And Others TITLE Prevention of Mental Disorders, Alcohol, and Other Drug Use in Children and Adolescents

Home , David Shaffer, Leon Eisenberg, Michael Rutter

DOCUMENT RESUME

ED 333 634 EC 300 387

AUTHOR Shaffer, David, Ed.; And Others TITLE Prevention of Mental Disorders, Alcohol, and Other Drug Use in Children and Adolescents. OSAP Prevention Monograph-2. INSTITUTION Alcohol, Drug Abuse, and Mental Health Administration (DHHS/PHS), Rockville, MD. Office for Substance Abuse Prevention.; American Academy of Child and Adolescent Psychiatry, Washington, DC. REPORT NO ADM-90-1646 PUB DATE 90 NOTE 484p. PUB TYPE Collecterl Works - General (020) -- Information Analyses (070) -- Guides - Non-Classroom Use (055)

EDRS PRICE MF02/13c20 Plus Postage. DESCRIPTORS Adolescents; *Alcoholism; At Risk Persons; Behavior Disorders; Children; Chronic Illness; *Delivery Systems; *Drug Abuse; *Emotional Disturbances; Intervention; Learning Disabilities; Mental Disorders; Parent Influence; *Prevention; Professional Education; Public Policy; Research Needs; Suicide; Theory Practice Relationship

ABSTRACT Compiled in this volume are summaries of the knowledge base oa prevention of alcohol and other drug use and mental disorders in children and adolescents. The papers address risk factors, preventive interventions, conceptual and methodological issues, epidemiology, identification, service delivery and treatment, research, and professional training. After an introduction by Irving Philips, papers with the following titles and authors are presented: "Commentary: The Integration of Problem and Prevention Perspectives: Mental Disorders Associated with Alcohol and Drug Use" (Morton M. Silverman); "Conceptual Issues in Prevention" (Arnold Sameroff and Barbara H. Fiese); "The Prevention of Child and Adolescent Disorders: From Theory to Research" (Raymond P. Lorion and others); "Prevention Programming as Organizational Reinvention: From Research to Implementation" (Richard H. Pric..and Raymond P. Lorion); "Public Policy: Risk Factor or Remedy?" (Leon Eisenberg); "Psychiatric Disorder in Parents as a Risk Factor for Children" (Michael Rutter); "Risks for Maladjustment Associated with Chronic Illness in Childhood" (I. Barry Pless and Terence M. Nolan); "Prevention of Psychiatric Morbidity in Children after Disaster" (Robert S. Pynoos and Kathi Nader); "Conduct Disorder: Risk Factors and Prevention" (D. R. Offord); "Prevention of Alcohol and Drug Abuse: A Critical Review of Risk Factors and Prevention StraAgies" (Karol L. Kumpfer); "Prevention Issues in Youth Suicide" (David Shaffer and others); "Prevention of Learning Disorders" (Archie A. Silver and Rosa A. Hagin); "Prevention of Psychiatric Disorders in Children and Adolescents: A Summary of Findiags and Recommendations from Project Prevention" (David Shaffer). Reference lists follow each chapter. (JDD) SCOPE OF INTEREST NOTICE 4071 The ERIC Facility hes assigned this document for processing to In our Judgment this document is also of interest to the ClearCc9- inghouses noted to the right Indexing should reflect their special points c.1 view

I 11 1 IIOil

V 5 DEPARTMENT OF EDUCATION Onse ii 1 elLn ill one!. S and imotovpment E MJCATIONAt RE SOuRC E S INF OF1MATION CE N1 PE PIC D

41/1"hs(In, unser,' has herh ,egyoclui ea as P, enven lihr" the oe,sot Orgar,tal.on gnat.no .rno.ove EtA 'tott ',antes nave um . nnamv ,ecnoduc ton ouant,

r1),IS iii w.e* lr Opn.unS slate() n1^....1.cu fnertt 00 "V ,,e(esSc ly ,eivesehl Of Ri posa on cr pone

I , I

I I I I I I I

BEST COPY AVAILABLE OSAP Prevention Monogi aph-2

PREVENTION OF MENTAL DISORDERS, ALCOHOL AND OTHER DRUG USE IN CHILDREN AND ADOLESCENTS Sponsors: Office for Substance Abuse Prevention American Academy of Child and Adolescent Psychiatry Editors: David Shaffer, M.B., B.S., F.R.C.P., F.R.C. Psych. Irving Philips, M.D. Norbert B. Enzer, M.D. Associate Editors: Morton M. Silverman, M.D. Virginia Anthony

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Alcohol, Drug Abuse, and Mental Health Administration Officc for Substance Abuse Prevention 5600 Fishers Lane Rockville, Maryland 20857 ii PREVENTION OF MENTAL DISORDERS

OSAP Prevention Monographs are prepared by the divisions of the Office for Substance Abuse Prevention (OSAP) and published by its Division of Com- munication Programs. The primary objective of the series is to facilitate the transfer of prevention and intervention technology between and among researchers, administrators, policymakers, educators, and providers in the public and private sectors. The content of state-of-the-art conferences, reviews of innovative or exenr airy programming models, and review of evaluative studies are important elements of OSAP's information dissemination mission. This publication is the product of a project on the prevention of alcohol, drug use, and mental disorders in children and adolescents supported by OSAP and the American Academy of Child and Adolescent Psychiatry. The presentations herein are those of the authors and may not necessarily reflect the opinions, official policy, or position of OSAP; the Alcohol, Drug Abuse, and Mental Health Administration; the Public Health Service;or the U.S. Department of Health and Human Services. OSAP obtained permission from the copyright holder to reproduce the tables on pages 196-203, 206, 210, and the figure on page 213. Further reproduction of these materials is prohibited without specific permission of the copyright holder.All other material in this volume, except quoted passages from copyrighted sources, is in the public domain and may be used reproduced without permission from OSAP or the authors.Citation of source is appreciated. OSAP Production Officer: Timothy F. Campbell DHHS Publication No. (ADM)90-1646 Printed 1989, Reprinted 1990

Project Officer: Mel Segal, M.S.W. OSAP Prevention Monograph Series Elaine M. Johnson, Ph.D. Director, OSAP Robert W. Denniston Director, Division of Communication Programs, OSAP Foreword

This volume is the second in a series of preventionmonographs of the Office for Substance Abuse Prevention and is cosponsoredby the American Academy of Child and Adolescent Psychiatry. It carefullydefines the knowledge base of prevention of and intervention in child and adolescentpsychiatric illnessei. Concern about the toll of alcohol and other drug use amongeildren and adolescents mirrors concern about other disorders.Chapters in this work examine alcohol and other drug problems, learningdisabilities, conduct disorders, public policy or lack thereof, psychiatric disorderdin parents, chronic posttraumatic stress disorderst and suicide. Chapters onconcept and methodology provide a necessary foundation on which the chapters ondisorders are developed. This volume is a tangible exhibit of our commitment toprevention and to the real hope it offers.It uniquely encompasses the many dimensions of our knowledge about treatment and the directions for futureresearch.This monograph informs child and adolescent psychiatrists, otherphysicians, educators, students, clinicians, policymakers, citizens, and parent groups. It illustrates our commitment to the transfer of information fromresearchers to clinicians and ultimately to front-line earlyintervention programs as the key to prevention. The understanding of each disordersheds light on the others and on the dynamics of a most important, yetvulnerable, group of children. We hope that the knowledge shared will stimulate and direct treatmentresearch and suggest a pathway for the future.

Elaine M. Johnson, Ph.D., Director Office for Substance Abuse Prevention Preface

The Office for Substance AbusePrevention (OSAP) and theAmerican Academy of Child and AdolescentPsychiatzy (AACAP) are pleased topublish this volume on the prevention of alcoholand other drug use and mentaldisorders in children aiud adolescents. It isthe culmination of ProjectPrevention: An Intervention Initiative, an interdisciplinaryproject developed by theAACAP to educate child and adolescentpsychiatrists and other mentalhealth professionals. Following a series of meetings, theProject Prevention SteeringCommittee outlined several' risk factors that areeasily identifiable and subject tomodifica- tion through preventive interventions.The steering committee thencommis- sioned experts to summarizeknowledge about these risk factors,the appropriate preventive interventions,and key conceptual andmethodological issues. These reviews summarize ourknowledge base for prevention efforts. Chapters 1-3 focus on prevention inchild and adolescent psychiatricdisorders including epidemiology, identification,and behavioral risk factors.Also addressed are dimensions for change inservice delivery and treatment,research, and training in the profession. The Pew Charitable Trusts, theIttleson Foundation, and the vanAmeringen Foundation, Inc., provided support forthis project. The AACAP is amedical association whose main objective is toprovide a national forum for the stimulation and advancement of medicalcontributions to the knowledge, diagnosis,and treatment of psychiatric illnessesof children and adolescents. PewCharitable Trusts, the Ittleson Foundation, andthe van Ameringen Foundation,Inc., recognized the importance of providingprofessionals and the public with an adequate knowledge base about preventiveintervention and therefore provided critical support for this importantproject. In the 'United States, 12 percent of our63 million children are affectedby an identifiable maladjustment. About 3 millionof this group suffer from serious emotional illness. Prevention offers thebest hope of alleviating the problemsof mental illness and alcohol and otherdrug use. Prevention is ultimately theonly logical solution to the problem of largenumbers of mentally ill and emotionally dysfunctional childreneven werethere enough caregivers, treatment programs, and supportservices. In this volume, major authoritiesreview selected areas of risk research toaid all readers who are planning orconducting programs for prevention; other chapters discuss general issues related totheoretical problems, research, and implementation. Reference lists followingeach chapter are complete and current, and the references are readilyavailable. Included are clear recommendations regarding implementationof preventive interventions for specific disorders and dysfunctions.

r) vi PREVENTION OF MENTAL DISORDERS

The publication of this information ina single volume makes a significant contribution to the broad distribution ofvaluable prevention information to colleagues ir the medical specialties, alliedprofessions, teaching, and training institutions, and the general public. OSAP and AACAP wish toexpress appreciation to the following individuals who have contributed to the important workof this project. Jerry M. Wiener, M.D., President American Academy of Child and AdolescentPsychiatry PREVENTION OF MENTAL DISORDERS vii

Project Prevention Steering CommitteeMembers David Shaffer, M.B., B.S., F.RC.P., F.RC.Psych., Project Chairman ColumLia University, New York Norbert Enzer, M.D., Project Co-Chairman Michigan State University, East Lansing Irving Philips, M.D., Project Co-Chairman University of California, San Francisco Thomas Anders, M.D. Brown University, Providence, Rhode Island Virginia Q. Anthony, BA Executive Director, American Academy ofChild and Adolescent Psychiatry, Washington, D.C. Michael Cohen, M.D. Albert Einstein College of Medicine, Montefiore Medical Center, New York Leonard Lawrence, M.D. University of Texas Health Science Center,San Antonio Raymond Lorion, Ph.D. Uniwasity of Maryland, College Park Frank Rafferty, M.D. Healthcare International, Austin, Texas Naomi Rae Grant, M.D. Children's Hospital of Western Ontario, London,Ontario, Canada Jon Shaw, M.D. National Institute of Mental Health, Washington,D.C. Archie Silver, M.D. University of South Florida, Tampa Morton Silverman, M.D. University of Chicago, Illinois Elizabeth Q. Bulatao, M.A. Project Prevention Coordinator American Academy of Child and AdolescentPsychiatry, Washington, D.C. viii PREVENTION OF MENTAL DISORDERS

Gvests at Steering Committee Meetings David Faso ler, M.D. Shelburn, Vermont George Comerci, M.D. Department of Pediatrics University of Arizona at Tucson

Financial Support Pew Charitable Trusts Rosman Siegel, Senior Program Associate Ittleson Foundation David Nee, Executive Director William Beaty (Executive Director until 1988) van Ameringen Foundation, Inc. Patricia Kind, President

Contributors Kathleen Bacon, Ph.D. Research Associate New York Psychiatric Institute New York, NY William W. Eaton, Ph.D. Associate Professor of' Mental Hygiene Johns Hopkins University Baltimore, MD Leon Eisenberg, M.D. Maude and Lillian Presley Professor andChairman Department of Social Medicine and Health Policy and Professor of Psychiatry Harvard Medical School Boston, MA Barbara H. Fiese, Ph.D. Assistant Professor of Psychology Syracuse University Syracuse, NY

9 PREVENTION OF MENTAL DISORDERS ix

Ann Garland, M.A. Research Associate New York Psychiatric Institute New York, NY Rosa A. Hagin, Ph.D. Professor Psychiatry Graduate School of Education Fordham University Lincoln Center, NY Karol L. Kumpfer, Ph.D. Department of Psychiatry and Social Research Institute University of Utah Salt Lake City, UT Raymond P. Lorion, Ph.D. Professor of Psychology and Director of Clinical Training University of Maryland College Park, MD Kathi Nader Director of Evaluation Prevention Intervention Program in Trauma, Violence, and Sudden Bereavement University of California Neuropsychiatric InstAute and Hospital Center for the Health Sciences Los Angeles, CA Terence M. Nolan, M.D., Ph.D. Senior Lecturer Melbourne University Department of Pediatrics Royal Children's Hospital Melbourne, Australia D.R. Offord, M.D. Department of Psychiatry McMaster University Hamilton, Ontario, Canada Irving Philips, M.D. Professor of Child Psychiatry Department of Psyc'aiatry University of California San Francisco, CA PREVENTION OF MENTAL DISORDERS

Barry Pleas, M.D. National Health Scientist Professor of Pediatrics McGill University Montreal, Canada Richard H. Price, Ph.D. Professor of Psychology and Director of Industrial-Organizational Psychology University of Michigan Ann Arbor, MI Robert S. Pynoos, M.D., M.P.H. Director, Prevention Intervention Programin Trauma, Violence, and Sudden Bereavement University of California Neuropsychiatric Institute and Hospital Center for the Health Sciences Los Angeles, CA Michael Rutter, C.B.E., M.D., F.R.S. Honorary Director MRC Child Psychology Unit Department of Child and Adolescent Psychiatry Institute of Psychology De Crespigny Park, Denmark Hill, London Arnold J. Sameroff, Ph.D. Profeesor of Psychiatry and Human Behavior Brown University and Director of Research Bradley Hospital East Providence, RI Dricl Shaffer, M.B., B.S., F.RC.P.,F.R.C., Psych, New York Psychiatric Institute Department cf Child Psychiatry New York, NY Archie A. Silver, M.D. Professor and Director Division of Child and Adolescent Psychiatry College of Medicine University of South Florida Tampa, FL PREVENTION OF MENTALDISORDERS xi

Morton M. Silverman, M.D. Associate Professor of Psychiatry Director, Student Counselingand Resource Service The University of Chicago Chicago, IL CONTENTS

Page

Introduction, Irving Philips Commentary: The Integrationof Problem and Prevention Perspectives:Mental Disorders Drug Associated with Alcohol and 7 Use, Morton M. Silverman

in Prevention, Chapter 1: Conceptual Issues 23 Arnold J. Sameroff andBarbara H. Fiese

Chapter 2: The Prevention ofChild and Adolescent Disorders:From Theory P. Lorion, to Research, Raymond 55 Richard H. Price, and WilliamW. Eaton

Chapter 3: PreventionProgramming as Organizational Reinvention: From Research toImplementation, 97 Richard H. Price and RaymondP. Lorion

Chapter 4: Public Policy:Risk Factor or Remedy? 126 Leon Eisenberg

Chapter 5: Psychiatric Disorderin Parents 167 as a Risk Factorfor Children, Michael Rutter

Chapter 6: Risks for MaladjustmentAssociated with Chronic Illness inChildhood, I. Barry Pless 191 and Terence M. Nolan

Chapter 7: Prevention ofPsychiatric Morbidity in Children After Disaster,Robert S. Pynoos 225 and Kathi Nader

r xiv PREVENTION OF MENTAL DISORDERS

Page Chapter 8: Conduct Disorder. Risk Factors and Prevention, D. R. Offord 273

Chapter 9: Prevention of Alcohol and Drug Abuse: A Critical Review of Risk Factors and Prevention Strategies, Karol L. Kurnpfer 309

Chapter 10: Prevention Issues in Youth Suicide

David Shaffer, Irving Philips, Ann Garland, and KathleenBacon ..373

Chapter 11: Prevention of Learning Discrders, Archie A. Silver and Rosa A. Hagin 413

Chapter 12: Prevention of Psychiatric Disorders in Children and Adolescents: A Summary of Findings and Recommendations from Project Prevention, David Shaffer 443

Index 457 Introduction

In pre-Salk days, the dread specterof poliomyelitis was always onthe horizon. Summers were looked to with awesomehorror; no one knew who wouldbe afflicted. The word itself conjureddread and fear. There was noprevention and no cure. Once theanterior horn cells of thespinal cord were involved, the healthy were made lame. There waslittle to be done except forcozsolidation, physical therapy, and orthopedic devices.On the scene came anurse-physical therapist with a new method toreduce the disability and, in some cases,"cure" the disorder. Many said thatSister Kenny was a miracle worker.Her methods consisted of massage and warmpacks. Some responded to herministrations; others, not at all. The cures werepublicized with fanfare. Her methods were controversial and often debated.In time, most studiesrevealed that she provided support, but little change,in the course of the disease.Some of the afflicted spontaneously recovered, as waswell known in the morbidhistory of polio, but many remainedunchanged. Nevertheless, she was anundaunted worker convinced of her methods.It was claimed that only she andher disciples were effective.The others did not understand orhave sufficient training or experience or were unwilling to acceptthe "new." Sister Kenny was acontrover- sial figurembunderstood andoften criticized. Nevertheless, shepersisted. The day the polio vaccine appeared,she disappeared, as did the su mmer scourge. Prevention was achieved. In thefield of child and adolescentpsychiatry, ilto Sister Kenny living and well? Of course, there are differencesbetween infectious disease andmental illnesses and emotionaldysfunctions. We cannot expectsomething as simple as a vaccine tosolve the fundamental problems ofmental illness. Prevention requires attention to a variety ofissues, including genetics, insults tothe central nervous system,nutrition, and physical and socialenvironmental factors. It will involve attention to thefamily and child-rearing pructicesand social institutions such as the public schoolsand day care or preschoolexperience.. It will involve the study of riskfactors with multiple causality andcomplex interventions. It will involve thedevelopment of social policy andreordering of priorities to emphasize a systemof values and ethics that is trulychild and family centered. There will be no"magic bullet," but an evolvingchange in our orientation to the child and family toprovide purpose and dignity in the developmental years. There have been gains over the years,but the number of mentally illchildren who remain untreated is visible among us.The hurt children and their families suffer alone, often unattended, and thenumbers are growing. The epidemiological data indicate that 10 percentof children and youth need mentalhealth services, and many believe this is anunderstatement. But, if we consider this figure a close approximation, it seemsunlikely that there will be sufficient psychiatric or mental health servicesfor children and youth to meet the overwhelming need. We continue to treatthe few, while the great majority

1 2 PREVENTION OF MENTAL DISORDERS

remain unserved. "We have developeda philosophical approach that emphasizes a triage mentality ratherthan one of spontaneously helpingthe afflicted. We are asked to provideBand-Aid treatment for serious problems, and there is little thought of prevention"(Philips 1985). The earlier roots of child psychiatrybegan to flower in the early 1930s with private foundations emphasizingsupport of prevention. Itwas their hope and intention that, if mental illness could bediagnosed and treated in childhood, the incidence in adults would diminish.To pursue this effort, demonstrationclinics were established in the mkjor cities of the United States.Traveling clinics with a social worker, psychologist, and psychiatristestablished child guidance centers. These professionals hopedto prove the worth of the clinics,convince the cities that they shouldcontinue them, andmove on. This was a noble thought, but it remainedan unrealized dream.Illness in children proved difficult to treat, and slrIrt-termtreatment became longer and longer.The enthusiasm of thatera did not reach its original objective, andthe model for prevention faded. Preventionwas relegated to secondary and tertiary modes. Little primary preventionwas achieved. We were treating the afflicted. We need to develop primaryprevention models. Thereare those who consider that primary prevention isa failure of social institutions. Thereare those who believe that social action andmikjor changes in the fundamentalprocesses of society need revision. There islittle question that if all of societywere well housed, well fed, and well educated;if each child were well parented;and if there were opportlinity for full employmenton the horizon, many mental health problems wouldb9 ameliorated.Obviously, Utopia is not in sight. Theonly way to make an impacton our society's mental health problems isto establish a comprehensive program forprevention. The complexity ofprograms of intervention is described in the chaptersto follow; for example, MichaelRutter, 'Psychiatric Disorder in Parentsas a Risk Factor for Children"; Leon Eisenberg, 'Public Policy: Risk Factoror Remedy?"; and Karol Kumpfer, 'Preventionof Alcohol and Drug Abuse: A CriticalReview of Risk Factors andPrevention Strategies." Child and adolescent psychiatryprograms have participated little in the preventive arena. Training programs, for the most part, do not provideex- perience in prevention,nor do they ha ve coordinated systems forint Nention. Most trainees have littleexperience in early intervention. Nocurriculum has been established to provide for preventionand in childpsychiatry, there isno mention of training for prevention. Some programs provide liaisonexpeTiences in consultation with pediatricsand family medicine andcommunity agencies, including the schools, but theseprograms allow for no coordinated experience. A curriculum in interventionand prevention is essentialif current knowledge in these areas is to be transmitted to future child and adolescentpsychiatrists. The research agenda is meager. Although substantive research findingshave been achieved, too few of theresults have been applied. (Thereare examples of applied research findings, suchas Kellam's (1972) work on the Woodlawn INTRODUCTION 3

children of psychotic project; Grunebaum's work(Grunebaum et al. 1978) on parents; the work of Rice etal. (1971) with children, ofhospitalized parents.) of equal Intervention programs, curriculumdesign, and a research agenda are importance in assuring preventionof its rightful place inchild and adolescent psychiatry. emphasis on the The research pendulum ofpsychiatry has shifted to an understanding of basic biologicalmechanisms.Molecular genetics seeks genotypes for the majormental illnesses.Any genotype has aphenotypic expression, which presents in aclimate favorable or unfavorable.No matter understanding of genotypes may be,the understanding of the how successful an need process of expressionin a social-culturalenvironment will continue to provide explanation, but aninterven- study. The biologic underpinnings may complexity in all its tim for prevention mustconsider the biopsychosocial dimensions, The chapters byLorion, Price, and Eaton;Sameroff and Fiese; and Offord illustrate this well. Research has establishedsignificant risk factors associatedwith suosequerat psychopathology. Risk correlateswith spectral outcomes. Highcorrelations exist between life events, trauma,and the emergence ofpsychopathology, as portrayed in figure 1. systems may We are beginning to learnwhat interventions and social support be effective in prevention;what fosters invulnerability;and what makes children vulnerable to outcomesthat are detrimental todevelopment. In this regard, prevention is in its infancy.Although there is much to learn, weknow of interventions that a'13effective; e.g., G. Caplan's(1961) work wi:`, premature infants and M.M. Weissman etal. (1986) epidemiologic studiesof depressed children. The chapters of thisvolume provide many additionalexamples. In an effort to move thefield of child and adolescentpsychiatry into greater participation in the area of prevention,in my presidential address tothe American Academy of Child andAdolescent Psychiatry, I suggestedthat we "begin to develop an old enterpriserecommended again and again but never and it was fully implementedprevention.Prevention transformed pediatrics, the impetus for the inception ofchild psychiatry. We will not develop inoculants or fluorides for mentalhealth, but we can have acomprehensive program of preventionthrough research and earlyintervention" (Philips 1985). As a result, a PreventionInitiative was undertaken bythe American Academy of Child and AdolescentPsychiatry (AACAP), It was supportedby public and private funds.* Asteering committee was formed. Inits deliberations, it decided to establish aknowledge base in areas of risk arbitrarily selected to determine what we knowand what has been effective,which eventuated into

*OSAP, the AACP P's AbramsonFund, the Pew Charitable rrusts, theIttleson Foundation, and the van AmeringenFoundation. 't 4 PREVENTION OF MENTALDISORDERS

Invulnerability Life Events/ Spectral Trauma

Risk Factors Outcomes 4kr4. Psychotic Parents Substance Abuse Depressed Parents Divorce Vulnerability Delinquency Suicide and Homicide Alcoholism and Drug Abuse in Family (Substance Abuse) School Dropout Conduct Disorders Teenage Pregnancy Sexual anu Physical Abuse Psychoses and Depression Loss and Separation Learning Disability Mental Retardation Chronic Illness Invincible and Invulnerable Poverty Children of Fate "Acts of God"

Figure L Correlations betweenlife events and psychopathology. this volume. It is nota totally comprehensive review of all of might have been appropriate, prevention work as but rather selectedones well referenced by a series of efforts to provide backgroundmaterial for this study. This these issuee. There volume addresses are also chapters that consider conceptualand social issues, theory to research, andresearch to practice. It providesa basis of our state of knowledge of whatwe have learned and know,as well as frontiers to be explored. It is anew beginning of an old enterpriseto develop a curriculum, examine interventions thatmay prove effective, and establisha research agenda. In a recent article (Bower1987), a metaphoricalrepresentation of the field was described. Bower quoted LutherWoodward, who described test of insanity: an old Cornish

The person to be tested isplaced in a smallroom facing a sink in which there is a spigot, a pail underneaththe spigot, anda ladle in the pail. The spigot is turned on, and the testes istold to krep the water from pail. The overflowing from the person who continues to ladle, howeverenergetically andsucoees- fully, without attendingto the flow from the spigot isjudged insane. It seems that in our profession we are ladlingvery rapidly, and the pail continues to overflow. This volume will be followed by an AACAP publicationthat will map outa program for child and adolescentpsychiatry regarding intervention, riculum, and a research agenda. a cur-

Irving Philips, M.D. University of California, SanFrancisco 1 6 INTRODUCTION 5

References

Bower, E. Prevention: A word whose time has come.American Journal of Orthopsychiatry 57:4-5, Jan. 1987. Caplan, G. Prevention of Mental Disorders inChildhood. New York: Basic Books, 1961. Grunebaum, H.; Cohler, B.; and Kauffman, C. et al.Children of depressed and schizophrenic mothers. Child Psychiatry Hum. Dev.8:219-228, 1978. Kellam, S.G.; Branch, J.D.; and Agrawal, K.C. et al.Woodlawn Mental Health Center: An evolving strategy for planning incommunity mental health. In Handbook of Community Mental Health, edited byGolann, S.E. and Eisdorfer, C. New York: Appleton-Century-Crofts, 1972. Philips, I.President's Plenary Address. The decay of optimism:The opportunity for change. Journal of the AmericanAcademy of Child and Adolescent Psychiatry 25:151-157, Mar. 1985. Rice, E.P.; Ekdahl, M.C.; and Miller, L. Children ofMentally Ill Parents. New York: Behavioral Publications, 1971. Weissman, M.M. Merikangas, KR.; and John, K. etal. Family-genetic studies of psychiatric disorders. Arch. Gen. Ps:/chiatry43:1104, Nov. 1986.

r COMMENTARY

The Integration of Problem and Prevention Perspectives: Mental Disorders Associated with Alcohol and Drug Use

Morton M. Silverman, MD. Associate Professor of Psychiatry Director, Student Counseling and Resource Service The University of Chicago Chicago, Illinois Introduction

It is the aim of this chapter to place the contributions of this monograph into both a broader public health perspectivc and an alcohol, other drug, andmental (ADM) disorders perspective. The prevalence data regarding ADMdisorders in children and adolescents will be br:Aly highlighted.This wilt introduce a discussion of comorbidity and multiple diagnoses in children and youth. After the presentation of the concept of multiple problem behaviors, fundamental prevention concepts and approaches will be presented from a public health perspective. The chapter will highlight the role of the school system in implementing and integrating various preventive intervention techniques and approaches targeting ADM disorders and dysfunctions. Finally, the contributions of Project Prevention will be placed in the broader context of national public health goals and objectives. Problem Perspectives Prevalence Although there have been some promising reductions recently in the use of alcohol and other drugs by adolescents as measured by the National High School Senior Survey (Johnston et al. 1987), problems of ADM disorders in children

This paper was prepared at the invitation of the Division of Communication Programs, Office for Substance Abuse Prevention, ADAMHA. It was written subsequent to the completion of Project Prevention and the editorial preparation of this monograph. r 8 PREVENTION OF MENTAL DISORDERS and youth remain prevalent and perplexing.A wide array of national data collection sources and scholarly reviews supportthe same conclusions: Asa nation we must seriously address the psychologicalneeds and chemical dependencies of our children and adolescents. Some facts and figures illustrate theseriousnees of the problem: Alcohol and other druguse significantly increases the risk of transmission of the human immunosuppressive virus(HIV) directly through the sharing of contaminated needles, throu&:isexual conficts with in- travenous drug users or other drug injectors, andthr t. igh in utero infection, and indirectly through adverseeffects on immune system functioning and the increased risk of unsafesexual practices (Petrakia 1988).

The use of cigarettes, alcohol, andmarijuana increases the risk ofuse of other illicit drugs. Theuse of these drugs is correlated with other health problems, including adolescent suicide,homicide, school dropouts, motor vehicle crashes, delinquency, andprecocious sexual activity andun- wanted pregnancy (NIDA National Surveyof Drug Abuse, 1985). Extrapolations from dataon drinking practices obtained from household probability surveys suggests that thereare approximately 6.6 million children of alcoholics under theage of 18. Although they are at increased risk for alcoholism,a large percentage do not develop this condition. They may, however, develop other drug abuseor mental disorders (NIAAA, Research on Children of Alcoholics GrantAnnouncement, 1989). Although State laws have made alcoholan illegal drug for people under 21 years of age, 36 percent of highschool seniore report that within the 2 weeks prior to being surveyed, theyhad five or more drinks ina row once or twice each weekend; 92 percent have hadexperience with alcohol and 66 percent have used in thepast month. Even more troublesome is that the High School Senior Survey dealswith mainstream youth and does not capture dataon alcohol and other drug use among school dropouts (NIDA National High SchoolSenior Survey, 1987). The 1985 Household Survey showed that illicit druguse among 19- to 21-year-old high school dropoutswas 67 percent higher than for high school graduates (NIDA National Household Survey, 1985). Seventy-seven percent of eighth-gradestudents have tried alcohol; of these, 55 percent report first trying it bysixth grade. Eighty-nine percent of tenth-grade students reporthaving tried an alcoholic beverage;of these, 69 percent report firstuse by eighth grade (National Adolescent School Health Survey, 1987). Fifteen percent of eighth-grade studentsreport having tried marijuana; of these, 44 percent report firstuse by sixth grade. Thirty-five percent of tenth-grade studentsAeport having triedmarijuana with 56 percent of COMMENTARY 9

them reporting first use by eighth grade (National Adolescent School Health Survey, 1987). Five percent of eighth-grade students and 9 percent of tenth-grade students report having tried cocaine. Two percent of eighth-grade students and 3 percent of tenth-grade students report having used cocaine during the past month (National Adolescent School Health Survey, 1987). In 1986, emergency rooms (ER) reported 119,263 drug abuse episodes; 13,343 (11.2 percent) of the episodes involved patients 10 to 17 years old. Approximately 6 out of 10 of the youth ER visits were related to a suicide attempt or gesture. The drugs mentioned most frequently by young ER patients were aspirin, acetaminophen, alcohol-in-combination, marijuana, and cocaine (NIDA Dawn Report, 1986). Suicide is now the second leading cause of death for Americans between the ages of 15 and 24 (National Center for Health Statistics, March 1989 report: Health United States: 1988). The first leading mortality i isk for this age group continues to be motor vehicle crashes, about half of which are linked with alcohol use (NIAAA-Sixtli Special Report to Congress). Depressive disorders are a major contributor to adolescent suicide, which has increased 300 percent during the past three decades. The increasing use of alcohol and drugs among youth may be a method of self-medication for depression. Conversely, this alcohol and other drug use may precipitate a depressive disorder. Deykin and colleagues interviewed 424 college students between the ages of 16 and 19 with a standardized epidemiological interview (the National Institute of Mental Health Diagnostic Interview Schedule, or DIS) to assess lifetime prevalence rates of major depressive disorder, alcohol abuse/dependence, and other drug abuse/dependence (Deykin et al. 1987). Lifetime prevalence rates were: major depressive disorder, 6.8 percent; alcohol abuse, 8.2 percent; other drug abuse, 9.4 percent. Subjects with a history of alcohol abuse were almost four times as li'.dy to report a history of major depression than other nonabusing adolescents, oat not more likely to report a history of other noneffective psychiatric disorders. Subjects with a history of drug abuse were more than three times as likely to report a history of major depression than other nonabusing adolescents, as well as more likely to report a history of other noneffective disorders. Data on age of illness onset indicated that for subjects reporting both disorders, a first episode of major depression usually preceded the emergence of alcohol abuse by more than 4 years, suggesting that young persons who develop a mood disorder in childhood or adolescence are much more likely than their peers to develop an alcohol or other drug use disorder. These findings are consistent with the hypothesis that adolescents begin using alcohol and illicit drugs to alleviate existing painful mood states (for example, depression, loneliness, or low self-esteem), but should not be taken out of context to mean that this isa final or best explanation oi' alcohol and other drug use (Kaplan 1977; Kandel 1982). 10 PREVENTION OF MENTAL DISORDERS

Summarizing the extensive epidemiologic and etiological research literature, nryfoos, in her rsport to the Carnegie Council on Adolescent Development, estimated that 3 million children and youth, aged 10 to 17, are already in serious trouble and are experiencing multiple problems resulting from alcohol and other drug use, unwanted pregnancy, school failure, and delinquency (Dryfoos 1987). An additional 4 milPon children and youth are estimated to be engaged in multiple problem behaviors (school failure, alcohol and other drug use, and early unprotected intercourse leading to childbearing) and are at high risk of serious health and social consequennes. Another one-fourth of this cohort (7 million)are at risk of the consequences of problem behaviors because they may experiment with smoking and drinking, may engage in sexual activity but use contraception, may be doing poorly in school but not failing, and may occasionally be truant or commit other minor offenses. The Carnegie Council report concludes that "the future of 1 in 4 of our youth is in jeopardy unless intensive interventions and treatment are initiated to ameliorate their problems" (Carnegie Council 1989). A number of risk factors seem to be linked with the subsequentemergency of multipis problem behaviors, including low academic achievement,susceptibility to peer influence, inadequate family management and parentalsuper- vision, nonconventionality, sensation-seeking behavior, early alcohol and other drug use (including tobacco), early aggressive and/or acting-out behavior, and diminished self-esteem and self-efficacy.Children are also found to be at increased risk when their attitudes toward educationare negative or when their adjust, nents to school are poor (Kellam et al. 1982). Comorbidity Contemplating prevention effortu for ALM disorders in children and adolescents necessitates clear etiologies for the development and onset of these disorders and dysfunctions. Epidemiological studies identifying theage of onset of major psychiatric disorders and dysfunctions, including alcohol and other drug use, suggest some overlapping ages of onset fora range of emotional disorders and behavioral dysfunctions (Robins et al. 1984; Regier et al. 1988; Johnston et al. 1988; Kandel et al. 1986; Kandel and Davies 1986). These common ages of onset suggest possible universal etiologies and environmental stimulants for expression of ADM disorders. The clinical and treatment litera- tures have coined the phrases "dual diagnosis" and "comorbidity" to describean increasingly common clinical presentation of individuals with both chemical dependencies and emotional/psychiatric disorders and dysfunctions (Ross et al. 1988; Mirin et al. 1988). The issue of "which came first" has takena back seat to the more urgent concern that these dual problems tend to coexist andserve to maintain each other's expression in pathological ai .! detrimentalways (Mirin 1984; Alterman 1985). Specifically, numerous epidemiologic and research studies during the past 15 years indicate that a large percentage of children and adolescents are st risk for developing multiple problems such as juvenile onset depression, alcohol and other drug use, suicidal behavior, dropping out of school, delinquency, running COMMENTARY 11 away from home, and unwanted pregnancies(Moskowitz and Jones 1988; Kellam et al. 1983; Kandel and Yamaguchi 1986; Donovan and Jessor 1985). A large number of youth are at risk of developing at least one, and probably more than one, of these serious health and social problems. The Anti-Drug Abuse Act of 1986 (Public Law 99-570) defines a "high risk youth" as "any individual who has not attained the age of 21 years who is at high risk of becoming or who has become a drug abuser or an alcohol LIbuser and who: (1) is identified as a child of a substance abuser; (2) is a victim of physical, sexual, or psychological abuse; (3) has dropped out of school; (4) has become pregnant; (6) is economically disadvantaged; On has committsx1 a violent or delinquent act; (7) has experienced mental health problems; (8) has attempted suicide; (9) has experienced long-term physical pain due to injury; or (10) has experienced chronic failure in school." Additionally, other high-risk situations, settings, and behaviors are of major concern, including being a child of a psychiatrically ill parent (Silverman, in press; Rutt9r, this volume) and being exposed to chronic family disruption and marital discord. The available evidence substantiating the relationship between alcohol and other drugs and behavioral problems, emotional problems, accidents and natural disasters, suicides, physical illnesses, and learning problems is summarized in this volume, and even a cupsory review strongly suggests that the available information is quite compelling in terms of identifying commonalities of comorbidity. All too often primary care r,hysicians and health professionals do not associate alcohol and other drug problems with these and other behavior problems (Coulehan et al. 1987; Kamerow (t al. 1986; Bridge et al. 1988). Clinically, we know that children and adolescents suffering from depression, physical problems, low self-esteem, and attention-deficit and hyperactivity disorders may turn to alcohol and other drugs for relief from emc ional pain, psychic discomfort, and feelings of low self-worth (Powers and Kutash 1980; Parker et al. 1987; Friedman et al. 1987). According to Macdonald, "adolescents with drug or alcohol problems appear in emergency rooms as victims of trauma, accidental overdose, or suicide attempts. More often, however, pediatricians see young abusers for routLe care or problems not usually thought of as drug-related. Fatigue, sore throat, cough, chest pain, abdominal pain, headache, and school or behavioral problems are the most common symptoms of drug use. Awareness of the epidemic and serious health consequences cf alcohol and other drug use should force the pediatrician to consider abuse seriously in all adolescents, especially those with suggestive symptoms" (Macdonald 1984). It is now wiAl known that some of the common side effects of the use of alcohol and other drugs include behavioral problems, physical problems, depressed mood, and misperceptions of self (poor body image and poor self-perception). 12 PREVENTION OF MENTAL DISORDERS

Recent studiee suggest an association betweenchronic physical illness and behavioral dysfunctions (Pless and Nolan, thisvolume). Othor researchsuggests an association between chronic physicalillness and increased alcohol and other drug use (Wells et al. 1988). Prevention Perspectives Public Health

Within the public health field, the goalof prevention encompasses both preventing negative outcomes and enhancingpositive results. In this context, Project Prevention struggled to addressacme underlying assumptions related to developing prevention programming: (1)we can accurately predict the future; (2) we can accurately identifypopulations who are now in need of preventive interventions; (3) we have available thespecific preventive interventions to prevent specific negative outcomes; (4) thepreventive interventionscan be eisectively and efficiently implemented;(5) no long-term negativeconsequences are associated with the intervention itself; and (6) predictednegative outcomes will turn out to be negative and will beof long-standing duration. Highlighted in these chaptersare the need for (1) increased specificity of target populations to receive the preventiveinterventions; (2) increased specificity of the preventive interventions;(3) increased specificity of theoutcome measures to be attained subsequent to the intervention;and (4) refinement of causal models that link these threevariables (target population, preventive intervention, outcome measures). Weare concerned not only about the at-risk problem that weare attempting to forestall, alleviate, modify,attenuate, or prevent, but also about the goal of enhancing,promoting, protecting, and maintaining mental health, mental well-being,and stab..ity of the individual over time and across many physical and mentalparameters. Key questions are: Whatare the risks? What is the likelihood of expression? What are the protective factors (individual,environmental, cultural) that will prevail? What are the preventive factors(individual, environmental, societal) that may be called into play? Evidencesuggests that the development of ADM disorders among youth is associated withmultiple risk and resiliency factors that are inherent within the individual(e.g., genetics, personality, physical health), the individual's environment (e.g.,family, peers) and the individual's interaction with his or her environment.Likelihood that a youngperson will use alcohol and other drugs appears to increaseas the number of risk factors increases and the number of resiliencyor protective factors decreases. Risk and resiliency factors affecting high-risk youthmay include immediate and extended family, peers, school, neighborhood,commun4, and the larger society.Special stresses and protective factorsmay be associated with the membership ofmany high-risk youth in racialor ethnic minoritiesAny preventive or treatment intervention is likely to bemore effective if it focuses on reducing thepower of 0 COMMENTARY 13 risk factors and increasing the potency of emiliency factors acrossseveral envizimmental levels.Preventive intervention models proposing to target single-risk factors are likely to be lees effective. Models in relatedhealth and safety fields address these issues (Institute of Medicine 1986). Some of these models rely heavily on identifying and quantifying host, agent, and environmental factors. In the ADM disorders prevention field, we have just begun to identify thoee risk factors that in combination will increase the likelihood of expression of a negative eventthe development of incapacitating symptoms, the expression of behavioral dysfunctions, the inappropriate use and misuae of licit and illicit drugs, or the movement toward self-injurious behaviors. The yet-to-be quantified host, agent, and environmental factors are those associated with the spectrum of outcomes that can occur once the process has begun. In summary, the ADM prevention field has begun to tentatively identify certain risk settings, risk behaviors, and risk situations that are predictive of specific outcomes. Furthermore, we are beginning to recognize that some of these predicted negative outcomes may, in fact, be transitory and may not. prevail over time. This spectrum of outcomes serves as a cautionary note as the ADM prevention field moves to increasing specificity of target population, preventive intervention, and outcome measurement. There does not seem to be a distinct relationship between a particular risk status and a distinct psychiatric or behavioral outcome. For example, not all children of alcoholics develop alcoholism, other drug use, or mental disorder. On the other hand, there do seem to exist certain basic ingredients and building blocks that are eesential to the general maintenance of a healthy physical and ennetional trajectory. Theseessential ingredients may be perceived as either protective factors or preventive factors, or both. Many of these ingredients are highlighted in this monograph. The prevention cf ADM disorders requires a multidisciplinary approach, which ic based on research findings from a number of interrelated scientific disciplinespopulation-bused epidemiology, genetics, clinical studies (including outco Ile and followup studit neurobiologi and biotechnology, and the identification of biological markers for mental disorders (Pardee ot al. 1989). Various models for the development and maintenance of ADM diaorders have been proposed in this volume and have appeared elsewhere (Silverman and Koretz 1989). Complicating the attempts to develop clear models is the knowledge that youth are often in developmental transition, resulting in a fluidity of behaviors, attitudes, experimentation, and values that may be tran- siently influenced by role models and environmental factors. For example, the common behaviors of curiosity and brief experimentationregarding licit and illicit drugs is worthy of further study and consideration prior to developing specific alcohol and other drug use prevention programs directed at all youth (Yamaguchi and .Kandel 1984).

r e 14 PREVENTION OF MENTAL DISORDERS

Integrating Interventions The emergence of the concepts of dual diagnosis and comorbidity have highlighted the need for more holistic thinking about problems encountered by children and adolescents that may be amenable to preventive interventions. The precise nature of the interrelationship have yet to be elucidated, butmany data support the pursuit of these relationships (Kandel and Yamaguchi 1985; Jessor and Jessor 1977; Kellam and Brown 1982; Robins and Przybeck 1985). Dryfoos' summary of the epidemiological literature fmds that 'Theemergence of common predictors of multiple problem behaviors lends force to theargu- ment that int xventions should focus more on the predictors of the behavior than on the behavior itself. These findings lead to the conclusion that enhancement of early schooling in preventing school failure should receive high priority not only from those interested in lowering the dropout rate, but also for those who are interested in preventing substance abuse, pregnancy, and delinquency." Eisenberg (this volume) agrees that we must look forcommon antecedents for these disorders. Others have argued similarly when discussing approaches to the prevention of youth suicide (Felner and Silverman 1988) and stress-related disorders (Bloom 1979). Numerous etiologic research studies indicate a wide variety of social,emotional, developmental, behavioral, and biologic factors thatmay place a child at risk of developing these disorders. Kerlin suggests that the aniecedents of multiple problem behaviors appear to be highly intercorrelated andmay form a constellation of precursors common to the emergence of these health problems (Kellam et al. 1982. )Some researchers suggest that the antecedents of drug and alcohol use, school dropouts, delinquency, anda host of other problems can be identified in the early elementary grades, long before the actual problemsare manifested (Elliott and Huizinga 1987; Hawkins et al. 1985). Dryfoos concludes that "early sexual activity, early childbearing, early initiation of smoking and alcohol use, heavy drug use, low academic achievement, school misbehavior, school drop-out and delinquencyare interrelated" (1987, pp. 39-40). The Carnegie Council report concludes that "four attrihutesemerge as characteristics of those young people who exhibit all of the behaviors: doing poorly in school, being a non-conformist, going around with friends who act out in the same ways, and having inattentive parents." This profile differs inmany important respects from the congressional definition of high-risk youth because it includes more of a psychological and familial context to understanding youth at risk Thus, one common findMg is that early school problems (e.g., attention deficits, learning disabilities, acting-out, and conduct problems)are often associated with many child and adolescent ADM disorders. This suggests the development of sites for the provision and application of preventive intervention programs in school settings. Of critical importance is that preventive interventions be tailored to the receptivity and developmental stages of thetarget

'7 COMMENTARY 15 audience. Also, the interventions must have high credibility with the target audience. Dryfoos emphasizes the school setting as the major locus for preventive intervention activities, arguing that enhancement of early schooling in preventing school failure should receive high priority as interventions that would prevent later disability and dysfunction across a range of bahaviors (Dryfoos, in press). Dryfoos emphasizes two mkjor points. First, it is critical to keep chiliren and adolescents in a school setting for purposes of delivering preventive vention messages addressing a range of risk-taking behaviors:alcohol and other drug use and abuse, early sexual activity, suicide and life-threatening activities, criminal activities, and so forth. Second, the fact that one remains in school and is expceed to the positive attributes of a school environment (e.g., education, socialization, peer support, and physical exercise) may well protect an adoleycent from engaging in certain risk-taking behaviors.A major assumption underlying this position is that school settings do not contribute to the risk. In other words, school environments must be deeigned to be health promotive and disease preventivenot risk enhancing. Recent programmatic developments for families and schools offer hope that the childhood antecedents of adolescent problems can be prevented or remediated. Promising technologies exist for improving parenting skills and strengthening family function, for improving instructional practices, and for restructuring schools to improve students' performance, adjustment, and commitment to education (Felner et al. 1982). Although students usually spend more of their waking and learning hours in the school environment than they do with their own families, a school-based program can not afford to be school limited. As one of the most essential institutions in a community, a school's well-being and its concerns interact with those of business, governmental, legal, religi3us, health care, service, and social groups that rlake up the community. All of these groups want to play a role in what goes Ohithin the school's walls and on its playgrounds, and all can supply talent, expertise, and resources (Hawkins and Lam 1987). Major conclusions from this monograph and that of other recent studies (e.g., Carnegie Council on Adolescent Development) are that a growing body of research has documented the statistical associations among so-called "problem behaviors": alcohol and other drug use, delinquent behavior, early childbearing, and their sequelae (e.g., alcohol-impaired driving, violence, early parenthood, and sexually transmitted diseases).It has become clear that each of these problems has common antecedents, the most consequential of which is school failure. Young people who use drugs, who commit delinquent acts, and who become parents at early ages are much more likely to have been failing in school or have already dropped out of school than those who avoid the behaviors. A young person who is simultaneously failing in school, using drugs, acting out, and having unprotected sexual intercourse is at risk. The prevalence of any one of these behaviors has concomitant risks and negative consequences; when 16 PREVENTION OF MENTAL DISORDERS they occur as a "package," the risk is greatly exacerbated. Early initiation of any one behavior often predicts the others; the first event in this progression is typically failure in school. It has been suggested that children who willengage in this range of high-risk behaviors can be identified as early as second grade. Research has shown that being in a low-income family and living ina poor neighborhood compounds children's risk for multiple problems. Particularly among disadvantaged children, it has been documented that early intervention can significantly improve outcomes.Early childhood education and social supports for families have been shown to measurably increasesuccess in the elementary school years, and that success, in turn, improves educationalper- formance over time (Dryfoos, in press). Children who have hadaccess to early interventions have been proven to have lower rates of alcohol and other drug use, delinquency, and pregnancy during adolescence.

