other approaches are precluded simply because they are not dogma, but rather it should promote a more realistic view of our methodological tools, and maybe a fuller understanding of why behavior therapy has advanced us so far. More importantly, it may prevent the premature application of technology to systems without an appropriate analysis, because we don’t have all the answers now, and if the trend continues, we never will. Some have said that this argument is largely academic. Given the difficulties of carrying out meaningful clinical research and the scarcity of resources available to support research, only a few will be doing research in any case. Let them fight about methodology. After all, we know that the modal number of publications for clinical psychologists is zero, despite the fact that the Doctor of Philosophy degree, still awarded by most schools, presumably indicates some competence in research, and is not any better for physicians or social workers. In fact, I believe that this is the last and best reason for re-emphasizing an approach more appropriate to behavior therapy since it is very clear from our experience of the last 20 years, despite the rise of behavior therapy and the clear demonstration that meaningful research can be done, that we’re not reaching the practitioners, and the practitioners are not part of this process. Carl Rogers said, at the turn of the last decade, that he doubted if any research had ever really affected his clinical practice, nor did he expect it to (see Bergin & Strupp, 1972). I doubt if many behavior therapists would go so far, but I think, if you examine the behavior of practicing clinicians, few, if any, of our clinical procedures are guided by the scientific training we received. And despite the recent advances in behavioral assessment, I doubt if even a small minority of behavior therapists administer more than an occasional questionnaire or a short period of self-monitoring to determine the effectiveness of their interventions, an impression confirmed in recent surveys by Ford and Kendall (1979) and Swann and MacDonald (1978). A letter several months ago in The Behavior Therapist from Mozer (1979) of Montana reflects many of these issues, and I choose to mention it since I have personally received many letters during my year as President of AABT that reflect similar concerns. Mozer and other practitioners like him are not complaining of the inability to do science in a private practice; most practitioners gave up that idea long ago. What he speaks of is the difficulty of using techniques which seem to be successful in journal articles in a clinical setting. Perhaps more importantly, he notes that he has become skeptical of the extent to which the scientific method can obtain useful information for therapists. What Mozer is talking about, of course, is the gap between clinical reality and the production and use of scientifically reasonable data, a gap which clinical psychologists in particular have been trying to bridge since the

Behavior Therapy: The Next Decade  39

Boulder conference over two decades ago noted that this would be desirable. But it hasn’t happened, and I don’t think it will happen as long as our science emphasizes factorial designs, multivariate statistics, and the .05 level of probability. For these will never be used in the private office or clinic. And yet, I really believe that the closing of this scientist–practitioner gap has to be accomplished, and will be accomplished; that a time is coming when our practitioners will not only readily consume the important advances in clinical research, but actually produce many of the advances themselves. This development, if it occurs, would counteract the growing shallowness of our science, multiply the data base on which our true advances are made by a factor far greater than we have thought possible, and make the scientist–practitioner a reality. There are two major factors which should make this possible. The first is the development of realistic and practical measures or assessment techniques for specific problems which are widely agreed on and easily used by every practitioner. The second is the growing pressure from government and society to be accountable for what we do on every single case we see. For if there is one crystal clear trend in the next decade, it is accountability; which will be achieved by demonstrating unambiguously to third-party payers and government agencies alike, that what we do is effective. These two factors, combined with the developing sophistication in using the basic ingredients of the intensive study of the single case, make it possible, I believe, for every practicing clinician to be a major part of the technique building and testing process. For the procedures involved in the single case approach to applied research are very close to the procedures used in the clinic, and its most basic ingredient, repeated measurement of target problems, is becoming more readily achievable by practitioners as new, easy to use measures are developed. Some of my good friends and colleagues have criticized the single case approach as being unable to close the scientist–practitioner gap because of the complexities involved in single case research (see Agras et al., 1979), but then, few have really tried. One enormous step forward that would greatly expand our data base is the use of the most rudimentary of single case designs, sometimes called the A-B design, where simple but reliable measures are administered for a short time prior to treatment, and then repeatedly during treatment and through a follow-up period. Normal demographic and life situation information routinely collected during the interview would be available to correlate with successes and failures. This would be a relatively small step for clinicians, bur one which would provide us with invaluable information on the generality of effect of our newly emerging technology, without losing sight of the individual. It would also be a small step to then go on to do functional analyses on certain series of cases if interesting questions arose, although this would not be routinely done nor expected. What is most needed, of course, and what is happening I think, is that realistic and practical measures repeatedly employed to assess progress over time in individual clients are now being developed and used. Several groups, including AABT, are now preparing task force reports on the most effective procedures and most relevant measures to use with specific problems. It shouldn’t be overlooked that third-party payers or others interested in this process, such as the people involved in the APA/CHAMPUS project, would be more than pleased with this development, and fortunately, or unfortunately depending

40  David H. Barlow

on your point of view, might quickly make it mandatory. This development would increase the data base of our science and provide the kind of information on hundreds, and even thousands, of clients heretofore only available in the most ambitious 10- or 15-year psychotherapy outcome study. What I’m advocating here, of course, is not a return to the type of percentage outcome study that was often reported in the 50s, where an ill-defined technique such as psychotherapy or behavior therapy was applied to a diffuse population and progress was assessed by some overall judgment of outcome or perhaps on some general personality measure. Rather, these would be a series of specific problems treated by specific, well-described techniques, using simple, practical, but widely agreed on measures of change. One small illustration of this can be found in a report of a recent series of cases by Maletzky (1980). During the last 10 to 12 years, Maletzky has treated some 186 exhibitionists. Progress was assessed in all cases by the now standard behavior and fantasy records, with corroborating reports from parents or relatives and, occasionally, the courts. In addition, penile plethysmograph data have been collected on the last 80 or so exhibitionists. The techniques used in treatment changed somewhat from patient to patient, based on individual behavioral analysis, but were always very specific and well described, and comprise what most people would agree would be a sophisticated behavioral approach to the treatment of exhibitionism. In addition to collecting these data, Maletzky also conducted several functional analyses using single subject or small group comparison approaches on certain aspects of his procedures in a “technique building” fashion. The basic findings are remarkable in that of the first 155 clients treated, 86% improved to the extent that all overt exhibitionistic behavior was eliminated. The results in the last 31 clients were even better, with 28 showing this degree of improvement. And the follow-ups, including objective physiological measures of arousal administered with equipment now readily available to practitioners, are continuing into their ninth year. This is an extraordinary series; but, more importantly, Maletzky has taken the time to collect the basic demographic data and to look for patterns among the 30 or so failures. Thus, he does not lose sight of the individual. I’m not doing justice to the richness of these data in this brief summary, but Maletzky, who is a full time private practitioner, closes the chapter in which he summarizes his results by saying, “Who knows, when the lions of clinical practice lie down with the lambs of academia, what brain children they may hatch.” I think that speaks for itself. I would predict that these data are going to have a major impact. Consider for a moment the impact of some other major series in the past where specific, well-described techniques were applied to a large series of clients with more or less specifically described problems. In 1958, Wolpe described in some detail the treatment of approximately 210 anxiety-based disorders, mostly phobias, with results that you all know well. Another series of carefully described procedures applied to specific objective problems was published by Masters and Johnson (1970). I would suggest that these two series have done more to influence our practice and our research than any more traditional research published anywhere, anytime. And these two series did not have the benefit of widely agreed on practical measures, nor did Wolpe or Masters and

Behavior Therapy: The Next Decade  41

Johnson really pull apart their package treatment. But those series had such an impact that others are pulling it apart for them. No one, least of all these investigators, would pretend for a minute that these data can establish causal or functionally analytic relationships. But, on the other hand, no one can deny the impact on the field; and yet I think they are in the best tradition of single case research. It seems to me that a priority of our training programs over the next decade should be to produce the kind of practitioner who can collect these data and contribute to our mutual data base, rather than to divorce him- or herself from any form of science once terminal degrees are received. I believe this can be done, and to the extent that a scientific empirical approach is the heart of behavior therapy, then behavior therapy will come to mean something to every practicing clinician because it will be a specific way of going about solving human problems rather than simply a collection of techniques which are easily co-opted into the eclectic, non-data-based psychotherapist’s armamentarium. Unless we incorporate these principles into our training programs and make a serious attempt to reach the clinicians, then behavior therapy may well be smothered and strangled by its own success long before our goals are reached. This approach, in my opinion, will be the reason for behavior therapy to continue to exist and to maintain its identity and its purpose through the next decade. Note Presented as the Presidential Address to the Association for Advancement of Behavior Therapy at the Annual Convention in San Francisco, December 1979. Requests for reprints should be sent to the author. Department of Psychology, State University of New York, Albany, NY, 12222.

References Agras, W. S., Kazdin, A. E., & Wilson, G. T. Behavior therapy: Toward an applied clinical science. San Francisco: W. H. Freeman & Company, 1979. Azrin, N. H. A strategy for applied research: Learning based but outcome oriented. American Psychologist, 1977, 32, 140–149. Bandura, A. Social learning theory. Eaglewood Cliffs, NJ: Prentice-Hall, 1977. Barlow, D. H. (Ed.). Directory of graduate study in behavior therapy. New York City: Association for Advertisement of Behavior Therapy, 1978. Barlow, D. H., & Hersen, M. Single case experimental designs: Uses in applied clinical research. Archives of General Psychiatry. 1973, 29, 319–325. Bergin, A. E., & Strupp, H. H. Changing frontiers in the science of psychotherapy. Chicago: Aldine, 1972. Birk, L., Stolz, S. B., Brady, J. P., Brady, J. V., Lazarus, A. A., Lynch, J. J., Rosenthal, A. J., Skelton, W. D., Stevens, J. B., & Thomas, E. J. Behavioral therapy in psychiatry. Washington, DC: American Psychiatric Association, 1973. Brown, B. S., Wienckowski, L. A., & Stolz, S. B. Behavior modification: Perspective on a current issue. DHEW Publication No. (ADM) 75–202. Washington, DC: U.S. Government Printing Office, 1975. Editorial comment. Science, January 26, 1979, 339.

42  David H. Barlow Ehrenfeld, D. The arrogance of humanism. New York: Oxford University Press, 1978. Ford, J. D., & Kendall, P. C. Behavior therapists’ professional behaviors: Converging evidence of a gap between theory and practices. The Behavior Therapist, 1979, 2(5), 37–39. Glasgow, R. E., & Rosen, G. M. Self-help behavior therapy manuals: Recent developments and clinical usage. Clinical Behavior Therapy Review, 1979, 1, 1–20. Herbert, E. W., Pinkston, G. M., Hayden, M. L., Sajwaj, T. E., Pinkston, S., Cordua, G., & Jackson, C. Adverse effects of differential parental attention. Journal of Applied Behavior Analysis, 1973, 6, 15–30. Hersen, M., & Barlow, D. H. Single case experimental designs: Strategies for studying behavior change. New York: Pergamon Press, 1976. Kazdin, A. E. History of behavior modification: Experimental foundation of contemporary research. Baltimore, MD: University Park Press, 1978a. Kazdin, A. P. A behavioral formulation of psychotherapy (review of Counseling and psychotherapy: A behavioral approach by E. L. Phillips). Contemporary Psychology, 1978b, 23, 237–239. Kazdin, A. E. Fictions, factions, and functions of behavior therapy. Behavior Therapy, 1979, 10, 629–654. Kazdin, A. E., & Wilson, G. T. Evaluation of behavior therapy: Issues, evidence, and research strategies. Cambridge, MA: Bollinger Publishing Company, 1978. Krasner, L. On the death of behavior modification: Some comments from a mourner. American Psychologist, 1976, 31, 387–388. Lang, P. J. Imagery in therapy: An information processing analysis of fear. Behavior Therapy, 1977, 8, 862–886. Lazarus, A. A. Has behavior therapy outlived its usefulness? American Psychologist, 1977, 32, 550–554. London, P. The end of ideology in behavior modification. American Psychologist, 1972, 27, 913–926. Luiselli, J. K., Marholin, D., Steinman, D. L., & Steinman, W. M. Assessing the effects of relaxation training, Behavior Therapy, in press. Maletzky, B. M. “Assisted” covert sensitization in the treatment of exhibitionism. In D. J. Cox & R. J. Daitzman (Eds.), Exhibitionism: Description, assessments, and treatment. New York: Garland Press, 1980. Marholin, D., Steinman, W. M., Luiselli, J. K., Schwartz, C. S., & Townsend, N. M. The effect of progressive muscle relaxation on the behavior of autistic adolescents: A preliminary analysis. Child Behavior Therapy, 1979, 1, 75–84. Masters, W., & Johnson, V. Human sexual inadequacy. Boston: Little, Brown & Company, 1970. Mavissakalian, M. R., & Barlow, D. H. (Eds.). Phobia: Psychological and pharmacological treatment. New York: Guilford Press, in press. Meehl, P. E. Theoretical risks and tabular asterisks: Sir Karl, Sir Ronald, and the slow progress of soft psychology. Journal of Consulting and Clinical Psychology, 1978, 46, 806–835. Mozer, M. H. Confessions of an ex-behaviorist. The Behavior Therapist, 1979, 2(3), 3. Paul, G. L., & Lentz, R. J. Psychosocial treatment of chronic mental patients: Milieu versus social-learning programs. Cambridge, MA: Harvard University Press, 1977. Stolz, S. B., & Associates. Ethical issues in behavior modification. San Francisco: Jossey-Bass, 1978. Strupp, H. H., & Hadley, S. W. Specific vs. nonspecific factors in psychotherapy. Archives of General Psychiatry, 1979, 36, 1125–1137. Swann, G. E., & MacDonald, M. L. Behavior therapy in practice: A national survey of behavior therapists. Behavior Therapy, 1978, 9, 799–807.

Behavior Therapy: The Next Decade  43 Turkat, I. D., & Feuerstein, M. Behavior modification and the public misconception. American Psychologist, 1978, 33, 194. Wilson, G. T. On the much discussed nature of the term “behavior therapy.” Behavior Therapy, 1978, 9, 89–99. Wolpe, J. Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press, 1958. Woolfolk, A. E., Woolfolk, R. L., & Wilson, G. T. A rose by any other name . . . : Labeling bias and attitudes toward behavior modification. Journal of Consulting and Clinical Psychology, 1977, 45, 184–191.

Article 2

Why Can’t We Be More Idiographic in Our Research? David H. Barlow 1 and Matthew K. Nock 2 Center for Anxiety and Related Disorders, Boston University and 2 Harvard University

1

Most psychological scientists make inferences about the relations among variables of interest by comparing aggregated data from groups of individuals. Although this method is unarguably a useful one that will continue to yield scientific advances, important limitations exist regarding the efficiency and flexibility of such designs, as well as with the generality of obtained results. Idiographic research strategies, which focus on the intensive study of individual organisms over time, offer a proficient and flexible alternative to group comparison designs; however, they are rarely taught in graduate training programs and are seldom used by psychological scientists. We highlight some of the unique strengths of idiographic methods, such as single case experimental designs, and suggest that psychological science will progress most efficiently with an increased use of such methods in both laboratory and clinical settings. Edward Tolman said to Gordon Allport, “I know I should be more idiographic in my research, but I just don’t know how to be,” to which Allport replied, “Let’s learn!” (Allport, 1962, p. 414). This sentiment was based on the fact that, whether it’s a laboratory rat or a patient in the clinic with a psychological disorder, it is the individual organism that is the principal unit of analysis in the science of psychology. The intensive study of the individual is associated with a hallowed tradition in scientific psychology. Indeed, the founders of experimental psychology including Fechner, Wundt, Ebbinghaus, and Pavlov studied individual organisms with scientific approaches that would be considered internally valid, and they strengthened these findings (and began to establish generality) through replication in other organisms (see Barlow, Nock, & Hersen, 2008). This scientific strategy, which is fully capable of establishing causal relations among variables, came to be known as the idiographic approach. Gordon Allport, in his area of social psychology, argued eloquently that the science of psychology should attend to the uniqueness of the individual organism (Allport, 1962). Rooted deep in the structural school of psychology, this

Why Can’t We Be More Idiographic in Our Research?  45

approach also was popular in more applied branches in psychology in the middle of the last century. Perhaps the biggest champion of an idiographic approach in clinical settings was Shapiro, who was advocating a scientific approach to the study of individuals with psychopathology as early as 1951 (e.g., Shapiro 1961, 1966). The idiographic approach perhaps reached its zenith in psychological science with the work of B. F. Skinner. In a famous quote, Skinner (1966) noted: “. . . instead of studying a thousand rats for one hour each or a hundred rats for ten hours each the investigator is more likely to study one rat for a thousand hours” (p. 21). Thus, Skinner and his colleagues in the animal laboratories are largely credited with developing and refining an experimental idiographic approach that came to be known as the experimental analysis of behavior. This idiographic approach represents a true scientific undertaking, as independent variables are manipulated in the context of carefully measured and repeatedly assessed dependent variables. This is in contrast to the alternative nomothetic experimental strategy, in which the researcher looks to assemble relatively large groups of individual organisms and, in the most straightforward application, examines the average response of the group to the introduction of some manipulation compared with the response to well-construed control conditions. The major differences between the idiographic and nomothetic traditions are, of course, approaches to intersubject variability and the generality of findings. As variability is often considerable among organisms, the task of any psychological scientist is to discover functional relations among independent variables over and above the welter of environmental and biological variables influencing the organism at any given point in time. A nomothetic approach makes an implicit assumption that much of this variability is intrinsic to the organism and uses sophisticated data analytic procedures to look for reliable effects over and above this “error.” Significant effects are then assumed to be more or less generalizable based on the number of individuals included in the experimental group and the representativeness of the population of such individuals (i.e., the use of random sampling). Of course, random sampling is seldom achieved in psychological research where, indeed, the goal is more often to strive for homogeneous samples in which the generality of findings can be very limited. Sidman (1960) made the following point a number of years ago when discussing approaches to variability: The rationale for statistical immobilization of unwanted variables is based on the assumed random nature of such variables. In a large group of subjects, the reasoning goes, the uncontrolled factor will change the behavior of some subjects in one direction and will affect the remaining subjects in the opposite away. When the data are averaged over all the subjects, the effects of the uncontrolled variables are presumed to add algebraically to zero. The composite data are then regarded as though they were representative of one ideal subject who had never been exposed to the uncontrolled variables at all. (p. 162)

46  David H. Barlow and Matthew K. Nock

Addressing the issue of the generality of findings, Sidman wrote the following: Tracking down sources of variability is then a primary technique for establishing generality. Generality and variability are basically antithetical concepts. If there are major undiscovered sources of variability in a given set of data, any attempt to achieve subject or principle generality is likely to fail. Every time we discover and achieve control of a factor that contributes to variability, we increase the likelihood that our data will be reproducible with new subjects and in different situations. Experience has taught us that precision of control leads to more extensive generalization of data. (p. 152) Although the use of the idiographic approach led to significant advances in the earliest days of laboratory-based experimental psychology, as well as during early translations of findings from psychological science to clinical applications in the middle of the last century, it is clear that the nomothetic strategy has become a dominant method to establish both internal and external validity over the past few decades (Kazdin, 2003; Nock, Janis, & Wedig, 2008). One reason for this development in applied settings was the beginning of funding of large randomized clinical trials (RCTs) by the National Institutes of Health. Many such studies require 10 or more years and many millions of dollars to perform one treatment trial. For instance, the National Institute of Mental Health (NIMH) funded the treatment of depression collaborative research program (Elkin et al., 1989). This study, which took 13 years to finish (1977–1990), was reminiscent of earlier efforts such as the Cambridge Somerville Youth Study conducted from 1935 through 1951, which divided delinquent boys into two groups—one treatment group and one group that received “treatment as usual” (McCord, 1978). The fact that there were no effects at 5, 10, 20, or 30 years did much to discourage efforts of this type for at least the next 30 years. In fact, results from the NIMH depression collaborative trial were not particularly revealing either, as no significant differences existed among treatment and comparison groups at any point in time. Nevertheless, this trial provoked useful comment and a great deal of controversy about strategic issues and the potential for improvement in the methodology of RCTs. These trials have improved to the point where they have become “the gold standard” for establishing causal relations between independent and dependant variables more generally, and data emanating from these trials have deep influences on health care practices (Barlow, 2004). But is something still lacking? Scientifically, relying on a relatively small group of researchers requiring enormous amounts of time and resources to perform a single treatment trial can be seen as an inefficient method of advancing knowledge. In applied clinical settings, clinicians often question the applicability of findings from RCTs to individuals seen in typical clinical settings. In other words, there is a strong perception that problems exist in generalizing a nomothetic result to an idiographic situation. The variety of forms that these arguments take are often cast as specific objections to RCT methodology, and these arguments have been detailed numerous times in the past decade (e.g., Persons & Silberschatz, 1998; Westen, Novotny, & Thompson-Brenner, 2004).

Why Can’t We Be More Idiographic in Our Research?  47

Rather than simply critiquing nomothetic methodologies, can we enrich these methodologies with a complementary focus on the individual? The fact is that we have a good idea of how to be more idiographic in our research. Although most psychological researchers have been trained in group comparison designs and have relied primarily on them, exciting advances have been made in the use of idiographic methodologies, such as the single-case experimental design (see Barlow et al., 2008). The flexibility and efficiency of these designs make them ideally suited for use by psychological scientists, clinicians, and students alike, given that they require relatively little time and few resources and subjects and yet they can provide strong evidence of causal relations between variables. The time now seems right to put more emphasis on idiographic strategies that can be integrated in a healthy way into existing nomothetic research approaches in both clinical and basic science settings. In clinical science, having established the effectiveness of a particular independent variable (e.g., an intervention for a specific form of psychopathology), one could then carry on with more idiographic efforts tracking down sources of intersubject variability and isolating factors responsible for this variability (Kazdin & Nock, 2003; Nock, 2007). Necessary alterations in the intervention protocols to effectively address variability could then be tested, once again idiographically, and incorporated into these treatments. Researchers in basic science laboratories could undertake similar strategies and avoid tolerating large error terms. Thus, all of psychological science, both basic and applied, would benefit. References Allport, G.D. (1962). The general and the unique in psychological science. Journal of Personality, 30, 405–422. Barlow, D.H. (2004). Psychological treatments. American Psychologist, 59, 869–878. Barlow, D.H., Nock, M.K., & Hersen, M. (2008). Single case experimental designs: Strategies for studying behavior change (3rd ed.). Boston: Allyn & Bacon. Elkin, I., Shea, M.T., Watkins, J.T., Imber, S.D., Sotsky, S.M., Collins, J.F., et al. (1989). National Institute of Mental Health Treatment of Depression Collaborative Research Program: General effectiveness of treatments. Archives of General Psychiatry, 46, 971–982, 983. Kazdin, A.E. (2003). Research design in clinical psychology (4th ed.). Boston: Allyn & Bacon. Kazdin, A.E., & Nock, M.K. (2003). Delineating mechanisms of change in child and adolescent therapy: Methodological issues and research recommendations. Journal of Child Psychology and Psychiatry, 44, 1116–1129. McCord, J. (1978). A thirty-year follow-up treatment effects. American Psychologist, 33, 284–289. Nock, M.K. (2007). Conceptual and design essentials for evaluating mechanisms of change. Alcoholism: Clinical and Experimental Research, 31, 4S–12S. Nock, M.K., Janis, I.B., & Wedig, M.M. (2008). Research design. In A.M. Nezu & M. Nezu (Eds.), Evidence-based outcome research: A practical guide to conducting randomized controlled trials for psychosocial interventions (pp. 201–218). New York: Oxford University Press. Persons, J.B., & Silberschatz, G. (1998). Are results of randomized controlled trials useful to psychotherapists? Journal of Consulting and Clinical Psychology, 66, 126–135. Shapiro, M.B. (1961). The single case in fundamental clinical psychological research. British Journal of Medical Psychology, 34, 255–263.

48  David H. Barlow and Matthew K. Nock Shapiro, M.B. (1966). The single case in clinical psychological research. Journal of General Psychology, 74, 3–23. Sidman, M. (1960). Tactics of scientific research: Evaluating experimental data in psychology. New York: Basic Books. Skinner, B.F. (1966). Operant behavior. In W.K. Honig (Ed.), Operant behavior: Areas of research and application (pp. 12–32). New York: Appleton-Century-Crofts. Westen, D., Novotny, C.M., & Thompson-Brenner, H. (2004). The empirical status of empirically supported psychotherapies: Assumptions, findings, and reporting in controlled clinical trials. Psychological Bulletin, 130, 631–663.

Treatment of Anxiety and Related Disorders

Article 3

Social Reinforcement in the Modification of Agoraphobia Stewart Agras, MD; Harold Leitenberg, PhD; and David H. Barlow, MA Burlington,VT

Although there is no convincing evidence from group outcome studies that psychotherapy is effective1, recent work2 suggests that groups of patients so treated show both negative and positive change when compared to untreated controls. Those treated tend to be widely dispersed from improved to worsened, while untreated control subjects usually show slight improvement and cluster about the mean. The variability in outcome following psychotherapy appears to depend on therapist and to a lesser extent on patient characteristics. Thus the evidence suggests that behavioral change occurs during psychotherapy, but is masked in large-sample statistical-outcome studies, where positive and negative therapeutic effects cancel out. An alternative research approach is the controlled study of the single case. This approach allows for detailed and sensitive investigation of the variables responsible for behavioral change in psychotherapy. Individualized and direct measures of pertinent symptomatic behaviors can be devised and monitored throughout an experimental therapy. The effect of a single therapeutic variable on neurotic behavior can then be determined by its introduction, removal, and reintroduction in sequence. Few attempts have been made to systematically change the behavior of a therapist during therapy with an individual patient, and to study the effect of this on both a clinically relevant and directly measurable aspect of the patient’s behavior. This paper describes an investigation in which therapists’ verbal behavior was varied so that the effect of selective positive reinforcement (social praise) on agoraphobic behavior could be studied. Selective reinforcement is both contingent upon and immediately follows a previously specified behavior of the subject. It is therefore different from noncontingent therapeutic support often loosely regarded as reinforcement. Moreover, its presence or absence can be clearly specified allowing for experimental manipulation of therapist behavior. Selective positive reinforcement was chosen for study because it has been shown to modify behavior; (a) in basic studies of animal behavior,3 and human verbal behavior4 and (b) in clinical studies of schizophrenia,5 mental retardation,6 and children’s disorders.7 It is therefore likely to be important in the modification of neurotic behavior.

52  Stewart Agras et al.

Procedure The subjects were three agoraphobic patients, two women (Ss 1 and 3) 23 and 39 years old, and one man (S2) 36 years old. They had been severely phobic for 1, 15, and 16 years respectively, all having numerous fears including fear of walking alone, traveling alone, crowds, illness, and death. Subjects 1 and 3 had been unable to leave their homes alone, while subject 2 had been able to manage a five-minute drive to work with difficulty. They were admitted separately to the University of Vermont Clinical Research Center. One of the central symptomatic behaviors in agoraphobia is the patient’s difficulty in leaving a dependent situation. Thus both time spent away, and distance walked alone from the Clinical Research Center, were used as indicators of phobic behavior. These were measured by mapping out a “course” from the center to downtown, landmarks being identified at 25-yard intervals for over a mile. The subjects were asked to stay on the course, to note the point at which they turned back, and were told, “We would like to see how far you can walk by yourself without experiencing undue tension. We find that repeated practice in a structured situation often leads to progress.” Two sessions were held each day in the first case, and five in the other two cases. Each session lasted half an hour, unless the patient stayed out longer than 20 minutes on the first trial, in which instance the session was ended on return. At the end of each trial the patient reported the point reached in his walk, which was noted by the therapist. Since much of the course was easily observable, frequent checks of the patient’s behavior were made throughout each phase of the study, which confirmed the accuracy of their verbal report of distance walked. The duration of each walk was timed by the therapist using a stop watch. The patient was not told how long he had been away from the center. During the baseline period the therapist maintained a pleasant relationship with the patient, but made no comment on the distance walked or time spent away. Reports of improvement made by the patient were also ignored. In the reinforcement phase the first trial of each day and all trials meeting the criterion were reinforced. The criterion was usually established as the mean between the previous criterion and the next highest trial, allowing behavior to be steadily shaped. Thus, if the patient had been reinforced at a criterion of 5 minutes, and spent 10 minutes away on the next trial, the criterion became 7.5 minutes. The patient now had to be away for at least 7.5 minutes to be reinforced. The subject was not told of changes in criterion. During the first two reinforcement periods, with S1, and both reinforcement periods, with Ss 2 and 3, time away, not distance was reinforced. During the last two reinforcement phases with S1, the criterion behavior was changed from time away to distance away. The patient was not told of this change. This was not repeated in the second and third patients since they were not able to stay in the research unit long enough. Reinforcement consisted of praise such as, “Good . . . you’re doing well . . . excellent” given with appropriate enthusiasm. During the reinforcement phase remarks made by the patient to the nursing staff about progress were also praised.

Social Reinforcement in the Modification of Agoraphobia  53

Nonreinforcement consisted of a return to baseline conditions, stopping selective praise, but taking especial care to maintain a generally pleasant, supportive attitude towards the patient. In this way a distinction between general support and selective social reinforcement could be made. Findings The main findings are shown in Figures 3.1 to 3.3. Three different patterns of behavior were found during the baseline period in response to therapeutic instructions. S1 showed no response while S2 showed an apparently sustained response. The baseline data for time away in S3, expressed as means for successive pairs of trials in minutes, were 2.8, 4.2, 6.3, 8.1, 5.1, and 3.3. This suggests a transient response to therapeutic instructions in this case. Following the introduction of reinforcement, which was provided on about 80% of trials, both time away and distance walked alone showed a sustained increase for each subject. Although only increasing time away from the center was reinforced in this phase, distance walked tended to follow the increase in time. When reinforcement was withdrawn there was either no further improvement or a decline in these measures. During this phase the subjects expressed feelings of depression and criticized themselves for doing poorly. Reinstatement of reinforcement led to a rapid reacquisition of the lost behavior in each subject and to a loss of the depressive statements. These findings demonstrate that social reinforcement in combination with therapeutic instructions is a

Distance Time S1

1,800

Mean Distance (yd)

1,600 1,400 1,200 1,000

A B C D E F

60 50

Time Reinforced No Reinforcement Time Reinforced Distance Reinforced No Reinforcement Distance Reinforced

40 30

800 20

600 400

Mean Time (min)

2,000

10

200 A

2 4 6 Baseline

B

C

D

E

F

8 10 12 14 16 18 20 22 24 26 28 30 32 34 Blocks of 4 Sessions

Figure 3.1 The Effect of Reinforcement and Nonreinforcement upon the Time Spent Away and Distance Walked from the Hospital by an Agoraphobic Patient (subject 1).

54  Stewart Agras et al.

powerful modifier of established neurotic behavior. The experimental design allowed the effects of selective positive reinforcement to be separated from those of instructions and general therapeutic support since both of the latter variables were present during the entire experimental period and progress varied with the introduction and removal of selective praise. Although it is possible that withdrawal of social praise was accompanied by a loss of general therapeutic support, every effort was made to control this. Thus the nurses and the study psychiatrist, who were not the therapists, gave the patient equal amounts of attention and support during the entire period. In addition, the therapists attempted to omit only the selective praise. The selective effect of reinforcement upon behavior was demonstrated in another way in the first patient. Figure 3.1 shows that the change in reinforcement criterion from time away to distance away in the second part of the experiment, resulted in an immediate decrease in time while distance increased. On the day of the change the patient complained that she “was doing very badly . . . and was not able to stay out as long as before” even though she increased her distance walked. This suggests that she was not aware of the change in reinforcement criterion, although her behavior responded appropriately. Following a steady increase in both distance walked alone and time spent away, reinforcement was withdrawn for the second time. At this point a marked spurt in performance occurred. This was similar to that shown by 1,200 Distance Time

40

S2

800

30

600 20

Mean Time (min)

Mean Distance (yd)

1,000

400 10

200

2

No Reinforcement

Reinforcement

Baseline

4

6

8

10

12

14

Reinforcement

16

Blocks of 5 Trials

Figure 3.2 The Effect of Reinforcement and Nonreinforcement upon the Performance of an Agoraphobic Patient (subject 2).

Social Reinforcement in the Modification of Agoraphobia  55

S2 when reinforcement for time was withdrawn, and both parallel the initial extinction behavior found in animals.8 The brief spurt was followed by a dramatic drop in performance almost replicating the initial symptomatic behavior of the patient. Reinstatement of reinforcement for distance walked led to rapid reacquisition of the lost behavior. Despite the use of apparently antitherapeutic strategies such as periods of nonreinforcement, the first patient was able to walk alone downtown by the end of the study which lasted for 96 days in her case. This behavior transferred to other situations so that she had no difficulty in going where she pleased. A brief regression in response to a life upset occurred three months later; however, follow-up one year after treatment showed her to be symptom free. Comment While it is surprising that a relatively weak reinforcer can modify longestablished neurotic behavior, the findings confirm and extend a recent study by Truax9 who analyzed tape recordings of a single long-term case treated with nondirective psychotherapy. He unexpectedly found that the therapist responded differentially to five of the nine patient behaviors studied and 700

Distance Time S3

600

25

20

400 15

300

Mean Time (min)

Mean Distance (yd)

500

10 200 No Reinforcement

Reinforcement

2 Baseline

4

6

8

Reinforcement

5

10

Blocks of 10 Trials

Figure 3.3 The Effect of Reinforcement and Nonreinforcement upon the Performance of an Agoraphobic Patient (subject 3).

56  Stewart Agras et al.

that the rate of emission of four of these behaviors increased during therapy, clearly demonstrating the presence of selective social reinforcement. Similarly, in an investigation of systematic desensitization therapy which according to Wolpe10 is based on the reciprocal inhibition of anxiety by relaxation, we found11 that the combination of therapeutically oriented instructions and verbal reinforcement accounted for much of the progress of a group of female subjects with snake fears. Indeed, a group treated with systematic desensitization, but not given therapeutic instructions or reinforcement, did not differ significantly in their progress from an untreated control group, suggesting that these variables play a major part in this therapy. These findings in both nondirective psychotherapy and systematic desensitization together with the experimental confirmation of the effect of social reinforcement described in this paper, imply that variables such as instructions and reinforcement may be present in most forms of psychotherapy. This may partly explain the puzzling observation that apparently very different therapeutic procedures produce similar results.12 The common assumption that neurotic behavior cannot be modified by variables that have been shown to control normal animal and human behavior in the laboratory seems to have little basis in fact. Thus the investigation of the effect of such variables appears to be useful for psychotherapy research. The present study also demonstrates the usefulness of the controlled study of the single case to psychotherapy research, in which single variables are introduced, removed, and reintroduced in sequence while measurable changes in a well-specified clinically relevant behavior are observed. If the variables responsible for behavioral change in psychotherapy can be isolated singly and in their various combinations, then it may be possible to redesign existing therapies to increase their efficacy, or to design new forms of therapy based upon research findings. Summary Social reinforcement in the form of praise was found to modify the severely agoraphobic behavior of three patients. Reinforcement was introduced, removed, and reintroduced, in sequence in each case. During a baseline period without reinforcement, therapeutic instructions by themselves produced a different response in each patient. Following the introduction of social reinforcement there was a steady increase in time spent away, and distance walked alone. Dropping reinforcement resulted in either no further increase or a decline in performance, while reinstatement led to rapid improvement. The effect of social reinforcement was separated from the effects of instructions and general therapeutic support since both of the latter variables were present during the entire experimental period. The use of a sequential design in individual patients to study the effect of variables which have been shown to modify behavior in circumstances other than psychotherapy, appears to be a fruitful approach to psychotherapy research.

Social Reinforcement in the Modification of Agoraphobia  57

Note This work was supported by Public Health Service Clinical Research Center grant FR-109, and National Institute of Mental Health research grant MH-13651. L. Thomson, BA, and G. Woodman, RN, assisted in the study.

References 1 Eysenck, H.J.: The Effects of Psychotherapy, New York. International Science Press, 1966. 2 Truax, C.B., and Carkhuff, R.R.: Toward Effective Counseling and Psychotherapy: Training and Practice, Chicago: Aldine Publishing Company, 1967. 3 Honig, W.K.: Operant Behavior: Areas of Research and Application, New York: Appleton-Century-Crofts, 1966. 4 Greenspoon, J.: “Verbal Conditioning and Clinical Psychology,” in Bachrach, A.J. (ed.): Experimental Foundations of Clinical Psychology, New York: Basic Books, Inc., Publishers, 1962, pp. 510–553. 5 Ayllon, T., and Azrin, N.H.: The Measurement and Reinforcement of Behavior of Psychotics, J Exp Anal Behav 8:357–383 (Nov) 1965. 6 Bijou, S.W., and Orlando, R.: Rapid Development of Multiple-Schedule Performances with Retarded Children, J Exp Anal Behav 4:7–16 (Jan) 1961. 7 Patterson, G.R., et al.: A Behaviour Modification Technique for the Hyperactive Child, Behav Res Ther 2:217–226 (Jan) 1965. 8 Thompson, T.: The Effect of Two Phenothiazines and a Barbiturate on Extinction-Induced Rate Increase of a Free Operant, J Comp Physiol Psychol 55:714–718 (Oct) 1962. 9 Truax, C.B.: Reinforcement and Nonreinforcement in Rogerian Psychotherapy, J Abnorm Psychol 71:1–9 (Feb) 1966. 10 Wolpe, J.: The Systematic Desensitization Treatment of Neuroses, J Nerv Ment Dis 132:189–203 (March) 1961. 11 Leitenberg, H., et al: The Contribution of Selective Positive Reinforcement and Therapeutic Instructions to Systematic Desensitization Therapy, J Abnorm Psychol to be published 1969. 12 Wilder, J.: Facts and Figures on Psychotherapy, J Clin Psychopath 7:311–347 (Oct) 1945.

Article 4

Behavioral Approaches to Anxiety Disorders: A Report on the NIMH-SUNY, Albany, Research Conference David H. Barlow and Barry E. Wolfe State University of New York at Albany and National Institute of Mental Health

Recent years have produced striking advancements in the discovery of effective psychologically based treatments for some anxiety disorders. If this progress continues during the next decade, our understanding of these disorders may be increased to the point where truly effective treatments and/or preventive efforts would be available for all anxiety disorders. Nevertheless at this juncture there is sizable disagreement among leading clinical investigators on how to proceed with research in this area. Thus, a National Institute of Mental Health–sponsored conference of leading investigators was convened to recommend research strategies that will advance the field toward a truly cumulative body of knowledge. Recommendations were made in the areas of general research strategies, assessment and classification, process research, outcome research, delivery of treatment, and dissemination of results. A summary of these recommendations is presented. In recent years, significant advances have been achieved in the psychosocial treatment of the various anxiety disorders. Promising behaviorally oriented treatments have been developed, particularly for agoraphobia and obsessivecompulsive disorders, two of the most intractable behavioral disorders, To capitalize on the research momentum generated thus far, the National Institute of Mental Health (NIMH) awarded a contract to the Research Foundation of the State University of New York (SUNY) at Albany to organize a research conference on behavior therapy for the anxiety disorders. The chief objective of this conference was to bring together the leading and most active researchers in the area to confront the existing methodological and conceptual problems in the research on behavioral treatments for two classes of anxiety disorders: (a) obsessive-compulsive disorders and (b) anxiety and phobic disorders. After considering the state of the art in the treatment of these disorders and defining current research needs, the participants were charged with recommending specific research strategies for advancing knowledge in the field.

Behavioral Approaches to Anxiety Disorders  59

The conference was held at SUNY, Albany, on April 28–30, 1980. In addition to David H. Barlow and Edward B. Blanchard, the principal investigators of the NIMH contract, and Barry E. Wolfe, who represented the NIMH, 20 of the leading clinical researchers in the field participated. After the conference was completed, a detailed final report was prepared containing the recommendations on which a consensus was achieved (NIMH, Note I). A summary of that final report is presented below. State of the Art A consensus quickly emerged on the state-of-the-art issue with respect to the treatment of the anxiety disorders. It was agreed that exposure treatments produce the most consistent improvement, particularly when applied to agoraphobia and obsessive-compulsive disorders. Results from numerous studies indicate that approximately 65%–75% of those clinical phobics who complete treatment show substantial, clinically significant improvement from these treatments, with positive effects enduring through follow-up of 4 to 9 years. On the basis of somewhat less extensive data, it was estimated that approximately 60%–70% of those obsessive-compulsive patients who complete treatment significantly improve, particularly with respect to their ritualistic behavior. It was acknowledged, however, with regard to both disorders, that the extent of the improvement is not always optimal and that a significant minority of patients relapse. In addition, there are a substantial number of patients who either refuse treatment or do not complete it. Despite the moderate success of exposure treatments with agoraphobia and obsessive-compulsive disorders, it was further acknowledged that the effectiveness of psychosocial treatment with several other anxiety disorders is either unknown or undemonstrated. These disorders include obsessional states, generalized anxiety disorders, or panic disorders, and a broad range of problems referred to as social anxieties or social phobias. General Research Strategies Before the participants deliberated at length on the details of specific research strategies, they arrived at a number of general research strategies that served as major choice points for deciding what is to be studied. 1.

The first consensus recommendation was that at the present time, there is no further need for straightforward outcome studies on the effects of psychologically based treatment of clinical phobias. Psychological treatments, specifically performance-based exposure procedures, have been shown to be effective in the treatment of phobic disorders. This conclusion does not imply that performance-based exposure procedures should not, in some instances, be supplemented by other cognitive, behavioral, or pharmacological treatments, since the evidence is not yet clear on the value of these additions to this therapeutic package, but exposure-based procedures should be strongly considered as a basis for any therapeutic intervention.

60  David H. Barlow and Barry E. Wolfe

 

Behavioral Approaches to Anxiety Disorders  61

the possibility of formal multicenter trials might be entertained for the rarer anxiety disorders, such as obsessional thinking, an equally fruitful approach would be the creative use of single case experimental designs in each of a number of centers with encouragement of close communication among centers on basic assessment tools and their administration. Process or outcome research should report data, wherever possible, from individual cases as well as from the usual descriptive statistical reporting. This should include repeated measurement over the course of the therapeutic intervention and periodic measurement through long-term follow-up, with the emphasis on clinically significant change. Particular attention should be paid to abrupt changes during the course of any therapeutic intervention, such as sudden relapses or improvements, in the hope of identifying clues to the functional relationship of these changes to environmental events. 7. It was further recommended that the accumulation of useful knowledge would be significantly aided by a pattern of allocation of resources emphasizing long-term funding of productive groups of clinical investigators in this area. This type of funding would allow long-term commitment to the area of anxiety disorders as well as careful planning of sequential studies, with significantly greater amounts of attention paid to questions of maintenance and generalization of treatment effects. This pattern of long-term funding, along with continued close communication and occasional collaboration among centers, was seen as a better utilization of funds than the alternative strategy of funding 2- or 3-year projects with little or no continuity. Specific Research Strategies The conferees offered specific research strategies in four separate areas: (a) assessment and classification, (b) process research, (c) outcome research, and (d) delivery of treatment and dissemination of results. In most instances the specific strategies apply to all disorders. Where a particular strategy concerns a specific disorder, the disorder will be specified. Assessment and Classification

1. The majority of conferees agreed that all future research involving either process or outcome questions within the area of anxiety disorders should assess the problem in each of three response systems, subjective, behavioral, and physiological, unless the investigator demonstrates that the question to be asked is not relevant to all three response systems. High priority should be given to research on the accuracy (i.e., reliability and validity) of methods for assessing anxiety disorders in these three response systems. Since assessing anxiety disorders in all three response systems is not feasible in every research center at this time, particularly in view of the technology required for assessing physiological response systems,

62  David H. Barlow and Barry E. Wolfe

2.

3.

4.

5.

resources should be provided within all projects to cover the cost of this venture whether the cost incurred involves basic hardware or personnel. Technology is sufficiently advanced to provide this equipment at a reasonable cost to all clinical research centers that desire it. Efforts to standardize assessments in the three response systems across research centers should be supported and given high priority. Some preliminary agreement exists in the area of clinical phobias on various subjective, behavioral, and physiological indices of this problem, but considerably more research is needed on the development of accurate measures of other problems, such as obsessive-compulsive disorders, generalized anxiety disorders, and social anxieties and phobias. Improved methods of assessing associated social and occupational disruptions as well as assessing the relationship of mood disorders to anxiety disorders should be developed. Among the many avenues in need of exploration is the role of the family or other closely associated people in the assessment of these problems. This recommendation includes a much fuller description of those entering treatment on a variety of dimensions. Such dimensions include work and social adjustment, depression, and family and marital adjustment. Particular emphasis was placed on family and marital adjustment and mood disorders, since there is evidence that these two factors relate in an integral way to anxiety disorders. Current scales to rate marital distress or adjustment were considered inadequate, and it was recommended that new procedures or measurement devices with more reliability and validity be developed for this purpose. Greatly improved methods of assessing obsessional thinking that take into account both conceptual and emotional processes are needed. It was noted by the conferees that current methods of monitoring obsessional thinking and associated behavioral events provide but a beginning in this underdeveloped area of assessment. Moreover, it was noted that measures of compulsions, although somewhat more advanced than measures of obsessional states, are less well developed than behavioral measures of phobic responses. Direct measurement of frequency and/or duration of rituals, preferably in the natural environment was encouraged. Behavioral diaries may be particularly useful here as a supplemental measure of ritualistic behavior. Another supplemental measure would be observation of the disrupting effects of rituals on common daily activities. Greater effort should be expended on developing methods for increasing compliance and accuracy in self-monitoring procedures. Further research is needed on the reliability and validity of the classifications of generalized anxiety disorder and social phobia and the relationship of these classifications to agoraphobia and other clinical phobias. This is a critical area of research necessary for the successful differential diagnosis of these several anxiety disorders. Elucidation of this relationship is also necessary to determine if subsequent process and outcome research should deal with these disorders as distinct entities or extensions of other anxiety disorders.

