Edited by Eric J. Mash and Russell A. Barkley THE GUILFORD PRESS New York London © 2003 The Guilford Press A Division of Guilford Publications, Inc. 72 Spring Street, New York, NY 10012 www.guilford.com All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher. Printed in the United States of America This book is printed on acid-free paper. Last digit is print number: 9 8 7 6 5 4 3 2 1 Library of Congress Cataloging-in-Publication Data Child psychopathology / edited by Eric J. Mash and Russell A. Barkley. — 2nd ed. p. cm. Includes bibliographical references and index. ISBN 1-57230-609-2 1. Child psychopathology. I. Mash, Eric J. II. Barkley, Russell A., 1949– . RJ499 .C4863 2002 618.92'89—dc21 2002009086 To our wives, Heather and Pat This page intentionally left blank About the Editors Eric J. Mash, PhD, is Professor in the Department of Psychology at the University of Calgary. He completed his undergraduate studies at City University of New York, his doctorate in clinical psychology at Florida State University, and his postdoctoral work at the Oregon Health Sciences University. Dr. Mash is a fellow of the American and Canadian Psychological Associations and has served as an editorial board member and consultant for numerous scientific and professional journals. His research interests are in child and adolescent psychopathology, assessment, and therapy, and he has published many books and journal articles on these topics. His research has focused on interaction patterns in families of children with different problems including attentiondeficit and oppositional disorders and children who have been maltreated. Russell A. Barkley, PhD, is Professor in the College of Health Professions at the Medical University of South Carolina, Charleston, South Carolina. He is a Diplomate in both Clinical Psychology and Clinical Neuropsychology, has written more than 200 scientific articles and book chapters dealing with ADHD and related topics, and is author, editor, or coeditor of 15 books. Dr. Barkley is the founding Editor of The ADHD Report, a newsletter for clinicians, and creator of seven professional videos, two of which have won national awards. He has served as President of the International Society for Research in Child and Adolescent Psychopathology and the Section of Clinical Child Psychology of the American Psychological Association (now Division 53). vii This page intentionally left blank Contributors Anne Marie Albano, PhD, Institute for the Study of Child and Adolescent Anxiety Disorders, New York University Child Study Center, New York, New York Joan Rosenbaum Asarnow, PhD, Department of Psychiatry, Neuropsychiatric Institute, University of California, Los Angeles, School of Medicine, Los Angeles, California Robert F. Asarnow, PhD, Department of Psychiatry, Neuropsychiatric Institute, University of California, Los Angeles, School of Medicine, Los Angeles, California Russell A. Barkley, PhD, College of Health Professions, Medical University of South Carolina, Charleston, South Carolina David H. Barlow, PhD, Center for Anxiety and Related Disorders, Department of Psychology, Boston University, Boston, Massachusetts Marcia C. Barnes, PhD, Department of Pediatrics, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada Carolyn Black Becker, PhD, Graduate School of Professional and Applied Psychology, Rutgers University, Piscataway, New Jersey ix Diane Benoit, MD, FRCPC, Department of Psychiatry, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada Kim B. Burgess, PhD, Department of Human Development, University of Maryland, College Park, Maryland Laurie Chassin, PhD, Department of Psychology, Arizona State University, Tempe, Arizona Bruce F. Chorpita, PhD, Department of Psychology, University of Hawaii, Honolulu, Hawaii Geraldine Dawson, PhD, Department of Psychology, University of Washington, Seattle, Washington David J. A. Dozois, PhD, Department of Psychology, University of Western Ontario, London, Ontario, Canada Elisabeth M. Dykens, PhD, Neuropsychiatric Institute, Department of Child Psychiatry, University of California, Los Angeles, Los Angeles, California Jack M. Fletcher, PhD, Department of Pediatrics/ Center for Academic and Reading Skills, University of Texas–Houston Health Science Center, Houston, Texas Kenneth E. Fletcher, PhD, University of Massachusetts Medical School, Worcester, Massachusetts Shelly Grabe, MA, Department of Psychological Sciences, University of Missouri–Columbia, Columbia, Missouri Constance Hammen, PhD, Department of Psychology, University of California, Los Angeles, Los Angeles, California Karen Heffernan, PhD, private practice, New York, New York Stephen P. Hinshaw, PhD, Department of Psychology, University of California, Berkeley, Berkeley, California Robert M. Hodapp, PhD, Department of Education, University of California, Los Angeles, Los Angeles, California Amy E. Kennedy, BA, Department of Human Development, University of Maryland, College Park, Maryland Kevin M. King, MA, Department of Psychology, Arizona State University, Tempe, Arizona Laura Grofer Klinger, PhD, Department of Psychology, University of Alabama, Tuscaloosa, Alabama Steve S. Lee, MA, Department of Psychology, University of California, Berkeley, Berkeley, California G. Reid Lyon, Child Development and Behavior Branch, National Institute of Child Health and Human Development, Bethesda, Maryland Karlen Lyons-Ruth, PhD, Department of Psychiatry, Harvard Medical School, Cambridge, Massachusetts Eric J. Mash, PhD, Department of Psychology, University of Calgary, Calgary, Alberta, Canada x Contributors Lizette Peterson, PhD, (deceased), Department of Psychological Sciences, University of Missouri–Columbia, Columbia, Missouri Kelle Reach, BA, Department of Psychological Sciences, University of Missouri–Columbia, Columbia, Missouri Peggy Renner, PhD, Department of Psychology, University of Alabama, Tuscaloosa, Alabama Jennifer Ritter, MA, Department of Psychology, Arizona State University, Tempe, Arizona Kenneth H. Rubin, PhD, Department of Human Development, University of Maryland, College Park, Maryland Karen D. Rudolph, PhD, Department of Psychology, University of Illinois, Champaign, Illinois Shannon L. Stewart, PhD, Child and Parent Resource Institute, Ministry of Community, Family, and Children’s Services, London, Ontario, Canada Ryan S. Trim, MA, Department of Psychology, Arizona State University, Tempe, Arizona Christine Wekerle, PhD, Department of Psychiatry, University of Toronto and Child Psychiatry Program, Centre for Addiction and Mental Health, Toronto, Ontario, Canada G. Terence Wilson, PhD, Graduate School of Applied and Professional Psychology, Rutgers University, Piscataway, New Jersey David A. Wolfe, PhD, Department of Psychology, The University of Western Ontario, London, Ontario, Canada Charles H. Zeanah, MD, Department of Psychiatry and Pediatrics, Tulane University School of Medicine, New Orleans, Louisiana xi Preface Research in child, adolescent, and developmental psychopathology continues to flourish, even more so than when the first edition of this text was published. Previously recognized disorders are even better delineated than they were only a few years ago, and a few new ones seem to have been discovered along the way. The publication rate in this field is extraordinary, with many journals now focusing exclusively on childhood mental illness and health, and numerous articles on children appearing each month in journals that were once the exclusive domains of adult psychopathology. To those of us who take a developmental view of psychopathology, this is a most gratifying state of affairs as we come to recognize the roots of many adult disorders in childhood and adolescence. The down side, of course, is that even the expert researchers in the various disorders that constitute this field find it harder than ever to keep abreast of research findings appearing at such a rapid clip. And woe to the clinical professionals who must deal with these childhood disorders: They may find themselves quickly and hopelessly behind in the advancements occurring in the understanding of these clinical conditions. Hence the need for a volume such as this, and especially for its second edition, to assist the clinical professional, student, and even expert in remaining current on child and adolescent psychopathological disorders. Now more than ever, the field of child psychopathology epitomizes the dynamic, accumulative, and self-correcting nature of the scientific enterprise, as new findings expand upon and are assimilated with the established facts in any given disorder. Often these new findings challenge older theoretical or conceptual assumptions or more explicit models of these disorders, at times even leading to small-scale paradigm shifts in perspective. In short, the literature on child and adolescent psychopathology is alive, well, prosperous, and rapidly advancing. Old questions undoubtedly get answered, but along the way those answers raise new questions for researchers to pursue in ever more complex programs of research on each of the childhood disorders covered here. Although the pace and excitement levels vary considerably across different areas of child psychopathology, within each area the eager anticipation of new knowledge remains palpable as new lines of research and methodologies— such as neuroimaging, behavioral and molecular genetics, structural equation modeling, and longitudinal designs—come to overlap old ones and so provide greater opportunities to better understand these disorders. The challenge remains for this second edition as it was for the first: How are we to capture the current status of this rapidly evolving field? Our answer was again to identify those experts who have dedicated their professional careers to these disorders, and let them—unfettered by fashion or the editors’ pet perspectives—tell us what they have learned. In other words, we tried to find the most knowledgeable professionals on particular disorders and asked them to provide up-to-date and comprehensive summaries of the nature of the disorders in which they have specialized. We asked only that their discussions be grounded in their respective bodies of scientific literature, eschewing clinical lore, dogmatic wisdom, the sayings of the guru du jour, or political agendas. We also asked that they set aside the concerns of assessment and treatment of their respective disorders, so as to have ample room for the burgeoning findings on the disorders themselves. These other topics are the focus of related books (Mash & Barkley, 1998; Mash & Terdal, 1997). In essence, each author or group of authors was once more challenged to answer these basic questions: “What do we know about this disorder?”, “What are the implications for future research into further understanding the disorder?”, and, just as important, “Where are the current limitations or gaps in our knowledge that deserve future attention?” If sound, scientifically grounded theoretical or conceptual models of the disorder exist, then these were also to be reviewed. In addressing these questions, the experts assembled here were directed to cover (1) the nature of the behavior, symptoms, and/or cognitive and emotional deficits that typify the core of each disorder; (2) a brief historical perspective; (3) any criteria that exist to establish its presence (diagnosis) and a candid appraisal of those criteria; (4) epidemiological knowledge pertaining to the prevalence, gender distribution, and ethnic and cultural factors associated with the disorder; (5) the developmental course and varied pathways shown xii Preface to be associated with the disorder; (6) the psychiatric, psychological, and social disorders or difficulties that most often coexist with the disorder (comorbidity); and (7) a survey of those things believed to give rise to the disorder (etiology). Once more, we believe that the many authors assembled here have done a marvelous job accomplishing their charge. We trust the reader will concur. As before, we are indebted to the professionals who agreed to write for this second edition on their respective disorders. We genuinely appreciate the substantial time commitment they have made to writing their chapters, many of which are major updates of their previous work. Many others deserve our gratitude as well, including Jeannie Tang, Judith Grauman, Marie Sprayberry, Carolyn Graham, Kim Miller, and Alison Wiigs, for shepherding the manuscript through the production process. Special thanks are also owed to our long-time friends and founders of The Guilford Press, Seymour Weingarten (Editor-in- Chief) and Bob Matloff (President), for more than 20 years of support for our various books, including this one. Last, but hardly least, we thank our families—Heather Mash, and Pat, Ken, and Steve Barkley—for relinquishing the family time such a project requires, and for their support, patience, and encouragement of our careers in this field. Eric J. Mash, PhD Department of Psychology University of Calgary Russell A. Barkley, PhD College of Health Professions Medical University of South Carolina REFERENCES Mash, E. J., & Barkley, R. A. (Eds.). (1998). Treatment of childhood disorders (2nd ed.). New York: Guilford Press. Mash, E. J., & Terdal, L. G. (Eds.). (1997). Assessment of childhood disorders (3rd ed.). New York: Guilford Press. Contents xiii III. INTRODUCTION ONE Child Psychopathology: A Developmental– 3 Systems Perspective Eric J. Mash and David J. A. Dozois III. BEHAVIOR DISORDERS TWO Attention-Deficit/Hyperactivity Disorder 75 Russell A. Barkley THREE Conduct and Oppositional Defiant Disorders 144 Stephen P. Hinshaw and Steve S. Lee FOUR Adolescent Substance Use Disorders 199 Laurie Chassin, Jennifer Ritter, Ryan S. Trim, and Kevin M. King III. EMOTIONAL AND SOCIAL DISORDERS FIVE Childhood Mood Disorders 233 Constance Hammen and Karen D. Rudolph SIX Childhood Anxiety Disorders 279 Anne Marie Albano, Bruce F. Chorpita, and David H. Barlow SEVEN Childhood Posttraumatic Stress Disorder 330 Kenneth E. Fletcher EIGHT Social Withdrawal in Childhood 372 Kenneth H. Rubin, Kim B. Burgess, Amy E. Kennedy, and Shannon L. Stewart IV. DEVELOPMENTAL AND LEARNING DISORDERS NINE Autistic Disorder 409 Laura Grofer Klinger, Geraldine Dawson, and Peggy Renner TEN Childhood-Onset Schizophrenia 455 Joan Rosenbaum Asarnow and Robert F. Asarnow ELEVEN Mental Retardation (Intellectual Disabilities) 486 Robert M. Hodapp and Elisabeth M. Dykens TWELVE Learning Disabilities 520 G. Reid Lyon, Jack M. Fletcher, and Marcia C. Barnes IV. INFANTS AND CHILDREN AT RISK FOR DISORDER THIRTEEN Disorder and Risk for Disorder during Infancy 589 and Toddlerhood Karlen Lyons-Ruth, Charles H. Zeanah, and Diane Benoit FOURTEEN Child Maltreatment 632 Christine Wekerle and David A. Wolfe VI. EATING AND HEALTH-RELATED DISORDERS FIFTEEN Eating Disorders 687 G. Terence Wilson, Carolyn Black Becker, and Karen Heffernan SIXTEEN Health-Related Disorders 716 Lizette Peterson, Kelle Reach, and Shelly Grabe Author Index 751 Subject Index 786 xiv Contents 1. A Developmental–Systems Perspective 1 I INTRODUCTION 2 I. INTRODUCTION This page intentionally left blank 1. A Developmental–Systems Perspective 3 CHAPTER ONE Child Psychopathology A Developmental– Systems Perspective Eric J. Mash David J. A. Dozois 3 This volume provides a comprehensive account of the characteristics, definitions, developmental course, correlates, causes, contexts, and outcomes of psychopathology in children.1 Our knowledge base of child and developmental psychopathology has grown exponentially over the past decade (Cicchetti & Cohen, 1995a, 1995b; Cicchetti & Sroufe, 2000; Mash & Wolfe, 2002; Ollendick & Hersen, 1998). New conceptual frameworks, research methods, and findings continue to advance our understanding of childhood disorders (Cicchetti & Rogosch, 1999; Rutter & Sroufe, 2000; Sameroff, Lewis, & Miller, 2000), as well as our ability to assess and treat children with problems (Mash & Barkley, 1998; Mash & Terdal, 1997a; Orvaschel, Faust, & Hersen, 2001; Shaffer, Lucas, & Richters, 1999). However, this knowledge base is compromised by the frequently atheoretical, unsystematic, and fragmented fashion in which research findings in child psychopathology have accrued, and by the conceptual and research complexities inherent in the study of such a rapidly changing and socially embedded organism as the child (Hinshaw, 2001; Jensen et al., 1993; Kazdin & Kagan, 1994). In this introductory chapter, we address several central themes and issues related to conceptualizing childhood dysfunction and its many determinants. In doing so, we provide a developmental– systems framework for understanding child psychopathology— one that emphasizes the role of developmental processes, the importance of context, and the influence of multiple and interacting events and processes in shaping adaptive and maladaptive development. FACTORS COMPLICATING THE STUDY OF CHILD PSYCHOPATHOLOGY Almost since modern views of mental illness began to emerge in the late 18th and early 19th centuries, far less attention has been given to the study of psychopathology in children than in adults (Silk, Nath, Siegel, & Kendall, 2000). For example, in 1812 Benjamin Rush, the first American psychiatrist, suggested that children were less likely to suffer from mental illness than adults, because the immaturity of their developing brains would prevent them from retaining the mental events that caused insanity (Silk et al., 2000). More recently, interest in the study of child psychopathology has increased dramatically. This is due to a growing realization that (1) many childhood problems have lifelong consequences and costs both for children and for society; (2) most adult disorders are rooted in early childhood con4 I. INTRODUCTION ditions and/or experiences; and (3) a better understanding of childhood disorders offers promise for developing effective intervention and prevention programs (National Advisory Mental Health Council [NAMHC] Workgroup, 2001). Issues concerning the conceptualization and definition of psychopathology in children continue to be vigorously debated. Such debates are fueled by the relative absence of well-controlled research studies with children as compared with adults. Until recently, much of the field’s accumulated knowledge about child psychopathology, its causes, and its outcomes was extrapolated from theory and research on adult disorders. For example, only in the last 5–10 years have childfocused models and research into such disorders as depression and anxiety emerged (Zahn-Waxler, Klimes-Dougan, & Slattery, 2000). Even in studies conducted with children, much of our knowledge is based on findings obtained at a single point in a child’s development and in a single context. Although useful, such findings provide still photographs of moving targets and fail to capture the dynamic changes over time that characterize most forms of child psychopathology (Achenbach & Dumenci, 2001; Lewis & Granic, 2000; Patterson, 1993). In addition, prior studies have not given sufficient attention to the social and cultural milieu in which atypical child development occurs (Cicchetti & Aber, 1998; Garc?a Coll & Garrido, 2000). Contextual models (e.g., Bronfenbrenner, 1977) and longitudinal approaches (e.g., Robins, 1966) have been available in the field of child study for some time. However, only in the past decade has the research enterprise taken seriously the need for developmentally sensitive systems-oriented models to account for the emergence of psychopathology in children (del Carmen & Huffman, 1996; Sameroff, 2000a), or the need to study developmental trajectories utilizing longitudinal methods (e.g., Emery, Waldron, Kitzmann, & Aaron, 1999; Hauser-Cram, Warfield, Shonkoff, & Krauss, 2001; Kotler, Cohen, Davies, Pine, & Walsh, 2001; Maughan & Rutter, 2001; Verhulst & Koot, 1991). The study of child psychopathology is further complicated by the facts that childhood problems do not come in neat packages, and that most forms of psychopathology in children are known to overlap and/or to coexist with other disorders (Angold, Costello, & Erkanli, 1999). For example, there is much overlap among such problems as violence, emotional and behavioral disorders, child maltreatment, substance abuse, delinquency, and learning difficulties (e.g., Greenbaum, Prange, Friedman, & Silver, 1991); between childhood anxiety and depression (e.g., Compas & Oppedisano, 2000; Seligman & Ollendick, 1998); and between reading disabilities and anxiety and depression (Willcutt & Pennington, 2000b). Many behavioral and emotional disturbances in children are also associated with specific physical symptoms and/or medical conditions (Egger, Costello, Erkanli, & Angold, 1999; Meltzer, Gatward, Goodman, & Ford, 2000). It is also the case that distinct boundaries between many commonly occurring childhood difficulties (e.g., noncompliance, defiance) and those problems that come to be labeled as “disorders” (e.g., oppositional defiant disorder) are not easily drawn (e.g., Loeber, Burke, Lahey, Winters, & Zera, 2000). Judgments of deviancy often depend as much on other child characteristics (e.g., age, sex, intelligence), the situational appropriateness of a child’s behavior, the social and cultural context in which judgments are made, and the characteristics and decision rules of adults who make these judgments as they do on any specific behaviors displayed by the child (Achenbach, 2000; Mash & Terdal, 1997b). There is a growing recognition that all current diagnostic categories of child psychopathology are heterogeneous with respect to etiology and outcome, and will need to be broken down into subtypes (Kagan, 1997). Although these diagnostic systems make some allowances for subtypes, designations are rudimentary at best, given the many different subgroups and types that have been identified for children with such disorders as attention-deficit/ hyperactivity disorder (ADHD), conduct disorder, oppositional defiant disorder, anxiety disorders, and mood disorders (e.g., Milich, Balentine, & Lynam, 2001). It has become increasingly evident that most forms of child psychopathology cannot be attributed to a single unitary cause. Although certain rare disorders (e.g., phenylketonuria, fragile-X mental retardation, or Rett’s disorder) may be caused by single genes, current models in behavioral and molecular genetics recognize that more common and complex disorders are the result of the operation of multigene systems containing varying effect sizes (Goldsmith, Gottesman, & Lemery, 1997; McGuffin, Riley, & Plomin, 2001; O’Conner & Plomin, 2000). Most forms of child psychopathology are polygenic, involving a number of susceptibility genes that interact with one 1. A Developmental–Systems Perspective 5 another and with environmental influences to result in observed levels of impairment (Rutter, 2000a; State, Lombroso, Pauls, & Leckman, 2000). Child and family disturbances are likely to result from multiple, frequently co-occurring, reciprocal, and