Dimensional Models of Personality Disorders: Refining the Research Agenda for DSM-V

DIMENSIONAL MODELS OF PERSONALITY DISORDERS Refining the Research Agenda for DSM-V

This page intentionally left blank

DIMENSIONAL MODELS OF PERSONALITY DISORDERS Refining the Research Agenda for DSM-V Edited by

Thomas A. Widiger, Ph.D. Erik Simonsen, M.D. Paul J. Sirovatka, M.S. Darrel A. Regier, M.D., M.P.H.

Published by the American Psychiatric Association Washington, D.C.

Note: The authors have worked to ensure that all information in this book is accurate at the time of publication and consistent with general psychiatric and medical standards, and that information concerning drug dosages, schedules, and routes of administration is accurate at the time of publication and consistent with standards set by the U.S. Food and Drug Administration and the general medical community. As medical research and practice continue to advance, however, therapeutic standards may change. Moreover, specific situations may require a specific therapeutic response not included in this book. For these reasons and because human and mechanical errors sometimes occur, we recommend that readers follow the advice of physicians directly involved in their care or the care of a member of their family. The findings, opinions, and conclusions of this report do not necessarily represent the views of the officers, trustees, or all members of the American Psychiatric Association. The views expressed are those of the authors of the individual chapters.

Copyright © 2006 American Psychiatric Association ALL RIGHTS RESERVED Manufactured in the United States of America on acid-free paper 10 09 08 07 06 5 4 3 2 1 First Edition Typeset in Adobe’s Frutiger and AGaramond. American Psychiatric Association 1000 Wilson Boulevard Arlington, VA 22209-3901 www.psych.org Library of Congress Cataloging-in-Publication Data Dimensional models of personality disorders : refining the research agenda for DSM-V / edited by Thomas A. Widiger ... [et al.]. — 1st ed. p. ; cm. Includes bibliographical references and index. ISBN 0-89042-296-6 (pbk. : alk. paper) 1. Personality disorders—Classification. 2. Diagnostic and statistical manual of mental disorders. I. Widiger, Thomas A. [DNLM: 1. Personality Disorders—classification. 2. Models, Psychological. WM 15 D582 2006] RC554.D555 2006 616.85'8—dc22 2006014661 British Library Cataloguing in Publication Data A CIP record is available from the British Library.

This text is dedicated to the memory of Jerry S. Wiggins

This page intentionally left blank

CONTENTS CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi DISCLOSURE STATEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv FOREWORD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Darrel A. Regier, M.D., M.P.H. PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi Thomas A. Widiger, Ph.D. Erik Simonsen, M.D. INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxv Erik Simonsen, M.D. Thomas A. Widiger, Ph.D.

1

ALTERNATIVE DIMENSIONAL MODELS OF PERSONALITY DISORDER: Finding a Common Ground. . . . . . . . . . . . . 1 Thomas A. Widiger, Ph.D. Erik Simonsen, M.D.

2

COMMENTARY ON WIDIGER AND SIMONSEN: Toward a Consensus Personality Trait Structure. . . . . . . . . . . . . . . . . . 23 Lee Anna Clark, Ph.D.

3

COMMENTARY ON WIDIGER AND SIMONSEN: Working Out a Dimensional Framework . . . . . . . . . . . . . . . . . . . . . . . 29 John M. Oldham, M.D.

4

COMMENTARY ON WIDIGER AND SIMONSEN: From ICD-10 and DSM-IV to ICD-11 and DSM-V . . . . . . . . . . . . . . . . 33 Charles B. Pull, M.D., Ph.D.

5

BEHAVIORAL AND MOLECULAR GENETIC CONTRIBUTIONS TO A DIMENSIONAL CLASSIFICATION OF PERSONALITY DISORDER. . . . . 39 W. John Livesley, M.D., Ph.D.

6

COMMENTARY ON LIVESLEY: Genetic Contributions to a Dimensional Classification: Problems and Pitfalls . . . . . . . . . . . . . . 55 Peter McGuffin, M.D., Ph.D.

7

NEUROBIOLOGICAL DIMENSIONAL MODELS OF PERSONALITY: A Review of Three Models . . . . . . . . . . . . . . . . . . 61 Joel Paris, M.D.

8

COMMENTARY ON PARIS: Personality as a Dynamic Psychobiological System . . . . . . . . . . . . . . . 73 C. Robert Cloninger, M.D.

9

COMMENTARY ON PARIS: The Problem of Severity in Personality Disorder Classification . . . . . . . 77 Peter Tyrer, M.D.

10 TEMPERAMENT AND PERSONALITY AS BROAD-SPECTRUM ANTECEDENTS OF PSYCHOPATHOLOGY IN CHILDHOOD AND ADOLESCENCE . . . . . . . . . . . . . . . . . . . . . . . . 85 Ivan Mervielde, Ph.D. Barbara De Clercq, Ph.D. Filip De Fruyt, Ph.D. Karla Van Leeuwen, Ph.D.

11 COMMENTARY ON MERVIELDE ET AL.: Toward a Developmental Perspective on Personality Disorders . . . . . 111 Rebecca L. Shiner, Ph.D.

12 PERSONALITY DIMENSIONS ACROSS CULTURES . . . . . . . . . . . . . . . 117 Jüri Allik, Ph.D.

13 COMMENTARY ON ALLIK: The Lexical Approach to the Study of Personality Structure. . . . . . . . 133 Michael Ashton, Ph.D.

14 COMMENTARY ON ALLIK: A Historical Perspective on Personality Disorder . . . . . . . . . . . . . . . . 139 Juan J. López-Ibor, M.D.

15 COMMENTARY ON ALLIK: Cross-Cultural Diagnosis of Personality Disorders. . . . . . . . . . . . . . . 143 Yueqin Huang, M.D., M.P.H., Ph.D. Siu Wa Tang, M.B., Ph.D., F.R.C.P.(C.)

16 CONTINUITY OF AXES I AND II:

Toward a Unified Model of Personality, Personality Disorders, and Clinical Disorders . . . . . . . . . . 149 Robert F. Krueger, Ph.D.

17 COMMENTARY ON KRUEGER: What to Do With the Old Distinctions. . . . . . . . . . . . . . . . . . . . . . . . 163 M. Tracie Shea, Ph.D.

18 COMMENTARY ON KRUEGER: Traits Versus Types in the Classification of Personality Pathology . . . . 167 David Watson, Ph.D.

19 DIMENSIONAL MODELS OF PERSONALITY DISORDER: Coverage and Cutoffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Timothy J. Trull, Ph.D.

20 COMMENTARY ON TRULL:

Drizzling on the 5 ± 3 Factor Parade . . . 189

Drew Westen, Ph.D.

21 COMMENTARY ON TRULL:

Just Do It: Replace Axis II With a Diagnostic System Based on the Five-Factor Model of Personality . . . 195 Paul Costa Jr., Ph.D.

22 COMMENTARY ON TRULL:

Reservations and Hopes . . . . . . . . . . . . 199

Carl C. Bell, M.D.

23 CLINICAL UTILITY OF DIMENSIONAL MODELS FOR PERSONALITY PATHOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Roel Verheul, Ph.D.

24 COMMENTARY ON VERHEUL: Focusing on the Clinician’s Need for a Better Model . . . . . . . . . . . . . 219 Erik Simonsen, M.D.

25 COMMENTARY ON VERHEUL: Clinical Utility of Dimensional Models for Personality Pathology . . . . 227 Theresa Wilberg, M.D., Ph.D.

26 PERSONALITY DISORDER RESEARCH AGENDA FOR DSM-V . . . . . . . 233 Thomas A. Widiger, Ph.D. Erik Simonsen, M.D. Robert F. Krueger, Ph.D. W. John Livesley, M.D., Ph.D. Roel Verheul, Ph.D. INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

This page intentionally left blank

CONTRIBUTORS Jüri Allik, Ph.D. Professor, Department of Psychology, University of Tartu, Estonia; Supervisor, Perception and Personality Research Team, The Estonian Center of Behavioral and Health Sciences, Tartu, Estonia Michael Ashton, Ph.D. Associate Professor, Department of Psychology, Brock University, St. Catharines, Ontario, Canada Carl C. Bell, M.D. C.E.O./President, Community Mental Health Council, Inc., Chicago, Illinois; Professor of Psychiatry and Public Health, Director of Public and Community Psychiatry, Department of Psychiatry, University of Illinois at Chicago Lee Anna Clark, Ph.D. Professor of Psychology, Department of Psychology, University of Iowa, Iowa City, Iowa C. Robert Cloninger, M.D. Wallace Renard Professor of Psychiatry and Genetics, Department of Psychiatry, Washington University Medical School, Saint Louis, Missouri Paul Costa Jr., Ph.D. Chief, Laboratory of Personality and Cognition, Gerontology Research Center, National Institute on Aging, National Institutes of Health, Baltimore, Maryland; Professor of Behavioral Biology, Department of Psychiatry and Behavioral Sciences, The Johns Hopkins University School of Medicine, Baltimore, Maryland Barbara De Clercq, Ph.D. Researcher, Department of Developmental, Personality and Social Psychology, Ghent University, Ghent, Belgium Filip De Fruyt, Ph.D. Professor, Department of Developmental, Personality, and Social Psychology, Ghent University, Ghent, Belgium xi

xii

Dimensional Models of Personality Disorders

Yueqin Huang, M.D., M.P.H., Ph.D. Professor, Institute of Mental Health, Peking University, Beijing, China Robert F. Krueger, Ph.D. Associate Professor, Department of Psychology, University of Minnesota, Minneapolis, Minnesota W. John Livesley, M.D., Ph.D. Professor, Department of Psychiatry, University of British Columbia, Vancouver, British Columbia, Canada Juan J. López-Ibor, M.D. Chairman, Institute of Psychiatry and Mental Health, Hospital Clínica San Carlos, Universidad Complutense, Madrid, Spain Peter McGuffin, M.D., Ph.D. Director and Professor of Psychiatric Genetics, Medical Research Council Social, Genetic, and Developmental Psychiatry Centre, Institute of Psychiatry, King’s College London, United Kingdom Ivan Mervielde, Ph.D. Professor, Department of Developmental, Personality and Social Psychology, Ghent University, Ghent, Belgium John M. Oldham, M.D. Professor and Chairman, Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, Charleston, South Carolina Joel Paris, M.D. Professor of Psychiatry, McGill University, Montreal, Quebec, Canada Charles B. Pull, M.D., Ph.D. Professor and Chairman, Department of Neurosciences, Centre Hospitalier de Luxembourg, Luxembourg, Grand-Duché de Luxembourg M. Tracie Shea, Ph.D. Professor, Department of Psychiatry and Human Behavior, Brown University Medical School, Veterans Affairs Medical Center, Providence, Rhode Island Rebecca L. Shiner, Ph.D. Associate Professor, Department of Psychology, Colgate University, Hamilton, New York

Contributors

xiii

Erik Simonsen, M.D. Associate Research Professor, University of Copenhagen, and Medical Director, Institute of Personality Theory and Psychopathology, Copenhagen, Denmark Paul J. Sirovatka, M.S. Associate Director for Research Policy Analysis, Division of Research/American Psychiatric Institute for Research and Education, Arlington, Virginia Siu Wa Tang, M.B., Ph.D., F.R.C.P.(C.) Professor and Head, Department of Psychiatry, The University of Hong Kong, Hong Kong, China Timothy J. Trull, Ph.D. Professor, Department of Psychological Sciences, University of Missouri at Columbia, Columbia, Missouri Peter Tyrer, M.D. Professor of Community Psychiatry, Department of Psychological Medicine, Imperial College London, United Kingdom Karla Van Leeuwen, Ph.D. Postdoctoral researcher, Department of Developmental, Personality and Social Psychology, Ghent University, Ghent, Belgium Roel Verheul, Ph.D. Professor of Personality Disorders, Department of Clinical Psychology, University of Amsterdam (UvA), Amsterdam, The Netherlands; Managing Director, Viersprong Institute for Studies on Personality Disorders (VISPD), Halsteren, The Netherlands David Watson, Ph.D. F. Wendell Miller Professor, Department of Psychology, University of Iowa, Iowa City, Iowa Drew Westen, Ph.D. Professor, Departments of Psychology and Psychiatry, Emory University, Atlanta, Georgia Thomas A. Widiger, Ph.D. Professor, Department of Psychology, University of Kentucky, Lexington, Kentucky Theresa Wilberg, M.D., Ph.D. Department for Research and Education, Psychiatric Division, Ulleval University Hospital, Oslo, Norway

This page intentionally left blank

DISCLOSURE STATEMENT The research conference series that produced this monograph is supported with funding from the U.S. National Institutes of Health (NIH), Grant No. U13MH067855 (Principal Investigator: Darrel A. Regier, M.D., M.P.H.). The National Institute of Mental Health (NIMH), the National Institute on Drug Abuse (NIDA), and the National Institutes on Alcohol Abuse and Alcoholism (NIAAA) jointly support this cooperative research planning conference project. The Workgroup/Conference on Personality Disorders is not part of the official revision process for the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-V), but rather is a separate, rigorous research planning initiative meant to inform revisions of psychiatric diagnostic classification systems. No private-industry sources provide funding for the research review. Coordination and oversight of the overall research review, publicly titled “The Future of Psychiatric Diagnosis: Refining the Research Agenda,” is provided by an Executive Steering Committee composed of representatives of the several entities that are cooperatively sponsoring the NIH-funded project. Present and former members are as follows: • American Psychiatric Institute for Research and Education—Darrel A. Regier, M.D., M.P.H.; support staff: William E. Narrow, M.D., M.P.H., Maritza Rubio-Stipec, Sci.D., Paul Sirovatka, M.S., Jennifer Shupinka, Rocio Salvador, and Kristin Edwards • World Health Organization—Benedetto Saraceno, M.D., and Norman Sartorius, M.D., Ph.D. (consultant) • National Institutes of Health—Wayne Fenton, M.D. (NIMH), Wilson Compton, M.D. (NIDA), and Bridget Grant, Ph.D. (NIAAA); NIMH grant project officers have included Bruce Cuthbert, Ph.D., Lisa Colpe, Ph.D., Michael Kozak, Ph.D., and Karen H. Bourdon, M.A. • Columbia University—Michael B. First, M.D. (consultant)

xv

This page intentionally left blank

FOREWORD Darrel A. Regier, M.D., M.P.H.

Dimensional Models of Personality Disorders: Refining the Research Agenda for DSM-V is the first in a series of volumes that collectively will summarize an international research-planning project undertaken to assess the status of scientific knowledge that is relevant to psychiatric classification systems and to generate specific recommendations for research to advance that knowledge base. As does the current volume, each forthcoming monograph in the series will report on a conference focused on a specific diagnostic topic or category, if not a delimited diagnosis. Titled “The Future of Psychiatric Diagnosis: Refining the Research Agenda,” the conference series is being convened by the American Psychiatric Association (APA) with the collaboration of the World Health Organization (WHO) and the U.S. National Institutes of Health (NIH). The APA/WHO/NIH conference series and monographs represent key elements in an extensive research review process designed to set the stage for the fifth revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM-V). APA intends that information and recommendations developed as part of this process also should be available to scientific groups who are updating other national and international classifications of mental and behavioral disorders. Within the APA, the American Psychiatric Institute for Research and Education (APIRE), under the direction of Darrel A. Regier, M.D., M.P.H., holds lead responsibility for organizing and administering the diagnosis research planning conferences. Co-sponsors, and members of the Executive Steering Committee for the series, include representatives of the WHO’s Department of Mental Health and Substance Abuse and of three NIH institutes that are jointly funding the project: the National Institute of Mental Health (NIMH), the National Institute on Drug Abuse (NIDA), and the National Institute on Alcohol Abuse and Alcoholism (NIAAA). APA published the fourth edition of the DSM in 1994 and a text revision in 2000. With a target publication date of 2011 for DSM-V, we initiated planning for the fifth revision 12 years in advance, with the aim of stimulating research that would address identified opportunities as well as gaps in nosological research. The initial step was a 1999 collaboration between APA and NIMH that led to prepaxvii

xviii

Dimensional Models of Personality Disorders

ration of six white papers that proposed broad-brush recommendations for research in key areas; topics included Developmental Issues, Gaps in the Current Classification, Disability and Impairment, Neuroscience, Nomenclature, and Cross-Cultural Issues. Each team that developed a paper included at least one liaison member from NIMH, with the intent—largely realized—that these members would integrate many of the workgroups’ recommendations into NIMH research programs. The volume A Research Agenda for DSM-V (Kupfer et al. 2002) more recently has been followed by a second compilation of white papers (Narrow et al., in press) that outline diagnosis-related research needs in the areas of gender, infants and children, and geriatric populations. As a second phase of planning, the APA leadership envisioned a series of conferences that would address specific diagnostic topics in greater depth, with conference proceedings serving as resource documents for groups involved in the official DSM-V revision process. A prototype symposium on mood disorders was held in conjunction with the XII World Congress of Psychiatry in Yokohama, Japan, in late 2002. Presentations addressed diverse topics in depression-related research, including preclinical animal models, genetics, pathophysiology, functional imaging, clinical treatment, epidemiology, prevention, medical comorbidity, and public health implications of the full spectrum of mood disorders. This pilot effort underscored the importance of structuring multidisciplinary research planning conferences in a manner that would force interaction among investigators from different fields and elicit a sharp focus on the diagnostic implications of recent and planned research. These emphases strongly influenced the proposal for the cooperative research planning conference grant that NIMH awarded to APIRE in 2003, with substantial additional funding support from NIDA and NIAAA. Eleven conferences funded under the grant are the basis for this and future monographs in this series, and represent a second major phase in the scientific review and planning for DSM-V. Finally, a third major component of advance planning is the DSM-V Prelude Project, an APA-sponsored Web site designed to keep the DSM user community and the public informed about research and other activities related to the fifth revision of the manual. An “outreach” section of the site permits interested parties to submit comments about problems with DSM-IV and suggestions for DSM-V. All suggestions are being entered into the DSM-V Prelude database for eventual referral to the appropriate DSM-V Work Groups. This site and associated links can be accessed at www.dsm5.org. The conferences that comprise the core activity of the second phase of preparation have multiple aims. One is to promote international collaboration among members of the scientific community in order to increase the likelihood of developing a future DSM that is unified with other international classifications. A second aim is to stimulate the empirical research necessary to allow informed decision-making regarding crucial diagnostic deficiencies identified in DSM-IV. A

Foreword

xix

third is to facilitate the development of consensus criteria that could be used by the research community as alternatives to the clinical DSM criteria for future research into the etiology and pathophysiology of mental disorders. Challenging as it is, this last objective reflects widespread agreement in the field that the wellestablished reliability and clinical utility of prior DSM classifications must be matched in the future by documenting the validity of diagnoses. Given the vision of an ultimately unified international classification system, members of the Executive Steering Committee have attached high priority to ensuring the participation of investigators from all parts of the world in the project. Toward this end, each conference in the series will have two co-chairs, drawn respectively from the United States and a country other than the United States; approximately half of the 25 experts invited to each working conference are from outside the United States; and half of the conferences are being convened outside the United States. Two leaders in the field of personality disorders research—Thomas Widiger, University of Kentucky, and Erik Simonsen, University of Copenhagen—agreed to help organize and co-chair the first conference. As evident in the contributions that follow, Drs. Widiger and Simonsen, working closely with the Executive Steering Committee, succeeded in inviting a stellar roster of participants for the conference. This monograph is the second of two reports derived from the conference on personality disorders. Earlier products included two successive issues of the Journal of Personality Disorders (Widiger and Simonsen 2005a, 2005b), which published eight of the papers presented in this monograph and which appear here with the permission of the journal.

References Kupfer DJ, First MB, Regier DA: Introduction, in A Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp xv–xxiii Narrow WN, First MB, Sirovatka P, et al. (eds): Age and Gender Considerations in Psychiatric Diagnosis: A Research Agenda for DSM-V. Arlington, VA, American Psychiatric Association (in press) Widiger TA, Simonsen E: Alternative dimensional models of personality disorder: finding a common ground. J Personal Disord 19:110–130, 2005a Widiger TA, Simonsen E: Personality disorder research agenda for DSM-V. J Personal Disord 19:315–338, 2005b

This page intentionally left blank

PREFACE Thomas A. Widiger, Ph.D. Erik Simonsen, M.D.

I

n 1999, a DSM-V Research Planning Conference was held under joint sponsorship of the American Psychiatric Association (APA) and the National Institute of Mental Health (NIMH), the purpose of which was to set research priorities that might affect future classifications (McQueen 2000). One impetus for this conference was frustration with the existing nomenclature: In the more than 30 years since the introduction of the Feighner criteria by Robins and Guze, which eventually led to DSM-III, the goal of validating these syndromes and discovering common etiologies has remained elusive. Despite many proposed candidates, not one laboratory marker has been found to be specific in identifying any of the DSM-defined syndromes. Epidemiological and clinical studies have shown extremely high rates of comorbidities among the disorders, undermining the hypothesis that the syndromes represent distinct etiologies. Furthermore, epidemiological studies have shown a high degree of short-term diagnostic instability for many disorders. With regard to treatment, lack of treatment specificity is the rule rather than the exception. (Kupfer et al. 2002, p. xviii)

DSM-V Research Planning Work Groups were formed to develop white papers to guide research in a direction that would maximize impact on future editions of the diagnostic manual. The Nomenclature Work Group, charged with addressing fundamental assumptions of the diagnostic system, concluded that it is “important that consideration be given to advantages and disadvantages of basing part or all of DSM-V on dimensions rather than categories” (Rounsaville et al. 2002, p. 12). The Nomenclature Work Group recommended in particular that initial efforts toward a dimensional model of classification be conducted with the personality disorders. “If a dimensional system of personality performs well and is acceptable to clinicians, it might then be appropriate to explore dimensional approaches in other domains” (Rounsaville et al. 2002, p. 13). The white papers developed by the DSM-V Research Planning Work Groups were published in an APA monograph, edited by Drs. Kupfer, First, and Regier (Kupfer et al. 2002). The white paper addressing personality disorders provided xxi

xxii

Dimensional Models of Personality Disorders

conceptual and empirical support for a dimensional model of classification (First et al. 2002). A closely related monograph concerned with dilemmas in psychiatric diagnosis (Phillips et al. 2003) was published by the APA the following year. This monograph included a chapter by Livesley (2003) that again provided conceptual and empirical support for a dimensional model of personality disorder classification. The white papers developed by the DSM-V Research Planning Work Groups are being followed by a series of international conferences that aim to further enrich the empirical data base in preparation for the eventual development of DSM-V (a description of this conference series can be found at www.dsm5.org). These conferences are being organized with the assistance and support of the World Health Organization and are co-funded by the NIMH, the National Institute on Alcohol Abuse and Alcoholism, and the National Institute on Drug Abuse. The principal purpose of each conference is to offer recommendations for, and stimulate the production of, research that would be of most use to the authors of DSM-V. The Executive Committee governing these conferences (i.e., Drs. William Compton, Bruce Cuthbert, Michael First, Bridget Grant, Darrel Regier, Benedetto Saraceno, and Norman Sartorious) decided to devote the first conference to reviewing the research and setting a research agenda that would be most useful and effective in leading the field toward a dimensional classification of personality disorder. This conference, “Dimensional Models of Personality Disorder: Etiology, Pathology, Phenomenology, and Treatment,” was held on December 1–3, 2004, at the APA headquarters in Arlington, Virginia. Members of the conference steering committee were Drs. Robert Krueger, John Livesley, Erik Simonsen (Co-Chair), Roel Verheul, and Thomas Widiger (Co-Chair). Topics covered were 1) alternative dimensional models of personality disorder (Drs. Thomas Widiger and Erik Simonsen), 2) behavioral genetics and gene mapping (Dr. John Livesley), 3) neurobiological mechanisms (Dr. Joel Paris), 4) childhood antecedents (Drs. Ivan Mervielde, Barbara De Clercq, Filip De Fruyt, and Karla Van Leeuwen), 5) crosscultural issues (Dr. Jüri Allik), 6) Axes I and II continuity (Dr. Robert Krueger), 7) coverage and cutoff points for diagnosis (Dr. Timothy J. Trull), and 8) clinical utility (Dr. Roel Verheul). The conference began with an introductory paper by Dr. Erik Simonsen, followed by the eight plenary papers that summarized the existing research and made recommendations for future research. Each plenary address was followed by brief discussant papers, provided by Drs. Michael Ashton, Carl Bell, Lee Anna Clark, Robert Cloninger, Paul Costa, Benjamin Greenberg, Deborah Hasin, Yueqin Huang, Juan J. López-Ibor, Peter McGuffin, John Oldham, Charles Pull, M. Tracie Shea, Rebecca Shiner, Erik Simonsen, Peter Tyrer, David Watson, Drew Westen, and Theresa Wilberg. Abbreviated versions of the plenary addresses were initially published across two special issues of the Journal of Personality Disorders (Widiger and Simonsen 2005). The Editor of Journal of Personality Disorders (Dr. John Livesley) and Guilford Publications graciously consented to allow abbreviated versions of the journal

Preface

xxiii

articles to be published together within this monograph. This monograph also includes abbreviated versions of the discussant papers by Drs. Michael Ashton, Carl Bell, Lee Anna Clark, Robert Cloninger, Paul Costa, Yueqin Huang, Juan J. LópezIbor, Peter McGuffin, John Oldham, Charles Pull, M. Tracie Shea, Rebecca Shiner, Erik Simonsen, Peter Tyrer, David Watson, Drew Westen, and Theresa Wilberg. One should not infer from the conference and this monograph publication that the existing categories of personality disorder will be replaced in the next edition of the diagnostic manual by a dimensional classification. There are arguments against making such a conversion. However, it is hoped that the publication of these papers will indeed provide further support for and stimulation toward the eventual adoption of a dimensional model of personality disorder by the American Psychiatric Association and the World Health Organization.

References First MB, Bell CB, Cuthbert B, et al: Personality disorders and relational disorders: a research agenda for addressing crucial gaps in DSM, in A Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp 123–199 Kupfer DJ, First MB, Regier DA: Introduction, in A Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp xv–xxiii Livesley WJ: Diagnostic dilemmas in classifying personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 153–190 McQueen L: Committee on Psychiatric Diagnosis and Assessment update on publications and activities. Psychiatric Res Rep 16(2):3, 2000 Phillips KA, First MB, Pincus HA (eds): Advancing DSM: Dilemmas in Psychiatric Diagnosis. Washington, DC, American Psychiatric Association, 2003 Rounsaville BJ, Alarcon RD, Andrews G, et al: Basic nomenclature issues for DSM-V, in A Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp 1–19 Widiger TA, Simonsen E: Introduction to the special section: the American Psychiatric Association’s research agenda for DSM-V. J Personal Disord 19:103–109, 2005

This page intentionally left blank

INTRODUCTION Erik Simonsen, M.D. Thomas A. Widiger, Ph.D.

Current Categorical Classification of Personality Disorders The personality disorder nomenclatures of the World Health Organization’s (WHO) International Classification of Diseases, 10th Revision (ICD-10; World Health Organization 1992) and the American Psychiatric Association’s (APA) Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR; American Psychiatric Association 2000) are quite similar. The notable exceptions are as follows: 1. ICD-10 schizotypy is consistent with DSM-IV (American Psychiatric Association 1994) schizotypal personality disorder but is included within the section for schizophrenia and schizotypal and delusional disorders. 2. DSM-IV narcissistic personality disorder is not included in ICD-10. 3. Some disorders have different names: DSM-IV borderline (ICD-10 emotional unstable), DSM-IV avoidant (ICD-10 anxious), and DSM-IV obsessive-compulsive (ICD-10 anankastic). Personality disorders are regarded as being among the more important diagnoses within the APA’s diagnostic nomenclature because they have the unique distinction of being placed on a separate diagnostic axis, thereby drawing the attention of many clinicians who may otherwise not have considered their presence. They were first placed on a separate axis in DSM-III (American Psychiatric Association 1980) in order to encourage clinicians to recognize the presence of maladaptive personality traits even when their attention is understandably drawn to a disorder of more immediate, pressing concern (Frances 1980; Spitzer et al. 1980): “This separation ensures that consideration is given to the possible presence of disorders that are frequently overlooked when attention is directed to the usually more florid Axis I disorder” (American Psychiatric Association 1980, xxv

xxvi

Dimensional Models of Personality Disorders

p. 23). The reason the authors of the multiaxial system of DSM-III wanted to draw attention to personality disorders was the “accumulating evidence that the quality and quantity of preexisting personality disturbance may...influence the predisposition, manifestation, course, and response to treatment of various Axis I conditions” (Frances 1980, p. 1050). DSM-IV and ICD-10 diagnose personality disorders categorically. The clinician determines whether the personality disorder symptomatology of a patient is sufficiently close to a particular diagnostic category to warrant the respective diagnosis. To facilitate the reliability and validity of this effort, DSM-IV includes specific and explicit diagnostic criteria sets (Spitzer et al. 1980). There are a number of advantages of the categorical approach to diagnosis (Frances 1993; Gunderson et al. 1991; Millon et al. 1996), including, for instance, their ease of use by clinicians who must make rapid diagnoses of a large number of patients whom they see briefly. The typologies are historically wellestablished syndromes, and they work as a standard reference for clinicians. All current and prior diagnostic systems have been categorical, and it would represent a major shift in clinical practice to convert to a dimensional model (Frances 1993). Clinicians must have a diagnostic system to guide their practice and facilitate their conceptualization of a patient’s pathology. They would otherwise be faced with repeated analyses that could not be generalized from one patient to another. Furthermore, clinicians are able to dig beyond the manifest and make interpretations from one behavior to another (Westen 1997). Finally, the typologies restore and recompose the unity of the self and constitute the patient as a person. Personality disorders, however, are among the more controversial and problematic disorders within the diagnostic manual. Maser et al. (1991) surveyed clinicians from 42 countries with respect to DSM-III-R (American Psychiatric Association 1987): “The personality disorders led the list of diagnostic categories with which respondents were dissatisfied” (p. 275). A number of reasons exist for this dissatisfaction with the existing diagnostic categories. We will summarize here some of the concerns regarding the categorical approach: 1) excessive diagnostic co-occurrence, 2) inadequate coverage, 3) heterogeneity within diagnoses, 4) arbitrary and unstable diagnostic boundaries, and 5) inadequate scientific base.

EXCESSIVE DIAGNOSTIC CO-OCCURRENCE ICD-10 and DSM-IV provide diagnostic criterion sets to help guide the clinician toward the correct diagnosis. The intention is to help determine which particular disorder is present, the selection of which would hopefully indicate the presence of a specific pathology that will explain the occurrence of the symptoms and suggest a specific treatment that would ameliorate the patient’s suffering (Frances et al. 1995). It is evident, however, that the diagnostic nomenclatures routinely fail in the goal of guiding the clinician to the presence of one specific personality dis-

Introduction

xxvii

order. Studies have consistently indicated that many patients meet diagnostic criteria for an excessive number of personality disorder diagnoses (Bornstein 1998; Lilienfeld et al. 1994; Livesley 2003; Widiger and Trull 1998). The maladaptive personality functioning of patients does not appear to be adequately described by a single diagnostic category.

INADEQUATE COVERAGE In addition to the problem of excessive diagnostic co-occurrence, there is the opposing or complementary problem of inadequate coverage. Clinicians provide a diagnosis of personality disorder not otherwise specified (PDNOS) when they determine that a person has a personality disorder that is not adequately represented by any one of the 10 officially recognized diagnoses (American Psychiatric Association 2000). PDNOS is often the single most frequently used diagnosis in clinical practice, one explanation for which is that the existing categories are not providing adequate coverage (Verheul and Widiger 2004). Westen and Arkowitz-Westen (1998) surveyed 238 psychiatrists and psychologists with respect to their clinical practice and reported that “the majority of patients with personality pathology significant enough to warrant clinical psychotherapeutic attention (60.6%) are currently undiagnosable on Axis II” (p. 1769). The clinicians reported the treatment of commitment, intimacy, shyness, work inhibition, perfectionism, and devaluation of others that were not well described by any of the existing diagnoses. One approach to this problem is to add more diagnostic categories, but there is considerable reluctance to do so, in part because this would have the effect of increasing further the difficulties of excessive diagnostic co-occurrence and differential diagnosis (Pincus et al. 2003).

HETEROGENEITY WITHIN DIAGNOSES There are also important differences among the persons who share the same personality disorder diagnosis. Patients with the same diagnosis will vary substantially with respect to which diagnostic criteria were used to make the diagnosis, and the differences are not trivial. For example, only a subset of persons who meet the DSM-IV criteria for antisocial personality disorder will have the prototypic features of the callous, ruthless, arrogant, charming, and scheming psychopath (Hare 2003) and there are even important differences among the persons who would be diagnosed as psychopathic (Brinkley et al. 2004). The same point can be made for all of the other personality disorders (Millon et al. 1996), such as the differentiation of borderline psychopathology with respect to the dimensions of affective dysregulation, impulsivity, and behavioral disturbance (Sanislow et al. 2002), and the differentiation of dependent personality disorder into submissive, exploitable, and affectionate variants (Pincus and Wilson 2001).

xxviii

Dimensional Models of Personality Disorders

ARBITRARY AND UNSTABLE DIAGNOSTIC BOUNDARIES An additional problem has been establishing a nonarbitrary boundary with normal personality functioning. The existing diagnostic thresholds lack a compelling rationale (Tyrer and Johnson 1996). In fact, no explanation or justification has ever been provided for most of them (Widiger and Corbitt 1994). The thresholds for the DSM-III schizotypal and borderline diagnoses are the only two for which rationales have been provided. The DSM-III requirements that the patient have four of eight features for the schizotypal diagnosis and five of eight for borderline (American Psychiatric Association 1980) were determined on the basis of maximizing agreement with similar diagnoses provided by clinicians (Spitzer et al. 1979). However, the current diagnostic thresholds for these personality disorders bear little relationship with the original thresholds established for DSM-III. Blashfield et al. (1992) reported a kappa of only –.025 for the DSM-III and DSM-IIIR schizotypal personality disorders, with a reduction in prevalence from 11% to 1%. Seemingly minor changes to diagnostic criterion sets have resulted in unexpected and substantial shifts in prevalence rates across each edition of the diagnostic manuals that profoundly complicate scientific theory and public health decisions (Blashfield et al. 1992; Narrow et al. 2002).

INADEQUATE SCIENTIFIC BASE The fifth and final issue that we want to highlight is an inadequate scientific base. Blashfield and Intoccia (2000) conducted a computer search and concluded that there were “five disorders (dependent, narcissistic, obsessive-compulsive, paranoid, and passive-aggressive) that had very small literatures, averaging fewer than 10 articles per year” (p. 473). “The only personality disorder whose literature is clearly alive and growing is that of borderline personality disorder” (Blashfield and Intoccia 2000, p. 473). They characterized the research literature concerning the dependent, narcissistic, obsessive-compulsive, paranoid, passive-aggressive, schizoid, and histrionic personality disorders as being “dead” or “dying” (p. 473). The conclusions of Blashfield and Intoccia (2000) might have been overly negative. Their search appears to have been confined to studies that could be identified by the general index phrase “personality disorders,” and they may have missed the considerable research literature concerning psychopathy (Hare 2003); the many studies on sociotropy, dependency, and attachment (Bornstein 1992); and the many studies concerning narcissism and conflicted self-esteem published in the general personality literature (Widiger and Bornstein 2001). Nevertheless, their warning is well taken: “Disorders with literature growth that is dead or dying are not succeeding at accumulating new empirical knowledge, nor are they likely to be stimulating substantial clinical interest” (Blashfield and Intoccia 2000, p. 473). Most of the individual disorders included in DSM-IV have research programs and clinicians uniquely devoted to understanding their particular etiology, pathol-

Introduction

xxix

ogy, or treatment. This does not appear to be the case for many of the individual personality disorders, although there are notable exceptions (e.g., Bornstein 1992). Relative to many of the other diagnoses within the ICD-10 and DSM-IV nomenclatures, little progress is being made with respect to understanding the neurobiology, genetics, developmental antecedents, or treatment implications for many of the individual personality disorder diagnoses.

Alternative Dimensional Approach to the Classification of Personality Disorders The question of whether personality disorders are discrete clinical conditions or arbitrary distinctions along dimensions of general personality functioning has been a long-standing issue (Blashfield 1984; Kendell 1975; Schneider 1923). Proposals for a dimensional model of personality disorder have been made throughout the history of the American Psychiatric Association’s and the World Health Organization’s diagnostic manuals (e.g., Eysenck 1970; Presly and Walton 1973; Tyrer and Alexander 1979). DSM-III was quite innovative in many respects (Frances 1980; Millon et al. 1996), but it continued to diagnose personality disorders categorically despite the improvements in validity and clinical utility that would be obtained through a dimensional model of classification (Cloninger 1987; Eysenck 1987; Frances 1982; Kiesler 1986; Livesley 1985; Walton 1986; Widiger and Frances 1985; Wiggins 1982). The authors of DSM-III-R attempted to address some of the problems inherent to the categorical model by using polythetic criterion sets in which multiple diagnostic criteria are provided, only a subset of which are necessary for the diagnosis (Widiger et al. 1988). Compelling proposals for a more fundamental shift in how personality disorders are classified and diagnosed, however, continued to be made (e.g., Benjamin 1996; Clark 1992; Cloninger et al. 1993; Costa and McCrae 1990; Livesley et al. 1992; Oldham et al. 1992; Pincus and Wiggins 1990; Siever and Davis 1991; Stone 1993; Trull 1992; Tyrer 1988; Widiger 1993). The DSM-IV Personality Disorders Work Group considered proposals to include a dimensional model of classification within the diagnostic manual (Widiger and Sanderson 1995), but its final decision was to limit the presentation to simply a few sentences within the text of DSM-IV (American Psychiatric Association 1994, pp. 633–634). Further arguments and proposals for a more fundamental shift to a dimensional model of classification, however, continued to be provided (e.g., Clark et al. 1997; Cloninger and Svrakic 1997; Tyrer and Johnson 1996; Widiger and Costa 1994). There are arguments against making a conversion to a dimensional model, including (but certainly not limited to) the potential disruption to clinical practice of making a radical shift in how personality disorders are conceptualized and diag-

xxx

Dimensional Models of Personality Disorders

nosed (e.g., see Frances 1993; Gunderson et al. 1991; Shedler and Westen 2004; Zimmerman 1988). Nevertheless, the field should be open to these alternative ways of enhancing clinical utility and improving the validity of our basic concepts in classification of personality disorder. It is our hope that the publication of these papers will indeed provide further support for and stimulation toward productive research that would lead to an eventual adoption by the APA and WHO of some kind of a dimensional model of personality disorder classification.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition. Washington, DC, American Psychiatric Association, 1980 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC, American Psychiatric Association, 1987 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Benjamin LS: Interpersonal Diagnosis and Treatment of Personality Disorders, 2nd Edition. New York, Guilford, 1996 Blashfield RK: The Classification of Psychopathology. Neo-Kraepelinian and Quantitative Approaches. New York, Plenum, 1984 Blashfield RK, Intoccia V: Growth of the literature on the topic of personality disorders. Am J Psychiatry 157:472–473, 2000 Blashfield RK, Blum N, Pfohl B: The effects of changing Axis II diagnostic criteria. Compr Psychiatry 33:245–252, 1992 Bornstein RF: The dependent personality: developmental, social and clinical perspectives. Psychol Bull 112:3–23, 1992 Bornstein RF: Reconceptualizing personality disorder diagnosis in the DSM-V: the discriminant validity challenge. Clin Psychol-Sci Pr 5:333–343, 1998 Brinkley CA, Newman JP, Widiger TA, et al: Two approaches to parsing the heterogeneity of psychopathy. Clin Psychol-Sci Pr 11:69–94, 2004 Clark LA: Resolving taxonomic issues in personality disorders. J Personal Disord 6:360– 378, 1992 Clark LA, Livesley WJ, Morey L: Personality disorder assessment: the challenge of construct validity. J Personal Disord 11:205–231, 1997 Cloninger CR: A systematic method for clinical description and classification of personality variants. Arch Gen Psychiatry 44:573–588, 1987 Cloninger CR, Svrakic DM: Integrative psychobiological approach to psychiatric assessment and treatment. Psychiatry 60:120–141, 1997 Cloninger CR, Svrakic DM, Przybeck TR: A psychobiological model of temperament and character. Arch Gen Psychiatry 50:975–990, 1993 Costa PT Jr, McCrae RR: Personality disorders and the five-factor model of personality. J Personal Disord 4:362–371, 1990

Introduction

xxxi

Eysenck HJ: A dimensional system of psychodiagnostics, in New Approaches to Personality Classification. Edited by Mahrer AR. New York, Columbia University Press, 1970, pp 169–207 Eysenck HJ: The definition of personality disorders and the criteria appropriate for their description. J Personal Disord 1:211–219, 1987 Frances AJ: The DSM-III personality disorders section: a commentary. Am J Psychiatry 137:1050–1054, 1980 Frances AJ: Categorical and dimensional systems of personality diagnosis: a comparison. Compr Psychiatry 23:516–527, 1982 Frances AJ: Dimensional diagnosis of personality—not whether, but when and which. Psychol Inq 4:110–111, 1993 Frances AJ, First MB, Pincus HA: DSM-IV Guidebook. Washington, DC, American Psychiatric Press, 1995 Gunderson JG, Links PS, Reich JH: Competing models of personality disorders. J Personal Disord 5:60–68, 1991 Hare RD: Hare Psychopathy Checklist Revised (PCL-R): Technical Manual. North Tonawanda, NY, Multi-Health Systems, 2003 Kendell RE: The Role of Diagnosis in Psychiatry. London, Blackwell Scientific, 1975 Kiesler DJ: The 1982 Interpersonal Circle: an analysis of DSM-III personality disorders, in Contemporary Directions in Psychopathology: Toward the DSM-IV. Edited by Millon T, Klerman G. New York, Guilford, 1986, pp 571–597 Lilienfeld SO, Waldman ID, Israel AC: A critical examination of the use of the term “comorbidity” in psychopathology research. Clin Psychol-Sci Pr 1:71–83, 1994 Livesley WJ: The classification of personality disorder, I: the choice of category concept. Can J Psychiatry 30:353–358, 1985 Livesley WJ: Diagnostic dilemmas in classifying personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 153–190 Livesley WJ, Jackson DN, Schroeder ML: Factorial structure of traits delineating personality disorders in clinical and general population samples. J Abnorm Psychol 101:432– 440, 1992 Maser JD, Kaelber C, Weise RF: International use and attitudes toward DSM-III and DSMIII-R: growing consensus in psychiatric classification. J Abnorm Psychol 100:271–279, 1991 Millon T, Davis RD, Millon CM, et al: Disorders of Personality: DSM-IV and Beyond. New York, John Wiley & Sons, 1996 Narrow WE, Rae DS, Robins LN, et al: Revised prevalence estimates of mental disorders in the United States: using a clinical significance criterion to reconcile 2 surveys’ estimates. Arch Gen Psychiatry 59:115–123, 2002 Oldham JM, Skodol AE, Kellman HD, et al: Diagnosis of DSM-III-R personality disorders by two semistructured interviews: patterns of comorbidity. Am J Psychiatry 149:213– 220, 1992 Pincus AL, Wiggins JS: Interpersonal problems and conceptions of personality disorders. J Personal Disord 4:342–352, 1990

xxxii

Dimensional Models of Personality Disorders

Pincus AL, Wilson KR: Interpersonal variability in dependent personality. J Pers 69:223– 251, 2001 Pincus HA, McQueen LE, Elinson L: Subthreshold mental disorders: nosological and research recommendations, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 129–144 Presly AS, Walton HJ: Dimensions of abnormal personality. Br J Psychiatry 122:269–276, 1973 Sanislow CA, Morey LC, Grilo CM, et al: Confirmatory factor analysis of DSM-IV borderline, schizotypal, avoidant and obsessive-compulsive personality disorders: findings from the Collaborative Longitudinal Personality Disorders Study. Acta Psychiatr Scand 105:28–36, 2002 Schneider K: The Psychopathic Personalities. Vienna, Austria, Deuticke, 1923 Shedler J, Westen D: Dimensions of personality pathology: an alternative to the five-factor model. Am J Psychiatry 161:1743–1754, 2004 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991 Spitzer R, Endicott J, Gibbon M: Crossing the border into borderline personality and borderline schizophrenia. Arch Gen Psychiatry 36:17–24, 1979 Spitzer RL, Williams JBW, Skodol AE: DSM-III: the major achievements and an overview. Am J Psychiatry 137:151–164, 1980 Stone MH: Abnormalities of Personality. Within and Beyond the Realm of Treatment. New York, WW Norton, 1993 Trull TJ: DSM-III-R personality disorders and the five-factor model of personality: an empirical comparison. J Abnorm Psychol 101:553–560, 1992 Tyrer P: What’s wrong with DSM-III personality disorders? J Personal Disord 2:281–291, 1988 Tyrer P, Alexander J: Classification of personality disorder. Br J Psychiatry 135:163–167, 1979 Tyrer P, Johnson T: Establishing the severity of personality disorder. Am J Psychiatry 153:1593–1597, 1996 Verheul R, Widiger TA: A meta-analysis of the prevalence and usage of the personality disorder not otherwise specified (PDNOS) diagnosis. J Personal Disord 18:309–319, 2004 Walton HJ: The relationship between personality disorder and psychiatric illness, in Contemporary Directions in Psychopathology. Toward the DSM-IV. Edited by Millon T, Klerman G. New York, Guilford, 1986, pp 553–569 Westen D: Divergences between clinical and research methods for assessing personality disorders: Implications for research and the evolution of Axis II. Am J Psychiatry 154:895–903, 1997 Westen D, Arkowitz-Westen L: Limitations of Axis II in diagnosing personality pathology in clinical practice. Am J Psychiatry 155:1767–1771, 1998 Widiger TA: The DSM-III-R categorical personality disorder diagnoses: a critique and an alternative. Psychol Inq 4:75–90, 1993

Introduction

xxxiii

Widiger TA, Bornstein RF: Histrionic, narcissistic, and dependent personality disorders, in Comprehensive Handbook of Psychopathology, 3rd Edition. Edited by Adams H, Sutker P. New York, Plenum, 2001, pp 507–529 Widiger TA, Corbitt E: Normal versus abnormal personality from the perspective of the DSM, in Differentiating Normal and Abnormal Personality. Edited by Strack S, Lorr M. New York, Springer, 1994, pp 158–175 Widiger TA, Costa PT Jr: Personality and personality disorders. J Abnorm Psychol 103:78– 91, 1994 Widiger T, Frances A: The DSM-III personality disorders: perspectives from psychology. Arch Gen Psychiatry 42:615–623, 1985 Widiger TA, Sanderson CJ: Towards a dimensional model of personality disorders in DSMIV and DSM-V, in The DSM-IV Personality Disorders. Edited by Livesley WJ. New York, Guilford, 1995, pp 433–458 Widiger TA, Trull TJ: Performance characteristics of the DSM-III-R personality disorder criteria sets, in DSM-IV Sourcebook, Vol 4. Edited by Widiger TA, Frances AJ, Pincus HA, et al. Washington, DC, American Psychiatric Association, 1998, pp 357–373 Widiger T, Frances A, Spitzer R, et al: The DSM-III-R personality disorders: an overview. Am J Psychiatry 145: 786–795, 1988 Wiggins JS: Circumplex models of interpersonal behavior in clinical psychology, in Handbook of Research Methods in Clinical Psychology. Edited by Kendall P, Butcher JN. New York, Wiley, 1982, pp 183–221 World Health Organization: The ICD-10 Classification of Mental and Behavioural Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland, World Health Organization, 1992 Zimmerman M: Why are we rushing to publish DSM-IV? Arch Gen Psychiatry 45:1135– 1138, 1988

This page intentionally left blank

1 ALTERNATIVE DIMENSIONAL MODELS OF PERSONALITY DISORDER Finding a Common Ground Thomas A. Widiger, Ph.D. Erik Simonsen, M.D.

The many limitations of the categorical model of personality disorder classification are well recognized. An obvious response to this recognition is the generation of proposals for dimensional classifications. If the authors of a future edition of the diagnostic manual shift toward a dimensional model, they will have quite a few alternative proposals to consider. The purpose of this presentation is twofold. We will first simply describe alternative proposals. More important, we will suggest that future research work toward an integration of these alternative models within a common hierarchical structure.

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 2, pages 110–130, 2005).

1

2

Dimensional Models of Personality Disorders

Eighteen Proposals One approach to converting to a dimensional classification of personality disorder is to simply provide a dimensional profile of the existing (or somewhat revised) diagnostic categories (Widiger and Sanderson 1995). Three such proposals have been developed by Oldham and Skodol (2000), Tyrer and Johnson (1996), and Westen and Shedler (2000). An advantage of this approach is that it would retain the existing diagnostic constructs (e.g., antisocial), thereby easing the transition to a dimensional classification. A limitation of this approach is that there might be more fundamental dimensions of maladaptive personality functioning that cut across the existing personality disorders, contributing to their diagnostic co-occurrence. A second approach is to reorganize the existing (and perhaps expanded) diagnostic criterion sets into more clinically useful and empirically valid dimensions of maladaptive personality functioning. Four such proposals have been developed by Clark et al. (in press; assessed using the Schedule for Nonadaptive and Adaptive Personality [SNAP]), Harkness and McNulty (1994; assessed using the Personality Psychopathology—Five [PSY-5]), Livesley (2003; assessed using the Dimensional Assessment of Personality Psychopathology—Basic Questionnaire [DAPPBQ]), and Shedler and Westen (2004; assessed using the Shedler and Westen Assessment Procedure–200 [SWAP-200]). The three clusters included within DSMIV-TR (American Psychiatric Association 2000) could be said to represent a fifth version of this proposal, although the DSM-IV clusters do not in fact reorganize the criterion sets into a more coherent structure. Clark et al. (in press) include within their factor analyses of personality disorder symptoms traitlike manifestations of anxiety and mood disorders, because the diagnostic co-occurrence of personality with disorders on Axis I of DSM-IV could be due to the presence of common, underlying dimensions of maladaptive personality functioning. A third approach to a dimensional model of personality disorder is to identify spectra of dysfunction that cut across the existing personality, mood, anxiety, substance use, and other diagnostic classes. Two such proposals have been developed by Siever and Davis (1991) and Krueger (2002). Personality disorders may not only be on a continuum with Axis I disorders, they may also be on a continuum with general personality functioning, contributing to the absence of a clear boundary between normal and abnormal personality functioning and to the presence of a considerable amount of personality disorder symptomatology within the general population (Livesley 2003; Widiger and Sanderson 1995). A fourth approach is to integrate the classification of personality disorders with dimensional models of general personality structure. There are quite a few dimensional models of personality (Wiggins 2003). Eight that have been related explicitly to the DSM-IV personality disorders are those of Cloninger (2000; assessed using the Temperament and Character Inventory [TCI]), Costa and McCrae (1990; assessed using the Neuroticism–Extraversion–Openness Per-

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder

3

sonality Inventory—Revised [NEO PI-R]), Eysenck (1987; assessed using the Eysenck Personality Questionnaire [EPQ] or the Eysenck Personality Profiler [EPP]), the interpersonal circumplex (IPC; Benjamin 1996; Wiggins 2003), Millon et al. (1996; assessed using the Millon Index of Personality Styles [MIPS]), Tellegen (Watson et al. 1999; assessed using the Multidimensional Personality Questionnaire [MPQ]), Tyrer (2000; assessed using the Personality Assessment Schedule [PAS]), and Zuckerman (2002; assessed using the Zuckerman-Kuhlman Personality Questionnaire [ZKPQ]).

Toward an Integration of Alternative Models These proposals do vary in their empirical support as a dimensional model of personality disorder (e.g., O’Connor and Dyce 1998), and one of them is likely to have more validity and clinical utility than any one of the other 17. The additional papers within this text address (in part) research concerning behavioral genetics, neurobiological mechanisms, childhood antecedents, cross-cultural application, continuity with Axis I, coverage, and clinical utility of the alternative models. However, it may also be true that some common ground can be found among them. It appears to be the case, at least to us, that none of the models lacks any limitations that could not at times be well compensated through an integration with another model. Each model will have some flaws and deficits, and each model will likely have at least some useful features. The optimal decision for the authors of a future edition of the diagnostic manual might not be a zero-sum game, where one model is victorious and all other models are abandoned. Rather, the ideal solution might be to develop a common, integrative representation that includes the important contributions and potential advantages of each of the models.

COMMON HIGHER-ORDER DOMAINS Fortunately, most of the alternative models do appear to be readily integrated within a common hierarchical structure (Bouchard and Loehlin 2001; John and Srivastava 1999; Krueger and Tackett 2003; Larstone et al. 2002; Livesley 2003; Markon et al. 2005; Trull and Durrett 2005; Widiger and Mullins-Sweatt 2005; Zuckerman 2002). This should not be surprising, given that most of them are attempting to do largely the same thing (i.e., identify the fundamental dimensions of maladaptive personality functioning that underlie and cut across the existing diagnostic categories). We suggest more specifically that all but a few of the personality traits and behaviors contained within the 18 proposed models could be organized within a more fully developed, hierarchical structure. At the highest level could be the two clinical spectra of Internalization and Externalization identified by Krueger (2002) and Achenbach (1966). Immediately beneath the two dimensions of Internalization and Externalization would be three to five broad domains

4

Dimensional Models of Personality Disorders

of personality functioning. Immediately beneath these broad domains would be personality trait scales, and at the lowest level would be the more behaviorally specific diagnostic criteria. Table 1–1 indicates how the broad domains of most of the models might be aligned with one another. We will discuss in more detail below the fit at this broad level of the DAPP-BQ, EPQ, IPC, MCMI-III, MPQ, NEO PI-R, PAS, PSY-5, SNAP, TCI, ZKPQ, and the Siever and Davis (1991) clinical spectra models. However, we want to acknowledge that a couple of the models do not as readily fit within this common structure, at least based on the existing research. For example, the SWAP-200 is not included in Table 1–1 because current research suggests that it does not have a congruent higher-order factor structure (Shedler and Westen 2004). The MIPS self–other, pleasure–pain, and active–passive polarity model of Millon et al. (1996) is not included, because its alignment with the other models is not readily apparent, and only one study has empirically related its polarities to any one of the other models (i.e., Millon [1994] reports correlations of the 6 MIPS scales with the NEO PI-R). On the other hand, factor analyses of the personality disorder scales of various editions of the Millon Clinical Multiaxial Inventory (MCMI-III; Millon et al. 1996) have produced solutions that do converge well with the four (or five) factor structure (e.g., Dyce et al. 1997; O’Connor and Dyce 1998; Simonsen 2005). An important focus of future research will be to determine whether the three (or six) polarities of the MIPS and the 12 scales of the SWAP-200 can be integrated with the higher-order structure of the DAPP-BQ, SNAP, MPQ, PSY-5, MCMI-III, IPC, NEO PI-R, EPQ, ZKPQ, and PAS, or whether the MIPS and SWAP-200 concern instead aspects of maladaptive personality functioning that are not commensurate with these other dimensional models. We do expect that a common structure is likely to be found, as the intention of these models is common: identify the fundamental dimensions of maladaptive personality functioning that underlie and cut across the existing diagnostic categories. We will also present below how the subscales of the SWAP-200 might in fact correspond with the subscales of the DAPP-BQ, EPP, MPQ, NEO PI-R, PAS, SNAP, and TCI.

Extraversion Versus Introversion It is evident from Table 1–1 that most if not all of the models include a domain that concerns Extraversion, also described as Sociability, Activity, Positive Emotionality, and (when keyed in the opposite direction) Inhibition, Introversion, or Withdrawal. This domain contrasts being gregarious, talkative, assertive, and active with being withdrawn, isolated, introverted, and anhedonic. The terms Extraversion and Positive Emotionality might appear to suggest different domains of personality functioning. However, many studies have confirmed that these are in fact the same domains (Bouchard and Loehlin 2001; Harkness et al. 1995; John and Srivastava 1999; Watson et al. 1994). The title Positive Affectivity is preferred

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder

5

by some, because it is believed that Positive Affectivity might be providing the motivating force for Extraversion, reflecting individual differences in a behavioral activation (or reward sensitivity) system (Depue and Collins 1999; Pickering and Gray 1999; Watson and Clark 1997). The interpersonal term of Extraversion is preferred by others in part because it is more simply descriptive of much of the behaviors that are included within the domain and it facilitates recognition of the association with the IPC domains of affiliation (Communion) and power (Agency) (Wiggins 2003). An important decision for the authors of the future edition of the diagnostic manual will be to select the optimal term(s) to characterize this (or any other) domain.

Antagonism Versus Compliance Most of the dimensional models also include traits referring to aggressive, dissocial, or antagonistic interpersonal relatedness at this higher-order level. This domain contrasts being suspicious, rejecting, exploitative, aggressive, antagonistic, callous, deceptive, and manipulative with being trusting, compliant, agreeable, modest, dependent, diffident, and empathic. This domain is represented more narrowly by the PSY-5 and the ZKPQ, as these versions of this domain are confined largely to interpersonal aggressiveness, whereas the other models include such additional components as mistrust, exploitation, suspiciousness, deception, and arrogance. Psychoticism from Eysenck’s dimensional model is not aligned perfectly with this domain, because he included within Psychoticism both interpersonal antagonism and impulsive disinhibition (Bouchard and Loehlin 2001; Clark and Watson 1999; Eysenck 1987; John and Srivastava 1999), comparable to the conceptualization of this domain by Siever and Davis (1991). A potential point of confusion of Table 1–1 is that Psychoticism scales are included in two different locations. This reflects the fact that this single term has been used to refer to quite different constructs. The Psychoticism of Eysenck’s (1987) EPQ is not the same as the Psychoticism of Harkness and McNulty’s (1994) PSY-5. As we just indicated, the Psychoticism of Eysenck (1987) refers to impulsive and aggressive behaviors, whereas the Psychoticism of the PSY-5 refers to cognitive and perceptual aberrations. The three-dimensional models of the MPQ and the SNAP do not include this domain of personality functioning at this higher-order level. The SNAP does include scales for mistrust, manipulativeness, and aggression, but these are placed within the domain of Negative Affectivity (Clark et al., in press), just as the MPQ includes an aggression scale within the domain of Negative Emotionality (A. Tellegen, N.G. Waller, “Exploring Personality Through Test Construction: Development of the Multidimensional Personality Questionnaire” [unpublished manuscript], Minneapolis, MN, 1987). Being mistrustful, aggressive, and manipulative does often (if not invariably) include a negative affect of angry hostility. However, joint-factor analyses of the DAPP-BQ and SNAP subscales have consistently

6

TABLE 1–1.

Alignment of alternative dimensional models: broad domains First

Second

Third

Fourth

DAPP-BQ

–Inhibition

Dissocial

Compulsivity

Emotional dysregulation

NEO PI-R

Extraversion

Antagonism

Conscientiousness

Neuroticism

SNAP and MPQ

Positive Affectivity

(Negative Affectivity)

Constraint

Negative Affectivity

PSY-5

Positive Aggressiveness Emotionality

Constraint

Negative Emotionality

Psychoticism

Agency

MCMI-III

–Withdrawn

EPQ and EPP

Extraversion

ZKPQ

Sociability

Communion Aggressiveness Constraint

Neuroticism

Psychoticism Aggression– Hostility

Neuroticism –Impulsive

Neuroticism

Activity PAS

–Withdrawn

Openness

Antisocial– Dependent

Inhibited

Dimensional Models of Personality Disorders

IPC

Fifth

Alignment of alternative dimensional models: broad domains (continued) First

Siever-Davis

TCI

(–Inhibition)

Second

Third Aggression– Impulsive

–Cooperativeness

Fourth

Fifth

Affective Instability Anxiety/ Inhibition

Cognitive– Perceptual

Persistence SelfHarm Novelty Seeking Directedness Avoidance

SelfTranscendence Reward Dependence

Note. Off-center scales lie between the domains defined by the adjoining columns. Italicized scales describe domains that are somewhat narrower in scope. Scales within parentheses are more strongly related to another domain. DAPP-BQ = Dimensional Assessment of Personality Psychopathology—Basic Questionnaire (Livesley 2003); EPQ = Eysenck Personality Questionnaire (Eysenck 1987); EPP = Eysenck Personality Profiler (Eysenck 1987); IPC = interpersonal circumplex model (Benjamin 1996; Wiggins 2003); MCMI-III = Millon Clinical Multiaxial Inventory—Third Edition (Millon et al. 1996); MPQ = Multidimensional Personality Questionnaire (Watson et al. 1999); NEO PI-R = Neuroticism–Extraversion–Openness (NEO) Personality Inventory—Revised (Costa and McCrae 1990); PAS = Personality Assessment Schedule (Tyrer 2000); PSY-5 = Personality Psychopathology—Five (Harkness and McNulty 1994); SNAP = Schedule for Nonadaptive and Adaptive Personality (Clark et al., in press); Siever-Davis = clinical spectra model (Siever and Davis 1991); SWAP-200 = Shedler and Westen Assessment Procedure–200 (Shedler and Westen 2004); TCI = Temperament and Character Inventory (Cloninger 2000); ZKPQ = Zuckerman-Kuhlman Personality Questionnaire (Zuckerman 2002).

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder

TABLE 1–1.

7

8

Dimensional Models of Personality Disorders

yielded a four-factor solution (Clark and Livesley 2002; Clark et al. 1996) that corresponds to the first four domains of Table 1–1. As indicated by Watson et al. (1994), “extensive data indicate that...the Big Three and Big Five models define a common ‘Big Four’ space” (p. 24), consisting of Negative Affectivity (Neuroticism), Positive Affectivity (Extraversion), Antagonism, and Constraint.

Constraint Versus Impulsivity All but a couple of the models also include a domain concerned with the control and regulation of behavior, referred to as Constraint, Compulsivity, or Conscientiousness, or, when keyed in the opposite direction, Impulsivity or Disinhibition. It contrasts being disciplined, compulsive, dutiful, conscientious, deliberate, workaholic, and achievement-oriented with being irresponsible, lax, impulsive, negligent, and hedonistic (Constraint, as assessed by the SNAP and MPQ, also contain aspects of Antagonism). Dimensional models that do not include this domain of personality functioning are the IPC and the PAS. Tyrer (2000) places the symptoms of the obsessive-compulsive (anankastic) personality disorder within his PAS Inhibited domain, which is otherwise defined largely by traits of anxiousness and dysphoria (i.e., a somewhat different meaning for the term Inhibition than is used, for instance, for the DAPP-BQ). The IPC does not include Constraint versus Disinhibition, as it is a two-dimensional model confined to domains of interpersonal relatedness. An additional example of the same term having different meanings is the Harm Avoidance of the TCI and the MPQ. Harm Avoidance in the TCI refers to an anxious behavioral inhibition, whereas Harm Avoidance in the MPQ is a quite different construct, referring to a low Constraint (or behavioral disinhibition that is potentially fearless). The differences between TCI Harm Avoidance and MPQ Harm Avoidance are so striking that they are in fact placed within different broad domains.

Emotional Dysregulation Versus Emotional Stability Finally, it is also evident from Table 1–1 that all but one of the models include a broad domain of emotional dysregulation, otherwise described as Negative Affectivity or Neuroticism. It contrasts feeling anxious, depressed, angry, despondent, labile, helpless, self-conscious, and vulnerable with feeling emotionally stable, selfassured, invulnerable, calm, glib, shameless, and invincible. The only model not to include this domain of personality functioning is again the IPC. This fourth domain is also somewhat more narrowly defined by Siever and Davis (1991), as they separate anxiousness from affective instability. In sum, the predominant models of normal and abnormal personality functioning appear to converge upon four broad domains of personality functioning (Bouchard and Loehlin 2001; John and Srivastava 1999; Krueger and Tackett 2003; Larstone et al. 2002; Livesley 2003; Markon et al. 2005; Trull and Durrett

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder

9

2005; Watson et al. 1994; Widiger and Mullins-Sweatt 2005; Zuckerman 2002) that can be described as Extraversion versus Introversion, Antagonism versus Compliance, Constraint versus Impulsivity, and Emotional Dysregulation versus Emotional Stability. The authors of these various models would not all agree on the best names for each domain, due in part to the fact that 1) no single name is likely to optimally describe an entire domain; 2) some models place more emphasis on the normal variants (e.g., NEO PI-R and TCI), whereas other models place more emphasis on the abnormal variants (e.g., DAPP-BQ and SNAP); and, finally, 3) the models vary in how broadly or narrowly they define each domain. We have provided tentative names for each domain that emphasize (for the most part) the maladaptive variants, as these would be of most relevance and interest to clinicians. In any case, the convergence among the alternative models with respect to the existence of the four domains is quite evident. Empirical support for the convergence of these models has been provided in quite a number of studies (e.g., Austin and Deary 2000; Clark et al. 1996; Deary et al. 1998; Dyce et al. 1997; Livesley et al. 1998; Markon et al. 2005; Mulder and Joyce 1997; O’Connor and Dyce 1998).

Unconventionality Versus Closedness to Experience Only three of the models include a fifth broad domain, characterized within the NEO PI-R as Openness to Experience, described as unconventionality by Tellegen and Waller (A. Tellegen, N.G. Waller, “Exploring Personality Through Test Construction: Development of the Multidimensional Personality Questionnaire” [unpublished manuscript], Minneapolis, MN, 1987), identified within the PSY-5 as psychoticism (i.e., illusions, misperceptions, perceptual aberrations, and magical ideation), and identified within the clinical spectra of Siever and Davis (1991) as cognitive–perceptual aberrations. There are also subscales within the SNAP (e.g., eccentric perceptions), the DAPP-BQ (perceptual cognitive distortion), the MPQ (absorption), the PAS (eccentricity, rigidity), the SWAP-200 (thought disorder, dissociation), and the TCI (transpersonal identification, spiritual acceptance) that relate conceptual and empirically to this domain (Bouchard and Loehlin 2001; Clark and Livesley 2002). A domain of Unconventionality (or Openness) is obtained in joint factor analytic studies that provide sufficient representation (e.g., Clark and Livesley 2002; Costa and McCrae 1990; Wiggins and Pincus 1989). However, it appears to be the case that when this domain is narrowly defined as simply cognitive–perceptual aberrations, scales to assess it either load on other factors (typically negative affectivity) or define a factor that is so small that it might not appear to be worth identifying (Austin and Deary 2000; Clark et al. 1996; Larstone et al. 2002). Openness to Experience is itself the fifth and smallest domain of the Five-Factor Model (Ashton and Lee 2001). It is also possible that cognitive–perceptual aberrations do not belong within a dimensional model of abnormal personality functioning, consistent with the World Health Organization’s (1992) inclusion of DSMIV schizotypal personality disorder as a variant of schizophrenia.

10

Dimensional Models of Personality Disorders

LOWER-ORDER TRAITS AND SYMPTOMS We now turn to the constructs and scales that might be included within each one of these four (or five) broad domains. This is facilitated by the fact that some of the dimensional models include lower-order scales that have been related empirically to one another. Consideration of the lower-order scales also provides a better understanding of the potential integration of the SWAP-200 within the hierarchical structure. Space limitations prohibit a presentation of lower-order scales for all four domains. Excluded from this presentation are the lower-order scales within the domain of Emotional Dysregulation versus Emotional Stability (this material is provided in the previously published version of this chapter [Widiger and Simonsen 2005]).

Constraint Versus Impulsivity Table 1–2 provides trait scales within a domain of Constraint versus Disinhibition (or Impulsivity). Normal and abnormal variants of constraint are readily identified, with a number of scales from the TCI, PAS, MPQ, and NEO PI-R that refer to normal, adaptive levels of Constraint (or Conscientiousness), such as dutifulness, conscientiousness, responsibility, ambitiousness, achievement, resourcefulness, deliberation, control, and self-discipline, and maladaptive variants of these traits emphasized by the DAPP-BQ, SWAP-200, EPP, and SNAP (i.e., compulsivity, obsessionality, workaholism, and propriety). We placed the TCI scales of perfectionism and work-hardened and the EPP scales of impulsivity, risk taking, and irresponsibility within the abnormal range, but these could just as well have been placed within the normal range. We are not suggesting, of course, that a future edition of the diagnostic manual provide all of the 30 scales included in Table 1–2. There is clearly substantial redundancy. Only a small subset would in fact be necessary, and an important issue for future research is which subset would represent the optimal choice. Considered in this selection might be extent of overlap, adequate coverage of the domain, representation of different dimensional models, clinical relevance, familiarity, and ease of use. An additional question for future research is the potential bipolarity of the hierarchical structure. The existing hierarchical structure of DSM-IV does not include any such bipolarities (e.g., there is no maladaptive variant of low compulsivity). However, we would suggest that this bipolarity is evident in other areas of medicine. For example, there are maladaptive consequences of both high and low blood pressure, with normal blood pressure occupying an intermediate ground. In addition, a bipolar structure appears to be inherent to any hierarchical organization of the adaptive and maladaptive personality scales included within the existing instruments. For example, in the higher-order structure of the SNAP, the SNAP Impulsivity scale loads negatively on the Constraint domain, whereas the

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder 11

TABLE 1–2.

Lower-order scales within a domain of Constraint versus

Impulsivity 1.

Version including all relevant scales from respective instruments Abnormally high traits DAPP-BQ: compulsivity SNAP: workaholism TCI: perfectionism, work-hardened SWAP-200: obsessionality Normal traits NEO PI-R: dutifulness, order, achievement-striving, selfdiscipline, deliberation, competence PAS: conscientiousness MPQ: achievement, control, traditionalism, harm avoidance SNAP: propriety TCI: resourcefulness, eagerness of effort, responsibility, ambitiousness, purposefulness Abnormally low traits SNAP: impulsivity TCI: impulsiveness, disorderliness PAS: irresponsibility, childishness, impulsiveness EPP: impulsivity, risk taking, irresponsibility

2.

Simplified version Abnormal Compulsivity Workaholism Obsessionality Impulsivity Irresponsibility Normal Dutifulness Achievement-striving Resourcefulness

Note. DAPP-BQ = Dimensional Assessment of Personality Psychopathology—Basic Questionnaire; EPP = Eysenck Personality Profiler; MPQ = Multidimensional Personality Questionnaire; NEO PI-R = Neuroticism–Extraversion–Openness (NEO) Personality Inventory—Revised; PAS = Personality Assessment Schedule; SNAP = Schedule for Nonadaptive and Adaptive Personality; SWAP-200 = Shedler and Westen Assessment Procedure– 200; TCI = Temperament and Character Inventory.

12

Dimensional Models of Personality Disorders

SNAP Propriety and Workaholism scales load positively (Clark et al., in press). A comparable finding occurs for other domains. SNAP Exhibitionism loads positively on the domain of Positive Affectivity, whereas Detachment loads negatively (Clark et al., in press), just as the DAPP-BQ Stimulus Seeking scale loads positively on this domain and the DAPP-BQ Social Avoidance scale loads negatively (Livesley et al. 1998). The same finding also occurs when higher-order structures are developed with the DSM-IV personality disorder criterion sets. For example, in one of the initial efforts to integrate empirically the dimensional and categorical models of personality and personality disorder, Wiggins and Pincus (1989) indicated how the histrionic and narcissistic personality disorders loaded positively on an Extraversion dimension, whereas schizoid loaded negatively. Dependent personality disorder loaded positively on an Agreeableness factor, whereas the antisocial, paranoid, and narcissistic disorders loaded negatively. Comparable findings have occurred consistently in subsequent higher-order organizations of the DSMIV personality disorder constructs (Coker et al. 2002; O’Connor and Dyce 1998). A future diagnostic manual can avoid the conceptual complexity of this bipolarity, however, by simply excluding it from the visual presentation of the scales, consistent with how the personality disorders are currently presented. Table 1–3 also provides a much simplified version of the Constraint-versus-Impulsivity domain. Only a small subset of the constructs are provided within this simplified version, and the bipolarity is removed from the visual presentation. Note, however, that the scales included in the simplified version do not represent our suggestion of which scales should be included or excluded. Its intention is only to indicate visually that the presentation of the scales need not be as complex or burdensome as might be implied by the complete list.

Extraversion Versus Introversion Table 1–3 provides possible scales for a domain of Extraversion versus Introversion. At one pole could be maladaptive scales from the DAPP-BQ (Stimulus Seeking), SNAP (Exhibitionism), TCI (Extravagance), and SWAP-200 (Histrionic Sexualization); at the opposite pole could be the maladaptive scales from the DAPP-BQ (Intimacy Problems, Social Avoidance), SNAP (Detachment), PAS (Aloofness), and SWAP-200 (Schizoid Orientation). In between could be the normal variants of these constructs, as represented by the NEO PI-R scales concerning Gregariousness, Assertiveness, Activity, Excitement-Seeking, and Positive Emotionality; the MPQ scales of Social Potency and Social Closeness; the TCI scales of Exploratory Excitability and Sociability; and the EPP scales of Sociability, Assertiveness, and Activity. It is also important to note that much of the existing personality disorder diagnostic criteria would be easily included within this hierarchical structure. Each of the abnormal trait scales would include items for their assessment, and in most in-

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder 13 stances these items would resemble closely the existing personality disorder diagnostic criteria. In fact, the existing personality disorder diagnostic criteria are already included within the DAPP-BQ, SNAP, and SWAP-200 scales. In sum, clinicians familiar with the existing diagnostic criterion sets would readily identify much (if not all) of the existing personality disorder symptoms within the dimensional hierarchical structure. The dimensional model will have simply reorganized the criterion sets into a more coherent and empirically supported structure. In addition, the diagnostic manual could go further by providing guidelines for a profile matching with which clinicians could recover the DSM-IV diagnostic constructs (e.g., antisocial or borderline). For example, what was identified in DSM-IV as schizoid personality disorder could be diagnosed by elevations on the detachment and aloofness scales. Research has indicated that these profile-matching algorithms reproduce well the findings that are currently obtained with the existing diagnostic constructs (e.g., Miller and Lynam 2003; Trull et al. 2003).

Antagonism Versus Compliance Table 1–3 also provides a description of how the respective lower-order personality trait scales from the PAS, DAPP-BQ, SNAP, TCI, NEO PI-R, MPQ, and SWAP200 might be aligned with one another within a domain of antagonism versus compliance. Scales from the NEO PI-R and TCI refer largely to normal variants (i.e., being trusting, compliant, straightforward, altruistic, modest, helpful, compassionate, sentimental, or empathic), whereas scales from the DAPP-BQ, SWAP200, PAS, and SNAP refer largely to abnormal, maladaptive variants of these same traits (i.e., being dependent, diffident, gullible, sacrificial, meek, docile, submissive, or self-denigrating). The Antagonism versus Compliance domain is useful in illustrating the close relationship of the normal and abnormal variants of these traits, as some scales are difficult to even classify (e.g., MPQ scale for aggression), consistent with the considerable amount of research indicating a continuum between normal and abnormal personality functioning (Cloninger 2000; Livesley 2001; Reynolds and Clark 2001; Saulsman and Page 2004; Trull and Durrett 2005; Tyrer 2001; Widiger and Costa 2002). A decision for the authors of a future edition of the diagnostic manual will be whether to include normal variants of each of the domains of personality functioning. There are arguments against doing so (e.g., the argument that the diagnostic manual is not for the purpose of describing normal psychological functioning). However, there are also compelling arguments for including normal personality scales—for example, their inclusion will allow for the provision of a more comprehensive description of a patient’s entire personality functioning; will facilitate an integration of the diagnostic manual with basic science research on general personality functioning; and may be helpful clinically by identifying adaptive personality traits that contribute to treatment responsivity.

14

Dimensional Models of Personality Disorders

TABLE 1–3.

Lower-order scales within the domains of Extraversion versus Introversion and Antagonism versus Compliance 1.

Extraversion versus Introversion Abnormally high traits DAPP-BQ: SNAP: TCI: SWAP-200:

stimulus seeking exhibitionism, (entitlement) extravagance (histrionic sexualization)

Normal traits NEO PI-R: MPQ: TCI: EPP: PAS:

gregariousness, assertiveness, activity, excitementseeking, positive emotionality (warmth) social potency, social closeness, well-being exploratory excitability, sociability, (attachment) sociability, assertiveness, activity (optimism)

Abnormally low traits DAPP-BQ:

intimacy problems, social avoidance, restricted expression SNAP: detachment PAS: aloofness, (shyness) TCI: (shyness) SWAP-200: schizoid orientation 2.

Antagonism versus Compliance Abnormally high traits DAPP-BQ:

narcissism, suspiciousness, interpersonal disesteem, conduct problems, passive oppositionality, rejection SNAP: mistrust, manipulativeness, aggression, entitlement MPQ: aggression, (alienation) PAS: suspiciousness, aggression, callousness SWAP-200: narcissism, psychopathy Normal traits NEO PI-R: TCI: MPQ:

trust, straightforwardness, altruism, compliance, modesty, tender-mindedness, agreeableness helpfulness, compassion, pure-hearted, sentimentality, empathy, social acceptance, attachment (social closeness)

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder 15

TABLE 1–3.

Lower-order scales within the domains of Extraversion versus Introversion and Antagonism versus Compliance (continued) 2.

Antagonism versus Compliance (continued) Abnormally low traits DAPP-BQ: SNAP: PAS: TCI:

diffidence, insecure attachment (dependency) dependence, submissiveness dependence

Note. Some scales are noted parenthetically because they include aspects of personality function from another domain. DAPP-BQ = Dimensional Assessment of Personality Psychopathology—Basic Questionnaire; EPP = Eysenck Personality Profiler; MPQ = Multidimensional Personality Questionnaire; NEO PI-R = Neuroticism–Extraversion–Openness (NEO) Personality Inventory—Revised; PAS = Personality Assessment Schedule; SNAP = Schedule for Nonadaptive and Adaptive Personality; SWAP-200 = Shedler and Westen Assessment Procedure–200; TCI = Temperament and Character Inventory.