Behavioral Approaches to Anxiety Disorders  63

The area of assessment research deserves high priority, since all other studies relating to outcome or basic mechanisms of action of any treatment depend on the adequacy of assessment technology. Levels of effort within this area will vary based on the specific recommendation proposed here. For example, one or two well-designed projects would be sufficient to develop further the necessary technology to carry out Recommendation 1. On the other hand, projects to promote a consistency of measurement operations among centers, as in Recommendation 2, or a determination of the reliability or validity of various classifications, such as generalized anxiety disorder, will have to be of sufficient scope that the results will be meaningful to all major centers working in the area. Process Research

Seven major recommendations relating to the advancement of process research were offered by the conferees. 1.

Within the general research plan, highest priority should be given to the process question, “Why do exposure-based treatments work?” This very broad question actually encompasses several subsidiary questions. First, there is a need to develop a heuristic theoretical framework for explicating the exposure-treatment process. A second issue concerns the specification of exactly what is “functional” exposure and whether cognitive and attentional avoidance are as important as behavioral avoidance. In view of the fact that exposure treatments are effective but often considerably less than totally effective, an important area of process research will be the determination of adjunctive factors that enhance the effectiveness of exposure treatment. 2. A second high priority process question involves the determination of the prevalence and implications of desynchrony in all of the anxiety disorders. Desynchrony refers to a lack of covariance among the various measures of anxiety taken across the three response systems. During any intervention, the extent of desynchrony should be explored as well as its implications for later outcomes and ultimate generalization among response systems. Although this effort is deemed feasible given the developments foreseen in the section on assessment and classification, it is likely to require considerable time and effort. The question may best be answered by looking at results across many research centers over a follow-up period of 2 to 4 years. 3. A considerable research effort is required for ascertaining cognitive changes that occur during any exposure procedure and the relationship of these changes to mood. This effort should precede or accompany attempts to determine the relative efficacy of cognitive therapies in conjunction with or compared with other therapeutic approaches in those anxiety disorders where this would be desirable. 4. Investigators are needed on the nature of panic, particularly within the context of agoraphobia. Such research would include the monitoring of panic attacks in three response systems over a period of time to determine the form, duration, and intensity of these attacks and their relationship to anxiety disorders.

64  David H. Barlow and Barry E. Wolfe

5. Another significant area of process research should focus on the elucidation of the interpersonal context of anxiety disorders, particularly the role that the family might play in the generation and maintenance of anxiety disorders. A special focus should be placed on research on the role of the family in generalizing and maintaining treatment effects. Research should also focus on their influence, positive or negative, when family members participate in the treatment process. 6. Still another important research question identified by conferees, but given somewhat less priority, concerns the further investigation of the interaction of pharmacological and exposure-based treatments in the varying anxiety disorders. Since the mechanisms of pharmacological treatments on anxiety disorders are not yet known, studies of this process in conjunction with exposure treatments might shed light on common or complementary mechanisms of action, which would then lead to further research on increased efficiency and efficacy of therapeutic interventions. 7. Finally, a recommendation was made for beginning efforts in the study of the natural history and epidemiology of anxiety disorders employing commonly used measures of anxiety collected in the general population. Data from these studies would lead ultimately to preventive efforts, but it is thought that our knowledge is too rudimentary at this time to carry out preventive efforts or to make funding of such efforts fruitful. These process recommendations were given such high priority that the conferees contended that no outcome research should be considered unless it includes significant process questions. The level of effort within this area should be the highest, yet the scope, in most instances, will not extend beyond the laboratories or clinics of individual groups of investigators. An exception would be research investigating desynchrony, which is probably best carried out in the context of outcome studies across many research grants. Outcome Research

The conferees generally agreed that outcome studies comparing exposure treatments to other forms of treatment or to no treatment should be accorded very low priority. Outcome studies that are conducted should contain significant process questions. Within this context, several areas were identified as being in need of further investigation using traditional outcome research strategies. 1. Research is needed on methods for reducing relapse and for increasing generalization of treatment effectiveness. Cost-effectiveness data relating to the treatments under study should also be collected. 2. A related recommendation was to encourage support for research carefully examining differences among those patients who succeed and those who fail, including not only those who relapse after treatment but also those who fail to complete treatment and those who refuse treatment at the outset. 3. There is a clear need for further studies on the separate and interactive effects of psychopharmacologic and exposure treatments.

Behavioral Approaches to Anxiety Disorders  65

4. In contrast to the area of clinical phobia, a clear priority exists for outcome research on the effects of various psychosocial treatments for generalized anxiety disorders and social phobias. Comparisons with both pharmacological treatments and with no treatment are needed in this area. 5. Outcome research is needed on the efficacy of psychosocial treatments for obsessional thinking. Because of the rarity of this disorder, the effects of various treatments could be explored using the research strategy of single case experimental designs. Although these recommendations within the category of outcome research should be accorded high priority, the overall research plan would call for integration of many of the process questions within the outcome research priorities. For instance, questions on the basic mechanisms of action of exposure treatments could be included within most of the outcome research categories, but particularly in the projects that are targeting individual differences in response to treatments as a major question. Similarly, questions on generalization and maintenance of treatments would include, at some point, studies of the role of the family in this process and the impact of treatment on the family and interpersonal system of the patient. In addition, any outcome studies comparing the efficacy of pharmacological and exposure treatments should also look at mechanisms of action during the process of treatment. Moreover, the issue of desynchrony also could be examined in the context of outcome research. Given the lengths of follow-up desirable and the necessity of examining such questions as generalization and maintenance inherent in a long-term effort, it is expected that this part of the research plan will occupy the better part of the next decade. Delivery of Treatment and Dissemination of Results

The majority of the conferees believed that our knowledge of the efficacy of exposure-based treatments was far enough advanced to recommend research designed to increase the efficiency and effectiveness of delivery of these treatments. Answers should be sought to the following four issues: (a) the relative effectiveness of nonprofessionals versus professionals participating in the various phases of the intervention effort, (b) the usefulness of self-help manuals combined with various levels of therapist intervention up to and including no therapist intervention, (c) factors increasing compliance with recommendations for treatment and particularly compliance with recommendations pertaining to maintenance and generalization of treatment after formal intervention is terminated, and (d) the cost-effectiveness of training various existing practitioners to carry out these programs versus the information of more-specialized “phobia clinics” in strategic geographical areas. In view of the fact that answers to these questions are likely to affect a large number of people, they are also afforded a high priority in the research plan. Moreover, since various forms of these delivery systems are already in effect in various parts of the country and in the world, it should be feasible to test existing forms of delivery as they now exist. Once again, in the interest of efficiency, significant process questions could be integrated into the testing of various forms of service delivery.

66  David H. Barlow and Barry E. Wolfe

Some Dissenting Positions The above-mentioned recommendations are those on which a consensus was achieved by the conferees. In this respect, they represent the best advice available for the field in terms of research directions and research questions that should be pursued. But the conference was much richer in its product, and the seeds sown during the meeting itself often came to fruition once it was over. On the strength of this premise, participants were asked, once they returned home, to provide a summary of their personal reactions to the conference and particularly their own research recommendations that may not have found their way into the consensus report. This assignment resulted in some additional research ideas but also some significant dissenting positions. Only some of the most striking dissents are presented here. Physiological Data

Although the consensus view was that data should be collected from the three response systems, cognitive, behavioral, and physiological, a minority of conferees questioned the relative usefulness of physiological data. Their view was that physiological data should not be collected in any routine or automatic way but rather the collection of such data should reflect the precise aspects of behavior that are expected to change or should be tied to an important theoretical question. Others felt that the yield obtained from physiological data does not justify the cost in time and expense. The unreliability of physiological data at present was cited by still others. Straightforward Outcome Studies on Clinical Phobias

Several conferees felt that we still have much to learn from straightforward outcome studies, even in the area of clinical phobias. One conferee felt that outcome studies were in fact premature, because the field lacks sufficient information describing the nature of the populations suffering from anxiety disorders or the nature of the specific anxiety disorder. These descriptive questions require answers before outcome studies should be encouraged. Another conferee saw little justification for supporting process research over outcome research or vice versa. “Let every flower bloom!” he suggested; “We do not know enough to establish restrictive priorities in the questions that need to be addressed and the manner in which they should be answered.” Exposure Treatments

A closer look at the improvement rates of exposure treatments with phobics suggested that the actual rate is closer to 49% than to the oft-quoted 75%. The lower figure takes into account those individuals who were unwilling or unable to complete treatment. A fairer appraisal of a treatment’s effectiveness, one participant argued, must include those patients who failed to complete treatment as well as completers.

Behavioral Approaches to Anxiety Disorders  67

Finally, let it suffice to mention a research recommendation suggested by one participant that was neither included in the consensus nor represents an actual dissenting position. It is just a very interesting suggestion. This participant thought that we could learn much from a comparison between individuals who function optimally under stress and those individuals who suffer from an anxiety disorder. Comparisons could be made on problem-solving styles, autonomic responsivity, and cognitive styles, among other variables. These dissenting positions suggest some of the richness of this conference, and they are presented to dissuade the reader from believing that complete unanimity was obtained with respect to the proposed directions of future research. Conclusions The NIMH-SUNY, Albany, conference proved to be a highly productive and rich-yielding conference, and it may be counted as successful in its effort to develop a broad but detailed research plan for advancing us toward the goal of developing efficient and effective psychosocial treatments of all of the anxiety disorders. The renewed focus on process questions and questions of the basic mechanisms of therapeutic action is timely. On the premise of adequate funding, all of the above recommendations seem to be quite within the field’s current conceptual and technical feasibility. One caveat was offered by the conferees regarding the successful implementation of these recommendations. Many of the questions defined cannot be answered in a short span of time, and it is anticipated that successful completion of the research plan will occupy the whole of the 1980s. It is the clear expectation of the conferees, however, that responses to the recommendations put forth above will go far in relieving the significant human suffering associated with anxiety disorders. On this prospect there was no dissent. Notes Editor’s Note. Although the present report of the Albany Research Conference deviates somewhat from the usual articles published in the Journal of Consulting and Clinical Psychology, it was the Editor’s opinion that it would be of interest to a large number of the readership. Thus, the manuscript was accepted for publication. Preparation of this article was supported in part by Contract RFP NIMH ER-79–003 from the National Institute of Mental Health. Conferees included W. Stewart Agras, Thomas Borkovec, John C. Boulougouris, Dianne L. Chambless, Paul M. G. Emmelkamp, Edna B. Foa, Michael G. Gelder, Marvin R. Goldfried, Alan J. Goldstein, R. Julian Hafner, Richard S. Hallam, Peter J. Lang, Harold Leitenberg, Issac Marks, Andrew Mathews, Matig Mavissakalian, Donald H. Meichenbaum, S. Rachman, G. Terence Wilson, and Charlotte M. Zitrin.

Reference Note National Institute of Mental Health. Final report of NIHM conference #REP NIMH ER-79–003, Behavior therapies in the treatment of anxiety disorders: Recommendations for strategies in treatment assessment research. Bethesda, Md.: National Institute of Mental Health, 1981.

Article 5

Behavioral Conception and Treatment of Panic David H. Barlow, Ph.D. Center for Stress and Anxiety Disorders. Department of Psychology. State University of New York at Albany

Behavior therapy “cut its teeth” on phobias. Since Wolpe’s (1958) pioneering book on the treatment of neurotic problems a variety of behavioral procedures have been tested and proven effective in the treatment of phobia. During the 1960’s and 1970’s scores of studies were performed isolating the effective ingredients of these treatments. The overriding conclusion that emerged from this systematic research during the 1970’s was that one particular procedure was common to all successful behavioral treatments for phobia (Mavissakalian & Barlow, 1981). This procedure, which came to be called exposure, consisted of motivating the patient to expose him or herself to a series of increasingly more difficult situations that provided the context for the phobic reactions and arranging conditions to facilitate this exposure. At present there are many psychologically based theories as to why this treatment is successful, including but not limited to extinction, habituation, emotional processing, and increases in perceived self-efficacy (Barlow & Mavissakalian, 1981; Barlow, in press). While none of these theories is fully supported as yet, there is clear evidence that exposure-based treatments are effective with phobias in general and agoraphobia with panic in particular. Some of these data are represented in Table 5.1 which summarizes the results of 24 controlled outcome studies testing exposure-based treatments for agoraphobia as reviewed by Jansson & Öst (1982). The data summarized in this review indicate that 70 percent of patients improved substantially with exposure-based treatments. The dropout rate averages 12 percent, although it is consistently higher for the type of exposure program which is administered in a very intensive fashion, for example, over a 2-week period, and much lower for programs administered in a more gradual self-paced manner (e.g., Barlow, O’Brien, & Last, 1984). At the same time this systematic research was ongoing, research from another perspective was isolating the phenomenon of panic. In 1959 when agoraphobics were still called schizophrenics by most clinicians because of their severe symptoms and marked social impairment, Donald Klein had a clinical bunch regarding the importance of panic. From this idea emerged a progression of descriptive and pharmacological research identifying panic as a central component in many of the anxiety disorders. This, in turn, led to revisions in the Diagnostic and Statistical Manual of Mental Disorders

Behavioral Conception and Treatment of Panic  69

(DSM-III) designating panic disorder as a separate diagnostic entity (American Psychiatric Association, 1980). Now the evidence is strong from both pharmacological and behavioral investigators that agoraphobias with panic and panic disorder do not differ in any substantial manner with the exception of the development of avoidance behavior as a complication in some, but not in all, people who initially suffer from unexpected panic attacks (Barlow, 1986). If panic is indeed the central feature of agoraphobia, two questions arise. Why are exposure-based treatments as successful as they are, and what happens to panic attacks during exposure? Because of a somewhat different conceptualization of this disorder, behavior therapists very seldom measured panic in past studies, although clinical reports would indicate that panic attacks did indeed decrease during successful exposure-based treatments. Nevertheless, more recently, clinical researchers investigating exposure-based treatments have measured panic more systematically, and the results are somewhat surprising. For example, Chambless, Goldstein, Gallagher, et al. (in press) assessed panic and avoidance behavior in a series of 35 patients receiving exposurebased treatments who were medication free at all times. They chose patients from a larger series who were medication free in order to observe more systemically the effects of exposure on panic. While this was not a controlled experimental study, the results, which are presented in Table 5.2, indicate that almost two-thirds of these 35 patients were free of panic attacks for at least a 1-week period at the 6-month follow-up. In fact, the results with panic seem to be somewhat better than the results for avoidance behavior where only 29 percent of the patients were free of avoidance. However, avoidance behavior was measured by a very conservative questionnaire which probably overestimates phobic avoidance. In any case, it seems that exposure-based treatments are consistently effective in reducing panic attack when this is measured, and that this is a very important result in exposure-based treatments. For example, Barlow et al. (1984), in a study examining the effectiveness of adding the spouse to exposure-based treatments, discovered that among those who responded to the treatment, significant reductions occurred in a composite measure of panic combining frequency, duration, and intensity of attacks. But for those who were judged to be nonresponders on a broadly based criteria, this measure of panic actually showed increases. One-third of the responders in this Table 5.1 Outcome of exposure treatments for agoraphobia based on 24 studies at follow-up of at least 5 months (range 5 months to 5 years)

Global Ethical Ratings (%) Improved (Median)

Range

Dropouts (Median)

Range

70

58–83

12

0–35

Adapted from Jansson & Öst, 1982.

70  David H. Barlow Table 5.2 Outcome of behavioral treatment of agoraphobia on avoidance and panic for 35 medication free patients at 6 months

Panic Attacks Yes 36%

Avoidance While Alone (as compared to normals) No 64%

Yes 71%

No 29%

Adapted from Chambless et al., in press.

study reported no panic attacks whatsoever after treatment. These results are presented in Table 5.3. These results have now been replicated in several other studies examining exposure-based treatments (Michelson, Mavissakalian, & Marchione, 1985: Ghosh & Marks, in press). Why do exposure-based treatments result in improvements in panic? To explain this one has to combine one old idea and one relatively new idea into a behavioral conceptualization of panic. This conceptualization leads to new behavioral treatments directed specifically at panic attacks which are in an early stage of development. Retrospective survey evidence strongly suggests that the first unexpected panic occurs in the context of some marked life stress (Roth, 1960; Doctor, 1982; Last, Barlow, & O’Brien, 1984). It now seems that the occurrence of panic attacks in the population is far more widespread then we expected. For example, Norton (1985) found that almost 35 percent of the population reported one or more panics in the past year as indicated in Table 5.4. Norton has now substantiated these data in a replication study and other data from around the world are beginning to appear supporting this finding (Klerman, 1985). If this is true, then one fascinating question concerns the differences between nonclinical people having occasional unexpected panics vs. those who end up in clinics seeking help for this problem. This latter group seems to comprise less than 5 percent of the population (Myers, Weissman, Tischler, et al., 1984). One possibility is that the latter group learns to fear the possibility of having another panic attack, a condition which seems to be present in almost all patients with panic disorder presenting at a clinic (Barlow, Vermilyea, Blanchard, et al., 1985). That is, some learning takes place in that the recent panicker acquires an extreme sensitivity to somatic events because of the marked fear that panic may reoccur. The occurrence of slight physiological sensations, even if for perfectly normal reasons, may then evoke marked fear in the patient. This fear activates the autonomic nervous system and increases the intensity of somatic sensations resulting in an increasing spiral of fear with panic at the top of the spiral. The marked fear concerning the possibility of having another panic attack seems particularly likely to develop when no ready explanation or cause is available to which to attribute the initial panic attack. Thus, patients develop a “phobia” of internal physical sensations that may represent the beginnings of another panic attack. This phobia increases anticipatory anxiety, as well as a variety of somatic events, which,

Behavioral Conception and Treatment of Panic  71 Table 5.3 Ratings of frequency duration and intensity of panic attacks for prior three days after couples treatment

Responders N = 18 Nonresponders N = 10

Pre

Post

3.22

2.27

3.30

4.13

p < .05

N = 6 Report no panic Adapted From Barlow, O’Brien, & Last, 1984.

Table 5.4 Panic in the normal population (n = 186)

One or More Panics in Last Year

34.4%

1–2 3–4 5+

17.2% 11.3% 6.0%

1 2 3+

17.2% 4.8% 2.1%

Panic Frequency Last Year

Panic Frequency Last 3 Weeks

Adapted from Norton et al., 1986.

paradoxically, insure that future panic attacks will occur in an ever-spiralling pattern. The behavioral conception of panic, then, involves a psychobiological process wherein persons susceptible to anxiety experience an initial panic attack while under stress and subsequently develop a fear of certain patterns of somatic sensations that might represent the beginnings of another panic through a process called interoceptive conditioning. The four steps involved in this process, whether in the laboratory or in the natural environment, are presented below. 1. 2. 3. 4.

The elicitation of physiological sensations. The strength of the subjects’ previously learned association between physiological sensations and psychological feelings of acute anxiety. The degree of current (baseline) anxiety. The degree of anxiety elicited by the experimental setting.

This, in fact, is the old idea referred to above since this pathway to panic was first published in 1964 (Breggin, 1964) as what was considered the best explanation for a long series of panic provocation studies in the laboratory using epinephrine.

72  David H. Barlow

Furthermore, this conceptualization of panic is very different from the “nonspecific stress” hypothesis often mentioned as a cause of laboratory panic. Interoceptive conditioning suggests that patients become acutely sensitized to specific patterns of internal physiological responses rather than to generalized arousal, and this seems to be well-illustrated in a very early study where panic was experimentally provoked in the laboratory (Lindemann & Finesinger, 1938). In this study 20 patients were administered either adrenaline or mecholyl, a parasympathetic stimulant with a very different pattern of physiological activation in a counterbalanced fashion. This was an extremely well-done study for the time. For example, verbatim accounts of panic attacks are provided which should convince even the most skeptical critic that what was being studied in 1938 was the same phenomenon of panic with which we are dealing today. Six of the patients panicked during adrenaline but not during mecholyl. Five panicked during mecholyl but not during adrenaline. Five panicked both during mecholyl and adrenaline, while four panicked under neither condition. Although studies such as these will have to be replicated using more current methodology, the results indicate that patients seem to learn to fear specific patterns of physiological responses. Based on this conceptualization, new behavioral methods for treating panic directly have been devised in our center and elsewhere. What these methods have in common is that the important cues for exposure in patients with panic are not external cues found in shopping malls or any location away from a safe place or person, but rather the interoceptive cues that they have come to fear. Treatment then revolves around reproducing the physiological sensations that are most feared by individual patients in order to produce extinction or emotional processing of fear cues. Types of procedures that should be effective in this regard are listed in Table 5.5 and preliminary evidence exists for the effectiveness of some of them (Barlow, Cohen, Waddell, et al., 1984; Waddell, Barlow, & O’Brien, 1984: Clark, Salkovskis, & Chalkley, 1985; Orwin, 1973; Beck & Emory, 1985). In our clinic, for example, we assess for patterns of fear by having patients engage in a variety of exercises designed to produce different patterns of physiological symptoms. To activate cardiovascular symptoms we use exercise. For respiratory symptoms we induce voluntary hyperventilation. For audiovestibular systems we induce dizziness. When feared patterns are identified, these become part of treatment. Clark et al. (1985) treated 18 patients identified as sensitive to the effects of hyperventilation. Eleven were agoraphobic while seven were panicking without avoidance behavior. Panics were essentially eliminated in all patients as represented in Figure 5.1 through repeated exposure to the effects of hyperventilation combined with other behavioral procedures. Barlow et al. (1984) treated patients with panic disorder in a small controlled pilot study and found significant reductions on panic compared to the control group. Paradoxically, some of the effectiveness of exposure with panic may result from repeated evocation of panic in a benign context with resulting extinction or cognitive reappraisal of the consequences of panic. While current evidence is mostly at the level of open clinical trials, preliminary results seem promising and controlled experimentation will be forthcoming.

Behavioral Conception and Treatment of Panic  73 Table 5.5 Psychological treatment for panic

1) Exercise 2) Breathing Hyperventilation 3) CO2 Inhalation 4) Imagined Physio Cues 5) Cognitive Reattribution 6) Relaxation

Mean number of panic attacks per week

12

Baseline Respiratory control

Situationals Non-situationals

10 8 6 N = 11

4 2 0 8

N=7

6 4 2 0 –3 –2 –1

1

2

3

4

5

6

7

Weeks

8

9

10 11 12 13

6 24 Months

Figure 5.1 Effect of Respiratory Control on Panic Attacks. From Clark, D.M., Salkovskis, P.M., and Chalkley, A.J. Respiratory control as a treatment for panic attacks. Journal of Behavior Therapy and Experimental Psychiatry. 16:23–30, 1985. Permission granted to reproduce this figure.

References American Psychiatric Association. DSM-III; Diagnostic and Statistical Manual of Mental Disorders. 3d ed. Washington, DC: The Association, 1980. Barlow, D.H. In defense of panic disorder with agoraphobia and the behavioral treatment of panic: A comment on Kleiner. The Behav. Therapist, 9:99–100, 1986. Barlow, D.H. (ed.), The Anxiety Disorders. New York: The Guilford Press. (In press.) Barlow, D.H., Cohen, A.S., Waddell. M., et al. Panic and generalized anxiety disorders: Nature and treatment. Behav. Ther., 15:431–449, 1984. Barlow. D.H., O’Brien, G.T., and Last, C.G. Couples treatment of agoraphobia. Behav. Ther., 15:41–58, 1984.

74  David H. Barlow Barlow, D.H., Vermilyea, J., Blanchard. E.B., et al. The phenomenon of panic. J. Abnormal Psych., 94:320–328, 1985. Beck, A.T., and Emory, G. Anxiety Disorders and Phobias: A Cognitive Perspective. New York: Basic Books, 1985. Breggin, P.R. The psychophysiology of anxiety with a review of the literature concerning adrenaline. J. Nerv. Ment. Dis., 139:558–568, 1964. Chambless, D.L., Goldstein, A.J., Gallagher, R., et al. Integrating behavior therapy and psychotherapy in the treatment of agoraphobia. Psychotherapy. (In press.) Clark, D.M., Salkovskis, P.M., and Chalkley. A.J. Respiratory control as a treatment for panic attacks. J. Behav. Ther. & Exp. Psychiat., 16:23–30, 1985. Doctor, R.M. Major results of a large-male pretreatment survey of agoraphobics. In: DuPont, R.L. (ed.), Phobia: A Comprehensive Summary of Modern Treatments. New York: Brunner/Mazel, Inc., 1982, pp. 203–230. Ghosh, A., and Marks, L.M. Self-treatment of agoraphobia by exposure. Behav. Ther. (In press.) Jansson, L., and Öst, L.G. Behavioral treatment for agoraphobia. An evaluative review. Clin. Psychol. Rev., 2:311–336, 1982. Klerman, G.L. Personal communication, (1985). Last, C.G., Barlow, D.H., and O’Brien, G.T. Precipitants of agoraphobia: Role of stressful life events. Psych. Reports, 54:567–570, 1984. Lindemann, E., and Finesinger, J. E. The effect of adrenaline and mecholyl in states of anxiety in psychoneurotic patients. Am· J. Psychiat., 95:353–370, 1938. Mavissakalian, M., and Barlow, D.H. (eds.). Phobia: Psychological and Pharmacological Treatment. New York: Guilford Press, 1981. Michelson, L., Mavissakalian, M., and Marchione, K. Cognitive and behavioral treatments of agoraphobia: Clinical, behavioral, and psychophysiological outcomes. J. Consult. Clin. Psychol., 53:913–925, 1985. Myers, J.K., Weissman, M.M., Tischler, C.E., et al. Six-month prevalence of psychiatric disorders in three communities. Arch. of Gen. Psychiat., 41:959–967, 1984. Norton, G.R., Dorward. J., and Cox, B.J. Factors associated with panic attacks in non-clinical subjects. Behav. Ther., 17:239–252, 1986. Norton, G.R., Harrison, B., Hauch, J., and Rhodes, L. Characteristics of people with infrequent panic attacks. J. Abnorm. Psychol., 94:216–221, 1985. Orwin, A. ‘The running treatment’: A preliminary communication on a new use for an old therapy (physical activity) in the agoraphobic syndrome. Brit. J. Psychiat., 122:175–179, 1973. Roth, M. The phobic anxiety-depersonalization syndrome and some general aetiological problems in psychiatry. J. of Neuropsychiat., 306:293–306, 1960. Waddell, M.T., Barlow, D.H., and O’Brien, G.T. A preliminary investigation of cognitive and relaxation treatment of panic disorder: Effects on intense anxiety vs. ‘background’ anxiety. Behav. Res. Ther., 22:393–402, 1984. Wolpe, J. (ed.), Psychotherapy by Reciprocal inhibition. Stanford, CA: Stanford University Press, 1958.

Article 6

The Process of Fear and Anxiety Reduction: Affective Therapy David H. Barlow Center for Stress and Anxiety Disorders. Department of Psychology. State University of New York at Albany Now that models of the process and etiology of panic and anxiety have been set forth, it is possible to examine the implications of these models for treatment. The enormous strides made in recent years in developing psychological and pharmacological treatments for anxiety have not, for the most part, been guided by our knowledge of emotions. Rather, most treatments have been developed serendipitously. An example cited earlier was the discovery of drug treatments for panic. Other treatments developed as outgrowths of general schools of psychotherapy or behavior therapy; in these cases, more attention was often devoted to the therapeutic approach than to the disorder in question. Of course, influential theorists and therapists such as Wolpe and Beck have made important contributions to our knowledge of emotion therapy and theory, since their therapeutic approaches were developed in the context of emotional disorders. For this reason, the major psychotherapeutic approaches to emotional disorders that have received empirical support are outgrowths of procedures developed by pioneers such as Wolpe and Beck. Nevertheless, our knowledge of the mechanisms of action involved in the modification of intense maladaptive emotional responding is rudimentary at best. In fact, in the entire area of psychotherapy, much is made of the importance of emotional expression, whether the disorders are emotional disorders or not. And yet few approaches are guided by a knowledge of emotion theory and fact. Greenberg and Safran (1987) have begun to rectify this deficit with an important and scholarly book. But their emphasis is not on emotional disorders, or on the modification of maladaptive emotions. The purpose of this article is to survey, briefly and critically, the strengths and weaknesses of current theories of fear or anxiety reduction. The survey is limited to theoretical conceptions that have attempted to account for the process of fear or anxiety reduction in general as it occurs in a number of different therapeutic techniques. Implications for the modification of affect from emotion theory are suggested. The chapter concludes with a general therapeutic approach to fear and anxiety reduction, based on the models of panic and anxiety developed earlier. Since anxiety and fear have not been distinguished in most models of therapeutic change (see

76  David H. Barlow

Chapter 1), I refer to the techniques surveyed in this article as “anxiety reduction procedures.” Differential therapeutic targets within fear (panic) and anxious apprehension based on the models described earlier are specified below. Theories of Fear and Anxiety Reduction Most of the development and evaluation of treatments for anxiety have taken place in the context of phobic disorders. An integral part of anxiety reduction procedures when treating phobic disorders is the process of exposure to fearful or anxiety-producing situations. A major difference among psychosocial treatments is the manner in which exposure is arranged, and the additional procedures (strong suggestion, relaxation, the presence of a therapist, modeling, etc.) that are implemented to facilitate exposure. But exposure is not a theory of therapeutic action; it is simply an observation of a common procedure in many treatments. While some in the past have treated exposure as an explanatory concept for anxiety reduction, objections arose to this atheoretical stance very early. For example, Rosenthal and Bandura (1978) criticized this explanation as deficient on several accounts. They pointed out that patients with similar intensities of anxiety often respond very differently to identical amounts of exposure to their feared objects or situations. They suggested that exposure does not take into account the information providing properties of the therapeutic situation, which may differ substantially from patient to patient. For example, watching an agoraphobic model enter a feared setting such as a department store and then faint may have an effect on the patient that is quite different from the effect of watching the same model enter the same department store and not faint. Nevertheless, the amount of actual exposure to the model in the department store is exactly equal. Wolpe (1976) also pointed out that most exposure-based treatments contain elements that are likely to inhibit anxiety; examples include the presence of a therapist or some other significant person or instructions to relax. Such elements, he argued, are more critical than exposure per se. This notion is difficult to untangle from the alternative explanation that these elements (e.g., presence of a therapist) merely serve to increase the probability that exposure rather than escape or avoidance will occur, therefore facilitating anxiety reduction. Some have even suggested that exposure does not necessarily play a major role in the process of anxiety reduction, and have cited other procedures as more essential. For example, de Silva and Rachman (1981) have pointed out several instances in which exposure does not seem necessary to explain anxiety reduction. At present, there is no clear answer to the question of why some people experience reductions in anxiety during therapy. At least five distinct theories have evolved to explain the general process of anxiety reduction. The theory of anxiety reduction underlying pure exposure is usually based on one of two closely related processes: habituation or extinction.

The Process of Fear and Anxiety Reduction  77

Habituation As early as 1966, Lader and Wing first proposed that what was then the most popular anxiety reduction procedure, systematic desensitization, could be accounted for through the process of habituation. “Habituation” refers to a decline in fearful reactions, particularly psychophysiological aspects of fearful reactions, with repetitive exposure to fear-provoking stimulation. The literature on habituation, although it has developed independently, resembles the literature on systematic desensitization as originated by Wolpe (1958) in many ways. For example, habituation occurs more readily if the subject’s baseline level of psychophysiological arousal is low. A corollary to this finding is that habituation will not occur if baseline arousal is too high. In this case, further increases in arousal, or “sensitization,” will occur under some conditions. Habituation, in its simplistic laboratory model, is also usually considered to be short-term, with physiological responses returning after a suitable rest period (Lader & Wing, 1966; Thompson & Spencer, 1966). This feature of habituation was overlooked by many early theorists, although it is also clear that habituation can occasionally be longer-lasting (Watts, 1979). The primary dependent variables in classical studies of habituation, as mentioned above, are psychophysiological measures such as galvanic skin response (GSR) during orienting responses. In the context of anxiety reduction, attempts to correlate individual rates of habituation to experimental stimuli (such as auditory tones) with rate of anxiety reduction are often not successful (e.g., Gillian & Rachman, 1974; Klorman, 1974). In other words, physiological measures may decrease, but other indices of anxiety, including subjective reports, do not change. There are several problems with habituation as a satisfactory theory of anxiety reduction. To take one example, most studies describing the treatment of agoraphobia have utilized prolonged direct in vivo exposure. Among the consequences of prolonged exposure are maximum anxiety and arousal, as, for example, would occur during a panic attack. A strict adherence to habituation theory would predict that sensitization would occur under these conditions. And yet, according to all the evidence to date, this does not seem to happen. Recently, habituation theories have been extended in such a way that decrements in response as a function of massed, high-intensity, long-term exposure can be accommodated. These “dual-process” theories describe a complex interaction between habituation and sensitization, in which prolonged high-intensity arousal will eventually reduce sensitization, allowing habituation to occur (Groves & Thompson, 1970; Watts, 1979). There is some emerging evidence from the animal laboratories that different neural processes underlie these two phenomena (Davis, Parisi, Gendelman, Tischler, & Kekne, 1982). Now, evidence emerging from studies examining an important new aspect of the anxiety reduction process also raises questions about habituation (as well as other theories) as a satisfactory explanation. After successful anxiety reduction, either in the laboratory or in the natural environment, a number of people—perhaps as many as 50%—demonstrate a return of anxiety as measured by self-report. Rachman and his colleagues have been studying this

78  David H. Barlow

phenomenon for years and refer to it as “return of fear” (Craske & Rachman, 1987; Grey, Rachman, & Sartory, 1981). What they have found, in general, is that subjects who later evidenced a return of fear showed the same amount of fear reduction on all measures of anxiety after treatment as those subjects who did not evidence a return of fear. The one consistent predictor of return of reports of fear appeared to be elevated heart rate measured before treatment. It is important to note that heart rate in both experiments was not assessed during treatment itself or in a fearful situation. Thus, heart rate was probably reflecting “anticipatory” anxiety. On the surface, it might seem that these data support habituation theory, which would predict that those subjects with higher physiological arousal to begin with might not “habituate” as quickly. However, the results differ on several counts from the pattern of results that would be predicted by habituation theory. First, habituation theory would predict that subjects with high heart rate would not show the same reduction in physiological measures of anxiety as those with low heart rate, and yet the results clearly indicate that they did show the same reductions in these measures during treatment. Second, when subjects have shown substantial anxiety reduction on both physiological and subjective measures, habituation theory would predict a return of fear on both measures after a suitable interval. But return-of-fear studies have found increases only in subjective measures. Craske and Rachman (1987) speculate that the explanation for the return-offear phenomenon may be found largely in the period after the fear has been reduced. They suggest, for example, that some form of disturbance of consolidation of information occurs at this time. Information-processing accounts of anxiety reduction are discussed further below. Extinction A similar process often cited as underlying the effects of exposure is extinction. “Extinction” refers to decrements in the strength of learned responses through repetition of unreinforced responding. In the context of anxiety reduction, this rather technical explanation means that subjects repeatedly encounter feared situations without experiencing the fearful consequences, whatever they may be. This is derived directly from laboratory work on classical fear conditioning. While the literature on extinction is broad and deep, most of the basic experimental work on extinction of fear has concerned itself with avoidance behavior as the primary dependent variable, as contrasted to the work on habituation, in which physiological indices of fear have been highlighted. But there are some subtle differences between habituation and extinction that may or may not be relevant to theories of anxiety reduction (Rachman, 1978). Basically, “habituation” refers to decrements in the strength of unlearned responses, probably controlled at a relatively low neurological level such as the brain stem (Davis et al., 1982; Groves & Lynch, 1972). “Extinction” refers to decrements in the strength of learned responses. Also, habituation is often thought to be temporary, with the return of response (usually GSR in humans) occurring after the habituation process has been discontinued for a suitable time. In other words, nothing is really “learned.” Extinction is considered

The Process of Fear and Anxiety Reduction  79

more permanent. Procedurally, repetitive exposure to the feared situation is sufficient for habituation to occur, although other factors (such as initially low psychophysiological activity) are desirable. In the extinction process, exposure must occur in the absence of the aversive consequence. More recently, procedural differences between habituation and extinction have been questioned, because it seems that both may be permanent under certain conditions such as habituation to a novel stimulus, as in the orienting response (Watts, 1979). But the notion that something is learned during the process is clearly associated with extinction and not habituation. Extinction has proved to be by far the most popular theory of anxiety reduction over the years, although, in fact, little work has appeared directly testing the adequacy of this theory in clinical populations. Nevertheless, an extinction-based model of anxiety reduction has several strengths. First, there is an extensive series of studies (reviewed in subsequent chapters) indicating that long-term intensive exposure, in which phobics are required to remain in the feared situation until anxiety decreases, is more effective than shorter or more widely spaced periods of exposure (e.g., Chaplin & Levine, 1981; Marshall, 1985; Stern & Marks, 1973). According to extinction theory, patients presumably learn, by staying in the situation until anxiety decreases substantially, that the feared consequences do not occur. Implicit in this explanation is the idea that new learning takes place—an idea that is also found in the extensive literature on extinction from the animal laboratories (e.g., Rescorla, 1979). The traditional account of precisely what is learned, stemming from Pavlov’s (1927) original work, is that avoidance responding is associated with a process of inhibition after repeated unreinforced trials. While there is some disagreement over exactly what is learned, there is agreement that extinction is not a passive process, and therefore that it has something in common with more cognitive models of anxiety reduction that stress information processing. In fact, some basic researchers are beginning to speculate about cognitive processes in animals to account more adequately for what is learned during extinction (Hulse, Fowler, & Honig, 1978), while other have speculated on similar cognitive processes in humans (Martin & Levey, 1985; Reiss, 1980). In 1973, Seligman and Johnston suggested that what is learned during the extinction of avoidance behavior associated with anxiety is discontinuation of outcome expectations. Outcome expectation is a cognitive construct that may be influenced by many factors in addition to length of exposure. In that sense, it is closely related to self-efficacy, another cognitive construct that is taken up below. In this regard, one could also postulate that what is learned during extinction in clinical populations is enhanced self-efficacy. But there are also pieces of evidence that undermine extinction theory. Foremost among these is the conflicting series of studies (described more fully in Chapter 11) demonstrating that anxiety reduction can occur even though subjects are allowed to escape from the feared situation early in the “trial” before reaching maximum anxiety (e.g., Agras et al., 1968; Emmelkamp, 1982; Rachman, Craske, Tallman, & Solyom, 1986). If subjects do not have a chance to “learn” that there are no untoward or aversive consequences to remaining in the feared situation, then it is not clear how classical extinction can occur.

80  David H. Barlow

Also problematic for extinction theory are some of the early studies utilizing implosion, in which the most terrible consequences of the feared situation were presented (at least symbolically), and yet anxiety decreased (Stampfl & Lewis, 1968). This outcome contradicts the basic premise of extinction, which is stimulus presentation without reinforcement (aversive consequences). Of course, it is possible that symbolic or imaginal consequences do not assume the same importance as real consequences. For example, imagining that one is fainting or losing control is obviously not the same as actually experiencing it. On the other hand, few if any agoraphobics ever really faint or lose control during their ventures out; most only imagine they are about to do so. Thus, the problems for extinction theory remain. One possible path out of this morass of conflicting evidence is to conclude that the feared consequences have been improperly conceptualized to begin with. It is possible—and indeed likely, according to the models of panic and anxiety presented in Chapters 6 and 7—that the feared consequence in panic disorder and perhaps other phobic disorders as well is the perceived effect from somatic cues previously associated with alarms. An early study by Watson and Marks (1971) provided intriguing hints that this is the case. In this study, phobics showed equivalent anxiety reduction from relevant imagery specific to their phobic condition and from irrelevant imagery (being eaten by tigers). But the imagery in both cases resulted in equal exposure to cues associated with increased arousal. Therefore, it is possible that patients “learned” that there is little to fear from this increased arousal. Nevertheless, findings from the series of studies mentioned above, demonstrating that escape or the experience of less than maximum arousal is as effective as intense exposure with maximum arousal, are still problematic for a theory of extinction based on exposure to arousal cues. Presumably, one should confront maximal arousal cues to “learn” that the consequences of arousal (e.g., dying or losing control in the case of panic patients) are not there. The fact that subjects do not seem to have to experience maximal arousal, during which the feared consequences are most likely to occur, again brings into question the adequacy of extinction accounts of fear reduction. Toughening Up A biological model of fear reduction exists that closely parallels habituation and extinction. In 1976, Weiss, Glazer, and Pohorecky suggested a biochemical explanation for the phenomenon of learned helplessness (referred to briefly in Chapter 7), based on changes in noradrenaline levels. In 1982, Weiss and colleagues elaborated on the specific biochemical mechanisms of action associated with these changes in noradrenaline levels. In this revised view, chronic stress, as simulated by exposure to repeated electric shocks in animals, results in a rise in tyrosine hydroxylase activity (an enzyme associated with noradrenaline synthesis). This, in turn, facilitates increased noradrenaline levels in the locus ceruleus. The important step in this process then follows: The increased noradrenaline causes a rise in autoreceptor-mediated inhibition in the locus ceruleus, and therefore a drop in noradrenergic transmission in the forebrain. To simplify this a bit, brief exposure to stressful or fearful stimuli will result

The Process of Fear and Anxiety Reduction  81

in overall increases in noradrenaline, whereas more chronic exposure to the stimuli will eventually result in overall decreases in noradrenaline. These biochemical changes are associated with a variety of behavioral effects in animals (Gray, 1985). Generally, these effects can be described under the heading of “resistance” or “toughness.” Specifically, animals will show greater tolerance for aversive stimuli; hence the name “toughening up.” In a fascinating preliminary experiment on the process of toughening up, Holt and Gray (1983) seemed to demonstrate that this result could be produced biologically as well as behaviorally. They electrically stimulated the septal area of the brain in rats. This process produced results that seemed to mimic the effects of toughening up. Specifically, animals with this stimulation also showed increased resistance to a variety of stimuli that would normally be considered anxiogenic. The data on toughening up demonstrate once again the influence of learning and experience on neurobiology, and the inevitable interaction between these two systems (e.g., Bandura et al., 1985). That is, the behavioral process of repeated exposure is associated with biochemical changes, which in turn are associated with further behavioral changes. Of course, it seems to be necessary for some “learning” to occur at each stage. Specifying and isolating a specific biochemical mechanism underlying the effects of exposure also have the potential advantage of explaining differential results of exposure-based treatments among patients with seemingly similar levels of anxiety. It is possible that individual differences at a neurobiological level result in different rates of toughening up. For example, some individuals may be unable to reach the same levels of receptor desensitization that are quickly reached by others. While this result is certainly what would be expected, further evaluation would be required to see whether the individual differences are substantial enough to account at least partly for the wide intersubject variability in response to exposure. It is also interesting, as noted in Chapters 5 and 11, that the mechanisms of action of some successful drug treatments for anxiety seem to involve roughly similar mechanisms of receptor desensitization and alteration of noradrenaline levels (e.g., Charney & Heninger, 1985). On the other hand, it would seem a bit naive to take the reductionistic view that all manifestations of anxiety—behavioral, physiological, and subjective—can be accounted for by alterations in levels of noradrenaline or related monoaminergic activity. In view of the enormous influences of learning and other environmental and cognitive events on anxiety, as reviewed in Chapter 5, the relative level of noradrenaline is most likely only one of many influences on the process of anxiety reduction. To take just one example, noradrenergic depletion has been cited as a possible neurobiological process associated with depression, as described in Chapter 7 (Gray, 1985; Schildkraut, 1965). Assuming that this is true, what accounts for positive outcomes associated with noradrenergic depletion in some patients, as seems to happen during exposure therapy, and negative outcomes (depression) in other contexts? In animal laboratories, the difference seems to be whether the encounter with the aversive stimulus is brief or prolonged. If it is brief, noradrenergic synthesis is increased in a seemingly compensatory fashion (Weiss et al., 1982). One speculative possibility is

82  David H. Barlow

that when toughening up occurs in therapy, the patient is in a context where behaviors associated with control and perception of control are systematically learned over a relatively prolonged period. When the individual is not in therapy, the same buffeting with aversive stimuli and the resulting noradrenergic depletion quickly result in depression. Of course, one still has to explain the initial development of depression versus anxiety in etiological accounts. That is, why do some individuals become depressed and others anxious as the result of seemingly identical stress-related experiences? As suggested in Chapter 7, both the emotional disorder and the associated action tendencies, as well as (possibly) differing monoaminergic activity, may be a function of differential developmental experiences with stress and perceptions of control. In any case, the biological process of toughening up is closely related to habituation. Both “theories” have little theoretical “meat,” and are just descriptions of processes. Both toughening up and habituation predict a return of fear if contact with the fearful object or situation ceases for a period of time. In other words, there is no specific provision for new learning or alteration of response patterns per se. Of course, it is an easy enough matter to consider that something is learned during these periods of low responding. What may be learned is a response that competes with anxiety. The addition of learning to toughening up and habituation would make habituation, extinction, and the underlying biological process of toughening up very similar in many ways, although one would expect some disagreement among various theorists over what specifically is learned. Nevertheless, the biochemical data on toughening up, while seemingly compatible with extinction (if one is willing to consider that something may be learned during the process of toughening up), also cannot account for the series of studies mentioned above. These studies demonstrated that short-duration exposure with less than full evocation of arousal seems as effective as intense exposure. In order to accommodate these data, one must go beyond the usual types of simple learning associated with extinction and consider more complex explanations. In reviewing some of the literature on toughening up and the various parameters that affect this phenomenon, Gray (1985) raises an important issue with clinical implications. Specifically, Gray describes a number of animal experiments indicating that the administration of drugs with arousal-reducing properties interferes with the process of toughening up. Essentially, animals are more sensitive to anxiety-producing situations after the removal of anxiolytic drugs than they would be if they had never received drugs at all, despite similar amounts of exposure to these situations. This evidence suggests that administration of drugs in the context of clinical exposure-based treatments may short-circuit the biological toughening-up process, and therefore may interfere with short-term anxiety reduction. This anticipates several clinical studies described in Chapter 11, in which the administration of drugs seemed to interfere with exposure-based treatments. On the other hand, a number of additional studies reviewed in that chapter do not indicate that combined drug-exposure treatments produce weaker effects then exposure-based treatments administered without drugs. But clinical studies, of necessity, are rather broad-based in their strategies; this raises