Toward the Year 2000

In order to address multiple problem behaviors of children andadolescents and their possible antecedent conditions, attention must be drawnto their importance. The Endings of the innovative Federal initiative, the 1990 Health Objectives for the Nation, have provided the needed documentation andsupport for the next steps in this processspecifically, the development ofa national program for the amelioration of these pk oblem behaviors (DHHS/PHS 1986). Such a potential national agenda is being formulated under the auspices ofthe Year 2000 Health Objectives for the Nation. This public health service project is designed to set health objectives for the Nation, particularly for mental health and alcohol and other drug problems. The monograph summarizes the extensive literature regarding theprevention of some negative outcomes, including learning C'3orders, alcohol andother drug use, suicide, and conduct disorders. Anyone or more of these outcomes may result from the three known environmental risk conditions reviewed in this monograph: being the child of a parent withan ADM disorder (Ruttur chapter); suffering from a chronic physical illness (Pless and Nolan chapter); andbeing exposed to natural disasters (Pynoos and Nader chapter). This isnot to suggest, however, a one-to-one association between the three known riskstatus conditions and these four preventable disorders. The exactnature of the ADM illness that is expressed is not specified because of the lack of precision inour etiological and causal models of psychopathology. Bridging thegape between the identification of risk status conditions and the pstablishment of theirrole in the development of negative outcomes is the work that challengesthe prevention research field. The challenges are to identify thecommon antecedents and precursors, the common environmental conditions, and the common solutions and interventions that prevent certain outcomes and promote others. COMMENTARY 17

The deliberations and directions highlighted in this monograph are relevant to current concerns about developing measurable health objectives inreducing alcohol-impaired driving, reducing fetal alcohol syndrome, enhancing awareness of the detrimental effects of drugs onphysical well-being, increasing the average age of first use, decreasing multipleproblem behaviors, decreasing the prevalence of depression, decreasing youk suicide, reducing stress-related disorders, increasing the number of physicians in training who are aware of multiple problem behaviors (especially alcohol and other drug use, depression, suicide, and strees), and increasing the awareness of professional health organizations in identifying and treating alcohol and other drug use. In fact, the Project Prevention effort directly relates to at least 7 of the 20 rational health priority areas for the year 2000. Conclusion

The work of Project Prevention represents true interdisciplinary concern for those with ADM disorders. The prevention of ADM disorders will not come about easily without coordinated, comprehensive, and collaborative efforts from many clinical and scientific disciplines. Preventionefforts have moved from being potential to possible to practical. The authors of this monograph suggest those pract* Ilinterventions that have probable benefits for selective conditions and can be implemented in therapeutic and clinical settings. The field of prevention has always been action oriented and future directed. The process has been one of searching for (1) universal concepts that explain the development of ADM disorders and dysfunctions; (2) universal, essential ingredients for preventive interventions that have broad-based effectiveness; and (3) universal risk factors that have negative impact for individuals or groups of individuals. From this search for universal concepts, theories, and essential ingredients will come the development of specific targeted preventive interventions for well-defined high-risk populations. Some of those risk factors, at-risk populations, and preventive interventions have been highlighted in this contribution to the ADM prevention field. 18 PREVENTION OF MENTAL DISORDERS

References

Alterman, A.I., ed. Substance Abuse and Psychopathology. New York:Plenum Press, 1985. Bloom, B.L. Prevention of mental disorders: Recent advancesin theory and practice. Community Mental Health Journal 15(3):179-191, 1679. Bridge, T.P.; Mirsky, A.F.; and Goodwin, F.K., eds.Psychological, neuropsychiatric, and substance abuse aspects of AIDS. In: Advances in Biochemical Psychopharmacology. Vol. 44. New York: Raven Press, 1988. Carnegie Council. 'Turning Point" Report. Carnegie Councilon Adolescent Development's Task Force on Education of Young Adolescents.Washington, D.C.: Carnegie Council on Adolescent Development, 1989. Coulehan, J.L.; Zettler-Segal, M.; Block, M.; McClelland,M.; and Schulberg, H.C.. Recognition of alcoholism and substance abuse inprimary care patients. Archives of Internal Medicine 147:349-352, 1987.

Deykin, E.Y.; Levy, J.C.; and Wells, V. Adolescent depression,alcohol and drug abuse. American Journal of Public Health 77:178-182,1987. Donovan, J., and Jessor, R. Structure of problem behavior inadolescence and young adulthood. Journal of Consulting and Clinical Psychology 35:6, 1985. Dryfoos, J.G."Youth at Risk: One in Four in Jeopardy." Workingpaper submitted to the Carnegie Councilon Adolescent Development, 1987. Dryfoos, J.G. Determinants of early childbearing and otheradolescent risk- taking behavior: A framework for prevention interventions.In: Lipsitt, L., and Mitnick, L., eds.Self Regulation, Impulsivity, and Risk-TakingBehavior: Causes and Consequences. Norwood, N.J.: Albex Publishing,in press. Elliott, D.S., and Huizinga, D. The relationship betweendelinquent behavior and alcohol, drug, and mental health problem behaviors. In:Juvenile Offenders with ADM Problems. DHHS Publication No. (ADM) 87-1503,1987. Felner, RD.; Ginter, M.; and Primavera, J. Primaryprevention during school transitions: Social support and environmental structure.American Journal of Community Psychology 10(3):277-290, 1982. Felner, R.D., and Silverman, M.M. "Primary Prevention:A Consideration of General Principles and Findings for the Prevention of YouthSuicide." Report of the Secretary's Task Forceon Youth Suicide, Vol III: Prevention and Inter- ventions in Youth Suicide. Washington, D.C.: U.S.Government Printing Of- fice, 1988. COMMENTARY 19

Friedman, A.S.; Utada, A.T.; Glickman, N.W.; and Morrissey, M.R. Psychopathology as an antecedent to, and as a "consequence" of, substance use in adolescence. Journal of Drug Education 17(3):233-244, 1987. Hawkins, J.D., and Lam, T. luacher practices, social development, and delinquency. In: Burchard, J.D., and Burchard, S.N., eds. Prevention of Delinquent Behaviors. Beverly Hills: Sage Publications., 1987. Hawkins, J.D.; Lishner, D.; and Catalano, R.F. Childhood predictors and the prevention of adolescent sulastance abuse. In: Jones, C.L., and Battjes, R.J., eds.: Etiology of Drug Abuse: Implications for Prevention. NIDA Research Monograph 56, DHHS Publication No. (ADM) 86-1335. Washington, D.C.: U.S. Government Printing Office, 1985. Institute of Medicine. Injury in America: A Continuing Public Health Problem. Washington, D.C.: National Academy Press, 1985. Jessor, R., and Jessor, S.L. Problem Behavior and Psychosocial Development: A Longitudinal Study of Youth. New York: Academic Press, 1977. Johnson, L.D.; O'Malley, P.M.; and Bachman, J.G. Illicit Drug Use, Smoking, and Drinking by America's High School Students, College Students, and Young Adults, 1975-1987. DHHS Publication No. (ADM) 89-1602. Washington, D.C.: U.S. Government Printing Office, 1988. Johnston, L.D.; O'Malley, P.M.; and Bachman, J.G. National Trench. in Drug Use and Related Factors Among High School Students and Young Adults, 1976-1986. DHHS Publication No. (ADM) 87-1535. Washington, D.C.: U.S. Government Printing Office, 1987. Kamerow, D.B.; Pincus, H.A.; and Macdonald, D.I. Alcohol abuse, other drug abuse, and mental disorders in medical practice.Journal of the American Medical Association 255:2054-2057, 1986. Kandel, D.B. Epidemiologic and psychosocial perspectives on adolescent drug use. Journal of the American Academy of Child Psychiatry 21:328-347, 1982. Kandel, D. B., and Davies, M. Adult sequelae of adolescen t depressive symptoms. Archives of General Psychiatry 43:255-265, 1986. Kandel, D.B.; Davies, M.; Karus, D.; and Yamaguchi, K The consequences in young adulthood of adolescent drug involvement: An overview. Archives of General Psychiatry 43:746-754, 1986. Kandel, D.B., and Yamaguchi, K Developmental patterns of the use of legal, illegal, and medically prescribed psychotropic drugs from adolescence to young adulthood. In: Jones, C.L., and Battjes, R.J., eds. Etiology of Drug Abuse: Implications for Prevention.DHHS Publication No. (ADM) 36-1335. Washington, D.C.: U.S. Government Printing Office, 1985. 20 PREVENTION OF MENTAL DISORDERS

Kaplan, H.B. Increase in self-rejection and continuing/discontinuing deviant response. Journal of Youth and Adolescence 6:77-87, 1977. Kellam, S.G., and Brown, C.H. Social, Adaptational and Psychological Antece- dents of Adolescent Psychopathology Ten Years Later. Baltimore: Jihns Hop- kins University Press, 1982. Kellam, S.G.; Brown, C.H.; Rubin, B.R.; and Ensminger, M.E. Paths leading to teenage psychiatric symptoms and substance use: Developmental epidemiological studies in Woodlawn. In: Guze, S.B.; Earls, F.J.; and Barrett, J.E., eds. Childhood Psychopathology and Development. New York: Raven Press, 1983. Kellam, S.G.; Stevenson, D.L.; and Rubin, B.R. How specificare the early predictors of teenage drug use? In: Harris, L.S., ed. Problems of Drug Depend- ence, 1982 Rockville, Md.: NIDA, 1982. Macdonald, D.I. Drugs, Drinking and Adolescents. Chicago: Year Book Medi- cal Publishers, Inc., 1984. Mirin, S.M., ed.Substance Abuse and Psychopathology. Washington, D.C.: American Psychiatric Press, Inc., 1984. Mirin, S.M.; Weiss, R.D.; and Michael, J. Psychopathology in substance abusers: Diagnosis and treatment, American Journal of Drug and Alcohol Abuse 14(2):139-157, 1988. Moskowitz, J.M., and Jones, R.Alcohol and drug problems in the schools: Results of a national survey of school administrators. Journal of Studieson Alcohol 49:299-305, 1988. Pardee, H.; Silverman, M.M.; and West, A. Prevention and the field ofmental health: A psychiatric perspective.In: Breslow, L.; Fielding, J.E.; and Love, L.B., eds. Annual Review of Public Health 10:403-422, 1989. Parker, DA; Parker, E.S.; Harford, T.C.; and Farmer, G.C.Alcohol use and depression symptoms among employed men andwomen. American Journal of Public Health 77(6):704-707, 1987. Petrakis, P.L., ed. "Acquired Immune Deficiency Syndrome and Chemical Dependency." Report of symposium sponsored by the American Medical Society on Alcoholism and Other Drug Dependencies, Inc. (AMSAODD), and the Na- tional Council on Alcoholism (NCA). DHHS Publication No. (ADM)88-1513. Washington, D.C.; U.S. Government Printing Office, 1988. Powers, R.J., and Kutash, I.L. Stress and alcohol. InternationalJournal of Addictions 20(3):461-482, 1985.

Regier, D.A.; Boyd, J.H.; Burke, J.D.; Rae, D.S.; et al. One-monthprevalence of mental disordeu in the United States.Archives of General Psychiatry 45:977-986, 1988. fl -) COMMENTARY 21

Robins, LN.; Helzer, J.E.; Weissman, M.M.; Orvaschel, H.; et al.Lifetime prevalence of specific psychiatric disorders in three sites. Archives ofGeneral Psychiatry 41:949-958, 1984. Robins, L N., and Przybeck, T.R Age of onset of drug use as a factor in drug and other disorders. In: Jones, C.L., and Battjes, R.J., eds. Etiology of Drug Abuse: Implications for Prevention. NIDA Research Monograph 56, Publication No. DHHS (ADM) 85-1335. Washington, D.C.:U.S. Government Printing Office, 1985. Ross, H.E.; Glaser, F.B.; and Gertnanson, T. The prevalence of psychiatric disorders in patients with alcohol and other drug problems. Archives of General Psychiatry 45:1023-1031, 1988. Silverman, M.M. Children of psychiatrically ill parents: A preventive perspective. Hospital and Community Psychiatry, in press. Silverman, M.M., and Koretz, D.S. Preventing mental health problems. In: Stein, R.E.K., ed.Caring for Children with Chronic Illness:Issues and Strategies. New York: Springer Publishing Company, 1989. U.S. Department of Health and Human Services/Public Health Service. The 1990 Health Objectives for the Nation: A Midcourse Review. Washington, D.C.: Office of Disease Prevention and Health Promotion, Public Health Service, 1986. Wells, KB.; Golding, J.M.; and Burnam, M.A Psychiatric disorder in a sample of the general population with and without chronic medical conditions. American Journal of Psychiatry 145(8):976-981, 1988. Yamaguchi, K., and Kandel, D.: Patterns of drug use from adolescence to young adulthood: U. Sequences of progression. American Journal of Public Health 74(7):668-672, 1984. CHAPTER 1

Conceptual Issues in Prevention

Arnold J. Sameroff, Ph.D. Director of Research Bradley Hospital and Professor of Psychiatry and Human Behavior Brown University Barbara H Fiese, Ph.D. Assistant Professor of Psychology Syracuse University The prevention of children's psychosocial disorders has not been an easily accomplished task. In a critical appraisal of such efforts, Rutter (1979) was led to conclude that our knowledge of the topic is limited and that there are few interventions of proven value. The two greatest myths reviewed by Rutter were the beliefs that there are single causes for psychiatric disorders and that these causes can be eliminated by treating the child. Whateversubstance can be found in this area of research points to multiple causation as the rule rather than the exception and the need for intervening in the child-rearing context. This chapter begins with an overview of traditional concepts of prevention. When these ideas are used to interpret causal factors in disease, a variety of paradoxes emerge that require for their understanding a contextual systems analysis of developmental processes. A transactional model is described that takes into account the mutual effects of context on child and child on context, in explaining behavioral outcomes. The transactional model is embedded in a regulatory system that is characteristic of all developmental processes. Based on the regulatory system, a number of prevention strategies aredescribed that are theoretically driven and enhance the possibility of providingoptimal outcomes for children. Although the primary concern within child psychiatry is for disorders in children that require treatmentfor example, attempted suicide, substance abuse, and conduct disordersa larger set of children's behaviors have not yet pi oducx1 a diagnosable condition but will lead to such problems inadulthood. Ile broader concerns of preventive efforts in child psychiatry must extend not oni; to those children who will coi,..e to psychiatric attention before adulthood, but &so to those who will arrive during adulthood. The models presented here have as their premise that there may be no difference in kind and perhaps in

This paper was prepared for the Project Prevention initiative of the American Academy of Child and Adolescent Psychiatry. 24 PREVENTION OF MENTAL DISORDERS timing between preventive efforts to eliminate early disordersand those to eliminate later ones. The principles of developmental psychopathology(Sroufe and Rutter 1984), upon which much of what follows is based, applyto clinical problems throughout a person's lifespan. Defining Prevention For the last 30 years there has beena division of prevention efforts into primary and secondary categories (Commissionon Chronic Illness 1957). Primary prevention is practiced before the biological originof the disease. Secondary prevention is practiced after the diseasecan be identified but before it has ceused suffering and disability. More recentlytertiary prevention has been added to the list (Leavell and Clark 1965). Tertiary preventionis practiced after suffering or disability has been experienced, and its goalis to prevent further deterioration. Though secondary and tertiary preventionmay be quite important, they do not have the glamour associated with primary prevention (Lamband Zusman 1979). However, glamour may not be the appropriatecriterion for evaluating the effectiveness of prevention. Gerdon (1983) has arguedthat the tripartite classification of prevention efforts isan artifact of the mechanistic conceptions of health and disease that characterized earlyeras when biomedical research was almost exclusively a laboratory activity. The growth of epidemiological research has introduced more complex causal models thatmay restructure approaches to prevention. Gordon maintained that the primary-secondarydistinction does not separate preventive strategies that have different epidemiologicaljustifications and that require different utilization strategies. A further problemis that, especially for the nonprofessional community, the terms implya preferred priority when only a qualitative distinction is intended.Cost-benefit analyses of many prevention efforts have found that the "secondary" treatment ofa high-risk group may be far more efficient than a "primary" universal treatment. (See the chapter byLorion, Price, and Eaton in this vulume fora more extended discussion of these points.) When one turns to the prevention of psychological disease,the complexity of the problem is further increased. Whereas clearlinkages have been found between some "germs" and spec 'cm biological disorders,this has not been true for behavioral disorders. Primary prevention of psychologicaldisorders in the sense of deterring a biological factor may have meaning ina very small percentage of eases, although these casesmay be the most severe and profound. On the other hand, behavioral disturbances in thevast majority of cases are the result of factors more strongly associatedwith the psychological and social environment than with any intrinsic characteristicsof the affected individuals. Primary prevention might be effective if clearcauses of developmental disorders could be identified; but if empirical evidencefor these connections cannot be found or if it can be demonstrated theoretically thatclear causes do notor cannot exist, the choice of prevention strategies must bereassessed.

rl I CONCEPTUAL ISSUES 25

Paradoxes of Prevention How is it possible for clear causes not to exist for developmentaldisorders? The answer lies in new approaches for understanding biologicaland behavioral development wherein causal analyses are based on a probabilisticinteraction of multiple factors. These approaches were required to explainparadoxes that resulted when linear causal models were the only explanations used for avariety of disorders.Examplas involve illnesses that were thought to have clear presumptive causes accepted generally for long periods of medicalhistory. Changing theories of the etiology of tuberculosis is a good case in point. Tuberculosis initially was diagnosed as an environmental disorder. It was caused by conditions that accompanied poverty, that is, poor nutrition, poorair, and poor health. To prevent tuberculosis, the patient was told to avoidsuch conditions of poverty and dwell in circumstances of good air, good nutrition,and a life devoid of stms. There wereclear epidemiological data to support the connection between the degree of disorder and the degree of poverty inthe population. This analysis, however, was proved to be an artifact whenthe tubercle bacillus was discovered. A causal mechanism was found wherebythe disorder could clearly be attributed to the action of a specific entity. Curewould result from an elimination of the entity, a triumph of the disease model ofillness. However, this disease model, while effective in curing the disease, has not been equally effective in preventing the disease. Tuberculosis is not causedby the tubercle bacillus alone. Only 5 4o 15 percent of individuals with a positive tuberculin test (i.e., those infected with the bacillus) ever become ill with tuberculosis (Edwards 1975). The bacillus is a necessary condition for this disorder but not a sufficient condition; something else is also necessary. The additional factor is lowered resistance resulting from poor nutrition, poor health, or poor air, all correlates of poverty. The germ alone cannot cause the disorder; poor resistance alone cannot cause the disorder, but a combination of the two can. The disease entity must be viewed in a context. If the context were different, the outcome would be different. Prevention of tuberculosis can be accomplished with no attention to the specific germ associated with the illness.Elimination of the disease can be achieved by eliminating the necessary context for the bacillusby eliminating factors that lower resistance. Eisenberg (1982) pointed out that Cie strongest evidence for the role of social factors in modifying disease is the marked decline in the mortality rate from tuberculceis in the United Kingdom during the 19th century before the causative organism was discovered. The rate fell from an estimated 400 per 100,000 in 1840 to 200 per 100,000 in 1880 in the absence of effective medical remedies. This reduction in susceptibility to infection has been attributed to improvements in nutrition, sanitary conditions, and living conditions, that is, a change in context (McKeown 1976). 26 PREVENTION OF MENTAL DISORDERS

For biological disturbances suchas tuberculosis, at least two factors and probably more are involved in producing clinicalsymptoms.Because the combination of these factors varies in each individual,the probability that a disorder will result varies. Whatappear to be symptoms with ciear under closer scrutiny are seen to be the result of"probabilistic interactions of multiple factors" (Go llin 1981). Causal analysesapplied to the understanding of behavioral disorders, and especially theirprevention, need to identify these multiple factors and the associated probabilitiesof their interactions. The goal of preventionprograms for mental health is to rzduce the incidence of behavioral problems in children by firstidentifying high-risk conditions that produce greater than average proportions ofchildren with learning problemsor emotional disturbances and then interveningto reduce the number of risk conditions or their effects. Before interventionsare instituted, there needs to be good evidence for the connection betweenrisk conditions and deviant outcome. Frequently, such evidence has not been available,seriously undermining the rationale for well-intentioned intervention efforts. Early intervention programswere based on stable models of development in which children who were assessedas doing poorly early in life were expected to continue to do poorly. The early childhoodeducation movement, as exemplified in the Head Start program,was designed to improve the learning and social competence of children during the preschoolyears with the expectation that these improvements would be maintainedinto later life.Unfortunately, followup research of such children has found onlymoderate gains in measurable intellectual competence being maintainedinto adolescence (Zigler and Trickett 1978) although there were reducedrates of school failure and need for special education (Lazar and Darlington 1982). From a different perspective, children whowere identified early in life as being at risk from biological circumstancessuch as birth complicationswere thought to have generally negativedevelopmental outcomes. On the contrary, longitudinal research in thisarea has demonstrated that the majority of children suffering from such biologicalconditions did not have intellectualor social problems later in life (Sameroff andChandler 1975). In both domains early characteristicsof the child have been overpowered by factors in the environmental context of development.Where family and cultural variables have fostered development, childrenwith severe perinatal complications have been indistinguishable fromchildren without complications. Wher these variables have hindered development,children from the best preschool intervention programs have developedsevere social and cognitive deficits later in life.

Two points emerge from this analysisthat have major implicationsfor prevention programs. The firsis that the child's level of competencyat any point in early development, wnetherreached through normal developmental processes or some special intervention efforts, is not linearlyrelated to, that is, predictive of, the child's competence laterin life. The second point is thatto CONCEPTUAL ISSUES 27 complete the predictive equation one needs to add the effects of contextthe child's social and family environmentthat foster or impede the continuing positive developmental course of the child.In short, prevention programs cannot be successful if changes are made only in the individual child. There have to be corollary changes in the environment that will enhance the existing competencies of the child and buffer the child from stressful life evsnts in the future. Representative Risk Factors Let us turn for a moment to reeearch aimed at identifying representative risk factors in the development of cognitive and social-emotional competent*. Such competencies of young children have been found to be strongly related to family mental health and social status (Broman et al. 1975; Golden and Birns 1976; Werner and Smith 1982). Efforts to prevent developmental devianciee must be based on an analysis of how families in different social classes differ on the characteristics that foster or impede psychological development in their children. These factors range from proximai variables like the mother's interaction with the child, to intermediate variables like the mother's mental health, to distal variables like the financial resources of the family. While causal models have been sought in which singular variables uniquely determine aspects of child behavior, a series of studies in a variety of domains have found that, except at the extremes of biological leviation, it is the number rather than the nature of risk factors that is the best determinant of outcome (Greenspan 1980; Parmelee and liaise,- 1973; Rutter 1979). In a study of several hundred 4-year-old children, Sameroff et al. (1987) assessed a set of 10 environmental variables that are correlates of socioeconomic status (SES), but not equivalents of SES. They tested whether poor development was a function of low SES or the compounding of environmental risk factors found in low-SES groups. The 10 environmental risk variables were chronk:ity of maternal mental illness; maternal anxiety; a parental perspectives score derived from a combination of measures that reflected rigidity or flexibility in the attitudes, beliefs, and values that mothers had about their children's development; spontaneous positive maternal interactions with their children during infancy; occupation of head of household; maternal education; disadvantaged minority status; family support; stressful life events; and family size. When these risk factors were related tcs social-emotional and cognitive competence scores, major differences were found between those children with low multiple-risk scores and those with high scores.In terms of intelligence, children with no environmental riska scored more than 30 points higher than children with eight or nine risk factors. Similarly, the range in scores on an assessment of the social and emotional competencies of the children showed a similar spread over two standard deviations. 28 PREVENTION OF MENTAL DISORDERS

Three conclusions from this study are relevant to prevention efforts. The first conclusion is that the social and family factors were explaining most of the variance in outcomes, whereas factors related to the child's behavior during the first year of life explained almost none. The second conclusion is that the number of risk factors was the prime determinant of outcome within each socioeconomic level, not the socioeconomic level itself.The third and most important conclusion for prevention strategies is that thesame outcomes resulted from different combinations of risk factors. No single factorwas regularly related to either poor or good outcomes. If this is thecase, it is unlikely that universal preventions can be found for the problems of children. The contrast is that unique analyses of risk factme will require unique sets of intervention strategies embedded in a developmental model of psychopathology. Transactional Model

A similar developmental model appears to apply ina number of scientific domains (Satneroff 198S1. In this model outcomesare a function of neither the individual t,aken alone nor the experiential context taken alone. Outcomesare a product of the combination of an individual and his or her experience. To predict outcome, a singular focus on the characteristics of the individual, in this case the child, frequently will be misleading. What needs to be added is an analysis and assessment of the experiences available to the child. A model of development that included both the child and the child'sexperi. ences was suggested by Sameroff and Chandler (1975; Sameroff 1975). In this "transactional model" the development of the childwas seen as a product of the continuous dynamie interactions of the child and the experience provided by the family and social context. What was innovative in the transactionalmodel was the equal emphasis placed on the effect of the childon the environment, so that the experienoas pre vided by the environmentwere not independent of the child. The child's previous behaviormay have been a strong determinant of current experiences. A diagram of such a modelcan be seen in figure 1.

Constitution

Environment Ei

Time 1161111=111RUNI 11!=11111,

Figure L Transactional model of child development. CONCEPTUAL ISSUES 29

The child's outcome at some point in time (Cs) is neither a function ofthe initial state of the child (Ci) nor the initial state of the environment (El), but a complex interplay of child and environment over time.Figure 2 shows an example of such a transactional outcome. A complicated childbirth may have made an otherwise calm mother somewhat anxious. The mother's anxiety during the first months of the child's life may have caused her to be uncertain and inappropriate in her interactions with the child.In response to such inconbistency, the infant may have developed some irregularities in feeding and sleeping patterns that give the appearance of a difficult temperament. This difficult temperament decreases the pleasure that the mother obtains from the child and so she tends to spend less time with the child.If adults are not interacting with the child, especially not speaking to the child, the child may not meet the norms for language development and score poorly on preschool language tests.

Mother

Child Language Delay

ti t t t 4 Time 2 3 5

Figure 2. Transactional outcomes in child development. What determined the poor outcome in this example? Was the poor linguistic performance caused by the complicated childbirth, the mother's anxiety, the child's difficult temperament, or the mother's avoidance of verbal interaction? If one were to design a prevention program for this family, where would it be directed and would it be defined as primary or secondary prevention? If one were to pick the most proximal cause, it would be the mother'savdidance of the child, yet one can see that such a view would be a gross oversimplification of a complex developmentd sequence. Would primary prevention be directed at eliminating the thi!d's difficult temperament or at changing the mother's reaction or at providing alternative sources of verbal stimulation for the child? Each of these would eliminate a deviation at some contemporary point in the developmental system; but would any of these efforts ensure the verbal competence of the child or, perhaps more important, ensure the continued progress of the child after the preventive effort was completed?

.1 I. 80 PREVENTION OF MENTAL DISORDERS

Constellation of Behaviors A number of empirically validated examples of transactionalprocesses are in development (see Sameroff 1986, 1987) but few among these are as yet directly pertinent to clinical child psychiatry. One of the most compelling data sets emerges from the work of Patterson and his colleagues in a series of studies on the origins of antisocial behavior in childhood (Patterson 1986). In the Patter- son model, children normally engage in some proportion of noncompliance activitiee.If parents are inept in disciplining their children, they createa context in which the child is reinforced for learning a set of coercive behaviors. Parent ineptitude is characterized by lack of monitoring, harsh discipline, lack of positive reinforcement, and lack of involvement with the child. The child develops noncompliant behaviors characterized by whining, teasing, yelling, and disapproval. These behaviors escalate parental negative coerciveresponses that promote further child noncompliance, eventuating in high-amplitudeag- gressive behaviors, including physical attack. The highuse of noncompliance with inept parents does not permit the child to learna set of social strategies that will be necessary with peers and in school. When these aggressive,noncompliant children enter the school setting they elicitpoor peer acceptance that maintains poor self-esteem and poor academic performance. This constellation of antisocial behavior, poor peer relations, and poor school achievement has been demonstrated by Patterson to unfold in the developmentalsequence of negative transactions described previously (see figure 3). The child's initialnoncom- pliance does not lead directly to antisocial behavior; rather it is theinept parenting response that converts age-appropriate expressions of autonomy into a coercive interactive style.

Parents pBehercive aviors

Child Coercive rAntisocial Behaviors Behaviors

t1 t t t 4 Time 2 3 t 5

Figure 3. Developmental sequence of negative transactions.

2 CONCEPTUAL ISSUES 31

Although the parents in the Patterson model are more blameworthy than in the temperament example, they, too, are embedded in transactional contexts with their own parents. Other research (Elder et al. 1983; Huesmann et al. 1983) has demonstrated cross-generational effect5 associated with antisocial child behavior. In two longitudinal studies, the poor disciplinary practices of the grandparents were related to antisocial behavior of the parents and the grandchild. Moreover, the child, parents, and grandparents are embedded in a social context that supporta these child-rearing strategies as a means of successfully adapting to a particular level of socioeconomic existence (Kohn 1969, 1973). The many points where deviancy is fostered in transactional models are also points at which prevention can be attempted. Differentiating among primary, secondary, and tertiary prevention may be inappropriate in such developmental models. Primary prevention is defined as havins temporal priority over secondary or tertiary efforts. In the preceding examples, primary prevention would have been directed at preventing birth complications in the first case and preventing children from saying "no" in the second case. But the proximal, most directly connected causes of the problem in the child are the last in the chainthe parental avoidance and coercive child-rearing strategies. It is necessary to find another way of thinking about such developmental progressions because the distal eventsfor example, birth complications and noncom- pliancelead to a variety of child outcomes, both good and bad; and the deviant outcomesfor example, language delay, antisocial behavior, and poor school achievement--can be caused by a variety of proximal determinants, many of which are not connected to the child's initial state. Biologic Transactions The transactional model, despite its novel name, is in reality not a new idea. It is merely a new emphasis on some very old traditions in developmental theory, especially theories of the dialectic in history and philosophy. A more cogent referent is theory and research in biology, where transactions are a recognized essential part of any developmental process. In the study of embryological development, for example, there are continuous transactions between the phenotype and the genotype (Ebert and Sussex 1970; Waddington 1957). A simple view of the action of genes is that they produce the parts that make up the organism. A brown eye gene may be thought to produce a brown eye.In reality there is a much more complex process of mutual determinism. The material in the fertilized egg cell turns on or off specific genes in the chromosomes. The turned-on genes initiate changes in the biochemicals in the cell. These changed biochemicals then act back on the genetic material, turning on or off more genes in a continuous process, and usually producing a well-developed organism. In certain circumstances, the illusion of a linear relationship exists between a particular gene and a particular feature of the phenotype, as in the case of eye

4 4-1 32 PREVENTION OF MENTAL DISORDERS color.In reality, however, determinism is never linear, because of the complexity of biological processes. What then creates the illusion? The answer is in the regulatory system that buffers development, what the embryologist Waddington (1957) described as "canalization." In all the complex interactions between genotype and phenotype is a regulatory system that monitors the developmental changes to assure that they stay within defined bounds. This regulatory system and the bounds are the result of an evolutionaryprocess that occurred across myriad generations and that now assures a particular outcome. With eye color, the system is hidden because it isso tightly buffered (i.e., regulated), so that if one knows the stvuctural genes one can generally predict the outcome. However, there are some simple examples in which the regulatory system is quite evident. In the case of identical twins, a single fertilized cell splits in two. The genetic regulatory system ensures that the outcome is not two half-sized children. Compensations are made so that the resulting infants will both be of normal size. In the case of genetic dominance, the result fora homozygous individual is the same as fora heterozygous one even though there is a clear difference in the quantity of genetic material If thereare two brown eye genes, the eyes are no browner than if there were only one. These examples are clear evidence for regulatory processes at the biological level. The genetic system never operates alone: It is always in an environment that isa mekior codeterminant of gene activity. What follows is evidence that analogous regulatory systems that direct development toward a particular set of outcomescan also be found at the social level. Understanding of the genetic regulatory system has offered the hopeof preventing a variety of physical disorders.Similarly, our increased understanding of the family and cultural regulatory systems will offer hope in the psychological domain, and perhaps, as some have suggested, improve physical health as well (Rodin 1986).