interacting risk factors, causal events, and processes (e.g., Eaves et al., 1997; Ge, Conger, Lorenz, Shanahan, & Elder, 1995; Rende, 1999; Rutter et al., 1997). Contextual events exert considerable influence in producing child and adolescent disorders—an influence that is almost always equivalent to or greater than those factors usually thought of as residing “within” the child (Caspi, Taylor, Moffitt, & Plomin, 2000; Reiss & Neiderhiser, 2000; Rutter, 2000b). Numerous determinants of child psychopathology have been identified, including genetic influences (e.g., State et al., 2000); hypo- or hyperreactive early infant dispositions (e.g., Hirshfeld, Biederman, Brody, & Faraone, 1997); insecure child–parent attachments (e.g., Bretherton, 1995; Sroufe, Carlson, Levy, & Egeland, 1999); difficult child behavior (e.g., Costello & Angold, 2001); social-cognitive deficits (e.g., Crick & Dodge, 1994; Schwartz & Proctor, 2000); deficits in social learning (e.g., Patterson, 1982; Patterson, Reid, & Dishion, 1992), emotion regulation (e.g., Keenan, 2000), and/or impulse control and response inhibition (Barkley, 1997; Nigg, 2000, 2001); neuropsychological and/or neurobiological dysfunction (e.g., Cicchetti & Cannon, 1999); maladaptive patterns of parenting (e.g., Lovejoy, Graczyk, O’Hare, & Neuman, 2000); parental psychopathology, such as maternal depressed mood (e.g., Goodman & Gotlib, 1999); parental or couple discord (e.g., Grych & Fincham, 2001); limited family resources and other poverty-related life stressors (e.g., Rutter, 1999); institutional deprivation (e.g., Kreppner et al., 2001); and a host of other potential factors. These factors cannot be understood in isolation, and for most disorders, research does not support granting central etiological status to any single risk or causal factor (e.g., Seifer, Sameroff, Baldwin, & Baldwin, 1992). Since the many causes and outcomes of child psychopathology operate in dynamic and interactive ways over time, they are not easy to disentangle. The designation of a specific factor as a cause or an outcome of child psychopathology usually reflects (1) the point in an ongoing developmental process at which the child is observed, and (2) the perspective of the observer. For example, a language deficit may be viewed as a disorder in its own right (e.g., mixed receptive–expressive language disorder), the cause of other difficulties (e.g., impulsivity), or the outcome of some other condition or disorder (e.g., autistic disorder). In addition, biological and environmental determinants interact at all periods of development. Dawson, Hessl, and Frey (1994), for example, noted that the characteristic styles parents use in responding to their infants’ emotional expressions may influence how patterns of cortical mappings and connections within the limbic system are established in the infants. Similarly, J. Hart, Gunnar, and Cicchetti (1995) reported that maltreated preschoolers showed reduced cortisol activity in response to stress relative to controls—a finding that suggests altered activity of the stress-regulating hypothalamic–pituitary– adrenocortical (HPA) system among children who have been maltreated. These and other findings suggest that early experiences may shape neural structure and function, which may then create dispositions that direct and shape a child’s later experiences and behavior (Cicchetti & Walker, 2001; Dawson et al., 1999; Glaser, 2000; Kaufman & Charney, 2001; Post & Weiss, 1997). As will be discussed throughout this volume, current models of child psychopathology seek to incorporate the role of evolved mechanisms, neurobiological factors, early parent–child relationships, attachment processes, a long-term memory store that develops with age and experience, micro- and macrosocial influences, cultural factors, age and gender, and reactions from the social environment as variables and processes that interact and transform one another over time. In short, then, current approaches view the roots of developmental and psychological disturbances in children as the result of complex interactions over the course of development between the biology of brain maturation and the multidimensional nature of experience (Cicchetti & Toth, 1997; Cicchetti & Tucker, 1994; Reiss & Neiderhiser, 2000; Rutter et al., 1997). The experience and the expression of psychopathology in children are known to have cognitive, affective, physiological, and behavioral components; in light of this, many differing descriptions and definitions of dysfunctionality in children have been proposed. As we discuss in a later section, a common theme in defining child psychopathology has been that of “adaptational failure” in one or more of these components or in the ways in which these components are organized and integrated (Sameroff, 2000a; Sroufe, 6 I. INTRODUCTION 1997; Sroufe & Rutter, 1984). Adaptational failure may involve deviation from age-appropriate norms (Achenbach, 2001), exaggeration or diminishment of normal developmental expressions, interference in normal developmental progress, failure to master developmental tasks, failure to develop a specific function or regulatory mechanism, and/or the use of non-normative skills (e.g., rituals, dissociation) as a way of adapting to regulatory problems or traumatic experiences (Fischer et al., 1997; Sroufe, 1997). A multitude of etiological models and treatment approaches have been proposed to explain and remediate psychopathology in children. Unfortunately, most of these have yet to be substantiated— or, for that matter, even tested (Kazdin, 2000, 2001). These models and approaches have differed in their relative emphasis on certain causal mechanisms and constructs, often using very different terminology and concepts to describe seemingly similar child characteristics and behaviors. Although useful, many of these models have been based on what seem to be faulty premises concerning singular pathways of causal influence that do not capture the complexities of child psychopathology (Kazdin & Kagan, 1994). In this regard, evolutionary models have emphasized the role of selection pressures operating on the human species over millions of years; biological paradigms have emphasized genetic mutations, neuroanatomy, and neurobiological mechanisms as factors contributing to psychopathology; psychodynamic models have focused on intrapsychic mechanisms, conflicts, and defenses; attachment models have emphasized the importance of early relationships and the ways in which internal representations of these relationships provide the foundation for constructing working models of self, others, and relationships more generally; behavioral/reinforcement models have emphasized excessive, inadequate, or maladaptive reinforcement and/or learning histories; social learning models have emphasized the importance of observational learning, vicarious experience, and reciprocal social interactions; cognitive models generally focus on the child’s distorted or deficient cognitive structures and processes; affective models have emphasized dysfunctional emotion-regulating mechanisms; and family systems models have conceptualized child psychopathology within a framework of intra- and intergenerational family systems and subsystems and have emphasized the structural and/or functional elements that surround family relational difficulties. The distinctiveness of each model mentioned above is in the relative importance it attaches to certain events and processes. However, it should be emphasized that despite these variations in the relative emphasis given to certain causes versus others, most models recognize the role of multiple interacting influences. For example, although differing in emphasis, social learning and affective models both place importance on the role of symbolic representational processes in explaining childhood dysfunction. There is a growing recognition of the need to integrate currently available models through intra- and interdisciplinary research efforts. Such integration generally requires looking beyond the emphasis of each single-cause theory to see what can be learned from other approaches, as well as a general openness to relating concepts and findings from diverse theories (cf. Arkowitz, 1992). Recent studies suggest that theoretical integration is becoming more common in psychopathology research (e.g., Beauchaine, 2001). Attachment theory has, for instance, been increasingly integrated with cognitive models (e.g., Ingram & Ritter, 2000). Theoretical integration is also apparent in studies combining proximal cognitive and interpersonal factors with distal variables, such as the early home environment and patterns of attachment (e.g., Lara, Klein, & Kasch, 2000). The link between cognitive and neuropsychological functioning is likewise being tested more frequently (e.g., Nigg, Blaskey, Huang-Pollack, & Rappley, 2002; Pine & Grun, 1999). Thus it appears as though researchers are beginning to recognize the importance of combining theoretical approaches, and are accepting the monumental task of incorporating increased complexity into their research designs. Interdisciplinary perspectives on child psychopathology mirror the considerable investment in children on the part of many different disciplines and professions. The study of the etiology and maintenance of psychopathology in children has been and continues to be the subject matter of psychology, medicine, psychiatry, education, and numerous other disciplines. Clearly, no one discipline has proprietary rights to the study of childhood disturbances. Each discipline has formulated child psychopathology in terms of its own unique perspective. Particularly relevant, in the context of this chapter, is that child psychopathology and normality in medicine and psychiatry 1. A Developmental–Systems Perspective 7 have typically been conceptualized and defined categorically in terms of the presence or absence of a particular disorder or syndrome that is believed to exist “within the child.” In contrast, psychology has more often conceptualized psychopathology– normality as representing extremes on a continuum or dimension of characteristics, and has also focused on the role of environmental influences that operate “outside the child.” However, the boundaries between categories and dimensions, or between inner and outer conditions and causes, are arbitrarily drawn, and there is a continuing need to find workable ways of integrating the two different world views of psychiatry/medicine and psychology (Richters & Cicchetti, 1993; Scotti & Morris, 2000; Shaffer et al., 1999). As the subsequent chapters in this volume attest, research into child psychopathology is accelerating at a remarkable rate. This in turn has resulted in a rapidly expanding and changing knowledge base. Each chapter in this volume provides a comprehensive review of current research and theory for a specific form of child psychopathology, and a discussion of new developments and directions related to this disorder. In the remainder of this introductory chapter, we provide a brief overview and discussion of the following: historical developments in the study of child psychopathology; epidemiological considerations; basic issues; approaches to the definition and classification of childhood disorders; common types of psychopathology in children; important philosophical and epistemological assumptions that have guided theory and research; predominant theories regarding etiology; and prevalent and recurrent conceptual and methodological issues that cut across the wide spectrum of disorders represented in this volume. Particular emphasis is given to concepts, methods, and strategies capturing the complexities, reciprocal influences, and divergent pathways that current models and research have identified as crucial for understanding child psychopathology. HISTORICAL CONTEXT FOR CHILD PSYCHOPATHOLOGY Brief Historical Overview Historical developments surrounding the emergence of child psychopathology as a field of study have been documented in a number of excellent sources and are considered only briefly here (see Achenbach, 1982; Cicchetti, 1990; Donohue, Hersen, & Ammerman, 2000; Kanner, 1962; Rie, 1971; Rubinstein, 1948; Silk et al., 2000). In general, the emergence of concepts of child psychopathology was inextricably related to the broader philosophical and societal changes in the ways children have been viewed and treated by adults over the course of history (Aries, 1962; Borstelmann, 1983; French, 1977; Postman, 1994). Several overlapping perspectives for conceptualizing and dealing with deviant child behavior emerged, including the religious, the legal, the medical, the social, and the educational (Costello & Angold, 2001). In ancient Greek and Roman societies, child behavior disorders were believed to result from organic imbalances, and children with physical or mental handicaps, disabilities, or deformities were viewed as sources of economic burden and/ or social embarrassment. As such, they were usually scorned, abandoned, or put to death (French, 1977). This mistreatment, by today’s standards, was common throughout the Middle Ages (A.D. 500–1300). In colonial America, as many as twothirds of all children died prior to the age of 5 years, and those who survived continued to be subjected to harsh treatment by adults. For example, the Massachusetts Stubborn Child Act of 1654 permitted a father to petition a magistrate to put a “stubborn” or “rebellious” child to death (fortunately, no sentences were carried out); in Massachusetts and elsewhere, mentally ill children were kept in cages and cellars into the mid- 1800s (Silk et al., 2000). The historical record indicates that prior to the 18th century, when references to disordered child behavior were made at all, they were usually presented in terms of the problem child’s behavior as inherently evil (Kanner, 1962). Bizarre behaviors in children were attributed to Satanic possession and evil spirits during the Spanish Inquisition, and both John Calvin and Martin Luther viewed mentally retarded children as filled with Satan. And, as noted by Rie (1971), “No distinct concept of disordered behavior in children could emerge so long as possession by the devil excluded other notions of causality” (p. 8). Although nearly all varieties of aberrant behavior in children have existed for millennia, the formal study of such behavior is relatively recent. Following a comprehensive review of historical developments in child psychopathology, Rie 8 I. INTRODUCTION (1971) concluded: “There is a consensus, then, about the absence of any substantial body of knowledge—prior to the twentieth century— concerning disordered behavior in childhood; about the inconsistencies and discontinuities of efforts on behalf of disturbed children; and about the relative absence of those professional specialties which now concern themselves with such problems” (p. 6). Rubinstein (1948) noted that (1) there was not a single article dealing with insanity in childhood in any of the first 45 volumes of the Journal of Insanity; (2) there was no discovery or theory of importance to child psychiatry in the American literature prior to 1900, and no research today stems from any of these writings; and (3) the only significant work with children prior to the 20th century focused on the care, treatment, and training of “mental defectives.” Increased concern for the plight and welfare of children with mental and behavioral disturbances was the result of two important influences. First, advances in general medicine, physiology, and neurology led to the reemergence of the organic disease model and a concomitant emphasis on more humane forms of treatment. Second, the growing influence of the philosophies of John Locke, Johann Pestalozzi, and Jean-Jacques Rousseau led to the view that children needed moral guidance and support. With these changing views came an increased concern for moral education, compulsory education, and improved health practices. These early influences also provided the foundation for evolving views of child psychopathology as dependent on both organic and environmental causes. Masturbatory Insanity: An Example Societal and clinical views regarding masturbation in children can be used to illustrate the ways in which conceptualizations of child psychopathology have changed over time, as well as several general issues related to its definition, study, and treatment. In addition to the historical significance of masturbation as the first disorder described as unique to children and adolescents (Rie, 1971), early conceptualizations of masturbatory insanity illustrate a view of mental illness as residing within the child (Cattell, 1938; Hare, 1962; Rees, 1939; Rie, 1971; Szasz, 1970). Society’s objections to masturbation originated from Orthodox Jewish codes and from Judeo- Christian dogmata (Patton, 1985; Szasz, 1970). It was not until the 18th century—with a decline in the domination of religious thought, coupled with the augmented influence of science—that masturbation came to be viewed as particularly harmful (Rie, 1971; Szasz, 1970). An anonymous clergyman who later became a physician wrote a dissertation entitled Onania, or the Heinous Sin of Self-Pollution (circa 1710, cited in Szasz, 1970). It was this manuscript that initially transformed the moral convictions regarding the wrongfulness of masturbation into a physiological explanation with severe medical ramifications. Following this exposition, numerous books appeared claiming that masturbation was a predominant etiological cause of both physical disease and mental illness. Thus the notion that sexual overindulgence was deleterious to one’s health was accepted, “virtually unaltered, first by the Church and then by Medicine” (Szasz, 1970, p. 182). Although the medical view of masturbation first emphasized the adverse impact upon physical health, the dominant thinking shifted by the middle of the 19th century to a focus on the presumed negative effects on mental health and nervous system functioning. By the latter part of the 19th century, masturbation was the most frequently mentioned “cause” of psychopathology in children. In fact, Spitzka (1890; cited in Rie, 1971) attributed at least 25% of all psychiatric cases to this etiological factor. Views of masturbatory insanity emerged and were maintained in the absence of any thought to the contrary, and without any consideration of the base rate of masturbation in the general population. Although interest in masturbatory insanity began to wane in the latter half of the 19th century, the argument endured (albeit in milder forms) during the early 20th century, when psychoanalytic theory gained rapid acceptance. Freud suggested that masturbation was one of the precipitants of neurasthenia, hypochondriasis, and anxiety neurosis (Rees, 1939). Apart from his own theories regarding the pathogenesis of neuroses, however, Freud did not present any real evidence for this view (Szasz, 1970). Eventually the notion of masturbatory insanity gave way to the concept of neurosis, but it was still not until much later in the 20th century that the misguided and illusory belief in a relationship between masturbation and mental illness was dispelled. As conceptualizations of childhood psychopathology evolved, and several variants of psychotherapy and residential treatments were developed (see Grellong, 1987, and Roberts & Kurtz, 1. A Developmental–Systems Perspective 9 1987), the search for determinants of psychiatric disorders in children became increasingly sophisticated, thorough, and systematic (Rie, 1971). With this increased refinement of theory and research, there remained only fragments of the etiological hypothesis of masturbation. For example, in some psychoanalytic circles, enuresis was thought to symbolize suppressed masturbation (Rees, 1939; Walker, Kenning, & Faust- Campanile, 1989). Eventually masturbation came to be viewed as entirely harmless (Szasz, 1970) and even as usefully adaptive (Baker, 1996). This brief historical review illustrates a number of points. First, it shows how the political and social climates influence our definitions of child psychopathology. The impact of religious thought was clearly reflected in the transformation from the moral judgment against the sins of the flesh, to the medical opinion that masturbation was harmful to one’s physical health, to the psychiatric assertion that sexual overindulgence caused insanity. Second, the review points out the need to be cognizant of the ways in which moral convictions, idiosyncratic definitions of normality or pathology, and personal expectations influence what investigators look for and ultimately find in the name of science. In the case of masturbation, misleading findings resulted because hypotheses were “tested” with a mentality of confirmation rather than falsification (see Maxwell & Delaney, 1990). Szasz (1970), in writing about the powerful authority of America’s historical psychiatric figures such as Benjamin Rush, noted that there is a tendency among scientists to “attend only to those of their observations that confirm the accepted theories of their age, and reject those that refute them” (p. 187). Third, masturbatory insanity illustrates the potential dangers that ensue when treatment decisions are made on the basis of deficient theoretical exposition and in the absence of empirical data. For example, early treatments consisted of clitoridectomies for women and spike-toothed rings placed on the penises of men (Szasz, 1970). Finally, the example of masturbatory insanity portrays the long-standing view of psychopathology as residing within the child and the essential neglect of the role of his or her surroundings, context, relationships, and the interactions among these variables. Current theory, research, and practice reflect a shift toward acknowledging developmental factors and including the family, peer group, school, and other sources of influence in conceptualizing and understanding child psychopathology (Luthar, Burack, Cicchetti, & Weisz, 1997; Mash & Wolfe, 2002). Additional developments have included an increased research emphasis on examining the interactions of multiple proximal and distal vulnerability factors (Ingram, Miranda, & Segal, 1998; Price & Lento, 2001), understanding psychopathology across the life span (Ingram & Price, 2001), identifying empirically supported treatments for various childhood problems (Kazdin & Weisz, 1998; Lonigan, Elbert, & Johnson, 1998), and a focus on prevention (Greenberg, Domitrovich, & Bumbarger, 2001; National Institute of Mental Health, 2001). SIGNIFICANCE OF CHILD PSYCHOPATHOLOGY There has been and continues to be a great deal of misinformation and folklore concerning disorders of childhood. Many unsubstantiated theories have existed in both the popular and scientific literatures. These have ranged from mid- 19th-century views that overstimulation in the classroom causes insanity (see Makari, 1993), to mid-20th-century views that