The Process of Fear and Anxiety Reduction 83

the possibility that some detrimental effects of combined treatment may exist in some individuals that are masked or lost in the overall variance. For example, few of these studies systematically evaluated patients after withdrawal of drugs. Cleaner experimental analyses of the interactions of exposure and drug treatments will be required to determine whether anxiolytic drugs interfere with the biological bases of long-term anxiety reduction in human clinical subjects. Extinction, Habituation, and Toughening Up: Summary Neither extinction, habituation, nor the neurobiological process of toughening up seems fully satisfactory at this time as an explanation of fear or anxiety reduction. Most likely, clinical exposure-based treatment and its underlying theoretical rationales of habituation, extinction, or toughening up are oversimplified means of explaining the reduction of anxiety, although each may make some contribution. In addition to poor specification of what is “learned” during exposure, none of these theories easily accommodates the notion of anxiety as a construct of three relatively independent response systems. For example, for both agoraphobics and obsessive-compulsives, exposure and the prevention of avoidance or escape seem particularly effective in reducing avoidance behavior; however, the subjective experience of anxiety, and to some extent physiological indices, do not always respond as well (Mavissakalian & Barlow, 1981a; Rachman et al., 1979). Any comprehensive theory of anxiety reduction will require an explanation of the effects of successful clinical procedures on the integrated construct of anxiety. It is not clear that this is yet forthcoming. As noted above, it may be possible to “habituate” physiologically to a feared stimulus but still to report overwhelming fear. Another common observation is that someone will approach a feared situation despite physiological arousal and the subjective experience of fear; when this happens, it has been characterized as “bravery” or “courage” (Rachman, 1978). Similarly, it seems possible to eliminate avoidance behavior and even subjective reports of fear without resulting habituation of physiological responses. It is not clear that anxiety (in the sense that we understand it) is really occurring when physiological responses remain high, but when avoidance behavior and subjective experiences of anxiety are reduced or eliminated. In the model presented in Chapter 7, this may be a very normal and relatively unemotional response to an upcoming event reflecting a preparatory set. Finally, any theory of anxiety reduction—particularly theories postulating gradual decrements in anxiety, such as habituation, extinction, or toughening up—must consider panic. The introduction of panic into a consideration of theories of anxiety reduction is akin to introducing the phenomenon of mass extinctions into the orderly theory of evolution. Until several years ago, natural selection occurring in a very gradual fashion seemed adequate to account for the current status of living organisms on earth. But evidence of mass extinctions produced by cataclysmic upheavals has altered and diminished our view of the importance of natural selection. Now Rachman and Levitt (1985), in a study described in Chapter 4, have pointed out how the

84  David H. Barlow

occurrence of unexpected panic can disrupt the process of anxiety reduction in simple phobics. In their experiment, the occurrence of unexpected panic resulted in marked return of anxiety and greater escape from the feared situation subsequent to the panic than did the occurrence of expected panic. It is difficult for extinction and related theories to deal with unexpected panics unless one considers that unexpected panics are in fact the feared consequences of exposure to certain objects or situations, as suggested above. For successful extinction to occur, then, patients should repeatedly confront feared situations without panicking. As Rachman and Levitt (1985) have demonstrated, this is hard to arrange. A related possibility mentioned above is that panic itself with its associated intense arousal is the feared “situation,” and that the aversive “consequences” of panic are the experience of loss of control sensations. In this sense, extinction can only occur if an unexpected panic or its equivalent is presented repeatedly until one “toughens up” to it or perhaps learns that the expected but never-experienced consequences of losing control do not occur. However, the current weakness of this argument in regard to the effectiveness of exposure to less than full arousal has been reviewed above. Thus, one is left, once again, with the possibility that something more complex is learned during exposure than simply a disconfirmation of feared consequences while experiencing maximally fearful conditions. Self-Efficacy A theory of anxiety reduction that meets this requirement is receiving increased attention as well as some empirical verification. This theory is Bandura’s model of perceived self-efficacy as a determinent of fear and avoidance. Essentially, Bandura (1977a) theorized that increasing one’s sense of self-efficacy or competence in mastering a feared situation is the single result of all successful anxiety reduction treatments. He observed that direct behavioral experience and accomplishment probably constitute the most powerful means of increasing one’s sense of self-competence and efficacy, but that many other procedures may also contribute to this perception. These may include verbal instructions to the effect that one will be successful, as well as watching other people perform successfully. All of these various inputs then summate to produce self-efficacy. Early experiments with analogue subjects seemed to confirm the predictive power of self-efficacy (Bandura, 1977a). But it was not long before critics such as Borkovec (1978) observed that the theory is unparsimonious and circular. For example, Borkovec suggested that self-efficacy is more likely to be a reflection of behavioral change mechanisms than to be the mediator of such change. Teasdale (1978) also suggested that it may be impossible to separate self-efficacy from outcome expectations. One sign of the health of a theory is the amount of attention and controversy it generates, and in this regard self-efficacy theory is doing very well indeed. Periodically, Bandura (1983, 1984) answers accumulated criticisms. In the meantime, evidence continues to accumulate on the usefulness of this construct. For example, Williams and colleagues (Williams, Dooseman, & Kleifield, 1984; Williams, Turner, & Peer, 1985) have provided some evidence with simple phobics that perceived self-efficacy predicts therapeutic outcome

The Process of Fear and Anxiety Reduction  85

more accurately than does arousal during treatment, anticipated danger, or perceived danger. These clinical investigators suggest that constructing a treatment to target perceived self-efficacy by arranging conditions to instill a sense of mastery will be more effective than will simple exposure-based procedures without these conditions. They provide some evidence from work with their simple phobics that supports this notion. Lee (1984) also found that self-efficacy was a better predictor of outcome in an analogue experiment with snake phobics than were reports of outcome expectations. On the other hand, Biran and Wilson (1981) and others (e.g., Feltz, 1982; Kendrick, Craig, Lawson, & Davidson, 1982), working with simple phobias and performance anxiety, found much more modest correlations between self-efficacy and eventual outcome of treatment. Early evidence contradicting the predictive power of self-efficacy was provided by Rachman (1978, 1983a). Rachman observed that while self-confidence predicted fearless performance in combat in a number of U.S. veterans in the Pacific Theater during World War II, approximately 17% of these soldiers continued to experience intense fear reactions during combat, despite a strong sense of self-competence or self-efficacy preceding combat. Like other studies cited earlier, this would seem to reflect discordance between the subjective experience of anxiety (or expectation of this experience) and responses in other systems during stressful events. Thus, desynchrony and discordance again pose some difficulties for any one postulated mechanism of action underlying current fear reduction procedures. It is also possible, of course, that the soldiers were experiencing occasional panic attacks or true alarms during particularly difficult combat circumstances. These alarms might have been fully attributed by them to the circumstances in which they found themselves, such as a close brush with death. In this case, it is within the realm of possibility that these “expected” panics did not affect subsequent anticipatory anxiety or performance, much as in the Rachman and Levitt (1985) analogue study with claustrophobics. Of course, like other theories of anxiety reduction, self-efficacy theory has yet to attempt to accommodate the newly discovered functional relationship between panic and anxiety. Also, Craske and Rachman (1987), in a study cited above, noted that perceptions of self-efficacy bore little relationship to the return of subjective reports of performance fears in musicians. High heart rate prior to treatment was the only successful predictor of return of subjective fear. Wilson (1984) observes that self-efficacy theory is undergoing occasional elaborations as new data appear. For example, he notes after reviewing some of Lee’s work that accurate feedback during performance is critical to the formation of realistic and useful perceptions of self-efficacy. Furthermore, it is often overlooked that Bandura (1983, 1984) emphasizes that self-efficacy is predictive of behavior rather than intensity of anxiety or arousal: [P]erceived self-efficacy does not include anxiety in either the definition or the measuring devices. Self-efficacy scales ask people to judge their performance capabilities and not if they can perform nonanxiously. Indeed, considering the confused relationship that exists between anxiety arousal and behavior (Barlow, Leitenberg, Agras, & Wincze, 1969; Leitenberg,

86  David H. Barlow

Agras, Butz, & Wincze, 1971; O’Brien & Borkovek, 1977; Orenstein & Carr, 1975; Schroeder & Rich, 1976), to include nonanxiety as a defining property of self-efficaciousness would diminish its predictive value. (Bandura, 1984, p. 238) Thus, it is possible that studies reporting a low correlation between anxiety (or panic) and self-efficacy may not necessarily contradict other studies showing that self-efficacy can predict the accomplishment of certain tasks in a behavioral frame of reference. In fact, the data reviewed above generally indicate that the relationship between self-efficacy and performance is the strongest, while the relationship between self-efficacy and anxiety or arousal is the weakest. This supports Bandura’s contention. The difficulty with restricting self-efficacy theory to predicting performance capabilities is that these capabilities play little or no role in many of the major clinical anxiety disorders. By all estimates, the most common clinical anxiety disorders are the anxiety states—panic disorder and generalized anxiety disorder. Of course, panic disorder may be accompanied by the complication of varying amounts of avoidance behavior, but unless the avoidance behavior is extensive and accompanied by an elaborate network of safety signals, this is not the focus of concern. Rather, the focus, in terms of both complaints by the patient and the target of treatment, is the anxiety itself. Furthermore, evidence is reviewed in Chapter 11 that the most important predictor of outcome in clinical populations with panic disorder with or without avoidance behavior (agoraphobia) is a reduction in the intensity of anxiety occurring during treatment (e.g., Barlow, O’Brien, & Last, 1984; Michelson Mavissakalian, Marchione, Dancu, & Greenwald, 1906). It may be possible to extend self-efficacy theory to include performance capabilities in dealing with intense anxiety and panic. Although the definition of “performance” is not particularly clear in this context, since there are few behavioral references, successful experiences of coping with the sudden increases in somatic sensations that form the beginning of the spiral into panic might somehow be formulated in self-efficacy terms. More recently, Bandura (1986) seems to have made this extension by suggesting that “inefficacy sometimes involves perceived vulnerability to total loss of personal control” (p. 426). With this extension, constructs such as perception of self-efficacy and a sense of control may blend together, although each has a different emphasis. A sense of predictability and controllability may be more useful in predicting fluctuations in intensity of anxiety, while self-efficacy relates more clearly to performance. To anticipate a later discussion a bit, a sense of control seems to play an important role in panic attacks (Rapee et al., 1986), and it is possible that this cognitive construct may describe more accurately one’s capacity to experience sensations of increased physiological arousal in a nonanxious manner. But the tests have not been done, and with the clear emphasis of self-efficacy theory on performance capabilities, it has been misleading to assume that self-efficacy predicts intensity of emotions—something Bandura himself (1984, 1986) would not claim. This state of affairs seems to provide yet another

The Process of Fear and Anxiety Reduction  87

example of the difficulties encountered in any theorizing about a construct such as anxiety, with multiple points of reference that necessarily differ in their expression from one clinical manifestation to another. Emotional Processing One of the newest theories of fear and anxiety reduction is based on the bioinformational theory of affect proposed by Lang (1977a, 1977b, 1979, 1985) and reviewed in some detail in Chapter 2. This is a model that encompasses desynchrony to some extent and puts more emphasis on emotional aspects of anxiety than on behavioral aspects such as avoidance behavior. This model underscores the importance of exposure, but observes that what is important about exposure, whether in vivo or in imagination, is the information processing that occurs when all components or elements of the affective structure are present. Like bioinformational theory itself, this derivative emphasizes physiological indices of anxiety as the most direct measure of the cognitive activity underlying anxious responding and anxiety reduction. In this sense, it resembles habituation in that ii requires the initial elicitation of, and gradual decrements in, physiological responding during exposure to a feared situation. What is different, however, is the suggestion that something is learned during this process. This model of anxiety reduction derived from Lang’s theorizing is called “emotional processing,” a term coined by Rachman (1980). Rachman (1900) and Foa and Kozak (1985, 1986) have elaborated on this model, providing greater specificity to the predictions. Briefly reviewing the underlying bioinformational theory of affect, Lang rejects the long-held view that images are basically “pictures in the mind” or iconic representations in storage, and that imagining involves the inward perceiving of these images. Lang cites Pylyshyn (1973), who suggested that images in the brain are more like elaborated descriptions. In short, the image is a functionally organized, finite set of propositions. In the context of emotional, fearful images, it is hypothesized that these images contain at least three fundamental classes of statements: stimulus propositions, response propositions, and meaning propositions. “Stimulus propositions” include those sensory (auditory, visual, tactile, etc.) details of the feared object or situation. “Response propositions,” on the other hand, are broken down into the specifics of the now familiar triple-responses associated with the stimuli and responses. “Meaning propositions” are the interpretations associated with the stimuli and responses. Fear, then, is a program for fear behavior existing in memory (much like a computer program), comprising stimulus, response, and meaning structures. This fear program must be fully accessed if any important therapeutic change is to occur. In other words, to modify fear or anxiety, the individual must first become fearful or anxious. Analyzing imagery-based anxiety reduction procedures in particular (but not limiting this model to those procedures), Lang proposes that different techniques, such as desensitization and imaginal flooding, may call forth different parts of these proportional statements. Thus, vividness of the image and intensity of the affect elicited are determined by the completeness of the propositional statement presented. For example, systematic desensitization, with

88  David H. Barlow

its emphasis on imagining feared situations arranged in a hierarchy of intensity, concentrates on stimulus propositions. Flooding, on the other hand, concentrates on response propositions, since details of anxious responding (“Your heart is racing; you feel like running”) are emphasized in imagination. The therapeutic effectiveness of an image, then, is determined in part by its propositional structure, the balance between stimulus and response elements, and the interrelated characteristics of vividness and affective intensity. Lang notes that the effectiveness of desensitization with specific fears, such as fear of small animals, may be due to the circumscribed nature of the stimulus propositions of these fears. The seemingly greater success of flooding with the more complex problems such as agoraphobia, in which anxiety associated with somatic responses is a major problem, may be due to the relative emphasis on response propositions (particularly visceral reactions) typically described in a flooding session. The greater success of in vivo exposure compared to imaginal exposure may be due to the more certain (although still not definite) elicitation of all stimulus, response, and meaning propositions comprising the phobia. The fundamental strategy of treatment is to “process” information comprising the emotional image, and the emotional image is, for Lang, the prototype fear image stored in long-term memory. This “template” can be elicited for emotional processing in a variety of ways, either imaginal or in vivo, but all aspects of it should be processed for successful therapy. This processing may make possible the formation of a new response prototype preliminary to overt behavior change. This new prototype would contain a less anxious or nonanxious response to the stimulus propositions. The Process of Anxiety Reduction During Emotional Processing

Lang does not actually suggest a specific process by which something new is learned, but Rachman (1980) and Foa and Kozak (1985, 1986) have begun to address this issue. In a stimulating heuristic, Rachman first suggested indices of unsatisfactory emotional processing that would be observable over time. These indices are listed in Table 6.l. The term “test probes” refers to the occasional presentation of aspects of the original emotional experience to determine whether unprocessed elements of the emotional template remain in memory. Rachman then suggested factors that may impede or promote the transformation or “neutralization” of emotion-provoking stimuli. These factors are presented in Table 6.2. Many of the listed factors that may promote processing are familiar ingredients of current therapies, particularly exposurebased therapy. But a crucial ingredient is autonomic reactivity. In other words, to modify an emotion one must elicit emotion, as indicated by autonomic indices in the emotional situation. This and other issues raised by Rachman are discussed further below. But Rachman provided only an outline of the therapeutic process, and did not specify therapeutic targets in this early paper. Foa and Kozak (1986) expand and extend some of these ideas, particularly in the context of exposure-based therapy. They suggest that the process of fear or anxiety evocation that is part of any exposure-based treatment must contain information that is incompatible with at least some of the elements that

The Process of Fear and Anxiety Reduction  89 Table 6.1 Indices of unsatisfactory emotional processing observable over time

A. Direct signs   1. Test probes elicit disturbances   2. Obsessions   3. Disturbing dreams   4. Unpleasant intrusive thoughts   5. Inappropriate expressions of emotion (as to time/place)   6. Behavioral disruptions, distress (e.g., crying)   7. Pressure of talk   8. Hallucinations (e.g., after bereavement)   9. Return of fear B. Indirect signs 10. Subjective distress 11. Fatigue 12. Insomnia 13. Anorexia 14. Inability to direct constructive thoughts 15. Preoccupations 16. Restlessness 17. Irritability 18. Resistance to distraction Note. Adapted from Rachman, S. (1980). Emotional processing. Behavior Research and Therapy, 18, 51–60. Copyright 1980 by Pergamon Journals, Ltd. Adapted by permission.

exist in the patient’s anxiety structure. These new elements must then be perceived and integrated into the existing structure for an emotional change to occur. Specifically, during exposure-based treatments, patients learn through short-term habituation of physiological responding (usually during a single session) that anxiety will not last forever. They learn through repeated exposure over several sessions that the probability of harm or danger is low. This particular change in the anxiety structure is reflected in between-session, long-term habituation. Negative affect or the “aversiveness” associated with a situation also diminishes with repeated exposure. In other words, they learn that the experience itself is not so bad. These three pieces of corrective information constitute what is “learned” during emotional processing, according to Foa and Kozak. In vivo exposure is a particularly good way to access the anxiety and transmit new information or elements, but, for various reasons, it may not always be successful at either accessing the anxiety structure in memory or communicating new information. For example, simple distraction or misinformation during exposure procedures should interfere with anxiety reduction, as also suggested by Rachman (1980). There is some evidence that this happens (e.g., Grayson, Foa, & Steketee, 1982; Sartory, Rachman, & Grey, 1982). In addition, the feared situation may be better accessed by imaginal presentations for a variety of reasons, such as technical difficulties that arise in reproducing the actual situation. Other factors that Foa and Kozak (and Rachman) suggest may interfere with emotional processing include the duration of exposure. For example, longer

90  David H. Barlow

exposure should produce more habituation, and consequent change in the fear structure should also facilitate change. Adequate matching of situations and information presented during exposure to the actual anxiety structure should also facilitate change. In addition, if the patient is depressed or is “cognitively avoiding” the feared stimulus during exposure by self-distraction, emotional processing will be incomplete. Finally, emotional processing will be hindered if initial arousal is so high that short-term habituation cannot occur, thus communicating to the patient that anxiety may continue indefinitely. Over the long term, high arousal is seen as reflecting continuing erroneous evaluation of the probability of danger or harm; this has been termed “overvalued ideation.” This model, as outlined by Rachman (1980) and Foa and Kozak (1986), has many implications for behavior change procedures. For example, it suggests that physiological responding must occur during therapy for the emotional image to be fully processed. Similarly, lack of autonomic reactivity, reflecting incomplete accessing of the anxiety structure among some patients, may explain the wide intersubject variability during otherwise similar exposure trials. While this model is certainly preliminary, it has the virtue of being testable. Evidence Supporting Emotional Processing Evidence is now beginning to appear supporting the theory of emotional processing, although the very nature of the theory, with its emphasis on deep cognitive structures and the relationship between these cognitive structures and psychophysiological responses, guarantees that the road to confirmation or disconfirmation will be long and hard. Lang (1979, 1985) has demonstrated, mostly with analogue subjects, that training these subjects to process response information as well as stimulus

Table 6.2 Factors that promote or impede emotional processing

Promote

Impede

Engaged exposures Calm rehearsals (especially of coping)

Avoidance behavior Agitated rehearsals (especially of not coping) Silence Distractions Poorly presented material Excessively brief presentations Inadequate practice Excessively large “chunks” Immobility Fatigue Irregularity of stimulation Unresponsive autonomic reactions

Talk Habituation mining Extinction No distractions Catharsis Vivid presentations Long presentations Repeated practice Relaxation Autonomic reactivity

Note. Adapted from Rachman, S. (1980). Emotional processing. Behaviour Research and Therapy, 18, 51–60. Copyright 1980 by Pergamon Journals, Ltd. Adapted by permission.

The Process of Fear and Anxiety Reduction  91

information leads to increased physiological activity during imagery. According to the theory, this suggests successful processing of the emotional image, although this fact alone may simply reflect the results of training in imagining response information. Recently, Robinson and Reading (1985) demonstrated with simple phobics that imagery with response propositions produced stronger physiological responses and increased concordance with already declining subjective measures of these phobias than did imagining of stimulus propositions alone. Once again, this suggests successful processing. Finally, this theory may accommodate the otherwise puzzling finding to be reviewed in Chapter 10 (J. A. Vermilyea, Boice, & Barlow, 1984)—namely, that initial high heart rate in feared situations predicted a positive outcome among a group of agoraphobics. This finding is reminiscent of analogue studies by Lang (1985) and others, in which higher physiological responding to feared images at pretreatment predicted better outcome from treatment. Presumably, what has been happening in these studies, according to emotional processing theory, is that subjects with higher physiological responses have been processing the anxiety cues more effectively and efficiently. This in turn has allowed for successful formation of new propositions (after the response has habituated) and the reduction of anxiety. Foa and her colleagues have also found initial physiological responding predictive of positive outcome from therapy (Foa & Kozak, 1985, 1986). But Foa and Kozak (1985) cite evidence from both basic and applied studies that high arousal during the initial stages of treatment may interfere with outcome. They suggest that moderate arousal that is neither too high nor too low is required for successful processing. But it will be very difficult to decide in a given individual what is “moderate” and what is not, particularly with physiological measures. Craske and Rachman (1987) also provide what may be contradictory evidence on the therapeutic predictive power of initial high arousal. They demonstrated that initial high heart rate was the one variable that predicted return of fear in a group of musicians suffering performance anxiety. In this study high heart rate did not predict response to treatment at posttest, but did predict the self-reported return of fear 3 months after treatment was completed. Of course, heart rate was not high in an absolute sense, as would be required to support the notion that moderate heart rate is optional for emotional processing. That is, heart rate in these individuals was not categorized as high, moderate, or low on the basis of predetermined criteria. Rather, individuals whose average heart rate was above the median evidenced return of fear. But this was also true in the J. A. Vermilyea et al. (1984) study, where high heart rate did predict successful outcome immediately after treatment (at posttest). There are many barriers to a full understanding of these results at present. For example, relatively high heart rate seemed to predict successful outcome in the J. A. Vermilyea et al. (1984) study, but not the Craske and Rachman (1987) study. However, there was some evidence of return of fear in individuals in the former study at follow-up. Thus, initial high heart rate may predict both positive responding after treatment and return of fear at follow-up when these issues are fully explored. An additional difficulty with a clean interpretation of the Craske and Rachman study is that heart rate was not measured during

92  David H. Barlow

the actual feared situation (in this case, during a musical performance); rather, heart rate was measured before the performance. Presumably, this reflected preparatory or anticipatory arousal. This arousal may have had little or no correlation with arousal during the feared situation itself, which might, in turn, have indicated more accurately whether the emotional structure was being processed. In the J. A. Vermilyea et al. (1984) study, on the other hand, initial heart rate was measured in the actual feared situation. In several earlier experiments, where return of fear was also observed, emotional processing received stronger support. Both Grayson et al. (1982) and Sartory et al. (1982) demonstrated that subjects who were distracted in some way during the exposure process showed less between-session habituation and a return of fear several days after individual exposure sessions. While it is not entirely clear how return of fear within a day or two relates to return of fear months later, the data deserve further investigation. Craske and Rachman’s (1987) observation that the process of consolidating new information after exposure seems to be disrupted in those whose anxiety later returns underlines the importance of these data. In other words, distraction somehow prevents full attention to and adequate processing of the fear structure. But there was no overt distraction in the Craske and Rachman experiment that could account for return of fear in some subjects and not in others. The theory would predict that some sort of covert distraction or some other factor inhibited effective processing in this instance. But both within-session and between-session habituation seemed to be going as well in those whose fear later returned as in those whose fear reduction was more permanent. The reasons for these discrepancies are not clear. Difficulties with the Theory Discordance

By now, the similarities that emotional processing shares with many elements of the habituation process are obvious: Each theory relies on a gradual decrement in physiological responding, both within and between sessions, for evidence of successful processing. Unfortunately, the theory also shares several weaknesses with habituation. Among these is one example of the phenomenon of courage, described above (Rachman, 1987). In this example, people are able to perform dangerous tasks very effectively and with little or no report of fear, despite the presence of marked physiological responding that would otherwise be indicative of fear, in a situation where fear is appropriate. This phenomenon is also apparent in patients, where it is the most common manifestation of posttreatment discordance among response systems (J. A. Vermilyea et al., 1984). Specifically, patients report substantial reductions in fear, but evidence no change whatsoever in physiological responding such as heart rate. For years, it was thought that these people were at risk for relapse. This would be consistent with emotional processing theory, since the assumption is made that the images have not been completely processed. However, there is little evidence as yet that this happens (Craske, Sanderson, & Barlow, 1987a; J. A. Vermilyea et al., 1984). Of course, studies required to document this fact are few and far between and very difficult to carry out. In any case,

The Process of Fear and Anxiety Reduction  93

there are many clinical examples of patients who seem to do very well, despite the continuing presence of high physiological responding in feared situations. In an early example, Leitenberg et al. (1971) described a claustrophobic who demonstrated increasing heart rate during treatment, which remained very high at a 1-year follow-up. Nevertheless, she was able to enter small enclosed places quickly and with seeming ease, and reported little subjective disturbance. What she said was that she had learned to be “brave” and that carrying out her formerly feared behavior was no longer a problem, despite awareness that her heart was beating faster. What is difficult about this situation for emotional processing theory is not that the woman was performing well in spite of her fear (something that might have been predicted by self-efficacy theory). Rather, the problem is that the woman evidenced high physiological responding without any report of fear at a 1-year follow-up. It goes without saying that this is an isolated case; one can find an anecdotal report to support or undermine any theoretical position. The full implications of posttreatment discordance for the theory of emotional processing will have to await new data. Lang, of course, is fully aware of the poor correlation among the three major response systems comprising an emotion, since he was the first to propose this trichotomy among response systems (Lang, 1968). As noted above, he would presumably point out that these examples of long-term discordance, which are otherwise clinically satisfactory, represent cases in which the phobic emotional propositions stored in memory are incompletely processed or somewhat weaker but are still present. Another possibility (mentioned in Chapter 2) is that in some cases, the propositions are so diffuse that arousal may no longer be strongly connected to the phobic structure. If future research determines that discordance after treatment predicts relapse, then these data not only will support the theory but will have substantial clinical implications. On the other hand, if it is enough to reduce subjective disturbance and behavioral manifestations of anxiety by, for example, instilling a sense of control despite the presence of physiological reactivity, then it is not clear that these ideas will be useful in the long run as a comprehensive theory of fear reduction in clinical situations. In other words, heightened physiological responding may not have any functional or clinical implications. Gradual Decrements in Physiological Responding

The requirement of a gradual decrement in physiological responding as evidence of emotional processing is also inconsistent with other data. Evidence has been presented in Chapter 5 (see Figure 5.3) that changes in heart rate are not a valid indicator of anxiety reduction in the sense of discriminant validity. We (Holden & Barlow, 1986) found that gradual decrements in heart rate occurred over time in both a group of agoraphobics and a matched group of normal controls who were simply tested at identical intervals. No differences were evident in their rate of habituation. The only clear differences that emerged were that agoraphobics registered stronger physiological responding during resting baseline; thus, somewhat higher responding continued all the way through treatment. But normals reported no fear whatsoever, despite

94  David H. Barlow

initial high heart rate and similar rates of habituation. If gradual decrements in physiological responding signify successful emotional processing and anxiety reduction, then what can be said about the group of normals showing a similar response? There are no easy explanations. The issue of response to “novelty,” which is considered by some to be closely related to fear (at least physiologically), does not seem explanatory: Normals were very quickly bored by the standardized behavioral approach test designed for agoraphobics, which consisted of walking away from a safe place. Identification of Fear Structures

Other issues arise at this early stage of analysis of emotional processing as a theory of anxiety reduction. In laboratory studies, Lang and his colleagues have been most successful at eliciting coherent and concordant responses, presumably reflecting feared prototypes, in simple phobics compared to social phobics or agoraphobics (Lang, 1985). Lang suggests that the emotional prototypes are generally less coherent or more diffuse for the latter groups. On the other hand, it is possible that Lang’s experiments have not tapped the emotional prototypes in these groups in the most effective manner. For example, the variety of laboratory panicogenic procedures described in Chapter 4 would seem far more appropriate then simply imagining moving away from a feared situation, which was often the test in early experiments. Rather than the diffuse, fluid response structures proposed by Lang for the anxiety states, the context of fear for patients with panic disorder may have to be specified more carefully. Only in this way can emotional processing theory be fully tested with these disorders. For a patient with generalized anxiety disorder, it is somewhat more difficult to envision emotional processing, since the response seems so diffuse that it is not clear what is being processed in terms of stimulus and response propositions. Here action tendencies and underlying arousal are associated, by current definition, with numerous situations and areas of functioning, making identification of stimulus propositions problematic. Non-Exposure-Based Treatments

In addition, Rachman (1985) points out a number of factors, including the provision of new information, that seem to result in reductions in fear without any exposure taking place whatsoever, Certainly, this raises some difficulty for emotional processing or, for that matter, for any of the theories mentioned above except perhaps self-efficacy theory. Rachman suggests, however, that emotional processing can easily incorporate this phenomenon. What is “processed” in this instance is new information. This new information contributes to the formation of a different propositional structure in relation to the object or situation that formerly evoked fear. If emotional processing means accessing the memory structure that contains the irrational fear and arranging for “reality testing” to occur, then it is entirely possible that simple provision of additional information will occasionally have some impact on that memory. This assumes that the memory is accessed to begin with. Of course, it seems

The Process of Fear and Anxiety Reduction  95

that non-exposure-based methods of fear reduction, such as the provision of new information, seldom have more than moderate effects on severe irrational fears (Barlow & Beck, 1984). But it is important to make this modification to emotional processing theory if it is to be comprehensive. This modification could be tested by observing rates of habituation and initial levels of arousal during these non-exposure-based methods. Psychophysiological Methods

A final difficulty with emotional processing, as with the bioinformational theory of affect from which it derives, is its reliance on psychophysiological data collected under a variety of different conditions. Studies reporting unreliability of these data, even in the best of circumstances, are reported in subsequent chapters (see Chapter 10). A number of different factors influence states of arousal. Also, a number of different factors in addition to arousal account for psychophysiological responding. Unfortunately, we have known for years that psychophysiological responses are not well correlated even with each other, reflecting, as they do, different bodily functions. Building a precise theory of anxiety reduction based on psychophysiological responding, no matter how promising, will be time-consuming and difficult. Summary of Emotional Processing

Despite the richness and comprehensiveness of Lang’s theory of emotion, the evidence supporting the derivative of this theory, emotional processing, as an adequate explanation for anxiety reduction is very sparse at this time. As noted above, the requirement of decrements in physiological responding to formerly feared situations does not necessarily occur, or if it does occur, it does not seem to be specific to the fear reduction process (e.g., Holden & Barlow, 1986). This leaves us with the findings on high physiological responsiveness before treatment predicting successful outcome. But once again, as reviewed above, the results are inconsistent. This has led Foa and Kozak (e.g., 1986) to propose that physiological responsiveness pretreatment can be neither too high nor too low, but must be somewhere in between. This may be either an elegant or a disingenuous solution, depending on future data, but it will be very difficult to disconfirm in view of the enormous variability and the individual nature of physiological responsiveness in terms of what is high, low, or moderate. Lang’s theories will generate years of research, but the derivative process of anxiety reduction may need considerably more elaboration before it forms an adequate explanation of anxiety reduction. Nevertheless, it is clearly the most comprehensive theory yet to appear, and the only one to rest firmly on a foundation of emotion theory. Affective Therapy Peter Lang was not the first to call for a coherent and consistent therapeutic approach to emotional disorders based on emotion theory. Emotion theorists (including Darwin, James, and Cannon) have for years suggested therapeutic

96  David H. Barlow

strategies for modifying maladaptive emotions, based on their theories (e.g., Izard, 1971). For a variety of historical reasons, these suggestions did not take root; thus, later elaborations into detailed coherent therapeutic approaches were not forthcoming. In 1981, Rachman, reacting to Zajonc’s comments on the primacy of affect reviewed in Chapter 2, suggested that conceptual development in the area of affect modification would be a necessary task for any psychotherapy or behavior therapy in the 1980s. Wilson (1982) suggested that attention to cognitive aspects of psychopathology reflected in the development of cognitive therapy would have to be accompanied by similar developments in the modification of affect. It is clear that a comprehensive approach to therapeutic intervention must consider affect, particularly in the context of the emotional disorders, where problems with affect predominate. As noted above, Greenberg and Safran (1987), in an important book, have outlined a general approach to psychotherapy that is based on emotion theory. The construction of Lang’s bioinformational theory of affect set the stage for the first theoretical statement on anxiety reduction to be based on a theory of emotion, as outlined above (Foa & Kozak, 1986; Rachman, 1980). Of course, creative and knowledgeable therapists approaching emotional disorders from a clinical perspective have attended to the implications of basic theories of emotion. In the best-known example, Aaron Beck based his theory of pathological anxiety as a function of distorted information processing on his clinical observations. With these exceptions, there have been few attempts to integrate emotion theory with therapy for emotional disorders. In preceding chapters, the beginnings of models of panic and anxious apprehension, based on present knowledge of emotion theory and associated cognitive-affective processes, have been presented. I have suggested that the phenomenon of panic (emotional alarm) is not in and of itself pathological unless it becomes associated with the process of anxious apprehension. But once it becomes associated with anxiety, panic requires therapeutic intervention. The model of the process of anxious apprehension is presented in Figure 7.6. This model, in conjunction with a specification of factors implicated in the etiology of anxious apprehension (Figure 7.7), suggests the possibility of intervening at many different points in the cycle of anxious apprehension. In other words, therapeutic procedures directed at any one of the many components of the loose cognitive-affective structure of anxious apprehension have the potential for therapeutic benefit if the negative feedback cycle is disrupted. However, in my view, some components of the cycle make more crucial contributions to the ongoing spiral of anxiety than others. Therefore, these components comprise more essential targets for therapeutic action. Targeting other components, although helpful, should not prove to be as efficient therapeutically in decreasing the intensity of the cycle. According to the reasoning developed in previous chapters, the core of anxiety disorders is negative affect—characterized in dimensional analyses of emotion as high arousal supporting a preparatory coping set, accompanied by a marked sense of uncontrollability and helplessness. This is experienced subjectively as unpleasant (i.e., it has a negative valance). The closely related cognitive process of attentional shifts to a self-evaluative focus is also part of the core of anxiety. The essential targets for therapeutic action are thus (1) efforts to cope,

The Process of Fear and Anxiety Reduction 97

as well as the emergency action tendencies associated with alarms; (2) a sense of lack of control; and (3) attentional shifts to self-evaluative affective content. Targeting a number of additional components of anxiety may be useful or helpful, but perhaps not as essential, according to both evidence and theory. Among these additional targets are (1) “hot” apprehensive cognitions; (2) hypervalent cognitive schemata and the associated cognitive-affective structure of attention narrowing; (3) associated coping skills, social support networks, or other factors capable of moderating the intensity of negative affect in response to personal stressors; and (4) physiological arousal and other somatic processes supporting the action tendencies. A truly comprehensive therapy should target all components. But I suggest that without marked change in the essential components, affective modification will not be successful. The remainder of this article consists of evidence for these assertions. Specification of these therapeutic components forms the framework for a discussion of effective therapeutic interventions for each anxiety disorder in subsequent chapters. Finally, it is important to note that individual therapeutic procedures are not necessarily associated with one component of anxiety or another. For example, popular procedures such as relaxation or coping skills training may not have their primary effect through decreasing arousal or teaching specific new skills. Rather, both procedures may instill a sense of control or may change action tendencies or foci of attention. Essential Targets for Change Action Tendencies

Baudura, one of the foremast proponents of cognitive explanations of behavior change, stated as early as 1977 in reference to anxiety reduction: On the one hand, explanations of change processes are becoming more cognitive. On the other hand, it is performance based treatments that are proving most powerful in effecting psychological changes. Regardless of the method involved, the treatments implemented through actual performance achieve results consistently superior to those in which fears are eliminated to cognitive representations of threat. (1977b, p. 78) For years, therapists of all persuasions have agreed that a necessary step in the treatment of clinical phobia is to encourage contact with the feared object or situation. The well-known statement of Freud on the necessity of having phobic patients eventually face their fears has often been used to support exposure-based treatments: “One can hardly ever master a phobia if one waits till the patient lets the analysis influence him to give it up. . . . one succeeds only when one can induce them through the influence of the analysis . . . to go about alone and struggle with the anxiety while they make the attempt” (Freud, 1919/1959, pp. 165–166). In fact, exposure to feared situations has become the sine qua non of any therapeutic approach to phobia, as reviewed above. Exposure has been viewed by many as simply a vehicle to reduce or eliminate “anxiety” connected with phobic situations. “Anxiety” usually refers to

98  David H. Barlow

subjective or physiological aspects of the affect in this instance, and habituation or extinction processes are usually invoked as mechanisms of action. But emotion theorists such as Lang or Izard have long recognized that emotions are primarily action tendencies (Izard, 1977). To the extent that one can counteract these action tendencies, emotions will change. As Izard (1971) points out, theories and evidence from emotion theory indicate that “the most efficient and generalized principles and techniques for emotion control [are] focused on the neuromuscular component of emotion. . . . striate muscle action can initiate, amplify, attenuate, or inhibit an emotion” (p. 415). In other words, “the individual learns to act his way into a new way of feeling.” (p. 410). It is possible that the crucial function of exposure, instead of facilitating extinction, is to prevent the action tendencies associated with fear and anxiety. As our measurement procedures become more precise and sophisticated, investigations are beginning to pinpoint the action tendencies associated with emotion. For example, Fridlund et al. (1986), in a study referred to above, determined that physiological elevation during anxiety did not represent generalized arousal, but rather specific action tendencies associated with the effect. Further evidence reviewed above indicates that the expression of action tendencies intensifies the specific emotion with which they are associated. Thus, allowing action tendencies associated with fear and anxiety to occur is countertherapeutic. What are these action tendencies? For the ancient, basic emotion of fear (panic), the behavior is clearly escape or flight. As noted in previous chapters, patients experiencing panic are preoccupied with an overwhelming urge to escape, which is usually irrational and maladaptive. An example is the fictional character Mrs. Thayer’s urge to rush headlong into the ocean during the winter blizzard, described in Chapter 1 (Fisher, 1978). There is nothing volitional about this action tendency. Preventing it may be the single most important, therapeutic means of directly countering panic. It is here that patients may be able to act their way into a new way of feeling. But I have noted above that anxious apprehension may be more central to anxiety disorders. What is the action tendency associated with the negative affective core of this cognitive-affective structure? In Chapter 7, this tendency is described as a preparatory coping set. This is the general “activation” observed by Fridlund et al. (1986) and Fowles (1986), which they have differentiated from generalized arousal. It is the chronic state of readiness that is referred to as “vigilance” in most accounts of anxiety. The purpose of chronic hyperarousal, mentioned so often in previous chapters, is to support this coping set. But it is not a specific behavior, such as escape; it is the readiness to behave. It is this activation that becomes an important target for change. It is possible that attention to action tendencies forms an important part of the treatment of other emotional disorders. For example, Beck, Rush, Shaw, and Emery (1979) and others (e.g., Lewinsohn & Lee, 1981) spend a considerable amount of time countering the tendencies of their depressed patients to behave in a passive, retarded, and apathetic manner. In any case, the majority of therapeutic approaches to anxiety disorders in general and phobia in particular have stumbled on the necessity of directly countering the action tendency of escape associated with panic. Escape behavior may be blatant, as is the case with phobic avoidance, or more subtle, as

The Process of Fear and Anxiety Reduction  99

with the variety of cognitive avoidance strategies seen in panic disorder (see Chapter 11). One must prevent escape or other avoidance responses, both cognitive and behavioral, and must strongly encourage approach behavior. Overwhelming evidence supporting the usefulness of this approach is presented throughout this book. Assuming that the prevention of chronic preparatory coping sets is also a primary goal in the modification of anxiety disorders, it is possible to interpret the role of other direct interventions a bit differently. For example, all emotion theorists believe that the physiological or neurobiological processes associated with varying emotions are there for a reason, which is to support behavior associated with the affect. If this is true, procedures directly targeting physiological or neurobiologcal aspects of anxiety may only be indirectly effective. That is, procedures such as relaxation, biofeedback, or anxiolytic medications may be effective only to the extent that they alter the action tendencies associated with chronic vigilance through undermining the supportive physiology. On the other hand, a procedure such as relaxation may be useful not because of any arousal-reducing properties, but because it directly substitutes a different action tendency for chronic vigilance. In summary, the overwhelming evidence from emotion theory is that an essential step in the modification of emotional disorders is the direct alteration of associated action tendencies. Laughter, humor, and associated facial expressions induced during successful paradoxical intention (Frankl, 1960)—a technique successfully used to counteract fear and anxiety (e.g.. Ascher, 1980)— may be effective not because of changes in self-statements, as is often assumed. Rather, prevention of behavioral responses (including facial expressions) associated with fear and anxiety, and the substitution of action tendencies associated with alternative emotions, may account for the effectiveness of this technique. Control

I have suggested that at the core of the affective component of the complex cognitive-affective structure of anxiety is a sense of uncontrollability and unpredictability. As noted in Chapter 2, “control” is one of the three essential elements of emotion, according to dimensional analyses (e.g., Lang, 1985). A sense of lack of control may also be a “meaning proposition” in the Langian sense. Furthermore, a useful tradition of experimental analyses of the consequences of lack of control has been established within the context of another emotional disorder. This is the literature on learned helplessness associated with depression (Abramson et al., 1978). Helplessness or a sense of uncontrollability is common to anxiety and depression; these differ in other ways, such as in basic action tendencies, as outlined in Chapter 7. Preceding chapters have also described the close relationship between anxiety and anger in terms of their affective components. That is, anxiety and anger (and perhaps excitement) seem to share similar action tendencies, but differ in the all-important feature of perceptions of control. Angry and excited people are emotional but perceive that they are in control; anxious people perceive themselves as lacking control. For this reason, altering perceptions of control becomes an essential part of the therapeutic effort directed at anxious apprehension.