Statutes, Stories, and Styles

Just as there is a biological organization, the genotype, that regulates the physical outcome of each individual, there isa social organization that regulates the way human beings fit into their society. This organization operatesthrough family and cultural socialization patterns and has been postulated tocompose an "environtype" (Sameroff 1985) analogous to the biological genotype.Th,.. importance of identifying the sources of regulation of human developmentis obvious if one is interested in manipulating that development,as in the case of prevention or intervention programs. The failures of such effortscan be understood only in terms of a failure to understand the regulatory system. Each individual's environtype contains these regulatory patterns. Theenvirontype is composed ofsubsystems that not only transact with the child but alsotransact with each other. CONCEPTUAL ISSUES 33

Bronfenbrenner (1977) has provided themost detailed descriptions of environmental organizations that impacton developmental processeel within categories of microsystems, mesosystems,exosystems, and macrosystems. The microostem is the immediate setting ofa child in an environment with par- features, activities, and roles (e.g., the homeor the school). The mesosys- tem comprises the relationships between the =Orsettings at a particular point ticular.in an individua"s lievelopment. The exosystem isan extension of the mesosys- tem that includes settings thai the childmay not be part of btit that affect the settings in which the child does participate (e.g.,the world of work and neighborhoods).Finally, the macrosystem includes the overarchinginstitutional patterns of the culture including the economic, social,and political systems of which the microsyntems, mesosystems,and exosyetems are concreteexpres- sions.Bronfenbrenner's ecological model has been fruitfullyapplied in the analysis of a number of clinical issues includingthe effects of child abuse (Belsky 1980) and divorce (Kurked 1981). The present discussion is restricted to levelsof environmental factors in the culture and the family. Developmental regulationsat each of these levels are carried in codes, the cultural code and thefamily code. These regulationsare encoded to direct cognitive and social-emotionaldevelopment so that the child ultimately will be able to filla role defined by society. Although the environtype can be conceptualizedindependent of the child, changes in the abilities of the developing childare major trigg3rs for regulatory changes and most likelywere mWor contributors to the evolution of a developmental agenda that is each culture's timetablefor developmental milestones. The cultural code is influenced bya variety of characteristics of society, including the customs, mores, belief patterns, andlegal system. We have given these cultural factors that directly impacton child development the generic label of statutes. These have a more formal and enduring characterthan the stories that are the primary regulatory factors in the family code. Finally, theseregulations must be carried out through the interaction of actualpeople, who modify the expression of the cultural and family codes by theirindividual styles, that is, the characteristics of their personality andtemperament. Most behavioral research on the effects of theenvironment have focueedon analyses of dyadic interaction patterns in whichlabels are placed on the participating individuals. Only recently havethese relationships themselves become empirical issues of inquiry. Parkeand Tinsley (1987), inan extensive review of family interaction research, have pointedto the important new trend of not only adding father-child interactionto the study of mother-child interaction, but also combining them into stu diesof triadic interactions and entire family behavioral patterns. The behavioral researchis slowly overcoming the technological difficulties embodied iv imalysesof multiple interacting individuals. Another growing empirical basecomes from the direction of beliefs rather than behavior (Sigel 1985).Investigators have become increasingly articulate at defining the dimensions of parentalbelief systems with the ultimate goal of describing the effecp of thesebelief systems on parent behavior 4 34 PREVENTION OF MENTAL DISORDERS and, ultimately, on child behavior. For the present, however, these research domains have provided primarily promissory notes of important future contributions to successful prevention efforts. Cultural Code The ingredients of the cultural code are the complex of characteristics that organize a society's ch ild-rearing system, incorpozating elements of socialization and education. It is beyond the scope of this chapter to elucidate the full range of cultural regulatory processes that are potentially relevant to prevention efforts. As a consequence only a few points are highlighted to flesh out the dimensions of the cultural code. Although the common biological characteristics of humans have produced similar developmental agendas in most cultures, there are differences in many major features that often ignore the biological status of the individual. In most cultures, formal education begins between the ages of 6 and 8 (Rogoff 1981), when most children have reached the cognitive ability to learn from such structured experiences. On the other hand, informal educationcan begin at many different ages, depending on the culture's attributions to the child. The Digo and Kikuyu are two East African cultures that 'wive different beliefs about infant capacities (deVries and Sameroff 1984). The Digo believe that infants can learn within a few months after birth and begin socialization at that time. The Kikuyu wait until the second year of life before they believe serious education is possible. Closer to home, some segments of middle-class parents have been convinced that prenatal experiences will enhance the cognitive development of their children. Such examples demonstrate the variability of human developmental contexts and the openness of the regulatory system to modification either by charlatans or mental health professionals. One of the mikjor contemporary risk conditions toward whichmany programs are being directed is the elimination of adolescent pregnancy. Alt Lough for certain young mothers the pregnancy is the outcome of individual factors, fora large proportion it is the result of a cultural code that defines maturity, family relationships, and socialization patterns with adolescent motherhoodas a normative ingredient. In such instances, to focuson the problem as one that res:des wholly at the individual level would seriously undercut effectivepreven- tive efforts. A broad v prevention requires an appreciation of the cultural context of developmei owever, from the perspective of child psychiatry, there is little role for prevention in programs at the societal level.Psychiatrists can participate in public health, educational, or politicalprograms; but such efforts do not maximize the unique contributions of training in child psychiatry. This uniqueness of concern, especially with psychodynamic issues, finds its most useful applicationathe level of the family regulatory system. CONCEPTUAL ISSUES 35

Family Code Just as cultural codes regulate the fit betweenindividuals and the social system, family codes regulate individualsin the family system. Family codes provide a source of regulation that allows a groupof individuals to form a coliective unit in relation to society as awhole. As cultural codes regulate development so that an individual may fill a role insociety, family codes regulate development to produce members that fulfill a rolein the family and ultimately are able to introduce newmembers into the shared system.Traditionally, new members are incorporated through birth andmarriage, although more recently remarriage has taken on a more frequent role inproviding new family members. An understanding of the family code andits regulatory principles provides a x'rairswork for identifying additional nodalpoints for prevention and intervention efforts. A more precise understanding ofthese principles may give way to more efficient means ofprevention. As each culture regulates family and childbehavior through various forms of statutes, the family regulates the child'sdevelopment through a variety of forms that can be organized into a generic categoryof stories. This category includes, in addition to actual stories, rituals,myths, and family paradigms (Reiss 1989). An unresolved issue that will haveimportance for intervention efforts is how these forms are transmitted within thefamily and especially their level of representation. We shall put this issue aside atthis point and restrict our discussion to the description of these familyregulatory forms. Rituals Rituals are the most clearly self-aware of thefamily regulatory farms. Rituals are practiced by the whole family and arefrequently documented. They may be times for takingphotographs, exchanging gifts, or preserving mementos. Ritual activities are by definition set off fromthe normal routine. The content of family rituals includes symbolic information aswell as important preparatory phases, schedules, and plans. These ritualshighlight role definition in the family. For example, at Thanksgiving thefather is seen as the head of the household who sits at the head of the table and carvesthe turkey. Rituals serve a regulatory function byassigning clear roles or tasks to each member of the family. To participate in the ritiml, each familymember must conform to the specific characteristics of the role. There is a developmental progression as children areable to enter more fully into family rituals. To participate inrituals, children must be able to understand role assignments and alter theirbehavior patterns to fit a new routine. Hudson and Nelson (1983) have demonstrated thatpreschool and first-grade children recall stories about birthday parties beforerecalling stories about routine events such as baking cookies. There is ahigh degree of saliency to ritu als that facilitate participation by childrenand encoding of family structure. 36 PREVENTION OF MENTAL DISORDERS

Stories Stories provide a second form for family regulations. Reiss (1989) madea distinction between the act of storytelling asa source of regulation and the content of stories as a source of information. The content of stories serves a communicative function at all social levels, from the cultural, tc the family, to dyads in the familyBefore written records, stories provided cultures with a means for passing down customs and taboos, thus regulating family members' behavior within the cultural code. Family stories also provide guidelines for individual conduct inside and outside the family. Family stories are defined by the4 shared or communal act of historical reconstruction (Reiss 1989). They are frequently transgenerational. The content of family stories includes descriptions of significant family members and events and highlights family customs and values. Family stories also include detailed information about role regulation, providing parents with modelsor guidelines for their own behavior as well as the behavior of their children. For example, a matriarchal family relegates the disciplinarian role to the mother. Family stories may include how the grandmother disciplined the mother. This story then serves as an example or justification to the next generation. The transgenerational component gives credence to the mother's roleas disciplinarian. As a regulatory source, family stories are conservatory. They havea stabiliz- ing effect by preserving important events and passing ona value system to the next generation (Reiss 1989). Family stories are fully self-aware and can be told by several members of the family. There is a strong developmental component to family stories. Asa source of regulation, the telling of stories is a major feature of early relationships between infant and family members. Ratner and Brunher (1977) have proposed that this early storytelling provides a framework for the learning of conversational turn taking and facilitates language development. During these earlyyears the child engages in storytelling by being a story-listener and will often encourage others to tell or read a story. It is interesting to note that this activity can be shared by all members of the family and across generations. Children are increasingly able to recall specific aspects of stories. Nelson and her colleagues have demonstrated that there isa developmental progression in the recall of stories (Nelson 1981; Nelson and Gruendel 1981).Preschool children readily talk about their experiential knowledge in scriptlike form, and these scripts affect the way in which children interpret and remember stories and everyday events (Nelson 1981; Nelson and Gruendel 1981). Childrenare receptive to hearing stories and organizing experiences along story lines, which provides parents with the opportunity topass down values through their storytelling. CONCEPTUAL ISSUES 37

Myths A third source of regulation in the family code arefamily myths. Family myths are beliefs that go unchallenged in spite ofreality (Lewis and Beavers 1976).Myths may have a traumatic origin and frequentlyhave a strong affective component (Kramer 1985). Family myths are not openfor discussion, nor are they readilyrecognized as distortions (Ferreira 1963). Some family myths help to regulate role definitions.For example, a traditional family may consider females as unable to handleprofessional responsibilities a the work world despite the fact that they areable to balance the family checkbook and organize a busy household.Family myths serve a regulatory function through processes like role inflation.Subtle aspects of a particular role may become inflated and incorporated intothe myth.For example, parents of a physically handicapped child maybelieve that the child is also cognitively handicapped despite examples ofthe child's intelligent behavior. A myth develops that casts the child in ahandicapped role that encompasses behaviors beyondphysical limitations.In the same context, another family may create a myth that their mentally retardedchild is unim- paired because of a bright-eyed appearance. Developmental problems can arise when the child must accept adistorted family myth to remain in the family or when the family imposes aninflated role on the child by creating a newmyth. In the first instance, sexually abused children or witnesses to parental abuse may constructcomplicated stories to deny the wrongdoing of the family member (Strauss et al.1980). In the second instance, a handicapped child may be treated as the youngestsibling despite birth order or chronological age (Sigel 1985). Paradigms Family paradigms are a fourth form of family regulations. Reissand his colleagues (Reiss 1981; Reiss et al. 1981) have described how familiesdevelop paradigms that include a set of core assumptions, convictions, orbeliefs that each family holds about its environment. Reiss et al. (1981) arguethat these paradigms generally persist for years and even generations and aremanifested "in the fleeting fantasies and expectations by all members of thefamily and, even more important, in theroutine action patterns of daily life." Basing their research on empirically derived dimenzions of configuration,coordination, and closure, these investigators have identified a four-categorytypology of paradigms, including environment-sensitive, consensus-aensitive,achievement-sensitive, and distance-sensitive families. Paradigms appear to be the form of family regulaticthat is the least articulated in awareness, although they can be expressed in family storiesand myths. The importance of family paradigms for prevention efforts isthat, although they can be identified only in the course of fam...ly problem-solving tasks, they are manifested in the relationships that family members,including 38 PREVENTION OF MENTAL DISORDERS children, form with other individuals and groups. Thus the normalor disturbed behavior of children must to some degree be interpreted asan outgrowth of the family paradigm (Reiss et al. 1981). Individual Code There is good evidence that individual behavior is influenced by the family context. When individuals operate as part of a family, the behavior of each individual is altered (Parke and Tinsley 1987), frequently withoutawareness of the behavioral change (Reiss 1981).However, there is also no doubt that individuals bring their own contributions to family interactions. The contribution of parents is much more complexly determined than that ofyoung children, given the multiple levels that contribute to their behavior. We have discussed the socializing regulations embodied in the cultural and family codes. We have not discussed the individualized interpretations that each parenting figure imposes on these codes. To a large extent these interpretationsare conditioned by both parents' past participation in their own family's coded interactions, but they are captured uniquely by each member of the family. These individual influences, which we have labeled style, further condition each parent'sresponses to the child. The richness of both health and pathoiogy embodied in these responses is well described in the clinical literature. In terms of early development, Freiberg and her colleagues (Freiberg et al. 1980) have providedmany descriptions of the attributions that parents bring to their parenting. These "ghosts" of unresolved childhood conflicts have been shown to "do their mischief according to a historical or topical agenda, specializing in suchareas as feeding, sleep, toilet-training or discipline, dependingupon the vulnerabilities of the parental past." The effect of parental pathology has long been recognizedas a contributor to the poor developmental status of children.While we acknowledge that influence, we must also be careful to add the contexts in which parental behavior is rooted, the family and cultural codes. To ignore these contexts would permit only limited additional success for prevention efforts that foundered whenthe child was the sole target of treatment. It is important to recognize theparent as a major regulating agent, but it is equally important to recognize that parental behavior is itself embedded in regulatory contexts. Regulations The description of the contexts of development isa necessary prologue to understanding psychiatric problems and the eventual design ofprevention programs. Once an overview of the complexity of systems is obtained,we can turn to the search for nodal points at which intervention strategiescan be directed,These points will be found in the interfacesamong the child, the family, and the cultural systems, especially where regulationsoccur. To complete the picture, we must elaborate on the complexity of regulatoryprocesses reflected in their timespan and in their level of representation.

ar ) CONCEPTUAL ISSUES 39

The cultural and family codes canbe broken down into sets of regulatory functions that operate across differentmagnitudes of time and will require different intervention strategies.The longest cycle is associatedwith the macroregulations that are part of aculture's "developmental agenda."The developmental agenda is a series of pointsin time when the environmentis restructured to provide different experiencesto the child. Age of weaning, toilet-training, schooling, initiation rites,and marriage are coded differentlyin each culture, but they provide thebasis for socialization in eachculture. The validity of such agendas is not in theirdetails, but in the fact that the culture is successfully reproduced in generationafter generation of offspring. Macro- regulations are known to socializedmembers of each culture. On a shorter time base areminiregulations that include the caregiving activities of the child's family. Suchactivities are feeding children when they awaken, changing diapers when they are wet,and keeping children warm. Such regulations are exemplified in the caregivingpractices of coeTtive parents in Patterson's (1986) work. Miniregulationsalso are known to m mbers of society and can be transmitted from member tomember. On the shortest time base aremicroregulations that refer to the momentary interactions between child and caregiver;others have referred to these as "behavioral synchrony" or "attunement"(Field 1979; Stern 1977). Microregula- tions are a blend of social and biologicalcodes because, although they may be brought to awareness, many of these activities appearautomatic. Toward the biological end are the caregiver's smile in responseto an infant's smile, and toward the socialized end are"microsocial" patterns of interaction that increase or decrease antisocialbehavior in the child (Patterson 1986). The three sources of regulation that havebeen outlined operate predominantly at different levels of the developmentalsystem Macroregulations are the modal form of regulations operating inthe cultural code. Cultural codes are written down add may be passed on toindividual members of society through the generic category of statutes inwhich we have included customs, norms, mores, and mythologies,in addition to actual laws that are aimed atregulating child health and education, The familydevelops their caretaking routines influenced by the transactions between thecultural and family codes, that is, between statutes and stories. As childrendevelop, they increasingly participate in these transactions that serve as afoundation fat social interaction. Families highlight the role defined for each child throughrituals and develop myths that further regulate the child's development.The style of the family members contributes to the way in which the regulationswill be carried out in relation to the individuality of each child. The operation of the family code ischaracterized by a series of' transactions. The parents may hold particular conceptsof development that influence their caretaking practices, As children are exposed todifferent role expectations and listen to the family stories, they make their owncontribution by their particular style, The child's acting-out of roles in thefamily is incorporated into family Ii 40 PREVENTION OF MENTAL DISORDERS

stories, rituals, and myths. By becomingan active transactor in the family code, the child's behavior may ultimately affectthe child-rearing practices of the parents and the creation of the code to be passed downto the next generation. The family code providesa bridge between the cultural code and the social development of the individual. Families developstrategies to negotiate culturally determined regulations. These strategiesmay be articulated by a family stud frequently include a planningcomponent. Areas in which families must develop strategies to deal with cultural codesinclude entering and leaving school, joining and leaving the work force,leaving and returning home, and marrying and setting up household (Hareven1984). The family codemay rely on the cultural call in initiating the timing of the strategies,but the family code will regulate the amount of disruptionexperienced by the family members. For example, many school systems holdopen houses for parents to enroll their children in school. Once the childrenare ennlled, they become memben of that institution, with little additional effort expendedby most families. The family code incorporates the developmentalagenda of the culture, and the child is able to assume a new role outside the familyinthis case, a student role. However, there are times when the familycode employs unsuccessful strategies in negotiating culturalregulations.Children who develop school phobias are not fulfilling the culturalstatute for school attendance. Inmany cases the child's failure to attend school is in parta failure of the family code to regulate the child's transition into school. Acommon scenario in such families begins with a child being fearful of leavinghome, is followed bya mother's failure to set limits for the child, and ends withthe child remaining home. The child may present with a host of physical symptoms thatencourage more attention from the mother and result in the child'sbeing able to maintaina close relationship to her (Weiner 1970). Atransaction develops in which the mother indulges the "sick" child, followed byunsuccessful attempts of the childto enter school. A strategy for dealing with the culturalstatute of school attendance does not develop, and the family miniregulationscome into conflict with the cultural macroregulat ions.

Jay Haley (1980) has eloqvently describedhow the family codemay unsuc- cessfully incorporate the cultural codeat the end of adolescence. Many families of disturbed young adults donot develop successful strategiesto encourage independence. When theyoung adult makes attempts to leave home, thefamily becomes disorganized to the point ofpreventing the young adult's successful transition to autonomy. The family developsstrategies to preserve the child's dependence; and these strategies conflictwith the cultural expectationsfor the young adult's indepenchnce.

Family codes are primarily expressedthrmigh miniregulsaions embodiedin caisgiving procedures. However, theseregulations are modified by feedback from the effect of microregulationson the child. Microregulations operateon the level of the individual and neednot be part of consciousawareness. A colicky infant, for example,may be difficult to comfort. This difficulttemperament

1 4, CONCEPTUAL ISSUES 41 influences the miniregulations of tbe family in caretalang activities such as regulating sleep time.Changes in microregulations may be evident as the mother receives less satisfaction from the child and spends less time in direct eye contact. The infant, in turn, may develop acharacteristic style of high activity level to gain the mother's attention. These changes in style may develop as part of a transaction that is regulated out of awarenessof either individual. Targeting Prevention Efforts

Given an understanding of the regulatory system for psychological development, what implications will it have for prevention? The primary application will be to analyze the etiology of deviant outcomes so that appropriate targets for intervention can be chosen. On the caregiver's side, one must analyze the factors that caused regulation to fail.These would include such factors as parents not knowing the cultural code or knowing the code but being unable to use it because of other demands for their time and resources.Such other demands may include the need to be away from the home to make a living, life-event stresses that interfere with their caregiving, or mental illness that diverts their attention from their children's needs to their own. The family code sets the stage for the interaction of members with the broader social world, as well as for fostering unique perspectives in developing children. Idiosyncratic family stories may limit or distort both the developmental aspirations of these children and their interactions with peers and other extended social groups. To the extent that a child may have a developmental problem unrelated to the family code, the code still may place limitations on the family's ability to recruit and use community resources for support and therapy. To the extent that the child's problems are an outgrowth of the family code, a more direct confrontation may be indicated as a prevention strategy. The parents may know the cultural code and have a family code well adapted to cultural statutes, but they may be confronted by a child who does not fit the code. A child with a handicap, or born prematurely, or with a difficult temperament would present such problems. In such a case, deviancy will be the outcome of a stress on the regulatory system. The prevention of deviancy will be a function of the identification of that stress, whether it comes from the child, the parents, or perhaps, the larger social context. The analysis of a variety of risk factors in the Sameroff et al. (1987) study showed that developmental outcomes for young children are multiply determined. No single factor wal always present or always absent in cases in which low levels of social-emotional and intellectual competency were found. In other words, no single factor could be identified as a cause.Given this array of variables, what actions can be taken to protect the child from their negative consequences? Certain of these variables are enduring characteristics of the family, for example, minority status and family size. Others are not in the usual domain of intervention, for example, stressful life events and marital status. .1,4cr. ) 42 PREVENTION OF MENTAL DISORDERS

Another set is highly unlikely to changeoccupation and educationallevel. What are left are the coping skills of the parents. These includethe psychological variables of mental health, parental perspectives, and parent-childinteraction patterns.These coping skills are aspects of whatwe have previously described as part of the cultural and family code. These codescompose the social regulatory system that guides children through their developmentand buffers them from those aspects of the broader environment withwhich they are not yet able to cope by themselves. An ideal developmental system isone in which the environtype and genotype are in harmony, where transactions occur in an orderly fashion between the typical growth milestones of the child and the familyand cultural codes. However, these codes can be organized toa greater or lesser degree, and they can be in conflict to a greater or lesser degree. To the extent that the cultural and family codes are unorganizedor deviant, then the possibilities of guiding the child through life are correspondingly impaired.Professional intervention is one aspect of the environtype thatwe are trying to enhance to support or replace other inadequate regulatory systems. Regulation Using S. transactional model,we now describe strategies that capitalize on the preceding analysis of regulatoryprocesses to provide appropriate targets for prevention efforts. The addition of the environtypeconcept to the transactional model requires altering figure 1. The initialpresentation of the transactional model emphasized the mutual regulations betweenchild and context. The addition of codes requireda refocusing on the continuities in the family, and also in the child. The enlarged regulation model is presentedin figure 4. A set of arrows leads from tLe child's initial state (CO to the child'sstate at succeeding points in time. This dimension refers to the continuity ofcompetency within the child. The line gets thicker as children grow older and learnmore skills for taking care of themselves and buffering themselves from stressfulexperiences. Another set of arrows leads from the parents' initialst&te WI) to the parents' state at succeeding points in time.This dimension refers to the continuity in the parents' understanding of the cultural codeand their competency at regulating their child's development. Thesets of vertical arrows refer to the actions of parents on children and, conversely, of childrenon parents. Prevention strategies must focuson the vertical arrows that mediate the regulatory functions. These strategies fallinto two categories affecting,respectively, the upward and downwardarrows. The upward arrows reflect the effects of the child on the parents. Theeeeffects can be changed by either changing the child or changing the parent's interpretationof the child. The downwardarrows reflect the effects of the parenton the child. Parent effects can be similarly CONCEPTUAL ISSUES 43 altered by either changingthe parent or changing thechild's interpretation of the parent. The followinganalysis will be based on thefirst three possibilities. redefinition, and reeduca- For simplicity, these havebeen labelled remediation, tion (Sameroff 1987).

Parents

Child

Time t t2 t3

Figure 4. Regulation transactionalmodel.

Remediation The strategy of remediation isthe prevention mechanismaimed at repairing or changing thechild, This strategy is based onthe idea that the psychological development of the child isdetermined hy the child's biological state.Thus by repairing biology, one cannormalize psychological functioning.Although there valid, for are manyphysical conditions for whichsuch an approach may be example, intestinal or cardiacanomalies, the vast mikjority ofbehaviorally disordered children are alb result oftransactional processes. In such cases,the intervention str IfA.,gy is to changethe child's effect on the parentalregulating system. Malnutrition in infancy may be agood example of how this strategyoperates. Although there was an earlyassumption that malnutrition ininfancy adversely affected later intelligence byreducing the number of braincells, longitudinal studies with appropriate control groupshave shown that lower later intelligence in malnourished infants is the consequenceof their poor environments, nottheir poor biology (Read1982).Cravioto and DeLicardie (1979)found that the

r..Li 44 PREVENTION OF MENTAL DISORDERS

behavioral effects of malnutritionwere most prevalent in families in which the mothers were passively traditional in theirchild care and provided little stimulation to their children.In a naturalistic study, Winick et al. (1975) compared two groups of Korean children whohad suffered severe malnutrition as infants. One group was raised by their parents inKorea and scored poorly on psychological tests given during adolescence. The secondgroup w,.s adopted by middle-class U.S. parents who hadno knowledge that the children had suffered from malnutritionas infants. The adopted group scored as well asor better than their U.S. contemporaries whentested. Poor later outcomes of malnourished infantswere thought to result from the infant's impaired attentionalprocesses, reduced social responsiveness, heightened irritability and inability totolerate frustration, low activity level, reduced independence, and diminishedaffect.In a study based on these hypotheses, Barrett et al. (1982) compareda group of ch ildren who had received caloric supplementation during infancywith a group that did not. The results were that better nutrition was associated with greater socialresponsiveness, more expression of affect, greater interest in the environment,and higher activity level at school age. Food supplementationof young infants interrupted the negative transaction in which theirlow energy levels failed to stimulate their parents to engage in adequate socialization.The failure to develop normal patterns of social interaction, especially withpeers, found for the malnourished control group was prevented for the supplementedgroup, Remediation as an intervention does nottamper with the cultural code, that is, the parental regulatory system. It changesthe child to better fit whatever the normative code is.For older children, both biochemical andbehavioral approaches to remediation have beenused. Hyperactive children who disturb their parents and teachersare frequently given medication to quiet them. The subdued child is better able to participatein normative interactions. Such children, as well as those with conductdisorders, have been given behavior therapies to modify their behaviorso that they will be less likely to elicit negative responses from others in their social context. Such positivechanges may be possible for some behavioral problems,but are less successful withmore biologically determined problems suchas handicaps or difficult temperaments. One cannot easily makea blind child see or a spina bifida child walk. One cannot easily stop some babies from havingirregular sleep and eating habitsor from having high levels of endogenously determinedcrying. In these cases a second strategythe strategy of redefinitionmustbe used to prevent later disabilities, Redefinition

Redefinition is required if the parents havedefined the child as deviant and are either unwilling or unable to engage in normal developmentalregulation, that is, caregiving. In thecase of children with handicapping conditions the source of the parents' reactions is fairlyeasy to identii, When the parents' CONCEPTUAL ISSUES 45 reactions are embedded in their family orindividual stories, however, the source may be more difficult todetermine. Parents of Sandicapped children may try toconvince society that the child cannot be maintained in the family and mustbe reared in a completely different setting (i.e., an institution), with anappropriate abnormal child-rearing program for their abnormalchild.In other cases, the parents may acceptthe responsibility for the physical care of theirchild but expect little in the way of a satisfactorypsychological relationship with the cl,;Id(Roskies 1972). The prevention effort with such familiesis directed at a redefinition of the situation, at identify.,4 for the parents thepossibilities for normal child-rearing within what appears to be a deviant situation.In the case of children with handieaps, the redefinition may involve arefocus on the possibility of normal corsitive and social-emotional development.In the case of a retarded child, such s child with Downsyndrome, the redefinition may involve afocus on the normal sequencing of development, albeit at aslower pace. The family may need to alter their role definitionsof a handicapped child to incorporate the child into their family stories.Modifications may be needed for the child to fully participate in familyrituals. For example, if a family ritual includes a high amount of physical activitysuch as a weekend football game, then the physically handicapped child's rolewill have to be redefined or the ritual will have to be altered. Theseredefinitions allow the family to admit the child to their ca agiving system. They allowthe parents to successfully experience raising their child within thecaregiving system they already know. They may need to learn some specialskills for feeding or positioning the child, but these are only var4ations of what theywould have done with a nonhand- icapped child. In the case of temperamental problems,the redefinitions may be simpler. When a colicky child who cries most ofthe time is perceived as emotionally deviant, the redefinition takes the forin ofindicating that colic ie only an extreme on a normal dimensionof individual differences (Thomas andChess 1980). Crying babies need not become mentallyill adults or even crying adults. Certainly, it is a greater strain on the regulatory systemto raise a handicapped or colicky infant, butthis does not mean that the regulatory systemis not adequate for this purpose. When the source of the parents' regulationdeviancy arises from within rather than without, the modification of theindicated family and individual stories fits well within the role of a child and adolescentpsychiatrist or other qualified mental health professional. In many such cases,the child can be an unwitting elicitor of caretaking deviancies by some physicalfeature, such as having the wrong gender, hair color, or eyecolor. Along tile lines of temperamental variation, the child may have behavioral characteristicsthat would not trigger deviant responses in most parents,but do in special cases in which they become enmeshed in family stories. But in most cases includedin this category, the attributions arise from within the parent independent of theinitial condition of the child. 46 PREVENTION OF MENTAL DISORDERS

General programs may be devised for generalredefinition problems (e.g., dealing with a handicappedor a difficult child); however, when the problem arises from a unique family code,a much more individualized preventive intervention strategy is necessary. The detectionof "ghosts in the nursery" (Fraiberg et al. 1980) requiresa much deeper penetration into the family code than would be possible in educationalor support programs. The strategy of redefinition is to interveneso that the parents will use their existing regulatory system to guide the child towardnormative developmental outcomes. Redefinition preventsan initial biological or attributed deficit from being converted into a later behavioralabnormality.Redefinition is a reasonable strategy when the parents havenormal child-rearing capacity and when they know the cultural code. But if theparents do not know this code, if there are major gape in their ability to raisea child at all, then redefinition would be an insufficient prevention strategy. Reeducation The third strategy, reeducation, simplyrefers to teaching parents how to raise children. Reeducation involves thereplacement of family myths with normative information. Itspurpose is to teach the cultural code that regulates the child's development from birth tomaturity. The most obvious targets for such prevention effortsare adolescent mothers. An increasing proportion of children are being born to teenage, unmarriedmothers who have few intellectual, social, or economicresources for raising their children. In thesecases, the intervention is aimed at training them howto be mothers. Thereare few normative strategies among these parents,and the child's survival is morea function of the child's resiliencyor the supporting social network than the parents' abilities (McDonough 1985).Other populations include parents of children at high risk because of eitherpsychosocial or biological factors. One technique that hasproven fruitful for early educational interventions is training parents to elicit infant behavior.Because of the lack of reflective capacity in very young children,a parent's behavior can elicit reflexive child responses that can immediately act as validators of theparent's caregiving behavior. With older children, behaviormodification stra egies workmore slowly with less guarantee of immediatefeedback that would reinforce the parent's feelings of efricacy. Widmayer and Field (1981) compared threegroups of low-income teenage mothers with preterm infants whoeither watched the administration ofthe Brazelton Neonatal Behavioral AssessmentScales and were trained to administer an adaptation of the scales duringtheir baby's first month,or who were only trained to administer the adaptation,or who did neither. Observations of later mother-child interactionssaw improvements in the interventiongroups, and the infants scored higheron later developmental assessments. Thesevery needy parents seemed to benefit greatlyfrom a targeted 1-month intervention CONCEPTUAL ISSUES 47 program. Positive effectsof similar interventions werefound in several studies of middle-class mothers of healthyfull-term babies (Liptak et al. 1983). Adolescent parents may be an easilyidentified target for education efforts, but there is similar need forknowledge even among middle-classprofessional parents. Changing environtypeshave produced many fathers andmothers who have had no experience taking care ofsiblings when growing up. Moreover,they are now separatedfrom their own parents, whoprovide the child-rearing training in more traditional societies.The difference between these twokinds of parents is that the middle-clessprofessionals usually will seek outthe information to educate themselves inhow to raise a child, whereas theteenage mother usually will not. In both casesinformation and training are necessary to equip the parents with thecultural code.For some parents, having the information available will be sufficient;for others, more intrusiveeducational efforts are necessary to preventdevelopmental deviancies. The three preventive interventionstrategies have been presented with examples that are directed at differentaspects of the regulatory system.A more comprehensive discussion would elaborate onthe full matrix of remediation, redefinition, and reeducation strategiesdirected toward altering microregulations, miniregulations, or macroregulationsin the family and culturalcodes. This matrix has been worked out forearly childhood interventims(Sameroff and Fiese 1989), but not for thewhole span of childhood included inthe concerns of child psychiatry. Traditionaltherapeutic efforts have focused onremediation and also on the fourth strategy (notpresented here), that of changing thechild's definition of the parents. However, theseeffiy.ts have typically not been carried out with a transactional process inmind. Using the transactionalregulatory model as a basis, the effectiveness or lackof effectiveness of many traditional intervention strategies can be reinterpreted. Conclusion

The preceding discussion has beenaimed at understanding the complexity of contextual influences on development.Through an ecological analysis, some aspects of the environtype werehighlighted as providing the regulatory framework for healthy child development.These factors included the cultural and family codes. The culture operatesthrough both formal and informal statutes, and the family functionsthrough a category of stories that includes rituals, myths, and paradigms. A case was made that the environment is anactive force in shaping outcomes. However, the shaping force is constrainedby the state and potentialities of the individual (Sameroff 1983). In an attemptto incorporate both aspects in a coherent model of developmert, theutility of the transactional model for designing programs to prevent cognitiveand social-emotional problems was explored. The development of theseproblems has been interpreted as deviations in a child-rearing regulatory system.The prevention of these problems 48 PREVENTION OF MENTAL DISORDERS

has been defined as the adjustment of the child to betterfit the regv 'atory system or the adjustr .ant of the regulatory system to better fitthe child. Developmental PF1 ytho pat hology has introducedan important reorientation in psychiatry. The principles of development that applyto the achievement of healthy growth are now seenas the same ones that apply to the achievement of illness (Sroufe and Rutter 1984).In this view, most illnessesare indeed achievements that result from the active strivings of individualsto reach an adaptive relation to their environment. Thenutrients or poisons thatex- perience provides will flavor that adaptation. Nocomplex human achievement has been demonstrated to arise without being influencedby experience. For young children, these experiences are either provided or arranged by the family. As children grow, the peergroup and school complement, supplement, and even supplant family influences in providing the regulatioasthat shape development. Within this regulatory framework, transactionsare ubiquitous. Whenever parents change their way of thinking aboutor behaving toward the child as a result of something the child does,a transaction occurs. Most of these transactions are normative within the existing culturalcode and facilitate development. Intervention becomes necessary only when thesetransactions are nonnormative. A normative event for which society is preparedis one in which the family registers the child in school. Society respondsby changing a large part of the child's environment through the provision ofa new physical environment, the school; new regulators of socialization, the teachers;and a new social network, the classmates. A nonnormative event for whichsociety may or may not be prepared is one in which the parent seeks professionalhelp for a deviant child. The degree of help thatcan be provided is a function of society's awareness of how development is regulated and the availabilityof resources for intervening. In our progress toward effective preventionprograms, we have reached a key theoretical breakthrough. The problems of childrenare no longer seen as restricted to children. Social experience isnow recognized as a critical component of all behavioral developments, both normaland abnormal. Unfor- tunately, we have not yet reached the level ofsophistication in theory and research that would connect each childhood problemwith a corollary regulatory problem. A more profound enigma is that therear^ rnany possible regulations to solve the same problem and, therefore,many po..Ae interventions. Future research needs tl test the relative efficacies ofinterventions at the individual, family, or cultural level. The role of health professionals is to preventdeviancies by intervening in the re!ntlation of child growth where the normativemechanisms are not functioning ade quately. The complex biological model thatcharacterizes modern understanding of the regulation of developmentseems an appropriate one for analyz- ing th 9 etiology of mental illness and retardation.It permits the understanding of prevention at anecessary level of sophistication so that appropriate targets can be identified for intervention. It helps us to understand whyinitial conditions do not determine outcomes, either positivelyor negatively. The model also CONCEPTUAL ISSUES 49 helps us to understand whyearly prevention efforts maynot determine later development where regulations can outcomes.There are many points in facilitate or retard the child's progress.The hopeful part of the modelis that these many points in time areopportunities for changing the courseof development. inadequate for both In sum, models that focus onsingular causal factors are the study and the manipulationof developmental outcomes.The evolution of living systems has provided aregulatory model that incorporatesfeedback mechanisms between the individualand regulatory codes. Thesecultural and genetic codes are the context ofdevelopment. By appreciating theworkings of this regulatory system, we canobtain a better grasp of the processof development and how to change it.

References

Barrett, D.E,; Radke-Yarrow,M.; and Klein, R.E. Chronicmalnutrition and child belie vior: Effects of earlycaloric supplemontation oncial and emotional functioning at school age. ChildDevelopment 18:541-556, 1982. Belsky, J.Child maltreatment: Anecological integration.Amoican Psychologist 35:430-435, 1980. Broman, S.H.; Nichols, P.L.; andKennedy, W.A. Preschool IQ:Prenatal and Early Development Correlates.New York: L Erlbaum Associatc3,1975. Bronfenbrenner, U. Toward anexperimental ecology of humandevelopment. American Psychologist 32:513-531,1977, illness in Commission on Chronic Illness.Prevention of chronic illness, chronic the United States, Vol. 1.Cambridge, MA: Harvard UniversityPress for the Commonwealth Fund, 1957. Cravioto, J., and DeLicardie, E.R.Nutrition, mental development, andlearning. In: Falhner, F., and Turner,J.M., eds. Human Growth. Vol.3. New York: Irenum Pros% 1979. eVries, M.W., and Sameroff,A.J. Culture and temperament:Influences on temperament in three EastAfrican societies.American Journal of Ortho- psychiatry 54:83-96, 1984. Ebert, J.D., and Sussex, I.M.Interacting Systems in Development.2d. ed. New York: HR & W, 1970, Edwards, P.Tuberculosis.In:Frankenburg, W.K., and Camp, B.W.,eds, Pediatric Screening Tests.Springfield, Ill,: Thomas, 1975.

f; 50 PREVENTION OF MENTAL DISORDERS

Eisenberg, L. Conceptual issueson biobehavioral interactions.In:Perron, D.L., and Eisenberg, L., eds. Infants at Risk forDevelopmental Dysfunctions. U.S. Health and Behavior: A Research Agenda InterimReport No. 4. Washington, D.C.: National Academy Press, 1982. Elder, G.H.; Caspi, A.; and Downey, G. Problembehavior in family relationships: A multigenerational analysis. In: Sorensen, A.;Weinert, F.; and Sher- rod, L., eds. Human Development: InterdisciplinaryPerspective. Hillsdale, N.J.: L Erlbaum Associates, 1983.pp. 93-118. Ferreira, A. Family myth and homeostasis. Archivesof General Psychiatry 9:457-463, 1963. Field, T.M. Interaction patterns of preterm andterm infants. In: Field, T.M.; Sostek, A.M.; and Schuman, H.H., eds. InfantsBorn at Risk: Behavior and Development. New York: SP Medical and Scientific Books,1979. Fraiberg, S.; Adelson, E.; and Shapiro, V.Ghosts in the nursery: A psychoanalytic approach to the problems of impairedinfant-mother relationships. In: Fraiberg, S., ed. Clinical Studies in InfantMental Health: The First Year of Lift. New York: Basic Books, 1980. Golden, M., and Birns, B. Social class and infantintelligence. In: Lewis, M., ed. Origins of intelligence: infancy and earlychildhood. New York: Plenum, 1976.

Gollin, E.S. Development and plasticity. In: Gollin,E.S., ed. Developmental Plasticity. New York: Academic Press, 1981.

Gordon, R.S. An operational classification of diseaseprevention. Public Health Reports 98:107409, 1983. Greenspan, S.I.Psychopathology and Adaptation in Infancyand Early Childhood: Clinical Infant Reports No. 1. NewYork: International Univer- sities Press, 1980.

Haley, J. Leaving Home. New York: McGraw-Hill,1980. Hareven, T.K. Themes in the historical developmentof the family. In: Parke, R.D., ed. Review of Child Development Research.Vol. 7. Chicago: University of Chicago Press, 1984.

Hudson, J., and Nelson, K.Effects of script structureon children's recall. Developmental Psychology 19:625-635, 1983.

Huesmann, Eron, L.D.; Lefkowitz, M.N.; and Walder,L.O. "The Stability of Aggression Over Time and Generations."Paper presented at the meeting of the Society for Research in Child Development,Detroit, Mich., April 1983. CONCEPTUAL ISSUES 51

Kohn, M.L. Class and Conformity: A Study inValues. Homewood, Ill.: Dorsey, 1969, Kohn, M.L. Social class and schizophrenia: A criticalreview and reformulation. Schizophrenia Bulletin 7:60-79, 1973. Kramer, J. Family myth and homeostasis.Archives of General Psychiatry 9:457-463, 1985. Kurked, L.A.An integrative perspective on children'sdivorce adjustment. American Psychologist 36:856-866, 1981. Lamb, H.R., and Zusman, J. Primary preventionin perspective. American Journal of Psychiatry 136(1):12-17, 1979. Lazar, I., and Darlington, R. Lasting effects ofearly education: A report from the consortium for longitudinal studies.Monographs of the Society for Research in Child Development, 47(195), 1982. Leavell, H.R., and Clark, E.G. PreventiveMedicine for a Doctor in His Com- munity: An Epidemiological Approach . 3ded. New York: McGraw-Hill, 1965. Lewis, J.M., and Beavers, W.K. No SingleThreadPsychological Health in Family Systems. New York: Brunner/Mazel,1976. Liptak, G.S.; Keller, B.13.; Feldman, A.W.; andChamberlin, R.W. Enhancing infant development and parent-practitionerinteraction with the Brazelton neonatal assessment scale. Pediatrics72:71, 1983. McDonough, S.C. Intervention program foradolescent mothers and their offspring. Journal of Children in ContempormySociety, 1985. McKeown, T.M. The Role of Medicine: Dream,Mirage or Nemesis? London: Nuffield Provincial Hospitals Trust, 1976. Nelson, K. Social cognition in a script framework.In: Flavell, J., and Ross, R, eds. Social Cognitive Development.Cambridge, England: Cambridge Univer- sity Press, 1981. Nelson, K., and Gruendel, J. Generalized eventrepresentations: Basic building blocks of cognitive development. In: Brown, A., andLamb, M., eds. Advances in Developmental Psychology. Vol. 1.Hillsdale, N.J.: L Erlbaum Associates, 1981. Parke, R.D., and Tinsley, B.J. Family interaction ininfancy. In: Osofsky, J., ed. Handbook of Infant Development. 2nded. New York:Wiley, 1987. pp. 579-641. Parmelee, A. H., and Haber, A. Who is the at riskinfant? Clinical Obstetrics and Gynecology 16:376-387, 1973. 52 PREVENTION OF MENTAL DISORDERS

Patterson, G.R. Performance models for antisocial boys.American Psychologist 41;432-444, 1986. Ratner, N., and Brunher, J.Games, social exchange and the acquisition of language. Journal of Child Language 5:391-401, 1977. Read, M.S. Malnutrition and behavior. AppliedResearch in Mental Retarda. tion 3:279-291, 1982. Reiss, D. The Family's Construction of Reality.Cambridge, Mass.: Harvard Univemity Press, 1981.

Reiss, D. The represented and practicing family: Contrastingvisions of family continuity. In: Sameroff, A.J., and Emde, RN., eds. RelationshipDisturbances in Early Childhood: A Developmental Approach. NewYork: Basic Books, 1989. Reiss, D.; Oliveri, M.E.; and Curd, K Family paradigmand adolescent social behavior. In: Grotevant, H.D., and Cooper, C.R., eds.Adolescent Development in the Family: New Directions for Child Development.Vol. 22. San Francisco: Jossey-Bass, 1983. pp. 77-91.

Rodm, J. Aging and health: Effects of thesense of control. Science 233:1271- 1275, 1986.

Rogoff; B. Schooling and the development of cognitive skills.In: Triandis, H.C., and Heron, A., eds. Handbook of Cross-CulturalPsychology: Developmental Psychology. Vol. 4. Boston: Allyn and Bacon, 1981. Reekie3, E.Abnormedity and Normality:The Mothering of Thalidomide Children. Ithaca, N.Y.; Cornell University Press,1972. Rutter, M. Protective factors in children'sresponses to stress and disadvantage. hr. Kent, M.W., and Rolf, J.E., eds. PrimaryPrevention of Psychopathology. Vol. 3. Social Competenoe in Children. Hanover,N.H.: University Press of New England, 1979. Sameroff, A.J.Early influences on development:Fact or Fancy? Merrill- Palmer Quarterly 21:267-294, 1975. Sameroff, A.J. Developmental systems: Contextsand evolution. In: Kessen, W., ed. History, Theories, and Methods. Vol. Iof Musson, P.H., ed. Handbook of Child Psychology. New York: Wiley, 1983. Sameroff, AJ, "Can Development Be Continuous?"Paper presented at the Annual Meeting of the American PsychologicalAssociation, Los Angeles, 1985. Sameroff, A.J.Env ironmental context of child development.Journal of Pediatrics 109;192-200, 1986.

Sarneroff, AnY. The ociaI context of development.In:Eisenberg, N., ed. Contemporary Topics in Developmental Psychology.New York; Wiley, 1987.