As we discuss further below, future research might indicate that some of our scale placements are inaccurate or at least not optimal. For example, a number of studies have placed Dependency scales within a domain of Neuroticism or Emotional Instability (e.g., Clark and Livesley 2002; Clark et al. 1996; De Clercq and De Fruyt 2003; Trull 1992). However, quite a few studies have placed Dependency scales within a domain of Agreeableness (e.g., Blais 1997; Coker et al. 2002; Costa and McCrae 1990; Dyce and O’Connor 1998; Haigler and Widiger 2001; Hyer et al. 1994; Lynam and Widiger 2001; Pincus and Gurtman 1995; Sprock 2002; Wiggins and Pincus 1989; Zuroff 1994). The IPC dimensional models consistently place Dependency scales within an Agreeableness (or Compliance) domain (Pincus and Gurtman 1995; Widiger and Hagemoser 1997), as the IPC dimensional models do not include a Neuroticism (or Negative Affectivity) domain. The inconsistency in placement is perhaps due largely to the complexity of the dependent personality disorder construct (i.e., involving traits of both Neuroticism and Agreeableness).

Conclusion and Recommendations In sum, we suggest that an important goal of future research will be the identification of a common ground among the alternative dimensional models of personality disorder. We recognize that future research will continue to focus on the particular strengths, nuances, and advantages of alternative models. The respective validity and clinical utility of each alternative model would be informative to the

16

Dimensional Models of Personality Disorders

authors of a future edition of the diagnostic manual in making decisions regarding which specific components of each model to include within an integrative structure. However, we would also encourage researchers to themselves consider the possibility of working toward a more unified, integrative model. Given that these models are all attempting to do largely the same thing, it would seem likely that they share common ground. It is possible that the authors of a future edition of a diagnostic manual will come to the decision that one particular model is preferable to all of the others. However, it is our opinion that it is unlikely that any one particular model will lack any redeeming or useful features or, conversely, that any one particular model will be without any meaningful faults or limitations. We would therefore suggest that research that leads to an integrative structure will be particularly informative to the authors of this future diagnostic manual. It is also possible that a common, integrative structure will not resemble closely the illustrative model we provided. Our integrative model is guided by a considerable amount of supportive research (e.g., Austin and Deary 2000; Clark et al. 1996; Clark et al. 1993; Costa and McCrae 1990; Deary et al. 1998; De Clercq and De Fruyt 2003; De Fruyt et al. 2000; Duijsens and Diekstra 1996; Dyce et al. 1997; Livesley et al. 1998; Markon et al. 2005; Mulder and Joyce 1997; O’Connor and Dyce 1998; Ramanaiah et al. 2002; Reynolds and Clark 2001; Schroeder et al. 1992; Trull 1992; Trull et al. 1995; Zuckerman 2002), but admittedly there are some studies that suggest that some of the existing dimensional models might not be well integrated within this structure (e.g., Shedler and Westen 2004) and that the placement of some models (e.g., PAS) is based on only a limited amount of research. Our placements of some of the PAS, DAPP-BQ, SNAP, TCI, NEO PI-R, MPQ, and SWAP-200 scales might also be disputed. Perhaps we have misunderstood the constructs assessed by these scales or the research that has been conducted to date. Some of the placements (e.g., those placed parenthetically) were difficult because the constructs appear to contain aspects from more than one domain. Future research could help determine how these alternative dimensional models of personality disorder could be best integrated into a common, unified, hierarchical structure. The devil, of course, could be in the details. It is apparent that the illustrative model includes considerable redundancy (a direct effect of the alternative efforts to describe a common ground) and uncertain labeling. If the authors of a future diagnostic manual prefer to use an integrative, hierarchical structure, they will need to decide which scales and constructs will be optimal for inclusion, and how best to represent them. Considered in this selection could be extent of overlap, adequate representation of different models, adequate coverage of the domain, clinical relevance, familiarity, and ease of usage. Last but not least is whether the diagnostic manual should include normal, adaptive traits. We again argue for the importance of their inclusion. The inclusion of normative, adaptive traits will facilitate the provision of a more comprehensive (and accurate) description of each

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder 17 patient’s general personality structure; it will facilitate an integration of the diagnostic manual with basic science research on general personality structure; and it will facilitate treatment decisions through the recognition of traits that contribute to an understanding of treatment responsivity. Even if the diagnostic manual does not explicitly include normal personality traits, it should be closely coordinated with them so that the American Psychiatric Association diagnostic manual of personality disorders is itself well integrated and coordinated with basic science research on general personality structure.

References Achenbach TM: The classification of children’s psychiatric symptoms: a factor-analytic study. Psychol Monogr 80:1–37, 1966 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Ashton MC, Lee K: A theoretical basis for the major dimensions of personality. Eur J Personality 15:327–353, 2001 Austin EJ, Deary IJ: The “four As”: a common framework for normal and abnormal personality? Personality and Individual Differences 28:977–995, 2000 Benjamin LS: Interpersonal Diagnosis and Treatment of Personality Disorders, 2nd Edition. New York, Guilford, 1996 Blais MA: Clinician ratings of the five-factor model of personality and the DSM-IV personality disorders. J Nerv Ment Dis 185:388–393, 1997 Bouchard TJ, Loehlin JC: Genes, evolution, and personality. Behav Genet 31:243–273, 2001 Clark LA, Simms LJ, Wu KD, Casillas A: Schedule for Nonadaptive and Adaptive Personality—2nd Edition (SNAP-2): Manual for Administration, Scoring, and Interpretation. Minneapolis, MN, University of Minnesota Press, in press Clark LA, Livesley WJ: Two approaches to identifying the dimensions of personality disorder: convergence on the five-factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 161–176 Clark LA, Watson D: Temperament: a new paradigm for trait psychology, in Handbook of Personality. Theory and Research, 2nd Edition. Edited by Pervin L, John O. New York, Guilford, 1999, pp 399–423 Clark LA, Livesley WJ, Schroeder ML, et al: Convergence of two systems for assessing specific traits of personality disorder. Psychol Assess 8:294–303, 1996 Clark LA, McEwen JL, Collard L, et al: Symptoms and traits of personality disorder: two new methods for their assessment. Psychol Assess 5:81–91, 1993 Cloninger CR: A practical way to diagnosis personality disorder: a proposal. J Personal Disord 14:99–108, 2000 Coker LA, Samuel DB, Widiger TA: Maladaptive personality functioning within the Big Five and the FFM. J Personal Disord 16:385–401, 2002 Costa PT Jr, McCrae RR: Personality disorders and the five-factor model of personality. J Personal Disord 4:362–371, 1990

18

Dimensional Models of Personality Disorders

Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO PI-R) and NEO FiveFactor Inventory (NEO-FFI) Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992 Deary IJ, Peter A, Austin E, et al: Personality traits and personality disorders. Br J Psychol 89:647–661, 1998 De Clercq B, De Fruyt F: Personality disorder symptoms in adolescence: a five-factor model perspective. J Personal Disord 17:269–292, 2003 De Fruyt F, van De Wiele L, van Heeringen C: Cloninger’s psychobiological model of temperament and character and the five-factor model of personality. Personality and Individual Differences 29:441–452, 2000 Depue RA, Collins PF: Neurobiology of the structure of personality: dopamine, facilitation of incentive motivation, and extraversion. Behav Brain Sci 22:491–569, 1999 Duijsens I, Diekstra RFW: DSM-III-R and ICD-10 personality disorders and their relationship with the big five dimensions of personality. Personality and Individual Differences 21:119–133, 1996 Dyce JA, O’Connor BP: Personality disorders and the five-factor model: a test of facet-level predictions. J Personal Disord 12:31–45, 1998 Dyce JA, O’Connor BP, Parkins S, et al: Correlational structure of the MCMI-III personality disorder scales and comparisons with other data sets. J Pers Assess 69:568–582, 1997 Eysenck HJ: The definition of personality disorders and the criteria appropriate for their description. J Personal Disord 1:211–219, 1987 Haigler ED, Widiger TA: Experimental manipulation of NEO PI-R items. J Pers Assess 77:339–358, 2001 Harkness AR, McNulty JL: The Personality Psychopathology Five (PSY-5): issue from the pages of a diagnostic manual instead of a dictionary, in Differentiating Normal and Abnormal Personality. Edited by Strack S, Lorr M. New York, Springer, 1994, pp 291–315 Harkness AR, McNulty JL, Ben-Porath YS: The Personality Psychopathology Five (PSY-5): constructs and MMPI-2 scales. Psychol Assess 7:104–114, 1995 Hyer L, Brawell L, Albrecht B, et al: Relationship of NEO-PI to personality styles and severity of trauma in chronic PTSD victims. J Clin Psychol 50:699–707, 1994 John OP, Srivastava S: The Big Five trait taxonomy: history, measurement, and theoretical perspectives, in Handbook of Personality: Theory and Research, 2nd Edition. Edited by Pervin LA, John OP. New York, Guilford, 1999, pp 102–138 Krueger RF: Psychometric perspectives on comorbidity, in Defining Psychopathology in the 21st Century. DSM-V and Beyond. Edited by Helzer JE, Hudziak JJ. Washington, DC, American Psychiatric Publishing, 2002, pp 41–54 Krueger RF, Tackett JL: Personality and psychopathology: working toward the bigger picture. J Personal Disord 17:109–128, 2003 Larstone RM, Jang KL, Livesley WJ, et al: The relationship between Eysenck’s P-E-N model of personality, the five factor model of personality, and traits delineating personality dysfunction. Personality and Individual Differences 33:25–37, 2002 Livesley WJ: Conceptual and taxonomic issues, in Handbook of Personality Disorders: Theory, Research, and Treatment. Edited by Livesley WJ. New York, Guilford, 2001, pp 3–38

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder 19 Livesley WJ: Diagnostic dilemmas in classifying personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 153–190 Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Lynam DR, Widiger TA: Using the five factor model to represent the DSM-IV personality disorders: an expert consensus approach. J Abnorm Psychol 110:401–412, 2001 Markon KE, Krueger RF, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 Miller JD, Lynam DR: Psychopathy and the five-factor model of personality: a replication and extension. J Pers Assess 81:168–178, 2003 Millon T: Millon Index of Personality Styles Manual. San Antonio, TX, Psychological Corporation, 1994 Millon T, Davis RD, Millon CM, et al: Disorders of Personality: DSM-IV and Beyond. New York, John Wiley & Sons, 1996 Millon T, Millon C, Davis RD: MCMI-III Manual. Minneapolis, MN, National Computer Systems, 1994 Mulder RT, Joyce PR: Temperament and the structure of personality disorder symptoms. Psychol Med 27:99–106, 1997 O’Connor BP, Dyce JA: A test of models of personality disorder configuration. J Abnorm Psychol 107:3–16, 1998 Oldham JM, Skodol AE: Charting the future of Axis II. J Personal Disord 14:17–29, 2000 Pickering AD, Gray JA: The neuroscience of personality, in Handbook of Personality: Theory and Research, 2nd Edition. Edited by Pervin L, John OP. New York, Guilford, 1999, pp 277–299 Pincus AL, Gurtman MB: The three faces of interpersonal dependency: Structural analysis of self-report dependency measures. J Pers Soc Psychol 69:744–758, 1995 Ramanaiah NV, Rielage JK, Cheng Y: Cloninger’s Temperament and Character Inventory and the NEO Five-Factor Inventory. Psychol Rep 90(3 Pt 2):1059–63, 2002 Reynolds SK, Clark LA: Predicting dimensions of personality disorder from domains and facets of the five-factor model. J Pers 69:199–222, 2001 Saulsman LM, Page AC: The five-factor model and personality disorder empirical literature: a meta-analytic review. Clin Psychol Rev 23:1055–1085, 2004 Schroeder ML, Wormworth JA, Livesley WJ: Dimensions of personality disorder and their relationship to the Big Five dimensions of personality. Psychol Assess 4:47–53, 1992 Shedler J, Westen D: Dimensions of personality pathology: an alternative to the five-factor model. Am J Psychiatry 161:1743–1754, 2004 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991 Simonsen E: Experiences in translating the Millon inventories in a European country, in Personality and Psychopathology. Edited by Strack S. New York, John Wiley & Sons, 2005, pp 390–402 Sprock J: A comparative study of the dimensions and facets of the five-factor model in the diagnosis of cases of personality disorder. J Personal Disord 16:402–423, 2002

20

Dimensional Models of Personality Disorders

Trull TJ: DSM-III-R personality disorders and the five-factor model of personality: an empirical comparison. J Abnorm Psychol 101:553–560, 1992 Trull TJ, Durrett CA: Categorical and dimensional models of personality disorder. Annu Rev Clin Psychol 1:355–380, 2005 Trull TJ, Useda JD, Costa PT Jr, et al: Comparison of the MMPI-2 Personality Psychopathology Five (PSY-5), the NEO-PI, and the NEO PI-R. Psychol Assess 7:508–516, 1995 Trull TJ, Widiger TA, Lynam DR, et al: Borderline personality disorder from the perspective of general personality functioning. J Abnorm Psychol 112:193–202, 2003 Tyrer P (ed): Personality Disorders. Diagnosis, Management, and Course, 2nd Edition. London, Arnold, 2000 Tyrer P: Personality disorder. Br J Psychiatry 179:81–84, 2001 Tyrer P, Johnson T: Establishing the severity of personality disorder. Am J Psychiatry 153:1593–1597, 1996 Watson D, Clark LA: Extraversion and its positive emotional core, in Handbook of Personality Psychology. Edited by Hogan R, Johnson J, Briggs S. New York, Academic Press, 1997, pp 767–793 Watson D, Clark LA, Harkness AR: Structures of personality and their relevance to psychopathology. J Abnorm Psychol 103:18–31, 1994 Watson D, Wiese D, Vaidya J, et al: The two general activation systems of affect: structural findings, evolutionary considerations, and psychobiological evidence. J Pers Soc Psychol 76:820–838, 1999 Westen D, Shedler J: A prototype matching approach to diagnosing personality disorders: toward DSM-V. J Personal Disord 14:109–126, 2000 Widiger TA, Costa PT Jr: Five factor model personality disorder research, in Personality Disorders and the Five Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 59–87 Widiger TA, Hagemoser S: Personality disorders and the interpersonal circumplex, in Circumplex Models of Personality and Emotions. Edited by Plutchik R, Conte HR. Washington, DC, American Psychological Association, 1997, pp 299–325 Widiger TA, Mullins-Sweatt SN: Categorical and dimensional models of personality disorder, in The American Psychiatric Publishing Textbook of Personality Disorders. Edited by Oldham JM, Skodol AE, Bender DS. Washington, DC, American Psychiatric Publishing, 2005, pp 35–53 Widiger TA, Sanderson CJ: Towards a dimensional model of personality disorders in DSMIV and DSM-V, in The DSM-IV Personality Disorders. Edited by Livesley WJ. New York, Guilford, 1995, pp 433–458 Widiger TA, Simonsen E: Alternative dimensional models of personality disorder: finding a common ground. J Personal Disord 19:110–130, 2005 Wiggins JS: Paradigms of Personality Assessment. New York, Guilford, 2003 Wiggins JS, Pincus AL: Conceptions of personality disorders and dimensions of personality. Psychol Assess 1:305–316, 1989 World Health Organization: (The ICD-10 Classification of Mental and Behavioural Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland, World Health Organization, 1992

Widiger and Simonsen: Alternative Dimensional Models of Personality Disorder 21 Zuckerman M: Zuckerman-Kuhlman Personality Questionnaire (ZKPQ): An alternative five-factorial model, in Big Five Assessment. Edited by de Raad B, Perugini M. Kirkland, WA, Hogrefe & Huber, 2002, pp 377–397 Zuroff DC: Depressive personality styles and the five-factor model of personality. J Pers Assess 63:453–472, 1994

This page intentionally left blank

2 COMMENTARY ON WIDIGER AND SIMONSEN Toward a Consensus Personality Trait Structure Lee Anna Clark, Ph.D.

W

idiger and Simonsen (Chapter 1 in this volume) have summarized a great deal of information to present a compelling, dimensional, hierarchical model of adaptive and maladaptive personality traits that can serve as the common basis for integration of at least 12 of 18 existing models of personality dysfunction. Lowerorder scales from a 13th model can be placed within the common structure, a 14th model emerges easily from combinations of the 4 higher-order dimensions of the common structure, and all that remain are 4 models tied directly to the DSM, which the common structure replaces. Thus, this common structure encompasses all currently well-known non-DSM models of personality pathology to some degree and, in most cases, to quite a large degree. Therefore, the primary intent of this comment is to further the thrust of Widiger and Simonsen’s argument by proposing solutions to certain difficulties they pose or do not address; offering a different viewpoint on the relation of measures to the higher-order structure; suggesting alternative placements of lower-order dimensions within the common structure; and discussing the next challenge facing the intertwined fields of personality and personality pathology. First, Widiger and Simonsen indicate that the clinical spectra of internalizing and externalizing exist at the highest level of the structure, and from them emerge the “three to five broad domains of personality functioning” (Chapter 1, pp. 3–4). However, they do not specify the relation of these three to five domains to the two 23

24

Dimensional Models of Personality Disorders

clinical spectra. I suggest that internalizing and externalizing are not a higher level of abstraction than the three to five higher-order domains, but rather result from combinations of these domains. Specifically, internalizing and externalizing are correlated through the common factor of Neuroticism/Negative Affectivity/Emotionality (hereafter, N/NA/NE). Internalizing combines N/NA/NE with low Extraversion/Positive Affectivity/Emotionality (E/PA/PE), whereas externalizing is a combination of Antagonism and Impulsivity, with N/NA/NE as a secondary, but still important, component (Krueger and Tackett 2003). In this context it is worth noting that selecting “the optimal term(s) to characterize this (or any other) domain” is more important politically than scientifically, and far less important than clarifying both the core and, to the extent possible, the boundaries of the respective domains. That is, the concepts Extraversion and Positive Emotionality both contribute to defining the same broad domain. Whether sociability, agency, or emotionality—or all three conjointly—will prove to be the core construct(s) of the domain is an empirical question whose answer awaits further research, and it will be more fruitful to devote our efforts to designing and conducting that research than to arguing over the domain name. Second, Widiger and Simonsen go to some pains to align 12 of the models in the common structure presented in their Table 1–1 (see Chapter 1, p. 7, for key to instrument abbreviations used here). Remarkably, I have only one quibble with their placement, and that is that the SNAP and MPQ structures are highly convergent with those of the EPQ and the EPP (Markon et al. 2005), which is not well reflected in the table. Specifically, the SNAP–MPQ Constraint (vs. Disinhibition) factor falls between the table’s Second and Third factors, just as do Eysenck’s (1997) Psychoticism and Siever and Davis’s (1991) Aggression-Impulsive. Similarly, I feel sure enough of the data to take public exception to only one placement in Table 1–3, and to offer a suggestion for Table 1–2. In Table 1–3, the parentheses around Entitlement belong under Antagonism versus Compliance, not under Extraversion versus Introversion (Clark et al., in press; Markon et al. 2005). In Table 1–2, “spontaneous” would be a good addition to the Normal list of the Simplified Version. I have some questions about a few other placements as well, but more data are needed to answer them. For a field about which many outside observers believe there is great chaos and contention, this level of agreement is nothing short of extraordinary. Much more important than any of these quibbles, however, are the realization that 1) there may be no need to shoehorn these measures into a single “three to five” domain structure and 2) the attention of the field needs now to turn to clarifying the lower levels of the hierarchy. I discuss each of these in turn. First, as Markon et al. (2005) demonstrate clearly, many of the measures in Table 1–1 fit quite neatly into a hierarchical structure with four levels of two to five factors, respectively (see Chapter 1, p. 7, for key to instrument abbreviations used here). The SNAP/MPQ/EPQ three factors spawn the four-factor DAPP-BQ/MCMI-III/

Clark: Commentary on Widiger and Simonsen

25

ZKPQ by the splitting of (EPQ) Psychoticism versus Constraint into Antagonism–Aggression and Conscientiousness versus Impulsivity. In turn, the fifth factor in five-factor models (e.g., PSY-5, NEO PI-R) emerges by (PSY-5) Psychoticism–Openness splitting off from a broader Extraversion factor. At the top of this hierarchy are Digman’s (1997) Alpha (N/NA/NE + EPQ Psychoticism vs. Constraint) and Beta (E/PA/PE + Openness) factors. Once this multilevel hierarchical structure is acknowledged, the need for Widiger and Simonsen to discuss how Eysenck’s Psychoticism and the three-dimensional models of the MPQ and SNAP relate to Antagonism versus Compliance (see Chapter 1 subsection titled “Antagonism Versus Compliance”) simply disappears. The answer is that they exist at the three-factor level, whereas Antagonism versus Compliance only emerges at the four-factor level. It is extremely important in this context not to begin the debate with whether the 3- or 4- or 5- or 20-factor model is the “right” or “true” model. It may well be that one, or some subset, of models ultimately will prove to be the most useful, for example, in terms of its reflecting most closely brain structures, genetic architecture, or cultural patterns of reinforcement. But these are larger empirical questions that cannot be resolved by debate or psychometric refinements, no matter how sophisticated. Nor can they be resolved if—either by explicit agreement or simply by default—all personality disorder research is conducted only at one level of the hierarchy. Rather, I contend that the field will be best served by a strong consensus that a relatively well-defined hierarchical structure exists, and that both existing and new measures should be examined in the context of the structure as a whole—unless, of course, the research is designed explicitly to test the robustness of the structure itself, either internally or in relation to a clear external set of correlates. We now have enough data to be reasonably sure that working within this broad structure will not lead us far astray for the foreseeable future, and that much is to be gained by everyone pulling in the same general direction with an interrelated set of tools. Accordingly, the next challenge facing the field of personality—both adaptive and maladaptive—today is not the higher-order structure, but rather clear, coherent, and consistent delineation of the lower-order structure of the model, from 6 through 20 or 30 factors. It is unlikely (and probably unnecessary as well) that every level of the structure can be spelled out with the same degree of precision that characterizes the 2- through 5-factor structures. However, for measures that fall into roughly the same level of the overall hierarchy, it would be useful to work toward a consensus set of dimensional constructs at that level. For example, the 15 scales of the SNAP might be studied conjointly with the 18 scales of the DAPP-BQ to arrive at a consensus set of lower-order constructs at this middle level of specificity. To illustrate, two DAPP-BQ scales—Diffidence and Insecure Attachment— both correlate moderately strongly with SNAP Dependency. A focused study could determine the empirical value of these scales, investigating such questions as whether SNAP Dependency lacks important content covered by one or the other

26

Dimensional Models of Personality Disorders

of the DAPP scales (and vice versa); whether the two DAPP scales have more, the same, or less predictive power than the one SNAP scale for specified relevant variables; and whether the two DAPP scales have differential correlates, justifying their existence as separate traits, or whether they each correlate similarly with the same set of variables, raising questions about the utility of having separate scales. Such investigations would clarify—and, ideally, yield a consensus regarding—each focused portion of the hierarchical structure. Furthermore, in cases such as dependency, for which the placement of the subdomain relative to the higher-order structure is unclear, such research would address this question. Moreover, this research will need to address additional questions such as those raised by Widiger and Simonsen, including whether the traits are bipolar in nature, how best to assess the continuum of traits from normality into psychopathology, and even how traits may combine to form “true” categorical entities which are more than the sum of their parts. Newer, more sophisticated methods such as item response theory (IRT) modeling and taxometrics will be of value in these investigations. The end result of such research might be 1) a single consensus set of X traits, with each instrument adding constructs it previously lacked, 2) a single consensus set of Y traits, with some realignment of the scales’ variables (e.g., the item content of the three scales Diffidence, Insecure Attachment, and Dependency might be realigned into a cleaner two-trait set), or 3) some other outcome not yet envisioned. In addition, we might expect increased clarity regarding where the resultant traits fit in the hierarchical structure and the nature of the traits, including their potential bipolarity, relations between normal and abnormal variants, and the possibility of emergent categories. Widiger and Simonsen suggest that a future DSM would provide a set of lowerorder traits such as those just discussed, and that the selection of traits to be included would be based on “extent of overlap, adequate coverage of the domain, representation of different dimensional models, clinical relevance, familiarity, and ease of use” (Chapter 1, p. 10). This implies that a future committee would be choosing from among a cacophony of extant variables and measures. What I am suggesting here is that determining an optimal set of lower-order traits, such as those shown in Widiger and Simonsen’s Tables 1–2 and 1–3, for inclusion in a future DSM is work that is too important to leave to a selection committee. Rather, we must take responsibility as researchers to provide the committee with sufficient data to determine which traits to include in the next DSM. Widiger and Simonsen have started us in the right direction. We must now take up the baton and do our part.

References Clark LA, Simms LJ, Wu KD, Casillas A: Schedule for Nonadaptive and Adaptive Personality. Minneapolis, MN, University of Minnesota Press, in press

Clark: Commentary on Widiger and Simonsen

27

Digman JM: Higher-order factors of the Big Five. J Pers Soc Psychol 73:1246–1256, 1997 Eysenck HJ: Personality and experimental psychology: the unification of psychology and the possibility of a paradigm. J Pers Soc Psychol 73:1224–1237, 1997 Krueger RF, Tackett JL: Personality and psychopathology: working toward the bigger picture. J Personal Disord 17:109–128, 2003 Markon KE, Krueger RF, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991

This page intentionally left blank

3 COMMENTARY ON WIDIGER AND SIMONSEN Working Out a Dimensional Framework John M. Oldham, M.D.

The paper by Widiger and Simonsen is a careful, elegant, and thoughtful review of 1) proposed modifications of the current DSM diagnostic system for the personality disorders, 2) potential ways to dimensionally reorganize personality disorder symptoms, 3) the relative merits of clinical spectra models, and 4) dimensional models that encompass general personality functioning. The authors propose four to five broad domains of personality functioning, emphasizing 1) Extraversion versus Introversion, 2) Antagonism versus Compliance, 3) Constraint versus Impulsivity, 4) Emotional Dysregulation versus Emotional Stability, and possibly 5) Unconventionality versus Closedness to Experience. Each domain is then displayed as a “bipolar system,” implying a continuum with normal personality traits in the midzone and pathology on the high and low ends of each continuum; estimates of which assessment instruments best capture which parts of each continuum are presented. Presumably, what would be measured would be behaviorally specific criteria reflecting an attempt to quantify the amount of a specific personality trait present or absent, which, if “too much” or “too little,” would define a component of pathology. For clinicians, some DSM-IV-TR (American Psychiatric Association 2000) personality disorder categories might be rather easily reconceptualized using one of the five domains proposed by Widiger and Simonsen. For example, schizoid personality disorder might naturally be understood as a condition characterized by 29

30

Dimensional Models of Personality Disorders

a high degree of introversion. Other familiar conditions, such as borderline personality disorder, might need to be “unpacked” into subtypes, such as one subtype characterized by impulsivity and another subtype characterized by emotional dysregulation. In other cases, such as narcissistic personality disorder, it is not clear how current DSM-defined categories could be translated into this dimensional model, or even whether there are sufficient data supporting its construct validity to justify doing so. Relying solely on an evidence base to establish validity is complicated, however, since research data that inform us about personality pathology are mostly derived from clinical populations. It may be misleading, therefore, to judge the relative “validity” of a given DSM-defined personality disorder based on the number of published research studies and clinical case reports on the disorder, since the volume of published studies in the literature may reflect who comes for treatment, or who comes to public or institutional attention. Borderline personality disorder, by this measure, would naturally be the subject of many studies, since these patients are highly prevalent within clinical populations. Patients with antisocial personality disorder, though not necessarily high on the treatment-seeking list, contribute quite significantly to social cost and burden, and they populate correctional systems at great expense. However, patients with paranoid or schizoid personality disorder may avoid treatment settings, even when significantly symptomatic, by the very nature of their conditions, as could be the case for other personality disorders such as dependent or obsessive-compulsive. In other cases, features of given personality disorders may be widely accepted as “facts,” even though the data vary greatly depending on the population studied. Borderline personality disorder, for example, is frequently described as a condition more common in women than in men (Gunderson 2001), whereas population-based studies report even distribution of the disorder between women and men (Torgersen 2001). We know that DSM-defined borderline personality disorder is an enormously heterogeneous category, perhaps partially explaining the apparent discrepancy among reported borderline personality disorder populations, and patterns of comorbidity could vary as well—high percentages of women with borderline personality disorder might have comorbid major depressive disorder and might be likely to seek treatment, whereas high percentages of men with borderline personality disorder might have comorbid antisocial personality disorder, and their more characteristic impulsivity could land them in correctional settings more often than treatment settings. It is important to underscore, as the authors point out, that DSM-IV-TR is not a textbook about the nonpathological biology and psychology of motivated human behavior. Instead, it is used to define illness (i.e., abnormality), which then presumably guides treatment planning. Practically and inevitably, as well, it also guides health policy, coding, reimbursement, and legal and forensic arguments. At least some of the DSM-defined categories of personality disorders—such as paranoid personality disorder, schizoid personality disorder, antisocial personality dis-

Oldham: Commentary on Widiger and Simonsen

31

order, and obsessive-compulsive personality disorder—have been stable and persistent, appearing in every edition of the DSM, perhaps reflecting a longitudinal clinical consensus that conveys some sense of “real world” validity. Other personality disorders have been proposed and later have been judged to be invalid (e.g., aesthenic personality disorder) or have been reconceptualized as Axis I disorders (e.g., cyclothymic disorder), and still others have been introduced in one of the later editions of the DSM (e.g., borderline personality disorder). One could argue that this process of reevaluation and revision is appropriate and is mirrored by a similar process for Axis I conditions. Whatever diagnostic system for the personality disorders eventually achieves consensus, there will still need to be a way to define the threshold or “cut point” that differentiates normal traits from pathology, and this applies for dimensional systems as well (Rothschild et al. 2003). Such a differentiation would have to be worked out if a system similar to that proposed by Widiger and Simonsen were adopted. For example, how much or how little of a particular trait on a continuum would define pathology? Would a hypothetical individual be judged to have “antagonistic personality disorder” if above and “compliant personality disorder” if below particular cut points on an Antagonism-versus-Dependency scale? If so, what evidence would guide such delineation? If, instead of radical revision, a modification of the current DSM-IV-TR system were introduced, a dimensional framework could be worked out (Oldham and Skodol 2000) identifying a designated number of criteria to define trait, subthreshold, threshold, or prototypic levels of a given personality disorder, as Widiger (1993) and Widiger and Sanderson (1995) previously proposed. Finally, new data are raising questions about the long-term stability of the personality disorders (Grilo et al. 2004; Paris and Zweig-Frank 2000; Zanarini et al. 2003), perhaps challenging the current inclusion of “stable and of long duration” as one of the DSM-IV-TR generic defining features for all personality disorders. The now clearly established fluctuation in reported levels of symptoms over time in patients with personality disorders might argue in favor of a dimensional, continuum model, to some degree analogous to the exacerbations and remissions of Axis I disorders. One could still use a stratified, combined, categorical/dimensional system, but the lack of longitudinal stability may argue against a fundamental difference between Axis I and Axis II disorders.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Grilo C, Sanislow CA, Gunderson JG, et al: Two-year stability and change of schizotypal, borderline, avoidant and obsessive-compulsive personality disorders. J Consult Clin Psychol 72:767–775, 2004

32

Dimensional Models of Personality Disorders

Gunderson JG: Borderline Personality Disorder: A Clinical Guide. Washington, DC, American Psychiatric Publishing, 2001 Oldham JM, Skodol AE: Charting the future of Axis II. J Pers Disord 14:17–29, 2000 Paris J, Zweig-Frank H: A 27-year follow-up of patients with borderline personality disorders. Psychiatr Q 71:291–307, 2000 Rothschild L, Cleland C, Haslam N, et al: A taxometric study of borderline personality disorder. J Abnorm Psychol 112:657–666, 2003 Torgersen S, Kringlen E, Cramer V: The prevalence of personality disorders in a community sample. Arch Gen Psychiatry 58:590–596, 2001 Widiger TA: The DSM-III-R categorical personality disorder diagnoses: a critique and an alternative. Psychol Inq 4:75–90, 1993 Widiger TA, Sanderson CJ: Toward a dimensional model of personality disorder, in The DSM-IV Personality Disorders. Edited by Livesley WJ. New York, NY, Guilford, 1995, pp 433–458 Zanarini MC, Frankenburg FR, Hennen J, et al: The longitudinal course of borderline psychopathology: 6-year prospective follow-up of the phenomenology of borderline personality disorder. Am J Psychiatry 160:274–283, 2003

4 COMMENTARY ON WIDIGER AND SIMONSEN From ICD-10 and DSM-IV to ICD-11 and DSM-V Charles B. Pull, M.D., Ph.D.

As a member of the overall World Health Organization (WHO) Task Force that coordinated the development of ICD-10 (World Health Organization 1992), the author of this discussant paper participated in many of the meetings that were convened to compare the successive drafts developed for the classification and description of personality disorders in ICD-10 and DSM-IV (American Psychiatric Association 1994). It became rapidly apparent to the members of the ICD-10 and the DSM-IV Task Forces that there was little evidence to rely on for making fundamental decisions, e.g., for deciding what categories of personality disorders to include in the ICD-10 or DSM-IV, or for choosing the number and content of individual diagnostic criteria defining each individual disorder. As such, members of the two tasks forces had to rely heavily on considerations that were based, not on scientific evidence, but on the traditions of various schools of psychiatry or psychology, as well as on the expert opinion of specialists. Many of the final decisions in either system were, in fact, consensus decisions taken by the respective ICD-10 or DSM-IV committees. The main data that were used to finalize the ICD-10 criteria sets came from the results of the International Pilot Study on Personality Disorders (IPSPD), which included centers from 11 countries in North America, Europe, Africa, and 33

34

Dimensional Models of Personality Disorders

Asia (Loranger et al. 1994; World Health Organization 1997). In the IPSPD, personality disorders were assessed with the International Personality Disorder Examination (IPDE). The IPDE is a semistructured diagnostic interview that contains a prescribed set of carefully selected and researched questions for the assessment of each criterion provided for each personality disorder in either system (draft criteria of ICD-10 and DSM-III-R [American Psychiatric Association 1987] criteria in the original version, final ICD-10 and DSM-IV criteria in the current version of the instrument). Although the principal objectives of the study were to determine the cultural applicability, user friendliness, and interrater reliability of the instrument, the results also provided information on the interrater reliability and temporal stability of personality disorder diagnoses, on the co-occurrence of personality disorders in the same patients, and on the overlap of ICD-10 and DSM-IV diagnoses. The results of the IPSPD revealed major difficulties in the diagnosis of personality disorders in both ICD-10 and DSM-III-R. First, administration of the IPDE took a long time (usually between 2 and 3 hours). Second, interrater agreement and temporal stability of IPDE diagnoses were far from perfect. For interrater agreement, the overall weighted kappa values for definite and definite/probable diagnoses were .57 and .69 in DSM-III-R and .65 and .72 in ICD-10, respectively. For temporal stability, the overall weighted kappa values for the definite and definite/probable diagnoses were .50 and .53 in DSMIII-R and .54 and .53 in ICD-10, respectively. Third, a substantial number of patients had more than one diagnosis of personality disorder. Of the patients with a DSM-III-R personality disorder diagnosis, 30.3% had more than one personality disorder, including 15.0% with two, 8.7% with three, and 6.6% with more than three disorders. Of the patients with an ICD10 personality disorder diagnosis, 33.9% had more than one type of disorder, including 20.1% with two, 9.5% with three, and 4.2% with more than one disorder. Fourth, there was substantial disagreement regarding the cases of personality disorders identified by ICD-10 and DSM-III-R. The question of whether the two classification systems identified the same patients as having a particular disorder was addressed by calculating an overall weighted kappa based on all of the disorders, regardless of whether they met the criterion of a 5% base rate. The kappa was .54, an indication of moderate agreement. The results of the IPSPD led to a number of modifications in the criteria that were retained for the diagnosis of the different personality disorders in ICD-10 and DSM-IV, concerning in particular the thresholds required for a positive diagnosis in either system, and the content or formulation of several of the initial criteria. There remained, however, substantial differences between the final ICD-10 and DSM-IV criteria for most of the personality disorders in either system (Pull and Pull 2002).

Pull: Commentary on Widiger and Simonsen

35

In some instances, the differences are such that the diagnosis of a given personality disorder may be met according to the criteria in one of the systems, while none of the criteria from the other system are met for the same disorder. An example of this extreme situation is provided by paranoid personality disorder (Table 4–1). In both ICD-10 and DSM-IV, a diagnosis of paranoid personality disorder can be made when (at least) 4 out of 7 criteria are met. Four of the ICD-10 criteria (criteria 1, 4, 6, and 7) are not part of the DSM-IV criteria, and in the same way, four of the DSM-IV criteria (criteria 2, 3, 4, and 6) are not part of the ICD-10 criteria. Since a diagnosis of paranoid personality disorder can be made when only ICD-10 criteria 1, 4, 6, and 7 (or DSM-IV criteria 2, 3, 4, and 6) are met, it is quite possible (at least on a theoretical level), to make a diagnosis of the disorder according to the criteria in either of the systems, even when no single criterion for the same disorder is present from the other. Similar considerations apply to the criteria of dissocial (ICD-10) and antisocial (DSM-IV) personality disorder. Other major differences between ICD-10 and DSM-IV personality disorders include discrepancies in the criteria for anankastic (ICD-10) and obsessive-compulsive (DSM-IV), anxious (ICD-10), and avoidant (DSM-IV) and dependent personality disorders. All along, the members of both the ICD-10 and the DSM-IV task forces were fully aware of the existence of dimensional models of personality disorder and, over the years when the two systems were developed, the option to replace diagnostic categories by dimensions was discussed many times. In both groups, however, the general opinion prevailed that it was too early to make this kind of radical change at that time. The diagnostic criteria defining the various types of personality disorder in the current classification systems translate enduring patterns of thinking, feeling, and behaving. According to the dimensional model, personality disorders represent maladaptive variants of personality traits. While it would be difficult to accommodate criteria defining patterns of behaviors within a dimensional model, most other current diagnostic criteria in ICD-10 and DSM-IV-TR could in fact easily be viewed as combinations of abnormal high or abnormal low traits or facets within a small number of fundamental dimensions or domains, such as those that are identified in the five-factor model (Widiger and Costa 2002). The masterly and comprehensive review by Drs. Simonsen and Widiger show that the field has considerably evolved during the past decade and that significant advances have been made in the development of dimensional models of personality as well as of personality disorder. Drs. Widiger and Simonsen have described 18 alternative proposals and suggested, quite convincingly, that these alternative models might be integrated within a common hierarchical structure. In fact, they provide an illustration of how this could be accomplished. Considering the data that Drs. Widiger and Simonsen have presented, the time may now be ripe to make those radical changes that could not be made a decade ago. The task ahead is obviously a formidable challenge, but certainly worth the try.

36

Dimensional Models of Personality Disorders

TABLE 4–1.

A comparison of ICD-10 and DSM-IV criteria for paranoid personality disorder ICD-10 criteria

DSM-IV criteria

At least four of the following must Four or more of the following: be present: 1. excessive sensitivity to setbacks _____ and rebuffs 2. tendency to bear grudges 5. persistently bears grudges, i.e., persistently, e.g., refusal to forgive is unforgiving of insults, insults, injuries, or slights injuries, or slights 3. suspiciousness and a pervasive 1. suspects, without sufficient tendency to distort experience by basis, that others are misconstruing the neutral or friendly exploiting, harming, or actions of others as hostile or deceiving him or her contemptuous 4. a combative and tenacious sense of _____ personal rights out of keeping with the actual situation 5. recurrent suspicions, without 7. has recurrent suspicions, justification, regarding sexual fidelity without justification, of spouse or sexual partner regarding fidelity of spouse or sexual partner 6. persistent self-referential attitude, _____ associated particularly with excessive self-importance 7. preoccupation with unsubstantiated _____ “conspiratorial” explanations of events either immediate to the patient or in the world at large _____ 2. is preoccupied with unjustified doubts about the loyalty or trustworthiness of friends or associates _____

3. is reluctant to confide in others because of unwarranted fear that the information will be used maliciously against him or her

37

Pull: Commentary on Widiger and Simonsen

TABLE 4–1.

A comparison of ICD-10 and DSM-IV criteria for paranoid personality disorder (continued) ICD-10 criteria

DSM-IV criteria _____

_____

4. reads hidden demeaning or threatening meanings into benign remarks or events 6. perceives attacks on his or her character or reputation that are not apparent to others and is quick to react angrily or to counterattack

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC, American Psychiatric Association, 1987 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Loranger AW, Sartorius N, Andreoli A, et al: The International Personality Disorder Examination: the World Health Organization/Alcohol, Drug Abuse, and Mental Health Administration International Pilot Study of Personality Disorders. Arch Gen Psychiatry 51:215–224, 1994 Pull CB, Pull MC. Conceptions typologiques des troubles de la personnalité: critères diagnostiques des troubles de la personnalité, in Les troubles de la personnalité. Edited by Féline A, Guelfi JD, Hardy P. Paris, France, Médecine-Sciences Flammarion, 2002, pp 81–97 Widiger TA, Costa PT: Five factor model personality disorder research, in Personality Disorders and the Five Factor Model of Personality, 2nd Edition. Edited by Costa PT, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 59–87 World Health Organization: The ICD-10 Classification of Mental and Behavioural Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland, World Health Organization, 1992 World Health Organization: Assessment and Diagnosis of Personality Disorders: The ICD10 International Personality Disorder Examination (IPDE). Edited by Loranger AW, Janca A, Sartorius N. Cambridge, UK, Cambridge University Press, 1997

This page intentionally left blank

5 BEHAVIORAL AND MOLECULAR GENETIC CONTRIBUTIONS TO A DIMENSIONAL CLASSIFICATION OF PERSONALITY DISORDER W. John Livesley, M.D., Ph.D.

The central argument advanced in this chapter is that it is feasible to construct an etiologically informed dimensional classification of personality disorder that addresses many of the limitations of current categorical systems. The evidence shows that personality disorders are best represented by behavioral continua continuous with normal personality variation and that current diagnoses are not natural kinds that “carve nature at its joints” but rather artifactual kinds—contrived constructs used to organize clinical information. When constructing a dimensional alternative, the challenge is to construct a system based on natural kinds. Accumulating knowledge on the genetic etiology of personality disorder makes this an achievable

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 2, pages 131–155, 2005).

39

40

Dimensional Models of Personality Disorders

goal. The intent of this chapter is not to review the genetics of personality in detail but to consider how genetic research may help taxonomic endeavors. For this reason, the chapter begins by briefly considering the classification of personality disorder and what is needed to build a valid system.

The Classification of Personality Disorder A dimensional classification requires two components (Livesley 2003). First, a definition of personality disorder is needed to diagnose the disorder because extreme variation alone does not necessarily imply disordered functioning (Wakefield 1992). Second, constructs are needed to represent individual differences in personality pathology. Ideally, the definition of personality disorder should be based on an understanding of disturbances in normal personality functioning. Consequently, the category of personality disorder is an artifactual kind—a construction for organizing information about mental disorders. Hence, genetic research probably has little to contribute to its definition. Genetic information can, however, make a substantial contribution to delineating the constructs needed to represent individual differences. Most models of individual differences assume that traits are hierarchically organized with secondary traits subdividing into multiple primary traits. Common models of personality such as Eysenck’s three-component model as assessed by the Eysenck Personality Questionnaire (EPQ; Eysenck and Eysenck 1992) and the five-factor model as assessed with the Neuroticism–Extraversion–Openness Personality Inventory—Revised (NEO PI-R; Costa and McCrae 1992) differ in the number and nature of the constructs used to represent personality. Nevertheless, there is considerable agreement that four secondary dimensions underlie personality disorder (Mulder and Joyce 1997). These dimensions may be labeled: anxious-submissive, psychopathic, socially withdrawn, and compulsive. They resemble the five-factor dimensions of neuroticism, (dis)agreeableness, introversion, and conscientiousness (clinical studies consistently fail to find a component corresponding to openness) and Eysenck’s neuroticism, extraversion, and psychoticism dimensions. The difference is that Eysenck considers compulsivity to be a facet of psychoticism rather than a secondary domain (Eysenck 1991, 1992). Despite this convergence, debate continues on the number of factors needed to represent broad differences in personality. Although this issue is relatively unimportant for many research purposes—because it is often useful to vary level of analysis according to the research question under investigation—it is important when constructing a dimensional classification because secondary traits are a useful way to organize personality descriptors. For this reason, we need to resolve discrepancies across models. Disagreement stems partly from the failure to differentiate secondary and primary traits. What is a primary trait to one theorist is a

Livesley: Behavioral and Molecular Genetic Contributions

41

secondary trait to others, as illustrated by the disagreement between Costa and McCrae and Eysenck over whether conscientiousness is a secondary domain or merely a facet of psychoticism. Moreover, broad agreement on the secondary dimensions masks disagreement on the primary traits defining each domain. For example, there is agreement that sociability and affiliation are major components of extraversion but not on the importance of agency, activation, impulsivity, sensation seeking, positive emotions, or optimism (Depue and Collins 1999). Similar problems occur with other secondary traits (e.g., whether impulsivity is a feature of neuroticism or of psychoticism). To construct a dimensional classification that integrates current models, we need to specify the difference between secondary and primary traits, determine the number and content of secondary domains, and develop a procedure for identifying and defining primary traits. The failure of standard psychometric studies to resolve these issues suggests that psychometric criteria alone may be insufficient for this purpose (Eysenck 1991, 1992). Hence, it seems appropriate to consider whether a genetic perspective can clarify the structure of personality and whether it is possible to construct what Tsuang and colleagues (1993) referred to as a “genetically informed nosology.” The foundations for such a system would be provided by evidence that 1) genetic factors have an extensive influence on personality disorder; 2) behaviors used to classify individual differences arise primarily from genetic influences and that the environment does not produce new personality structures, or at least not structures that are pertinent to classification; and 3) the classification incorporates phenotypes that reflect the etiological structure of personality disorder.

Genetic Approaches to Personality Considerable progress has been made recently in molecular genetic and behavioral genetic research on personality. The search for personality genes was stimulated by Cloninger’s (1987; Cloninger et al. 1993) innovative hypotheses about the neurotransmitters associated with personality including the putative association between novelty seeking and dopamine. The dopamine receptor DRD4 was known to exist in short and long forms with the shorter form coding for a receptor that is more efficient in binding dopamine. Cloninger and colleagues (1996) hypothesized that individuals with the long allele seek novelty to increase dopamine release. Consistent with the hypothesis, the initial reports showed that novelty-seeking scores were significantly higher in individuals with the long form of the allele (Benjamin et al. 1996; Ebstein et al. 1996). Unfortunately, subsequent studies yielded inconsistent results and meta-analyses concluded that the mean association was nonsignificant and similar in magnitude to the association with other temperament scales, namely, harm avoidance and reward dependence (Kluger et al. 2002; Schinka et al. 2002). As a result, Kluger and colleagues suggested that “researchers in the field have to

42

Dimensional Models of Personality Disorders

assess whether or not variable small true association between DRD4 and NS is theoretically meaningful” (Kluger et al. 2002, p. 715). Studies using other measures and candidate genes also produced conflicting results. Despite extensive investigation of associations between polymorphisms and personality, meta-analyses find scant evidence of significant effects (Munafò et al. 2003). Whatever relationships exist, most effects are probably small and relatively nonspecific. Future classifications may well be profoundly influenced by specification of the genes of personality. However, the absence of consistent findings limits the immediate nosological value of molecular genetic studies. A second important development in molecular genetic studies of personality is work on gene–environment interaction in the development of antisocial behavior by Caspi and colleagues (2002). They noted that although in boys, childhood maltreatment is a major risk factor for conduct disorder, antisocial behavior, and violent offending, not all maltreated children develop antisocial behavior. Based on observations that the monoamine oxidase A gene (MAOA) is associated with aggressive behavior, they hypothesized that the MAOA genotype can modulate the influence of maltreatment. In a large sample, they found that maltreated children with the genotype that conferred high levels of MAOA expression were less likely to develop antisocial behavior. Identification of a gene that modulates the effects of developmental adversity suggests that future classifications of mental disorders will probably need a provision for coding genotypes that are associated with increased likelihood of disorder and those that are protective. Indeed, Gottesman (2002) recently envisioned a sixth axis for DSM-V to record relevant genotypic information. In contrast to molecular studies, behavioral genetic research is beginning to provide the information needed to construct a dimensional classification of personality disorder. Behavioral genetics has progressed beyond simple heritability analyses that estimate the magnitude of genetic and environmental influences on personality to the use of more sophisticated multivariate genetic techniques that permit exploration of the etiological factors responsible for observed patterns of trait covariation. Although it has been suggested that the advent of molecular genetics has rendered behavioral genetics less useful because behavioral genetics can only infer genetic effects using well-established statistical methods whereas molecular genetics specify the actual genes associated with phenotypic variability (Faraone et al. 1999), behavioral genetic methods can play a useful role in clarifying personality phenotypes.

Genetic Structure of Personality Behavioral genetic research provides convincing evidence of extensive genetic influences on individual differences in normal and disordered personality. Heritability is typically estimated in the 40%–60% range and environmental influences are largely confined to nonshared effects. All traits are subject to genetic influences, nonherit-

Livesley: Behavioral and Molecular Genetic Contributions

43

able traits have not been identified (Plomin et al. 1990), heritability does not differ significantly across traits, and heritability estimates are not appreciably influenced by method of measurement: the results are similar when based on ratings by cotwins (Heath et al. 1992) and peers (Riemann et al. 1997). Estimates of the broad heritability of Eysenck’s dimensions of neuroticism, extraversion, and psychoticism are 36%, 53%, and 39%, respectively (Tambs et al. 1991). Studies of the NEO PIR yielded values of 41%, 55%, 58%, 41%, and 37% for neuroticism, extraversion, openness, agreeableness, and conscientiousness, respectively (Jang et al. 1996). The temperament and character domains of Cloninger’s Temperament and Character Inventory (TCI; Cloninger et al. 1994) are equally heritable with estimates ranging from 49% for self-directedness to 34% for novelty seeking (Ando et al. 2004). Similarly, Torgersen and colleagues (2000) reported that the heritability of DSM-III-R (American Psychiatric Association 1987) personality disorders ranged from 79% to 28%. The heritability of secondary factors of personality disorder traits assessed with the DAPP were 53%, 50%, 51%, and 38% for emotional dysregulation, dissocial behavior, inhibitedness, and compulsivity, respectively, and the heritability of the 18 basic traits ranged from 56% to 35% (Jang et al. 1996; Livesley et al. 1993). Heritability estimates are similar throughout the trait hierarchy: 24 of the 30 NEO PI-R facets are heritable (Jang et al. 1996) as are the 24 facets of the TCI (Ando et al. 2004) and all but 3 of the 69 of the subtraits defining the 18 basic DAPP-BQ dimensions (Jang et al. 1996). Evidence that all personality disorder traits are heritable provides the basis for a genetically informed nosology: evidence that some traits were forged by experience would reduce the taxonomic value of genetic information. However, the results of heritability studies have limited value in organizing a classification because they only provide information on relative contributions of genes and environment to a given trait. They do not provide information on whether the trait is subject to a single or multiple sources of genetic influence or whether genetic influences are specific to that trait or shared with other traits. This information, which is needed to organize traits on the basis of a shared etiology, is provided by multivariate genetic analyses that explore the etiology of the covariance structure underlying multiple traits at different levels of the trait hierarchy (Carey and DiLalla 1994). Multivariate genetic analyses extend univariate analysis of etiological influences on a single trait to estimate genetic and environmental contributions of the covariation between two or more traits. The degree to which genetic and environmental effects on two variables are correlated is indexed by genetic and environmental correlation coefficients. Genetic and environmental correlation or covariance matrices may then be factored to provide information on underlying structures (Crawford and DeFries 1978). The approach permits estimation of separate genetic and environmental common and specific factor loadings. Information about common sources of genetic variation underlying a set of traits is useful in clarifying the secondary structure of personality and information about specific sources of variance

44

Dimensional Models of Personality Disorders

should help to clarify the relationship between secondary and primary traits and determine whether primary traits are etiologically distinct from secondary traits. Multivariate analyses are also useful in exploring the relationship between the phenotypic and etiological structure of personality. A crucial question when developing a dimensional classification is the extent to which trait structure reflects an underlying biological structure. This question may be explored by comparing matrices of genetic and environmental correlations with a matrix of observed or phenotypic correlations computed between subscales of a personality measure. High congruency between genetic and observed structures permits the use of genetic information to help resolve seemingly intractable problems of the number and contents of secondary domains and develop a classification based on natural kinds.

Genetic Contributions to Trait Covariance The evidence indicates substantial convergence between the phenotypic and genetic structure of personality. In an early investigation of the genetic structure underlying a set of personality traits, Loehlin (1987) identified four factors from the genetic correlation matrix computed for item clusters from the California Psychological Inventory that resembled Neuroticism, Extraversion, Openness, and Conscientiousness of the five-factor model. The apparent congruence between the phenotypic and genetic structure of the five-factor model was subsequently confirmed using the NEO PI-R. Jang and colleagues (2002) reported that analysis of the genetic covariance matrix for NEO PI-R facets yielded five factors. Congruence coefficients computed between genetic factors and published normative structure ranged from .92 to 70. A similar study by Ando and colleagues obtained congruence coefficients greater than .95 (J. Ando, S. Yamagata, A. Suzuki, et al.: “Cross-National Generalizability of Personality Trait Structure: Psychometric and Biometrical Support From Europe, Asia, and North America” [unpublished manuscript]). These values were probably higher than those in the previous study because genetic structures were compared with phenotypic structures based on the study samples rather than normative structure. There is also high congruence between genetic and phenotypic factor structures of the DAPP-BQ (Livesley et al. 1998): congruence coefficients were .97, .97, .98, and .95 for emotional dysregulation, dissocial, inhibition, and compulsivity, respectively. The TCI is an exception to this pattern: evaluations of the extent to which the seven dimensional phenotypic structure of the TCI reflects the genetic architecture of personality have yielded mixed results. Genetic analyses of the seven dimensions of the TCI show considerable overlap among them. The temperament and character traits are not as etiologically distinct as the theory postulates: temperament accounts for 26%, 37%, and 10% of additive genetic variance in the character dimensions of self-directedness, cooperativeness, and self-transcendence,

Livesley: Behavioral and Molecular Genetic Contributions

45

respectively (Gillespie et al. 2003). Also, the assumption that temperament traits are based on independent genetic factors was not supported for reward dependence—28% of the variance in this dimension was explained by other temperament scales and the genetic correlation with novelty seeking was .42. Despite these findings, Gillespie and colleagues concluded that the seven component structure is justified by evidence of considerable specific genetic variance for all dimensions. This conclusion differs from that of Ando and colleagues (2002). Unlike Gillespie et al., they examined the genetic structure of the TCI facets scales employing the methods used to examine the structure of the NEO-PI-R and DAPPBQ. A five-factor model provided the best fit. A subsequent study using a larger sample (Ando et al. 2004) yielded a four-component structure in which temperament and character scales often loaded on the same component. Ando and colleagues (2004) used these results to reorganize the temperament scales to provide a better match to genetic structure. Genetic intercorrelations among the revised scales were small, ranging from .02 to –.23. The lack of congruence between genetic and hypothesized phenotypic structure of the TCI probably occurs because the phenotypic structure of TCI is theoretically rather than empirically derived. The phenotypic analyses of the TCI do not support the hypothetical seven-factor structure (Herbst et al. 2000) and suggest that the seven dimensions fit the fivefactor framework (De Fruyt et al. 2000). Overall, the evidence suggests high correspondence between genetic and empirically derived phenotypic factor structures—a finding with important implications for classification.

Nosological Implications of Phenotype– Genotype Congruence The higher-order genetic analyses of normal and disordered personality traits suggest that a few general genetic factors account for observed trait covariation and that personality is subject to extensive pleiotropic effects (a single genetic entity influences distinct phenotypes). The high genotype–phenotype correspondence for personality disorder contrasts with the poor genotype–phenotype correspondence observed with many psychiatric disorders (Merikangas 2002). Evidence that trait structure primarily reflects genetic influences forms an additional component to the foundation for a genetically informed classification. Although environmental influences on personality traits are similar in magnitude to genetic influences, the congruence of genetic and phenotypic structures suggests that these effects do not change the structure of trait covariation. Instead, environmental events and the extensive interplay between genes and environment probably consolidate pleiotropic effects. Since environmental influences do not give rise to new traits or substantially change trait structure, they have fewer implications for conceptualizing this level of a personality nosology.

46

Dimensional Models of Personality Disorders

When discussing the potential contributions of genetic research for psychiatric classification, Merikangas (2002) suggested that phenomena such as pleiotropy compromise the nosological value of genetic findings. While this may be the case with disorders associated with genes of large effect, with personality disorder, pleiotropic effects may help to organize the secondary structure of the system along etiological lines and resolve persistence problems with domain definition.

Defining Secondary Traits A few general genetic factors may account for the secondary structure of personality. This would suggest that a secondary domain could be defined as a cluster of primary traits that share a common genetic influence that is distinct from the general genetic factors influencing other secondary domains (Livesley et al. 2003). Defined in this way, the secondary domains would “carve nature at its joints” because they involve etiologically and functionally distinct behaviors. This does not mean that the secondary domains are distinct in the sense that a given individual will only exhibit pathology in a single secondary domain. Rather, they are distinct in the sense that respiratory and cardiac disorders are distinct—they involve anatomically and functionally separate systems even though they may show symptomatic overlap and co-occur. With this approach, the goal would be to define a set of secondary domains with minimal genetic intercorrelations. Ando and colleagues’ (2004) reorganization of the TCI to yield scales with genetic intercorrelations ranging from .02 to –.23 and the low genetic correlations among DAPP secondary domains (range: –.10 to .21 before scale modification using genetic information) suggest that this is an achievable goal. This definition of a secondary domain provides a genetic criterion to supplement the usual statistical criteria to determine the number and content of secondary constructs. The secondary structure of a classification would be determined by the number of general genetic factors required to account for the variation among a comprehensive set of primary traits representing individual differences in personality pathology. The location of a given primary trait within the trait hierarchy would be determined by its loadings on common genetic factors and the pattern of genetic correlations with other primary traits. For example, disagreement persists as to whether sensation seeking and impulsivity are distinct traits belonging to the separate domains of neuroticism and extraversion as in the NEO PI-R (Costa and McCrae 1992) or whether they are related and part of the psychoticism/dissocial domain (Eysenck 1992; Livesley et al. 1998). These would be resolved by examining the genetic correlations between these traits, their genetic correlations with the primary traits defining these domains, and their loadings on the respective general genetic factors.

Livesley: Behavioral and Molecular Genetic Contributions

47

Genetic Influences on Primary Traits Identification of a few general genetic factors that account for trait covariation raises questions about whether these factors also explain all sources of genetic influence. This in turn raises questions about the etiology of primary traits and their role in models of normal and disordered personality. Originally, primary traits were thought to be heritable simply because they are components of secondary traits that are heritable (Loehlin 1982). This assumption is questioned by the finding that basic traits have a specific heritable component when the effects of higherorder factors are partialled out and standard heritability analyses are applied to the residuals (Jang et al. 1998; Livesley et al. 1998). For example, 11 of 18 traits assessed by the DAPP-BQ had substantial residual heritability ranging from .48 for conduct problems to .26 for intimacy problems (Livesley et al. 1998). These findings were confirmed with multivariate genetic analyses of the subscales defining the 18 basic dimensions of the DAPP-BQ (Livesley et al. 2003). For 12 scales, a single common genetic factor was identified in addition to the general genetic factor that contributes to trait covariation. The remaining scales were influenced by two or three factors. The 30 NEO-PI-R facet scales showed even greater evidence of specificity of genetic influences, with 26 scales showing residual heritability ranging from .21 to .37 (Jang et al. 1998). The specific nature of genetic influences emphasizes the importance of primary traits for a dimensional classification. A set of precisely defined primary traits would facilitate biological research by providing more homogeneous targets than the traits that have been the focus of attention. It would also enhance clinical utility by increasing the value of the classification in treatment planning (Livesley 2003). Psychosocial and pharmacological interventions increasingly focus on specific components of personality pathology rather than global diagnoses. This does not mean, however, that the secondary trait level is unimportant. Secondary traits impose meaningful structure on what would otherwise be merely a list of primary traits and for many purposes secondary traits are the most appropriate level of description and evaluation. However, given the importance of primary traits, greater attention needs to be given to how primary traits should be defined and assessed. Despite the etiological importance of primary traits, the establishment of a comprehensive set of primary traits has been a challenge for personality research (Costa and McCrae 1998). A behavioral genetic approach would begin by defining a primary trait as a class of behaviors that index a single, specific source of genetic variance. Personality measures are usually constructed using psychometric criteria to select items that form homogeneous scales. Incorporation of a genetic perspective into this process would also involve selecting items that index a specific genetic dimension. Two methods are available for this purpose. The first is to use multivariate genetic analysis at the item level. Heath and colleagues (Heath and Martin 1990; Heath et al. 1989a, 1989b) used this approach to analyze the genetic

48

Dimensional Models of Personality Disorders

structure of the EPQ. They showed that Neuroticism and Extraversion were influenced by common genetic and environmental factors, indicating that these scales are etiologically homogeneous. Less evidence was found for a common genetic factor for Psychoticism (Heath and Martin 1990; Heath et al. 1989a, 1989b). Items related to cruelty, suspiciousness, and hostility loaded positively on a general genetic factor, and items related to conventional behavior loaded negatively. These findings are consistent with phenotypic analyses showing that the Psychoticism scale is heterogeneous (Heath and Martin 1990; Heath et al. 1988). A second method would be to select items based on their association with the underlying genetic factor. This approach is made possible by the development of methods to estimate genetic and environmental factor scores (Sham et al. 2001; Thomis et al. 2000). With this method, factor scores would be computed for genetic factors contributing to a set of scales and correlating each item with each genetic component. For example, disagreements about the nature of impulsivity could be explored by estimating the genetics underlying a set of impulsivity scales and related measures. Item factor score correlations could then be used to select items that index a specific genetic factor. The value of this approach is that it fits well with standard item analysis: items would be selected according to their correlation with the total scale score and the genetic factor score to foster phenotypically and genetically coherent scales.

Toward a Genetically Informed Taxonomic Research Strategy A genetically informed dimensional classification for DSM-V could be developed using a “winner takes all” approach in which different models are evaluated using genetic and phenotypic criteria. The problem with this approach is that the different models not only have some features in common but also incorporate specific constructs not included in other models. None offer a comprehensive account of normal and disordered personality. Even the more comprehensive models of normal personality do not provide an empirically derived set of primary traits that capture all aspects of personality pathology that clinicians consider important. These considerations suggest that a better approach would be to construct an integrated taxonomy by using the results of genetic research to help identify similarities and resolve discrepancies across models. Although a genetically informed approach is likely to facilitate taxonomic endeavors, it needs to be considered within the context of the overall framework for developing a classification. The optimal approach probably remains the construct validation approach used in test construction (Blashfield and Livesley 1991; Livesley and Jackson 1992). Traditionally, construct validation has relied on psychometric methods to develop valid measures. Genetic methods could, however, be

Livesley: Behavioral and Molecular Genetic Contributions

49

readily incorporated into the approach. As conceptualized by Loevinger (1957), construct validity has three components. First, a theoretical classification needs to be constructed that defines diagnostic entities and items. This would be based on the findings of genetic and phenotypic analyses of normal and disordered personality with especial emphasis on ensuring comprehensive representation of clinical concepts. Agreement across models makes it possible to develop a preliminary theoretical taxonomy based on four secondary factors that combines elements of different models and provides a comprehensive representation of the domain. In addition to substantial evidential support, the four factor structure has the advantage of incorporating biological distinctions into the basic structure of the system. The structure also resembles the more prevalent and valid DSM-IV (American Psychiatric Association 1994) diagnoses of borderline, antisocial, schizoid-avoidant, and obsessive-compulsive personality disorders, thereby creating a degree of continuity with the present classification. Greater continuity may be difficult to achieve without sacrificing validity given the serious flaws with current categorical models. The next step would be to determine the primary traits that define each secondary trait. Here there are considerable divergences across models. For this reason, the initial list of primary traits should be comprehensive and incorporate considerable redundancy by including all primary or facet traits from the different models. The second step in construct validation is to demonstrate that diagnostic items combine empirically to form the diagnostic entities proposed in the theoretical taxonomy. This would involve a combination of phenotypic and genetic analyses, including genetic item analysis, of data collected within a genetically informed design such as a twin study. Third, relationships between constructs and external criteria need to be established, although the use of genetic methods would reduce the need for some forms of external validation. The three stages form an iterative process in which empirical evaluations of the theoretical structure are used to modify this structure. Successive iterations lead increasingly to a valid system.

Conclusion The essential argument in this paper is it is possible to construct an etiologically informed classification of personality disorder by incorporating behavioral genetic methods within the construct validation framework. Genetic research does, however, raise some conceptual problems that need to be addressed. One is the relative importance of secondary and primary dimensions—an issue that is basic to the organization and application of a dimensional system. Attention needs to be given to the circumstances that require an assessment at the primary as opposed to the secondary trait level. Should the diagnostic process focus on evaluating only primary traits so that secondary trait assessment is simply a summary of primary trait endorsement? Or, should the system incorporate diagnostic items to assess second-

50

Dimensional Models of Personality Disorders

ary traits directly since this more parsimonious approach may provide sufficient information for many purposes? This essay has focused almost exclusively on the implications of the findings of genetic research for classification. However, behavioral genetic research also highlights the importance of the environment. Indeed, environmental influences account for the greatest part of the variation in personality traits. The ability to estimate genetic and environmental factor scores raises the issue of whether diagnostic items should reflect genetic or environmental influences. The use of diagnostic items with high genetic loading may be preferable for molecular genetic and other biological investigations. On the other hand, items saturated with environmental variance may be more useful when investigating psychosocial factors and perhaps treatment outcomes. Although this seems a major conceptual issue, it may not be a practical problem since it seems difficult to identify items that capture only genetic or environmental variance. Finally, there is the problem of gender differences. There is good evidence of sex differences in personality disorder phenotypes and gender-specific genetic and environmental influences. It is not clear whether these differences affect the structures that would form the basis of a classification. These unresolved conceptual problems should not, however, detract from the recognition that the information is available to construct an integrated dimensional classification that reflects the phenotypic and etiological structure of personality disorder. The major constructs are well established and the procedures for integrating different systems are readily available and relatively straightforward.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC, American Psychiatric Association, 1987 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Ando J, Ono Y, Yoshimura K, et al: The genetic structure of Cloninger’s seven-factor model of temperament and character in a Japanese sample. J Pers 70:583–609, 2002 Ando J, Suzuki A, Yamagata S, et al: Genetic and environmental structure of Cloninger’s temperament and character dimensions. J Personal Disord 18: 379–393, 2004 Benjamin J, Li L, Patterson C, et al: Population and familial association between the D4 dopamine receptor gene and measures of novelty seeking. Nat Genet 12:81–84, 1996 Blashfield RK, Livesley WJ: A metaphorical analysis of psychiatric classification as a psychological test. J Abnorm Psychol 100:262–270, 1991 Carey G, DiLalla DL: Personality and psychopathology: genetic perspectives. J Abnorm Psychol 103:32–43, 1994 Caspi A, McClay J, Moffitt T: Role of genotype in the cycle of violence in maltreated children. Science 297:851–854, 2002

Livesley: Behavioral and Molecular Genetic Contributions

51

Cloninger CR: A systematic method for clinical description and classification of personality variants. Arch Gen Psychiatry 44:573–588, 1987 Cloninger CR, Svrakic DM, Pryzbeck TR: A psychobiological model of temperament and character. Arch Gen Psychiatry 50:975–990, 1993 Cloninger CR, Przybeck T, Svrakic D, et al: The Temperament and Character Inventory (TCI): A Guide to Its Development and Use. St. Louis, MO, Center for Psychobiology and Personality, Washington University, 1994 Cloninger CR, Adolfsson R, Svrakic NM: Mapping genes for human personality. Nat Genet 12:3–4, 1996 Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO-PIR) and the NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992 Costa PT Jr, McCrae RR: Six approaches to the explication of facet-level traits: examples from conscientiousness. Eur J Personality 12:117–134, 1998 Crawford CB, DeFries JC: Factor analysis of genetic and environmental correlation matrices. Multivar Behav Res 13:297–318, 1978 Depue RA, Collins PF: Neurobiology of the structure of personality: dopamine, facilitation of incentive motivation, and extraversion. Behav Brain Sci 22:491–569, 1999 Ebstein RP, Novick O, Umansky R, et al: D4DR exon III polymorphism associated with the personality trait of novelty seeking in normal human volunteers. Nat Genet 12:78–80, 1996 Eysenck HJ: Dimensions of personality: 16, 5 or 3? Criteria for a taxonomic paradigm. Pers Individ Dif 12:773–790, 1991 Eysenck HJ: A reply to Costa and McCrae: P or A and C—the role of theory. Personality and Individual Differences 13:867–868, 1992 Eysenck HJ, Eysenck SBG: Manual of the Eysenck Personality Questionnaire—Revised. San Diego, CA, Education and Industrial Testing Service, 1992 Faraone SV, Tsuang MT, Tsuang DW: Genetics of Mental Disorders. New York, NY, Guilford, 1999 De Fruyt F, van de Wiele L, van Heeringen C: Cloninger’s psychobiological model of temperament and character and the five-factor model of personality. Personality and Individual Differences 29:441–452, 2000 Gillespie NA, Cloninger CR, Heath AC, et al: The genetic and environmental relationship between Cloninger’s dimensions of temperament and character. Personality and Individual Differences35:1931–1946, 2003 Gottesman I: Defining genetically informed phenotypes for the DSM-V, in Descriptions and Prescriptions: Values, Mental Disorders, and the DSMs. Edited by Sadler JZ. Baltimore, MD, Johns Hopkins University Press, 2002, pp 291–300 Heath AC, Martin NG: Psychoticism as a dimension of personality: a multivariate genetic test of Eysenck and Eysenck’s psychoticism construct. J Pers Soc Psychol 58:111–121, 1990 Heath AC, Jardine R, Eaves LJ: The genetic structure of personality, I: phenotypic factor structure of the EPQ in an Australian sample. Personality and Individual Differences 9:59–67, 1988

52

Dimensional Models of Personality Disorders

Heath AC, Jardine R, Eaves LJ, et al: The genetic structure of personality, II: genetic item analysis of the EPQ. Personality and Individual Differences 10:615–624, 1989a Heath AC, Eaves LJ, Martin NG: The genetic structure of personality, III: multivariate genetic item analysis of the EPQ scales. Personality and Individual Differences 10:877– 888, 1989b Heath AC, Neale MC, Kessler RC, et al: Evidence for genetic influences on personality from self-reports and informant ratings. J Pers Soc Psychol 63:85–96, 1992 Herbst JH, Zonderman AB, McCrae RR, et al: Do dimensions of the temperament and character inventory map a simple genetic architecture? Evidence from molecular genetics and factor analysis. Am J Psychiatry 157:1285–1290, 2000 Jang KL, Livesley WJ, Vernon PA: Heritability of the big five personality dimensions and their facets: a twin study. J Pers 64:577–591, 1996 Jang KL, McCrae RR, Angleitner A, et al: Heritability of facet-level traits in a cross-cultural twin study: support for a hierarchical model of personality. J Pers Soc Psychol 74:1556–1565, 1998 Jang KL, Livesley WJ, Angleitner A, et al: Genetic and environmental influences on the covariance of facets defining the domains of the five factor model of personality. Personality and Individual Differences 33:83–101, 2002 Kluger AN, Siegfried Z, Ebstein RP: A meta-analysis of the association between DRD4 polymorphism and novelty seeking. Mol Psychiatry 7:712–717, 2002 Livesley WJ: Diagnostic dilemmas in the classification of personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips K, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 153–189 Livesley WJ, Jackson DN: Guidelines for developing, evaluating, and revising the classification of personality disorders. J Nerv Ment Dis 180:609–618, 1992 Livesley WJ, Jang KL, Jackson DN, Vernon PA: Genetic and environmental contributions to dimensions of personality disorder. Am J Psychiatry 150:1826–1831, 1993 Livesley WJ, Jang KL, Vernon PA: The phenotypic and genetic architecture of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Livesley WJ, Jang KL, Vernon PA: The genetic basis of personality structure, in Comprehensive Handbook of Psychology, Vol 9: Personality and Social. Edited by Millon T, Lerner MJ. New York, NY, Wiley, 2003, pp 59–83 Loehlin JC: Are personality traits differentially heritable? Behav Gen 12:417–428, 1982 Loehlin JC: Heredity, environment, and the structure of the California Psychological Inventory. Mult Behav Res 22:137–148, 1987 Loevinger J: Objective tests as instruments of psychological theory. Psychol Rep 3:635–694, 1957 Merikangas KR: Implications of genetic epidemiology for classification, in Defining Psychopathology in the 21st Century. Edited by Helzer JE, Hudziak JJ. Washington, DC, American Psychiatric Publishing, 2002, pp 185–209 Mulder RT, Joyce PR: Temperament and the structure of personality disorder symptoms. Psychol Med 27:1315–1325, 1997 Munafò MR, Clark TG, Moore LR, et al: Genetic polymorphisms and personality in healthy adults: a systematic review and meta-analysis. Mol Psychiatry 8:471–484, 2003

Livesley: Behavioral and Molecular Genetic Contributions

53

Plomin R, Chipuer HM, Loehlin JC: Behavioral genetics and personality, in Handbook of Personality: Theory and Research. Edited by Pervin LA. New York, NY, Guilford, 1990, pp 225–243 Riemann R, Angleitner A, Strelau J: Genetic and environmental influences on personality: a study of twins reared together using the self- and peer report NEO-FFI scales. J Pers 65:449–476, 1997 Schinka JA, Letsch EA, Crawford FC: DRD4 and novelty seeking: results of meta-analyses. Am J Med Genet 114:643–648, 2002 Sham PC, Sterne A, Purcell S, et al: GENESiS: creating a composite index of the vulnerability to anxiety and depression in a community-based sample of siblings. Twin Research 3:316–322, 2001 Tambs K, Sundet JM, Eaves L, et al: Pedigree analysis of Eysenck Personality Questionnaire (EPQ) scores in monozygotic twin families. Behav Genet 21:369–382, 1991 Thomis MA, Vlietnick RF, Maes HH, et al: Predictive power of individual genetic and environmental factor scores. Twin Research 3:99–108, 2000 Torgersen S, Lygren S, Oien PA, et al: A twin study of personality disorders. Compr Psychiatry 41:416–425, 2000 Tsuang MT, Faraone SV, Lyons MJ: Identification of the phenotype in psychiatric genetics. Eur Arch Psychiatry Clin Neurosci 243:131–142, 1993 Wakefield JC: The concept of mental disorder: on the boundary between biological facts and social values. Am Psychol 47:373–388, 1992

This page intentionally left blank

6 COMMENTARY ON LIVESLEY Genetic Contributions to a Dimensional Classification: Problems and Pitfalls Peter McGuffin, M.D., Ph.D.

Dr. Livesley has provided a scholarly and thoughtful account of the possible utility of molecular and behavioral genetic approaches in arriving at a new classification of personality disorder, one that is based on dimensions and takes etiological factors into account. The article assumes in the reader a fairly sophisticated understanding of quantitative genetics but does not go into detail about the complications and difficulties that arise from behavioral genetic models and the problems that may occur if some of the implicit simplifying assumptions are incorrect. This brief commentary will attempt to explicate such problems in a nontechnical way for the general reader.

Heritability as an Indicator of Validity Heritability is strictly defined as the proportion of phenotypic variance that is explained by additive genetic factors. It has been suggested in the past (e.g., Farmer et al. 1987) that finding a definition of a disease that describes an entity with high heritability may be one way of validating diagnostic criteria. Livesley points out that diagnostic systems based on high genetic loading are likely to be preferable for studies where the aim is to discover the biological or molecular basis of a disorder, but that one may also wish to explore diagnostic items that are more highly loaded on environmental influences for studies of psychosocial etiology. The inherent 55

56

Dimensional Models of Personality Disorders

limitation in estimating heritability is that the results are applicable only to a particular population at a specific time point. Although there is evidence, for example, that disorders such as schizophrenia, defined using modern criteria, are roughly as heritable in Europe as they are in Japan (Cardno and Gottesman 2000), one cannot assume that uniformity of heritability across all populations for all disorders necessarily applies. Some societies may show greater variability than others in relation to exposure to environmental factors. Societies in which there is much environmental variation will then tend to have higher phenotypic variation overall, and there will tend to be lower heritability than would be found in a society where the environmental variation is more restricted. Other complications arise because of nonadditive effects. Most of the quantitative genetic models applied in the studies reviewed by Livesley, especially those looking at multivariate effects, assume genetic additivity—that is, that genes of small effect sum together to contribute to a dimension—termed in quantitative genetics a liability—to a disease. However, genes may interact within a locus—the phenomenon of dominance—or between loci—the phenomenon called epistasis. If it is assumed that there is no dominance or epistasis when in fact these phenomena exist, heritability will tend to be overestimated from twin data. It is also generally assumed in basic twin models that genes and environment coact—that is, that they combine in a straightforward additive way. This, again, may not be the case. In particular, gene–environment interaction and gene–environment covariation may occur.