100  David H. Barlow

Altering perceptions of uncontrollability is also important from an etiological perspective, as outlined in Figure 7.7. These perceptions seem to comprise one of the major etiological contributions to the formation of the cognitive-affective structure of anxiety. That is, experiences of lack of control during development in biologically susceptible individuals may be a crucial etiological step in the origins of anxiety. For this reason, changing perceptions of helplessness is central to any therapeutic endeavor. It should be clear from the preceding comments that altering action tendencies may be a good beginning to altering a sense of control. Modification of action tendencies should have an impact on the affective core of anxiety in general. Nevertheless, there are many instances where preventing phobic (or cognitive) avoidance with exposure-based procedures is only partially effective and sometimes totally ineffective. It is possible that some of these exposure-based treatments may have been ill-designed. For example, I have suggested in Chapter 6 that when learned alarms are involved in an anxiety disorder, it may be essential to prevent escape from somatic cues associated with panic, in addition to escape from any external situations associated with phobic behavior. Nevertheless, these therapeutic attempts will not be completely successful unless the patient begins to feel in control of potential upcoming events, whether environmental or somatic. This may explain the puzzling series of studies with phobics described in Chapter 11 and alluded to above, demonstrating that anxiety reduction can occur even if subjects escape from the feared situation before reaching maximum anxiety. These studies are puzzling because one of the basic assumptions of extinction (or emotional processing) is that the patient must learn that there are no untoward or aversive consequences to remaining in the feared situation. But this would seem difficult if the patient does not stay in the situation long enough to learn this. It is possible that what is “learned” during these experiences is that events are not out of control. That is, whether aversive consequences occur or not, or whether unwanted physiological arousal occurs or not, the individual is in control of his or her world. This may be true even if the sense of control is illusory (Alloy et al., 1981). In this conceptualization, it is not important if high arousal supporting a coping set remains, as long as the action tendencies themselves are altered and a sense of control is instilled. Discordance, as indicated by the continuation of high physiological activation, is relatively unimportant. Of course, in the majority of cases, one would not expect the supportive physiology to continue on indefinitely in an autonomous manner if other, more essential aspects of the affect have been eliminated. There is additional evidence that what may be learned during the variety of treatment procedures for anxiety disorders is a sense of control. The centrality of the issue of control is also evident in emerging data on the emotional effects of expected versus unexpected noxious events, particularly alarms (reviewed in Chapter 4). For example, Rachman and Levitt (1985) have indicated quite clearly that the occurrence of false or learned alarms does not necessarily increase anxiety subsequently. Only if the alarms are unexpected does anxiety subsequently increase. Similarly, Rapee et al. (1986) have demonstrated that provocation of panic in the laboratory through CO2 procedures seem to be

The Process of Fear and Anxiety Reduction  101

a function of whether one expects the various somatic sensations associated with CO2 inhalation or not. If the effects are fully described in advance, reports of panic do not occur. In other words, if events are seen as predictable or controllable, there is little or no anxiety. Sanderson et al. (1987c) confirmed this finding by directly manipulating a sense of control, which affected reports of anxiety and panic. Finally, what may be the strongest evidence for the importance of a sense of control comes not from the anxiety disorders, but from the stress disorders. Although a subset of patients with stress disorders seem to have a strong sense of mastery and control, as exemplified in Type A personalities, a large number of individuals within this heterogeneous group of disorders seem to resemble patients with anxiety disorders in many essential ways (see Chapter 9). For example, patients with chronic tension headache and irritable bowel syndrome, to take two common stress disorders, present with a great deal of emotionality. This emotionality can be characterized as anxiety surrounding their somatic symptoms. One of the most popular and effective treatments for these disorders is biofeedback targeted to the specific disturbances. But one of the most puzzling findings from biofeedback research is that the mechanism of action does not seem to be in the reduction of the physiological response presumably underlying the symptoms. For example, biofeedback of tension headaches has as its goal the reduction of muscle tension in crucial areas of the face and head. But recent investigations have shown that reductions in headache when they occur are not associated with specific reductions in muscle tension in these areas (Andrasik & Holroyd, 1980; Holroyd et al., 1984). This raises a major paradox in the field of biofeedback today. What accounts for therapeutic success when it occurs? One possibility is that these patients undergoing the technically sophisticated treatment of biofeedback, with all of its electronic gadgetry, begin to develop the sense that their unpredictable and uncontrollable aversive somatic experiences are at last coming under some sort of control. As Blanchard points out in discussing his clinical success with irritable bowel syndrome (Neff & Blanchard, 1987), it seems that at last these patients begin to feel some control over an area of their functioning that has been unpredictable and uncontrollable for years. In view of the seeming similarity in the etiology of anxiety and stress disorders, it is possible that common mechanisms of action underlie the success of the variety of treatments used for both disorders. Self-Focused Attention

The last aspect of treatment seen as essential in the modification of anxiety is the necessity of shifting attention away from an internal affective self-evaluative focus. In Chapter 7, I have observed that one of the major differences between “normal” and pathological anxiety involves the persistence of an internal self-evaluative focus of attention. In the model presented in Figure 7.6, this leads to further intensification of arousal, eventual disruptions in performance, and the creation of the vicious cycle of anxious apprehension. Furthermore, patients fail to habituate when in a self-evaluative attentional mode. Patients without pathological anxiety, on the other hand, may

102  David H. Barlow

experience similar levels of arousal and even unpleasant affect, but their attention is easily shifted when necessary to the task at hand or to a more objective nonaffective focus, and “anxiety” disappears. Evidence presented above indicates that distraction during exposure-based therapies may interfere with the process of anxiety reduction (e.g., Gravson et al., 1982; Sartory et al., 1982). This is also consistent with an emotional processing view of anxiety reduction (Foa & Kozak, 1986). But it is possible that one of the essential steps in successful anxiety reduction is the redirection of attention to more external or mechanical aspects of the anxious situation, perhaps because it contributes to the development of a sense of control. This does not involve diverting one’s attention away from the situation or escaping it. Rather, the focus of one’s attention is on other aspects of the situation or more mechanical stimulus aspects of one’s responding. Once again, this process may have been implicated in Rapee et al.’s (1986) demonstration that providing full information on the sensory qualities of the somatic effect of CO2-induced panic reduced reports of panic and anxiety. This different focus of attention may have contributed to a greater sense of control. Suls and Fletcher (1985a) have reviewed studies on differential focus of attention in the context of health psychology. They also conclude that focusing on the sensory or mechanical aspects of the stressor and one’s reaction to it (“nonavoidant strategy”) will be more adaptive in the long run than distracting oneself from the situation (“avoidant strategy”). Certainly, the development of an external focus of attention is crucial to increasingly effective performance in the situation, which, of course, is another essential step in breaking the cycle of anxious apprehension. I have hypothesized in Chapter 7 that shifts in attention to an internal self-evaluative affective focus, while originally a cognitive component of the complex cognitive-affective structure of anxiety, eventually becomes an integral part of the negative affective core of anxiety. In view of the enormously important role this component seems to play, according to hard evidence from my colleagues’ and my sexually dysfunctional patients (see Chapter 7), it seems that targeting self-focused attention is essential during therapy. Helpful Targets for Change Targeting a number of additional components of the cognitive-affective structure of anxiety may well be helpful or useful, but is probably not sufficient to effect major changes in anxiety, in the present view. Among these components are self-statement or the “hot” apprehensive cognitions associated with intense worry, and the process of attention narrowing that increases the salience and intensity of these “hot” cognitions. Evidence mentioned above and reviewed in some detail below indicates that altering anxious cognitions per se, either as an exclusive treatment or as an addition to more basic treatments, seems to have little effect on anxiety. As Bandura eloquently pointed out a decade ago, the process of emotional change does not seem to be accomplished through “cognitive interventions,” by which he meant altering self-statements or the like. Simple persuasion is seldom effective. On the other hand, some evidence is beginning to appear that these cognitive components to treatment may serve

The Process of Fear and Anxiety Reduction  103

something of a protective function in terms of preventing relapse (Marshall, 1985). The precise role of cognitive interventions in the anxiety disorders awaits future research, and it may well be that these procedures will prove to be helpful in some aspects of the therapeutic endeavor. It is unlikely, however, that they will be as essential as those described above. Similarly, intervening in social support networks or teaching coping skills may well be useful, particularly in terms of preventing relapse. In some future community mental health endeavor, teaching effective coping skills may even prevent the development of severe anxiety. Nevertheless, it is unlikely that these efforts in themselves will strongly affect the structure of anxious apprehension unless they contribute in a major way to the modification of action tendencies, alterations of focus of attention, or the instilling of a sense of control. Of course, the development of strong and adequate coping procedures will probably have a major effect on a sense of control in the long run. In fact, teaching adequate coping strategies may be one of the most powerful therapeutic approaches available to us (e.g., Goldfried, 1986). But this strategy will probably be useful in the direct modification of anxiety only to the extent that a sense of control is instilled. Finally, as mentioned above, altering physiological or neurobiological responding, which seems to play such an important role in supporting the action tendencies of various emotions, is more or less helpful in some disorders. But this strategy in isolation is not seen as essential. The example of prolonged heightened physiological responding, despite elimination of clinical “anxiety,” has been mentioned above. Almost all individuals, including people without “anxiety,” respond to challenges or stressors with arousal. This arousal serves a very functional purpose in terms of increasing the efficiency of performance, even at high levels, and is seldom accompanied by “anxiety.” Therefore, this is not an essential target of therapeutic change in anxiety disorders. Rather, direct attacks on the action tendencies associated with this supportive physiology seem more crucial. The Process of Fear and Anxiety Reduction There is one exception to this guideline, based on evidence developed in Chapters 4 and 5. Panic attacks seem to “spike off ” a platform of high baseline anxiety and arousal. Therefore, reduction or elimination of somatic responding may reduce the cues that trigger the alarm. But whether this is done through medication or through psychological procedures such as relaxation, it is unlikely that any permanent effect will occur unless the patient also develops a sense of control concerning future false alarms, along with different action tendencies. Otherwise, anxiety and panic will recur as soon as the medication is stopped or during the next stressful event when relaxation or meditation procedures are not immediately available. This is not to say that pharmacological approaches to anxiety may not form a very important component of treatment. Succeeding chapters review the role that these agents can play in successful therapy. The point here is that the long-term beneficial effects of either pharmacological or psychological procedures aimed at reducing arousal, activation, or specific underlying monoaminergic

104  David H. Barlow Table 6.3 Components of any affective therapy

A. Essential targets for change 1. Action tendencies 2. A sense of uncontrollability-unpredictability 3. Self-focused attention B. Helpful but not essential targets for change 1. “Hot” apprehensive cognitions 2. Hypervalent cognitive schemata and attention narrowing 3. Coping skills and social support 4. Elevated physiological responding and altered neurobiological functions

processes associated with anxiety may be initially helpful, but are not necessarily essential. Both essential and helpful components of any affective therapy are outlined in Table 6.3. With these principles in mind, it is now possible to apply these guidelines to individual anxiety disorders—a task that occupies the remainder of this book. First, however, Chapter 9 briefly discusses principles underlying the current nomenclature for specific anxiety disorders as represented in DSM-III-R. New definitions for the anxiety disorders found in this revision of DSM-III are placed in the context of the models of anxiety and panic presented above.

Article 7

A Comparison of Alprazolam and Behavior Therapy in Treatment of Panic Disorder Janet S. Klosko and David H. Barlow Center for Stress and Anxiety Disorders, State University of New York at Albany

Robin Tassinari Center for Stress and Anxiety Disorders, State University of New York at Albany, and Albany Medical College

Jerome A. Cerny Indiana State University

The results of a clinical outcome study (N = 57) comparing behavior therapy directed at panic disorder (panic control treatment [PCT]) with alprazolam were reported. These conditions were compared with a medication placebo and a waiting-list control group. Patterns of results of measures on panic attacks, generalized anxiety, and global clinical ratings reveal that PCT was significantly more effective than placebo and waiting-list conditions on most measures. The alprazolam group differed significantly from neither PCT nor placebo. The percentage of clients completing the study who were free of panic attacks following PCT was 87%, compared with 50% for alprazolam, 36% for placebo, and 33% for the waiting-list group. Since alprazolam may work more quickly than PCT but may also interfere with the effects of behavioral treatment, these data suggest a series of studies on the feasibility of integrating these treatments and on the precise patterns and mechanisms of action of various successful treatment approaches to panic disorder. Recently a large-scale, multicenter study examined the effectiveness of alprazolam for panic disorder (Ballenger et al., 1988). In that study, over 500 patients were assigned randomly to alprazolam or placebo. Assessment measures included prospective self-monitoring of panic attacks. The study yielded a complex pattern of results, but the central finding was that 55% of alprazolam subjects and 32% of placebo subjects who began the study were panic-free following 8 weeks of treatment or at the time they dropped out (endpoint analysis). Because a significantly larger number of placebo subjects dropped out

106  Janet S. Klosko, et al.

(44% of those placebo subjects who completed at least 3 weeks of treatment dropped out before Week 8 compared with 8% of alprazolam subjects), an analysis of those who completed the 8-week study was also reported. Of those subjects who were completers, 59% of alprazolam subjects were free of panic versus 50% of placebo subjects, a nonsignificant difference. Because the latter result may favor placebo by highlighting placebo responders and because the former result may favor alprazolam by underestimating the number of placebo subjects who might have responded if they had stayed in treatment, Ballenger et al. suggested that the true effects of alprazolam fall somewhere between these two methods of reporting results. Several investigators have reported uncontrolled clinical replication series that suggest the effectiveness of behavioral or cognitive–behavioral treatments targeting panic attacks directly rather than agoraphobic avoidance (e.g., Clark, Salkovskis, & Chalkley, 1985). In these uncontrolled reports, from 80% to 100% of clients were free from panic after treatment. In two preliminary controlled studies of behavioral treatments for panic disorder (Barlow et al., 1984; Beck, 1988), clients improved significantly more than waiting-list controls. Now the results from a large-scale treatment outcome study of panic disorder are available (Barlow, Craske, Cerny, & Klosko, 1989). In that study, a newly developed treatment for panic disorder focusing on exposure to somatic sensations associated with panic attacks was evaluated. This exposure procedure, combined with cognitive therapy directed at catastrophic thoughts associated with panic attacks, was compared with a relaxation condition in which clients were taught to use relaxation in a cue-controlled manner whenever they began to feel anxious or panicky. In a third condition, these two treatments were combined. All three conditions were compared with a waiting-list control. Results indicated that over 87% of those who completed treatment and received exposure to somatic sensations plus cognitive therapy, either alone or in combination with relaxation, were free of panic attacks during a 2-week period after treatment. This compares with 60% in the relaxation condition and 35% in the waiting-list control group. Only those groups receiving exposure plus cognitive therapy were significantly different from waiting-list control subjects. However, a significantly greater number of subjects dropped out of the relaxation condition. Therefore, if one includes dropouts in the final analysis, the percentage of patients who were panic-free remains at approximately 80% in the exposure plus cognitive therapy conditions but drops to 40% in the relaxation condition, which is not significantly different from the waiting-list control group. In view of the effectiveness of this behavioral approach and of alprazolam for panic attacks, the purpose of this study is to evaluate the relative effectiveness of each treatment in one setting using a group of clients diagnosed in an identical manner with outcome measured in precisely the same way. Both medication and placebo as well as a waiting-list group were included to provide benchmarks against which to assess therapeutic improvement. This study was seen as a precursor to studying possible integration or coordination of these treatments in panic disorder patients.

A Comparison of Alprazolam and Behavior Therapy  107

Method Subjects

Our subject selection, exclusionary criteria, and general procedures followed closely the detailed protocol developed for the Upjohn cross-national study on panic disorder (Ballenger et al., 1988). Subjects between 18 and 65 years of age were drawn from the pool of patients presenting to the Phobia and Anxiety Disorders Clinic of the Center for Stress and Anxiety Disorders at the State University of New York at Albany. Although most panic disorder patients who qualified for the study presented with no more than mild agoraphobic avoidance, several with moderate to extensive agoraphobic avoidance were included. Each patient was administered the Anxiety Disorders Interview Schedule–Revised (ADIS-R; Di Nardo et al., 1988), a structured interview that has been shown to be a reliable instrument for diagnosing panic disorder (Barlow. 1988). The ADIS-R administrators were advanced clinical psychology graduate students and postdoctoral or licensed clinical psychologists. All had completed extensive training that culminated in requiring trainees to match the diagnoses of experienced interviewers on 3 consecutive patients who were seen by both the trainees and the experienced interviewer. Each patient received a primary diagnosis of panic disorder, with a clinician’s severity rating of at least 4 on a 0–8 scale in which 0 = none, 4 = definitely disturbing/disabling, and 8 = very severely disturbing/disabling. Thus subjects not only had to meet the Diagnostic and Statistical Manual of Mental Disorders (DSM-III; American Psychiatric Association, 1980) definition for a case but also had to present with at least moderate severity. Only subjects who were panicking actively were included; that is, subjects who reported at least one panic attack in the week before starting treatment on the weekly record self-monitoring form (described later). Finally, subjects were between 18 and 65 years of age. Exclusion criteria. Subjects were excluded who had begun pharmacotherapy or psychotherapy in the past 6 months: subjects who had been either in drug or psychotherapeutic treatment more than 6 months were excluded unless they agreed to stop such treatment for the duration of the study. Subjects were excluded who had been on 4 mg or more of alprazolam for any 3-week period and were nonresponders, who displaced evidence of benzodiazepine hypersensitivity, or who had undergone cognitive-behavioral therapy for anxiety at any time. Exclusionary criteria also included (a) females who were pregnant or lactating or at risk to become pregnant; (b) subjects with significant medical problems, as determined by history, medical report, and laboratory values; (c) subjects with a history of psychotic disorder or dementia; (d) subjects with a history of alcohol or other substance abuse within the last 6 months; and (e) subjects with current or past bipolar disorder. Subjects with major depression were excluded only if depression predominated over panic disorder at the time of presentation and if depression preceded panic disorder chronologically. Subjects with acute suicidal ideation were excluded.

108  Janet S. Klosko, et al.

Measures

Psychophysiological, clinical assessment, and medical assessment measures were administered to all subjects before and after treatment. Sell-monitoring measures were administered throughout treatment. Psychophysiological measures and results are reported elsewhere (Klosko, 1987). Self-monitoring measures. Subjects engaged in daily self-monitoring with the weekly record, a form constructed by clinic staff over a period of years with the goal of providing the most useful information about panic attacks while ensuring maximum compliance (see Barlow & Cerny, 1988). Subjects were instructed to record on the diary the following information about each discrete episode of anxiety experienced that day rated 4 or higher on the 0–8 scale: (a) date and time of onset and offset of the anxiety episode, (b) maximum level of anxiety during the episode, (c) whether the subject considered this episode of anxiety a panic attack (in accord with the revised edition of the DSM–III [DSM–III-R; American Psychiatric Association, 1987], subjects were instructed and trained to define a panic attack as the sudden onset of intense fear, accompanied by at least four characteristic panic symptoms; the attack had to peak within 10 min), and (d) whether the subject usually considered the situation or context in which the episode occurred to be anxiety provoking or non-anxiety provoking (in other words, was the episode situational or cued?). Data from self-monitoring served to measure anxiety episodes and panic attacks, both spontaneous and situational. Because subjects were trained to separate panic attacks from episodes of anxiety that began more slowly and typically lasted longer, this procedure ensured a conservative measure of panic. Spontaneous attacks were defined as those that occurred in situations that subjects customarily rated as non-anxiety provoking; situational attacks occurred in situations that subjects customarily rated as anxiety provoking. Subjects were instructed to define anxiety provoking in relative fashion: A crowded grocery store might not produce anxiety for most people but might well produce anxiety for the particular subject. Clinical assessment measures. Clinical assessment measures from the ADIS-R (Di Nardo et al., 1988) included global clinical severity ratings, panic ratings, the Hamilton Anxiety Rating Scale, and the Hamilton Rating Scale for Depression. Medical assessment and procedures. The study psychiatrist obtained medical histories from all subjects and administered a brief physical examination, which consisted of a medical history, vital signs, and examination of head and neck, chest, abdomen, extremities, skin, and neurological signs. If a more thorough physical examination was warranted, the subject was referred to an appropriate physician. Pretreatment and post-treatment blood samples were obtained from subjects in the three treatment groups (alprazolam, placebo, and cognitive-behavioral therapy) for clinical laboratory determinations of medical exclusion criteria, as well as pretreatment and posttreatment plasma benzodiazepine screens. Subjects in the three treatment groups withdrew from prestudy medications under the supervision of the study psychiatrist. Adherence to drug withdrawal

A Comparison of Alprazolam and Behavior Therapy  109

was determined by analyses of plasma benzodiazepine screens. Subjects remained off medication for at least 7 days before administration of psychophysiological, self-report, and self-monitoring measures and random assignment to one of the three treatment groups. Treatment

Alprazolam and placebo treatment groups. Subjects received 15 individual treatment sessions in weekly meetings with a study psychiatrist experienced in alprazolam treatment of panic disorder. Medication was supplied by the Upjohn Company in matching 1-mg tablets, packaged in matching bottles containing sufficient medication for 1 week, and was administered double-blind. Treatment followed the protocol used in the Upjohn multicenter trial (Ballenger et al., 1988), adapted to 15 weeks. Medication was gradually increased following a standardized but flexible schedule until maximum benefit was achieved or dose-limiting side effects occurred. At least three attempts were made to titrate the medication upward to at least 6 mg per day. Dosage was advanced to a maximum of 10 mg per day if required. (See Ballenger et al., 1988, for a more detailed description of drug dose and regimen procedures.) The psychiatrist was instructed to limit interactions with subjects to discussion of clinical history, explanation of panic disorder, discussion of medication effects and side effects, and general support. At the beginning of the 13th week of treatment, the psychiatrist began to taper doses of medication at a rate no faster than one tablet every 3 days. The psychiatrist continued meeting with subjects until they had stopped taking medication completely or, if they were unable to withdraw from medication, until they were restabilized on the study medication once again. At this point treatment was considered to be over, and the psychiatrist completed the post-study termination record. If subjects wished to continue medication use, they were given appropriate referrals. Behavior therapy treatment group. Subjects received 15 individual sessions of an integrated cognitive-behavioral treatment for panic disorder (panic control treatment, PCT) in weekly meetings. Therapists were either PhD psychologists or advanced doctoral students who had been trained in the application of the treatment by observation and practice with corrective feedback. A detailed treatment manual was used, and supervision was provided on a weekly basis to ensure the correct application of therapeutic procedures. Treatment comprised a rationale and education about panic disorder and emphasized exposure to interoceptive (somatic) cues; cognitive approaches, progressive relaxation training, and respiration training (to slow breathing rate) were also included. (The detailed session-by-session treatment protocol is presented in Barlow & Cerny, 1988; an updated protocol suitable for distribution to clients for use under clinical supervision is now available from the Center for Stress and Anxiety Disorders, State University of New York at Albany.) All therapy sessions were tape-recorded and checked for treatment integrity. Waiting-list control group. Subjects were placed on a 15-week waiting list for treatment. They were told that they might contact the clinic by telephone during this time if they felt the need and that we would contact them

110  Janet S. Klosko, et al.

approximately weekly by telephone. Although inclusion criteria required all subjects to have been stabilized on medication, waiting-list subjects were not required to withdraw from medication. Therefore, this waiting-list group might also be considered a minimal treatment condition, thereby providing a more conservative comparison with other treatment conditions. Treatment integrity. The integrity of treatment delivery was examined by means of ratings of the content of 25% of PCT sessions. (All PCT sessions were audiotaped.) All drug treatment sessions were directly observed and rated. For PCT sessions, verbalizations were checked as belonging to one of nine categories, including information and rationale, assigning/discussing behavioral tasks, and so forth. In addition, for both PCT and drug conditions, raters recorded any verbalization that was inappropriate (e.g., off-target, alternative therapeutic techniques). Raters were also asked to determine the particular treatment condition and which of the three phases of treatment yielded the sample. In all cases, raters correctly identified the treatment condition represented by the sample as well as the treatment phase from which the sample came. Across all conditions, there were only two instances of inappropriate material, both of which referred to nontargeted problem areas and not to the application of inappropriate treatment techniques. Thus, in both drug and PCT conditions, treatment was delivered as intended. Results Attrition

Out of 69 initial subjects, 57 subjects completed the study, and 12 subjects dropped out. A higher rate of dropout was observed in the placebo group compared with the other three groups. One subject out of 17 (5.9%) dropped from the alprazolam group, 7 out of 18 (38.9%) from the placebo group, 3 out of 18 (16.7%) from the PCT group, and 1 out of 16 (6.3%) from the waiting-list group. A chi-square analysis on these dropout frequencies was significant, x2(3, N = 69) = 8.75, p < .05. Separate chi-squares on each pair of groups showed significant differences between the placebo and alprazolam groups. x2(1, N = 35) = 3.69, p < .05; the placebo and PCT groups, x2(1, N = 36) = 3.01 p < .05; and the placebo and waiting-list groups, x2(1, N = 34) = 3.45 p < .05. Subjects who dropped from the study were questioned about their reasons. Of the 7 subjects who dropped from the placebo treatment group, 3 reported they disliked the side effects of the study medication, and 3 developed intense panic attacks in the 1st week of treatment that the psychiatrist regarded as intolerable, causing him to drop them from the study (1 worked with dangerous equipment, 1 reported suicidal ideation, and 1 reported she felt unable to continue in her current state). The 7th subject reported she changed her mind about study participation and returned her first bottle unopened. Two of the 3 PCT dropouts decided they did not have the necessary time to devote to therapy. One who was unemployed when the study began started working and dropped out after 3 weeks, and a 2nd subject’s duties increased, causing

A Comparison of Alprazolam and Behavior Therapy  111

her to drop out after 5 weeks. A 3rd wished to resume medication and dropped out after 5 weeks. The waiting-list dropout reported insufficient time to complete the periodic measures and dropped after 8 weeks. Those who dropped from the study were compared with study completers on major pretreatment variables. No major differences emerged on demographic, self-monitoring, medical, or clinical assessment variables related to panic disorder including Hamilton Anxiety and Depression scales. However, a few significant differences did appear in several tangential questionnaires not reported in this manuscript (see Klosko, 1987). Since all placebo subjects dropped from the study before completion of 3 weeks of treatment, endpoint analyses were not conducted (e.g., as were conducted in the Ballenger et al., 1988, study). It is likely that the status of dropouts when they left the study is most adequately represented by their scores on pretreatment assessment measures. To be conservative, other dropouts were also assumed to be unimproved. Pretreatment Characteristics of Treatment Groups

Analyses were conducted on the 57 subjects who completed treatment to describe pretreatment characteristics and to identify pretreatment differences that might have occurred among the four groups. As appropriate, analyses used were either chi-squares (or Fisher’s exact tests) or one-way analyses of variance (ANOVAS) across treatment groups. Demographic characteristics. Fifteen men (26%) and 42 women (74%) completed the study. Mean age was 37 years (SD = 11.04). Thirty-one subjects (54%) were married, 19 (33.3%) were single, 6 (11%) were divorced, and 1 (2%) was separated. One dropout from the placebo group was Black; all other subjects were White. Analyses across groups of all demographic characteristics were nonsignificant. Clinical assessment measures. All clinical assessment measures were derived through administration of the ADIS-R. Table 7.1 displays some diagnostic characteristics including extent of agoraphobic avoidance. Analyses across groups on all diagnostic characteristics including number and type of additional diagnoses were nonsignificant. Subjects endorsed a mean number of DSM-III-R panic symptoms of 9.7 (SD = 2.3). Mean intensity on 0–4 rating scales, where 0 equals none and 4 equals very severe, was 1.5 (SD = .5) for somatic symptoms and 1.8 (SD = .9) for cognitive symptoms. (Symptoms not endorsed were assigned an intensity of 0, accounting for relatively low severities.) Mean Hamilton Anxiety Rating Scale score was 17.8 (SD = 5.4); mean Hamilton Rating Scale for Depression score was 12.44 (SD = 6.3). Analyses across groups on all these clinical assessment measures were nonsignificant. Medication use. Table 7.1 also displays prestudy medication use. (One should note that, in the PCT treatment group, one subject reported use of both benzodiazepines and beta blockers, and in the waiting-list group, 3 subjects reported use of more than one medication.) Chi-square analyses were conducted on frequencies of use of each medication and use of medication generally. Significant results were obtained for use of benzodiazepines. x2(3,

4 0 0 0 4

68.8 00.0 00.0 00.0 68.8ab

11 0 0 0 11

3 7 1

5.36 .674

31.3 56.3 12.5

5 9 2

n

5.38 .885

%

n

36.4 00.0 00.0 00.0 36.4a

27.3 63.6 9.1

%

Placebo (n = 11)

Note. Groups with different subscripts are significantly different, p < .05.

Variable Diagnosis Panic disorder, uncomplicated Panic disorder, mild avoidance Panic disorder, moderate to extensive avoidance Global Clinical Severity ratings of panic disorder diagnoses M SD Medication use Benzodiazepines Antidepressants Beta blockers Antipsychotics Total medication use

Alprazolam (n = 16)

6 0 1 0 6

5.00 .845

2 13 0

n

PCT (n = 15)

Table 7.1  Pretreatment diagnostic characteristics and medication use of treatment groups

40.0 00.0 6.7 00.0 40.0a

13.3 86.7 00.0

%

12 2 2 0 13

5.47 .990

2 11 2

n

Waiting list (n = 15)

80.0 13.3 13.3 00.0 86.7b

13.3 73.3 13.3

%

33 2 3 0 34

5.30 .865

12 40 5

n

Total (N = 57)

57.9 3.5 5.3 00.0 59.6

21.1 70.2 8.8

%

A Comparison of Alprazolam and Behavior Therapy  113

N = 57) = 7.84, p < .05, and use of medication generally, x2(3, N = 57) = 7.59, p < .05. For use of benzodiazepines, separate chi-squares (or Fisher’s exact tests) on each pair of groups, including the waiting-list versus the placebo groups, showed no significant differences; for use of medication generally, separate chi-squares (or Fisher’s exact tests) showed significant differences between the PCT and waiting-list groups, x2(1, N = 30) 5.17, p < .05, and placebo and waiting-list groups, x2(1, N = 26) = 5.04, p < .05. Results were nonsignificant for antidepressants, beta blockers, and antipsychotics. Subjects from the three treatment groups (alprazolam, placebo, and PCT) were withdrawn from current medications as part of pretreatment medical procedures. They remained medication-free for approximately 1 week and were administered plasma benzodiazepine screens to ensure compliance with medication withdrawal before treatment began. Of the 42 subjects who were administered plasma benzodiazepine screens, no medication was detected in 37 (88.1%). Four subjects (9.5%) gave samples in which desmethyldiazepam was detected. Because this substance can remain in the blood longer than 1 week after discontinuation, these subjects were permitted to start treatment. One blood sample from a therapy subject was lost. However, this subject reported no medication use pretreatment, and her posttreatment sample was clean. Chi-square analyses of pretreatment samples were all nonsignificant. Self-monitoring measures. In the 2-week period pretreatment, on the weekly record, subjects reported a mean number of panic attacks per week of 2.0 (SD = 1.9). Mean number of spontaneous attacks was 1.1 (SD = 1.4). Mean intensity of attacks, on a 0–8 scale on which 0 equals none, 4 equals moderate, and 8 equals as much as one can imagine, was 4.5 (SD = 2.2). Subjects reported a mean of 1.9 (SD = 1.9) anxiety episodes in the 2-week pretreatment period. Mean intensity was 3.39 (SD = 2.3). One-way ANOVAS on frequency and intensity of panic attacks and anxiety episodes across groups were all nonsignificant. Posttreatment Assessment General Approach

The strategy for identification of pretreatment differences among groups, with separate chi-squares or ANOVAS on each pretreatment measure, represented a liberal approach, and although a large number of tests were conducted, the alpha level remained at .05. Nevertheless, few pretreatment differences were uncovered. In view of this pretreatment equivalence of groups, the general strategy for analyzing posttreatment data was administration of multivariate analyses of variance (MANOVAS) on posttreatment measures. In an effort to account for baseline (pretreatment) values, repeated-measures MANOVAS were also conducted on all analyses. Since the pattern of results was identical in each instance, we have chosen to present results of MANOVAS on posttreatment variables, because this approach provides the most straightforward identification of post hoc differences among pairs of means. Clinical assessment, medical assessment, self-monitoring, and psychophysiological measures were

114  Janet S. Klosko, et al.

analyzed separately. Within each type of measure, variables were grouped thematically. Examples of the various themes included global panic disorder severity, panic, general anxiety, and depression. To explore further the nature of MANOVAS with statistically significant results, univariate ANOVAS were conducted upon each dependent variable included in the analysis. We performed post hoc comparisons among all pairs of means, using Duncan’s Multiple Range Test (MRT; p < .05). Dose

Alprazolam was increased rapidly to maximum average daily dose of 4.60 mg (SD = 1.82). Although every attempt was made to reach 6.0 mg, in some cases the clinical response was satisfactory at a lower dose; in others, side effects precluded higher dosage. Maximum average daily dose of placebo was somewhat higher at 5.08 (SD = 2.65). Posttreatment Clinical Assessment Measures

Posttreatment clinical assessment measures were gathered through administration of a short form of the ADIS-R. The ADIS-R administrators were blind to group assignment. Global clinical ratings. Clinical ratings of severity of panic disorder, assigned by ADIS administrators before and after treatment, represent the most global study measure. Ratings were based on levels of psychological distress and interference in functioning produced as a function of panic disorder. The top of Table 7.2 presents posttreatment global clinical ratings. A one-way ANOVA on posttreatment clinical ratings was significant, F(3, 53) = 4.12, p < .01. Results of Duncan’s MRT indicated that the alprazolam and PCT treatment groups were significantly more improved than the waiting-list group. To conduct a test of end state functioning, study completers were grouped according to whether they had obtained posttreatment global clinical ratings of clinical versus nonclinical severity as defined on the ADIS. They were grouped as subjects with low end state functioning if they obtained ratings greater than or equal to 4 (clinical severity) or as subjects with high end state functioning if they obtained ratings less than 4 (nonclinical severity). Approximately half of study completers obtained high end state functioning. The next part of Table 7.2 displays the number and percentage of subjects with low and high end state functioning. A chi-square analysis of these frequencies was significant, x2(3, N = 57) = 8.62, p < .05. Separate chi-squares showed the PCT group was significantly different from the waiting-list group, x3(1, N = 30) = 6.56, p < .01. Since the placebo group had a disproportionate number of dropouts, it is reasonable to argue that analysis of end state functioning that includes only study completers represents a distortion of results. Given the reasons and the rapidity (within the first 3 weeks) with which most subjects dropped from the study, it is likely that, at time of study withdrawal, dropouts maintained their pretreatment low end state functioning status. The bottom of Table 7.2 presents subjects who obtained high versus low end state functioning, with

A Comparison of Alprazolam and Behavior Therapy  115

dropouts included. A chi-square analysis of these frequencies was significant, x2(3, N = 69) = 7.86, p < .05. Separate chi squares showed once again that the PCT group was significantly different from the waiting-list group, x2(1, N = 34) = 4.65, p < .05. Measures of general anxiety. Univariate ANOVA on Hamilton Anxiety Rating Scale scores was significant, F(3, 53) = 3.19. p < .05. Results of Duncan’s MRT showed that the PCT group was significantly more improved than the waiting-list group. Means and standard deviations of pretreatment and posttreatment scores on the Hamilton Anxiety Rating Scale were as follows: alprazolam mean, pretreatment = 18.75 (SD = 5.74), posttreatment = 13.68 (SD = 7.02); placebo mean, pretreatment = 17.36 (SD = 5.59), posttreatment = 13.36 (SD = 5.63); PCT mean, pretreatment = 16.93 (SD = 5.43), posttreatment = 9.80 (SD = 5.31); and waiting-list mean, pretreatment = 18.07 (SD = 5.22), posttreatment = 16.27 (SD = 4.71). Posttreatment Medication Use Measures

As noted, alprazolam and placebo subjects who were able to withdraw completely from medication did so; subjects who experienced difficulty were permitted to resume stable dosage, at or near levels they received during the study, before

Table 7.2 Posttreatment clinical assessment measures of treatment groups: global clinical ratings

Variable Sample size (n) Completers Total Disorder severitya M SD Endstate nonclinical severityb Completers n % Total n %

Alprazolam

Placebo

PCT

Waiting list

Total

16 17

11 18

15 18

15 16

57 69

3.56a 1.90

3.55ab 1.51

2.73a 1.53

4.80b 1.47

3.67 1.76

8 50.0

5 45.5

11 73.3b

3 20.0b

27 47.4

8 47.1ab

5 27.8ab

11 61.1a

3 18.8b

27 39.1

Note. PCT = panic control treatment. Subscripts indicate that values differed significantly at p < .05. a Panic disorder severity was assessed using the Anxiety Disorder Interview Schedule–Revised (ADIS-R). b End state functioning was assessed on the ADIS-R by assigning ratings of clinical or nonclinical severity.

116  Janet S. Klosko, et al.

posttreatment assessment. There were several justifications for this method. It soon became apparent that many subjects were unwilling to withdraw from their study medication. Rather than continue tapering indefinitely for these subjects, tapering was stopped. It seemed methodologically unsound to conduct posttreatment assessments on medication subjects while they were in states of withdrawal; hence, subjects who could not withdraw were stabilized once again before undergoing assessments. In fact, only 1 out of 16 subjects withdrew completely from alprazolam. The remaining 15 were quickly stabilized at or near their study dosage level. In contrast, 2 subjects out of 11 in the placebo group were “unable to withdraw” and were stabilized at study dosage levels. Subjects were not pressured to withdraw from medication. Subjects who experienced even mild difficulty were permitted to resume their study dosage levels very quickly. No subject exhibited extreme withdrawal symptoms. Thus, it is unlikely that attempts to taper alprazolam exerted any negative effect upon their scores on posttreatment assessment measures. This assertion is supported by analyses of the panic data preceding attempts to withdraw subjects from their study medication compared with posttreatment. Panic measures were taken from the last available weekly record before patients were instructed to begin taper withdrawal. Subjects in the alprazolam group had a mean number of panic attacks of .94 (SD = 1.61) and a mean intensity of panic of 2.06 (SD = 2.91). These figures are slightly higher than the mean number and intensity of panics for alprazolam subjects posttreatment (see Table 7.3). Approximately 63% of alprazolam subjects were panic-free before taper withdrawal. Therefore, before taper withdrawal, slightly fewer alprazolam subjects were experiencing panic attacks in comparison with posttreatment, but those subjects who were still experiencing panic were having more frequent and severe attacks in comparison with posttreatment. Two procedures were used to determine posttreatment medication use: self-report during posttreatment ADIS administration and collection of plasma benzodiazepine screens for tests of medication content. The latter procedure applied to subjects in the alprazolam, placebo, and PCT treatment groups only. Posttreatment, 29 subjects reported medication use. The reports were consistent with protocol with the exception of 2 placebo subjects who reported taking benzodiazepines despite plasma benzodiazepine screens that confirmed they were on placebo. These were the 2 subjects who also were unable to withdraw from placebo and were restabilized and thus were the only 2 placebo subjects still taking drugs. Of the 42 subjects who completed one of the three active treatments, 28 had posttreatment plasma benzodiazepine screens. Blood tests were missing or not available for 14 subjects. No deviations from protocol were revealed through plasma benzodiazepine screens. Psychiatrist ratings of medication effectiveness. As part of posttreatment assessment, the psychiatrist rated the effectiveness of the study medication for each medication group subject through use of the poststudy termination record. While remaining blind to group assignment, the psychiatrist compared the study medication with “usual drug treatment of this disorder.” Ratings ranged from noneffective (1) to much more effective (6). The mean rating the psychiatrist assigned to subjects in the alprazolam group was 4.13 (SD = .72);

Table 7.3 Posttreatment clinical and self-monitoring measures of panic and panic frequency in each group

Measure

Alprazolam

Placebo

PCT

Waiting list

Total

n = 16

n = 11

n = 15

n = 15

(N = 57)

Clinical ratings Number of symptoms endorsed M SD Intensity of somatic symptoms M SD Intensity of cognitive symptoms M SD

6.06ab 4.28

6.90ab 3.11

4.40a 2.82

7.64b 2.02

6.20 3.34

1.03a .65

.89ab .46

.56b .52

1.15a .39

.91 .55

.67ab 1.04

.085ab .50

.47a .44

1.29b 1.03

.91 .55

Self-monitoring of panic attacks Frequency per week: total M SD Frequency per week: spontaneous M SD Intensity M SD

.51a .81

.56ab .85

.20a .65

1.72b 2.73

.76 1.62

.24 .63

.32 .60

.03 .13

.65 .96

.31 .68

1.66a 2.08

1.78ab 2.04

.71a 1.99

3.53b 3.01

1.92 2.50

Frequency of panic attacks Experienced zero panic attacks n % Experienced zero spontaneous attacks n % Experienced zero cued (situational) attacks n %

8 50.0ab

4 36.4a

13 86.7b

5 33.3a

30 52.6

10 62.5

6 54.5

13 86.7

6 40.0

35 61.4

9 56.3

7 63.6

14 93.3

10 66.7

40 70.2

Note. Groups with different subscripts are significantly different at p < .05.

118  Janet S. Klosko, et al.

the mean rating he assigned to subjects in the placebo group was 2.73 (SD = 1.10). A one-way ANOVA on ratings of medication effectiveness was significant, F(1, 25) = 15.97, p < .001. Panic ratings. A MANOVA was conducted on posttreatment panic ratings from the ADIS-R, with number of symptoms endorsed, mean intensity of somatic symptoms, and mean intensity of cognitive symptoms as dependent measures. It was significant, Pillais F(9, 159) = 2.14, p .25 for 8 of 9 omnibus PDSS analyses of covariance, suggesting little heterogeneity. Subject Disposition, Dosing, and Laboratory Analyses

Acute completion rates did not differ significantly across treatment groups. Among the 8 noncompliant patients was 1 patient receiving imipramine alone who violated protocol by using benzodiazepines at rates exceeding those

Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder  131 Table 9.1 Baseline analyses*

Variable

CBT Alone Imipramine Placebo CBT Plus CBT Plus Total† Alone Alone Imipramine Placebo

No. of subjects Female sex, % Age, mean (SD), y White race, % Currently married, % Duration of illness, mean (SD), y PDSS average item score, mean (SD)§ Current major depression, %

77 63.2 37.5 (10.9) 89.0 52.1

83 60.2 35.5 (9.7) 89.0 45.1

24 75.0 34.2 (9.7) 91.3 39.1

65 64.1 34.1 (11.4) 95.3 48.4

63 58.1 37.8 (11.3) 90.3 58.1

312 62.5 36.1 (10.7) 90.8 49.7

P Value‡

.67 .23 .71 .45

 6.61  6.38 (8.55) (7.54)

 6.64   (10.4)

 5.50  6.33 .96 (8.17) (8.43)

   1.88 (0.56)

 1.86  1.86 (0.52) (0.57)

 1.74  1.83 .58 (0.51) (0.54)

23.7

29.2

27.0

30.1

26.6

27.1

.92

CBT indicates cognitive-behavioral therapy; placebo, pill placebo plus medical management; and PDSS, Panic Disorder Severity Scale. No. of subjects for each cell may vary slightly from the number of subjects in each treatment arm due to occasional missing data (exact numbers available on request).

*

Total of all 5 treatments combined.

†

P value for omnibus analysis of variance (for continuous measures) or 5 × 2 Freeman-Halton exact test (for categorical measures). ‡

A higher PDSS score indicates greater severity.

§

allowed, and 1 patient receiving CBT who began nonprotocol antidepressant treatment during acute treatment. The figure also shows that 82% to 90% of maintenance-eligible patients completed maintenance across active treatments, compared with 33% of placebo patients. The omnibus test for this analysis was significant. Pairwise comparisons were significant for all active treatments vs placebo, including CBT alone (P = .006), imipramine alone (P = .007), CBT+imipramine (P = .001), and CBT+placebo (P = .004). Patients receiving medication reported taking doses of 175 to 180 mg/d across treatments by week 6 and 214 to 239 mg/d by week 12. Mean (SD) imipramine+desipramine plasma levels at week 6 were 219 (186) ng/dL and 223 (127) ng/dL for CBT+imipramine and imipramine alone, respectively, and 225 (148) ng/dL and 263 (156) ng/dL at week 12. Rates of urine samples that tested positive for benzodiazepines were low. At week 6, 3 (1.5%) of 197 samples tested positive (1 CBT+imipramine, 1 CBT, 1 imipramine), and at week 12, 3 (1.8%) of 164 (1 CBT+imipramine, 1 CBT, 1 imipramine) tested positive. Rates of missing urine samples were not significantly different across treatments.

132  David H. Barlow, et al.

CBT Alone and Imipramine Alone vs Placebo. In the acute ITT analysis, both CBT alone and imipramine alone were superior to placebo for the PDSS continuous measure (Table 9.2). Both treatments had significantly fewer dropouts for lack of efficacy than did placebo (4/18 for CBT and 10/30 for imipramine vs 8/10 for placebo [Figure 9.1]), and the imipramine group had significantly more dropouts for adverse effects than did the placebo group. Maintenance ITT analysis (Table 9.2) confirms the superiority of both CBT alone and imipramine alone to placebo. CBT vs Imipramine. We found no significant difference between CBT alone and imipramine alone in the acute ITT or completer analyses or in the maintenance ITT or ITC analyses (Table 9.2). As expected, significantly more patients in the imipramine group than in the CBT group dropped out because of adverse effects (Figure 9.1). The follow-up analyses (Table 9.2) show trends favoring CBT over imipramine. Combined CBT+Imipramine vs Single Treatment. We hypothesized that CBT+imipramine would be better than all 3 of the relevant comparison groups. Table 9.2 shows that CBT+imipramine was superior to CBT alone on 1 of 3 ITT analyses and 1 of 3 completer analyses but was not superior to CBT+placebo. In the maintenance ITT analysis (Table 9.2), combined treatment was better in the PDSS average analysis than in all 3 comparison treatments (there by meeting our criteria for superiority) and better than CBT and CBT+placebo on ITC analyses. Intent-to-continue analyses showed that responders to imipramine, with or without CBT, fared significantly worse in the no-treatment follow-up period than those who received either CBT alone or CBT+placebo (Table 9.2). After treatment discontinuation in the follow-up ITT analyses, the treatments that continued to show evidence of superiority to placebo were CBT alone (Table 9.2) and CBT+placebo (statistically significant for all 3 measures). Quality of Response

In a secondary analysis restricted to responders based on the CGI definition (Table 9.3), acute imipramine responders had significantly lower PDSS average scores than acute CBT responders, indicating a higher quality of response. Maintenance responders showed the same pattern at the trend level. Responders to combined CBT+imipramine had higher-quality responses than CBT responders at acute and maintenance points, as well as a higher quality of response than patients taking CBT+placebo at maintenance. Timing of Loss of Response

Timing of loss of response could be determined in 4 of 5 imipramine followup completers, losing response during months 3, 4, 5, and 6 (1 case each) and in 1 of 2 CBT follow-up completers, losing response during month 3. Among 9 CBT+imipramine follow-up completers losing response, relapse occurred during months 2, 4, and 5 (2 cases each), months 1, 3, and 6 (1 case each), and in the CBT+placebo relapser during month 1.