inadequate parenting causes autism (Bettelheim, 1967) or that chemical food additives cause hyperactivity (Feingold, 1975). In addition, many of the constructs used to describe the characteristics and conditions of psychopathology in children have been globally and/or poorly defined (e.g., “adjustment problem,” “emotional disturbance”). Despite the limitations, uncertainties, and definitional ambiguities that exist in the field, it is also evident that psychopathology during childhood represents a frequently occurring and significant societal concern that is gradually coming to the forefront of the political agenda. In the United States, the approach of the new millennium witnessed the first Surgeon General’s report on mental health (U.S. Public Health Service, 1999), which was followed by White House meetings on mental health in young people and on the use of psychotropic medications with children. A Surgeon General’s conference on children’s mental health resulted in an extensive report and recommendations (U.S. Public Health Service, 2001a), a similar report on youth violence (U.S. Public Health Service, 2001b), and a “blueprint” for research on child and adolescent mental health (NAMHC Workgroup, 2001).2 10 I. INTRODUCTION Increasingly, researchers in the fields of developmental psychopathology, child psychiatry, and clinical child psychology are considering the social policy implications of their work and striving to effect improvements in the identification of and services for youths with mental health needs (Cicchetti & Toth, 2000; Weisz, 2000). Greater recognition is also being given to factors that contribute to children’s successful mental functioning, personal well-being, productive activities, fulfilling relationships, and ability to adapt to change and cope with adversity (Cicchetti, Rappaport, Sandler, & Weissberg, 2000; Thompson & Ontai, 2000; U.S. Department of Health and Human Services, 2000b; U.S. Public Health Service, 2001a). The growing attention to children’s mental health problems and competencies arises from a number of sources. First, many young people experience significant mental health problems that interfere with normal development and functioning. As many as 1 in 5 children in the United States experience some type of difficulty (Costello & Angold, 2000; Roberts, Attkisson, & Rosenblatt, 1998), and 1 in 10 have a diagnosable disorder that causes some level of impairment (Burns et al., 1995; Shaffer et al., 1996). These numbers probably underestimate the magnitude of the problem, since they do not include a substantial number of children who manifest subclinical or undiagnosed disturbances that may place them at high risk for the later development of more severe clinical problems. For example, McDermott and Weiss (1995) reported that of the children in their national sample who were classified as adjusted, 34.4% were classified as being only “marginally” adjusted. In addition, although not meeting formal diagnostic criteria, many subclinical conditions (e.g., depressed mood, eating problems) are also associated with significant impairment in functioning (e.g., Angold, Costello, Farmer, Burns, & Erkanli, 1999; Lewinsohn, Striegel-Moore, & Seeley, 2000). Evidence gathered by the World Health Organization (WHO) suggests that by the year 2020, childhood neuropsychiatric disorders will rise by over 50% internationally, to become one of the five most common causes of morbidity, mortality, and disability among children (U.S. Public Health Service, 2001a). Second, a significant proportion of children do not grow out of their childhood difficulties, although the ways in which these difficulties are expressed change in both form and severity over time (Offord et al., 1992). Even when diagnosable psychopathology is not evident at later ages, a child’s failure to adjust during earlier developmental periods may still have a lasting negative impact on later family, occupational, and social adjustment. And some forms of child psychopathology— for example, an early onset of antisocial patterns of behavior in boys—can be highly predictive of various negative psychosocial, educational, and health outcomes in adolescence and adulthood (see Hinshaw & Lee, Chapter 3, this volume). Third, recent social changes and conditions may place children at increasing risk for the development of disorders, and also for the development of more severe problems at younger ages (Duncan, Brooks-Gunn, & Klebanov, 1994; Kovacs, 1997). These social changes and conditions include multigenerational adversity in inner cities; chronic poverty in women and children; pressures of family breakup, single parenting, and homelessness; problems of the rural poor; direct and indirect exposure to traumatic events (e.g., terrorist attacks or school shootings); adjustment problems of children in immigrant families; difficulties of Native American children; and conditions associated with the impact of prematurity, HIV, cocaine, and alcohol on children’s growth and development (McCall & Groark, 2000; National Commission on Children, 1991; Shonkoff & Phillips, 2000). In addition to sociocultural changes, medical advances associated with higher rates of fetal survival may also contribute to a greater number of children showing serious behavior problems and learning disorders at a younger age. Fourth, for a majority of children who experience mental health problems, these problems go unidentified: Only about 20% receive help, a statistic that has not changed for some time (Burns et al., 1995). Even when children are identified and receive help for their problems, this help may be less than optimal. For example, only about half of children with identified ADHD seen in realworld practice settings receive care that conforms to recommended treatment guidelines (Hoagwood, Kelleher, Feil, & Comer, 2000). The fact that so few children with mental health problems receive appropriate help is probably related to such factors as a lack of screening, inaccessibility, cost, a lack of perceived need on the part of parents, parental dissatisfaction with services, and 1. A Developmental–Systems Perspective 11 the stigmatization and exclusion often experienced by these children and their families (Hinshaw & Cicchetti, 2000; Kroes et al., 2001). Empirically supported prevention and treatment programs for many childhood disorders are only now becoming available (Kazdin & Weisz, 1998; Lonigan et al., 1998), and there is a pressing need for the development and evaluation of prevention and intervention programs that are grounded in theory and research on child development in general, and developmental psychopathology in particular (Greenberg et al., 2001; Kazdin, 2001; Kurtines & Silverman, 1999; NAMHC Workgroup, 2001; Rapport, 2001).3 Fifth, a majority of children with mental health problems who go unidentified and unassisted often end up in the criminal justice or mental health systems as young adults (Loeber & Farrington, 2000). They are at much greater risk for dropping out of school and of not being fully functional members of society in adulthood; this adds further to the costs of childhood disorders in terms of human suffering and financial burdens. For example, average costs of medical care for youngsters with ADHD are estimated to be double those for youngsters without ADHD (Leibson, Katusic, Barberesi, Ransom, & O’Brien, 2001). Moreover, allowing just one youth to leave high school for a life of crime and drug abuse is estimated to cost society from $1.7 to $2.3 million (Cohen, 1998). Finally, a significant number of children in North America are being subjected to maltreatment, and chronic maltreatment during childhood is associated with psychopathology in children and later in adults (Emery & Laumann- Billings, 1998; MacMillan et al., 2001). Based on a review of the evidence, De Bellis (2001) has proposed that the psychobiological outcomes of abuse be viewed as “an environmentally induced complex developmental disorder” (p. 539). Although precise estimates of the rates of occurrence of maltreatment are difficult to obtain, due to the covert nature of the problem and other sampling and reporting biases (see Cicchetti & Manly, 2001; Mash & Wolfe, 1991), the numbers appear to be large. Nearly 3 million suspected cases of child abuse and neglect are investigated each year by child protective service agencies, and about 1 million children in the United States were confirmed as victims of child maltreatment in 1998 (U.S. Department of Health and Human Services, 2000a). It has been estimated that each year as many as 2,000 infants and young children die from abuse or neglect at the hands of their parents or caregivers (U.S. Advisory Board on Child Abuse and Neglect, 1995). Moreover, many reports of “accidental” injuries in children may be the result of unreported mistreatment by parents or siblings (Peterson & Brown, 1994). It would appear, then, that the total number of children who show adverse psychological and physical effects of maltreatment in North American society is staggering. EPIDEMIOLOGICAL CONSIDERATIONS Prevalence Epidemiological studies seek to determine the prevalence and distribution of disorders and their correlates in particular populations of children who vary in age, sex, socioeconomic status (SES), ethnicity, or other characteristics (Costello & Angold, 2000). The overall lifetime prevalence rates for childhood problems are estimated to be high and on the order of 14–22% of all children (Rutter, 1989). Rutter, Tizard, and Whitmore (1970), in the classic Isle of Wight Study, found the overall rate of child psychiatric disorders to be 6–8% in 9- to 11-year-old children. Richman, Stevenson, and Graham (1975), in the London Epidemiological Study, found moderate to severe behavior problems for 7% of the population, with an additional 15% of children having mild problems. Boyle et al. (1987) and Offord et al. (1987), in the Ontario Child Health Study, reported that 19% of boys and 17% of girls had one or more disorders. Many other epidemiological studies have reported similar rates of prevalence (e.g., Brandenburg, Friedman, & Silver, 1990; Costello, Farmer, Angold, Burns, & Erkanli, 1997; Earls, 1980; Hewitt et al., 1997; Lapouse & Monk, 1958; MacFarlane, Allen, & Honzik, 1954; Shaffer et al., 1996; Verhulst & Koot, 1992; Werner, Bierman, & French, 1971). Perhaps the most consistent general conclusions to be drawn from these studies are that prevalence rates for childhood problems are generally high, but that rates vary with the nature of the disorder; the age, sex, SES, and ethnicity of the child; the criteria used to define the problem; the method used to gather information (e.g., interview, questionnaire); the informant (e.g., child, parent, teacher); 12 I. INTRODUCTION sampling considerations; and a host of other factors. Age Differences Bird, Gould, Yager, Staghezza, and Camino (1989) reported no significant age differences for children aged 4 to 16 years in the total number of Diagnostic and Statistical Manual of Mental Disorders, third edition (DSM-III) disorders diagnosed at each age. Some studies of nonclinical samples of children have found a general decline in overall problems with age (e.g., Achenbach & Edelbrock, 1981), whereas similar studies of clinical samples have found an opposite trend (e.g., Achenbach, Howell, Quay, & Conners, 1991). Some studies have reported interactions among number of problems, age, sex of child, problem type, clinical status, and source of information (e.g., Simonoff et al., 1997). For example, Achenbach et al. (1991) found that externalizing problems showed a decline with age relative to internalizing problems, but only for those children who had been referred for treatment; Offord, Boyle, and Racine (1989) found complex interactions between age and sex of the child, with the results also depending on whether the informant was a child, parent, or teacher. These and other findings raise numerous questions concerning age differences in children’s problem behaviors. Answers to even a seemingly simple question such as “Do problem behaviors decrease (or increase) with age?” are complicated by (1) a lack of uniform measures of behavior that can be used across a wide range of ages; (2) qualitative changes in the expression of behavior with development; (3) the interactions between age and sex of the child; (4) the use of different informants; (5) the specific problem behavior(s) of interest; (6) the clinical status of the children being assessed; and (7) the use of different diagnostic criteria for children of different ages. Notwithstanding these difficulties, both longitudinal (MacFarlane et al., 1954) and crosssectional (Achenbach & Edelbrock, 1981; Achenbach et al., 1991) general population surveys are informative in depicting changes in the proportions of specific parent-, teacher-, or childreported problem behaviors with age (e.g., “hyperactive,” “argues,” “cries”), as well as the manner in which the age changes vary as a function of problem type, sex, and clinical status of the child. However, it should be emphasized that general age trends are based on group statistics, which may obscure the nonlinear and nonnormative changes that often occur for individual children. In addition, general surveys do not provide information concerning the processes underlying age changes. Studies of change in individual children over time and of the context in which this change occurs are needed if such processes are to be understood (e.g., Bergman & Magnusson, 1997; Francis, Fletcher, Stuebing, Davidson, & Thompson, 1991). Socioeconomic Status Although most children with mental health problems are from the middle class, mental health problems are overrepresented among the very poor. It is estimated that 20% or more of children in North America are poor, and that as many |as 20% of children growing up in inner-city poverty are impaired to some degree in their social, behavioral, and academic functioning (Duncan et al., 1994; Institute of Medicine, 1989; Schteingart, Molnar, Klein, Lowe, & Hartmann, 1995). Lower-SES children have been reported to display more psychopathology and other problems than upper-SES children (e.g., Keenan, Shaw, Walsh, Delliquadri, & Giovannelli, 1997; Samaan, 2000). However, although the reported relationships between SES and child psychopathology are statistically significant, the effects are small and should be interpreted cautiously (Achenbach et al., 1991). More importantly, global estimates of SES often tell us little about the associated processes through which SES exerts its influence on a child. Knowledge of such processes is needed to inform our understanding of disorders. For example, the effects of SES on aggression can be explained mostly by stressful life events and by beliefs that are accepting of aggression (Guerra, Tolan, Huesmann, Van Acker, & Eron, 1995). The impact of socioeconomic disadvantage on children derives from the fact that SES is a composite variable that includes many potential sources of negative influence (Bradley, Corwyn, McAdoo, & Garc?a Coll, 2001). In addition to low income, low SES is often characterized by low maternal education, a low level of employment, single-parent status, parental psychopathology, limited resources, and negative life events (e.g., poor nutrition, exposure to violence). Since overall indices of SES may include one or more of these variables in any given study, the relation1. A Developmental–Systems Perspective 13 ship that is reported between SES and child psychopathology may vary as a function of the particular index used, as well as ethnic factors (McLeod & Nonnemaker, 2000). In short, SES may serve as a proxy or indicator of other more active factors that influence risk for child psychopathology. Some research findings in child psychopathology are confounded by a failure to control for SES. For example, although physically abused children show higher levels of externalizing problems than nonabused children (Mash, Johnston, & Kovitz, 1983), it is not clear that physical abuse and externalizing problems are associated when the effects of SES are controlled for (Cummings, Hennessy, Rabideau, & Cicchetti, 1994; Wolfe & Mosk, 1983). The relationships among SES, maltreatment, and behavior disorders are further complicated by other findings that the effects of physical abuse on internalizing disorders may be independent of SES, whereas the effects of abuse on externalizing disorders may be dependent on SES-related conditions (Okun, Parker, & Levendosky, 1994). Sex Differences Although sex differences in the expression of psychopathology have been formally recognized since the time that Freud presented his views at the beginning of the 20th century, psychopathology in girls has received far less research attention than psychopathology in boys (Bell-Dolan, Foster, & Mash, in press; Eme, 1979). In the past, many studies have either excluded girls from their samples entirely or have examined all children together without considering findings for girls separately. For example, until recently there were relatively few studies on disruptive behavior disorders in girls (e.g., Moffitt, Caspi, Rutter, & Silva, 2001; Silverthorn & Frick, 1999; Zoccolillo, 1993). This omission was related to the perception that such disorders are much more common in boys than in girls; to sampling biases in which boys, who are more severely disruptive, are also more likely to be referred and studied; and to the use of inclusionary diagnostic criteria that were derived and validated largely from studies with boys (Spitzer, Davies, & Barkley, 1990). Research has confirmed that there are important differences in the prevalence, expression, accompanying disorders, underlying processes, outcomes, and developmental course of psychopathology in boys versus girls (Eme, 1979, 1992; Hops, 1995; Keenan & Shaw, 1997; Willcutt & Pennington, 2000a; Zahn-Waxler, 1993). ADHD, autism, childhood disruptive behavior disorders, and learning and communication disorders are all more common in boys than girls, whereas the opposite is true for most anxiety disorders, adolescent depression, and eating disorders (Hartung & Widiger, 1998). Although these sex differences are well established, their meaning is not well understood. For example, it is difficult to determine whether observed sex differences are a function of referral or reporting biases, the way in which disorders are currently defined, differences in the expression of the disorder (e.g., direct vs. indirect aggressive behavior), sex differences in the genetic penetrance of disorders, or sex differences in biological characteristics and environmental susceptibilities. All are possible, and there is a need for research into the processes underlying observed differences. Clearly the mechanisms and causes of sex differences may vary for different disorders (e.g., ADHD vs. depression), or for the same disorder at different ages (e.g., child vs. adolescent obsessive–compulsive disorder or early- vs. late-onset conduct disorder). Early research into sex differences focused mainly on descriptive comparisons of the frequencies of different problems for boys versus girls at different ages. In general, differences in problem behaviors between the sexes are small in children of preschool age or younger (e.g., Briggs-Gowan, Carter, Skuban, & Horwitz, 2001; Gadow, Sprafkin, & Nolan, 2001), but become increasingly common with age. For example, Weisz and Suwanlert (1989) studied children in the United States and Thailand, and found that boys were rated higher than girls on every problem for which there was a significant sex difference— including total problems, undercontrolled problems, overcontrolled problems, and culturespecific problems. Across cultures, boys have been found to display more fighting, impulsivity, and other uncontrolled behaviors than girls (Olweus, 1979). It has been found that boys show greater difficulties than girls during early or middle childhood, particularly with respect to ADHD and disruptive behavior disorders (MacFarlane et al., 1954). Girls’ problems may increase during adolescence, with higher prevalence rates for depression and dysphoric mood from midadolescence through adulthood. For example, conduct disorder and hyperactivity have been found to be 14 I. INTRODUCTION more frequent in 12- to 16-year-old boys than girls, whereas emotional problems have been found to be more frequent for girls than boys in this age group (Boyle et al., 1987; Offord et al., 1987). In addition, early signs of aggression have been found to predict later antisocial behavior for boys but not for girls (Tremblay et al., 1992). However, not all studies have reported significant sex differences in overall rates of problem behavior (e.g., Achenbach & Edelbrock, 1981; Velez, Johnson, & Cohen, 1989), and even when significant overall sex differences have been found, they tend to be small and to account for only a small proportion of the variance. It has also been found that although there is a much larger predominance of externalizing problems in boys and of internalizing problems in adolescent girls in samples of children who are referred for treatment, sex differences in externalizing versus internalizing problems are minimal in nonreferred samples of children (Achenbach et al., 1991). Comparisons of the behavioral and emotional problems in boys versus girls over time can provide useful information about sex-related characteristics. However, taken in isolation, such global comparisons do not address possible qualitative differences in (1) expressions of psychopathology in boys versus girls; (2) the processes underlying these expressions; (3) the long-term consequences of certain behaviors for boys versus girls; and/or (4) the impact of certain environmental events on boys versus girls (Zahn-Waxler, 1993). As noted by Hops (1995), it seems likely that “the pathways from childhood to adolescence and adult pathology are age and gender specific and that these differences may be the result of different social contexts that nurture the development of health or pathology for female and male individuals” (p. 428). In addition to differential socialization practices, there are likely to be differences in the expression and outcome of psychopathology in boys versus girls as a function of biologically based differences. For example, in a study of the psychophysiology of disruptive behavior in boys versus girls, Zahn-Waxler, Cole, Welsh, and Fox (1995) found that disruptive girls showed high electrodermal responding relative to disruptive boys and were also highly activated by a sadness mood induction. These investigators suggested that girls’ disruptive behavior may be more closely connected than boys’ disruptive behavior to experiences of anxiety. Other research has found that increases in depression in females during adolescence are related mostly to accompanying changes in levels of estrogen and androgen (Angold, Costello, Erkanli, & Worthman, 1999). It is also possible that