.1r CONCEPTUAL ISSUES 63

Sameroff, A.J., and Chandler, M.J.Reproductive risk and the continuumof caretaking casualty. In: Horowitz,F.D.; Hetherington, M.; Scarr-Salapatek, S.; and Siegel, G., eds. Review ofChild Development Research. Vol. 4.Chicago: University of Chicago Press, 1975. Sameroff, A.J., and Fieee, B.H. Transactionalregulation and early interaction. In: Meisels, S.J.; and Shonkoff,J.P., eds. Early Intervention: AHandbook of Theory, Practice, and Analysis.Cambridge, England: Cambridge University Press, 1989. Sameroff, A.J.; Seifer, R.; Barocas, B.;Zax, M.; and Greenspan, S. IQ scoresof 4-year .old children: Social-environmentalrisk factors. Pediatrics, 1987. Sigel, E. Parental belief systems: Thepsychological consequences for children. Hillsdale, N.J.: L Erlbaum Associates, 1985. Sroufe, L.A., and Rutter, M. The domainof developmental psychopathology. Child Development 55:17-29, 1984. Stern, D. The First Relationship:Infant and Mother.Cambridge, Mass.: Harvard University Press, 1977. Strauss, MA.; Gelles, R.J.; and Steinmetz,S.K. Behind Closed Doors: Violence in the American Family. New York:Doubleday, 1980. Thomas, A., and Chess, S. The EVnarnicsof Psychological Development. New York: Brunner/Mazel, 1980. Waddington, C.H. The Strategy of the Genes.London: Allen & Unwin, 1957. Weiner, I.B. Psychological Disturbancein Adolescence. New York: Wiley,1970. Werner, E.E., and Smith, R.S. Vulnerablebut invincible: A longitudinal study of resilient children and youth. NewYork: McGraw Hill, 1982. Widmayer, S.M., and Field, T.M.Effects of Brazelton demonstrations for mothers on the development of preterminfants. Pediatrics 72:711-714, 1981. Winick, M.; Meyer, K.; and Harris, R.Malnutrition and environmental enrichment by early adoption. Science190:1173-1175, 1975. Zigler, E., and Trickett, P.K. IQ, socialcompetence, and evaluaiion of early childhood intervention programs.American Psychologist 33.789-799, 1978. CHAPTER 2

The Preventionof Child and AdolescentDisorders: From Theory to Research

Raymond P. Lorion, Ph.D. Professor of Psychology andDirector of Clinical Training University of Maryland, CollegePark

Richard H. Price, Ph.D. Professor of Psycholoiy and Directorof Industrial-Organizational Psychology University of Michigan, Ann Arbor

William W. Eaton, Ph.D. Associate Professor of Mental Hygiene Johns Hopkins University, Baltimore,Maryland

A defining goal of this volumeis to offer a paradigmwithin which the prevention of child and adolescentdisorders can be conceptualized,researched, and translated into viable andeffective intervention strategies.Our participation in the effort reflects ourfirm conviction that without aparadigm, the scientific accumulation ofknowledge is, at best, inefficientand, at worst, impossible. In his seminal discussionof the role of the paradigm inthe history of the physical sciences, kuhn(1970) made clear the limits ofaparadigmatic research: All of the facts that could possiblypertain to the development of agiven science are likely to seem equallyrelevant. As a result, earlyfact-gathering is a far more nearly randomactivity than the ono that subsequentscientific development makes familiar. Furthermore,in the absence of a reasonfor seeking some particular form of morerecondite information, early fact- gathering is usually restricted to thewealth of data that lie ready tohand. The resulting pool of facts containsthose accessible to casualobservation and experiment together with someof the more esoteric dataietrievable from established crafts.

This paper was prepared for theProject Prevention initiative of theAmerican Academy of Child and AdolescentPsychiatry. 56 PREVENTION OF MENTAL DISORDERS

But though this sort of fact-gathering has been essentialto the origin of many significant sciences . .it produces a morass. One somehow hesitates to call the literature that results scientific. (Kuhn 1970,pp. 15-56) By contrast, a paradigm serves ae,a conceptual framework within which a phenomenon of interest is understhod in terms of its evolution, maintenance, and alteration. Moreover, the availability of sucha framework enables those interested in the phenomenon to organize and integrateits established knowledge base, identify and prioritize its unanswered questions, and applyor refine its procedures for resolving those questions. In effecta paradigm allows for, indeed requires, the systematic accumulation ofknowledge because it provides criteria for assessing eachnew fact's relevance to and consistency with the framework's theoretical foundation. Similarly, it contributesto the development and validation of measurement procedures by providing the "nomological network" necessary for construct validation (Cronbachand Meehl 1955). It also informs the design of necessary methodological innovations.For these reasons, it is important for researchon the prevention of child and adolescent disorders to evolve paradigmatically. Sameroff and Fiese (chapter 1 of this volume) described thetheoretical base for the paradigmatic approach presented here.The defining tenets of the transactional model (originally described by Sameroff andChandler 1975) argue that much human behaviorwhether cognitive, emotional,or physicalis determined jointly by characteristics of the individualsand of the environments in which they live. By definition, the theory is rootedfirmly in an ecological perspective. As such, it recognizes both individual and environmentalcon- tributors to developmental outcomes. Moreover, itappreciates the continuous synergistic interrelationships between these two msjorinfluences.Conse- quently, the transaction model posits that individualcharacteristics (e.g., a genetically based vulnerability ora particuler temperamental predisposition) in many cases are manifested only under specific environmentalconditions (e.g., in the presence of povertyor familial instability or in a setting that demands precise visual-motor coordination). It simultaneouslyrecognizes that the environment both shapes and is shaped by the individualswho inhabit it. Central to the model is an appreciation of the continuousnature of this process. Thus, throughout a person's lifespan psychological andbehavioral status is influenced by contextual factors, which themselvesare influenced by the individual. Sameroff and Fiese described the resulting interrelationshipebetween organism and environment in terms of "transactions."The proposed paradigm holds that the capacity to understand and predicthuman outcomes depends on the identification, analysis, and ultimate understandingof such transactions. It is assumed that with such understanding willcome significant increments in the ability to predict and control causal and etiologicalcontributors to child and adolescent disorders. As will become evident, we view the identification,measurement, and ultimate control of such contributors to adaptive andmaladaptive functioning as FROM THEORY TO RESEARCH 67 the sine qua non of preventive intervention research.The transactional model argues that risk for disorder can arise fromthree sources: (a) characteristics of the individual, which are causally linked or predispoee oneto disorder or dysfunction; (b) characteristics of the environment thatimpinge on the individual; and (c) particular combinations of individualand environmental characteristics. Together, these elements contribute to the sequencesof events that precede and evolve into functional ordysfunctional behavior. Such sequences define the "etiological chains"preceding and maintaining the emotional and behavioral outcomes that prevention efforts aredesigned to avoid. As explained subsequently in this chapter, we believe that theidentification of one or more of the links in such chains is a necessaryprecondition both for the design of interventions and, typically, for the selection of itspotential recipients. Thus, we concur with Sameroff and Fiese's position that seriousquestions must be raised about the justification of applying apreventive intervention in the absence of at least partial knowledge of a disorder'setiology and of those factors associated with risk for that outcome (Lorion1985). This chapter focuses on methodological strategies forgenerating and assessing such a knowledge base. In addition, the chapter addresses the issue oftranslating such knowledge into risk assessment and reduction (i.e., preventiveinterventions) procedures testable by preventive trials. Related to the appreciation of etiological chains andtheir associated risk factors is the importance for prevention researchers ofunderstanding the influence of base rates on the application of their interventionsand the assessment of the effectiveness of these interventions.For the most part, child and adolescent disorders are relatively rare events affecting, ingeneral, fewer than one of ten children. Certaincategories of dysfunction, for example, attention deficit disorders, occur more frequently but sti,i reachcomplete diagnostic manifestation in, at most, only one of four children. Otherdisorders, such as autism and other childhood psychoses, are much more rare(Achenbach 1982). Such population base rates are of limited use in planningpreventive interventions. Because the likelihood of occurrence of specificdisorders is rarely constant across all segments of the population, it isessential that one be able to estimate that likelihood for the recipients of one's intervention.Otherwise, it is impossible to interpret the preventive results of theintervention. For example, a rate of postintervention disorder significantly below thepopulation rate may, in fact, exceed the rate at which thedisorder typically occurs in the subgroup researched. By contrast, a rate of postinterventiondisorder that exceeds the population base rate may, in fact, be significantlybelow that rate typically found in the subgroup involved. In either case, the actualpreventive value of the intervention will not be recognized. The section onstatistical considerations discusses in detail the care that must be taken toaddress the "base rate" issue in prevention research. Research on the prevention of child and adolescent disordersis necessarily developmental in nature for two reasons. As explained in detailelsewhere (Lorion 1986; 1987a) and in this chapter, we believe thatpreventive interven- j 58 PREVENTION OF MENTAL DISORDERS

tions must be designed to impact on theprocesses leading to the outcomes to be avoided rather than on the outcomes themselves. Knowledge of relevant etiological chains must inform the design of prevention interventions. Under- standing the elements of such chains necessarily involves the acquisition of knowledge about multiple temporal characteristics of disorder, including the following: The chronological sequence among the links of sucha chain. The latency between the occurrence ofprecursor events or conditions and the manifestation of detectable signs or symptoms of dysfunction. The point(s) along the identified etiological chain at whichan intervention can and should be initiated. The appropriate duration of an intervention. The latency of intervention effects (i.e., how long before effectsappear). The expected duration of intervention effects. For example,if no signs appear for 5 years after intervention, can one claim an intervention "success"? We believe that it is important to distinguish between preventivetrials and clinical trials. As stated, we perceive the formeras inherently developmental and longitudinal in nature; such is not necessarily thecase with respect to clinical trials. A clinical trialassesses the effects of an intervention on the reduction or removal of a measurable condition that ispresent in all of the subjects included in the trial. By contrast, preventive trialsassess the capacity of an intervention to avoid partor all of the elements of an etiological chain and moot notably the occurrence of the disorderor dysfunction of interest. In effect, therefore, a successful preventive interventionreplaces a potentially pathogenic developmental process with its nonpathogenicor normative counterpart. We are tempted to argue that documentation ofsuch an exchange of developmental sequences represents the operationaldefinition of an effective preventive intervention. Finally, to be of practical significance, the interventionmust be adoptable; that is, its procedures must be transferable from thesetting in which they were developed to other settings and situations. Forthat reason, consideraimn is given to the mejor issues concerning the identificationof effective and essential intervention components. We believe such "process"or "formative" evaluation procedures are important components of the analysisof program "adoptability." The latter term refers to theease and effectiveness with which an intervention evolves from being a research protocol examinedwithin a preventive trial toa functioning program thatcan be applied in a standardized fashion under varying situational conditions. Sensitivity to adoptabilitythroughout the latter FROM THEORY TO RESEARCH 59 stages of the research effort, we believe, can significantlyenhance the im- mediacy with which demonstrably effective strategies canbe applied to the alleviation of actual needs under real-life circumstances. Each topic discussed thus far is considered in detail inthe pages that follow. Space does not permit comprehensive consideration ofall topics, which are discussed further by Felner et al. (1983); Roberts andPeterson (1984); Quay (1987); Steinberg and Silverman (1987); and Rickel andAllen (1988).In addition, other readings are identified in the comprehensiveannotated bibliography prepared by Buckner et al. (1985). Our intent in thischapter is not to discuss the evaluation of preventive interventions whoseeffectivegess has already been ascertained through preventive trials. Rather, ouremphasis is on the development and application of the knowledge bases onwhich interventions will be designed and tested. Thus, we examine ways tothink about and design research the results of which will inform the development ofviable prevention strategies and eventuate in rigorous preventive trials.Once developed and validated, such strategies need to be examined systematicallyin terms of their adoptability to other settings and different populations, with bothsummative and formative evaluations. Discussion of those procedures ispresented by Price and Lorion in chapter 3 of this voiume. Outcomes of Preventive Interventions

Two approaches for classifying preventive outcomes havereceived considerable attention by the mental health disciplines. The approachproposed by Caplan is discussed here. The alternative offered by Gordon (1983) is described at the beginning of the next section. More than two decades ago, Caplan (1964) urged adoption of the classictriad of prevention outcomes used by public health practitioners. Theseprevention effortsprimary, secondary, and tertiaryare discussed here in reverse order. Tertiary Prevention Tertiary prevention refers to efforts that avoid the sequelae of established disorders.Tertiary efforts seek to minimize the long-term and secondary consequences of disorder, including those related tochronicity and to participation in a treatment protocol. Many (e.g., Albee 1982, 1983, 1986; Bloom1984; Cowen 1983) have argued that tertiary efforts are, in fact, not preventive because they are initiated only after a disorder has been established. From one perspective, this is undoubtedly true. Yet it also assumes that the individual is irreversibly altered by the disorder and that having the disorder is a continuous state. The transactional model, however, posits that the presence orabsence of symptoms depends on the occurrence of specific transactions between the individual and the environment.If pathogenic transactions can be avoided, then not all episodes need occur. Thus, tertiary prevention includes efforts to avoid recurrent episodes of presumably chronic conditions (e.g., conduct disor-

7 t) 60 PREVENTION OF MENTAL DISORDERS dere or chemical dependency) or of the significant sequelae that often accompany dysfunction (e.g., the behavioral concomitants of learning disorders). For this reason, we question the categorical elimination of the pursuit of such outcomes from "true" prevention efforts. Given the considerable demand on the nation's mental health and human service resources, national prevention goals should include the avoidance of repeated institutional placement; of long-term dependence on treatment services; and of the negative concomitants of cognitive, emotional, and behavioral handicapping conditions. Of course, such inclusion blurs the distinction between treatment and prevention.Nevertheless, the potential benefits of designing viable strategies that control such burdensome sequelae of chronic disorders may justify the resulting conceptual imprecision. Furthermore, tertiary efforts directed at one condition may have preventive consequences for another. To date, evidence of the value of their strategy has primarily involved physical conditions (e.g., control of hyper- tension to prevent heart attacks and arteriosclerosis). Its applicability to emotional and behavioral disorders (e.g., remediation of learning disabilities to prevent alcohol and other drug abuse), however, appears worthy of investigation. Secondary Prevention Effective secondary efforts reduce the number of active cases of a condition in the population, that is, its "prevalence"the total number of cases (or proportion of the population) in existence at a given time. Prevalence reflects the combined con# ributions of "incidence" (i.e., the rate at which new cases develop during a specified period of ex_ osure) and the average duration of a disorder or condition.Secondary efforts reduce prevalence by affecting chronicity. Strategies that affect duration early in the genesis of disorder are called secondary prevention by public health practitioners.Such strategies are designed to interrupt the continued evolution of pathogenic processes and thereby to avoid the complete clinical manifestation of the disorder. By effecting preclinical states, secondary efforts preclude the need for standard treatment procedures. Secondary prevention outcomes depend on the successful design and application of psychometrically sound screening procedures that detect individuals experiencing preclinical states. Secondary prevention includes effective early intervention strategies that abort what would otherwise be a pathogenic sequence leading to an eetablished disorder. The many advantages of secondary prevention efforts over tertiary procedures include the avoidance of seriot... levels of dysfunction, the minimization of the secondary consequences of dysfunction, and the option of applying relatively inexpensive and nonin- trusive interventions. Successful secondary prevention strategies give truth to the apholism that a "stitch in time saves nine." Primary Prevention Even though secondary prevention efforts are preferable to the application of treatments to established conditions, such strategies do not protect 71 FROM THEORY TO RESEARCH 61 individuals from experiencing preclinical states, nor do they avoid the onset of pathogenic sequences. As a consequence, some consider them not truly "preventive" (e.g., Bloom 1984; Cowen 1983) and others consider them only questionably distinct from treatment (Albee 1986).The third public health category of prevention, primary prevention, represents the ideal for all. Designed to reduce prevalence by lowering the incidence rate, primary prevention efforts reduce the number of individuals in whom the relevant pathogenic sequence is initiated.In essence, primary prevention efforts replace one developmental process with another. They share thischaracteristic with all other forms of prevention. They are distinguished, however, by their intent to avoid entirely- the onset of the pathogenic sequence. As noted by Catalano and Dooley (1980), onset may be avoided either reactively through the elimination of apathological sequence or proactively through the initiation of apositive adaptation-producing sequence.In either case, as discussed later in this chapter ("Analysis of Intervention Components" and "Statistical Considerations"), it is the demonstration of that change in the onset ot processes that operationally defines the distinction between primary and secondary prevention. If equally effective, both would achieve comparable results in terms of reducing the number of cases requiring treatment. In addition, however, the primary prevention group would not have experienced relevant preclinical states. Thus, Caplan's proposed adoption by the mental health disciplines of public health's triad of preventive outcomes offers three categories of intervention to apply in responding to child and adolescent disorders, In effect, one can avoid onset, the exacerbation of preclinical states, or the sequelae of treatment and chronicity. Caplan's suggestion that the mental health disciplines model their preventive efforts after those traditionally used by public health practitioners no doubt contributed to the disciplines' increasinginvolvement in the design of preventive interventions during the past two decades. At this time, however, it is appropriate to examine the value of these categories to produce further progress in the pursuit of effective preventiveinterventions,In our view, inadequate attention has been paid to the fact that selection among these alternativel presumes the availability of quite distinct bodies of knowledge: In the absence of knowledge of a disorder's causes and/or of the individual, familial, and environmental conditions for its manifestations, the initiation of a primary prevention effort appears premature. Similarly, if one is ignorant of the preliminary manifestations of a targetdisorder, unable to systematically detect their presence, or incapable of altering their evolution, one is unprepared to attack a problem at the secondary level.Finally, if we are unaware of how a specific skill develops and is maintained in the everyday environment, enhancement efforts may need to bc deferred, (Lorion 1983, p. 257) As noted by Sameroff and Fiese, application of the public health classificatory schema to emotional and behavioral disorders creates some problems. For example, unintended though it may be, the labels "primary," "secondary," and "tertiary" imply priority among those alternativee. Yet, selection among them

P"4 62 PREVENTION OF MENTAL DISORDERS

should be based on characteristics of the disorder to be avoided and the logistics of program implementation. This issue is discussed insome detail later in this chapter. At this point, we simply encourage the reader to remainopen minded about all three forms of preventive intervention. We alsopropose that the debate about their respective merits be deferred until specific examples of demonstrably effective strategies are available for direct comparison. It should also be noted that the public health categories of preventive interventions relate to a particular model of how disease evolves. Specifically, as pointed out by Sameroff and Fiese, the categories are best applied to disorders that follow a simple mechanistic and linearprocess from onset to clinical manifestation.This model is best illustrated by a viral infection, suchas measles, in which an agent (i.e., the virus) invades the host and thereafter follows a predictable course leading to a diagnosable disease state. To theextent that the transactional model accurately portrays the evolution of emotional and behavioral disorders, however, questions ariseas to the consequent meaning of "onset" and the "course of disease evolution." Onset "Onset" implies that there existsan identifiable point in time before which the disease process was not operative in the organism and after which itwas. Identifiable perhaps in the case of the measles virus, onsetseems much less easily established in the transactional model, in which disorderrepresents the consequence of a sequence of increasingly disruptive transactions between the organism and its environment. Where along thatsequence does one place onset? Given the fact that early stages of suchsequences appear with much greater frequency in the population thon cases of established disorder, the criteriafor defining a "case," that is, for thetinguishing itspresence or absence, must be carefully operationalized. Yet, as noted by Long (1986), the disability associtited with emotional and behavioral disorders more often than not developsalong a continuum "from barely identifiable behavior to disabling symptomsand breakdown" (p. 827). Where along that continuum doesone place the onset of disorder per se? The definition of onset in terms of the initiation ofa pathogenic process raises significant conceptual and measurement challenges for prevention researchers, because at that point, by definition, thereare no individual signs of disorder or dysfunction. By contrast, the definition of onset interms of the presence of "preclinical signs and symptoms" appears to preclude theachievement of primary prevention goals. Moreover, the definition of onset ofa pathogenic sequence is complicated by the fact that often individual or environmental characteristics that precede many disorders do not necessarilyresult in disordor (Sameroff and Chandler 1975). For thatreason, such characteristics cannot be used exclusively to define onset. In a sense, the transactional perspective presentsa deterministic view of pathogenesis; that is, one's vulnerability to disorder isa function of one's FROM THEORY TO RESEARCH 63 experience before that given point in time. In otherwords, the appearance of symptoms reflects both historical and contemporaryfactors.Transactional processes, however, do net. cease atthat point. Rather, transactions between a "child with a disorder" and the environment determinewhether and how disorder is maintained, exacerbated, or alleviated. Adoption of thetransactional model requires that we not lose sight of its ongoing,dynamic nature. Whereas in some instances a "pathogen" can be identifiedforexample, the loss of a parent through death or divorcein others the sequenceof life experiences per se appears pathogenic. In the former instance, onemight fix onset at the time of the event; in the latter, the conceptof onset seems lees applicable. As one reviews the range of emotional andbehavioral disorders that affect children and adolescents, it becomes apparentthat a few (e.g., reactive depression or alcohol and other drug use) may have identifiableonsets, whereas most (e.g., attention deficit disorder, autism,childhood schizophrenia) may not. That the issue of onset is important in conceptualizingpreventive interventions and in designing methodologies for their evaluation isemphasized by Albee (1982): Primary prevention efforts are aimed at reducing the incidenceof mental disturbance in groups of people. Incidence refers to the total number of new cases appearingwithin a specified time period.Therefore, it is important in determining incidence to be able to tell when acondition actually begins. The more vague the time of onset, the more difficultit is to measure incidence, and the greater the temptation to useprevalence, the total number of cases that exist within a specified time as a measure of rate. (p. 1045) Albee (1982, 1986), among others (e.g., Cowen 1986; Long 1986), argues against using prevalence as an outcome for preventiveinterventions. Under- s'Andably, they view avoiding onset as le ideal goal of prevention.Their position, however, leaves unanswered the alternative to influencingprevalence in those cases in which onset cannot be fixed. Within the categoryof significant disorders without definable onset, one must include attention-deficitdisorder, specific learning disabilities, nonreactive depressions, and mostchildhood psychoses.Together these condifions constitute the majority of child and adolescent dysfunctions.For that reason alone, reducing their prevalence should represent an acceptable goal of preventive efforts. Strategies that interrar t the progression of an ongoing pathogenic sequence have important heuristic benefits. As stated elsewhere: In contrast to an apparent trend to dismiss the real and potential contribution of secondary efforts as insignificant in relation to the promise of primary approaches, we argue that secondary efforts offer both an opportunity for immediate reductions in rates of disorder and the potential for highlighting promising routes for primary preventive efforts.Logis- tically, secondary efforts involve technologies that are closely related to 64 PREVENTION OF MENTAL DISORDERS

those available to most traditionally trained mental health professionals (i.e., screening, diagnosis and treatment).Moreover, those served by secondary efforts are deemed by the general public to be appropriate for the receipt of publicly funded services. Although they are less intense and debilitating than those of traditional clientele, the needs of secondary prevention target populations can be documented and, in a world of increasingly limited human service resources, justified for both humane and economic reasons. Secondary preventive efforts reduce human suffering at minimal cost (Cowen, 1973).Overall, the advantages of active secondary preventive efforts are clear. We can carry out such efforts now and, in the process, gain both information and.. .important credibility for the concept of prevention. (Lorion and Lounsbury 1982, pp. 28-29) Rather than emphasizing the theoretical distinctions between primary (i.e., incidence focused) and secondary (i.e., preva'gence focused) preventive efforts, vv.-a propose that their overlapping value for emotional and behavioral disorders be appreciated. We further propose that they be understood and operationalized within the transactional perspective. In doing so, we believe that prevention researchers will be confronted with challenges that, when resolved, will make possible an acceleration of the pace at which positive evidence of prevention's impact is found and reported in the literature. Central to these challenges is the need to understand that the target of preventive efforts is theprocesses that lead to disordered states rather than the states themselves. In explaining the transactional perspective, Sameroff and Fiese described behavioral andemotional status at any point in time as reflecting the product of synergistic exchanges among individual and environmental factors over time. By understanding behavior as the observable manifestation ofa continuous sequence of underlying normative or pathological processes, one gains insight into whatwe believe will be a productive approach to conceptualizing preventive interventions. If, as proposed, diagnosable conditions represent the product and ongoing occurrence of such processes or pathogenic sequences, the focus of preventive interventions should be on the avoidance or alteration of thosesequences. In effect, within this framework, primary prevention would be defined operationally as the replacement of a pathogenic sequence thatwas to have occurred with a normative sequence. In this instance, however, the problem of defining onset remains a serious obstacle. Secondary prevention, by contrast, would involve the alteration of an ongoing pathogenic process so that its anticipated pathological outcome is replaced by a normative transactional pattern. Focusing on the prevention of sequences or processesas the key to avoiding diagnosable end states increases our ability toassess the effectiveness of interventions. In essence, one now has the potential toassess repeatedly the degree to which the preventive goal is being achieved and to identify individual, environmental, and programmatic elements that contribute toor detract from that effort. Adopting terminology introduced by Price (1982),one can label the diagnosable end state as the "distal outcome," that is, the behavioral state that occurs at the end of the pathogenic sequence. Prior to that point, however,one 76 FROM THEORY TO RESEARCH 65 assumes that there exist (to paraphrase Long 1986)"increasingly identifiable behaviors" representing intermediary steps between normal functioning and pathology. Such intermediary steps can serve as marker indicators signaling that either a new sequence has been initiated or an existing one has changed in a positive direction. Again, using Price's terminology, welabel the assessment of such steps as "proximal outcomes." Thus, selection between primary and secondary approaches will be determined not in terms of one's preference but instead in terms of the completeness with which one can articulate the relevant proximal outhomes leading to the distal outcome of interest. Ultimately the decision among alternative prevention strategies must be based on the availability of psychometric procedures to differentiate cases, that is, "those who do or do not have the disease" (Bloom 1984, p. 198). Course of Disease Evolution Complicating this seemingly simple discrimination is the inevitability of the course of disease evolution alluded to earlier. Emotional and behavioraldysfunctions rarely have a single identifiable cause.Reciprocally, an identified causal factor, for example, a critical life event or stressful transition (Felner 1984), can affect an individual's vulnerability for a number of disorders. Within the transactional perspective, one would interpret this "spectrum of dige.,ers" effect as reflecting both the sensitivity of a range of individual vulnerabilities to certain events and the influence of environmental factors in determining how that vulnerability manifests itself. This explains, in part, why an identifiable perceptual-motor deficit may look like an isolated reading impairment, a conduct disorder, or an affective disturbance. Often, a combination of these deficits will be present (Ross 1977; Silver 1984). Inevitability is an important concern because conditions defined as end states for children and adolescents (e.g., attention-deficit disorders, conduct disorders) also frequently represent markers for adult disorders (Kellam et al. 1983; Small 1973). Of course, justifiable questions have been raised about the predictability of adult disorder based on the occurrence of childhood conditions (e.g., Kohlberg et al, 1972; Rutter 1972). Consistent with the transactional perspective, at least some portion of the reported discontinuity can be explained by such environmental events as the availability of adequate social support networks (Heller et al. 1986; Rutter in press). An additional portion will undoubtedly be explained by the fact that certain disorders (e.g., specific learning disabilities) depend on the presence of specific environmental ceendition.s for their manifestation (e.g., assignment of a child with an auditm7 pervaptual impairment to a classroom that emphasiza3 a phonetic rather than a sight-word approach to reading). Thus, the prevention-oriented researcher and clinician is confronted by a most complex situation when focusMg on child and adoiesoent disorders. Specific outcomes can represent the result of quite distinct pathogenic sequences. Rarely will a single cause be found for a single condition. More likely, disorder will appear fiAlowing the occurrence of a sequence (If increasingly problematic behaviors. As the individual dii3p1ays each of the behaviors char- 86 PREVENTION OF MENT AL DISORDERS acterizing the sequence, the environment's response to that behavior will influence positively or negatively the likelihood that a subsequent pathogenic step will be taken. Rather than a deterministic process, the sequence, in fact, is stochastic. As Sameroff and Chandler (1975) and Sameroff and Fiese report, one can document numerous examples of individuals whose evident risk for disorder was overcome by a nurturant or responsive environment. In effect, at multiple points along a rathogenic trajectory, the introduction of thenecessary and appropriate regulatory process can induce the likelihood of dysfunction. Similarly, the emergence of a required skill (e.g., social competence) byan individual can increase the environment's support and subsequent assistance to avoid disorder.Hence, throughout the process, the potential exists for increasing or decreasing the likelihood of a negstive outcome. It seems reasonable to assume that the accuracy of risk assessment and the potential for determining who will or will not eventually manifest the disorder is a function of an individual's location along the pathogeniccourse. Thus, we hypothesize that one's risk level rdative to the population at largecan be assessed in terms of the number of proximal outcomes displayed. Thus, based on the number of developmental markers passed, one can describe an individual or subgroup as being at increased or decreased risk for an outcome(s) of concern. Thus, the earlier one intervenes in thesequence, the less certain it is that the recipient of one's effort will actually manifest the disorder ofconcern. Converse- ly, the later one intervenes, the greater is the certainty that dysfunction will occur if nothing is done. At the same time, however, by that point, both the individual and those around him or her have already suffered toa considerable extent; and the likelihood of reversing the outcomemay be quite low. Although the relation is not statistically established,we suggest that predictive certainty and preventive potential are inversely related; that is,as one increases the other decreases and vice versa. If accurate, this assumption imposeson prevention researchers the responsibility for selecting recipients of their interventionsin as sensitive and timely a manner as possible. Selection of Intervention Recipients

The transactional perspective makes evident the complexity of predicting specific risk for child and adolescent disorders. The base rateat which such disorders ocnur in the general population ranges froma high of 20 to 25 percent for attention-deficit disorders and learning disabilities (McGuinness1985) to well below 1 percent for childhood autism and schizophrenia (Achennach1982). It appears that childhood depression and alcohol and other druguse fall between these extremes.Given the infrequency of the occurrence of emotional and behavioral disordets, it is both inefficient and potentially riskyto apply proven . tive efforts in a wholesale manner. In fact, insome instances, the major risk faced by the recipients of prevrmtive interventions relatesto the possibility of iatrogenic effects from the intervention itself (Lorion 1987a). Forexample, Gersten et al. (1078) reported that premature (i.e., before itwas developmentally FROM THEORY TO RESEARCH 67 appropriate) application of a secondary preventiveintervention to reduce problems related to anxiety and aggression increasedtheir occurrenee in the experimental population compared with their control group peers.Recognition of the possibility that preventive interventions mayhave unintended negative consequences should lead thoseimplementing such interventions to include in their evaluative deeign strategies for detecting suchunintended outcomes. Appreciation of prevention's iatrogenic potential alsodemands that we ensur that the recipients of such interventions are atsufficient risk to justify their involvement in the preventive intervention. Furtherdiscussion of selection issues is provided in a later section ("StatisticalConsiderations") in the examination of the effect of low base rates on the power ofevaluative designs for preventive interventions. Gordon's Categories of Prevention The importance of targeting preventive efforts toappropriate recipients is underscored by Gorden (1983). He argued that themultiple distinctions between infectious disorders and behavioraldysfunctions justify replacing the public health categories of preventive interventions (i.e.,primary, secondary, and tertiary) with categories reflecting who receives theintervention. Gordon's proposal also highlights the aforementioned factthat preventive interventions can have iatrogenic consequences.In Gordon's view, the risk of such consequences should be consideredalong with the recipient's vulnerability for the disorder in justifying the implementation of any preventiveintervention. Universal Interventions Gerdon labeled as "universal" those interventionsdesigned for reception by all segments of the population. Examples include suchpublic health strategies as clean water regulations,fluoridation, and immunization requirements fur school entry. Other examples include the use of publicservice announcements to influence attitudes and behaviors aboutalcohol and drugs; revision of primary-grade curricula to include training in interpersonalskills; and parenting workshops to reduce child abuse andenhance children's self-esteem. Universal interventions, according to Gordon, are massdistributed and therefore tend to have relatively low cost per unit of service.They tend also to be minimally intrusive and nonspecific and may therefore belimited in their capacity to produce dramatic change. This limitation reflects,in large part, their general fir-us and their need to be acceptable to and safefor wide segments of the population. Frequently the direct effects of universalinterventim on the inckience or prevalence of disorder are difficult to measure. Selected interventions Whereas universal interventions are made available to all,"selected" interventions are directed toward those segments of the populationcharacterized by epidemiologically established risk factors (e, g., offspring of a teenagemother or alcoholic parent). Such individuals are known to be at increased riskfor one or

fr 68 PREVENTION OF MENTAL DISORDERS more identifiable disorders. Knowledge of the intervention recipients' above- average risk level for specific outcomes enables the intervention's designer to sharpen the focus and intensity of the effort. This increased focusmay also increase the intervention's iatrogenic potential, causing some recipients to respond negatively to elements of the intervention. For thatreason, its application is limited to those at enhanced risk.For those, it is assumed that the potential for the preventive intervention to have negativeconsequences is outweighed by demonstrated likelihood that the participants' risk for the disorder of concern will be reduced. Inclusion in selected interventions is based not on characterist:cs of specific individuals but ratheron the presence of demographic or experiential characteristics (e.g., loss ofa parent or assignment to a foster home) associated with increased risk for population subgroups. The goal of such interventions is to respond early enough to reduce the level ofrisk and thereby decrease the incidence of disorder in the vulnerable subgroup. Public health examples of such strategies include targeted antismoking efforts (e.g., for spouses of victims of cardiovascular disordersor parents of asthmatic children) and influenza immunization programs for the elderly.Additional examples include programs for the children of alcoholicor depressed parents, infant stimulation programs for the offspring of schizophrenicor adolescent mothers, and stress inoculationprograms for children scheduled for elective surgery. In effect, selected programs are designed to reduce the occurrence of an identified disorder in a subgroup of the population at increased risk for that disorder. Indicated Interventions

Gordon's universal and selected categories involve interventions targetedto groups rather than to individuals. His third category, "indicated," refers to strategies designed to reverse, in specific individuals,an already initiated pathogenic sequence. Similar to secondary preventive interventions, indicated interventions require the availability of sensitive screening proceduresto identify individuals who are displaying preclinical signs of emotionalor behavioral disorders.Indicated interventions are provided to specific individuals with specific indexes of dysfunction.Because of their specificity and intensity, indicated interventions are likely to have higher potential for iatrogeniceffects than either universal or selected strategies. Gordon argued that thisincreased risk for negative effects is balanced both the immediate needs of program participants and by their epidemiologically established risk formore serious subsequent dysfunction. If successful, indicated interventions addressboth of these needs. Thus, the early signs are alleviated and the underlyingpathogenic process is arrested. An important difference between the public health preventioncategories and those offered by Gordon is the latter's emphasison the intervention's recipients rather than on its intended outcomes. A second difference isits ac- knowledgment of the potential for iatrogenic effects thataccompany preventive

7 ; FROM THEORY TO RESEARCH 69 interventions. Without intending to exaggerate the likelihood ofsuch outcomes, it is important to appreciate the fact that they can occur. Asstated elsewhere: To assume (as opposed to demonstrate) that preventive strategieswill have only positive or, at worst, neutral consequences represents anaive and irresponsible position.Itis inconceivable that an intervention which is designed to avoid or limit the impact of a pathological process or to generate heretofore absent inter or intrapersonalcompetencies could not be recognized as also able to cause negative outcomes(Lorion 1983, p. 252). Attending to the base rate of child and adolescent disorders providesinsight into the issue.If, for example, 6 percent of the population may eventuallydisplay an affective disorder and an intervention toprevent such disorders were applied to the general population, the only risk for 94 percent of itsrecipients is for a negative response to the intervention. Evidence of the reality of such iatrogenie effects in children has been reported by Gersten et al. (1978)and Lorion et a. (1974). McCord (1978) also provided an interesting report on thelong-term consequences of a program to reduce delinquencyrisk in male adolescents. Urging that prevention researchers appreciate the negative potential of their interventions is intended as a plea for caution and not as a call for a moratoium on preventive trials. We believe that suchcaution should be displayed itwo ways. First, as discussed in detail in chapter 3 byPrice and Lorion, evaluation designs should include the capacity to assess both intended and unintended outcomes.Through such research, we expect that negative effects can be identified, understood, and minimized. The second approach to minimizing iatrogenic effects is to gather sufficient epidemiological data about the risk factors associated with emotional and behavioral dysfunctions to allow forthe careful selection of intervention recipients. Ideally, use of risk factor information should result in the identification of potential program participants whose risk for the outcome to be prevented exceeds the population base rate. Obvious- ly, were everything known about the etiology and evolution of the disorder and were one to defer intervention until the occurrenceof a sufficient number of marker signs along the pathogenic sequence, the accuracy of one's selection would increase substantially.In the absence of those options, however, the challenge confronting prevention resc:archers is to identify a sufficient number of risk factors or marker signs so that interventions can be designed that either avoid the onset of or interrupt as soon as possible the pathogenic sequence relevant to the disorder to be avoided. Sources of Ilisk A common pathogenic sequence can precede a number of clinical manifestations of disorder. Concretely this means that certain individual characteristics (e.g., genetically determined vulnerabilities; personality characteristics) and environmental characteristics (e.g., poverty, familial instability, the occurrence of a critical life event) uniquely and in combination can place a person at risk 70 PREVENTION OF MENTAL DISORDERS for emotional and behavioral disorder. The factors that determine how and when the disorder will manifest itself are, for the most part, unknown. The transactional perspective, however, provides a framework within which such information can be acquired and understood. It does so by identifying both potential sources of risk and potential mechanisms whereby the risk factors contribute to the development of pathology. Within Sameroffs framework, the classic public health triad of agent, host, and environment are important elements of the concept of risk. The first of these, the agent, refers to the specific cause of the disorder. The transactional model asserts that potential causes can include characteristics of the individual, of the environment, or of their combined influence. Examples of characteristics of the individual that cause or contribute to pathology include genetic and congenital factors related to such disorders as mental retardation, autism, schizophrenia, attention-deficit disorders, specific learning disabilities, affective disorders, and alcoholism. Also included within this category of risk factors are characteristics, in part physiological and in part psychological, related to temperament, tolerance of stress, intellectual and interpersonal skills, and perceptual acuity. A second source of risk relates to characteristics of the environment thatcan contribute to emotional and behavioral dysfunction. Examples include personal (e.g., an abusive parent) and impersonal (e.g., an auto accident)causes of physical injury, particularly to the cortex; intense stress froma demanding physical setting; characteristics of the familial and peer environment; and the experience of significant life events (e.g., loss of parent through deathor divorce; moving to a new city; or having a chronically ill sibling). The third source of risk involves the consequence of combinations of individual and environmental factors. Examples include havinga schizophrenic or depressed mother and losing one's father through death; being highly sensitive to stress and the offspring of a highly mobile family; havinga visual perceptual impairment and being assigned to a classroom in which the primary means of information exchange include reading and working at the chalkboard; living in a neighborhood that associates masculinity with athletic skill and having poor motor coordination. The transactional paradigm hypothesizes that behavioral and emotional status at any point in time represents the consequence of prior and continuing synergistic interactions between individual and environmental characteristics. As a consequence, it must be recognized that in manycases an individual's risk level varies over time. As Bell (1986) accurately remindedus, depending on individual, familial, and environmental circumstances,a child may be protected against or vulnerable to emotional and behavioral disorder. Hencewe must design our intervention and evaluation techniques with such risk variAility mind. We must also appreciate the fact that in some instances, the influence of a risk factor on adjustment may be direct. For example, genetic or congenital events will result in mental retardation or an organic brain syndrome, Inmost

) FROM THEORY TO RESEARCH 71 instances, however, the extent of dysfunctionand disability, if any, depends on the sequence of synergistic events that lead up toand follow the risk factor. Thus, for example, research findings reportedby Garmezy and Rutter (1983) inform us that certain developmental experiences serve aspositive risk factors or buffers, which decreaseone's vulnerability to stressful life eventsand reinforce one's psychological resilience and capacityfor adaptive coping. Similarly, the availability of social support during andafter the occurrence of a risk- producing event has important implications for itspathogenic impact (Heller et al. 1986). In fact, Thoits (1986) proposed that"social support might be usefully reconceptualized as coping assistance, or the activeparticipation of significant others in an individual's stress managementefforts" (p. 417). The identification of risk factors represents amajor challenge for the mental health research disciplines. If obtained, such information mayprovide important insights into underlying pathogenic processeeand allow for the usto of what Gordon (1983) labeled as "selected" interventionstrategies. As the criteria for the selection of program recipients increase, sotoo does the power of its evaluative design. This latter point is discussed morecompletely in the section entitled "Statistical Considerations." Case-Control Study An important epidemiological strategy for riskfactor identification is the case-controt, study.Also commonly called a retrospectivestudy, this study follows a paradigm that prock9eds from effect to cause.In the typical case-control study, individuals with a particular condition ordisease (i.e., the "cases") are selected for comparison with a ser;es of individualsin whom the condition or disease is absent (i.e., the "controls"). Casesand controls are compared with respect to existing or past attributes or exposuresthought to be relevant to the development of the condition or disease understudy (Schlesselman 1982). The logic of the case-control study is toproceed from an outcome, effect, or disease to an efficient and broad search forantecedents that may be causes. When little is known or agreed on about thedisease except that it is to be avoided, his manner of exploration is particularlylogical. The second important part of the logic is the use of the controls toestablish a base rate of frequency for comparison. The case-control study isefficient particularly in rare diseases, because cases can be located through clinics orhospitals. An important result of the case-control study, as well asother epidemiological strategies, is the determination of "attributable" risk.Attributable risk is "the maximum proportion of a disease that can be attributedto a characteristic or etiologic factor; alternatively, it is considered theproportional decrease in the incidence of a disease if the er. Are population were nolonger exposed to the suspected etiological agent" (Lilienfeld and Lilienfeld1980, p. 217).At- tributable risk quantifies the importance of individualrisk factors in a manner that is easy to understand and allows us to rankthem in terms of importance. For example, significant early case-control study onsmoking and cancer showed 72 PREVENTION OF MENTAL DISORDERS that about 35 percent of the lung cancer cases in the United States wouldnot have occurred if no one smoked. The case-control study may not be as well adapted to the transactional paradigm as it is to epidemiologic studies in the medical paradigm. Forone thing, the medical paradigm works best in thepresence of a single clearly definable outcome. In the transactional paradigmwe are typically interested in more than one outcome, and part of our research effort is to learnmore how to define the concept of outcome. The ease-control study allows searching throughout the entire life of the individual before the study, andn this sense it is very efficient for an exploratory endeavor. Butany attempt to link variables over time in the type of synergistic interactions that are centrally embedded in the transactional paradigm is awkward. This restricted abilit search for and inz.lei temporal associ2tions is a crucial weakness for the case-control design in developmental studies. Transactional Sequence The transactional perspective has important implications for risk-factor research. Implicit in this paradigm is the recognition thatone must simultaneously examine individual, environmental, and transactional elementsas pmcursors t- amotional and behavioral disorders. The paradigm also suggests that one can concepter' -.43 risk exclusively in terms of one's potentialfor the end-state condition of .....erest or in terms of one's potential for experiencingone or more of f:ie proximal or marker outcomes defining its pathogenic path. In other words, the accuracy with whichwe choose recipients of our preventive interventions may be increased if we use early tisk factors, be they demogreahic characteristics or "barely identifiable behaviors" (Long 1986,p. 327), as predictors of subsequently appearing precursors.If, as proposed by Sameroff and Fiese, maladaptive states evolve sequentially, the paradigmsuggests that end-state conditions will be avoided through replacementor alteration of an otherwise occurring pathogenic sequence.Not unlike the child's game of hopscotch or the game of chess, our successmay depend on our capacity to anticipate 4he next one or two jumps. The foregoing conceptualization of risk-factor analysis imposesa heavy burden on the mental health research disciplines to expand their understanding of the natural evolutionary history of emotional and behavioralfunctioning. Such a knowledge base would, in fact, representa specific focusing of mental health research. Such research appropriately fits within the domainoi developmental psychopathology (Achenbach 1982). Rather than examine howdisorder manifests itself at varyir g developmental stages, however, this fieldshould also consider the pathogenic course by \ one proceeds from normative to dysfunctional emotional or behe7ioral84...au& In essence such rese rch would identify the sequence of stages through which disorder e7olves end,relatedly, should inform us immensely about the etiological paths precedingspecific dysfunctions. Research can assist in understand iag themechanisms whereby certain common antecedents serve as predecessors for multinledysfunctions.