83

Intentionto-continue in maintenance

312

.13

64.1

61.9

57.1

41

39

170

NA

79.5 79.5

73.2

CGI response rate, % 78.0

PDSS response rate, % 37.5

37.5

87.8

90.0

82.1

76.3

79.6

77.7

.06

NA

.03

9

.10

.02

.03

40

54.8

50.0

.12

41

37.5

60.3

PDSS average item 0.76 (0.77) 0.54 (0.72) 1.03 (0.84) 0.29 (0.60) 0.67 (0.67) 0.60 (0.71) .005 score, mean (SD)§

No. of subjects

48.2

CGI response rate, % 53.9

21.7

Maintenance Treatment Analyses

45.8

48.7

NA

.09

.32

NA

.03

.05

63

80.6

PDSS response rate, %

65

86.7

.009 .02

24

89.1

73.7

PDSS average item 1.14 (0.74) 1.05 (0.77) 1.52 (0.90) 0.88 (0.74) 0.99 (0.70) 1.06 (0.76) .003 score, mean (SD)§

77

No. of subjects

Acute intentionto-treat

64.3

80.0

NA

78.4

CGI response rate, % 74.5

84.4 NA

.06

.02

38.5

.01

74.5

67.3

PDSS response rate, %

213 .17

45 .03

Acute Treatment Analyses 51 14 47

0.95 (0.65) 0.75 (0.65) 1.15 (0.86) 0.60 (0.61) 0.72 (0.62) 0.79 (0.66) .003

NA

.60

.16

NA

.75

.41

NA

.52

.13

NA

.13

.006

NA

.86

.34

NA

.78

.22

.32

.10

.15

.22

(Continued )

NA NA

.08

.001 .07

NA NA

.18

.02

NA NA

.06

.002 .23

P Value I vs P C vs P I vs C C + I vs C+I C+I vs C + P‡ vsC‡ I‡

PDSS average item score, mean (SD)

Total

56

CBT Plus CBT Plus Imipramine Placebo

No. of subjects

Imipramine Placebo Alone Alone

Acute completers

CBT Alone

Measure

Analysis

Table 9.2 Treatment and follow-up analyses*

Measure

CBT Alone

Imipramine Placebo Alone Alone

CBT Plus CBT Plus Imipramine Placebo

Total

100

100 51.7

83.3

83.3 70.4

70.5 .02

.01

50.0

116

60.0

30

60.0

30

.003

.003

CGI response rate, % 82.1

3

43.3

42.2

PDSS response rate, % 85.2

25

50.0

46.8

1.11 (1.02) 0.53 (0.70) 0.71 (0.87) .02

28

56.3

Follow-up Analyses

13.0

57.1

PDSS average item 0.56 (0.72) 0.81 (0.90) –0.11 (0) score, mean (SD)§

No. of subjects

37.8

CGI response rate, % 42.1

13.0

19.7

CGI response rate, % 31.9

13.0

9.1 26.3

25.0 41.0

41.0

28.0

27.6

.03

.01

.55

.34

NA

.53

.52

.20

.02

.09

.09

.05

.12

NA

.11

.007

.59

.28

.04

.12

.08

NA

.01

.06 .01

.23

.63

.87

NA

1.00

1.00

.10

.01

.02

.02

.72

.30

.03

.03

.59

.56

.43

.41

.41

NA NA

.02

.009 .58

.01

.13

.04

.004 .01

§

Baseline adjusted means.

‡

All 3 comparisons of C + I > C + P, C + I > C, and C + I > I were required to indicate superiority of combined treatment.

† Pairwise comparisons are for 2-tailed Fisher exact tests; analyses were for continuous data, as per first footnote. Boldface indicates P values significant at ≤.05; gray tint indicates therapy listed on top is better than therapy listed below (consistent with hypotheses); boxed values indicate bottom therapy is better than top therapy (counter to hypotheses) at either significant or trend levels. Exact P values are shown, rounded to 2 significant digits when P ≥ .01, otherwise to 3 significant digits. NA indicates post hoc pairwise comparisons not applicable because of nonsignificant omnibus test results.

*See first 3 footnotes to Table 9.1. P value for omnibus analysis of the continuous measure used analysis of covariance with baseline as covariate or repeated-measures analysis of variance when analysis of covariance assumptions were not met. CGI indicates Clinical Global Impression Scale; I, imipramine; P, placebo; C, CBT; C + I, CBT + imipramine; and C + P, CBT+ placebo.

19.7

PDSS response rate, % 32.4

Follow-up No. of subjects 73 77 24 59 62 295 intention-to-treat PDSS average item 1.33 (0.93) 1.45 (0.83) 1.62 (0.77) 1.45 (0.90) 1.18 (0.92) 1.37 (0.89) .12 score, mean (SD)§

Intentionto-continue in follow-up

37.8

PDSS response rate, % 39.5

.05

.04

P Value I vs P C vs P I vs C C + I vs C+I C+I vs C + P‡ vsC‡ I‡

Maintenance No. of subjects 77 83 24 65 63 312 intention-to-treat PDSS average item 1.18 (0.86) 1.14 (0.87) 1.54 (0.83) 0.78 (0.86) 1.08 (0.79) 1.09 (0.86) .001 score, mean (SD)§

Analysis

Table 9.2 (Continued)

40

9

Placebo Alone 41

CBT Plus Imipramine 39

CBT Plus Placebo 170

Total

32

31

3

36

31

133

15

3

15

25

81

0.26 (0.30) 0.15 (0.21) –0.05 (0.0) 0.20 (0.26) 0.20 (0.24) 0.19 (0.25) .39

23

*First 4 footnotes to Table 9.2 apply.

    Follow-up

No. of subjects

Pairwise Comparisons

.37

.66

NA NA

.81

.08

NA

.09

.01

NA

.006

.34

NA

.001

.01

NA

.11

.83

P Value I vs P C vs P I vs C C + I vs C+I vs C C+I vs I C+P

0.69 (0.41) 0.47 (0.45) 0.77 (0.58) 0.48 (0.50) 0.56 (0.43) 0.58 (0.45) .03

41

Imipramine Alone

    Maintenance 0.49 (0.43) 0.32 (0.42) 0.26 (0.50) 0.19 (0.33) 0.45 (0.42) 0.35 (0.41) .02

No. of subjects

    Acute

No. of subjects

CBT Alone

Mean (SD)

Table 9.3  Analysis of PDSS average Item score for responders*

136  David H. Barlow, et al.

Comment Our results demonstrate that both imipramine and CBT are better than pill placebo for treatment of PD. Imipramine produced a superior quality of response, but CBT had more durability and was somewhat better tolerated. In our study, ITT placebo response did not differ from active treatment on acute-phase assessment when CGI was used to determine responder status. By contrast, the 7-item PDSS successfully discriminated between conditions. Moreover, the placebo response in the ITT analysis of 37.5% based on CGI criteria after 3 months drops to 13% after 9 months of treatment. Several studies have made similar observations in the treatment of depression and PD.38–41 While attrition in the placebo group may have compromised comparisons at the end of maintenance, placebo response is weak in magnitude and transient in duration.1,24 There are several differences between active treatments. Although no differences emerged on a priori planned completer or ITT analyses, patients treated with imipramine and designated responders based on CGI criteria following the acute phase showed significantly more improvement on the PDSS than patients who responded to CBT. A trend level of significance remained at the end of maintenance. Thus, among those who did well with either treatment, patients receiving imipramine responded more completely. However, at follow-up, patients who had received CBT alone maintained their improvement significantly better (4% relapse) than those treated with imipramine (25% relapse) based on PDSS responder criteria. We discontinued imipramine by tapering during a 1- to 2-week period, following standard practice at the time. Relapses in medication-treated patients appeared to be evenly distributed across follow-up, suggesting that withdrawal played little role in the results. We did not aim to determine optimal tapering strategy or to ascertain the duration of medication maintenance that produces the best long-term outcome. Our results indicate the need for such work. Findings of high acceptability and durability of CBT are consistent with previous reports,16 although attrition in the CBT-alone group was higher than reported previously.6,14,15,42 Adverse effects with imipramine and relapse following discontinuation are also consistent with previous reports.38–40 However, our results that show a superior quality of response with imipramine among responders to both treatments in the acute phase underscore the need to reduce attrition and develop optimal maintenance and discontinuation procedures for those receiving medication. Acute coadministration of imipramine and CBT resulted in limited benefit over monotherapy. Adding medication to CBT achieved significantly better results than CBT alone at postacute assessment on some measures, but this combination was never better than CBT+placebo. By the end of maintenance, CBT+imipramine was superior to both CBT alone and CBT+placebo (as well as imipramine alone) on the PDSS average measure. However, this robust combination treatment produced the highest relapse rate at follow-up assessment. Surprisingly, the addition of CBT to imipramine did not mitigate relapse following medication discontinuation; addition of imipramine appeared to reduce the long-term durability of CBT. More work is needed to elucidate this result. A selection effect may account for the relatively poor

Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder  137

outcome of combined treatment after discontinuation. Combined treatment could conceivably select patients who had been in particular need of longterm treatment from the beginning. We have not been able to detect important differences at baseline on variables in Table 9.1 between patients who did and did not enter the follow-up phase, but these analyses do not exclude the possibility of selection on an unmeasured prognostic factor. There are several limitations of this study. First, to avoid the complications and possible confounding factors of adding exposure-based interventions to each condition, we enrolled patients with only limited degrees of phobic avoidance, and results are generalizable only to this group. Second, it could be argued that we underestimate the benefits of medication by using a tricyclic antidepressant instead of an SSRI. Current recommendations43 (not yet in place when this study began) consider SSRIs to be the first-line medication. While there is little question that SSRIs are more convenient and have a more limited adverse-effect profile, studies examining differences from tricyclic antidepressants do not consistently find higher efficacy for SSRIs. In fact, of 4 randomized studies, none found significant differences in the end-of-study acute ITT analyses.44–47 One study found evidence for a more rapid response for the SSRI,44 and another found that the tricyclic antidepressant but not the SSRI was more effective than placebo.46 Moreover, there have also been refinements in CBT since we began our study. Thus, we believe it likely that results of a similar comparative study using an SSRI would not differ substantially from ours. Third, at the time we designed this study, a consensus existed that an adequate dose of imipramine should be at least 200 mg/d. A recent study, however, suggests that imipramine/desipramine plasma levels of about 150 mg/ mL may be optimal in treating PD.48 Hence, it is conceivable that we underestimate the therapeutic potential of imipramine in this study. Finally, the use of nonstudy medication is a possible confounding factor to any anxiety study. We made the decision to permit limited use of benzodiazepines, because we sought to simulate clinical practice. Rates of urine samples that tested positive for benzodiazepine use among the 5 treatments were equivalent and low. Only 1 subject’s data were censored because of excessive benzodiazepine use. Thus, we believe benzodiazepine use did not play a significant role in our results. This study represents, to our knowledge, the first multicenter trial comparing medication and psychosocial therapies and their combination for PD. Prior studies contrasting the 2 approaches have been criticized because of possible investigator-allegiance bias in study design, implementation, and/or analysis. Our study sites included 2 with psychotherapy and 2 with pharmacotherapy expertise. In this context, the absence of site differences in ITT outcome supports the important implication that both types of treatment should be transportable to most clinical settings and confirms the generalizability of the results. References 1 Keller MB, Yonkers KA, Warshaw, MG, et al. Remission and relapse in subjects with panic disorder and panic with agoraphobia. J Nerv Ment Dis. 1994; 182:290–296. 2 Robins LN, Regier DA, eds. Psychiatric Disorders in America: The Epidemiologic Catchment Area Study. New York: The Free Press; 1991.

138  David H. Barlow, et al. 3 Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Arch Gen Psychiatry. 1994; 51:8–19. 4 Sherbourne CD, Wells KB, Judd LL. Functioning and well-being of patients with panic disorder. Am J Psychiatry. 1996; 153:213–218. 5 Klerman GL, Weissman MM, Ouellette R, et al. Panic attacks in the community. JAMA. 1991; 265:742–746. 6 Katon W. Primary care–psychiatry panic disorder management module. In: Wolfe BE, Maser JD, eds. Treatment of Panic Disorder: A Consensus Development Conference. Washington, DC: American Psychiatric Press Inc; 1994:4–56. 7 Beitman BD, Thomas AM, Kushner MG. Panic disorder in the families of patients with normal coronary arteries and non-fear panic disorder. Behav Res Ther. 1992; 30:403–406. 8 Walker EA, Roy-Byrne PP, Katon WJ, et al. Psychiatric illness in irritable bowel syndrome. Am J Psychiatry. 1990; 147:1656–1661. 9 Katerndahl DA, Realini JP. Where do panic attack sufferers seek care? J Fam Pract. 1995; 40:237–243. 10 Swinson RP, Cox BJ, Woszczyna CB. Use of medical services and treatment for panic disorder with agoraphobia and for social phobia. CMAJ. 1992; 147:878. 11 Leon AC, Portera L, Weissman MM. The social costs of anxiety disorders. Br J Psychiatry Suppl. 1995; 166:19–22. 12 Salvador-Carulla L, Segui J, Fernandez-Cano P, Canet J. Costs and offset effect in panic disorders. Br J Psychiatry Suppl. 1995; 166:23–28. 13 Hofmann S, Barlow DH. The costs of anxiety disorders. In: Miller NE, Magruder, KM, eds. Cost-Effectiveness of Psychotherapy: A Guide for Practitioners, Researchers and Policymakers. New York: Oxford University Press Inc; 1999:224–234. 14 Barlow DH, Craske MG, Cerny JA, Klosko JS. Behavioral treatment of panic disorder. Behav Ther. 1989; 20:261–282. 15 Clark DM, Salkovskis PM, Hackmann A, et al. A comparison of cognitive therapy, applied relaxation and imipramine in the treatment of panic disorder. Br J Psychiatry. 1994; 164:759–769. 16 Barlow DH, Lehman CL. Advances in the psychosocial treatment of anxiety disorders. Arch Gen Psychiatry. 1996; 53:727–735. 17 Zitrin CM, Klein DF, Woerner MG. Treatment of agoraphobia with group exposure in vivo and imipramine. Arch Gen Psychiatry. 1980; 37:63–72. 18 Zitrin CM, Klein DF, Woerner MG, Ross DC. Treatment of phobias, I: comparison of imipramine hydrochloride and placebo. Arch Gen Psychiatry. 1983; 40:125–138. 19 Mavissakalian M, Michelson L. Two-year follow-up of exposure and imipramine treatment of agoraphobia. Am J Psychiatry. 1986; 143:1106–1112. 20 Marks IM, Gray S, Cohen D, et al. Imipramine and brief therapist-aided exposure in agoraphobics having self-exposure homework. Arch Gen Psychiatry. 1983; 40:153–162. 21 Mavissakalian MR, Perel JM. Imipramine treatment of panic disorder with agoraphobia. Am J Psychiatry. 1995; 152:673–682. 22 Lydiard RB, Brawman-Mintzer O, Ballenger JC. Recent developments in the psychopharmacology of anxiety disorders. J Consult Clin Psychol. 1996; 64:660–668. 23 Liebowitz MR, Barlow DH. Panic disorder: the latest on diagnosis and treatment. J Pract Psychiatry Behav Health. 1995; 1:10–19. 24 Wolfe BE, Maser JD. Origins and overview of the consensus development conference on the treatment of panic disorder. In: Wolfe BE, Maser JD, eds. Treatment of Panic Disorder: A Consensus Development Conference. Washington, DC: American Psychiatric Press Inc; 1994:3–16.

Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder  139 25 Mavissakalian MR. Antidepressant medications for panic disorder. In: Mavissakalian MR, Prien RF, eds. Long-term Treatment of Anxiety Disorders. Washington, DC: American Psychiatric Press Inc; 1996:265. 26 Black DW, Wesner R, Bowers W, Gabel J. A comparison of fluvoxamine, cognitive therapy, and placebo in the treatment of panic disorder. Arch Gen Psychiatry. 1993; 50:44–50. 27 Marks IM, Swinson RP, Basoglu M, et al. Alprazolam and exposure alone and combined in panic disorder with agoraphobia. Br J Psychiatry. 1993; 162:776–787. 28 Oehrburg S, Christiansen PE, Behnke K, et al. Paroxetine in the treatment of panic disorder. Br J Psychiatry. 1995; 167:374–379. 29 Sharp DM, Power KG, Simpson RJ. Fluvoxamine, placebo, and cognitive behaviour therapy used alone and in combination in the treatment of panic disorder and agoraphobia. J Anxiety Disord. 1996; 10:219–242. 30 Woods SW, Sholomskas DE, Shear MK, et al. Efficient allocation of patients to treatment cells in clinical trials with more than two treatment conditions. Am J Psychiatry. 1998; 155:1446–1448. 31 Di Nardo PA, Barlow DH. Anxiety Disorders Interview Schedule–Revised: (ADIS-R). San Antonio, Tex: Graywind Publications Inc/The Psychological Corp; 1988. 32 Di Nardo PA, Moras K, Barlow DH, et al. Reliability of DSM-III-R anxiety disorder categories. Arch Gen Psychiatry. 1993; 50:251–256. 33 Hofmann SG, Barlow DH, Papp LA, et al. Pretreatment attrition in a comparative treatment outcome study on panic disorder. Am J Psychiatry. 1998; 155:43–47. 34 Barlow DH, Craske MG. Mastery of Your Anxiety and Panic, II. San Antonio, Tex: Graywind Publications Inc/The Psychological Corp; 1994. 35 Shear MK, Brown TA, Barlow DH, et al. Multicenter collaborative Panic Disorder Severity Scale. Am J Psychiatry. 1997; 154:1571–1575. 36 Guy W. ECDEU Assessment Manual for Psychopharmacology, Revised. Washington, DC: US Government Printing Office; 1976. DHEW publication ADM 76–338. 37 Klein DF, Ross DC. Reanalysis of the National Institute of Mental Health Treatment of Depression Collaborative Research Program General Effectiveness Report. Neuropsychopharmacology. 1993; 8:241–251. 38 Quitkin FM, Rabkin JG, Ross D, Stewart JW. Identification of true drug response to antidepressants. Arch Gen Psychiatry. 1984; 41:782–786. 39 Quitkin FM, Rabkin JD, Markowitz JM, et al. Use of pattern analysis to identify true drug response. Arch Gen Psychiatry. 1987; 44:259–264. 40 Stewart JW, Quitkin FM, McGrath PJ, et al. Use of pattern analysis to predict differential relapse of remitted patients with major depression during 1 year of treatment with fluoxetine or placebo. Arch Gen Psychiatry. 1998; 55:334–343. 41 Lepola UM, Wade AG, Leinonen EV, et al. A controlled, prospective, 1-year trial of citalopram in the treatment of panic disorder. J Clin Psychiatry. 1998; 59: 528–534. 42 Telch MJ, Lucas JA, Schmidt NB, et al. Group cognitive-behavioral treatment of panic disorder. Behav Res Ther. 1993; 31:279–287. 43 American Psychiatric Association, for the Work Group on Panic Disorder. Practice guideline for the treatment of patients with panic disorder. Am J Psychiatry. 1998; 155(suppl 5):1–34. 44 Lecrubier Y, Bakker A, Dunbar G, Judge R, for The Collaborative Paroxetine Panic Study Investigators. A comparison of paroxetine, clomipramine and placebo in the treatment of panic disorder. Acta Psychiatr Scand. 1997; 95:145–152. 45 Bystritsky A, Rosen RM, Murphy KJ, et al. Doubleblind pilot trial of desipramine versus fluoxetine in panic patients. Anxiety. 1994–1995; 1:287–290.

140  David H. Barlow, et al. 46 Nair NP, Bakish D, Saxena B, et al. Comparison of fluvoxamine, imipramine, and placebo in the treatment of outpatients with panic disorder. Anxiety. 1996; 2:192–198. 47 Wade AG, Lepola U, Koponen HJ, et al. The effect of citalopram in panic disorder. Br J Psychiatry. 1997; 170:549–553. 48 Uhlenhuth EH, Matuzas W, Warner TD, Thompson PM. Growing placebo response rate. Psychopharmacol Bull. 1997; 33:31–39.

Article 10

Toward a Unified Treatment for Emotional Disorders David H. Barlow, Laura B. Allen, and Molly L. Choate Boston University

Over 40 years of development of cognitive behavioral approaches to treating anxiety and related emotional disorders have left us with highly efficacious treatments that are increasingly widely accepted. Nevertheless, these manualized protocols have become numerous and somewhat complex, restricting effective training and dissemination. Deepening understanding of the nature of emotional disorders reveals that commonalities in etiology and latent structure among these disorders supercedes differences. This suggests the possibility of distilling a set of psychological procedures that would comprise a unified intervention for emotional disorders. Based on theory and data emerging from the fields of learning, emotional development and regulation, and cognitive science, we identify three fundamental therapeutic components relevant to the treatment of emotional disorders generally. These three components include (a) altering antecedent cognitive reappraisals; (b) preventing emotional avoidance; and (c) facilitating action tendencies not associated with the emotion that is dysregulated. This treatment takes place in the context of provoking emotional expression (emotional exposure) through situational, internal, and somatic (interoceptive cues), as well as through standard mood-induction exercises, and differs from patient to patient only in the situational cues and exercises utilized. Theory and rationale supporting this new approach are described along with some preliminary experience with the protocol. This unified treatment may represent a more efficient and possibly a more effective strategy in treating emotional disorders, pending further evaluation. In the 1960s, cognitive behavioral approaches to treating emotional disorders such as anxiety and mood disorders began to emanate from basic psychological science; specifically, theories and data pertaining to learning, emotional development and regulation and, somewhat later, cognitive science. To justify a relatively radical new and unified psychological approach to treating emotional disorders, it is important to provide some background.

142  David H. Barlow, et al.

In the late 1950s and early 1960s, treatments began to deviate from a common general psychotherapeutic approach by directly targeting specific psycho-pathology, such as phobias. These treatments represented the beginnings of behavior therapy. This trend is probably best represented by Wolpe’s systematic desensitization, as well as the development of situational exposure for phobic behavior (Agras, Leitenberg, & Barlow, 1968; Marks, 1971; Wolpe, 1958). Wolpe’s development of systematic desensitization intrigued many, particularly those who foresaw the promise of translational research from basic behavioral science to the clinic. Furthermore, his descriptions of systematic desensitization, operationalized as they were, provided a large boost to attempts to demonstrate its efficacy empirically. This allowed investigators to begin the task of refocusing psychotherapy research from an emphasis largely on process to one on outcomes (Barlow & Hersen, 1984; Hersen & Barlow, 1976). Unfortunately, systematic desensitization had only limited efficacy in the clinic (in contrast to its success with college sophomores with snake fears) and then only for some types of specific phobia. Attempts to apply systematic desensitization to more complex clinical conditions such as agoraphobia were unsuccessful (Barlow, 1988, 1994). In the mid-1960s, we began experimenting with some different therapeutic strategies in which we encouraged individuals with what we would now call panic disorder with agoraphobia (PDA) to expose themselves to real-life frightening situations (e.g., Agras et al., 1968). At the same time, Isaac Marks in London was experimenting with similar procedures (e.g., Marks, 1971). This approach was innovative at the time because conventional wisdom held that experiencing anything more than small doses of anxiety might result in some harm to the patient, an idea based on prevailing theories. Gradually, through the 1970s, programmatic research revealed that situational exposure did not require the various trappings that were then associated with it (e.g., relaxation, contingent social reinforcement) and that it could be implemented relatively rapidly. By the 1980s, it was clear that we had an effective treatment for phobic behavior, but it was at best a blunt instrument. Best estimates of outcomes by the 1980s suggested that 60% to 75% of people receiving exposure-based treatments for PDA showed some clinical benefit. However, in a review of 24 studies, it was clear that as many as 35% of those entering treatment received little or no benefit (Jansson & Öst, 1982). Furthermore, of those receiving some benefit, only 10% to 20% could be said to approach normal functioning. The remainder continued to suffer from anxiety, panic, and residual avoidance. During this period of time, several important developments occurred in the treatment of anxiety and related emotional disorders. First, investigators began to focus on mechanisms of action and theories of behavior change. In considering exposure procedures, it was clear that “exposure” was simply a dry theoretical description of a process, with no heuristic value. Various theories of fear reduction began to be investigated, including habituation, extinction, and more cognitively based theories such as changes in selfefficacy (Bandura, 1977), modifying cognitive schemas (Beck et al., 1985), and emotional processing accounts (Foa & Kozak, 1986; Lang, 1979; Rachman,

Toward a Unified Treatment for Emotional Disorders  143

1980). Second, we learned a great deal about the nature of emotional disorders from ongoing studies of psychopathology. These theories of behavior change and increased knowledge of psychopathology led directly to new interventions. In addition to exposure-based procedures, cognitive therapy, first developed to treat depression, became a staple of treatments for anxiety disorders (Beck, 1972; Beck et al., 1985). In addition to situational exposure, we developed interoceptive exposure initially targeting PDA, recognizing that the context of anxiety and fear was internal as well as external (Barlow, 1988). Yet another important development was the beginnings of research on outcomes of individual cognitive behavioral therapy (CBT) protocols, targeting specific anxiety (or other) disorders (e.g., Barlow, Hayes, & Nelson, 1984). An important consequence was the growing realization that meaningful research on outcomes required the generation of detailed individual therapeutic manuals so that subsequent clinical research efforts could attempt to replicate the therapeutic procedures. As a result, psychological treatments were increasingly characterized by individual protocols that contained specific strategies such as cognitive restructuring, coping skills, and, where necessary, situational and interoceptive exposure procedures targeted to specific forms of psychopathology. These treatments were then tested empirically in a variety of formats, uses, and settings. Current Status of Treatment Early evidence on the efficacy of these CBT protocols led directly to largescale clinical trials, often conducted across several sites in order to include a large N and control for allegiance effects. For example, one large clinical trial tested the effectiveness of CBT, medication, and their combination as treatments for panic disorder (Barlow, Gorman, Shear, & Woods, 2000). A second study tested the effectiveness of cognitive behavioral group therapy (CBGT) compared to medication and several placebo groups, including a psychological placebo, for social phobia (Heimberg et al., 1998; Liebowitz et al., 1999). A third study looked at the separate and combined effects of a psychological treatment, medication, and their combination for chronic major depressive disorder (MDD; Keller et al., 2000). In this large study, patients received either nefazodone, a CBT constructed specifically for chronically depressed patients, or their combination. In yet another large, recently completed multitrial, the effects of clomipramine, intensive behavior therapy consisting of exposure and response prevention, and a combination were compared for the treatment of obsessive-compulsive disorder (OCD; Kozak, Liebowitz, & Foa, 2000). It is not our intention within the scope of this review to provide detailed outcomes from each individual protocol for specific disorders. Nevertheless, some general conclusions flow from the reports mentioned above and other studies, at least for adults (Barlow, 2001; Nathan & Gorman, 2002). Approximately 50% to 80% of patients undergoing treatment for one or more of the emotional disorders achieve “responder” status, with the definition of “responder” necessarily differing somewhat from study to study. In most of these cases, the individual has made a clinically significant improvement, although they may not be “cured” (symptom-free). These outcomes are typically better than a

144  David H. Barlow, et al.

credible alternative psychological treatment or “placebo” in every anxiety disorder, with information on this issue less certain for the mood disorders. Thus, these results indicate that “common factors” of positive expectancies, remoralization, and a strong therapeutic alliance, while contributory to outcomes, are substantially enhanced by the addition of psychological treatments, at least for the anxiety disorders (Barlow, 2001; Nathan & Gorman, 2002). Finally, it seems clear from the multisite studies mentioned above that psychological and pharmacological treatments achieve approximately equal efficacy immediately after treatment is concluded (except possibly for OCD, where the psychological treatment seems more efficacious), but that psychological treatments are more enduring after treatment is discontinued. The evidence also suggests that, in those disorders where the question has been evaluated, simultaneously combining drug and psychological treatments does not confer a substantial advantage, with the possible exception of MDD. Sequential combinations of treatments, on the other hand, are more promising (Barlow, 2002; Nathan & Gorman, 2002). Nevertheless, a number of significant limitations to current treatments exist. Obviously, there are still a considerable number of patients who do not respond well to this type of remedy, and the reasons for their lack of response are not yet known. Thus, although treatment is effective for many people, there is plenty of room for improvement. Another problem that has become apparent with manualized treatments is that there are simply too many of them. Clinicians must use separate handbooks, workbooks, and protocols for each disorder. Not only can this be quite costly, but it can take a significant amount of training to become adequately familiar with each of the distinct protocols. Finally, because the protocols are somewhat complex, dissemination of treatment to providers becomes an obstacle (Barlow, Levitt, & Bufka, 1999). For example, in the area of depression, a recent NIMH task force specified as a priority for treatment development the need for more “user-friendly” protocols (Hollon et al., 2002). Unless these treatments become more “user-friendly” as recommended, it is unlikely that most nonresearch clinicians will have a sufficient understanding of, or access to, these empirically supported techniques for the emotional disorders. The Nature of Emotional Disorders and “Negative Affect Syndrome” One argument for a unified treatment approach to emotional disorders is the facilitation of dissemination and training focused on a single set of therapeutic principles rather than diverse protocols. A second, more fundamental argument concerns emerging research and theory based on conceptions of the major emotional disorders that emphasize their commonalities rather than their differences. These arguments point to major developments in the areas of phenomenology and nosology, with a particular focus on comorbidity, all suggesting considerable overlap among disorders. Additionally, the observed effects of current psychological treatments on comorbid conditions and the nonspecificity of treatment response support this overlap. Equally important is emerging research on the latent structure of dimensional features of

Toward a Unified Treatment for Emotional Disorders  145

emotional disorders. Finally, a body of evidence supports commonalities in the etiology of emotional disorders, which has been summarized recently in the form of a new model referred to as “triple vulnerabilities” (Barlow, 1991, 2000, 2002). Each of these will be briefly reviewed in turn. For purposes here, the focus will be on anxiety and unipolar mood disorders (MDD and dysthymia). However, we envision that the principles elucidated may be applicable more broadly to psychopathology in which negative affect plays a functional role, including bipolar, somatoform, dissociative, and anger-related disorders, as well as eating disorders. Overlap Among Disorders

Currently the evidence strongly suggests considerable overlap among the various anxiety and mood disorders. At the diagnostic level this is most evident in the high rates of current and lifetime comorbidity (e.g., Brown, Campbell, Lehman, Grisham, & Mancill, 2001; Kessler et al., 1996, 1998). We have collected data on the percentages of additional diagnoses in patients who have been diagnosed with a principal anxiety or mood disorder (Brown et al., 2001). These data were derived from a large sample: 1,127 patients who were carefully diagnosed with the Anxiety Disorders Interview Schedule for DSM-IV— Lifetime Version (ADIS-IV-L; Di Nardo, Brown, & Barlow, 1994). As noted in that article, these summaries are most likely conservative due to limits in generalizability such as the nature of inclusion-exclusion criteria used. Overall, results indicate that 55% of patients with a principal anxiety or mood disorder had at least one additional anxiety or depressive disorder at the time of assessment, and this rate increases to 76% when additional diagnoses occurring at any time during the patient’s life, including currently (i.e., lifetime diagnoses), are considered. To take one example, of patients diagnosed with PDA, 60% out of 324 patients were determined to meet criteria for an additional anxiety or mood disorder, breaking down to 47% with an additional anxiety disorder and 33% with an additional mood disorder. When lifetime diagnoses are considered, the percentages rise to 77% experiencing any anxiety or mood disorder, breaking down to 56% for any anxiety disorder and 60% for any mood disorder. The principal diagnostic categories of posttraumatic stress disorder (PTSD), MDD, dysthymia, and generalized anxiety disorder (GAD) were associated with the highest comorbidity rates. For specific patterns of comorbidity associated with each diagnoses, see Brown et al. (2001). Further evidence of the conservative nature of this estimate is present in findings on the comorbidity of GAD and mood disorders. This is due to hierarchical exclusions remaining in the DSM-IV. For instance, when adhering strictly to DSM-IV diagnostic rules, the comorbidity of dysthymia and GAD was 5%. However, when the hierarchical rule that GAD should not be assigned when occurring exclusively during a course of a mood disorder was suspended, the comorbidity estimate increases to 90%. These data also ignore the presence of subthreshold symptoms that did not meet diagnostic thresholds for one disorder or another. There are several possible explanations for these high rates of comorbidity that we have reviewed extensively elsewhere (Brown & Barlow, 2002).

146  David H. Barlow, et al.

Among these are trivial problems with overlapping definitional criteria; artifactual reasons, such as differential base rates of occurrence in our setting; and the possibility that disorders are sequentially related and that the features of one disorder act as risk factors for another disorder. For example, depression seems to follow PDA, and panic disorder seems to follow PTSD. But another more intriguing explanation is that this pattern of comorbidity argues for the existence of what has been called a “general neurotic syndrome” (Andrews, 1990, 1996; Tyrer, 1989). Under this conceptualization, heterogeneity in the expression of emotional disorder symptoms (e.g., individual differences in the prominence of social anxiety, panic attacks, anhedonia, etc.) is regarded as trivial variation in the manifestation of a broader syndrome. This, in turn, is consistent with models we have developed that anxiety and mood disorders emerge from shared psychosocial and biological/genetic diatheses. If this is the case, then a unified treatment protocol cutting across current diagnostic categories to address core features of the emotional disorders could be a more parsimonious, and, perhaps, powerful option. The Effects of Psychological Treatments on Comorbid Disorders, and Nonspecificity of Treatment Response

Other findings supporting our contention of a “general neurotic syndrome” or “negative affect syndrome” (NAS), as we prefer to call it, include the observation that psychological treatments for a given anxiety disorder produce significant improvement in additional comorbid anxiety or mood disorders that are not specifically addressed in treatment (Borkovec, Abel, & Newman, 1995; Brown, Antony, & Barlow, 1995). For example, we examined the course of additional diagnoses in a sample of 126 patients who were being treated for PDA at our center. At pretreatment, 26% had an additional diagnosis of GAD, but the rate of comorbid GAD declined significantly at posttreatment to 9%, and remained at this level at a 2-year follow-up. Whether this represents the generalization of elements of treatment to independent facets of both disorders, or a way of effectively addressing “core” features of emotional disorders, is not significant to our purpose here. In both cases, the efficiency of a unified treatment protocol is suggested. The fact that a wide range of emotional disorders (e.g., MDD, dysthymia, OCD, and PDA) respond approximately equivalently to antidepressant medications has also been interpreted as indicating a shared pathophysiology among these symptoms (e.g., Hudson & Pope, 1990). Also, Tyrer et al. (1988) treated 210 outpatients with GAD, PDA, or dysthymia with either drug, placebo, CBT, or a self-help program. Although some differences were noted at posttreatment as a function of treatment condition (e.g., some drugs were less effective than other drugs and psychological treatments), no diagnostic group differences were observed. This suggested to Tyrer et al. (1988) that the differential diagnosis of anxiety and mood disorders does not provide a sound basis for treatment prescription. It should be noted, of course, that these data are only suggestive, because the current generation of more powerful psychological treatments was not utilized. But it does appear to offer support for the above conceptions.

Toward a Unified Treatment for Emotional Disorders  147

Latent Structure of the Emotional Disorders

There is wide agreement that DSM-IV represents the zenith of a “splitting approach” to nosology, with the obtained advantage of high rates of diagnostic reliability. But there is growing suspicion that this achievement has come at the expense of diagnostic validity, and that the current system, as suggested above, may be erroneously distinguishing categories that are minor variations of broader underlying syndromes. This would not imply a return to a nonempirical system of classification based on theories of etiology. Rather, this thinking points to a quantitative approach using structural equation modeling to examine the full range of anxiety and mood disorders without the constraints of artificial categories, given their strong relationship and potential overlap (Brown, Chorpita, & Barlow, 1998; Chorpita, Albano, & Barlow, 1998; Clark & Watson, 1991; Watson, Clark, & Harkness, 1994). We have been studying this question for the last 10 years (e.g., Brown et al., 1998; Zinbarg & Barlow, 1996) and have confirmed, with some modifications, the tripartite model of emotional disorders first proposed by Clark and Watson (1991). Some of our findings are presented in Figure 10.1. One of the intriguing and important findings from this line of research is that mood disorders show greater overlap with certain anxiety disorders such as GAD than do other anxiety disorders, supporting and reinforcing the commonalities of depression and anxiety at a phenomenological level (Brown et al., 1998; Clark, Steer, & Beck, 1994; Mineka, Watson, & Clark, 1998). The findings from Brown et al. (1998) using a sample of 350 patients with DSM-IV anxiety and mood disorders confirmed a hierarchical structure. In this structure, negative affect and positive affect emerged as higher-order factors to the DSM-IV disorder factors, with significant paths from negative affect to each of the five DSM-IV factors, and significant paths from positive affect to the mood disorders and social phobia factor only. In this model, autonomic arousal, which we consider to represent the phenomenon of panic, emerges as a lower-order factor with significant paths from PDA and GAD (where the relationship was negative). These findings indicate that the “key features” of the DSM anxiety and mood disorders cannot be collapsed indiscriminately into an NAS. But it seems safe to conclude that what is common outweighs what is not (DSM-IV factors). Our view, then, is that DSM-IV emotional disorder categories do not qualify in any sense as real entities (Kendell, 1975) but do seem to be useful concepts or constructs that emerge as “blips” on a general background of NAS. It also appears increasingly likely for a variety of reasons, including the findings from genetics (e.g., Kendler, 1996; Kendler, Heath, Martin, & Eaves, 1987), that DSM-V will turn to a more dimensional description of these phenomena that would reinforce conceptions of common underlying components (Kupfer, First, & Regier, 2002). Etiology

Elsewhere, we have elaborated on an interacting set of vulnerabilities or diatheses relevant to the development of anxiety, anxiety disorders, and related emotional disorders. This “triple vulnerabilities” theory encompasses a generalized biological vulnerability, a generalized psychological vulnerability, and

148  David H. Barlow, et al. –.36*

PA –.29* .67* .33*

.45* DEP

.50*

NA –.28* .43*

.74* .65* .18* .58*

GAD

.81* PD/A

–.22*

.67*

.31*

.76* OCD .02

SOC –.02

.11* AA

Figure 10.1 Structural Model of Interrelationships of DSM-IV Disorder Constructs and Negative Affect, Positive Affect, and Automatic Arousal. From “Structural Relationships Among Dimensions of the DSM-IV Anxiety and Mood Disorders and Dimensions of Negative Affect, Positive Affect, and Autonomic Arousal,” by T. A. Brown, B.F. Chorpita, & D.H. Barlow, 1998, Journal of Abnormal Psychology, 107, pp. 179–192. Copyright 1998 by the American Psychological Association. Reprinted with permission. Note. PA, positive affect; NA, negative affect; DEP, mood disorders; GAD, generalized anxiety disorder; PD/A, panic disorder with/without agoraphobia; OCD, obsessive-compulsive disorder; SOC, social phobia; AA, autonomic arousal.

a specific psychological vulnerability emerging from early learning (Barlow, 2000, 2002). A generalized biological vulnerability involves nonspecific genetic contributions to the development of anxiety and negative affect. Much of the research on this generalized biological vulnerability has focused on temperaments labeled “anxiety,” “neuroticism,” “negative affect,” or “behavioral inhibition.” Although the relationships among these closely related traits and temperaments have yet to be fully worked out, it is likely that each partially represents a common theme associated with a biological vulnerability to develop emotional disorders generally (Barlow, 2000, 2002). Additionally, early life experiences under certain conditions contribute to a generalized psychological vulnerability or diathesis to experience anxiety and related negative affective states (Chorpita & Barlow, 1998). It is this set of experiences that produces a sense of uncontrollability that seems to be at the core of negative affect and derivative states of anxiety and depression. If these two vulnerabilities happen to line up, and are potentiated by the influence of life stress, the likely result are the clinical syndromes of GAD and/or depressive disorders as outlined in Figure 10.2. Notice that false alarms (panic attacks) may occur as a function of stressful life events, facilitated by high levels of baseline anxiety, and emerging

Toward a Unified Treatment for Emotional Disorders  149

as a function of these synergistic generalized vulnerabilities. But these false alarms are not in themselves necessarily implicated in a clinical disorder. For that to occur, an additional layer of a more specific psychological vulnerability must be considered. In this conception, certain learning experiences seem to focus anxiety on specific life circumstances; that is, these circumstances or events become imbued with a heightened sense of threat or danger. For example, specific early learning experiences seem to determine whether individuals may view somatic sensations, intrusive thoughts, or social evaluation as specifically dangerous (Barlow, 2002; Bouton, Mineka, & Barlow, 2001). It is this specific psychological vulnerability that, when coordinated with the generalized biological and psychological vulnerabilities mentioned above, seems to contribute to the development of discrete anxiety disorders such as social phobia, OCD, panic disorder, and specific phobias, as represented in Figure 10.3. Evidence for this model has been reviewed in detail elsewhere (Barlow, 2000, 2002; Bouton et al., 2001; Chorpita & Barlow, 1998). While future research will determine the validity of this model, it is consistent with the emerging phenomenological evidence reviewed above on the overriding importance of common factors in the genesis and presentation of emotional disorders. Implications for Treatment Growing evidence on the unifying principles of emotional disorders (Barlow, 1991, 2002) with a focus on common underlying mechanisms suggests the possibility of distilling a set of psychological procedures that would comprise a unified intervention for emotional disorders. In 1988 one of us proposed, following the emotion theorists (e.g., Izard, 1971), that a coherent and consistent therapeutic approach to emotional disorders would ultimately be based on emotion theory and evolving knowledge of the modification of emotional states. Following Lang (1968), Rachman (1981), and Wilson (1982), whose early speculations were influential, components of any affective therapy were outlined (see Table 10.1). Based on theoretical and empirical work at that time, and a lineage of knowledge dating back to Darwin (1872), there seems no quicker or more powerful way to change emotional expression than to modify action tendencies associated with a specific emotion. Other essential targets for change included increasing a fundamental sense of controllability and predictability over events in one’s environment, and decreasing the powerful avoidant strategy of focusing attention on non-task-related consequences of excessive emotional activity (neurotic self-preoccupation). Other targets for change that were speculated as being insufficient but perhaps helpful inasmuch as they facilitated change in the essential targets included focusing directly on emotional cognitions, coping skills, social support networks, and heightened arousal. Evidence for these assertions was reviewed in Barlow (1988). Development of these ideas was, for the most part, put aside during the 1990s as we concentrated on DSM-IV, large clinical trials, and other tasks, but these ideas were not entirely ignored by others. For example, Marsha Linehan brought the concept of modifying action tendencies (opposite action tendencies) to good use in her development of dialectical behavior therapy (DBT; Linehan, 1993). For the past 18 months, we have turned our attention once again to these fundamental themes, taking advantage of substantial progress

150  David H. Barlow, et al. Synergistic Vulnerabilities Biological Vulnerabilities Generalized Psychological Vulnerability Stress

False Alarms (Panic)

Generalized Anxiety Depression

Figure 10.2 Diathesis-Stress Model of the Development of Generalized Anxiety and Depression. From “Unraveling the Mysteries of Anxiety and its Disorders from the Perspective of Emotion Theory,” by D.H. Barlow, 2000, American Psychologist, 55, pp. 1247–1263. Copyright 2000 by the American Psychological Association. Reprinted with permission.

Synergistic Vulnerabilities Biological Vulnerabilities Generalized Psychological Vulnerability Specific Psychological Vulnerability (Focus of anxiety: e.g., physical sensations are dangerous; social evaluations are dangerous; bad thoughts are dangerous) Stress Panic Disorder Social Phobia OCD

False Alarms (Panic)

Figure 10.3 Triple Vulnerabilities Model in the Development of Certain Anxiety Disorders. From “Unraveling the Mysteries of Anxiety and its Disorders from the Perspective of Emotion Theory,” by D.H. Barlow, 2000, American Psychologist, 55, pp. 1247–1263. Copyright 2000 by the American Psychological Association. Reprinted with permission.

in a number of related areas over the past decade. This progress has occurred in the context of developments in modern learning theory and cognitive neuroscience, as well as our greatly increased knowledge of naturally occurring processes in the regulation of emotional expression. We will briefly touch on developments in each of these areas. Modern Learning Theory and Cognitive Neuroscience In an article exploring theoretical conceptualizations of the etiology of anxiety disorders, Bouton et al. (2001) suggest that early panic attacks result in

Toward a Unified Treatment for Emotional Disorders  151 Table 10.1 Components of any affective therapy

A. Essential targets for change 1. Action tendencies 2. A sense of uncontrollability-unpredictability 3. Self-focused attention B. Helpful but not essential targets for change 1. “Hot” apprehensive cognitions 2. Hypervalent cognitive schemata and attention narrowing 3. Coping skills and social support 4. Elevated physiological responding and altered neurobiological functions Note. From Anxiety and Its Disorders:The Nature and Treatment of Anxiety and Panic, by D.H. Barlow, 1988, New York: The Guilford Press. Copyright 1988 by The Guilford Press. Reprinted by permission.

conditioned associations between the attack and a variety of interoceptive and exteroceptive cues. When some of these cues are elicited in a nonpanic situation, a constellation of behavioral and physiological responses arise, which we collectively call “anxiety.” As such, anxiety is a state of relatively low-level arousal preparing us for possible future danger. Panic, on the other hand, is associated with a surge in autonomic arousal enabling immediate action (fight-flight). These mechanisms are not necessarily mutually exclusive. In fact, modern learning theory suggests that anxiety may initiate panic because anxiety, too, can become a conditioned stimulus. It seems that small physiological, behavioral, and emotional changes can become associated with an extreme fear reaction (such as panic) with or without conscious knowledge of the cues. These conditioned events can begin to influence behavior at a subconscious level, such that strengthening of the association between physical and emotional cues and panic begins to occur (LeDoux, 1996). Brain imaging techniques have offered support for this interpretation, even indicating differences in the neurobiological bases of conscious and unconscious conditioning processes (Öhman, 1999). Future work in understanding the actual functional brain changes in emotional disorders continues, and researchers are now using imaging strategies to examine changes in brain function following cognitivebehavioral treatments for anxiety (e.g., Furmark et al., 2002). Thus, the future investigation of emotional disorders and their treatment will involve a comprehensive study of the psychological, emotional, and neurobiological correlates of emotion-based conditioning procedures. Work in this area has already influenced our knowledge of emotional cues from a variety of external and internal contexts that must be considered in treatment to maximize effects. Emotion Regulation

A particularly important concept for understanding emotional disorders is that of emotion regulation (Brenner & Salovey, 1997; Mayer & Salovey, 1997). By this, we are referring to the strategies individuals use to influence the occurrence, experience, intensity, and expression of a wide range of emotions

152  David H. Barlow, et al.

(Frijda, 1986; Masters, 1991; Richards & Gross, 2000). Emotion regulation and dysregulation seem to play an important role in emotional disorders (and other psychopathology), and levels of positive and negative emotions as well as their functional relationships often differ depending on the particular disorder (Gross & Levenson, 1997; Rottenberg & Gross, 2003; Thayer, 2000). Emotional disorders seem characterized to some degree by attempts to control both positive and negative emotions in a variety of contexts as outlined below. Individuals concerned about the expression and experience of their feelings may attempt to suppress, hide, or ignore them, with unintended consequences (Gross & Levenson, 1997; Pennebaker, 1997). This is because excessive attempts to control emotional experience lead to an increase in the very feelings the individuals are attempting to regulate, as demonstrated by attempts to control emotions after initial panic attacks (Craske, Miller, Rotunda, & Barlow, 1990). Furthermore, the degree to which one attempts to control emotions is somewhat related to the degree of intensity to which an individual experiences negative emotions, which can quickly become overwhelming (Lynch, Robins, Morse, & Krause, 2001). These overpowering experiences often lead to attempts at thought suppression as a convenient and accessible way to reduce emotional responsiveness. It is this pattern that may erupt in a vicious cycle of increased physiological and emotional arousal, leading to more unsuccessful attempts at suppression, which in turn contributes to growing psychological distress. Thus, it is clear that future treatments for emotional disorders must focus on this issue and develop treatments specifically targeting emotion dysregulation. A Unified Treatment

Based on theory and practice described above, over the past year we have distilled three fundamental therapeutic components that currently comprise our unified treatment approach to emotional disorders. Following a standard psychoeducational phase common to all psychotherapeutic approaches, these three components include (a) altering antecedent cognitive reappraisals—an intensive emotion-regulation procedure that directly facilitates the next two steps in treatment; (b) preventing emotional avoidance—a broad-based effort that goes well beyond traditional attempts to prevent behavioral avoidance in phobic disorders by targeting cognitive, behavioral, and somatic experiential avoidance; and (c) facilitating action tendencies not associated with the emotion that is disordered. This treatment takes place in the context of provoking emotional expression (emotion exposure) through situational, internal, and somatic (interoceptive) cues, as well as through standard mood-induction exercises, and differs from patient to patient only in the situational cues and exercises utilized. Notice also that “exposure” is not conceptualized as a mechanism of action. Rather, successfully provoking emotions is considered a setting condition in order to implement the essential treatment components. Of course, we recognize that the “mere-exposure” paradigm (Zajonc, 2001) has some emotion-regulating properties itself in terms of producing positive affect, even if the exposure is subliminal, most likely due to classical conditioning. These mechanisms may apply to emotion exposure as well.