for some disorders (e.g., ADHD), girls may require a higher genetic loading for the disorders than boys before the disorders are likely to express themselves (Rhee, Waldman, Hay, & Levy, 1999). There may also be differences in the processes underlying the expression of psychopathology and distress in boys versus girls. For example, findings suggest that the an adolescent’s emergent sexuality may create special difficulties with the parent of the opposite sex, and that distress in adolescent males may be particularly disruptive for mothers and daughters (Ge et al., 1995). Others studies have found that daughters of depressed mothers may be at greater risk than sons for the development of internalizing disorders (Gelfand & Teti, 1990) and that sons of fathers showing avoidant patterns of adjustment to marital distress may be particularly susceptible to internalizing disorders (Katz & Gottman, 1993). Finally, depression in adolescent females has been found to be strongly associated with maternal depression, whereas a lack of supportive early care appears to be more strongly associated with depression in adolescent males (Duggal, Carlson, Sroufe, & Egeland, 2001). It has also been found that the types of childrearing environments predicting resilience to adversity may differ for boys and girls. Resilience in boys is associated with households in which there is a male model (e.g., father, grandfather, older sibling), structure, rules, and some encouragement of emotional expressiveness. In contrast, resilient girls come from households that combine risk taking and independence with support from a female caregiver (e.g., mother, grandmother, older sister) (Werner, 1995). Zahn-Waxler et al. (1995) refer to the “gender paradox of comorbidities,” which is that although the prevalence of disruptive behavior is lower in females than in males, the risk of comorbid conditions such as anxiety is higher in female samples. In explaining this paradox, Zahn-Waxler et al. (1995) suggest that girls’ heightened level of interpersonal sensitivity, caring, and empathy may be a protective factor in insulating them from developing antisocial behavior. At the same time, girls’ overreceptivity to the plight of others, and their reluctance to assert their own needs in situations involving conflict and distress, may elevate their risk for the development of internalizing problems. However, the relations between gender and 1. A Developmental–Systems Perspective 15 comorbidity are likely to vary with the disorders under consideration, the age of the child, the source of referral, and other factors. For example, in contrast to Zahn-Waxler et al. (1995), Biederman et al. (2002) found that girls with ADHD had a significantly lower rate of comorbid major depression than did boys with ADHD. Although findings relating to sex differences and child psychopathology are complex, inconsistent, and frequently difficult to interpret, the cumulative findings from research strongly indicate that the effects of gender are critical to understanding the expression and course of most forms of childhood disorder (Bell-Dolan et al., in press; Kavanagh & Hops, 1994). It is particularly important to understand the processes and mechanisms underlying these gender effects, and to recognize that biological influences and differential socialization practices are likely to interact throughout development in accounting for any differences between the sexes that are found. Rural versus Urban Differences Although there is a general belief that rates of child behavior disorder are higher in urban than in rural areas, research findings in support of this view are weak and/or inconsistent. Findings from the Isle of Wight, Inner London Borough, and Ontario Child Health Studies reveal prevalence rates of problem behavior that were higher for urban than rural children (Offord et al., 1987; Rutter, 1981). On the other hand, in a crosscultural investigation, Weisz and Suwanlert (1991) found few differences in parent or teacher ratings of child problems as a function of rural versus urban status in either of the cultures that were studied (United States and Thailand). In a detailed analysis that controlled for the effects of SES and ethnicity and also looked at gradations of urbanization, Achenbach et al. (1991) found few differences in children’s behavior problems or competencies as a function of rural versus urban status, although there was a significant but very small effect indicating higher delinquency scores for children in urban environments. These investigators concluded that earlier findings of higher rates of problem behavior in urban than in rural areas “may have reflected the tendency to combine areas of intermediate urbanization with large urban areas for comparison with rural areas as well as a possible lack of control for demographic differences” (p. 86). Even in studies in which rural versus urban differences have been found, for the most part these differences were associated with economic and cultural differences between sites, and not with urbanization per se (Zahner, Jacobs, Freeman, & Trainor, 1993). Ethnicity and Culture Ethnicity Numerous terms have been used to describe ethnic influences. These include “ethnicity,” “race,” “ethnic identity,” “ethnic orientation,” “acculturation,” “bicultural orientation,” and “culture.” As pointed out by Foster and Martinez (1995), there is a need to recognize the diversity of terminology that has been used in describing ethnicity, and the fact that these terms refer to related but different things. Despite the growing ethnic diversity of the North American population, ethnic representation in research studies and the study of ethnicity-related issues more generally have received relatively little attention in studies of child psychopathology (Garc?a Coll, Akerman, & Cicchetti, 2000; U.S. Public Health Service, 2001c). In lamenting this state of affairs, Foster and Martinez (1995) state: “The underrepresentation of children from diverse backgrounds is accompanied by a dearth of empirical literature on the origins, correlates, and treatment of child psychopathology in different ethnic groups within the United States. Instead, investigators have based theories of child behavior, both normal and deviant, on data drawn largely from European-American culture” (p. 214). Research into child psychopathology has generally been insensitive to possible differences in prevalence, age of onset, developmental course, and risk factors related to ethnicity (Kazdin & Kagan, 1994), and to the considerable heterogeneity that exists within specific ethnic groups (Murry, Bynum, Brody, Willert, & Stephens, 2001; Murry, Smith, & Hill, 2001). In addition, few studies have compared ethnic groups while controlling for other important variables, such as SES, sex, age, and geographic region. In recent comparisons that have controlled for these variables, African American and Hispanic American children are identified and referred at the same rates as other children, but they are much less likely to actually receive specialty mental health services or psychotropic medications (Garc?a Coll & Garrido, 2000). European American and Native American children have been found to display similar mental health problems with the 16 I. INTRODUCTION exception of substance abuse, where rates are higher for Native American youngsters (Costello, Farmer, & Angold, 1999). Some studies that have included a small number of African American children in their samples have reported somewhat higher rates of externalizing problems for this group (Costello, 1989; Velez et al., 1989). However, other studies with much larger national samples that included European American, African American, and Hispanic American children have reported either no or very small differences related to race or ethnicity when SES, sex, age, and referral status were controlled for (Achenbach & Edelbrock, 1981; Achenbach et al., 1991; Lahey et al., 1995). So, although externalizing problems have been reported to be more common among African American children, this finding is probably an artifact related to SES. Externalizing disorder is associated with both ethnicity and SES, and since there is an overrepresentation of minority status children in low-SES groups in North America, caution must be exercised in interpreting the relationships among SES, ethnicity, and aggression (Guerra et al., 1995; Lahey et al., 1995). Ethnicity has not been found to be strongly associated with risk for eating disorders (Leon, Fulkerson, Perry, & Early-Zald, 1995), although differences between European Americans and other groups have been reported for such subclinical eating disturbances as dietary restraint, ideal body shape, and body dissatisfaction (Wildes & Emery, 2001). Differing patterns of substance abuse as a function of ethnicity have also been reported (Catalano et al., 1993). More research is needed, but these and other findings suggest that the effects of ethnicity are likely to vary with the problem under consideration and its severity. As is the case for SES and sex differences, global comparisons of the prevalence of different types of problems for different ethnic groups are not likely to be very revealing. On the other hand, studies into the processes affecting the form, associated factors, and outcomes of different disorders for various ethnic groups hold promise for increasing our understanding of the relationship between ethnicity and child psychopathology (e.g., Bird et al., 2001; Bradley, Corwyn, Burchinal, McAdoo, & Garc?a Coll, 2001). Culture The values, beliefs, and practices that characterize a particular ethnocultural group contribute to the development and expression of childhood distress and dysfunction, which in turn are organized into categories through cultural processes that further influence their development and expression (Harkness & Super, 2000; Wong & Ollendick, 2001). Through shared views about causality and intervention, culture also structures the way in which people and institutions react to a child’s problems. Since the meaning of children’s social behavior is influenced by cultural beliefs and values, it is not surprising that the form, frequency, and predictive significance of different forms of child psychopathology vary across cultures, or that cultural attitudes influence diagnostic and referral practices (Lambert & Weisz, 1992). For example, shyness and oversensitivity in children have been found to be associated with peer rejection and social maladjustment in Western cultures, but with leadership, school competence, and academic achievement in Chinese children in Shanghai (Chen, Rubin, & Li, 1995). Similarly, Lambert and Weisz (1989) found that overcontrolled problems were reported significantly more often for Jamaican than for American youngsters—a finding consistent with Afro-British Jamaican cultural attitudes and practices that discourage child aggression and other undercontrolled behavior, and that foster inhibition and other overcontrolled behavior. Weisz and Sigman (1993), using parent reports of behavioral and emotional problems in 11- to 15-year-old children from Kenya, Thailand, and the United States, found that Kenyan children were rated particularly high on overcontrolled problems (e.g., fears, feelings of guilt, somatic concerns), due primarily to numerous reports of somatic problems. In this mixed-race sample, whites were rated particularly high on undercontrolled problems (e.g., “arguing,” “disobedient at home,” “cruel to others”). Weisz and Suwanlert (1987) compared 6- to 11-year-old children in the Buddhist-oriented, emotionally controlled culture of Thailand with American 6- to 11-yearolds. Parent reports revealed Thai–U.S. differences in 54 problem behaviors, most of which were modest in magnitude. Thai children were rated higher than American children on problems involving overcontrolled behaviors such as anxiety and depression, whereas American children were rated higher than Thai children on undercontrolled behaviors such as disobedience and fighting. Weisz and Suwanlert (1991) compared ratings of behavior and emotional problems of 2- to 1. A Developmental–Systems Perspective 17 9-year-old children in Thailand and the United States. Parents and teachers in Thailand rated both overcontrolled and undercontrolled problems as less serious, less worrisome, less likely to reflect personality traits, and more likely to improve with time. These findings suggest that there may be cultural differences in the meanings ascribed to problem behaviors across cultures. Findings from these and other studies suggest that the expression of, and tolerance for, many child behavioral and emotional disturbances are related to social and cultural values. The processes that mediate this relationship are in need of further investigation. In this regard, it is important that research on child psychopathology not be generalized from one culture to another, unless there is support for doing so. There is some support for the notion that some processes—for example, those involved in emotion regulation and its relation to social competence—may be similar across diverse cultures (Eisenberg, Pidada, & Liew, 2001). The rates of expression of some disorders, particularly those with a strong neurobiological basis (e.g., ADHD, autistic disorder), may be less susceptible to cultural influences than others. However, even so, social and cultural beliefs and values are likely to influence the meaning given to these behaviors, the ways in which they are responded to, their forms of expression, and their outcomes. An important distinction to be made with respect to cross-cultural comparisons is whether or not there are real differences in the rates of the disorder, or differences in the criteria used to make judgments about these problems. For example, Weisz and Suwanlert (1989) compared the teacher-reported behavioral/emotional problems of Thai and U.S. children (ages 6–11 years). It was found that Thai teachers were confronted with students who were more prone to behavioral and emotional problems at school than were teachers in the United States, but that they applied different judgments to the behaviors they observed. Cultural factors are known to influence not only informal labeling processes but formal diagnostic practices as well. For example, reported prevalence rates of ADHD in Britain are much lower than in the United States, because of differences in the way in which diagnostic criteria for ADHD are applied in the two countries. Such differences in diagnostic practices may lead to spurious differences in reported prevalence rates for different forms of child psychopathology across cultures. Cross-cultural research on child psychopathology would suggest that the expression and experience of mental disorders in children are not universal (Fisman & Fisman, 1999). Patterns of onset and duration of illness and the nature and relationship among specific symptoms vary from culture to culture, and across ethnic groups within cultures (Hoagwood & Jensen, 1997). However, few studies have compared the attitudes, behaviors, and biological and psychological processes of children with mental disorders across different cultures. Such information is needed to understand how varying social experiences and contexts influence the expression, course, and outcome of different disorders across cultures. For example, greater social connectedness and support in more traditional cultures, and greater access to resources and opportunities in industrialized societies, are examples of mechanisms that may alter outcomes across cultures. Sensitivity to the role of cultural influences in child psychopathology has increased (Evans & Lee, 1998; Lopez & Guarnaccia, 2000), and is likely to continue to do so as globalization and rapid cultural change become increasingly more common (Garc?a Coll et al., 2000). BASIC ISSUES IN CHILD PSYCHOPATHOLOGY Several recurrent and overlapping issues have characterized the study of psychopathology in children (Rutter & Garmezy, 1983; Rutter & Sroufe, 2000). A number of these are highlighted in this section, including (1) difficulties in conceptualizing psychopathology and normality; (2) the need to consider healthy functioning and adjustment; (3) questions concerning developmental continuities and discontinuities; (4) the concept of developmental pathways; (5) the notions of risk and resilience; (6) the identification of protective and vulnerability factors; and (7) the role of contextual influences. Psychopathology versus Normality Conceptualizing child psychopathology and attempting to establish boundaries between what constitutes abnormal and normal functioning are arbitrary processes at best (Achenbach, 1997). Traditional approaches to mental disorders in children have emphasized concepts such as symptoms, diagnosis, illness, and treatment; by doing 18 I. INTRODUCTION so, they have strongly influenced the way we think about child psychopathology and related questions (Richters & Cicchetti, 1993). Childhood disorders have most commonly been conceptualized in terms of deviancies involving breakdowns in adaptive functioning, statistical deviation, unexpected distress or disability, and/or biological impairment. Wakefield (1992, 1997, 1999a) has proposed an overarching concept of mental disorder as “harmful dysfunction.” This concept encompasses a child’s physical and mental functioning, and includes both value- and science-based criteria. In the context of child psychopathology, a child’s condition is viewed as a disorder only if (1) it causes harm or deprivation of benefit to the child, as judged by social norms; and (2) it results from the failure of some internal mechanism to perform its natural function (e.g., “an effect that is part of the evolutionary explanation of the existence and structure of the mechanism”; Wakefield, 1992, p. 384). This view of mental disorder focuses attention on internally evolved mechanisms—for example, executive functions in the context of self-regulation (Barkley, 2001). Nevertheless, as pointed out by Richters and Cicchetti (1993), this view only identifies the decisions that need to be made in defining mental disorders; it does not specify how such decisions are to be made. As is the case for most definitions of mental disorder that have been proposed, questions related to defining the boundaries between normal and abnormal, understanding the differences between normal variability and dysfunction, defining what constitute “harmful conditions,” linking dysfunctions causally with these conditions, and circumscribing the domain of “natural” or of other proposed mechanisms are matters of considerable controversy (Lilienfeld & Marino, 1995; Richters & Cicchetti, 1993).4 Categories of mental disorder stem from human-made linguistic distinctions and abstractions, and boundaries between what constitutes normal and abnormal conditions, or between different abnormal conditions, are not easily drawn. Although it may sometimes appear that efforts to categorize mental disorders are carving “nature at its joints,” whether or not such “joints” actually exist is open to debate (e.g., Cantor, Smith, French, & Mezzich, 1980; Lilienfeld & Marino, 1995). However, clear joints do not necessarily need to exist for categorical distinctions to have utility. For instance, there is no joint at which one can carve day from night; yet distinguishing the two has proven incredibly useful to humans in going about their social discourse and engagements. Likewise, although the threshold for determining disorder from merely high levels of symptoms may be fuzzy, it could be stipulated as being at that point along a dimension where impairment in a major, culturally universal life activity befalls the majority of people at or exceeding that point. Thus, despite the lack of clear boundaries between what is normal and abnormal, categorical distinctions are still useful. Healthy Functioning The study of psychopathology in children requires concomitant attention to adaptive developmental processes for several reasons. First, judgments of deviancy require knowledge of normative developmental functioning, both with respect to a child’s performance relative to sameage peers and with respect to the child’s own baseline of development. Second, maladaptation and adaptation often represent two sides of the same coin, in that dysfunction in a particular domain of development (e.g., the occurrence of inappropriate behaviors) is usually accompanied by a failure to meet developmental tasks and expectations in the same domain (e.g., the nonoccurrence of appropriate behaviors). It is important to point out, however, that adaptation should not be equated with the mere absence of psychopathology. Kendall and his colleagues (Kendall, Marrs-Garcia, Nath, & Sheldrick, 1999; Kendall & Sheldrick, 2000), for instance, contend that it is important to use normative comparisons to evaluate treatment outcome; they suggest that improvement involves falling within a certain range of healthy functioning, in addition to the amelioration of one’s symptom presentation. Moreover, adaptation involves the presence and development of psychological, physical, interpersonal, and intellectual resources (see Fredrickson, 2001). Third, in addition to the specific problems that lead to referral and diagnosis, disturbed children are likely to show impairments in other areas of adaptive functioning. For example, in addition to their core symptoms of impulsivity and inattention, children with ADHD also show lower-thanaverage levels of functioning in their socialization, communication, and activities of daily living (e.g., Stein, Szumowski, Blondis, & Roizen, 1995). Fourth, most children with specific disorders are known to cope effectively in some areas of their 1. A Developmental–Systems Perspective 19 lives. Understanding a child’s strengths informs our knowledge of the child’s disorder and provides a basis for the development of effective treatment strategies. Fifth, children move between pathological and nonpathological forms of functioning over the course of their development. Individual children may have their “ups and downs” in problem type and frequency over time. Sixth, many child behaviors that are not classifiable as deviant at a particular point in time may nevertheless represent less extreme expressions or compensations of an already existing disorder or early expressions of a later progression to deviant extremes as development continues (Adelman, 1995). Finally, no theory of a childhood disorder is complete if it cannot be linked with a theory of how the underlying normal abilities develop and what factors go awry to produce the disordered state. Therefore, understanding child psychopathology requires that we also attend to these less extreme forms of difficulty and develop more complete models of the normal developmental processes underlying the psychopathology. For these and other reasons to be discussed, the study of child psychopathology