Nonadditive Gene–Environment Interplay Gene–environment interaction is present if a particular genotype confers special sensitivity to a particular environmental effect. Several lines of evidence suggest that gene–environment interaction may exist for personality traits, particularly those associated with antisocial behavior. For example, Cadoret et al. (1995) studied adoptees raised apart from parents who did or did not have a history of antisocial behavior or substance abuse. The authors found that although family disharmony predicted antisocial behavior in the adoptees, the effect was small and nearly negligible in those without a biological antisocial background and marked in those whose biological parents were themselves antisocial. A somewhat different pattern, but again one suggesting gene–environment interaction, was found recently by Button et al. (2005), who studied symptoms of conduct disorder or antisocial behavior in a population-based sample of adolescent twins. The specific environmental factor was parent-reported family disharmony. There were main effects of both family discord and genes but also evidence of a highly significant interaction between these two factors. Such studies of twins and adoptees deal with inferred, unobserved genotypes, but, as Livesley notes, studies now have looked at a specific genotype and a measured

McGuffin: Commentary on Livesley

57

environment. The specific genotype in question is monoamine oxidase A (MAOA), an X-linked gene that has a common variation in the promoter region that confers either high or low enzyme activity. Previous work on a family segregating a mutated form of MAOA that completely disrupts enzyme activity and on knockout mice lacking the MAOA gene has indicated that complete absence of MAOA activity is associated with aggressive behavior. However such mutations in humans appear to be extremely rare, and the evidence as to whether the promoter variant affects levels of aggression in humans has been inconsistent. Caspi et al. (2002) studied a cohort of men from New Zealand for whom there were excellent longitudinal data, including information on convictions and antisocial or aggressive behavior up to the age of 26 years. These investigators found that although there was no main effect of MAOA genotype on antisocial behavior, regardless of how it was measured, there was a striking interaction effect between early maltreatment and MAOA, in that only those with a low-activity genotype showed a significant association between antisocial or aggressive behavior and early maltreatment. This pattern, using somewhat different environmental measures, has subsequently been replicated by an independent study conducted in the United States (Foley et al. 2004). Gene–environment correlation for personality traits has yet to be equally well explored, but there are three possible mechanisms by which it can arise (Plomin et al. 2001). The first is passive—that is, parents pass on both genes and environmental influences to their offspring. Second, particular behaviors that are genetically influenced may evoke certain responses from parents or other adults involved in a child’s upbringing or education. Thus, a well-behaved child may evoke more favorable and encouraging responses from adults than a frequently naughty child, toward whom carers may be more punitive. Third, there may be active gene–environment covariation such that, for example, a child who is intellectually curious may seek out environments that facilitate learning more often than does a child who is innately less open to new ideas. Part of the reason that gene–environment covariation has been little studied in behavioral genetics is that until comparatively recently, methods have not been available to reliably detect it. Such difficulties are now gradually being overcome, and, for example, twin models that incorporate gene–environment covariation have been developed (Purcell and Sham 2002).

Are the Extremes the Same as the Middle? A final complication to address in applying genetic approaches to the classification of personality is crucial to the question of whether categories really can be abolished in favor of dimensions. In fact, modern quantitative genetic studies in general make an assumption that disorders are nothing more than the extremes of dimensions. Thus, the prevalent model for familial diseases is the liability threshold model (Falconer 1965; Reich et al. 1972), in which it is assumed that multiple en-

58

Dimensional Models of Personality Disorders

vironmental and genetic effects contribute to a continuum of liability that tends to have a normal distribution, and that only those individuals who at some point exceed a certain threshold manifest the disorder. Relatives of affected individuals have on average an increased liability compared with the population mean, and thus more of them lie beyond the threshold for being affected. Knowing the proportion of the population affected and the proportion in a certain class of relatives allows a correlation in liability to be calculated. This is a useful measure of familiality that forms the starting point for structural equation model fitting. For most diseases, liability is unobserved, but for personality disorders there is at least the potential that personality dimensions can be sufficiently well measured that scores on a trait or set of trait measures will be more useful than simply categorizing individuals as affected or unaffected. This, however, presupposes that the same etiological factors operate at the ends of a distribution as are operating in the middle, which may not always be the case. Although this author knows of no examples relating to personality, there are instances for other behavioral traits, such as language acquisition, where the lower end of the distribution shows higher heritability than is found in the middle (Dale et al. 1998).

Conclusion Dr. Livesley’s general thesis that genetic studies are likely to be of help in improving our understanding of personality disorder and arriving at a sounder, more etiologically based classification must be seen as broadly correct. The general approach offers one of the best ways forward, and it is apparent that genetics cannot be ignored in attempts to devise an improved classificatory scheme for personality disorders. Nevertheless, there are complications and inherent limitations in the genetic approach that need to be taken into account and can potentially lead to major errors if they are overlooked.

References Button TM, Scourfield J, Martin N, et al: Family dysfunction interacts with genes in the causation of antisocial symptoms. Behav Genet 35:115–120, 2005 Cadoret RJ, Yates WR, Troughton E, et al: Genetic-environmental interaction in the genesis of aggressivity and conduct disorders. Arch Gen Psychiatry 52:916–924, 1995 Cardno AG, Gottesman II: Twin studies of schizophrenia: from bow-and-arrow concordances to star wars Mx and functional genomics. Am J Med Genet (Semin Med Genet) 97:12–17, 2000 Caspi A, McClay J, Moffit TE, et al: Role of genotype in the cycle of violence in maltreated children. Science 297:851–854, 2002

McGuffin: Commentary on Livesley

59

Dale PS, Simonoff E, Bishop DV, et al: Genetic influence on language delay in two-yearold children. Nat Neurosci 1(4):324–328, 1998 Falconer DS: The inheritance of liability to certain diseases, estimated from the incidence among relatives. Ann Hum Genet 29:51–76, 1965 Farmer AE, McGuffin P, Gottesman II: Twin concordance for DSM-III schizophrenia. Scrutinizing the validity of the definition. Arch Gen Psychiatry 44:634–641, 1987 Foley DL, Eaves LJ, Wormley B, et al: Childhood adversity, monoamine oxidase a genotype, and risk for conduct disorder. Arch Gen Psychiatry 61:738–744, 2004 Plomin R, DeFries JC, McClearn GE, et al: Behavioral Genetics, 4th Edition. New York, Worth, 2001 Purcell S, Sham P: Variance components models for gene-environment interaction in quantitative trait locus linkage analysis. Twin Research 5:572–576, 2002 Reich T, James JW, Morris CA: The use of multiple thresholds in determining the mode of transmission of semi-continuous traits. Ann Hum Genet 36:163–184, 1972

This page intentionally left blank

7 NEUROBIOLOGICAL DIMENSIONAL MODELS OF PERSONALITY A Review of Three Models Joel Paris, M.D.

Background and Guiding Principles Personality disorders can be understood as reflecting pathological amplifications of personality trait profiles (Costa and Widiger 2002; Depue and Lenzenweger 2001; Livesley et al. 1998; Paris 2003; Siever and Davis 1991). All of psychopathology, whether on Axis I or Axis II, need not necessarily fall into categories, and could reflect interactions between a few broad dimensions (Krueger 1999). Historically, there have been two approaches to defining personality dimensions. One is primarily empirical, based on the results of factor analysis of selfreport data, yielding broad traits that are relatively independent of each other and that can be further subdivided into narrower facets or subtraits. Building a model of disorders using this approach could be described as “bottom-up,” and this method characterizes the Five-Factor Model (FFM; Costa and Widiger 2002) as well as the work of other researchers (Clark and Livesley 2002; Livesley et al. 1998).

This chapter is an abbreviated version of a paper first published in the Journal of Personality Disorders (Volume 19, Issue 2, pages 156–170, 2005).

61

62

Dimensional Models of Personality Disorders

Alternatively, one might develop a “top-down” approach, linking traits with neurobiological measures. Gray (1991) proposed one of the earliest of these systems, using three brain systems: 1) approach (response to positively reinforcing stimuli), 2) fight–flight (response to negatively reinforcing stimuli), and 3) behavioral inhibition. However, although this model resembles some of the schema to be reviewed here, it has been tested primarily in animal models and has not been operationalized for humans using standard measurements.

CLONINGER MODEL The original model proposed by Cloninger (1987) was tridimensional, describing three dimensions of temperament, each of which was hypothesized to be linked to a specific neurotransmitter system: Novelty Seeking, linked to dopamine; Harm Avoidance, linked to serotonin; and Reward Dependence, linked to norepinephrine. A revised model (Cloninger et al. 1993) included a fourth temperamental dimension of Persistence, as well as three “character” dimensions: Self-Directedness, Cooperativeness, and Self-Transcendence, with Self-Directedness and Cooperativeness being broad measures of personality disturbance. The investigators proposed that the three characterological dimensions are less rooted in temperament. All seven dimensions can be assessed with a self-report instrument, the Temperament and Character Inventory (TCI). Because a large amount of data support the FFM as reflecting the basic structure of personality (Costa and Widiger 2002), one might ask whether the TCI has any specific advantages. Temperamental and characterological profiles derived from the model can be used to describe many of the same phenomena as Axis II diagnoses (Svrakic et al. 2002), but this has also been shown for the FFM (Costa and Widiger 2002). The temperamental factors in this model generally resemble four of the five factors on the FFM (Extraversion, Neuroticism, Agreeableness, and Conscientiousness), although they do not map personality in quite the same way. The factors in the TCI are not fully independent of each other (Herbst et al. 2000), but this is also true of the FFM (Costa and Widiger 2002). There is substantial overlap between these two systems (Macdonald and Holland 2002), and the TCI can be factor-analyzed to yield the FFM (Ramanaiah et al. 2002). The main question is whether the Cloninger model can truly be described as neurobiological. Whereas the original theory on the links between temperament, neural pathways, and neurotransmitters were very specific, the evidence in support of these associations has not been consistent. Perhaps this lack of supporting data explains why these links have not been emphasized in later publications (e.g., Cloninger 2004). In retrospect, one might conclude that the original schema was heuristic but speculative, and that it has not been shown to be consistent with the ways behavioral systems are organized at the neurobiological level.

Paris: Neurobiological Dimensional Models of Personality

63

Nonetheless, a large amount of research has been published using the Cloninger schema, largely because the scales purport to provide a measure of temperament. In this review, my focus will be on whether biological research supports the model. Peripheral measures of neurotransmitters can be used to test neurobiological associations with personality dimensions. A few studies support a relationship between peripheral measures of serotonin receptor binding and TCI Harm Avoidance (Nelson et al. 1996; Peirson et al. 1999). This finding is consistent with frequent reports of a relationship between serotonin levels and impulsivity (Coccaro 2004). However, Gerra et al. (1999) reported that Novelty Seeking was associated with plasma norepinephrine (rather than dopamine, as predicted by the theory). A second method involves molecular genetics. A number of studies have looked for linkages between candidate genes and the TCI scales. Quite a few have reported negative or unclear results (Gebhart et al. 2000; Ham et al. 2004; Herbst et al. 2000; Jonsson et al. 2003; Suzuki et al. 2003; Thierry et al. 2004). In a multivariate study of 59 traits against all seven dimensions, Comings et al. (2000) noted that their findings showed “different ratios of functionally related groups of genes, and of different genotypes of the same genes, for different traits” (p. 375). In relation to Novelty Seeking, some molecular genetic studies have supported the model, while others have not. A relationship between the D4 dopamine receptor gene (DRD4) and Novelty Seeking was reported by Cloninger et al. (1996) and replicated by Benjamin et al. (1996), by Ebstein et al. (1996), and by KeltikangasJarvinen et al. (2003). However, this association was not found by Ebstein et al. (1997), Gebhardt et al. (2000), Herbst et al. (2000), Malhotra et al. (1996), Pogue-Geile et al. (1998), or Vandenbergh et al. (1997). Ronai et al. (2001) found a relation between the DRD4 and Novelty Seeking, as the theory would predict, but the link was also significant only in females. Benjamin et al. (2000) suggested that more variance might be accounted for by considering interactions between the serotonin transporter gene (5-HTT) and DRD4, indicating that this trait may not be related to a single neurotransmitter system. Finally, reviews (Lusher et al. 2001; Patterson et al. 1999) and meta-analyses (Kluger et al. 2002; Munafò et al. 2003; Schinka et al. 2002) of the relationship between DRD4 and Novelty Seeking suggest that there is at best only a weak relationship between them. In relation to Harm Avoidance, there is some evidence for a link with 5-HTT, the serotonin transporter gene (Melke et al. 2003). However, other studies (Ham et al. 2004; Jonnson et al. 2003; Kusumi et al. 2002; Samochowiec et al. 2001; Soyka et al. 2002) failed to confirm this relationship. One study (Lee et al. 2003) found that the relationship between 5-HTT and Harm Avoidance might apply only to women. This observation would be in concordance with recent evidence that serotonergic pathways in the brain are different in men and women (Soloff 2004). There have been fewer studies of Reward Dependence, and no confirmation that it is related, as the theory predicts, to norepinephrine. In another report that

64

Dimensional Models of Personality Disorders

points to the importance of gender differences in neurobiological correlates of personality, Itoh et al. (2004) found a relationship between brain-derived neurotrophic factor and Reward Dependence that was significant only in females. The other dimensions of the model have an even smaller base of empirical data. A sib-pair study (Farmer et al. 2003) was supportive of the overall model in that it suggested that the character traits of the TCI (particularly Self-Directedness) have genetic loadings that could be accounted for by the original group of temperamental dimensions. However, in the study by Comings et al. (2000), the DRD4 gene was linked to Self-Transcendence, suggesting that the “character” dimensions must also reflect temperament. Behavior genetics has provided support for the temperamental nature of the original four dimensions: Stallings et al. (1996) used a twin sample to show that the four dimensions in the Cloninger model are genetically independent; in a Japanese sample, Ando et al. (2002), using multivariate genetic analysis, found no significant associations between Novelty Seeking, Harm Avoidance, and Reward Dependence, as the theory predicts, while the genetic components of Persistence, Self-Directedness, and Cooperativeness were derived from the temperament dimensions. However, Ando et al. (2004) also found that the character scales had a similar genetic contribution as the temperamental scales. In addition, multivariate analyses obtained four factors that did not correspond to the current TCI. In neuroimaging findings, a functional magnetic resonance imaging study (Gusnard et al. 2003) found an association between the Persistence dimension and activity in orbitofrontal and adjacent ventromedial cortex. In a positron emission tomography (PET) study of normal volunteers, Youn et al. (2002) noted specific differences in activity in different brain regions in relation to TCI scores. Suhara et al. (2001) found differences in D2 receptor binding in insular cortex related to Novelty Seeking, a finding that supports the original theoretical model. Neurophysiological markers suggest that Self-Directedness correlates with event-related brain potentials (Vedeniapin et al. 2001). A similar finding was reported by Hansenne et al. (2000), but the authors noted a failure to find strong correlations. In summary, research findings concerning links between the Cloninger model and neurobiological variables are highly inconsistent and often negative. Although the model has generated a large body of research, theoretical concepts have not been supported by data. Thus, the claim that the Cloninger model is neurobiological is not justified by the existing evidence. It shows no superiority to the FFM in describing personality or personality disorders.

DEPUE MODEL Depue and Lenzenweger (2001) have developed an overall model of personality and personality disorders. They propose that personality is built on four higher-

Paris: Neurobiological Dimensional Models of Personality

65

order trait dimensions, the first three of which are associated with specific transmitters that modulate brain systems, although they acknowledge that personality traits cannot be modeled on single neurotransmitters. The higher-order traits that appear in virtually every taxonomy are Extraversion and Neuroticism. The FFM and most other schema include dimensions of Agreeableness and Conscientiousness. Depue and Lenzenweger divided Extraversion into two components, one related to Affiliation (sociability), and the other to Agency (social dominance), and these facets can in turn be divided into lower-order components. Depue and Lenzenweger distinguish between Neuroticism (anxiety) and Harm Avoidance (fear), which some behavioral genetic studies (Tellegen et al. 1988) have indicated to be independent traits. Depue and Lenzenweger also argue that Impulsivity is a heterogeneous cluster of lower-order traits, as opposed to a single higher-order trait, redefining this construct as nonaffective constraint to separate it from behavioral disinhibition that is not related to emotional states. Depue and Lenzenweger (2001) propose a biological basis for Extraversion and Neuroticism. Extraversion would be related to dopaminergic brain systems. Depue and Collins (1999) have provided a detailed neurobiological theory of this trait, which they consider to be a measure of positive incentive motivation. Depue and Collins note parallels between Extraversion (particularly its Agency component) and a mammalian behavioral approach system based on positive incentive motivation, implicating specific neuroanatomic networks and modulatory neurotransmitters in its processing. Thus, the corticolimbic-striatal-thalamic network carries out functions that include 1) integrating salient incentive context in the medial orbital cortex, amygdala, and hippocampus; 2) encoding intensity of incentive stimuli in a motive circuit composed of the nucleus accumbens, ventral pallidum, and ventral tegmental area dopamine projection system; and 3) creating an incentive motivational state that can be transmitted to the motor system. Individual differences in the functioning of this network could arise from functional variation in the ventral tegmental area dopamine projections directly involved in coding the intensity of incentive motivation. The other aspect of Extraversion is Affiliation, which concerns neurobiological processes that promote longer-term affective bonds. Research by Insel (1997) suggests that neuropeptides such as oxytocin and vasopressin play important roles in this process. In the Depue model, Neuroticism and Harm Avoidance are hypothesized to be independent, but with norepinephrine playing a role in both. One key anatomical structure here is the locus coeruleus, which innervates all brain regions. It is suggested that fear responses are coordinated by the amygdala, while anxiety is related to a different subcortical structure (the bed nucleus of the stria terminalis). Nonaffective Constraint describes the response threshold of brain systems. It is hypothesized to be related to serotonergic brain systems, associated with the wide innervation of brain structures by the dorsal raphe.

66

Dimensional Models of Personality Disorders

Depue and Lenzenweger (2001) describe how these five dimensions can account for features of the personality disorder diagnoses. They suggest that this model can explain a number of puzzling phenomena about personality disorders, ranging from gender differences in prevalence to unclear relationships with genes and biological markers. We require much more data before we can evaluate this model. Like the Cloninger model, it is consistent with some current research but may not account for inconsistent findings, not to speak of research that is likely to appear in the future. Although there is much evidence that dopamine is related to reward systems, we need more confirmation about whether it is the major neurotransmitter associated with Extraversion and whether the specific pathways suggested for its action are operative and unique to this trait. In summary, the Depue model is well thought out but has not generated specific instruments for measuring the precise dimensions it postulates. With the exception of preliminary studies (e.g., Lenzenweger et al. 2004), the model has not been the subject of a large body of empirical research.

SIEVER AND DAVIS MODEL The model of Siever and Davis (1991) was designed to describe neurobiological dimensions that underlie all categories of psychopathology, both on Axis I and Axis II. Four dimensions are described: 1. Cognitive/Perceptual—based on brain systems for attention and response. The mechanism is hypothesized to be dopaminergic. Evidence derives from studies of schizophrenic and schizotypal patients, with abnormalities in homovanillic acid (HVA) related to other biological markers, such as eye-movement dysfunction. 2. Impulsivity/Aggression—based on brain systems related to the capacity to inhibit behavior. The mechanism is hypothesized to be serotonergic. Evidence derives from the relationships between impulsivity and aggression and measures of 5-hydroxyindoleacetic acid (5-HIAA), neuroendocrine challenge tests, and PET studies. 3. Affect Regulation—based on brain systems related to stability of mood. In the original publication, it was proposed that this system relates to noradrenergic– cholinergic balance, but this hypothesis may have been discarded due to lack of supporting evidence. 4. Anxiety/Inhibition—based on brain system for assessing danger, and associated with high arousal. In the original publication, it was proposed that this system relates to reduced dopaminergic and increased serotonergic activity. The Siever and Davis model focuses on clinical phenomena rather than normal variations, and there is no theory to establish linkages between them. There

Paris: Neurobiological Dimensional Models of Personality

67

remains a lack of consistent evidence that neurotransmitter variations consistently correspond to these trait dimensions. Few specific empirical tests have been applied to the model. Like the Cloninger model, the Siever and Davis model is somewhat simplistic in linking behavioral traits to neurotransmitter systems. The best and most robust data support a relationship between impulsivity/aggression and serotonin (Coccaro 2004; Mann 1998; Paris et al. 2004; Siever et al. 1998), although even here, correlations have not demonstrated a simple linear relationship. On the other hand, there is evidence disconfirming the proposal that affect regulation in personality disorders is related to a defect in noradrenergic–cholinergic balance (Paris et al. 2004). Depue and Lenzenweger (2001) point out that the dimensions seen in psychopathology may not be primary, but rather might reflect secondary interactions between more basic processes. Moreover, levels of neurotransmitters such as serotonin are not specific to one dimension but have widespread effects on behavior, including impulsivity, aggression, arousal, and emotional regulation. The Siever and Davis (1991) model has not generated standard instruments for measuring its dimensions, although it has stimulated research, using other measures on the dimension of affective instability (Henry et al. 2001). Like the Cloninger model, most of the hypothesized relationships between dimensions and neurotransmitters have not been supported by research, with the important exception of links between impulsivity and central serotonin activity.

What Research Is Needed to Develop a Better Model? Researchers will continue to look for associations between trait dimensions and neurobiological measures. It is possible that this approach could bear fruit. But it is also possible that it will continue to provide us with ambiguous data. Although a few findings in the literature have been robust (e.g., links between serotonin with impulsive aggression or harm avoidance), most have not shed great light on these relationships. It may well be that our understanding of brain mechanisms is at too early a stage to conduct this research program successfully. Until we know more about the neuroscience of emotions and behaviors, attempts to develop a neurobiological model linked to personality dimensions are premature. Neurobiological research has rarely supported simple reductionistic models. There is virtually no brain function that is strictly limited to one brain site or to one type of neuron: modulation and interaction are the rule for virtually anything one wishes to study (Andreasen 2001). We are a long way from being able to map out such brain systems, either anatomically or functionally. The question is whether the data are sufficient to show that any specific trait is strongly linked to

68

Dimensional Models of Personality Disorders

any specific biological mechanism. It is not sufficient that links be plausible; they must also be solidly established. The literature on neurotransmitters reveals an astonishing level of complexity. The monoamines, which have been the main subject of research, serve to modulate the effects of other neurons that use glutamine and gamma-aminobutyric acid (GABA) as transmitters (Cooper et al. 2003), so that their effects on behavior are far from linear; the same receptors could have entirely different effects in different brain locations, and serotonin may have as many as 15 receptor sites (Kroeze et al. 2002). These findings make it unlikely that we can find one-to-one correspondence between any single neurotransmitter and any single neurophysiological mechanism, not to speak of behavioral traits. Moreover, if such associations are found in the future, they are likely to be related to narrower dimensions rather than the broader traits that are currently being investigated (Livesley et al. 1998). Thus, although it is useful in principle to define traits that have a “genetic architecture,” it is not likely that we will find robust correlates associated with just a few dimensions, given the multiplicity and complexity of brain systems. Even if we could identify traits that have a specific genetic architecture, it does not follow that they will be associated with unique neurotransmitters. Neuroscience has tended to focus on neurochemistry, probably because this is the area in which the most dramatic progress has been made. But neurotransmitters have different effects, depending on brain anatomy and physiology. Future models will probably depend on advances in the understanding of neurocircuitry. Finally, the models reviewed here need to be compared with other measures of personality dimensions, some of which have their own neurobiological literature. For example, impulsive aggression, which may be related to Novelty Seeking and Harm Avoidance, has shown consistent (although not always strong) relationships with serotonergic dysfunction and with polymorphisms of the 5-HTT gene, at least when operationally defined by self-report measures (such as the Barratt Impulsivity Scale; Coccaro 2004). Personality traits are powerful constructs that predict a wide range of behavioral patterns, but there is little evidence that personality dimensions correspond in any predictable way to brain systems (Matthews et al. 2003; Zuckerman 1991). Instead, traits can be thought of as complex outcomes of interactions between many systems. In summary, although research should continue on the relationship between personality traits and neurobiology, it is premature for psychiatry to adopt a neurobiologically based system to describe the disorders currently classified on Axis II. If we decide to replace categories with dimensions, we would probably be better off using a factor analytically derived schema. Moreover, these dimensions could be revised at some later point in the light of further progress in research on the links between brain and behavior.

Paris: Neurobiological Dimensional Models of Personality

69

References Ando J, Ono Y, Yoshimura K, et al: The genetic structure of Cloninger’s seven-factor model of temperament and character in a Japanese sample. J Pers 70:583–609, 2002 Ando J, Suzuki A, Yamagata S, et al: Genetic and environmental structure of Cloninger’s temperament and character dimensions. J Personal Disord 18:379–393, 2004 Andreasen NC: Brave New Brain: Conquering Mental Illness in the Era of the Genome. New York, Oxford University Press, 2001 Benjamin J, Patterson C, Greenberg BD, et al: Population and familial association between the D4 receptor gene and measures of novelty seeking. Nat Genet 12:81–84, 1996 Benjamin J, Osher Y, Kotler M, et al: Association between tridimensional personality questionnaire (TPQ) traits and three functional polymorphisms: dopamine receptor D4 (DRD4), serotonin transporter promoter region (5-HTTLPR) and catechol-Omethyltransferase (COMT). Mol Psychiatry 5:96–100, 2000 Clark LA, Livesley WJ: Two approaches to identifying the dimensions of personality disorder: convergence on the five-factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 161–176 Cloninger CR: A systematic method for clinical description and classification of personality variants. A proposal. Arch Gen Psychiatry 44:573–588, 1987 Cloninger CR: Feeling Good. New York, Oxford University Press, 2004 Cloninger CR, Svrakic DM, Pryzbeck TR: A psychobiological model of temperament and character. Arch Gen Psychiatry 50:975–990, 1993 Cloninger CR, Adolfsson NM, Svrakic DM: Mapping genes for human personality. Nat Genet 12:3–4, 1996 Cloninger CR, Svrakic NM, Svrakic DM: Role of personality self-organization in development of mental order and disorder. Dev Psychopathol 9:881–906, 1997 Coccaro EF: Intermittent explosive disorder and impulsive aggression: the time for serious study is now. Curr Psychiatry Rep 6:1–2, 2004 Comings DE, Gade-Andavolu R, Gonzalez N, et al: A multivariate analysis of 59 candidate genes in personality traits: the temperament and character inventory. Clin Genet 58:375–385, 2000 Comings DE, Gonzales N, Saucier G, et al: The DRD4 gene and the spiritual transcendence scale of the character temperament scale. Psychiatr Genet 10:185–189, 2000 Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2003 Costa PT Jr, Widiger TA (eds): Personality Disorders and the Five Factor Model of Personality, 2nd Edition. Washington, DC, American Psychological Association, 2002 Depue RA, Collins PF: Neurobiology of the structure of personality: dopamine, facilitation of incentive motivation, and extraversion. Behav Brain Sci 22:491–517, 1999 Depue RA, Lenzenweger M: A neurobehavioral dimensional model, in Handbook of Personality Disorders: Theory, Research, and Treatment. Edited by Livesley WJ. New York, Guilford, 2001, pp 137–176

70

Dimensional Models of Personality Disorders

Ebstein RP, Novick O, Umansky R, et al: Dopamine receptor (D4R) exon III polymorphism associated with the human personality trait of novelty seeking. Nat Genet 12:78–80, 1996 Ebstein RP, Gritsenko I, Nemanov L, et al: No association between the serotonin transporter gene regulatory region polymorphism and the Tridimensional Personality Questionnaire (TPQ) temperament of harm avoidance. Mol Psychiatry 2:224–226, 1997 Farmer A, Mahmood A, Redman K, et al: A sib-pair study of the Temperament and Character Inventory scales in major depression. Arch Gen Psychiatry 60:490–496, 2003 Gebhardt C, Leisch F, Schussler P, et al: Non-association of dopamine D4 and D2 receptor genes with personality in healthy individuals. Psychiatr Genet 10:131–137, 2000 Gerra G, Avanzini P, Zaimovic A, et al: Neurotransmitters, neuroendocrine correlates of sensation-seeking temperament in normal humans. Neuropsychobiology 39:207–213, 1999 Gray JA: Neural systems, emotion, and personality, in Neurobiology of Learning, Emotion and Affect. Edited by Madden J. New York, Raven, 1991, pp 273–306 Gusnard DA, Ollinger JM, Shulman GL, et al: Persistence and brain circuitry. Proc Natl Acad Sci U S A 100:3479–3484, 2003 Ham BJ, Kim YH, Choi MJ, et al: Serotonergic genes and personality traits in the Korean population. Neurosci Lett 354:2–5, 2004 Hansenne M, Pitchot W, Pinto E, et al: P300 event-related brain potential and personality in depression. Eur Psychiatry 15:370–377, 2000 Henry C, Mitropoulou V, New AS, et al: Affective instability and impulsivity in borderline personality and bipolar II disorders: similarities and differences. J Psychiatr Res 35:307– 312, 2001 Herbst JH, Zonderman AB, McCrae RR, et al: Do dimensions of the temperament and character inventory map a simple genetic architecture? Evidence from molecular genetics and factor analysis. Am J Psychiatry 157:1285–1290, 2000 Insel T: A neurobiological basis of social attachment. Am J Psychiatry 154:726–735, 1997 Itoh K, Hashimoto K, Kumakiri C, et al: Association between brain derived neurotrophic factor 196 G/A polymorphism and personality traits in healthy subjects. Am J Med Genet 124B:61–63, 2004 Jonsson EG, Burgert E, Crocq MA, et al: Association study between dopamine D3 receptor gene variant and personality traits. Am J Med Genet 117B:61–65, 2003 Keltikangas-Jarvinen L, Elovainio M, Kivimaki M, et al: Association between the type 4 dopamine receptor gene polymorphism and novelty seeking. Psychosom Med 65:471– 476, 2003 Kluger AN, Siegfried Z, Ebstein RP: A meta-analysis of the association between DRD4 polymorphism and novelty seeking. Mol Psychiatry 7:712–717, 2002 Kroeze WK, Kristiansen K, Roth BL: Molecular biology of serotonin receptors structure and function at the molecular level. Curr Top Med Chem 2:507–528, 2002 Krueger RF: The structure of common mental disorders. Arch Gen Psychiatry 56:921–926, 1999 Kusumi I, Suzuki K, Sasaki Y, et al: Serotonin 5-HT(2A) receptor gene polymorphism, 5HT(2A) receptor function and personality traits in healthy subjects: a negative study. J Affect Disord 68:235–241, 2002

Paris: Neurobiological Dimensional Models of Personality

71

Lee HJ, Lee HS, Kim YK, et al: D2 and D4 dopamine receptor gene polymorphisms and personality traits in a young Korean population. Am J Med Genet 121B:44–49, 2003 Lenzenweger MF, Clarkin JF, Fertuck EA, et al: Executive neurocognitive functioning and neurobehavioral systems indicators in borderline personality disorder: a preliminary study. J Personal Disord 18:421–438, 2004 Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Lusher JM, Chandler C, Ball D: Dopamine D4 receptor gene (DRD4) is associated with novelty seeking and substance abuse: the saga continues. Mol Psychiatry 6:497–499, 2001 MacDonald DA, Holland D: Examination of relations between the NEO Personality Inventory—Revised and the Temperament and Character Inventory. Psychol Rep 91:921–930, 2002 Malhotra A, Virkkunen M, Ronney W, et al: The association between the dopamine D4 receptor (D4DR) 16 amino acid repeat polymorphism and novelty seeking. Mol Psychiatry 1:388–391, 1996 Mann JJ: The neurobiology of suicide. Nat Med 4:25–30, 1998 Matthews G, Deary IJ, Whiteman MC: Personality Traits, 2nd Edition, New York, Cambridge University Press, 2003 Melke J, Westberg L, Nilsson S, et al: Polymorphism in the serotonin receptor 3A (HTR3A) gene and its association with harm avoidance in women. Arch Gen Psychiatry 60:1017–1023, 2003 Munafò MR, Clark TG, Moore LR, et al: Genetic polymorphisms and personality in healthy adults: a systematic review and meta-analysis. Mol Psychiatry 8:471–484, 2003 Nelson EC, Cloninger CR, Pryzbeck TR, et al: Platelet serotonergic markers and Tridimensional Personality Questionnaire measures in a clinical sample. Biol Psychiatry 40:271–278, 1996 Paris J: Personality Disorders Over Time: Precursors, Course, and Outcome. Washington, DC, American Psychiatric Press, 2003 Paris J, Zweig-Frank H, Ng F, et al: Neurobiological correlates of diagnosis and underlying traits in patients with borderline personality disorder compared with normal controls. Psychiatry Res 121:239–252, 2004 Patterson AD, Sunohara GA, Kennedy JL: Dopamine D4 receptor gene: novelty or nonsense? Neuropsychopharmacol 21:3–16, 1999 Peirson AR, Heuchert JW, Thomala L, et al: Relationship between serotonin and the Temperament and Character Inventory. Psychiatry Res 89:29–37, 1999 Pogue-Geile M, Ferrell R, Deka R, et al: Human novelty-seeking personality traits and dopamine D4 receptor polymorphisms: a twin and genetic association study. Am J Med Genet 81:44–48, 1998 Ramanaiah NV, Rielage JK, Cheng Y: Cloninger’s temperament and character inventory and the NEO Five-Factor Inventory. Psychol Rep 90:59–63, 2002 Ronai Z, Szekely A, Nemoda Z, et al: Association between novelty seeking and the -521 C/ T polymorphism in the promoter region of the DRD4 gene. Mol Psychiatry 6:35–38, 2001

72

Dimensional Models of Personality Disorders

Samochowiec J, Rybakowski F, Czerski P, et al: Polymorphisms in the dopamine, serotonin, and norepinephrine transporter genes and their relationship to temperamental dimensions measured by the Temperament and Character Inventory in healthy volunteers. Neuropsychobiology 43:248–253, 2001 Schinka JA, Letsch EA, Crawford FC: DRD4 and novelty seeking: results of meta-analyses Am J Med Genet 114:643–648, 2002 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991 Siever LJ, New AS, Kirrane R, et al: New biological research strategies for personality disorders, in Biology of Personality Disorders. Edited by Silk KR. Washington, DC, American Psychiatric Press, 1998, pp 27–62 Soloff P: Impulsivity and suicide: review and update of research in BPD. Presentation to the European Congress on Personality Disorders, Zaragoza, Spain, June 2004 Soyka M, Preuss UW, Koller G, et al: Dopamine D4 receptor gene polymorphism and extraversion revisited: results from the Munich gene bank project for alcoholism. J Psychiatr Res 36:429–35, 2002 Stallings MC, Hewitt JK, Cloninger CR, et al: Genetic and environmental structure of the Tridimensional Personality Questionnaire: three or four temperament dimensions? J Pers Soc Psychol 70:127–140, 1996 Suhara T, Yasuno F, Sudo Y, et al: Dopamine D2 receptors in the insular cortex and the personality trait of novelty seeking. Neuroimage 13:891–895, 2001 Suzuki E, Kitao Y, Ono Y, et al: Cytochrome P450 2D6 polymorphism and character traits. Psychiatr Genet 13:111–113, 2003 Svrakic DM, Draganic S, Hill K, et al: Temperament, character, and personality disorders: etiologic, diagnostic, treatment issues. Acta Psychiatr Scand 106:189–195, 2002 Tellegen A, Lykken DT, Bouchard TJ, et al: Personality similarity in twins reared apart and together. J Pers Soc Psychol 54:1031–1039, 1998 Thierry N, Willeit M, Praschak-Rieder N, et al: Serotonin transporter promoter gene polymorphic region (5-HTTLPR) and personality in female patients with seasonal affective disorder and in healthy controls. Eur Neuropsychopharmacol 14:53–58, 2004 Vandenbergh DJ, Zonderman AB, Wang J, et al: No association between novelty seeking and dopamine D4 receptor (DRD4) exon III seven repeat alleles in Baltimore Longitudinal Study of Aging participants. Mol Psychiatry 2:417–419, 1997 Vedeniapin AB, Anokhin AP, Sirevaag E, et al: Visual P300 and the self-directedness scale of the Temperament and Character Inventory. Psychiatry Res 101:145–156, 2001 Youn T, Lyoo IK, Kim JK, et al: Relationship between personality trait and regional cerebral glucose metabolism assessed with positron emission tomography. Biol Psychol 60:109–120, 2002 Zuckerman M: The Psychobiology of Personality. New York, Cambridge University Press, 1991

8 COMMENTARY ON PARIS Personality as a Dynamic Psychobiological System C. Robert Cloninger, M.D.

The choice of a model for describing human personality is clinically important because personality defines a perspective that shapes the way clinicians understand and treat their patients. Consequently, there is not one standard by which to judge alternative approaches, because people differ in their interests and perspectives about personality and psychopathology. Clinicians become familiar with particular ways of evaluating and treating patients, and it requires effort to switch to alternative models of personality and psychopathology. Nevertheless, there is reason to be optimistic, because alternative approaches show extensive convergence in ways that may allow a widespread consensus in improving the assessment of personality. Trait descriptions of the observable personality differences between individuals became dominant before there was any possibility of identifying specific psychological processes underlying motivated behavior by functional brain imaging and related techniques. Psychodynamic psychiatrists and social-cognitive psychologists have long criticized personality theorists for focusing on differences between the traits of individuals while treating the mind as a black box. The concern of clinicians about between-person models of traits is the absence of attention to the dynamic processes within the person that predict response to treatment and personality development within a social context. When I developed the Temperament and Character Inventory (TCI), I deliberately measured psychobiologically defined traits that provided an explanation of the dynamics of personality development, thereby providing a model of both within-person and between-person differences. Subsequent research has confirmed strong correlations (r > .7) between 73

74

Dimensional Models of Personality Disorders

TCI dimensions and the activation of specific brain networks under psychologically specific conditions, as described in detail elsewhere (Cloninger 2004). In order to account for both within-person and between-person differences, it is necessary to recognize the nonlinearity of the relations between temperament and character and to decompose traits like Extraversion into multiple processes that have distinct psychology, biology, and sociology. For example, NEO Personality Inventory—Revised (NEO PI-R; Costa and McCrae 1992) Neuroticism correlates strongly with both high TCI Harm Avoidance and low TCI Self-Directedness, even though these two TCI dimensions are correlated with activity in different brain networks. Fortunately, the between-person model of traits measured by the NEO PI-R and the within-person model of dynamic processes measured by the TCI are strongly convergent statistically (Table 8–1). Table 8–1 summarizes the correlations between the seven TCI higher-order dimensions and the five NEO PI-R higher-order factor scores, along with the multiple correlations of each trait with the scores from the other model. The agreement is strong for all traits, except NEO Openness and TCI Self-Transcendence, which show moderate overlap. When facet scales are used, the multiple correlations of individual NEO scales predicted by TCI subscales are .71 to .80; those of TCI scales predicted by NEO facets are .64 to .80. Such agreement is remarkable, given that the tests were administered more than a year apart in a community sample of 662 individuals in Oregon by Goldberg (1999). Therefore, we can confidently say that both the NEO and TCI provide a reliable and thorough coverage of human personality, although they organize the self-reported information according to different perspectives. The difference in the perspective provided by the TCI can be important clinically. For example, TCI Self-Directedness, but not Harm Avoidance, predicts response to cognitive-behavioral therapy (Cloninger 2000). Essentially, character traits like Self-Directedness measure higher-order cognitive processes involving the prefrontal cortex, whereas temperament traits like Harm Avoidance depend on emotional processes involving limbic networks. These processes are clinically dissociated in that whereas all individuals with a personality disorder are low in SelfDirectedness, some may be in the anxious cluster (i.e., high in Harm Avoidance) while others may be low in anxiety-proneness (i.e., low in Harm Avoidance). Conversely, all individuals with an anxiety disorder are high in Harm Avoidance, but they may vary in degree of maturity of their character. The clinical importance of the within-person dynamic perspective of the TCI is also shown by findings that the TCI temperament dimensions have a simple and direct relationship to the traditional clusters of personality disorder recognized in DSM-IV (American Psychiatric Association 1994). The four TCI temperaments define individuals who are high in Harm Avoidance (i.e., pessimistic and anxietyprone, as in the anxious or DSM cluster C); high in Novelty Seeking (i.e., impulsive and anger-prone, as in the erratic or DSM cluster B); low in Reward Depen-

Correlation coefficients and multiple-correlation coefficients of TCI scales with NEO PI-R scales (N = 662) TCI variables

NEO PI-R variables

NS

HA

RD

PS

SD

CO

ST

Mult-R

N E

.06 .40

.63 –.55

–.02 .52

–.20 .40

–.62 .25

–.28 .19

.06 .22

.75 .77

O

.43

–.25

.25

.07

.12

.20

.37

.54

CN

–.34

–.26

.00

.51

.41

.15

–.10

.70

A

–.23

.02

.40

.01

.31

.61

.20

.66

.65

.76

.68

.60

.67

.65

.45

Mult-R

Cloninger: Commentary on Paris

TABLE 8–1.

Note. Temperament and Character Inventory (TCI) variables: NS=Novelty Seeking; HA=Harm Avoidance; RD=Reward Dependence; PS=Persistence; SD=Self-Directedness; CO=Cooperativeness; ST=Self-Transcendence. NEO Personality Inventory—Revised (NEO PI-R) variables: N=Neuroticism; E=Extraversion; O=Openness; CN =Conscientiousness; A=Agreeableness.

75

76

Dimensional Models of Personality Disorders

dence (i.e., aloof and cold-hearted, as in the odd or DSM cluster A); and high in Persistence (i.e., determined and perseverative, if a fourth cluster for obsessive personalities is included) (Cloninger 2000). In contrast, Extraversion is a particularly bad fit to the erratic or impulsive cluster. As can be seen in Table 8–1, NEO Extraversion is more strongly correlated with measures of sociability and warmth (as measured by high TCI Reward Dependence) than with high TCI Novelty Seeking, which defines the B cluster well, along with its comorbidity with substance dependence. Fortunately, in defining consensus measures of personality, we can be guided by the convergence of clinical tradition with the results of psychobiological research.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Cloninger CR: A practical way to diagnosis of personality disorder: a proposal. J Personal Disord 14:99–108, 2000 Cloninger CR: Feeling Good: The Science of Well-Being. New York, Oxford University Press, 2004 Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO PI-R) and NEO FiveFactor Inventory (NEO-FFI) Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992 Goldberg LR: A broad-bandwidth, public domain, personality inventory measuring the lower-level facets of several five-factor models, in Personality Psychology in Europe, Vol 7. Edited by Mervielde I, Deary I, DeFruyt F, et al. Tilburg, The Netherlands, Tilburg University Press, 1999, pp 7–28

9 COMMENTARY ON PARIS The Problem of Severity in Personality Disorder Classification Peter Tyrer, M.D.

The valuable articles in this series omit one consideration that is very important in clinical practice: They do not address the question of severity. In this paper, I argue that the measure of severity, using what are described as hybrid models, is a critical component of practice and can be recorded easily using standard systems, both existing and planned. In arguing this case, I will use an exemplar, the Personality Assessment Schedule (PAS), mainly because we have so much data from this instrument. I would emphasize, however, that other assessment procedures can be easily adapted to produce similar severity assessments.

Origins of Personality Assessment Schedule The Personality Assessment Schedule was developed in 1976 and published in 1979. It rates 24 personality traits on a nine-point dimensional scale, depending on the degree of social maladjustment created by the personality characteristics.

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 309–314, 2005).

77

78

Dimensional Models of Personality Disorders

This has an algorithm leading to a four-cluster classification (Tyrer and Alexander 1979) but also has modifications to allow International Classification of Diseases, 10th Revision (ICD-10; World Health Organization 1992) and Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV; American Psychiatric Association 1994) diagnoses to be created. In addition to recording categories, we also recorded severity in our earlier studies, using a five-point scale derived from our algorithm (no personality disorder, personality difficulty, personality disorder, severe personality disorder, and gross personality disorder) (Tyrer et al. 1990). However, we quickly realized from analyzing the first sets of data in clinical practice that this was not a true dimensional scale. “Severe personality disorder” in one personality domain was less of a handicap in short- and long-term outcome and treatment response than were relatively milder personality disorders when they covered more than one cluster (what Oldham and his colleagues call “extensive personality disorder”) (Oldham and Skodol 2000; Oldham et al. 1992). We therefore adjusted this to a four-point (later a five-point) scale of severity (Table 9–1), in which the separation of simple and complex personality disorder is determined by the number of clusters containing one or more personality disorders. The lowest level of personality abnormality—personality difficulty—is accounted for by subthreshold personality disorders (i.e., those that do not quite meet the criteria using standard procedures). Since the original publication in 1996, we also have added an extra level—severe personality disorder—which includes those with the most antisocial personalities who are at risk to a much wider group in society. There is no good theoretical reason why overlapping dimensions of personality should be equated with severity, but in practice this seems to have empirical justification. Because of the degree of overlap between the criteria for existing personality disorders (Livesley et al. 1992), the choice of clusters for combining severity was a natural decision. We wanted to extend the clusters to four to accommodate the original ones identified in the PAS (passive-dependent, sociopathic, withdrawn, and inhibited), but we felt it necessary to stick to the original DSM cluster model (in which cluster C includes both passive-dependent and inhibited groups) in fitting the system to existing structures. Some details of studies with the PAS are illustrated in Table 9–2 using the severity coding system in a wide range of patients. These give at least face validity to the system, as there is a graded increase in the influence of personality pathology in all studies.

STUDY 1—NOTTINGHAM STUDY OF NEUROTIC DISORDER This is a study in which patients with anxiety and depressive disorders (generalized anxiety disorder, dysthymic disorder, and panic disorder) have their personality assessed at the beginning of treatment and treatment impact determined at intervals

79

Tyrer: Commentary on Paris

TABLE 9–1. Level of severity

Dimensional system of classifying personality disorders Description

Definition by categorical system

0

No personality disorder

1

Personality difficulty

2

Simple personality disorder

3

Complex (diffuse) personality disorder

4

Severe personality disorder

Does not meet actual or subthreshold criteria for any personality disorder Meets subthreshold criteria for one or several personality disorders Meets actual criteria for one or more personality disorders within the same cluster Meets actual criteria for one or more personality disorders within more than one cluster Meets criteria for creation of severe disruption to both individual and to many in society

Source.

After Tyrer and Johnson 1996 and Tyrer 2000, pp. 129–130.

up to 12 years later. Persistent social dysfunction (Seivewright et al. 2004) and general outcome (Tyrer et al. 2004a) were well predicted by personality status. Unsurprisingly, most people with personality disorder at baseline had abnormalities in the cluster C grouping, but in follow-up there was a considerable shift from cluster C to cluster A (Seivewright et al. 2002). Despite this, the severity coding at baseline was a remarkably robust predictor of outcome and was of much greater value than the category coding (incidentally suggesting that those who change in personality status when severely disordered do so by crossing clusters rather than improving to no personality disorder). The patients with complex personality disorder had the worst outcome and after 12 years were marginally worse than at baseline, illustrating the contention that if personality disorder is ignored, patients will apparently have “resistant” affective disorders.

STUDY 2 AND 3—PERSONALITY DISORDER IN RECURRENT PSYCHOSIS Personality is more difficult to assess in recurrent psychotic disorders than in many other patients but we have also found exactly the same findings in this group. Those with complex personality disorders respond less well to treatment and have more problems in the community with regard to violence and other antisocial behavior leading to contacts with the police (see Table 9–1) (Gandhi et al. 2001;

80

TABLE 9–2.

Examples of severity model of recording personality abnormality in predicting the effect of personality abnormality in the presence of other symptoms and behavior No personality disorder

Personality difficulty

Simple personality disorder

Complex (diffuse) personality disorder

1.7

2.1

2.3

3.7

<0.001

6.3

7.9

9.2

12.1

<0.001

Clinical population

Specific outcome

Authors

Anxiety and depressive disorders (outcome after 12 years) (Study 1)

Global outcome (NDOS) Social function

Tyrer et al. 2004b Seivewright et al. 2004

Recurrent psychotic illness (Study 2)

Mean police contacts Gandhi et al. over 1 year 2001

0.05

0.13

0.34

0.39

<0.001

Psychosis (UK700 study) (Study 3)

Duration of inpatient Tyrer and treatment (median) Seivewright 2000 Quality of life Tyrer and (LQOLP) Seivewright 2000

8 months

8 months

10 months

13.5 months

0.002

4.72

4.51

4.45

4.45

0.002

Note.

Percentage with self- Tyrer et al. harm over 1 year 2004b

20.5

37.6

51.3

LQOLP = Lancashire Quality of Life Profile (Oliver et al. 1997); NDOS = Neurotic Disorder Outcome Scale (Tyrer et al. 2004a).

55.1

<0.001

Dimensional Models of Personality Disorders

Recurrent self-harm (Study 4)

P

Tyrer: Commentary on Paris

81

Moran et al. 2003). There is also a negative impact of personality with regard to duration of inpatient care and quality of life (Tyrer and Seivewright 2000). Again, there is a clear advantage in separating personality status into four areas of severity.

STUDY 4—INFLUENCE OF PERSONALITY DISORDER ON REPETITION OF SELF-HARM In this study, 480 patients were recruited into a randomized controlled trial of psychological treatment (brief manual-assisted cognitive-behavior therapy) or treatment as usual for recurrent self-harm (at least one previous episode prior to index episode). The results (Tyrer et al. 2004b) showed that personality status was the most significant predictor of repeat self-harm; this was not confined to those with borderline personality disorder but covered the whole range. Another study (Verheul et al. 2003) suggested that severity of disorder may be important in predicting the outcome of dialectical behavior therapy.

Dangerous and Severe Personality Disorder In the United Kingdom, or more specifically in England only, a new diagnosis of dangerous and severe personality disorder has been introduced (Maden and Tyrer 2003). Despite the fact that this has been introduced almost by diktat and with no consultation with professional experts it does emphasize the importance of severity in the description of mental disorders. The same pressure to identify “severe mental illness” has influenced services in many countries and we need to acknowledge this in our classification structures (Tyrer 2004). In both the general clinical and lay mind there is uniformity of focus; we need to know who have the most severe personality disorders and how to attend to their special needs.

Implications The PAS, like many other instruments, has a factor analytical structure that identifies four clear higher-order factors—passive dependent, sociopathic, anankastic, and schizoid features—that are virtually identical to the “big four” identified in other models (Widiger and Simonsen 2005), with particular concordance with Livesley’s emotional dysregulation, dissocial, inhibitedness, and compulsivity dimensions (Livesley et al. 1998), and Mulder and Joyce’s (1997) four A’s—i.e., antisocial, asocial, asthenic, and anankastic groups. In DSM-V, this must be taken into account if it is to serve the interests of clinicians. A simple classificatory rule to score severity could be made for all instruments along the lines suggested above in which the higher-order dimensions would represent the groups for grading. As

82

Dimensional Models of Personality Disorders

Verheul (2005) emphasizes, it is clinical utility that will carry the day in clinical practice and this paper argues that recording the severity of personality disorder is an absolutely essential component of clinical utility.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Gandhi N, Tyrer P, Evans K, et al: A randomised controlled trial of community-oriented and hospital-oriented care for discharged psychiatric patients: influence of personality disorder on police contacts. J Personal Disord 15:94–102, 2001 Livesley WJ, Jackson DN, Schroeder ML: Factorial structure of traits delineating personality disorders in clinical and general population samples. J Abnorm Psychol 101:432– 440, 1992 Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Maden T, Tyrer P: Dangerous and severe personality disorders: a new personality concept from the United Kingdom. J Personal Disord 17:489–496, 2003 Moran P, Walsh E, Tyrer P, et al: The impact of comorbid personality disorder on violence in psychosis—report from the UK700 trial. Br J Psychiatry 182:129–134, 2003 Mulder RT, Joyce PR: Temperament and the structure of personality disorder symptoms. Psychol Med 27:99–106, 1997 Oldham JM, Skodol AE, Kellman HD, et al: Diagnosis of DSM-III-R personality disorders by two semistructured interviews: patterns of comorbidity. Am J Psychiatry 149:213– 220, 1992 Oldham JM, Skodol AE: Charting the future of Axis II. J Personal Disord 14:17–29, 2000 Oliver J, Huxley P, Priebe S, et al: Measuring the quality of life of severely ill people using the Lancashire Quality of Life Profile. Soc Psychiatry Psychiatr Epidemiol 32:76–83, 1997 Seivewright H, Tyrer P, Johnson T: Change in personality status in neurotic disorders. Lancet 359:2253–2254, 2002 Seivewright H, Tyrer P, Johnson T: Persistent social dysfunction in anxious and depressed patients with personality disorder. Acta Psychiatr Scand 109:104–109, 2004 Tyrer P: Challenges for the future, in Personality Disorders: Diagnosis, Management, and Course, 2nd Edition. Edited by Tyrer P. London, Arnold, 2000, pp 126–132 Tyrer P: Getting to grips with severe personality disorder. Crim Behav Ment Health 14:1– 4, 2004 Tyrer P, Alexander J: Classification of personality disorder. Br J Psychiatry 135:163–167, 1979 Tyrer P, Johnson T: Establishing the severity of personality disorder. Am J Psychiatry 153:1593–1597, 1996 Tyrer P, Seivewright H: Studies of outcome, in Personality Disorders: Diagnosis, Management, and Course, 2nd Edition. Edited by Tyrer P. London, Arnold, 2000, pp 105–125 Tyrer P, Seivewright N, Ferguson B, et al: The Nottingham Study of Neurotic Disorder: relationship between personality status and symptoms. Psychol Med 20:423–431, 1990

Tyrer: Commentary on Paris

83

Tyrer P, Seivewright H, Johnson T: The Nottingham Study of Neurotic Disorder: predictors of 12-year outcome of dysthymic, panic and generalised anxiety disorder. Psychol Med 34:1385–1394, 2004a Tyrer P, Tom B, Byford S, et al: Differential effects of manual assisted cognitive behavior therapy in the treatment of recurrent deliberate self-harm and personality disturbance: the POPMACT study. J Personal Disord 18:82–96, 2004b Verheul R: Clinical utility of dimensional models for personality pathology. J Personal Disord 19:283–302, 2005 Verheul R, van den Bosch LMC, Koeter MJW, et al: Dialectical behaviour therapy for women with borderline personality disorder: 12-month, randomised clinical trial in The Netherlands. Br J Psychiatry 182:135–140, 2003 Widiger TA, Simonsen E: Alternative dimensional models of personality disorder: finding a common ground. J Personal Disord 19:110–130, 2005 World Health Organization: The ICD-10 Classification of Mental and Behavioural Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland, World Health Organization, 1992

This page intentionally left blank

10 TEMPERAMENT AND PERSONALITY AS BROADSPECTRUM ANTECEDENTS OF PSYCHOPATHOLOGY IN CHILDHOOD AND ADOLESCENCE Ivan Mervielde, Ph.D. Barbara De Clercq, Ph.D. Filip De Fruyt, Ph.D. Karla Van Leeuwen, Ph.D.

Although it is widely recognized that adult personality disorders have their roots in a variety of genetic, temperamental, and developmental factors, there has been remarkably little research examining the childhood and adolescent antecedents of personality disorders across more than one personality disorder. Several factors

This chapter is an abbreviated version of a paper published in the Journal of Personality Disorders (Volume 19, Issue 2, pages 171–201, 2005). Part of the research reported in this paper was supported by a Ghent University Ph.D. research grant awarded to B. De Clercq (Grant No. 011D0201) under the supervision of Filip De Fruyt and Ivan Mervielde.

85

86

Dimensional Models of Personality Disorders

contribute to this lack of research on childhood antecedents of personality disorders. The categorical conceptualization of personality disorders as adopted in DSM-IV (American Psychiatric Association 1994) and its predecessors is one of the major factors that impedes research on broad-spectrum childhood and adolescent antecedents of these disorders. Finding clear and unambiguous antecedents for the overly specific personality disorders is elusive because the effects of classic antecedents such as genetic, temperamental factors as well as the impact of environmental factors such as parental neglect and emotional and sexual abuse are nonspecific and hence related to several personality disorders. The DSM-IV classification of disorders in childhood and adolescence is restricted to Axis I psychopathology; therefore, it is not surprising that disruptive behavior disorders such as conduct disorder and oppositional defiant disorder, mood disorders, anxiety disorders, and attention-deficit disorders have been targeted as antecedents of adult psychopathology. However, the Axis I childhood and adolescence disorders are primarily related to their adult Axis I counterparts, and, to a lesser extent, to adult personality disorders. Childhood and adolescent temperament and personality are probably among the best candidates as general broad-spectrum developmental antecedents for adult personality disorders (Krueger and Tackett 2003; Shiner and Caspi 2003; Warner et al. 2004). According to Krueger and Tackett (2003): Our field is currently enjoying a renaissance of research linking personality and psychopathology. During much of the recent past, the personality and psychopathology literatures proceeded mostly in parallel.. .. Current research might therefore be conceptualized as a rapprochement of the personality and psychopathology literature.. .. Nevertheless, the specialized focus of much current research is naturally accompanied by the risk of neglecting the bigger picture. (p. 109)

The temperament literature and, to a far lesser extent, the recent personality literature tend to focus on narrow-spectrum traits; therefore, there is a need for models that are more comprehensive and attend to the bigger picture by emphasizing broad conceptualizations of temperament/personality and psychopathology in childhood and adolescence. The goal of this chapter is to search for broad and common dimensions that represent individual differences in childhood and adolescent models of temperament, personality, and psychopathology.

Temperament Models Researchers interested in individual differences among children and especially among young children have conceptualized these differences in terms of temperamental characteristics. Temperament is traditionally distinguished from personality because it refers to stable individual differences that appear from birth onward and that presumably have a strong genetic or neurobiological basis. Theorists differ in

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

87

their emphasis on the role of emotional processes, stylistic components, and attentional processes as the core of temperament. The New York Longitudinal Study (NYLS) was a milestone for introducing the concept of individual differences in developmental psychology and pediatrics (Thomas and Chess 1977). Thomas and Chess developed a system of nine categories to classify behaviors relevant for child development: activity level, rhythmicity, approach-withdrawal, adaptability, threshold of responsiveness, intensity of reaction, quality of mood, distractibility, attention span, and persistence. This system was further operationalized in several questionnaires and rating forms to be used by parents or teachers of infants, preschool, and school-age children. Presley and Martin (1994) compared the dimensions recovered from item-level factor analysis of several instruments and concluded that there was very little evidence supporting the Thomas and Chess nine-factor structure, although some of the original dimensions appeared to be strongly represented. Buss and Plomin (1975) initially distinguished four temperamental dimensions: Emotionality, Activity, Sociability, and Impulsivity. Emotionality is roughly equivalent to distress. It involves intense activation of the sympathetic nervous system and hence high emotional arousal. The rate and amplitude of speech and movement, displacement of body movements, and duration of energetic behavior best measure Activity, the second temperamental dimension. Sociability refers to the preference for being with others, the need to share activities and to receive rewarding attention as the result of social interaction. Questionnaires based on the Buss-Plomin model have been used extensively in genetic studies of behavior. A meta-analysis of eight of these studies by Goldsmith et al. (1997) showed weighted intraclass correlations of .57 to .66 for monozygotic twins and below .15 for dizygotic twins. The key concepts in the temperament model developed by Rothbart and Derryberry are Reactivity and Self-Regulation. These researchers asserted that temperament models should be based on the assumption that personality differences arise in part from the reactivity of underlying neural systems (Derryberry and Rothbart 1997). Behaviorally, temperament can be observed across all ages as differences in patterns of emotionality, activity and attention. Motivational as well as attentional systems are considered to provide the link relating specific neural systems to the major dimensions of temperament. To assess some of the behavioral and emotional indicators of the neurological systems, Rothbart and Ahadi (1994) constructed the Child Behavior Questionnaire (CBQ). Factor analysis of the 15 CBQ scales revealed three factors with some minor cross-cultural variations. The Surgency factor encompassed approach, high-intensity pleasure, smiling and laughter, high activity level, impulsivity, and shyness. This factor can be identified as a Positive Emotionality/Extraversion factor. The second factor, Negative Affectivity, included discomfort, fear, anger, sadness, and soothability/falling reactivity and was clearly related to Neuroticism or Negative Emotionality. The third factor, Effortful Control, combined inhibitory control, attentional focusing, low-intensity pleasure, and perceptual

88

Dimensional Models of Personality Disorders

sensitivity and may be linked to the Conscientiousness factor of the FFM. Goldsmith and Campos (1982) defined temperament as individual differences in the probability of experiencing and expressing primary emotions and arousal. The basic emotions that form the content dimensions of this model are anger, sadness, fear, joy and pleasure, disgust, interest, and surprise. Goldsmith constructed the Toddler Behavior Assessment Questionnaire (TBAQ), a caregiver report with acceptable convergent and discriminant validity (Goldsmith et al. 1997; Lemery et al. 1999). It consists of five independent scales: Activity Level, Pleasure, Social Fearfulness, Anger Proneness, and Interest/Persistence. Recent research on the Goldsmith and Campos model has been mainly confined to behavior genetic analyses. Goldsmith et al. (1997) conducted a classical intraclass correlation analysis with monozygotic and dizygotic twins, showing substantial monozygotic as well as dizygotic intraclass correlations for all five scales. This brief review of temperament models shows that each model has a set of scales that partially overlap with those proposed by other theorists. Moreover, even within a single temperament model, it is often necessary to construct different, age-specific measures to cope with the expanding behavioral repertoire of the children. Part of the problem in comparing the temperament models is that Thomas and Chess, and Rothbart and Derryberry, developed many scales that are moderately correlated, while Buss and Plomin used fewer but independent scales and Goldsmith preferred to integrate scales from different questionnaires into independent composites. Mervielde and Asendorpf (2000) compared the relatively independent dimensions emerging from the four temperament models. Emotionality is clearly present in each of the four models as a higher-order dimension. Given the emphasis on distress and negative emotions, this dimension can also be labeled Negative Emotionality. The second consistent dimension refers to sociability versus social inhibition or shyness and may be best captured with the more general label Extraversion. Activity is also prominent as an independent dimension in three of the four models. Finally, Persistence is present in two of the four models and hence might turn out to be a minor dimension, at least for preschool and school-age children. Given the greater comparability of dimensions emerging from the use of broader, more or less independent dimensions, future research might benefit from including higherorder factors or composite scores from at least two temperament models in studies of the structure, stability and predictive validity of temperament.

Personality Models PERSON-CENTERED APPROACH The person-centered approach emphasizes the importance of the person as the major unit of analysis and the study of the structure of personality profiles across

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

89

variables; hence, the resultant structure is a property of persons and not of the variables (Mervielde and Asendorpf 2000). According to Block and Block (1980), the dimensions ego control and ego resiliency are important to distinguish among three personality types: undercontrolled, overcontrolled, and resilient. Ego control refers to the tendency to contain versus express emotional and motivational impulses. Both extremes of the ego-control dimension can be considered rather maladaptive. Ego resiliency refers to the tendency to respond flexibly rather than rigidly to changing situational demands or particularly stressful situations; hence, low ego resiliency can be considered a maladaptive personality type. Personality types can be obtained by Q-factor analysis of Q sorts, as shown by Robins et al. (1996) and Hart et al. (1997). They found that only three Q-factors were replicable in their studies of 13-year-old North American boys and 7-year-old Icelandic children: Resilient, Overcontrolled, and Undercontrolled. Asendorpf et al. (2001) showed that the three personality types can also be recovered from replicated cluster analysis of Big Five ratings of 12-year-old children. Each of the types has a distinct Big Five profile. Resilient children score above average on Agreeableness, Extraversion, Conscientiousness, and Openness and below average on Neuroticism. In other words, Resilient children are well adapted and have a desirable Big Five personality profile. Overcontrolled children score more than half a standard deviation above the mean on Neuroticism and substantially below the mean on Extraversion. Undercontrolled children score substantially below the mean on Agreeableness and Conscientiousness; hence, Overcontrolled and Undercontrolled types have a personality profile that is less desirable and rather maladaptive. The replicability of the three types, at least in adult and adolescent samples, recently has become a controversial issue (Costa et al. 2002; De Fruyt et al. 2002). Moreover, head-to-head comparisons of the predictive validity of the person-centered versus the variable-centered approach show that types have no incremental validity beyond what can be predicted on the basis of a variable-centered approach (Asendorpf 2003). Although the assessment of personality types is a clear example of a categorical approach to childhood and adolescent personality, it remains important to realize that the characteristic features of the three types can be captured and described in terms of dimensional models of personality (Van Leeuwen et al. 2004a).

VARIABLE-CENTERED APPROACH Organizing personality traits as part of the Big Five, or the Five-Factor Model, is gradually recognized as the preferred strategy for representing the structure of adaptive or normal personality traits in adulthood (Costa and McCrae 1992). The five broad factors emerging from this research are commonly referred to as Extraversion, Agreeableness, Conscientiousness, Emotional Stability or Neuroticism, and Intellect or Openness. In the past decades, several types of research have shown that the Big Five can also be used as a model to represent individual differences in children.

90

Dimensional Models of Personality Disorders

Teacher and Parent Ratings Although, early in his career, Digman was convinced that at least 10 common factors could be extracted from teacher ratings of children on the Cattell scales, subsequent reanalysis (Digman and Takemoto-Chock 1981; Goldberg 2001) recovered the five-factor structure from teachers’ nominations of sixth-grade children on the Cattell scales. Mervielde et al. (1995) confirmed the validity of the Big Five as a model for individual differences among school-age children, using a set of 25 bipolar scales derived from Goldberg’s markers for the Big Five. Principal components analysis of ratings of kindergarten children (ages 4–6 years) yielded four of the five factors, including a combined Conscientiousness/Intellect/Openness factor. The complete five-factor structure emerged from the ratings of primary-school children. Several studies have recovered the Big Five from scales or item sets that were not constructed as measures of the Big Five, thus providing evidence that “prestructuring” is not a prerequisite for the emergence of a five-factor structure (Digman and Shmelyov 1996; John et al. 1994).

Natural-Language Personality Descriptions Building on the success of the Five-Factor Model and the lexical approach to detecting broad and cross-culturally replicable dimensions of adult personality, Kohnstamm et al. (1998) collected oral parental personality descriptions of 2,416 children ages 2–12 years as part of a collaborative international research project conducted in Belgium, China, Germany, Greece, Holland, Poland, and the United States. A literal transcription of each oral interview was segmented in small units that described individual differences and assigned them to 14 major categories, including the Big Five. Across countries, 76%–85% of the parental descriptors were classified as indicators of the Big Five. Extraversion was the most frequently used category in each of the seven countries, and Agreeableness was the second most frequently used category; Conscientiousness and Emotional Stability attracted on average slightly less than 10% of the descriptors; and the percentage of descriptors coded as indicating Openness/Intellect varied from 11% in Greece to more than 21% in the United States. To check the content validity at the level of the individual parent, the sampling frequencies for the Big Five categories were assessed for each of the parents. Analysis of the combined data set from the seven countries showed that almost 68% of the parents referred to at least four of the Big Five in describing their child, hence illustrating that these dimensions represent salient personality description categories. Taking into account the fact that parents are a primary source of information for psychiatric assessment of psychopathology in children, this analysis of natural-language descriptions indicates that the Big Five could offer a comprehensive and useful framework for structuring the way that parents naturally describe their children’s personality traits.

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

91

Hierarchical Personality Inventory for Children The construction of the Hierarchical Personality Inventory for Children (HiPIC) is based on Flemish data collected as part of the international research program on parental descriptions of child personality (Kohnstamm et al. 1998). The set of more than 9,000 Flemish parental descriptions was organized into 100 clusters, covering three age groups (5–7, 8–10, and 11–13 years). Three age-specific item sets were developed by writing 2–4 items per cluster, resulting in three preliminary inventories. Principal components analyses at the item level indicated that for each age level, the first five principal components tended to group items according to the Big Five. Because of the substantial overlap in content, the three age-specific item sets were integrated into a single instrument measuring individual differences in personality of primary-school children (Mervielde and De Fruyt 2002). The HiPIC groups 144 items at the highest level into five domains: Conscientiousness, Benevolence, Extraversion, Imagination, and Emotional Stability. The five broad domains were further subdivided into 18 facets, each measured by 8 items. The factor structure of the 18 facets was cross-validated in a sample of 719 twins and siblings rated by both parents (Mervielde and Asendorpf 2000). The four Conscientiousness facets (Achievement Motivation, Concentration, Perseverance, and Orderliness) were relatively pure markers of the first principal component. Egocentrism and Irritability were the highest-loading facets and the purest markers for the Benevolence factor, whereas Compliance operated more like a blend of Benevolence and Conscientiousness. Dominance and Altruism combined a primary loading on Benevolence with a substantial secondary loading on Extraversion. Shyness was the highest-loading facet of Extraversion, with a moderate loading on Emotional Stability. Two Extraversion facets, Optimism and Expressiveness, broaden the scope of this factor and can be considered as indicators of Positive Emotionality. Although Activity is a separate dimension in several temperament models, the related Energy facet did not acquire the status of a separate factor in the item set derived from free parental descriptions. Creativity and Curiosity are the defining facets of the Imagination factor, together with Intellect. Finally, Emotional Stability turns out to be the smallest HiPIC component that is subdivided in an Anxiety facet and one that measures self-confidence. Although the broad factors emerging from the several types of variable-centered studies of children are comparable to the Big Five typically appearing in studies of adult personality, the study of natural parental descriptions—and, in particular, the construction of the HiPIC—revealed a rather broad Agreeableness factor that was labeled Benevolence to mark this difference. The Benevolence factor combines facets such as Egocentrism and Compliance, which are comparable to typical adult Agreeableness facets, but also the Dominance facet, which is related to the Assertiveness facet marking the Extraversion factor in the Neuroticism–Extraversion–Openness Personality Inventory—Revised (NEO PI-R). Finally, Irritability

92

Dimensional Models of Personality Disorders

is a facet of Benevolence, whereas the related facet of Angry Hostility is part of the Neuroticism factor in the adult NEO PI-R.

A Common Taxonomy of Temperament and Personality Traits Based on an extensive review of the literature, Shiner (1998), Shiner and Caspi (2003), and Caspi et al. (2005) proposed a preliminary taxonomy of personality differences in childhood and adolescence. Given the rather diverse theoretical frameworks that inspired the different temperament and personality models, the development of such a taxonomy serves several purposes. It improves communication among researchers who use different concepts to refer to related phenomena and helps to integrate research findings stemming from different research traditions and disciplines, such as personality and developmental psychology. Moreover, it guides the evaluation of the status of time-honored and newly developed instruments by locating their position in the hierarchical taxonomy and relating them to the variables postulated at different levels of the taxonomy. As is evident from the wealth of research that was generated by the growing consensus on the Five-Factor Model, the development of a common taxonomy not only enables integration of diverse research findings but also is a major source for generating new hypotheses and exciting new research about individual differences. The latest version of this taxonomy (Caspi et al. 2005) has two levels. The highest level postulates five higher-order traits: Extraversion/Positive Emotionality, Neuroticism/Negative Emotionality, Conscientiousness/Constraint, Agreeableness, and Openness-to-Experience/Intellect. As is evident from the labels for the higher-level traits, they combine the major dimensions emerging from studies of temperament and the adult Big Five. Both the temperament and the Big Five studies show consistent evidence for individual differences in Extraversion/Positive Emotionality. This “superfactor” encompasses several lower-order traits. Social Inhibition or Shyness refers to feelings of discomfort in the presence of strangers and over the course of development, becomes differentiated from shyness with known others. Sociability refers to the preference to be with others; it should not be considered as the opposite of shyness but rather as a distinct lower-order trait that correlates with the expression of positive emotions. The taxonomy also emphasizes Dominance as a distinct lower-order trait for Extraversion/Positive Emotionality. This corresponds with the location of Assertiveness as a facet of Extraversion in the Five-Factor Model, as well as with the positioning of this lower-order trait in several lexical studies. In the empirically derived HiPIC taxonomy, Dominance is located as a facet of Agreeableness, loading items such as “acting as the boss,” presumably reflecting the salient negative connotation of dominance, at least in the eyes of most parents. The final lower-

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

93

order trait of the first factor of this taxonomy is labeled Energy/Activity Level. Although this lower-order trait is clearly a major dimension in most temperament models, its designation as a lower-order trait is warranted because over the course of development, the salient motor activity of infants is gradually transformed into social activity indicative of the broad Extraversion trait. The Neuroticism/Negative Emotionality higher-order trait refers to emotional reactions and, in particular, to Negative Emotionality. As is evident from our review, Negative Emotionality is a prominent dimension in each of the four reviewed temperament models. It is also an important factor in adult personality models such as the Five-Factor Model and is related to virtually every personality disorder (Widiger et al. 2002). This superfactor is further subdivided into Anxious Distress, a tendency to experience negative emotions mainly targeted at the self, such as anxiety, guilt, and fear. This lower-order trait seems to match the Anxiety facet of the higher-order Emotional Stability factor in the HiPIC. The lower-order trait Irritable Distress refers to distress that is directed toward others, such as anger, frustration, hostility, and irritability. As mentioned before, in the HiPIC model, which is derived from parental descriptions of children, this other-directed form of distress primarily loads the Benevolence factor and is best captured by the Irritability facet. Early in development, the two forms of distress may be difficult to distinguish; hence, both forms of distress may follow different pathways, whereby anxious distress eventually leads to internalizing disorders and irritable distress to externalizing disorders. The third higher-order trait of the taxonomy, labeled Conscientiousness/Constraint, taps individual differences in self-control versus behavioral impulsivity, attention, achievement motivation, orderliness, responsibility, and conventionality. In the Thomas and Chess temperament model, this factor is referred to as Persistence, whereas in the Rothbart-Derryberry model its central role is emphasized by labeling it as Effortful Control. It should be noted, however, that although Conscientiousness/Constraint does not emerge as a major higher-order dimension in the Buss-Plomin model and the Goldsmith-Campos model, it is clearly present in all HiPIC-based analyses of the structure of individual differences in children and adolescents. Hence, once again, this may suggest that Conscientiousness/Constraint is less well differentiated from other higher-order dimensions at an early age. The analysis of parental descriptions in different countries showed that the frequency of parental descriptions indicating Conscientiousness dramatically increases when children enter primary education (Kohnstamm et al. 1998). The third higher-order personality trait is further subdivided in six lower-order traits. Attention is a prominent lower-order trait in many temperament models, and low scores on this trait could be regarded as a step in the developmental pathway toward attention-deficit/ hyperactivity disorder. Self-Control versus Behavioral Impulsivity is a major trait that marks the development of externalizing problem behavior. Achievement Motivation—and, in particular, the lack of it—may contribute to various educational problems, such as lack of interest in formal education, skipping school, and even-

94

Dimensional Models of Personality Disorders

tually dropping out. Orderliness reflects a propensity to be neat, clean, and organized (versus sloppy and disorderly) and forms the empirical core of most factoranalytically derived models of Conscientiousness, including the HiPIC model. Responsibility reflects a blend of Conscientiousness and Agreeableness and ranges from the tendency to be reliable and dependable to being undependable. Finally, Conventionality taps the tendency to uphold traditions and societal norms and serves as one of the strongest predictors of avoiding risky behaviors such as excessive drug and alcohol consumption. The fourth dimension of this taxonomy, Agreeableness, is not present as a major dimension in any of the reviewed temperament models but is a prominent dimension in Big Five–inspired research for both adults and children. Children high on Agreeableness are cooperative, kind, compliant and considerate, whereas those scoring low on this higher-order trait tend to be egocentric, aggressive, rude, and antagonistic. This factor is the broadest and largest dimension emerging from the research on parental descriptions; it presumably reflects the parental concern with managing the child’s behavior and parent–child conflict. Shiner and Caspi (2003) distinguished Antagonism and Prosocial Tendencies as lower-order traits but emphasized that altruistic or Prosocial Tendencies do not necessarily constitute the opposite pole of Antagonism. Caspi et al. (2005) include Openness-to-Experience/Intellect as the fifth higher-order trait in their most recent proposal for the taxonomy. This brings the taxonomy fully in line with the Big Five factors emerging from lexical research, with the Five-Factor Model (Costa and Widiger 2002), and with the previously presented and empirically derived HiPIC model. Openness refers to imaginative, creative, and aesthetically sensitive behavior, while Intellect taps features such as being quick to learn, clever, and insightful. Both aspects are clearly represented in the HiPIC Imagination factor, which has three facets: creativity, curiosity, and intellect. As previously noted, the Intellect component may be difficult to distinguish from Openness and from Conscientiousness, particularly at younger ages. This overview of temperament and personality models of individual differences in childhood reveals that the apparent gap between the temperament models and the Big Five–inspired personality models is more apparent than real. Three major temperament dimensions—Emotionality, Extraversion, and Persistence—have clear counterparts in Big Five research conducted both with children and with adults. Although Agreeableness is not a prominent dimension in the temperament models, it is an important superfactor in the personality literature for both children and adults. Openness is the most controversial dimension, because it is entirely ignored by temperament models but still emerges as a significant but narrow dimension describing differences among children as assessed by the cross-cultural study on parental descriptions and the Big Five–inspired research with children and adolescents. As suggested by the analysis of teacher ratings of children ages 3–12 years (Mervielde et al. 1995), creativity and curiosity may only be readily distinguished

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

95

from educational achievement by the end of the primary-school years and hence emerge as a definite separate dimension during adolescence (McCrae et al. 2002).

Dimensional Psychopathology Models The study of psychopathology in childhood and adolescence has been largely inspired by the success of DSM as a system for classification of adult psychopathology. Although there is little doubt that school-age children and adolescents can suffer from psychiatric impairments, the debate about the validity of the diagnosis of these disorders in preschool children and infants is not closed. A Task Force on Research Diagnostic Criteria has recently proposed research diagnostic criteria for infants and preschool children for various disorders (Scheeringa et al. 2003). Critics of this approach would argue that this categorical conceptualization of psychopathology for children suffers from the same problems that are typical for the diagnosis of adult Axis II disorders, such as high rates of comorbidity, heterogeneity within disorder categories, arbitrary thresholds, and lack of clear boundaries between Axis I and II disorders as well as between maladaptive and normal personality (Kupfer et al. 2002; Widiger and Clark 2000). These problems might become even more important when a categorical approach is adopted to study psychopathology in childhood, because 1) there is a tendency to take adult syndromes as targets and to stretch the criteria for younger ages; 2) the limited language and cognitive abilities of developing children prohibit the use of extensive structured interviews and necessitate the reliance on parents and teachers as informants; 3) the restricted behavioral repertoire of children results in broader or more overlapping categories; 4) developmental changes—and, in particular, heterotypical developmental patterns—change the significance and meaning of behavioral indicators and symptoms over the course of development; and 5) it is difficult to distinguish between young children’s stable symptoms and transient responses to environmental adversity. Finally, most models for individual differences in temperament and personality are clearly dimensional models; therefore, it seems appropriate to concentrate on dimensional models for psychopathology in childhood and adolescence. Two dimensional models will be discussed: the Achenbach System of Empirically Based Assessment (ASEBA) and the Dimensional Personality Symptom Item Pool (DIPSI), a new dimensional model for assessment of maladaptive traits in children and young adolescents that is currently being developed at Ghent University (De Clercq et al. 2006).