5: 4 FROM THEORY TO RESEARCH 73

We propose ",at such research emphasize theidentification of the measurable manifestat that define the nature and sequence of thedevelopment of psychopathurogy. Thus the defining purpose of suchresearch would be to delineate the common and unique components of thetransactional sequences leading to and maintaining normal and disordered states.In Kuhn's (1970) term, its focus would be on the "work of normalscience," that is, on demonstra- ting ad. enhancing the paradigm's capacity to explainthe phenomenon of interest (i.e., human behavior). Of course, such sequences will not be delineatedquickly. The design of preventive interve.stions, however, does not (indeedcannot!) have te await completion of that task.Preventive interventions must be designed, implemented, and evaluated' based on the best availableinformation. In our view, the conduct of such interventions, in fact, cancontribute tc the development of the requisite data base for an informed science ofbehavioral pathogenesis. In their work with delinquents, Loeber et al. (1984)provided a useful outline for planning prevention research at this stage of our knowledge: One can think of prevention as a three-stage process:the first step is the identification of etiological variables, especially thoselending themselves to change; the l'acond step is to use the variables toidentify children at risk for deliuquency or antisocial life styles; and the third stepis the implementation of intervention strategies designed to change theetiological variables, thereby reducing the child's risk of engaging in acriminal or antisocial career (p. 9). Our discussion thus far has emphasi7ed the first step, thatis, identification of etiological factors. As noted, epidemiological strategiesoffer an assortment of methodologies for acquiri g such information. Itsavailability both informs us about the etiology ande---olution of the disorder of interest and provides indexes for step 2, that is, the selection of those at risk for suchdisorders. The use of such indexes is necessaryfor the application of the interventions belong- ing to Gordon's (1983) "selected" and "indicated" categories. Application of ILtIrventions Selected Interventions The first of these categories, selected interventions, isapplied to participants chosen because they share epidemiological!v identified characteristics as. sociated with the occurrencei the disorder(s) of interest.Thus, a child or adolescent is targeted because he or she belongs to ademographic subgroup in which the occurrence of the emotional or behavioral disorder tobe prevented exceeds the populati'm base rate. Examples of such indexesinclude being the offspring of an adoleecent mother (Achenbach 1982), living in afamily that has experienced or is abou:: to experience marital separation ordivorce (Sbillberg and Garrison 1985), being scheduled for pediatric surgery(Tadmor et al. 1986), experiencing a critical life transitim (Felner 19'34), or having a parentwith a history of xhizophrenia or affective disorder (Rutter in press). 74 PREVENTION OF MENTAL DISORDERS

Selected interventions are applied to individuals only because they belong to a specific population subgroup. As such one cannot assume that any specific individual is at risk; only that members of the subgroup are at increased risk to experience the disorder or dysfunction. Moreover, it should be recognized that some portion (the majority, perhaps) of those chosen for the intervention would not experience the disorder even without the intervention.By definition, selected interventions do not differentiate which members ofa subgroup are themselves at increased risk relative to other subgroup members. Improvement in our ability to maximize the number of high-risk individuals in our intervention samples becomes possible, however, by combining risk factors, especially independent risk factors. Specifically, for example,we know that children of depressed parents are at increased risk for experiencing depressive symptomatology. Does that risk increase if the child lives with butone parent? If so, does it matter with which parent the child lives? Is the risk increased even further if the child is a member ofa minority group, economically disadvantaged, and with a limited number of extended family members available? Responses to such questions will,we believe, significantly improve the identification of those at significant risk and thereby minimize the iatrogenic potential of applied interventions. Obtaininganswers to such questions requires the design and conduct of a series of epidemiological case-controlstudies examit.ing each of these possibilities in a systematic fashion. An important byproduct of such research, we believe, will be heuristically valuableinsights into the etiology of the disorders and dysfunctions under study. Indicated Interventions As noted, indicated interventions focuson specific individuals who are appropriate for participation in a preventive intervention because of Voepresence of one or more signs of an ongoing pathogenicsequence. Such individuals are typically identified through the use of screening procedures. Thenumber of such instruments available for children is quite extensive,as reflected in compendia such as those by Johnson and Bommarito (1971), Johnson(1976), and Walker (1973). Among the widely usedmeasures for children's disorders are the Denver Developmental Screeni .g Test. (Frankenburg and Dodds 1967); the Child Behavior Check List (Achenbach 1979); andthe Coopersmith (1967) Self-Concept Test. Such strategies allow for the efficient identification of patternsof behavior and emotional expression displayed by children who subsevently experiencediag- nosable disorder or serious dysfunction. Again, however,one is confronted with the fact that the predictiveaccuracy of such measures is less than optimal. Thus, for some at-risk children, the major risk factoT they face involves theiatrogenic consequencee of the preventive intervention. Gersten et al. (1978) argued that the latter risk increases whenone does not appreciate that in many cases the frequency with which "early signs"appear in the general population exceeds the rate at which disorder subsequentlyoccurs in the populations measured. This occurs simply because many such signs naturally dissipate withage. FROM THEORY TORESEARCH 75

the efficiency of Loeber et al. (1984) proposed auseful strategy for combining procedures with the increased accuracyof individualized assess- mass screening each ment procedures. Theirsolution involves use of a seriesof multiple "gates," maximize the risk level ofidentified individuals. Todo of which is designed to need of so, each gate isdesigned so that the numbernf individuals deemed in with delinquency-prone the intervention decreases.The work of Loewer et al. adolescents provides a goodexample of this procedure.Initially a teacher rating 1). Children measure was usedto identify children atrisk for delinquency (gate ratoff on this rating scale wereidentified as at scoring beyond a predetermined however, the inves- high risk. Rather thanassip them to an intervention, tigators used this identificationto select childrenwhose parents (typically the in gate 2, i.e., a phoneinterview about the child's mother) would then participate home. Based on such responsiveness to parentalrestrictions and behavior at interviews, Loeber et al.identified a subgroup ofadolescents meeting risk adolescents were then criteria on both gates of thescreening procedures. These interview lesigned invited to participate individuallyin a structured diagnostic The interview to detect specificallyselected signs of delinquency proneness. responses werethen used to select interventionrecipients. procedure of Loeber et al.improved the ac- The multiple-gating screening 24.5 to curacy of theirefforts to identify children atrisk for delinquency (from 56.3 percent). As discussedin the section on statisticalconsiderations, this research design will, in increment markedly increasesthe likelihood that one's effect. simultaneously, itenables the prevention fact, detect an experimental and program implementerto shift from a selectedto an indicated strategy thereby justifiably to increasethe intensity and specificityof the intervention procedures. Having documentedthe presence of certainpathogenic signs in specific individuals through theinterview, the program designers canproceed with increased certainty aboutthe participants' need forthe intervention. Loeber et al. The transactional frameworkallows for a modification of ;:ie relatively brief procedure. Rather than applyingall gates sequentially within a period, it is possible to coordinatetheir application with theanticipated appearance of specificmarker behaviors along thepathogenic sequence. For example, using archival records onemight identify a sample ofchildren experiencing anoxia and otherprenatal and perinatal complications.One would also attempt to assess environmentalfactors such as the parents'social cultural, and economic status; thestability of their marriage; andtheir capacity to tolerate the stress associatedwith raising a child atdevelopmental risk. Upon entry to nursery school,this sample may be screenedusing parent and teacher ratings to assess each child'sperceptual, motor, cognitive, andinterpersonal the family and its development.Simultaneously, one would again assess response to the child.Children or families displayingdeficits in one or more of these areas may be selected forsubsequent screening prior toschool entry. At that time, children continuingto display developmentaldeficits can be individually assessed with acomprehensive battery of procedures toidentify children at enhanced risk forattention-deficit disorder and specificlearning n r- 76 PREVENTION OF MENTAL DISORDERS disabilities (e.g., Lorion et al. 1987; Lorion et al. 1984).Assessment of familial variables and, at this point, of the parent-child interactioncan provide important information relevant to the determination of risk and the designof an intervention. A similar assessmentsequence might be developed to monitor children's psychosocial development, display of aggression,or involvement in predelin- quent behaviors. We hope that for many emotional and behavioral disordersit will become increasingly possible to identify the chronologythat defines the relevant pathogenic sequence and to design asseesmentsequences that monitor children's progress along those paths. Through theapplication of such ongoing monitoring, we believe that prevention reeearchers will finda solution to the aforementioned base rate problem, select with increased precisionthe recipients of their interventions, and learn important information aboutthe etiology and pathogenic sequences their interventionsare designed to influence. Timing of Interventions Finally, it is important that we underline the relationship betweenthe causal or etiological chain that defines the pathogenic process to be avoidedor altered and the choice among potential risk factors. Thetransactional model argues that, for the most part, emotional and behavioral disordersresult from complex sequences of transactional exchanges between the individual and the environment. An individual's probability of experiencing the outcome ofconcern varies at different points along suchsequences. If we identify those points along the sequence where the likelihood of the negative outcome increases sharply,we would have important clues about when to intervene.Similarly, given the multitude of factors that contribute toan individual's risk at each point along the sequence, identification of those points at whicha limited number of factors are most salient provides insight into which factors to address. At suchin- stances, it is quite likely that the attributable risk assignedto those factors will increase sharply. These two functions, that is,an increase in the level of risk and the emergence of a limited number of risk factors, will potentiallycoincide in time. If so, theee may be the critical points at which to apply preventive interventions. Itseems reasonable to assume that such interventions would betargeted to effect the aforementioned highly salient risk factors. We predictthat such targeted applications would have a high probability forsuccess. As noted previously, however, we must caution that the transactionalmodel does not presume continuity of risk over time. As the pathologicalor normative iature of the transactions changes, so too does individual risk. Formost of us, "being at risk" is a state condition determined by the nature oftransactional relationships between the individual and the environment. FROM THEORY TO RESEARCH 77

Analysis of Intervention Components

Preventive Thais At this point it is important to distinguish the intentof this section from that of chapter 3 by Price and Lorion. Here we examineissues relevant to the outcomes of preventive trials. We assume that suchtrials are conducted during the development of a preventive intervention andprovide the empirical basis for its subsequent application in the general community.Preventive trials focus on such questions as the effectsof the intervention on its recipients; the differential effectiveness of its components; the differential responseof population subgroups to the intervention; and the assessmentof the intervention's temporal characteristics (e.g., latency between theintervention and the appearance of its effects; duration ofthose effects; requisite duration of involvement in the intervention). By contrast, the Price and Lorion chapter discussespreventive applications the dissemination of demonstrably effective interventions todiverse populations in a variety of organizational settings andconditions.These applications confront those interested in the adoption of viable programswith a set of challenges distinct from those addressed by the developer of anintervention. Specifically, an application typically must be justified with the resultsof a needs assessment documenting the target population'sappropriateness for the intervention. Second, it is necessary to design, conduct, andanalyze participant selection procedures. Such procedures usually involve sume formof risk assessment screening using either demographic or individualindicators. The application of an intervention also involves the development ofprocedures to monitor the fidelity with which the original program components arereproduced in subsequent applications.In many cases, information relevant to program fidelity is collected simultaneously with other relevant managementinforma- tion (e.g., which staff deliver which 3ervices to whom underwhat conditions). Numerous management information systems exist that can servethis purpose. Finally, most applications will require some form of at least periodicdetermina- tion of the intervention's eermomic costs and achievement ofintended goals. As noted, peeventive trials are conducted during the developmentof an intervention. By necessity such trials are field based and require thatthe researcher be sensitive to the unique demands of the settingsinvolved in the trial. Discussions of such demands are provided by Cowen (1978);Cowen et al. (1974); Lorion (1978, 1983); Munoz et al. (1979); Price and Smith (1985);and Price et al. (1980). Technical considerations of the methodologicaldemands of such research are provided by Amabile and Stubbs (1982); Cookand Campbell (1979); Fairweather and Tornatzky (1977); and Selltiz et al. (1976). Interestod readers are encouraged to review these .3ources as they plan the designof their preventive trial.

(C.-3 78 PREVENTION OF MENTAL DISORDERS

The conduct of preventive trials requires theconsideration of multiple issues. Assuming that the intervention protocol hasbeen selected, the researcher must design the evaluative study to allow for the detectionof its proximal (immediate) and distal (long-term) outcomes (Price 1982).By necessity, this requires that the design be longitudinal and involvemeasurements at multiple points in time. Data collection procedures must there. allow for the monitoring of the processes under study throughout the evaluative period andbe capable of differentiating change resulting from theintervention from change due to maturation. Within the framework of "true"experiments, in which there is random assignment to experimental andcontrol conditions, the analysis of causal links is reasonably direct. Under fieldconditions, however, it is frequently necessary to conduct research withoutthe rigor of a randomized control design.In such instances, "quasiexperiments"are necessary to obtain reasonable confirmation of one's inferencesabout causal relationships (Campbell and Stanley 1963; Cook and Campbell1979). Maturational factors add to the complex psychometricchallenges confronting those whose interventionsare targeted to young children (e.g., during the preschool years) in the hopes of avoidingemotional and behavioral disorders in the preadolescent, adolescent,or adult years. For example, measures currently exist that allow for the continuous monitoringof relevant emotional, behavioral, or interpersonal functioning over suchan extended period. In part, this reflects the multiple qualitative metamorphosesthrough which phehomena suchas depressive affect, anxiety, feelings of rejection,self-esteem, and interpersonal competence pass with maturation. It also reflectspast difficulties in obtaining support for and carrying out the requisite long-termscale development studies. We hope that one byproduct of theincreased value placed by Federal policymakers in recentyears on preventive efforts (Department of Health and Human Services 1984, 1986) will bea resurgence of interest in and support for such studies. In the interim, researchers must carefullyselect a set of measures that provide the closest approximation to sig-Afficantpoints along the developmental path of interest. For example,a succession of different measurementprocedures (e.g., observation, analogue situations,teacher ratings, peer ratings, and self-ratings) may benecessary to monitor interpersonal effectivenessfrom preschool through adolescence. Similarly,anxiety may be assessed using ob- server ratings, physiological measures, and self-ratings aloneor in varied combinations throughout the periodof study. Each element of the evaluation battery must itself meet minimalpsychometric criteria. Moreover, throughout the evaluation period, the replacementof one measure for another must be scheduled to allow for overlap of bothprocedures. Such overlap will enablethe researcher to document the degree ofinterrelationship among themeasures. Interpretation of such correlational findings,of course, will both add to and depend on the adequacy of the constructvalidity of both measures. FROM THEORY TO RESEARCH 79

Verification of the interventions'preventive effect is likely tobe expected within relatively short increments ofat most several years.This fact reflects both the reluctance of funding sourcesto provide support withoutregular documentation that their intended purposes arebeing achieved and the impor- the tance that preventive interventionsbe continuously monitored to ensure appropriateness of their spplicationto their recipients.Consequently, the evaluation of effectiveness islikely to occur in a stepwise mannerin which theoretically determined "marker"points are selected.Identification of the sequence of markerpoints depends, of course, onthe adequacy of one's knowledge of the temporal course orpathogenesis by wit h disorder evolves. do the At each such point, three questionsshould be asked: (a) To what extent intervention and nonintervention(i.e., control) groups differ? (b) Towhat extent does the nonintervention groupcontinue to display evidence of being atrisk for the disorder or dysfunction ofinterest? (c) What, if any, iatrogeniceffects can be associated with the intervention? By repeatedly asking such questionsat each marker point, theresearcher should be able to chart the differentialdevelopmental paths of the intervention and nonintervention groups.Some of the latter group's pathwill presumably reflect a relatively uninterruptedpathogenic sequence, whereas n Jetof the former's will increasingly approximateage-appropriate patterns. For ahighly effective intervention by the end ofthe intended followup period,the nonintervention group's rate of evidenceddisorder or dysfunction will approximatethe level originally anticipated onthe basis of available epidemiologicalevidence Confirmation of each of thoseassumptions constitutes evidenceof the intervention's preventive effectiveness. Since intervention is rarely equallyeffective for all who receive it (Paul1967), examination of the factors thatdistinguish who does and does notrespond positively is an important programdevelopment step. By "respondingpositive- iy." we mean ren aaling alower-than-exrected risk for the unwanted outcome. The value of this step lies in theinformation it provides about necessary revisions in program elements, inparticipant selection criteria, or both.An efficient way to obtain such informationis to search for individual orenviron- mental variables that define groupswherein the effects of the intervention are found to be very strong or veryweak, Through 'le subsequent sequenceof comparative analyses, one is able to generate aseries of hypotheyes about the demographic and individual characteristicsef those for whom the intervention is most appropriate. By conducting aseries of studies that examine these hypotheses specifically, the programdeveloper is able to refine systematically the criteria by which intervention targets areselected. Process Studies Just as not all recipients respond equallyto an i, tervention, not all program components contribute equally to theachievemeet of intervention effects. Process studies are a means by which oneattempts to differentiate effectiveand ineffective program components.Ineffective components include those rare 0 80 PREVENTION OF MENTAL DISORDERS

instances in whichone or more procedures have negativeconsequences and those more frequently observedinstances in which proceduresare superfluous. As discussed in chapter 3, thoseresponsible for the delivery of preventive interventions are continuously underpressure to minimize the costs of such efforts and the additional burden theyplace on institutional staff.This is particularly true when the intervention'sprocedures must be incorporated within a setting's already demandingschedule (e.g., a classroom, daycare center, pediatric surgical unit,or social service program for the placement of foster children).For that reason, one must be especiallyconcerned about imposing any unnecessary demandson the staff of such settings. Several strategies allowone to examine the differential contribution of program elements. The most sophisticated of theseinvolves the random assignment of program recipients toone of multiple groups, each of which involves specific combinations ofprogram elements. By subsequently contrasting the 1-esponse ofeach subgroup to these various combinations,the program developer is ultimately able to determine,on the basis of scientifically derived information, the optimal program design.The format for such a study is comparableto that used in the comparative evaluationof therapeutic procedures, that is,a clinical trial. A second strategy for identifyingeffective program elements involves the careful documentation of the specificprocedures used with eachprogram recipient. These recordscan then be systematically examined and different combinations of program elements identifiedon a post hoc basis for comparison. Although this procedurecan be a useful means of generating hypotheses about which combinations of proceduresappear most promising, it cannot providean unambiguous basis for sucha conclusion.Its post hoc nature and related absence of random assignment ofrecipients to alternative conditions makesit impossible for one to determine preciselythe factors that contribute to observed differences. For thisreason, this strategy should be used sparingly andits findings appreciatedas tentative. A third procedure to determine theoptimal combination ofprogram elements involves the sequential introductionof intervention procedures. Thirstrategy ii applicable in three instancesin which the number ofprogram recipients available is not adequate to allowfor the simultaneous comparisonof multiple program strategies.In planning the sequential approach,the program developer needs to identify the intendedoutcomes of the intervention precisely. He or she must then identify theminimal num!.er ofprogram elements assumed necessary to achieve these objectives. This set ofprocedures will constit ate the ieitial intervention approach.If the intended objectivesare achieved, the program developer may decide that the task iscomplete or that superfluous procedures shouldnow be identified. In the lattercase, the inclusion of some form of management informationsystem or service provider's log in theinitial preventive trial can provide the requisiteinformation for identifying procedures to eliminate in subsequent trials.If desired effecte continue to beobserved, it FROM THEORY TO RESEARCH 81 can be assumed thatthoee procedures were, in fact, unnecessary.By contrast, if the consequences of the secondintervention's procedures are lessthan desired, the program developer is likely toreintroduce the withdrawn procedures. A similar iterative procedure iscarried out if the initial preventivetrial fails to produce the intended outcomes.In such an instance, the task confrontingthe program developer is toidentify precisely which outcomes were notachieved and then design additionalprocedures expected to achieve thoseoutcomes. Throughout this sequential development process,the originally intended effects must be constantly referenced asthe baseline against which allinterventions are compared.Ultimately the developer will be requiredto select the set of procedures that simultaneously provideethe closest approximation to the desired outcomes, is likely to be acceptable,and can thereby be reliably adopted within the setting. The simultaneous consideration ofdiscrete program strategies represents the most efficient andscientifically valid approach. Other strategies reviewed represent at best limitedapproximations of the desired program evaluation approach. They enable programdevelopers to test their best guesses about what works and does not work.They also provide a means by which program can be developedwith limited resources and within therealities of the pressure associated with a settingcommitted to involving iteelf in preventive efforts now. Monitoring Programs Those who cifign preventive interventionsshould appreciate that the level of developmen t of an intervention willinfluence its accep:ability within a setting and the fidelity with which its procedures areimplemented. In their discussion of the distinction between "manifest"and "true" adoptions, Rappaport et al. (1979) explained how agency staff respond tothe mandate that program procedures must be followed differentlyduring the development and application phases. Specifically, as the interventionbecomes an institutionalized part of a setting's activities, staff may feel increasinglycomfortable modifyieg the program elements that theybelieve either need to be changed or must besacrificed in reeponse to a newlyintroduced demand for their time. It isincumbent on those responsible for the application of preventiveinterventions to appreciate this understandable tendency of staff toshape activities in their own way. For that reason, program developers shouldincorporate procedures to monitor program fidelity periodically.Program developers must recognize that if a program is to achieve itsintended outcomes, its procedures mustbe followed. Thus, those procedures must be kept as simpleand clear as feasible. A related distinetion between evolving andestablished preventive interventions involves tho attitudes of those responsiblefor service delivery. Occasional- ly, staff involved in ',,he early stages of programdevelopment may be highly skeptical and unhappy that they must "wastetheir time" in this manner. Support for an intervention among program andinstitutional staff is usually 82 PREVENTION OF MENTAL DISORDERS

not automaticand, in fact,unproven interventions are rarely imposedon unwilling staff. More typically, volunteersare sought who are interested (for their own reasons) in contributingto the development of an intervention that they feel is "long overdue." Theirconcern with the emotional or behavioral need addressed by the intervention frequentlyresults in their enthusiastic participation in program development efforts.The influence of such enthusiasmon identified program effects is difficultto assess directly.It will, however, manifest itself in subsequent applicationsthat seem unable to replicate the original results. On the other hand,demonstrable effectiveness of the intervention gives it credibility, which is likelyto have its own consequenceson program outcomes. Although we still know far too littleabout such aspects of program adoption, an appreciation of their potentialrole in the application of "experimen- tar proceduresmay prepare us for unanticipated difficulties. Statistical Considerations

Thus far we have examined the methodologicalimplications of applying to prevention efforts the transactional modeldescribed in Sameroff and Fiese's chapter. We believe that this model offersprevention researchers a paradigmatic framework within which to conceptualizethe developmental nature of preventive efforts.Rather than simply comparing thepreintervention and postintervention status of participants andtheir nonintervention controls,we encourage prevention researchers to design studies thatfocus on the distinct developmental paths experienced bythese groups.By definition, effective prevention efforts should change thedevelopmental histories of those who receive them. In our view, documentation ofsuch changes definesa successful preventive intervention. Power of Experimental Procedures Numerous statistical challengesare associated with the analysis of such developmental changes. Firstamong these, we believe, are the implications of the preventive paradigm for thepower of the experimental procedures. Tech- nically, power refers to the potentialof the scientific procedures being usedto detect a "true" effect. In thiscase, the question is whether the experimental design being used is capable of detectinga genuine preventive effect. If so, the design will enable us to recognizereductions in the prevalence of the disorder or dysfunction of interest and attribute, withinthe limits of scientific certainty, that reduction to the intervention.Hence, power refers to the design's potential for ruling out alternative explanatoryhypotheses regardingconsequences observed followingan intervention. Several factors contribute to thepower of any research design. Firstamong these is alpha, that is, the level ofstatistical significance to be attainedbefore a label of "effective" is deemed justified. Typically,this level is set at p< .05, which means that in only 5instances out of every 100 studieswill one Ri FROM THEORY TO RESEARCH 83 inaccurately conclude that an interventionhad an effect when in fact it didnot. Since power is related to aipha, it ispneaible to increase the power of ourdesign by accepting a less stringent levelthan .05 (e.g., p < 10 or .20).In so doing, the investigator increases thelikelihood of erroneously concludingthat an intervention is effective when, in reality,it is not. Were there no costsassociated with disseminating an ineffectiveintervention, one might concludethat there is little to lose by adopting thehighest alpha level possible. Asnoted, however, all such programs tend to havecostswhether economic, ideologic, or iatrogenic. At this point in their history,peevention efforts in mental health cannot afford to be representedby ineffective interventions.As we strive for acceptance and credibility in our ownranks (e.g., Cumming% 1972;Lamb and Zusman 1979, 1982; Sanford 1965),proclaiming minimally usefulstrategies as models of prevention's potential can onlyweaken our position in the long run. Thus, we urge readers to exercise extremecaution in adopting an alpha above the p < .05 level, Sample Size A second factor closely related to the powerof our experimental procedures involves the size of the sample used toexamine the intervention's effects. Typically, one assumes that to maximize power onemust use the largest y assible sample. To determine the sample size necessary, oneneed only refer to one of many availablestatistics texts (e.g., Hays 1981, p.252) to secure the requisite formula for estimating sample size.Prevention research, however, must confront the complexity added by the lowbase rate at which disorders occur in any intervention recipients. The problem isthat one cannot determine with certainty the actual required size ofone's sample. The reason for thisuncertainty is the base rate problem discussedearlier. By definition, no individualparticipat- ing in a preventive intervention is anactual "case." Rather, the individual's inclusion in the intervention reflects either anepidemiologically defined level of risk for experiencing the disorder(s)ofconcern (i.e., to use Gordon's 1983 term, for a "selected intervention") ordemonstration of one or more precursors ofthe disorder(s) (i.e., for an "indicetedintervention"). In either case, it is evident that thetotal sample of individuals receiving the intervention does not represe lit the numberof individuals potentially responsive to that intervention. That nemberis, we believe, a function of the base rateat which one expects the disorder ordysfunction to appear in the population from which the intervention's sample was selected.For example, if our intervention sample includes a total of 100individuals and the population base rate is 6 percent, the power of our designwill be significantly less. Power isreduced as the proportion of"cases" to the total sample decresees. The logic underlying the assumption justpresented is as followsFirst, the statement that a population has anepidemiologically derived base rate for some disorder of, for example, 6 percent reflectspopulation rather than individual risk.Therefore, one might conclude that 94 percentof the members of that population are not likely to manifest the disorder.One must then conclude that 84 PREVENTION OF MENTAL DISORDERS

the base rate also represents the proportionof an intervention sample that has the potential (by virtue of its risk for experiencing thedisorder) for responding to the intervention. Because the base rate merelyinforms us of the proportion of a population or sample likely to manifest disorderand does not identify specific individuals at risk, it isnecessary to apply our interventions to many more individuals than are expected to display the disorder. Yetin attempting to assess the effects of our intervention,we must appreciate the fact that the effects that appear, if any, will be reflected ina reduction of the base rate below the 6-percent level.Thus, with any given sample the availablenumber of individuals whose developmental pathscan be altered in the predicted direction cannot exceed the base rate multiplied by the sample size. The validity of this reasoning has seriousimplications for the design of prevention research. In effect, it necessitates thatwe apply our interventions to very large samples of individuak, (toensure that an adequate number can respond to the procedures). An alternativeapproach would be to refineour risk assessment procedures (e.g., using the multiple-gatingprocedure described earlier) so that we can increase substantiallythe selectivity with whichwe recruit participants and thereby the baserate at which we anticipate the occurrence of disorder in the samples used. Eitherway, the demand on the prevention researcher is clear: Because onlya portion of our sample is likely to manifest the disorder and therefore has thepotential to respond to our inter- -ention, we must design our research accordingly. Developmental Continua Base rates are rarely available fora number of points along two relevant developmental continua thatwe feel are important in attempting to determine the minimal number of subjects needed fora preventive trial. The first of these continua involves the pathogenic path describedearlier. We proposed thet the individual's level of risk for displaying thedisorder of concern dependson where along that continuum the individual isat a given point in time. Presume oly, the more precursor steps heor she has displayed and the more markers heor she has passed, the greater the risk.If this assumption is valid, theaone approach to maximizing thepower of one's intervention study is to apply indicated procedures.In that way. one is likely to include inthe study individuals whose level of risk exceeds that of thepopuiation from which theywere drawn. The exact size of that risk level, however,will be unknown until the completion of the study. At that time, it should bepossible to estimate that level by examining the prevalence of thedisorder in the control sample and comparing it to that of the population. The second relevant developmentalcontinuum is that involving the lifespan. Rarely is this factor considered in availablebase rate data. Specifically, it refers to the developmental status of the childor adolescent for whom we are attempting to estimate risk. Obviously chronologicalage is one relevant index. Equally relevant is the child or adolescent's levuof cognitive, emotional, and psychosocial maturation. As we increaseour understanding of how such developmental ; ti FROM THEORY TO RESEARCH 85 characteristics interact with demographicfactors in the determinationof individual risk levels, we willachieve corresponding increasesin our euiiity to accurately select recipients for ourinterventiom Attrition Before leaving the topic of sampling, wemust emphasize that oneof the most underestimated threats to validinference in the conduct ofpreventive trials is attrition in the experimental trialsample or control group.At first glance, it the primary problem associatedwith attrition has to do with may appear that of attrition are much loss of sample size andstatistical power; but the problems more complex anddifficult to resolve. Sample attrition can occur atmultiple points in a preventivetrial for either the experimental or control group.For example, attrition can occur(a) at the during the initial point of entry or recruitmentinto a preventive trial, (b) assessment, (c) during theintervention itself resulting in a"low dosage" intervention, or (d) at any of anumber of followup points inthe assessment of apparent when we outcome. The magnitude ofthese difficulties becomes more recognize that an initial sampleof 1,000 children with a20-percent attrition rate at each of four observationpoints will yield a sample ofonly 512 children at the final observationpoint. However, the problems onlybegin with loss of sample size.As Cook (1985) observed, other forms of attrition arelikely to create biases in theoutcomes of preventive trials. For example,in smoking preventiontrials, students who are most likely to be heavysmokers are leabt likely to stayin a particular study irrespective of the treatment towhich they were originallyassigned. We may describe this as risk-correlatedattrition, and it also meansthat preventive interventions with this kind ofattrition problem will be testedwith populations external validity of the who need them the least.This produces a threat to the trial. There are even more problematictypes of attrition with whichto deal in preventive trials. For example,treatment-correlated attrition is a majorprob- lem in some preventive trials.To continue the smokingexample, students who drop out of a study experience smoking preventiontreatments are more likely to when compared to studentsin a no-treatment control group(Cook 1985). Though the reasons for this areunclear, it is more likely thatthey drop out of the subsequent followupobservations rather than out ofthe prevention treatment itself. This causesmajor problems in analysis.The analytic methods required here involve analyzingthe trial as a quasi-experiment orlongitudinal survey rather than a truefield experiment, reducing thecertainty of causal inference considerably. A final form of attrition has todo with hard-to-reach populationsthat become increasingly difficu1t to reach overtime . For example, inner-cityminority 86 PREVENTION OF MENTALDISORDERS

children from low-income familiesmay, for a variety of reasons, experience increased mobilityor less strong attachments to socialinstitutions in the community such as schools, renderingthem more difficult to followin long-term preventive trials. This requires substantial efforts and ingenuityon the part of the investigator to maintain contact with the sample to collectfollowup data. Followup

Although careful followup isa costly and time-consuming effort,the results can be rewarding from a scientificpoint of view. An exampleis the Perry Preschool Project (Berrueta-Clementet al. 1984), which involved of qu ality preschool the provision expel .ances to urban minority ch ildren ina randomized field experiment. The samplewas followed for 15 years withvery little attrition, primarily because of the tenacityand commitment of the research results of this prevontive trial staff. The are quite encouragingeven though the sample size in this particular trialwas not large. In short, sample attritionat various points ina preventive trial is a largely underestimated, but extremely serious, threat to validinference. Though the use of complex multivariats methodsfor estimating the effects possible, it can only give of attrition is us lower- or upper-bound estimateson the effects of attrition on results of thetrial; and suchan approach is no substitute for minimizing attrition througheffective recruitment methods efforts. and followup The power of a research design is not determined,,occlusively by alpha and sample size.Also relevant is thepsychometric quality of the procedures used to measurement assess the effects of the intervention.If, as suggested, those procedures are applied to themonitoring of an individual's developmental path, then progress along a applied measurementprocedures mustassess that progress in a reliable and validmanner. Unreliability wil: lower the power of efforts to assess statistical an intervention's effects. Invalidityin the measurement procedures results incomplex problems in determining evaluetive design and in the power of one's integrating the meaning ofone's findings. Thus,from the onset,one is confronted with the limitations of the psychometricprocedures for assessing thepresence of established disorder addition, one is confronted (e.g., Kleinmuntz 1982).In with the challengeof assessing emotional behavioral processes and as they evolve over extended periodsof time. Thus far, the mental health scienceshave not made extensive measurement fronts. progress in either of these In our view, adoptionof the transactionalparadigm requires thatprogram developers develop sophisticated ing changes in multiple techniques forassess- developmentalprocesses simultaneously and forderiv- ing appropriatemathematical models of suchchanges. In the interim we urge that those involved inthe design of preventive interventions ensure that their measurement proceduresare psychome rically rigorous. As a first steptoward that goal,one must determine that ment procedures measure- are reliablethat theymeasure the phenomenon ofinterest FROM THEORY TO RESEARCH 87 in a stable and consistent manner. For preventionresearch, measurement procedures must be designed to assess the status of processesover time.Thus, in addition to whatever index of internal consistencythe researcher selects to assess the structural integrityof the measure, some evidence of temporal stability must also be demonstrated. This requirement maybe met through a test-retest analysis using a conceptuallyappropriate intertest interval. The length of this interval is particularly critical in preventionresearch when the intervention is introduced during one developmentalperiod (e.g., early childhood) and its consequences are expected to occur one or moreperiods later (e.g., preadolescence). Yet to be resolved are the criteriaby which a researcher can apply measurement procedures acrosssuch durations with confidence. Equally problematic for the prevention researcher is thechallenge of monitoring a developmental process over several developmentalstages. As cognitive, emotional, or behavior processes mature, they undergoqualitative changes that may require multiple distinctmeasurement approaches for their assessment. The challenge to the psychometrician is to design assessmentstrategies that mirror the natural developmental path of such processes.What is probably required are batteries of conceptually linked measures that monitorthe maturation of significant dimensions of emotional, behavioral, anapsychological growth. Pragmatically, components le such batteries wouldoverlap at critical developmental junctures to ensure the accurate assessment ofunderlying processes during transitional sta03.An illustration of this approach is repre- ernted by the Wechsler series of measures to assess intellectualfunctioning from preschool through adulthood, i.e., the Wechsler Preschool and PrimaryScale of Intelligence (WPPSI), the Wechsler Intelligence Scale for ChildrenRevised (WISC-R), and the Wechsler Adult Intelligence Scale Revised (WAIS-R) (Settler 1982). To design such sequential assessment procedures,preventionresearchers should identify or conduct sufficientgenerativestudies (Cowen 1980) to flesh out the necessary developmental paths. If thetransactional perspective does indeed have paradigmatic potential, it must inform the questions tobe addressed by such research. Obviously, it will require the simultaneousexamina- tion of individual and environmental factors and,nec,?ssarily, their interplay. Beyond its psychometric contributions, such reeearch will alter insignificant ways the conceptual processesemployed in the design of preventive interventions. Rather than being driven teleologically by theoutcome(s) to be avoided, such interventions should be guided by the developmental sequences tobe achieved. Considerable validation effort will be required before suck, measurement tools are readily available. Most criticalwill be the documentation of the construct validity (Cronbach an1 Meehl 1966) of such measures. Such validation studies, in turn, can confirm the predictive utility of the transactional framework itself. The defining characteristic of an effective preventive intervention is the documentation of changes in the developmental experiences of its recipients. 88 PREVENTION OF MENTAL DI3ORDERS

For some, this will involve the replacementof a nathogenic sequence with its normative counterpart. In such instances,the onsdt of a disorderedprocess will have been avoided. For others,screening procedures will detect thepresence of early indicants of disorderor dysfunction and provide suggestive evidencethat a pathogenic sequence is underway. We emphasize"suggestive" because of the frequency with which early signs of disorderappear in the normal population (Gersten et al. 1978).Letter reversals, for example,are present in many learning-disabled children during the preschoolandIvimary grade years. Their utility as an indicator of risktor learning disabilities, however, issignificantly limited by the frequency with whichthey appear in the nonlearning-dissbled population throughout much of thesame period.The statistical chall.TIge confronting us is to design assessmentprocedures that enable researcherto contrast the developmental experiencesof those receiving the preventiveintervention and their appropriatecontrols. In effect, what is requiredis a set of procedures that contrast individualstatus at multiple points in time andthat continually contrast wherean individual is developmentally with whereone would expect himor her to be. As indicated, two distinct differencescores are relevant to the research question. Thefirst of these has to do with therelative position of tho intervention andnonintervention groups. One wouldexpect that the former will displayan increasingly normative developmentalpattern. By contrast, the control group is expected to display evidence of thepathogenic process to be avoided. In our view,a second contrast is also important. By examining the distinct developmentalpaths experienced by those withinthe intervention group who door do not respond to the intervention,we are likely to learn much about contributors to pathology. Wr, are also likely to learnmuch about our interventions andconsequently about their optimalprescriptive application.

Concept of Proof inPrevention Research

Having designed and implementeda preventive intervention basedon the transactional paradigm, how is itssuccess to be confirmed? The viewwe have offered throughout this chapteeis that the sinequa non of success is evidence of an altered developmental path or experience. Clearly, much remains tobe accomplished in the design ofstatistical analytic proceduresto assess such change precisely. Others approachthe task not unlikea psychotherapy outcome study. In their view, the questionto be answered is, Howare the intervention and nonintervention groups different at the end of the evaluationperiod? Consequently, they designa pretest and posttest study (with potentiallytwo or more posttests as followups) andassess the differential status of bothgroups at some defined point in time. The limitationof this approach lies in itsinability to provide other than themost minimal data about howthe intervention influenced the lives of its recipients.All we learn from such analysesis whether the groups differ interms of central tendency andvariation. Because of the aforementioned base rate issue,researchers are also frequentlyunable to secure FROM THEORY TO RESEARCH 89 large enough differences in those truly atrisk to offset the limited amount of change possible in those who would nothave become pathological. Elsewhere (Lorion 1987b) we have discussed thisproblem in some detail. It is appropriate at this point, however, to emphasizethat documentation of the success of a preventive intervention should includeevidence both that the disorder of concern did not occurand that developmentally appropriatebehaviors either appefx Or reappear. This issue is critically importar t to preventionresearch. If our critique of group comparison studies onlow-base-rate disorders is valid, it is likely that considerable effort and resources have been investedin studies that, from the outset, were incapable of answering thequestion asked. Therefore, it is possible that a number of "negative" studies, in fact, neverrepresented a reasonable test of the intervention, for they failed toappreciate that the study of rare events requires large samples. Consequently, therelative dearth of solid scientific evidence for the attainability of preventive objectives maybe somewhat artifactual. This is unfortunate for two reasons.First, it means that valuable research resources have been wasted.Second, it has provided critics of prevention with unjustified support for their argument that the mentalhealth disciplines are not ready to undertake preventivetrials (Lamb and Zusman 1982). After more than a decade of debating the merits ofinvestigating preventive strategies, we and our supporters are under increasing pressure tosubstantiate our claims with facts. No longer can we attempt to justify theimplementation of interventions, at either the programmatic or policy level,merely on the basis that they are good for people.We have argued effectively that preventive interventions needed to be added to the armament of mentalhealth professionals. We based this position on the long-eetabliehed publichealth truism that no significant contributor to human morbidity and mortality has everbeen controlled by treatment, but only through effectiveprevention. We argued that the prevalence of emotional and behavioral disorder inthe nation's child and adolescent population far exceeded the existing oranticipated pool of service providers (Albee 1959; Cowen 1973; Glidewell andSwallow 1969). We urged policymakers to make the design and application of preventivestrategies a national priority (Klein and Goldston 1977) and basedthis on the otherwise uncontrollable costs of meeting related human needs. Incombination, these arguments served us well and produced substantial andseemingly continuing increases in the availability of funds to develop and assesssuch interventions (Department of Health and Human Services 1984, 1986). Inreality, however, they also put us on the spot. Having won the argument, we must nowdeliver! Much of this chapter has focused on how we might usethe transactional paradigm in designing research that provides the information necessaryto document and expand the potential of preventive interventions tosustain these promises. We have argued that the most theoretically andscientifically valid manifestation of that promise is represented by evidencedocumenting that our interventions have replaced or modified significantdevelopmental processes. We assume that effective prevention implies that a childeither will or will not 90 PREVENTION OF MENTAL DISORDERS

experience differently events in hisor her life that threaten emotional and behavioral adaptation. Recipient., ofpreventive services should undergo different developmental trajectories thanthey would have without theintervention. In effect, the transactionalperspective argues that the achievementof such changes will have bothimmediate and long-term implicationsfor the child's or adolescent's functioning.

However accurate such speculationmay be (and only extensive researchover many years will enable us to make thatjudgment), its limitedcurrency for policymakers and the general publicmust be noted. As argued elsewhere (Lorion 1983, 1985, 1987a),at its most basic level theconcept of prevention demands a relatively simple anddirect kind of evidence of itsefficacy; that is, do fewer casee of a disorder or class of disorders exist in thepopulation? We must appreciate that bothour supporters and critiesare justified in their insistence that we address thisquestion in as direct and convincinga manner as possible. Whatever other benefitsmay be achieved by our efforts, they must ultimately result in a reduction inthe prevalence of the conditionsthat initially justified the intervention. Thus,for all of us the real proof ofa preventive intervention's value and basisfor deserving a share of availablehuman service resources must be evidence that something has been prevented. Shouldwe succeed, there is little doubt thatprevention-related activities will have tributed in a substantial con- way to both the science and servicedelivery of the mental health disciplines.

References

Achenbach, T.M. The ChildBehavior Profile: An empiricallybased system for assessing children's behavioralproblems and competencies.International Jour- nal of Mental Health 7:24-42,1979. Achenbach, T.M. Developmental Psychopathology. 2d ed. NewYork: Wiley, 1982.