requires an understanding of both abnormal and healthy functioning. As noted by Cicchetti and Richters (1993), “it is only through the joint consideration of adaptive and maladaptive processes within the individual that it becomes possible to speak in meaningful terms about the existence, nature, and boundaries of the underlying psychopathology” (p. 335). To date, far greater attention has been devoted to the description and classification of psychopathology in children than to healthy child functioning; to nonpathological psychosocial problems related to emotional upset, misbehavior, and learning; or to factors that promote the successful resolution of developmental tasks (Adelman, 1995; Sonuga-Barke, 1998). In light of this imbalance, there is a need for studies of normal developmental processes (Lewis, 2000), for investigations of normative and representative community samples of children (Ialongo, Kellam, & Poduska, 2000; Kazdin, 1989), and for studies of “resilient” children who show normal development in the face of adversity (Masten, 2001). Developmental Continuities and Discontinuities A central issue for theory and research in child psychopathology concerns the continuity of disorders identified from one time to another and the relationship between child, adolescent, and adult disorders (Caspi, 2000; Garber, 1984; Kazdin & Johnson, 1994; Rutter & Rutter, 1993; Sroufe & Jacobvitz, 1989). Over the past two decades, research into early attachment has stimulated general interest in the roles of relational processes and internalized representational systems as the bases for understanding continuities and discontinuities in psychopathology over time and across generations (Cassidy & Shaver, 1999; Lyons-Ruth, 1995; Sroufe, Duggal, Weinfeld, & Carlson, 2000). Some childhood disorders, such as mental retardation and autistic disorder, are chronic conditions that will persist throughout childhood and into adulthood. Other disorders, such as functional enuresis and encopresis, occur during childhood and only rarely manifest themselves in adults (Walker et al., 1989). And other disorders (e.g., mood disorders, schizophrenia, generalized anxiety disorder) are expressed, albeit in modified forms, in both childhood and adulthood and exhibit varying degrees of continuity over time. Evidence in support of the continuity between child and adult disorders is equivocal and depends on a number of methodological factors related to research design, assessment instruments, the nature of the study sample, and the type and severity of the disorder (Garber, 1984). In general, the literature suggests that child psychopathology is continuous with adult disorders for some, but not all, problems. As we discuss below, there is evidence that appears to favor the stability of externalizing problems over internalizing problems. However, previous findings may reflect the severity and pervasiveness of the disorders assessed, referral biases, and the fact that longitudinal investigations of children with internalizing and other disorders are just beginning to emerge. For example, one study found that firstgrade anxious symptoms predicted levels of anxious symptoms and adaptive functioning in fifth grade (Ialongo, Edelsohn, Werthamer-Larsson, Crockett, & Kellam, 1995). In another report, early-onset bulimia nervosa was associated with a 9-fold increase in risk for late-adolescent bulimia nervosa and a 20-fold increase in risk for adult bulimia nervosa (Kotler et al., 2001). The possible mechanisms underlying the relationships between early maladaptation and later disordered behavior are numerous and can operate in both direct and indirect ways (Garber, 1984; Rutter, 1994a; Sroufe & Rutter, 1984). Some examples of direct relationships between early and 20 I. INTRODUCTION later difficulties include (1) the development of a disorder during infancy or childhood, which then persists over time; (2) experiences that alter the infant’s or child’s physical status (e.g., neural plasticity), which in turn influences later functioning (Courchesne, Chisum, & Townsend, 1994; Johnson, 1999; Nelson, 2000); and (3) the acquisition of early patterns of responding (e.g., compulsive compliance, dissociation) that may be adaptive in light of the child’s current developmental level and circumstances, but may result in later psychopathology when circumstances change and new developmental challenges arise. Some examples of indirect associations between child and adult psychopathology may involve early predispositions that eventually interact with environmental experiences (e.g., stressors), the combination of which leads to dysfunction. For example, Egeland and Heister (1995) found that the impact of day care on disadvantaged high-risk children at 42 months of age was related to the children’s attachment quality at 12 months of age, with securely attached children more likely to be negatively affected by early out-of-home care. Other examples of indirect links between child and adult disturbance include (1) experiences (e.g., peer rejection) that contribute to an altered sense of self-esteem (DuBois & Tevendale, 1999), or that create a negative cognitive set, which then leads to later difficulties; and (2) experiences providing various opportunities or obstacles that then lead to the selection of particular environmental conditions, and by doing so guide a child’s course of development (Rutter, 1987; Sroufe & Rutter, 1984). Research efforts have focused not only on the continuities and discontinuities in childhood disorders, but also on the identification of factors that predict them. One factor that has been studied in the context of conduct disorder is age of onset, with early onset usually viewed as the occurrence of conduct disorder symptoms prior to age 12 years (Loeber & Dishion, 1983; O’Donnell, Hawkins, & Abbott, 1995). It has been found that early onset of symptoms is associated with higher rates and more serious antisocial acts over a longer period of time for both boys and girls (Lavigne et al., 2001). However, psychosocial variables that are present prior to and following onset may influence the seriousness and chronicity more than age of onset per se does (Tolan & Thomas, 1995). A question that needs to be addressed is this: Does early age of onset operate in a causal fashion for later problems, and if so, how? Another issue is whether the causal processes that are associated with an early onset of a disorder (e.g., depression) are different from those that serve to maintain the disorder. Even then, the specification of an age of onset need not be made so precisely that it creates a false distinction that only valid cases meet that precise threshold, as may have happened with ADHD (see Barkley, chapter 2, this volume). Such efforts to impose precision where none exists may have backfired in hampering studies of teens and adults having the same disorder who cannot adequately recall such a precise onset, and in presuming that cases having qualitatively identical symptoms and impairments but later onsets are invalid instances of a disorder. Although research supports the notion of continuity of disorders, it does not support the continuity of identical symptoms over time (i.e., “homotypic correspondence”). Continuity over time for patterns of behavior rather than for specific symptoms is the norm. For example, although externalizing disorders in boys are stable over time, the ways in which these behavioral patterns are expressed are likely to change dramatically over the course of development (Olweus, 1979). Even with wide fluctuations in the expression of behavior over time, “children may show consistency in their general adaptive or maladaptive pattern of organizing their experiences and interacting with the environment” (Garber, 1984, p. 34). Several research findings can be used to illustrate this notion of consistent “patterns of organization.” For example, early heightened levels of behavioral inhibition may affect later adjustment by influencing the way in which a child adapts to new and unfamiliar situations and the ensuing person–environment interactions over time (Kagan, 1994a). Another example of a consistent pattern of organization involves early attachment quality and the development of internal working models that children carry with them into their later relationships (Bowlby, 1988; Goldberg, 1991). Internal working models of self and relationships may remain relatively stable over time, at the same time that the behavioral expressions of these internal models change with development. From a neuroscientific perspective, Pennington and Ozonoff (1991) argue that certain genes and neural systems also play a significant predisposing role in influencing the continuity of psychopathology, and that the “discontinuities at one level of analyses—that of observable behavior—may mask continuities at 1. A Developmental–Systems Perspective 21 deeper levels of analysis; those concerned with the mechanisms underlying observable behavior” (p. 117). Given that developmental continuity is reflected in general patterns of organization over time rather than in isolated behaviors or symptoms, the relationships between early adaptation and later psychopathology are not likely to be direct or uncomplicated. The connections between psychopathology in children and adults are marked by both continuities and discontinuities. The degree of continuity–discontinuity will vary as a function of changing environmental circumstances and transactions between a child and the environment that affect the child’s developmental trajectory. Developmental Pathways The concept of “developmental pathways” is crucial for understanding continuities and discontinuities in psychopathology. Such pathways are not directly observable, but function as metaphors that are inferred from repeated assessments of individual children over time (Loeber, 1991). A pathway, according to Loeber (1991), “defines the sequence and timing of behavioral continuities and transformations and, ideally, summarizes the probabilistic relationships between successive behaviors” (p. 98). In attempting to identify developmental pathways as either “deviant” or “normal,” it is important to recognize that (1) different pathways may lead to similar expressions of psychopathology (i.e., “equifinality”); and (2) similar initial pathways may result in different forms of dysfunction (i.e., “multifinality”), depending on the organization of the larger system in which they occur (Cicchetti & Rogosch, 1996; Lewis, 2000; Loeber, 1991). Research findings related to child maltreatment provide an example of a possible developmental pathway. It has been found that physically abused children are more likely to develop insecure attachments, view interpersonal relationships as being coercive and threatening, become vigilant and selectively attend to hostile cues, instantly classify others as threatening or nonthreatening, and acquire aggressive behavioral strategies for solving interpersonal problems (see Cicchetti & Manly, 2001). These children bring representational models to peer relationships that are negative, conflictual, and unpredictable. They process social information in a biased and deviant manner, and develop problems with peer relationships that involve social withdrawal, unpopularity, and overt social rejection by peers (Dodge, Pettit, & Bates, 1994). In another example of a developmental pathway, the diagnosis of conduct disorder typically precedes the initiation of use of various substances, and this use in turn precedes the diagnosis of alcohol dependence in adolescents (Kuperman et al., 2001). The systematic delineation of developmental pathways not only offers several advantages for the study of the etiology and outcomes of childhood disorders, but may also suggest strategies for intervention. Loeber (1991, p. 99) describes these advantages as “attempts to capture the changing manifestations and variable phenotype of a given disorder” over time. In this way, the study of developmental pathways includes etiological considerations, the assessment of comorbidities as they accrue over time, and a sensitivity to diverse outcomes (e.g., White, Bates, & Buyske, 2001). Risk and Resilience Previous studies of child psychopathology focused on elucidating the developmental pathways for deviancy and maladjustment to the relative exclusion of those for competency and adjustment (but see Luthar, 1993; Rutter, 1985, 1987, 1994b; and Rutter & Rutter, 1993, for exceptions). However, a significant number of children who are at risk do not develop later problems. There is a growing recognition of the need to examine not only risk factors, but also those conditions that protect vulnerable children from dysfunction and lead to successful adaptations despite adversity (Cicchetti & Garmezy, 1993). “Resilience,” which refers to successful adaptations in children who experience significant adversity, has now received a good deal of attention (Luthar, Cicchetti, & Becker, 2000). Early patterns of adaptation influence later adjustment in complex and reciprocal ways. Adverse conditions, early struggles to adapt, and failure to meet developmental tasks do not inevitably lead to a fixed and unchanging abnormal path. Rather, many different factors, including chance events and encounters, can provide turning points whereby success in a particular developmental task (e.g., educational advances, peer relationships) shifts a child’s course onto a more adaptive trajectory. Conversely, there are numerous events and circumstances and underlying dynamic biological systems that may deflect the child’s 22 I. INTRODUCTION developmental trajectory toward maladaptation (e.g., a dysfunctional home environment, peer rejection, difficulties in school, parental psychopathology, intergenerational conflict, and even late-onset genetic effects). Although the term “resilience” has not been clearly operationalized, it is generally used to describe children who (1) manage to avoid negative outcomes and/or to achieve positive outcomes despite being at significant risk for the development of psychopathology; (2) display sustained competence under stress; or (3) show recovery from trauma (Werner, 1995). Risk is usually defined in terms of child characteristics that are known to be associated with negative outcomes—for example, difficult temperament (Ingram & Price, 2001; Rothbart, Ahadi, & Evans, 2000)—and/or in terms of a child’s exposure to extreme or disadvantaged environmental conditions (e.g., poverty or abuse). Individual children who are predisposed to develop psychopathology and who show a susceptibility to negative developmental outcomes under high-risk conditions are referred to as “vulnerable.” Genetic makeup and temperament are two factors that are presumed to contribute to susceptibility for children who are exposed to high-risk environments (Rutter, 1985; Seifer, 2000). Research on resilience has lacked a consistent vocabulary, conceptual framework, and methodological approach (Luthar et al., 2000; Rutter, 2000c; Zimmerman & Arunkumar, 1994). It is particularly important to ensure that resilience is not defined as a universal, categorical, or fixed attribute of a child, but rather as a number of different types of dynamic processes that operate over time. Individual children may be resilient in relation to some specific stressors but not others, and resilience may vary over time and across contexts (Freitas & Downey, 1998). As noted by Zimmerman and Arunkumar (1994, p. 4), “research on resiliency can only identify those particular risk circumstances when environmental conditions, individual factors, and developmental tasks interact to help children and adolescents avoid negative consequences.” Fortunately, models of resilience have increasingly begun to address the complex and dynamic relationships between the child and his or her environment, to incorporate the theoretical and empirical contributions of developmental psychology, and to acknowledge the multiple factors related to normal and deviant behavior (Glantz & Johnson, 1999; Walden & Smith, 1997; Tebes, Kaufman, Adnopoz, & Racusin, 2001). One problem in research on resilience has been an absence of agreed-upon criteria for defining positive developmental outcomes (see Kaufman, Cook, Arny, Jones, & Pittinsky, 1994, for a review of the ways in which positive outcomes in studies of resilience have been operationalized). For example, there is currently debate as to whether the criteria for defining resilience and adaptation should be based on evidence from external criteria (e.g., academic performance), internal criteria (e.g., subjective well-being), or some combination of these (see Masten, 2001). Variations across studies in the source of information (e.g., parent or teacher); the type of assessment method (e.g., interview, questionnaire, observation); the adaptational criteria used; and the number and timing of assessments can easily influence the proportion of children who are designated as resilient or not in any particular investigation (Kaufman et al., 1994; Masten, 2001). And there is also some confusion about and circularity in how the term “resilience” has been used, in that it has been used to refer to both an outcome and to the cause of an outcome. Several different models of resilience have also been proposed, the most common ones being a compensatory model, a challenge model (e.g., stress inoculation), and a protective-factors model (Garmezy, Masten, & Tellegen, 1984). Years of research suggest that resilience is not indicative of any rare or special qualities of the child per se (as implied by the term “the invulnerable child”), but rather is the result of the interplay of normal developmental processes such as brain development, cognition, caregiver–child relationships, regulation of emotion and behavior, and the motivation for learning (Masten, 2001). Some researchers have argued that resilience may be more ubiquitous than previously thought, and that this phenomenon is part of the “ordinary magic” and makeup of basic human adaptation (Masten, 2001; Sheldon & King, 2001). It is when these adaptational systems are impaired, usually through prolonged or repeated adversity, that the risk for childhood psychopathology increases. Protective and Vulnerability Factors Various protective and vulnerability factors have been found to influence children’s reactions to potential risk factors or stressors. These include factors within the child, the family, and the community (Osofsky & Thompson, 2000; Werner & 1. A Developmental–Systems Perspective 23 Smith, 1992). Common risk factors that have been found to have adverse effects on a child encompass both acute stressful situations and chronic adversity; they include such events as chronic poverty, serious caregiving deficits, parental psychopathology, death of a parent, community disasters, homelessness, reduced social support, decreased financial resources, family breakup, parental marital/ couple conflict, and perinatal stress (Deater- Deckard & Dunn, 1999; Rutter, 1999; Tebes et al., 2001; Walden & Smith, 1997). Protective factors within a child that have been identified include an “easy” temperament (i.e., a child who is energetic, affectionate, cuddly, goodnatured, and/or easy to deal with), which makes the child engaging to other people; early coping strategies that combine autonomy with help seeking when needed; high intelligence and scholastic competence; effective communication and problem-solving skills; positive self-esteem and emotions; high self-efficacy; and the will to be or do something (Fredrickson, 2001; Gilgun, 1999; Werner, 1995). An example of a possible protective factor within the child is seen in findings that high vagal tone and vagal suppression—taken as indices of a child’s ability to regulate emotion via self-soothing, focused attention, and organized and goal-directed behavior—can buffer children from the increases in externalizing behaviors, internalizing behaviors, and social problems often associated with exposure to parental marital/couple hostility and discord (Katz & Gottman, 1995) or parental problem drinking (El-Sheikh, 2001). At a family level, protective factors that have been identified include the opportunity to establish a close relationship with at least one person who is attuned to the child’s needs, positive parenting, availability of resources (e.g., child care), a talent or hobby that is valued by adults or peers, and family religious beliefs that provide stability and meaning during times of hardship or adversity (Werner & Smith, 1992). Protective factors in the community include extrafamilial relationships with caring neighbors, community elders, or peers; an effective school environment, with teachers who serve as positive role models and sources of support; and opening of opportunities at major life transitions (e.g., adult education, voluntary military service, church or community participation, a supportive friend or marital/relationship partner). In summary, early patterns of adaptation influence later adjustment in complex and reciprocal ways. Adverse conditions, early adaptational struggles, and failure to meet developmental tasks do not inevitably lead to a fixed and unmalleable dysfunctional path. Rather, as noted earlier, many different factors can act to alter a child’s developmental course for the better. Conversely, numerous events and circumstances may serve to alter this course for the worse. The interrelated issues of developmental continuities– discontinuities, developmental pathways, risk and resilience, and vulnerability and protective factors are far from being resolved or clearly understood. The multitude of interdependent and reciprocal influences, mechanisms, and processes involved in the etiology and course of child psychopathology clearly suggest a need for more complex theories (e.g., chaos theory, nonlinear dynamic models) (Barton, 1994; Glantz & Johnson, 1999; Gottman, Guralnick, Wilson, Swanson, & Murray, 1997; Haynes & Blaine, 1995), research designs, and data-analytic strategies (Kazdin & Kagan, 1994; Mash & Krahn, 2000; Richters, 1997). Contextual Influences Messick (1983) cogently argues that any consideration of child psychopathology must consider and account for three sets of contextual variables: (1) the child as context—the idea that unique child characteristics, predispositions, and traits influence the course of development; (2) the child of context—the notion that the child comes from a background of interrelated family, peer, classroom, teacher, school, community, and cultural influences; and (3) the child in context—the understanding that the child is a dynamic and rapidly changing entity, and that descriptions taken at different points in time or in different situations may yield very different information. Research has increasingly come to recognize the reciprocal transactions between the developing child and the multiple social and environmental contexts in which development occurs (Cicchetti & Aber, 1998; Deater-Deckard, 2001). Understanding context requires a consideration of events that impinge directly on the child in a particular situation at a particular point in time; extrasituational events that affect the child indirectly (e.g., a parent’s work-related stress); and temporally remote events that continue to affect the child through their representation in the child’s current cognitive–affective data base. Defining context has been, and continues to be, a matter of some complexity. The context of 24 I. INTRODUCTION maltreatment provides an illustration of difficulties in definition. Maltreatment can be defined in terms of its type, timing, frequency, severity, and chronicity in the family (e.g., Manly, Kim, Rogosch, & Cicchetti, 2001). Each of these parameters and their interaction may contribute to child outcomes, but in different ways. For example, Manly, Cicchetti, and Barnett (1994) studied different types of maltreatment and found that outcomes generally did not differ for children who were categorized as neglected versus abused. However, a regression analysis indicated that neglect accounted for more of the variance in child problems than other types of abuse did. In this study, sexually abused children were also found to be more socially competent than children exposed to other forms of maltreatment. This may reflect a lack of chronicity associated with sexual abuse, or it may suggest that problems related to sexual abuse may not reveal themselves until later periods in a child’s development, when issues concerning sexuality become more salient. Other studies have found that psychological maltreatment and emotional abuse account for most of the distortions in development attributed to maltreatment in general, and have the most negative consequences for a child (Crittenden, Claussen, & Sugarman, 1994). The example of maltreatment illustrates how contexts for development encompass heterogeneous sets of circumstances, and how child outcomes may vary as a function of (1) the configuration of these circumstances over time, (2) when and where outcomes are assessed, and (3) the specific aspects of development that are affected. More precise definitions are needed if the impact of maltreatment, or for that matter any contextual event (e.g., parent disciplinary styles, family support, intellectual stimulation), is to be understood. Even for those forms of child psychopathology for which there are strong neurobiological influences, the expression of the disorder is likely to interact with contextual demands. For example, Iaboni, Douglas, and Baker (1995) found that although the overall pattern of responding shown by children with ADHD was indicative of a generalized inhibitory deficit, the self-regulatory problems of these children became more evident with continuing task demands for inhibition and/or deployment of effort. Likewise, tasks having high interest value or high external incentives may moderate these children’s typically deficient performance on less interesting or low incentive tasks (Carlson & Tamm, 2000; Slusarek, Velling, Bunk, & Eggers, 2001). Child psychopathology research has increasingly focused on the role of the family system, the complex relationships within families, and the reciprocal influences among various family subsystems (Fiese, Wilder, & Bickham, 2000). There is a need to consider not only the processes occurring within disturbed families, but the common and unique ways in which these processes affect both individual family members and subsystems. Within the family, the roles of the mother–child and marital/couple subsystems have received the most research attention to date, with less attention given to the roles of siblings (Hetherington, Reiss, & Plomin, 1994) and fathers (Lamb & Billings, 1997; Phares & Compas, 1992). For the most part, research into family processes and child psychopathology has not kept pace with family theory and practice, and there is a need for the development of sophisticated methodologies and valid measures that will capture the complex relationships hypothesized to be operative in disturbed and normal family systems (Bray, 1995; Bray, Maxwell, & Cole, 1995). This task is complicated by a lack of consensus concerning how dysfunctional or healthy family functioning should be defined, what specific family processes are important to assess (Bray, 1994; Mash & Johnston, 1995), or the extent to which such measures of family environment reflect true environmental effect or shared genetic influences between parent and child (Plomin, 1995). DEFINING CHILD PSYCHOPATHOLOGY There has been, and continues to be, a lack of consensus concerning how psychopathology in children should be defined (Silk et al., 2000; Sonuga-Barke, 1998). Although the situation is improving, comparisons of findings across studies are extremely difficult to make, because of the idiosyncratic ways in which samples of children have been constituted. For example, children described as “hyperactive” in previous studies have varied widely with respect to their symptoms and conditions, problem severities, comorbidities, and levels of cognitive functioning. More recently, researchers and clinicians have come to define child psychopathology using stan1. A Developmental–Systems Perspective 25 dardized diagnostic systems such as DSM-IV (American Psychiatric Association [APA], 1994, 2000) and the International Classification of Diseases, 10th revision (ICD-10; WHO, 1992). The diagnostic criteria utilized in DSM-IV are the ones most commonly used in North America, and these are presented for the individual disorders described in each of the chapters of this volume. However, the increased use and acceptance of DSM-IV should not be taken as an indication of widespread agreement regarding the fundamental nature of what constitutes psychopathology in children or the specific criteria used to define it (cf. Achenbach, 1997; Cantwell, 1996; Follette & Houts, 1996; Scotti, Morris, McNeil, & Hawkins, 1996). In many ways, the increased use of DSMIV seems to reflect a degree of resignation on the part of researchers and clinicians concerning the prospects for developing a widely agreed-upon alternative approach, combined with a growing consensus regarding the need to achieve a greater level of standardization (albeit an imperfect one) in defining childhood disorders. Several fundamental questions have characterized most discussions concerning how child psychopathology should be defined: 1. Should child psychopathology be viewed as a disorder that occurs within the individual child, as a relational disturbance, as a reaction to environmental circumstances, or as some combination of all of these? 2. Does child psychopathology constitute a condition qualitatively different from normality (aberration), an extreme point on a continuous trait or dimension, a delay in the rate at which a normal trait would typically emerge, or some combination of the three? How are “subthreshold” problems to be handled? 3. Can homogeneous disorders be identified, or is child psychopathology best defined as a configuration of co-occurring disorders or as a profile of traits and characteristics? 4. Can child psychopathology be defined as a static entity at a particular point in time, or do the realities of development necessitate that it be defined as a dynamic and ongoing process that expresses itself in different ways over time and across contexts? 5. Is child psychopathology best defined in terms of its current expression, or do definitions also need to incorporate nonpathological conditions that may constitute risk factors for later problems? There are currently no definitive answers to these questions. More often, the way in which they are answered reflects theoretical or disciplinary preferences and specific purposes and goals (e.g., defining samples for research studies, or determining program or insurance eligibility). Psychopathology as Adaptational Difficulty As we have noted earlier, a common theme in defining child psychopathology has been that of adaptational difficulty or failure (Garber, 1984; Mash, 1998). Sroufe and Rutter (1984) note that regardless of whether “particular patterns of early adaptation are to a greater or lesser extent influenced by inherent dispositions or by early experience, they are nonetheless patterns of adaptation” (p. 23). Developmental competence is reflected in a child’s ability to use internal and external resources to achieve a successful adaptation (Masten & Curtis, 2000; Waters & Sroufe, 1983), and problems occur when the child fails to adapt successfully. Even with wide variations in terminology and proposed explanatory mechanisms across theories, there is general agreement that maladaptation represents a pause, a regression, or a deviation in development (Garber, 1984; Simeonsson & Rosenthal, 1992). In conceptualizing and defining psychopathology as adaptational difficulty, it is also essential to conceptualize and identify the specific developmental tasks that are important for children at various ages and periods of development, and the many contextual variables that derive from and surround the child (Garber, 1984; Luthar et al., 1997; Mash, 1998). In this regard, the study of psychopathology in children and the study of development and context are for all intents and purposes inseparable (Cicchetti & Aber, 1998). In determining whether a given behavior should be considered to be deviant in relation to stage-salient developmental issues, Garber (1984) stresses the need to understand several important parameters. The first, “intensity,” refers to the magnitude of behavior as excessive or deficient. The second, “frequency,” refers to the severity of the problem behavior, or how often it does or does not occur. Third, the “duration” of behavior must be considered. Some difficulties are transient and spontaneously remit, whereas others persist over time. To these parameters, we would add a qualitative parameter reflecting how 26 I. INTRODUCTION grossly atypical the behavior may be (e.g., some of the complex compulsions seen in Tourette’s disorders), such that even low-intensity, lowfrequency, and short-duration behavior may be so bizarre as to constitute “psychopathology.” It is crucial that the intensity, frequency, duration, and atypicality of the child’s behavior be appraised with respect to what is considered normative for a given age. The final parameter of deviance concerns the “number of different symptoms” and their “configuration.” Each of these parameters is central to research and theory, and to one’s specific definition of adaptational failure, regression, stagnation, or deviation. Social Judgment The diagnosis of psychopathology in children is almost always a reflection of both the characteristics and behavior of the child and of significant adults and professionals (Lewis, 2000). Research findings utilizing behavior problem checklists and interviews indicate that there can be considerable disagreement across informants (e.g., parents, teachers, professionals) concerning problem behaviors in children (Achenbach, McConaughy, & Howell, 1987; Feiring & Lewis, 1996). Mothers typically report more problems than do fathers (e.g., Achenbach et al., 1991), and across a range of domains, teachers identify more problems than other informants do in assessing the same domains. For example, in a study with maltreated children, only 21% of children were classified as resilient by teachers, whereas 64% of children were so classified based on reports from other sources (Kaufman et al., 1994). Issues regarding disagreement–agreement among informants are complicated by the fact that the amount of agreement will vary with the age and sex of the child (Offord et al., 1989), the nature of the problem being reported on (e.g., internalizing vs. externalizing), the method used to gather information (e.g., interview vs. questionnaire), and the informants being compared. For example, Tarullo, Richardson, Radke-Yarrow, and Martinez (1995) found that both mother– child and father–child agreement was higher for preadolescent than for adolescent children and, in a meta-analysis, Duhig, Renk, Epstein, and Phares (2000) reported higher mother–father agreement for externalizing than for internalizing problems. Disagreements among informants create methodological difficulties in interpreting epidemiological data when such data are obtained from different sources, and also in how specific diagnoses are arrived at in research and practice. Also of importance is how disagreements among informants are interpreted. For example, disagreements may be viewed as (1) reflections of bias or error on the part of one informant; (2) evidence for the variability of children’s behavior across the situations in which they are observed by others; (3) lack of access to certain types of behavior (i.e., private events) on the part of one informant; (4) denial of the problem; or (5) active distortion of information in the service of some other goal (e.g., defensive exclusion, treatment eligibility). Parental psychopathology may “color” descriptions of child problems—as may occur when abusive or depressed mothers provide negative or exaggerated descriptions of their children (Gotlib & Hammen, 1992; Mash et al., 1983; Richters, 1992), or when dismissive/avoidant adult informants deny the presence of emotional problems at the same time that professionals observe a high level of symptoms (Dozier & Lee, 1995). These latter types of problems in reporting may be especially likely, given the frequent lack of correspondence between the expression and the experience of distress for many child and adult disturbances. Hypothesized relationships between parental psychopathology and reports of exaggerated child symptoms have received mixed support. For example, some studies have failed to find evidence for distorted reports by depressed mothers (Tarullo et al., 1995). TYPES OF CHILD PSYCHOPATHOLOGY The types of problems for which children are referred for treatment are reflected in the different approaches that have been used to conceptualize and classify these problems. Among the more common of these approaches are the following: 1. General and specific behavior problem checklists, which enumerate individual child symptoms— for example, the Child Behavior Checklist (Achenbach, 1991) and the Children’s Depression Inventory (Kovacs & Beck, 1977). 2. Dimensional approaches, which focus on symptom clusters or syndromes derived from behavior problem checklists—for example, the Child Behavior Checklist and Profile (Achenbach, 1993). 1. A Developmental–Systems Perspective 27 3. Categorical approaches, which use predetermined diagnostic criteria to define the presence or absence of particular disorders—for example, the DSM-IV (APA, 1994) and ICD- 10 (WHO, 1992). 4. A multiple-pathway, developmental approach, which emphasizes developmental antecedents and competencies both within the child and the environment that contribute to (mal)adjustment and (mal)adaptation (Sroufe, 1997). Issues related to the use of these different classification approaches are discussed in a later section of this chapter. What follows is a brief overview of the types of problem behaviors, dimensions, and disorders that occur during childhood and that are the topics of this volume’s other chapters. Individual Symptoms The individual behavioral and emotional problems (i.e., symptoms) that characterize most forms of child psychopathology have been found to occur in almost all children at one time or another during their development (e.g., Achenbach & Edelbrock, 1981; Achenbach et al., 1991; MacFarlane et al., 1954). When taken in isolation, specific symptoms have generally shown little correspondence to a child’s overall current adjustment or to later outcomes. This is the case even for many symptoms previously hypothesized to be significant indicators of psychopathology in children—for example, thumbsucking after 4 years of age (Friman, Larzelere, & Finney, 1994). Usually the age-appropriateness, clustering, and patterning of symptoms are what serve to define child psychopathology, rather than the presence of individual symptoms. Many of the individual behavior problems displayed by children referred for treatment are similar to those that occur in less extreme forms in the general population or in children of younger ages. For example, Achenbach et al. (1991) found that although referred children scored higher than nonreferred children on 209 of 216 parent-rated problems, only 9 of the 209 items showed effects related to clinical status that were considered to be large (accounting for more than 13.8% of the variance), according to criteria specified by Cohen (1988). To illustrate the kinds of individual symptoms that are more common in referred than in nonreferred children, individual parent-reported symptoms that accounted for 10% or more of the variance in clinical status in the Achenbach et al. (1991) study are shown in Table 1.1. It can be seen that even the problems that best discriminated between referred and nonreferred children are relatively common behaviors that occur to some extent in all children—they are not particularly strange or unusual behaviors. In addition, most individual problem behaviors (approximately 90% of those on behavior problem checklists) do not, by themselves, discriminate between groups of clinicreferred and nonreferred children. Nondiscriminating items include some problems for children in both groups that are relatively common (e.g., “brags,” “screams”) and others that occur less frequently (e.g., “sets fires,” “bowel movements outside the toilet”). Dimensions of Child Psychopathology A second approach to describing child psychopathology identifies symptom clusters or “syndromes” derived through the use of multivariate statistical procedures, such as factor analysis or cluster analysis (e.g., Achenbach, 1993, 1997; McDermott, 1993; McDermott & Weiss, 1995). Research has identified two broad dimensions of child psychopathology—one reflecting “externalizing” or “undercontrolled” problems, and the other reflecting “internalizing” or “overcontrolled” problems (Reynolds, 1992). The externalizing dimension encompasses behaviors often thought of as directed at others, whereas the internalizing dimension describes feelings or states that are commonly viewed as “innerdirected.” Within the two broad dimensions of externalizing and internalizing disorders are specific subdimensions or syndromes. Some subdimensions of child psychopathology that have commonly been identified in research are presented in Table 1.2. They include “withdrawn,” “somatic complaints,” “anxious/depressed,” “social problems,” “thought problems,” “attention problems,” “delinquent behavior,” and “aggressive behavior” (Achenbach, 1993). Examples of the specific problem behaviors constituting each of these subdimensions are also included in Table 1.2. The particular subdimensions that are identified may vary from study to study as a function of the item pool from which they are derived, the age and sex of children in the sample, the methods of assessment, and the informants. 28 I. INTRODUCTION Taxometric efforts have also described groups of children in terms of consistently identified profiles of scores on the various syndromes (Achenbach, 1993). Such profiles have been reliably identified and appear to have promise in addressing problems related to comorbidity (see the section on comorbidity, below). At present, however, our nomenclature for describing these profiles is limited, and they have yet to be widely validated or used in clinical research and practice. Categories of Child Psychopathology The DSM-IV diagnostic system (APA, 1994, 2000) provides comprehensive coverage of the general types of symptom clusters displayed by children characterized as having mental disorders. To illustrate, DSM-IV categories that apply to children are listed in Tables 1.3 to 1.6. These tables are not intended to be exhaustive of all DSM-IV diagnoses that may apply to children. Rather, they are intended to provide an overview of the range and variety of disorders that typically occur during childhood. Specific DSM-IV disorders and their subtypes are discussed in detail in the subsequent chapters of this volume. Table 1.3 lists the DSM-IV categories for developmental and learning disorders, including mental retardation, pervasive developmental disorders (e.g., autistic disorder), specific problems related to reading and mathematics, and communication difficulties. Many of these disorders constitute chronic conditions that often reflect deficits in capacity rather than performance difficulties per se. Table 1.4 lists DSM-IV categories for other disorders that are usually first diagnosed in infancy, childhood, or adolescence. These disorders have traditionally been thought of as first occurring in childhood or as exclusive to childhood and as requiring operational criteria different from those used to define disorders in adults. Table 1.5 lists disorders that can be diagnosed in children or adolescents (e.g., mood disorders, anxiety disorders), but that are not listed in DSMIV as distinct disorders first occurring during childhood, or requiring operational criteria that are different from those used for adults. In many ways, the DSM-IV distinction between child and adult categories is an arbitrary one; it is more a reflection of our current lack of knowledge concerning the continuities between child and adult disorders than of the existence of qualitatively distinct conditions. Recent efforts to diagnose ADHD in adults illustrate this problem. Although the criteria for ADHD were derived from work with children, and the disorder is included in the “infancy, childhood, or adolescence” section of DSM-IV, these criteria are being used to diagnose adults even though they do not fit the expression of the disorder in adults very well. The more general issue here is whether there is a need for separate diagnostic criteria for children versus adults, or whether one can use the same criteria by adjusting them to take into account differences in developmental level. For instance, the childhood category of overanxious disorder in DSM-III-R (APA, 1987) was subsumed under the category of generalized anxiety disorder in DSM-IV (APA, 1994). With this TABLE 1.1. Individual Parent-Rated Problems Accounting for More than 10% of the Variance in Clinical Status of Children Aged 4–16 Poor school work (19%)a,b Can’t concentrate, can’t pay attention for long (18%)b Lacks self-confidence (17%)b Punishment doesn’t change his/her behavior (17%)b Disobedient at home (15%)b Has trouble following directions (15%)b Sad or depressed (15%)b Uncooperative (14%)b Nervous, high-strung, or tense (14%)b Feels he/she can’t succeed (13%) Feels worthless or inferior (13%) Disobedient at school (13%) Easily distracted (13%) Lies (13%) Looks unhappy without good reason (13%) Fails to finish things he/she starts (12%) Defiant (12%) Doesn’t get along with other kids (12%) Has a hard time making friends (12%) Doesn’t seem to feel guilty after misbehavior (12%) Needs constant supervision (12%) Sudden changes in mood or feelings (12%) Angry moods (11%) Impulsive or acts without thinking (11%) Irritable (11%) Temper tantrums or hot temper (10%) Does things slowly and incorrectly (10%) Loses train of thought (10%) Loss of ability to have fun (10%) Passive or lacks initiative (10%) Note. Data from Achenbach, Howell, Quay, and Conners (1991, pp. 107–115). aNumber in parentheses indicates the percentage of variance accounted for by this problem behavior. bItems accounting for 14% or more of the variance are designated as having a large effect size, according to criteria presented by Cohen (1988). 