ACHENBACH SYSTEM One of the most widely used instruments for the broad-spectrum screening of children’s and adolescents’ emotional and behavioral problems is Achenbach’s em-

96

Dimensional Models of Personality Disorders

pirically based assessment system (Achenbach 1991), consisting of the Child Behavior Checklist (CBCL), the Teacher Report Form (TRF) and the Youth SelfReport (YSR). Achenbach developed eight cross-informant syndrome constructs labeled Anxious/Depressed, Withdrawn, Somatic Complaints, Social Problems, Thought Problems, Attention Problems, Delinquent Behavior (rule-breaking behavior), and Aggressive Behavior. Two second-order scales were derived: Internalizing, grouping the syndrome scales Anxious/Depressed, Withdrawn, and Somatic Complaints, and Externalizing, grouping the syndrome scales Delinquent Behavior and Aggressive Behavior. The Internalizing and Externalizing scales are often used in research and provide measures of emotional symptoms and antisocial/conduct problems, respectively. The CBCL is a typical example of a bottom-up approach in which taxonomic constructs are derived from multivariate analyses of large samples. The term syndrome in this assessment system refers to concurrent and associated problems that are reported for children and adolescents. This inductive approach stands in contrast with the top-down or deductive approach adopted for the construction of the DSM (Achenbach 1995). Although the empirically based taxonomies and the DSM nosology adopt a different scientific methodology, the resulting taxonomies are clearly related (Ferdinand et al. 2004). The latest version of Achenbach’s empirically based assessment system further elaborates and strengthens this link by complementing the syndrome scales with a set of DSM-oriented scales comprising the items that a substantial majority of child psychiatrists rated as very consistent with particular DSM-IV categories. Both the DSM-oriented scales and the empirically based scales show significant associations with DSM-IV clinical diagnoses and other standardized rating scales (Achenbach et al. 2003). The new DSM-oriented scales are an improvement over the original CBCL syndrome scales because they better correspond to current conceptualizations of dimensions of child psychopathology and include items that reflect more closely the DSM-IV criteria. This extension of the classical CBCL with DSM-oriented scales not only documents the overlap between both systems but bridges the gap between the bottom-up and the top-down approach; it also opens new perspectives to search for a set of common dimensions that adequately represent both systems in an integrated and comprehensive model for childhood psychopathology.

DIPSI TAXONOMY OF CHILDHOOD PERSONALITY SYMPTOMS Although it is widely acknowledged that personality disorders as conceptualized in DSM-IV may have antecedents in childhood temperament and personality, most research on the antecedents of adult psychopathology focuses on the relationship between childhood emotional and behavioral problems and Axis I psychopathology (Roza et al. 2003). The previously described construction of DSM-oriented

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

97

scales for the Achenbach system is a clear example of the primary focus on Axis I pathology. However, there is a growing interest for research on the developmental antecedents of Axis II personality disorders (Kasen et al. 1999; Ramklint et al. 2003). Although adherents of the spectrum hypothesis such as Widiger and Clark (2000) suggest that personality disorders are extreme variants of normal adaptive personality traits, the current version of the DSM provides only specific categories and criteria for the diagnosis of these disorders in adulthood. This is remarkable, because the stability of personality across the life span has now been well documented (Roberts and Del Vecchio 2000; Roberts et al. 2001) and hence it is reasonable to propose that the extreme manifestations of personality traits in adulthood should have clear counterparts at earlier stages in development.

Assembly of the Dimensional Personality Symptom Item Pool Taking into account recent formulations of the spectrum hypothesis as well as the growing evidence on the stability of personality across the life span, our research group set out to fill in this gap by constructing an item pool with personality symptoms targeted at school-age children (ages 6–14 years). In line with the spectrum hypothesis, 333 extreme variants (maladaptive items) were written covering both poles of the 144 items from the previously described HiPIC. Given that the fifth factor remains the most controversial, is absent from the entire temperament literature, and only plays a minor role in predicting adult psychopathology, we decided not to include maladaptive variants primarily loading this factor. Those who advocate replacing the DSM-IV classification with an alternative system based on the five-factor model are often countered with the criticism that this model is not clinically relevant. Critics contend that the model reduces personality description to concepts derived from factor-analysis of self- and peer ratings provided by laypersons lacking exposure to the kind of personality pathology that professionals encounter in their daily practice (Shedler and Westen 2004). Given that the HiPIC is based on parental description of child personality, our approach is surely liable to be criticized for the same reasons. Therefore, it was decided to screen two instruments that assess the DSM-IV criteria for personality disorders in adulthood, for child-relevant items: The Structured Clinical Interview for DSM-IV Axis II personality disorders (SCID-II) and the Assessment of DSMIV Personality Disorders (ADP-IV) (Schotte et al. 1998). As the result of this screening, 83 ADP-IV items and 90 SCID-II items were added as potential descriptors of child personality symptoms, bringing the total number of DIPSI descriptors to 503. This total set of items was subsequently classified by the authors in 44 categories and four to six items for each category were written, resulting in the first version of the DIPSI, containing 256 items. Additional construction strategies leading to the final DIPSI are more extensively described in De Clercq et al. (2006).

98

Dimensional Models of Personality Disorders

Psychometric Properties of the DIPSI The reliability of the DIPSI categories was determined with a sample of 205 children referred for psychosocial problems (118 boys, 87 girls; mean age 118.92 months; age range 5.5–14 years). The initial alpha coefficients for the 44 categories were improved by reassigning the DIPSI items. The revised version of the DIPSI included 222 items, organized in 40 trait-pathology categories, with alpha coefficients ranging from .57 (Social avoidance) to .89 (Depressive traits), and with a median value of .82. An exploratory factor analysis of the 40 categories suggested either a two- or a four-factor solution, explaining 44.4% and 57.6% of the total variance, respectively. In decreasing order of explained variance, the two-factor solution represented an Externalizing (23.5%) and an Internalizing (20.9%) trait factor, whereas the four factors were labeled as Disagreeableness (22.7%), Emotional Instability (18.3%), Introversion (9.0%), and Compulsivity (7.6%).

Content Validity To examine the comprehensiveness of the 40 DIPSI categories, we compared them with the scales from Livesley’s Dimensional Assessment of Personality Psychopathology—Basic Questionnaire (DAPP-BQ; Livesley 1990), a dimensional instrument for adult personality pathology. This comparison showed a striking correspondence between the DIPSI categories and the DAPP-BQ scales. For all DAPPBQ scales, the DIPSI provides parallel categories, except for Intimacy problems and Self-harming behaviors. In the DIPSI, Intimacy problems are partly assigned to Social avoidance. A self-harm category is absent, because such items were intentionally excluded from the item pool, given the young age of the primary target group (early and middle childhood). In addition to analogues of the DAPP-BQ, the DIPSI contains supplementary categories grouping Hyperactive traits, Impatience, Distraction, Hyperexpressivity, Depressive traits, Obsessive traits, and Unforgivingness.

Concurrent Validity As a dimensional measure of childhood psychopathology, the DIPSI is expected to share common variance with the Dutch CBCL (Verhulst et al. 1996), the most widely used instrument for screening children on psychopathology. The Disagreeableness factor of the four-factor solution is strongly related to the CBCL syndromes Aggressive Behavior (.83), Delinquent Behavior (.61), and Attention Problems (.55) and has a moderate correlation with Social Problems (.24). Emotional Instability is positively related to all the CBCL syndromes, except to Delinquent and Aggressive Behavior. Introversion shows a strong positive correlation with Withdrawn behavior (.52) and is also correlated with Delinquent Behavior (.28) and Anxious/Depressed (.27). Compulsivity is positively correlated with Somatic Complaints (.24) and Anxious/Depressed (.22) and negatively correlated with Attention problems (–.20). The higherorder DIPSI Internalizing factor correlates with CBCL Withdrawn Behavior, Somatic

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

99

Complaints, Anxious/Depressed, and Social and Thought Problems, whereas the DIPSI Externalizing factor significantly predicts Attention Problems, Delinquent Behavior, and Aggressive Behavior. The DIPSI two-factor solution shows a .84 correlation between the higher-order DIPSI Externalizing factor and CBCL Externalizing problems and a .73 correlation between DIPSI Internalizing and CBCL Internalizing problems. These results support the concurrent validity of the higher-order DIPSI factors and suggest that maladaptive child personality dimensions are differentially related to symptomatic behavior as captured by the CBCL syndromes. Although threequarters of the DIPSI categories and items are based on extreme variants of HiPIC adaptive traits, the current analysis shows a high degree of overlap between the two higher-order DIPSI factors and the broad-spectrum Internalizing and Externalizing factors of the CBCL. This corroborates the spectrum hypothesis that childhood psychopathology can be conceptualized as a manifestation of extreme variants of individual differences in temperament/personality as measured with the HiPIC, a comprehensive measure of adaptive individual differences in children.

The Relationship Between Personality and Psychopathology In line with our aim to pay sufficient attention to the broader picture, we conclude this chapter with data gathered by our research group that show relationships between—on the one hand—personality in childhood and adolescence measured with instruments such as the HiPIC and the NEO PI-R, which capture adaptive or normal personality, and—on the other hand—the two-dimensional psychopathology models and DSM-IV–based instruments, which measure personality disorders.

RELATIONSHIPS BETWEEN THE HIPIC, THE CBCL, AND THE DIPSI The spectrum hypothesis would lead one to expect substantial correlations between the HiPIC, a measure of adaptive personality differences, and the CBCL and the DIPSI measuring child problem behavior and maladaptive traits. One would further expect the pattern of correlations to be similar in a nonreferred sample of children and in a clinic-referred sample, but to find differences between nonreferred and clinic-referred groups in the magnitude of the correlations. These relationships were investigated in two samples: 1. The nonreferred general-population sample of children (276 boys, 320 girls; mean age 10.9 years, range 7–15 years), rated by their mothers, is partly from a longitudinal study investigating parenting, personality characteristics, and children’s problem behavior (Van Leeuwen et al. 2004b).

100

Dimensional Models of Personality Disorders

2. The clinic-referred sample consisted of 205 mothers and children (118 boys, 87 girls; mean age: 9.9 years; range: 5–14 years) recruited from various outpatient mental health services. The top section of Table 10–1 shows the correlations between HiPIC domain scores and CBCL total scores on Internalizing and Externalizing problems for both samples. This part of the table suggests substantial negative correlations between CBCL Externalizing and HiPIC Benevolence and Conscientiousness for both samples and negative correlations between CBCL Internalizing and HiPIC Emotional Stability and Extraversion. In the general-population sample, the HiPIC domains are differentially related to the higher-order CBCL dimensions but the discrimination is not clear-cut, because both higher-order factors show moderate correlations with the other HiPIC domains. In the clinic-referred sample, however, the pattern of correlations is much more differentiated; CBCL Internalizing is clearly and exclusively related to Emotional Stability and Extraversion, whereas Externalizing shares substantial variance with Benevolence and Conscientiousness. Apparently the same basic pattern is present in both types of samples, but the relationships are stronger and more specific in the clinical sample. From the viewpoint of the spectrum hypothesis, one would expect clear mean-level differences between a general population and a clinical sample (on both personality and psychopathology measures), but what is observed here is that the strength of the personality–psychopathology relationships depends on the nature of the sample. The bottom six rows of Table 10–1 report the relationships between the HiPIC domain scores and the DIPSI factor scores for the four- and the two-factor solutions extracted from maternal ratings of children in the clinical sample. The HiPIC domain Extraversion is negatively correlated with DIPSI Introversion and Emotional Instability and positively correlated with DIPSI Disagreeableness. HiPIC Benevolence is strongly and negatively related to DIPSI Disagreeableness, whereas HiPIC Emotional Stability is strongly and negatively related to Emotional Instability and moderately to Disagreeableness. Conscientiousness is positively related to Compulsivity and negatively related to Disagreeableness. Imagination is negatively related to every DIPSI factor except Compulsivity. Overall, this pattern of correlations reflects what is to be expected given that three-fourths of the DIPSI items are extreme variants of HiPIC items. The pattern of correlations between the HiPIC and the two-factor DIPSI scores (Internalizing and Externalizing) reported in the last two rows of the table is rather similar to the pattern observed in the top four rows. These results support the validity of the higher-order DIPSI factors and suggest not only that maladaptive child personality dimensions are differentially related to HiPIC personality dimensions but also that the higher-order factors extracted from the DIPSI show a pattern of correlation with HiPIC personality that is rather similar to the pattern observed for the CBCL Internalizing and Externalizing factors.

101

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

TABLE 10–1.

Relationships between HiPIC domains and CBCL and DIPSI higher-order dimensions HiPIC domains BEN

CON

EMO

EXT

IMA

–.26*** –.63***

–.24*** –.40***

–.53*** –.12***

–.32*** .06

–.22*** –.13***

–.08 –.70*** –.10 .10

.03 –.36*** –.08 .08

–.61*** –.01 –.07 –.69***

–.39*** .16* –.51*** –.20*

–.06 –.14* –.19* –.16*

–.77*** .04 .09 –.78***

–.56*** .58*** .21** –.56***

.16* –.03 –.67*** .16*

.38* .12 –.34*** .31***

–.14* .45*** –.10 –.15

General-population sample 1 CBCL Internalizing CBCL Externalizing Clinic-referred sample2 CBCL Internalizing CBCL Externalizing DIPSI Introversion DIPSI Emotional Instability DIPSI Disagreeableness DIPSI Compulsivity DIPSI Internalizing DIPSI Externalizing

Note. HiPIC =Hierarchical Personality Inventory for Children; CBCL = Child Behavior Checklist; DIPSI = Dimensional Personality Symptom Item Pool. HiPIC domains: BEN = Benevolence; CON = Conscientiousness; EMO = Emotional Stability; EXT = Extraversion; IMA = Imagination. 1General-population sample (n=578); mean maternal ratings. 2Clinic-referred sample (n=205); maternal ratings. *P<0.05; **P<0.01; ***P<0.001.

RELATIONSHIP BETWEEN FFM PERSONALITY DIMENSIONS AND PERSONALITY DISORDERS IN ADOLESCENCE Saulsman and Page (2004) conducted a meta-analysis of relationships between FFM domains and personality disorders in adults. Five main conclusions can be drawn from this meta-analysis: 1. Neuroticism and Agreeableness are most consistently related, in terms of direction and magnitude of the correlation, to personality disorders. 2. Extraversion and Conscientiousness are also substantially related to personality disorders, but the direction of the correlation is disorder-dependent. 3. Openness to Experience is not prominently associated with personality disorder variance.

102

Dimensional Models of Personality Disorders

4. Domain–disorder relationships are similar across clinical and nonclinical samples and relatively independent of the measures used to assess traits and personality disorders. 5. FFM traits demonstrate both statistical significance and practical usefulness in describing personality disorder variance, despite their aclinical item content. De Clercq and De Fruyt (2003) and De Clercq et al. (2004) recently examined self-rated FFM personality–personality disorder relationships in two independent samples of adolescents. Both of these studies adopted a different measure to assess personality in adolescents, using the NEO PI-R and the previously described HiPIC, and both used the ADP-IV (Schotte et al. 1998) to assess Axis II personality disorders. Table 10–2 compares the significant personality–personality disorder correlations culled from these studies and that of Saulsman and Page (2004). An indication of the similarity of personality–personality disorder correlations across the three studies was obtained by correlating meta-analysis results with the 2003 and 2004 results across disorders and the five domains. This analysis reveals correlations of .83 (P < 0.001) between meta-analysis results and the NEO PI-R study, .84 (P<0.001) between meta-analysis patterns and the HiPIC study, and .86 (P< 0.001) between the personality–personality disorder pattern observed in both studies with adolescents or between HiPIC- and NEO PI-R– based personality–personality disorder correlations. Overall, these data suggest that the personality–personality disorder correlations observed in the three studies are remarkably similar. A fine-grained analysis by personality factor shows the highest personality–personality disorder correspondences across the three studies for Extraversion and Conscientiousness. In sum, the conclusions of Saulsman and Page (2004) are thus to a large extent applicable to adolescence: 1. Agreeableness and Neuroticism show uniformly negative and positive associations, respectively, with personality disorders. 2. Conscientiousness shows negative correlations with some personality disorders but a positive correlation with obsessive-compulsive disorder. 3. Openness to Experience does not substantially contribute to explaining personality disorder variance. The major discrepancy from the adult association patterns was found for Extraversion, which was positively related to histrionic and narcissistic personality disorders in adults but did not show similar associations in adolescents.

Patterns of significant relationships between FFM personality dimensions and Axis II personality disorders in adulthood and

adolescence Five-Factor Model (FFM) personality dimensions Neuroticism

Extraversion

Openness

M

A1

A2

M

A1

A2

M

Paranoid

+

+

+

–

–

–

–

Schizoid

+

+

–

–

–

–

–

–

–

Schizotypal

+

+

Antisocial

+

+

Borderline

+

+

+

–

+

+

+

Histrionic Narcissistic

+

A1

Agreeableness

A2

– +

+

+

–

– +

+

+

Avoidant

+

+

+

–

–

–

–

–

Dependent

+

+

+

–

–

–

–

–

Obsessive-Compulsive

+

+

+

–

–

–

–

–

Conscientiousness

M

A1

A2

M

A1

–

–

–

–

–

–

–

–

–

A2

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

+

–

–

–

+

–

– +

Notes. +/–=significant positive or negative relationship at P≤ 0.01; M=meta-analytic adult data based on Saulsman and Page 2004a; A1=adolescent data based on De Clercq and De Fruyt 2003; A2 = adolescent data based on De Clercq et al. 2004.

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

TABLE 10–2.

103

104

Dimensional Models of Personality Disorders

The Broad Perspective Although different FFM measures and psychopathology measures are used in the studies discussed above, a general pattern of associations between personality and psychopathology emerges. Two broad dimensions can perhaps best represent the highest-level psychopathology: Internalizing and Externalizing. Individual differences in personality and temperament can be reasonably well covered by the five higher-order traits proposed in the common taxonomy that are clearly related to the factors of the Five-Factor Model (i.e., Neuroticism, Extraversion, Agreeableness, Conscientiousness, and Openness). Figure 10–1 presents a bold proposal for how personality and psychopathology are related at the highest level, particularly in studies of children and adolescents. The left part of the figure illustrates the ties between the variable-centered and the person-centered approaches to the study of individual differences. Overcontrollers can be profiled as subjects who score low on Emotional Stability and Extraversion. Undercontrollers, by contrast, tend to score low on Agreeableness and Conscientiousness. Both types are less well adapted and show more problem behavior than Resilient children (Van Leeuwen et al. 2004a). As was evident from our studies on the FFM, the four dimensions are clearly related to dimensions emerging from psychopathology studies. At the highest level, Agreeableness and Conscientiousness are clearly related to Externalizing, as measured with either the CBCL or the DIPSI, whereas high Neuroticism and low Extraversion are the typical correlates of Internalizing problem behavior. The four lower-order DIPSI dimensions listed on the right provide a bridge between the highest-level psychopathology dimensions and the four FFM dimensions. Moreover, the four DIPSI dimensions are strikingly similar to the dimensions proposed by Livesley and colleagues (Livesley 1990; Livesley et al. 1992) to organize the scales and subcategories of the DAPP-BQ, an instrument validated mainly in adult samples. This correspondence between dimensions of childhood and adult psychopathology is, of course, a large step toward an even broader, across-the-life span perspective on the relationships between both disciplines. The task we set out to accomplish was to show evidence for a dimensional conceptualization of personality/temperament and psychopathology in childhood. The higher-order dimensions of Internalizing and Externalizing are familiar to child psychologists because of their prominent role in the CBCL; hence, critics of this approach could argue that these dimensions may be suitable for the conceptualization of childhood psychopathology but not for adult psychopathology. However, a number of recent studies, published for the most part in psychiatric journals by Krueger and his Minnesota Twin Study group (Krueger 1999; Krueger et al. 1998, 2001; Tackett et al. 2003), have argued that in addition to their utility in organizing Axis II personality disorders, the Internalizing and Externalizing dimensions are also primary contenders for organizing Axis I pathology. These studies once again broaden the perspective, emphasizing that these two broad dimen-

TYPES

RESILIENTS

High

FFM/BIG FIVE

TYPES

Emotional Stability Low

High

Extraversion

High

Openness

High

Agreeableness

Low

CBCL-2/DIPSI-2

Emotional Instability

INTERNALIZING

Introversion

Low

High Conscientiousness Low

FIGURE 10–1.

OVERCONTROLLERS

DIPSI-4

UNDERCONTROLLERS

Disagreeableness

EXTERNALIZING

Compulsivity

Linking person- and variable-centered approaches with psychopathology models.

FFM = Five-Factor Model (“Big Five”); DIPSI = Dimensional Personality Symptom Item Pool; CBCL = Child Behavior Checklist.

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

MALADAPTIVE TRAITS

ADAPTIVE PERSONALITY TRAITS

105

106

Dimensional Models of Personality Disorders

sions not only have proven utility in organizing relatively stable personality-related psychopathology across the life span, but also have potential utility in organizing the more transient Axis I psychopathology.

References Achenbach TM: Integrative Guide for the 1991 CBCL-4–18, YSR, and TRF Profiles. Burlington, VT, University of Vermont, Department of Psychiatry, 1991 Achenbach TM: Empirically based assessment and taxonomy: applications to clinical research. Psychol Assess 7:261–274, 1995 Achenbach TM, Dumenci L, Rescorla LA: DSM-oriented and empirically based approaches to constructing scales from the same item pools. J Clin Child Adolesc Psychol 32:328–340, 2003 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Asendorpf JB: Head-to-head comparison of the predictive validity of personality types and dimensions. Eur J Pers 17:327–346, 2003 Asendorpf JB, Borkenau P, Ostendorf F, et al: Carving personality description at its joints: confirmation of three replicable personality prototypes for both children and adults. Eur J Pers 15:169–198, 2001 Block JH, Block J: The role of ego-control and ego-resiliency in the organization of behavior, in Minnesota Symposium on Child Psychology, Vol 13. Edited by Collins WA. Hillsdale, NJ, Lawrence Erlbaum, 1980, pp 39–101 Buss AH, Plomin R: A temperament theory of personality development. New York, Wiley, 1975 Caspi A, Roberts BW, Shiner RL: Personality development: stability and change. Annu Rev Psychol 56:56:453–484, 2005 Costa PT Jr, Herbst JH, McCrae RR, et al: The replicability and utility of three personality types. Eur J Pers 16:S73–S87, 2002 Costa PT Jr, McCrae RR: The five-factor model of personality and its relevance to personality disorders. J Personal Disord 6:343–359, 1992 Costa PT Jr, Widiger TA (eds): Personality Disorders and the Five-Factor Model of Personality. Washington, DC, American Psychological Association, 2002 De Clercq B, De Fruyt F: Personality disorder symptoms in adolescence: a five-factor model perspective. J Personal Disord 17:269–292, 2003 De Clercq B, De Fruyt F, Van Leeuwen K: A little five lexically based perspective on personality disorder symptoms in adolescence. J Personal Disord 18:477–496, 2004 De Clercq B, De Fruyt F, Van Leeuwen K: The structure of maladaptive personality traits in childhood: a step toward an integrative developmental perspective for DSM-V. J Abnorm Psychol 115, 2006 De Fruyt F, Mervielde I, Van Leeuwen K: The consistency of personality type classification across samples and five-factor measures. Eur J Pers 16:S57–S72, 2002 Derryberry D, Rothbart MK: Reactive and effortful processes in the organization of temperament. Dev Psychopathol 9:633–652, 1997

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

107

Digman JM, Shmelyov AG: The structure of temperament and personality in Russian children. J Pers Soc Psychol 71:341–351, 1996 Digman JM, Takemoto-Chock NK: Factors in the natural-language of personality—re-analysis, comparison, and interpretation of 6 major studies. Multivar Behav Res 16:149– 170, 1981 Ferdinand RF, Visser JH, Hoogerheide KN, et al: Improving estimation of the prognosis of childhood psychopathology; combination of DSM-III-R/DISC diagnoses and CBCL scores. J Child Psychol Psychiatry 45:599–608, 2004 Goldberg LR: Analyses of Digman’s child-personality data: derivation of Big-Five factor scores from each of six samples. J Pers 69:709–743, 2001 Goldsmith HH, Campos JJ: Toward a theory of infant temperament, in The Development of Attachment and Affiliative Systems. Edited by Emde RN, Harmon RJ. New York, Plenum, 1982, pp 161–193 Goldsmith HH, Buss KA, Lemery KS: Toddler and childhood temperament: expanded content, stronger genetic evidence, new evidence for the importance of environment. Dev Psychol 33:891–905, 1997 Hart D, Hofmann V, Edelstein W, et al: The relation of childhood personality types to adolescent behavior and development: a longitudinal study of Icelandic children. Dev Psychol 33:195–205, 1997 John OP, Caspi A, Robins RW, et al: The “little five”: exploring the nomological network of the five-factor model of personality in adolescent boys. Child Dev 65:160-178, 1994 Kasen S, Cohen P, Skodol AE, et al: Influence of child and adolescent psychiatric disorders on young adult personality disorder. Am J Psychiatry 156:1529–1535, 1999 Kohnstamm GA, Halverson CF Jr, Mervielde I, et al: Parental Description of Child Personality: Developmental Antecedents of the Big Five? Mahwah, NJ, Lawrence Erlbaum, 1998 Krueger RF: The structure of common mental disorders. Arch Gen Psychiatry 56:921–926, 1999 Krueger RF, Tackett JL: Personality and psychopathology: working toward the bigger picture. J Personal Disord 17:109–128, 2003 Krueger RF, Caspi A, Moffitt TE, et al: The structure and stability of common mental disorders (DSM-III-R): a longitudinal-epidemiological study. J Abnorm Psychol 107:216– 227, 1998 Krueger RF, McGue M, Iacono WG: The higher-order structure of common DSM mental disorders: internalization, externalization, and their connections to personality. Pers Indiv Differ 30:1245–1259, 2001 Kupfer DJ, First MB, Regier DA (eds): A Research Agenda for DSM-V. Washington, DC, American Psychiatric Association, 2002 Lemery KS, Goldsmith HH, Klinnert MD, et al: Developmental models of infant and childhood temperament. Dev Psychol 35:189–204, 1999 Livesley WJ: Dimensional Assessment of Personality Pathology—Basic Questionnaire. British Columbia, Canada, University of British Columbia, 1990

108

Dimensional Models of Personality Disorders

Livesley WJ, Schroeder ML, Jackson DN: Factorial structure of traits delineating personality disorders in clinical and general-population samples. J Abnorm Psychol 101:432– 440, 1992 McCrae RR, Costa PT Jr, Terracciano A, et al: Personality trait development from age 12 to age 18: longitudinal, cross-sectional, and cross-cultural analyses. J Pers Soc Psychol 83:1456–1468, 2002 Mervielde I, Asendorpf JB: Variable-centred and person-centred approaches to childhood personality, in Advances in Personality Psychology, Vol 1. Edited by Hampson SE. Philadelphia, PA, Taylor & Francis, 2000, pp 37–76 Mervielde I, De Fruyt F: Assessing children’s traits with the hierarchical personality inventory for children, in Big Five Assessment. Edited by De Raad B, Perugini M. Seattle, WA, Hogrefe & Huber, 2002, pp 129–146 Mervielde I, Buyst V, De Fruyt F: The validity of the Big-Five as a model for teacher ratings of individual differences among children aged 4–12 years. Personality and Individual Differences 18:525–534, 1995 Presley R, Martin RP: Toward a structure of preschool temperament—factor structure of the temperament assessment battery for children. J Pers 62:415–448, 1994 Ramklint M, von Knorring AL, von Knorring L, et al: Child and adolescent psychiatric disorders predicting adult personality disorder: a follow-up study. Nord J Psychiatry 57:23–28, 2003 Roberts BW, Caspi A, Moffitt TE: The kids are all right: growth and stability in personality development from adolescence to adulthood. J Pers Soc Psychol 81:670–683, 2001 Roberts BW, Del Vecchio WF: The rank-order consistency of personality traits from childhood to old age: a quantitative review of longitudinal studies. Psychol Bull 126:3–25, 2000 Robins RW, John OP, Caspi A, et al: Resilient, overcontrolled, and undercontrolled boys: three replicable personality types. J Pers Soc Psychol 70:157–171, 1996 Rothbart MK, Ahadi SA: Temperament and the development of personality. J Abnorm Psychol 103:55–66, 1994 Roza SJ, Hofstra MB, van der Ende J, et al: Stable prediction of mood and anxiety disorders based on behavioral and emotional problems in childhood: a 14-year follow-up during childhood, adolescence, and young adulthood. Am J Psychiatry 160:2116–2121, 2003 Saulsman LM, Page AC: The five-factor model and personality disorder empirical literature: a meta-analytic review. Clin Psychol Rev 23:1055–1085, 2004 Scheeringa M, Anders T, Boris N, et al: Research diagnostic criteria for infants and preschool children: the process and empirical support. J Am Acad Child Adolesc Psychiatry 42:1504–1512, 2003 Schotte CKW, de Doncker D, Vankerckhoven C, et al: Self-report assessment of the DSMIV personality disorders. Measurement of trait and distress characteristics: the ADPIV. Psychol Med 28:1179–1188, 1998 Shedler J, Westen D: Dimensions of personality pathology: an alternative to the five-factor model. Am J Psychiatry 161:1743–1745, 2004 Shiner RL: How shall we speak of children’s personalities in middle childhood? A preliminary taxonomy. Psychol Bull 124:308–332, 1998

Mervielde et al.: Temperament/Personality as Antecedents of Psychopathology

109

Shiner RL, Caspi A: Personality differences in childhood and adolescence: measurement development, and consequences. J Child Psychol Psychiatry 44:2–32, 2003 Tackett JL, Krueger RF, Sawyer MG, et al: Subfactors of DSM-IV conduct disorder: evidence and connections with syndromes from the Child Behavior Checklist. J Abnorm Child Psychol 31:647–654, 2003 Thomas A, Chess S: Temperament and Development. New York, Brunner/Mazel, 1977 Van Leeuwen K, De Fruyt F, Mervielde I: A longitudinal study of the utility of the resilient, overcontrolled, and undercontrolled personality types as predictors of children’s and adolescents’ problem behaviour. Int J Behav Dev 28:210–220, 2004a Van Leeuwen K, Mervielde I, Braet C, et al: Child personality and parental behavior as moderators of problem behavior: variable- and person-centered approaches. Dev Psychol 40:1028–1046, 2004b Verhulst FC, Van der Ende J, Koot HM: Handleiding voor de CBCL/4–18 [Manual of the CBCL/4–18]. Rotterdam, The Netherlands, Erasmus Universiteit, 1996 Warner MB, Morey LC, Finch JF, et al: The longitudinal relationship of personality traits and disorders. J Abnorm Psychol 113:217–227, 2004 Widiger TA, Clark LA: Toward DSM-V and the classification of psychopathology. Psychol Bull 126:946–963, 2000 Widiger TA, Trull TJ, Clarkin JF, et al: A description of the DSM-IV personality disorders with the five-factor model of personality, in Personality Disorders and the Five-Factor Model of Personality. Edited by Costa PT Jr, Widiger TA. Washington DC, American Psychological Association, 2002, pp 89–99

This page intentionally left blank

11 COMMENTARY ON MERVIELDE ET AL. Toward a Developmental Perspective on Personality Disorders Rebecca L. Shiner, Ph.D.

I

n recent years, researchers studying normal temperament and personality development have addressed several issues with striking relevance for understanding potential childhood antecedents of adult personality disorders and personality pathology in youths: the commonalities between temperament and personality, the structure of individual differences in children and adolescents, and the stability of personality across the life span. In this commentary, I provide a brief review of each of these issues, and I highlight avenues for integrating knowledge about normal and abnormal personality development.

Commonalities Between Temperament and Personality Traits Temperament researchers generally agree that temperament includes individual differences that emerge early in life, encompass differences in emotional processes,

This chapter is an abbreviated version of a paper first published in the Journal of Personality Disorders (Volume 19, Issue 2, pages 202–210, 2005).

111

112

Dimensional Models of Personality Disorders

and have a biological basis. Typically, personality is understood as emerging later than temperament and as encompassing a wider range of individual differences. Although temperament traits and personality traits are sometimes seen as quite distinct, recent empirical work has demonstrated that childhood temperament traits have much in common with adult personality traits, such as the Big Five. Temperament in infancy and early childhood is influenced by environmental experiences, not only by heredity (Emde and Hewitt 2001), and nearly all selfreported and observed personality traits in adulthood show moderate genetic influences (Bouchard 2004). Thus, most contemporary temperament researchers recognize that childhood temperament is influenced by both heredity and experience (Rothbart and Bates 1998), just as is true for adult personality. Temperament and personality traits share another important feature: Many traits from both domains are characterized by specific, habitual positive and negative emotions (Caspi and Shiner 2006; Rothbart and Bates 1998).

A Common Structure for Personality in Childhood, Adolescence, and Adulthood As Mervielde et al. (Chapter 10 in this volume) note, temperament and personality are also linked in another way: the two domains include many similar traits and share a similar structure. There is now convincing evidence that individual differences in childhood and adolescence share a similar structure to personality in adults, beginning sometime during the preschool years. At the level of broad traits, this structure includes the Big Five traits of Extraversion, Neuroticism, Conscientiousness, Agreeableness, and Openness-to-Experience (Caspi and Shiner 2006; Caspi et al. 2005). Evidence that children and adolescents exhibit traits encompassed in the Big Five model has been provided from a variety of sources—factor analytic studies of personality questionnaires from parents, teachers, and adolescents (Caspi and Shiner 2006); temperament questionnaire studies (Rothbart and Bates 1998); and research on behavioral-task and observational measures of children’s temperaments and personalities (Shiner 1998). Like adult personality traits, children’s individual differences appear to be organized hierarchically (Caspi and Shiner 2006; Putnam et al. 2001). Covariation among specific behavioral descriptors (e.g., tense, fearful) is explained by lower-order traits, and the covariation among these narrow, lower-order traits (e.g., anxiety, sadness) is explained by broad, higher-order traits (e.g., Neuroticism). Below, I describe briefly the nature of the higher-order Big Five traits in children and adolescents, and I note likely lower-order components. Readers are referred to Caspi and Shiner (2006) for a more complete description of this taxonomy, the early childhood antecedents of these traits, and the processes underlying the various traits.

Shiner: Commentary on Mervielde et al.

113

Extraversion in children and adolescents is characterized by tendencies to be sociable, expressive, high-spirited, lively, socially potent, physically active, and energetic. In contrast, introverted youths are quiet, inhibited, and lethargic. Extraversion encompasses the lower-order traits of sociability (talkativeness, gregariousness) and energy/activity level. Another lower-order trait—social inhibition—is related to both Extraversion and Neuroticism; this trait includes feelings of discomfort and reluctance to act in novel situations. Children and adolescents who are high on Neuroticism are described as anxious, vulnerable, tense, easily frightened, “falling apart” under stress, guilt-prone, moody, low in frustration tolerance, and insecure in relationships with others. Fewer descriptors define the lower end of this dimension; these include traits such as stability, being “laid back,” adaptability in novel situations, and the ability to “bounce back” after a bad experience. The lower-order traits associated with Neuroticism in children are somewhat unclear. Research on negative emotions suggests that Neuroticism may include at least three distinct but related lower-order traits in youths: fear, anxiety, and sadness (Muris et al. 2001). Two other lower-order traits are likely to be related to both Neuroticism and (low) Agreeableness: anger/ irritability and alienation/mistrust. Highly Conscientious children and adolescents are described as responsible, attentive, persistent, orderly and neat, planful, possessing high standards, and thinking before acting. Children low on this trait are depicted as irresponsible, unreliable, careless, distractible, and quitting easily. Conscientiousness appears to include a number of lower-order components in childhood: attention, self-control, achievement motivation, orderliness, and responsibility. Agreeableness includes a variety of traits seen as very important by psychologists and psychiatrists; yet, historically, these traits have been left out of temperament models. The high end of Agreeableness includes descriptors such as considerate, empathic, generous, gentle, kind, and protective of others, whereas the low end includes tendencies toward being aggressive, rude, spiteful, stubborn, bossy, cynical, and manipulative. Childhood Agreeableness also includes being willing to accommodate others’ wishes rather than forcing one’s own desires and intentions on others; for children, this aspect of the trait also involves how manageable the child is for parents and teachers. Agreeableness includes several lower-order components: prosocial tendencies (also potentially called helpfulness or nurturance), antagonism versus peacefulness, and willfulness versus flexibility. Agreeableness in children may also include modesty versus conceitedness and integrity (i.e., being honest, principled, and sincere). In Big Five studies, children who are high on Openness-to-Experience are described as eager and quick to learn, clever, knowledgeable, perceptive, imaginative, curious, and original. It appears that Openness can be measured reliably by at least age 6 or 7 years, but the childhood trait does not appear to be as broad as some adult conceptualizations of Openness (Caspi and Shiner 2006). The lower-order

114

Dimensional Models of Personality Disorders

components of Openness in childhood are not yet clear, but 1) intellect and 2) curiosity and creativity have received some support. Although much more work remains to be done to understand the structure of personality in childhood and adolescence, the traits described in this taxonomy provide a useful starting point for conceptualizing pathological personality development in childhood and adolescence. The traits described in this taxonomy may well need to be supplemented by additional individual differences such as identity, attachment style, and defense mechanisms (Cohen and Crawford 2005), but this taxonomy already includes many personality patterns that are likely to be related to the emergence of pathological personality functioning.

Stability of Personality Across the Life Span A recent meta-analysis by Roberts and DelVecchio (2000) provided a comprehensive empirical evaluation of the extent to which temperament and personality are stable across the life course. This meta-analysis included studies from birth through old age that reported correlations between dispositional measures taken at least 1 year apart (the average time between assessments was 6.8 years). The following estimated population cross-time correlations for dispositional measures were obtained: 0–2.9 years = .35; 3.0–5.9 years = .52; 6.0–11.9 years = .45; 12.0– 17.9 years =. 47; 18.0–21.9 years = .51; 22–29 years = .57; 30–39 years = .62; 40– 49 years = .59; 50–59 years = .75; and 60–73 years = .72. These results document several patterns that are relevant for the development of personality disorders. Individual differences show more modest continuity during infancy and toddlerhood and then show a rather large increase in stability during the preschool years; personality remains moderately stable throughout childhood and adolescence. There is a gradual linear increase in stability from late adolescence through the adult years, with strong stability achieved by around age 50 years. Thus, there is already considerable personality stability by age 3 years but also considerable change occurring well into the adult years. It is not clear whether these conclusions apply to personality pathology as well, but some recent work suggests similar moderate stability of personality disorder symptoms in adolescence and early adulthood (Johnson et al. 2000).

Directions for Future Work on the Development of Personality Disorders All of the previously described research suggests that childhood personality functioning can and should be integrated into developmental research and applied work on personality disorders. In particular, much more needs to be learned about

Shiner: Commentary on Mervielde et al.

115

the ways that children’s early traits may interact with their experiences to shape pathological personality functioning. Children’s early traits may shape their later experiences in a number ways: by shaping children’s effective experience of the environment, by evoking certain patterns of responses from adults and other children, and by leading children to select particular environmental experiences (Caspi and Shiner 2006; Shiner and Caspi 2003). Youths’ personalities also predict their later mastery of and struggles with important developmental tasks such as forming friendships and romantic relationships and achieving success in school and work (Caspi and Shiner 2006; Shiner and Caspi 2003). In turn, youths’ relative success or failure at such adaptive tasks is likely to shape their personality functioning (Shiner and Masten 2002). However, it is critical to emphasize that, as Thomas and Chess pointed out several decades ago, children’s temperaments may yield positive or pathological outcomes, depending on the “goodness of fit” between the children’s temperament and the environmental context (Thomas et al. 1963). Some recent work has begun to substantiate this claim; for example, early childhood social inhibition may develop into more pathological forms of shyness when parents are overprotective or derisive (this pattern and others are reviewed in Gallagher 2002). In short, a richer understanding of normal personality development has the potential to yield a more complete developmental perspective on personality disorders in both childhood and adulthood.

References Bouchard TJ: Genetic influence on human psychological traits: a survey. Current Directions in Psychological Science 13:148–151, 2004 Caspi A, Shiner RL: Personality development, in Handbook of Child Psychology, 6th Edition, Vol 3: Social, Emotional, and Personality Development. Edited by Damon W, Lerner R (Series) and Eisenberg N (Volume). New York, Wiley, 2006, pp 300–365 Caspi A, Roberts BW, Shiner RL: Personality development: stability and change. Annu Rev Psychol 56:453–484, 2005 Cohen P, Crawford T: Developmental issues: personality disorder in children and adolescents, in The American Psychiatric Publishing Textbook of Personality Disorders. Edited by Oldham JM, Skodol AE, Bender D. Washington, DC, American Psychiatric Publishing, 2005, pp 171–185 Emde RN, Hewitt J: Infancy to Early Childhood: Genetic and Environmental Influences on Developmental Change. New York, Oxford University Press, 2001 Gallagher KC: Does child temperament moderate the influence of parenting on adjustment? Dev Rev 22:623–643, 2002 Johnson JG, Cohen P, Kasen S, et al: Age-related change in personality disorder trait levels between early adolescence and adulthood: a community-based longitudinal investigation. Acta Psychiatr Scand 102:265–275, 2000 Muris P, Schmidt H, Merckelbach H, et al: The structure of negative emotions in adolescents. J Abnorm Child Psychol 29:331–337, 2001

116

Dimensional Models of Personality Disorders

Putnam SP, Ellis, LK, Rothbart MK: The structure of temperament from infancy through adolescence, in Advances in Research on Temperament. Edited by Eliasz A, Angleneitner A. Miami, FL, Pabst Science, 2001, pp 165–182 Roberts BW, DelVecchio WF: The rank-order consistency of personality traits from childhood to old age: a quantitative review of longitudinal studies. Psychol Bull 126:3–25, 2000 Rothbart MK, Bates JE: Temperament, in Handbook of Child Psychology, 5th Edition, Vol 3: Social, Emotional, and Personality Development. Edited by Damon W (Series) and Eisenberg N (Volume). New York, Wiley, 1998, pp 105–176 Shiner RL: How shall we speak of children’s personality traits in middle childhood? A preliminary taxonomy. Psychol Bull 124:308–322, 1998 Shiner RL, Caspi A: Personality differences in childhood and adolescence: measurement, development, and consequences. J Child Psychol Psychiatry 44:1–31, 2003 Shiner RL, Masten AS: Transactional links between personality and adaptation from childhood through adulthood. Journal of Research in Personality 36:580–588, 2002 Thomas A, Chess S, Birch H, et al: Behavioral Individuality in Early Childhood. New York, New York University Press, 1963

12 PERSONALITY DIMENSIONS ACROSS CULTURES Jüri Allik, Ph.D.

Personality Research Across Cultures Many popular psychological assessment instruments, originally developed in English, have been translated into numerous languages and are now commonly used throughout the world. Most of these translations were made with an explicit or at least tacit assumption that the core psychological constructs assessed by the measures substantively transcend human language and culture. The generalizability across languages and cultures, however, was in most cases presumed, not demonstrated. Therefore, it is not surprising that some researchers have expressed concern with this assumption and with practices guided by it (Shweder 1991). Skeptics have questioned, for instance, whether the uncritical extension of “Western” ways of thinking to the rest of the world should serve as standard procedure in psychological science. From a cultural constructionist point of view, all personality models are based on conceptions of personhood and standards of culturally

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 212–232, 2005). I thank Jeff McCrae and Dave Schmitt for discussions, collaboration, and sharing of their data. Preparation of this paper was supported by the Estonian Science Foundation and the Estonian Ministry of Science and Education.

117

118

Dimensional Models of Personality Disorders

appropriate behavior that have evolved, as a rule, in Anglo-American societies. It may be even more so concerning personality disorders which are, as some researchers believe, quintessential cultural products that owe their meaning and descriptive content to a distinctive cultural tradition (Fabrega 1994). It is also claimed that studies of mental health are culturally insensitive partly due to uncritical transfer of tools and concepts from one language and culture, typically English and United States, to other countries (Rogler 1999). In 1997, McCrae and Costa made a bold proposal about cross-cultural generalizability of the five-factor model of personality. Studying six translations of the NEO Personality Inventory—Revised (Costa and McCrae 1992) into German, Portuguese, Hebrew, Chinese, Korean, and Japanese they observed that all translations showed similar structures after varimax rotation of 5 factors and almost identical factor structure when the varimax solution was targeted towards the original American factor structure. For some reasons, people from different places around the world are inclined to think that individuals who, for example, talk much are at the same time optimistic and bubble with happiness, or that those who often get disgusted with other people also are inclined to feel inferior to others. Because the samples studied represented highly diverse cultures with languages from five distinct language families, McCrae and Costa (1997) proposed that personality trait structure is universal. The proposal that the personality structure is largely or even completely independent from the culture (McCrae and Costa 1996, 1999) was rather unorthodox and was expectedly met with a skepticism and denial (Allik and McCrae 2004a; Toomela 2003). It is also important to notice that although the replicability of the pattern of covariations across cultures turned out to be a rather robust phenomenon (cf. Rolland 2002), this fact does not necessarily imply that all and finest details are always replicated in every new culture.

Problems in Comparing Personality Trait Scores Across Cultures For psychologists seeking to investigate personality traits across cultures, one of the more inconvenient problems has centered on whether personality trait scales possess conceptual equivalence across cultures or not (van de Vijver and Leung 1997, 2000). Particularly troublesome has been establishing whether the mean scores across different cultures show metric or scalar equivalence (Byrne and Campbell 1999; Little 2000). That is, when comparing the mean scores of different cultures on a personality trait scale, any observed differences may be due not only to a real cultural disparity on some personality trait, but also to inappropriate translations, biased sampling, or the nonidentical response styles of people from different cultures (Bijnen and Poortinga 1988; Heine et al. 2002; van de Vijver and Leung 1997).

Allik: Personality Dimensions Across Cultures

119

Among the more common methods for establishing the cross-cultural comparability of personality trait measures is to first show that the trait scales contained in the measures are internally reliable across all targeted languages and cultures. A second frequently employed technique is to demonstrate a high degree of factorial structure invariance across different linguistic and cultural contexts (Barrett et al. 1998; Caprara et al. 2000; Rolland 2002). Metric equivalence can also be established through differential item analysis and bilingual administrations. Historically, if trait scales from a personality measure showed high internal reliability, invariant factor structure, and item equivalence across different languages and cultures, comparing the mean scores across cultures was often deemed a reasonable next step (Steel and Ones 2002; van de Vijver and Leung 2001). However, even with evidence of reliability, factor invariance, and item equivalence, problems can remain in how to metrically interpret mean-level differences in personality traits across cultures. Another way to increase confidence in the cross-cultural comparability of personality measures is to show that the mean levels of different assessment instruments intended to measure the same construct, or approximately the same construct, are highly correlated across multiple languages or cultures. For example, if two conceptually similar personality trait scales are used in a large number of different cultures, a positive association between the mean levels of those trait scales across the broad set of cultures would provide evidence that both measures are tapping the same underlying construct (Campbell and Fiske 1959).

Large-Scale Cross-Cultural Studies There have been many studies in which pairs of cultures have been compared on personality measures (cf. Katigbak et al. 1996, 2002), but there have been only a few in which a sufficiently broad sample of cultures was examined providing a statistically meaningful comparison. One of the first comprehensive personality trait measures to enjoy worldwide popularity and a fairly large number of translations into different languages was Eysenck’s Personality Questionnaire (EPQ; Eysenck and Eysenck 1975). In 1984, mean-level trait scores from 25 countries were made available (Barrett and Eysenck 1984). Ten years later, the number of countries in which three broad personality traits—Neuroticism, Extraversion, and Psychoticism—were measured by the EPQ was expanded to 37 (Lynn and Martin 1995). Soon it was found that the internal reliability and factorial structure of the EPQ across languages and cultures appeared to be replicable (Barrett et al. 1998). Although evidence of the cross-cultural generalizability of the EPQ seemed overwhelming, there was still some space for reservations. Some of these doubts were met by van Hemert and colleagues (2002), who critically reanalyzed available EPQ data both at the individual and the country

120

Dimensional Models of Personality Disorders

level. Many previous studies were dropped because of insufficient information (e.g., the number of women and men was not specified). The final set contained studies in 38 countries with a total of 68,374 respondents. Using multilevel factor analysis, they found that the EPQ has different factorial structure at both levels. In some countries (China, India, Japan, and Uganda) the pattern of correlation between EPQ scales differed from those found elsewhere. For the exploration of the meaning of the EPQ scales, country-level correlations were found with a variety of country characteristics, such as Hofstede’s measures of cultural differences and Diener’s of subjective well-being. A disappointing result of this analysis was that Lynn and Martin’s (1995) findings were not replicated, as no expected correlations were found, for example, between Extraversion and the death rate. The most comprehensive instrument thus far designed to measure the Big Five, or the Five-Factor Model (FFM), is the NEO Personality Inventory—Revised (NEO PI-R; Costa and McCrae 1992). Recently, the NEO PI-R was translated into many different languages and administered to samples from over two dozen countries. In 2001, NEO PI-R data from 26 countries or cultural regions became available for the research community (McCrae 2001) and the database was soon expanded by 10 additional cultures covering 5 major language families: IndoEuropean, Uralic, Altaic, Dravidian, and Sino-Tibetian (McCrae 2002). In every culture and language that has been studied, the trait scales of the NEO PI-R have displayed adequate levels of internal reliability, and the factorial structure of the NEO PI-R has been considered robust (McCrae 2001, 2002). Direct comparisons of the NEO PI-R with the EPQ have suggested that translations of both instruments provide reasonably comparable estimates of mean levels of extraversion and neuroticism across cultures. For example, the mean-level scores of neuroticism and extraversion as measured by the NEO-PI-R and the EPQ were significantly correlated across 18 nations (r=.80 and r= .51, respectively) (McCrae 2002). Thus, if a nation scored relatively high on the EPQ Neuroticism and Extraversion scales, it was likely to score high on the NEO PI-R Neuroticism and Extraversion scales as well. For this result, however, it was necessary to omit EPQ data from India as a suspected outlier. These empirical findings, although limited to 18 cultural regions, can be taken as supportive evidence that at least two of the Big Five dimensions, Neuroticism and Extraversion, can be comparably measured across human languages and cultures. Perhaps David Schmitt conducted the largest to date cross-cultural study of personality as a part of the International Sexuality Description Project, a collaborative effort of more than 100 social, behavioral, and biological scientists from 56 countries (Schmitt et al. 2003). All studied samples were administered the Big Five Inventory (BFI) of personality traits (Benet-Martínez and John 1998). The BFI, a short version of the Big Five family questionnaires, was translated from English into 29 languages and administered to 17,837 individuals from 56 nations. Results indicated that the five-dimensional structure of the BFI was highly robust across

Allik: Personality Dimensions Across Cultures

121

major regions of the world (D.P. Schmitt, J. Allik, R.R. McCrae, V. Benet-Martínez, L. Alcalay, L. Ault, et al., “The Geographic Distribution of Big Five Personality Traits: Patterns and Profiles of Human Self-Description Across 56 Nations,” submitted for publication). However, the correlation between BFI scales and their counterparts from EPQ and NEO PI-R was disappointingly low. Although the EPQ and the BFI do not conceptualize Neuroticism and Extraversion in a completely identical way, it is reasonable to expect at least modest convergent correlations. There were 26 countries for which the mean scores of Neuroticism and Extraversion were measured by both EPQ (Lynn and Martin 1995; van Hemert et al. 2002) and BFI. As expected, the Neuroticism scales of the two instruments were significantly correlated (r=.49, P=0.01) but the correlation between BFI Extraversion scale and its EPQ counterpart was disappointingly low and did not reach statistical significance (r= .18). The problems with convergent correlations were not only due to conceptual differences between measuring instruments. The convergent correlation was also absent when two conceptually similar Big Five instruments, BFI and NEO PI-R, were compared. In an overlapping set of 27 cultures, correlations between BFI and NEO PI-R scales were possible. Although the convergent correlations of BFI and NEO PI-R nation-level scores on Neuroticism (r=.45, P < 0.05), Extraversion (r=.44, P < 0.05), and Conscientiousness (r=.45, P<0.01) were significant, two remaining convergent correlations, Agreeableness (r=.22) and Openness (r=.27), failed to reach the level of statistical significance. There were greater problems with the discriminant correlations. For example, the BFI Openness scale was very strongly correlated with the NEO PI-R Extraversion scale (r=.73, P <0.001) rather than with its intended counterpart. Thus, two independent measures of the Big Five (the BFI and the NEO PI-R) demonstrated only partial cross-cultural agreement. Indeed, in some cases, parallel measures were rather consistent. According to both the BFI and NEO PI-R, for example, Japan’s level of Neuroticism was among the highest of all cultures, and according to the EPQ, Japan’s Neuroticism was the third highest (Lynn and Martin 1995). In most other cases, however, agreement between parallel personality measures across cultures was modest at best (Schmitt et al., submitted for publication).

GEOGRAPHICAL DISTRIBUTION OF PERSONALITY TRAITS How are personality traits distributed throughout the world? Until recently, it was impossible to give any informed answer to such questions. However, the first known attempt to examine a systematic pattern in the world-wide distribution of personality traits was already rather rewarding (Allik and McCrae 2004b). Although the translation quality of the NEO PI-R varied considerably, and some of the studied cultures were represented by very small (less than 100) and convenient (e.g., only college students) samples, the NEO PI-R data set provided strong and reliable evidence that the geographical distribution of the mean-level trait scores

122

Dimensional Models of Personality Disorders

produced meaningful patterns. Not only were the mean-level personality trait scores predictably related to other culture-level indicators, such as Hofstede’s dimensions of culture (Hofstede and McCrae 2004; McCrae 2002), but their distribution in geographic space seemed to have regular, systematic patterns. Neighboring countries tended, as a rule, to have similar personality means, and regions separated geographically or historically had less similar means on personality trait scales (Allik and McCrae 2004b). The results of multidimensional scaling demonstrate a clear contrast of European and American cultures with Asian and African cultures: the former were higher in Extraversion and Openness to Experience and lower in Agreeableness. A second dimension reflected differences in psychological adjustment, with the largest correlations with Neuroticism and Hofstede’s Uncertainty Avoidance. This pattern indicates that the distribution of self-reported personality traits is organized geographically.

Personality Dimensions Across Cultures: Problems and Challenges THE CULTURE-LEVEL FACTOR STRUCTURE When the number of studied cultures becomes large enough, it is possible to aggregate data within one culture and to treat aggregates as single cases for further cross-cultural analysis. Hofstede called this ecological factor analysis and used it to identify dimensions of culture (Hofstede 1983). Initially, it was thought that the close resemblance of the culture-level factor structure to the individual-level factor structure is surprising and tells us something meaningful about generalizability across cultures (Allik and McCrae 2002; McCrae 2002). In reality, however, the resemblance of these two different-level factor structures is quite expectable due to purely statistical reasons. For example, when in a large data set all subjects are reassigned randomly to arbitrary groups (“cultures”) it is very likely that such groupings retain the individual-level factor structure. This result becomes less mysterious if we think that when two variables covary, groups that happen for any reason to be high on one will tend also to be high on the other. When group-level data are analyzed, these two variables will still covary. Thus, the replication of the individual-level factor structure on the ecological level of analysis, where each culture is represented as a single case is not surprising, but rather all deviations from the exact replication. These are deviations from the individual-level factor structure that are informative, suggesting that grouping—in this case, belonging to different cultural groups—adds something beyond inter-individual variation. This addition, of course, can be due to measurement error (e.g., different sampling procedures in different cultures) or due to meaningful cultural differences. Culture-level factor analyses of personality traits have revealed only modest deviations from the

Allik: Personality Dimensions Across Cultures

123

individual-level factor structure (McCrae 2001, 2002). This seems to indicate that if the culture-level aggregation adds anything that goes beyond interindividual variation then it is relatively modest in its size.

COMPREHENSIVENESS OF THE BIG FIVE It has been argued that the administration of an instrument like the NEO PI-R in non-Western and especially illiterate societies will not necessarily lead to the similar factor structure. The main reason is that non-Western cultures may lack the notion of personality as an individual who is clearly distinctive from others. Indeed, Luria’s expeditions to the mountain regions of Uzbekistan and Kirghizia in 1931–1932 showed that inhabitants of villages who had no formal schooling had difficulty or were incapable of analyzing their subjective qualities and personality (Luria 1976). They frequently found it much easier to characterize other people than themselves and instead of describing their own personality they referred to the group to which they belong, or they described external circumstances rather than their personality traits or character (p. 144 ff.). However, illiteracy is not in itself an obstacle for an appropriate personality description, because questionnaires can be administered orally to the participants (e.g., Lima 2002). Also, inability to describe one’s own personality does not imply that personality is absent or that external observers who have mastered the analysis of personality with sufficient sophistication cannot provide consistent descriptions of their targets. It is also questioned whether the Big Five is sufficient or not to describe personality traits in all cultures. Cultures may differ in traitedness, and some traits typical to Western cultures are not appropriate, for example, to Eastern ones (cf. Church 2000). This proposal is a cultural-level equivalent of Allport’s idiographic hypothesis (Allport 1937): not all traits are equally applicable to all cultures. For instance, it was found among Chinese students that the Interpersonal Relatedness factor, which was absent in the Big Five, could not be consistently explained by a combination of factors (Cheung et al. 2001). However, this supposedly indigenous Interpersonal Relatedness factor was replicated in European–American samples (Cheung et al. 2003; Lin and Church 2004), indicating that these dimensions are not unique to Chinese populations. Contrary to previous interpretations of the Interpersonal Relatedness dimension in terms of interdependent self-construals, the dimension was only modestly correlated with relational and collective aspects of self, two aspects of interdependent self-construals (Lin and Church 2004).

STEREOTYPES AND EXPERT OPINIONS Can one interpret mean personality trait scores as an indicator of common personality features, and can one compare cultures on these traits? The aggregate personality scores vary systematically across cultures and demonstrate interpretable cor-

124

Dimensional Models of Personality Disorders

relations with other culture-level indicators. This seems to suggest that aggregate personality scores make sense. However, these scores did not match the intuitive assessments of both laypersons and a panel of expert cross-cultural judges: Japan, for example, showed a low score for Conscientiousness, despite the widespread perception that the Japanese are an industrious people. Perhaps only a lack of sufficient ethnographic knowledge prevents us from being too surprised that the most purposeful and strong-willed individuals, according to their own self-reports at least, are Telugu and Marathi Indians (McCrae 2002). In order to collect expert opinions, McCrae (2001) asked eight prominent cross-cultural psychologists to identify the personality factor that had been used to rank 26 cultures based on their mean NEO PI-R scores. He asked, for example, which personality factor is lowest among Hong Kong Chinese and South Koreans but highest among Norwegians and Americans. Rather surprisingly, these experts all considered this a difficult task and were unable to identify factors at a better-than-chance level. The most focused and systematic attempt to compare national stereotypes by means of self-reported personality traits was undertaken by Church and Katigbak (2002). In this study, 43 judges who had lived in both the Philippines and the United States for a considerable time rated whether Filipinos or Americans would tend to show a particular trait more. As they wrote, “The backgrounds of our bicultural judges would seem to be fairly ideal for their task. Thus, if their judgments of average cultural differences are not valid, it is not clear whose judgments would be” (Church and Katigbak 2002, p. 149). These bicultural judges were in high agreement with each another but were not consistent with the mean NEO PI-R profiles (Church and Katigbak 2002). These negative findings suggest that mean trait levels and national stereotypes, even held by experts, are not necessarily compatible. This is, of course, puzzling although there are several obvious reasons for the discrepancy. One attractive interpretation, proposed previously, is to discount beliefs about national character as a form of mythology. National stereotypes may be historical accidents, or self-serving attributions, or totems that serve the function of cementing group identity, rather than veridical accounts of aggregate personality traits (Allik and McCrae 2002). This interpretation runs against numerous facts where national stereotypes are well motivated and may reflect actual differences between groups (Peabody 1985). Another possibility is that except in extreme cases, it is in principle difficult if not impossible to reach correct judgments of mean personality profiles. The actual differences are rather small for most traits, and it may be impossible for individuals to conceive them accurately. Studies show that cross-cultural differences in the mean values are rather small compared to inter-individual differences within each culture (McCrae 2001, 2002). The distance between standings on the basic personality dimensions of any two individuals randomly selected from the same culture is almost certainly larger than the difference between any two arbitrarily chosen on the same personality dimension.

Allik: Personality Dimensions Across Cultures

125

SUBSTANCE VERSUS STYLE A regularity seems to emerge from recent large-scale cross-cultural studies of personality. The strongest correlation with culture level indicators such as gross national product (GNP) is observed not with the mean scores of personality traits but with the scales supposedly measuring social desirability and different indicators of response bias. For example, van Hemert and colleagues (2002) found that the most striking finding was a substantial negative correlation between the EPQ Lie scale and indicators of economic prosperity: people from more rich and wealthy countries are less prone to conformity, and their answers depended less on the approval of other people. Thus, it is likely that socially desirable responding is not a style of responding that jeopardizes cross-cultural equivalence of other personality dimension; rather, social desirability may be a relatively stable personality characteristic that is influenced by cultural and socioeconomic conditions (Poortinga et al. 2002). In order to evaluate quality and consistency of NEO PI-R data collected from 50 cultures, McCrae and Terracciano developed an index based on six indicators (e.g., number of response protocols with more than 40 missing responses, acquiescence bias, whether the native or the second language was used) (McCrae et al. 2005). When countries were listed in descending order on the data quality index, the entries at the top of the list were from affluent, mostly Western nations, whereas those at the bottom were from underdeveloped nations. It seems that in less-affluent countries, careless or acquiescent responding and failure to understand the nuances of language are more frequent than in better-educated and richer countries. Indeed, the rank-order correlations between the data quality index and the Cronbach alphas for the Big Five dimensions were in the range of .63 to .81. Perhaps the most informative research in this regard is a recent study in which the Rosenberg Self-Esteem Scale (RSES) was translated into 28 languages and administered to 16,998 participants across 53 nations (Schmitt and Allik 2005). It was shown that with few exceptions (primarily due to a single item), the factor structure of the RSES was invariant across nations. As expected, the RSES scores correlated significantly negatively with Neuroticism and positively with Extraversion, within nearly all nations, providing some support for the cross-cultural equivalence of global self-esteem. All nations scored, on average, above the theoretical midpoint of the RSES, indicating that generally positive self-evaluation may be culturally universal. Although psychometric studies have generally supported the unidimensionality of the RSES (Carmines and Zeller 1979; O’Brien 1985; Marsh 1996), there is still a pervasive tendency to respond to negatively worded items slightly differently than to positively worded items (Benson and Hocevar 1985; Marsh 1986). If people from certain types of cultures respond differently to the phrasing of negative items, this would indicate that direct cross-cultural comparisons on the RSES are confounded by a negative-item bias. Although positively and negatively worded items of the RSES were strongly correlated on both intracultural and intercultural levels

126

Dimensional Models of Personality Disorders

and were similarly related to external personality variables, the difference between aggregates of positive and negative items was clearly smaller in more developed nations. Although national self-esteem levels had neither relation to the Human Development Index (HDI) nor to its three principal components—life expectancy at birth, adult literacy rate, and standard of living measured by GDP per capita, the difference between sums of positively and negatively worded items was strongly correlated with HDI (r=–.53, P<0.001) and several other national-level indicators. Thus, in developed countries, where people have better education and live longer, a message formulated in negative wording has a slightly different meaning compared with the same message formulated in positive wording. The results of this study suggest that comparing the raw scores of the RSES across cultures has limited value unless the inherent bias related to the different functioning of positively and negatively worded items has been taken into account (Schmitt and Allik 2005). Together with other findings, this result seems to suggest that it is not always the mean scores of personality traits per se that have a strong and systematic relation to the national-level indicators. In many cases, these are the “secondary” or higher-order properties of the personality questionnaires that have pervasive associations with how much people in their respective countries are educated, how long they are expected to live, and to what extent their economic life is secured. For example, McCrae (2001, 2002) observed that the standard deviations around the national mean scores are clearly ordered: the variation between individual mean scores is higher in economically advanced and lower in stagnated countries. Interestingly, the magnitude of gender differences follows a similar pattern, with the largest gender differences in European countries and the smallest in Asian (Costa et al. 2001). Analogously, measures of social desirability (van Hemert et al. 2002), acquiescence bias (Smith 2004), and negative-item bias (Schmitt and Allik 2005) were strongly and systematically related to national indicators of socioeconomic development, often even more powerfully than the mean levels of personality traits themselves. This fact can be interpreted as a nuisance of intercultural comparisons that impedes comparability of cultures with respect to personality traits (Poortinga et al. 2002). According to another interpretation, however, responding in a socially favorable manner and other response biases have substantive cultural meaning and cannot be eliminated after correction for response style (Smith 2004). Although cultures may differ in response style or self-presentation strategies, it is also possible that the observed regularities represent real differences in personality. It is possible, for example, that a readiness to agree with expressed opinions and a cautiousness toward disparaging self-descriptions are, like sociability and conservativism, some aspects of the basic personality traits. For currently poorly understood reasons, these traits are more accurately “leaking out” through what is known as the acquiescence or the negative-item bias than through the mean levels themselves. Unfortunately, our knowledge in this area is seriously limited by a lack of relevant facts, and only future studies can build a stronger basis for conclusions.

Allik: Personality Dimensions Across Cultures

127

Conclusion The data reviewed above seem to point in the direction of the psychic unity of humankind, at least with regard to underlying personality dispositions. Although cultural differences appear to be enormous, anthropologists are arguing that there is a common basic culture of all humanity in all places and in all periods: “I would argue that in a real sense there is only one culture—the culture of humankind— and that societal differences with respect to cultural items are small” (D’Andrade 2001, p. 253). It is likely that cultural unity is based, at least partly, on the psychic unity of all people. Perhaps due to this unity, the quality of translation of personality questionnaires, conditions of their administration, and sampling that seldom represented a given culture as a whole had little effect on the results: the dimensional structure of personality—five relatively independent groups of covarying traits—can be easily recovered from data that are not necessarily optimal for it, not only in Western cultures but also in non-Western cultures, including Zimbabwe, Philippines, Malaysia, and China (McCrae 2002). Today there is much more justification than there was 10 years ago for agreeing with the hypothesis that the pattern of covariation among personality traits is universal and extends across languages and cultures (McCrae and Costa 1997). This also means that the FFM, which is the best summary of their pattern of covariation, is generalizable to psychiatric samples as well (Widiger and Costa 2002; Widiger and Trull 1992). Unfortunately, the number of cross-cultural comparisons of psychiatric samples is still very small. The first results, however, are promising and demonstrate that the FFM retains its main properties also in non-Western psychiatric samples, such as populations in China (Yang et al. 1999). In contrast to the generalizability of the FFM across languages and culture, the comparison of the mean trait scores is much more problematic. The first wave of the large-scale, cross-cultural projects revealed several paradoxes. Although the mean trait scores for different cultures were geographically organized, the emergent pattern was not so easily interpretable. It contradicted not only the layperson’s intuition but also the knowledge of experts (Church and Katigbak 2002). Take, for example, an extravert who, according to the NEO PI-R manual, is sociable, assertive, talkative, and likes large groups and gatherings (Costa and McCrae 1992). Which nation fits best with this psychological portrait? It is not very likely that candidates for the most extraverted nations include any of the Scandinavian countries; according to widespread stereotypes, Scandinavians are typically quiet and reserved. In fact, according to self-reports, the most extraverted individuals live in Denmark, Norway, and Sweden (McCrae 2002). Thus, stereotypes do not agree with rankings of the mean scores on personality traits. Another surprising result that emerged from the first wave of studies is a relatively low agreement between parallel studies and instruments. The convergent correlations were disappointingly low, even between two instruments belonging to

128

Dimensional Models of Personality Disorders

the same family of personality measures. Particularly problematic were, however, discriminant correlations: it was not uncommon for one scale from one instrument to have the strongest correlation not with its counterpart from another questionnaire but rather with some other scale. Even the pattern of convergentdiscriminant correlations between two of the Big Five questionnaires, NEO PI-R and BFI, was rather confused (Schmitt et al., submitted for publication). One possible explanation for this poor agreement is that the mean-level differences in personality traits across cultures are very small in magnitude. For example, the NEO PI-R factor means of 36 cultures had standard deviations that were about one-third the magnitude of individual differences within cultures (McCrae 2002). Approximately the same ratio of between- to within-culture variation was obtained among 56 cultures where the BFI was administered (Schmitt et al., submitted for publication). This means that even if the personality measures were perfectly accurate, knowing a person’s nationality or cultural background would tell us rather little about his or her personality (Allik and McCrae 2002). The magnitude of the cross-cultural differences in the trait mean scores is smaller or comparable to the overall measurement accuracy, which can be compromised by many factors, including biased sampling and small differences in translation, to say nothing about random measurement errors. Consequently, it is not very likely that the same, “true” ordering of cultures on any of the Big Five personality traits would be exactly preserved in two parallel studies. If these differences are small, it is also improbable that people could accurately notice these tiny differences between different groups and aggregate them into stereotypes that reflect true standings on personality dimensions. Perhaps it may be possible in a few extreme cases, when two compared cultures are extraordinarily polarized—such as, for example, North Americans and Japanese on the scale of self-esteem. But in all other cases, the differences in the mean levels of personality traits may be too small to be consistently reproduced on two parallel occasions. Nevertheless, some more integral (e.g., the similarity between personality profiles) or subtle (e.g., the disparity between positively and negatively worded items) features might reveal systematic relationships with relevant socioeconomic and geographic variables. Fortunately, this may be good news for cross-cultural studies in general and for cross-cultural studies of personality disorders in particular. The prospect that for a proper psychological assessment, in both normal and psychiatric samples, it would be necessary to develop culture- or even subculture-specific norms looks frightening, not only because of the enormous amount of work required to develop these norms but also because of the increasing fragmentation and particularization of psychological measurement. A relatively modest size of cross-cultural differences in mean values may imply that a reasonable scalar equivalence can be achieved and that all individuals, irrespective of their language and culture, can be represented in a common metric.

Allik: Personality Dimensions Across Cultures

129

References Allik J, McCrae RR: A Five-Factor Theory perspective, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic/ Plenum, 2002, pp 303–322 Allik J, McCrae RR: Escapable conclusions: Toomela 2003 and the universality of trait structure. J Pers Soc Psychol 87:261–265, 2004a Allik J, McCrae RR: Towards a geography of personality traits: patterns of profiles across 36 cultures. J Cross Cult Psychol 35:13–28, 2004b Allport GW: Personality: A Psychological Interpretation. New York, Holt, Rinehart Winston, 1937 Barrett P, Eysenck SBG: The assessment of personality factors across 25 countries. Personality and Individual Differences 5:615–632, 1984 Barrett PT, Petrides KV, Eysenck SBG, et al: The Eysenck Personality Questionnaire: an examination of the factorial similarity of P, E, N, and L across 34 countries. Personality and Individual Differences 25:805–819, 1998 Benet-Martínez V, John OP: Los cinco grandes across cultures and ethnic groups: multitrait multimethod analysis of the Big Five in Spanish and English. J Pers Soc Psychol 75:729–750, 1998 Benson J, Hocevar D: The impact of item phrasing on the validity of attitude scales for elementary school children. J Edu Meas 22:231–240, 1985 Bijnen EJ, Poortinga YH: The questionable value of cross-cultural comparisons with the Eysenck Personality Questionnaire. J Cross Cult Psychol 19:193–202, 1988 Byrne BM, Campbell TL: Cross-cultural comparisons and the presumption of equivalent measurement and theoretical structure: a look beneath the surface. J Cross Cult Psychol 30:555–574, 1999 Campbell DT, Fiske DW: Convergent and discriminant validation by the multitrait-multimethod matrix. Psychol Bull 56:81–105, 1959 Caprara GV, Barbaranelli C, Bermudez J, et al: Multivariate methods for the comparison of factor structures in cross-cultural research: an illustration with the Big Five Questionnaire. J Cross Cult Psychol 31:437–464, 2000 Carmines EG, Zeller RA: Reliability and Validity Assessment. Beverly Hills, CA, Sage, 1979 Cheung FM, Leung K, Zhang JX, et al: Indigenous Chinese personality constructs—Is the five-factor model complete? J Cross Cult Psychol 32:407–433, 2001 Cheung FM, Cheung SF, Leung K, et al: The English version of the Chinese Personality Assessment Inventory. J Cross Cult Psychol 34:433–452, 2003 Church AT: Culture and personality: toward an integrated cultural trait psychology. J Pers 68:651–703, 2000 Church AT, Katigbak MS: The Five-Factor Model in the Philippines: investigating trait structure and levels across cultures, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic/Plenum, 2002, pp 129–154

130

Dimensional Models of Personality Disorders

Costa PT Jr, McCrae RR: Revised NEO Personality Inventory NEO PI-R and NEO FiveFactor Inventory NEO-FFI Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992 Costa PT Jr, Terracciano A, McCrae RR: Gender differences in personality traits across cultures: robust and surprising findings. J Pers Soc Psychol 81:322–331, 2001 D’Andrade R: A cognitivist’s view of the units debate in cultural anthropology. Cross-Cult Res 35:242–257, 2001 Eysenck HJ, Eysenck SBG: Manual of the Eysenck Personality Questionnaire. London, Hodder & Stoughton, 1975 Fabrega H: Personality disorders as medical entities—a cultural interpretation. J Personal Disord 8:149–167, 1994 Heine SJ, Lehman DR, Peng K, et al: What’s wrong with cross-cultural comparisons of subjective Likert scales: the reference-group problem. J Pers Soc Psychol 82:903–918, 2002 Hofstede G: Dimensions of national cultures in fifty countries and three regions, in Expiscation in Cross-Cultural Psychology. Edited by Deregowski J, Dziurawiec S, Annis RC. Lisse, Netherlands, Swets & Zeitlinger, 1983, pp 335–355 Hofstede G, McCrae RR: Personality and culture revisited: linking traits and dimensions of culture. Cross Cult Res 38:52–88, 2004 Katigbak MS, Church AT, Akamine TX: Cross-cultural generalizability of personality dimensions: relating indigenous and imported dimensions in two cultures. J Pers Soc Psychol 701:99–114, 1996 Katigbak MS, Church AT, Guanzon-Lapena MA, et al: Are indigenous personality dimensions culture specific? Philippine inventories and the five-factor model. J Pers Soc Psychol 821:89–101, 2002 Lima MP: Personality and culture: the Portuguese case, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic/ Plenum, 2002, pp 249–260 Lin EJL, Church AT: Are indigenous Chinese personality dimensions culture-specific? An investigation of the Chinese Personality Assessment Inventory in Chinese American and European American samples. J Cross Cult Psychol 35:586–605, 2004 Little TD: On the comparability of constructs in cross-cultural research: a critique of Cheung and Rensvold. J Cross Cult Psychol 31:213–219, 2000 Luria AR: Cognitive Development: Its Cultural and Social Foundations. Cambridge, MA, Harvard University Press, 1976 Lynn R, Martin T: National differences for thirty-seven nations in extraversion, neuroticism, psychoticism and economic, demographic and other correlates. Personality and Individual Differences 19:403–406, 1995 Marsh HW: Positive and negative global self-esteem: a substantively meaningful distinction or artifactors? J Pers Soc Psychol 70:810–819, 1996 Marsh HW: Negative item bias in ratings scales for preadolescent children: a cognitivedevelopmental phenomenon. Dev Psychol 22:37–49, 1986 McCrae RR: Trait psychology and culture: exploring intercultural comparisons. J Pers 69:819–846, 2001

Allik: Personality Dimensions Across Cultures

131

McCrae RR: NEO-PI-R data from 36 cultures, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic/Plenum, 2002, pp 105–125 McCrae RR, Costa PT Jr: Towards a new generation of personality theories: theoretical context for the Five-Factor Theory, in The Five-Factor Model of Personality: Theoretical perspective. Edited by Wiggins JS. New York, Guilford, 1996, pp 51–87 McCrae RR, Costa PT Jr: Personality trait structure as a human universal. Am Psychol 52:509–516, 1997 McCrae RR, Costa PT Jr: A Five-Factor theory of personality, in Handbook of Personality. Theory and Research, 2nd Edition. Edited by Pervin LA, John OP. New York, Guilford, 1999, pp 139–153 McCrae RR, Terracciano A, Khoury B, et al: Universal features of personality traits from the observer’s perspective: data from 50 cultures. J Pers Soc Psychol 88:547–561, 2005 O’Brien EJ: Global self-esteem scales: unidimensional or multidimensional? Psychol Rep 57:383–389, 1985 Peabody D: National Characteristics. Cambridge, MA, Cambridge University Press, 1985 Poortinga YH, van de Vijver FJR, van Hemert DA: Cross-cultural equivalence of the Big Five, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic/Plenum, 2002, pp 281–302 Rogler LH: Methodological sources of cultural insensitivity in mental health research. Am Psychol 54:424–433, 1999 Rolland J-P: The cross-cultural generalizability of the Five-Factor Model of personality, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic/Plenum, 2002, pp 7–28 Schmitt DP, Allik J: Simultaneous administration of the Rosenberg Self-Esteem Scale in 53 nations: exploring the universal and culture-specific features of global self-esteem. J Pers Soc Psychol 89:623–642, 2005 Schmitt DP, Alcalay L, Allik J, et al: Universal sex differences in the desire for sexual variety: tests from 52 nations, 6 continents, and 13 islands. J Pers Soc Psychol 85:85–104, 2003 Shweder RA: Thinking Through Cultures: Expeditions in Cultural Psychology. Cambridge, MA, Harvard University Press, 1991 Smith PB: Acquiescent response bias as an aspect of cultural communication style. J Cross Cult Psychol 35:50–61, 2004 Steel P, Ones DS: Personality and happiness: a national-level analysis. J Pers Soc Psychol 83:767–781, 2002 Toomela A: Relationships between personality structure, structure of word meaning, and cognitive ability: a study of cultural mechanisms of personality. J Pers Soc Psychol 85:723–735, 2003 van de Vijver FJR, Leung K: Methods and data analysis of comparative research, in Handbook of Cross-Cultural Psychology, Vol 1: Theory and Method. Edited by Berry JW, Poortinga YH, Pandey J. Boston, MA, Allyn & Bacon, 1997, pp 257–300 van de Vijver FJR, Leung K: Methodological issues in psychological research on culture. J Cross Cult Psychol 31:33–51, 2000

132

Dimensional Models of Personality Disorders

van de Vijver FJR, Leung K: Personality in cultural context: methodological issues. J Pers 69:1007–1031, 2001 van Hemert DA, van de Vijver FJR, Poortinga YH, et al: Structural and functional equivalence of the Eysenck Personality Questionnaire within and between countries. Personality and Individual Differences 33:1229–1249, 2002 Widiger TA, Costa PT Jr: Five factor model personality disorder research, in Personality Disorders and the Five Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 59–87 Widiger TA, Trull TJ: Personality and psychopathology: an application of the five-factor model. J Pers 60:363–393, 1992 Yang J, McCrae RR, Costa PT Jr, et al: Cross-cultural personality assessment in psychiatric populations: the NEO-PI-R in the People’s Republic of China. Psychol Assess 11:359–368, 1999

13 COMMENTARY ON ALLIK The Lexical Approach to the Study of Personality Structure Michael Ashton, Ph.D.

The Lexical Approach to Personality Structure To find a cross-culturally generalizable structural model for personality traits, we need to analyze variable sets that have two important properties. First, the variables should be indigenous to the culture in which they are to be analyzed; that is, we should study personality traits that are recognized within the culture in question. Second, the variables should provide a broad and even representation of the full domain of personality variation, rather than a special emphasis on traits that strongly define the factors of any hypothesized model of personality structure. The method by which personality researchers have obtained variable sets that meet the above requirements—that is, of being indigenous to each culture and of being neutral with regard to proposed structural models—has been to use the lexical approach to personality structure (e.g., Goldberg 1981; Saucier and Goldberg 1996). In this strategy, researchers use the personality lexicon of a given language to derive a large set of personality variables. Specifically, they begin by examining

This chapter is an abbreviated version of a paper first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 303–308, 2005).