Albee, G.W. Mental Healthand Manpower T..?nds. New 1959. York: Basic Books, Albee. G.W. Preventing psychopathologyand promoting humanpotential. American Psychologist 37(9):1043-1050,1982, Albee, G.W. The argument for primary prevention. In:Marlowe, H.A., Jr., and Weinberg, R.B., eds. Primary Prevention:Fact or Fallacy.Tampa, Fla.: University of South FloridaPress, 1983. Albee, G.W. Advocatesand adversaries of prevention.In: JesIer, M. and Goldston, S.E., eds. A Decadeof Progress in PrimaryPrevention.. Hanover, N.H.: University Press ofNew England, 1986.pp. 309-332. FROM THEORY TO RESEARCH 91

Amabile, T.M., and Stubbs, M.L. Psychological Research in the Classroom. New York: Pergamon Press, 1982. Bell, R.Q. "Changing Psychological Risk. Implications for Mass-Targeted Men- tal Health Intervention and Promotion." Invited address presented at the Annual Meeting of the American Psychological Association, Washington, D.C., August 1986. Bcrrueta-Clement, J.R.; Schweinhart, L.J.; Barnett, W.S.; Epstein, A.S.; and Weikart, D.P. Changed Lives: The Effects of the Perry Preschool Program on Youths Through Age 19. Monograph on the High Scope Educational Research Foundation, No 8. Ypsilanti, Mich.: 1984. Bloom, B.L.Community Mental Health: A General Introduction.2d ed. Monterey, Calif.: Brooks/Cole, 1984. Buckner, J.C. ; Trickett, E.J.; and Corse, S.J. Primary Prevention in Mental Health: An Annotated Bibliography. Washington, D.C.: U.S. Govt. Print. Off., 1985. Campbell, D.T., and Stanley, J.C.Experimental and Quasi-Experimental Designs for Research. Boston: Houghton-Mifflin, 1963. Caplan, G. Principles of Preventive Psychiatry. New York: Basic Books, 1964. Catalano, R., and Dooley, D.Economic change in primary prevention.In: Price, R.H.; Ketterer, R.F.; Bader, B.C.; and Monahan, J., eds. Prevention in Mental Health: Research, Policy, and Practice. Beverly Hills, Calif.:Sage, 1980. pp. 21-40. Cook, T.D. Priorities in reeearch in smoking prevention. In: Prevention Re- search Deterring Drug Abuse Among Children and Adolescents. NIDA Monograph Series, DHHS Publication No. (ADM) 85-1334. Washington, D.C.: U.S. Govt. Print. Off., 1985. pp. 197-219. Cook, T.D., and Campbell, D.T. Quasi-Everimentation: Design and Analysis Issues for Field Settings. Chicago: Rand McNally, 1979. Coopersmith, S. The Antecedents of Self-Esteem. San Francisco: Freeman, 1967. Cowen, &L. Social and community interventions. Annual Review of Psychology 24:423-472, 1973. Cowen, E.L. Some problems in community program evaluation. Journal of Consulting and Clinical Psychology 46:792-805, 1978. Cowen, E.L. The wooing of primary prevention. American Journal of Com- munity Psychology 8:258-284, 1980. 92 PREVENTION OF MENTAL DISORDERS

Cowen, E.L. Primary prevention inmental health: Past, present, and future. In: Felner, R.D., Jason, LA, Moritsugu,J.N., and Farber, S.S., eds. Preventive Psychology: Theory, Research, and Practice.New York: Pergamon Press, 1983. pp. 11-25.

Cowen, E.L. Primary preventionin mental health: Tenyears of retroepect and ten years of prospect. In: Kessler,M., and Goldston, S.E., eds. A Decadeof Progress in Primary Prevention.Hanover, N.H.:University Press of New England, 1986. pp. 346.

Cowen, E.L.; Lorion, RP.; andDorr, D. Research in the communitycauldron: A case report. Canadian Psychologist15:313-325, 1974. Cronbac..., L.J., and Meehl, P.E.Construct validity in psychologicaltests. Psychological Bulletin 42:281-302, 1955.

Cummings, E. Primary prevention:More cost than benefit? In: Gottesfeld,N., ed. The Critical Issues of CommunityMental Heal,th. New York:Behavioral Publications, 1972. pp. 161-174.

Department of Health and HumanServices. Prevention Activities of theAlcohol, Drug Abuse and Mental HealthAdministration, Fiscal Year 1983,Report to Congress. Washington, D.C.:U.S. Public Health Service, 1984.

Department of Health and Human Services.Prevention Activities of the Alcohol, Drug Abuse and Mental HealthAdministration, Fiscal Year 1984,Report to Congress. Washington, D.C.:U.S. Public Health Service, 1986. Fairweather, G.W., and Tornatzky,L.G.Experimental Methods for Social Policy Research. New York:Pergamon Press, 1977.

Felner, RD. Vulnerability in childhood:A preventive framework forunder- standing children's efforts tocope with life stresses and transitions.In: Roberta, M.C., and Peterson, L.,eds. Prevention of Problemsin Childhood: Psychological Research and ApplicationsNew York: Wiley, 1984.pp. 232-265. Feiner, RD.; Jason, L.; Moritsugu,J.; and Farber, 5.5., eds.Preuentive Psychol- oc: Theory, Research and Practice. New York:Pergamon Press, 1983.pp. 251-272.

Frankenburg, ond Dodds, J.B. The DenverDevelopmental Screening Test. Journal of Pediatrics71:181-191, 1967.

Garmezy, N., and Rutter, M., eds.Stress, Coping, and Developmentin Children. New York: McGraw-Hill, 1983.

Gersten, J.C.; Latimer, T.S.; andSimcha-Fagen, O. Developmentalpatterns of types of behavioral disturbance andsecondary prevention. InternationalJour- nal of Mental Health 7(3-4):132-149,1978. 4 FROM THEORY TO RESEARCH 93

Glidewell, J.C., and Swallow, C.S. The Prevalence of Maladjustment in Elemen- tary Schools: A Report Prepared for the Joint Commission on the Mental Health of Children. Chicago: University of Chicago Press, 1969. Gordon, R. An operational definition of prevention. Public Health Reports 98:107-109, 1983. Hays, W.L.Statistics for Psychologists. New York:Holt, Rinehart and Winston, 1981. Heller, K; Swindle, R.W.; and Dusenbury, L. Component social support processes:Comments and integration.Journal of Consulting and Psychology 54(4):466-470, 1986. Johnson, 0.G. Tests and Measurements in Child Development: Handbook 2. Vols. 1 and 2. San Francisco: Jossey-Bass, 1976. Johnson, 0.G., and Bommarito, J.W. Tests and Measurements in Child Develop- ment: Handbook I. San Francisco: Jossey-Bass, 1971. Kellam, S.G.; Brown, C.H.; Rubin, B.R.; and Ensminger, M.E. Paths leading to teenage csychiatric symptoms and substance use: Developmen'al epidemiological studies in Woodlawn. In: Guge, S.B.; Earls, FAT.; and Barrett, I.E., eds. Childhood Psychopathology and Devdopment. New York: Raven Press, 1983. pp. 17-51. Klein, D.C., and Goldston, S.E., eds. Primary Prevention: An Idea Whose Time Has Come. Washington, D.C.: U.S. Govt. Print. Off., 1977. Kleinmuntz, B.Personality and Psychological Assessment. New York: St. Martin's Press, 1982. Kohlberg, L.; La Crosse, J.; and Ricks, D. The predictability of adult mental health from childhood behavior. In: Workman, B.B., ed. Manual of Child Psychopathology. New York: McGraw-Hill, 1972. pp. 1217-1284. Kuhn, T.S. The Structure of Scientific Revolutions, Chicago: University of Chicago Press, 1970. Lamb, H.R., and Zusman, J.Primary prevention in perspective. American Journal of Psychiatry 136(1):12-17, 1979.

Lamb, H.R., and Zusman, J.The seductiveness of primary prevention.In: Perlmutter, F.D., ed. Mental Health Promotion and Primary Prevention, 1982. pp. 19-30. Lilienfeld, A.M., and Lilienfeld, D.E. Foundation of Epidemiology. 2d ed. New York Oxford UniversAy Press, 1980. Loeber, R.; Dishion, T.H ; and Patterson, G.R. Multiple-gating: A multistage assessment procedure for identifying youths at risk for delinquency. Journal of Research in Crime and Delinquency 21:7-32, 1984, 4 I ./ 94 PREVENTION OF MENTAL DISORDERS

Long, B.B. The prevention of mental-emotional disabilities.: A report froma National Mental Health Association Commission.Atnerican Psychologist 41(7):825-829, 1986. Lorion, R.P. Productivity and health care. Strategies for tho assessment of needs, resources and services. Social Policy 8(1):17-27, 1978. Lorion, R.P. Environmental approaches to prevention: The dangers of imprecision. Prevention in Human Services 4(1/2):193-205, 1983. Lorion, R.P. Evaluating preventive interventions: Guidelines for the serious social change agent. In: Feiner, RD.; Jason, L.; Moritsugu, J.; and Farber, S.S., eds. Preventive Psychology:Theory, Research, and Practice in Community Intervention. Nc tv York: Pergamon Press, 1985. pp. 251-272. Lorion, R.P. The other side of the coin: The potential for negativeconsequences of preventive intervention.In:Steinberg, J.A., and Silverman, M.M., eds. Preventing Mental Disorders: A Research Perspective. Washington, D.C.: U.S. Govt. Print. Oft, 1987a. Lorion, R.P. Methodological challenges in prevention research. In: Steinberg, J.A., and Silverman, M.M., eds.Preventing Mental Disorders: A Research Perspective. Washington, D.C.: U.S. Govt. Print. Oft, 1987o. Lorion, R.P., Cowen, E.L., and Caldwell, R.A. Problem types of children referred to a school-based mental health program: Identification and outcome. Journal of Consulting and Clinkal Psycholoatv 42:491-496, 1974. Lorion, R.P.; Hightower, A.D.; Work, W.C.; and Shockley, P.The Basic Acaemmic Skills Enhancement Program: Translating prevention theory into action research. Journal of Community Psychology 15(1), 1987. Lorion, R.P., and Lounsbury, J.W. Conceptual and methodological considerations in evaluating preventive interventions. In: Tash, W.R., and Stahler, G., eds. Innovative Approaches to Mental Health Evaluation. New York: Academic Press, 1982. pp. 23-59. Lorion, R.P.; Work, W.C.; and Hightower, A.D. A school-based, multi-level preventive intervention: Issues in program development and evaluation.Per- sonnel and Guidance Journal 52:-79-484, 1984. McCord, J. A thirty-year fullow-up of treatment effects. AmericanNychologist 33:284-289, 1978.

McGuinness, D. When Children Don't Learn. New York: Basic Books,1985. Munoz, R.F.; Snowden, L.R.; and Kelly, J.G., eds.Social and Psychological Research in Community Settings: Designing and ConductingPrograms for Social and Personal Well-Being. San Francisco:Jossey-Bass, 1979. FROM THEORY TO RESEARCH 95

Paul, G.L. Strategy of outcome researchin psychotherapy. Journal of Consult- ing Psychology 52:109-118, 1967.

Price, RH. "Four Domains of PreventionResearch." Paper presented to the National Institute of Mental HealthConference on Prevention Research. Aus- tin, Tex., February 22-24, 1982.

Price, RH.; Ketterer, R.F.; Bader, B.C.;and Monahan, J., eds. Preventionin Community Mental Health: Research,Policy and Practice. Beverly Hills,Calif.: Sage, 1980.

Price, RH., and Smith, S.S. A Guideto Evaluating Prevention Programs in Mental Health. Washington, D.C.: U.S.Govt. Print. Off., 1985. Quay, H.C., ed. Handbook of JuvenileDelinquency. New York: Wiley, 1987. Rappaport, J.; Seidman, E.; and Davidson,W.S. Demonstration research and manifest versus true adoption. In: Munoz,R.F.; Snowden, L.R.; and Kelly, J. G., eds. Social and Psychological Researchin Community Settings: Designingand Conducting Programs for Social andPersonal Well-Being. San Francisco: Jossey-Bass, 1979. pp. 101-144.

AV., and Allen, L. Preventionof Childhood Disorders. BeverlyHills, Calif.: Sage, 1988.

Roberts, M.C., and Peterson, L., eds.Prevention of Problems in Childhood: Psychological Research and Applications.New York: Wiley, 1984. Ross, AO. Learning Disability:The UnreaVzed Potential. NewYork: Mc- Graw-Hill, 1977.

Rutter, M. Relationships betweenchild and adult psychiatric disorders:Some research considerations. ActaPsychiatrica Scandinavica 48:3-21,1972. Rutter, M. Psychiatric disorders inparents as a risk factor for children,in press. Sameroff, A.J., and Chandler, M.J.Reproductive risk and the continuumof caretaking casualty. In: Horowitz,F.D.; Hetherington, M.; Scarr-Salapatek, S.; and Siegel, G., eds. Reviewof Child Development Research.Vol. 4. Chicago: University of Chicago Press, 1975.

Sanford, N. The prevention of mentalillness. In: Wolman, B., ed. Handbook of Clinical Psychology. New York:McGraw-Hill, 1965.pp. 1378-1400. Settler, J.M. Assessment ofChildren's Intelligence and SpecialAbilities. 2d ed. Boston: Allyn and Bacon, 1982.

Schlesselman, J .J. Case-Control Studies:Design, Conduct Analysis. New York: Oxford UniversityPress, 1982. r 4 '96 PREVENTION OF MENTAL DISORDERS

Selltiz, C.; Wrightsman, LS.; and Cook, S.W. Research Methods in Social Relations. New York: Holt, Rinehart and Winston, 1976. Silver, LB. The Misunderstood Child. New York: McGraw-Hill, 1984. Small, L. Neuropsychodiagnosis in Psychotherapy. New York: Brunner/Mazel, 1973. Steinberg, J.A., and Silverman, M.M., eds. Preventing Mental Disorders: A Research Perspective. Washington, D.C.: U.S. Govt. Print. Off., 1987. Stohlberg, A.L., and Garrison, KM. Evaluating a primary prevention program for children of divorce. American Journal of Community Psychology 13(2):111- 124, 1985. Tadmore, C.S. ; Bar-Maor, il A; Birkham, J.; Sho-Shany, G.; and Hofman, J.E. "Pediatric Surgery: A Preventive Intervention Approach to Enhance Mastery of Stress." Unpublished manuscript, University of Haifa, 1985. Thoits, P.A. Social support as coping assistance. Journal of Counseling and Clinical Psychology 54(4):416-423, 1986. Walker, D.K.Socioemotional Measures for Preschool and Kindergarten Children: A Handbook. San Francisco: Jossey-Bass, 1973. CHAPTER 3

Prevention Programming as Organizational Reinvention: From Research to Implementation

Richard H. Price, Ph.D. University of Michigan, Ann Arbor

Raymond P. Lorion, Ph.D. University of Maryla td, College Park In this chapter we argue that the successful implementation of a model prevention program is a form of organizational reinvention (Rice and Rogers 1980). The term organizational reinvention captures tile process we wish to cibacribe for several reasons. First, the implementation of a model prevention program is inherently organizational in nature.The proce9s involvea the orchestration of internal and external organizational resources; the scanning of the organizational environment; the focusing of program goals and objectives; and finally, implementing a model program and monitoring the program for fidelity to its original goals. These activities depend heavily on internal project organization and the receptiveness of the host organization. Second, the idea of organizational reinvention emphasizes the fact that all model prevention programs have both core and adaptive features. The adaptive features can be adjusted to fit local circumstances and require invention and inventiveness on the part of the implementer.However, reinvention also requires sensitivity to the need for fidelity in the implementation and n.onitor- ing of the core features of the praven+ ive program. Finally, the idea of organizational reinvention implies that imphmenting the model prevention program involves the replication of a program already tested in a previous preventive trial. The program is not entirely new, but has a core technology, which is to be reproduced as faithfully as possible. Achievement of this objective is no small accomplishment. Unfortunately, both developers of social innovations and practitioners frequently underestimate the challenges associated with the task of implementing ilese innovations in their local settings. As mental health professionals learn how to identify risk factors and design effective preventive trials, they must appreciate the wisdom of Embry (1984), who noted: "The field of education is littered with the corpses of proven

This paper was prepared for the Project Prevention initiative of the American Academy of Child and Adolescent Psychiatry.

I 98 PREVENTION OF MENTAL DISORDERS

innovations. Solving a problem is not usually enough to ensure that the solution will be widely implemented" (p. 82). We hope that this chapter will assist readers to avoid that fate for their preventive efforts. In one sense, this chapter is a continurtion of the earlier chapter by Lorion, Price, and Eaton. Once a preventiv: Lrial has provided convincing evidence of effectiveness and consequently identified a "model" to be implemented, the logical next step is to reinvent that model program in a real-life setting with its attendant demands. To clarify this process from one perspective, we begin by examining the local implementation context and the role of the innovator. We then consider a number of discrete steps in the implementation process, including (a) environmental scanning and the initial development of local linkages; (b) organizational focusing and the setting of local objectives; (c) actual implementation, balancing the needs for fidelity and adaptation; and (d) the establishment ofa monitoring system to document that the preventive trial continues to explicitly state objectives focused on proximal outcomes of the program. Preventive programs, if effectively implemented and widely disseminated, will become part of what Sameroff and Fiese (this volume) referred toas "macroregulations." As Sameroff and Fiese observed, such macroregulations represent a culture's "developmental agenda" whereby at various points in time the environment is restructured to provide socially desired experiences fora child (e.g., kindergarten entry). Many preventive programscan and should be designed to complement and enhance society's developmental agenda for children.In fact, in our view, shaping macroregulations to enhance human development is the major objective of preventive interventions. The implementation of a preventive program frequently occurs insome form of social organization. To draw examples from other chapters in this volume, adolescent suicide prevention efforts typically occur in high schools. Alcohol and other drug use prevention efforts frequently involve families,peer groups, or schools, separately or in combination. Delinquency preventionprograms have been targeted at the workplace, the school, or the juvenile courts.Finally, preventive programs aimed at children of emotionally disturbed parentscan be implemented collaboratively with psychiatric hospitals, juvenileor family courts, and foster care and adoption agencies. If the implementation of a preventive program isan act of organizational reinvention that always occurs in the context ofsome social organization, then consideration of organizational readiness for prevention efforts ismost important. Understandably, many organizations are not prepared to adoptpreven- tive programs, no matter how persuasive the evidence from earlier controlled trials. Thus, we would argue that the successful implementation ofa preventive program is no accident. Rather, it results from the convergence of multiple factors, which include critical characteristics of the role of the innovator, the readiness of the host organizatior ,d the appreciation and implementation of a series of steps that root the innovation in the organization. In the following FROM RESEARCH TO IMPLEMENTATION 99 section, we consider implementation activities in the context of the overall prevention research process. Implementation ContJxt Price (1983) described the prevention-research process in terms of the sequential completion of four interrelated stages (see figure 1).In this model, prevention research moves through a series of successive developmental stages, each of which is cyclical and iterative. Thus, each stage has its own cycle of activities, which may or may not ultimately yield a successful productwhich, in turn, serves as the input for the next stage.

Survey Change Descriptive echnoiogy Problem Research Definition INNOVATION PROBLEM ANALYSIS DESIGN Technology Modifiable and Selection Descriptive Selection Risk Factors of Change Factors Risk Factor Research

Widespread Innovation Adoption Design

Innovation Prototype

Broad Context Diffusion Assessment of Standardized Version Determine( Evaluation INNOVATION User of FIELD DIFFUSION Population Innovation TRIAL

Partial,41 \ Diffusion Implementation With Continuing of Innovation Modification

Figure L The prevention-research process (from Price 1983). The initial stage of the prevention-research process, problem analysis, is represented in the upper left-hand portion of the diagram. This stage involves the completion of epidemiological studies aimed at identifying modifiable risk factors. If successful, the results of such research inform the development ofa preventive innovation, which, once designed, is tested in a preventive field trial. Ultimately, the process producer! a prototype or model program.The availability of a model program necessitates initiation of the fir al stage of implementation and dissemination, innovative diffusion, which,one hopes, will be followed by widespread preventive effects. 100 PREVENTION OF MENTAL DISORDERS

A number of chapters in this volume deal with risk factors and theory, and therefore address problems that fall in the upper left-hand portion of this schematic framework. The chapter on methodological problems by Lorion, Price, and Eaton focuses primarily on C...., demands of innovation designs and field trials.But as we indicated earlier, successful preventive trials do not automatically lead to effective implementation. Achievement of this objective requires consideration of unique demands, many of which involve restructuring elements of the organizations in which the interventions are to be embedded. This chapter examines those demands, beginning with an overview of the reinvention process and the concept of "soft technologies." The implementation of a preventive intervention involves the installation and maintenance of a soft technology. A technology is described as soft when it involves the establishment and stabilization of social transactions inan organizational context, for example, to meet preventive goals. Both the technology of the intervention (which inevitably involves transactions among individuals) and the organizational context in which it is installed are soft in that bothare subject to a variety of dynamic organizational and cultural forces thatcan act either to protect and strengthen the innovation or to undermine and distort it. Implementing a preventive intervention is quite different from, f3r example, adding a new circuit board to a computer. A unique set of skills is required to translate an effective preventive trial into a viable community-based and -supported intervention program. The successful application of these skills, however, requires a responsive setting; that is, the host organization must be prepared for and capable of adapting to the addition of theprogram. The setting conditions that define this state are discussed later under the heading "Or- ganizational Readiness." When these conditions are met, the opportunity exists for program diffusion, that is, for achieving the necessary reinvention. That process involves completion of the following steps: (a) environmental scanning and the establishment of initial interorganizational linkages; (b) organizational focusing activities, including the formulation of goals and objectives; (c) implementation that is subject tc a number of factors that may facilitateor hinder the process; and (d) effective monitoring of program outputs.Because innovators typically work with empirically proven interventions, considerably more emphasis is placed on local target selection, organizational arrangements, and goals and objectives. Having an intervention already tested ina preventive trial also means that we can focus more on the evaluation of proximal outcomes to assure fidelity in program delivery. The Innovator and the Host Organization

At least two critical ingredients can increase the likelihood of successfully implementing a preventive program. The first relates to the roleattributes or skills of the innovator; the second to the organizational madiness ofthe host

1 11 FROM RESEARCH TO IMPLEMENTATION 101 organization. We view each as a necessary but not sufficient condition for successful program implementation. Furthermore, although we discuss them independently, these two factors obviously interact over time. Role Attributes cf a Successful Prevention Innovator Let us consider first some of the role attributes or skills characteristic of a successful innovator. These attributes are important additions to the specific professional training of the individual innovator. Child psychiatrists, for example, may have a considerable fund of knowledge about the mental health of children, various modes of adaptation they display, and characteristics of settings and relationships that promote optimal psychosocial development. Child psychiatrists also him., an understanding of the characteristics of children at risk for various negative health and mental health outcomes. For successful program implementation, however, this array of knowledge and skills should be complementei with the following additional specific skills. Goal Orientation The implementation of a preventive program requires that the innovating individual have relatively specific goals concerning the proximal outcomes of the preventive intervention. For example, in a program designed to reduce the incidence of learning disorders, the attainment of a specific set of academic skills by the target group by a defined point in time (e.g., beginning of fourth grade) might represent such a proximal outcome. Of course, a variety of accommoda- tions and adaptations will be required as the program is implemented. But having a clear understanding of the proximal outcomes to be achieved and the sequence of steps by which they will be achieved enables the innovator to distinguish those aspects of the program that must be maintained intact from those that can be adapted to the particular local context. Orientation to External Relationships Typically, program implementers must be able to reach out beyond the host organization. They must identify supporters and champions in the environment As well as "gatekeepers" who have socially sanctioned control over access to critical reeources needed to make the program effective. Tly e critical resources may be potential program recipients, access to the media, ..ommunity support for access to the schools, sources of local funding, or any of a variety of other needed items. An important aspect of this roln attribute is the capacity to be sensitive t the "strength of weak ties" (Granovetter 1973). Frequently those ties may be indirect connections among key individuals in the community who can bring needed resources and support to a program. Thus, for example, it may not be the personal acquaintance of an innovator, but a "friend of a friend" (Boissevain 1974; Sarason 1976) who is critical in implementing a preventive program. 102 PREVENTION OF MENTAL DISORDERS

1 esource Sensitivity The ability to mobilize needed resources and to know where they exist in the local community is an important attribute of successful innovators (Price 1986). Such people seem intuitively attuned to the availability of criticalTesources in the form of funding or influence. An h.-portant aspect of mource sensitivity is an awareness of potential threats to needed resources as well as to their potential availability. Participative Orientation The implementation of a preventive program is by definition a participative enterprise. The implementers must recognize that successful innovations always require the involvement of others in program development and ideally make them major stakeholders in its success (Tornatzky et al. 1983). Successful innovators resist the temptation to "go around" an individualor organization and attempt to engage the person or group at some level in moving the preventive enterprise forward. Entrepreneurial Orientation

Mental health professionals vary considerably in their willingness togo outside their own familiar professional setting to obtain theresources and access needed for a preventive program. Some professionals believe that their professional training Nmoves them from the necessity toengage in such activities; others believe that it entitles them to neededresources.What must be developed, however, is an appreciation of the art ofn. ptiation and coalition building. Frequently, a team approach is useful in developinga preventive program precisely because some team members are more accustomed toan entrepreneurial approach than othets. Cultural Sensitivity

Because the implementation of a preventive programoccurs in a local context, the innovator must be culturally sensitive to differences in assumptions,world views, and expectations that are likely tooccur in diverse communities. For example, how - local community and school greeta program that screens children for learning disabilities may differ substantially dependingon that community's past experience with similarprograms, whether such programs have treated recipients with respect and caring, and theperceived ethnic sensitivity of the innovator.Similarly, programs designed to reduce such outcomes as teenage pregnancy, adolescent suicide, the risk foremotional disorder in children of disturbed parents, and conduct disordersor alcohol and other drug use in adolescents allcan be highly volatile in a local community situation. Frequently, reducing risk factors is notseen .as a neutral act by some community members. Innovators who forget this doso at their own peril.

.411 f. FROM RESEARCH TO IMPLEMENTATION 103

Organizational Readiness Just as innovators with certain skills are morelikely to achieve program implementation goals, preventive interventions are also morelikely to succeed in host organizations that have certain attributes.D'Aunno (1986) described a set of characteristics related to anorganization's readiness to implement a preventive program. D'Aunno's discussion sensitizes us tocritical aspects of host organizations that should be assessed before programimplementation. A brief summary of D'Aunno's model for organizationalreadiness is given in figure 2. The remainder of this section briefly reviews the nAjorelements of this model.

(1) Support/Demands from Environment of the Treatment Unit 'State and Regional agen.lies 'Advocacy Groups

(4) Resources Available (5) Organizational (2) Awareness and Acceptance to Support Prevention Structures and of Problem by Host Org.inization 'Funds 10, Services that Facilitate the Facilities Adoption of 'Expertise Innovations

(3) Attitudes, Beliefs, and Practices 00 of Statf

Readiness to Launch and Support Innovative Prevention Programs

Figure 2. A model of organizational readiness to launch prevention programs (adapted from D'Aunno 1988). D'Aunno argued that organizational readiness relates in part to support and demands from the larger organizational environment to adopt an innovative prevention program (Van de Ven 1986). Advocacy groups in the local community, region, or state, for example, may provide support and incentives to implement a prevention program, as well as financial resources, information, or technical assistance. Environmental support couldinclude political pressure. The capacity of such pressure to affect organizational readiness should be appreciated and exploited. D'Aunno also observed that local awareness and acceptance of the problem to be addressed by the program is a second critical ingredient.Preventive programs aimed, for example, at teenage suicide ordrug use are unlikely to encounter receptive audiences in the host organization if the local community

1 1 4 104 PREVENTION OF MENTALDISORDERS

is either unaware ofor unwilling to acknowledge the problems.In such instances, activities to enhance communityawareness of and pressure to respond to targeted problems may be needed before a program isimplemented. A third critical ingredient thatboth increasesawareness and acceptance and is likely to increase theresources available for program implementation do with the attitudes, beliefs, has to and practices of the localhost organization's staff or members. If the problem to be preventedis treated as a taboo topicto be avoided, it will hinder the hostorganization's readiness toimplement preventive programming. D'Aunno argued thatsupports and demands from theenvironment and attitudes and beliefs oflocal organizational membersboth affect the awareness and acceptance of the problemto be prevented and, inconjunction with problemawareness, may contribute to the likelihood available to support that resources will be a preventive intervention. Theseresources, of course, include such critical ingredientsas funds, personnel, expertise, facilities, materials. and

Yet another aspect of the host,aganization needs to be examinedin assessing organizational readiness, that is,organizational structures and in place that will facilitate services already preventive innovations.For example, D'Aunno explained that organizations are more likely to launch innovativeprevention programs if they already have organizationalstructures and services available to facilitate such adoption (Galbraith 1982; Van de Ven1986). Organizations that have already created roles to reach out into the communityare more likely to adopt innovative preventionprogramming. Finally, organizations ready have developed cooperative that al- relationships with relevantreferral agencies are also more likely to innovate. In addition, a growingbody of organizationalresearch and theory has identified snme of thecharacteristics of organizations likelihood of adapting to that increase their new conditions and demands fromtheir environment (Hasenfeld 1083; Katz and Kahn 1P78; Lawrence and Lorsch1967; Mintzberg 1979; Thompson 1967).For example, organizations flexibility in their structures with higher levels of for communication, coordination,and decisionmaking and flexibility in thedefinition of work roles to adapt and to implement are more likely to be able preventive programs. Organizationswith relatively high levels of communicationamong staff members and administrators (Georgopoulos 1986) are alsomore likely to adapt and toengage in coordination by mutual adjustment, adjustingon the spot rather than relying through bureaucracy. on coordination Thus far, we have outlined some of the role characteristics ofpotentially successful preventive innovatorsand identifiedsome of the characteristics of host organizations thatare more likely to adopt preventive it is unlikely that most innovations. Though innovators or host organizationswill possess all these attributes, assessing andmaximizing their plementation procesq presence before beginning the im- can contribute significantly to thelikelihood of implementation success. For example, examining the organizationalreadiness of a school 115 FROM RESEARCH TOIMPLEMENTATION 105 along the dimensions we have justdescribed may lead to the choice of oneschool over another forbeginning one's preventive program.As discussed below, the selection of a host organizationwith a high likelihood of success canhave important positive impacts onsubsequent efforts with more problematichost organizations. The organizationslcharacteristics described previouslyprovide the innovator with a practicalchecklist for selecting among potential targethost organizations. Our discussion of the role attributesof preventive innovators can also serve as a checklist orinventory of useful skills andorientations.It should be recognizad that no single individualis likely to or needs to possess all theeerole characteristics to achieve successfulinnovations. Rather, we believe thatthe implementation of a preventive interventionin a local community setting typically requires a team rather than anindividual enterprise. The aforementioned checklist can be used to determinewhether the team's skills and orientations, in aggregate, provide theneeded role orientations and skillsfor successful innovation. Having reviewed critical aspects oforganizational readiness and role characteristics for effective implementation, we nowdescribe a series of steps that increase the likelihood of successfulimplementation. Environmental Scanning andInitial Linkages

Assuming the requisite role characteristicsand a veasonable level of organizational readiness, the stage is set for thefirst step in implementation. This step involves scanning theenvironment to collect critical planningdata and establishing initial linkages with critical actors,both in the host organization and in the larger environment. We assumethat the implementation team has already identified (a) a population subgroupthey believe to be at risk, (b) a broad model for intervention, and (c) somepreliminary ideas about the nature of the negative health nnd mental healthconditions to be prevented. At this stage, the orientation of the project team mustbe both outward toward the target population and agencies that may play arole in the project, and inward toward the host organization wherecritical decisions about implementationwill be made. Data-based planning and project implementationis necessarily a continuous process. Questions willbe initially formulated and thenreformulated. Alter- natives for implementation will be evaluated,sorted, and then reevaluated in the light of new information. This processis both continuous and interactive. Plans developed privately by a singleindividual are rarely likely to be implemented. Typically, they suffer from theabsence of the diverse perspectives available from the various staff members inthe implementation team and members of the host organization. For these reasons,the initial planning stages in the project should be interactive. Infact, we encourage that they be conducted within a workshop format. 106 PREVENTION OF MENTALDISORDERS

Planning Workshop The rationale underlyingthe workshop suggestion informed planning is simple.If a well- group is assembled, the workshopprovides an initialoppor- tunity tocompare their perceptions of the program is to be implemented. local environment inwhich the The workshop formatalso providesan opportunity to initiate linkagesbetween the collaborators in the development who will later be involved and implementationof the local prevention 1 providesa worksheet for planning project. Table questions outlined a prevention ?roject workshop.The on the worksheet can be usedto L.volve a numberof staff members in a planningsession and to mobilize The planning session their participation andinterest. car also be broadened toinclude a widerrange of stakeholders, includingmembers or the target providers, or volunteer groups, other human-service groups in the community.Before the planners focus narrowly on the needsof the target population, should be formulated a broad picture of the project in an intensiveplanning session with the members of the host project staff, the organization, and, perhaps,members of other knowledgeable groups in thecommunity. Table I. A Worksheetfor Planninga Prevention Project foran Identified Target Population 1.Describe the target population (age,sex, socioeconomic characteristics, geographic distribution). 2.Identify major stresses affecting the targetpopulation. 3. What problemswithin the target populaion should bereduced or eliminated witha prevention intervention? 4. What skills does the target populationneed to developto cope? 5.Identify the agenciesor groups in the community planning for this that must be involvedin target population. Whichperson(s) need to beinvolved? 6. What steps willbe taken tosecure the interest and communitygroups or agencies? cooperation of the 7.Establish several tentative objectives forthe interventionproject. 8.Identify intervention strategies to achievethese objectives. 9. How will the program be evaluated to identifyneeded administrative changes while theproject is underway? 10. How will the project be evaluatedto determine theextent to which intervention objectiveshave been wet forthe target group? 11. What level ofresources (information, will be needed? money, support, space, expertise) What sources for theseresources should be approached?

11.7 FROM RESEARCH TO IMPLEMENTATION 107

Rarely will the product of such a workshop be a detailed program plan. Instead, an outcome of the workshop should be a substantial reduction in 12`3ff uncertainty about the general direction of the project. Furthermore, this initial effort will almost certainly uncover a number of barriers, problems, and 'questions that must be answered before additiunal planning can proceed. The data collected from such a local workshop can be combined with other data coAected by methods described later; and all these datacan provide critical information for later development of locally adapted project goals and objectives. Collecting data that focuses both on the host organization and its envi:on- ment is clearly not a neutral act. Data collection in any organizationoccurs in a context of strongly held beliefs, vigorously protectAd territory, partisan viewpoints, and, at best, partially open communication. It is not surprising that many projects begin to fail early in their development because of' the difficulty of getting adequate information about the host organization and other aspects of the program's environment that are necessary for effective implementation. Murrell (1976) recommended that initial environment scant 'ag activities be judged by two criteria.First, the information needed should be collected by unbiased methods; second, the information should subsequently be used in program operation. Much of his discussion focused on the frequently ignored issue oforganizational contextin which these initial environmental scanning activities are carried out. Murrell suggested that detailed discussion and interaction between the project staff and the director of the host organization shouldoccur well before data collection begins.This early interaction is intended to maximize the likelihood that the information obtained from data collection will actually be used to make project decisions. Murrell suggested that the project staff should conduct what he calls an "in-loop assessment" of the host organization director and sponsor. This internal examination focuseson needs and resources in the sponsoring organization and should be done collaboratively with the directorof that organization, wherever possible. Knowledge about several criticalissues should result from such a review. First, the prevention project staff should know the linst organization's feared risks. For instance, the collection of information can".risky to an administrator. Certain types of information (e.g.,consumer satisfaction or staff efficiency), if collected and made public,can threaten the stability of the host organization or fundingsources. It must be recognized that although needs-asseesment data can documenta gap in an organization's service delivery network, it can also be interpretedas evidence of that network's failure to meet its responsibilities to the community. Thus, those interestedin program organization should be aware that some agencies, particularly if their funding sources are "soft" or dependenton public good will, may be reluctant to release certain types of information deemed relevant toprogram development. A second consideration involves weighing the perceived risksof gaining information against the costs of various methods of data collection. Sometypee of data collection, such as social indicator dataor census data, present relatively limited risk to most organizations. On the other hand, conductinga community '1 1 S 108 PREVENTION OF MENTAL DISORDERS forum about the problem to be prevented may raise a number of uncomfortable questions for the host organization and, therefore, may be justifiably perceived as risky. This does not mean that such a forum should not be undertaken. A third issue to address in the in-loop assessment has to do with the perceived benefits of information from the host organization's point of view. What does the host organization expect to gain from the implementation of the prevention project in its home territory? These expected benefits should be taken into account during the environmental scann ing activity so that data relevant to that perceived benefit for the sponsor or host organizationcan be collected. For example, if the host organization believes that the proposedprogram will contribute to its reputation as an innovative organization,some measure of the community's opinion of its status as a resource might be includ9d. Following such an in-loop assessment, it is essential to plan with the host the specification of critical choice points in the data collection process and to involve the host organization or sponsor in data analysis and interpretation. Methodological Options in Environmental Scanning

There is no single beet method oL collecting data about the local environment in which a prevention program is to be implemented. Arange of methodological options, however, can be considered. Each of these options has its respective strengths and weaknesses. Knowing those strengths and weaknesses is critiail in making intelligent choices about what kinds of dataare genuinely useful in program planning. Following are six data-collecting strategies that can aid in planning. In each (=se, we briefly deecribe the method, commenton some of the underlying assumptions associated with the strategies, and mentionsome likely data sources as well as advantages and disadvantages of each method. 1. Analysis of sociodemographic and health statistics. This methodinvolves compiling data from public records and making inferences aboutcom- munity needs based on these findings. For example, police records might be examined to assess the frequency and seriousness of juvenileoffenses and to identify the geographical area in which to implementa program for the reduction of delinquency. Theuse of archival sources assumes that demographic and other social characteristics relate to the mental health of children. Available sources of such data irclude city planning departments, funding agencies, health departments, police records, local and state educational agencies, and the census department. The advantages of archival searches are that such dataare readily available and a wide range of information can thereby be obtained. There is no simpleway, however, to relate these data to mental healthprograms and to prevention programming plans. Moreover, this method fails to consider citizeninput on the meaning 9f the data.

1 1 9 FROM RESEARCH TO IMPLEMENTATION 109

2.Client use of agency services.This method involves examining previous patterns of use of agency services by clients andinferring on that basis the community's future needs. For example, caseloads ofdrug abuse treatment clinics could be examined to assess theneed for services designed to prevent alcohol or other drug use in theschools or to respond to the needs of the children or siblings of the clinic's clientele.Examination of service records assumes that all community needs are coming to agencyattention and that high utilization reflects high need.Of course, this is not necessarily true. The typical data source for thismethod is agency records detailing the number of clients served by various programs over a particular period of time. An advantage of this method is thatall agencies have records available in some form, and at a relativelylow cost. Disad- vantages of this method are that its findingsreflect demand other than need and that the findings may, in fact, reflect factors such aspublicity, service costs, or availabilityor perhaps the referral patternsof other agencies. 3.Analysis of existing serviceresources in the community. This method invohes a count of the type and capacity of relevant servicesin the community that are available to meet client needs. Forexample, telephone hotlines and emergency room services in a community might beidentified to assess the degree to which suicide prevention resourcesalready exist in settings accessible to youth. This method assumes that information about the availability of other rekources is relevant toinferring community needs. Possible sources of information about other services include service directories, associations of agencies, existing agency relationships, staff knowledge, and client referral patterns. Theadvantages of resource analysis methods are that they help avoid duplication of services and can be completed at relatively low cost. Their disadvantages are that they do not provide a direct indication of acommunity's need for prevention services, they require consideration of' multiple information sources, and the reliability is difficult to assess. 4. Citizen survey. To achieve the most exacting estimate,this method involves interviewing a stratified random sample of citizens by mail, telephone, or personal contact. The content of the interview focuses on the respondent's mental health needs, problems in living, and knowledge of service availability.For example, a random sample of seventh-grade children could be anonymously surveyed to assess the degree of self- reported alcohol and other drug use as well as knowledge of such use by peers. Survey use assumes that the self-reportof needs or problems is a valid indicator of' a subgroup's or community's mental health needs. The advantage of surveys is that they provide relatively reliable, direct, and wide-ranging information. The primary disadvantages of the method are that it is expensive and time consuming and requires sampling expertise.

1 2 0 110 PREVENTION OF MENTAL DISORDERS

5. Key informant interviews. This methodinvolves selecting nomore than 10 to 15 knowledgeable community members,such as police officers, public health muses, long. time residents,or agency personnel, to interview about their perceptions of mental health needsof a segment of the community. For example, problems of children of mentallyill parents could be assemod by interviewing knowledgeable socialworkers, nurses, psychologists, psychiatrists, teachers, and daycare workers.The key informant technique assumes that those interviewed havean accurate picture of the problems of communitygroups or areas.Data sources includerepre- sentatives frc_. existing networks of communityaction groups, service agencies, field workers, and citizensgroups. This method is quick and economical, and it may lead to otherinformation sources. Insome cases, it also provides informationon the political climate of the community, that is, its readiness to accept the proposedintervention. The disadvantage of this method is that itmay not be truly representative of thegroups in need, because the perceptinne nf key informantsmay be biased or selective. 6. Community forum. This method involvesholding a well-publicized public meeting at which a wide range of participantsdiscuss their perceptions of community needs. For example,a cowmunity forum could be heldon the needs for additionalprograms for learning-disabled childrenor for the children of depressed mothers. Use ofthe community forumassumes that participants are representative ofor understand the groups in need and that their perceptionsare accurate indicators of such need. Thus, the data source is citizen's public statements. The advantagesof a community forum are that it is economical,produces diverse input, andcan be a catalyst for citizen act'ion. The disadvantageis that a large forum limits individual participation. Frequently,however, these community forums are poorly attended and highly reactive. Theymay also raise unrealistic expectations about whatpr ograms can accomplish. As a group, these strategies forenvironmentai scanning havetwo important characteristics. First,a number of them provide data about the contextof the local program to be implemented.Furthermore, contact with the datasource is, in many cases,a critical opportunity to link with membersof the host organization or other agenciesor with groups whose collaboration andsupport are needed for program success.This is particularly true, ofcourse, when agencies are contacted toassess service utilization patternsor available service resources in the community. Similarly, keyinformant interviews andcommunity forums can representimportant initial contacts withthe host community. Even the analysis of archivalhealth statistics can provide initiallinks with critical agencies suchas local public health or police departments.