Toward a Unified Treatment for Emotional Disorders  153

Antecedent Cognitive Reappraisal

Since the 1970s, and under the enormous influence of cognitive therapy as innovated by Aaron T. Beck (Beck, 1972; Beck, Rush, Shaw, & Emery, 1979), clinicians have focused on the appraisals and judgments that individuals with emotional disorders make regarding external events, as well as their own efficacy in coping with these events. This approach was first developed in the context of depression. In this context Beck outlined the well-known “cognitive triad” in which individuals maintained negative beliefs about their own self, the world, and the future. Subsequently, this notion was extended to anxiety disorders (Beck et al., 1985). It wasn’t long before the importance of appraisals focused on internal events, such as physical sensations and emotions, began to be recognized. In panic disorder, this was first suggested by Goldstein and Chambless (1978), who outlined a “fear of fear” model of panic disorder. In this model, conscious negative appraisals of somatic and affective manifestations of the experience of “fear” as dangerous became an important focal point. Subsequently, the concept of interoceptive conditioning originally developed by Razran (1961) was applied to anxiety disorders to describe a relatively less conscious process by which internal cues, often generated by emotional activation, could trigger anxiety and panic, with the implication that direct exposure to emotion-associated internal somatic cues would be an important part of treatment (Barlow, 1988). The notion of appraising internal cognitive and emotional states in a negative way was extended to other anxiety disorders such as OCD (Steketee, 1993) and GAD (Craske, Rapee, Jackel, & Barlow, 1989; Wells et al., 1995). Cognitive therapy focuses on evaluating the rationality of these negative appraisals and substituting more realistic evidence-based appraisals in their place. On one level, this may seem very much as an attempt to eliminate or suppress negative thoughts and immediately replace them with more adaptive or realistic appraisals, and it has been used in this way, as lucidly noted by Hayes, Strosahl, and Wilson (1999). But, in fact, this process can be conceptualized from an emotion-regulation perspective as altering antecedent reappraisals of threat and negativity. Support for this subtle but crucial reconceptualization emerges from the emotion-regulation literature, where evidence clearly exists that reappraisal of both internal and external threat and danger before the fact (that is, before heightened levels of negative emotion are provoked) has a salutory effect on the later expression of negative emotion, as noted above (Gross, 1998; Richards & Gross, 2000; Thayer, 2000). The importance of antecedent reappraisal versus more reactive strategies was first outlined in 1991 by Masters (1991). In a program of research, Gross (1998) has found that antecedent cognitive reappraisal does, in fact, reduce subjective experience of negative emotion. For individuals with emotional disorders, we first demonstrated this phenomenon in 1989 (Sanderson, Rapee, & Barlow, 1989) in the context of manipulating antecedent appraisals of control over a threatening situation. In this experiment, patients with PDA were told that they would be able to control—in a CO2 inhalation paradigm—the flow of CO2 by turning a dial when a light was illuminated. For half of the patients the light was never

154  David H. Barlow, et al.

illuminated, leading them to believe they had no control over the sensation. For the other half, the light did come on, indicating that the dial was operative and leading them to believe that they had some control over the situation. In fact, the dial had no bearing on CO2 flow and thus provided only an illusion of control (perceived control). Nor did patients ever really attempt to use the dial. Nevertheless, patients in the perceived control condition reported significantly fewer panic attacks and less emotion in general than those in the nocontrol group, despite receiving the same amount of CO2. More recently, Telch and colleagues have demonstrated this principle. For example, patients with specific phobias who were instructed to focus on their identified core threats and who received brief guidance on threat reappraisal evaluation did significantly better than those simply receiving exposure exercises without reappraisal (Kamphuis & Telch, 2000; Sloan & Telch, 2002). In our hands, we have adapted cognitive interventions to focus on two fundamental antecedent misappraisals: first, the probability of a negative event happening (probability overestimation), and second, the consequences of that negative event if it did happen (catastrophizing; Barlow & Craske, 2000; Craske, Barlow, & O’Leary, 1992). We have now extended these concepts to the full range of emotional disorders. Now Hariri and colleagues (Hariri, Bookheimer, & Mazziotta, 2000; Hariri, Mattay, Tessitore, Fera, & Weinberger, 2003), building on work by Damasio and colleagues (Bechara, Damasio, Damasio, & Lee, 1999; Damasio, 1994), have begun to elucidate the neural circuits underlying this process using functional magnetic resonance imaging (FMRI). They demonstrated that activation of the right prefrontal and anterior cingulate cortices during conscious evaluation and appraisal of emotional stimuli modulates and regulates bilateral amygdala responding. These findings also suggest that implementing antecedent cognitive reappraisal strategies is an important step in altering emotional responding. Emotional Avoidance

There is also evidence that many emotional disorders are related to attempts to down-regulate (avoid) excessive unexpected emotional experiences. Examples of this process in the context of depression, anger, and excitement (mania), as well as fear, are provided in Barlow (2002). Preliminary data also exist on the prevalence of the occurrence of unexpected and sometimes distressing emotions in the nonclinical population (Craske, Brown, Meadows, & Barlow, 1995). Specifically, over 300 undergraduates were surveyed with a questionnaire to determine the prevalence of both cued and uncued panic attacks, anger outbursts, episodes of sadness, and surges of excitement, as well as the degree of worry or distress over the recurrence of each type of emotional experience. While fully 32% of the sample reported no uncued emotional experiences, the fact that 68% of this sample reported experiencing at least one uncued emotion in the previous 3-month interval was surprising. It is also noteworthy that while 10% of the population experienced uncued panic, which is consistent with other surveys, as many as 34% of the sample reported uncued depressive episodes, and many found these episodes very distressing. The fact that distress was associated with a substantial proportion of these unexpected episodes

Toward a Unified Treatment for Emotional Disorders  155

implies attempts to regulate these emotional experiences through suppression (although this was not specifically tested) with resulting consequences. These results, of course, are consistent with observations of emotional avoidance in other disorders. For example, Roemer, Litz, Orsillo, and Wagner (2001) reported that veterans with PTSD were more likely to report intentionally withholding their emotions, both positive and negative, than were veterans without PTSD. In a related study of rape-related PTSD, Orsillo, Roemer, and Litz (2001) found that women with PTSD described their sexual assault experiences with fewer fear words than did women without PTSD, although women with PTSD displayed higher levels of arousal. Evidence on the deleterious use of avoidant techniques extends to calming procedures so much a part of our own earlier protocols for treating anxiety and panic (Barlow & Cerny, 1988). Specifically, when calming techniques such as relaxation and breathing control are conceptualized to the patient as a specific strategy for reducing negative emotions and distress, in which the focus is to “cope” with the emotions and distress (rather than as a noncontingent calming exercise), the results seem counterproductive. For example, Schmidt et al. (2000) concluded that breathing retraining did not add any clear benefits to a treatment package consisting of education, cognitive restructuring, and exposure-based techniques for patients with panic disorder. In fact, a trend in the data indicated that patients who received breathing retraining showed lower end-state functioning on both self-report and clinician-rated measures. Similar results have been obtained from our prior work evaluating distraction strategies (Craske, Street, & Barlow, 1989; Craske, Street, Jayaraman, & Barlow, 1991; Kamphuis & Telch, 2000). And, of course, the deleterious effects of downregulating (avoiding) emotion through reliance on talismen or safety signals have been conclusively demonstrated (Salkovskis, Clark, Hackman, Wells, & Gelder, 1999; Sloan & Telch, 2002; Wells et al., 1995). Finally, laboratory studies are directly demonstrating the effects of avoiding or suppressing emotion. For example, Feldner, Zvolensky, Eifert, and Spira (2003) divided nonclinical subjects into high or low emotional avoiders and subjected them to 4 breaths of 20% CO2-enriched air. Half in each group were instructed to inhibit negative emotional reactions, the other half to simply observe their emotional response. High emotional avoiders reported greater distress and anxiety, whether suppressing or not, compared to low avoiders. In our laboratory, Levitt, Brown, Orsillo, and Barlow (in press) divided 60 patients with PDA into three groups, each of whom listened to a 10-minute audiotape describing one of two emotion-regulation strategies (acceptance or suppression) or a neutral narrative. Patients then underwent a 15-minute 5.5% CO2 challenge. Following this challenge, they were asked to participate in a second challenge. The acceptance group was significantly less anxious and less avoidant than either the suppression or control groups in terms of subjective anxiety during the CO2 challenge and willingness to participate in a second challenge. Now we have also developed some direct evidence of the maladaptive use of emotion-regulation strategies in patients with a wide range of emotional disorders (Campbell-Sills, Barlow, Brown, & Hofmann, 2003). In this study, 60 patients who met diagnostic criteria for an anxiety or mood disorder and 30 individuals with no history of emotional disorders experienced an induction

156  David H. Barlow, et al.

of a negative emotion by watching an emotional film. In Study 1, the spontaneous emotion appraisals and emotional-regulation strategies were observed in both the clinical sample and the control sample. The patients in the clinical sample reported significantly different emotional appraisals and emotionalregulation strategies than nonclinical participants. Consistent with data reported above, clinical participants reported greater anxiety focused on the occurrence of emotions, as well as less emotional clarity. They also endorsed more reliance on maladaptive emotion-regulation strategies (e.g., suppression, cognitive rehearsal). The patients also rated their resulting emotions as less acceptable and engaged in more emotional suppression. Higher levels of suppression were associated, in turn, with elevated heart rate during the emotion induction, as well as inhibited recovery from subjective distress, skin conductance, and finger temperature changes after the induction. In the second study, patients were instructed to engage in either emotion suppression activities during the emotional induction exercise, or emotion acceptance activities. Suppression participants failed to recover from subjective distress after the induction, and they manifested a different heart rate pattern than acceptance participants. Specifically, when patients were instructed to suppress their emotions, once again heart rate actually increased from anticipation to termination of the film, while heart rate in the acceptance group decreased during this period. Thus, patients with emotional disorders endorsed more negative emotion appraisals and utilized counterproductive emotion-regulation strategies compared to individuals without disorders. Modifying Emotional Action Tendencies

As Izard pointed out in 1971, theories and evidence from emotion theory indicate that “the most efficient and generalized principles and techniques for emotion control [are] focused on the neuromuscular component of emotion . . . [S]triate muscle action can initiate, amplify, attenuate, or inhibit an emotion” (p. 415). In other words, “[T]he individual learns to act his way into a new way of feeling” (p. 410). As I suggested in 1988, it is possible that the crucial function of exposure in the treatment of phobic disorders is to prevent the action tendencies associated with fear and anxiety and facilitate different action tendencies (Barlow, 1988). For example, Fridlund, Hatfield, Cottam, and Fowler (1986) determined that physiological elevation during anxiety did not represent generalized arousal, but rather specific action tendencies associated with anxiety. I also speculated (Barlow, 1988) that it is possible that attention to action tendencies forms an important part of the treatment of other emotional disorders. For example, Beck, Rush, Shaw, and Emery (1979) and others (e.g., Lewinsohn & Lee, 1981) spent a considerable amount of time countering the tendencies of their depressed patients to behave in a “passive, retarded, and apathetic manner” (p. 312). More recently, behavioral activation has become a central and defining principle in some new treatments for depression that show considerable promise (Jacobson, Martell, & Dimidjian, 2001). In fact, these strategies have a long history. Laughter, humor, and associated facial expression induced during successful paradoxical intention techniques (Frankl, 1960), a technique successfully used to counteract fear and anxiety

Toward a Unified Treatment for Emotional Disorders  157

in previous decades (e.g., Ascher, 1980), may be effective not because of the induction of cognitive changes, as was often assumed, but rather because of the prevention of behavioral responses, including facial expressions, and the substitution of action tendencies associated with alternative emotions. As noted above, Linehan (1993) has adapted this strategy creatively and with good effect to patients with borderline personality disorder (which, with its core feature of the avoidance and intolerability of negative affect, may be fundamentally a severe emotional disorder). Also Hayes, in his creative conceptualization of acceptance and commitment therapy (ACT; Hayes et al., 1999), has underscored the importance of encouraging action as an alternative coping strategy, with the purpose of instituting a sense of control as opposed to focusing on decreasing unwanted internal events. Applications to Emotional Disorders

As reviewed above, it is our contention that these three basic therapeutic principles can be applied, with relatively minor modifications, to each of the emotional disorders. Antecedent cognitive reappraisal in each of the emotional disorders is easily categorized into overestimating the probability of a negative event happening (probability overestimation) and exaggerating the consequences of that negative effect if it did happen (catastrophizing). These concepts are relatively well-known by cognitive behavioral therapists and need no further elaboration as applied to individual disorders. Tables 10.2 and 10.3 provide some examples of both the general implementation of strategies for preventing emotional avoidance and modifying action tendencies across emotional disorders, as well as some specific adaptations of these strategies that seem useful when dealing with individual emotional disorders. We have reviewed evidence above on the variety of cognitive and behavioral avoidance strategies—including cognitive rituals, distraction, emotion suppression, etc., as noted in Table 10.2—that cut across emotional disorders. Providing examples across some specific emotional disorders, it is clear that the lives of individuals with PDA revolve around avoiding intense emotions, particularly fear, and associated somatic sensations. This also extends to contexts or situations that produce strong emotional and somatic responding. We have referred elsewhere to this pattern of avoidance as “interoceptive avoidance,” and have developed a questionnaire to assess the extent of its presence (Brown, White, Forsyth, & Barlow, 2004; Rapee, Craske, & Barlow, 1994/1995). Cognitive and behavioral rituals have long been recognized as avoidant strategies in OCD. More recently, we have recognized similar features associated with GAD in the form of “worry behaviors,” which are fundamentally checking rituals (Craske et al., 1992), as well as the behavioral pattern of “perfectionism,” particularly negative/maladaptive perfectionism. This behavioral pattern seems to reflect a tendency to impose control over perceived uncontrollable daily events in one’s life, thereby reducing distress and negative affect (Frost, Heimberg, Holt, Mattia, & Neubauer, 1993; Scott & Cervone, 2002). Similarly, for MDD, behavioral tendencies toward withdrawal seem clearly associated with avoiding interactions and contexts that may provoke negative affect. In modifying the behavioral action tendencies driven by fundamental emotions, the first step is to provoke the emotions insofar as possible, usually through

158  David H. Barlow, et al. Table 10.2 Preventing avoidance

General

Safety signals

Specific Water bottles

PDA

Cell phones

Social phobia

Medications

Specific phobia

Cognitive rituals Distraction

OCD GAD

Rationalization

PTSD

Emotion suppression Worry-ruminationrehearsal

MDD

Avoidance of: somatic sensations— ”emotional” experiences Avoidance of: performanceinteractions Avoidance of: objectsituation Behavioral rituals Perfectionism, worry-worry behaviors Avoidance of: strong affect-trauma cues Withdrawal

Note. PDA = panic disorder with agoraphobia; OCD = obsessive-compulsive disorder; GAD = generalized anxiety disorder; PTSD = posttraumatic stress disorder; MDD = major depressive disorder.

emotion-inducing exposure-based procedures (see Table 10.3). Adopting strategies that encourage experience of the emotion without engaging in the associated action tendencies (accepting the emotion) is a very basic strategy in this regard. When applied to specific disorders, emotional and behavioral activation, especially in situational context, becomes a particularly powerful tool. Modifying action tendencies often, but not always, coincides with and complements techniques to prevent avoidance of emotional expression. In fact, preventing avoidance in some emotional contexts, such as anxiety and fear, is one way to modify action tendencies. Thus, for most phobic situations it is clear that encouraging approach behavior in place of avoidance is an important step. But this strategy of modifying action tendencies also might involve provoking a different emotion, or facial expression, with its associated action tendencies in the phobic context (e.g., laughter or humor as in paradoxical intention). In GAD, implementing this strategy might involve actively prescribing “nonperfect” behavior at home or in the workplace. Passivity and detachment may be more appropriate actions in the context of anger, or more positive, active coping skills to counter withdrawal in depression. It is also clear that individuals may initiate different action tendencies (e.g., approach) during emotion induction while simultaneously engaging in substantial avoidance behavior (e.g., distraction). Thus, attention to both strategies is essential. At present we are in the beginning process of evaluating the efficacy and feasibility of this protocol. Thus far, we have collected data from several groups of 6

Toward a Unified Treatment for Emotional Disorders  159 Table 10.3 Modifying action tendencies

General

Specific

Emotion exposure Emotion recognition

GAD: emotional arousal MDD: behavioral activation, increased coping behavior PDA: approach behavior, activation of fearrelated somatic sensations

Facilitate emotion experiencing and acceptance

Note. GAD = generalized anxiety disorder; MDD = major depressive disorder; PDA = panic disorder with agoraphobia.

to 8 patients with heterogeneous emotional disorders, including such disparate principal diagnoses as PTSD, MDD, GAD, and a variety of phobic disorders. While final results await more systematic data collection and experimentation, patients in these groups are doing as well as or a bit better than patients in more homogeneous groups. More importantly, patients express their understanding of the fundamental similarities in their experiences, despite different diagnoses and somewhat different presenting symptoms. Confirming this impression through systematic and rigorous experimentation is the next step, as well as further elucidating mechanisms of therapeutic action. In conclusion, it is of interest to note that the first author had the good fortune to be “present at the creation” of behavior therapy (at least as a student), at a time when the concept of “neuroses” was in ascendance and long-term depth psychotherapy was the “unified” approach to these problems. Now the junior authors are beginning their careers with concepts of “negative affect syndrome” and a proposed unified psychological approach. While ironic on the face of it, the fundamental differences in these two seemingly similar conceptions represent, in a very real way, the revolution of the last 40 years. For now we have a broader and deeper understanding of the psychopathology of emotional disorders that is based on the slow but inexorable process of science. And now we have successful and effective ways—empirically derived, but also firmly grounded in theory—to address the suffering occasioned by negative affect in its many manifestations. The ability to submit the many and varied hypotheses generated over the years to the scientific method, as well as the fact that those generating the hypotheses were often proven wrong along the way, resulting in mid-course corrections, has brought us to where we are today. Most of us going down this road will be wrong again in the future and it is possible that many of the ideas presented in this article will not be sustained. But we will all be better off finding that out, and in a fundamental sense, this is the fulfilled promise of CBT that differentiates us from our forebears. Notes Dr. Barlow was invited to contribute an article about his research program on the occasion of his receiving an award for Outstanding Contribution by an Individual for Research Activities at the 2001 convention of the Association for Advancement of Behavior Therapy. This article,

160  David H. Barlow, et al.

by Barlow, Allen, and Choate, was received in response to this request. It underwent a modified editorial review process, similar to the one used for AABT presidential address articles. The Editor and Associate Editors congratulate Dr. Barlow on his richly deserved award. This work was supported in part by NIMH Award 5 RO1 MH45965, “Treatment of Panic Disorder: Long-Term Strategies.” References Agras, W. S., Leitenberg, H., & Barlow, D. H. (1968). Social reinforcement in the modification of agoraphobia. Archives of General Psychiatry, 19, 423–427. Andrews, G. (1990). Classification of neurotic disorders. Journal of the Royal Society of Medicine, 83, 606–607. Andrews, G. (1996). Comorbidity in neurotic disorders: The similarities are more important than the differences. In R. M. Rapee (Ed.), Current controversies in the anxiety disorders (pp. 3–20). New York: The Guilford Press. Ascher, L. M. (1980). Paradoxical intention. In A. Goldstein & E. B. Foa (Eds.), Handbook of behavioral interventions: A clinical guide. New York: Wiley. Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Barlow, D. H. (1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: The Guilford Press. Barlow, D. H. (1991). Disorders of emotion. Psychological Inquiry, 2, 58–71. Barlow, D. H. (1994). Effectiveness of behavior treatment for panic disorder with and without agoraphobia. In B. E. Wolfe & J. D. Maser (Eds.), Treatment of anxiety and panic: A Consensus Development Conference (pp. 105–120). Washington, DC: American Psychiatric Press. Barlow, D. H. (2000). Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory. American Psychologist, 55, 1247–1263. Barlow, D. H. (Ed.). (2001). Clinical handbook of psychological disorders: A step by-step treatment manual (3rd ed.). New York: The Guilford Press. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). New York: The Guilford Press. Barlow, D. H., & Cerny, J. A. (1988). Psychological treatment of panic. New York: The Guilford Press. Barlow, D. H., & Craske, M. G. (2000). Mastery of your anxiety and panic (MAP3): Client workbook for anxiety and panic (3rd ed.). San Antonio, TX: Graywind/Psychological Corporation. Barlow, D. H., Gorman, J. M., Shear, M. K., & Woods, S. W. (2000). Cognitive behavioral therapy, imipramine, or their combination for panic disorder: A randomized controlled trial. Journal of the American Medical Association, 283, 2529–2536. Barlow, D. H., Hayes, S. C., & Nelson, R. O. (1984). The scientist-practitioner: Research and accountability in clinical and educational settings. New York: Pergamon Press. Barlow, D. H., & Hersen, M. (1984). Single case experimental designs: Strategies for studying behavior change (2nd ed.). New York: Pergamon Press. Barlow, D. H., Levitt, J. T., & Bufka, L. F. (1999). The dissemination of empirically supported treatments: A view to the future. Behaviour Research and Therapy, 37, S147–S162. Bechara, A., Damasio, H., Damasio, A. R., & Lee, G. P. (1999). Different contributions of the human amygdala and ventromedial prefrontal cortex to decision-making. Journal of Neuroscience, 79, 5473–5481.

Toward a Unified Treatment for Emotional Disorders  161 Beck, A. T. (1972). Depression: Causes and treatment. Philadelphia: University of Pennsylvania Press. Beck, A. T., & Emery, G., with Greenberg, R. L. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New York: The Guilford Press. Borkovec, T. D., Abel, J. L., & Newman, H. (1995). Effects of psychotherapy on comorbid conditions in generalized anxiety disorder. Journal of Consulting and Clinical Psychology, 63, 479–483. Bouton, M. E., Mineka, S., & Barlow, D. H. (2001). A modern learning-theory perspective on the etiology of panic disorder. Psychological Review, 108, 4–32. Brenner, E. M., & Salovey, P. (1997). Emotion regulation during childhood: Developmental, interpersonal, and individual considerations. In P. Salovey & D. J. Sluyter (Eds.), Emotional development and emotional intelligence: Educational implications (pp. 168–195). New York: Basic Books. Brown, T. A., Antony, M. M., & Barlow, D. H. (1995). Diagnostic comorbidity in panic disorder: Effect on treatment outcome and course of comorbid diagnoses following treatment. Journal of Consulting and Clinical Psychology, 63, 408–418. Brown, T. A., & Barlow, D. H. (2002). Classification of anxiety and mood disorders. In D. H. Barlow (Ed.), Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). New York: The Guilford Press. Brown, T. A., Campbell, L. A., Lehman, C. L., Grisham, J. R., & Mancill, R. B. (2001). Current and lifetime comorbidity of the DSM-IV anxiety and mood disorders in a large clinical sample. Journal of Abnormal Psychology, 110, 49–58. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179–192. Brown, T. A., White, K. S., Forsyth, J. P., & Barlow, D. H. (2004). The structure of perceived emotional control: Psychometric properties of a revised anxiety control questionnaire. Behaviour Therapy, 35, 75–99. Campbell-Sills, L., Barlow, D. H., Brown, T. A., & Hofmann, S. G. (2004). The relevance of emotion regulatory processes to anxiety and mood disorders. Manuscript submitted for publication. Chorpita, B. F., Albano, A.M., & Barlow, D. H. (1998). The structure of negative emotions in a clinical sample of children and adolescents. Journal of Abnormal Child Psychology, 107, 74–85. Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety: The role of control in the early environment. Psychological Bulletin, 124, 3–21. Clark, D. A., Steer, R. A., & Beck, A. T. (1994). Common and specific dimensions of selfreported anxiety and depression: Implications for the cognitive and tripartite models. Journal of Abnormal Psychology, 103, 645–654. Clark. L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 103, 103–116. Craske, M. G., Barlow, D. H., & O’Leary, T. (1992). Mastery of your anxiety and worry. San Antonio, TX: Graywind/Psychological Corporation. Craske, M. G., Brown, T. A., Meadows, E. A., & Barlow, D. H. (1995). Uncued and cued emotions and associated distress in a college sample. Journal of Anxiety Disorders, 9, 125–137. Craske, M. G., Miller, P. P., Rotunda, R., & Barlow, D. H. (1990). A descriptive report of features of initial unexpected panic attacks in minimal and extensive avoiders. Behaviour Research and Therapy, 28, 395–400.

162  David H. Barlow, et al. Craske, M. G., Rapee, R. M., Jackel, L., & Barlow, D. H. (1989). Qualitative dimensions of worry in DSM-III-R generalized anxiety disorder subjects and nonanxious controls. Behaviour Research and Therapy, 27, 397–402. Craske, M. G., Street, L. L., & Barlow, D. H. (1989). Instructions to focus upon or distract from internal cues during exposure treatment of agoraphobic avoidance. Behaviour Research and Therapy, 27, 663–672. Craske, M. G., Street, L. L., Jayaraman, J., & Barlow, D. H. (1991). Attention versus distraction during in vivo exposure: Snake and spider phobias. Journal of Anxiety Disorders, 5, 199–211. Damasio, A. R. (1994). Descartes’ error. New York: Grosset/Putnam. Darwin, C. R. (1872). The expression of the emotions in man and animals. London: John Murray. Di Nardo, P. A., Brown, T. A., & Barlow, D. H. (1994). Anxiety Disorders Interview Schedule for DSM-IV: Lifetime Version (ADIS-IV-L). San Antonio, TX: Psychological Corporation. Feldner, M. T., Zvolensky, M. J., Eifert, G. H., & Spira, A. P. (2003). Emotional avoidance: An experimental test of individual differences and response suppression using biological challenge. Behaviour Research and Therapy, 41, 403–411. Foa, E. B., & Kozak M. J. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20–35. Frankl, V. E. (1960). Paradoxical intention: A logotherapeutic technique. American Journal of Psychotherapy, 14, 520–535. Fridlund, A., J., Hatfield, M. E., Cottam, G. L., & Fowler, J. C. (1986). Anxiety and striate muscle activation: Evidence from electromyographic pattern analysis. Journal of Abnormal Psychology, 95, 228–236. Frijda, N. H. (1986). The emotions. Cambridge, UK: Cambridge University Press. Frost, R. O., Heimberg, R. G., Holt, C. S., Mattia, J. I., & Neubauer, A. L. (1993). A comparison of two measures of perfectionism. Personality and Individual Differences, 14, 119–126. Furmark, T., Tillfors, M., Marteinsdottir, I., Fischer, H., Pissiota, A., Langstrom, B., & Fredrikson, M. (2002). Common changes in cerebral blood flow in patients with social phobia treated with Citalopram or cognitive-behavioral therapy. Archives of General Psychiatry, 59, 425–433. Goldstein, A. J., & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 47–59. Gross, J. J. (1998). Antecedent- and response-focused emotion regulation: Divergent consequences for experience, expression, and physiology. Journal of Personality and Social Psychology, 74, 224–237. Gross, J. J., & Levenson, R. W. (1997). The acute effects of inhibiting negative and positive emotion. Journal of Abnormal Psychology, 106, 95–103. Hariri, A. R., Bookheimer, S. Y., & Mazziotta, J. C. (2000). Modulating emotional responses: Effects of a neocortical network on the limbic system. Neuroreport, 11, 43–48. Hariri, A. R., Mattay, V. S., Tessitore, A., Fera, F., & Weinberger, D. R. (2003). Neocortical modulation of the amygdala response to fearful stimuli. Biological Psychiatry, 53, 494–501. Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). Acceptance and commitment therapy: An experiential approach to behavior change. New York: The Guilford Press. Heimberg, R. G., Liebowitz, M. R., Hope, D. A., Schneier, F. R., Holt, C. S., Welkowitz, L. A., Juster, H. R., Campeas, R., Bruch, M. A., Cloitre, M., Fallon, B., & Klein, D. F. (1998). Cognitive behavioral group therapy vs. phenelzine therapy for social phobia: 12-week outcome. Archives of General Psychiatry, 55, 1133–1141.

Toward a Unified Treatment for Emotional Disorders  163 Hersen, M., & Barlow, D. H. (1976). Single case experimental designs: Strategies for studying behavior change. New York: Pergamon Press. Hollon, S. D., Munoz, R. F., Barlow, D. H., Beardslee, W. R., Bell, C. C., Guillermo, B., Clark, G. N., Francois, L. P., Kazdin, A. E., Kohn, L., Linehan, M. M., Markowitz, J. C., Milkowitz, D. J., Persons, J. B., Niederehe, G., & Somers, D. (2002). Psychosocial intervention development for the prevention and treatment of depression: Promoting innovation and increasing access. Biological Psychiatry, 52(6), 610–630. Hudson, J. I., & Pope, H. G. (1990). Affective spectrum disorder: Does antidepressant response identify a family of disorders with a common pathophysiology? American Journal of Psychiatry, 147, 552–564. Izard, C. E. (Ed.). (1971). The face of emotion. New York: Appleton-Century-Crofts. Jacobson, N. S., Martell, C. R., & Dimidjian, S. (2001). Behavioral activation treatment for depression: Returning to contextual roots. Clinical Psychology: Science and Practice, 8, 255–270. Jansson, L., & Öst, L.-G. (1982). Behavioral treatments for agoraphobia: An evaluative review. Clinical Psychology Review, 2, 311–336. Kamphuis, J. H., & Telch, M. J. (2000). Effects of distraction and guided threat reappraisal on fear reduction during exposure based treatments for specific fears. Behaviour Research and Therapy, 38, 1163–1181. Keller, M. B., McCullough, J. P., Klein, D. N., Arnow, B., Dunner, D. L., Gelenberg, A. J., Markowitz, J. C., Nemeroff, C. B., Russell, J. M., Thase, M. E., Trivedi, M. H., & Zajecka, J. (2000). A comparison of nefazodone, the cognitive behavioral-analysis system of psychotherapy, and their combination for the treatment of chronic depression. New England Journal of Medicine, 342, 1462–1470. Kendell, R. E. (1975). The role of diagnosis in psychiatry. Oxford: Blackwell. Kendler, K. S. (1996). Major depression and generalized anxiety disorder: Same genes (partly), different environments—revisited. British Journal of Psychiatry, 168 (Suppl. 30), 68–75. Kendler, K. S., Heath, A. C., Martin, N. G., & Eaves, L. J. (1987). Symptoms of anxiety and symptoms of depression: Same genes, different environments? Archives of General Psychiatry, 44, 451–457. Kessler, R. C, Nelson, C. B., McGonagle, K. A., Lui, J., Swartz, M., & Blazer, D. G. (1996). Comorbidity of DSM-III-R major depressive disorder in the general population: Results from the National Comorbidity Survey. British Journal of Psychiatry, 168, 17–30. Kessler, R. C., Stang, P. E., Wittchen, H. U., Ustan, T. B., Roy-Byrne, P. P., & Walters, E. E. (1998). Lifetime panic-depression comorbidity in the National Comorbidity Survey. Archives of General Psychiatry, 55, 801–808. Kozak, M. J., Liebowitz, M. R., & Foa, E. B. (2000). Cognitive behavior therapy and pharmacotherapy for OCD: The NIMH-sponsored collaborative study. In W. K. Goodman, M. V. Rudorfer, & J. D. Maser (Eds.), Treatment challenges in obsessive-compulsive disorder: Contemporary issues in treatment (pp. 501–530). Mahwah, NJ: Erlbaum Associates. Kupfer, D. J., First, M. B., & Regier, D. A. (Eds.). (2002). Research agenda for DSM-V. Washington, DC: American Psychiatric Press. Lang, P. J. (1968). Fear reduction and fear behavior: Problems in treating a construct. In J. M. Shlien (Ed.), Research in psychotherapy (Vol. 3) (pp. 90–102). Washington, DC: American Psychological Association. Lang, P. J. (1979). A bio-informational theory of emotional imagery. Psychophysiology, 16, 495–512. LeDoux, J. E. (1996). The emotional brain: The mysterious underpinnings of emotional life. New York: Simon & Schuster.

164  David H. Barlow, et al. Levitt, J. T., Brown, T. A., Orsillo, S. M., & Barlow, D. H. (in press). The effects of acceptance versus suppression of emotion on subjective and psychophysiological response to carbon dioxide challenge in patients with panic disorder. Behavior Therapy, 35, 747–766. Lewinsohn, P.M., & Lee, W. M. L. (1981). Assessment of affective disorders. In D. H. Barlow (Ed.), Behavioral assessment of adult disorders. New York: The Guilford Press. Liebowitz, M. R., Heimberg, R. G., Schneier, F. R., Hope, D. A., Davies, S., Holt, C. S., Goetz, D., Juster, H. R., Lin, S. H., Bruch, M. A., Marshall, R. D., & Klein, D. F. (1999). Cognitive-behavioral group therapy versus phenelzine in social phobia: Long term outcome. Depression and Anxiety, 10, 89–98. Linehan, M. (1993). Skills training manual for cognitive behavioral treatment of borderline personality disorder. New York: The Guilford Press. Lynch, T. R., Robins, C. J., Morse, J. O., & Krause, E. D. (2001). A mediational model relating affect intensity, emotion inhibition, and psychological distress. Behavior Therapy, 32, 519–536. Marks, I. M. (1971). Phobic disorders four years after treatment: A prospective follow-up. British Journal of Psychiatry, 129, 362–371. Masters, J. C. (1991). Strategies and mechanisms for the personal and social control of emotion. In J. Garber & K. Dodge (Eds.), The development of emotion regulation and dysregulation: Cambridge studies in social and emotional development (pp. 182–207). New York: Cambridge University Press. Mayer, J. D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey & D. J. Sluyter (Eds.), Emotional development and emotional intelligence: Educational implications (pp. 3–34). New York: Basic Books. Mineka, S., Watson, D., & Clark, L. A. (1998). Comorbidity of anxiety and unipolar mood disorders. Annual Review of Psychology, 49, 377–412. Nathan, P. E., & Gorman, J. M. (Eds.). (2002). A guide to treatments that work. London: Oxford University Press. Öhman, A. (1999). Distinguishing unconscious from conscious emotional processes: Methodological considerations and theoretical implication. In T. Dalgleish & M. Power (Eds.), Handbook of cognition and emotion (pp. 321–352). Chichester, England: Wiley. Orsillo, S. M., Roemer, L., & Litz, B. T. (2001). Narrative analysis of emotional processing in PTSD. Manuscript submitted for publication. Pennebaker, J. W. (1997). Therapeutic process. Psychological Science, 8, 162–166. Rachman, S. J. (1980). Emotional processing. Behaviour Research and Therapy, 18, 51–60. Rachman, S. J. (1981). The primacy of affect: Some theoretical implications. Behaviour Research and Therapy, 19, 279–290. Rapee, R. M., Craske, M. G., & Barlow, D. H. (1994/1995). Assessment instrument for panic disorder that includes fear of sensation-producing activities: The Albany Panic and Phobia Questionnaire. Anxiety, 1, 114–122. Razran, G. (1961). The observable unconscious and the inferable conscious in current Soviet psychophysiology: Interoceptive conditioning, semantic conditioning, and the orienting reflex. Psychological Review, 68, 81–150. Richards, J. M., & Gross, J. J. (2000). Emotion regulation and memory: The cognitive costs of keeping one’s cool. Journal of Personality and Social Psychology, 79, 410–424. Roemer, L., Litz, B. T., Orsillo, S. M., & Wagner, A. W. (2001). A preliminary investigation of the role of strategic withholding of emotions in PTSD. Journal of Traumatic Stress, 14, 143–150. Rottenberg, J., & Gross, J. J. (2003). When emotion goes wrong: Realizing the promise of affective science. Clinical Psychology: Science and Practice, 10, 227–232.

Toward a Unified Treatment for Emotional Disorders  165 Salkovskis, P.M., Clark, D. M., Hackman, A., Wells, A., & Gelder, M. G. (1999). An experimental investigation of the role of safety-seeking behaviours in the maintenance of panic disorder with agoraphobia. Behaviour Research and Therapy, 37, 559–574. Sanderson, W. C., Rapee, R. M., & Barlow, D. H. (1989). The influence of an illusion of control on panic attacks induced via inhalation of 5.5% carbon dioxide–enriched air. Archives of General Psychiatry, 46, 157–164. Schmidt, N. B., Woolaway-Bickel, K., Trakowski, J., Santiago, H., Storey, J., Koselka, M., & Cook, J. (2000). Dismantling cognitive-behavioral treatment for panic disorder: Questioning the utility of breathing retraining. Journal of Consulting and Clinical Psychology, 68, 417–424. Scott, W. D., & Cervone, D. (2002). The impact of negative affect on performance standards: Evidence for an affect-as-information mechanism. Cognitive Therapy and Research, 26, 19–37. Sloan, T., & Telch, M. J. (2002). The effects of safety-seeking behavior and guided threat reappraisal on fear reduction during exposure: An experimental investigation. Behaviour Research and Therapy, 40, 235–251. Steketee, G. (1993). Treatment of obsessive compulsive disorder. New York: The Guilford Press. Thayer, R. E. (2000). Mood regulation and general arousal systems. Psychological Inquiry, 11, 202–204. Tyrer, P. J. (1989). Classification of neurosis. Chichester, England: Wiley. Tyrer, P. J., Seivewright, N., Murphys, S., Ferguson, B., Kingdon, D., Barczak, B., Brothwell, J., Darling, C., Gregory, S., & Johnson, A. L. (1988). The Nottingham study of neurotic disorder: Comparison of drug and psychological treatments. Lancet, 2, 235–240. Watson, D., Clark, L. A., & Harkness, A. R. (1994). Structures of personality and their relevance to psychopathology. Journal of Abnormal Psychology, 103, 18–31. Wells, A., Clark, D. M., Salkovskis, P.M., Ludgate, J., Hackmann, A., & Gelder, M. (1995). Social phobia: The role of in-situation safety behaviors in maintaining anxiety and negative beliefs. Behavior Therapy, 26, 153–161. Wilson, G. T. (1982). Psychotherapy process and procedure: The behavioral mandate. Behavior Therapy, 13, 291–312. Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. Current Directions in Psychological Science, 10, 224–228. Zinbarg, R. E., & Barlow, D. H. (1996). Structure of anxiety and the anxiety disorders: A hierarchical model. Journal of Abnormal Psychology, 105, 181–193.

Nature, Diagnosis, and Etiology of Anxiety and Related Disorders

Article 11

The Phenomenon of Panic David H. Barlow, James Vermilyea, Edward B. Blanchard, Bonnie B. Vermilyea, Peter A. Di Nardo, and Jerome A. Cerny Department of Psychology and Center for Stress and Anxiety Disorders. State University of New York at Albany

One hundred eight patients complaining of anxiety were assessed using a structured interview and were categorized imo the various Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (DSM-III) anxiety disorder and major depression categories. The incidence and nature of panic were examined in a detailed fashion in these patients and included a search for any difference as a function of whether panic was predictable or cued (e.g., panic when encountering a phobic object or situation) or unpredictable. At least 83% of individuals in each diagnostic category admitted to having had a panic attack. Frequency of DSM-III panic attacks distinguished among categories, but few differences emerged among diagnostic categories on severity of the 12 DSM-III symptoms associated with panic. Symptom severity was also similar for patients reporting either predictable or unpredictable panic. However, patients with unpredictable panic reported a significantly greater number of the 12 symptoms than did those with predictable panic. These data point to differences between these two types of panic and underline the importance of discovering the nature of these differences. The phenomenon of panic has, in a brief period of time, achieved a central role within the anxiety disorders. Recent conferences and books devoted to the psychopathology and treatment of the anxiety disorders have focused on panic from both an etiological and a treatment perspective (Barlow & Beck, 1984; Klein & Rabkin, 1981; NIMH, 1983; Tuma & Maser, 1985). Much of this increased attention is due to the delineation of the concept of panic that is found in the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (DSM-III; APA, 1980). Panic is defined as “the sudden onset of intense apprehension, fear or terror often associated with feelings of impending doom” (p. 230). Sudden onset has come to mean 10–15 min during which panic will reach a peak. There is growing evidence that panic and generated anxiety may be qualitatively different. Data supporting this difference come from reports of different

170  David H. Barlow, et al.

developmental antecedents (Raskin, Peeke, Dickman, & Pinkster, 1982), different patterns of familial aggregations (Crowe, Noyes, Pauls, & Slymen, 1983; Harris. Noyes, Crowe, & Chaundry, 1983), and from reports of possibly different genetic contributions (Torgersen, 1983). Finally, patients with Panic Disorder (PD) demonstrate a stronger somatic component of their anxiety than patients with Generalized Anxiety Disorder (GAD) on both detailed physiological assessments and questionnaire measures of anxiety (Barlow et al., 1984). In proposed revisions to DSM-III (Spitzer & Williams, 1983), panic assumes an even greater centrality, with the concept of agoraphobia receding in prominence. Yet we know little about panic. DSM-III implicitly recognizes the ubiquity of panic by noting that panic attacks may accompany withdrawal from substances such as barbiturates, intoxication due to caffeine or amphetamines, and so forth. More important, it is noted that panic attacks may occur in the context of simple or social phobias, but only in the specific phobic situations. That is, there are clear cues for the panic attacks. This is in contrast to spontaneous or unpredictable panics such as those occurring in PD or Agoraphobia With Panic and occasionally other anxiety disorders and affective disorders, where clear cues or triggers are not readily identifiable. Yet the impression has grown that unpredictable panics differ in ways other than lack of cues from predictable panics. Apart from various speculations, this is due largely to one study (Zitrin, Klein, & Woerner, 1978; Zitrin, Klein, Woerner, & Ross, 1983). In this study, a differential response to treatment with imipramine was reported in agoraphobics with panic and in mixed phobics (phobics with unpredictable panic but who lack a broad pattern of avoidance) versus simple phobics. That is, agoraphobics and mixed phobias with (unpredictable) panic who were treated with a variety of structured and unstructured, exposure-based procedures improved somewhat more if they also received imipramine instead of a placebo. This was not true for the simple phobics. Thus the differential drug response suggests qualitative differences between predictable and unpredictable panic. Although this study has been criticized elsewhere on methodological grounds (Emmelkamp, 1982; Marks, 1981; Mathews, Gelder, & Johnston, 1981; Wilson, in press), the voluminous amount of data collected, including direct estimates of spontaneous panic by both patients and therapists and independent evaluators, and the large number of patients included, makes this a landmark contribution. Nevertheless, a close look at the results indicates no drug-diagnosis interactions for spontaneous panic across these three groups. That is, based on the major statistical analytic technique, an analysis of covariance, there was no differential improvement from imipramine among patients with predictable versus unpredictable panic, although, as the authors pointed out, the error variance was large. More fine-grained analyses of drug effects within the diagnostic categories (despite the lack of an interaction) repeats that some measures of panic within the agoraphobia and mixed-phobia categories show a drug effect (most often at the .10 level of significance), whereas others do not. Similar analyses of drug effects in simple phobics either do

The Phenomenon of Panic  171

not show this effect or, more often, are not reported. One ether study also reported differences in drug effects among agoraphobics and simple phobics (Sheehan, Ballenger, & Jacobsen, 1980), but panic was not directly measured, and only nine specific phobics were treated. Overall, these results provide only modest support for a differential drug response on predictable versus unpredictable panic. Inferring the nature of psychopathological states by observing treatment effects is, of course, a weak experimental approach, subject to a logical fallacy (post hoc ergo propter hoc). A more straightforward approach is to directly examine the phenomenon of both predictable and unpredictable panic across various diagnostic categories. Method Subjects

The subjects for this study were 108 individuals who presented for treatment at an anxiety disorders research clinic. Most patients were referred by other professionals or were self-referred. The sample contained 31 men and 77 women, ranging in age from 20 to 67 years with a mean age of 35.3. The breakdown of the sample by primary diagnosis is presented in the results section. Procedures

All individuals inquiring about treatment at the clinic were initially screened over the telephone by one of the professional staff. This brief interview was designed to screen out grossly inappropriate cases such as individuals who were currently alcoholics or who gave clear evidence of having psychotic disorders. All patients were then interviewed on two separate occasions by two of the professional staff using a structured interview schedule, the Anxiety Disorders Interview Schedule (ADIS). We have previously reported on the reliability of differential diagnoses within the anxiety disorders obtained with the instrument for the first 60 patients (Di Nardo, O’Brien, Barlow, Waddell, & Blanchard, 1983) as well as for the 108 patients reported here (Barlow, 1983, in press). The two interviews were conducted within 3 weeks for 61% of the patients, however up to 4–9 weeks intervened in a few cases. (See Di Nardo et al., 1983, for procedural details.) All patients gave written informed consent for these assessments. During each interview, patients were asked about the occurrence of any panic attacks and about their onset and frequency. Specifically, each patient was asked, “Have you had times when you felt a sudden rush of intense fear or anxiety or feelings of impending doom?” If the patient admitted to having had a panic attack, he or she was then asked about each of the 12 symptoms of panic attacks listed in DSM-III (pp. 231–232), including the presence or absence and degree of intensity of each symptom using the individual’s typical

172  David H. Barlow, et al.

or most recent attack as a point of reference. Each symptom was rated by the interviewer on a 5-point scale of seventy (0 = absent, 4 = very severe). Spontaneity of the panic attack was ascertained by careful questioning regarding the circumstances surrounding the panic and by a determination of whether the attacks occurred only in a specific phobic situation or if they ever occurred unpredictably. Following the two interviews, staff arrived at a consensus diagnosis. Results Characterization of the Sample and Presence of Panic

Table 11.1 lists the frequencies of primary diagnosis represented by the sample along with the mean age and sex distribution for each diagnostic category. Also included in this table are the percentages of each subgroup who reported having had a panic attack and the percentage who met the DSM-III criteria for PD. Specifically, reporting that they had a panic attack means that the patient answered yes to the question, “Have you ever had times when you felt a sudden rush of intense fear or anxiety or feelings of impending doom?” “Diagnosis met except for panic frequency” means that the patient reported panic and also reported at least 4 of the 12 symptoms listed as associated with panic in DSM-III (see Table 11.2). “Diagnosis on DSM-III panic criteria” means that patients reported panic, met the four symptom criteria, and reported at least three panics in a 3-week period during the past year. It should he noted that 6 patients eventually received a consensus diagnosis of major depressive disorder. We recognize that these patients are not in any way a random sample of major depression, because they were referred to our clinic specifically for anxiety symptoms. Nevertheless, their data are included for comparison purposes. As is obvious from Table 11.1, panic is a ubiquitous problem among patients with anxiety disorders, with at least 83% of patients in any diagnostic category reporting having at least one panic attack. A similar observation was made by Sheehan and Sheehan (1982). Furthermore, almost all patients who reported a panic attack also reported at least 4 of the 12 symptoms necessary to qualify as a DMS-III panic attack. Only when one reaches the frequency criterion do differences among diagnoses emerge. Thus the value of the frequency criterion in DSM-III is apparent in that the percentage of patients with social phobia, simple phobia, and GAD who met the frequency criterion is markedly lower than for PD. One should not confuse the data in this table with data on unpredictable or spontaneous panic versus predictable panic, which appear in Tables 11.3 and 11.4. For example, simple phobics might report a lower frequency of panic because they successfully avoid specific cues or triggers for panic, such as bridges or planes. In my case, most of the panics that did occur in the categories of simple and social phobias were predictable, cued, panics. The apparent anomaly in Table 11.1 whereby only 82% of patients with a diagnosis of PD meet the full DSM-III criteria including frequency arose because the diagnostic category to which a patient is assigned is a consensus diagnosis arrived at by the staff after presentation of data from both

89

84

50

98

74

82

100

100

29

75

83 83

83

†

83

83

†

x2(6, N = 108) = 14.57*

x2(6, N = 108) = 9.50

x2(6, N = 108) = 7.27

F(6.101) = 4.64**

Note. Categories sharing superscripts are not significantly different by Duncan’s multiple range test. † As were too low to analyze because of missing data on panic frequency. However, for the 3 obsessive-compulsives and 3 depressed patients for whom we had complete data, 100% met the criteria. GAD = Generalized Anxiety Disorder. *p < .03, **p < .001.