133

134

Dimensional Models of Personality Disorders

the dictionary of a given language to generate a complete list of the personalitydescriptive words (generally adjectives) familiar to speakers of that language. Next, they obtain self- or peer ratings on those personality adjectives from a large sample of respondents who are fluent in that language. The researchers then factor-analyze these ratings and inspect the adjective content of the obtained factors in order to identify the personality dimensions that are of greatest subjective importance within that culture. Finally, by comparing the sets of factors obtained in lexical studies conducted independently in various languages, the researchers may identify personality dimensions that are cross-culturally generalizable.

The Cross-Culturally Replicated Structure of Personality From Lexical Investigations During the past two decades, standard lexical studies of personality structure have been conducted in several different languages, mostly (but not entirely) of the Indo-European family. A recent published review of these investigations (Ashton et al. 2004b), as performed in seven languages—Dutch, French, German, Hungarian, Italian, Korean, and Polish—showed that the factor structures obtained in these studies were substantially similar across languages. In addition, more recent investigations in the English language (Ashton et al. 2004a, using 1970s archival data whose variable sets were too large to have been factor analyzed at the time) have also produced a similar structure. The factor solution that has replicated across these languages is in many ways quite similar to that specified by the Five-Factor Model (FFM), which itself originated in early, small-scale lexical studies in the English language (McCrae 1989). For example, these various languages all contain factors corresponding to Extraversion, Conscientiousness, and (to a somewhat lesser extent) Openness to Experience. However, there are also some important differences between the crossculturally replicated structure and the FFM. First, personality traits similar to the Angry Hostility facet of the FFM—that is, variables such as ill-temper and choler—have not loaded on a Neuroticism factor, but instead have loaded squarely at the low pole of an Agreeableness factor, alongside traits such as quarrelsomeness and stubbornness. Second, the Neuroticism factor, which, as noted above, is not defined by Angry Hostility, also includes some variables not directly represented in the FFM version of the factor. Specifically, this cross-cultural, lexical version of Neuroticism is defined not only by anxiety and vulnerability but also by sensitivity, sentimentality, dependence, and physical fearfulness. Together, these traits might suggest the name Emotionality. Third, traits such as slyness (or deceit) and pretentiousness (or conceit) have defined the low pole of a factor of their own, separate from the Agreeableness fac-

Ashton: Commentary on Allik

135

tor defined by anger-related traits. The additional factor, whose high pole has been labeled Honesty–Humility, thus represents a sixth dimension of personality. To summarize, the cross-culturally replicated structure of personality traits—as derived from indigenous variables selected to be neutral with regard to proposed models—has many aspects in common with the Five-Factor Model. However, the differences suggested by the lexical investigations are important, for at least two reasons. First, although some aspects of Honesty–Humility are incorporated within the FFM, others are not. For example, (lack of ) greed or status seeking is not directly represented, nor is (lack of ) exploitation of others. These important aspects of personality will thus be only partially captured by the FFM, even though they are core elements of the cross-language structure. Second, the reduction of three lexical factors (the cross-cultural variants of Emotionality, Agreeableness, and Honesty-Humility) into two dimensions (FFM Neuroticism and Agreeableness) will diminish our ability to discriminate among certain aspects of personality pathology. Consider three personality-relevant disorders that, although not listed on Axis II, are included within the DSM system as “disorders diagnosed in childhood.” In terms of defining content, separation anxiety disorder corresponds to Emotionality, oppositional defiant disorder corresponds to low Agreeableness, and many features of conduct disorder (conning, theft, instrumental bullying) correspond to low Honesty–Humility. If we reduce Emotionality, Agreeableness, and Honesty–Humility to Neuroticism and Agreeableness, these near-isomorphisms will be lost. Despite the above cautions, the findings of lexical investigations of personality should be encouraging to researchers who wish to establish a cross-culturally generalizable structural model that could summarize normal and abnormal personality variation. The solutions obtained in lexical personality research are largely compatible with the frameworks that have been proposed as a means for integrating structures obtained in investigations of traits relevant to personality pathology. As recently outlined by Widiger and Simonsen (Chapter 1 in this volume), these various proposed structures suggest a common set of four factors that includes dimensions of Extraversion, Constraint, Antagonism, and Emotional Dysregulation. This fourdimensional space is nicely subsumed within the solution obtained across cultures in the lexical studies described above: Extraversion and Constraint correspond directly to two of the obtained lexical factors, and the plane formed by Antagonism and Emotional Dysregulation is captured by the three lexical dimensions of Agreeableness, Emotionality, and Honesty–Humility. Beyond providing this general support for the proposed model of normal and abnormal personality, the cross-culturally replicated results from lexical studies of personality structure can also help to resolve some issues regarding the placement of specific traits within that model. For example, as noted above, investigations based on indigenous personality variables of diverse languages suggest that angerrelated traits are associated more with Antagonism than with Emotional Dysregu-

136

Dimensional Models of Personality Disorders

lation, and that traits related to dependency are associated more with Emotional Dysregulation than with low Antagonism. In addition, these investigations also suggest an important distinction between those aspects of Antagonism that involve anger and oppositionality, on the one hand, and exploitation and entitlement, on the other hand. The incorporation of these findings from cross-cultural personality research should help to refine the proposed model of normal and abnormal personality still further, and thereby sharpen the effectiveness of that model in explaining personality pathology.

Summary The main points of this commentary can be summarized as follows: 1. The lexical approach to personality involves identification of personality traits indigenous to a given culture, by finding the familiar personality-descriptive adjectives of a given language. Factor analyses of ratings on these adjectives reveal the subjectively important personality dimensions of that language and ultimately allow identification of a set of personality dimensions that are replicated across cultures. 2. The cross-culturally replicated structures obtained from lexical investigations in several languages suggest strong similarities to the FFM. Three important differences are also observed: 1) anger-related traits define (low) Agreeableness, not Neuroticism; 2) Neuroticism (i.e., Emotionality) includes traits such as sensitivity, dependence, and physical fearfulness; and 3) traits such as slyness/deceit and pretentiousness/conceit define a sixth factor, (low) Honesty–Humility. 3. The lexically derived structural model of personality variation also supports the main features of the proposed set of higher-order domains that are common to several models of normal and abnormal personality. The findings from cross-cultural investigations of indigenous personality descriptors may be useful in deciding on the classification of some traits related to Antagonism and to Emotional Dysregulation, and also suggest a division (or subdivision) of the Antagonism domain.

References Ashton MC, Lee K, Goldberg LR: A hierarchical analysis of 1,710 English personalitydescriptive adjectives. J Pers Soc Psychol 97:707–721, 2004 Ashton MC, Lee K, Perugini M, et al: A six-factor structure of personality-descriptive adjectives: solutions from psycholexical studies in seven languages. J Pers Soc Psychol 86:356–366, 2004

Ashton: Commentary on Allik

137

Goldberg LR: Language and individual differences: the search for universals in personality lexicons, in Review of Personality and Social Psychology, Vol 2. Edited by Wheeler L. Beverly Hills, CA, Sage, 1981, pp 141–165 McCrae RR: Why I advocate the five-factor model: joint analyses of the NEO-PI with other instruments, in Personality Psychology: Recent Trends and Emerging Directions. Edited by Buss DM, Cantor N. New York, Springer-Verlag, 1989, pp 237–245 Saucier G, Goldberg LR: The language of personality: lexical perspectives on the five-factor model, in The Five-Factor Model of Personality: Theoretical Perspectives. Edited by Wiggins JS. New York, Guilford, 1996, pp 21–50

This page intentionally left blank

14 COMMENTARY ON ALLIK A Historical Perspective on Personality Disorder Juan J. López-Ibor, M.D.

A principal problem with personality disorders is the definition of their nature. From a pragmatic point of view, the question may be avoided in clinical settings. Discussing how diseases could be defined, von Weizsäcker (1987) reached the conclusion that a “patient is a person who goes to the physician.” From this perspective, the principal role of the doctor is not to look for the cutoff point between health and disease; the patient already has done that in seeking medical attention. Rather, the task of the physician is to reach a diagnosis in order to implement a treatment. In psychiatric practice, personality disorders may serve as a catch-all diagnosis for the use of those patients who do not fit in other diagnostic categories. Of course, outside of everyday clinical practice—for example, in forensic settings, or in the education system—a perspective that defines a disorder on the basis of an individual’s help-seeking behavior is highly unsatisfactory. In this brief commentary, I will review different perspectives which, historically, have been used to conceptualize what we today call personality disorders. One initial basis for conceptualizing personality disorders is statistical abnormality. The long-standing tradition of the “psychology of differences” has described temperaments and characters that differentiate individual. Some individuals are significantly more “different” than others, to the extent that they collide with accepted social norms. Cullen, in 1777, introduced the notion of neuroses, defined as preternatural reactions. Whereas this term encompassed nonorganic conditions, over the course of time some of these were dropped as organic expla139

140

Dimensional Models of Personality Disorders

nations were discovered (e.g., Parkinson’s disease). The contemporary connotation of “preternatural” would be statistically abnormal. Kurt Schneider (1971) utilized this approach when defining abnormal personalities as extreme forms of normality from a statistical point of view. For Schneider, not all abnormal personalities are morbid. Creative persons, geniuses, and others have traits that are not common, but they are not clinical cases. This criterion is a weak one, difficult to use outside the cultural settings where it is applied, and heavily value-loaded. A second approach is to consider personality disorders as disturbances of behavior. That is the concept of moral insanity of Prichard (1835). Initially, the word “moral” connoted what we would call today behavior or habits, such as in moral treatment. But the term very quickly received a negative value. For Benjamin Rush (1812), these patients showed “an innate moral depravation.” In the 19th century, the notion of degeneration incorporated a negative moral connotation which was not in the mind of Morel (1859) when he first used it. Degeneration, which is the opposite of evolution as used by Darwin, refers to a process by which more severe forms of mental disorder appear in successive generations. A third perspective views abnormal behavior—i.e., that which collides with the norms and habits of the society in which a person with a personality disorder lives—as maladaptive, and distinct from criminal behavior. Pinel (1801) described the case of a young, spoiled adolescent who, in a fit of rage, threw a maid into a pit. In the subsequent court case, Pinel made the plea that although the young man had no symptoms of a mental disorder, his behavior was so purposeless that he had to be considered insane. Pinel used two terms to describe this kind of insanity: manie sans delire and folie raisonante, which could be translated into modern English as “mental disease without symptoms of mental disease” and “sane insanity.” Cleckley (1941) reached the same conclusion: the behavior of psychopaths is so maladaptive that it can only be the manifestation of a hidden insanity. The psychopathic behavior is a “mask of insanity.” Other early nosologists considered the (statistically) abnormal personality to be the consequence of an underlying morbid process. Juan Huarte de San Juan (1575) described some morbid features of personality; however, in premodern times, the distinction between the melancholic temperament and diseases induced by the black bile, for example, was not clear. In any event, Schneider (1971) followed this approach when he observed that some abnormal personalities could be morbid, or psychopathic, when the abnormality led to suffering. With this interpretation, we have closed a circle, because it is the suffering which leads a person to the doctor. For Schneider, only mental diseases with an underlying brain disease, degeneration, trauma, or malformation were real diseases. The rest, the abnormal variations of being—namely neurotic and personality disorders—were not real diseases. In Schneider’s view, they were, so to speak, “metaphors” of diseases. For Kranz (1936), a pupil of Schneider, in forensic settings, psychopaths should be considered “as if ” they were real mental patients.

López-Ibor: Commentary on Allik

141

One weakness of Schneider’s approach is that in the case of sociopaths (dissocial personalities) who lack a personal sense of suffering, he considered the suffering induced. For Schneider (1928), a “psychopath is an abnormal personality which, due to his/her abnormality, suffers or induces suffering” (p. iii). If suffering is a criterion to define a disease, in this case it is the one who suffers from the induced suffering who should be considered as the patient, and not the agent. Today there is more scientific evidence available to achieve a more precise concept of personality disorder. Neuroscience, in a powerful way, offers a means of reaching beyond the dualistic approach of Schneider, and delves into the brain mechanisms involved in behavior, both normal and abnormal. A recent focus on the neurobiology of temperament has been very productive; in this context, studies now have shown that human behavior, in contrast to animal behavior, shows a greater repertoire and independence from basic mechanisms of temperament. In personality disorders, this repertoire is limited, sometimes reduced to a minimum set. This notion is present in the current definitions of personality disorders as they appear in ICD-10 (World Health Organization 1992) and DSM-IV (American Psychiatric Association 1994), and should be verified on the basis of ongoing neurobiological research. A few years ago, for example, our group showed that platelet monoamine oxidase (MAO) levels in three groups of individuals (controls, bullfighters, explosives experts) with different amounts of impulsivity varied along a spectrum, whereas in a group of pathological gamblers it was not only low, but the standard deviation was also very small. We interpreted these results as the expression of dimensional phenomena in normal behavior, with categorical differences with the morbid condition. One implication of this finding is that the reduced range of neurobiological response patterns limits the freedom of the individual with the condition by reducing choices available to him. López-Ibor Sr. (1951) suggested that in forensic settings, the psychiatric expert should first assess the level of freedom that the subject has toward him- or herself. Cloninger (1996) arrived at the same conclusion, as suggested in his proposed criteria for personality disorder, in which subjects with the condition(s) were rated low in the following domains: Cooperativeness, Self-Transcendence and Self-Directiveness and capacity to control, regulate, and adapt behavior.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Cleckley H: The Mask of Sanity. London, Henry Kimpton, 1941 Cloninger CR: A psychobiological model of temperament and character: fundamental findings for use in clinical practice, in New Research in Psychiatry. Edited by Häfner H, Wolpert EM. Göttingen, Germany, Hogrefe & Huber, 1996, pp 95–106

142

Dimensional Models of Personality Disorders

Cullen W: First Lines of the Practice of Physic. London, Henry Kimpton, 1777 Huarte de San Juan J: Examen de Ingenios Para las Ciencias. Madrid, Spain, Esteban de Torre, 1977 (originally published 1575) Kranz H: Lives of Criminal Twins. Berlin, Germany, Springer, 1936 López Ibor JJ Sr: La responsabilidad penal del enfermo mental (Discurso para la recepción pública de Académico Electo y contestación de. P. Laín Entralgo). Instituto de España. Real Academia Nacional de Medicina, Madrid, Spain, Cosano, 1951 Morel B: Traité des dégéneréscences physiques, intellectuelles et morales de l’espèce humaine. Paris, France, Bailliere, 1859 Pinel PH: Traité médico-philosophique sur l’alienation mentale. Paris, France, Richard, Caille et Ravier, 1801 Prichard JC: A Treatise on Insanity and Other Disorders Affecting the Mind. London, Sherwood, Gilbert & Piper, 1835 Rush B: Medical Inquiries and Observations Upon the Diseases of the Mind. Philadelphia, PA, Kimber & Richardson, 1812 Schneider K: Die psychopathischen Personlichkeiten. Vienna, Austria, Franz Deuticke, 1923 Schneider K: Klinische Psychopathologie (Neunte, unveränderte Auflage). Stuttgart, Germany, Georg Thieme Verlag, 1971 von Weizsäcker V: Gesammelte Schriften in zehn Bänden. Herausgegeben von Peter Achilles, Dieter Janz, Martin Schrenk und Carl Friedrich von Weizsäcker. Suhrkamp Verlag Frankfurt am Main Band 6 Stücke einer medizinischen Anthropologie. Bearbeitet von Peter Achilles unter Mitwirkung von Dieter Janz, Mechthilde Kütemeyer, Wilhelm Rimpau, Walter Schindler & Martin Schrenk, 1987 World Health Organization: The ICD-10 International Classification of Mental and Behavioral Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland, World Health Organization, 1992

15 COMMENTARY ON ALLIK Cross-Cultural Diagnosis of Personality Disorders Yueqin Huang, M.D., M.P.H., Ph.D. Siu Wa Tang, M.B., Ph.D., F.R.C.P.(C.)

Studies on Personality Disorders in China Psychiatry, as a medical specialty based on Western methodology, began in China around the beginning of the twentieth century. The topic of personality disorders began to attract some attention in the late seventies, coinciding with the opening of China to the outside world, and has gained momentum in recent years. There have been two large-scale epidemiological surveys in Chinese in recent years (Chen et al. 1986; Zhang et al. 1998), reporting on the results of the World Health Organization (WHO)–sponsored psychiatric epidemiology study in China. Among all psychiatric disorders, prevalence rates for personality disorder were the lowest, at 0.013%, with no cases found in the rural areas. A cross-sectional study that sampled 2,205 university freshmen reported that the prevalence rate of personality disorders was 2.5% (Huang et al. 2000). In a survey of senior high school students in Beijing, Huang and colleagues found a low PD prevalence rate of 1.8% (Huang et al. 2002). In evaluating the very low prevalence of personality disorders reported in China, one may consider several interesting possibilities. First, the Chinese surveying instruments may be inefficient in detecting personality disorder cases as defined by Western-derived diagnostic instruments. A second possibility is that the Chinese have a unique perception of personality disorders, different than that of Americans and Europeans. A third possibility is that there may be a lack of enthu143

144

Dimensional Models of Personality Disorders

siasm among Chinese surveyors toward personality disorder diagnosis. The most interesting possibility, however, would be that the Chinese do, in fact, have a lower number of personality disorder cases compared to Western countries. Chinese psychiatrists in China offer various explanations for the lack of interest or research focused on personality disorders. First, there is a lack of resources committed to the study of personality disorders in China. Second, this topic is not the focus of Chinese psychiatry. Many psychiatrists believe that the study of personality disorders should be left to psychologists, social workers, or educators. Finally, some Chinese psychiatrists believe that Western-derived diagnostic categories of personality disorder are difficult to apply to the Chinese. This lack of enthusiasm may change in the near future. Increasing interaction with American or European psychiatrists may encourage more attention to this area of psychiatry. Also, the strong economic growth in China may begin to generate additional resources for the Chinese mental health service and, thus, funds for research. The recent international efforts in improving the diagnostic instruments for personality disorders may also facilitate further work in China.

Cross-Cultural Diagnosis of Personality Disorders On the surface, it seems that every society should be able to identify those whose behavior or personalities are “unacceptable.” Using Tyrer’s criteria, it should be possible to further qualify whether these unacceptable personality traits are longterm, harmful to self or others, and/or deeply ingrained and hardly changeable, and qualitatively different from societal norms. Instruments for diagnosing personality disorders so far have not been shown to be highly reliable. Different diagnostic instruments seemed to show significant disagreement on personality disorder cases (Zimmerman 1994). The difficulty in diagnosing personality disorders is magnified when one attempts a cross-cultural comparison of statistics. Acceptability of behavior is very much culture-bound. Indeed, DSM-IV (American Psychiatric Association 1994), in its general diagnostic criteria for a personality disorder, specially states that the behavior in question “deviates markedly from the expectations of the individual’s culture.” There is no lack of literature on the importance of contextualization in assessing behavior cross-culturally. Put simply, the variation in societal norms among different cultures causes the same behavior to be seen as abnormal or tolerated within a given society. When a behavior is not perceived as grossly abnormal, then it may not be viewed as evidence of a personality disorder. Social or occupational impairment are similarly dependent on the element of abnormality. If the behavior were not perceived as grossly deviant, then functional impairment would not necessarily follow.

Huang and Tang: Commentary on Allik

145

What is or is not a culturally acceptable behavior is not an easy question to answer, as so many variables are involved. It is not simply what a physician or an interviewer from that culture sees as inappropriate or appropriate. Even a structured, standardized diagnostic interview may leave too much flexibility to physicians or interviewers with regard to cultural appropriateness. Obviously, personality disorders are much less well defined than schizophrenia, and the definitions are much more subject to cultural factors. The decision of DSM-IV to place emphasis on the cultural element in the diagnosis of personality disorders, the work of the National Institute of Mental Health–sponsored group on culture and diagnoses, and the World Health Organization international study on personality disorders (Loranger et al. 1994) are all constructive steps toward making the diagnosis of personality disorders more feasible in a culturally diverse world. Nonetheless, the development and validation of specific and stringent guidelines, culturally tuned diagnostic criteria and interview schedules, as well as appropriate translation of these into the native languages, are all extremely time-consuming and complex. Before this can happen, the blind application of cultural “yardsticks” runs the risk of subjective interpretation of personality disorders in the name of culture. In essence, this will allow a clinician, not wanting to label certain behavior as abnormal for whatever reasons, to avoid doing so. Moreover, our study on parental rearing behaviors indicated that it is common to encounter situations in which people from the same culture disagree about whether or not certain behaviors are acceptable within their own culture (Huang et al. 1996). This is frequently encountered in a society undergoing rapid changes when successive generations adopt different behavioral codes in adapting to societal changes. In these countries, many members of the older generation may hold very different views on the acceptability of certain behaviors compared with younger members. Thus, a culturally “tuned” diagnostic instrument is very much needed for cross-cultural comparison of personality disorder statistics. Before the arrival of a valid and reliable culturally tuned instrument for cross-cultural re-survey, it would be fair to say that application of current diagnostic/survey tools or instruments without appropriate modification may not yield valid prevalence data. However, it still does not explain the exceptionally low prevalence rate of personality disorders in China.

Future Research From the above discussion, it becomes apparent that there are some important and major cultural differences between the Chinese and their Western counterparts concerning their views on personality and personality disorders. The data of Hwu et al. (1983), together with that of Hyler et al. (1990), suggest that administration of a questionnaire with high sensitivity, followed by standardized diagnostic interviews with high specificity, is the preferable method of

146

Dimensional Models of Personality Disorders

surveying personality disorders in a foreign culture. To obtain meaningful statistical data in China for comparison with Western statistics, new diagnostic and survey instruments are definitely required. Tyrer et al. (1991) pointed out the interesting nature of the DSM group of personality disorders, in that their origins were mostly heterogeneous. In any case, their origins were all European-American and the interpretation is heavily “loaded.” It is obviously unreasonable, in the absence of research data, to assume a priori that the manifestation or presentation of the different types of DSM-IV personality disorders will be the same in Asians, Europeans, and Americans. The concept of borderline, histrionic, narcissistic, and dependent personality disorders particularly may find difficulties in acceptance by the Chinese at this point, whereas paranoid, schizoid, schizotypal, antisocial, and obsessive-compulsive personality disorders may be easier to define cross-culturally. Therefore, stringent descriptions of the behaviors in question and the avoidance of highly culturally subjective or culturally ambiguous terms and statements would help (Tang and Huang 1995). It would be most interesting for psychiatrists as well as for other behavioral scientists if it were true that personality disorders are indeed rare in China. Either rarity or overexpression of personality disorders in one culture versus another, when properly defined and confirmed, would offer many valuable research opportunities. Personality disorders are promising diagnostic candidates for studying the interplay of genetic, hormonal, early childhood rearing behaviors, and family and societal influences and forces. We look forward to a new generation of transcultural psychiatric research to answer some of these most intriguing questions.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Chen C, Zhang W, Shen Y, et al: Data analysis of an epidemiological study on mental disorders, drug and alcoholic dependence and personality disorders. Chinese Journal of Neurology and Psychiatry 19:70–72, 1986 Huang Y, Someya T, Takahashi S, et al: A pilot evaluation of the EMBU Scale in Japan and the United States. Acta Psychiatr Scand 94:445–448, 1996 Huang Y, Liu X, Li L, et al: An epidemiological study of personality disorder and related risk factors in undergraduate students. Chinese Journal of Psychiatry 33:44–46, 2000 Huang Y, Liu B, Liu Z, et al: A cross-sectional study of personality dysfunction among students of first-grade senior high schools in Beijing. Zhonghua Liu Xing Bing Xue Za Zhi 23:338–340, 2002 Hwu HG, Yeh EK, Chen CT, et al: An applicability study of the Chinese modification of Diagnostic Interview Schedule (DIS-CM). Bulletin of Chinese Society of Neurology and Psychiatry 9:136–145, 1983

Huang and Tang: Commentary on Allik

147

Hyler SE, Rieder RO, Williams JBW, et al: Validity of the Personality Diagnostic Questionnaire—Revised: comparison with two structured interviews. Am J Psychiatry 147:1043–1048, 1990 Loranger AW, Sartorius N, Andreoli A, et al: The International Personality Disorder Examination. Arch Gen Psychiatry 51:215–224, 1994 Tang S, Huang Y: Diagnosing personality disorders in China. International Medical Journal 2:291–297, 1995 Tyrer P, Casey P, Ferguson B: Personality disorder in perspective. Br J Psychiatry 159:463– 471, 1991 Zhang WX, Shen YC, Li SR, et al: Epidemiological study on mental disorder in seven regions in China. Chinese Journal of Psychiatry 31:69–71, 1998 Zimmerman M: Diagnosing personality disorder. Arch Gen Psychiatry 51:225–245, 1994

This page intentionally left blank

16 CONTINUITY OF AXES I AND II Toward a Unified Model of Personality, Personality Disorders, and Clinical Disorders Robert F. Krueger, Ph.D.

I

n DSM-IV-TR (American Psychiatric Association 2000), mental disorders are divided into two groups: clinical disorders (CDs) and personality disorders (PDs). CDs are recorded on the first of the five axes used to classify patients, and PDs are described on the second of the five axes. DSM-IV-TR describes this arrangement as linked to the importance of considering the possible presence of a PD even when a potentially more florid CD is present. In addition, DSM-IV-TR is careful to note that the separation of mental disorders into CDs and PDs need not imply fundamental distinctions in terms of “pathogenesis or range of appropriate treatment” (p. 28) for PDs and CDs. The purpose of this paper is to consider this arrangement in terms of both theory and data, in an attempt to inform the construction of future editions of the DSM. Shall we continue to separate mental disorders into CDs and PDs? Or has the time come to reconsider this arrangement, in light of recent research developments? Although DSM-IV-TR is careful to avoid overstating distinctions between CDs and PDs, the separation of mental disorders across Axes I and II continues to

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 233–261, 2005). Preparation of this manuscript was supported in part by U.S. Public Health Service grant MH65137.

149

150

Dimensional Models of Personality Disorders

influence the way these disorders are often conceived of in both research and clinical practice. That is, CDs and PDs are often conceived of as more distinct than they actually are, at least based on current evidence. In addition, I will argue that the closeness of the connection between CDs and PDs can be understood when viewed through the prism of general personality functioning. Basic research in personality and individual differences provides a framework that is very useful in understanding psychopathology. Integration of this basic research with research on constructs from the DSM provides an important avenue for linking heretofore somewhat distinct literatures, as well as for helping to link science and practice. Indeed, the focus of this conference on dimensional models, many of which derive from basic research in personality and individual differences, is a strong sign that a more integrated perspective on basic personality constructs and DSM constructs is both desirable and feasible. The state of the evidence regarding putative bases for distinguishing between CDs and PDs is reviewed in an extended version of this paper (Krueger 2005). Herein, I review evidence regarding the joint structure of normal and abnormal personality. I then discuss how personality constructs are also connected to CDs, and outline how these lines of evidence could be brought together under a more comprehensive model of personality and psychopathology.

The Joint Structure of Normal and Abnormal Personality PDs and CDs are not well distinguished in terms of stability, age at onset, treatment response, insight, comorbidity and symptom specificity, or etiology (Krueger 2005). The most promising general direction for research would therefore focus on understanding how and why the PD and CD domains are so interconnected. A better understanding of connections between PDs and CDs may be provided by a focus on how PDs and CDs are both connected to the structure of personality. A first issue, then, is how personality is connected with PDs—that is, What is the joint structure of normal and abnormal personality? A bewildering array of personality constructs—as well as models for organizing those constructs—have been proposed to account for human personality variation. In addition, literatures on both normal and abnormal personality variation have evolved somewhat independently (Strack and Lorr 1994). In spite of this, increasing evidence points to the feasibility of developing an empirically based model of personality that simultaneously incorporates both normal and abnormal variation (Trull and Durrett 2005). One approach to delineating such an integrative model focuses on points of intersection between various normal and abnormal personality models that have been articulated in the literature. Markon et al. (2005) report two studies that take

Krueger: Continuity of Axes I and II

151

this approach. The first study focused on joint structural modeling of constructs measured by the DAPP (Livesley and Jackson, in press), the Eysenck Personality Questionnaire (EPQ and EPQ-R; Eysenck and Eysenck 1975; Eysenck et al. 1985), the Multidimensional Personality Questionnaire (MPQ; Tellegen 1985), variants of the NEO-PI broad domain scales (NEO-PI, NEO-PI-R, and NEOFFI; Costa and McCrae 1985, 1992), and the Temperament and Character Inventory and its predecessors (TCI and TPQ; Cloninger 1987; Cloninger et al. 1993). Markon et al. (2005) took a meta-analytic approach to assembling a matrix of correlations among the 44 scales derived from all of these inventories, combining data from 52 different studies. Structural modeling of the meta-analytically derived matrix yielded the following conclusions. First, the data indicated no more than five major factors underlying variation in the 44 scales. Second, these five factors (neuroticism, agreeableness, conscientiousness, extraversion, and openness) strongly resembled the factors of the five-factor model (FFM) that has been suggested as a useful framework for understanding PDs (see Costa and Widiger 2002). Nevertheless, additional analyses also supported the existence of meaningful factors above the level of the five factors. Specifically, the four-factor level resembled four-factor models often articulated in the personality and psychopathology literature (e.g., Livesley et al. 1998; O’Connor and Dyce 1998; Watson et al. 1994), in that openness did not emerge as a separate factor, but combined with extraversion in a broader dimension of positive emotionality. The three-factor level resembled the three-factor models of theorists such as Clark and Watson (1999), Eysenck (1994), and Tellegen (1985), with dimensions of negative emotionality (neuroticism), disinhibition (a combination of disagreeableness and unconscientiousness) and positive emotionality. Finally, the two-factor level resembled the two factor model described by Digman (1997), with one factor (alpha) combining neuroticism, agreeableness, and conscientiousness, and the other factor (beta) combining extraversion and openness. The second study presented in Markon et al. (2005) replicated this hierarchical model using a specific sample of participants (as opposed to a meta-analytic data set) who completed a somewhat different set of instruments: the NEO PI-R facet scales (Costa and McCrae 1992), the EPQ-R (Eysenck and Eysenck 1975; Eysenck et al. 1985), the Schedule for Nonadaptive and Adaptive Personality (SNAP; Clark 1993), and the Big Five Inventory (BFI; John et al. 1991). This suggests that the hierarchical model identified in the meta-analytic portion of the research can also be identified in a specific sample, and that the model is robust to variation in the instruments and scales used to measure personality. The Markon et al. (2005) findings both reinforce and extend some basic conclusions about the structure of normal and abnormal personality. The findings indicate that normal and abnormal personality constructs can be integrated within the same structural model (see also O’Connor and Dyce 2001). Rather than representing highly separate domains of human individual differences, normal and

152

Dimensional Models of Personality Disorders

abnormal personality measures can be located in the same factor space. In addition, the findings speak to the importance of the FFM in understanding the joint structure of normal and abnormal personality. The FFM appears to represent the “base” of the higher-order structure that links normal and abnormal personality, as there was no compelling evidence for factors beyond the five. However, there was evidence for structures above the five, and these structures resemble major conceptions of personality that complement the FFM conception. This helps in integrating various models and addressing questions of which dimensional model might be most empirically compelling. One concern that is sometimes raised regarding conversion from the categorical model of PDs in the DSM to a dimensional model is the question of which dimensional model should be chosen (cf. Livesley 2001). The Markon et al. (2005) results suggest that the FFM level of the hierarchy is the basic organizing framework of choice because there was no compelling evidence for higher-order structures beyond the five, and structures above the five can be understood as combinations of the FFM domains. Nevertheless, it would be incorrect to conclude that the FFM level of the personality hierarchy provides all the information necessary to capture personality variation (cf. Harkness 1992). Structures above the FFM, as well as the facet level scales that delineate the FFM domains, are also theoretically and clinically important. For example, the domain of disinhibition, which is above the FFM level and combines disagreeableness and unconscientiousness, is closely linked to antisocial behavior (e.g., Lynam and Derefinko 2006). In addition, many of the scales in the Markon et al. (2005) analyses contained substantial amounts of residual variance that could not be accounted for by the higher order factors. For example, the DAPP Self-Harm scale was a marker of the broad neuroticism domain of the FFM (more so than of any of the other domains), but the majority of the variance in DAPP Self Harm was unique to this scale, and not accounted for by the FFM domains. Self harm is of obvious clinical importance and, inasmuch as it can not be captured entirely by the broader neuroticism domain, it represents an example of a specific, facet-level construct below the level of the FFM that might be important to include in a complete system of abnormal personality description. Indeed, the issue of which facet-level constructs should be included in a comprehensive system of abnormal personality description is an important topic for continued research and discussion. One thing that seems clear is that the ability of the FFM to capture the clinical PD constructs described in the DSM is enhanced by a focus on the facet level of the FFM, as specific facets have differential relevance to specific DSM PDs (see, e.g., Table 6.1 in Widiger et al. 2002). Although the Markon et al. (2005) analyses demonstrate that the facet-level scales in diverse inventories can be organized into the FFM domains, such analyses do not provide guidance regarding which facets are most optimal for clinical description, and the facet-level scales included in the inventories linked by the FFM domains remain distinctive. Consider the earlier example of self harm, which is a facet in the DAPP

Krueger: Continuity of Axes I and II

153

but not, e.g., in the NEO PI-R. Some might argue that self harm is too specific a construct, and might be collapsed into a broader personality facet, or that the DAPP Self-Harm scale lacks sufficient variability to constitute a major facet level construct (e.g., the scale was omitted from analyses presented in Livesley et al. 1998, owing to low variability). Others might argue that the overwhelming clinical importance of self harm justifies its inclusion as a facet-level construct in an official nosology. The point, however, is that this kind of discussion is vital in moving towards a consensus set of facet level abnormal personality constructs suitable for inclusion in future revisions of the DSM. The detailed information about personality functioning contained in facet-level constructs is important in connecting broad domains of personality functioning with the richness of the clinical phenomena encountered in PDs (see Shedler and Westen 2004 for a related perspective). Developments in methodology can also aid in informing these discussions. For example, the general analytic approach to delineating the broad domains that organize personality variation (factor analysis) may not be well suited to delineating fine-grained distinctions between personality constructs; continued methodological development in this area is highly pertinent to better understanding the facet level structure of abnormal personality (e.g., Bacon 2001). In addition, many factor analysis models in current use are formulated such that they do not contain parameters describing the location of variables along latent dimensions. Typical models are capable of identifying the strength of the relationships between the variables and the factors (so-called “factor loadings”) but they are not formulated to also identify the locations of the variables along the factors. Models that also include location parameters have potential to enrich our conceptualizations of personality and psychopathology. For example, Krueger et al. (2004) used this type of model to illustrate how alcohol problems can be well-conceptualized in a dimensional framework, as problems ranging from those that were normative to those that were pathological were found to lie along a dominant dimension of severity. Finally, dimensions of personality such as the FFM domains may delineate the space in which personality is best described, but there can still be points of greater and lesser density in this space, a possibility that can be investigated using new developments in the modeling of multivariate data (Muthén 2002). These regions of greater density (if they exist) might be thought of as frequently encountered personality configurations, and some of these configurations might be of unusual clinical importance. For example, Hicks et al. (2004b) used this kind of approach (a model-based cluster analysis) to demonstrate that prisoners identified as psychopathic by the Psychopathy Checklist—Revised (Hare 1991) consisted of two groups with very different personality profiles. The first group resembled the “primary” psychopath, with unusually low stress reactivity, whereas the second group resembled the “secondary” psychopath, with unusually high levels of aggression. In sum, the FFM appears to represent a compelling model for organizing normal and abnormal personality variation, including the variation captured by the

154

Dimensional Models of Personality Disorders

PD categories described in DSM-IV-TR (Costa and Widiger 2002; Lynam and Widiger 2001; Miller et al. 2004; Trull et al. 2003; Warner et al. 2004). Further discussion and research could profitably focus on delineating the optimal facets for inclusion in official nosologies such as the DSM (even if these facets are well organized into the FFM domains), the ability of different facets to cover the entire range of normal and abnormal personality, and the possibility that certain personality configurations occur more frequently than others, and might be of particular clinical importance.

Toward Novel Models of Personality and Psychopathology Reconceptualizing PDs in personality dimensional terms represents a good start on providing a solid empirical footing for future editions of the DSM. Yet certain conundrums will remain even if PDs are reconfigured in personality trait terms. These conundrums pertain to CDs, as DSM-defined PDs, as well as dimensional personality traits, are both closely connected with CDs (for recent reviews, see Clark 2005; Krueger and Tackett 2003). What is ultimately needed is a model that can make sense of the connections linking all these domains (CDs, PDs, and the structure of normal and abnormal personality). The beginnings of such a model are provided by work on the structure of common mental disorders. The motivation behind this work has been to provide an understanding of the reasons why the mental disorders defined in official nosological systems such as the DSM and ICD are frequently comorbid. Rather than viewing comorbidity as an artifact or a nuisance, this work approaches comorbidity as a reliable empirical observation in need of an explanatory model. The focus of this work has been primarily on comorbidity among mental disorders commonly observed in epidemiological samples (unipolar mood, anxiety, substance use, and antisocial behavior disorders); in the DSM nosology, these disorders are mostly conceptualized as CDs (with the exception of antisocial personality disorder). These disorders represent key targets because of their prevalence and clear public health relevance. Our earlier work on the link between personality and these disorders in epidemiological samples (e.g., Krueger 1999a) yielded evidence of systematic links between “normal range” personality constructs and these disorders, both cross-sectionally and longitudinally. Specifically, the mood and anxiety disorders were associated with high levels of Neuroticism/Negative Emotionality, whereas the substance use and antisocial behavior disorders were associated with high levels of Neuroticism/Negative Emotionality and high levels of disinhibition. Given this pattern of findings, comorbidity among common mental disorders makes sense when thought of in terms of the personological underpinnings of

Krueger: Continuity of Axes I and II

155

these disorders. Krueger et al. (1998) demonstrated that unipolar mood and anxiety disorders were indicators of a latent internalizing propensity, and substance dependence and antisocial behavior disorders were indicators of a latent externalizing propensity (see also Krueger 1999b). Moreover, these propensities were highly stable over time, a finding replicated by Vollebergh et al. (2001). Recent evidence indicates that the internalizing–externalizing structure can also be observed in the primary care setting in numerous countries around the globe, and that the internalizing spectrum also appears to encompass somatoform syndromes (Krueger et al. 2003). Putting the personality findings together with the findings on the structure of mental disorders, neuroticism/negative emotionality appears to provide the personological basis for internalizing psychopathology, and negative emotionality paired with disinhibition appears to provide the personological basis for externalizing psychopathology. Thus, the connections between personality and psychopathology make psychological sense. Negative Emotionality is internalized given normative levels of disinhibition and presents as unipolar mood and anxiety disorder. If negative emotionality is paired with high levels of disinhibition, the presentation tends more toward externalizing (substance use and antisocial behavior) problems. These observations provide the outlines of a hierarchically organized spectrum model of common mental disorders that also extends to encompass the link between personality and psychopathology. The findings suggest that unipolar mood disorders, anxiety disorders, and Negative Emotionality form a coherent group of constructs, and substance use disorders, antisocial behavior disorders, and disinhibitory personality traits form a related (owing to the role of negative emotionality in both internalizing and externalizing problems) but also distinguishable (owing to the unique role of disinhibition in externalizing problems) group of constructs. This conceptualization is bolstered by research on the genetic underpinnings of the connections between psychopathological syndromes and personality traits. That is, unipolar mood and anxiety problems share significant genetic variance with the personality trait of Neuroticism; in addition, substance use and antisocial behavior problems share significant genetic variance with an unconstrained, impulsive personality style (see Krueger and Tackett 2003 for a review). This conceptualization is also bolstered by research on the genetic underpinnings of the structure of psychopathology. Kendler et al. (2003) recently presented a study showing that the observed phenotypic structure of common mental disorders closely mirrors the underlying genetic architecture of these constructs. Kendler et al. (2003) studied the genetic and environmental underpinnings of the comorbidity among seven syndromes (major depression, generalized anxiety disorder, phobia, alcohol dependence, other drug abuse or dependence, adult antisocial behavior, and conduct disorder) that delineate the internalizing and externalizing spectra. Two general genetic fac-

156

Dimensional Models of Personality Disorders

tors were found, the first related primarily to the internalizing disorders (major depression, generalized anxiety disorder, phobia) and the second related primarily to the externalizing disorders (alcohol dependence, other drug abuse or dependence, adult antisocial behavior, and conduct disorder). As noted by Kendler et al. (2003), “These results suggest strongly that genetic factors are largely responsible for the pattern of comorbidity that results in the two frequently co-occurring clusters of internalizing and externalizing disorders” (p. 935). In addition to these broad genetic factors, Kendler et al. (2003) also documented specific etiologic contributions that distinguish disorders within the broad internalizing and externalizing spectra. For example, finer-grained modeling of the internalizing syndromes of major depression, generalized anxiety disorder, panic disorder, animal phobia, and situational phobia revealed evidence for genetically distinguishable (but correlated) subfactors within the broader internalizing domain, with one subfactor loading more on depression and generalized anxiety disorder (with panic disorder as a weaker marker), and the other loading more on animal phobia, and situational phobia (cf. Krueger 1999b; Watson 2005). Specific genetic factors also contributed to alcohol and drug abuse/dependence, above and beyond the contribution of the overarching genetic externalizing factor. In addition, environmental factors were important contributors to distinguishing between closely related syndromes. For example, conduct disorder showed a unique contribution from the shared family environment (cf. Hicks et al. 2004a; Krueger et al. 2002). These specific etiological contributions in the context of broad, genetically coherent spectra are important phenomena because they speak to the hierarchical organization of the structure of common mental disorders. The idea behind this work has never been that there are only two constructs of relevance in understanding common mental disorders (internalizing and externalizing). Rather, the idea is that the internalizing and externalizing constructs provide the broad organizational schema and sources of genetic coherence for this domain, and are therefore major sources of the comorbidity among common mental disorders; in addition to this, other constructs (e.g., unique genetic and environmental events) explain how syndromes closely connected by shared genetic etiology come to be distinguishable. Put somewhat differently, both “lumping” and “splitting” perspectives on the organization of psychopathology are partially correct, and they can be reconciled by adopting a dimensional-hierarchical model of etiologic contributions within this domain that recognizes etiologic factors at continually varying levels of specificity versus breadth (cf. Krueger and Piasecki 2002). Integrating this ongoing work on the structure of mental disorders into the DSM system represents a complex challenge, but it is a challenge that seems worth pursuing if the goal is to place the organization of the DSM on solid empirical footing. One way in which the organizational structure of future editions of the DSM could reflect the empirical structure of mental disorders would be to organize the syndromes that have been studied to date into internalizing and external-

Krueger: Continuity of Axes I and II

157

izing sections. The organization of internalizing syndromes has been particularly problematic in DSM (see, e.g., Widiger 2003). Watson (2005) has recently articulated an approach to resolving some of these problems. Watson’s (2005) approach focuses on delineating facets within the internalizing spectrum in terms of the distinction between disorders that are more distress-related (e.g., major depression and generalized anxiety disorder), those that are more fear-related (e.g., panic disorder and phobias), and those that involve bipolarity of mood (e.g., bipolar I and II, cyclothymia). This reorganization reflects current knowledge of the structure of internalizing syndromes better than the putative distinction between mood and anxiety disorders (cf. Kendler et al. 2003; Krueger 1999b). Organizing externalizing syndromes into a coherent section also requires rethinking some basic aspects of the current organization of the DSM. The externalizing spectrum encompasses problems that are currently spread throughout the DSM, across sections covering substance-related disorders, disorders usually first diagnosed in infancy, childhood, or adolescence (conduct disorder), and the section on PDs (antisocial PD). Yet empirical evidence continues to speak to the coherence of the externalizing spectrum, as well as the dimensional nature of this spectrum. For example, Krueger et al. (2005) studied the comorbidity among the syndromes of conduct disorder, adult antisocial behavior, alcohol dependence, marijuana dependence, and drug dependence in a large, representative sample of adults. A series of models were fit to the data to ask if the comorbidity among these syndromes could be better accounted for in terms of a set of categories, vs. in terms of a coherent dimension of liability to experience multiple disorders in the externalizing realm. The data better supported a dimensional conceptualization, with an overarching dimension of externalizing liability connecting the disorders. A model specifying five separate categories of substance disorders and antisocial behavior disorders was untenable. Another challenging issue relates to disorders described in the DSM that have not been studied in the internalizing-externalizing framework. The internalizingexternalizing framework provides only the beginnings of a comprehensive model of clinical psychopathology because many important forms of psychopathology have not been studied in the context of the internalizing-externalizing framework. This owes primarily to the limitations of existing data. For example, data on psychotic disorders are collected in large-scale investigations of comorbidity, but these disorders are conceptualized in such a way that they have very little variance in the population at large. This makes studying relations between these disorders and common mental disorders infeasible because there are too few cases of less common disorders—much less observations of patterns of comorbidity linking less common disorders with more common disorders—to allow for reliable conclusions to be drawn regarding less common disorders. Two strategies are likely to be productive in overcoming this obstacle to broadening the internalizing–externalizing framework. A first strategy would focus on

158

Dimensional Models of Personality Disorders

integrating novel domains of psychopathology into the framework in samples where prevalence is enhanced (e.g., in samples from psychiatric clinics). For example, Krueger et al. (2003) were able to broaden the internalizing spectrum to include somatoform disorders by studying the model in primary care samples. A second strategy would focus on broadening conceptualizations of less common syndromes. For example, rather than focusing on dichotomous CDs such as schizophrenia, research could focus on dimensions of psychosis such as positive symptoms, negative symptoms, and disorganization. Emerging work on dimensional approaches to psychosis (e.g., van Os et al. 1999) could be integrated with ongoing work on the internalizing and externalizing dimensions that appear to underlie common mental disorders. Another issue relates to the placement of personality constructs in a system that recognizes the internalizing and externalizing spectra, as these spectra transcend personality and psychopathology. It might make the most sense to ultimately focus not on the putative PD–CD distinction, but rather, on the generation of key facetlevel constructs that cover the range of symptomatology and traits currently spread across Axes I and II of DSM-IV-TR. This idea was well-articulated by Widiger and Clark (2000, p. 954) who wrote, “Ultimately, as one builds toward DSM-V, what may emerge is a structured set not of categorical diagnoses but of component dimensions, a set of symptom-cluster building blocks from which the panoply of diagnoses could be constructed.” Based on our current understanding, personalityrelated facets (building blocks) appear well organized by the FFM domains (although more work is needed to identify the most optimal facets for clinical purposes), and common psychopathological symptoms appear well organized by the internalizing-externalizing domains. With regard to the latter, the internalizing and externalizing domains have emerged from analyses of DSM categories, and hence, more work is needed that breaks these diagnoses down into their component symptom dimensions. As noted by Widiger and Clark (2000), models for the fundamental dimensions that constitute the internalizing domain continue to be developed (Mineka et al. 1998; Watson 2005), and efforts to delineate fundamental dimensions of externalizing are also underway (Krueger et al. 2005). Clearly, however, much work is needed before a reorganization of the DSM along the lines adumbrated here is possible. Thus, for the time being, some intermediate steps can be suggested. First, as we build toward DSM-V, we should consider converting the existing PD section to a system of facet-level constructs organized by the FFM domains, and pursue research and discussion on the most optimal facets, as well as the most optimal cut points on those facets to distinguish between personality and its pathological manifestations (cf. Livesley 2001). That is, adoption of a dimensional approach to the description of personality in the DSM does not mean abandoning a distinction between normality and abnormality. To the contrary, it sharpens discussions about how dimensions of symptoms and traits are related to concepts such as mental disorder and psychopathology. It

Krueger: Continuity of Axes I and II

159

necessitates research and discussion focused on the links between these concepts. For example, even if alcohol problems are dimensional in nature, such that there are no natural cut points demarcating heavy drinking, abuse, and dependence (Krueger et al. 2004), there are levels of alcohol problems beyond which society and professionals would deem intervention warranted. Understanding where relevant cut points lie would involve research on the way in which symptomatology per se is linked to consequences (e.g., social and occupational dysfunction), and discussion with professionals and policy makers about levels of consequences that societies are unwilling to tolerate. Second, we should consider reorganizing sections and disorders described in the DSM-IV-TR to recognize the internalizing and externalizing spectra, and pursue research and discussion on the most optimal organization of diagnoses within these spectra. Finally, we should encourage further research linking cutting edge developments in methodology with novel ideas about how to describe and organize personality and psychopathology constructs. For example, Widiger and Clark (2000) envisioned the creation of a “diagnostic table of the elements” for future editions of the DSM. What are the main dimensions that would organize such a table? Do these represent some meshing between the FFM and the internalizing and externalizing spectra? What are the quantitative properties that distinguish one “element of psychopathology” from another? Open-minded, creative research that asks probing and novel questions but sticks close to the data in pursuing these questions has real potential to ultimately result in a diagnostic system that is empirically supported, useful in the clinic, and inspires research that leads to better prevention and treatment of mental disorders.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Bacon DR: An evaluation of cluster analytic approaches to initial model specification. Structural Equation Modeling 8:397–429, 2001 Clark LA: Manual for the Schedule for Nonadaptive and Adaptive Personality. Minneapolis, MN, University of Minnesota Press, 1993 Clark LA: Temperament as a unifying basis for personality and psychopathology. J Abnorm Psychol 114:505–521, 2005 Clark LA, Watson D: Temperament: a new paradigm for trait psychology, in Handbook of Personality: Theory And Research, 2nd Edition. Edited by Pervin LA, John OP. New York, Guilford, 1999, pp 399–423 Cloninger CR: A systematic method for clinical description and classification of personality variants. Arch Gen Psychiatry 44:573–588, 1987 Cloninger CR, Svrakic DM, Przybeck TR: A psychobiological model of temperament and character. Arch Gen Psychiatry 50:975–990, 1993

160

Dimensional Models of Personality Disorders

Costa PT Jr, McCrae RR: The NEO Personality Inventory Manual. Odessa, FL, Psychological Assessment Resources, 1985 Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992 Costa PT Jr, Widiger TA (Editors): Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Washington, DC, American Psychological Association, 2002 Digman JM: Higher-order factors of the Big Five. J Pers Soc Psychol 73:1246–1256, 1997 Eysenck HJ: Normality-abnormality and the three factor model of personality, in Differentiating Normal and Abnormal Personality. Edited by Strack S, Lorr M. New York, Springer, 1994, pp 3–25 Eysenck HJ, Eysenck SBG: Manual of the Eysenck Personality Questionnaire. London, Hodder & Stoughton, 1975 Eysenck SBG, Eysenck HJ, Barrett P: A revised version of the Psychoticism scale. Pers Indiv Differ 6:21–29, 1985 Hare RD: Manual for the Hare Psychopathy Checklist—Revised, 1st Edition. Toronto, Canada, Multi-Health Systems, 1991 Harkness AL: Fundamental topics in the personality disorders: candidate trait dimensions from lower regions of the hierarchy. Psychol Assessment 4:251–259, 1992 Hicks BM, Krueger RF, Iacono WG, et al: Family transmission and heritability of externalizing disorders: a twin-family study. Arch Gen Psychiatry 61:922–928, 2004a Hicks BM, Markon KE, Patrick CJ, et al: Identifying psychopathy subtypes on the basis of personality structure. Psychol Assessment 16:276–288, 2004b John OP, Donahue EM, Kentle RL: The Big Five Inventory—Versions 4a and 5. Berkeley, CA, University of California, Berkeley, Institute of Personality and Social Research, 1991 Kendler KS, Prescott CA, Myers J, et al: The structure of genetic and environmental risk factors for common psychiatric and substance use disorders in men and women. Arch Gen Psychiatry 60:929–937, 2003 Krueger RF: Personality traits in late adolescence predict mental disorders in early adulthood: a prospective-epidemiological study. J Pers 67:39–65, 1999a Krueger RF: The structure of common mental disorders. Arch Gen Psychiatry 56:921–926, 1999b Krueger RF: Continuity of Axes I and II: toward a unified model of personality, personality disorders, and clinical disorders. J Personal Disord 19:233–261, 2005 Krueger RF, Piasecki TM: Toward a dimensional and psychometrically informed approach to conceptualizing psychopathology. Behav Res Therapy 40:485–499, 2002 Krueger RF, Tackett JL: Personality and psychopathology: working toward the bigger picture. J Pers Disord 17:109–128, 2003 Krueger RF, Caspi A, Moffitt TE, et al: The structure and stability of common mental disorders (DSM-III-R): a longitudinal-epidemiological study. J Abnorm Psychol 107:216–227, 1998 Krueger RF, Hicks BM, Patrick CJ, et al: Etiologic connections among substance dependence, antisocial behavior, and personality: modeling the externalizing spectrum. J Abnorm Psychol 111:411–424, 2002

Krueger: Continuity of Axes I and II

161

Krueger RF, Chentsova-Dutton YE, Markon KE, et al: A cross cultural study of the structure of comorbidity among common psychopathological syndromes in the general health care setting. J Abnorm Psychol 112:437–447, 2003 Krueger RF, Nichol PE, Hicks BM, et al: Using latent trait modeling to conceptualize an alcohol problems continuum. Psychol Assessment 16:107–119, 2004 Krueger RF, Markon KE, Patrick CJ, et al: Externalizing psychopathology in adulthood: a dimensional-spectrum conceptualization and its implications for DSM-V. J Abnorm Psychol 114:537–550, 2005 Livesley WJ: Conceptual and taxonomic issues, in Handbook of Personality Disorders. Edited by Livesley WJ. New York, Guilford, 2001, pp 3–38 Livesley WJ, Jackson D: Manual for the Dimensional Assessment of Personality Pathology—Basic Questionnaire. Port Huron, MI, Sigma Press, in press Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Lynam DR, Derefinko KJ: Psychopathy and personality, in Handbook of Psychopathy. Edited by Patrick C. New York, Guilford, 2006, pp 133–155 Lynam DR, Widiger TA: Using the five-factor model to represent the DSM-IV personality disorders: an expert consensus approach. J Abnorm Psychol 110:401–412, 2001 Markon KE, Krueger RF, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 Miller JD, Reynolds SK, Pilkonis PA: The validity of the five-factor model prototypes for personality disorders in two clinical samples. Psychol Assessment 16:310–322, 2004 Mineka S, Watson D, Clark LA: Comorbidity of anxiety and unipolar mood disorders. Ann Rev Psychol 49:377–412, 1998 Muthén BOP: Beyond SEM: general latent variable modeling. Behaviormetrika 29:81– 117, 2002 O’Connor BP, Dyce JA: A test of models of personality disorder configuration. J Abnorm Psychol 107:3–16, 1998 O’Connor BP, Dyce JA: Rigid and extreme: a geometric representation of personality disorders in five-factor model space. J Pers Soc Psychol 81:1119–1130, 2001 Shedler J, Westen D: Refining personality disorder diagnosis: integrating science and practice. Am J Psychiatry 161:1350–1365, 2004 Strack S, Lorr M (eds): Differentiating Normal and Abnormal Personality. New York, Springer, 1994 Tellegen A: Structures of mood and personality and their relevance to assessing anxiety with an emphasis on self-report, in Anxiety and the Anxiety Disorders. Edited by Tuma AH, Maser JD. Hillsdale, NJ, Lawrence Erlbaum, 1985, pp 681–706 Trull TJ, Durrett CA: Categorical and dimensional models of personality disorder. Annu Rev Clin Psychol 1:355–380, 2005 Trull TJ, Widiger TA, Lynam DR, et al: Borderline personality disorder from the perspective of general personality functioning. J Abnorm Psychol 112:193–202, 2003 van Os, J, Gilvarry C, Bale R, et al: A comparison of the utility of dimensional and categorical representations of psychosis. Psychol Med 29:595–606, 1999

162

Dimensional Models of Personality Disorders

Vollebergh WAM, Iedema J, Bijl RV, et al: The structure and stability of common mental disorders: the NEMESIS study. Arch Gen Psychiatry 58:597–603, 2001 Warner MB, Morey LC, Finch JF, et al: The longitudinal relationship of personality traits and disorders. J Abnorm Psychol 113:217–227, 2004 Watson D: Rethinking the mood and anxiety disorders: a quantitative hierarchical model for DSM-V. J Abnorm Psychol 114:522–536, 2005 Watson D, Clark LA, Harkness AR: Structures of personality and their relevance to psychopathology. J Abnorm Psychol 103:18–31, 1994 Widiger TA: Personality disorder and Axis I psychopathology: the problematic boundary of Axis I and Axis II. J Personal Disord 17:90–108, 2003 Widiger TA, Clark LA: Toward DSM-V and the classification of psychopathology. Psychol Bull 126:946–963, 2000 Widiger TA, Trull TJ, Clarkin JF, et al: A description of the DSM-IV personality disorders with the Five-Factor Model of Personality, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 89–99

17 COMMENTARY ON KRUEGER What to Do With the Old Distinctions M. Tracie Shea, Ph.D.

Dr. Krueger’s integrative review of a large body of empirical and conceptual work addressing relationships between Axis I, or clinical disorders, and Axis II, personality disorders, raises important considerations. One is the absence of support for distinctions that have been presumed to distinguish clinical disorders from personality disorders. Another is the proposal that personality dimensions are key to understanding the links and co-occurrences among the clinical disorders and personality disorders. These among other issues raised and discussed in this paper have direct implications for the structure of the next DSM, including whether to continue the Axis I–Axis II distinction, and for rethinking the optimal organizational scheme for all mental disorders. In my comments below I attempt to provide additional context for such directions, mostly in the form of considering aspects of the “old” that may be important to consider in any integrative “new” scheme for mental disorders. Soon after the introduction of Axis II, the question of how to understand its association with Axis I arose. One of the rationales for Axis II was to draw attention to personality disorders: “The listing of Personality Disorders...on a separate axis ensures that consideration will be given to the possible presence of personality disorders...that might otherwise be overlooked when attention is drawn to the usually more florid Axis I disorders” (American Psychiatric Association 1994, p. 26). The wisdom of this division has become increasingly questioned, in part due to the consistent findings of untenably high rates of overlap and co-occurrence. In the absence of support for most of the presumed distinctions such as sta163

164

Dimensional Models of Personality Disorders

bility, as noted by Krueger, the rationale for the distinction represented by the two Axes becomes even more questionable. Several models have been offered to explain the high rates of co-occurrence (e.g., Klein et al. 1993). One of these models describes maladaptive personality traits as conveying a vulnerability to clinical disorders. Conceptually, this vulnerability model is consistent with the idea of a separate axis for personality disorders. Dr. Krueger’s review is consistent with a “spectrum” model, with different disorders across Axis I and Axis II linked by shared latent, genetically based dimensions. The spectrum notion has often been proposed as explaining links between pairs of disorders, such as schizotypal and schizophrenia, or borderline personality disorder and bipolar disorder. Siever and Davis (1991) were the first to explicitly offer a model of common dimensions underlying multiple disorders across Axis I and II, although other dimensional schemes had linked personality dimensions with Axis I disorders (e.g., Clark et al. 1994). Dr. Krueger’s review integrates the very large literature on the structure of personality and personality disorders with the findings from a more limited number of studies of structure underlying disorders primarily categorized on Axis I, in essence broadening the findings from personality into the domain of all mental disorders. Although he suggests the two broad constructs of internalizing and externalizing as an organizing scheme, he also emphasizes that this is simply the broadest level of a hierarchical organization of dimensions. Thus, this organizational scheme is very flexible; the issue is not one of determining the correct number of dimensions, but rather what level of specificity may be optimally useful. The appeal and advantage of this approach is its integrative nature, and the strong empirical base of support. I believe that it is difficult to defend the current distinction embodied by the two axes for the many reasons articulated in Dr. Krueger’s paper. Having said this, however, some of the original conceptual distinctions, while not well served in the current system, may nonetheless be important to consider. One distinction is the notion of “symptom” versus “trait.” A prototypical notion of Axis I disorders emphasizes symptoms that appear in circumscribed episodes, in contrast to the more enduring maladaptive “traits” of the personality disorders. In addition to the difference in temporal pattern, the use of the term symptom connotes something that is an indicator of an underlying disease or disorder, whereas the term trait is usually defined in terms of an enduring propensity to show a characteristic and distinguishing pattern of behaviors, attitudes, or feelings. As the Axis I disorders were more explicitly defined and studied over time, it became clear that many of the symptoms tended to be chronic and persistent. Furthermore, many of the maladaptive “traits” of the personality disorders have been shown to be less persistent than initially assumed (McGlashan et al. 2005). Consequently, rates of similarly defined “remission” do not distinguish personality disorders from the Axis I clinical disorders. In fact, most of the anxiety disorders have been shown to have lower rates of remission than the personality disorders (Shea

Shea: Commentary on Krueger

165

and Yen 2003). Thus, in reality, many “symptoms” shade into “traits” in terms of persistence, and stability does not serve as a useful distinction as the Axis I and II disorders are currently defined. Despite this broad lack of distinction, however, there clearly are instances of more temporally restricted episodes of symptoms that are markedly distinct from the typical baseline of individuals. For example, despite the evidence that mood disorders are more chronic and recurring than originally believed, for many the course is marked by distinct episodes. Similar points can be made with regard to more distinct episodes of mania and psychosis. The dimension of Neuroticism/Negative Affectivity may underlie the propensity toward mood disorders, but are the episodes simply a quantitative increase in intensity, or is there a qualitative difference in the pathophysiology of severe episodes of depression? Clinically, a severe episode of depression characterized by psychomotor retardation and other vegetative symptoms appears very different from a chronic dysthymic presentation. There are thus differences in the types of symptoms as well as the temporal course. The point here is not that these clinical phenomena are incompatible with genetically based cross-cutting dimensional schemes, but rather that we need to consider how the phenomena of episodes would be presented within a diagnostic scheme organized around dimensional constructs. As unsatisfactory as the Axis I–Axis II distinction is, its absence could result in a lack of attention to such temporal distinctions that are as yet not clearly understood. In a sense, the division of the axes has served to represent, albeit inadequately, a distinction between symptom—an indicator of an underlying disease or disorder— and trait—a persistent distinguishing quality not presumed to be an indicator or marker of something else. The current criteria sets in both Axis I and Axis II are a mixture of these constructs. A second distinction between Axis I and II is the interpersonal nature of the personality disorder criteria. Most of these criteria explicitly concern interactions with others. In contrast, although their presence may have consequences for interpersonal functioning, Axis I criteria tend not to be directly interpersonal. Several theoretical and treatment approaches for the personality disorders conceptualize the range of maladaptive patterns of Axis II as reflective of interpersonal processes. The interpersonal approach of Lorna Benjamin (1996) is perhaps the most explicit of these: she explains each of the DSM-IV-TR (American Psychiatric Association 2000) personality disorder criteria in terms of their interpersonal origins and current interpersonal functions. Many of the Axis I symptoms are viewed as a consequence of disturbances in attachment or interpersonal patterns. In contrast, the Siever and Davis (1991) model views many of the maladaptive interpersonal behaviors as long-term strategies aimed at coping with the core biological deficit. Again, the point is not that this distinction warrants a separate Axis, or that these conceptualizations are incompatible, but rather that the personality disorders have drawn attention to the interpersonal nature and expression of personality disturbances, a focus that may be important to maintain.

166

Dimensional Models of Personality Disorders

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Benjamin LS: Interpersonal Diagnosis and Treatment of Personality Disorders. New York, Guilford, 1996 Clark LA, Watson D, Mineka S: Temperament, personality, and the mood and anxiety disorders. J Abnorm Psychol 103:103–116, 1994 Klein MH, Wonderlich S, Shea MT: Models of the relationship between personality and depression: toward a framework for theory and research, in Personality and Depression: A Current View. Edited by Klein MH, Kupfer DJ, Shea MT. New York, Guilford, 1993, pp 1–54 McGlashan TH, Grilo CM, Sanislow CA, et al. Two-year prevalence and stability of individual criteria for schizotypal, borderline, avoidant, and obsessive-compulsive personality disorders. Am J Psychiatry 162:876–882, 2005 Shea MT, Yen S: Stability as a distinction between Axis I and Axis II disorders. J Personal Disord 17:373–386, 2003 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991

18 COMMENTARY ON KRUEGER Traits Versus Types in the Classification of Personality Pathology David Watson, Ph.D.

I

n his target paper, Krueger provides an incisive and thought-provoking comparison of the Axis I clinical syndromes and the Axis II personality disorders in DSMIV (American Psychiatric Association 1994). I basically concur with his major conclusions and see no reason to quibble about minor points of disagreement. Accordingly, I simply will amplify a few of these points and explicate certain implications that are not fully developed in his paper. Krueger’s paper can be divided into two main sections. First, he examines several putative bases for distinguishing between the Axis I and Axis II disorders. These proposed factors include temporal stability, age at onset, response to treatment, level of insight, and etiology. Krueger’s review clearly establishes that none of these factors provides a clear and unambiguous basis for distinguishing between Axis I and Axis II disorders. For instance, he shows that both types of disorders display comparable levels of temporal stability over time. Accordingly, he concludes that the current Axis I/Axis II distinction is untenable and calls for the development of a unifying structure that would recognize and model the significant commonalities among these disorders. I agree that the current classification of the Axis I and Axis II disorders is empirically unjustified and scientifically untenable. Krueger does not address an even more basic issue, however, which is whether any of these factors actually could be used as an empirical basis for classifying disorders in a substantially revised taxonomy. Put differently, could factors such as stability, age at onset, and etiology be 167

168

Dimensional Models of Personality Disorders

used as a meaningful basis for classifying mental disorders? In some cases, the answer clearly is “No.” Most notably, in light of the enormous evidence that already is available to us, it clearly would be fruitless to try to classify disorders on the basis of genetic versus environmental etiology. In other instances, however, these putative factors actually might provide a suitable basis for meaningful classification. For example, some disorders do appear to be characterized by a much greater level of insight than others. The problem—as Krueger cogently discusses in his paper— is that an insight-based classification scheme would completely cut across the current distinction between the Axis I and Axis II disorders and ultimately would lead one to sort these disorders quite differently. This analysis points to a larger problem that one faces in implementing the type of approach that Krueger used in his paper: Many of the current DSM-IV disorders are misclassified and have been placed on the wrong axis. Their current placement is a historical accident that cannot be justified on rational or empirical grounds. For instance, generalized anxiety disorder (GAD) is characterized by chronic and pervasive worrying. It tends to arise very early in life (according to DSM-IV, many individuals with GAD “report that they have felt anxious and nervous all their lives” [American Psychiatric Association 1994, p. 434]), is strongly stable over time, and can be very difficult to treat. It essentially appears to be a personality disorder that arbitrarily has been classified instead as an Axis I anxiety disorder. As another example, the generalized form of social phobia—which also is classified as an anxiety disorder—overlaps extensively with avoidant personality disorder and displays many of the features that traditionally are associated with Axis II (Widiger 1992). Krueger’s analysis, therefore, ultimately leads one into a pervasive problem within the DSM—namely, the relative lack of attention that has been given to the higher-level organization of the taxonomy (Clark et al. 1995). Why is GAD classified as an anxiety disorder rather than a personality disorder? Why is body dysmorphic disorder—which involves a preoccupation with an imagined defect in one’s appearance—classified as a somatoform disorder, given that the common feature in this diagnostic class is described as “the presence of physical symptoms that suggest a general medical condition” (American Psychiatric Association 1994, p. 445)? Although great care has gone into the specification of individual disorders within DSM-IV, far less thought has been given to the higher-level organization of these syndromes into diagnostic classes. As I argue elsewhere (Watson 2005), we now have sufficient data to eliminate this haphazardly constructed, phenomenologically based system and replace it with an empirically grounded structure that reflects the actual commonalities among disorders. Krueger’s review clearly establishes the need for such a scientifically based nosology. In the second section of his paper, Krueger presents a hierarchical dimensional structure of normal and abnormal personality that is based on extensive factor analytic evidence (Markon et al. 2005). I believe this structural scheme provides an

Watson: Commentary on Krueger

169

excellent starting point for a scientifically based nosology of personality pathology to supplant the current Axis II. Of course, a hierarchical dimensional model would represent a radical departure from the current collection of categorical personality disorders. An alternative approach is badly needed, however, and would represent a substantial improvement over the flawed current system. Personality psychologists traditionally distinguish between two basic kinds of dispositional constructs: traits and types (Carver and Scheier 2004). The adoption of a hierarchical dimensional scheme essentially would represent a shift from a typological to a trait-based approach. To appreciate fully the nature of this shift, it is useful to examine two key ways in which typological and trait constructs differ from one another. First, types are assumed to be discrete categories of membership that are distinct and discontinuous: Any given individual either fits—or does not fit—the type. In sharp contrast, trait approaches emphasize the existence of continuous quantitative variations along an underlying dimension of individual differences. As one example, Jung originally proposed that individuals could be classified as either of two basic types: introverts or extraverts. Subsequent data, however, established that most individuals actually fall somewhere between the extremes of introversion and extraversion. Consequently, Jung’s original typological view gradually was supplanted by continuous, trait-based models of the construct. Second, traits are defined as homogeneous response dispositions. Traits cause people to act, think, and feel consistently across different contexts and situations. For instance, individual differences in extraversion lead some people to be sociable, assertive, and enthusiastic and others to be more reserved, submissive, and subdued. In contrast, typological constructs can be quite heterogeneous and do not necessarily have the kind of internal coherence that characterizes traits. For example, the coronary-prone “Type A” personality traditionally was defined by three distinct—and relatively independent—characteristics: impatience, competitiveness, and anger/hostility. Similarly, the current Axis II syndromes represent extremely heterogeneous sets of characteristics that lack any semblance of internal coherence. Indeed, many of the current personality disorders actually subsume characteristics reflecting different higher-order traits. As one example, borderline personality disorder combines a diverse array of attributes reflecting neuroticism/ negative emotionality (e.g., affective instability, dysphoria), disagreeableness/antagonism (e.g., anger, fighting), and impulsivity/low conscientiousness (e.g., recklessness, spending sprees). It should be noted that these heterogeneous characteristics would be assessed and analyzed separately in a trait-based approach. Proponents of typologies assume—either implicitly or explicitly—that these heterogeneous constructs contain surplus meaning—that is, that they reflect important interactive or configural effects that are missed in a simple linear traitbased approach. In the case of Type A, for instance, it initially was assumed that this particular constellation of characteristics was “more than the sum of its parts”;

170

Dimensional Models of Personality Disorders

in other words, it was argued that this particular combination of tendencies caused significant health problems. It is noteworthy, however, that this assumption of surplus meaning rarely has been explicitly tested in the long history of typological constructs. Furthermore, when it actually has been tested, it almost never turns out to be justified. For example, the Type A literature made a tremendous advance when investigators dropped this assumption of surplus meaning and began to examine the three components separately. Among other things, this trait-based approach established that anger/hostility is the most important contributor to coronary heart disease (Thoresen and Powell 1992). The current Axis II system now has been in place for more than two decades, and we still lack any compelling evidence that would justify the continued existence of these heterogeneous typological syndromes. In response to the accumulating data, mainstream personality psychologists increasingly have abandoned heterogeneous typological schemes in favor of more theoretically parsimonious—and empirically useful—trait-based models. It is high time for the DSM to follow suit.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Carver CS, Scheier MF: Perspectives on Personality, 5th Edition. Boston, MA, Allyn & Bacon, 2004 Clark LA, Watson D, Reynolds S: Diagnosis and classification of psychopathology: challenges to the current system and future directions. Annu Rev Psychol 46:121–153, 1995 Markon KE, Krueger RK, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 Thoresen CE, Powell LH: Type A behavior pattern: new perspectives on theory, assessment, and intervention. J Consult Clin Psychol 60:595–604, 1992 Watson D: Rethinking the mood and anxiety disorders: a quantitative hierarchical model for DSM-V. J Abnorm Psychol 114:522–536, 2005 Widiger TA: Generalized social phobia versus avoidant personality disorder: a commentary on three studies. J Abnorm Psychol 101:340–343, 1992

19 DIMENSIONAL MODELS OF PERSONALITY DISORDER Coverage and Cutoffs Timothy J. Trull, Ph.D.

In order to justify and to facilitate adoption of a dimensional model of personality disorder, both the psychopathologists who study these disorders and the clinicians who assess and treat these conditions must embrace the dimensional perspective. Some feel that dimensional models are more cumbersome and less user-friendly. Many prefer categorical systems, because clinical decisions are often categorical as well (e.g., to diagnose or not; to treat or not). However, as is well known, dimensional systems can be converted to suit one’s “categorical” needs, provided the appropriate cutoffs, as well as decision algorithms, are available. Furthermore, to the extent that dimensional models provide a more reliable and valid description of patients, clinicians and researchers alike will prefer these models. Finally, as discussed later in this paper, guidelines for how to use dimensional models to characterize psychopathology are beginning to appear (Trull and Durrett 2005). These efforts will make it more likely that mental health researchers and professionals will be open to considering alternatives to the current DSM system of describing personality pathology.

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 262–282, 2005).

171

172

Dimensional Models of Personality Disorders

Two major challenges that dimensional models must address if they are to be considered viable alternatives to the present categorical system for diagnosing personality disorders are coverage and cutoffs. Regarding coverage, one of the major criticisms of the DSM-IV (American Psychiatric Association 1994) personality disorders classification system is its limited ability to characterize varieties of personality pathology that come to the attention of mental health professionals (Pincus et al. 2003). Thus, to address this concern, any alternative dimensional system must adequately cover and characterize conditions that are frequently encountered in clinical practice. Regarding cutoffs, how does one make the judgment that personality pathology is sufficiently present to warrant clinical attention? In other words, at what point do we consider personality traits to be indicators of personality pathology or disorder? These issues will be discussed and recommendations for future research in these areas will be offered. A more detailed discussion of these issues and recommendations has been published previously (see Trull 2005).

Coverage To what extent can DSM-IV and ICD-10 (World Health Organization 1992) personality disorders and symptoms be represented by dimensional models of personality disorder? One very important issue that must be addressed is that of coverage. Coverage refers to the extent to which a model or system of personality pathology adequately represents those conditions or symptoms that are frequently encountered by clinicians and studied by psychopathologists.

DSM-IV-TR Most would agree that the 10 official personality disorders (PDs) presented in the DSM-IV-TR (American Psychiatric Association 2000) do not represent all forms of personality pathology that the clinician is likely to encounter and to treat (Shedler and Westen 2004). Personality disorder not otherwise specified (PD-NOS) is typically the most frequently assigned PD diagnosis, suggesting a need for considering other varieties of personality pathology in our diagnostic nomenclature (Clark et al. 1995). A diagnosis of PD-NOS indicates that the clinician believes that a personality disorder is present but the patient does not meet the diagnostic threshold for one or more of the “official” personality disorders. This state of affairs suggests that the DSM-IV personality disorders do not provide adequate coverage of the varieties of personality pathology that clinicians encounter in their practice (Clark et al. 1995; Verheul and Widiger 2004). There may be important varieties of personality pathology that have been over-looked because of an adherence to the DSM diagnostic system.

Trull: Coverage and Cutoffs in Dimensional Models

173

ALTERNATIVE MODELS OF PERSONALITY PATHOLOGY Shedler and Westen’s SWAP Model Perhaps the most direct criticism of the coverage of the DSM-IV system of personality disorders has been leveled by Westen and colleagues. Westen and ArkowitzWesten (1998) reported that only 39.4% of patients who were being treated for personality pathology qualified for an Axis II diagnosis, suggesting that the current diagnostic system does not cover the majority of personality pathology conditions that lead patients to seek treatment. To address this and other limitations of the DSM system of personality disorder diagnosis, the Shedler-Westen Assessment Procedure-200 (SWAP-200) was developed (Westen and Shedler 1999a, 1999b). The authors argue that the SWAP-200 system of describing personality pathology is to be preferred because it provides a more rich clinical description of the DSMIV disorders (descriptions that overlap with DSM-IV, yet expand the descriptions in important and clinically meaningful ways), and the SWAP-200 system has helped identify varieties of personality pathology that are overlooked by the DSMIV system (Shedler and Westen 2004). The authors of SWAP-200 report that the item set “subsumes Axis II criteria” (Shedler and Westen 2004). To date, the only published assessments of the correspondence between the SWAP-200 items or dimensions and the DSM personality disorders have involved examining the Q-sorts of clinicians’ ratings of patients that met diagnostic criteria for each of the DSM-IV personality disorders (i.e., actual patients) as well as “prototypes” of each DSM-IV personality disorder (e.g., Shedler and Westen 2004). The results of these studies suggest that clinicians tend to favor a more expanded description of the DSM-IV personality disorders.