121 FROM RESEARCH TO IMPLEMENTATION 111

Organizational Focusing: Mobilizing Internal Resources

The process of environmental scanningprovides program implementers with a wide range ofobservations and data relevant for programplanning. Unfor- tunately, it also frequently leaves theimplementation team with more options than it can choose. Thus, the next stageof development requires the innovation team to define specific program goalsand objectives, which serve a dual purpose. First, they focus the project team and,in some cases, members of thehost organization, on a limited number of specificactivities aimed at carrying out the steps necessary for programimplementation. We refer to this as "organize- tional focusing." If properly formulated, thedevelopment of goals and objectives also provides specific measurable programoutcomes. These "proximal outcomes" (Price 1987) reflect the short-termobjectives to be achieved by the prevention program. In the local implementation of model programs,long-term objectives seldom can he measured. Theselonger teem, "distal" objectives are typicallymeasured in the preventive trial from which themodel program was developed (Lorion, Price, and Eaton, this volume). Assumingthat there is reasonably strong evidence of a causal connection betweenproximal and distal outcomes, the achievement of proximal outcomes serves as anindicator that the program is ftanctioning as it should in its new hostorganization. Consider the following example. If, as someresearchers suggest (Kumpfer, this volume), cigarette smoking is a"gateway drug" to more serious drug use problems, then reduction in the number of newcigarette smokers in ju nior high school can serve as a proximal indicator to monitor analcohol or other drug use prevention program Of course, anyproximal indicator is useful only if it can be observed reliably and if there isreasonably strong evidence of a causal connection between it anddistal outcome of concern. Thus, by focusing on goals and objectives, theinnovator gains an opportunity to distill and reorganize informationcollected during environmental scanning. The formnlation of pre\ ...ntion project goals serves acrit.Cal planning function and represents an important conceptual stagein the imprsementation of prevention programs. This planning function mustbe a collaborative undertaking. The formulation of goals and objectives canand should bring people from the project and host organization together and serve as anopportunity both for clarifying concepts and for mobilizing the energies ofstaff and administrators toward a common goal. This is a critical linkbetween activities of the environmental scanning, on the one hand, and programimplementation, on the other. The approach we describe here has beendeveloped by Reddin (1971) and Burian and his colleagues (1979) and has also beendescribed in Price and Smith (1985).

'") 1r)I 112 PREVENTION OF MENTAL DISORDERS

Developing a Problem Statement

Organizational focusing begins withformulating a problem statement.The problem statement, ina sense, reinvents the original preventivetrial, because similar problem statementshave been formulated during theoriginal preventive trial.

At least three elementsmust be present to definea problem. First, a desired end state or prevention goal should be selected. Second,information concerning the situation or circumstances of a target group should beavailable.And, finally, the sought-after differencebetween the current situationand the prevention goal or end statemust be specified. An exampleof a problem statement is given in table 2. Variousindicators of the present conditionare given on the left-hand side of theproblem workup. On the farright-hand side, the ideal state of affairsor goal is stated in general terms. Finally,barriers that stand in the way of reaching thegoal are listed in thk middleof the table. This problem statement illustrateshow data from the environmentalscanning stage can be incorporated. The barriers listedin table 2 represent hypothesesabout major problems to be removedto reach the desired pmventiongoal.

Table 2. Problem WorkupExample Present Condition Barriers Problem Resolution Goal

Family life stress Sound family life in Washtenaw County in Washtenaw County Indicators: Number of divorces Deficiencies in in 1985 intrafamilial communication skills Number ofsepara- Inadequate prepara- tions in 1985 tion for marriage Number of domestic- Underemployment disturbance police and unemployment calls in 1985

Number of child- Deficiencies in neglect referrals community supports in 1985 for one-parent families FROM RESEARCH TOIMPLEMENTATION 113

This analytic approach toprevention program planning represents astrategy for removing the barriers to thedesired end state. In a sense,then, one can say barriers to a desired that the prevention program issimply aimed at removing prevention goal. Forming Goals andObjectives There is a sharp distinctionbetween prevention project goalsand project objectives. Project goals are generalstatements that specify thesought-after condition or state of affairs. On theother hand, a project objectiis "a specific statement of the outcomes thatindicate progress toward the goal orthe removal of barriers within a specifictime frame" (Price and Smith1985, pp. 47-48). In this approach to planning, nodirect measure of program goals ismade. Instead, the focus is on specific attainmentof specific operational objectivesthat define the path toward the goal. Formulating explicit, well-designedproject objectives is difficult, time consuming, frustrating, and absolutelyessential to effective programplanning. Each objective formulated for aparticular prevention program mustspecify four elements: (a) the result to beachieved; (b) the criteria by which wewill know the result has been achieved;(c) the timeframe within which itis expected that the result will be attained; and(d) the target group or objecttoward which the program effort is aimed.Table 3 illustrates examples ofthese elements of objectives written for an elementarydrug use prevention program.

Table 3. Example of a PreventionProgram Objective

Target: 80 percent of seventh and eighthgraders at Wilson School

Result: Knowledge of harmful effc of drugs

Criterion: The XYZ test

Time: By the end of the 1986-1987 school year 80 percent of the Objective: "By the end of the 1986-1987 school year, seventh and eighth graders at WilsonSchool will have gained an acceptable level of knowledgeof the harmful effects of drugs as measured by theXYZ test."

1P,1 114 PREVENTION OF MENTAL DISORDERS

This sort of organizationalfocusing activity is extremely valuablein the process of program implementation. It should beundertaken collaboratively with the innovationteam, and the results should be sharedwith the host organization. The objectives, if properlyformulated, specify the elements of the program and indicate the proximal outcomesthat can be used to monitor program implementation. Though we can state objectivesprecisely, we may not alwaysknow at the outset that they can be completed. Writing achievable objectives dependson considerable prior knowledge ofthe phenomena underconsideration, the capacities of project staff to achieveprogrammatic workover time, and a number of other elements about whichthere may be some uncertainty.Even if prior knowledge is imperfect, the focusingfunction involved in formulatinggoals and objectives should not be neglected.As more realistic estimatesof progress ale obtained, objectives can be revised. In fact, such review andpossibly revision should be scheduled regularly,perhaps annually, at the timeof reformulating budgets and renewing organizationalarrangements for program continuation. Implementation

The time for implementation isat hand when the innovatoris armed with appropriate role characteristics;the host organization isat a reasonable level of readiness; planning data have been collected throughenvironmental scanning; and goals and objectiveshave been defined that takeinto account local variation and uniqueness. The necessary steps have now been taken toreinvent the original prevention modelthat constitutes the innovationin this process. We now consider factors that the research literaturesuggests enhance the probability of successfulimplementation andsome of the dilemmas faced by the implementer in attempting tobalance program fidelity with adaptation. the need for local Tornatzky and his colleagues (1983)suggested that implementation thought of as "a host of activities can be which take place betweenadoption (some point of organizational commitmentto the innovation) and theptmanent incorpora- tion of the innovation into the organization's repertoire ofpractices" (p. 131). A number of researchers have conducted empirical studies anddeveloped theories about factors that affect thelikelihood of succeesfulimplementation. Facilitating Factors

Experimental studies varyingthe conditions under whichmental health innovations have been implementeti (Fairweather et al. 1974;Tornatzky et al. 1980) reveal that implementationsuccess depends on the degree of participative decision-raking in the hostorganization and whetherintervention and consultation techniques emphasizeface-to-face interadion. Insupport of this, Corbett and Guttinger (1977) reportedthat team, mther than in a workshop training iddividual, involvement session increased the likelihoodof an effective implemen- FROM RESEARCH TOIMPLEMENTATION 115

of local tation,In addition, Yin (1979,1980) observed that the existence initiative in the developmentof an innovative program isrelated to implementation success. Similarly,Yin et al. (1977) reported thatpractitioner involvement in implementationplanning, in this case involvingteacher participation during strategy sessions, andearly practitioner experience withinnovative technologies aided later implemerattioneftbrts. Stolz (1981, 1984) has alsoreviewed the literature in this area.She reported personal that the strongest single variableinfluencing the diffusion uf models is interaction between the agencydecision-maker and a colleague who promotee the use of the model. Herstudies suggested that individualcontacts and individual personalities are crucialin determining whether aparticular technology would be used. The findingsreported by Patton et al. (1977) supportthis view. Theso researchersinterviewed decision-makers and foundthat the single most important element inthe utilization and implementation processwas the "personal factor," which appeared tobe "made up of equal parts ofleadership, interest, enthusiasm, determination,commitment, aggressiveness, andcaring" (Patton et al. 1977, p. 155). Stolz (1984) wisely observed thatthe personal factor described heresheuld not be misinterpreted asevidence that successful innovators areunusual individuals or that nothing short ofextraordinary personal resources and energy is effective. Rather, as otherresearch has shown, the adoption ofinnovations frequently result from the efforts of a single , d their successful implementation individual who works with localorganizations and maintains a personal relationship with them over time.An example of how this processhas been institutionalized is the use of locallybased agricultural experts who give farmers information on the latest developmentsin agriculture.These Agriculture Extension Service agents, throughtheir friendships with local farmers, effectively disseminate improvedfarming techniques (Knott andWildavsky 1980; Lawler 1982). Finally, Rothman et al. (1976)concluded from research on innovationthat innovations allowing for a "trialrun" have a higher adoption rate thaninnova- tions requiring total acceptance oradoption without such an anticipatorytrial. Rothman et al translated thisgeneralization into the following practitioner principle: Practitioners wishing to promote aninnovation in a general target system should develop it initiallyin a partial segment of that target system. Table 4 summarizes the facilitatingfactors described here. This summary table, though primarily descriptive, canbe thought of from a prescriptive point of view. That is, in actuallyimplementing an innovation, attempting tomaxi- mize each of these factors should beimportant to the ultimate success of the program.

1 r f; 116 PREVENTION OF MENTAL DISORDERS

Table 4. Facilitating Factorsin Program Implementatkm

Participative decision-making inthe host organization Use of face-to-face interactionin communication Stimulation of local initiative inimplementation Early involvement of practitionersin the process Personal interaction betweenhost organization decision- makers and program implementer

High levels of personal commitmentby implementer Personal relationships establishedover long periods of time Partial adoption precedingfull implementation

Fidelity Versus Adaptation Thus far we have suggested that the development ofan increased level of participation among those whowill implement the innovationis likely to increase its success.Furthermore, numerousresearch studies suggest increasing the that sense of program "ownership" by the hostorganization is critical to implementationsuccess (Tornatzky et al. 1983).If this is true, confronted with we are a dilemma, because increasing thesense of ownership may also mean that local participants will seekto modify theprogram model in ways that address local needs anddemands but thereby reduceits effectiveness. A substantial controversy hasarisen in the field concerning (1984) observed; this issue. As Stolz

In the study of knowledgediffusion and utilization,a major controversy centers on whether to insist thata model be used in a formas close to the originally tested formas possible (fidelity), or toencourage organizations to modify innovations(adaptation) (Emshoff 1982; 1982). On the Roitman and Mayer one hand, some believe thatany modification of a model would dilute its effectivenessin some unpredictable use of a model requires consistent way. In this view, any application across settingi.Others, however, contend that because every setting is different, modelsmust be adapted from their originalform to be suitable forthe organizational, political, social and economiccharacteristics of the tion, in this new setting. Adapta- sense, is not just a cost-saving compromise.Adaptation (or reinvention, see F. Rice and Rogers 1980)means creative changeeneces- sary to adapt the model to local ci;._mnstances. Many authorssuggest that 1 7 FROM RESEARCH TO IMPLEMENTATION 117

human service providers have such a strong preference foradapting models and that models may not get used unless some adaptation is permitted (Fawcett et al. 1980, p. 238). As discussed, we believe that both fidelity and adaptation have their respective strengths and weaknesses. Striving for fidelity presumablyenhances the likelihood of success, because the model program, not some variation, was originally deemed effective. The disadvantage of striving for fidelit) is that the difficulty of implementation increases if the host organization or ?ractitioners must adapt to the innovation rather than the reverse. At the same time, adaptation has its own advantages and disadvantages. Seekins and Fawcett (1984) argued that permitting some adaptation of a program model promotes rapid implementationand diffusion of the intervention. The obvious disadvantage, however, is that some adaptations may actually decrease the effectiveness of the innovation and thereby reduce the likelihood of further dissemination. A number of strategies have been used by program designers to preserve fidelity (Stolz 1984). Some develop certification programs so that the implemented program is certified as "real" only if certain criteria are met. Others attempt to control adaptation by not permitting any modifications of procedures for an initial period of time. Still others select only cooperative sites or reject host organizations that seem intent on making significant adaptations to the program. Stolz argued that the evidence indicating that adapting programs reduces actual effectiveness is mixed.Roitman and Mayer (1982), for example, reported a correlation of approximately 0.4 between fidelity to the original model and effectiveness of the program. Although this is a correlation of moderate magnitude, it suggests some relationship between program modification and effectiveness. We believe that the appropriate resolution of the fidelity versus adaptation dilemma requires the distinction between the core elements of the intervention and the adaptive characteristics that can be adjusted to local circumstances. For example, a preventive intervention designed to reduce alcohol and other drug use problems in ju nior high school students may invIlve a well-articulated curriculum of ten training sessions with learning objectives for each session. Presumably, this is the core technology of the intervention, and the achievement of the learning objectives represents a check on the degree to which the program has been effectively implemented. On the other hand, the nature of the site in which the program is delivered, timing and scheduling issues, who delivers the program, and other factors may be modifiable with less risk to thefidelity of the program. Prevention programs are rarely described in ways that easily allow for distinguishing between what is "core technology" and what is not.In many cases, a model program has not been evaluated in ways thatidentify essential

1 2' S 118 PREVENTION OF MENTAL DISORDERS program elements.Nevertheless, a distinction betweencore and adaptive characteristics of the innovation is critical ifprogram implementation is both to be effective and to obtain the original prevention goals.

Closing the Feedback Loop: Monitoring for Fidelity

As stated earlier, preventive interventionsare soft technologies. Political and economic forces in the host organization,as well as in the larger environment, tend to squeeze, distort, and otherwise alter preventionprograms, sometimes reducing their effectiveness. Consequently, it isimportant to collect information that documents how the program is being implemented.Such information can serve as a gyroscope to keep the program on course in the face of organizational pressures to distort it in variousways. Data of this sort close the feedback loop and allot' those delivering theprogram to alter various aspects of the program that may have drifted from their originally plannedcourse. It should be emphasized that this sort of datacollection is quite different from tho outcome-oriented data collected in thecontext of a preventive trial (Lorion, Price, and Eaton, this volume). Its intent isto assess the degree to which the program meets short-term and proximal objectives that have been shownin the preventive trial to be linked to longer-term preventiveoutcomes. To the degree that it does, we can have some confidence thatthe program is being implemented in such a way that its desired preventiveimpacts are likely. Though there is a wide array of data thatwould be useful to collect about program implementation, two types of informationare particularly important. First, documentation that thecore technology of the preventive program is being delivered as intended is critical information.For example, a program designed to increase the learning skills of childrenat risk for developmental disabilities may require 200 hours of active learning experiences with personalcomputers as a critical core component. Data indkating that childrenin the program receive less than half that amount would showthat the core technology is not in place. A second major type ofprogram monitoring data involves tracking the attainment of the program's proximal objectives.Recall our earlier example of t;igarette smoking as a gateway substance.If smoking precedes Use of other drugs, reducing the incidence of cigarettesmokingshould be a proximal outcome for a program targeted toa broad array of alcohol and other druguse, health, and mental health outcomes. A number of different datasources can provide information that indicates how a program is being delivered. Frequentlythese data are available from the host agency in routineprogram records, including the. following: FROM RESEARCH TO IMPLEMENTATION 119

Characteristics of the target populatior Thesedata include demographic characteristics, special needs and problems,and characteristics that may affect receptivity to the preventiveinterventionall important data sources for programmonitoring.If the program is being delivered to populations whose characteristics arestrikingly different from those of the target population in the originalpreventive trial, preventive impact is obviously uncertain. Measures of program delivery objectives.These include indicators of successful administration of specific componentsof the program, attendance information, or other data concerningthe degree to which the program is delivered asdesigned. Staffing patterns.These may vary and, in some cases, maybe an important core element of the program.If so, such staffing information should have been specified in the descriptionof the original preventive trial, and information about staff backgroundand education, previous experience and skills, selection, supervision,and communication flow should be monitored in the replication. Proximal program outcomes. Examples includesmoking incidence in drug-use prevention, school absence indelinquency prevention, suicidal ideation in suicide prevention, or academicachievement scores in programs to prevent learningdisabilitiee. Cost data, This informationper serviceepisode, per client, or per unit of risk resolutioncan be collected in many cases.Cost data are important to monitor, both for the purposesof future budget planning and for program justification. The costefficiency as well as efficacy of a program are considerations that cannotbe ignored. Broskowski et al. (1978) offered several principlesconcerning implementation of program monitoring systems. Theyargued that one important prereq- uisite for any monitoring system is a thoroughunderstanding of the multiple decisions that managers and program deliverystaff must make for daily operations and future planning. It may be useful tofollow and observe staff over time to better understandthe environments in which they operate and the types of decisions they make.Similarly, in-depth interviews with agency directors and staff can be informative. Examiningthe minutes of meetings and other records also, frequently, provides importantinformation. A second principle suggested by Broskowski etal. (1978) is to secure the commitment of management in tne host organization.Knowing about managerial decision-making does not assure the actual useof information ,,ollected in the context of program monitoring. Asking programdelivery personnel or managers of the host organization to comment onthe form in which information is provided frequently can improve the likelihoodof its use for program monitoring and stabilizaion. 120 PREVENTION OF MENTAL DISORDERS

Finally, ensuring staff participation andcooperation is crucial, both in the design of a monitoring system andin its implementation. One strategyto increase participation is to provide frequentreports to staff and offer continual feedback, asking them to suggestimprovements in the reporting system they must use. Finally, developing a monitoringsystem incrementally is important. By working in stages, staffcan spend time determining what are essentialdata to be collected routinely and whatcan be obtained on an ad hoc basis. In addition, using an incremental approach allowsthose who are developinga program monitoring system to obtainresponses from staff on initial versions of the system, and on subpartsas they are implemented.

Conclusion

Successful implementation ofa prevention program is a form of organizational reinvention. Furthermore,successful implementation dependson the particular role attributes of implementersand the organizational readiness ofthe host organization. The actualprocess of implementation begins with environmental scanning and establishinginitial linkages with the environment.Data collected in the context of scanningprovide an opportunity fororganizational focusing, which involves settingobjectives and mobilizing staff and otherinter- nal resources towardprogram implementation. We have argued thata number of organizational factors facilitatesuccessful implementation, particularly enhancingparticipation, obtaining the commitment of the implementers, and developingpreventive interventionson a partial basis before full implementationoccurs. We also suggested thatevery implementation of a preventiveprogram involves balancing a desire for fidelity versus the need to adapt theprogram to local circumstances. Makinga diatinc- tion between core technology andadaptive features ofa program increases the probability of successful implementationand, at the same time, maintainsthe likelihood of program impact. Finally, the implementation ofa prevention program involves monitoring of critical data to determinewhether the program is being deliveredas intended. These data, when fed back to theprogram, can be used as a basis forprogram adjustments to maintain fidelity. Theyallow the program to continueto onerate as an effective social system, involvinga vulnerable population and a supportive, growth-enhancing environment. Thatis, after all, the ultimate goal ofpreventive research and action.

1 31 FROM RESEARCH TO IMPLEMENTATION 121

References

Boiss

Hasenfeld, Y. Human Service Organizations. Englewood Cliffs, N.J.: Prentice- Hall, 1983. Katz, D., and Kahn, R.L.The Social Psychology of Organizations.2d ed. New York: Wiley, 1978. Knott, J., and Wildavsky, A.If dissemination is the solution, what is the problem? Knowledge: Creation, Diffusion, Utilization 1:537-538, 1980. Lawler, RE., Ill. Strategies for improving the quality of work life. American Psychologist 37:486-493, 1982. Lawrence, P.R., and Lorsch, J.W.Organization and Environment. Boston: Harvard Business School, Division of Research, 1967. Mintzbe-g, H. The Structuring of Organizations. Englewood Cliffs,N.J.: Pren- tice-Hall, 1979. Murrell, S.A. Eight process steps for converting needsassessment data into program operations. In: Bell, R.A.; Sundel, M.; Apontie, J.F.; and Murrell, S.A., eds. Needs Assessment in Health and Human Services. HealthResources Administration, DHEW. Washington, D.C.: U.S. Govt. Print. Off.,1976. Patton, M.Q.; Grimes, P.S.; Guthrie, KM.; Brennan, N.J.; French,RD.; and Blyth, D.A. In search of impact: An analysis of the utilization offederal health evaluation research.In: Weiss, C.11., ed.Using Social Research in Public Policymaking. Lexington, Mass.: Lexington Books, 1977. Price, R.H. The education of a prevention psychologist. In:Feiner, R.D.; Jason, L.; Moritsugu, J.; and Farber, S.S., eds. PreventivePsychology: Theory, Re- search, and Practice in Community Intervention. Elmsford,N.Y.: Pergamon Press, 1983. pp. 290-296.

Price, R.H. Education for prevention. In: Kessler, M.,and Goldston, S.E., eds. A Decade of Progress in Primary Prevention. NewEngland Universities Press, 1986. (Hanover, New Hampshire, and London, England) Price, RH. Linking intervention research and riskfactor research. In: Stein- berg, J.A., and Silverman, M.M., eds. PreventingMental Disorders: A Research Perspective. DHHS Publication No. ADM87-1492.Washington, D.C.:U.S. Govt. Print. Off., 1987. pp. 48-56.

Price, RH., and Smith, S.S. AGuide to Evaluating Prevention Programsin Mental Health. Rockville, Md.: National Instituteof Mental Health, 1985.

Reddin, W .J. Effective Management by Objectives. New YorkMcGraw-Hill, 1971.

Rice, R.E., and Rogers, E.M. Reinvention in theinnovation process. Knowledge: Creation, Dif fusion, Utilization 1:499-514, 1980. FROM RESEARCH TO IMPLEMENTATION 123

Roitman, D.B., and Mayer, J.P. Fidelity and reinvention in the implementation of innovations.In:Blakely, C.H., chair."Adoption, implementation, and Routinization of Organizational Innovations." Symposium presbutedat the Annual Meeting of the American Psychological Association, Washington,D.C., 1982.

Rothman, J.; Erlich, J.L.; and Teresa, J.G. Promoting Innovation andChange in Organizations and Communities: A Planning Manual. New York:Wiley, 1976.

Sarason, S.B. Community psychology networks, and Mr. Everyman.American Psychologist 31:317-328, 1976. Seekins, T., and Fawcett, S.B.Planned diffusion of social technologies for community groups. In: Paine, S.C.; Bellamy, G.T.; and Wilcox, B., eds.Human Services That Work: From Innovation to Standard Practice.Baltimore, Md.: Paul H. Brooks Publishing, 1984. Stolz, S.B. Adoption of innovations from applied behavioralresearch: "Does anybody care?" Journal of Applied Behavior Analysis14:491-505, 1981. Stolz, S.B. Dissemination of standardized human service models.In: Paine, S.C.; Bellamy, C.T.; and Wilcox, B., eds. Human ServicesThat Work: From Innovation to Social Process. Elmsford, NY: Pergamon Press,1980. Stolz, S.B. Dissemination of standardized human servicemodels. In: Paine, S.C.; Bellamy, C.T.; and Wilcox, B., eds. Human ServicesThat Work: From Innovation to Standard Practice. Baltimore, Md.: PaulH. Brooks Publishing, 1984.

Thompson, J.D. Organizations in Action. New York: McGraw-Hill,1967. Tornatzky, L.G.; Fergus, E. 0.; Avellar, J.W.; and Fairweather,G.W. Innovation and Social Process. Elmsford, N.Y.: Pergamon Press,1980.

Tornatzky, L.G.; Fergus, E.O.; Avellar, J.W.; and Fairweather,G.W.The Process of Technological Innovation: Reviewing the literature.Washington, D.C.: National Science Foundation, 1983.

Van de Ven, A.H. Central problems in themanagement of innovation. Manage- ment Science 32(5):590-608, 1986.

Yin, R.K. Changing Urban Bureaucracies. Lexington,Mass.: Lexington Books, 1979.

Yin, R.K. Studying tit, Implementation ofPublicPrograms. TR-352-577. Gold- en, Colo.: Solar Energy Research Institute, 1980. Yin, R.K.: Heald, K.A.; and Vogel, M.E. TinkeringWith the System. Lexington, Mass.: Lexington Books, 1977.

, ,'4 4 CHAPTER 4

Public Policy: Risk Factor or Remedy?

Leon Eisenberg, MD. Maude and Lillian Pressley Professor and Chairman Department of Social Medicine and Health Policy and Professor of Psychiatry Harvard Medical School

The things that you're liable To read in the Bible They ain't necessarily so, Ira Gershwin Everyone "knows" that an ounce of prevention is worth a pound of cure and that a stitch in time saves nire. Why, then, is prevention given such short shrift in public policy, in the health habits of the public, in the education of medical students, and in the practice of medicine? For one thing, what everybody knows is not always true. If the n taphoric ounces, pounds, and stitches of the aphorisms are converted into dollar costs for prevention and cure, the sayings hold true only for some preventive interventions and are clearly false for many others; if they are converted into indexes of health outcome (morbidity and mortality), the bottom line becomes much more favorable; that is, much of the time, effective prevention offers betkr health at additional cost. Unless we specify the target, the units of assessment, and the social context, we cannot evaluate efficacy in unambiguous terms. The IOUs issued in the early days of the mental hygiene movement, which promised that mental illness would be eliminated by education about human relationships, proved to lx unredeemable. The result was widespread disbelief in preventive psychiatry (Eisenberg 1962). The fact is that a number of sources of psychiatric morbidity in the United States have been markedly reduced, Depressed and demented patients with pellagra no longer crowd the wards of state institutions; yet, during the first two decades of this century, the number of cases of pellagra 'vas estimated at 500,000 (Roberts 1920), of whom 40 percent displayed "mental symptoms" and 4 percent

This paper was prepared for the Project Prevention initiative of the American Academy of Child and Adolescent Psychiatry. 1 ,ul 126 PREVENTION OF MENTAL DISORDERS were committed as "insane" (Singer 1915).Research by Goldberger (1915) established the dietary cause of the disease; the tide of pellagra abated as social and economic progress ended the dependence of Southern tenant farmers on milled corn as a dietary staple (Sydenstricker 1958). Congenital paresis became a rarity with the effective treatment of syphilis. Vaccination against measles and rubella has markedly reduced brain damage resulting from uncontrolled infection in infants and children (Gruenberg et al. 1986).In contrast, the pathogenesis of such psychiatric disorders as schizophrenia and Alzheimer's disease are so little understood that we lack plausible rationales on which to design preventive programs. Between the proud successes and the areas of ignorance is a large gray zone: areas of public policy where knowledge suffices for effective intervention, but social action moves by fits and starts or not at all. Were sex education and family-planning clinics to be provided in public schools (Kenney 1986), current methods of contraception could reduce teenage pregnancy and its consequent neuropsychiatric and psychosocial morbidity for mothers and infants; were housing laws enforced and lead in ambient air controlled by environmental regulations, the toll of lead encephalopathy in children could be contained. Public health programs lag behind what is possible because the political will to act on the available knowledge base is lacking (Richmond and Kotelchuck 1985). Thus, despite prevailing cynicism about prevention in psychiatry, much mental disease has been prevented, more could be averted with what is now known, and additional accomplishments await the new knowledge basic research can provide. In formulating health policy, judgments about investing in prevention will differ depending on the criteria for outcome; that is, is the goal the reduction of mortality and morbidity or the reduction of health care costs? The goals (and the values that underlie them) must be stipulated if we are to achieve a common understanding of the potential for, and the limits to, prevention. To place the issues in context, this chapter will consider the following topics: the economics of prevention, alternative paradigms for prevention, assessing side effects, decisionmaking in a pluralistic society, and public policy on psychiatricre- search. The Economics of Prevention

There's no such thing as a free lunch. American folk saying For most physicians, it is taken as a given that preventing disease is one of the ch ief aims of medicine, even if practice is not always in accord with principle. When Government is asked to provide the funds for,say, a populationwide immunization program, it often relies on an economic cost-benefit analysis in arriving at a decision (Russell 1986). It asks: Will the investment in prevention PUBLIC POLICY 127 produce net budget savings by lowering subsequent costs for medical care? The analyst assigned to the task faces an important decision at the outset. Is the calculus to be limited to the direct costs (for physician visits, hospitalization, special schooling, and institutionalization), or should it also include the indirect costs resulting from disease and its sequelae (wages lost from work missed and earnings lost over a lifetime because of incapacity or death)? And should an attempt be made to transform pain and suffering into monetary terms? The inclusion of indirect costs in the cost-benefit equation provides a far more comprehensive answer, but it poses a bureaucratic dilemma. Direct costs are immediately evident in the increased expenditures for the new preventive initiative; the effect on health care costs will become apparent more slowly. The savings in indirect cosiz do not appear in the health budget per se; accrue over a longer time period; are not easily credited to an, initiativeundertaken yoars earlier; and are likely to be less compelling to elected officials preoccupied with this year's budget, the one by which taxpayers will judge them. Koplan and White (1986) carried out cost-benefit analyses of the impact of measles and rubella vaccination for one birth cohort: the 3.5 million infants born in 1981, 95 percent of whom would have been infected by early adulthood in the absence of immunization. The computation included both direct and indirect costs. For measles vaccination, total costs were $51.1 million (including vaccine side effects and cases not averted); without vaccination, costs from disease would have totalled $745 million; the benefit-to-cost ratio is 14.6 to 1. If the calculation is limited to the direct impact on the health budget, the benefit-to-cost ratio becomes 3 to 1. For rubella, the $54.8 million spent on a vaccination program saved $611.6 million in disease costs, a benefit-to-cost ratio of 11.2 to 1. If the calculation is limited to direct costs, the ratio becomes 2.6 to 1. Thus, by either system of reckoning, each vaccination program is an enor- mous bargain. As to health effects, the gains from each are even more impressive: 3.3 million cases of measles, 360 deaths, 1,100 cases of measles encephalitis and 30 cases of subacute sclerosing panencephalitis avoided in the first instance; 1.5 million cases of rubella, and 2,000 cases of congenital rubella syndrome (including 400 deaths and 300 cases of mental retardation) averted in the second. The campaign to eliminate rubella is informative about additional aspects of policy. the choice of an appropriate prevention strategy and the inclusion of a system to monitor compliance. U.S. policy is based on attaining universal (or near universal) immunization at preschool age, with a secondary program for susceptible adolescent and adult women (Bart et al. 1985).With effective implementation of this policy, the incidence of rubella has declined by 99 percent since 1969, the year the vaccine was licensed (Centers for Disease Control 1987). Between 1969 and 1980, cases of congenital rubella syndrome (CRS) continued to occur in appreciable number; since then, the annual rate has been reduced to one-sixth of the 1969 total, as herd immunity has taken effect and as the previously immunized have begun to reach childbearing age. In contrast, CRS

1 128 PREVENTION OF MENTAL DISORDERS

persists in the United Kingdom, which opted for a strategy of selective protection of adolescent girls rather than universal immunization (Editorial 1987). The choice by British authorities was based on their limitedsuccess with populationwide vaccination (even today the uptake of measles vaccine is only 68 percent). Until vaccine uptake levels exceed 90 percent, the incidence of CRScan show a paradoxical increase. Because the virus circulatesmore slowly in a partially immunized population than in an unimmunizedone, age at disease shifts toward the reproductive years. It is not enough to havea potent vaccine; to be effective, it requires an appropriate strategy foruse. That strategy, in turn, rests on a system for monitoring uptake, enforced by law in the United States at time of school entry. In the case of vaccination programs for adults, anticipated dollar savings prove to be illusory.Vaccinating all Americans age 65 and older against influenza in a typical year in the 1960s would have resulted ina net additional cost to the health budget of $600-$700 per year of life gained if 50-percent vaccine efficiency is assumed and of $300-$400 at a 70-percent efficiency. (SeeKlarman and Guzick 1976, who cite J. Kavet in Influenza and Public Policy,1972.) To take a second example, communitywideuse of polyvalent pneumococcal vaccine (pneumovax) for persons aged 45 to 64 would cost about $5,700per year of healthy life gained; if pneumovax were to be administeredto persons age 65 and older, the cost would decrease to $1,000per year of healthy life gained (Willems and Sanders 1981).Is $5,700 per year of life gained, or $1,000,or $500 a reasonable expenditure? And from whose standpoint? Theanswer will differ, not only in relation to the value placed ona year of life, but also to the availability of resources. The cost of administeringpneumovax would exceed the total health care budget in many developing countries; for the UnitedStates, universal pneumovax immunization for those 65 and older wouldrequire some 0.2 percent of current national health care expenditures. Nor does the economic analysis ene llere. The elderlypatients for whom vaccination protects against prematura death (an event without "cost" to the health care system or to economic productivity becausemost are retired) live into additional years of risk for other diseasee that willentail medical costs. If costs from disability during theseyears are entered into the computation, the "price" per year of life gained from influenza vaccinationincreases manyfold (Office of Technology Assessment 1981).Indeed, from a macroeconomic perspective (Gori and Richter 1978), to the extent thatpreventive measures increase the number of elderly Americans, they will addto social security costs and to medical costs because of the infirmities ofage. None of this diminishes the case for prevention whenit is weighed on a scale of humane values; health benefitsmay well be worth added costs. But it does illustrate the hazard of accepting official rhetoricthat makes cost control the primary justification for prevention,as President Carter (1979) did when he wrote that health promotion and disease prevention "cansubstantially reduce both the suffering of our people and the burdenon our expensive system of PUBLIC POLICY 129 medical care." What is proffered on the flawed premise of cost control may just as easily be denied when it becomes evident that it adds to expenses. The danger of relying on a claim of cost savings as a principal justification for prevention applies no less to children than to the elderly. Although vaccination against measles and rubella (and other childhood infectious diseases as well) yields dollar savings as well as health benefits, this is not true for disorders of much lower prevalence and preventive methods less dramatic in their effects. Such a situation is illustrated by an economic analysis of screening programs to detect asymptomatic lead poisoning in preschool children (Berwick and Komaroff 1982).Is the investment in screening and treating more or less "costly" than paying for medical care, special education, and institutionalization in the absence of screening? For screening by the free erythrocyte protoporphyriii assay, total dollar costs are lower than those in its absence only when the prevalence of lead poisoning among preschoolers in the community is 7 percent or higher. When prevalence is less, it is "cheaper" for the health budget to provide care after the fact. It is clearly not cheaper for the poisoned children and their families, whose costs in the form of encephalopathy, learning disabilities, mental retardation, and personal suffering are not weighed in the formal analysis. There is an even more fundamental issue at stake. Screening is designed to detect toxic lead levels early in order to minimize biological effects; it does not reduce initial risk. And that risk is present at blood levels much lower than those previously thought to be necessary (Davis and Svensgaard 1987). Primary prevention (controlling lead in industrial effluents, requiring lead-free gasoline, and mandating the prophylactic rehabilitation of housing stock contaminated by lead paint) has been deemed "too costly" by industry and has not been required by Government. A meaningful cost-benefit analysis of prevention must distinguiA between the costs borne by victims, those borne by the health budget, and those assignable to the commercial ventures that contribute to the ubiquity of lead in the environment. Cost-benefit analysis can be a useful and informative exercise. There is, for every prevention program, a point at which the size of the investment will not be warranted if the health benefit is limited to relatively few. Further, it is important to compare the costs of alternative methods for achieving the same goal.It is morally unacceptable, however, to base decisions on economic considerations alone, without taking into account human costs and the differential distribution of benefits and costs in the various sectors of the population. 130 PREVENTION OF MENTAL DISORDERS

Paradigms for Prevention

A world ends when its metaphor has died. Archibald Mac Leigh The protection against disease afforded by vaccination is commonly taken to be the ultimate paradigm for prevention.The World Health Organization (WHO) campaign against smallpox provides the most spectacular example of its success. In 1967, when the campaign was initiated, 15-20 million new cases and some 2 million deaths were occurring annually in 31 countries with endemic foci; the last clinical case was reported in Somalia in October 1977 (Bremen and Arita 1980). The variola virus now exists only in WHO-certified,Bemire con- tainment laboratories in Atlanta and Moscow. Now that it has proven possible, by recombinant DNA techniques, to insert segments of the variolagenome int,) noninfectious bacterial plasmids, there is no longer a scientific need for maintaining stocks of the virus (Dumbell 1987). For the first time in history, a mtkjor disease has been entirely eliminated by planned human action; clinical smallpox will notoccur again unless the virus is deliberately reintroduced into the environmentas a biological weapon. Worldwide economic benefits are estimated at more than $1 billion eachyear; the swings to the United States each month (bectrise vaccination isno longer necessary) exceed the total U.S. contribution to the WHO effort (Henderson 1987). How applicable is this model as a prototype for prevention? Its generalizability is severely limited because of the unique features that characterize the epidemiology of smallpox: Transmission is person-to-person and relatively slow; there is no known animal reservoir for the virus; there is no human carrier state (i.e., a person who is asymptomatic and infectious); patients are infectious only when the rash appears and only until the last scab has separated; only one serotype exists, and immunity following vaccinationor recovery from infection is long lasting; furthermore, immune persons can be recognized by visual inspection because of permanentscars from vaccination or infection. This unique combination of attributes made it possible to eliminate the disease when three developments were pitin place: (a) the production of a heat-stable freeze-dried vaccine that retainetl its potency in tropical climates; (b) a new public health strategy, once mass immunization had reduced incidence to low levels, based on the isolation of identifiedcases and the prompt vaccination of all susceptible contacts; and (c)an international commitment for the necessary funding to the WHO. No other infectious disease has the precise combination of features that made the eradication of smallpox possible.Although measles and poliomyelitis viruses have no animal reservoir, carrier states do exist, the vaccinesare less immunogenic, and determination of immune status dependson access to medical records or serologic testing. Despite thesuccess of vaccination in markedly reducing the incidence of these diseases, their complete eliminationcontinues PUBLIC POLICY 131 to elude ublic health authorities; periodic recrudescences continue (Gustafson et aL. 1987; Nkowane et al. 1987). Other infectious diseases present even more difficult challenges because of high rates of antigenic mutation, primary vaccine failure, the waning of immunity with time, transmissibility through food or water, and animal reservoirs for the agent. There is a still more fundamental limitation to the vaccine model of prevention. It aprlies superbly well when the cause of a disease is exposure to a specific transmissible agent capable of inducing long-iived immunity to later challenge. It is irrelevant to intoxications (e.g., lead poisoning) in which host resistance against reexposure is weaker rather than stronger after an initial insult because of cumulativo burden. An equally telling instance is provided by the nutritional needs of infants and children. An adequate diet provides "specific protection" against malnutrition. When food is supplied in time and in the appropriate dose, it is the "specific medical treatment" for incipient malnutrition. Yet, proper feeding at age 1 provides no immunity against an episode of malnutrition at age 2 or age 5 if starvation occurs. Moreover, prolonged protein-calorie deprivation in early life hot only results in an increase in immediate morbidity and mortality among affected infants, but it also leads to retarded cognitive and social development, particularly when it is conjoined with disadvantageous family circumstances (Richardson 1976). Under such conditions, protein-calorie renourishment does not suffice for repair, an observation that highlight; the interaction between biological and social insults to the developing organism. The likelihood of recovery is enhanced if long-term social stimulation is added to sustained food supplementation (Dabbing 1987). However complete the recovery from the initial episode, susceptibility to malnutrition is lifelong, although its effectsvary with age.The vaccine paradigm is entirely inapplicable to developmental processes. In like fashion, the psychosocial needs of infant, child, and adolescent differ during development, but those needs must be met throughout the life-span (Eisenberg 1977). Development is an epigenetic process; successful negotiation of one stage makes successful passage through the nextmore likely, but does not assure it. The child faces new adaptive challenges at each higher level of behavioral organization and requires new as well as continuou6 inputs. Thus, although a pathologic event can derail the development process,no single event can confer enduring protection against the vicissitudes of subsequent experience. Nevertheless, the search for "psychosocial immunization" continues. The concept of imprinting, derived from methodologically flawed observations of a limited set of avian and ungulate species (Gottlieb 1976), entered the clinical literature as "bonding" theory: Immediate postnatal skin-to-skincon- tact between infant and parent is held to be essential to the full development of parenting behavior. The very term bonding suggestsa permanent epoxy junc- tion. Klaus and Kennell (1976) have argued: "There isa sensitive period in the first minutes or hours of life during which it is necessary that the mother and 141 132 PREVENTION OF MENTAL DISORDERS father have cicee contact with their neonate for later development to be optimal" (italics added). That contention, coinciding with a thrust from the women's movement for the humanization of the birthing process, helped give impetus to laudable changes in obstetrical practice. However, not only is there no compelling evidence of long-term effects (Lamb and Hwang 1982), but the concept of bonding has aroused guilt and distress in parents who believe they have lost a crucial moment that can never be recap- tured if early contact has been precluded by medical necessity or hospital rules. Child abuse has been attributed to a failure of bonding; yet when the hypothesis was put to test in a comparison of abused with well cared-for infants born to primiparous mothers of low socioeconomic status, there wereno discernible differences betweer, thc two groups in the amount of early contact (Egeland and Vaughn 1981). The emphasis on bonding has led some (Lozoff and Brittenham 1979) to infer that current patterns of infant care, consequentupon new roles for women, may be harmful to both infants and mothers, a serious allegation to level without strong evidence at a time when half of all mothers witha child under 1 year of age are in the labor force. And there isa further cost: Each unwarranted claim adds to the widespread skepticism that greets proposalsfor psychosocial intervention, even when they are well founded. The vaccine model has no applicability to the noninfectious diseases that have become the major health problems in the industrialized world: heart disease, cancer, stroke, and violence (accidents, homicide, and suicide). They are multi- factorial in causation and often have genetic anlagen. Theyare behaviorally mediated; that is, they occur at higher frequencyas a consequence of smoking, drinking, overeating, and other health-injurious be'haviors. That behavior isa product of social forces: the dominant culture; effects of class and caste; and the actions (and inactions) of government as it fostersone or another lifestyle by its regulatory and tax policies, either by designor inadvertently. For example, legislation mandating a 55-mile-per-hour speed limit,a measure taken to conserve gasoline at the time of the OPEC oil crisis, haltkA and reversed what had been a steady increase in highway fatalitiesover several decades. Now that the crisis no longer pervades public consciousness and speed limits havebeen raised, a resumption of the previous pattern of increasing deathsis now inevitable. With the possible exceptior -fcancers that are of viral origin, such as hepatocellular carcinoma (Beasley and Hwang 1984;Sherman and Shafritz 1 and are thus preventable by immunization with hepatitisB vaccine, p.ition of the "diseases of affluence" demands major changes in lifestylein large segments of the population, a very different challenge for publichealth from one-shot, high-tech interventions. Not only is the learch fora "vaccine" futile, but also it distracts attention andresources from the much more difficult, long-term effort that is needed. 142 PUBLIC POLICY 133