33

85

100

1 5 45.3c

Obsessive– Major depressive Statistical comparison compulsive episode

8 2 4 4 36.1ab 28.5a

98

6 11 34.8a

10 9 28.8a

2 39 36.2ab

Men (n) Women (n) Mean age (in years) Reports of panic (%) Diagnosis met except for panic frequency (%) Diagnosis on DSM-III Panic Criteria (%)

2 5 44.6bc

Agoraphobia with Social phobia Simple phobia Panic disorder GAD panic

Variable

Table 11.1 Sample descriptors and incidence of panic across diagnostic categories

68.0ab

61.3a

85.6b

1.33 0.67 1.42 1.58abc

1.38 0.66 1.19 0.97a

1.51 1.61 1.74 2.11bc

1.33 1.67 0.92 1.17 0.58* 0.83 0.92 1.50

Simple phobia

1.33 2.03 0.31 0.90 0.91* 0.88 0.22 0.94

Social phobia

1.71 2.35 1.31 1.35 2.04b 1.28 0.76 1.71

Agoraphobia with panic

83.3b

1.44 1.15 1.71 2.50c

1.29 2.21 1.29 1.06 l.76b 1.83 1.00 1.41

58.3a

0.78 0.94 1.11 .89a

1.67 1.67 1.00 0.83 1.39ab 0.78 0.78 1.11

Panic disorder GAD

90.2b

1.60 1.10 2.00 1.50ab

1.80 2.10 1.00 1.50 l.70ab 1.80 1.10 2.10

Obsessive-compulsive

61.6a

1.60 0.10 1.10 1.10ab

1.10 0.96 0.90 0.70 0.70a 0.90 0.80 1.70

Major depressive episode

0.76 3.02 1.33 5.53a

0.66 2.05 2.30 0.74 5.04a 1.86 1.72 2.12

F

6,91 6.21a

6.91 6.91 6.91 6.90

6.90 6.91 6.91 6.91 6.91 6.91 6.91 6.91

df

Note. Means sharing superscripts are not significantly different by Duncan’s multiple range test. Severity ratings were made for the following scale: 0 = none, 1 = mild, 2 = moderate, 3 = severe, and 4 = very severe. Table of SDs are available from the first author. GAD = Generalized Anxiety Disorder. *p < .001.

Dyspnea (0–4) Palpitations (0–4) Chest pain (0–4) Choking (0–4) Dizziness (0–4) Unreality (0–4) Parasthesias (0–4) Hot or cold flushes (0–4) Sweating (0–4) Fainting (0–4) Shaking (0–4) Fear of going crazy, losing control (0–4) Average percentage of symptoms endorsed

Symptoms

Table 11.2 Average severity ratings of specific panic symptoms across diagnostic categories

The Phenomenon of Panic  175 Table 11.3 Average percentage of symptoms endorsed and average severity ratings of four groups of patients reporting predictable or unpredictable panics

Symptom

Group 1: Group 2: Group 3: Group 4: df Cued Mixed Agoraphobia Panic disorder panic (n = 11) with panic (n = 17) (n = 27) (n = 41)

Average 58.23a percentage of symptoms endorsed Dyspnea 1.04 Palpitations 1.46 Chest pain 0.37a Choking 0.74 Dizziness 0.72a Unreality 0.65a Parasthesias 0.30 Hot or cold 0.70a flushes Sweating 0.98 Faintness 0.56a Shaking 0.93 Fear of going 0.63a crazy, losing control

F

66.50a

85.60b

83.29b

3.85 11.63**

1.29 1.77 0.86ab 0.95 1.00a .77ab 0.73 1.36b

1.67 2.29 1.28b 1.32 1.99b 1.24bc 0.74 1.67b

1.75 2.21 1.29b 1.06 1.76b 1.82c 1.00 1.41b

3.91 2.31 3.92 3.77 3.92 5.31* 3.92 1.60 3.92 10.82** 3.92 4.76* 3.92 3.47 3.92 6.23**

0.95 0.73a 1.14 1.41b

1.48 1.57b 1.70 2.06bc

1.44 1.15ab 1.71 2.50c

3.92 1.85 3.92 5.64** 3.92 4.02 3.91 15.19**

Note: Means sharing subscripts are not significantly different by Duncan’s multiple range test. Tables of SDs are available from the first author. *p < .004. **p < .001.

interviewers. The data for the last row of Table 11.1 are the result of averaging the reported frequency of panic attacks elicited by the two interviews. Thus the average from these two interviews might be slightly below the DSM-III criterion of three panics in 3 weeks, but consideration of the total clinical picture warranted the diagnosis. In like fashion, the 29% of GADs who met the criteria is also based on an average of the two separately elicited sets of frequency data. An interesting trend emerges from consideration of the sex distribution of patients with PD and GAD. Although women outnumber men in the diagnosis of PD in our sample by almost 2 to 1, we find the reverse distribution for GAD. This trend reaches the p < .10 level of significance, x2(1, N = 29), = 2.77. Panic Attack Symptoms

Interrater reliability of panic symptom severity ratings was calculated using the following formula: agreements/agreements + disagreements. Ratings within one point of each other were considered agreements unless one of the scores was zero, in which case the ratings were scored as a disagreement. The

17 21 9 12 12 12 6 15 16 11 15 9

63 78 33 44 44a 44 22 56 63 41 56 35a

8 8 6 7 7 5 5 9 9 7 8 10

n

n

%

Mixed (n = 11)

Cued (n = 27)

80 89 55 64 64a 45 45 82 82 64 73 91bc

% 37 40 31 30 39 28 26 35 35 31 36 37

n 90 98 76 73 95b 68 63 85 93 76 88 90c

%

Agoraphobia with panic (n = 41)

Note: Categories sharing superscripts are not significantly different based on orthogonal planned comparisons. *p < .001.

Dyspnea Palpitations Chest pain Choking Dizziness Unreality Parasthesias Hot or cold flushes Sweating Faintness Shaking Fear of going crazy, losing control

Symptom

14 15 12 10 17 16 12 14 14 11 16 17

n

82 88 71 54 100b 94 71 82 94 65 94 100c

%

Panic disorder (n = 17)

3 3 3 3 3 3 3 3 3 3 3 3

df

Table 11.4 Number of patients and percentage of group endorsing each symptom across groups having predictable or unpredictable panic

0.93 8.33 13.13 1.43 30.96* 13.06 14.22 8.75 0.95 8.50 13.03 38.26*

x2

The Phenomenon of Panic  177

mean percentage of interrater agreement was 76.96% (SD = 8.77) with a range from 70.59% to 94.12%. For those individuals who admitted to having had a panic attack, the severity ratings of the 12 DMS-III symptoms of panic are listed in Table 11.2. These ratings were each subjected to one-way analysis of variance (ANOVA) using the Bonferroni technique to adjust experiment-wise alphas (Marascuilo & Levin, 1983). On two symptoms, there were significant differences in the severity ratings: dizziness, and fear of going crazy or losing control. Patients with PD and Agoraphobia With Panic report more severe dizziness than do patients with social phobia, simple phobia, or major depressive episodes. Those with GAD or obsessive-compulsive disorder do not differ from either of these two groups. For the symptom of fear of going crazy or of losing control, patients with PD have the highest average severity rating, midway between moderate and severe. This rating is significantly higher than that given patients with social phobia, GAD, or major depressive disorder. The average rating for those patients with Agoraphobia With Panic also exceeds the scores of those with social phobia and GAD. The overall severity scores are somewhat low due to the fact that many patients who reported panic and at least 4 of the 12 symptoms, nevertheless, did not have all the symptoms. Therefore, a number of zeros on each particular symptom lowered the average severity rating. The difference in percentage of symptoms endorsed in those patients reporting panic in each diagnostic group are found in the last row of Table 11.2. Agoraphobics With Panic, PDs, and Obsessive-Compulsives all report a high percentage of the 12 symptoms (85.6%, 83.3%, and 90.2% respectively). This is a significantly higher percentage than all other diagnostic groups except simple phobics. In general, few differences in panic phenomenology emerge between diagnostic categories On only 2 out of 12 symptoms, as well as the average percentage of symptoms endorsed, are there any differences. When differences exist, Agoraphobics With Panic, PDs, and Obsessive-Compulsives tend to score slightly higher on symptom severity and percentage of symptoms reported, whereas social phobics clearly score lower. The preceding analyses obscure a direct comparison between predictable and unpredictable panic. Although Agoraphobics With Panic and PDs suffer from unpredictable panic by diagnostic definition, other patients, who do not meet DSM-III criteria for those two diagnostic categories, also experience unpredictable panic occasionally (Barlow, Di Nardo, Vermilyea, Vermilyea, & Blanchard, 1984; Zitrin et al., 1980, 1983). Typically, these may be simple or social phobics who, if they experience panic, report that it occurs in the presence of their feared situations: however, they also report recurring spontaneous or unpredictable panic. Some patients with obsessive-compulsive disorder, GAD, or major depressive disorder also report unpredictable panics. In the case of major depressive disorders, these unpredictable panics were considered part of the depressive syndrome until recently (e.g., Leckman, Weissman, Merikangas, Pauls, & Prusoff, 1983; Leckman, Merikangas, Pauls, Prusoff, & Weissman, 1983). To contrast predictable and unpredictable panics more directly, patients who had never reported spontaneous panic were compared to three groups

178  David H. Barlow, et al.

who did report spontaneous panic. Patients with a diagnosis of depression were not included in order to restrict the analyses to the anxiety disorders. Group 1 consists of patients who reported never experiencing an unpredictable panic. Group 2 consists of simple and social phobics as well as patients with GAD who reported experiencing at least one unpredictable panic during their lifetime. This group has been referred to as mixed by previous investigators (e.g., Zitrin et al., 1983). Group 3 consists of Agoraphobics With Panic, and Group 4 consists of PDs. These data are presented in Table 11.3, which lists the severity ratings of the 12 DMS-III symptoms of panic for those individuals who admitted having had a panic attack in the four groups. As in Table 11.2, these ratings were subjected to a one-way ANOVA, and alpha levels were adjusted using the Bonferroni technique. As one can see, Agoraphobics With Panic and PDs endorse significantly more of the 12 symptoms than do the mixed group or the cued panic group. Both PDs and Agoraphobics With Panic endorse over 80% of the 12 symptoms. In this analysis, unpredictable panics also achieve significantly higher severity ratings on most of the symptoms. This is particularly true for the Agoraphobics With Panic and the PDs, with the mixed groups occupying a somewhat intermediate position. However, if one recalculates average severity ratings based on only those patients who actually reported having each symptom, a different picture emerges. In this analysis, no differences among symptoms emerged, and in no case do symptoms among the group of PD patients exceed severities of the cued panic group. This result indicates that almost all of the differences found in Table 11.3 are due to the tendency for those experiencing unpredictable panics to endorse a larger number (a greater percentage) of the 12 symptoms, thereby inflating some of the severity scores, Because differences emerged in the number of symptoms endorsed. Table 11.4 examines the patterning of symptoms across predictable and unpredictable panic with chi-square analyses. Again, the Bonferonni technique was used to adjust for experiment-wise alpha inflation. The pattern of chi-square tests matches the result found in Table 11.3, which again suggests that patients with Agoraphobia With Panic and PD endorsed almost every symptom compared to patients with predictable panic. Because of the highly conservative nature of the Bonferroni adjustment, however, significant differences among groups emerged only on the symptoms of dizziness and fear of going crazy or losing control. Discussion These data confirm a DSM-III convention that panic occurs across the anxiety disorders but in some categories, such as simple phobia, clear cues exist for panic. Furthermore, whether cues exist or not, panic attacks in almost all cases meet the DSM-III criteria in that at least 4 of the 12 panic symptoms are reported. Only in frequency do differences among groups begin to emerge, with agoraphobics and PDs reporting the highest frequencies relative to simple and social phobics and GADs. It should also be noted that in the small number of obsessive-compulsive and major depressive patients for whom frequency

The Phenomenon of Panic  179

data were available, all reported panic. Once again the depressed patients were referred to our clinic for anxiety problems and therefore represent a skewed sample of major depression. Furthermore, if one looks at the symptomatology of panic across diagnostic categories irrespective of the presence of cues, few differences emerged. Agoraphobics With Panics, PDs, and Obsessive-Compulsives reported somewhat more severe symptomatology, but only on 2 of the 12 symptoms. This seems due to the tendency of these three groups to endorse a greater number of the 12 symptoms than did the other groups. But differences in panic among diagnostic categories are confounded by the fact that both predictable and unpredictable panics occurred in all diagnostic groups. The implications of the ubiquity of panic for currently accepted diagnostic systems are spelled out elsewhere (Barlow, 1985; Barlow et al., 1984; Cerny, Himadi, & Barlow, in press). One gets a clearer view of differences in these two hypothetically important types of panic by contrasting the symptomatology within each type. Here, no differences in severity emerge, but differences in the number of symptoms reported are clear, because more people with unpredictable panic report symptoms such as dizziness, parasthesias, shaking, pain in chest, and tear of going crazy or loss of control. In fact, because over 90% of patients with unpredictable panic (PDs and Agoraphobia With Panic) reported loss of control and dizziness to symptoms and because these two features seem to discriminate unpredictable from predictable panics so well, these factors may emerge as defining characteristics of unpredictable panic. It is possible, of course, that these data support a qualitative difference among these types of panic due, perhaps, to differential underlying biological processes (see, e.g., Klein & Rabkin, 1981). Another possibility is that the difference lies in the cues that serve as the major defining characteristic of predictable versus unpredictable panic. That is, the cues for panic in simple and social phobia are obvious and readily reported by the patient. Cues in obsessive-compulsive disorder are also usually obvious and are tied to identifiable, aversive, intrusive thoughts or dangerous or contaminating external situations that are not readily escapable. However, cues in PD and Agoraphobia With Panic are not always so obvious, although panic attacks within agoraphobia often become associated with leaving a safe place or a safe person. It is possible that the inability to identify consistent cues is due to inadequate measurement of panic. As noted above, most clinicians and researchers to date have simply asked the patient to recall panic attacks occurring during the past week or perhaps during a longer period of time. At the same time, these patients are asked to recall antecedents. From the point of view of psychological measurement, a more satisfactory procedure would involve prospective self-monitoring of panic so that antecedents or cues can be recorded at the time of the panic over a period of days or weeks. For example, a depressed patient presented as an agoraphobic at our clinic with reports of panic associated with leaving the house, going shopping, and so forth. A closer prospective analysis revealed that these panic attacks were triggered by having to make a decision, however small, concerning making out a grocery list, whether go to a friend’s house or not, and so on. The cues for these panics, then, were

180  David H. Barlow, et al.

consistent with a diagnosis of major depression. Our own impression is that cues for panic in both Agoraphobics With Panic and PDs are usually associated with mild exercise, sexual relations, sudden temperature changes, stress, or other cues that alter physiological functioning in some discernable way, albeit outside the patient’s awareness. The fact that cues associated with altered physiological functioning may trigger “unpredictable” panics through a process of interoceptive conditioning has long been considered a possibility (Evans, 1972; Mineka, in press; Razran, 1961; Ackerman & Sachar, 1974). Although this hypothesis awaits confirming empirical evidence, it would account for the well-known marked sensitivity to physiological changes and bodily sensations on the part of patients with unpredictable panics (see, e.g., Chambless, 1982; Chambless & Goldstein, 1981). Because symptoms of panic in their most severe form are also cues for panic in a milder form, an ever-spiraling repetitive cycle of panic attacks occurs that can by cued by changes in humidity, temperature, mild excessive stress, or even relaxation (Cohen, Barlow, & Blanchard, 1985; Heide & Borkovec, 1983, 1984). The fact that these cues are subtle accounts for the unpredictability of the panic by the patient and the endorsement of feelings of going crazy or losing control. A successful search for antecedents to unpredictable panics as well as a careful analysis of the effects of both pharmacological and behavioral treatments on panic will require sophisticated prospective monitoring of panic attacks and their potential antecedents. Note This research was supported by National Institute of Mental Health Grant MH 39096.

References Ackerman, S. H., & Sachar, E. J. (1974). The lactate theory of anxiety: A review and reevaluation. Psychosomatic, Medicine, 36, 69–79. American Psychiatric Association. (1986). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington D.C.: Author. Barlow, D. H. (1983, October). The classification of anxiety disorders. Paper presented at the Conference of DSM-III: An interim appraisal. Sponsored by the American Psychiatric Association, Washington D.C. Barlow, D. H. (in press). The dimensions of anxiety disorders. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders. Hillsdale, NJ: Erlbaum. Barlow. D. H., & Beck, J. G. (1984). The psychosocial treatment of anxiety disorders: Current status, future directions. In J. B. W. Williams & R. L. Spitzer (Eds.), Psychotherapy research: Where are we and where should we go? (pp. 29–66). New York: Guilford Press. Barlow, D. H., Cohen, A. S., Waddell, M. T., Vermilyea, B., Klosko, J. S., Blanchard, E. B., & Di Nardo, P. A. (1984). Panic and generalized anxiety disorders: Nature and treatment. Behavior Therapy, 15, 431–449. Barlow, D. H., Di Nardo, P. A., Vermilyea, J. A., Vermilyea, B. B., & Blanchard, E. B. (1984). The classification of anxiety disorders: Co-morbidity and depression within the anxiety disorders. Manuscript submitted for publication. Cerny. J. A., Himadi, W., & Barlow, D. H. (in press). Issues in diagnosing anxiety disorders. Journal of Behavioral Assessment.

The Phenomenon of Panic  181 Chambless, D. I., (1982). Characteristics of agoraphobics. In D. L. Chambless & A. J. Goldstein (Eds.), Agoraphobia (pp. 1–18). New York: Wiley. Chambless, D. L., & Goldstein, A. J. (1981). Clinical treatment of agoraphobia. In M. Mavissakalian & D. H. Barlow (Eds.), Phobia (pp. 103–144). New York: Guilford Press. Cohen, A. S., Barlow, D. H., & Blanchard, A.J. (1985). Psychophysiology of relaxationassociated panic attacks. Journal of Abnormal Psychology, 9, 96–101. Crowe, R. R., Noyes, R., Pauls, D. L., & Slymen, D. (1983). A family study of panic disorder. Archives of General Psychiatry, 40, 1065–1069. Di Nardo, P., O’Brien. G. T., Barlow, D. H., Waddell, M. T., & Blanchard, E. B. (1983). Reliability of DSM-III anxiety disorder categories using a new structured interview. Archives of General Psychiatry, 40, 1070–1075. Emmelkamp, P. M. G. (1982). Phobic and observe–compulsive disorders: Theory, research, and practice. New York: Plenum Press. Evans, I. M. (1972). A conditioning model of a common neurotic pattern—Fear of fear. Psychotherapy: Theory, Research and Practice, 9, 238–241. Harris, E. L., Noyes, R., Crowe, R. R., & Chaundry, D. R. (1983). Family study of agoraphobia. Archives of General Psychiatry, 40, 1061–1064. Heide, F. J., & Borkovec, T. D. (1983). Relaxation-induced anxiety: Paradoxical anxiety enhancement due to relaxation training. Journal of Consulting and Clinical Psychology, 51, 171–182. Heide, F. J., & Borkovec, T. D. (1984). Relaxation-induced anxiety: Mechanisms and theoretical implications. Behaviour Research and Therapy, 22, 1–12. Klein, D. F., & Rabkin, J. (Eds.), (1981). Anxiety: New research and changing concepts. New York: Raven Press. Leckman, J. F., Merikangas, K. R., Pauls, D. L., Prusoff, B. A., & Weissman, M. M. (1983). Anxiety disorders associated with episodes of depression: Family study data contradict DSM-III convention. American Journal of Psychiatry, 140, 880–882. Leckman, J. F., Weissman, M. M., Merikangas, K. R., Pauls, D. L., & Prusoff, B. A. (1983). Panic disorder and major depression. Archives of General Psychiatry, 40, 1055–1060. Marascuilo, L. A., & Levin, J. R. (1983). Multivariate statistics in the social sciences: A researcher’s guide. Monterey, CA: Brooks/Cole. Marks, I. (1981). Cure and care of neuroses: Theory and practice of behavioral psychotherapy. New York: Wiley. Mathews. A. M., Gelder. M. G., & Johnston, D. W. (1981). Agoraphobia: Nature and treatment. New York: Guilford Press. Mineka, S. (in press). Animal models of anxiety-based disorders: Their usefulness and limitations. In A. H. Turna & J. D. Maser (Eds.), Anxiety and the anxiety disorders. Hillsdale, NJ: Erlbaum. NIMH. (1983, September). NIMH Clinical Research Branch Conference on Anxiety and Anxiety Disorders. Bethesda, MD: Sterling Forest Conference Center. Raskin, M., Peeke, A. V. S., Dickman, W., & Pinkster, A. (1982) Panic and generalized anxiety disorders: Developmental antecedants and precipitants. Archives of General Psychiatry, 39, 687–689. Razran, G. (1961). The observable unconscious and the inferable conscious in current Soviet psychophysiology: Interoceptive conditioning, semantic conditioning, and the orienting reflex. Psychological Review, 68, 81–150. Sheehan, D. V., Ballenger, J., & Jacobsen, G. (1980). Treatment of endogenous anxiety with phobic, hysterical, and hypochondriacal symptoms. Archives of General Psychiatry, 37, 51–59.

182  David H. Barlow, et al. Sheehan, D. V., & Sheehan, M. S. (1982). The classification of anxiety and hysterical states. Part I. Historical review and empirical delineation. Journal of Clinical Psychopharmacology, 2, 235–243. Spitzer, R. L., & Williams, J. B. W. (1983, September). Proposed revisions of the DSM-III Classification of anxiety disorders based on research and clinical experience. Paper presented at the conference on Anxiety and Anxiety Disorders, sponsored by the National Institute of Mental Health. Tuxedo, New York. Torgersen, S. (1983). Genetic factors in anxiety disorders. Archives of General Psychiatry, 40, 1085–1089. Tuma, A. H., & Maser, J. D. (in press). Anxiety and the anxiety disorders. Hillsdale, NJ: Erlbaum. Wilson, G. T. (in press). Fear reduction methods and the treatment of anxiety disorders. In G. T. Wilson, C. M. Franks, K. D. Brownell, & P. C. Kendall (Eds.), Annual review of behavior therapy. New York: Guilford Press. Zitrin, C. M., Klein, D. F., & Woerner, M. G. (1978). Behavior therapy, supportive psychotherapy, imipramine, and phobias. Archives of General Psychiatry, 35, 307–321. Zitrin, C. M., Klein, D. F., & Woerner, M. G. (1980). Treatment of agoraphobia with group exposure in vivo and imipramine. Archives of General Psychiatry, 37, 63–72. Zitrin, C. M., Klein, D. F., Woerner, M. G., & Ross, D.C. (1983). Treatment of Phobias I: Comparison of imipramine hydrochloride and placebo. Archives of General Psychiatry, 40, 125–139.

Article 12

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference David H. Barlow Center for Stress and Anxiety Disorders, State University of New York at Albany

Recent research findings have implicated, in a preliminary manner, five factors that seem to differentiate sexually functional subjects from sexually dysfunctional subjects suffering from inhibited sexual excitement. These factors include differences in affect during sexual stimulation, differences in self-reports of sexual arousal and perception of control over arousal, distractibility during sexual stimulation, and differential sexual responding while anxious. These findings suggest a working model of sexual dysfunction that is based on cognitive interference and anxiety. Implications of this model for the treatment of sexual dysfunction are suggested. Sex therapy has changed very little since Masters and Johnson (1970). Although direct and brief behavioral approaches to psychogenic sexual dysfunction began to appear in the late 1950s and early 1960s (Brady, 1966; Lazarus, 1963; Wolpe, 1958), Masters and Johnson’s work influenced psychotherapeutic approaches to this problem in a manner unprecedented for other psychological disorders. It is little wonder that innovations since that time have been limited largely to small variations of the Masters and Johnson approach reflecting the inclinations of various sex therapists (e.g., Hartmann & Fithian, 1972; Kaplan, 1981). Typically, all of these approaches, including the original Masters and Johnson techniques, contain a variety of techniques based loosely on hypothetical causes of sexual dysfunction, such as performance anxiety. The success rates of these approaches and the commonalities among them have been reviewed many times (e.g., Cooper, 1981; Crown & D’Ardenne, 1982; Kuriansky & Sharpe, 1981; Marks, 1981; Mills & Kilmann, 1982). It seems that with or without variations in the basic treatment, approximately one half to two thirds of sexually dysfunctional patients will show some improvement immediately following treatment. For erectile dysfunction, however (e.g., inhibited sexual excitement in males from diagnoses in Diagnostic and Statistical Manual of Mental Disorders, DSM-III; American Psychiatric Association, 1980), the figure can be as low as 30% (Crown & D’Ardenne, 1982). Furthermore, Levine

184  David H. Barlow

and Agle (1978) pointed out that these “improved” patients with erectile dysfunction are quite unstable in their sexual functioning over time and are certainly not cured. These more pessimistic results for erectile dysfunction are paralleled by increasing reports from sex therapists of difficulties in replicating the success of Masters and Johnson, even in the area of erectile dysfunction where some of their highest failure rates were originally reported (e.g., Zilbergeld & Evans, 1980). In any case, in recent years interest in surgical intervention for erectile dysfunction has increased. Surgical alternatives to the treatment of erectile difficulties have become particularly popular with the advent of diagnostic techniques such as evaluating nocturnal penile tumescence (NPT; Karacan, 1970). The procedures are thoroughly reviewed elsewhere in this series. But, as Fisher et al. (1979) pointed out, recordings of NPT are only an aid in clinical diagnosis and not a definitive criterion. Indeed, many patients, later proven to be psychogenically impotent, display depressed NPTs, whereas others with known organic etiologies seemingly display normal NPTs. In addition, it has become an increasingly common clinical report that even patients suffering from erectile difficulties with known organogenic etiology, such as diabetes, hypospadias, alcohol neuropathy, or other peripheral neuropathies, are capable of erectile response (Sakheim, 1984) and will on occasion respond to psychological sexual therapy. As Beutler and Gleason (1981) and others pointed out, organic and psychogenic etiologies can coexist and mask one another. Nevertheless, those approaching the problem from a surgical versus a psychological perspective continue to accuse each other of doing unnecessary surgery on the one hand or performing lengthy, unnecessary psychological therapy on the other. This state of affairs seems a clear indication of the continuing unevenness of research in sexual dysfunction in general and inhibited sexual excitement specifically. That is, most research to date has been focused on treatment outcome rather than more basic knowledge in the area of psychology, psychopathology, or physiology. The fact is, little is known about the physiology of the sexual response on the one hand, for example, the hemodialysis of penile erections (Bennett, 1982; Schiavi, 1981), or the psychological mechanisms of action responsible for erectile difficulties or related sexual dysfunctions on the other. More recently, data have accumulated that contradict, in part, some of the prevailing notions of the causes of sexual dysfunction and inhibited sexual excitement in particular, on which most treatments are based. This article reviews new data on factors contributing to erectile dysfunction in men, and to some extent, inhibited sexual excitement in women and suggests a working model of sexual dysfunction that may have substantial implications for treatment. Before describing this model, a brief review of the relation of anxiety as well as a variety of cognitive sets to sexual arousal is necessary. Role of Anxiety Anxiety is considered to play a role in the development and maintenance of sexual dysfunctions for both men and women by writers of virtually every theoretical persuasion. For example, both Fenichel (1945) and Wolpe (1958)

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  185

have considered anxiety to be a contributing factor to the development of the various types of sexual dysfunctions in men and women. More recently, Masters and Johnson (1970) and Kaplan (1974, 1981) emphasized the role of anxiety in the short-term treatment of sexual dysfunctions. Masters and Johnson (1970) underlined the importance of performance fears as a major concern to individuals and couples experiencing sexual dysfunctions. Fears of inadequacy often coexist with performance fears deterring effective sexual functioning according to Masters and Johnson. Kaplan (1974, 1981) has also given fear of failure a primary role in the development of sexual anxieties and has added other fears, such as demands for performance on the part of the partner and excessive need to please partners, to the list of sexual anxieties. These sexual anxieties are seen as preventing an individual from experiencing sexual arousal and, in fact, as inhibiting autonomic nervous system functioning to such an extent that physiological arousal is impossible (Kaplan, 1974, 1981). For both Masters and Johnson and Kaplan, a major task of therapy is overcoming performance fears and feelings of sexual inadequacy as well as other sources of sexual anxiety. In a similar vein, as early as 1958 Wolpe suggested the use of systematic desensitization in the treatment of sexual dysfunctions, with the elimination of anxiety as the goal of treatment. The fact that anxiety is thought to be the most important etiological factor in sexual dysfunction is surprising because the evidence is based purely on clinical inference and not on empirical data. Schiavi pointed out as late as 1976 that there were “no studies of men with erectile dysfunctions that have assessed the pattern of autonomic arousal and penile tumescence in response to erotic stimulation” (Schiavi, 1976, p. 564). In fact, there is both clinical and experimental evidence that anxiety can facilitate sexual arousal as indicated by physiological assessment (Beck & Barlow, 1984; Norton & Jehu, 1984). For example, from a clinical perspective, Ramsey (1943) conducted a study on adolescent boys and reported that approximately 50% of the boys noted erections from some type of nonerotic stimulus. The situation in which the nonerotic responses occurred usually involved elements of fear, excitement, or other emotional situations. These included such events as accidents, near accidents, being chased by the police, fear of being punished, and so forth. Bancroft (1970) noted similar examples. The theoretically interesting account of male sexual performance during sexual threats and molestation by women, reported by Sarrel and Masters (1982), adds to this clinical evidence. In this report, men were able to perform sexual intercourse (and therefore were able to attain adequate erection) repeatedly despite ongoing direct threats with knives and other weapons if they failed. This phenomenon has also been noted in some paraphilias as well, because many exhibitionists or voyeurs, for example, find it impossible to get sexually aroused and erect without first experiencing fear or anxiety concerning being apprehended. Stoller (1976) provided a theoretical rationale for this in his analysis of sexual excitement. From a basic physiological perspective both Barfield and Sachs (with rats; 1968) and Redmond, Kosten, and Reiser (with adult males; 1982) noted that severely elevated levels of “anxiety” were associated with increased sexual arousal and/or activity. This topic has also been examined from a social

186  David H. Barlow

psychological perspective. A number of studies (i.e., Berscheid & Walster, 1974; Brehm, Gatz, Goethals, McCrimmon, & Ward, 1978; Dutton & Aron, 1974; Riordan, 1979) have found that experimentally induced anxiety can increase interpersonal or sexual attraction. More recently, the relation of anxiety to sexual arousal has been examined directly in the laboratory. Hoon, Wincze, and Hoon (1977) preexposed sexually functional women to 2-min film sequences designed to elicit either anxiety or neutral responses and presented them in counterbalanced order while measuring sexual arousal. The results indicated that exposure to an anxiety-producing film subsequently produced increases rather than decreases in sexual arousal during the viewing of an erotic film as compared with preexposure to a neutral film. Wolchik et al. (1980) replicated this study, in part, with sexually functional men. Once again it was demonstrated that a validated anxiety-producing preexposure film could enhance rather than inhibit sexual arousal. Despite this diverse literature, Wolpe (1978, p. 453) pointed out that increases in sexual arousal subsequent to an anxiety-producing event such as a film, as in the Hoon et al. (1977) experiment, might be due to an “anxiety relief” phenomenon rather than a facilitatory effect of anxiety. To address this issue, Barlow, Sakheim, and Beck (1983) examined the effects of anxiety simultaneous with sexual arousal in young sexually functional volunteer men. Anxiety was manipulated by a shock threat during an explicit, erotic film. Two types of experimental instructions were employed; each was signaled by a different light during the film. In one condition subjects were told that there was a 60% chance of shock if they did not achieve adequate arousal (contingent threat). In the second condition subjects were told that the chance of shock was unrelated to their arousal or any other response (noncontingent threat). These two forms of shock threat instruction were used in an attempt to produce personal performance pressures or demands (contingent) and to differentiate this performance anxiety from more generalized anxiety (noncontingent). These two conditions were compared with a no-shock condition, that is, a light signaling no shock while watching the erotic movie. The results indicated that noncontingent anxiety increased sexual arousal, compared with the no-shock condition. But, in an unexpected development, even the demand condition where subjects were told there was a 60% chance they would be shocked if they did not achieve adequate arousal increased sexual arousal when compared with the no-shock condition. In fact, this produced the highest overall sexual arousal of all three conditions (see Figure 12.1). Despite these data, it is still possible that sexually dysfunctional subjects could react very differently from sexually functional subjects to erotic stimuli while experiencing anxiety. We now have completed an experiment testing this exact same paradigm with sexually dysfunctional and matched sexually functional men (Beck, Barlow, Sakheim, & Abrahamson, 1984b). Our previous findings were partially replicated in that noncontingent threat produced significantly more sexual arousal than did a no-shock control condition for our sexually functional men. Unlike the previous report, however, contingent shock threat did not elevate sexual arousal over the control condition (although it did not inhibit sexual arousal either). It is possible that this represents an age difference. The previous experiment was conducted with relatively young volunteers with a mean age of 26 years old. The sexually functional subjects

Penile Circumference Change Baseline (mm)

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  187 30

Contingent Shock Threat Noncontingent Shock Threat No Shock Threat

25 20 15 10 5

1

2

3

4

5

6 Epochs

7

8 9 Signal Light Off

10

11

12

Figure 12.1 Average penile circumference change in millimeters (mm.) for each 15-s epoch during each of three conditions: no shock threat, noncontingent shock threat, and contingent shock threat. (From “Anxiety Increases Sexual Arousal” by D.H. Barlow, D.K. Sakheim, and J.G. Beck, 1983, Journal of Abnormal Psychology, 92, p. 52. Copyright 1983 by the American Psychological Association. Reprinted by permission.)

in this experiment, on the other hand, were matched with our older sexually dysfunctional population with a mean age of 38 years old. But the more important finding was that unlike our sexually functional subjects, dysfunctional subjects evidenced significantly less sexual arousal during shock threat conditions. Thus, anxiety seems to affect functionals and dysfunctionals in very different ways. To reiterate, the first major difference between our sexually functional and sexually dysfunctional subjects is that anxiety, operationalized as shock threat, increases sexual arousal in sexually functional subjects and decreases sexual arousal in sexually dysfunctional subjects. We will return to this issue later. All of this must be seen, of course, in the context of anxiety as a construct (Barlow, Mavissakalian, & Schofield, 1980; Lang, 1968; Mavissakalian & Barlow, 1981) in which three separate response systems exist that may not be perfectly correlated. It is possible, therefore, that the physiological or somatic component of anxiety, when increased, facilitates sexual arousal in an additive way, as suggested in the emotional transfer literature (e.g., Zillman, 1983) because these two responses share many common components (e.g., increased heart rate and increased blood pressure), whereas the cognitive component of anxiety, such as distracting performance-related negative cognitions, might have a different effect. Role of Cognitive Interference If autonomic arousal or the physiological component of anxiety is not fully correlated with other response systems, as the literature in anxiety disorders

188  David H. Barlow

suggests (Barlow & Mavissakalian, 1981), then it is possible that other response systems and particularly the cognitive response system are of functional significance in sexual dysfunction. In fact, a literature is beginning to emerge suggesting that cognitive processes influence sexual arousal in a major way. For example, Briddell et al. (1978) demonstrated the strong effect of expectancy on sexual responding in sexually functional subjects, a finding that has since been replicated in part (Lansky & Wilson, 1981). Several years ago Laws and Rubin (1969) and Henson and Rubin (1971) demonstrated that sexually functional subjects could suppress erections while watching erotic films if asked to do so. Abel, Barlow, Blanchard, and Mavissakalian (1975) demonstrated that a group of men with homosexual arousal patterns could also suppress erections. Cerny (1978) replicated this finding, in part, with sexually functional women. This phenomenon has been followed up in subjects with other sexual orientations (Abel, Blanchard, & Barlow, 1981). In any case, when this suppression occurs, presumably the mechanism by which subjects suppress erections is self-distraction or a shift in attention. Geer and Fuhr (1976) experimentally isolated the effects of distraction alone by noting decrements in sexual arousal in normal volunteers using a dichotic listening paradigm. They had subjects listen to increasingly distracting cognitive tasks in one ear involving adding numbers and so forth while listening to erotic audiotapes in the other. Using a similar procedure, Farkas, Sine, and Evans (1979) found the same effect of distraction in normal volunteers who were watching explicit erotic videotapes. There is a wide agreement among clinicians that several types of cognitive activities have a marked effect on sexual arousal and therefore may play an important role in the genesis and/or maintenance of sexual dysfunction. It is possible that these cognitive activities operate through the mechanism of cognitive interference, or distraction. For example, Kaplan (1974) outlined various types of distracting thoughts that interfere with potency among many of her clients. Masters and Johnson (1970, p. 10) discussed distraction as an inhibitor of potency, calling it “spectatoring.” In other places these well-known clinicians discussed other potentially competing cognitions, using such labels as performance demand characteristics or failure self-statements. Of course, this hypothesis requires a direct examination of the effects of cognitive interference on sexual arousal in both sexually functional and sexually dysfunctional subjects. We have now completed an experiment testing the effects of neutral distraction on sexual arousal in sexually dysfunctional and matched sexually functional men (Abrahamson, Barlow, Sakheim, Beck, & Athanasiou, 1985). In this experiment subjects viewed an erotic film while listening to an unrelated audiotape portraying a nonsexual passage from a popular novel. Subjects were told they would be questioned on material from the audiotape. Both groups achieved adequate and equivalent levels of penile responding under the no-distraction condition, and the sexually functional subjects evidenced significant detumescence during distraction. But, in a surprising development, sexually dysfunctional subjects were not affected by distraction and maintained tumescence (see Figure 12.2). Nevertheless, sexually dysfunctional subjects consistently underestimated their erections on a continuous subjective measure of arousal, a point to which we will return later.

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  189 Functionals (N = 10)

Strain Gauge: Circumference Change From Baseline (mm)

16

Dysfunctionals (N = 10)

Key

15

Distraction No Distraction

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0

2

4

6

8 10 12 14 16 18

0

2

4

6

8 10 12 14 16 18

Epochs (= 10 s)

Figure 12.2 Mean Strain Gauge Responding Across Subjects by Epoch During Distraction and No Distraction: Sexually Functional (left) Versus Sexually Dysfunctional (right) Subjects. (From “Effects of Distraction on Sexual Responding in Functional and Dysfunctional Men” by D.J. Abrahamson, D.H. Barlow, D.K. Sakheim, J.G. Beck, and R. Athanasiou, 1985, Behavior Therapy, 16, p. 509. Copyright 1985 by the Association for Advancement of Behavior Therapy. Reprinted by permission.)