Livesley’s 18-Dimension Model Over the course of the past two decades, Livesley and colleagues have developed and refined self-report items to assess traits relevant to personality pathology. Subsequent data collections and psychometric analyses resulted in a model of personality disorder that contains 18 trait dimensions and is operationalized through a 290-item self-report measure, Livesley and Jackson’s (in press) Dimensional Assessment of Personality Pathology—Basic Questionnaire (DAPP-BQ). The majority of research on Livesley’s model concerns how it relates to other models of personality (e.g., Clark et al. 1996; Jang and Livesley 1999; Jang et al. 1999; Schroeder et al. 1992), as well as the factor structures and genetic heritabilities of DAPP-BQ scores in both clinical and nonclinical populations (e.g., Jang and Livesley 1999; Jang et al. 1998, 1999; Livesley et al. 1992, 1993, 1998). Four higher-order factors of personality disorder features are reliably extracted in factor analytic studies: emotional dysregulation, dissocial behavior, inhibitedness, and compulsivity. Interestingly, these resemble higher-order dimensions of “normal” personality (Widiger 1998). Two studies have assessed the relations between

174

Dimensional Models of Personality Disorders

DAPP-BQ scores and the DSM personality disorders (Bagge and Trull 2003; Pukrop et al. 2001). As expected, the four higher-order factors of the DAPP-BQ were related in predictable ways to the DSM personality disorders (Bagge and Trull 2003; Pukrop et al. 2001). Furthermore, Bagge and Trull (2003) found strong support for selected DAPP-BQ lower-order traits as unique and specific predictors of DSM personality disorder symptoms.

Clark’s SNAP Model The development of Clark’s model of personality pathology, operationalized through the Schedule for Nonadaptive and Adaptive Personality (SNAP; Clark 1993), was similar in many ways to that of Livesley, and, interestingly, the dimensions of personality pathology included in Clark’s model have been shown to be both conceptually and empirically similar to those of Livesley’s model (Clark and Livesley 2002; Clark et al. 1996). In developing her model, Clark (1993) asked clinicians to sort DSM-III-R (American Psychiatric Association 1987) personality disorder criteria as well as traitlike manifestations of some Axis I disorders into conceptually similar symptom clusters. These 22 symptom clusters were subsequently factor analyzed, yielding 12 dimensions of maladaptive personality functioning. These 12 dimensions of maladaptive personality functioning have been shown to be related to several models of higher-order personality dimensions, including that of Watson and Tellegen (Negative Affectivity, Positive Affectivity, Constraint), the five-factor model (Clark and Livesley 2002), and Livesley’s model (Clark et al. 1996). For example, Clark et al. (1996) conducted a joint factor analysis of the 18 dimensions of DSM-IV symptoms identified by Livesley et al. (1998), along with the 22 symptom scales included within the three-dimensional model of Clark (1993). They concluded that their analyses yielded a four-factor solution “which corresponded to the well-established dimensions of neuroticism, introversion, (dis)agreeableness (aggression-hostility), and (low) conscientiousness (impulsive sensation seeking)” (Clark et al. 1996, p. 300). There have been relatively few studies published that directly examined the relations between SNAP scores and the DSM-IV personality disorders. Existing data suggest that the SNAP may be particularly adept (compared to general personality measures) at tapping into maladaptive variants of personality traits that are characteristic of the DSM personality disorders (e.g., Reynolds and Clark 2001). One potential limitation of the SNAP model is that it may map on to the DSM personality disorder constructs too well (given the original method of item generation; Clark and Livesley 2002), limiting its ability to identify and define nonDSM varieties of personality pathology. However, one advantage of the SNAP model is that it also includes major “normal” personality traits (i.e., Negative Affectivity, Positive Affectivity, Constraint).

Trull: Coverage and Cutoffs in Dimensional Models

175

Cloninger’s Seven-Factor Model Cloninger’s (1987) original psychobiological model of personality was revised by Cloninger and colleagues (1993) to include dimensions of both temperament (heritable styles or biases in information processing by the perceptual memory system) and character (individual differences in self-concepts). The four temperament dimensions included in Cloninger’s model reflect individual differences in associative learning in response to novelty (Novelty Seeking), to danger or punishment (Harm Avoidance), and to reward (Reward Dependence). In addition, a fourth dimension of temperament reflects individual differences in perseverance despite frustration and fatigue (Persistence). In contrast to temperament, character traits represent conscious insight learning or reorganization of self-concepts. The three dimensions of character identified by Cloninger’s model reflect individual differences in the degree to which one tends to view oneself as an autonomous individual (Self-Directedness), as an integral part of humanity and society (Cooperativeness), and as an important component of all things/the Universe (Self-Transcendence). The Temperament and Character Inventory (TCI; Cloninger et al. 1994), a self-report questionnaire, measures these dimensions and identifies facets of each dimension. Svrakic and colleagues (1993) tested several hypotheses concerning the relations between Cloninger et al.’s (1993) seven-factor model and DSM personality disorder diagnoses and symptoms. Results were generally supportive of the hypotheses that low Self-Directedness and low Cooperativeness will indicate the presence of general personality disorder, whereas scores on the temperament dimensions will distinguish among specific PDs and PD clusters. A number of additional studies have investigated various aspects of Cloninger’s theory and of his instruments that operationalize dimensions of temperament and character (Cloninger 1998). Studies have assessed the relations between individual personality disorders, personality disorder clusters or factors, and scores on Cloninger’s dimensions of temperament or character (Bayon et al. 1996; Goldman et al. 1994; Mulder and Joyce 1997; Mulder et al. 1994; Nagoshi et al. 1992; Starcevic et al. 1995). In general, these studies have found mixed support for Cloninger’s predictions. Less support has been garnered for Cloninger’s predictions of temperament scores’ relations to individual personality disorder categories, whereas these studies found stronger support for predictions regarding temperament and personality disorder cluster score relations. However, more recent studies show mixed support for the predicted relations between TCI dimensions and PD clusters (Ball et al. 1997; Bejerot et al. 1998; de la Rie et al. 1998; Svrakic et al. 2002).

The Five-Factor Model The five major domains of the Five-Factor Model (FFM; John 1990) of personality are typically referred to as 1) neuroticism versus emotional stability, 2) extraversion versus introversion, 3) openness versus closedness to experience, 4) agree-

176

Dimensional Models of Personality Disorders

ableness versus antagonism, and 5) conscientiousness versus negligence. The most popular operationalization of the FFM is the NEO Personality Inventory—Revised (NEO-PIR; Costa and McCrae 1992). Approximately 15 years ago, researchers began to explore whether and how the FFM might also be applied to issues relating to various forms of psychopathology. In the case of personality pathology, the question of whether this model of personality was applicable to the DSM personality disorders seemed obvious to ask (e.g., Trull 1992; Wiggins and Pincus 1989). Over the past decade, there has been a host of studies that assessed the relations between FFM constructs and personality disorders (Widiger and Costa 2002). Studies have found general support for the relevance of the FFM to the full range of personality disorders (Widiger and Costa 2002). Most recently, Saulsman and Page (2004) concluded from their meta-analysis of studies that assessed the relations between the DSM personality disorders and the FFM that “the results showed that each [personality] disorder displays a five-factor model profile that is meaningful and predictable given its unique diagnostic criteria” (p. 1055). There are two particularly appealing aspects to the FFM: its inclusion of both normal and abnormal traits (like the SNAP) and, a unique feature, its item content independence from the symptoms of the DSM personality disorders. This latter feature allows for a more independent assessment of the relation between personality traits and personality disorders as well as for the possibility of characterizing abnormal personality pathology not listed in DSM-IV.

INTEGRATION OF MAJOR DIMENSIONAL MODELS The discussion above may leave the reader with the impression that one must choose one of the models mentioned above to the exclusion of others. However, this is not the case. There is general consensus for the relevance of at least four higherorder domains of personality functioning that are clearly related to personality pathology: Neuroticism/Negative Affectivity/Emotional Dysregulation; Extraversion/ Positive Emotionality; Dissocial/Antagonistic Behavior; and Constraint/Compulsivity/Conscientiousness (Trull and Durrett 2005). A focus on these four higherorder dimensions may, in fact, have the advantage of directing our attention to the area of overlap between “normal” and “abnormal” personality (Markon et al. 2005). Personality research has established psychosocial, neurobiological, and genetic correlates of these traits. The incorporation of these findings from personality research can greatly inform the study of the etiology, assessment, and treatment of personality pathology (Trull and Durrett 2005).

Cutoffs Even if a consensus develops on which dimensional personality trait model is most relevant to the quest to better characterize personality pathology, one extremely

Trull: Coverage and Cutoffs in Dimensional Models

177

important question remains: Using a dimensional model of personality and personality pathology, how can it be determined whether a “disorder” of personality is present? This is a complex issue, involving a consideration both of trait elevations as well as of impairment or dysfunction. If an individual cannot be characterized as “extreme” on one or more personality traits but does show impairment, then it does not make much sense to say that she or he has a personality disorder. Dysfunction exists but it does not appear to be due to personality style or traits. Similarly, even with elevations on personality traits, if impairment, distress, or dysfunction is not present, then we would not conclude that a person has a personality disorder. This individual’s personality style may seem atypical (i.e., statistically deviant), but these traits do not lead to impairment or dysfunction.

PERSPECTIVES ON DEFINING ELEVATED TRAITS First, for dimensional models to gain some favor, empirically defensible cutoffs to define “significant elevations” must be developed. There are multiple ways in which one might decide that a trait is elevated. One simple way, a method used by psychometricians for decades, is to focus on “statistical deviance”—for example, a score on a trait that is 2 standard deviations above or below the mean score in a target population. If, for example, a measure of the trait of neuroticism has a mean score of 50 and a standard deviation of 10, then scores of ≥70 or ≤30 would be considered deviant, significant enough to warrant further evaluation. Personality disorder might then be identified much like Minnesota Multiphasic Personality Inventory–2 (MMPI-2) code types, in which trait scales are assigned numbers or codes that then are used to describe elevations for the individual. Although this method is appealing in its simplicity and objectivity, it does have limitations. First, statistical deviance alone does not constitute mental disorder. Some examples of statistical deviance may in fact be desirable if they are not accompanied by disability or impairment (e.g., high levels of agreeableness). Further, the method itself does not address whether one deviant score is sufficient or whether a number of deviant trait scores are necessary to entertain the possibility that a personality disorder may be present. This method is agnostic as to which trait combinations define personality pathology and which combinations are most likely to be encountered in clinical settings. Thus, this may create a gap between theories of personality pathology, current conceptualizations of personality disorders, and a dust-bowl empirical method of identifying and defining deviant personality traits. Finally, cutoffs may vary according to the trait in question and according to the clinical decision being made.

PERSPECTIVES ON DYSFUNCTION AND IMPAIRMENT An independent evaluation of distress or impairment will be required because high or low levels of certain traits do not necessarily indicate maladaptivity. Most agree that personality pathology is not simply a composite of elevations on certain per-

178

Dimensional Models of Personality Disorders

sonality traits; personality traits are judged to be maladaptive only if they cause significant distress or impairment. The crucial issue then becomes what constitutes “significant” distress or impairment? DSM-IV included general diagnostic criteria for personality disorder to encourage clinicians to consider this issue when making a judgment about personality disorder diagnosis. Although most see this as a very important addition to the nomenclature (e.g., Livesley 1998, 2003), it is not clear that these general criteria have made much of an impact. This likely is because whereas criteria such as these cite domains of functioning to be considered, they do not actually quantify or operationalize what constitutes clinically significant distress or impairment (see criterion C of the general diagnostic criteria for a personality disorder [American Psychiatric Association 2000, p. 689]). What constitutes clinically significant levels of maladaptivity? Wakefield and First (2003) argue that the crucial feature of a mental disorder diagnosis is that of dysfunction. In the case of personality pathology, it is not enough to have extreme elevations on personality traits. Rather, these extreme levels must also be accompanied by dysfunction of one or more psychological processes (e.g., cognitive, motivational, behavioral, emotional, or some other psychological mechanism). In their view, the judgment of the “clinical significance” of the symptoms should not rest on the severity of the symptoms, because this is circular. Rather, one should consider the context of the symptoms (are they exaggerated or unexpected and indicative of a dysfunction of psychological processes?) to make the judgment of clinical significance. In a sense, then, some judgment concerning a pathological psychological process is required. However, how to assess such a dysfunction reliably and validly in a psychological process is not specified. From another perspective, Narrow et al. (2002) have explored how one might define clinically significant impairment using items that are frequently administered in epidemiological studies. These authors demonstrated how the prevalence rates of mental disorders assessed by major epidemiological studies were drastically reduced when attempts were made to operationalize the DSM clinical-significance criterion and apply this to the assessment of mental disorder (as opposed to simply considering whether the correct number of criteria were met). Many have been critical and skeptical of the apparently high rates of mental disorders in the general population that have been reported in the Epidemiologic Catchment Area (ECA) study and the National Comorbidity Study (NCS). By considering respondents’ answers to questions about whether one had communicated with doctors and professionals about the symptoms, whether more medication had been taken because of the symptoms, and whether the symptoms interfered with one’s life or activities, Narrow et al. (2002) found that rates for mental and addictive disorders dropped from 28.1% to 18.5% (1-year prevalence). The implications are that many of those who meet diagnostic symptom criteria for a mental disorder may not have either a functional impairment or a perceived need for treatment. Should these individuals be diagnosed with a mental disorder in the traditional sense?

Trull: Coverage and Cutoffs in Dimensional Models

179

Although certain aspects of Narrow and colleagues’ approach have been criticized (Wakefield and Spitzer 2002), their analysis does raise the issue of how to define clinically significant impairment and how to incorporate this with an evaluation of symptoms (or, in the case of the present analysis, maladaptive personality traits or features). Wakefield and Spitzer (2002) advocated an independent evaluation of disability in individuals, one that focuses on the disability that is due to mental illness symptoms. It is important that such an evaluation be independent of the severity of symptoms or traits that define personality pathology. The Global Assessment of Functioning scale (GAF; American Psychiatric Association 2000), although admirable in some ways, is problematic in this regard. Specifically, use of the GAF scores to define clinically significant impairment is circular, or at the very least redundant, because the GAF scores are dependent on the severity of the symptoms and behaviors that are used to define the diagnostic construct. What is needed is an independent evaluation of functioning, in one or more domains (e.g., personal, social, occupational), that is not defined by symptom severity.

EXAMPLES OF DIAGNOSING PERSONALITY DISORDER USING A DIMENSIONAL MODEL OF PERSONALITY With these issues in mind, we can now turn to an evaluation of proposals for how one might use dimensional models to diagnose personality disorder. Several personality disorder researchers have, in the context of one of the personality models reviewed above, presented options for making a diagnosis of personality disorder using a dimensional model of personality. It is worth noting at the outset that only a few of these proposals explicitly call for combining a description of the traits of personality or personality pathology with an evaluation of dysfunction or impairment.

Shedler and Westen Westen and Shedler (2000) proposed a prototype-matching approach to describing personality pathology. Recently outlined in Shedler and Westen (2004), their proposal suggests that the diagnostic manual present a brief narrative description (instead of a specific diagnostic criteria set) of a prototypic case of each personality disorder. The task of the clinician is then to use a 5-point scale to indicate the degree to which the current patient’s presentation “matches” the prototype for each disorder (1 = description does not apply; 2 = only minor features of prototype; 3 = significant features of prototype; 4 = strong match, patient has the disorder; and 5 = exemplifies the disorder, prototypic case). The descriptions of each personality disorder come from the SWAP-200, whose items describe both DSM-IV personality pathology and personality pathology that is not included in the diagnostic manual. To date, however, all of the studies using the SWAP-200 have been conducted on patients that were well known to the clinicians. Therefore, in addition

180

Dimensional Models of Personality Disorders

to not prescribing an independent evaluation of dysfunction or disability, a limitation of this method is the lack of evidence that it can be reliably and validly accomplished in the first intake encounter with a patient (when it is most likely that a diagnosis will be needed).

Cloninger Cloninger (2000) described how one might, using his framework, diagnose personality pathology. According to Cloninger, all personality disorders are characterized by low scores in Self-Directedness, most are also characterized by low Cooperativeness, and the more severe personality disorders (like borderline personality disorder) are characterized by low Self-Transcendence. Cloninger proposes a two stage process in which general criteria for personality disorder are evaluated (low Self-Directedness, low Cooperativeness, low Affective Stability, and low Self-Transcendence). If at least two of the general criteria are deemed present, then the patient is subsequently rated on the dimensions of Novelty Seeking, Harm Avoidance, and Reward Dependence to subtype the personality pathology. For example, borderline personality disorder is characterized by high Novelty Seeking, high Harm Avoidance, and low Reward Dependence. This system would provide eight subtypes of personality pathology, five of which are recognized as DSM-IV personality disorders (antisocial, histrionic, borderline, obsessional, schizoid, and avoidant). Again, the major limitation to this method is the lack of an independent evaluation of impairment.

Livesley Livesley (2003) suggested a two-step approach to diagnosing personality disorder. In the first step, a diagnosis of general or generic personality disorder is made. According to Livesley, there are three important indicators of a diagnosis of personality disorder that should be assessed independently from individual differences in personality traits: 1) a failure to establish and maintain stable representations of self and of others; 2) interpersonal dysfunction; and 3) a failure to develop prosocial behavior and cooperative relationships. Livesley (2003) argues that the presence of general personality disorder should be represented on Axis I. The second step involves presenting, on Axis II, an individual’s dimensional scores on lowerorder personality traits that have been shown to be descriptive of personality variation (e.g., the 18 traits of the DAPP-BQ). Thus, Livesley’s (2003) proposed system would provide the categorical diagnosis of personality disorder as well as a dimensional characterization of the patient’s personality pathology. To date, however, no research has been conducted that assesses whether this method can be used reliably and validly by clinicians, and it is not clear how best to define and measure the important constructs mentioned in step 1 (i.e., a failure to establish and maintain stable representations of self and of others; interpersonal dysfunction; and a failure to develop prosocial behavior and cooperative relationships).

Trull: Coverage and Cutoffs in Dimensional Models

181

Five-Factor Model Perhaps the most elaborate “diagnostic” system proposed is that for the FFM. Widiger, Costa, and McCrae (2002) proposed a four-step procedure for the diagnosis of personality disorders using the FFM (because the fourth step is optional and not directly relevant to this paper’s focus, only the first three steps will be discussed): 1. In step 1, a description of the individual using FFM language (five trait domains and 30 trait facets) is provided. This results in a fairly comprehensive characterization of the person’s adaptive as well as maladaptive personality traits. 2. In step 2, the evaluator identifies the personal, social, and occupational problems associated with the individual’s characteristic personality traits. A list of “problems in living” commonly associated with each of the 60 poles of the 30 facets of the FFM has been provided by Widiger et al. (2002, pp. 438–442). 3. Step 3 involves a determination of whether the problems reach a clinically significant level of impairment. The authors give an example of this step using GAF scores. To date, no research has been conducted on these steps outlined to make a personality disorder diagnosis from the perspective of the FFM. Furthermore, this proposal, like others above, does not specifically define “clinically significant” outside of suggesting that GAF scores might be used for this purpose.

Recommendations for Future Research This review of the issues of coverage and cutoffs in the context of dimensional models of personality pathology has revealed several areas that need research attention.

COVERAGE • It should not be assumed that a given dimensional model accounts for DSMIV personality disorder pathology. Empirical demonstrations to this effect are necessary. Although such coverage may be optimal, the fact that a given model does not account for one of the DSM-IV personality disorders should not, in and of itself, be seen as a fatal flaw of the model. It is important to remember that the DSM-IV conceptualization of personality disorders simply represents one alternative model, not a gold standard. • Which personality pathology traits does the DSM-IV model cover well, and where are its deficiencies? Judging from comorbidity rates and patterns, there appears to be a great deal of redundancy in the DSM-IV model. Perhaps more attention and effort should be devoted to identifying other forms of personality pathology that clinicians and psychopathologists encounter but that are not represented well in the DSM-IV system.

182

Dimensional Models of Personality Disorders

• Coverage issues will likely depend on which level of the trait hierarchy we focus on. Although a focus on the lowest level of personality traits (first-order) will result in the greatest differentiation among personality pathology styles, such a system may prove too cumbersome and perhaps fail to recognize higher-order composites of personality pathology that have genetic or biological correlates.

CUTOFFS • Should the same cutoffs be used for all traits, or will this vary depending on the trait in question or on the clinical decision that is being considered? For example, the cutoff score for a trait like self-harm might be set lower than one for a trait like affective instability because the consequences of the former are likely to be more serious than those of the latter. In addition, a different statistical cutoff might be necessary when one is deciding whether to medicate a patient with affective instability versus when one is deciding whether or not to enroll an individual in dialectical behavior therapy; a lower cutoff score might be desirable for the latter versus the former. • Another issue concerns how best to define and to measure clinically significant dysfunction. As mentioned, existing measures of dysfunction often conflate symptom severity and impairment. Instead, it is preferable for the evaluations of symptoms and traits to be independent of the evaluation for dysfunction (Lehman et al. 2002). DSM-IV Axis V is limited in this regard, and research is needed to develop and validate independent measures of dysfunction or impairment. • Finally, it will be necessary to demonstrate that clinicians can make reliable and meaningful distinctions between normal and abnormal personality. Research is needed on the various proposals that have been offered to aid in the diagnosis of personality disorder, using dimensional models.

A Proposal As a starting point to encourage research on these issues, I offer the following proposal for conducting research on the use of a dimensional model of personality to diagnose personality disorder. This proposal is informed by the convergence of evidence supporting four major higher-order dimensions of personality and personality pathology (e.g., Trull and Durrett 2005), as well as attempts to systematize how one might proceed from a dimensional assessment of personality to a diagnosis of personality disorder (e.g., Widiger et al. 2002). Table 19–1 presents a summary of the proposal as well as the research issues that need to be addressed concerning each feature. As is clear from this and other reviews, there is general agreement in the field that four higher-order personality traits can reasonably represent those features as-

Trull: Coverage and Cutoffs in Dimensional Models

183

TABLE 19–1.

A proposal for using personality dimensions to diagnose personality disorder in DSM-V • Major dimensional models of personality and personality pathology converge on four higher-order personality traits: 1. Neuroticism/Negative Affectivity/Emotional Dysregulation 2. Extraversion/Positive Affectivity 3. Dissocial/Antagonism 4. Constraint/Compulsivity/Conscientiousness • Define the primary (first-order) traits that best define these higher-order dimensions. That is, which set of primary traits best, and most efficiently represent these higher-order domains of personality? This task can be accomplished through reviews and re-analyses of existing data sets. –

These primary traits must demonstrate clinical utility and predictive validity.

–

“Best” items of these primary traits can be identified through reviews and re-analyses of existing data sets.

• Match (conceptually and empirically) current DSM-IV personality disorder symptoms to these primary traits. This can be accomplished through reviews and re-analyses of existing data sets. This matching 1) will ensure a “transition” from the current diagnostic system of DSM-IV to a dimensional classification system; 2) will likely help us understand issues related to Axis II comorbidity (e.g., several symptoms from two disorders are strongly associated with the same primary trait[s]); and 3) will help us identify personality traits that are not well represented in the current Axis II diagnostic system. –

Example: DAPP Affective Lability ➠ BPD affective instability

• Define trait elevations empirically. What constitutes a trait elevation? This can be addressed by consulting test manuals (e.g., cutoff scores based on normative data). –

Consider that cutoff scores may vary by gender and culture.

• Identify significant problems or difficulties that are associated with each primary trait. An initial list has been provided by Widiger et al. (2002, pp. 438–442). –

Example: NEO Assertiveness ➠ domineering, pushy, bossy, authoritarian

184

Dimensional Models of Personality Disorders

TABLE 19–1.

A proposal for using personality dimensions to diagnose personality disorder in DSM-V (continued) • Construct and adopt an independent assessment of impairment and dysfunction. It is important that this measure cover domains of functioning relevant to personality pathology (e.g., personal, social, interpersonal, occupational functioning) and that it be independent of the personality traits and the problems associated with those traits. In other words, the goal is to operationalize the degree to which the problems associated with the trait constitute impairment or dysfunction. • From this dimensional perspective, a diagnosis of personality disorder (PD) becomes: –

PD = elevation on one or more personality trait + evidence of dysfunction

sociated with personality and personality pathology. As a next step, it is important to identify the primary (first-order) personality traits that best define these higherorder domains, and to do so in an economical way. Data sets that include many of the models and measures discussed in this review can be used for this purpose (e.g., see Markon et al. 2005). Throughout this process, it will also be important to ensure that the primary traits identified have demonstrated predictive ability (predict relevant clinical outcomes) and clinical utility (are useful in clinical conceptualizations and treatment planning). Again, data on these issues exist and can be consulted. Next, it is important to “match” current DSM-IV PD symptoms with these primary traits. This will accomplish several goals: 1) inform and ease the transition from the current system to a dimensional system of PD diagnosis; 2) aid our understanding concerning the comorbidity of PD diagnoses and symptoms (because they are associated with the same traits); and 3) help us identify forms of personality pathology that are not well represented in the current diagnostic system. Once these initial steps have been completed, we turn to the tasks of defining trait elevations, identifying problems associated with the primary traits that were identified, and constructing an independent measure of impairment and dysfunction. Elevations on traits can be defined empirically based on normative data; test manuals routinely provide this information. Cutoffs may, however, vary as a function of gender and culture. Potential problems that are associated with each primary trait can be identified and lists compiled (e.g., see Widiger et al. 2002), making this dimensional system more reliable (versus having individual clinicians do this) and more clinician/user-friendly. The construction of an independent measure of impairment and dysfunction represents a tall hurdle. The measure, on the one hand, must cover domains of functioning relevant to personality pathology (e.g., personal, social, interpersonal, occupational) but, on the other hand, be

Trull: Coverage and Cutoffs in Dimensional Models

185

more than simply a relisting of the problems associated with the traits or more than an index of symptom severity (like the GAF scale). Using this dimensional system, a personality disorder diagnosis (that may simply be labeled as “personality disorder,” with the maladaptive primary traits as specifiers) is defined as an elevation on one or more primary personality traits and evidence of impairment or dysfunction associated with the expression of the trait(s).

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC, American Psychiatric Association, 1987 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Bagge CL, Trull TJ: DAPP-BQ: factor structure and relations to personality disorder symptoms in a nonclinical sample. J Personal Disord 17:19–32, 2003 Ball SA, Tennen H, Poling JC, et al: Personality, temperament, and character dimensions in substance abusers. J Abnorm Psychol 106:545–553, 1997 Bayon C, Hill K, Svrakic DM, et al: Dimensional assessment of personality in an outpatient sample: Relations of the systems of Millon and Cloninger. J Psychiatr Res 30:341– 352, 1996 Bejerot S, Schlette P, Ekselius L, et al: Personality disorders and relationship to personality dimensions measured by the Temperament and Character Inventory in patients with obsessive-compulsive disorder. Acta Psychiatr Scand 98:243–249, 1998 Clark LA: Manual for the Schedule for Nonadaptive and Adaptive Personality. Minneapolis, MN, University of Minnesota Press, 1993 Clark LA, Livesley WJ: Two approaches to identifying the dimensions of personality disorder: convergence on the five-factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 161–176 Clark LA, Watson D, Reynolds S: Diagnosis and classification of psychopathology: challenges to the current system and future directions. Annu Rev Psychol 46:121–153, 1995 Clark LA, Livesley WJ, Schroeder ML, et al: The structure of maladaptive personality traits: convergent validity between two systems. Psychol Assess 8:294–303, 1996 Cloninger CR: A systematic method for clinical description and classification of personality variants. Arch Gen Psychiatry 44:573–588, 1987 Cloninger CR: The genetics and psychobiology of the seven-factor model of personality, in Biology of Personality Disorders. Edited by Silk KR. Washington, DC, American Psychiatric Press, 1998, pp 63–92 Cloninger CR: A practical way to diagnosis personality disorder: a proposal. J Personal Disord 14:99–108, 2000 Cloninger CR, Svrakic DM, Przybeck TR: A psychobiological model of temperament and character. Arch Gen Psychiatry 50:975–990, 1993

186

Dimensional Models of Personality Disorders

Cloninger CR, Przybeck T, Svrakic D, et al: The Temperament and Character Inventory (TCI): A Guide to Its Development and Use. St Louis, MO, Center for Psychobiology and Personality, Washington University, 1994 Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO PI-R) and NEO FiveFactor Inventory (FFI) Manual. Odessa, FL, Psychological Assessment Resources, 1992 de la Rie SM, Duijsens IJ, Cloninger CR: Temperament, character, and personality disorders. J Personal Disord 12:362–372, 1998 Goldman RG, Skodol AE, McGrath PJ, et al: Relationship between the Tridimensional Personality Questionnaire and DSM-III-R personality traits. Am J Psychiatry 151:274– 276, 1994 Jang KL, Livesley WJ: Why do measures of normal and disordered personality correlate? A study of genetic comorbidity. J Personal Disord 13:10–17, 1999 Jang KL, Livesley WJ, Vernon PA: A twin study of genetic and environmental contributions to gender differences in traits delineating personality disorder. Eur J Personality 12:331–344, 1998 Jang KL, Livesley WJ, Vernon PA: The relationship between Eysenck’s P-E-N model of personality and traits delineating personality disorder. Personality and Individual Differences 26:121–128, 1999 John OP: The “Big Five” factor taxonomy: dimensions of personality in the natural language and in questionnaires, in Handbook of Personality Theory and Research. Edited by Pervin L. New York, Guilford, 1990, pp 66–100 Lehman AF, Alexopoulos GS, Goldman H, et al: Mental disorders and disability: time to reevaluate the relationship? In A Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp 201–218 Livesley WJ: Suggestions for a framework for an empirically based classification of personality disorder. Can J Psychiatry 43:137–147, 1998 Livesley WJ: Diagnostic dilemmas in classifying personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 153–189 Livesley WJ, Jackson DN: Dimensional Assessment of Personality Pathology—Basic Questionnaire. Port Huron, MI, Research Psychologists Press (in press) Livesley WJ, Jackson DN, Schroeder ML: Factorial structure of traits delineating personality disorders in clinical and general population samples. J Abnorm Psychol 101:432– 440, 1992 Livesley WJ, Jang KL, Jackson DN, et al: Genetic and environmental contributions to dimensions of personality disorder. Am J Psychiatry 150:1826–1831, 1993 Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Markon KE, Krueger RF, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 Mulder RT, Joyce PR: Temperament and the structure of personality disorder symptoms. Psychol Med 27:99–106, 1997

Trull: Coverage and Cutoffs in Dimensional Models

187

Mulder RT, Joyce PR, Cloninger CR: Temperament and early environment influence comorbidity and personality disorders in major depression. Compr Psychiatry 35:225– 233, 1994 Nagoshi CT, Walter D, Muntaner C, et al: Validation of the Tridimensional Personality Questionnaire in a sample of male drug users. Personality and Individual Differences 13:401–409, 1992 Narrow WE, Rae DS, Robins LN, et al: Revised prevalence estimates of mental disorders in the United States: using a clinical significance criterion to reconcile 2 surveys’ estimates. Arch Gen Psychiatry 59:115–123, 2002 Pincus HA, McQueen LE, Elinson L: Subthreshold mental disorders: nosological and research recommendations, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 129–144 Pukrop R, Gentil I, Steinbring I, et al: Factorial structure of the German version of the Dimensional Assessment of Personality Pathology—Basic Questionnaire in clinical and nonclinical samples. J Personal Disord 12:226–246, 2001 Reynolds SK, Clark LA: Predicting dimensions of personality disorder from domains and facets of the five-factor model. J Pers 69:199–222, 2001 Saulsman LM, Page AC: The five-factor model and personality disorder empirical literature: a meta-analytic review. Clin Psychol Rev 23:1055–1085, 2004 Schroeder ML, Wormworth JA, Livesley WJ: Dimensions of personality disorder and their relationships to the big five dimensions of personality. Psychol Assess 4:47–53, 1992 Shedler J, Westen D: Refining DSM-IV personality disorder diagnosis: integrating science and practice. Am J Psychiatry 161:1350–1365, 2004 Starcevic V, Uhlenhuth EH, Fallon S: The Tridimensional Personality Questionnaire as an instrument for screening personality disorders: use in patients with generalized anxiety disorders. J Personal Disord 9:247–253, 1995 Svrakic DM, Whitehead C, Przybeck TR, et al: Differential diagnosis of personality disorders by the seven factor model of temperament and character. Arch Gen Psychiatry 50:991–1000, 1993 Svrakic DM, Draganic S, Hill K, et al: Temperament, character, and personality disorders: etiologic, diagnostic, treatment issues. Acta Psychiatr Scand 106:189–195, 2002 Trull TJ: DSM-III-R personality disorders and the five-factor model of personality: an empirical comparison. J Abnorm Psychol 101:553–560, 1992 Trull TJ: Dimensional models of personality disorder: coverage and cutoffs. J Personal Disord 19:262–282, 2005 Trull TJ, Durrett C: Categorical and dimensional models of personality disorders. Annu Rev Clin Psychol 1:355–380, 2005 Verheul R, Widiger TA: A meta-analysis of the prevalence and usage of the personality disorder not otherwise specified (PDNOS) diagnosis. J Personal Disord 18:309–319, 2004 Wakefield JC, First MB: Diagnostic dilemmas in classifying personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 23–56

188

Dimensional Models of Personality Disorders

Wakefield JC, Spitzer RL: Why requiring clinical significance does not solve epidemiology’s and DSM’s validity problem, in Defining Psychopathology in the 21st Century. Edited by Helzer JE, Hudziak JJ. Washington, DC, American Psychiatric Association, 2002, pp 31–40 Westen D, Arkowitz-Westen L: Limitations of Axis II in diagnosing personality pathology in clinical practice. Am J Psychiatry 155:1767–1771, 1998 Westen D, Shedler J: Revising and assessing Axis II, Part I: developing a clinically and empirically valid assessment method. Am J Psychiatry 156:258–272, 1999a Westen D, Shedler J: Revising and assessing Axis II, Part II: toward an empirically based and clinically useful classification of personality disorders. Am J Psychiatry 156:273– 285, 1999b Westen D, Shedler J: A prototype matching approach to diagnosing personality disorders: toward DSM-IV. J Personal Disord 14:109–126, 2000 Widiger TA: Four out of five ain’t bad. Arch Gen Psychiatry 55:865–866, 1998 Widiger TA, Costa PT Jr: Five-Factor Model personality disorder research, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 59–87 Widiger TA, Costa PT Jr, McCrae RR: Proposal for Axis II: diagnosing personality disorders using the Five-Factor Model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 431–456 Wiggins JS, Pincus A: Conceptions of personality disorder and dimensions of personality. Psychol Assess 1:305–316, 1989 World Health Organization: The ICD-10 Classification of Mental and Behavioural Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland, World Health Organization, 1992

20 COMMENTARY ON TRULL Drizzling on the 5 ± 3 Factor Parade Drew Westen, Ph.D.

The Five-Factor Model (FFM) and related trait models (hereafter, the 5 ± 3) tap important dimensions of personality that can be seen across samples, instruments, and cultures. The four-factor structure suggested by Trull (Chapter 19 in this volume), like the similar structure suggested by Widiger and Simonsen (Chapter 1 in this volume), is a very sensible extrapolation of the existing data. My goal in this commentary is not to rain on the 5 ± 3 factor parade but rather to drizzle a little, focusing on the issues of coverage and cutoff points raised by Trull. I address three issues: 1) the importance of distinguishing between categorical/dimensional versus person-centered/variable-centered diagnosis; 2) the extent to which the 5 ± 3 are as adequate for clinical purposes; and 3) potential advantages of prototype diagnosis.

Categorical Versus Dimensional and PersonCentered Versus Variable-Centered Diagnosis Trull, like most contributors to this volume, generally assumes that a dimensional model will be a trait model. It is important, however, to distinguish two antinomies that are easily conflated: categorical versus dimensional, and person-centered (describing kinds of people) versus variable-centered (describing kinds of traits). We are accustomed to linking categorical diagnosis to person-centered approaches, as in the categorical typology of personality disorders in DSM-IV (American Psychiatric Association 1994); and to linking dimensional classification with variable189

190

Dimensional Models of Personality Disorders

centered approaches, as in the tradition of trait psychology. However, one can just as easily dimensionalize a typological system by viewing personality styles or configurations as prototypes or ideal types that a patient can approximate more or less. In revising DSM-V, should we abandon person-centered approaches, even if dimensionalized? This seems to me an empirical question that requires comparing the two kinds of diagnosis on reliability, validity, and clinical utility. Although trait approaches tend to convey as much or more information as typological approaches, some typological approaches have proven useful, such as Moffitt and colleagues’ (1996) early- versus late-onset delinquency, characterized by phenotypically similar behavior but tremendous differences in etiology and prognosis (see also Hicks et al. 2004). Several research groups have now also distinguished subtypes of eating-disordered patients who are indistinguishable in their eating behavior and eating disorder diagnoses but differ in adaptive functioning, developmental history, family history, and treatment response in just the ways suggestive of a valid taxonomic distinction (see Thompson-Brenner and Westen 2005; Westen and Harnden-Fischer 2001).

The Goals of Classification and the Question of Content Validity A second issue pertains to the goals of classification. How we classify, and whether our measures and models have content validity (an important issue raised by Trull), depends on our goals. From the point of view of trait psychology, the convergence of many aspects of the FFM with trait models developed more recently to cover the realm of personality pathology (Widiger and Simonsen, this volume) suggests that these models have indeed covered well the universe of traits they aspire to cover. The question is whether they have equally well covered the domains of import for diagnosing and treating personality in practice (clinical utility). Although the goals of clinical assessment and research assessment overlap, they are not identical (see Westen et al. 2002). (In what follows, I use examples from the Shedler and Westen Assessment Procedure–200 [SWAP-200] Q-sort [Shedler and Westen 2004; Westen and Shedler 1999], a personality disorder instrument designed for clinically experienced observers rather than for self-report, to illustrate what may not be adequately covered for clinical purposes by the 5 ± 3.) What is figure to clinicians often is ground to trait researchers. In wrestling with how to define dysfunction, Trull (this volume) suggests that “extreme levels [of traits] must also be accompanied by dysfunction of one or more psychological processes (e.g., cognitive, motivational, behavioral, emotional, or some other psychological mechanism),” defined independently of personality (traits). Yet such functions, processes, and mechanisms are precisely what clinicians consider most clinically relevant about personality, because these processes concern what is work-

Westen: Commentary on Trull

191

ing, what is not working, and what needs to be a target of intervention. Consider the following item from the SWAP-200: “Tends to become attached quickly or intensely; develops feelings, expectations, etc. that are not warranted by the history or context of the relationship.” This item reflects theory (attachment theory, or, more broadly, evolutionary theory), empirical observation of children (as long ago as World War II) whose disrupted attachments appeared to produce serious personality pathology, and clinical observation of a characteristic seen in many personality disorder patients. Yet it has no obvious counterpart in the 5 ± 3. Clinicians attend to internal states and their transformations, both of which are absent from the FFM and many (but not all) related models. Consider the following: “Tends to ‘catastrophize’; is prone to see problems as disastrous, unsolvable, etc.” This is a construct central to both cognitive and psychodynamic theories of personality but missing from trait models, which tend to be silent on how people process information. Clinicians also attend to what Mischel and Shoda (1995) call “if/then contingencies”—that is, behaviors or mental processes that are contingent on some eliciting event, which are difficult to capture in adjectival terms. For example, “When distressed, perception of reality can become grossly impaired (e.g., thinking may seem delusional).” Although Axis II is substantially stronger in this respect than the FFM, I suspect one of the major reasons that borderline and paranoid personality disorders show such high comorbidity is the failure to distinguish the chronic suspiciousness of the paranoid patient (well indexed by the 5 ± 3 facet of mistrust) from the contingent malevolent expectations of the borderline patient (which have no counterpart in any trait model). In addition, clinicians attend to distinctions between implicit (unconscious) and explicit (conscious) processes, a distinction central to contemporary research in cognitive neuroscience and social psychology but absent from trait psychological approaches (see Westen 1998b). For example, a large body of research now suggests that motives can be implicit or explicit, and that implicit and explicit motives have different antecedents in childhood and different consequences in adulthood (McClelland et al. 1989). Or consider grandiosity in narcissistic patients that is elicited by threats to self-esteem. It can be indispensable clinically to distinguish between the implicit feeling of being small, inferior, or “dissed” and its explicit transformation of grandiosity. More generally, clinicians attend to functional domains (see Westen 1998a), only some of which are well represented in the 5 ± 3, and only some of which are likely to be addressed by models based in self-reports (see Westen et al. 2006). Consider, for example, the following domains of functioning: • Integrity of thought processes: “Thought processes or speech tend to be circumstantial, vague, rambling, digressive, etc.”; “Tends to think in concrete terms and interpret things in overly literal ways; has limited ability to appreciate metaphor, analogy, or nuance.”

192

Dimensional Models of Personality Disorders

• Cognitive style: “Tends to perceive things in global and impressionistic ways (e.g., misses details, glosses over inconsistencies, mispronounces names).” • Affect regulation (implicit and explicit): “Tends to think in abstract and intellectualized terms, even in matters of personal import”; “Attempts to avoid feeling helpless or depressed by becoming angry instead.” • Complexity of representations of people: “Appears unable to describe important others in a way that conveys a sense of who they are as people; descriptions of others come across as two-dimensional and lacking in richness.” • Capacity for self-reflection: “Has the capacity to recognize alternative viewpoints, even in matters that stir up strong feelings.” • Beliefs and feelings toward the self: “Has a deep sense of inner badness; sees self as damaged, evil, or rotten to the core (whether consciously or unconsciously).” • Sexuality: “Has difficulty directing both tender feelings and sexual feelings toward the same person (e.g., sees others as nurturing and virtuous or sexy and exciting, but not both).” It is difficult to see how a model built largely from adjectives used by lay people could approach the complexity of clinical language that has evolved over a century of treating patients with personality pathology. Indeed, this statement is a straightforward extrapolation from the lexical hypothesis that underlies the FFM (Shedler and Westen 2004).

Prototype Diagnosis In addressing cutoffs, Trull suggests developing a measure of disability independent of personality. The notion of a personality health-sickness scale that does not confound symptoms, states, and traits (as does the GAF) is certainly sensible. The fact that one could devise such a measure that is independent of personality, however, suggests precisely what is limited about trait models from a clinical standpoint: They leave out many of the personality processes that can interfere with the capacity to love, work, and find satisfaction in life. A prototype-matching approach applied to person-centered (typological) diagnosis does not face the same difficulty because the more a person matches a pathological prototype, the more pathological they are on that dimension (psychopathy, narcissism, etc.). One of the advantages of the simple 5-point prototype matching procedure we have proposed (see Westen and Bradley 2005; Westen and Shedler 2000; Westen et al. 2006) is that clinicians rate the extent to which the patient matches each prototype (dimensional diagnosis) but for purposes of communication can consider a rating of 4 or 5 to indicate “caseness” (categorical diagnosis) and 3 to indicate “features.” Although Trull appropriately wonders if clinicians can rate such prototypes reliably (or whether prototype diagnosis would be

Westen: Commentary on Trull

193

a “throwback” to DSM-II [American Psychiatric Association 1968]), data just collected find interrater reliability averaging .70 from clinicians’ ratings of paragraphlong personality disorder prototypes based on the first 4–5 psychotherapy hours. With respect to clinical utility, Spitzer and colleagues (personal communication, December 2004) recently conducted a “nonpartisan” study of different proposals for Axis II for DSM-V. They found that experienced psychiatrists and psychologists consistently rated prototype approaches to diagnosis more clinically useful than trait models, including the FFM.

Conclusion There are four ways we might integrate a 5 ± 3 trait model into a clinically useful revision of the DSM. The first would be to use an instrument such as the SWAP200 to identify personality process correlates of traits. The second would be to identify the traits that in combination constitute personality constellations such as those represented on Axis II. The third would be to factor-analyze SWAP-200 (or its latest iteration, SWAP-II) and major 5 ± 3 item sets jointly to see whether new dimensions are necessary or useful. A final possibility would be to supplement a dimensionalized typology with a set of functional domains for clinicians to rate (e.g., affect regulation, impulse regulation) and to use well-validated trait dimensions where they adequately cover a given domain. Whether one of these approaches might be more useful will depend on its clinical utility as well as its structural elegance.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 2nd Edition. Washington, DC, American Psychiatric Association, 1968 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Hicks BM, Markon KE, Patrick CJ, et al: Identifying psychopathy subtypes on the basis of personality structure. Psychol Assess 16:276–288, 2004 McClelland DC, Koestner R, Weinberger J: How do self-attributed and implicit motives differ? Psychol Rev 96:690–702, 1989 Mischel W, Shoda Y: A cognitive-affective system theory of personality: reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychol Rev 102:246–268, 1995 Moffitt T, Caspi A, Dickson N, et al: Childhood-onset versus adolescent-onset antisocial conduct problems in males: natural history from ages 3 to 18 years. Dev Psychopathol 8:399–424, 1996 Shedler J, Westen D: Dimensions of personality pathology: an alternative to the Five Factor Model. Am J Psychiatry 161:1743–1754, 2004

194

Dimensional Models of Personality Disorders

Thompson-Brenner H, Westen D: A naturalistic study of psychotherapy for bulimia nervosa, part 1: comorbidity and therapeutic outcome. J Nerv Ment Dis 193:573–584, 2005 Westen D: Case formulation and personality diagnosis: two processes or one? in Making Diagnoses Meaningful. Edited by Barron J. Washington, DC, American Psychological Association, 1998a, pp 111–138 Westen D: The scientific legacy of Sigmund Freud: toward a psychodynamically informed psychological science. Psychol Bull 124:333–371, 1998b Westen D, Bradley RG: Prototype diagnosis of personality, in Handbook of Personology and Psychopathology. Edited by Strack S. New York, Wiley, 2005, pp 238–256 Westen D, Harnden-Fischer J: Personality profiles in eating disorders: rethinking the distinction between Axis I and Axis II. Am J Psychiatry 165:547–562, 2001 Westen D, Shedler J: Revising and assessing Axis II, part 1: developing a clinically and empirically valid assessment method. Am J Psychiatry 156:258–272, 1999 Westen D, Shedler J: A prototype matching approach to diagnosing personality disorders toward DSM-V. J Personal Disord 14:109–126, 2000 Westen D, Heim AK, Morrison K, et al: Simplifying diagnosis using a prototype-matching approach: implications for the next edition of the DSM, in Rethinking the DSM: A Psychological Perspective. Edited by Beutler LE, Malik ML. Washington, DC, American Psychological Association, 2002, pp 221–250 Westen D, Gabbard GO, Blagov P: Back to the future: personality structure as a context for psychopathology, in Personality and Psychopathology. Edited by Krueger RF, Tackett JL. New York, Guilford, 2006, pp 335–384

21 COMMENTARY ON TRULL Just Do It: Replace Axis II With a Diagnostic System Based on the Five-Factor Model of Personality Paul Costa Jr., Ph.D.

T

rull has carefully reviewed the issues of coverage and cutoffs in his consideration of dimensional models of personality pathology and his paper contains an excellent agenda for future research. I am optimistic, however, that we can advance a clinical agenda right now, because over the past 70 years, much work has gone into conceptualizing and assessing personality, and this has resulted in a consensus toward adopting the Five-Factor Model (FFM). Although the FFM is based on general personality traits, it has also been linked to clinical psychopathologic indices and outcomes (Widiger and Costa 2002). The time has come to change Axis II of the DSM to reflect a dimensional view of personality. The Axis II diagnostic system is based on the notion that some psychiatric disorders are not acute episodes of mental disorder but are instead relatively chronic problems in living (appearing in adolescence or early adulthood) that are manifestations of basic tendencies or traits of the individual. The major problem is that not one of the 10 DSM-IV (American Psychiatric Association 1994) categories of personality disorder or any of the personality disorder criteria sets map directly onto a valid and comprehensive general model of personality structure. Axis II’s many problems include the lack of empirical basis for the disorders selected, the artificiality of the diagnostic thresholds, serious comorbidity with other Axis II and Axis I disorders, poor interjudge and interinstrument reliability, and temporal instability (see McCrae et al. 2005). 195

196

Dimensional Models of Personality Disorders

My colleagues and I (Widiger et al. 2002) have suggested that we should construct a new system as an alternative rather than attempt to patch up the next DSM Axis II system. This makes good sense, especially since Axis II psychopathology is supposed to be a reflection of personality, and it would seem logical to base its classification on the structure of personality itself. There are at least five compelling reasons to adopt the FFM as the foundation of Axis II personality disorder diagnosis (McCrae et al. 2005). First, with regard to coverage, the FFM is comprehensive. Originating in studies of lay terms extracted from the dictionary, the FFM also emerged from a series of comparative studies showing that this model included virtually all the personality constructs identified by other models (Markon et al. 2005; O’Connor 2002). When fleshed-out by the more specific facets that define each factor, the FFM provides a very broad and systematic basis for describing personality traits and personality pathology. Second, its developmental course is known. Because the developmental course and stability of personality traits have been well documented in longitudinal studies, the FFM provides a basis for distinguishing Axis I from the more chronic Axis II problems. If personality were not stable, then chronic problems could not be personality-related. Because there are also small but predictable changes in trait levels with aging, we can make predictions about the changing prevalence of personality-related problems. Third, its origins are known. Abundant evidence points to the fact that much, if not most, of the variance in adult personality traits is genetic in origin. Fourth, the FFM is universal. The universality of the FFM dimensions makes it possible to coordinate the DSM with the International Classification of Diseases (ICD). According to Five-Factor Theory (McCrae and Costa 1999), the way traits are expressed, their characteristic adaptations or maladaptations, allows for the fact that personality-related problems in living may sometimes differ dramatically across cultures. Nevertheless, the traits of the FFM are themselves universal, observed in recognizable form in every culture in which they have been investigated. Fifth, the FFM can be validly assessed by multiple methods. The reliability and validity of a number of measures of the FFM are already established. For all these reasons, the FFM can serve as the basis of a reformulation of Axis II for DSM-V (for a review, see McCrae et al. 2005). In adopting the FFM as a conceptual framework for this new Axis II, it is important to remember that personality traits are neither normal nor abnormal. But they are associated with characteristic adaptations that depending upon contextual factors, including social roles and norms, subcultural and cultural, may lead to serious problems in living. Axis II should be devoted to the assessment of personality and the identification of problems in living associated with these personality factors. This FFM reformulation of personality disorders dispenses with the problem of defining trait thresholds or cutoffs. As personality traits are continuously graded and there is no optimal cutoff point, no single value or range of values can infal-

Costa: Commentary on Trull

197

libly identify disordered personality. This is why trying to define exact cutoff points, as is currently the case, is misleading and likely to be fruitless. Instead, I recommend using personality scores and profiles as cues (not cutoffs) to guide the clinician to explore manifestations of high or low traits in spheres of functioning that are likely to lead to problems in living. Empirical research can determine optimal cue levels for different traits and populations. For example, an elevation of 1 standard deviation (SD) in Neuroticism is more likely to lead to problems in living than an elevation of 1 SD in Openness to Experience. The clinician could then determine whether the problem(s) in living is/are severe enough to constitute significant impairment in intrapsychic, interpersonal, or occupational domains. The Global Assessment of Functioning (GAF) would serve as a guiding metric or quantitative representation of impairment. It is important to remember that the GAF would not be used as a tool to assess impairment, but as a guide. An assessment of impairment is first and foremost a clinical judgment and not simply the result from one test. Having thus disposed of issues regarding content, coverage, and cutoffs, we can now proceed to Just Do It: Replace the 10 personality disorder categories in Axis II with the 10 diagnoses associated with the high and low poles of each of the five factors of the FFM. The basic ideas are relatively simple and straightforward. Generally speaking, one starts with a comprehensive personality description, and then, one identifies problematic expressions of that personality description or profile in various domains of functioning. Finally, the clinician would assess the clinical significance of the personality-related problems in living to decide whether a personality disorder diagnosis is justified. To recap the four-step process of personality disorder diagnosis (Widiger et al. 2002): • Step 1—Provide a description of the person’s personality traits with respect to the 5 domains and 30 facets of the FFM. • Step 2—Identify the problems, difficulties, and impairments that are secondary to each trait. • Step 3—Determine whether the impairments are clinically significant and, if so, diagnose the FFM-related disorders. • Step 4—Determine whether the constellation of FFM traits matches sufficiently the profile for a particular personality disorder pattern. This four-step process may represent the best tool in the description of personality psychopathology to date.

198

Dimensional Models of Personality Disorders

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Markon KE, Kruger RF, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 McCrae RR, Costa PT Jr: A five-factor theory of personality, in Handbook of Personality: Theory and Research, 2nd Edition. Edited by Pervin L, John OP. New York, Guilford, 1999, pp 139–153 McCrae RR, Löckenhoff CE, Costa PT Jr: A step toward DSM-V: cataloging personalityrelated problems in living. Eur J Personality 19:269–286, 2005 O’Connor BP: A quantitative review of the comprehensiveness of the Five-Factor Model in relation to popular personality inventories. Assessment 9:188–203, 2002 Widiger TA, Costa PT Jr: Five-factor model personality disorder research, in Personality Disorders and the Five Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 59–87 Widiger TA, Costa PT Jr, McCrae RR: A proposal for Axis II: diagnosing personality disorders using the five-factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 431–456

22 COMMENTARY ON TRULL Reservations and Hopes Carl C. Bell, M.D.

Personality Disorder Not Otherwise Specified (PD-NOS) Several authors have noted that the frequency with which PD-NOS is assigned as a diagnosis suggests the need for considering other varieties of personality pathology. However, the DSM-IV Guidebook (Frances et al. 1995) noted that the NOS diagnosis is the most maligned diagnosis in the DSM. Because it is quite difficult to diagnose a patient’s specific personality disorder definitively after 1 or 3 hours of a clinical interview, the authors suggest that PD-NOS is in fact a very judicious diagnosis. Thus, I have reservations about the proclamation that the frequency of PD-NOS diagnosis is an indication that the DSM diagnostic system is “broken.”

Coverage Another suggestion for overhauling the personality disorder classification in the DSM stems from a perception that the coverage in the current system may not be adequate. Coverage was a concern in DSM-I and DSM-II. However, thanks to Millon’s (1969) work and advocacy, the coverage in DSM-III and DSM-IV is much more comprehensive. Millon’s (1969) model consisted of a two (active/passive) by four (Detached, Dependent, Independent, Ambivalent) theoretical grid that encompassed several other previously overlooked categorical diagnostic crite199

200

Dimensional Models of Personality Disorders

ria. Considering the impact of this line of work on DSM-III and DSM-IV, and considering the clinical utility of the three personality clusters, it is my opinion that coverage in the current DSM is adequate, with one notable exception, that of the potential utility of a pathological-bias personality category or some consideration of pathological bias in a dimensional model. Finally, the lack of clarity regarding the contribution of genetics and other biological contributions—for example, structural brain issues—makes it difficult to know what our categories should cover in this regard.

Diversity Experiment The vast majority of participants in the Personality Disorders Work Group are from the continents of North America and Europe. There was only one participant from Asia, and none from South America, Australia, and Africa. This lack of diversity was pointed out to be of some concern considering the finding that ecotones (a term from the field of ecology that connotes a place where landscapes meet—like field with forest, or grassland with desert) are places of increased richness and diversity. To that point, while I appreciate the conference organizers’ efforts to convene an international conference, the lack of diversity evident in the lack of participants from nonwhite continents, and the resultant emphasis on European and European-American values and philosophy gave me cause for concern as to whether or not the conference would meet the definition of an ecotone; if not, then the opportunity for creativity and innovation would be limited. There are models in which efforts are largely fragmented into resource silos in which activity is driven by different languages and different goals. There also are models that share a common mission and language, are outcome-driven and evidence-based, and have the feature of synchronicity, which maximizes resources. The outstanding work that is currently going on seems to have elements of fragmentation and conflict but also seems to have elements of synchronicity, and the more that DSM-V research agenda efforts are aligned, the greater the likelihood of generating a diagnostic classification system that is clinically useful and valuable for research.

Coverage and Cutoff Considerations With regard to the issue of coverage and cutoffs, there must be consideration of the difference between universals and cultural specifics. It may be that some personality characteristics are more welcomed by all cultures and others may only be welcomed by specific cultures. Thus, where to place cutoffs to designate pathology should vary by culture. Issues of gender, class, and acquired biology may also vary across cultures and, thus, are factors that need to be considered with regard to cutoffs.

Bell: Commentary on Trull

201

In medicine, “many things cause a fever.” My earlier work on children exposed to violence and trauma (Bell 2004b) suggests that children who undergo chronic violence and trauma are at risk for major shifts in personality functioning. Behavior is multiply determined, and personality disorders clearly appear to be the outcome of multiple factors as opposed to coming from a single origin. The discussion regarding coverage and cutoffs is extremely complex. Following a recipe or DSM criteria is simple, but building a moon rocket or developing a biopsychosocial understanding and treatment plan for patients is complicated, and raising a child, changing behavior, or doing personality research is complex. In addition to the complexity, the challenge for personality disorders work is how to construct a diagnostic system that is culturally sensitive and will meet the needs of both researchers and clinicians.

Diaspora Meeting In November 2002, the Massachusetts General Hospital Department of Psychiatry and the Social Sciences Research Council–Mellon Mays Fellowship Program sponsored a “Diaspora Meeting” that brought together prominent psychiatrists from Barbados, Belize, Brazil, Canada, England, Haiti, Jamaica, Kenya, Martinique, Nigeria, Panama, Senegal, and Uganda, as well as prominent African American psychiatrists. One outcome of this conference was a recommendation that the issue of racism or pathological bias should be examined in the next revision of DSM (Alarcón et al. 2002; Bell 2004a, 2004b; First et al. 2002). Currently, investigators at the University of California–Los Angeles (Dunbar 1995, 1997, 2003) and New York University are studying “pathological bias” as a personality disorder question. The concept of the dimensions of Openness (Flynn 2005) or Neuroticism fits quite well into this area of discourse.

Ten Commandments of Implementing Change Finally, I want to leave with you the “Ten Commandments of Implementing Change” (Jick 1999) in the hopes that we will be able to have science—instead of politics—drive the contents of the personality disorder section of DSM-V: 1. 2. 3. 4. 5. 6. 7.

Analyze the organization and its need for change. Create a shared vision and common direction. Separate from the past. Create a sense of urgency. Support a strong leader role. Line up political sponsorship. Craft an implementation plan.

202

Dimensional Models of Personality Disorders

8. Develop enabling structures. 9. Communicate, involve people, and be honest. 10. Reinforce and institutionalize change.

References Alarcón RD, Bell CC, Kirmayer LJ, et al: Beyond the funhouse mirrors: research agenda on culture and psychiatric diagnosis, in American Psychiatric Association Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp 219–281 Bell CC: Taking Issue—Racism—A Mental Illness? Psychiatr Serv 55:1343, 2004a Bell CC: The Sanity of Survival: Reflections of Community Mental Health and Wellness. Chicago, IL, Third World Press, 2004b Dunbar E: The prejudiced personality, racism and anti-Semitism: the PR scale forty years later. J Pers Assess 65:270–277, 1995 Dunbar E: The relationship of DSM diagnostic criteria and Gough’s Prejudice Scale: exploring the clinical manifestations of the prejudiced personality. Cult Divers Ment Health 3:247–257, 1997 Dunbar E: Symbolic, relational, and ideological signifiers of bias-motivated offenders: toward a strategy of assessment. Am J Orthopsychiatry 73:203–211, 2003 First MB, Bell CC, Cuthbert B, et al: Personality disorders and relational disorders: a research agenda for addressing critical gaps in DSM, in American Psychiatric Association Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp 123–199 Flynn FJ: Having an open mind: the impact of openness to experience on interracial attitudes and impression formation. J Pers Soc Psychol 88:816–826, 2005 Frances A, First MB, Pincus HA: DSM-IV Guidebook. Washington, DC, American Psychiatric Press, 1995, pp 59–60 Jick TD: Implementing Change. Harvard Teaching Note, N9–114, 1999 Millon T: Modern Psychopathology. Philadelphia, PA, WB Saunders, 1969

23 CLINICAL UTILITY OF DIMENSIONAL MODELS FOR PERSONALITY PATHOLOGY Roel Verheul, Ph.D.

Many clinicians and researchers continue to favor a categorical system of personality diagnosis. This is remarkable, given that the deficiencies and limitations of DSM-IV (American Psychiatric Association 1994) Axis II—such as excessively overlapping categories, arbitrary cutoffs, classificatory dilemmas, loss of information, inadequate coverage, lack of reliability, and limited clinical utility—outnumber its favorable characteristics—such as familiarity and communication (Ball 2001; Livesley and Jackson 1992; Widiger 1992; Widiger and Frances 2002). Overwhelming evidence indicates that personality pathology is better conceptualized dimensionally (e.g., Widiger 1992). According to Widiger and Clark (2000), the field must therefore move toward more complex, multilevel hierarchical models in which groups of symptoms are classified at varying levels of specificity. There are few, if any, authors who object to this. However, there is still considerable debate about which dimensional system is most valid and useful. Recently, First et al. (2004) proposed to use clinical utility (in addition to diagnostic validity) as the primary driving force behind future revisions. Interestingly,

This chapter is an abbreviated and adapted version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 283–302, 2005).

203

204

Dimensional Models of Personality Disorders

DSM-IV and DSM-IV-TR (American Psychiatric Association 2000) mention clinical utility as their primary rationale or highest priority, yet the focus of the DSM-IV literature reviews and field trials was almost exclusively on diagnostic validity rather than clinical utility. Furthermore, several authors have been critical of the extent to which the authors of DSM-IV based decisions on expert opinions rather than systematic evidence (e.g., Clark et al. 1995). A concern common to many critiques of the DSM is whether the decisions have reflected simply the biased perspectives of a small group of persons (Widiger and Clark 2000). In fact, modifications to the original Axis II as introduced in DSM-III (American Psychiatric Association 1980) have been rather modest, despite numerous and important criticisms regarding its clinical usefulness. This chapter is a critical review on clinical utility of dimensional models for the classification of personality disorders. The concept of clinical utility will be defined and dissected into its components, followed by a framework for evaluating and investigating the clinical utility of various dimensional classification models for personality pathology.

Definition of Clinical Utility Clinical utility is “the extent to which DSM assists clinical decision makers in fulfilling the various clinical functions of a psychiatric classification system” (First et al. 2004, p. 947). First and colleagues (2004) recognized the multifaceted character of clinical utility by distinguishing between its various components (e.g., conceptualizing diagnostic entities, communicating clinical information, choosing effective interventions, predicting future treatment needs). This review is consistent with their approach, yet considers even more facets. Distinction is made between elements of diagnostic validity that are conditional to clinical utility (i.e., adequate coverage, consistency with etiological and change models) and components of clinical utility in the narrow sense of the term (i.e., user acceptability and accuracy, professional communication, interrater reliability, subtlety of diagnosis, and clinical decision making). The elements of diagnostic validity will be discussed only briefly because they are the focus of other contributions, while the components of clinical utility will each be discussed more extensively. For each component, arguments in favor of dimensional classification will be provided and research that would optimally inform the development of the future DSM is suggested.

Elements of Diagnostic Validity Several authors have noted that a clinically useful diagnostic system should encompass the spectrum of personality pathology seen in clinical practice, and each diagnosis should be associated with unique and theoretically meaningful correlates,

Verheul: Clinical Utility of Dimensional Models

205

antecedents, and sequelae (e.g., Livesley and Jackson 1992; Shedler and Westen 2004). Simultaneously, it has been recognized that the current categorical system fails to do so. First, DSM-IV provides only partial coverage of clinically relevant personality pathology (Verheul and Widiger 2004; Westen and Arkowitz-Westen 1998). Improving coverage is actually one of the major arguments in favor of dimensional systems for classification (Widiger and Mullins-Sweatt 2005). Second, the process of diagnosing psychopathology includes linking the pathology to theoretically meaningful correlates and antecedents rather than just providing a classification (Livesley and Jackson 1992; Shedler and Westen 2004). The fact that clinicians diverge from DSM-IV is perhaps partly accounted for by the fact that DSM-IV constructs do not correlate differentially and meaningfully with etiognostic information. Most, if not all, authors have argued that the taxometric findings with respect to personality disorder support a dimensional interpretation (e.g., Widiger and Frances 2002), and many are convinced that purely dimensional systems will prove to show relative specific relationships with genetic, neurobiological and perhaps even developmental correlates (e.g., Livesley et al. 1998). Third, the classification of personality disorder should be meaningfully linked with prognostic information and thus consistent with up-to-date data and theory about its changeability. Recent theorizing and empirical evidence primarily supports a contextualist perspective proposing that personality traits are multiply determined, and that learning experiences and the individual’s social environment exert an important influence on traits. Substantial evidence suggests that personality and personality disorder change over time and keep changing throughout adulthood (Srivastava et al. 2003). Furthermore, personality disorders are remarkably responsive to psychotherapeutic treatment (e.g., Perry et al. 1999). These findings are consistent with both behavior genetic findings pertaining to the importance of environmental influences on personality development and current biological thinking about the plasticity of brain structures as a result of the modification of gene expression (Kandel 1998). Importantly, various dimensional systems might have different implications with respect to the conceptualization and assessment of change. For example, the Five-Factor Model (FFM) consists of five bipolar dimensions, with maladaptive variants at each of the two poles of these dimensions (Widiger et al. 2002). Under the assumption that average scores on the FFM reflect normality and therefore the goal of treatment, clinical improvement could be measured by the extent to which someone moves to the middle of a dimension. However, this assumption is highly problematic, because statistical deviance alone is neither a necessary nor sufficient criterion for disorder (Livesley and Jang 2000). Consequently, clinical usage of the FFM would probably require identification of clinically significant problems, difficulties, and impairments that are secondary to each trait (Widiger et al. 2002). In contrast, most other dimensional models include (unipolar) dimensions focusing on the maladaptive and pathological manifestations of personality. It is easier

206

Dimensional Models of Personality Disorders

to conceptualize change within dimensional models of personality disorder symptoms than it is within dimensional models of general personality functioning. For example, an individual with very high conscientiousness on the FFM and very high compulsivity on the Dimensional Assessment of Personality Pathology—Basic Questionnaire (DAPP-BQ; Livesley et al. 1991), who has been successfully treated for his obsessive-compulsive personality disorder, might retain a very high score or show only marginal change on conscientiousness while losing a very high compulsivity score.

Components of Clinical Utility USER ACCEPTABILITY AND ACCURACY A classification system can be theoretically sound, valid and provide perfect coverage, yet be completely worthless when it is not used at all or when it is not used correctly. Achieving user acceptability and accuracy in the application of diagnostic criteria is critically important since nonutilization and incorrect application eliminate any potential benefits that might result from DSM changes (First et al. 2004). Ease of use and familiarity often have been referred to as strengths of the current system (e.g., Widiger and Frances 2002). Indeed, the categorical approach is consistent with the neo-Kraepelinian emphasis on identifying homogeneous and distinct syndromes, yet this does not necessarily imply user acceptability and accuracy. For example, it has been observed that the current Axis II algorithm for diagnostic decisions—counting personality disorder symptoms—diverges from methods that clinicians use in real-world practice. Westen (1997) demonstrated that clinicians, irrespective of theoretical persuasion, prefer instead to make Axis II diagnoses by listening to how patients describe interpersonal interactions and observing their behavior with the interviewer. An additional explanation for why the DSM-IV guidelines with respect to Axis II diagnostic criteria are relatively ignored is that the system is too complicated for many clinicians. In fact, proper assessment of Axis II requires the administration of a semistructured interview evaluating approximately 100 diagnostic criteria (Zimmerman 1994). This is time-consuming and requires additional clinical training. But even with adequate training, Axis II includes some complexities that only few clinicians know how to handle, such as the general diagnostic criteria, overlapping categories, and the diagnosis of personality disorder not otherwise specified (PDNOS). According to the manual, for instance, any person who meets the general diagnostic criteria does in principle meet criteria for a personality disorder. The general diagnostic criteria were introduced in DSM-IV to facilitate the differentiation of clinically meaningful personality pathology from remarkable personality traits and transient, symptom-induced behavioral changes. Relatively few diagnosticians, even those using semistructured interviews, systematically ap-

Verheul: Clinical Utility of Dimensional Models

207

ply these general criteria (Livesley and Jang 2000). Furthermore, it has been observed that clinicians tend to make only one diagnosis or to consider one diagnosis as the primary one, whereas the system explicitly allows multiple diagnoses in order to obtain a comprehensive description of the patient’s pathology (Westen 1997). Finally, PDNOS as a category is often incorrectly used when actually two or more specific diagnoses would apply and make a more specific and comprehensive clinical picture (Verheul and Widiger 2004). Empirical information about the potential user acceptability of the respective dimensional models is scarce. Using both prototypic versus nonprototypic case vignettes, Sprock (2003) asked psychologists to rate the level of diagnostic confidence in providing diagnoses based on the categorical DSM-IV, a hybrid model (i.e., a dimensionalized version of DSM-IV), and four dimensional models (i.e., the FFM, Cloninger et al.’s [1993] seven-factor model, Siever and Davis’ [1991] four-factor model, and the interpersonal circumplex model). For prototypic cases, diagnostic confidence was significantly higher for the categorical model than for the hybrid and dimensional models. For nonprototypic cases, however, similar diagnostic confidence ratings were reported for the categorical, hybrid, and twodimensional models (i.e., the FFM and Siever and Davis’ model), and these were all higher than for Cloninger’s and the interpersonal circumplex model. Since most patients in real-world practice are not prototypic, these findings are not supportive for the belief that categorical systems are easier to use than dimensional ones. In sum, the complexity of the DSM-IV classification of personality disorders evidently interferes with user accuracy, and preliminary evidence indicates that dimensional systems do not necessarily introduce more complexity. However, any proposal for revision—whether dimensional or categorical—would require that its correct application can be readily learned by clinicians and is feasible within a reasonable amount of time. In a field trial, user accuracy could be examined by comparing clinician-based diagnoses with expert-based, interview-based, or consensusbased diagnoses. Furthermore, it would be important to ensure that proposed changes are in line with overall clinical sensibilities to avoid nonutilization as is often the case with DSM-IV. User acceptability could be measured by 1) surveying users’ reactions to a presentation of the proposed changes, 2) assessing acceptability in the context of actual use (e.g., in a field trial), and 3) measuring the changes’ effects on ease of use (e.g., by timing the duration of assessment procedures).

PROFESSIONAL COMMUNICATION It is of utmost importance that a classification system allow for effective professional communication (Clark et al. 1995; Sprock 2003). Ease in communication has often been referred to as one of the advantages of the categorical DSM system (Widiger and Frances 2002). Partly consistent with this idea, Sprock (2003) reported that clinicians typically find communication with purely dimensional

208

Dimensional Models of Personality Disorders

models more difficult than with categorical and hybrid models. However, this apparent disadvantage of dimensional models is mitigated by several factors. First, Sprock’s (2003) findings might partly be accounted for by unfamiliarity with the various purely dimensional models. Second, communication does not necessarily take place at the data collection level. For example, clinicians do not typically communicate about personality disorders at the diagnostic criterion level, nor is it likely that they would communicate at the facet level when using, for instance, the FFM or the DAPP-BQ. More likely, clinicians will summarize detailed facet-level information into a limited set of scores on a higher-order factor level, or will translate dimensional information into categorical diagnoses (e.g., Widiger et al. 2002). Third, practitioners often like detailed information. In fact, the DSM system is often criticized for its simplicity. For example, in clinical practice, a patient would be described as having a severe borderline personality disorder with narcissistic, antisocial, and avoidant traits rather than as only having a borderline personality disorder. Finally, many patients meet criteria for PDNOS. Clinicians often struggle with this label as it does not tell much about the patient’s pathology. In sum, despite the criticisms, it is highly questionable that dimensional models will actually yield communicative problems. It is probably more important that clinicians perceive a system as clinically useful; the motivation to get used to and communicate with that system will then follow automatically. Yet this does not relieve the developers of DSM-V of the obligation to optimize communicative aspects of a new system.

INTERRATER RELIABILITY A relatively high inference level characterizes the DSM-IV criteria for personality disorders and reliable assessment requires substantial training and exercise. It is widely accepted that semistructured interview schedules are the preferable tool for Axis II assessment. However, systematic usage of these instruments is most often encountered in research settings and only occasionally in clinical practice. Few clinicians have been trained in using semistructured interviews, partly because clinicians prefer to diverge from the DSM diagnostic criteria (Westen 1997) and also because use of these instruments is highly time-consuming and therefore often regarded as unfeasible. It is also widely accepted that clinical diagnoses made without structured instruments lack sufficient interrater reliability (e.g., Heumann 1990). Therefore, insufficient reliability is a major handicap of the current categorical system. The question is whether dimensional models will yield more reliable diagnoses, as is often assumed. The answer will strongly depend on whether clinicians will actually be using structured instruments to obtain reliable diagnostic information, and this will likely depend on whether administration of these instruments is feasible and credible in real-world practice. Most dimensional models can be

Verheul: Clinical Utility of Dimensional Models

209

measured by self-report questionnaires, yet it remains to be seen whether clinicians are willing to rely on self-report information for classificatory purposes. Sprock (2003) computed interrater reliabilities for both categorical diagnoses and dimensional ratings obtained without using structured instruments. Interrater reliability for categorical diagnoses was good in prototypic cases (median kappa = .81) but poor in nonprototypic cases (median kappa = .49). Furthermore, intraclass correlations for the dimensional ratings were moderate for the hybrid model (.64) and the FFM (.62), but at unacceptable levels for the other dimensional models (range: .44 to .51). Thus, clinician-based dimensional ratings are not necessarily much better than categorical ratings. In contrast, interrater reliability of dimensional ratings using the Structured Interview for the Five-Factor Model (SIFFM; Trull et al. 2001) is much better. Self-report ratings are not subject to rater bias, but are instead subject to other biases (e.g., trait–state contamination) that often interfere with test–retest reliability (Zimmerman 1994). In sum, DSM-IV yields unreliable diagnoses, especially when clinician-based and in non-prototypic cases. Interrater reliability of dimensional models will strongly depend on the willingness to use structured instruments, which in turn will depend on whether their administration is feasible and credible in real-world practice. As sufficient interrater agreement is a prerequisite for a valid and useful classification system, any proposal for revision—whether categorical or dimensional—should at least evidently improve the interrater reliability of clinicianbased diagnoses. Furthermore, the current problem of inter-instrument disagreement originated from the development of multiple concurrent interview schedules. It would be highly recommendable for future revisions to obtain expert consensus about standard measures or at least develop explicit guidelines for assessment. The possible development of a completely new system for classifying personality pathology would create a unique chance in this regard.

SUBTLETY OF DIAGNOSIS Several authors have demonstrated that dimensional models permit greater subtlety of diagnosis than what can be derived from a category-based system. Categorical diagnoses based on polythetic criteria sets are characterized by clinical heterogeneity. For example, there are many different ways to meet the criteria for borderline personality disorder, yet only one diagnostic label (i.e., presence of the disorder) is given to characterize all of these cases. In contrast, several authors have pointed to the utility of dimensional models in providing a very detailed and specific description of clinical cases. For example, Stone (2002) emphasized the FFM’s capacity to detail both negative (maladaptive) and positive (adaptive) aspects of personality, thereby allowing to set down in a convenient way many of the patient’s relevant strengths and weaknesses. Furthermore, Lynam (2002) argued that the FFM provides a specific representation of psychopathy, and helps to make

210

Dimensional Models of Personality Disorders

sense of the factor structure of psychopathy, the diversity of psychopathic deficits, the concept of successful psychopath, and the patterns of comorbidity surrounding psychopathy. These are just two examples underlining the level of detail and richness of case descriptions based on dimensional models such as the FFM. When comparing the respective dimensional models, there are several important points to make. First, the level of subtlety of any dimensional model is likely to be associated with the number of lower-order factors. The number of facets of the NEO Personality Inventory—Revised (NEO PI-R; Costa and McCrae 1992a) (i.e., 30 facets) and the Temperament and Character Inventory (TCI; Cloninger 2000) (i.e., 25 facets) outnumbers those of other instruments such as the DAPPBQ (i.e., 18 facets) or Schedule for Nonadaptive and Adaptive Personality (SNAP; Clark 1993a) (i.e., 22 facets). However, the difference can be attributed at least partly to facets that might not be relevant or easily understood from a clinical point of view (e.g., the facets of openness to experience [NEO PI-R] and Self-Transcendence [TCI]), whereas the number of facets of the emotional dysregulation domain is higher for both DAPP-BQ and SNAP than for either the NEO PI-R or the TCI. Second, the items of the NEO PI-R—and to a somewhat lesser extent the TCI—are formulated relatively neutrally as the instruments are typically meant to cover the entire range of individual differences, whereas the DAPP-BQ and SNAP primarily include items that refer to the pathological manifestations of personality. The importance of coverage of the pathological range is self-evident, and many clinicians have argued that information about the individual’s strengths or positive traits is also important for making clinical decisions (e.g., Stone 2002). This implies that we are in need of a system covering both the adaptive and the maladaptive range. It might therefore be worthwhile to examine whether the level of richness, detail, and relevance of clinical descriptions improve when integrating or collapsing different models. In sum, dimension-based classification systems are generally superior to category-based systems in terms of subtlety of diagnosis. Important issues are the coverage of the adaptive versus maladaptive range, the inclusion of positive in addition to negative traits, and the inclusion of lower-order in addition to higher-order factors. The optimum level of detail provided by any classification system will eventually also be related to issues such as feasibility, preferred level for communication, added value for clinical decision making, and so forth. Research can help to determine the optimal level of comprehensiveness and detail. First, research could make use of item response theory methods to compare different models in terms of their coverage of adaptive and maladaptive aspects. Second, at the lower-order level, clinician surveys (e.g., identifying clinically relevant dimensions, and comparing different models with respect to the level of richness, detail and relevance of clinical descriptions of case vignettes) can perhaps provide valuable information over and above the already large database of studies examining associations between dimensional models and DSM-IV. Finally, exploratory and confirmatory factor analyses

Verheul: Clinical Utility of Dimensional Models

211

could be used to evaluate whether the number of facets at the lower-order level is higher for an integrated model than any one model separately.