1. A Developmental–Systems Perspective 29 TABLE 1.2. Commonly Identified Dimensions of Child Psychopathology and Examples of Items Reflecting Each of the Dimensions Withdrawn Would rather be alone Refuses to talk Secretive Shy, timid Stares blankly Sulks Underactive Unhappy, sad, depressed Withdrawn Somatic complaints Feels dizzy Overtired Aches, pains Headaches Nausea Eye problems Rashes, skin problems Stomachaches Vomiting Anxious/depressed Lonely Cries a lot Fears impulses Needs to be perfect Feels unloved Feels persecuted Feels worthless Nervous, tense Fearful, anxious Feels too guilty Self-conscious Suspicious Unhappy, sad, depressed Worries Harms self Thinks about suicide Overconforms Hurt when criticized Anxious to please Afraid of mistakes Social problems Acts too young Too dependent Doesn’t get along with peers Gets teased Not liked by peers Clumsy Prefers younger children Overweight Withdrawn Lonely Cries Feels unloved Feels persecuted Feels worthless Accident-prone Thought problems Can’t get mind off thoughts Hears things Repeats acts Sees things Strange behavior Strange ideas Stares blankly Harms self Fears Stores up things Attention problems Acts too young Can’t concentrate Can’t sit still Confused Daydreams Impulsive Nervous, tense Poor school work Clumsy Stares blankly Twitches Hums, odd noises Fails to finish Fidgets Difficulty with directions Difficulty learning Apathetic Messy work Inattentive Underachieving Fails to carry out tasks Delinquent behavior Lacks guilt Bad companions Lies Prefers older kids Runs away from home Sets fires Steals at home Swearing, obscenity Truancy Alcohol, drugs Thinks about sex too much Vandalism Tardy Aggressive behavior Argues Brags Mean to others Demands attention Destroys own things Destroys others’ things Disobedient at school Jealous Fights Attacks people Screams Shows off Stubborn, irritable Sudden mood changes Talks too much Teases Temper tantrums Threatens Loud Disobedient at home Defiant Disturbs others Talks out of turn Disrupts class Explosive Easily frustrated Note. Dimensions are based on analyses across informants (e.g., parents, teachers, and children) and assessment methods (Child Behavior Checklist, Youth Self-Report Form, and Teacher Report Form). Adapted from Achenbach (1993, pp. 41–43). Copyright 1993 by T. M. Achenbach. Adapted by permission. 30 I. INTRODUCTION TABLE 1.3. DSM-IV Categories for Developmental and Learning Disorders Usually First Diagnosed in Infancy, Childhood, or Adolescence Mental retardation Mild, moderate, severe, profound, severity unspecified Learning disorders Reading disorder Mathematics disorder Disorder of written expression Learning disorder not otherwise specified Motor skills disorder Developmental coordination disorder Communication disorders Expressive language disorder Mixed receptive–expressive language disorder Phonological disorder Stuttering Communication disorder not otherwise specified Pervasive developmental disorders Autistic disorder Rett’s disorder Childhood disintegrative disorder Asperger’s disorder Pervasive developmental disorder not otherwise specified change, the number of criteria required for children to meet this diagnosis was also altered. Finally, Table 1.6 lists DSM-IV categories for other conditions that are not defined as mental disorders, but that may be a focus of clinical attention during childhood or adolescence. The categories that are included are the ones that seem especially relevant to children, in that they emphasize relational problems, maltreatment, and academic and adjustment difficulties. APPROACHES TO THE CLASSIFICATION AND DIAGNOSIS OF CHILD PSYCHOPATHOLOGY The formal and informal classification systems that have been used by psychiatrists, psychologists, and educators to categorize the different forms of child psychopathology have played a central role in defining the field. For example, in referring to these systems, Adelman (1995) states: “They determine the ways individuals are described, studied, and served; they shape prevailing practices related to intervention, professional TABLE 1.4. DSM-IV Categories for Other Disorders Usually First Diagnosed in Infancy, Childhood, or Adolescence Attention-deficit and disruptive behavior disorders Attention-deficit/hyperactivity disorder Predominantly inattentive type Predominantly hyperactive–impulsive type Combined type Attention-deficit/hyperactivity disorder not otherwise specified Disruptive behavior disorders Conduct disorder Oppositional defiant disorder Disruptive behavior disorder not otherwise specified Feeding and eating disorders of infancy or early childhood Pica Rumination disorder Feeding disorder of infancy or early childhood Tic disorders Tourette’s disorder Chronic motor or vocal tic disorder Tic disorder not otherwise specified Elimination disorders Encopresis Enuresis Other disorders of infancy, childhood, or adolescence Separation anxiety disorder Selective mutism Reactive attachment disorder of infancy or early childhood Stereotypic movement disorder Disorder of infancy, childhood, or adolescence not otherwise specified training, and certification; and they influence decisions about funding. It is not surprising, therefore, that debates about classification schemes, specific diagnostic procedures, and the very act of labeling are so heated” (p. 29). Although early conceptualizations of psychopathology included underdeveloped and global descriptions of childhood disorders (e.g., “adjustment problem”), this state of affairs has been steadily improving. Nevertheless, problems and issues in describing and classifying childhood disorders continue to plague the field (e.g., Quay, Routh, & Shapiro, 1987). As noted by Rutter and Garmezy (1983), “All too frequently findings have been inconclusive because the measures employed have been weak, nondiscriminating, or open to systematic bias. Similarly, comparisons 1. A Developmental–Systems Perspective 31 between studies have often been vitiated because cases have been defined differently, because the settings have been noncomparable, or because the measures focused on different aspects of behavior” (p. 865). There is general agreement in medicine, psychiatry, and psychology regarding the need for a system of classifying for childhood disorders. However, major areas of contention have arisen around such issues as which disorders should be included in the system, what the optimal strategies are for organizing and grouping disorders, and what specific criteria should be used to define a particular disorder (Achenbach, 1985; Achenbach & Edelbrock, 1989; Mash & Terdal, 1997a; Sonuga-Barke, 1998). The two most common approaches to the diagnosis and classification of child psychopathology involve the use of (1) “categorical” classification systems that are based primarily on informed clinical consensus, an approach that has dominated and continues to dominate the field (APA, 1994, 2000); and (2) empirically based “dimensional” classification schemes derived through the use of multivariate statistical techniques (Achenbach, 1993, 1997). In addition, alternative and/ or derivative approaches to classification have been proposed to address perceived deficiencies associated with the use of categorical and dimensional approaches. These have included developmentally based measures (Garber, 1984; Mohr & Regan-Kubinski, 1999; Sroufe, 1997); laboratory and performance-based measures (Frick, 2000); prototype classification (Cantor et al., 1980); and behavioral classification based on behavioral excesses, deficits, and faulty stimulus control (Adams, Doster, & Calhoun, 1977; Kanfer & Saslow, 1969; Mash & Hunsley, 1990). Although each of these alternative approaches has something to offer to the classification of childhood disorders, they are generally underdeveloped and unstandardized, and have not been widely accepted or used in either research or practice. To date, no single classification scheme for childhood disorders has established adequate reliability and validity (Cantwell, 1996; Mash & Terdal, 1997a). Many researchers and clinicians continue to express concerns that current diagnostic and classification systems (1) underrepresent disorders of infancy and childhood; (2) are inadequate in representing the interrelationships and overlap that exist among many childhood disorders; (3) are not sufficiently sensitive to the developmental, contextual, and relational parameters that are known to characterize most forms of psychopathology in children; and (4) are heterogeneous with respect to etiology (Jensen & Hoagwood, 1997; Kagan, 1997). TABLE 1.5. Selected Categories for Disorders of Childhood or Adolescence That Are Not Listed Separately in DSM-IV as Those Usually First Diagnosed in Infancy, Childhood, or Adolescence Mood disorders Depressive disorders Major depressive disorder Dysthymic disorder Bipolar disorders Anxiety disorders Specific phobia, social phobia, obsessive–compulsive disorder, posttraumatic stress disorder, acute stress disorder, generalized anxiety disorder, anxiety disorder due to . . . (specific medical condition) Somatoform disorders Factitious disorders Dissociative disorders Sexual and gender identity disorders Eating disorders Sleep disorders Schizophrenia and other psychotic disorders Substance-related disorders Impulse-control disorders not elsewhere classified Adjustment disorders Personality disorders TABLE 1.6. Selected DSM-IV Categories for Other Conditions That May Be a Focus of Clinical Attention during Childhood or Adolescence, but Are Not Defined as Mental Disorders Relational problems Relational problem related to a general mental disorder or general medical condition Parent–child relational problem Partner relational problem Sibling relational problem Relational problem not otherwise specified Problems related to abuse or neglect Physical abuse of child Sexual abuse of child Neglect of child Bereavement Borderline intellectual functioning Academic problem Child or adolescent antisocial behavior Identity problem 32 I. INTRODUCTION Categorical Approaches Categorical approaches to the classification of childhood disorders have included systems developed by the Group for the Advancement of Psychiatry (1974), the WHO (1992), the APA (1994), and the Zero to Three/National Center for Clinical Infant Programs (1994). Although a detailed review of all these systems is beyond the scope of this chapter, a brief history of the APA’s development of the DSM approach is presented to illustrate the issues associated with categorical approaches, the growing concern for more reliable classification schemes for childhood disorders, and the evolving conceptualizations of childhood disorders over the past 50 years. Also, the Diagnostic Classification of Mental Health and Developmental Disorders of Infancy and Early Childhood, or Diagnostic Classification: 0–3 (DC:0–3; Zero to Three/National Center for Clinical Infant Programs, 1994), is described to illustrate a categorical approach that attempts to integrate developmental and contextual information into the diagnosis of infants’ and young children’s problems. Development of the DSM Approach One of the first efforts to collect data on mental illness was in the U.S. census of 1840, which recorded the frequency of a single category of “idiocy/insanity.” Forty years later, seven categories of mental illness were identified: dementia, dipsomania, epilepsy, mania, melancholia, monomania, and paresis (APA, 1994). Much later (in the 1940s), the WHO classification system emerged with the manuals of the ICD, whose 6th revision included, for the first time, a section for mental disorders (APA, 1994; Cantwell, 1996). In response to perceived inadequacies of the ICD system for classifying mental disorders, the APA’s Committee on Nomenclature and Statistics developed the DSM-I in 1952 (APA, 1952). There were three major categories of dysfunction in the DSM-I—“organic brain syndromes,” “functional disorder,” and “mental deficiency” (Kessler, 1971)—under which were subsumed 106 categories (by contrast, DSM-IV consists of 407 separate categories; Cantwell, 1996). The term “reaction” was used throughout the text, which reflected Adolf Meyer’s psychobiological view that mental illness involves reactions of the personality to psychological, social, and biological factors (APA, 1987). Children were virtually neglected in the early versions of DSM, with most childhood disorders relegated to the adult categories (Cass & Thomas, 1979; Silk et al., 2000). In fact, DSM-I included only one child category of “adjustment reactions of childhood and of adolescence,” which was included under the heading of “transient situational disorders.” As reflected in the use of the term “reaction,” psychoanalytic theory had a substantial influence on the classification of both child and adult psychopathology (Clementz & Iacono, 1993). In part, this was due to the fact that the first classification system to focus on childhood psychopathology was developed by Anna Freud in 1965 (see Cantwell, 1996). Although the term “reaction” was eliminated from DSM-II (APA, 1968), a separate section was reserved for classifying neuroses, and diagnoses could be based on either an assessment of the client’s presenting symptomatology or inferences about his or her unconscious processes (Clementz & Iacono, 1993). Once again, apart from conditions subsumed under the adult categories, DSM-II gave little recognition to childhood difficulties except for mental retardation and schizophrenia—childhood type (Cass & Thomas, 1979). As a formal taxonomy, DSM-III (APA, 1980) represented a significant advance over the earlier editions of the DSM. The first and second editions contained only narrative descriptions of symptoms, and clinicians had to draw on their own definitions for making a diagnosis (APA, 1980). In DSM-III, these descriptions were replaced by explicit criteria, which in turn enhanced diagnostic reliability (Achenbach, 1985; APA, 1980). Moreover, unsubstantiated inferences that were heavily embedded in psychoanalytic theory were dropped; more child categories were included; a multiaxial system was adopted; and a greater emphasis was placed on empirical data (Achenbach, 1985). These changes reflected the beginnings of a conceptual shift in both diagnostic systems and etiological models away from an isolated focus of psychopathology as existing within the child alone, and toward an increased emphasis on his or her surrounding context. DSM-III was revised in 1987 (DSM-III-R) to help clarify the numerous inconsistencies and ambiguities that were noted in its use. For example, empirical data at that time did not support the category of attention deficit disorder without hyperactivity as a unique symptom cluster (Routh, 1990), and this category was removed from DSM-III-R. DSM-III-R was also developed 1. A Developmental–Systems Perspective 33 to be polythetic, in that a child could be diagnosed with a certain subset of symptoms without having to meet all criteria. This was an important change, especially in light of the heterogeneity and rapidly changing nature of most childhood disorders (Mash & Terdal, 1997a). Relative to its predecessors, far greater emphasis was also placed on empirical findings in the development of the DSM-IV, particularly for the child categories. In order to bridge the planned 12-year span between the DSM-IV and DSM-V, a revision (DSM-IV-TR) of the DSM was published in 2000 (APA, 2000). The DSM-IV-TR was limited to text revisions (e.g., associated features and disorders, prevalence) and was designed mainly to correct any factual errors in DSM-IV, make sure that information is still current, and incorporate new information since the time the original DSM-IV literature reviews were completed in 1992. Substantive changes in diagnostic criteria were not considered or made; nor were there any changes in relation to new disorders or subtypes. Thus DSM-IV and DSM-IV-TR are equivalent with respect to specific diagnostic criteria. DSM-IV is a multiaxial system that includes five different axes. Axis I is used to report clinical disorders and other conditions that may be a focus of clinical attention. The various Axis I diagnostic categories that apply to infants, children, and adolescents have been listed in Tables 1.3 to 1.6 of this chapter. Axis II includes personality disorders and mental retardation. The remaining axes pertain to general medical conditions (Axis III), psychosocial and environmental problems (Axis IV), and global assessment of functioning (Axis V). Although DSM-III-R (APA, 1987) and DSMIV (APA, 1994) include numerous improvements over the previous DSMs—with their greater emphasis on empirical research, and more explicit diagnostic criteria sets and algorithms— criticisms have also been raised (e.g., Mohr & Regan-Kubinski, 1999; Nathan & Lagenbucher, 1999; Sonuga-Barke, 1998; Sroufe, 1997). One major criticism is the static nature of DSM categories, especially when one considers the dynamic nature of development in children (Mash & Terdal, 1997a; Routh, 1990). Another source of dissatisfaction is that the DSM-IV categorical scheme may contribute minimally to meeting children’s needs. For example, it may be necessary for a child to meet specific diagnostic criteria for a learning disability in order to qualify for a special education class. However, if the child’s problems are subclinical, or the child’s problems relate to more than one DSM category, then he or she may be denied services (Achenbach, 2000). However, even if one were to adopt a more dimensional approach to classification, there would nonetheless continue to be a categorical interpretation of the data (e.g., distinguishing between individuals who require help and those who do not) (Sonuga-Barke, 1998). Another problem with DSM-IV relates to the wording and the lack of empirical adequacy for certain criterion sets. For example, the words “often” in the criteria for ADHD and conduct disorder, and “persistent” and “recurrent” in the criteria for separation anxiety disorder, are not clearly defined. This ambiguity poses a particular problem when one considers that the primary sources of assessment information are often a child’s parents, whose perception and understanding of these terms may be idiosyncratic or inaccurate. This ambiguity and other factors may contribute to the unreliability or unsuitability of the DSM for diagnosing certain childhood disorders (e.g., Nicholls, Chater, & Lask, 2000). A further difficulty with DSM-IV diagnostic criteria is the lack of emphasis on the situational or contextual factors surrounding and contributing to various disorders. This is a reflection of the fact that DSM-IV continues to view mental disorder as individual psychopathology or risk for psychopathology, rather than in terms of problems in psychosocial adjustment. One problem with respect to the atheoretical nature of DSM is that it has perhaps mistakenly fostered the assumption that a description of symptoms is sufficient for diagnosis, without taking into account natural history, psychosocial correlates, biological factors, or response to treatment (Cantwell, 1996). However, the consideration in DSM-IV of such factors as culture, age, and gender associated with the expression of each disorder is laudable, as is the increased recognition of the importance of family problems and extrafamilial relational difficulties. The changes in the DSMs from 1952 to 2000 reflect increasing diagnostic accuracy and sophistication. The transition from “reactive” diagnoses (DSM-I) and the virtual neglect of childhood criteria (DSM-I, DSM-II) to an increased number of child categories, more explicit criteria, and multiaxial evaluation (DSM-III, DSM-III-R), and then to an even greater emphasis on empirical research to guide nomenclature as well as the increased awareness (and inclusion) of contextual 34 I. INTRODUCTION and developmental considerations (DSM-IV, DSM-IV-TR), exemplify important shifts in how psychopathology in children has come to be conceptualized. However, along with increased complexity has come a new set of problems. For example, the extent to which comorbidity is an artifact of the DSM’s polythetic criteria or truly differentiated nosological entities is unclear (Angold, Costello, & Erkanli, 1999; Nottelmann & Jensen, 1995), or whether the pendulum has swung too far from not recognizing psychopathology in children to identifying and diagnosing too much (Silk et al., 2000). It is also the case that ongoing changes in diagnostic criteria based on new findings and other considerations (e.g., eligibility for services) are likely to influence prevalence estimates for many childhood disorders. For example, current estimates of autistic disorder are about three times higher than previous ones (Fombonne, 1999; Tanguay, 2000); this increase is primarily due to a broadening of the criteria used to diagnose autism, as well as increased recognition of milder forms of the disorder (Bryson & Smith, 1998; Gillberg & Wing, 1999). There is also ongoing debate about whether Asperger’s disorder is a variant of autism or simply describes higherfunctioning individuals with autism (Schopler, Mesibov, & Kunce, 1998; Volkmar & Klin, 2000). The resolution of this debate and prevalence estimates for both autism and Asperger’s disorder will depend on how the diagnosis of Asperger’s disorder is used, since no “official” definition for this disorder existed until it was introduced in DSM-IV (Volkmar & Klin, 1998). Development of the DC:0–3 System In addition to the limitations noted above, DSMIV does not provide in-depth coverage of the mental health and developmental problems of infants and young children, for whom family relationships are especially salient. To address this perceived deficiency, the DC:0–3 was developed by the Diagnostic Classification Task Force of the Zero to Three/National Center for Clinical Infant Programs (Zero to Three/National Center for Clinical Infant Programs, 1994). DC:0–3 is intended to provide a comprehensive system for classifying problems during the first 3–4 years of life (Greenspan & Wieder, 1994; Lieberman, Wieder, & Fenichel, 1997). Unlike DSM-IV, DC:0–3 is based on the explicit premise that diagnosis must be guided by the principle that all infants and young children are