Thus, these data replicate the effects of neutral distracting stimuli previously found on young sexually functional subjects with our older sexually functional subjects. But once again sexually dysfunctional subjects seem to respond in a markedly different and somewhat surprising way. Thus, the second major difference between our sexually functional and dysfunctional subjects is that neutral distraction decreases sexual arousal in sexually functional subjects but does not alter sexual arousal in dysfunctional subjects. In view of these findings using neutral distracting stimuli (e.g., listening to unrelated material, adding numbers, etc.), an examination of the effects of other more clinically relevant cognitive sets becomes particularly interesting. We will note here that there seems to be some surprising parallels between these more clinically relevant cognitive sets and the effects of neutral distraction. In one experiment from our laboratory, sexual arousal was examined in sexually functional subjects and sexually dysfunctional subjects while viewing a film in which the partner was perceived as highly sexually aroused or not very aroused (along with an ambiguous condition; Beck, Barlow, & Sakheim,

190  David H. Barlow

1983). Clear group differences emerged when subjects were focusing on a partner who was highly aroused. In this condition, which subjects later reported as very “demanding,” sexually functional subjects showed increases in sexual arousal, whereas sexually dysfunctional subjects demonstrated markedly lower sexual arousal. In a further study, Abrahamson, Barlow, Beck, Sakheim, and Kelly (1985) replicated these findings using audiotaped erotic stimuli rather than videotapes. What is interesting about these studies again is a differential response between matched sexually functional subjects and dysfunctional subjects. Also, post hoc analyses indicate that sexually dysfunctional men seemingly begin concentrating on distracting performance-related concerns when viewing a more responsive partner, whereas sexually functional controls attend to and concentrate on erotic cues under the same conditions. In an interesting and important experiment along the same lines, Heiman and Rowland (1983) also reported a differential response between sexually functional and dysfunctional men who received either sensate focus, which required focusing on internal sensations of arousal, or performance demand instructions prior to listening to an erotic audiotape. Sexually functional men showed greater responsivity to the erotic audiotape when the tape was preceded by demand instructions that asked subjects to become as aroused as possible and maintain it. Sexually dysfunctional men, on the other hand, showed higher responding following sensate focus instructions. Once again, when focusing directly on sexual and performance-related cues, sexually functional men achieve more arousal (relative to focusing on less intense or alternative cues), whereas sexually dysfunctional men evidence less arousal. One additional experiment (Sakheim, Barlow, Beck, & Abrahamson, 1984) bears on these findings. The purpose was to examine the effect of availability of genital focus on subjective and physiological responding. Young sexually functional male volunteers were shown erotic films of varying intensity. In one session the subject was not allowed to view his genital responding (a sheet covered the genital area), whereas in the second session he was able to visually attend to tumescence. At low and moderate levels of erotically intense stimuli (films), subjects displayed less responding when genital cues were visually available compared with when they were not available. But during higher erotic intensity, significantly more tumescence was displayed when cues were available. In view of the previous findings, it might be predicted that sexually dysfunctional subjects would show the opposite results in that opportunities to view sexual responsiveness might produce distracting performance-related concerns. Thus, sexually functional subjects and sexually dysfunctional subjects react very differently not only to neutral distraction but also to a variety of more sexually related cognitive sets that can be described as either sexual-performance related (e.g., demand) or non-sexual-performance related (e.g., sensate focus). This constitutes the third major difference between sexually functional and dysfunctional subjects. It is possible that performance-related demands distract sexually dysfunctional subjects from erotic cues, whereas these same subjects seem relatively unaffected by non-performance-related self-focus (sensate focus) or by neutral distractors. Sexually functional subjects, on the other hand, find their sexual arousal facilitated by sexual-performance-related cognitive states and expectancies but are distracted by non-performance-related

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  191

self-focus or by neutral distractors. Further evidence bearing on those hypotheses is presented next. Interaction of Anxiety and Cognitive Interference Of course, emotional arousal and cognitive processes seldom exist in the type of isolation made possible by experimental manipulations. Ultimately, the interaction of these variables will have to be studied, and some preliminary data are now available. The next experiment seems to provide some insight into the relation among anxiety, cognitive processes, and sexual performance. Young sexually functional male volunteers were presented with erotic audiotapes simultaneous with four levels of shock threat (no shock, half tolerance, tolerance, and twice tolerance; Beck, Barlow, Sakheim, & Abrahamson, 1984a). As a further check on their level of attention to the audiotapes, a sentence recognition task administered with signal-detection methodology was carried out immediately following presentation of the stimuli. Essentially, the results indicated that shock threat lowered sexual responding particularly during half tolerance and tolerance shock; however, during the twice tolerance condition sexual responding returned somewhat and approached responding under the no-shock condition. Performance on the sentence recognition task improved during half tolerance and tolerance shock and then deteriorated somewhat during the twice tolerance condition. Thus, performance on the sentence recognition task forms an inverted U-shaped function under increasing intensities of shock threat (see Figure 12.3). This finding, of course, is as old as the Yerkes-Dodson Law (1908). Sexual arousal as measured by erectile response, however, mirrors this inverted U. In other words, the better that subjects do on the sentence recognition task (presumably as a function of greater attention to the sentences), the less sexual arousal. No such sentence recognition task was included in our previous shock threat experiments where we noticed a marked increase in sexual responding (at least when shock threat was not contingent) in our sexually functional subjects. What this experiment suggests in a very preliminary way is that anxiety, defined here as physiological arousal induced by shock threat, will improve performance (or concentration) in an inverted U-shaped function. But if this performance focus is nonsexual, then sexual arousal will suffer proportionally. Thus, when one allows sexually functional subjects to focus on sexual cues while inducing anxiety, they will do better (up to a point). But because shock threat seems to interfere with sexual arousal in dysfunctional subjects even when no overt nonsexual task is available for focus, it is possible that sexually dysfunctional subjects are already focusing on or attending to something other than erotic cues. A second experiment seems to support this notion. Beck (1984) showed sexually functional subjects and matched sexually dysfunctional subjects a series of erotic films while either threatened with shock in a contingent paradigm (you may be shocked if you do not achieve at least 60% arousal) or not. In addition, they were told to focus either on internal sensations (sensate focus) or actual erectile responses (spectator focus) very closely. Essentially, when sexually functional men were threatened with shock during either sensate focus or spectator focus, their sexual arousal decreased. Presumably, once again they were doing better at attending to the task at the expense of

192  David H. Barlow Mean Strain Gauge Values Mean Sentence Recognition (d') Values

1.00

25 Strain Gauge: MMS Circumference Change From Baseline (mm)

0.75 20

0.50 0.25

15

Sentence Recognition: d´

1.25

10

5

0 No shock

1/2 Tolerence

Tolerence

2x Tolerence shock

Figure 12.3 Mean Strain Gauge Responding, Averaged Across Stimulus Duration, and Mean Sentence Recognition Values During Four Shock Threat Conditions. (From “A Cognitive Processing Account of Anxiety and Sexual Arousal: The Role of Selective Attention, Thought Content, and Affective States” by J.G. Beck, D.H. Barlow, D.K. Sakheim, and D.J. Abrahamson, 1984, paper presented at the annual convention of the American Psychological Association, Toronto, Ontario, Canada.)

their sexual arousal. Sexually dysfunctional men, on the other hand, did not show this effect and, in fact, evidenced their highest level of sexual arousal while threatened with shock during the spectator focus condition. This experiment, of course, does not have a no-distraction control, but nevertheless the results parallel data from our experiment on the effects of neutral distraction on sexually functional and sexually dysfunctional men. That is, sexually functional men do worse (evidence lower sexual arousal) when given a task that removes their focus somewhat from the direct processing of erotic cues in the film. But sexually dysfunctional men either show no effect or do slightly better when distracted from processing erotic cues. The tentative implication from these studies is that sexual arousal in sexually functional men is decreased by any number of tasks that compete with processing of erotic stimuli and performance-related sexual cues and that either sexual arousal or performance on nonsexual tasks is enhanced by increased anxiety depending on one’s cognitive focus. For sexually dysfunctional subjects, however, performance-related sexual demands decrease sexual arousal perhaps through a process of distraction (based on data from simultaneous measures of cognitive processes and

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  193

post hoc debriefing) and that threat of shock during performance demands also further decreases sexual arousal possibly by enhancing concentration on the nonerotic content of their attentional focus. Presumably, this focus centers on sexual-performance-related concerns, fears of scrutiny, failure, ridicule, and so forth. Introducing a competing nonerotic task for dysfunctional subjects does not further reduce sexual arousal and in some cases disinhibits it slightly. Subjective and Affective Differences Between Sexually Functional and Dysfunctional Subjects Several other differences emerged from these experiments comparing sexually functional and sexually dysfunctional subjects. First, a pattern of results became evident suggesting different affective responses in a sexual context between these two groups. Thus far, both Heiman and Rowland (1983) as well as Abrahamson, Barlow, Beck, et al (1985) and Beck (1984, in her dissertation) have observed generally negative affective responses in a sexual context (viewing erotic films) for sexually dysfunctional men, whereas more generally positive affective responses are evidenced by sexually functional men. These negative affective responses, best characterized as depression, seem situationally specific to a sexual context and may be a result of perceived inadequate responding or expectancies of inadequate responding in these sexually dysfunctional subjects. But an intriguing possibility is that these negative affective responses predate the dysfunction and contribute to its etiology. For example, Byrne has been investigating personality variables that are dispositional in terms of responding adequately to erotic cues. The trait that emerges falls along the dimension he termed erotophilia-erotophobia (Byrne, 1977, 1983a, 1983b). The negative affective responses of sexually dysfunctional men place them within the erotophobic end of the dimension. Negative affective responses may contribute to an avoidance of erotic cues and thereby facilitate some sort of cognitive interference produced by focusing on nonerotic cues. These different affective responses constitute a fourth difference between sexually functional and sexually dysfunctional subjects. Another intriguing phenomenon is apparent when one examines the relation of self-reported arousal to penile erection. Unlike sexually functional subjects and organogenically impotent men, psychogenically sexually dysfunctional men consistently underreport sexual arousal. That is, at the same levels of erectile response, psychogenically dysfunctional men will report far less sexual arousal than will sexually functional or organogenically dysfunctional men (Sakheim, 1984). This also seems true for dysfunctional women (Morokoff & Heiman, 1980). Finally, sexually functional and sexually dysfunctional men also seem to differ in their ability to suppress sexual arousal (assessed both physiologically and subjectively) by any cognitive means (e.g., Beck, Barlow, & Sakheim, 1982). In this preliminary report, sexually dysfunctional men perceived themselves as maintaining less control over their sexual arousal relative to sexually functional men. That is, although both sexually functional and dysfunctional men were able to suppress their arousal when asked, sexually functional men reported accurately that they were suppressing their arousal and described various cognitive means by which they accomplished this task. Sexually dysfunctional men, on the other

194  David H. Barlow

hand, were not aware that they were successfully suppressing their arousal and were unable to perceive any means by which they either were or might be able to control their sexual arousal. Thus, a fifth difference between sexually functional and dysfunctional subjects is found in differential estimates of control of sexual arousal and the tendency of sexually dysfunctional subjects to underestimate their sexual arousal. Working Model of Sexual Dysfunction Thus far then, five factors seem to differentiate sexually functional subjects from sexually dysfunctional subjects. First, sexually dysfunctional subjects consistently evidence negative affect in the sexual context, whereas sexually functional subjects display more positive affect. Second, dysfunctional subjects consistently underreport their levels of sexual arousal and generally evidence diminished perceptions of control over their arousal. Third, dysfunctional men are not distracted by non-sexual-performance-related stimuli in that they evidence no decrease in erectile response, whereas sexually functional subjects are distracted and show decreases in sexual response. Fourth, dysfunctional men are distracted by performance-related sexual stimuli, whereas the sexual arousal of sexually functional men is enhanced. Finally, anxiety inhibits sexual arousal in dysfunctional subjects but facilitates it in sexually functional subjects. These preliminary findings, taken together, suggest to us a working model of psychogenic sexual dysfunction. In this model, we suggest that a cognitive interference process interacting with anxiety is responsible for sexual dysfunction, specifically, inhibited sexual excitement in men and women and possibly other related forms of sexual dysfunction. The nature of this cognitive process in dysfunctional subjects seems to revolve largely around focusing on or attending to a task-irrelevant context. This focus then becomes driven by the physiological aspects of arousal that clinicians more commonly refer to as anxiety which, in turn, results in further deterioration in sexual performance. Sexually functional subjects, on the other hand, focus on and process erotic cues without difficulty. Anxiety may enhance this processing (up to a point) and therefore may facilitate sexual arousal in sexually functional subjects. These feedback loops for sexually functional and sexually dysfunctional subjects are depicted schematically in Figure 12.4. For dysfunctional subjects, this inappropriate attentional focus seems to involve attending to the consequence of not performing or some other issue not directly related to the erotic cues. For this reason, neutral distracting stimuli not involving performance-related cues have no inhibiting effect on the sexual arousal of dysfunctional subjects (unlike sexually functional subjects). Dysfunctional subjects are already distracted. The development of these alternative and competing task-irrelevant thoughts seem rooted in one’s learning history, perhaps in the developed trait of erotophobia previously described, but the reasons why some people become dysfunctional with this process, whereas others with similar learning histories may not, is not yet clear. The similarities of this hypothetical psychopathological process to other performance anxieties subsumed under the heading of social

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  195 FUNCTIONALS (Positive Feedback Loop)

DYSFUNCTIONALS (Negative Feedback Loop) Explicit or implicit demands for sexual performance (e.g., a responsive partner or other contexts) leading to public expectation of performance (erection)

Positive affect and expectancies, accurate reporting of erection, perception of control

APPROACH

Attentional focus on erotic cues Increased autonomic arousal

AVOIDANCE

Negative affect and expectancies, inaccurate and underreporting of erection, perceived lack of control

Attentional focus on public consequences of not performing or other nonerotic issues Increased autonomic arousal

Increasingly efficient attentional focus on erotic cues

Increasingly efficient attentional focus on consequences of not performing (etc.)

Functional performance

Dysfunctional performance

Figure 12.4 Model of Erectile Dysfunction.

phobia, where similar competing cognitions seem responsible for performance deficits, has not escaped us (Beck & Barlow, 1984). It need not be noted that this model has only preliminary support from data collected thus far. Many conclusions depend on contrasting data sets from different experiments. Some aspects of the model have not yet been tested on clinical subjects, and appropriate comparisons between clinical subjects and volunteer normal subjects are therefore not yet available. Furthermore, most of the data that do exist were collected from men, although experiments with women as subjects provide data that are consistent with this model in each instance (Cerny, 1978; Hoon et al., 1977; Morokoff & Heiman, 1980). But, this model has important assessment and treatment implications if further support is forthcoming. For example, detailed assessment of affective and cognitive functioning in a sexual context may predict degree of improvement as well as long-term outcome. Aspects of treatments successful for clinical phobia and social phobia in particular may become increasingly important (Barlow & Beck, 1984). These treatments may focus increasingly on cognitive change and allocation of attention, although performance-based exercises may still be the most effective way to accomplish this. Emphasizing anxiety reduction operationalized as decreasing physiological arousal may be counterproductive in view of the effects of anxiety on sexually functional subjects. The methodologies now exist to address these questions. Note Preparation of this article and much of the research reported herein was supported by National Institute of Mental Health Grant MH33553.

196  David H. Barlow

References Abel, G. G., Barlow, D. H., Blanchard, E. B., & Mavissakalian, M. (1975). Measurement of sexual arousal in male homosexuals: Effects of instructions and stimulus modality. Archives of Sexual Behavior, 4, 623–630. Abel, G. G., Blanchard, E. B., & Barlow, D. H. (1981). The effects of stimulus modality, instructional set and stimulus content on the objective measurement of sexual arousal in several paraphilias. Behaviour Research and Therapy, 19, 25–34. Abrahamson, D. J., Barlow, D. H., Beck, J, G., Sakheim, D, K., & Kelly, J. P. (1985). The effects of attentional focus and partner responsiveness on sexual responding: Replication and extension. Archives of Sexual Behavior, 14, 361–371. Abrahamson, D. J., Barlow, D. H., Sakheim, D. K., Beck, J. G., & Athanasiou, R. (1985). Effects of distraction on sexual responding in sexual functional and sexually dysfunctional men. Behavior Therapy, 16, 503–515. American Psychiatric Association. (1980). Diagnostic and Statistical manual of mental disorders (3rd ed.). Washington, DC: Author. Bancroft, J. (1970). Disorders of sexual potency. In O. Hill (Ed.), Modern trends in psychosomatic medicine (pp. 246–259). New York: Appleton-Century-Crofts. Barfield, R., & Sachs, B. (1968). Sexual behavior: Stimulation by painful electric shock to skin in male rats. Science, 161, 392–395. Barlow, D. H., & Beck, J. G. (1984). The psychosocial treatment of anxiety disorders: Current status, future directions. In J. B. W. Williams & R. L. Spitzer (Eds.), Psychotherapy research: Where are we and where should we go? (pp. 29–66). New York: Guilford Press. Barlow, D. H., & Mavissakalian, M. (Eds.). (1981). Directions in the assessment and treatment of phobia: The next decade. In Phobia: Psychological and pharmacological treatment (pp. 199–245). New York: Guilford Press. Barlow, D. H., Mavissakalian, M., & Schofield, L. (1980). Patterns of desynchrony in agoraphobia. Behaviour Research and Therapy, 18, 441–448. Barlow, D. H., Sakheim, D. K., & Beck, J. G. (1983). Anxiety increases sexual arousal. Journal of Abnormal Psychology, 92, 49–54. Beck, J. G. (1984). The effect of performance demand and attentional focus on sexual responding in functional and dysfunctional men. Unpublished doctoral dissertation, State University of New York at Albany. Beck, J. G., & Barlow, D. H. (1984). Current conceptualizations of sexual dysfunction: A review and an alternative perspective. Clinical Psychology Review, 4, 363–378. Beck, J. G., Barlow, D. H., & Sakheim, D. K. (1982, August). Sexual arousal and suppression patterns in functional and dysfunctional men. Paper presented at the annual convention of the American Psychological Association, Washington, DC. Beck, J. G., Barlow, D. H., & Sakheim, D. K. (1983). The effects of attentional focus and partner arousal on sexual responding in functional and dysfunctional men. Behaviour Research and Therapy, 21, 1–8. Beck, J. G., Barlow, D. H., Sakheim, D. K., & Abrahamson, D. J. (1984a, August). A cognitive processing account of anxiety and sexual arousal: The role of selective attention, thought content, and affective states. Paper presented at the annual Convention of the American Psychological Association, Toronto, Ontario, Canada. Beck, J. G., Barlow, D. H., Sakheim, D. K., & Abrahamson, D. J. (1984b, November). Sexual responding during anxiety: Clinical versus non-clinical patterns. Paper presented at the 18th annual convention of the Association for the Advancement of Behavior Therapy, Philadelphia. Bennett, A. H. (1982). Management of male impotence. Baltimore, MD: Williams & Wilkins. Berscheid, E., & Walster, E. (1974). A little bit about love. In T. Huston (Ed.), Foundations of interpersonal attraction (pp. 355–381). New York: Academic Press.

Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference  197 Beutler, L., & Gleason, D. (1981). Integrating advances in the diagnosis and treatment of male potency disorders. Journal of Urology, 126, 338–342. Brady, J. (1966). Brevital relaxation treatment of frigidity. Behaviour Research and Therapy, 4, 71–77. Brehm, J., Gatz, M, Goethals, G., McCrimmon, W., & Ward, L. (1978). Physiological arousal and interpersonal attraction. JSAS Catalog of Selected Documents in Psychology, 8, 63. Briddell, D. W., Rimm, D. C., Caddy, G. R., Krawitz, G., Sholis, D., & Wunderlin, R. J. (1978). The effects of alcohol and cognitive set on sexual arousal to deviant stimuli. Journal of Abnormal Psychology, 87, 418–430. Byrne, D. (1977). Social psychology and the study of sexual behavior. Journal of Personality and Social Psychology, 3, 3–30. Byrne, D. (1983a). The antecedents, correlates, and consequents of erotophobia-erotophilia. In C. M. Davis (Ed.), Challenges in sexual science (pp. 53–75). Philadelphia: Society for the Scientific Study of Sex. Byrne, D. (1983b). Sex without contraception. In D. Byrne & W. A. Fisher (Eds.), Adolescents, sex, and contraception (pp. 2–31). Hillsdale, NJ: Erlbaum. Cerny, J. A. (1978). Biofeedback and the voluntary control of sexual arousal in women. Behavior Therapy, 9, 847–855. Cooper, A. J. (1981). Short-term treatment in sexual dysfunction: A review. Comprehensive Psychiatry, 22, 206–217. Crown, S., & D’Ardenne, . (1982). Symposium on sexual dysfunction: Controversies, methods, results. British Journal of Psychiatry, 140, 70–77. Dutton, D. G., & Aron, A. P. (1974). Some evidence for heightened sexual attraction under conditions of high anxiety. Journal of Personality and Social Psychology, 30, 510–517. Farkas, G. M., Sine, L. F., & Evans, I. M. (1979). The effects of distraction, performance demand, stimulus explicitness and personality on objective and subjective measures of male sexual arousal. Behaviour Research and Therapy, 17, 25–32. Fenichel, O. (1945). The psychoanalytic theory of neurosis. New York: W. W. Norton. Fisher, C., Schiavi, R. C., Edwards, A., Davis, D. M., Reitman, M., & Fine, J. (1979). Evaluation of nocturnal penile tumescence in the differential diagnosis of sexual impotence. Archives of General Psychiatry, 36, 431–437. Geer, J., & Fuhr, R. (1976). Cognitive factors in sexual arousal: The role of distraction. Journal of Consulting and Clinical Psychology, 44, 238–243. Hartmann, W., & Fithian, M. (1972), Treatment of sexual dysfunction: A bio-psychosocial approach. Long Beach, CA: Center for Marital and Sexual Studies. Heiman, J. R., & Rowland, D. L. (1983). Affective and physiological sexual patterns: The effects of instructions on sexually functional and dysfunctional men. Journal of Psychosomatic Research, 27, 105–116. Henson, D. E., & Rubin, H. B. (1971). Voluntary control of eroticism. Journal of Applied Behavior Analysis, 4, 37–44. Hoon, P., Wincze, J., & Hoon, E. (1977). A test of reciprocal inhibition: Are anxiety and sexual arousal in women mutually inhibitory? Journal of Abnormal Psychology, 86, 65–74. Kaplan, H. S. (1974). The new sex therapy. New York: Brunner/Mazel. Kaplan, H. S. (1981). The new sex therapy: Active treatment of sexual dysfunctions. New York: Brunner/Mazel. Karacan, I. (1970, April). Clinical value of nocturnal erection in the prognosis and diagnosis of impotence. Medical Aspects of Human Sexuality, 4, 27–34. Kuriansky, J. B., & Sharpe, L. (1981). Clinical and research implications of the evaluation of women’s group therapy for anorgasmia: A review. Journal of Sex and Marital Therapy, 7, 268–277.

198  David H. Barlow Lang, P. J. (1968). Fear reduction and fear behavior: Problems in treating a construct. In J. M. Shlien (Ed.), Research in Psychotherapy (Vol. 3, pp. 90–102). Washington, DC: American Psychological Association. Lansky, D., & Wilson, G. T. (1981). Alcohol, expectations, and sexual arousal in males: An information processing analysis. Journal of Abnormal Psychology, 90, 35–45. Laws, D., & Rubin, H. (1969), Instructional control of autonomic sexual response. Journal of Applied Behavior Analysis, 2, 93–100. Lazarus, A. (1963). The treatment of chronic frigidity by systematic desensitization. Journal of Nervous and Mental Diseases, 136, 272–278. Levine, S. B., & Agle, D. (1978). The effectiveness of sex therapy for chronic secondary psychological impotence. Journal of Sex and Marital Therapy, 4, 235–258. Marks, I. M. (1981). Review of behavioral psychotherapy, II: Sexual disorders. American Journal of Psychiatry, 138, 750–756. Masters, W., & Johnson, V. (1970). Human sexual inadequacy. Boston: Little, Brown. Mavissakalian, M., R. & Barlow, D. H. (Eds.). (1981). Phobia: Psychological and pharmacological treatment. New York: Guilford Press. Mills, K. H., & Kilmann, P. R. (1982). Group treatment of sexual dysfunctions: A methodological review of the outcome literature. Journal of Sex and Marital Therapy, 8, 259–296. Morokoff, P. J., & Heiman, J. R. (1980). Effects of erotic stimuli on sexually functional and dysfunctional women: Multiple measures before and after therapy. Behaviour Research and Therapy, 18, 127–137. Norton, G. R., & Jehu, D. (1984). The role of anxiety in sexual dysfunctions: A review. Archives of Sexual Behavior, 2, 165–183. Ramsey, G. (1943). The sexual development of boys. American Journal of Psychology, 56, 217. Redmond, D. E., Kosten, T. B., & Reiser, M. F (1982). Spontaneous ejaculation associated with anxiety: Psychophysiologic considerations. American Journal of Psychiatry, 140, 1163–1166. Riordan, C. (1979). Interpersonal attraction in aversive situations. Unpublished doctoral dissertation, State University of New York at Albany. Sakheim, D. K. (1984). Waking assessment of erectile potential: The validation of a laboratory procedure to aid in the differential diagnosis of psychogenic and organic impotence. Unpublished doctoral dissertation, State University of New York at Albany. Sakheim, D. K., Barlow, D. H., Beck, J. G., & Abrahamson, D. J. (1984). The effect of an increased awareness of erectile cues on sexual arousal. Behaviour Research and Therapy, 22, 151–158. Sarrel, D. M. & Masters, W. H. (1982), Sexual molestation of men by women. Archives of Sexual Behavior, 11, 117–131. Schiavi, R. (1976). Sex therapy and psychophysiological research. American Journal of Psychiatry, 133, 562–566. Schiavi, R. C. (1981). Male erectile disorders. Annual Review of Medicine, 32, 509–520. Stoller, R. (1976). Sexual excitement. Archives of General Psychiatry, 33, 899–909. Wolchik, S. A., Beggs, V., Wincze, J. P., Sakheim, D. K., Barlow, D. H., & Mavissakalian, M. (1980). The effects of emotional arousal on subsequent sexual arousal in men. Journal of Abnormal Psychology, 89, 595–598. Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford, CA: Stanford University Press. Wolpe, J. (1978). Comments on “A test of reciprocal inhibition” by Hoon, Wincze, & Hoon. Journal of Abnormal Psychology, 87, 452–454. Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to rapidity of habit-formation. Journal of Comparative Neurology and Psychology, 18, 459–482. Zilbergeld, B., & Evans, M. (1980, August). The inadequacy of Masters and Johnson. Psychology Today, 14, 29–43. Zillman, D. (1983). Transfer of excitation in emotional behavior. In J. T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A sourcebook (pp. 215–240). New York: Guilford Press.

Index

AABT see Association for the Advancement of Behavior Therapy ABCT see Association for Behavioral and Cognitive Therapy Abel, G. G. 188 Abrahamson, D. J. 190, 193 action tendencies 82, 94, 97 – 100, 103, 149, 152; emotion-related 20; modifying emotional 156 – 9 ADIS see Anxiety Disorders Interview Schedule affective therapy 20, 21, 95 – 7, 104, 149 Agle, D. 185 agoraphobia 6, 13, 21, 35 – 6, 58, 86, 88, 171; classifications 62; context 63; exposurebased treatments 59, 68; findings 53 – 5; in vivo exposure 77; mild 127; panic disorder with 142, 170, 177, 178, 179; procedure 52 – 3; social reinforcement 51 – 7; symptoms 69, 123; systematic desensitization 142; treatment 123 – 4 Agras, S.: “Social Reinforcement in the Modification of Agoraphobia” 51 – 7; University of Mississippi Medical Center 7, 8, 9, 10, 122; University of Vermont 5, 122 AHCPR see Agency for Health Care Policy and Research alarm theory 80, 97, 100, 103; see also false alarms; learned alarms Allen, L. B.: “Toward a Unified Treatment for Emotional Disorders” 141 – 65 Allport, G. 44 Allprazolam (Xanax) 20, 105 – 21 American Psychological Association (APA) 16, 417, 434; CHAMPUS project 39 – 40 American Psychological Association, Task Force on Psychological Intervention Guidelines: 15, 23

amygdala 154 antecedent cognitive reappraisal 141, 152, 153 – 4, 157 anxiety disorders: assessment and classification 61 – 3; behavioral approaches 58 – 67; delivery of treatment and dissemination of results 65; dissenting positions 66 – 7; double 213 – 16; exposure treatments 66 – 7; general research strategies 59 – 61; outcome research 64 – 5; pathological 96, 101; physiological data 66; process research 63 – 4; sexual dysfunction 184 – 7, 191 – 3; state of the art 59; straightforward outcome studies 66; see also generalized anxiety disorder Anxiety Disorders Interview Schedule (ADIS) 14, 114, 116, 145, 171; Lifetime Version (ADIS-IV-L) 145; Revised (ADIS-R) 107, 108, 111, 114, 118, 127 anxiety reduction 30, 75, 76, 77, 78 – 9, 80, 81, 82, 83 – 4, 85, 87, 93, 94, 95, 96, 97, 100, 102, 195; during emotional processing 88 – 90; process of fear 103 – 4; see also fear and anxiety reduction anxiety research clinic 122 – 4 arousal 40, 80, 82, 84, 85, 90, 91 – 2, 93, 94, 95, 96, 99, 101 – 2, 103; adequate 186; autonomic 147, 151, 185, 187; emotional 191; generalized 72, 98, 156; heightened 149; high-intensity 77; homosexual 188; hyper 98; physiological 77, 78, 83, 86, 97, 100, 152, 185; PTSD 155; sexual 14, 21, 22, 183, 184, 185 – 7, 188 – 90 Association for the Advancement of Behavior Therapy (AABT) 4, 18, 32, 33, 35, 38, 39, 159, 160 Association for Behavioral and Cognitive Therapy (ABCT) 4, 18

i2  Index aversiveness 89 avoidance behavior 69, 72, 79, 83, 86, 87, 158 Azrin, N. H. 35, 37 Bancroft, J. 185 Bandura, A. 8, 33, 76; emotional change 102; self-efficacy 84, 85 – 6 Barfield, R. 185 Barlow, D. H. 120, 154; Albany years 12 – 16; Boston: early years 2 – 4; Boston: later years 16 – 18; Brown years 10 – 12; Center for Stress and Anxiety Disorders 123; exposure-based treatments 69; Mississippi years 7 – 10; Vermont years 5–7 Barlow, D. H., works of 101, 155, 188; “Behavioral Approaches to Anxiety Disorders: A Report on the NIMHSUNY, Albany, Research Conference” 58 – 67; “Behavioral Conception and Treatment of Panic” 68 – 74; “Behavior Therapy: The Next Decade” 31 – 43; “Causes of Sexual Dysfunction: The Role of Anxiety and Cognitive Interference” 183 – 95; “A Comparison of Alprazolam and Behavior Therapy in Treatment of Panic Disorder” 105 – 21; “CognitiveBehavioral Therapy, Imipramine, or Their Combination for Panic Disorder: A Randomized Controlled Trial” 125 – 40; “The Development of an Anxiety Research Clinic” 122 – 4; “The Phenomenon of Panic” 169 – 82; “The Process of Fear and Anxiety Reduction: Affective Therapy” 75 – 104; “Social Reinforcement in the Modification of Agoraphobia” 51 – 7; “Toward a Unified Treatment for Emotional Disorders” 141 – 65; “Why Can’t We Be More Idiographic in Our Research? 44 – 8 Beck, J. G. 96, 98, 153, 156, 186, 190, 191, 193 behavior therapy: next decade 31 – 43; technique building 18, 34 – 5, 37, 39, 40 benzodiazepine 120, 127, 130, 131, 137, 204; high-potency 20; and beta blockers 111; hypersensitivity 107; plasma screens 108, 109, 113, 116; urine screens 129 Beutler, L. E. 184 bioinformational theory 87, 95, 96 BIS see behavioral inhibition system Blanchard, E. B. 8, 12, 59; Center for Stress and Anxiety Disorders 123; homosexual arousal patterns 188; irritable bowel

syndrome 101; “The Phenomenon of Panic” 169 – 82 Borkovec, T. D. 84 Bouton, M. E. 150 – 1 Briddell, D. W. 188 Brownell, K. D. 11 Byrne, D. 193 CALM see Coordinated Anxiety Learning and Management calming procedures 155 Cambridge Somerville Youth Study 46 Cautela, J. R. 2 – 4, 5, 6, 122 categorical classification 17, 23 Cerny, J. A. 188; “A Comparison of Alprazolam and Behavior Therapy in Treatment of Panic Disorder” 105 – 21; “The Phenomenon of Panic” 169 – 82 CGI see Clinical Global Impression Scale Chambless, D. L. 69, 153 Choate, M. L.: “Toward a Unified Treatment for Emotional Disorders” 141 – 65 Clark, D. M. 72 Clark, L. A. 147 Clinical Global Impression Scale (CGI) 125 – 6, 129, 132 – 4, 136 cognitive-behavioral therapy (CBT) 107, 108, 118 cognitive-behavioral therapy, imipramine, or their combination for panic disorder 125 – 40; assessment 129; baseline, site, and stratum analyses 130; methods 126 – 30; quality of response 132; results 130 – 5; statistical analyses 129 – 30; study design 126 – 7; subject disposition, dosing, and laboratory analyses 130 – 2; subjects 127; therapists 127 – 8; timing of loss of response 132 – 5; treatment conditions 128 – 9 cognitive interference: sexual dysfunction 22, 183 – 95 combat 85 comorbidity 144, 145 – 6 control 99 – 101; lack of 97, 99, 100, 195; see also Anxiety Control Questionnaire; panic control treatment; uncontrollability Cottam, G. L. 156 courage 83, 92 Craske, M. G. 14, 15, 78, 85, 91 – 2 CS see conditioned stimuli Darwin, C. R. 95, 149 Davison, J. 7

Index  i3 DCS see D-cycloserine depression 22, 46, 53, 81, 107, 123, 158, 178, 195; comorbidity 146; comorbid unipolar 127; control 99; emotional avoidance 154; exposure-based procedures 143; major 127, 169, 172, 179, 180; sexually dysfunctional 193; stratum effect 130; trait 245; treatment 136, 144, 153, 156; unipolar 127; versus anxiety 82, 147, 148; see also Hamilton Rating Scale for Depression de Silva, P. 76 desynchrony 63, 64, 65, 85, 87 Diagnostic and Statistical Manual of Mental Disorders (DSM) 302; II 6; III 13, 14, 68 – 9, 107, 169, 170; III 104, 169, 170, 171, 172, 175, 177, 178, 183; III-R 14, 22, 108, 111; IV 14, 17, 104, 123, 145, 147, 149 diagnostic reliability 147 differential attention 35 Di Nardo, P. A.: “The Phenomenon of Panic” 169 – 82 discordance 85, 92 – 3, 100 DSM see Diagnostic and Statistical Manual of Mental Disorders Duncan’s Multiple Range Test 114, 115, 118 Ehrenfeld, D. 32 Eifert, G. H. 155 Eisler, R. 122 Emery, G. 98, 156 emotional alarm 96 emotional avoidance 154 – 6 emotional disorders 141 – 65; action tendencies 156 – 7; antecedent cognitive reappraisal 153 – 4; applications 157 – 9; current status of treatment 143 – 4; effects of psychological treatments on comorbid disorders, and nonspecificity of treatment response 146; emotional avoidance 154 – 6; emotion regulation 151 – 2; etiology 147 – 9; latent temperamental structure 147, 357 – 61; modern learning theory and cognitive neuroscience 150 – 9; nature 144 – 9, 361 – 5; overlap among disorders 145 – 6; treatment implications 149 – 50; unified treatment 152; see also anxiety disorders; depression; fear; fear and anxiety reduction; negative affect emotional processing 87–95; difficulties with the theory 92–5; discordance 92–3; evidence supporting 90–2; gradual decrements in physiological responding

93–4; identification of fear structures 94; process of anxiety reduction during 88–90 emotion regulation 151 – 2 erectile dysfunction 183 – 4, 185 Erickson, M. 33 Evans, I. M. 188 exhibitionism 40, 197 experiential avoidance 152 exposure-based treatments 60, 63, 64, 65, 68 – 70, 76, 81, 82, 83, 85, 88 – 9, 97, 100, 102, 137, 142, 143, 155, 158, 170; agoraphobia 59 extinction 36, 55, 68, 72, 76, 78 – 80, 82, 83 – 4, 98, 100, 142 false alarms 100, 103, 148 – 9 Farkas, G. M. 188 fear 12, 19, 52, 60, 70, 72, 108, 143, 151, 154, 155, 156 – 7, 158, 169, 171, 172, 177, 178; failure 185, 193; inadequacy 185; performance 185; snake 56, 142; see also agoraphobia fear and anxiety reduction 3, 6, 20, 35, 75 – 104, 142; action tendencies 97 – 9; affective therapy 95 – 7; control 99 – 101; emotional processing 87 – 95; essential targets for change 97 – 102; extinction 78 – 80, 83 – 4; habituation 77 – 8, 83 – 4; helpful targets for change 102 – 3; lack of control 97, 99, 100; non-exposure-based treatments 94 – 5; process of 103 – 4; psychophysiological methods 95; selfefficacy 84 – 7; self-focused attention 101 – 2; theories 76; toughening up 80 – 4 fear of fear 153 Feldner, M. T. 155 Fenichel, O. 184 Feuerstein, M. 32 – 3 fight-or-flight behavior 98, 151 Fisher, A. 37, 184 Fisher exact tests 111, 113, 130 flooding 87 – 8 Foa, E. B. 87, 88, 89, 90, 91, 95 Ford, J. D. 38 Fowler, J. C. 156 Franks, C. 33 Freeman-Halton test 130 Fridlund, A. J. 98, 156 Fuhr, R. 188 Gallagher, R. 69 galvanic skin response (GSR) 77, 78 Geer, J. 188 generalized anxiety disorder 13, 14, 15, 21, 59, 60, 62, 63, 86, 120, 170, 186;

i4  Index autonomic arousal 147, 151, 185, 187; comorbidity 145 general neurotic syndrome 146 Glazer, H. I. 80 Gleason, D. 184 Goldstein, A. J. 69, 153 Gorman, J. M.: “Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder: A Randomized Controlled Trial” 125 – 40 Gray, J. A. 81, 82 Grayson 92 Graywind Publications 15, 23 Greenberg, R. L. 75, 96 Gross, J. J. 153 GSR see galvanic skin response habituation 77 – 8, 83 – 4, 87, 89, 90, 92, 93, 94, 95, 98, 142 Hamilton Anxiety Rating Scale 108, 111, 115, 120 Hamilton Rating Scale for Depression 108, 111 Hariri, A. R. 154 Hatfield, M. E. 156 Heiman, J. R. 190, 193 heart rate, high 78, 85, 91, 94 helplessness 80, 96, 99 Henson, D. E. 188 Herbert, E. W. 35 Hersen, M. 8, 9, 37 Holt, C. S. 81 Hoon, E. 186 Hoon, P. 186 idiographic approaches 9 – 10, 44 – 7 imipramine 20, 125 – 9, 130, 131 – 6, 137 intent-to-treat (ITT) 125, 129, 130, 132, 136, 137 interoceptive conditioning 71, 72 Izard, C. E. 98, 156 Jansson, L. 68 Johnson, V. 22, 40 – 1, 183, 184, 185, 188, 198 Johnston, D. W. 79 Kaplan, H. S. 185, 188 Kazdin, A. E. 32 Kelly, J. P. 190 Kendall, P. C. 38 Klein, D. 68 Klosko, J. S. 20; “A Comparison of Alprazolam and Behavior Therapy in Treatment of Panic Disorder” 105 – 21

Kosten, T. B. 185 Kozak, M. J. 87, 88, 89, 90, 91, 95 lack of control 97, 99, 100, 199 Lang, P. J. 36, 87 – 8, 90, 91, 93, 94, 95, 96, 98, 99, 149 Laws, D. 188 Lazarus, A. A. 32 learned alarms 100 “learning-theory” based approaches 3 Leitenberg, H. 93, 122; “Social Reinforcement in the Modification of Agoraphobia” 51 – 7 Levine, S. B. 184 Levitt, J. T. 83 – 4, 85, 100, 155 Linehan, M. M. 149, 157 Litz, B. T. 155 London, P. 32 MacDonald, M. L. 38 Maletzky, B. M. 40 Marks, I. M. 80, 142 Masters, J. C. 153 Masters, W. 22, 40, 183, 184, 185, 188, 198 Mavissakalian, M. 188 McHugh, R. K. 24 meaning propositions 87 – 8, 99 mechanisms of change 20 Meehl, P. E. 37 Miller, D. G. 11 Miller, P.M. 8, 122 Mineka, S. 22 Mozer, M. H. 38 – 9 National Alliance for Research in Schizophrenia and Depressive Disorders (NARSAD) 123 National Institute of Mental Health (NIMH) 10, 13, 14, 16, 18, 19, 20, 21, 33, 46, 123; -SUNY, Albany conference 58 – 67; task force 144 National Institutes of Health (NIH) 46; conference 20; grants 13 negative affect 1, 6, 89, 96, 97, 98, 102, 145, 146, 147, 148, 157, 159, 193, 194 NIH see National Institutes of Health NIMH see National Institute of Mental Health Nock, M. K. 19; “Why Can’t We Be More Idiographic in Our Research? 44 – 8 nocturnal penile tumescence 184 nomothetic approaches 9, 18, 45 – 7 non-exposure-based treatments 94 – 5 noradrenaline levels 80 – 1 Norton, G. R. 70 nosology 7, 14, 144, 147

Index  i5 obsessive-compulsive disorders (OCD) 58, 59, 83, 177, 178 – 9; antidepressant medications 146; cognitive and behavioral rituals 157; D-cycloserine 450; unified treatment 143, 144, 149, 153, 157 OCD see obsessive-compulsive disorders omnibus testing 130, 131 Orsillo, S. M. 155 panic attacks: spontaneous 22, 108, 113, 118, 170, 172, 177 – 8; unexpected 70, 84, 126 panic control treatment (PCT) 20, 105, 109, 110, 111, 113, 114, 115, 116, 118, 119 panic disorder 200 – 2: with agoraphobia 142, 143, 145, 146, 147, 153, 155, 157, 170, 177, 178, 179; behavioral conception and treatment 68 – 74 panic disorder, comparison of alprazolam and behavior therapy in treatment of 105 – 21; alprazolam and behavior therapy measurement 118 – 19; alprazolam dosage 114; attrition 110 – 11; measures 108 – 9; measures of panic attacks 118; method 107 – 10; posttreatment assessment 113 – 19; posttreatment clinical assessment measures 114 – 15; posttreatment medication use measures 115 – 18; pretreatment characteristics of treatment groups 111 – 13; results 110 – 13; selfmonitoring measures 118 – 19; subjects 107; treatment 109 – 10 panic disorder, cognitive-behavioral therapy, imipramine, or their combination for 125 – 40; assessment 129; baseline, site, and stratum analyses 130; methods 126 – 30; quality of response 132; results 130 – 5; statistical analyses 129 – 30; study design 126 – 7; subject disposition, dosing, and laboratory analyses 130 – 2; subjects 127; therapists 127 – 8; timing of loss of response 132 – 5; treatment conditions 128 – 9 Panic Disorder Severity Scale (PDSS) 125 – 6, 129 – 36 panic phenomenon 169 – 82; characterization of the sample and presence 172 – 8; method 171 – 2; procedures 171 – 2; results 172 – 8; subjects 171 pathological anxiety 96, 101

PDA see panic disorder: with agoraphobia PCT see panic control treatment PDSS see Panic Disorder Severity Scale performance anxiety 85, 91, 183, 186 Phobia and Anxiety Disorders Clinic 12, 16, 107 Pohorecky, L. A. 80 posttraumatic stress disorder (PTSD) 145, 146, 155, 159 President’s Commission on Mental Health 18, 33 PTSD see posttraumatic stress disorder Rachman, J. 14, 91 – 2, 94; affective therapy 149; affect modification 96; anxiety reduction 88, 89 – 90; claustrophobia 85; emotional processing 87, 89 – 90, 94; exposure-based treatments 76, 92; false or learned alarms 100; habituation 77 – 8; high arousal 91 – 2; non-exposurebased treatments 94; self-efficacy 85; unexpected panic 83 – 4 Ramsey, G. 185 randomized clinical trials (RCTs) 19, 46 Rapee, R. M. 14, 100 – 1, 102 Razran, G. 153 RCTs see randomized clinical trials Redmond, D. E. 185 Reiser, M. F. 185 response propositions 87 – 8, 91, 94 Reyna, L. 3 Robinson, J. L. 91, 359 Roemer, L. 155 Rogers, C. 38 Rosenthal, M. 76 Rowland, D. L. 190, 193 Rubin, H. B. 188 Rush, A. J. 98, 156 Sachs, B. 185 Sajwaj, T. E. 35 Sakheim, K. 186 same-sex arousal patterns 7 Sanderson, W. C. 101 Sarrel, D. M. 185 Schedule for Affective Disorders and Schizophrenia (SADS) 14 Schiavi, R. C. 184, 185 schizophrenia 51; see also National Alliance for Research in Schizophrenia and Depressive Disorders (NARSAD); Schedule for Affective Disorders and Schizophrenia (SADS) Schmidt, N. B. 155 SD see systematic desensitization

i6  Index self-efficacy 36, 68, 79, 84 – 7, 93, 94, 142 self-focused attention 101 – 2 Seligman, M. E. P. 79 sexual arousal 14, 21, 22, 183, 184, 185 – 7, 188 – 90 sexual dysfunction 10, 14, 21, 22, 123, 183 – 98; anxiety 184 – 7; anxious arousal 22; cognitive interference 187 – 91; interaction of anxiety and cognitive interference 191 – 3; subjective and affective differences between sexually functional and dysfunctional subjects 193 – 4; working model 194 – 5 Shapiro, M. B. 45 Shaw, B. F. 98, 156 Shear, M. K.: “Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder: A Randomized Controlled Trial” 125 – 40 Sidman, M. 45 – 6 Sine, L. F. 188 Skinner, B. F. 5, 32, 34, 45 Spira, A. P. 155 spontaneous attacks 22, 108, 113, 118, 170, 172, 177 – 8 State University of New York at Albany, National Institute of Mental Health Conference 58 – 67 Steinman, W. 35 stimulus propositions 87, 88, 91, 94 Stoller, R. 185 Stress Disorders Clinic 12 Strosahl, K. D. 153 Strupp, H. H. 37 Swann, G. E. 38 systematic desensitization (SD) 3, 4, 19, 56, 77, 87 – 8, 142, 185 Tassinari, R.: “A Comparison of Alprazolam and Behavior Therapy in Treatment of Panic Disorder” 105 – 21 Teasdale, J. D. 84 test probes 88 timing of loss of response 132 – 5 Tolman, E. 44

toughening up 80 – 4 Turkat, I. D. 32 – 3 Upjohn Pharmaceutical Company 20, 107, 109, 119 uncontrollability 22, 96, 99, 100, 148 unexpected panic attacks 70, 84, 126 unipolar mood disorders 145 Vermilyea, B.: “The Phenomenon of Panic” 169 – 82 Vermilyea, J. A. 91, 92; “The Phenomenon of Panic” 169 – 82 Vermont Clinical Research Center 52 veterans 85, 155 vigilance 98; chronic 99 Wagner, A. W. 155 Watson, D. 80, 147 Weiss, J. M. 80 Williams, J. B. W. 84 – 5 Wilson, G. T. 33, 34, 85, 96, 149 Wilson, K. G. 153 Wincze, J. P. 186 Wolfe, B. E.: “Behavioral Approaches to Anxiety Disorders: A Report on the NIMH-SUNY, Albany, Research Conference” 58 – 67 Wolpe, J. 4, 5, 35 – 6, 68, 122; affective therapy 75; anxiety-based disorders 40, 185; conditioning theories 268; exposure-based treatments 76; habituation 77; Psychotherapy by Reciprocal Inhibition 3; sexual dysfunction 184 – 5, 186; systematic desensitization 19, 56, 77, 142, 185 Woods, S. W.: “Cognitive-Behavioral Therapy, Imipramine, or Their Combination for Panic Disorder: A Randomized Controlled Trial” 125 – 40 Woolfolk, A. E. 33 Woolfolk, R. L. 33 Zajonc, R. B. 96 Zvolensky, M. J. 155