CLINICAL DECISION MAKING The extent to which a taxonomy is able to direct clinical decision making is perhaps the most important determinant of clinical utility (First et al. 2004). An important argument often mentioned in favor of a categorical model for classification is that clinical decisions are categorical (Widiger and Frances 2002). Indeed, apologists for categorical diagnoses have claimed that the DSM system has clinical utility in so far as it is valuable in formulating cases and planning treatment, yet there is actually little evidence for these assertions. Diagnostic categories such as borderline personality disorder have served as a stepping-stone for clinical intervention strategies, but cannot predict clinical interventions or outcomes except in the most general terms (Livesley 2001; Sanderson and Clarkin 2002; Spitzer 1998). Clark (1993b, p. 101) put it very clearly: Whereas it is indisputable that certain clinical decisions are categorical (e.g., whether or not to hospitalize, whether or not to prescribe medication), such dichotomous decisions in the treatment of personality disorders may be the exceptions rather than the rule. Rather, treatment decisions typically involve consideration of the appropriate degree of various therapist behaviors: where and how firmly to set limits, how much to permit dependence versus encourage independence, how and when to confront versus support in the face of anxieties, and so forth. For such decisions, diagnostic labels are of far less use than richer personality descriptions.

Clinical decision making refers to determining 1) the necessity and the benefit of treatment, 2) the type of treatment that is most likely to be effective and efficient, and 3) the type of interventions that are likely to be most helpful within the treatment model. First, decision making about the necessity of treatment requires simple, binary information about whether the observed problems are sufficiently severe to warrant treatment. Furthermore, although clinicians are not inclined to recommend no treatment and rarely do so for patients applying for help, several considerations might lead to prescription for no treatment and facilitate realistic expectations for degree of clinical change (Harkness and Lilienfeld 1997). It has been noted that the FFM may be of assistance in isolating those patients for whom treatment is contraindicated or who are unlikely to benefit from treatment (Miller 1991; Sanderson and Clarkin 2002). Second, when treatment is indicated, the choice and planning of treatment modality involves four broad or macro-treatment decisions regarding 1) setting (e.g., inpatient, day hospital, or outpatient), 2) format (e.g., individual, group, or family therapy), 3) major strategies and techniques (e.g., choice of theoretical model), 4) duration (e.g., crisis intervention, short-term, or long-term therapy)

212

Dimensional Models of Personality Disorders

and frequency of appointments (Livesley 2003; Sanderson and Clarkin 2002). Empirical data supportive of differential treatment selection based on personality factors are scarce, yet there is a growing number of predictor studies and clinical reports suggesting that personality dimensions can be helpful. For example, it has been suggested that partial hospitalization programs are particularly effective in treating severely disturbed patients (Bateman and Fonagy 2004; Livesley 2003). Furthermore, the optimal duration of treatment might depend on the severity, breadth and complexity of problems, poor premorbid functioning and adjustment, and poor “patient enabling factors” for treatment such as motivation for change, psychological mindedness, and interpersonal functioning (Lambert and Anderson 1996; Sanderson and Clarkin 2002). It has been noted, from the perspective of the FFM, that the ideal patient for brief treatment would show isolated but significant elevations on neuroticism, high openness to activities, ideas, and emotions, high warmth, and high agreeableness (Sanderson and Clarkin 2002). Next, the choice of theoretical model might primarily depend on the level of extraversion and openness to experience. Extraverts are sociable, talkative, and demonstrative, and might find therapies that require interpersonal interaction congenial (Costa and McCrae 1992b). Miller (1991) has noted that both clientcentered therapy and psychoanalysis require considerable spontaneous speech from the patient and are difficult for introverts. Introverted patients may instead prefer and benefit more from behavior or Gestalt therapy, in which the therapist has a more active role (Costa and McCrae 1992b). Individuals who are closed to experience are conventional in their beliefs and will probably prefer directive psychotherapies that offer sensible advice, behavioral techniques that teach concrete skills, or client-centered therapies that provide emotional support. In contrast, patients high in openness to experience are much more willing to consider novel ideas; hence Gestalt or psychoanalytic approaches may appeal to them (Costa and McCrae 1992b). Most of these ideas are still untested hypotheses and cutoffs for decision making are lacking. Nevertheless, dimensional models do seem to provide a comprehensive framework for research on the relation between individual differences and treatment outcomes. The FFM has been the one that has received most attention so far. In addition to the four macrotreatment decisions, several authors have pointed to the possible value of dimensional models in facilitating therapeutic tailoring or microtreatment decisions, including goal setting, matching patient to therapist characteristics, handling of transference and counter-transference phenomena, and degree of therapy directiveness (Livesley 2003; Miller 1991; Sanderson and Clarkin 2002). Whereas the information at the higher-order level of hierarchical dimensional models might help to direct the general therapeutic approach and dosage of treatment, information at the lower-order level is considered useful in identifying important themes and targets for change. For example, dialectical behavior therapy specifically targets self-harm, suicidal behavior, and other potentially self-damaging

Verheul: Clinical Utility of Dimensional Models

213

behaviors, rather than borderline personality disorder per se (Verheul et al. 2003). Similarly, mentalization-based treatment specifically targets mentalizing capacities, emotion regulation, and effortful control (Bateman and Fonagy 2004). Pharmacological interventions also tend to target behavioral clusters such as affective lability, cognitive disorganization, or impulsive behavior rather than diagnoses, and dosage typically depends on symptom levels (Soloff 1998). In sum, clinician reports and circumstantial evidence support the notion that a dimensional classification would be more consistent with treatment planning and goal setting than category-based systems. It might be tempting to demand empirical evidence before adopting a dimensional system for classification. For example, First et al. (2004) proposed to evaluate the extent to which diagnostic changes increase the degree of adherence to practice guidelines. They specifically suggest that “the ideal design would document an increased incidence of the desired clinical decision in a randomized controlled trial in which patients are randomly assigned to one or two groups: patients diagnosed according to the existing diagnostic classification and patients diagnosed according to the proposed changes” (p. 951). Although I would not principally object to these kinds of studies, there are several important points to make. First, such studies might be a bridge too far for the field of personality disorders where practice guidelines are scarce, not very specific, and perhaps not undisputed. Second, demanding such studies for revisions would be unprecedented and increase the threshold for any further revisions to an almost inaccessible level. Third, most revisions will likely result in subtle and gradual changes of clinical decision making that would require very large sample sizes to pick up in randomized studies. The most viable alternative to large-scale treatment studies is to investigate clinical decision making in clinicians using case vignettes that are described with various models, and to compare across models the clinicians’ performance (including case-based reasoning and decision making processes) with a panel of experts. In addition, a systematic literature review of studies on the prognostic validity of personality factors would be very informative.

Conclusion In summary, clinical utility is defined as the extent to which DSM assists clinical decision makers in fulfilling the various clinical functions of a psychiatric classification system. Distinction is made between diagnostic validity elements conditional to utility (i.e., coverage, and consistency with etiological and change models) and components of clinical utility in the narrow sense of the term (i.e., acceptability, accuracy, communication, reliability, subtlety, and decision making). Clinical utility is often considered the driving force behind the respective revisions of the DSM, yet it has been difficult to obtain reasonable levels of clinical utility. This

214

Dimensional Models of Personality Disorders

chapter presents evidence that a dimensional diagnostic system will substantially improve clinical utility, especially with respect to coverage, reliability, subtlety, and clinical decision making. Finally, some evidence suggests that the purely dimensional models, particularly the DAPP-BQ, SNAP, and FFM, outperform the socalled hybrid models (i.e., dimensional profiling of categories) with respect to coverage, consistency with etiology, subtlety, and clinical decision making. The hybrid models based on DSM-IV are of course relatively easy to disseminate and implement in clinical practice. Even the DSM-IV chapter on personality disorders would require only limited revision. This might seem attractive, but it also reflects that it is highly questionable whether the introduction of hybrid models would change anything. In addition, it remains to be seen whether dimensional profiling of existing categories would retain information more optimally, be more consistent with real-world practice, and facilitate clinical decision making. Finally, assessment according to this system would be as time-consuming and complex as the current categorical system. The latter concern might not apply to Shedler and Westen’s (2004) prototypical matching approach. An important conclusion from this review is that most components of clinical utility are lacking strong evidence across the broad range of models. In general, this implies that there is still much research work to do in case future revisions are indeed aimed at substantially increasing clinical utility. Several research options that might inform future revisions have been suggested, including 1) surveying user acceptability, accuracy of application, and adequate professional communication, 2) testing improvements in the interrater reliability of clinician-based diagnoses, 3) item response theory analyses into the coverage of the adaptive and maladaptive range of traits, 4) clinician surveys to identify clinically relevant dimensions and compare the level of richness and clinical relevance of different models, 5) exploratory and confirmatory factor analysis to compare the lower-order structure across models, and 6) comparing clinical decision making processes according to various models. Because each model has both advantages and limitations, it would be interesting to include an integrated or collapsed model in comparative studies. An important reason why many personality disorder experts reluctantly continue to favor the categorical system relates to a perception that there is still very poor consensus among the alternative dimensional models (Ball 2001). In reality, there is a remarkable level of consensus among personality (trait) researchers regarding the structure of personality. At the top of the trait hierarchy, there is compelling evidence for a fairly limited number (5 ± 2) of higher-order dimensions, although there is not perfect unanimity about the exact number, name, or specific features for each dimension (Ball 2001; Widiger and Mullins-Sweatt 2005). Furthermore, these broad dimensions are each defined by a number of lower-order traits, which are expressed in varying levels of adaptiveness or maladaptiveness. These notions are shared by many, if not all, personality researchers. The greatest challenge is to determine and define the facets at the lower-order level. To increase

Verheul: Clinical Utility of Dimensional Models

215

consensus, it might be important to depart not from particular instruments such as the NEO-PI-R or DAPP-BQ, but instead from the general theoretical framework as described above. The integration of various models, combined with a multifaceted clinical utility perspective as described in this paper, might facilitate the development of a dimensional diagnostic system that can count on the “popular vote” among clinicians. Whatever dimensional model for the classification of personality pathology is chosen in the future, it cannot entirely replace a categorical system (Frances 1993). For legal, medical, and administrative purposes, it may be necessary to obtain a formal DSM-IV diagnosis (Widiger et al. 2002). “Many persons are concerned that the formal, authoritative recognition and funding of personality disorders would be undermined by a dimensional model. If personality disorders are simply maladaptive variants of normal traits, and all persons have some degree of a personality disorder, then a diagnosis of personality disorder might lose credibility as a public health issue” (Widiger 1993, p. 139). Last but not least, some clinical decision making requires binary information or at least cutoff points. The introduction of any dimensional system should, therefore, be accompanied by a categorical one, and the two are preferably maximally compatible while minimally redundant. Several proposals have been put forth (e.g., Trull, Chapter 19 in this volume), but a proper comparison with respect to applicability, feasibility, and clinical and policy implications is still missing. This issue should certainly be included in the field trials.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition. Washington, DC, American Psychiatric Association, 1980 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Ball SA: Reconceptualizing personality disorder categories using personality trait dimensions: introduction to special section. J Pers 69:147–153, 2001 Bateman A, Fonagy P: Psychotherapy of Borderline Personality Disorder: Mentalization Based Treatment. New York, Oxford University Press, 2004 Clark LA: Manual for the Schedule of Nonadaptive and Adaptive Personality. Minneapolis, MN, University of Minnesota Press, 1993a Clark LA: Personality disorder diagnosis: limitations of the five-factor model. Psychol Inq 4:100–104, 1993b Clark LA, Watson D, Reynolds S: Diagnosis and classification of psychopathology: challenges to the current system and future directions. Ann Rev Psychol 46:121–153, 1995

216

Dimensional Models of Personality Disorders

Cloninger CR: A practical way to diagnose personality disorders. J Personal Disord 14:99– 108, 2000 Cloninger CR, Svrakic DM, Przybeck TR: A psychobiological model of temperament and character. Arch Gen Psychiatry 50:573–588, 1993 Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992a Costa PT Jr, McCrae RR: Normal personality assessment in clinical practice: the NEO Personality Inventory. Psychol Assess 4:5–13, 1992b First MB, Pincus HA, Levine JB, et al: Clinical utility as a criterion for revising psychiatric diagnoses. Am J Psychiatry 161:946–954, 2004 Frances AJ: Dimensional diagnosis of personality—not whether, but when and which. Psychological Inquiry 4:110–111, 1993 Harkness AR, Lilienfeld SO: Individual differences science for treatment planning: personality traits. Psychol Assess 9:349–360, 1997 Heumann KA: Reliability of categorical and dimensional judgments of personality disorders. Am J Psychiatry 147:498–500, 1990 Kandel ER: A new intellectual framework for psychiatry. Am J Psychiatry 155:457–469, 1998 Lambert MJ, Anderson EM: Assessment for the time-limited psychotherapies, in American Psychiatric Press Review of Psychiatry, Vol 15. Edited by Dickstein LJ, Riba MB, Oldham JM. Washington, DC, American Psychiatric Press, 1996, pp 23–42 Livesley WJ: Commentary on reconceptualizing personality disorder categories using trait dimensions. J Pers 69:277–286, 2001 Livesley WJ: Practical management of personality disorder. New York, Guilford, 2003 Livesley WJ, Jackson DN: Guidelines for developing, evaluating, and revising the classification of personality disorders. J Nerv Ment Dis 180:609–618, 1992 Livesley WJ, Jang KL: Toward an empirically based classification of personality disorder. J Personal Disord 14:137–151, 2000 Livesley WJ, Jackson DN, Schroeder ML: Dimensions of personality pathology. Can J Psychiatry 36:557–562, 1991 Livesley WJ, Jang KL, Vernon PA: The phenotypic and genetic architecture of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Lynam DR: Psychopathy from the perspective of the five-factor model of personality, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 325–348 Miller TR: The psychotherapeutic utility of the five-factor model of personality: a clinician’s experience. J Pers Assess 57:415–433, 1991 Perry JC, Banon E, Ianni F: Effectiveness of psychotherapy for personality disorders. Am J Psychiatry 156:1312–1321, 1999 Sanderson C, Clarkin JF: Further use of the NEO-PI-R personality dimensions in differential treatment planning, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 351–376

Verheul: Clinical Utility of Dimensional Models

217

Shedler J, Westen D: Refining personality disorder diagnosis: integrating science and practice. Am J Psychiatry 161:1350–1365, 2004 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991 Soloff P: Algorithms for pharmacological treatment of personality dimensions: symptomspecific treatments for cognitive-perceptual, affective, and impulsive-behavioral dysregulation. Bull Menninger Clin 62:195–214, 1998 Spitzer RL: Diagnosis and need for treatment are not the same. Arch Gen Psychiatry 55:120, 1998 Sprock J: Dimensional versus categorical classification of prototypic and nonprototypic cases of personality disorder. J Clin Psychol 59:991–1014, 2003 Srivastava S, John OP, Gosling SD, et al: Development of personality in early and middle adulthood: set like plaster or persistent change? J Pers Soc Psychol 84:1041–1053, 2003 Stone MH: Treatment of personality disorders from the perspective of the five-factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 405–430 Trull TJ, Widiger TA, Burr R: A structured interview for the assessment of the five-factor model of personality: facet-level relations to the Axis II personality disorders. J Pers 69:176–198, 2001 Verheul R: Clinical utility of dimensional models for personality pathology. J Personal Disord 19:283–302, 2005 Verheul R, Widiger TA: A meta-analysis of the prevalence and usage of personality disorder not otherwise specified (PDNOS). J Personal Disord 18:309–319, 2004 Verheul R, Bosch LMC van den, Koeter MWJ, et al: Dialectical behavior therapy for women with borderline personality disorder: 12-month, randomised clinical trial in The Netherlands. Br J Psychiatry 182:135–140, 2003 Westen D: Divergences between clinical and research methods for assessing personality disorders: implications for research and the evolution of Axis II. Am J Psychiatry 154:895–903, 1997 Westen D, Arkowitz-Westen L: Limitations of Axis II in diagnosing personality pathology in clinical practice. Am J Psychiatry 155:1767–1771, 1998 Westen D, Shedler J: A prototype matching approach to diagnosing personality disorders: toward DSM-V. J Pers Disord 14:109–126, 2000 Widiger TA: Categorical versus dimensional classification: implications from and for research. J Pers Disord 6:287–300, 1992 Widiger TA: Reply to commentators: from B to Z. Psychological Inquiry 4:135–141, 1993 Widiger TA, Clark LA: Toward DSM-V and the classification of psychopathology. Psychol Bull 126:946–963, 2000 Widiger TA, Frances A: Toward a dimensional model for the personality disorders, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 23–44

218

Dimensional Models of Personality Disorders

Widiger TA, Costa PT Jr, McCrae RR: A proposal for Axis II: diagnosing personality disorders using the five-factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 431–456 Widiger TA, Mullins-Sweatt SN: Categorical and dimensional models of personality disorder, in The American Psychiatric Publishing Textbook of Personality Disorders. Edited by Oldham JM, Skodol AE, Bender DS. Washington, DC, American Psychiatric Publishing, 2005, pp 35–53 Zimmerman M: Diagnosing personality disorder: a review of issues and research methods. Arch Gen Psychiatry 51:225–245, 1994

24 COMMENTARY ON VERHEUL Focusing on the Clinician’s Need for a Better Model Erik Simonsen, M.D.

Roel Verheul has written a thoughtful and comprehensive review of the clinical utility of the dimensional models for personality disorders. I share his expectations for an improvement of the classification system, but I also have my questions and doubts of where to go. What would my colleagues in clinical psychiatry say if we were to adopt a dimensional system in the near future? What comes to my mind?

Clinical Utility and the Role of Theory In general, clinicians permit theoretical considerations (whether cognitive, dialectical, psychodynamic, interpersonal, familial, or biogenic) to guide their therapeutic approaches to personality disorders. Probably no other diagnosis within psychiatry lays claim to more theories on development, pathogenesis, and treatment than personality disorders. This both manifests the richness of thought by theorists in the field and signifies a weakness of a clear and empirical science. Some clinicians hesitate becoming involved in the personality disorder field because of its relative newness and the confusing, conflicting mixture of its theories, even though the newness and complexity are challenges that stimulate curious, creative thinkers. Some clinicians handle complex treatment decisions simply by placing a label on their patient, while others use a personality disorder diagnosis as a means to justify refraining from treating their patients. Obviously, with the growing evidence that personality disorder patients are treatable, there is a need for guidelines. Models 219

220

Dimensional Models of Personality Disorders

are one way of conceptualizing the key issues. Models rest on concepts and hypothetical constructs but are less detailed than theories. Concepts are powerful factors in clinical practice for they guide how we understand the patient’s pathology. Moreover, concepts will, to a certain extent, help explain matters to the patient, and thereby become part of the patient’s learning and healing process. For the patient, concepts may play a crucial role in the process of understanding his or her problems. For example, consider a hyperactive patient, one who engages in many different activities at the same time. The clinician can introduce the term “impulsivity” to the patient. There are several different ways of conceptualizing the next step. The patient may be told that his actions are caused by an inborn temperament and should be addressed with a biological intervention. Or impulsivity may be explained as signifying the patient’s lack of control over his thoughts, hence suggesting the potential usefulness of a cognitive model. Yet again, impulsivity may be described as a way of denying more hidden and important aspects of the patient’s life by introducing concepts such as anxiety or defense mechanism. In each case, the idea of impulsivity may be seen as a concept or a dimension in a model, but the clinician’s selection and interpretation is highly dependent on theoretical preference and can lead to specific treatment approaches. Models are simplifications of a much more complicated scientific question: how do I conceptualize and organize robust empirical findings by placing them in a theoretical context? Clinicians need a theory to organize all variables and available information, to analyze these in the context of dynamics within the patient and his environment, and to guide treatment (Magnavita 2004). With an appropriate model, the clinician would ask: What are the most powerful concepts helpful in assessing and treating personality disorders, and how do I put them together in a theory? If we provide an alternative way of conceptualizing the patient’s personality pathology, we should, at best, also provide a clinical theory within the framework of natural science.

Clinical Utility and the Doctor–Patient Relationship Clinicians often evaluate personality pathology, by simply listening to patients describe their problems and observing their interpersonal interactions including behavior toward the interviewer rather than from a particular theoretical influence or by employing Axis II categories. It is likely that the same clinical approaches will be used when employing a formal dimensional model; indeed, these will be necessary for the clinician to obtain a meaningful understanding of the patient’s constellation of symptoms. Thus, the average clinician, even one who is not familiar with research on the different dimensional models, will combine all the varied ingredients comprising the dimensions to build a clinical impression. Regardless of

Simonsen: Commentary on Verheul

221

whether dimensions, categories, or prototypes are used, clinicians will strive for a common, unified theory, in one “language” to have their assessment, classification, and therapeutic techniques to hang together in a logical way. The clinician applies this complex gestalt.

Clinical Utility and Clinical Decision Making The single most important question for clinicians is the decision as to whether the patient has a personality disorder or not. This requires a clear definition of a specific group of disorders, and a categorical delineation of these disorders from others. Arriving at a decision might then require a clinician to differentiate a particular prototypical group of personality disorders or to specify several unipolar or bipolar dimensions. This step in the process should be useful, too, in deciding whether to treat the admitted patient, assessing the risk of treating or not treating, and determining the preferred approach to treating a patient (e.g., whether the approach should be supportive or confrontational) and the patient’s amenability to treatment (e.g., motivation for change). Numerous concepts come into play here. In making these determinations, a hierarchical understanding of higher- and lower-order dimensions or traits may be inadequate; the clinician may need to more deeply integrate surface phenomena—that is, ascertain how dimensions or traits are linked and how they are experienced by the patient. The patient’s sense of self, or selfawareness, will attach significance to the network of dimensions or traits. In other words, to grasp the patient as a person, it is necessary to introduce a higher level of abstraction that goes beyond traits and dimensions based on the direct observational data. This level of analysis and synthesis makes theory relevant, as it offers the means of connecting what is otherwise piecemeal and scattered. The heterogeneity of the personality disorder diagnosis often is a problem for decision making, whether the approach is categorical or dimensional. Although there are 256 ways of meeting criteria for borderline personality disorder, the patient receives only one diagnosis. Similarly, there are many ways of earning a high score on Neuroticism in the Five-Factor Model, as will be found by employing its facets. In this regard, dimensional facets are similar to the DSM diagnostic criteria. There is little difference between the facets that underlie dimensions and the diagnostic criteria that underlie categories. Clinicians must also decide how to handle a given patient’s particular problem. Patients with personality pathology present in diverse settings, including the emergency ward (suicidal behavior), jail (aggressive behavior), marital therapy (interpersonal problems), social welfare system (lack of involvement), or student counseling (narcissism, perfectionism). It is very unlikely that a single set of concepts or models will be useful in all the conditions of personality pathology that lead to these diverse settings. It is more likely that basic concepts will need to be tailored

222

Dimensional Models of Personality Disorders

to a particular setting to understand a patient’s problem and to make an appropriate clinical decision. In some instances it will be more useful to quantify the patient’s problems (e.g., avoidance, aggression, self-harm), while in others the less measurable, qualitative aspects of personality (e.g., unstable self, projection) will be more prominent and inform decisions about clinical approaches.

Clinical Utility and How to Quantify the Polarity in Dimensions Clinicians may find it easier to employ unipolar dimensions, which are closer to the categorical system. From a statistical point of view, this would be more rigorous and would simplify the task of quantifying the illness and making treatment decisions. Moreover, unipolar dimensions probably would facilitate dynamic interpretations of pathology. This approach is exemplified in analyses of the Millon Clinical Multiaxial Inventory (MCMI) and Minnesota Multiphasic Personality Inventory (MMPI) profiles (self-report inventories) or the “dimensions” within the DSM-IV (American Psychiatric Association 1994) personality disorder categories (number of criteria met). It is clinically useful to make dynamic interpretations using both the higher and lower scores on all the scales. The same experiences are drawn from the use of other “unipolar” instruments, such as the Dimensional Assessment of Personality Psychopathology (DAPP) and the Schedule for Nonadaptive and Adaptive Personality (SNAP). On the other hand, many clinicians would agree that the bipolarity would offer measures useful in gauging a patient’s strengths and weaknesses. But introducing concepts such as ego strength and normality into a model of psychopathology might be problematic for a classification system of diseases.

Clinical Utility and Assessment Tools A number of assessment tools have been constructed from the categorical classification of personality disorders based on a polythetic paradigm. Research on formal instruments has shown that self-report inventories are only modestly reliable in assessing personality and personality disorders. Assessment using dimensional models is nearly always done by inventories. The basis for each of the dimensional models has been established mainly through the use of their correspondent questionnaires, which might raise questions regarding validity. Furthermore, a clinician’s application of personality disorder diagnoses is most often a decision based on a clinical interview; self-reports are only used occasionally. Dimensions intended for clinical use should be defined at least as precisely as are all the diagnostic criteria of DSM categories, which is not a very high standard. Is it easier to define

Simonsen: Commentary on Verheul

223

conscientiousness or constraint dimensionally than to define compulsivity categorically? Clinicians would think that the reliability of personality diagnoses will not be easier to achieve, if or when they are to move from diagnostic criteria of the disorders to a number of alternative definitions of clinical dimensions.

Clinical Utility and Measuring Change and Outcome It would be a significant step forward if a dimensional model could capture personality features more precisely than categories (and their respective diagnostic criteria sets) so as to appraise precise changes related to specific interventions. However, if we are to move toward dimensional models based on mathematics and statistical procedures, there is a danger of creating a closed system, which would exclude clinical concepts embedded in other theoretical models. Also, there are numerous alternative ways of employing dimensions to appraise psychotherapeutically driven changes in personality. A variety of instruments designed for outcome research in the psychodynamic school—for example, the Karolinska Psychodynamic Profile (Weinryb et al. 1997), the Dutch Developmental Profile (Abraham and van Dam 2004), and Kernberg’s Personality Organization (Kernberg 1967)— are dimensional (theoretical and clinical) instruments that can be used in assessing treatment progress.

Dimensions, Categories, Domains: Mutually Exclusive? One alternative to a more radical conversion is, of course, a further development of the current diagnostic prototypes and by severity of personality pathology (Oldham and Skodol 2000; Tyrer and Johnson 1996, Widiger and Sanderson 1995). These suggestions point to the fact that qualitative distinctions in the categorical approach, and quantitative distinctions in the dimensional, need not be framed as being opposed or mutually exclusive. Whether the focus is on qualitative or quantitative distinction, there also needs to be a common ground of how to describe personality pathology. In the ICD-10 (World Health Organization 1992), the general description of personality pathology is tied to concepts of cognition, interpersonal relationship, emotionality, gratification, and control of impulses. These are important clinical domains that most practicing psychologists and psychiatrists routinely employ. These clinical domains may have greater likelihood of continued acceptance as compared to the theoretical or research-based dimensions under consideration. Any dimensional approach has limitations. Because some of the models are based on factor analytic techniques, which are designed to extract independent

224

Dimensional Models of Personality Disorders

factors, the dimensional approach tends to fractionate the intrinsic unity of personality into separate rows and columns and uncoordinated traits. But human nature and behavior do not work that way. They behave and react as highly interrelated and coordinated dynamic systems. And some kind of grouping in categories articulated in domains will probably always be needed, when clinicians communicate short information about their patients

How Can We Improve Clinical Utility in the Future? The Need for Further Research It will be important for the primary focus of research questions on dimensional approaches to maximize clinical utility, not to prove the worth of any specific model. Advocates of a specific model should be asked to demonstrate empirically the model’s strength (and weaknesses) in achieving useful clinical ends. It will be important for the field to be open to such research when clinically relevant research questions have been clearly formulated. An integrated model that coordinates several models should be developed, to be validated with regard to clinical utility, as well as to other significant aspects worthy of validation (e.g., genetic, neurobiological, developmental). An aim of future studies should be to determine which dimensional system (e.g., those of Oldham, Shedler and Westen, Costa and McCrae, Livesley, Tyrer, Clark, Cloninger, Millon) can add to our current knowledge of personality disorders. One or several of these (or a new integrated system) may, in the long run, add to the categorical approach or become an alternative. Most importantly, not the least for the World Health Organization, these new very promising alternative dimensional models need to be tested for their clinical utility in a transcultural context. It will be quite a challenge to convince psychiatrists worldwide that improved alternatives exist to the well-established medical approach to classification of personality disorders. Therefore, it will be important to anchor further research in the alternative (and integrated) model(s) in an international collaboration research network that involves clinicians throughout the world. How to maximize clinical utility is a question that clinicians certainly should help us to answer.

References Abraham R, van Dam Q: The Developmental Profile: its use in establishing indications for psychoanalytic treatment. Bull Menninger Clin 68:115–136, 2004 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Kernberg O: Borderline personality organization. J Am Psychoanal Assn 15:641–685, 1967

Simonsen: Commentary on Verheul

225

Magnavita JJ: The relevance of theory in treating personality dysfunction, in Handbook of Personality Disorders, 5th Edition. Edited by Magnavita JJ. Hoboken, NJ, John Wiley & Sons, 2004 Oldham JM, Skodol AE: Charting the future of Axis II. J Personal Disord 14:17–29, 2000 Tyrer P, Johnson T: Establishing the severity of personality disorder. Am J Psychiatry 153:1593–1597, 1996 Weinryb RM, Rössel RJ, Åsberg M, et al: The Karolinska Psychodynamic Profile (KAPP). Psychoanal Psychol 14:495–515, 1997 Widiger TA, Sanderson CJ: Towards a dimensional model of personality disorders in DSMIV and DSM-V, in The DSM-IV Personality Disorders. Edited by Livesley WJ. New York, Guilford, 1995, pp 433–458 World Health Organization: The ICD-10 Classification of Mental and Behavioural Disorders. Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland: World Health Organization, 1992

This page intentionally left blank

25 COMMENTARY ON VERHEUL Clinical Utility of Dimensional Models for Personality Pathology Theresa Wilberg, M.D., Ph.D.

As interest builds in developing a dimensional diagnostic system for personality disorders in DSM-V, it will be necessary to decide whether to choose a diagnostic system that builds upon, or represents a further development of, our current diagnostic prototypes, the so-called hybrid models, or, alternatively, to shift toward a more radical dimensional approach, represented by the pure dimensional models. As described by Roel Verheul (Chapter 23 in this volume), DSM-IV Axis II has been criticized for poor clinical utility, a weakness that may continue to adhere to the hybrid models. Whereas it is easy to join this criticism, one should be careful not to overlook the potential benefits of the current system with regard to clinical decision making. A total DSM-IV (American Psychiatric Association 1994) personality disorder criteria profile of a patient comprises a large amount of valuable information that, in the minds of clinicians, will start a train of associations regarding personality type and psychological processes, including defense styles and treatment processes. Moreover, the number of criteria within the specific personality disorder categories is associated with dysfunction and symptomatic distress (Cramer et al. 2003), whereas the total number of fulfilled personality disorder criteria provides information—possibly an underestimated amount—on the extent of personality pathology. The total number of fulfilled criteria combined with a criteria profile and a Global Assessment of Functioning rating may serve as a useful basis for assessing the severity of the disorder, for selecting broad categories of 227

228

Dimensional Models of Personality Disorders

treatment modalities, and for making prognoses (Joyce 2000; Ogrodniczuk et al. 2001; Wilberg et al. 2003), even if this approach is less powerful with regard to specific treatment interventions. One could ask, however, if the lack of empirical knowledge of treatment response and prognosis for most personality disorder categories (other than borderline, schizotypal, and, to some degree, antisocial personality disorder) has less to do with a weakness of the classification system and diagnostic prototypes than with a bias, in both clinical and research literature, against focusing on these specific disorders. It is encouraging that this state of affairs is gradually changing, as avoidant and obsessive personality pathology are slowly beginning to attract the attention of researchers (Skodol et al. 2002). In our search for a dimensional classification model, we may run the risk of underestimating the importance of the knowledge and familiarity inherent in the current clinical tradition, and the pure dimensional models still have a way to go to catch up with the so-called hybrid models at this point. But, of course, both the pure and the hybrid models need to be validated in terms of treatment response and longer-term prognosis. Studying these aspects of clinical utility is a time-consuming endeavor which, given the time schedule for DSM-V, will require extensive research collaboration to enroll samples large enough to allow comparison between models. Many will agree that when a dimensional system is introduced, it should include some kind of categorical approach. From a clinical perspective, establishing thresholds for “caseness” is essential for internal as well as external communication purposes. Moreover, as data from Norwegian specialized treatment settings show, the distinction between a diagnosis of personality disorder versus no personality disorder is a powerful dichotomy when it comes to treatment prognosis (Karterud et al. 2003; Wilberg et al. 2003), even if recent research has cast some doubt on this distinction concerning the prognosis of coexisting mood and anxiety disorders (Dressen and Arntz 1998; Mulder 2002). The personality disorder/no personality disorder dichotomy may seem unsound, as it covers an underlying dimensional grading of personality pathology. However, clinicians are familiar with dealing with disorders as categories, while at the same time they have an understanding of psychopathology as continuous or dimensional phenomena. This bearing may be facilitated by underscoring the conceptual nature of a diagnosis at this level. The view that diagnoses are concepts, rather than natural kinds, is a stance that perhaps is most pronounced within the prototype tradition of classification (Horowitz et al. 1981). One contribution to a definition of “caseness” would be to face up to the challenge of developing more clearly defined general criteria for personality disorder, criteria that can be assessed reliably in clinical practice, independent of the more temperamentally or biologically based personality traits that will describe the specific types of disorder. Whereas extremeness of traits has received much attention, the inflexibility and nonadaptivity of traits has not been extensively focused upon.

Wilberg: Commentary on Verheul

229

The distinction and relationship between basic trends and characteristic adaptations seem essential for understanding personality pathology. Important work needs to be done in this area. One line of research would be a thorough investigation of the large proportion of patients in clinical practice who receive the vaguely defined diagnosis of personality disorder not otherwise specified, or those patients considered as having personality problems not included in the present classification system (Verheul and Widiger 2004; Westen and Arkowitz-Westen 1998). However, there are already proposals for criteria or domains to be included in general criteria of personality disorder that also could serve as a starting point for such research (Cloninger 2000; Livesley 1998; R. Verheul, unpublished data, November 2002). These proposals comprise a broad range of areas, such as structural aspects of the self system, self and other representations, attachment capacity, reflective functioning, and integration of normative values. To understand these aspects of personality disorders, we will need concepts of inner psychological processes, like defense mechanisms or early maladaptive schemes. Advancements in this area probably have been hampered by the ambitions of an atheoretical approach in previous versions of DSM, leading to an avoidance of descriptions and conceptualization of psychological mechanisms and processes. Maybe the time now has come to agree on a common language for the inner life of a person that is sufficiently broad and elastic to enhance the possibility of a consensus, yet specific enough to be found meaningful by most clinicians. Another important point involves the stability of personality pathology. Clinicians are interested in the degree to which personality disorders can change, and how this occurs. Recent research from clinical and nonclinical samples indicate that dysfunctional traits and personality disorder diagnoses are not as stable over time, as was previously assumed (Grilo et al. 2004; Lenzenweger et al. 2004; Zanarini et al. 2003). At present, we do not really know what the more stable and unstable aspects of personality disorder are. The study of stability and change in personality pathology may be strengthened by a more pronounced acknowledgement of the importance of the interpersonal context of the subject, for the underlying traits and vulnerabilities to become expressed in dysfunctional behaviors and symptoms. Moreover, research on change in personality disorder has suffered from lack of an established time frame with regard to a definition of recovery from a personality disorder diagnosis. At the same time, in order to arrive at an empirically based definition of recovery or remission, the candidate models should be subjected to longer-term follow-up studies. One hypothesis is that those temperamentally based traits that are useful for characterizing specific types of disorders may not be equally helpful for assessing change in personality pathology. Defining core pathology could therefore also be a way to delineate aspects of personality that may be more amenable to change, although this is certainly an empirical issue. The dual purpose of a classification system, being an instrument for both research and clinical practice, may in most instances be unproblematic. However, at

230

Dimensional Models of Personality Disorders

least theoretically, one could think of aspects of classifications models where research and clinical considerations are not concurrent. For example, a classification model aimed at detecting and adjusting to the biological underpinning of personality disorders may not be equally suitable for everyday clinical practice. A research agenda for DSM-V that includes investigation of clinical utility with regard to the different models would help us become aware of possible areas of divergent interests between research and clinical work, where priorities and compromises will have to be made.

References American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 Cloninger CR: A practical way to diagnosis personality disorder: a proposal. J Personal Disord 14:99–108, 2000 Cramer V, Torgersen S, Kringlen E: Personality disorders, prevalence, socio-demographic correlations, quality of life, dysfunction, and the question of continuity. Persönlichkeitsstörungen 7:189–198, 2003 Dressen L, Arntz A: The impact of personality disorders and treatment outcome of anxiety disorders: best-evidence synthesis. Behav Res Ther 36:483–504, 1998 Grilo CM, Sanislow CA, Gunderson JG, et al: Two-year stability and change of schizotypal, borderline, avoidant, and obsessive-compulsive personality disorders. J Consult Clin Psychol 72:767–775, 2004 Horowitz LM, Post DL, French RD, et al: The prototype as a construct in abnormal psychology, II: clarifying disagreement in psychiatric judgments. J Abnorm Psychol 90:575– 585, 1981 Joyce AS: Partial hospital group therapy programs for patients with personality disorders. Int J Group Psychother 50:531–536, 2000 Karterud S, Pedersen G, Bjordal E, et al: Day hospital treatment of patients with personality disorders. Experiences from a Norwegian treatment research network. J Personal Disord 17:173–193, 2003 Livesley WJ: Suggestions for a framework for an empirically based classification of personality disorder. Can J Psychiatry 43:137–147, 1998 Mulder RT: Personality pathology and treatment outcome in major depression: a review. Am J Psychiatry 159:359–371, 2002 Ogrodniczuk JS, Piper WE, Joyce AS, et al: Using DSM Axis II information to predict outcome in short-term individual psychotherapy. J Personal Disord 15:110–122, 2001 Skodol AE, Gunderson JG, McGlashan TH, et al: Functional impairment in patients with schizotypal, borderline, avoidant, or obsessive-compulsive personality disorder. Am J Psychiatry 159:276–283, 2002 Verheul R, Widiger TA: A meta-analysis of the prevalence and usage of the personality disorder not otherwise specified (PDNOS) diagnosis. J Personal Disord 18:309–319, 2004

Wilberg: Commentary on Verheul

231

Westen D, Arkowitz-Westen L: Limitations of Axis II in diagnosing personality pathology in clinical practice. Am J Psychiatry 155:1767–1771, 1998 Wilberg T, Karterud S, Pedersen G, et al: Outpatient group psychotherapy following day treatment of patients with personality disorders. J Personal Disord 17:510–521, 2003 Zanarini MC, Frankenburg FR, Hennen J, et al: The longitudinal course of borderline psychopathology: 6-year prospective follow-up of the phenomenology of borderline personality disorder. Am J Psychiatry 160:274–283, 2003

This page intentionally left blank

26 PERSONALITY DISORDER RESEARCH AGENDA FOR DSM-V Thomas A. Widiger, Ph.D. Erik Simonsen, M.D. Robert F. Krueger, Ph.D. W. John Livesley, M.D., Ph.D. Roel Verheul, Ph.D.

The purpose of this chapter, authored by the Steering Committee of the American Psychiatric Association’s (APA) conference on dimensional models of personality disorder (Widiger and Simonsen 2005b), is to provide a summary of the findings and recommendations of the conference. The conference (“Dimensional Models of Personality Disorder: Etiology, Pathology, Phenomenology, and Treatment”) was devoted to reviewing the existing research and setting a future research agenda that would be most effective in leading the field toward a dimensional classification of personality disorder. Topics covered included 1) alternative dimensional models of personality disorder, 2) behavioral and molecular genetics, 3) neurobiological mechanisms, 4) childhood antecedents, 5) cross-cultural issues, 6) Axes I and II

This chapter is an abbreviated version of a paper with the same title first published in the Journal of Personality Disorders (Volume 19, Issue 3, pages 315–338, 2005).

233

234

Dimensional Models of Personality Disorders

continuity, 7) coverage and cutoff points for diagnosis, and 8) clinical utility. Our summary will be organized with respect to the eight topics that were addressed.

Alternative Dimensional Models Widiger and Simonsen (2005a) summarized in their presentation 18 alternative ways in which personality disorders could be converted to a dimensional model of classification. They called for an integration of these alternative models within a single, common hierarchical model. At the highest level could be the two clinical spectra of internalization and externalization identified by Krueger (1999, 2002). Immediately beneath would be four to five broad domains of personality functioning. Further below would be personality trait scales, and at the lowest level would be the more behaviorally specific diagnostic criteria. Widiger and Simonsen (2005a) suggested that all but one or two of the existing dimensional models of personality and personality disorder could be well represented within this common hierarchical structure. They illustrated how most of the scales of the dimensional models developed by Clark (Clark et al., in press), Cloninger (2000), Costa and McCrae (1992), Eysenck (1987), Harkness and McNulty (1994), the interpersonal circumplex (Wiggins 2003), Livesley (2003), Millon et al. (1996), Shedler and Westen (2004), Tellegen (Watson et al. 1999), Tyrer (2000), and Zuckerman (2002) could be well integrated within four broad domains of adaptive and maladaptive personality functioning (i.e., emotional dysregulation versus emotional stability, constraint versus impulsivity, extraversion versus introversion, and antagonism versus compliance). They further indicated how the existing personality disorder diagnostic criteria would be readily incorporated within this hierarchical structure and how the existing personality disorder constructs (e.g., antisocial or borderline) could be recovered through diagnostic algorithms using the personality trait scales. However, they also acknowledged that the placement of scales from some of the models (e.g., the scales from the Temperament and Character Inventory of Cloninger [2000] and the Personality Assessment Schedule of Tyrer [2000]) was based on only a limited amount of research. They suggested that one important focus of future research will be to determine whether the 3 (or 6) polarities of the Millon Index of Personality Styles (Millon 1994) and the 12 scales of the Shedler and Westen Assessment Procedure—200 (Shedler and Westen 2004) can also be integrated with the higher-order structure or whether they concern instead aspects of maladaptive personality functioning that are not commensurate with the other dimensional models. Some of the specific scale placements of Widiger and Simonsen (2005a) might also be disputed. Future research could help determine how these alternative dimensional models of personality disorder could be best integrated into a common hierarchical structure.

Widiger et al.: Personality Disorder Research Agenda for DSM-V

235

Not only is it important to determine where the scales and constructs should be included but also how many to include within each broad domain and how best to identify them. Considered in these decisions could be extent of overlap, adequate representation of different models, adequate coverage of the domain, clinical relevance, familiarity, and ease of usage. It will also be important to further consider the question of whether a fifth domain of unconventionality (e.g., cognitive-perceptual aberrations and peculiar, eccentric behaviors) also warrants inclusion. Last, but not least, is the fundamental question of whether also to include normal, adaptive traits. It is possible that the diagnostic manual simply should not include any reference to normal (healthy) psychological functioning. Conversely, it also is possible that the inclusion of normative, adaptive traits will facilitate the provision of a more comprehensive (and accurate) description of each patient’s general personality structure, the integration of the diagnostic manual with basic science research on general personality structure, and the facilitation of treatment decisions through the identification of traits that contribute to treatment responsivity.

Behavioral and Molecular Genetic Contributions Livesley (2005) summarized the existing behavioral and molecular genetic research. Although it is likely that a future edition of the diagnostic manual will (and should) incorporate a provision for encoding genotypes, there is currently only very limited data supporting a relationship between genetic polymorphisms and specific personality traits. There is more potential and existing empirical support for these associations than has been or likely will be obtained for the DSM-IV (American Psychiatric Association 1994) personality disorder diagnostic categories (Jang et al. 2001) but as yet it is unlikely that the next edition of the diagnostic manual could be guided explicitly by the molecular genetic research. Of more immediate relevance for developing a dimensional classification of personality disorder in DSM-V would be behavioral genetic research that seeks to explicate the genetic and environmental structure underlying phenotypic variation (e.g., Ando et al. 2004; Jang et al. 2002). A dimensional classification of personality disorder requires information about the etiological factors responsible for observed patterns of covariation. The development of multivariate genetic techniques permits the consideration of these findings to form the foundation for an etiologically based classification. Multivariate genetic analyses extend univariate analyses of genetic influences on a single trait to estimate genetic and environmental influences on the covariation among two or more traits. Genetic and environmental covariation matrices can then be factored to provide information on the structures underlying the covariation. The results of the behavioral genetic research can be used to refine personality phenotypes and to construct a genetically informed nosology. A

236

Dimensional Models of Personality Disorders

set of genetically defined primary traits would facilitate molecular research by providing targets with more homogeneous genetic variance. Molecular genetic research can then be used to fine-tune this nosology and provide further validation. Existing multivariate genetic research, using a wide variety of measures, suggests that a few general genetic factors account for the observed patterns of trait covariation (Livesley 2005). The remarkably consistent finding that four secondary traits (i.e., emotional dysregulation, constraint/conscientiousness, antagonism/ dissocial, and inhibition/introversion) are sufficient to represent personality disorder provides an initial structure for the system (Livesley et al. 1998). This suggests that personality is subject to extensive pleiotropic effects in which a single genetic entity influences distinct phenotypes. The high genotype–phenotype correspondence that occurs with studies of dimensional models of adaptive and maladaptive personality functioning contrasts with the relatively poor geneotypic–phenotypic correspondence observed with many existing diagnostic categories (Merikangas 2002). Evidence that trait structure primarily reflects genetic influences forms a valuable part of the foundation for a genetically informed classification. We are in fact in a position to develop a classification that is structured along etiological lines; in this sense the diagnosis of personality disorders is perhaps in a stronger position than most other mental disorders. Nevertheless, genetic research also raises conceptual problems that will need to be addressed in future research. For instance, the issue of the relative importance of primary and secondary traits is basic to the way a dimensional classification is organized and used. Primary traits are generally assumed to be part of a secondary trait domain, and secondary domains are generally assumed to be equal in breadth and to be defined by the same number of primary traits. However, it seems improbable that such models reflect the genetic architecture of personality disorder. Primary traits are likely to have unique genetic variance, and some may not even be part of a secondary domain.

Neurobiological Mechanisms Paris (2005) began his review by noting that existing research has failed to identify any consistent biological factors that correlate with the current diagnostic categories. As Paris (2000, 2003) has suggested elsewhere, the existing diagnostic categories are too heterogeneous to have biological coherence, as defined by consistent and specific relations to biological markers. In fact, there has been little to no effort even to attempt to explicate the neurobiology of some of the current diagnostic categories (e.g., histrionic, narcissistic, dependent, and obsessive-compulsive; Blashfield and Intoccia 1999). Paris (2005) reviewed the existing molecular genetic, behavioral genetic, neuroimaging, pharmacological responsivity, and additional human and animal

Widiger et al.: Personality Disorder Research Agenda for DSM-V

237

research relevant to neurobiological mechanisms for dimensional models of personality and personality disorder that have placed a particular emphasis on a neurobiological foundation. He focused on the dimensional models of Cloninger (2000), Depue (Depue and Collins 1999; Depue and Lenzenweger 2001), and Siever (Siever and Davis 1991). He noted that the empirical literature is quite extensive, but this research has also yielded quite inconsistent results. For example, the conclusions of some meta-analytic reviews of the molecular genetic research has reached positive conclusions. In an extensive meta-analysis involving 5,629 subjects, Sen et al. (2004) concluded that “there is a strong association between the serotonin transporter promoter variant and neuroticism...and that nonreplications are largely due to small sample size and the use of different inventories” (p. 85). Paris also suggested that there was consistent empirical support for an association of serotonergic functioning with harm avoidance and impulsive aggression. Nevertheless, other meta-analyses of molecular genetic research concerning (for example) novelty seeking has yielded more negative conclusions (e.g., Kluger et al. 2002; Schinka et al. 2002). Paris (2005) concluded that it might be premature or unrealistic to attempt to build a dimensional model of maladaptive personality functioning solely or even primarily on the basis of hypothesized neurobiological mechanisms. This effort might be premature in that our understanding of brain mechanisms is at too early a stage to conduct this research effectively. Until we know more about the basic science of emotions and behaviors, attempts to develop a neurobiological model linked to personality dimensions might be premature. The effort might also be unrealistic, given that no brain function is strictly limited to one site or to one neurotransmitter; modulation and interaction are the rules rather than the exceptions (Andreasen 2001). For example, the monoamines, which have been a main subject of research, serve to modulate the effects of other neurons that use glutamine and gamma-aminobutyric acid (GABA) as transmitters (Cooper et al. 2003). Their effects on behavior are far from linear, with the same receptors potentially having entirely different effects in different brain locations, depending on brain anatomy as well as physiology. Serotonin may have as many as 15 receptor sites (Kroeze et al. 2002). These findings make it unlikely that one will find a one-to-one correspondence between any single neurotransmitter and any single neurophysiological mechanism, not to speak of behavioral traits (Paris 2005). Minimally, future research is advised to focus on narrower personality trait dimensions, rather than the broad domains that have currently been the focus of research. In the meantime, Paris advises that it would probably be better to use a factor analytically derived personality schema rather than one based on a particular neurobiological hypothesis. The dimensions obtained from this research could be revised at a later point in time on the basis of the brain–behavior research.

238

Dimensional Models of Personality Disorders

Childhood Antecedents The review by Mervielde et al. (2005) began with the point that although it is widely recognized that adult personality disorders have their roots in a variety of developmental and temperamental factors, there has been remarkably little research examining the childhood and adolescent antecedents of the DSM-IV personality disorders (e.g., Johnson et al. 2000). Only one of the 10 personality disorders included in DSM-IV refers to childhood antecedents (American Psychiatric Association 2000). This is in stark contrast to the extensive research on the relationship of childhood temperament with adult personality structure (Roberts and DelVecchio 2000). Child and adolescent temperaments are probably among the best candidates as general broadband developmental antecedents for adult personality disorders (Krueger and Tackett 2003; Shiner and Caspi 2003). Researchers interested in individual differences among children and especially among young children have generally conceptualized these differences in terms of temperamental characteristics. Temperament is traditionally distinguished from personality as it refers to stable individual differences that appear from birth onward that presumably have a strong genetic or neurobiological basis. Mervielde et al. (2005) summarized the predominant temperament models, particularly those of Chess and Thomas (1996), Buss and Plomin (1984), Derryberry and Rothbart (1997), and Goldsmith and Camos (1982). They suggested that the alternative models can be integrated within four basic temperaments: 1) Emotionality (negative affectivity, anxious or irritable distress), 2) Extraversion (sociability versus social inhibition or shyness), 3) Activity (activity level, surgency), and 4) Persistence (task persistence, effortful control). Mervielde et al. (2005) further related these fundamental temperaments to adult personality traits on the basis of the considerable number of studies that have been conducted, including their own extensive peer nomination and childhood lexical studies. For example, Kohnstamm et al. (1998) collected unstructured parental personality descriptions of 2,416 children ages 2–12 years as part of a collaborative international research project conducted in Belgium, China, Germany, Greece, Holland, Poland, and the United States. Using the 9,000 free parental descriptions provided by Flemish parents, Mervielde and De Fruyt (2002) developed the Hierarchical Personality Inventory for Children (HiPIC). Analysis of its items led to a five-factor structure that they identified as Conscientiousness, Benevolence, Extraversion, Emotional Stability, and Imagination. They indicated how this classification integrated well the person-centered approach often used in childhood studies (e.g., overcontrollers are low in Emotional Stability and Extraversion, whereas undercontrollers are low on Agreeableness and Conscientiousness). Mervielde et al. (2005) indicated further that their childhood personality model is congruent with the recent conclusions of Shiner and Caspi (Shiner 1998; Shiner and Caspi 2003), who also reviewed the vast temperament literature and

Widiger et al.: Personality Disorder Research Agenda for DSM-V

239

suggested that much of the apparently disparate temperaments appear to be well organized within four broad domains of Extraversion or Positive Emotionality (i.e., social inhibition, sociability, dominance, energy/activity level), Neuroticism or Negative Emotionality (i.e., anxious distress, irritable distress), Conscientiousness or Constraint (i.e., attention, inhibitory control, achievement motivation), and Agreeableness (i.e., antagonism, prosocial tendencies). The only difference was the absence of an imagination dimension, which they suggested could reflect that preschool teachers do not generally distinguish curiosity and creativity from conscientiousness. In any case, Caspi et al. (2005) now do include an Openness dimension in their more recent reviews of the childhood research. Mervielde et al. (2005) concluded that it is evident that a general pattern of associations between temperament, personality, and psychopathology emerges across a wide variety of measures. The highest level is perhaps best represented by two broad domains of internalization and externalization (Achenbach 1995; Krueger and Tackett 2003). Individual differences can be reasonably well covered by the four higher-order traits proposed by Shiner and Caspi (2003) of neuroticism, extraversion, agreeableness, and conscientiousness. The integration of the child and adult research within a common model provides a conceptual basis for understanding the well-documented stability of personality across the life span (Caspi et al. 2003; Roberts and DelVecchio 2000). They suggested that an important contribution of future research will be to further integrate the existing personality disorder constructs of DSM-IV with the childhood temperament and adult personality literature in order to achieve a sound scientific basis for understanding the childhood antecedents of adult personality disorders.

Cross-Cultural Issues Cross-cultural studies of personality structure are fraught with difficulties, not the least of which is the expense. A few studies have considered the application of the DSM-IV personality disorder nomenclature within cultures notably different from the predominant western society in which it was largely created (e.g., Grilo et al. 2003) but there appears to be only one systematic multinational study, in which the International Personality Disorder Examination (IPDE) was administered in 14 mental health centers located in 11 different countries of North America, Europe, Africa, and Asia (Loranger et al. 1994). This is in contrast to the extensive research on the universality of general personality structure. Allik (2005) indicated that one of the first comprehensive personality trait measures to enjoy worldwide popularity and a fairly large number of translations into different languages was the Eysenck Personality Questionnaire (EPQ; Eysenck and Eysenck 1975), which assesses the broad domains of Neuroticism, Extraversion, and Psychoticism. Extensive data have been reported on its cross-

240

Dimensional Models of Personality Disorders

cultural generalizability (mean scores and trait structure) in a study including 38 countries and 68,374 participants (Lynn and Martin 1995). More recently, McCrae (2002) reported on the generalizability of the five factors of Neuroticism, Extraversion, Openness, Agreeableness, and Conscientiousness across 36 different countries involving five major language families (Indo-European, Uralic, Altaic, Dravidian, and Sino-Tibetian). McCrae et al. (2005) subsequently replicated the cross-cultural generalization using peer reports of 11,985 target individuals obtained in 50 different societies. Schmitt and colleagues (2003) conducted the largest cross-cultural study to date as part of the International Sexuality Description project, which includes 100 scientists from 56 countries. They administered in this project the Big Five Inventory (Benet-Martinez and John 1998), translated into 29 languages and administered to17,837 participants from 56 different countries. Results indicated that the five-dimensional structure was highly robust across major regions of the world, including North America, South America, Western Europe, Eastern Europe, Southern Europe, the Middle East, Africa, Oceania, South-Southeast Asia, and East Asia (Schmitt et al. 2003). The results of these (and other) studies have generally provided strong support for the universality of personality structure across a wide array of cultures and languages. There are also interesting data with regard to the relationship of these personality traits to cultural indicators, such as gross national product and Hofstede’s dimensions of culture (Hofstede and McCrae 2004). Allik (2005) indicated that it is also worth noting that the above findings are largely consistent with emic, lexical studies of the trait terms indigenous to a particular language. Some might argue that consistency of the trait structure of an inventory across many languages does not necessarily indicate that the personality traits included within the inventory are universal in their interest, relevance, or importance. Emic studies approach the question of universality by determining whether a comparable personality structure emerges from analyses of constructs or measures that are indigenous to each culture. For instance, Saucier and Goldberg (2001) summarized the results of lexical studies of the trait terms indigenous within 13 different languages (English, German, Dutch, Czech, Polish, Russian, Italian, Spanish, Hebrew, Hungarian, Turkish, Korean, and Filipino). They reported that the personality trait domains of Extraversion (Surgency or Positive Emotionality), Agreeableness, and Conscientiousness emerged consistently. The fourth and fifth factors of Emotional Instability and Openness also emerged across all languages, although not as consistently as the first three. Ashton and Lee (2001) extended these analyses, providing a systematic review of all indigenous lexical studies to date, and concluded that the findings have “repeatedly produced variances of Surgency (I), Agreeableness (II), Conscientiousness (III), and Emotional Stability (IV), the first four factors of the well-known Big Five” (p. 328). Nevertheless, more work needs to be done. Allik (2005) encouraged in particular further research on cross-cultural differences in response style (e.g., tendency

Widiger et al.: Personality Disorder Research Agenda for DSM-V

241

to describe oneself in a socially desirable manner or the tendency to acquiescence in response to inquiries). For instance, Schmitt and Allik (2005) reported Rosenberg Self-Esteem findings involving translations into 28 languages and administered to 16,998 participants across 53 countries. As expected, self-esteem correlated negatively with Neuroticism and positively with Extraversion within nearly every nation. In addition, all nations scored, on average, above the theoretical midpoint, suggesting that a generally positive self-evaluation might be culturally universal. There was no relationship between national self-esteem components of the Human Development Index (e.g., life expectancy, adult literacy rates, standard of living). However, there was a relationship of these indices with response styles. Persons from developed countries (where there is better education and longevity) appear to understand negatively worded items in a different manner than positively worded items. There appears to be a need for additional cross-cultural studies using additional methodologies, such as observer ratings and semistructured interviews. Allik (2005) also emphasized the importance of understanding the variation in mean levels across cultures. Some of the findings are difficult to explain given their inconsistency with objective measures of cultural stereotypes (e.g., relatively lower scores on conscientiousness for Japanese). On the other hand, the differences that are obtained might be so small that they will be largely negligible for cross-cultural personality description and may not even replicate in subsequent studies. The relatively modest size of the cross-cultural differences in scale elevations could in fact suggest that a reasonable scalar equivalence can be achieved and all individuals, regardless of language and culture, can be represented in a common metric (Allik 2005). Nevertheless, the implications of these differences in mean levels for a universal measure of maladaptive personality functioning is not yet clear. It is at least evident, however, that a strong scientific base for developing a universal dimensional model of personality disorder can be provided through its integration with the extensive cross-cultural study of general personality structure.

Continuity of Axis I and Axis II Personality disorders are placed on a separate axis of DSM-IV-TR (American Psychiatric Association 2000). The separate axis placement has received some criticism (Livesley 2003; Widiger 2003). Even if personality disorders are moved to Axis I in DSM-V, there would also remain the concern that there might not be any meaningful or clear boundary between personality and other mental disorders. Krueger (2005) reviewed the evidence for putative bases for distinguishing between personality and other mental (clinical) disorders, including temporal stability, age at onset, treatment response, insight, diagnostic co-occurrence, and etiology. He concluded that the personality and clinical disorders are not well dis-

242

Dimensional Models of Personality Disorders

tinguished in terms of any one (or more) of them. He suggested that the most promising general direction for future research would therefore be an understanding of why and how the personality and clinical disorders are so interconnected. He further suggested that a better understanding of this connection would be through an integration with the structure of general personality. Increasing evidence points to the feasibility of developing an empirically based model of personality that simultaneously incorporates both normal and abnormal variation (Trull and Durrett 2005). For example, Markon et al. (2005) provided a joint structural model of the constructs assessed with the Dimensional Assessment of Personality Pathology—Basic Questionnaire (DAPP-BQ; Livesley and Jackson, in press), the Eysenck Personality Questionnaire (EPQ; Eysenck and Eysenck 1975), the Multidimensional Personality Questionnaire (MPQ; Tellegen, in press), the NEO Personality Inventory—Revised (NEO PI-R; Costa and McCrae 1992), and the Temperament and Character Inventory (TCI; Cloninger 2000). Markon et al. (2005) first used a meta-analytic approach to assembling a matrix of correlations among the 44 scales derived from all of these inventories obtained from 52 prior studies. Structural modeling yielded the following conclusions. First, the data indicated that no more than five major factors underlie variation in the 44 scales. Second, these five factors (Neuroticism, Agreeableness, Conscientiousness, Extraversion, and Openness) strongly resembled the domains of the five-factor model (Costa and Widiger 2002). Further analyses, however, also supported the existence of meaningful factors above the level of the five; specifically, the four-factor level resembled four-factor models often articulated in the personality and psychopathology literature (e.g., Livesley et al. 1998; O’Connor and Dyce 1998; Watson et al. 1994). The three-factor level resembled the three factors of Clark and Watson (1999), Eysenck (1987), and Tellegen (in press), with the dimensions of Negative Emotionality, Disinhibition (a combination of disagreeableness and unconscientiousness), and Positive Emotionality. Finally, the two-factor model resembled the two-factor model of Digman (1990), with one factor (alpha) combining Neuroticism, Agreeableness, and Conscientiousness; the other factor (beta) combining Extraversion and Openness. The second study of Markon et al. (2005) replicated this hierarchical model using a new, unique sample of participants who completed the NEO PI-R (Costa and McCrae 1992), the EPQ-R (Eysenck et al. 1985), the Schedule for Nonadaptive and Adaptive Personality (Clark et al., in press), and the Big Five Inventory (John et al. 1991). The Markon et al. (2005) results suggest that the five-factor level of the hierarchy is the basic organizing framework of choice because there was no compelling evidence for higher-order structures beyond the five, and structures above the five can be understood as combinations of the five-factor domains. Nevertheless, the analyses also indicated that structures above the five factors, as well as the facet level scales that delineate the five domains, are also theoretically and clinically important. For example, the domain of Disinhibition, which is above the five-factor

Widiger et al.: Personality Disorder Research Agenda for DSM-V

243

level (combining disagreeableness and unconscientiousness), is closely linked to antisocial behavior (e.g., Lynam and Derefinko, in press). In addition, many of the scales in the Markon et al. analyses contained substantial amounts of residual variance that could not be accounted for by the higher-order factors. For example, the DAPP-BQ (Livesley and Jackson, in press) Self-Harm scale was a marker for the broad domain of Neuroticism, but the majority of the variance in DAPP-BQ SelfHarm was unique to this scale. Self-harm is of obvious clinical importance and inasmuch as it cannot be captured entirely by the broad Neuroticism domain, it represents an example of a specific, facet-level construct below the five broad domains that might be important to include in a complete system of normal–abnormal personality description. Indeed, the question of which facet-level constructs should be included in a comprehensive system of normal–abnormal personality description is an important topic for continued research and discussion. The detailed information about personality functioning contained in facet-level constructs is important in connecting broad domains of personality with the richness of clinical phenomena encountered with personality disorders (Shedler and Westen 2004). New developments in the modeling of multivariate data (Muthen 2002) can also be used to identify points of greater and lesser density within the multivariate space. These regions of greater density (if they exist) might be thought of as frequently encountered personality configurations, and some of these configurations might be of particular clinical importance (e.g., Hicks et al. 2004). Exploration at the higher-order level is helpful in leading toward a potential integration of the personality and Axis I disorders. Diagnostic co-occurrence is not necessarily an artifact or nuisance; it may represent instead a reliable empirical observation in need of an explanatory model. Diagnostic co-occurrence makes sense when thought of in terms of the personological underpinnings of these disorders (Krueger and Tacket 2003). Krueger et al. (1998) demonstrated that the co-occurrence of unipolar mood and anxiety disorders appears to reflect a latent internalizing propensity, and the co-occurrence of substance dependence and antisocial behavior disorders a latent externalizing propensity. This internalizing–externalizing structure can also be observed in the primary care setting in numerous countries around the globe, with the internalizing spectrum also encompassing somatoform disorders (Krueger et al. 2003). Putting the personality findings together with the findings on the structure of mental disorders, Neuroticism (Negative Affectivity) appears to provide the personological basis for internalizing psychopathology, and Negative Emotionality paired with Disinhibition the personological basis for externalizing psychopathology. Thus, the connections between personality and psychopathology make psychological sense within a common, hierarchical, integrative dimensional model. This conceptualization is further supported by behavioral genetic research that has obtained consistent findings (Kendler et al. 2003; Krueger and Tackett 2003). Much work, however, is needed before a complete reorganization of the diagnostic manual along these lines is possible. Krueger (2005) suggested that we first

244

Dimensional Models of Personality Disorders

consider converting the existing personality disorder section to a system of facetlevel constructs organized in terms of five broad domains (Markon et al. 2005) and pursue research and discussion of the most optimal facets for representing these domains in clinical practice. Second, we should consider reorganizing sections and disorders described within the diagnostic manual to recognize the internalizing and externalizing spectra, and pursue research and discussion on the most optimal organization of diagnoses within these spectra. Finally, we should encourage further research linking cutting-edge developments in methodology with novel ideas about how to describe and organize personality and psychopathology constructs. Openminded, creative research that asks probing and novel questions but sticks close to the data in pursuing these questions has real potential to ultimately result in a diagnostic system that is empirically supported, useful in the clinic, and inspires research that leads to better prevention and treatment of mental disorders.

Coverage and Cutoff Points Trull (2005) considered two major challenges that must be addressed by a dimensional model of personality disorder if it is to be considered a viable alternative to the present categorical system: coverage and cutoff points. A dimensional model must adequately cover the existing conditions that are currently seen in clinical practice, and provide a meaningful basis with which to diagnose the presence of personality disorder. It is evident that the existing diagnostic manual fails to provide adequate coverage of the maladaptive personality functioning seen in clinical practice (Verheul and Widiger 2004; Westen and Arkowitz-Westen 1998), yet at the same time it contains substantial redundancy, evident in part by the excessive diagnostic cooccurrence (Trull and Durrett 2005). Concerns regarding coverage generally have been addressed in existing research through determining whether a proposed dimensional model would adequately include the personality disorder symptomatology that is currently included within the existing criterion sets. In some instances, the answer to this question is rather straightforward because the dimensional models were derived in large part through analyses of the existing diagnostic criteria (e.g., Clark et al., in press; Livesley and Jackson, in press; Tyrer 2000; Westen and Shedler 2000). The five-factor model was developed instead to assess general personality structure, but Trull (2005) noted that more than 50 studies have now documented that the existing diagnostic criteria can be well understood from the perspective of this model (Livesley 2001; Saulsman and Page 2004; Widiger and Costa 2002). There are now even studies demonstrating that the extent to which a person matches a five-factor profile for a DSM-IV personality disorder has as much convergent and discriminant validity as a direct measure of that respective personality disorder (e.g., Miller et al. 2001; Trull et al. 2003).

Widiger et al.: Personality Disorder Research Agenda for DSM-V

245

A potential advantage of dimensional models of personality disorder relative to the existing diagnostic categories is the ability to characterize unique or idiosyncratic personality profiles that are currently not well represented by one of the existing diagnostic categories. One of the more common diagnoses within clinical practice has been personality disorder not otherwise specified, due in large part to the failure of the existing set of 10 diagnoses to provide adequate coverage of the maladaptive personality traits that are seen in clinical practice (Westen and Arkowitz-Westen, 1998; Verheul and Widiger 2004). All of the existing dimensional models of personality disorder are likely to provide better coverage than the existing diagnostic categories through individualized profile descriptions that will be relatively unique to each patient and through the inclusion of additional traits not included within any one of the existing diagnostic categories. Even the dimensional models of personality disorder that were derived in large part through analyses of the symptoms included within the existing categories have gone beyond the categories to include additional symptomatology (e.g., Clark et al., in press; Livesley and Jackson, in press; Westen and Shedler 2000). Nevertheless, it would be useful for future studies to document empirically that the inclusion of these additional traits (maladaptive and adaptive) do in fact provide incremental validity for clinical assessments and increase the coverage of cases currently receiving a diagnosis of personality disorder not otherwise specified. An issue not well addressed empirically by the existing research is how to use a dimensional model of personality disorder to make the clinical distinction between the presence versus absence of a personality disorder. Existing research documents that there does not appear to be a clear, qualitative distinction between normal and abnormal personality functioning (one of the arguments favoring a dimensional model), but the absence of any apparent or obvious point of distinction then makes it imperative that concerted effort be given toward developing a reasonable or meaningful basis for any such distinction. Cutoff points will be needed for clinical decisions (e.g., whether to provide treatment, medication, or insurance coverage). A purported advantage of dimensional models is a flexibility of cutoff points for these different social and clinical decisions. Nevertheless, the existing research has not provided any explicit support for how these decisions could or should be made. A few of the alternative models have included proposals for how such a distinction could be made. Westen and Shedler (2000) suggest using a relatively subjective prototypal matching procedure whereby the clinician uses his or her professional judgment in determining whether the personality traits of a patient are sufficiently close enough to a brief, narrative description of a prototypic case to warrant a diagnosis. Cloninger (2000) emphasizes low scores on his scale of SelfDirectedness, and secondarily low scores on Cooperativeness, Affective Stability, and/or Self-Transcendence, as a basis for determining whether a personality disorder is present. If a personality disorder is present, then other TCI scales would be

246

Dimensional Models of Personality Disorders

used to identify the specific features of the personality disorder. Livesley (2003) suggests that the initial assessment for the presence of a personality disorder should focus on a failure to establish and maintain stable representations of the self and others; interpersonal dysfunction; and/or a failure to develop prosocial behavior and cooperative relationships. The procedure of Widiger et al. (2002) is to first describe the person’s personality profile (including normal and abnormal traits), and then determine whether the maladaptive personality traits reach a clinically significant level of impairment. These are all admirable proposals, but they all share the feature that there is little to no direct empirical research on the reliability, validity, or utility of their clinical application. It is not known whether these proposed distinctions would resemble closely the diagnostic thresholds that are currently provided in DSM-IV or the diagnostic thresholds that are generally used by clinicians when they provide a diagnosis of personality disorder, not otherwise specified. In defense of the absence of this research, it should be noted that there has also been little to no research on the diagnostic thresholds currently provided in DSMIV (Samuel and Widiger 2006). What is needed in future research is a more concerted effort to identify optimal cutoff points for alternative clinical and social decisions. A personality disorder diagnoses could be made at that point in which there is evidence of a clinically significant level of impairment to social or occupational dysfunction, but as yet there is no clear consensus as to what constitutes a clinically significant impairment (Wakefield and First 2003). It is unlikely that a cutoff point could be based simply on statistical deviance, as each personality dimension will have different implications for impairment (e.g., a lower cutoff point would likely be used for a dimension of self-harm than for a dimension of social avoidance). In addition, it might be useful to disengage personality assessment from the determination of the presence of a personality disorder. This will facilitate the consideration of adaptive personality functioning and a more explicit, differentiated assessment of clinical impairment (Lehman et al. 2002). Different cutoff points likely will be necessary for alternative clinical decisions; for example, when to recommend medication, hospitalization, or entrance into a specialized treatment program (e.g., dialectical behavior therapy). In sum, a dimensional model of classification has much to offer for an improvement in how a personality disorder is diagnosed, but a conversion to a dimensional model of classification would be facilitated by studies that are explicitly concerned with the identification of the optimal cutoff points for these clinical decisions.

Clinical Utility As stated in DSM-IV, the “highest priority has been to provide a helpful guide to clinical practice” (American Psychiatric Association 2000, p. xxiii). Clinical utility has always been an important priority for the authors of the diagnostic manual, but

Widiger et al.: Personality Disorder Research Agenda for DSM-V

247

Verheul (2005) suggested that increased attention needs to be given to the obtaining of empirical data that will document explicitly how revisions to the diagnostic manual would address or improve matters of clinical utility (First et al. 2004). Verheul (2005) endorsed the definition of clinical utility offered by First et al. (2004); that is, “the extent to which DSM assists clinical decision makers in fulfilling the various clinical functions of a psychiatric classification system” (p. 947). First, Verheul identified elements of diagnostic validity that are conditional to clinical utility, including adequate coverage, consistency with developmental and etiological models, and consistency with models of course and change. Second, Verheul identified components that concern directly matters of clinical utility, including user acceptability and accuracy, professional communication, interrater reliability, subtlety of diagnosis, and clinical decision making. Achieving user acceptability and accuracy in the application of diagnostic criteria are critically important since nonutilization and incorrect application would eliminate any potential value that would be provided by a more valid diagnostic system. Current research suggests that the existing diagnostic criterion sets are in fact not being used effectively in clinical practice. Empirical information about the potential user acceptability of alternative dimensional models of classification, however, is scarce (Sprock 2003). It is also of considerable importance for a classification to allow for effective professional communication. It has been suggested that the existing diagnostic categories are preferable to the more complex dimensional models of classification because they allow for the communication of a considerable amount of information through the provision of a single diagnostic label. However, it also appears to be the case that the single diagnostic labels are inadequate in their description and coverage, and clinicians may in fact prefer to at least have the opportunity to obtain and consider a more extensive personality description (Zimmerman and Mattia 1999). Studies which assess and compare alternative dimensional models and the existing diagnostic categories with respect to effectiveness of professional communication would be of considerable benefit to the authors of a future diagnostic manual. The existing diagnostic categories can be assessed with adequate to high levels of interrater reliability, but it is not at all clear that adequate levels of reliability are being obtained in general clinical practice. Existing research has indicated that dimensional classifications generally obtain higher levels of reliability than categorical diagnoses, but these direct comparisons have been confined largely to studies that have used structured instruments. It is unclear whether clinicians would in fact use structured assessment instruments (Widiger and Samuel 2005) and, if they do not, whether a dimensional model of classification would still result in an improvement in reliability. Several authors have demonstrated that dimensional models permit greater subtlety of diagnosis than what can be derived from a category-based system (e.g., Stone 2002). This strength, however, is closely related to the complexity and detail

248

Dimensional Models of Personality Disorders

of the information provided. The optimum level of detail provided by any classification system will be related to issues of feasibility, preferred level for communication, and added value for clinical decision making. Studies which address specifically whether the additional information provided by a dimensional classification is actually helpful for clinical decisions would be of considerable benefit. The extent to which a taxonomy is able to direct clinical decision making is perhaps the most important component of clinical utility (First et al. 2004; Verheul 2005). Proponents of the existing diagnostic categories argue correctly that there currently exists a substantial amount of clinical literature regarding methods of treatment that are coordinated explicitly with these categories. However, a dimensional model that retains the existing personality disorder symptoms as lowerorder (behavioral) manifestations of respective personality trait scales would be able to draw upon this literature (Widiger and Simonsen 2005a). In addition, the existing diagnostic categories may not in fact be tremendously helpful to clinicians (Livesley 2001; Verheul 2005). For instance, many treatment decisions require more subtle, quantitative assessments (e.g., when and how to firmly set treatment limits, or how much to permit dependence or encourage independence) that go beyond the broad and heterogeneous diagnostic categories (Clark 1993). It is even unclear that the existing diagnostic categories are in fact helpful for such basic decisions as to whether to hospitalize or medicate a patient (Verheul 2005). Treatment decisions involve quite a few levels of consideration, including setting (e.g., inpatient, day hospital, or outpatient), format (e.g., individual, group, or family), major strategies or techniques (including choice of which theoretical model to guide conceptualization and intervention), duration (e.g., crisis intervention, short term, or long term) and frequency of appointments, and medication selection. There are a growing number of studies and clinical reports that suggest that an assessment of personality dimensions can be helpful in making these decisions (e.g., Lambert and Anderson 1996; Miller 1991; Sanderson and Clarkin 2002). There are further data and clinical reports to suggest the possible value of dimensional models in facilitating therapeutic tailoring or microtreatment decisions, including goal setting, matching patient to therapist characteristics, handling of transference, and degree of therapy directiveness (Livesley 2003; Verheul 2005). In any case, clinicians would find a conversion to a dimensional model of classification easier to accept if they were trained adequately with respect to this literature, or if it was somehow more effectively disseminated to them. In sum, there is much to be optimistic about concerning the actual and potential clinical utility of a dimensional model of personality disorder classification. Nevertheless, it is also apparent that a considerable amount of additional research in this area will be of tremendous importance. Of most use will be field trials in which the existing diagnostic categories and alternative dimensional models of classification are compared directly with respect to their clinical utility (e.g., ease of use, professional communication, interrater reliability, subtlety of diagnosis,

Widiger et al.: Personality Disorder Research Agenda for DSM-V

249

clinical decision making). Any comparison to the existing diagnostic categories will be hindered by the fact that the clinicians would have been trained with the current diagnostic system and may even be largely unfamiliar with the alternative dimensional models. Nevertheless, this research would still be of use in field testing alternative proposals and in alerting the authors of a future diagnostic manual to potential problems with the application of a dimensional model. Field trials can, of course, be quite expensive and labor-intensive. Short of this approach could be studies which survey clinicians as to their opinions regarding ease of usage, communication, and treatment decisions, either through direct questioning or in response to case vignettes.

Conclusion Participants at the conference largely endorsed a conversion to a dimensional model of classification. The conference presentations were quite helpful in documenting the advantages of and empirical support for an integrative, hierarchical dimensional classification of personality disorder, consisting of four or five broad domains of personality functioning. This research included a considerable number of studies addressing behavioral genetics, developmental antecedents, cross-cultural application, convergence with other mental disorders, coverage, and clinical utility. The participants were also helpful in generating ideas for research that would facilitate a conversion to a dimensional model of classification. More specifically, additional research is needed to help define more specifically the optimal number and content of the facets included within the four (or five) broad domains that are common across the alternative models. The alternative dimensional models have generated quite a number of possible scales to include in a dimensional model of personality disorder. Not all of them can be included, and research is needed to help lead the field toward their optimal integration within a common hierarchical structure. This research should include genetically defined primary traits that would lead to an etiologically defined classification that could in turn facilitate molecular research by providing targets with a more homogeneous genetic variance. Considered as well would be studies addressing the adequacy of the phenotypic representation of alternative dimensional models, coherence and coverage of a respective domain, familiarity and ease of usage for clinicians, and relevance to treatment decisions. The broad domains that have been identified to date do appear to be well coordinated with the basic temperaments identified in childhood, providing thereby not only a classification of personality disorder with explicit childhood antecedents but also one that is well coordinated with basic science research on childhood development. Nevertheless, it will be important in future research to further articulate the transition from the normal to the abnormal variants of these personality dimensions through childhood development.