active participants in relationships within their families. Hence descriptions of infant–caregiver interaction patterns, and of the links between these interaction patterns and adaptive and maladaptive patterns of infant and child development, constitute an essential part of the diagnostic process. In explicitly recognizing the significance of relational problems, DC:0–3 includes a relationship disorder classification as a separate axis (Axis II) in its multiaxial approach (Axis I, primary diagnosis; Axis III, medical and developmental disorders and conditions; Axis IV, psychosocial stressors; Axis V, functional emotional developmental level). The diagnosis of relationship disturbances or disorders is based on observations of parent– child interaction and the parent’s verbal report regarding his or her subjective experience of the child. Relational difficulties are rated with respect to their intensity, frequency, and duration, and classified as perturbations, disturbances, or disorders. In making the DC:0–3 Axis II relationship disorder diagnosis, three aspects of the relationship are considered: (1) behavioral quality of the interaction (e.g., sensitivity or insensitivity in responding to cues); (2) affective tone (e.g., anxious/ tense, angry); and (3) psychological involvement (e.g., parents’ perceptions of the child and of what can be expected in a relationship). Axis V of DC:0–3, functional emotional development level, includes the ways in which infants or young children organize their affective, interactive, and communicative experiences. Axis V assessment is based in large part on direct observations of parent–child interaction. The various levels include social processes such as mutual attention, mutual engagement or joint emotional involvement, reciprocal interaction, and affective/ symbolic communication. Problems may reflect constrictions in range of affect within levels or under stress, or failure to reach expected levels of emotional development. DC:0–3 is of note in recognizing (1) the significance of early relational difficulties; (2) the need to integrate diagnostic and relational approaches in classifying child psychopathology (Lyons-Ruth, 1995); and (3) the need to apply both quantitative and qualitative criteria in describing relational problems. In addition, the dimensions and specific processes that are used for classification (e.g., negative affect, unresponsivity, uninvolvement, lack of mutual engagement, lack of reciprocity in interaction) include those that have been identified as important in many develop1. A Developmental–Systems Perspective 35 mental and clinical research studies on early relationships, and the system is decidedly more sensitive to developmental and contextual parameters than DSM-IV. However, although promising, DC:0–3 is relatively untested, was generated on the basis of uncontrolled clinical observations, is of unknown reliability and validity, and suffers from many of the same criticisms that have been noted for DSM-IV (Eppright, Bradley, & Sanfacon, 1998). Nevertheless, the scheme provides a rich descriptive base for exploring the ways in which psychopathology is expressed during the first few years of life, and it calls attention to the need to examine potential continuities between early problems and later individual and/or family disorders (Keren, Feldman, & Tyano, 2001; Thomas & Clark, 1998; Thomas & Guskin, 2001). Dimensional Approaches Dimensional approaches to classification assume that a number of relatively independent dimensions or traits of behavior exist, and that all children possess these to varying degrees. These traits or dimensions are typically derived through the use of multivariate statistical methods, such as factor analysis or cluster analysis (Achenbach, 1993). Empirically derived schemes are more objective, are potentially more reliable, and allow for a greater description of multiple symptom patterns than clinically derived classification systems. However, there are also a number of problems associated with their use, including the dependency of the derived dimensions on sampling, method, and informant characteristics, and on the age and sex of the child (Mash & Terdal, 1997a). As a result, there can be difficulties in integrating information obtained from different methods, from different informants, over time, or across situations. Dimensional approaches have also shown a lack of sensitivity to contextual influences, although there have been efforts to develop dimensional classification schemes based on item pools that include situational content (e.g., McDermott, 1993). The growth in the use of multivariate classification approaches in child and family assessment has been fueled by the extensive work of Thomas Achenbach and his colleagues (see the Achenbach System of Empirically Based Assessment [ASEBA]: http://www.ASEBA.org) with the various parent, teacher, youth, observer, and interview versions of the Child Behavior Checklist and Profile (Achenbach, 1993), and by the development of similar assessment batteries (e.g., the Behavior Assessment System for Children [BASC]: Kamphaus et al., 1999; Reynolds & Kamphaus, 1992). For a comprehensive discussion of these approaches and the use of empirically derived classification schemes more generally, the reader is referred to Achenbach (1985, 1993), Hart and Lahey (1999), and Mash and Terdal (1997a), It should also be noted that there has been a trend toward greater convergence of the categorical and dimensional approaches to classification. Many of the items that were retained in DSM-IV child categories were derived from findings from multivariate studies, and the process that led to the development of DSM-IV treated most childhood disorders as dimensions, albeit the use of cutoff scores on item lists arbitrarily created categories out of these dimensions (Spitzer et al., 1990). Performance-Based Diagnostic Information Performance-based information and/or observational measures provide additional sources of diagnostic information that may be sensitive to differences among children exhibiting similar self- or other-reported symptoms (Frick, 2000; Kazdin & Kagan, 1994). These measures assess children’s performance on standardized tasks, usually ones that reflect basic biological, cognitive, affective, or social functioning. For example, tasks involving behavioral observations of fear and avoidance, recall memory under stressful conditions, delayed response times to threatening stimuli, and the potentiation of the blink reflex following exposure to a threatening stimulus have all been suggested as potentially useful in diagnosing groups and/or subgroups of children with anxiety disorders (Kazdin & Kagan, 1994; Vasey & Lonigan, 2000). Similarly, tests of behavioral inhibition (e.g., the stop-signal paradigm) and tasks involving sustained attention (e.g., the continuous-performance test) have proven useful with children with ADHD (Rapport, Chung, Shore, Denney, & Isaacs, 2000). Measures of low resting heart rate as an early biological marker for later aggressive behavior (Raine, Venables, & Mednick, 1997); facial emotion recognition tasks and gambling tasks in identifying children with psychopathic tendencies (Blair, Colledge, & Mitchell, 2001; Blair, Colledge, Murray, & Mitchell, 2001); and a variety of cognitive tasks 36 I. INTRODUCTION for children with autism (Klinger & Renner, 2000) have also been found to have diagnostic value. A study by Rubin, Coplan, Fox, and Calkins (1995) illustrates the utility of performance-based diagnostic information. These researchers differentiated groups of preschool children based on the two dimensions of “emotionality” (i.e., threshold and intensity of emotional response) and “soothability” (i.e., recovery from emotional reaction based on soothing by self and others), and on their amount of social interactions with peers. Children’s dispositional characteristics and behavioral styles were used to predict outcomes. Asocial children with poor emotion regulation had more internalizing problems. In contrast, social children with poor emotion regulation were rated as having more externalizing difficulties. When behavioral and emotional dimensions were incorporated into classification, it was possible to make finer predictions—for example, that only a certain type of asocial children (i.e., reticent children with poor emotion regulation) would display later problems. The use of performance-based measures in diagnosis is predicated on the availability of reliable and valid performance indicators for groups of children with known characteristics. Although such data are available in varying amounts for a wide range of disorders, there is a need to validate such findings for the purposes of diagnosis and against other sources of information. It is also the case that performance criteria for these measures are based on information obtained from children who were themselves previously identified using other diagnostic procedures. This raises the question of nonindependence and representativeness of data sources. There is also little normative information available regarding the base rates of children in the general population who exhibit certain patterns of responding on these tasks. ISSUES IN CLASSIFICATION Categories, Dimensions, or Both? Psychological studies of child psychopathology have tended to conceptualize behavior, affect, and cognition on quantitative/continuous dimensions, whereas child psychiatry has tended to conceptualize child psychopathology in categorical terms. Both approaches are relevant to classifying childhood disorders, in that some disorders may be best conceptualized as qualitatively distinct conditions and others as extreme points on one or more continuous dimensions. Kazdin and Kagan (1994) argue for greater research attention to qualitatively distinct categories of disorder, based on illustrative findings from studies suggesting that the emotional arousal generated by unfamiliarity, threat, and attack is not a continuous dimension, and that it is possible to identify different subgroups of aggressive children based on varying levels of adrenaline in their urine. There is currently little agreement as to which childhood disorders are best conceptualized as categories and which as dimensions. It has been suggested that many childhood disorders, such as anxiety, depression, ADHD, and the disruptive behavior disorders, appear to reflect dimensions of personality rather than categorical problems (e.g., Werry, 2001). For example, childhood ADHD symptom clusters of inattention–disorganization and hyperactivity–impulsivity have been found to be related to adult personality dimensions of low conscientiousness and low agreeableness, respectively (Nigg et al., 2001). Even a disorder such as autism, which has traditionally been viewed as “categorical” in nature, can be conceptualized as an extreme on a continuum of social behavior (Baron-Cohen, 2000). For dimensional disorders, children who score just below the cutoff for a diagnosis may one day meet criteria, and often show impairment comparable to that of children who score above the cutoff. Similarly, those above the cutoff may one day move below it. Since any classification scheme represents a construction rather than a reality, it seems unlikely that most disorders will fall neatly into one designation or the other (Lilienfeld & Marino, 1995). Whether or not particular conditions are construed as qualitatively distinct categories, as continuous dimensions, or as both will probably depend on the utility, validity, and predictive value of particular groupings and subgroupings for certain purposes related to understanding and remediating child psychopathology. Research into such subgroupings is just beginning to emerge (e.g., Kendall, Brady, & Verduin, 2001). Regardless of the particular approach one adopts for the classification of childhood psychopathology, diagnostic decisions need to be based on a comprehensive assessment of the individual 1. A Developmental–Systems Perspective 37 child—one that incorporates sensitivity to and understanding of the complexity of multiple antecedents, developmental considerations, comorbidity, continuity–discontinuity, and the constantly changing nature of the child (Orvaschel, Ambrosini, & Rabinovich, 1993). Comorbidity An issue that has important ramifications for theory and research in defining and classifying child psychopathology is comorbidity (Achenbach, 1995; Angold, Costello, & Erkanli, 1999; Carey & DiLalla, 1994; Caron & Rutter, 1991; Sonuga-Barke, 1998). “Comorbidity” generally refers to the manifestation of two or more disorders that co-occur more often than would be expected by chance alone. For example, although the base rates for ADHD and conduct disorder in the general population are less than 10% for each disorder, epidemiological studies have found that among children diagnosed with ADHD, approximately 50% are also diagnosed with conduct disorder (Kazdin & Johnson, 1994; Loeber & Keenan, 1994). Comorbidity has been reported to be as high as 50% in community samples and even higher in clinic samples (Anderson, Williams, McGee, & Silva, 1987; Bird et al., 1988; Caron & Rutter, 1991). Some of the more commonly co-occurring child and adolescent disorders include conduct disorder and ADHD, autistic disorder and mental retardation, and childhood depression and anxiety. There is continuing debate regarding the definition and nature of “comorbidity” (Angold, Costello, & Erkanli, 1999; Blashfield, McElroy, Pfohl, & Blum, 1994; Caron & Rutter, 1991; Lilienfeld, Waldman, & Israel, 1994; Meehl, 2001; Robins, 1994; Rutter, 1994b; Sameroff, 2000a; Spitzer, 1994; Widiger & Ford-Black, 1994). Some researchers contend that the term is wholly inadequate, because it does not distinguish accurately between manifest conditions seen in organic medicine (e.g., diseases) and latent conditions described in mental health (e.g., syndromes and disorders (Lilienfeld et al., 1994). Others argue that the dispute over whether one should use the term “comorbidity,” “co-occurrence,” or “covariation” is largely a semantic one (Rutter, 1994b; Spitzer, 1994; Widiger & Ford- Black, 1994). Several possible reasons why comorbidity may be exaggerated or artificially produced have been identified in the literature (Angold, Costello, & Erkanli, 1999; Caron & Rutter, 1991; Lilienfeld et al., 1994; Rutter, 1994b; Verhulst & van der Ende, 1993). There may be a sampling bias that occurs whenever there are fewer numbers of individuals who are referred to clinics than who exhibit a given disorder. In such cases, the clinic samples will contain a disproportionately large number of subjects who display comorbid conditions. This phenomenon occurs because the probability of being referred to mental health services is higher for a child with a comorbid condition than for a child with only one disorder. Related to this sampling bias are various other referral factors that may inflate the degree of cooccurring disorders among clinic samples. Clinics that and clinicians who specialize in treating more complicated cases, for example, may be more likely to receive referrals in which comorbid conditions are present. In addition, children with internalizing difficulties such as depression are more likely to be referred by their parents or the school system if they also show externalizing symptoms, largely because externalizing problems are viewed as more disruptive by referral sources. Comorbidity may also reflect various sources of nosological confusion arising from the manner in which different childhood disorders have been conceptualized and organized. For instance, Widiger and Ford-Black (1994) claim that excessive rates of co-occurrence seemed to appear concomitantly with the changes that occurred in DSM-III (e.g., increased coverage, divisions of diagnostic categories, the provision of separate and multiple axes). Another example is that DSM-IV makes it possible to have multiple diagnoses in the absence of multiple syndromes (Cantwell, 1996; Robins, 1994). One source of confusion stems from the overlapping criterion sets within contemporary classification schemes. In DSM-IV, diagnoses are based on a set of polythetic criteria that includes specific symptom constellations. In many cases, the presence of concomitant symptoms of a different kind are ignored, resulting in an increased likelihood that the accompanying symptoms will be represented in a different diagnostic category (Caron & Rutter, 1991). Sonuga-Barke (1998) argues, however, that although earlier diagnostic systems steered clear of comorbidity by using a hierarchical set of exclusionary criteria, “these approaches were abandoned because they clearly led to a misrepresentation of the 38 I. INTRODUCTION structure of disorder” (p. 119). For example, they led to low base rates of disorders and poor interrater agreement. Apart from the various artifactual contributors to comorbidity, there are also indicators in support of “true” comorbidity (Rutter, 1994b). It is possible that general propensities toward and/or struggles with adaptation are at the core of every disorder, but how the phenotype is expressed is contingent upon a myriad of environmental conditions and person–environment interactions (Caron & Rutter, 1991). Consistent with this notion, Lilienfeld et al. (1994) maintain that comorbidity in childhood disorders may be partly a function of developmental level—that is, of underlying processes that have not yet achieved full differentiation. Differing rates of comorbidity with age may also reflect the fact that the appearance of one disorder or problem may precede the appearance of the other, as is the case for anxiety preceding depression (Brady & Kendall, 1992) or for impulsivity preceding attentional problems (E. L. Hart et al., 1995). Still another possibility is that comorbidity reflects “a more amorphous early expression of psychopathology in young children that does not crystallize into more definitive psychopathology until later in life” (Cantwell, 1996, p. 4). Comorbidity can also arise as a result of a causal association in which the severity of one disorder may lead to or greatly increase the later risk for another disorder (e.g., ADHD and oppositional defiant disorder) or a shared underlying cause, such as common genetic effects (e.g., conduct disorder and depression) or shared environmental effects (oppositional defiant disorder and conduct disorder). In summary, it would appear that some cases of comorbidity are the result either of ambiguity in the definition of dysfunctionality that is used, or of artifactual/methodological issues. However, as Kazdin and Kagan (1994) note, “the broader point is still relevant and not controverted with specific diagnostic conundrums—namely, multiple symptoms often go together in packages” (p. 40). This is not to suggest that all disorders cluster together into packages; rather, the fact that many frequently do has important implications for how child psychopathology is conceptualized and treated. The complexity of comorbidity behooves researchers to move beyond singular models and to examine multiple expressions, etiologies, and pathways of childhood dysfunction (Burt, Krueger, McGue, & Iacono, 2001; Kazdin & Johnson, 1994). THEORY AND CHILD PSYCHOPATHOLOGY The Role of Theory in Child Psychopathology Every step in the research process is influenced by the investigator’s preconceptions and ideologies (Kuhn, 1962; Maxwell & Delaney, 1990). As the history of child psychopathology has shown, an overemphasis on a grand theory or explanatory model in the absence of data can perpetuate false ideas and seriously impede our understanding of childhood disorders. On the other hand, “data gathering in the absence of hypotheses can become an inconsequential exercise in gathering inconsequential facts” (Rutter & Garmezy, 1983, p. 870). The value of theory lies not just in providing answers but also in raising new questions, which arise not only from addressing new problems but also from looking at familiar problems in different ways. One cannot consider theory, research, and practice in childhood psychopathology without also having some understanding of the underlying philosophical and epistemological assumptions that have guided work in this area. In this context, Overton and Horowitz (1991) discuss four levels of science: (1) epistemology; (2) guidelines, rules, and definitions of scientific knowing; (3) metatheoretical principles; and (4) theory. The first level, “epistemology,” defined as a theory about the nature of knowledge itself, has to do with the general rules of science, the metatheoretical assumptions about the nature of humankind, and the specific theoretical models and research designs that arise out of such assumptions. One epistemological stance (i.e., “realism”) asserts that knowledge exists independently of one’s own perceptual and cognitive processes (Maxwell & Delaney, 1990; Overton & Horowitz, 1991). “Logical positivism,” a view that has guided most of our past and present research efforts in child psychopathology, reflects this stance. A second philosophical position is that of “rationalism.” Rationalists contend that the knower of scientific knowledge actively constructs what is known (Maxwell & Delaney, 1990). Instead of there being a fixed and absolute knowledge base to unveil, rationalists assume that knowledge derives from the exercise of relating and interpreting observables to latent constructs (Overton & Horowitz, 1991). Within this metatheoretical position, there lies a continuum between the 1. A Developmental–Systems Perspective 39 belief at one end that our knowledge base will always be uncertain, and the conviction at the other end that some universal truth must lie beyond our interpretive schemes. At the second level of scientific knowledge— that of “guidelines, rules, and definitions”—it becomes evident that epistemology exerts a strong influence. Logical positivism, for instance, distinguishes scientific knowledge from knowledge that accumulates from other modes of knowing by requiring that all theoretical constructs be reducible to stable, objective, and observable knowledge (Maxwell & Delaney, 1990; Overton & Horowitz, 1991). This view maintains that theoretical constructs are to be mathematically related (via correspondence rules) to directly observable behavior and events. Theory, under this argument, advances by means of the empirical method. A hypothesis is tested and when enough hypotheses have been independently and empirically supported, generalizations can be made (via the inductive process) to form a theoretical model. At the third level of scientific knowledge identified by Overton and Horowitz (1991), “metatheoretical principles” guide the development of more specific theories. Two metaphors have been dominant in guiding scientific metatheory: the “machine” and the “organic” metaphors (Overton & Horowitz, 1991; Simeonsson & Rosenthal, 1992). The machine metaphor adopts a metatheor