250

Dimensional Models of Personality Disorders

The broad domains also appear to have strong etic and emic cross-cultural support, providing the potential for a truly universal classification of personality disorder. Nevertheless, it will be important for future research to explore additional emic and etic methodologies, particularly studies including semistructured interview and peer assessments. It will also be helpful to specify more concretely how a universal method of diagnosis would be implemented (e.g., could or should cutoff points for diagnosis be common across different cultures and social settings). The broad domains are also well coordinated with the existing research explaining the common latent structure underlying the co-occurrence of personality with other mental disorders, particularly when this structure is understood hierarchically. Above the four or five broad domains appear to be two broad constructs of internalization and externalization, whereas beneath them are the more specific facets of normal and abnormal personality structure that will be of most immediate and direct clinical relevance. Additional research is needed though to help define more precisely how to integrate the classification of other mental disorders with structural models of personality. A dimensional model of personality disorder has considerable potential in improving the coverage of the maladaptive personality functioning seen in clinical practice. Nevertheless, it will be useful for future studies to document empirically that the inclusion of the additional maladaptive and adaptive personality traits within proposed models do in fact provide incremental validity for clinical assessments and increase the coverage of cases currently receiving a diagnosis of personality disorder not otherwise specified. An issue not well addressed currently by the existing research is how to use the dimensional models to make the clinical distinction between normal and abnormal personality functioning. A potential advantage of a dimensional model is flexibility in how cutoff points are used, but existing research has not yet demonstrated empirically how or where these cutoff points are best placed for different clinical decisions. Finally, additional research is needed on matters of clinical utility. A dimensional model of personality disorder would help address many of the problems that are currently problematic in clinical practice (e.g., inadequate coverage, unstable diagnostic boundaries, and excessive diagnostic co-occurrence), but it will also be important to demonstrate empirically whether and how a dimensional model of classification will improve ease of usage, reliability of unstructured clinical assessments, professional communication, and various treatment decisions. Field studies with clinicians in general practice would be informative, as well as survey and clinical vignette studies involving practicing clinicians. A consideration of utility in the context of general public health care concerns and costs will also be of particular importance. For instance, demonstrations of how a dimensional classification could facilitate the consideration of personality disorders in public health care coverage would be particularly informative.

Widiger et al.: Personality Disorder Research Agenda for DSM-V

251

References Achenbach TM: Empirically based assessment and taxonomy: applications to clinical research. Psychol Assess 7:261–274, 1995 Allik J: Personality dimensions across cultures. J Personal Disord 19:212–232, 2005 American Psychiatric Association: Diagnostic and statistical manual of mental disorders, 4th Edition. Washington, DC, American Psychiatric Association, 1994 American Psychiatric Association: Diagnostic and statistical manual of mental disorders, 4th Edition, Text Revision. Washington, DC, American Psychiatric Association, 2000 Ando J, Suzuki A, Yamagata S, et al: Genetic and environmental structure of Cloninger’s temperament and character dimensions. J Personal Disord 18:379–393, 2004 Andreasen NC: Brave New Brain: Conquering Mental Illness in the Era of the Genome. New York, Oxford University Press, 2001 Ashton MC, Lee K: A theoretical basis for the major dimensions of personality. Eur J Personality 15:327–353, 2001 Benet-Martinez V, John OP: Los cinco grandes across cultures and ethnic groups: multitrait multimethod analysis of the Big Five in Spanish and English. J Pers Soc Psychol 75:729–750, 1998 Blashfield RK, Intoccia V: Growth of the literature on the topic of personality disorders. Am J Psychiatry 157:472–473, 2000 Buss AH, Plomin R: Temperament: Early-Developing Personality Traits. Hillsdale, NJ, Lawrence Erlbaum, 1984 Caspi A, Harrington H, Milne B, et al: Children’s behavioral styles at age 3 are linked to their adult personality traits at age 26. J Pers 71:494–513, 2003 Caspi A, Roberts BW, Shiner RL: Personality development: stability and change. Annu Rev Psychol 56:453–484, 2005 Chess S, Thomas A: Temperament: Theory and Practice. New York, Brunner/Mazel, 1996 Clark LA: Personality disorder diagnosis: limitations of the five-factor model. Psychol Inq 4:100–104, 1993 Clark LA, Watson D: Temperament: a new paradigm for trait psychology, in Handbook of Personality. Theory and Research, 2nd Edition. Edited by Pervin L, John O. New York, Guilford, 1999, pp 399–423 Clark LA, Simms LJ, Wu KD, et al: Manual for the Schedule for Nonadaptive and Adaptive Personality (SNAP-2). Minneapolis, MN, University of Minnesota Press, in press Cloninger CR: A practical way to diagnosis personality disorders: a proposal. J Personal Disord 14:99–108, 2000 Cooper JR, Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology, 8th Edition. New York, Oxford University Press, 2003 Costa PT Jr, McCrae RR: Revised NEO Personality Inventory (NEO PI-R) and NEO FiveFactor Inventory (NEO-FFI) Professional Manual. Odessa, FL, Psychological Assessment Resources, 1992 Costa PT Jr, Widiger TA (eds): Personality Disorders and the Five Factor Model of Personality, 2nd Edition. Washington, DC, American Psychological Association, 2002 Depue RA, Collins PF: Neurobiology of the structure of personality: dopamine facilitation of incentive motivation and extraversion. Behav Brain Sci 22:491–569, 1999

252

Dimensional Models of Personality Disorders

Depue RA, Lenzenweger M: A neurobehavioral dimensional model, in Handbook of Personality Disorders. Theory, Research, and Treatment. Edited by Livesley WJ. New York, Guilford, 2001, pp 137–176 Derryberry D, Rothbart MK: Reactive and effortful processes in the organization of temperament. Dev Psychopathol 9:633–652, 1997 Digman JM: Personality structure: emergence of the five-factor model. Annu Rev Psychol 41:417–440, 1990 Eysenck HJ: The definition of personality disorders and the criteria appropriate for their description. J Personal Disord 1:211–219, 1987 Eysenck HJ, Eysenck SB: Manual of the Eysenck Personality Questionnaire. London, Hodder & Stoughton, 1975 Eysenck SBJ, Eysenck H, Barrett P: A revised version of the psychoticism scale. Personality and Individual Differences 6:21–29, 1985 First MB, Pincus HA, Levine JB, et al: Using clinical utility as a criterion for revising psychiatric diagnoses. Am J Psychiatry 161:946–954, 2004 Goldsmith HH, Camos JJ: Toward a theory of infant development, in The Development of Attachment and Affiliative Systems. Edited by Emde RN, Harmon RJ. New York, Plenum, 1982, pp 161–193 Grilo CM, Anez LM, McGlashan TH: The Spanish Language Version of the Diagnostic Interview for DSM-IV Personality Disorders: development and initial psychometric evaluation of diagnoses and criteria. Compr Psychiatry 44:154–161, 2003 Harkness AR, McNulty JL: The Personality Psychopathology Five (PSY-5): issue from the pages of a diagnostic manual instead of a dictionary, in Differentiating Normal and Abnormal Personality. Edited by Strack S, Lorr M. New York, Springer, 1994, pp 291–315 Hicks BM, Krueger RF, Iacono WG, et al: The family transmission and heritability of externalizing disorders: a twin study. Arch Gen Psychiatry 61:922–928, 2004 Hofstede G, McCrae RR: Personality and culture revisited: linking traits and dimensions of culture. Cross-Cult Res 38:52–88, 2004 Jang KL, Vernon PA, Livesley WJ: Behavioural-genetic perspectives on personality function. Can J Psychiatry 46:234–244, 2001 Jang KL, Livesley WJ, Angleitner A, et al: Genetic and environmental influences on the covariance of facets defining the domains of the five factor model of personality. Personality and Individual Differences 33:83–101, 2002 John OP, Donahue EM, Kentle RL: The Big Five Inventory—Versions 4a and 5. Berkeley, CA, University of California, Berkeley, 1991 Johnson JG, Cohen P, Smailes EM, et al: Adolescent personality disorders associated with violence and criminal behavior during adolescence and early adulthood. Am J Psychiatry 157:1406–1412, 2000 Kendler KS, Prescott CA, Myers J, et al: The structure of genetic and environmental risk factors for common psychiatric and substance use disorders in men and women. Arch Gen Psychiatry 60:929–937, 2003 Kluger AN, Siegfried Z, Ebstein RP: A meta-analysis of the association between DRD4 polymorphism and novelty seeking. Mol Psychiatry 7:712–717, 2002

Widiger et al.: Personality Disorder Research Agenda for DSM-V

253

Kohnstamm GA, Halverson CF, Mervielde I, et al (eds): Parental Descriptions of Child Personality. Developmental Antecedents of the Big Five? Mahwah, NJ, Lawrence Erlbaum, 1998 Kroeze WK, Kristiansen K, Roth BL: Molecular biology of serotonin receptors structure and function at the molecular level. Curr Top Mol Chem 2:507–528, 2002 Krueger RF: The structure of common mental disorders. Arch Gen Psychiatry 56:921–926, 1999 Krueger RF: Psychometric perspectives on comorbidity, in Defining Psychopathology in the 21st Century. DSM-V and Beyond. Edited by Helzer JE, Hudziak JJ. Washington, DC, American Psychiatric Publishing, 2002, pp 41–54 Krueger RF: Continuity of Axes I and II: toward a unified model of personality, personality disorders, and clinical disorders. J Personal Disord 19:233–261, 2005 Krueger RF, Tackett JL: Personality and psychopathology: working toward the bigger picture. J Personal Disord 17:109–128, 2003 Krueger RF, Caspi A, Moffitt TE, et al: The structure and stability of common mental disorders (DSM-III-R): a longitudinal-epidemiological study. J Abnorm Psychol 107:216– 227, 1998 Krueger RF, Chentsova-Button YE, Markon KE, et al: A cross-cultural study of the structure of comorbidity among common psychopathological syndromes in the general health care setting. J Abnorm Psychol 112:437–447, 2003 Lambert MJ, Anderson EM: Assessment for the time-limited psychotherapies, in American Psychiatric Press Review of Psychiatry, Vol 15. Edited by Dickstein LJ, Riba MB, Oldham JM. Washington, DC, American Psychiatric Press, 1996, pp 23–42 Lehman AF, Alexopoulos GS, Goldman H, et al: Mental disorders and disability: time to reevaluate the relationship? in A Research Agenda for DSM-V. Edited by Kupfer DJ, First MB, Regier DA. Washington, DC, American Psychiatric Association, 2002, pp 201–218 Livesley WJ: Conceptual and taxonomic issues, in Handbook of Personality Disorders: Theory, Research, and Treatment. Edited by Livesley WJ. New York, Guilford, 2001, pp 3–38 Livesley WJ: Diagnostic dilemmas in classifying personality disorder, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 153–190 Livesley WJ: Behavioral and molecular genetic contributions to a dimensional classification of personality disorder. J Personal Disord 19:131–155, 2005 Livesley WJ, Jackson D: Manual for the Dimensional Assessment of Personality Pathology—Basic Questionnaire. Port Huron, MI, Sigma Press, in press Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998 Loranger AW, Sartorius N, Andreoli A, et al: The International Personality Disorder Examination. The World Health Organization/Alcohol, Drug Abuse, and Mental Health Administration international pilot study of personality disorders. Arch Gen Psychiatry 51:215–224, 1994 Lynam DR, Derefinko KJ: Psychopathy and personality, in Handbook of Psychopathy. Edited by Patrick C. New York, Guilford, in press

254

Dimensional Models of Personality Disorders

Lynn R, Martin T: National differences for thirty-seven nations in extraversion, neuroticism, psychoticism, and economic, demographic and other correlates. Personality and Individual Differences 19:403–406, 1995 Markon KE, Krueger RF, Watson D: Delineating the structure of normal and abnormal personality: an integrative hierarchical approach. J Pers Soc Psychol 88:139–157, 2005 McCrae RR: NEO PI-R data from 36 cultures, in The Five-Factor Model of Personality Across Cultures. Edited by McCrae RR, Allik J. New York, Kluwer Academic, 2002, pp 105–125 McCrae RR, Terracciano A, Khoury B, et al: Universal features of personality traits from the observer’s perspective: data from 50 cultures. J Pers Soc Psychol 88:547–561, 2005 Merikangas KR: Implications of genetic epidemiology for classification, in Defining psychopathology in the 21st Century: DSM-V and Beyond. Edited by Helzer JE, Hudziak JJ. Washington, DC, American Psychiatric Publishing, 2002, pp 195–209 Mervielde I, De Fruyt F: Assessing children’s traits with the Hierarchical Personality Inventory for Children, in Big Five Assessment. Edited by De Raad B, Perugini M. Seattle, WA, Hogrefe & Huber, 2002, pp 129–146 Mervielde I, De Clercq B, De Fruyt F, et al: Temperament, personality, and developmental psychopathology as childhood antecedents of personality disorders. J Personal Disord 19:171–201, 2005 Miller JD, Lynam DR, Widiger TA, et al: Personality disorders as extreme variants of common personality dimensions: can the five-factor model adequately represent psychopathy? J Pers 69:253–276, 2001 Miller TR: The psychotherapeutic utility of the five-factor model of personality: a clinician’s perspective. J Pers Assess 57:415–433, 1991 Millon T: Millon Index of Personality Styles Manual. San Antonio, TX, Psychological Corporation, 1994 Millon T, Davis RD, Millon CM, et al: Disorders of Personality: DSM-IV and Beyond. New York, John Wiley & Sons, 1996 Muthen BO: Beyond SEM: General latent variable modeling. Behaviormetrika 29:81–117, 2002 O’Connor BP, Dyce JA: A test of models of personality disorder configuration. J Abnorm Psychol 107:3–16, 1998 Paris J: The classification of personality disorders should be rooted in biology. J Personal Disord 14:127–136, 2000 Paris J: Personality Disorders Over Time: Precursors, Course, and Outcome. Washington, DC, American Psychiatric Press, 2003 Paris J: Neurobiological dimensional models of personality: a review of the models of Cloninger, Depue, and Siever. J Personal Disord 19:156–170, 2005 Roberts BW, DelVecchio WF: The rank-order consistency of personality traits from childhood to old age: a quantitative review of longitudinal studies. Psychol Bull 126:3–25, 2000 Samuel D, Widiger TA: Differentiating normal and abnormal personality: the DSM perspective, in Differentiating Normal and Abnormal Personality, 2nd Edition. Edited by Strack S, Lorr M. New York, Springer, 2006, pp 165–183

Widiger et al.: Personality Disorder Research Agenda for DSM-V

255

Sanderson CJ, Clarkin JF: Further use of the NEO PI-R personality dimensions in differential treatment planning, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 351–376 Saucier G, Goldberg LR: Lexical studies of indigenous personality factors: premises, prospects, and products. J Pers 69:847–880, 2001 Saulsman LM, Page AC: The five-factor model and personality disorder empirical literature: a meta-analytic review. Clin Psychol Rev 23:1055–1085, 2004 Schmitt DP, Allik J: Simultaneous administration of the Rosenberg Self-Esteem scale in 53 nations: exploring the universal and culture-specific features of global self-esteem. J Pers Soc Psychol 89:623–642, 2005 Schmitt DP, Alcalay L, Allik J, et al: Universal sex differences in the desire for sexual variety: tests from 52 nations, 6 continents, and 13 islands. J Pers Soc Psychol 85:85–104, 2003 Schinka JA, Letsch EA, Crawford FC: DRD4 and novelty seeking: results of meta-analyses. Am J Med Genet 114:643–648, 2002 Sen S, Burmeister M, Ghosh D: Meta-analysis of the association between a serotonin transporter polymorphism (5-HTTLPR) and anxiety-related personality traits. Am J Med Genet 127B:85–89, 2004 Shedler J, Westen D: Dimensions of personality pathology: an alternative to the five-factor model. Am J Psychiatry 161:1743–1754, 2004 Shiner RL: How shall we speak of children’s personalities in middle childhood? A preliminary taxonomy. Psychol Bull 124:308–332, 1998 Shiner RL, Caspi A: Personality differences in childhood and adolescence: measurement, development, and consequences. J Child Psychol Psychiatry 44:2–32, 2003 Siever LJ, Davis KL: A psychobiological perspective on the personality disorders. Am J Psychiatry 148:1647–1658, 1991 Sprock J: Dimensional versus categorical classification of prototypic and nonprototypic cases of personality disorder. J Clin Psychol 59:991–1014, 2003 Stone M: Treatment of personality disorders from the perspective of the five-factor model, in Personality Disorders and the Five-Factor Model, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 405–430 Tellegen A: Multidimensional Personality Questionnaire. Manual for Administration, Scoring, and Interpretation. Minneapolis, MN, University of Minnesota Press, in press Trull TJ: Dimensional models of personality disorder: coverage and cutoffs. J Personal Disord 19:262–282, 2005 Trull TJ, Durrett CA: Categorical and dimensional models of personality disorder. Annu Rev Clin Psychol 1:355–380, 2005 Trull TJ, Widiger TA, Lynam DR, et al: Borderline personality disorder from the perspective of general personality functioning. J Abnorm Psychol 112:193–202, 2003 Tyrer P (ed): Personality Disorders: Diagnosis, Management, and Course, 2nd Edition. London, Arnold, 2000 Verheul R: Clinical utility of dimensional models of personality pathology. J Personal Disord 19:283–302, 2005

256

Dimensional Models of Personality Disorders

Verheul R, Widiger TA: A meta-analysis of the prevalence and usage of the personality disorder not otherwise specified (PDNOS) diagnosis. J Personal Disord 18:309–319, 2004 Wakefield JC, First MB: Clarifying the distinction between disorder and nondisorder: confronting the overdiagnosis (false-positives) problem in DSM-V, in Advancing DSM: Dilemmas in Psychiatric Diagnosis. Edited by Phillips KA, First MB, Pincus HA. Washington, DC, American Psychiatric Association, 2003, pp 23–55 Watson D, Clark LA, Harkness AR: Structures of personality and their relevance to psychopathology. J Abnorm Psychol 103:18–31, 1994 Watson D, Wiese D, Vaidya J, et al: The two general activation systems of affect: structural findings, evolutionary considerations, and psychobiological evidence. J Pers Soc Psychol 76:820–838, 1999 Westen D, Arkowitz-Westen L: Limitations of Axis II in diagnosing personality pathology in clinical practice. Am J Psychiatry 155:1767–1771, 1998 Westen D, Shedler J: A prototype matching approach to diagnosing personality disorders: toward DSM-V. J Personal Disord 14:109–126, 2000 Widiger TA: Personality disorder and Axis I psychopathology: the problematic boundary of Axis I and Axis II. J Personal Disord 17:90–108, 2003 Widiger TA, Costa PT Jr: Five-factor model personality disorder research, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 59–87 Widiger TA, Samuel DB: Evidence-based assessment of personality disorders. Psychol Assess 17:278–287, 2005 Widiger TA, Simonsen E: Alternative dimensional models of personality disorder: finding a common ground. J Personal Disord 19:110–130, 2005a Widiger TA, Simonsen E: Personality disorder research agenda for DSM-V. J Personal Disord 19:315–338, 2005b Widiger TA, Costa PT Jr, McCrae RR: A proposal for Axis II: diagnosing personality disorders using the five factor model, in Personality Disorders and the Five-Factor Model of Personality, 2nd Edition. Edited by Costa PT Jr, Widiger TA. Washington, DC, American Psychological Association, 2002, pp 431–456 Wiggins JS: Paradigms of Personality Assessment. New York, Guilford, 2003 Zimmerman M, Mattia JI: Differences between clinical and research practices in diagnosing borderline personality disorder. Am J Psychiatry 156:1570–1574, 1999 Zuckerman M: Zuckerman-Kuhlman Personality Questionnaire (ZKPQ): an alternative five-factorial model, in Big Five Assessment. Edited by de Raad B, Perugini M. Kirkland, WA, Hogrefe & Huber, 2002, pp 377–397

INDEX Page numbers printed in boldface type refer to tables or figures. Acceptable behavior, culture-bound, 144–145 Accuracy, of dimensional models, 206–207 ACD-IV (Assessment of DSM-IV Personality Disorders), 97 Achenbach System of Empirically Based Assessment (ASEBA), 95–96 Achievement alignment in dimensional models, 10, 11, 12 childhood models, 93–94 Acquiescence bias, cross-cultural studies, 126 Activity alignment in dimensional models, 4, 6, 12, 14, 41 childhood models, 91, 93 research agenda, 238–239 temperament models, 87–88 Adaptive traits consistent delineation, 25–26 cutoff points, 178–179, 201, 229 dimensional models diagnosis, 179–181 integration into dimensional models, 16–17, 23, 176 research agenda, 234–235, 242–243, 245–246, 250 subtlety of diagnosis, 209–211, 247–248 Adolescence, antecedents in, xxix, 3, 85–86

antisocial personality disorder, 42 broad associations, 104, 105, 106, 112 dimensional models, 95–99 of personality, 88–92 FFM and, 101–102, 103 relationship between, 99–100, 101 personality disorders development, 111–115 taxonomy proposal, 92–95 temperament as, 86–88 Adoptees, heritability studies, 56 ADP-IV, in adolescent personality studies, 102 Aesthenic personality disorder, 31 Affective instability, 5, 6–7, 8 Affect regulation classification schemes, 178, 191–192 neurobiological model, 66–67 Affiliation, 5, 6, 65 Age at onset, 150 Agency alignment in dimensional models, 5, 6, 41 neurobiological model, 65 Aggression alignment in dimensional models, 5, 6–7, 13, 14 alternative placement, 24–25 monoamine oxidase A gene associations, 42, 57 neurobiological model, 66–67

257

258

Dimensional Models of Personality Disorders

Agreeableness alignment in dimensional models, 14, 15, 40 childhood models, 89–92, 94, 112–113 cross-cultural studies, 121–122, 134–136 FFM vs. personality disorders metaanalysis, 101–102, 103 neurobiological model, 65 psychopathology associations, 104, 105, 106 Alcohol dependence, externalizing construct, 155–157, 159 Allik, Jüri, cross-cultural models, 117–128 diagnostic applications, 143–146 historical perspectives, 139–144 lexical approach, 133–136, 240 Altruism, 13, 14, 91 Ambitiousness, 10, 11, 12 American Psychiatric Association (APA) personality disorder categories, xxv, xxix Steering Committee, 233 Amygdala, 65 Anankastic trait, 81 Anger childhood models, 92 cross-cultural studies, 134–135 temperament models, 88 Type A personality and, 169–170 Antagonism alignment in dimensional models, 5, 6–7, 8–9, 29, 31 alternative placement, 24–25 childhood models, 94 cross-cultural studies, 136 lower-order traits, 13, 14–15, 15 research proposal, 183 Antecedents, of psychopathology. See also Childhood antecedents developmental, xxix, 3, 85–99 for diagnostic validity, 205

Antisocial personality disorder dimensional framework, xxiii, 12, 30–31 externalizing construct, 155–159 gene–environment interaction, 42 heritability, 56–57 recurrent psychosis and, 79–80 Anxiety alignment in dimensional models, 7, 40 childhood models, 91, 93 cross-cultural studies, 134 neurobiological model, 65–66 psychobiological processes, 74, 75 Anxiety disorders factor analysis, 2 genetics, 156 internalizing construct, 154–159 misclassification, 168 Nottingham study, 78–79, 80 Anxiousness, 6, 8–9 Approach, neurobiological models, 62 Arousal, temperament models, 88 ASEBA (Achenbach System of Empirically Based Assessment), 95–96 Ashton, Michael, 133 Assertiveness alignment in dimensional models, 12, 14 childhood models, 91–92 Assessment of DSM-IV Personality Disorders (ACD-IV), 97 Assessment tools, clinical utility and, 222–223 Attachment, insecure alignment in dimension models, 12, 13, 14–15, 15 consistent placement in hierarchy, 25–26 Attachment theory, 114, 191 Avoidant personality disorder, xxv, 168

Index Axis I disorders childhood vs. adult, 86, 96–97 adaptive vs. maladaptive, 104, 105, 106 exacerbation vs. remissions, 31 personality disorders continuum, 2–3, 57–58, 95, 149–150 clinical utility and, 227–230 joint structure, 150–154 novel unified models, 154–159 old distinctions, 163–165 research agenda, 241–244 trait vs. type classification, 167–170 Axis II disorders, 97, 135. See also Personality disorders clinical disorders continuum, 2–3, 57–58, 95, 149–150 clinical utility and, 227–230 joint structure, 150–154 novel unified models, 154–159 old distinctions, 163–165 research agenda, 241–244 trait vs. type classification, 167–170 Five-Factor Model replacement, 195–197 Axis VI, for genotypic information, 42 Behavior acceptable, as culture-bound, 144–145 classification schemes, 178, 190 treatment directed at, 212–213, 246 Behavioral genetics, 39–50 molecular studies vs., 42, 50 overview, xxix, 3, 39–40, 49–50 personality approaches, 41–42 personality disorder classification, 40–41, 205 personality structure and, 42–44 phenotype–genotype congruence, 45–46 primary traits influenced by, 40, 47–48, 236

259 problems and pitfalls, 55–58 research agenda, 235–236, 249 secondary trait definitions, 40, 46, 65 structure perspectives, 42–44 taxonomic research strategy and, 43, 48–49 trait covariance and, 43, 44–45, 50, 56 Behavioral inhibition, neurobiological models, 62 Behavior therapy, 212 dialectical, 212–213, 246 Beliefs toward self, 192 Bell, Carl C., 199 Benevolence, childhood models, 91–93, 100, 101 Between-person models, 73–74 BFI. See Big Five Inventory (BFI) Bias cross-cultural studies, 123–124, 127, 241 diagnostic prototypes, 228 Big Five Inventory (BFI) childhood personality/temperament, 89–94, 105, 112 cross-cultural studies, 120–121, 123, 128 joint structural model, 151 Bipolar system, 29, 222 lower-order traits, 10, 12, 26 Body dysmorphic disorder, 168 Borderline personality disorder arbitrary diagnostic boundaries, xxviii dimensional framework, 30–31 heterogeneous components, xxvii ICD-10 vs. DSM-IV, xxv Brain potentials, self-directedness correlations, 64 Brain systems genetic plasticity, 205 neurobiological models, 62, 65, 67–68 research agenda, 200, 237 Buss-Plomin model, of temperament, 87–88, 93

260

Dimensional Models of Personality Disorders

Caseness clinical utility, 30, 34, 228–229 diagnostic applications, 192–193, 223–224 examples, 179–180 research applications, 209, 213, 245, 250 Catastrophize, 191 Categorical model, xxv dimensional model alternative, xxix–xxx, 1–17 integration with dimensional models, 3–15, 152 limitations, xxvi–xxix, 1, 169, 205 reorganization, 2, 29, 243–244 revision, 2, 29, 31 Task Force, 33–34 CBCL. See Child Behavior Checklist (CBCL) CBQ (Child Behavior Questionnaire), 87 Change measuring clinical, 223 ten commandments of implementing, 201–202 Character, 74, 175 Child Behavior Checklist (CBCL), 96, 104, 105 DIPSI comparisons, 98–100, 101 Child Behavior Questionnaire (CBQ), 87 Childhood antecedents, 85–86 antisocial personality disorder, 42 broad associations, 104, 105, 106, 112 dimensional models, 95–99 personality as, xxix, 3, 88–92 FFM and, 101–102, 103 relationship between, 99–100, 101 personality disorders development, 111–115 research agenda, 238–239, 249 taxonomy proposal, 92–95 temperament as, 86–88 Clark, Lee Anna, 23, 174

Classification nomenclature alternative dimensional approach, xxix–xxx, 1–17 construct component, 40–41 content validity goals, 190–192, 229 coverage concerns, 171–179, 181–182, 200–201 current, xxv–xxix, 168–169 cutoff points, 200–201, 244–246, 250 excessive co-occurrence, xxvi–xxvii, 163–165, 228, 243, 250 heterogeneity within diagnoses, xxvii, 30, 146, 169, 221 inadequate coverage, xxvii, 199–200, 205, 244–247 inadequate scientific base in, xxviii–xxix, 146 personality disorder definition, 40 personality disorder severity in, 77–82, 79 research agenda, 241–246 Task Force, 33–34 Clercq, Barbara de, 85 Clinical case reports. See Caseness Clinical decision making, 211 clinical utility and, 210, 221–222, 228, 247 dimensional models and, 210–213 research agenda, 228–229, 247–249 Clinical disorders, 149 Axis II disorders continuum, 2–3, 57–58, 95, 149–150 clinical utility and, 227–230 joint structure, 150–154 novel unified models, 154–159 old distinctions, 163–165 research agenda, 241–244 trait vs. type classification, 167–170 Clinical improvement. See also Treatment response conceptualization and assessment, 205–206

Index Clinical spectra models. See Spectrum model Clinical utility, of dimensional models, 203–215 accuracy, 206–207 assessment tools and, 222–223 clinical decision making, 210–213, 221–222, 228, 247 commentary on, 219–224, 227–230 components, 206–213 definition, 204, 213 diagnosis subtlety, 209–211 diagnostic validity vs., 204–206 doctor–patient relationship and, 220–221 hybrid, 209, 214, 228 interrater reliability, 34, 208–209, 247 measuring change, 223 for personality pathology, 227–230 polarity quantification, 10, 12, 26, 29, 222 practice perspectives, xxix–xxx, 2–3, 15, 47, 82 professional communication, 207–208, 224, 247 research agenda, 224, 246–249 summary overview, 203–204, 213–215 theory role, 219–220 treatment outcomes and, 223 user acceptability, 206–207 Cloninger’s Seven-Factor model, 62–64, 141 coverage and cutoff points, 175 diagnostic examples, 180 Closedness, 6, 9, 29. See also Openness to Experience Cluster A disorders, 79 Cluster analysis, model-based, 153 Cluster B disorders, 76 Cluster C disorders, 78–79 Coding, DSM-IV-TR as tool, 30–31

261 Cognition alignment in dimensional models, 7, 9 childhood psychopathology, 95 classification schemes, 178, 190–192, 223 neurobiological model, 66 Collaboration, international, for research, 200, 224 Communication, professional, with dimensional models, 207–208, 224, 247 Communion, 5, 6 Comorbidities novel models, 154–155, 191 old distinctions, 163–165, 228 patterns, 30, 34, 150, 154 phenotypic structure, 155–156 Compassion, 13, 14 Competitiveness, Type A personality and, 169 Complex (diffuse) personality disorder, 78–79, 79 Compliance alignment in dimensional models, 5, 6–7, 9, 29 alternative placement, 24–25 childhood models, 91 lower-order traits, 13, 14–15, 15 Compulsivity alignment in dimensional models, 40 psychopathology associations, 104, 105, 106 research proposal, 183 stability vs. change, 206 Concepts. See Theory Concurrent validity, of DIPSI, 98–99 Conduct disorders. See also Antisocial personality disorder externalizing spectrum, 157 genetics, 47, 56, 155–156

262 Conscientiousness alignment in dimensional models, 6, 10, 11, 12 alternative placement, 25, 40 childhood models, 89–91, 92–94, 100, 101, 112–113 cross-cultural studies, 121, 124, 134 FFM vs. personality disorders metaanalysis, 101–102, 103 neurobiological model, 65 psychopathology associations, 104, 105, 106 research proposal, 183 stability vs. change in, 206 as trait vs. type, 169 Consensus dimensional model, 23–26, 31, 214 ICD-10 and DSM-IV categories, 33 Conservatism, cross-cultural studies, 126 Constraint alignment in dimensional models, 6, 8–9, 29 alternative placement, 24–25 childhood models, 92–94 lower-order traits, 10, 11, 12 nonaffective, 65 research proposal, 183 Construct validity dimensional models, 30–31, 49, 224 personality disorder classification, 40–41, 48–49 Content validity, of DIPSI, 98 Control alignment in dimensional models, 10, 11, 12 childhood models, 89 temperament models, 87–88 Conventionality, childhood models, 94 Co-occurrence, diagnostic excessive, xxvi–xxvii, 163–165 research agenda, 228, 243, 250

Dimensional Models of Personality Disorders Cooperativeness alignment in dimensional models, 5, 6–7, 9 childhood models, 94 genetic variations, 45 psychobiological processes, 74, 75 Coping, as maladaptive, 165 Correlates/correlations, for diagnostic validity, 25–26, 205 Corticolimbic-striatal-thalamic network, 65 Costa, Paul, Jr., 195 Countertransference, 212 Covariance, trait children and adolescents, 112 culture-related personality, 118–119, 127 gene–environment variables, 43, 50, 56 genetic contributions, 44–45, 47 Coverage classification inadequacy, xxvii, 199–200, 205, 244–247 by dimensional models, 171–176, 200–201, 250 research recommendations, 181–182, 183–184, 184 Creativity, childhood models, 91, 94 Cross-cultural models, of personality, 3, 117–128 Big Five Inventory, 120–121, 123, 128 childhood perspectives, 87, 90, 94 culture-level factor structure, 122–123 diagnostic applications, 143–146, 200–201 dimension challenges, 122–126 expert opinions, 123–124 large-scale studies, 119–122 lexical approach, 134–136, 240 measurement scale correlations, 127–128 research agenda, 239–241, 249–250

Index research review, 117–118, 127 stereotypes, 123–124 substance vs. style, 125–126 trait comparison problems, 118–119 Cultural dimensions, of personality. See Cross-cultural models Curiosity, childhood models, 91, 94 Cutoff points classification, xxvii, 171–176, 205, 244–247 dimensional models, 171–172, 176–181, 250 Five-Factor Model handling, 196–197 research recommendations, 182, 183–184, 184, 200–201 Cyclothymic disorder, 31 Dangerous personality disorder, 81 DAPP-BQ. See Dimensional Assessment of Personality Psychopathology— Basic Questionnaire (DAPP-BQ) Data quality index, cross-cultural studies, 125 Data sets, for identifying primary traits, 184 Decision making, clinical, 211 clinical utility and, 210, 221–222, 247 dimensional models and, 210–213 research agenda, 228–229, 247–249 Defense mechanisms, 114, 220, 227 Degeneration, evolution vs., 140 Deliberation, 10, 11, 12 Delusions, 191 Dependency alignment in dimension models, 12, 13, 14–15, 15, 31 consistent placement in hierarchy, 25–26 Dependent personality disorder, xxvii, xxviii, 12, 30 Depressive disorders genetics, 155–156 internalizing construct, 156–159

263 Nottingham study, 78–79, 80 old distinctions, 165 Depue’s model, 64–66 Detachment, 12–13, 14 Developmental model, of psychopathology. See also Childhood antecedents antisocial personality disorder, 42 broad associations, 104, 105, 106, 112, 249–250 commentary, 111–115 dimensional models, 95–99 future research directions, 114–115 overview, xxix, 3, 85–86 personality and, 88–92, 112 common structure, 112–114 FFM and, 101–102, 103 relationship between, 99–100, 101 stability vs. change, 31, 97, 106, 114, 150, 165, 205, 229 taxonomy proposal, 92–95 temperament and, 86–88, 111 Deviant behavior as culture-bound, 144 statistical, 177, 197 Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR) alternative dimensional categories, xxix–xxx. See also Dimensional models categorical personality disorders, xxv–xxvi. See also Categorical model childhood psychopathology assessment tools, 96–97 clinical utility, 206–215 consensus on, 33–34 criteria sets, xxvi. See also Diagnostic criteria sets diagnostic validity, 203–206

264

Dimensional Models of Personality Disorders

Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR) (continued) dissatisfaction with current, xxvi–xxix, 168–169, 171–172 ICD-10 categories vs., xxv–xxvi, 223 illness defined by, 30–31, 149 personality continuum, 2–3, 57–58, 95, 149–150 personality integration into, 154–159 implementing change for, 201–202 internalizing vs. externalizing syndromes organization, 156–159 separate axis, xxv–xxvi sixth axis, 42 Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, (DSM-V), 33 Diagnostic confidence, 207 Diagnostic co-occurrence excessive, xxvi–xxvii, 163–165 research agenda, 228, 243, 250 Diagnostic criteria sets clinical utility, 206–215 coverage and cutoffs, 171–179, 199–201 cross-cultural, 143–146 dimensional domains, 35, 179–181 clinical utility, 227–230 DSM-IV, xxvi behaviorally specific, 4, 29 cutoff points, 244–246, 250 excessive co-occurrence, xxvi–xxvii, 163–165, 228, 243, 250 heterogeneity within, xxvii, 30, 146, 169, 221 inadequate coverage, xxvii, 199, 205, 244–247 polythetic, xxix revision, 2, 29, 31

threshold differences, 34–35 unstable boundaries, xxviii, 13 higher-order traits, 3–5, 6–7, 8–9 high inference level, 208–209 ICD-10, 33–34 for infants and preschool children, 95 interpersonal nature, 165 lower-order traits, 10, 11, 12–13, 14–15, 15 misclassification, 168 reorganization, 2, 29, 243–244 research agenda, 241–246 validity, 203–204 elements of, 204–206 Diagnostic subtlety, dimensional models, 209–211, 247–248 Diagnostic table of the elements, 159 Diagnostic validity correlates for, 25–26, 205 of DSM-IV-TR, 203–204 elements of, 204–206 Diaspora Meeting, 201 Diffidence, 13, 15, 25–26 Dimensional Assessment of Personality Psychopathology—Basic Questionnaire (DAPP-BQ), 2 clinical utility, 206, 208, 210, 214–215, 222 consensus on structure, 24–26 coverage and cutoff points, 173–174 diagnostic examples, 180 DIPSI comparisons, 98, 104 heritability studies, 43, 45, 47 higher-order domains, 4, 6, 8–9 joint structural model, 151–153 lower-order traits, 10, 11, 12–13, 14–15, 15 Dimensional models, 1–17 arguments favoring, 205 categorical model vs., xxv–xxix change conceptualization, 205–206

Index clinical utility, xxix–xxx, 2–3, 15, 47, 82, 203–215 conclusive recommendations, 15–17 consensus personality trait structure, 23–26, 31, 214 construct component, 40–41 coverage and cutoffs, 171–179, 199–201, 205 research recommendations, 181–182 framework, 29–31 functional domains, xxix, 2–3, 29, 40, 191–193 higher-order domains, 3–5, 6–7, 8–9 integration, 3–15, 152, 176 limitations, 223–224 lower-order domains, 10, 11, 12–13, 14–15, 15 neurobiological, 61–68 number of dimensions vs. specificity level, 164 personality disorder definition, 40 proposals, 2–3 of psychopathology, 95–99 research agenda, 30, 214–215, 234–235 proposal, 182–185, 183–184 validity, xxix–xxx, 2–3, 15 construct vs. evidence, 30–31, 49, 224 Dimensional Personality Symptom Item (DIPSI) Pool, for childhood psychopathology, 95–99 assembling, 97 CBCL comparisons, 98–100, 101 HiPIC comparisons, 99–100, 101 psychometric properties, 98 taxonomy, 95–97 validity, 98–99 Dimensional profiling, 3, 246 clinical utility, 209, 214, 228–230 consensus on structure, 24, 214

265 heritability studies, 43, 48 higher-order domains, 4, 6 lower-order traits, 10, 11, 12–13, 14, 15 DIPSI. See Dimensional Personality Symptom Item (DIPSI) Pool Disability, independent evaluation, 179–180 Disinhibition alignment in dimensional models, 5, 6–7, 8, 24 lower-order traits, 10, 11, 12 neurobiological model, 65–66 Distress, 93, 157, 191 Diversity experiment, of Work Group, xxix, 200 Doctor–patient relationship, 220–221 Dominance childhood models, 91–92 genetic, 56 social alignment in dimensional models, 5, 6, 41 neurobiological model, 65 Dopamine extraversion related to, 65–66 novelty-seeking associations, 41–42 Dopamine receptor gene, neurobiological models, 41–42, 63–64 DRD4 gene, 41–42, 63–64 Drug dependence. See Substance use disorders DSM-IV-TR. See Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR) DSM-V (Diagnostic and Statistical Manual of Mental Disorders, 5th Edition), 33 Dutch Developmental Profile, 223 Dutifulness, 10, 11, 12

266 Dysfunction classification goals, 190–192, 229 clinically significant, 177–178, 181, 244–246 cutoffs, 176–177, 200–201 diagnosing, research proposal, 182, 183–184, 184 perspective on, 177–179 Dysthymic disorder Nottingham study, 78–79, 80 old distinctions, 165 Eating behaviors/disorders, 190 ECA (Epidemiologic Catchment Area), 178 Ecological factor analysis, 122 Economic prosperity, cross-cultural studies, 125–126 Ecotone, 200 Education level, cross-cultural studies, 123, 126 Effortful control, temperament models, 87–88 Egocentrism, childhood models, 91, 94 Ego control, 89 Ego resiliency, 89 Emotional Dysregulation alignment in dimensional models, 6, 8–9, 29 cross-cultural studies, 135–136 lower-order traits, 10, 15 PAS concordance, 81–82 research proposal, 183 Emotionality childhood models, 89, 94, 112 classification schemes, 178, 190, 223 cross-cultural studies, 135–136 research agenda, 238–239 temperament models, 87–88, 112 Emotional Stability alignment in dimensional models, 6, 8–9, 15, 29

Dimensional Models of Personality Disorders childhood models, 89–91, 93, 100, 101 lower-order traits, 10 psychopathology associations, 104, 105, 106 Empathy, 13, 14 Empirical data for diagnostic validity, 195, 197, 205, 207 dimensional models based on, 30, 49, 207 research agenda, 224, 242–243, 245, 247 taxonomic research, 167–168 for treatment response, 228–230 Encoding, of incentive stimuli, 65 Energy, childhood models, 91, 93 Entitlement, 14, 24 Environment–gene interaction. See Gene–environment interaction Epidemiological surveys, psychiatric comorbidity observed, 154 Western vs. Chinese, 143–144 Epidemiologic Catchment Area (ECA), 178 Episodes, phenomena of, 165, 168, 243 clinical utility, 221, 228 Epistasis, 56 EPP. See Eysenck Personality Profiler (EPP) EPQ. See Eysenck Personality Questionnaire (EPQ) Evidence-based validity diagnostic, 195–197, 205, 207 dimensional models, 30, 49, 207, 224 prognosis data, 228–230 research agenda, 224, 242–243, 245, 247 taxonomic research, 167–168 Evolutionary theory, 140, 191 Excitement-seeking, 12, 14 Exhibitionism, 12, 14

Index Expert opinion cross-cultural studies, 123–124 forensics, 141 Exploitation of others, cross-cultural studies, 135 Exploratory excitability, 12, 14 Expressiveness, childhood models, 91 Externalization clinical domain level, 23–24 genetics, 155–156 as higher-order trait, 3–4 psychopathology associations, 104, 105, 106 novel models, 155–159 research agenda, 234, 239 Extravagance, 12, 14 Extraversion, 169 alignment in dimensional models, 4–5, 6, 8–9, 12, 29 Alpha vs. Beta factors, 25 alternative placement, 24–25, 40–41 childhood models, 89–94, 100, 101, 112–113 cross-cultural studies, 120–122, 125, 127 FFM vs. personality disorders metaanalysis, 101–102, 103 heritability, 48 lower-order traits, 12–13, 14 neurobiological model, 65–66 psychobiological processes, 74, 75 psychopathology associations, 104, 105, 106 research proposal, 183, 238–239 temperament models, 87–88, 92–93 treatment decisions based on, 212 Eysenck Personality Profiler (EPP), 3 consensus on structure, 24 heritability studies, 43, 48 higher-order domains, 4, 6 lower-order traits, 10, 11, 12–13, 14, 15

267 Eysenck Personality Questionnaire (EPQ), 3 consensus on structure, 24 cross-cultural studies, 119–121, 239 higher-order domains, 4–5, 6 joint structural model, 151 personality construct variables, 40 Facet-level scales joint structural model, 152–154 professional communication about, 208 research agenda, 243–244 subtlety of diagnosis, 210–211 Factor analyses Child Behavior Questionnaire, 87 childhood psychopathology, 97 cross-cultural studies, 119–120, 122, 134, 136 ecological, 122 lower-order structure, 210–211, 214 in PAS, 81–82 personality disorders, 2, 9, 61, 168 limitations, 223–224 personality variations and, 153, 210 Factor loading, 153 Familial diseases, 57. See also Genetics; Heritability Fear/fearfulness alignment in dimensional models, 7, 11 cross-cultural studies, 134 neurobiological model, 65 temperament models, 88 Feelings, classification schemes, 192 FFM. See Five-Factor Model (FFM) Field studies, 250 Fight–flight response, neurobiological models, 62 Five-Factor Model (FFM), 8–9, 23, 24 Axis II replacement with, 195–197 change conceptualization, 205–206

268

Dimensional Models of Personality Disorders

Five-Factor Model (FFM) (continued) childhood applications, 89–90, 92–94, 104, 105, 112 clinical utility, 205–212, 214 coverage and cutoff points, 175–176 cross-cultural studies, 118, 120, 127, 134–135 diagnostic examples, 181 genetic studies, 40, 45 joint structural model, 151–154, 159 meta-analyses of adult disorders, 101–102, 103 neurobiological models, 61–62, 64–65 related trait models and, 189–193 research agenda, 242–243 5-HTT gene, 63, 237 Forensics, 141 DSM-IV-TR as tool, 30–31 Freedom of choice, neurobiological limitations, 141 Fruyt, Filip de, 85 Functional domains, of dimensional models, xxix, 2–3, 29, 40, 191–193 GAF (Global Assessment of Functioning Scale), 179, 181, 185, 192, 197, 227 Gamma-aminobutyric acid (GABA), 68, 237 Gender-specific phenotypes, 50 neurobiological models, 63–64 Gene–environment interaction comorbidities and, 155–156 fruitlessness, 168 heritability, 42, 43, 48, 50, 55, 112 nonadditive, 56–57 personality disorder classification and, 205 Genetic additivity, 56 Genetic factor scoring, of primary traits, 48, 236 Genetic loading, 50, 55, 64

Genetics. See also Behavioral genetics common mental disorders, 155–156 episodic phenomena, 165 molecular, 42, 50 neurobiological models, 63 of personality, xxix, 3, 39–50, 200, 205 research agenda, 235–236, 249 spectrum model, 164 Genotype nonadditive interplay, 56–57 nosology implications, 45–46, 49 research agenda, 236 Gestalt therapy, 212 Global Assessment of Functioning Scale (GAF), 179, 181, 185, 192, 197, 227 Glutamine, 68 Goal-setting, treatment, 212–213 Goldsmith-Campos model, of temperament, 88, 93 Gratification, 223 Greed, cross-cultural studies, 135 Gregariousness, 12, 14 Gross national product (GNP), crosscultural studies, 125–126 Harm avoidance alignment in dimensional models, 7, 11 neurobiological model, 65, 237 psychobiological processes, 74, 75 serotonin level, 63, 68 HDI (Human Development Index), 126, 241 Health policy, DSM-IV-TR as tool, 30–31 Helpfulness, 13, 14 Heritability, of personality traits, 42–44 primary vs. secondary, 46–48, 236 strict definition, 55 as validity indicator, 55–56 Heterogeneity, within diagnoses, xxvii, 30, 146, 169, 221

Index 5-HIAA (5-hydroxyindoleacetic acid), 66 Hierarchical Personality Inventory for Children (HiPIC), 91–92, 97, 238 adolescent personality studies, 102 DIPSI comparisons, 99–100, 101 Hierarchical structure, of personality traits, 3, 10, 13, 40 etiological contributions, 156, 167 joint modeling, 151–152 lower level clarifications, 24–26 novel model, 168–169 psychopathology integration, 155–159, 176 research agenda, 234, 250 Higher-order trait/domains common, 3–5, 6–7, 8–9 consensus on, 214–215 developmental perspectives, 85–106, 112–115 homogeneous vs. heterogeneous, 169–170 joint structural model, 151–153 neurobiological model, 65–66 PAS identification, 81–82 psychopathology model integration, 154–159, 176 relation of measurements, 23–26 research agenda, 168, 242–243 proposal on, 182–183, 182–185 subtlety of diagnosis, 210, 247–248 treatment directed at, 212 HiPIC. See Hierarchical Personality Inventory for Children (HiPIC) Hippocampus, 65 Histrionic sexualization, 12, 14 Hofstede’s measure, of cultural differences, 120, 122 Homogenous response dispositions, 169 Homovanillic acid (HVA), 66 Honesty-humility, cross-cultural studies, 135–136

269 Hostility childhood models, 92 cross-cultural studies, 134 Type A personality and, 169–170 5-HTT gene, 63, 68, 237 Huang, Yueqin, 143 Human Development Index (HDI), 126, 241 Hybrid models, clinical utility, 209, 214, 228 5-Hydroxyindoleacetic acid (5-HIAA), 66 Hypotheses, theoretical, clinical utility and, 219–220 ICD-10. See International Classification of Diseases, 10th Revision (ICD-10) ICD-11 (International Classification of Diseases, 11th Revision), 33 Identity, in children, 114 Illiteracy, 123, 126 Imagination, childhood, 91, 94, 100, 101, 239 Impairments classification goals, 190–192, 229 clinically significant, 177–178, 181, 244–246 cutoffs, 176–177, 200–201 diagnosing, research proposal, 182, 183–184, 184 perspective on, 177–179 Impatience, Type A personality and, 169 Impulsivity alignment in dimensional models, 5, 6–7, 8–9, 29 alternative placement, 24, 41 childhood models, 89, 93 heritability, 46, 48 lower-order traits, 10, 11, 12 neurobiological model, 63, 65–67, 237 temperament models, 87 as trait vs. type, 169, 220, 223 Incentive stimuli, encoding of, 65

270 Individual differences, in personality disorders, 139–140 joint structural model, 151–152 Infant personality traits, 95, 112 Inhibition alignment in dimensional models, 4, 6–7 behavioral, neurobiological models, 62, 66 social, childhood models, 91, 92 Insanity, 140 Insight-based classification, 168 Integrated dimensional models, 3–15, 152, 176 Integrative model, of normal/abnormal personality novel, 154–159 old distinctions, 163–165 overview, 3–17, 150–154, 176 trait vs. type classification, 167–170 Intellect, childhood models, 89–90, 92, 94 Internalization clinical domain level, 23–24 genetics, 155–156 as higher-order trait, 3–4 psychopathology associations, 104, 105, 106 novel models, 154–159 research agenda, 234, 239 International Classification of Diseases, 10th Revision (ICD-10) consensus on, 33–34 criteria sets finalization for, 33–34 DSM-IV personality disorder categories vs., xxv–xxvi, 223 International Classification of Diseases, 11th Revision (ICD-11), 33 International Personality Disorder Examination (IPDE), 34, 239 International Pilot Study on Personality Disorders (IPSPD), 33–34

Dimensional Models of Personality Disorders International Sexuality Description Project, 120–121, 240 Interpersonal circumplex (IPC), 3, 234 higher-order domains, 4–5, 6, 8 Interpersonal nature, of personality disorder criteria, 165, 181, 223, 246 Interpersonal relatedness, cross-cultural studies, 123 Interrater reliability dimensional models, 208–209 personality disorder diagnoses, 34, 247 Interviews, standardized, 97, 145–146, 208, 241 Intimacy problems, 12, 14, 47 Introversion, 169 alignment in dimensional models, 4–5, 6, 9, 29, 40 lower-order traits, 12–13, 14 psychopathology associations, 104, 105, 106 treatment decisions based on, 212 IPC. See Interpersonal circumplex (IPC) IPDE (International Personality Disorder Examination), 34, 239 IPSPD. See International Pilot Study on Personality Disorders (IPSPD) Irritability, childhood models, 91–93 IRT (item response theory), 26 Item factor scoring, of primary trait genetics, 48, 184 Item response theory (IRT), 26 Joint structural model of normal/abnormal personality, 150–154 of personality and psychopathology, 154–159 Karolinska Psychodynamic Profile, 223 Kernberg’s Personality Organization, 223 Krueger’s unified model, 149–159 old distinctions, 163–165

Index research agenda, 233–250, 241–242 trait vs. type classification, 167–170 Language acquisition, 58, 95 Language diversity, 117–118, 125, 145 lexical approach, 133–134, 240 research agenda, 90, 200, 240 Legal arguments, DSM-IV-TR as tool, 30–31 Lexical approach, to personality structure, 134–136 in childhood, 90, 94 cross-cultural replication, 134–136 research agenda, 240–241 Liability threshold model, of familial diseases, 57–58 Life expectancy, cross-cultural studies, 126 Life span, personality stability across, 31, 97, 106, 114, 150, 165, 205, 229 Linear approach, to traits, 169 Livesley’s model of adult psychopathology. See Dimensional Assessment of Personality Psychopathology—Basic Questionnaire (DAPP-BQ) coverage and cutoff points, 173–174 diagnostic examples, 180 genetic, 39–50 problems and pitfalls, 55–58 research agenda, 233–250 Location parameters, personality construct models, 153 Locus coeruleus, 65 López-Ibor, Juan J., 139 Lower-order trait/domains, 10, 11, 12–13, 14–15, 15 alternative placements, 23–25 consensus on, 214–215 consistent hierarchical delineation, 25–26 developmental perspectives, 85–106, 112–115 joint structural model, 151–153

271 psychopathology model integration, 154–159, 176 research agenda, proposal on, 182–183, 182–185 subtlety of diagnosis, 210–211, 247–248 treatment directed at, 212 Macrotreatment decisions, 211–212, 248 Magnetic resonance imaging, personality trait associations, 64 Maladaptive traits cutoff points, 178–179, 200–201, 229 dimensional models diagnosis, 179–181 historical perspectives, 140, 164–165 research agenda, 234–235, 242–243, 245–246, 250 subtlety of diagnosis, 209–211, 247–248 Maltreatment, childhood, antisocial personality disorder, 42 Manipulativeness, 5 MAOA gene, 42, 57 MAO (monoamine oxidase) levels, 141 McGuffin, Peter, 55 MCMI-III (Millon Clinical Multiaxial Inventory), 4, 6, 222 Medial orbital cortex, 65 Meekness, 13, 14 Melancholy, 140 Mental disorders. See also Psychopathology DSM-IV-TR groups, 30–31, 149 historical perspectives, 139–141 spectrum model, 155–156 internalizing vs. externalizing syndromes, 156–159 Mentalization-based treatment, 213 Mervielde model, of developmental psychopathology, 85–106 commentary, 111–115

272

Dimensional Models of Personality Disorders

Meta-analyses FFM domain and personality disorders, 101–102, 103 personality constructs, 151 personality genetics, 41–42, 237 research agenda, 237, 242 temperament and personality stability, 114 Metric equivalence, of culture-related traits, 118–119 Microtreatment decisions, 212, 248 Millon Clinical Multiaxial Inventory (MCMI-III), 4, 6, 222 Millon Index of Personality Styles (MIPS), 3, 4, 234 Minnesota Multiphasic Personality Inventory (MMPI), 222 Mistrust, 5, 14, 191 Molecular genetics behavioral genetic research vs., 42, 50 neurobiological models, 63, 237 research agenda, 235–236 Monoamine oxidase A gene (MAOA), 42, 57 Monoamine oxidase (MAO) levels, 141, 237 Mood disorders, unipolar factor analysis, 2 internalizing construct, 154–159 old distinctions, 165 Moral insanity, 140 Morbid features, of personality, 140–141 Motivation childhood models, 89, 93 classification schemes, 178, 190–191 temperament models, 87 Multidimensional Personality Questionnaire (MPQ), 3, 8, 151 consensus on structure, 24–25, 151 lower-order traits, 10, 11, 12–13, 14, 15

Multivariate analyses personality domain modeling based on, 153, 243 personality trait genetics, 43, 47 agenda for, 235–236 at item level, 47–48, 56, 64 Narcissistic personality disorder, xxv, xxviii, 30 National Comorbidity Study (NCS), 178 National Institute of Mental Health, 145 Natural-language descriptions, of childhood personality, 90 Negative Affectivity alignment in dimensional models, 5, 6, 8–9, 15 alternative placement, 24 joint structural model, 151–152 old distinctions, 165 research proposal, 183 temperament models, 87 Negative Emotionality alignment in dimensional models, 6 alternative placement, 24 childhood models, 92–93 novel models, 154–155 psychopathology associations, 154–155 temperament models, 87 as trait vs. type, 169 Negative-item bias, cross-cultural studies, 126 NEO PI-R. See Neuroticism– Extraversion–Openness Personality Inventory—Revised (NEO PI-R) Neurobiological dimensional models, 61–68 background principles, 61–62 Cloninger’s, 62–64 Depue’s, 64–66 developmental research, 67–68 research agenda, 236–237

Index Siever and Davis’s, 66–67 of temperament, 141 Neurobiology, xxix, 3 Neuroendocrine challenge tests, 66 Neuroimaging, personality trait associations, 64 Neuroscience, 141 Neurotic disorder, Nottingham study, 78– 79, 80 Neuroticism alignment in dimensional models, 5, 6, 8, 15 alternative placement, 24, 40 childhood models, 89, 92–93, 112–113 cross-cultural studies, 120–122, 125, 134–136 FFM vs. personality disorders metaanalysis, 101–102, 103 heritability, 46, 48 neurobiological model, 65–66, 237 novel models, 154–155 old distinctions, 165 psychopathology associations, 154–155 research proposal, 183 temperament models, 87 as trait vs. type, 169 Neuroticism–Extraversion–Openness Personality Inventory—Revised (NEO PI-R), 2–3 adolescent personality studies, 102 childhood personality studies, 91–92, 99 clinical utility, 210, 215 consensus on structure, 25 correlation coefficients with TCI, 74, 75 coverage and cutoff points, 176 cross-cultural studies, 118, 120–121, 124–125, 127–128 heritability studies, 43, 45–47

273 higher-order domains, 4, 6, 9 joint structural model, 151, 153 lower-order traits, 10, 12–13, 14, 15 personality construct variables, 40, 91 Neurotransmitters. See also Dopamine; Serotonin complexity of, 68 peripheral measures, 63 personality traits modeled on, 65–67 New York Longitudinal Study (NYLS), 87 Nonaffective constraint, 65 Noradrenergic–cholinergic balance, 66–67 Normality, personality disorders as extreme of, 57–58, 140 Norms, societal variation, 144–145. See also Cross-cultural models Nosology childhood psychopathology assessment tools, 96 comorbidity tendencies, 154 genetically informed, 41–43, 45–46 historical perspectives, 140, 168 scientifically based, 169 Nottingham study, 78–79, 80 Novelty-seeking alignment in dimensional models, 7 genetic variations, 45 neurotransmitter associations, 41–42, 63, 68 psychobiological processes, 74, 75 Nucleus accumbens, 65 NYLS (New York Longitudinal Study), 87 Obsessionality, 10, 11, 12 Obsessive-compulsive personality disorder, xxv, xxviii, 30–31 Oldham, John M., 29 Openness to Experience alignment in dimensional models, 6, 9 alternative placement, 25, 40 childhood models, 89–90, 92, 94, 112–114

274

Dimensional Models of Personality Disorders

Openness to Experience (continued) cross-cultural studies, 121–122, 134 FFM vs. personality disorders metaanalysis, 101–102, 103 psychopathology associations, 104, 105, 106 treatment decisions based on, 212 Optimism, 41, 91 Orderliness, childhood models, 94 Others, representations of, 192, 229 Outcome research. See also Treatment response clinical utility and, 223, 228 Overcontrollers, 89, 104, 105 Oxytocin, 65 Panic disorder genetics, 156 Nottingham study, 78–79, 80 Paranoid personality disorder dimensional framework, 30, 191 ICD-10 vs. DSM-IV criteria, 35, 36–37 inadequate scientific basis, xxviii Parental descriptions of childhood personality, 90–91, 94 of childhood psychopathology, 95 Parental rearing, 145 Parent–child conflicts, 94 Paris, Joel neurobiological model review, 61–68 personality disorder severity, 77–82 psychobiological systems, 73–76 Partial hospitalization programs, 212 PAS. See Personality Assessment Schedule (PAS) Passive-aggressive personality disorder, xxviii Passive dependency, 81 PDNOS. See Personality disorder not otherwise specified (PDNOS) Peer ratings, cross-cultural studies, 134 People descriptions, classification schemes, 192

Perception alignment in dimensional models, 7, 9 neurobiological model, 66 of reality, 191 Persistence psychobiological processes, 7, 75 research agenda, 238–239 temperament models, 88, 93–94 Personality cross-cultural comparisons, 118–119. See also Cross-cultural models dynamic psychobiological system, 73–76 genetic approaches, 41–42 genetic structure, 42–44 joint structure of normal/abnormal, 150–154 life span stability vs. change, 31, 97, 106, 114, 165, 205, 229 person- vs. variable-centered model, 88–89, 189–190, 229 psychopathology relationship with, 99–100, 101 broad perspective, 104, 105, 106, 112, 153, 249–250 novel models, 154–159 temperament vs., 86–87, 111–112 Personality Assessment Schedule (PAS) dangerous/severe personality disorder, 81 disorder severity classification, 77, 79 higher-order domains, 4, 6 implications, 81–82 lower-order traits, 10, 11, 12–13, 14–15, 15 Nottingham study, 78–79, 80 origins, 3, 77–78 psychosis studies, 79, 80, 81 research agenda, 234 self-harm repetition study, 80, 81 Personality constructs clinical utility and, 219–220 factor analysis, 153 integrative model, 150–154, 176

Index research agenda, 234–235 Personality difficulty, 78, 79 Personality disorder not otherwise specified (PDNOS), xxvii, 172, 206–208, 245 clinical utility, 199, 229 Personality disorders. See also Axis II disorders Axis I disorders continuum, 2–3, 57–58, 95, 149–150 clinical utility and, 227–230 joint structure, 150–154 novel unified models, 154–159 old distinctions, 163–165 research agenda, 241–244 traits vs. types, 167–170 categorical model, xxv–xxix. See also Categorical model childhood and adolescent antecedents, 85–106. See also Developmental model coverage and cutoffs, xxvii, 171–181, 199–200, 205 research agenda, 181–182, 200–201, 244–247, 250 diagnosis of four-step process, 197 subtlety of, 209–211, 247–248 dimensional model, xxix–xxx, 1–17. See also Dimensional models historical perspective, 139–141 International Pilot Study, 33–34 interpersonal nature, 165 levels of, 31 meta-analyses of FFM domains and, 101–102, 103 nomenclature for. See Classification nomenclature public health relevance, 154 severity classification, 77–82, 79 stability vs. change, 31, 97, 106, 114, 150, 165, 205, 229

275 as trait extremes, 57–58, 140 trait vs. type classification, 167–170 Personality Disorders Work Group, xxix, 200 Personality functioning childhood models, 114–115 dimensional models, xxix, 2–3, 29, 40, 191–193 mental disorders connection, 150, 153 cutoff points, 176–181, 200–201 novel models, 154–159 perspectives on, 177–179 Personality Psychopathology–Five (PSY-5), 2, 25 higher-order domains, 4–5, 6, 9 Personality structure childhood models, 88–92, 112–114 consensus on, 23–26, 31 integration into dimensional models, 2, 16–17, 176 integrative model of normal/abnormal, 150–154, 164, 176 lexical approach, 134–136 in childhood, 90, 94 cross-cultural replication, 134–136, 240 Personality traits covariance in children and adolescents, 112 culture-related, 118–119, 127 genetic contributions, 43, 44–45, 50, 56 data sets for identifying, 184 disorder symptoms vs., 164–165 cutoff points, 176–181, 200–201 elevated, defining significant, 177–178, 183–184 geographical distribution, 121–122 heritability, 42–44 primary vs. secondary, 46–48, 236 strict definition, 55 as validity indicator, 55–56

276

Dimensional Models of Personality Disorders

Personality traits (continued) higher-order, 3–5, 6–7, 8–9 measurement relationships, 23–26 linear approach, 169 lower-order, 10, 11, 12–13, 14–15, 15 placement alternatives, 23–26 primary vs. secondary, 40–41, 46–48, 153, 236 research proposal, 183–184, 184 social context, 73, 205 surplus meaning, 169–170 taxonomy proposal, 92–95 type vs., 167–170, 190, 192 Person-centered model of personality, 88–89, 189–190, 229 prototype-matching, 192–193 PET (positron emission tomography), personality trait associations, 64, 66 Pharmacological interventions, 47, 213, 248 Phenomenological system, 165, 168, 243 clinical utility, 221, 228 Phenotype analyses, 45, 49–50 agenda for, 236, 249 common mental disorders, 155–156 gender-specific, 50, 63–64 heritability definition, 55–56 nosology implications, 45–46 Phobias, 156, 168 Pleasure, temperament models, 88 Pleiotropy, 45–46, 236 Population-based studies, 30, 56 Positive Affectivity alignment in dimensional models, 5, 6, 8, 12, 14 alternative placement, 24 research proposal, 183 Positive Emotionality alignment in dimensional models, 4–5, 6 alternative placement, 24, 41

childhood models, 91–93 joint structural model, 151 temperament models, 87, 92–93 Positron emission tomography (PET), personality trait associations, 64, 66 Power, 5, 6 Preschool children, diagnostic criteria, 95 Pretentiousness, cross-cultural studies, 134–135 Preternatural reactions, 139–140 Primary traits genetic influences, 40–41, 47–48, 236 research proposal, 183–184, 184 Professional communication, with models, 207–208, 224, 247 Prognostic information clinical utility, 228–229 diagnostic validity, 190, 205 Propriety, 10, 11, 12 Prosocial tendencies, childhood models, 94 Prototypic cases clinical utility, 228–229 decision making and, 221, 228–229 diagnostic applications, 192–193, 223–224 examples, 179–180 research applications, 209, 213, 245, 250 PSY-5. See Personality Psychopathology— Five (PSY-5) Psychoanalysis, 212 Psychobiological system, personality as, 73–76 Psychodynamic systems, 223–224 Psychological processes, classification schemes, 178, 190–192, 229 Psychology of differences, 139–140 Psychometric studies, 41, 48, 98, 125 Psychomotor retardation, 165 Psychopathic personality heterogeneous components, xxvii, 40, 95 historical perspectives, 140–141

Index Psychopathology coverage, xxvii, 172–179, 200–201, 205, 244–247, 250 cutoff points, 176–181, 200–201 definitional nature, 139 dimensional models, 95–99 clinical utility, 227–230 research proposal, 182, 183–184, 184 etiology of, 156, 167, 190 internalizing–externalizing framework, 154–159 personality relationship with, 99–100, 101 broad perspective, 104, 105, 106, 112, 153, 249–250 novel models, 154–159 primary vs. secondary, 153, 183, 236 Psychopathology Checklist—Revised, 153 Psychosis old distinctions, 165 recurrent, PAS for, 79, 80, 81 Psychosocial interventions, 47 Psychoticism alignment in dimensional models, 5, 6, 9 alternative placement, 24–25, 40–41 heritability, 48 Public health, disorders relevant to, 154 Pull, Charles B., 33 Q-factor analysis, 89 Q-sort, in personality types, 89, 190 Qualitative distinctions, of dimensions, categories, and domains, 223–224 Quantitative studies of dimensions, categories, and domains, 223–224 genetics, 56–57 Questionnaires, standardized, 145–146 Reactivity, in temperament, 87

277 Reality perception, 191 Recovery, 229. See also Treatment response Reimbursement, DSM-IV-TR as tool, 30–31 Related trait models, Five-Factor Model, 189–193 Relationships, developmental perspectives, 115 Remission, 164 Representations of others, 192, 229 of self, 229 Research, 233–250 Axis I and Axis II continuity, 241–244 clinical utility focus, 227–230 childhood antecedents, 238–239 clinical utility, 224, 246–249 conference agenda, 233–234, 249 coverage, 181–182, 200–201, 244–246 cross-cultural issues, 239–241 diagnostic cutoff points, 182, 200–201, 244–246 dimensional models, 30, 214–215, 234–235 proposal, 182–185, 183–184 generating ideas for, 249–250 genetics, behavioral and molecular, 235–237 inadequate status, xxviii–xxix, 146 integrated models, 15–17 international collaboration network, 200, 224 language diversity, 200, 240 neurobiological model, 67–68, 236–237 outcome, 223 Resiliency, childhood models, 89 Resourcefulness, 10, 11, 12 Response dispositions, homogenous, 169 Responsibility, 10, 11, 12

278

Dimensional Models of Personality Disorders

Reward dependence alignment in dimensional models, 5, 7 dopamine and, 66 genetic variations, 45 norepinephrine and, 63–64 psychobiological processes, 74, 75 Rosenberg Self-Esteem Scale (RSES), 125 Rothbart-Derryberry model, of temperament, 87–88, 93 Sane insanity, 140 Scalar equivalence, of culture-related traits, 118–119 Schedule for Nonadaptive and Adaptive Personality (SNAP), 2 clinical utility, 193, 210, 214, 222 consensus on structure, 24–26 coverage and cutoff points, 174 higher-order domains, 4, 6, 8–9 joint structural model, 151 lower-order traits, 10, 11, 12–13, 14–15, 15 Schizoid orientation alignment in dimensional models, 12–13, 14 PAS identification, 81 Schizotypal personality disorder arbitrary diagnostic boundaries, xxviii, 13 dimensional framework, 29–30, 164 ICD-10 vs. DSM-IV, xxv, 9 Schmitt, David, 120–121 SCID-II (Structured Clinical Interview for Axis II disorders), 97 Scientific methods childhood psychopathology assessment, 96, 141, 168 for implementing change, 200–201, 223 inadequate status, xxviii–xxix, 146 Secondary traits behavioral genetics, 40–41, 46, 236

definitional constructs, 46 Self, representations of, 229 Self-denigrating, 13, 14 Self-directedness alignment in dimensional models, 7 brain potential correlations, 64 genetic variations, 45 psychobiological processes, 74, 75 Self-discipline, 10, 11, 12 Self-esteem, 229 cross-cultural studies, 125–126, 128 Self-evaluation. See Self-reports Self-harm, repetitive facet-level construct, 152–153 PAS for, 80, 81 Self-reflection, capacity for, 192 Self-regulation, in temperament, 87 Self-reports cross-cultural studies, 122–124, 126, 134 research agenda, 240–241 of personality disorders, 61, 209, 222 Self-transcendence, 7, 45 psychobiological processes, 74, 75 Semistructured interviews, 97, 145–146, 208, 241 Sensation seeking, 41, 46 Serotonin, impulsivity/aggression related to, 66–68, 237 Serotonin receptor binding, peripheral measures of, 63 Serotonin transporter gene (5-HTT), 63, 237 Seven-Factor model, Cloninger’s, 62–64, 141 coverage and cutoff points, 175 diagnostic examples, 180 Severe personality disorder, 78, 79, 81 Severity of illness research agenda, 244–246 treatment decisions based on, 211–212

Index Sexuality classification schemes, 192 international research, 120–121, 240 Shea, M. Tracie, 163 Shedler and Westen Assessment Procedure—200 (SWAP-200), 2 coverage and cutoff points, 173 diagnostic examples, 179–180 lower-order traits, 10, 12–13, 14, 15 research agenda, 234 subscale alignment potential, 4 Shiner, Rebecca L., 111 Shyness, childhood models, 91, 92 Siblings. See also Twins personality studies, 91 Siever-Davis model consensus on structure, 24 higher-order domains, 4, 7, 8–9 neurobiological, 66–67 Simonsen, Erik, xxv, 1, 219 Simple personality disorder, 78, 79 Slyness, cross-cultural studies, 134 SNAP. See Schedule for Nonadaptive and Adaptive Personality (SNAP) Sociability alignment in dimensional models, 4, 6, 12, 14, 40 neurobiological model, 65 psychobiological processes, 7, 75 research proposal, 183 temperament models, 87–88, 92 Social avoidance/closeness, 12, 14 Social-cognitive psychologists, 73 Social context acceptable behavior, 144–145, 250 personality traits, 73, 181, 191, 205, 246 treatment decisions based on, 212 Social desirability, cross-cultural studies, 125–126 Social dominance alignment in dimensional models, 5, 6, 41

279 neurobiological model, 65 Social potency, 12, 14 Societal norms, variation in, 144–145. See also Cross-cultural models Socioeconomic status, cross-cultural studies, 125–126 Sociopathy, 81, 141 Somatoform syndromes, 155, 158, 168 Spectrum model, of mental disorders, 2–4, 24, 155–156 hypothesis on, 29, 97, 141 internalizing vs. externalizing syndromes, 155–159 support for, 164, 204–205 Statistical deviance, 177, 197 Statistical models, for outcome measurement, 223 Status-seeking, cross-cultural studies, 125–126, 135 Stereotypes, cross-cultural, 123–124, 127, 241 Stimulus-seeking, 12 Structural equation models, 58 Structural model, joint of normal/abnormal personality, 150–154 of personality and psychopathology, 154–159 Structured Clinical Interview for Axis II disorders (SCID-II), 97 Submissiveness, 13, 15, 40 Substance use disorders externalizing construct, 155–157, 159 factor analysis, 2 genetics, 56, 156 Subtlety of diagnosis, 209–211, 247–248 Suffering, 140–141 Surplus meaning, of traits, 169–170 SWAP-200. See Shedler and Westen Assessment Procedure—200 (SWAP-200)

280 Symptoms clinically significant, 177–178, 181, 244–246 diagnosing, research proposal, 182, 183–184, 184 within general population, 2–3 stability vs. change, 31, 97, 106, 114, 150, 165, 205 traits vs., 164–165, 229 Tang, Siu Wa, 143 Taxonomic research behavioral genetics, 43, 48–49 clinical utility, 211–213 empirical basis, 167–168 lack of higher-level organization, 168 TBAQ (Toddler Behavior Assessment Questionnaire), 88 TCI. See Temperament and Character Inventory (TCI) Teacher ratings of childhood personality, 90, 94 of childhood psychopathology, 95–96 Temperament Cloninger’s model, 62–64, 73 coverage and cutoff points, 175 genetics, 41–42, 45 individual differences, 139–140 models, 86–88 neurobiology, 141 nonlinear relation to character, 74 personality vs., 86–87, 111–112 research agenda, 238–239 taxonomy, 92–95 Temperament and Character Inventory (TCI), 2 clinical utility, 210 correlation coefficients with NEO PI-R scales, 74, 75 coverage and cutoff points, 175 heritability studies, 43, 45 higher-order domains, 4, 7, 8–9

Dimensional Models of Personality Disorders joint structural model, 151 lower-order traits, 10, 11, 12–13, 14–15 neurobiological models, 62–64 psychobiological defined traits, 73–74 research agenda, 234, 245 Temporal stability old distinctions, 165 personality disorder diagnoses, 34, 97, 106, 114 Ten Commandments of Implementing Change, 201–202 Theory clinical utility role, 219–220, 228–229 diagnostic validity, 205 Thomas-Chess model, of temperament, 87–88, 93 Thought process, integrity in classification schemes, 178, 190–192 Three-Factor Model, 5, 8, 23, 24, 40 Toddler Behavior Assessment Questionnaire (TBAQ), 88 Traits. See Personality traits Transference, 212 Treatment directiveness, 212–213 Treatment duration, 211–212, 248 Treatment format, 211–212 Treatment frequency, 212, 248 Treatment goals, 212–213 Treatment planning DSM-IV-TR as tool, 30–31, 47 macro-decisions, 211–212, 248 micro-decisions, 212––213, 248 research agenda, 247–249 theoretical concepts, 220, 227–228 Treatment recommendations, 211 clinical utility and, 210, 221–222, 247 dimensional models and, 210–213 research agenda, 247–249 Treatment response, 150 clinical utility and, 223, 228 conceptualization and assessment, 205–206

281

Index Treatment-seeking behavior, 30, 139, 221 Treatment settings, 211–212, 248 clinical utility and, 221–222 Treatment techniques, 211–212, 248 Trull, Timothy J., 171, 189, 195, 199, 219, 227 Trust/trusting, 5, 14, 191 Twins heritability studies, 56 personality studies, 91 psychopathology studies, 104 temperament studies, 88 Type A personality, 169 Typology, of personality disorders, 169 traits vs., 167–170, 190, 192 Tyrer, Peter, 77 Uncertainty avoidance, cross-cultural studies, 122 Unconventionality, 7, 9, 29 Undercontrollers, 89, 104, 105 Unified model, of Axis I and Axis II disorders, 149–159 old distinctions, 163–165 trait vs. type classification, 167–170 Unipolar system, 29, 222 User acceptability, dimensional models, 206–207 measurement of, 207 Validity concurrent, of DIPSI, 98–99 construct of dimensional model, 30–31, 49, 224 of personality disorder classification, 40–41, 48–49 content, of DIPSI, 98 diagnostic correlates for, 25–26, 205 of DSM-IV-TR, 203–204 elements of, 204–206

of dimensional model, xxix–xxx, 2–3, 15 construct vs. evidence, 30–31, 49, 224 evidence-based diagnostic, 195–197, 205, 207 dimensional models, 30, 49, 207, 224 prognosis data, 228–230 research agenda, 224, 242–243, 245, 247 taxonomic research, 167–168 Van Leeuwen, Karla, 85 Variable-centered model, of personality, 89–92, 189–190 Vasopressin, 65 Vegetative symptoms, 165 Ventral pallidum, 65 Ventral tegmental area, 65 Verheul, Roel, 203, 233 Watson, David, 167 Westen, Drew, 189 Widiger-Simonsen model, xxv, 1–17 coding criteria commentary, 33–37 cross-cultural studies, 135 research agenda, 233–250 toward consensus on, 23–26 working out framework, 29–31 Wilberg, Theresa, 227 Withdrawal, 4, 6, 40 Within-person models, 74 Workaholism, 10, 11, 12 World Health Organization (WHO) personality disorder categories, xxv, xxix, 9, 244 Task Force, 33, 145 Youth Self-Report (YSR), 96 Zuckerman-Kuhlman Personality Questionnaire (ZKPQ), 3 higher-order domains, 4–5, 6