Assessing Side Effects

"I can't believe that," said Alice. "Can't you?" the Queen said in a pitying tone. "Try again; draw a long breath, and shut your eyes." Lewis Carroll, Through the Looking Glass The promise of prevention is alluring. After all, what can be wrong with an effort to prevent disease, so long as it is well meant, even if it may not attain its goal? The short answer is a great deal. Not only may the effort fail to prevent disease, it may cause disease in those who are not included within its scope; it can harm those who are included; and it may have negative effects on the public at large. Toxicity is not always foreseeable, but it is imperative (a) that all proposals be scrutinized for potential side effects before they are implemented and (b) that evaluation schemes include surveillance for side effects. Failure to attain the expected goal of prevention represents a waste of resources and entails an opportunity lost by preampting alternative uses. Those presumed to be benefiting may be denied more effective interventions; the community at large may neglect further stud7 of the problem on the presump- tion that a solution has been found. The Cambridge-Somerville Delinquency Prevention Project (Powers and Witmer 1951), the largest undertaking of its kind, operated from 1938 to 1945.Its design incorporated the accepted methodology of the day. Individual guidance, counseling, and case work therapy was provided to several hundred underprivileged boys. At the completion of the study, the counselors concluded that the program had brought substantial benefit to about two-thirds of the boys included; more than half of the boys themselves reported that they had been helped. The study was unique for the 1940s in that it maintained an untreated control group, whose members had been matched individually with those in the experimental group. A comparison of outcome for experimentals and controls, based on the number of court appearances and the number of offenses, demonstrated no significant between-group differences (Teuber and Powers 1953). Absent a control group, the impression of clinicians would have gone unchallenged. The study was not a wasted effort precisely because it did include a control group; the findings forced a reconsideration of the received wisdom in the field of delinquency control. However, the experimental subjects not only derived no benefit, but 30 years later, the followup data reported by McCord (1978) imply that they may even have suffered negative side effects. A prevention program can harm those who are not included. Rubella vaccination, to employ an instance cited earlier, illustrates this phenomenon. Rubella vaccine affords protection primarily to the offspring of those to whom it is given by preventing CRS in that generation. It may harm the offspring of those to whom it is not given because it postpones the age at natural infection among the unvaccinated. If it is offered to young children, and levels of uptake are low (i.e., less than 90 percent), a failed eradication program can cause 134 PREVENTION OF MENTAL DISORDERS increased rates of CRS. More cases will now appear among the infants of the unvaccinated, although those of the vaccinated population will have been protected (Knox 1984).For pertussis vaccine, the opposite sc-mario can be constructed. The greater the extent of immunization in the community, the less the risk of infection for the unimmunized. Because the vaccine carries a small but not negligible risk, the child best off would be the one left unvaccinated, and therefore running no risk of a vaccine reaction, when all other children had been immunized. The flaw in the scenario is that it applies equally to each child in the community; if all parents refuse vaccination for their children on such grounds, all children are at far greater risk because the complications of pertussis far exceed in frequency and severity the complications of immunization. An ill-designed prevention program can harm those for whom it is designed to benefit. Efforts to control drug use, to screen for sickle cell disease, and to prevent suicide illustrate the range of risks. During the perceived drug crisis of the late 1960s and early 1970s, an aroused public insisted that something must be done. The authorities agreed that illicit drugs are dangerous to health, and citizens believed that warning children and adolescents about the risk should, on logical grounds, deter use. In consequence, school-based drug education programs were rapidly put into place in many states. Because education must be good for children and, at the least, can do no harm (or so it was thought) evaluation was rarely built into school programs. When evaluations were finally undertaken, the findings were dismaying. Not only did "education" prove to be ineffective, but, among some subgroups, it was associated with increased drug use (Durrell and Bukoski 1984). By then, however, the politicians and the media had moved on to other themes and prevention of drug use was put on the back burner. In the 1980s, the drug problem seemed to mmerge. This time health educators were aware of the ineffectiveness and the hazards of purely didactic instruction. The new strategies emphasized peer support, skills in "saying no" to solicitations for drug use, and learning alternative ways of coping. These methods had proven highly useful in school-based antismoking programs (Bot- yin and Eng 1982; Schinke et al. 1985; Telch et al. 1982). Although the most recent outcome studies are somewhat more encouraging than earlier efforts, gains have been modest at best (Schaps et al. 1986). Urgent as the prevention of drug abuse may be, we have yet to learn how to do it effectively. Newborn screening for sickle cell anemia poses different risks for the families of infants found to be positive for the disease or the trait. The infant with sickle cell anemia will benefit only if parents understand the meaning of the diagnosis and if medical care is available from physicians knowledgeable about treatment. Even though early diagnosis, followed by appropriate medicalcare, including prophylaxis against pneumococcal infection, brings clear beneflt, parentscon- tinue to encounter inadequately informed physicians (Rowley and Huntzinger 1983). Further, as in the case of every recessive genetic disorder, testing parents

144 PUBLIC POLICY 135 carries the risk of discovering that the husband is notthe father of the child and placing the marriage at hazard. The infant at greatestrisk for harm without offsetting benefit is the one with sickle cell trait. Undue parentalapprehension may lead to a vulnerable-childsyndrome (Green and Solnit 1964); as an adult, the individual may be denied employment, insurance, or amilitary career. A screening program in Greece ended in tragedy for femalecarriers; young women ideutified as having the trait were no longer consideredmarriageable ar 1 had to leave their village (National ResearchCouncil 1975). Because blacks are the major population reservoir for the sickle gene in theUnited States, the genetic counseling that accompanies scmening has led to charges ofgenocide, when community leaders have not been consulted. Effectivefollowup, appropriate counseling, access to competent medical care, and participationby the community are essential for a viable program. Shaffer, Garland, and Bacon (this volume) have provided an incisivereview of suicide prevention programs, but the methodologic difficultiesin evaluation merit additional emphasis. What makes the challenge ofevaluation daunting is that suicide is a condition of high severity but lowincidence. Although suicide is second only to accidents as a cause of death in malesbetween 15 and 19, the mortality rate in 1983 was 14 per 100,000. Consider the difficulty inevaluating a preventive strategy thought tobe capable of reducing that rate by 25 percent and thus saving 425 lives. To detect the anticipated effect (at asignificance level of 0.05) and to be equally certain (at a 95-percent confidence level)that a real effect has not been missed, the study design would requireexperimental and control samples of 2.6 million adolescent males. If the investigator werewilling to accept a 20-percent probability of a type II error (i,e., a powerof 0.8), the requisite sample size would be 1.6 million in each group. Similar statistical problems bedevil screening for suicide. Suppose the impossible: a test with a sensitivity of 100 pbrcent (i.e., a test able toidentify all potential suicides) and a specificity of 99 percent (i.e., a test yielding only one false positive in every 100 normals). No such test is remotely possible.Yet, even if there were one, it wouldidentify the 14 positives in a population of 100,000 adolescent males at a cost of 1,000 false positives (i.e., 1 percent of the sample), a ratio of 71 false positives for every true positive. Because the label "high risk for suicide" would justify vigorous measures directed at each person identified, even so remarkably accurate a screening test would be entirely unsuitable as a public health measure because of the toxicity for false positives. Precisely such an analysis has demonstrated the folly of requiring' premarital screening for HIV infection as a madition for obtaining marriage licenses (Cleary et al. 1987). Finally, a preventive effort can be on target the individuals it enrolls and nonetheless lead to an increase in the prevalence of the hazard it is designed to avert. Because it has been evident for many years that teenagedrivers contribute disproportionately to automobile accidents, high school driver education courses have been introduced in a number ofjurisdictions. Outcome assessments have demonstrated a modest reduction in crashes per licenseddriver 136 PREVENTION OF MENTAL DISORDERS

among course graduates in comparison to those licensed without such training. Despite this apparent beneficial result, the net effect has been balefillWhy? Driver education courses increase the number of teenaged driverson ,,ne road; in consequence, they lead to an overall increase in fatal motor vehicleaccidents (Robertson and Zador 1978). Robertson (1980) compared communities that retained andothers that dis- continued high school driver educationcourses in Connecticut after the state terminated funding. The elimination of thesecourses resulted in a substantial reduction in early licensure of 16- to 17-year olds anda parallel net reduction in serious crashes involving adolescent drivers in thecommunities that had abandoned the courses. In that state, approximatelyone in five teenagers licensed to drive at 16 will be the driver ofa vehicle in a crash causing injury or more than $400 in property damage before age 18. The relevant policy question then becomes, not whether driver trainingcourses can reduce accident rates by graduates, but whether the "right" to a driving license before 18warrants the hazards, either to the young driversor to the other citizens who use the highways. That premature introduction ofprograms purported to prevent disease can be hazardous to health is no argument against prevention;it does highlight the need for (a) a thoughtful examination of potential hazardsbefore undertaking new initiatives and (b) rigorous design to assess negativeas well as positive effects when new programs are introduced. Policymaking in a Pluralistic Society

"Would you tell me, please, whichway I ought to go from here?" 'That depends a good dealon where you want to get to," said the Cat.

"I don't much care where... solong as I get somewhere," Alice added asan explanation. "Oh, you're sure to do that," said the Cat, "if you only walk long enough." Lewis Carroll, Alice's Aduentures in Wonderland

In a remarkably perceptive articleon the political roots of the Community Mental Health Centers Act, Alfred Freedman (1967)pointed out that understanding policy formation in the United Statesrequires an awareness of those features of our society that distinguishus from other modern industrialized societies: pluralism, pragmatism,a focus on short-term goals, problem solving by accretion, and unevenness of developmentin different jurisdictions. More than a century ago, de Tocqueville recognizedpluralism as the defining characteristic of American society. Thevery checks and balances of our Constitution were expressly conceived to prevent power from falling intothe hands of a sovereign ruler.However, there is an inherent contradictionbetween our

1 fi PUBLIC POLICY 137 historical constitutional system, whichworks toward the fragmentation of power, and the growingrole of technology and industrialrationalization, which work toward its consolidation. We yearnfor the democracy of the New England town meeting even as the sheersize of our population decreases the powerof the individual. We worship pragmatism and we abhortheory. We focus on short-term goals and abjure long-range planningbecause we distrust systems and systematizers, The typical American response to asocial problem is to create a new structure to deal with it while leavingexisting institutions in place. In consequence,the new is impeded fromthe outset by the inertia of the old.Local control in a pluralistic society affords communitiesself-determination at the price of marked unevenness in policy from onecommunity to another. Splendid services may be found in one area at the same timethat access to care is abysmal in a second. There is no unambiguous answer to thequestion: What is U.S. policy on preventing or treating mental disorders?Rather, there is a nested set of competing and complementary policies that varywith local option. In the political process, the would-bereformer must weigh the odds for success if the attempt is boldagainst the greater certainty of opting forsmaller changes at the margins.Rudolf Klein (1972) contrasts the "optimizing, rationalizing" model of policymaking in ademocracy with the "satisficing" model. The former calls for constructing anefficient and logical design based on coherent theory andrisks falling afoul of the political support forexisting structures. The latter emphasizes what is"good enough"; it is cautious, incremental, and based on compromise dictatedby the conflictiag claims of competing constituencies. If the first coursegambles for the greater gain, it does so at the risk of totaldefeat; if the second is shrewder at calculating theodds for success, it gives up on fundamental reformand settles in advance for smaller gains. Incrementalists argue that in politics theperfect is the enemy of the good. Klein defines the issues in the following terms: The problem for policy makersand those who try to assessthe outcome of the processis to know whether the rightbalance has been struck between overestimating the frictional costa, and thusmissing an opportunity for improvement, and underestimating thefrictional costs, and thus creating a situation of opposition to evolvingchange. There are additional factors that lead to the divergencebc tween the attitudes toward policy formulation among academics onthe one hand and among Government officials on the other (Kash and Ballard 1987).The political agenda is such that decisions must be made in the faceof limited and often unreliable data. The academic can afford to wit:, until the facts arein. The elected official needs answers now and is impatient with thescientist's reiteration of the necessity for more research. In the political process,there is no way not to decide; not to act is to act because it preserves the status quo.Politicians tend to operate within a timefrarne set by the nextelection; yet the impact of policy extends well beyond the usual term of office.Scientists complain that research 7 138 PREVENTION OF MENTAL DISORDERS

findings are ignored, that debatesproceed in disregard of data,and that politicians are unwilling to submitsocial policies to empirical trial bycomparing alternatives in separate geographicareas. Where, however, are the constituencies willing to serveas controls or experimentals?All too often, research findings are cited in the politicaldebate when they support prior beliefand ignored when they contrw:ict it,but it should be acknowledgedthat selective citation is hardly unique to thepolitical arena. Research findings indeed oftenprove irrelevant to the politicalprocess. In part, this may be because data from socialresearch rarely settle matters beyond doubt; more often thereason is that larger political considerationspredominate. In 1965, as part of the Waron Poverty, the Johnson Administrationestablished a natioual Head Start program to providepreschool education for economically disadvantaged children toreverse the developmental attrition associatedwith poverty (Eisenberg and Earls 1975).As program costs mounted,a contract to evaluate the effectiveness of HeadStart was awarded to theWestinghouse Learning Corporation (1969). Becausethe early increases in IQ that hadbeen reported for disadvantaged childrenenrolled in Head Startwere not sustained after they entered the primarygrades, the report concluded thatthe program was a failure. The finding was used to reinforceefforts to curtail funding. The campaign failed; appropriationsfor Head Startwere sustained by the Congress and have continued to increaseeach year to the present. The methodology of the Westinghousestudy was seriously flawed. Itlumped data from programs ot eery different quality, used IQ change rather thanschool progress as the criterion for outcome, and ignoredthe empowerment of parents and the provision of healthcare for children. A decade later,longitudinal studies (Berrueta-Clementet al. 1934; Lazar et aL 1982) haveprovided impressive evidence ofprogram effectivenees: better schoolprogress, fewer dropouts, less delinquency, anda better work record after high school. TheWestinghouse research findings had little impacton a political debate that was part ofa much larger national agenda. HeadStart has a large politicalconstituency in local communities. Its emphasison the children of the poor draws supporteven from those who are dubious aboutother welfare transfer payments. Just this year, the Committeefor Economic Development (1987),an organize- tion of more than 200 business executives and educators, issueda clarion call for preventing educational failureby providing early interventionprograms, by restructuring the public schools,and oy developing retentionand reentry programs. The committee calledon the Nation to "give the highest priority early and sustained intervention to in the lives of disadvantagedchildren." The committee justified its positionby noting the threatto the competitive position of American industry if the UnitedStates does not produce literateworkers with problem-solving skills.It recognized thecost of early intervention but concluded:

1Li PUBLIC POLICY 139

If the nation defers the expense of preventive programsduring the formative years, it will incur much higher and moreintractable costs for older children that have already experienced failure.Even so, we cannot limit our efforts to only one groupof disadvantaged children; both economic and humanitarian considerations impel us to find ways toexpand our preventive efforts, improve bp:ic education for allstudents, and enhance the chances of those in and out of school who have alreadybeen failed by the system (p. 19). If intervention in early childhood is tenable associal policy because it is consonant with American mores, preconceptionaland prenatal care remain controversial because they contravene the beliefs heldby a highly vocal and politically astute minority. The incidence of teenage pregnancyand low-birthweight infants are two interconnected public health problemswith high risk for maternal and neonatal mortality and neuropsychiatricmorbidity in both mothers and infants (Eisenberg 1987). Amongindustrialized countries, the United States has the highest teenage pregnancy, abortion,and birth rates because U.S. teenagers have the lowest rate of contraceptive use(Jones et al. 1985). Although the percentage of unmarried adolescent womenhaving had intercourse is higher by half in Sweden than in theUnited States, Swedish teenage pregnancy rates are lower by half becauseSweden provides a compul- sorysex education curriculum inits schouls, closely linked to contraceptive cliniC services. U.S. studies provide evidence that school clinics lead tolower birth rates among secondary school students (Kenney 1986)and that they are associated with a delay in the age at which coitus is initiated (Zabin etal. 1986). These facts have not yet been incorporated into thefashioning of State or Federal policy on teenage pregnancy (National Reseamh Council 1987). Low birth weight is a major determinant of neonatal mortality,total infant mortality, and developmental retardation among the infants who survive(Mc- Cormick 1986). The Institute of Madicine (I0M) (1985) has estimatedthat current rates of low birth weight in the UnitedStates could be reduced by 15 percent among whites and 12 percent among blacks if all womenbegan prenatal care in the first trimester of pregnancy and continued toreceive care through delivery. Yet, since 1978 the proportion of women in the United States not receiving care until the third trimester or receiving no careat all has remained unchanged. What has been missing is a national commitment to abolish the barriers to care. The problem persists, not because the knowledge needed for action is lacking, but because the social will to act has not yetbeen mobilized (Richmond and Kotelchuck 1986). Though further research is necessary to improve on present capabilities, the prime needis political: to create a national consensus on the importance of universal access to prenatal care. In addition to targeted programs, Government policies that influence the availability of health care to the poor have an important impact on child development. Following the passage of Medicaid legislation in 1965, there was a marked improvement in health care amongthose living in poverty; the number of physician visits per person per year among the poor increased from 3.8 in 1AU 140 PREVENTION OF MENTAL DISORDERS

1964 to 6.2 in 1978; similar changeswere seen in rates of hospital use. Concomitantly, infant mortality rateswere reduced almost by half during that period, whereas they had plateaued between 1960 and 1964(Rogers et al. 1982). In the past 4 years, cutbacks in health care financing have resultedin a decrease in the number of physician visits, particularly markedamong low income and minority families (Johnson Foundation 1987).The number of Americans without health insurance has risen to 35 million. Thue flu, the policiee examined have been direct ly concernedwith health care services. However, Federal and State policies thatinfluence the general social welfare of the population have a major impacton the public health. Indeed, McKeown (1976) has marshalled persuasive evidence thatthe mikjor gains in health over the past two centuries reflectimprovements in overall living standards rather than medical advancesper se. Consider the effect of tax policyon the disposable income available to the various sectors of our population.Over the past decade, the percentage of earnings consumed by Federal taxes has increasedby 20 percent for the poorest one-tenth of Americans and has decreased byan equivalent percentage among the richest one-tenth according toa recent analysis by the Congressional Budget Office; 80 percent of families haveseen their real incomes decline in the past 10 years (Associated Press 1987). One in four of the nation's 62 million childrenunder 18 lives with onlyone parent, almost always the mother. Of children livingin such female-headed households, more than a thirdare reared in poverty; for black families, the proportion in poverty is twice as high (U.S. Bureauof the Census 1987). Single mothers, beset by poverty as wellas lack of social support, have difficulty meeting the needs of their children. Developmentallyappropriate day care services for these and other youngsters inthe United States are lamentably inadequate (Johnson 1987; Zigler and Gordon1982). The need for day care is not limited to singleparent families. Whetheror not the family is intact, half of all mothers witha child less than 1 year of age are in the labor force. But can daycare for such young infants be "developmentally appropriate"? In earlier decades, studiesundertaken to examine this question found few significant differences between childrenreared at home or in daycare; more recent studies suggest that even relatively good infantday care may be associated with insecure attachment andpersonality deviations (Gambleand Ziegler 1986). The issue remains controversial;results vary from study to study and are cited selectively ;Phillips et al. 1987);many rely heavily on a laboratory measure (Ainsworth's Strange Situation) whose long-termpredicGve validity remains to be established. But there isa deeper problem. What is largely missing is attention to individual differencesin intrafamilial experiencesas well as to differences in child care experiences (McCartneyand Galanopoulos 1988). It may well be the interaction betweenthe two that is decisive for developmental outcome. Nonetheless, it would be irresponsibleto dismiss out of handconcerns PULLIC POLICY 141 about deleterious effects of infant day care. The mattermerits high priority on the research agenda of child psychiatry. It would be equally irresponsible to ignore the currentsocial context of infant day care, a context in which mothers may need towork to provide adequate family income. The alternatives that few politicians seemwilling to consider are paid infant-care leavesand family allowances to permit mothers (or fathers) who prefer to stay home to do so vAthout beingreduced to poverty, a policy approach commonplace in most European countries.Until the United States movee in that direction, asit almost certainly must, the crying need b for thA best day care we know how to designcarethat should be made available without regard to income but whose cost can be indexed toability to pay by means of direct subsidyand income tax deductions. It has become part of the conventional wisdom to assertthat Government welfare policies are primarily responsible for the decline in theproportion of intact families and "legitimate" births among the poor,particularly the black poor. However, carefulexamination, State by State, of the correlations among Aid for Families with Dependent Children supportlevels, marital dissolution, and out-of-wedlock births by Bane and Ellwood (1983) hasrevealed that welfare is not the underlying cause of the dramatic change infamily structure.If welfare policy is not the cause, neither is its current version("workfare") a viable solution to a problem whose source lies in increasing blackmale joblessness. Wilson (1987) has explored the relationship between employmentand marriage by computing a "male marriageable pool index": the numberof employed men per 100 women of the same ageand race. In the mid-1960s, the black index was about the same as thewhite; in the years since, it has plummeted, particularly in areas of the country that experienced economic stagnation precisely the areas in which marked increases among blackfemale-headed families have occurred. Industries where blacks had beenheavily employed (automobile, textile, rubber, steel, and meat packing) have borne thebrunt of plant shutdowns and deindustrialization. In 1974, 46 percent of youngblack males (age 20-24) were in the higher paying semiskilled and skilledblue collar positions; by 1986, that figure had dropped to 26 percent. Considerthe marriage prospects for a young black woman growiiig up on Chicago'sSouth Side with only 18 employed males for every 100 females age 16 or over. As Wilson(1987) concluded: 'The tragic decline of intact black households cannot be divorced from the equally tragic decline in the black male 'marriageable pool' in any serious policy deliberations on the plight of poor American families" (p. 106). Wilson's analysis uf the situation of disadvantaged Americans led him to conclude that a resolution of the growing crisis demands a comprehensive program of economic and social reform; macroeconomicpolicies to promote balanced economic growth and create a tight labor market, a national labor- market strategy, child support assurance, a child care strategy, and a family allowance program. Countries that rely the least on public assistance (Sweden, West Germany, and France) employ alternative and less invidious methods for

1e-1 142 PREVENTION OF MENTAL DISORDERS

income transfer family allowances, housing subsidies,child support and unemployment assistance, child care services, and daycare programs. Moreover, these countries emphasize labor market policies designedto enhancA high employment (Kamerman and Kahn 1982). Government policy has decisive effectson the social conditions in which children livethe context that determines theirlife char-es. Those concerned with preventing developmental attritionamong children and with optimizing the likelihood of favorable adolescent outcomescannot limit their focus to health services per se, importantas those services are. We must become forceful advocalas for social justice. Twentyyears ago, the Kerner Commission (1968) concluded: This nation will deserve neither safetynor progress unless it can demonstrate the wisdom and the will to undertakedecisive action against the root causes of racial disorder (p. 34). We have yet to learn that lesson.

Public Policyon Psychiatric Research

Human knowledge and humanpower meet in one; for where the cause is not known, theeffect cannot be produced. Nature to be commandedmust be obeyed. Francis Bacon Our hope of havingmore effective methods for prevention in the future depends on systematic researchon (a) the epidemiology and the pathogenesis of mental disorders and (b) controlledcommunity trials of promising preventive methods. What has been the pattern of publicsupport fcr medical research in the United States? By far the largest source of funding forresearch in child mental health is the Federal Government via two leadagencies:file Alcohol, Drug Abuse, and Mental Health Administration (ADAMHA)and the National Institute of Child Health and Development (NICHD). Becausethis year marks the centenary of the National Institutes of Health (NIH),a significant event in the history of American medicine, it is appropriateto review the pattern of support for biomedical research in general beforeexamining the fraction devotedto mental health. Although a National Institute of Healthwas first created in 1930 (as successor to the Hygienic Laboratory established at the U.S.Marine Hospital on Staten Island in 1887), its initial fundingwas quite modest. Not until the years following the Second World War didFederal support for medical research become substantial. Between 1956, whenFederal appropriations for NIHwere $98 million, and 1959, they tripledin resporto the efforts of a new health science coalition (Shannon 1987); bythe late 1960s, funding passed the PUBLIC POLICY 143 unprecedented $1 billion mark. It began to slow in the late 1960s; appropriations actually fell (in dollars corrected for inflation) during somebudget years in the 1970's and again in the early 1980s. Allocations for health researchand development from all sources in the United States remained about level in constant dollars between 1975 and 1983 (Office of ProgramPlanning and Evaluation 1986:4). At the initiative of the Congress, NIH funding over the past 5 years once again attained sustained growth, amounting to 70 percent in dollars appropriated and 28 percent in constant dollars; it reached $6.2 billion for fiscal year (FY) 1987 (Wyngaarden 1987).lic.,,Ni;iver, FY 1988 research budgets will be hard hit by Gramm-Rudman deficit-reduction legislation; dollars available will be les, 'than those appropriated for FY 1987 despite earlier congressional actions toward a 10-percent increase. The National Institute of Mental Health (NIMH), a component of NIH from its founding in 1946 until it was split off in 1974 together with the National Institute on Alcohol Abuse and Alcoholism (NIAAA) and the National Institute for Drug Abuse (NIDA) to form ADAMHA, began more modestly. The NIMH research budget did not reach $100 million until 1966. Had the NIMH research budget kept pace with inflation or paralleled the growth at NIH during the 1970s and 1980s, it would have exceeded $300 million by 1983 (Institute of Medicine 1984); however, because of the lower priority assigned to mental health, the actual allocation was $158 million, one-sixth of the amount awarded to the National Cancer Institute and one-fourth of that awarded to the National Heart, Lung, and Blood Institute in that year. Are there rational guidelines to determine appropriate resource allocations for medical research? In the heady years of the 1950s and early 1960s in the United Statrl, with the gross natior.r..; produd (GNP) increasing each year, little thought was given to the sustainable limits to expansion in the research enterprise. In the late 1970s and 1980s, at a time of budget deficits, a slowdown in growth of the GNP, and an unfavorable trade balancethe question of limits became prominent in policy debates. In 1976, the President's Panel (Murphy and Ebert 1976) argued:

In other fields of technological endeavor . .. it is customary to invest between 5 and 10 percent of the total budget on research and development.

...At the present time the health industry as a whole invests a considerably smaller percentage in research. ...While 5% would represent an abruptly large increase if committed overnight, it seems to us a rational percentage to head toward as a long-range goal. In 1976, the total investment in health research and development amounted to 3.6 percent of total health costs; the estimate for 1985 was 3.1 percent (Office of Program Planning and Evaluation 1986). U.S. health care expenditures for 1987 have been estimated at more than $500 billion. If the panel's recommendation for a 5 percent setaside for health research had been in effect, that would have justified an allocation of some $25 billion. What are the actual figures likely to be? During the past Vi3 144 PREVENTION OF MENTAL DISORDERS decade, the NIH budget has provided from 35 percent to 40 percent of all national support for health research and development, with other Federal sources providing 15 percent to 20 percent and industry about 30 percent to 39 percent (Office of Program Planning and Evaluation 1986). If similar ratios obtain in 1987 (an uncertain assumption) at an NIH budget of $6.2 billion, total support will equal some $16 billion, about 3.2 percent rather than 5 percent of "industry" costs. Within the health research budget, how are priorities to be assigned for allocations to particular disease problems? A rational approach would be based on a close analysis (a) of the scientific opportunity for discovery in a given area (the, availability of promising new concepts and reliable methods to explore them) and (b) of the health burden produced by the diseases under consideration. The Board on Mental Health and Behavioral Medicine of the IOM (1984) has made a persuasive case that psychiatric research is grossly underfunded in relation to progress in neuroscience and social science as well as to the health burden produced by mental disorders. In 1980, mental disorders entailed direct health care costa of $20 billion (without taking into account their contributions to morbidity from cirrhosis, drunk driving crashes, chronic pain syndromes, etc.), exoeeded in aggregate expense only by costs resulting from circulatory and digestive diseases. A 5 percent set-side rule would have warranted $1 billion for ADAMHA research; the actual figure did not reach half that amount for all three institutes under its aegis until 1987.

Table 1. Research on Mental Health (In Thousands)

Fiscal Year NIMH NIDA NIAAA Total

1983 $158,711 $47,501 $33,022 $239,234 1984 173,109 55,540 42,590 271,239 1985 193,328 63,760 i, i,5O9 305,597 1986 204,148 70,553 54,372 329,073 1987 246,746 133,100 71,235 451,081

Note: NIMH = National Institute of Mental Health; NIDA= National Institute for Drug Abuse; NIAAA = National Institute on Alcohol Abuse and Alcoholism.

I r; 1 PUBLIC POLICY 145

The IOM called for appropriations of$300 million for NIMH and $100 million for each of the other two institutes (in 1983dollars). Actual appropriations from 1983 through 1987 ere listed in table 1(Regier 1987). For 1987, ADAMHA research budgets, converted into 1983 dollarsby using the NIH Biomedical Researk:m and Development Price Index, were theequivalent in 1983 dollars of $198 million for NIMH, $107 million for NIDA, and$57 million for NIAAA, not quite three-fourths of the total the IOM hadrecommended 4 years earlier. And what of research in child mentalhealth? Each of the three ADAMHA institutes supports child mental health research.Table 2 lists amounts allocated between 1983 and 1987.

Table 2. Research on Child Mental Health (In Thousands)

Institute 1983 1984 1985 1986 1987

Mental health $34,194 $34,475 $42,527 $44,327 $43,380 Drug abuse 3,881 5,035 5,288 5,600 5,900 Alcohol 1,800 tuil 4,400 5,400 7,900 ADAMlig Total$39,875 $43,510 $52,215 $55,327 $57,180

aADAMHA = Alcohol, Drug Abuse, and Mental Health Administration.

Table 3 reports total NICHD appropriations for the same fiscalperiod, as well as those for behavioral researchand training, the component most closely related to mental health (Krasnegor 1987). Table 4 lists the ADAMHA allocations for child mentalhealth research and the NICHD allocations for behavioral research for the period1983 through 1987, the 1987 figures being provisional. They represent just under13 percent of the total ADAMHA and 21 percent of the total NICHDbudgets for FY 1987. The apparent increase, however, is deceptive. When thetotals are converted into constant 1983 dollars by use of the NIH BiomedicalResearch and Develop- ment Index, which takes inflation in the cost of doingresearch into account, the combined figures are as shown in table 5. Thus, the increase overthe 6-year period is 18 percent in constant dollars rather than the 48 percentsuggested by the dollars appropriated that are listed in table 4.

1 5 5 146 PREVENTION OF MENTALDISORDERS

Table 3. National Instituteof Child Health andDevrlopment Support for BehavioralResearch and Traini (In Thousands)

Fiscal Year Behavioral as Total Support Behavioral Support % of Total

1983 $208,482 $ 38,118 18 1984 225,605 43,513 19 1985 258,749 60,198 19 1986 257,563 52,876 21 1987 273,015 58,062 21

How do those amounts compare to what would have beenjustified by the formula suggested earlier:a 5-pereent setsside basedon health care costs? Unfortunately, thereare no data available on the totalcosts entailed by treating childhood psychiatric morbidity. The Office of TechnologyAssessment study (Dougherty et al. 1987) of childmental health services the total costs for was unable to ascertain care despite searching available datasets. Of the totals allocated for research, what proportion was assignedto prevention? The only agency reporting data on prevention research is NIMH. As notedin table 6, the allocation for prevention hasbeen about $5 millionper year in recent years, some 12-16 percent of the total NIH budget for child mental healthresearch.

Table 4. Research on Child Mental Health andDevelopment (In Thousands)

Year ADAMHA N1CHD Total

1983 $ 39,875 $ 38,118 $ 77,993 1984 43,510 43,513 87,023 1985 52,215 50,198 102,413 1986 55,327 52,876 108,203 1987 57,180 58,062 115,242

Note: ADAMHA= Alcohol, Drug Abuse, and Mental Health Administration; NICHD= National Institute of Child Healthand Development. PUBLIC POLICY 147

Table 5. Research in Constant Dollars (In Thousands)

Year Total (ADAMHA & NICHD)

1983 $ 77,993 1984 81,712 1985 90,977 1986 91,947 1987 92,297

Note: ADAMHA = Alcohol, Drug Abuse, and Mental HealthAdministration; NICHD = National Institute of Child Health and Development

These facts make it abundantly clear that the U.S. national effortin research on the prevention of mental disordersfalls far short, when judged against either the public health burden or the scientific opportunitiesprovided by the promising new developments reviewed elsewhere in this-7olume. Against those benchmarks, what is warranted is a research budget for ADAMHA on theorder of $1 billion rather than the $500 million slated for 1987(before Gramm- Rudman cuts were implemented). Though this account has been limited to expenditures internal to thehealth sector, the fight for melical research and the applicationof its findings demands attention to broader questions of tax policy and publicexpenditures.For example, doubling the excise tax on cigarettes to 32 cents a packwould yield an additional $3.1 billion in Federal revenues yearly; raising the excise tax onbeer and wine to the rate for distilled spirits would yield $5.8 billion more.In view of the evidence that substantial price increases decreasethe consumpt:on of these substances (Warner 1986), such a tax policy in itself would promotehealth at the same time it would make available tax revenuessufficient to more than double total current NIH and ADAMHA budgets. If cigarette and alcohol tax rates were to be indexed to the rates of inflation, the$1 billion generated in each of the next 3 years would exceed the expected annual increasesin NIH and ADAMHA budgets. On the expenditure side, we must recognize that what is spent for defense or to pay the interest on the Federaldeficit is simply not available for the improvement of health care (Eisenberg 1984). 148 PREVENTION OF MENTAL DISORDERS

Table 6. Child Research, National Institute of Mental Health, 1984-1987 (In Thousands)

Type of Research 1984 1985 1986 1987

CADR.Ba $ 7,298 $ 9,442 $10,396 $13,283 Clinical centers 519 836 585 580 Prevention 4,699 5,022 4,476 4,902 Epidemiology 1,132 2,090 2,169 2,212 Basic and developmental 7,156 8,780 15,849 10,524 Special 3,593 5,285 4,000 4,300 Intramural _5.01ii 5,536 3,426 _Mg Totals $29,386 $36,991 $40,901 $39,590

°All research projects listed underthe Child and Adolescent DisordersResearch Branch (CADRB) deal with DSM-III or DSM-III-Rdisorders. Summary

Policies designed to promote healthand to prevent disease should be evaluated by the health outcomesthey yield in the population.Advocating prevention on the promise of costcontainment is a dangerous strategy.Preven- tion is sometimes less costly forthe Government budget thancure; more often, effective prevention offers better healthat additional public cost. Few quick fixes lie in store for thepsychiatric disorders of children. Inlarge part, their prevention requires thecreation of a social environmentresponsive to the long-term psychobiologicalneeds of the developing child.That will be achieved neither cheaplynor easily because it demands confrontingand overcoming inequities and prejudicesdeeply entrenched inour society.

I ri PUBLIC POLICY 149

Good intentions are not enough. Prematureimplementation of well-meant but ill-designed programs may exacerbate the veryproblems they are intended to correct.Effective prevention requires a firm knowledgebase, a carefully considered political strategy for implementation, and thecreation of social consensus on the necessity for action. If we are to have better programs tomorrow than we canmount today, the United States must invest more heavily in psychiatric researchthan it has done in ita recent history. Given the illnest, burdenproduced by mental disorders and the scientific opportunities for discovery, thereis strong justification for doubling current Federal allocations for psychiatric research. As child and adolescent psychiatrists, we carry a specialresponsibility as citizens precisely because of what we know about thedevelopmental needs of children. To act in accord with the moral basis of ourprofession, we must become advocates for child health in every public forum available to us.That role has a long tradition in public health. Hermann M.Biggs, commissioner for health for New York State, wrote in the department's monthlybulletin in 1911 (Winslow 1929): Disease is largely a removable evil.It continues to afflict humanity, not only because of incomplete knowledge of its causes andlack of adequate individual and public hygiene, but also because it is extensivelyfostered by harsh economic and industrial conditions and by wretchedhousing in congested communities. These conditions and consequently thediseases which spring from them can be removed by better socialorganization. No duty of society, acting through its governmental agencies, is paramountto this obligation to attack the removable causes of disease. . .Public health is purchasable.

r") 150 PREVENTION OF MENTAL DISORDERS

References

Associated Press. Tax burden is found to rise forpoor in U.S. The New York Times, N o v.12, 1987, p. A-26,

Bane, M.J., and Elwood, D.P. The Dynamicg ofDependence: The Routes to Self-Sufficiency. Washington, D.C.: U.S. Departmentof Health and Hurnan Services, 1983. Bart, K.J.; Orenstein, W.A.; Preblud, S.R.; and Hinmen,AR. Universal immunization to interrupt rubella.Reviews of Infectious Disease 7(Supple- ment):5177-184, 1985. Beasley, R.P., and Hwang, L.Y.Hepatocellular carcinoma and hepatitis-B virus. Seminars on Liver Disease 4:113-121, 1984. Berrueta-Clement, J.R.; Schweinhart, L.J.; Barnett,W.S.; Epstein, AS.; and Weikart, D.P. Changed Lives: The Effects of thePerty Preschool Programon Youths Through Age 19. Ypsilanti, Mich.: TheHigh/Scope Press, 1984. Berwick, D.M.. and Komaroff, A.L. Costeffectiveness of lead screening. New England Journal of Medicine 306:1392-1398,1982. Botvin, G.J., and Eng, A. The efficacy ofa multicomponent approach to the prevention of cigarette smoking. Preventive Medicine11:199-211, 1982. Bremen, J.G., and Arita, I.The confirmation and maintenance ofsmallpox eradication. Science 164:262-270, 1980.

Carter, J. Preface to Healthy People: TheSurgeon General's Reporton Health Promotion and Disease Prevention. Washington,D.C.: U.S. Gavt. Print. Off., 1979, p. vii. Centers for Disease Control. Rubella andcongenital rubellaUnited States, 1984-1986. Journal of the American MedicalAssociation 258:2491-2500, 1987. Cleary, P.D.; Barry, M.J.; Mayer, K.J.;Brandt, A.M.; Gostin, L.; and Fineberg, H.V. Compulsory pre-marital screeningfor human immunodeficiency virus. Journal of the. American Medical Association258:1757-1762, 1987. Committee for Economic Development.Children in Need:Investment Strategies for the Educationally Disadvantaged.New York: Committee for Economic Development, 1987.

Davis, J.M., and Svensgaard, D.J.Lead and child development.Nature 329:297-300, 1987.

Dobbing, J. ed. Early Nutrition andLater Achievement. London: Academic Press, 1987.

i PUBLIC POLICY 151

Dougherty, D.M.; Saxe, L.M.; Cross, T.; and Silverman, N. Children's Mental Health; Problems and Services. A Report by the Office of TechnologyAssess- ment. Durham, N.C.: Duke University Press, 1987. Dumbell, K. What should be done about smallpox virus? Lancet ii:957-958, 1987. Durell, J., and Bukoski, W. Preventing substance abuse: Thestate of the art. Public Health Reports 99:23-31, 1984. Editorial. Immunisation policy: Recipes forsuccess. Lancet ii:78-80, 1987. Egeland, B., and Vaughn, B. Failure of "bond formation"as a cause of abuse, neglect and maltreatment. American Journal of Orthopsychiatry51:78-84, 1981.

Eisenberg, L. Preventive psychiatry. Annual Review of Medicine13:343-359, 1962.

Eisenberg, L. Development as a unifying concept in psychiatry.British Journal of Psychiatry 131:225-237, 1977. Eisenberg, L. Rudolf Ludwig Karl Virchow, whereare you now that we need you? American Journal of Medicine 77:524-532,1984. Eisenberg, L. Preventive pediatrics: The promise andthe peril. Pediatrics 80:415-422, 1987.

Eisenberg, L., and Eavls, F.J. Poverty, social depreciationand child development. In: Hamburg, D.A., ed. American Handbook ofPsychiatry, Vol. VI. New York: Basic Books, 1975, pp. 275-291.

Freedman, A.M. Historical and political roots of theCommunity Mental Health Centers Act. American Journal of Orthopsychiatry37:487-494, 1967. Gamble, T.J., and Ziegler, E. Effects of infant daycare: Another look at the evidence. American Journal of Orthopsychiatry 56:26-42,1986. Goldberger, J.; Waring, C.H.; and Willets, D.G. Theprevention of pellagra: A test of diet among institutional inmates. Public HealthReports 30:3117-3131, 1915. Gori, G.B., and Richter, B.J. Macroeconomicsof Jisease prevention in the United States. Science 200:1124-1130, 1978.

Gottlieb, G.Early development of species-specific perceptionin birds.In: Gottlieb, G., ed. Neural and Behavioral Specificity.New York: Academic Press, 1976, pp. 237-280. Green, M., and Solnit, A.Reactions to the threatened loss ofa child: A vulnerable child syndrome. Pediatrics 34:58-66,1964. 152 PREVENTION OF MENTAL DISORDERS

Gruenberg, E.M.; Lewis, C.; and Goldston, S.E.,eds. Vaccinating Against Brain Syndromes: The Campaign Against Measles andRubella. New York: Oxford University Press, 1986. Gustafson, T.L.; Lievens, kW.; Brunell, D.A.; Moellenberg,R.G.; Buttery, C.M.G.; and Schulster, L.M. Measles outbreak in afully immunized secondary school population. New England Journal of Medicine316:771-774, 1987. Henderson, DA. Smallpox eradication: A WHO successstory. World Health Forum 8:283-292, 1987. Institute' of Medicine. Research on Mental Illness andAddictive Disorder: Progress and Prospects. Washington, D.C.: NationalAcademy Press, 1984. Institute of Medicine. Preventing Low Birth Weight.Washington, D.C.: Na- tional Academy Press, 1985. Johnson Foundation Special Report. Amess to HealthCare in the United States: Results of a 1986 Survey. Princeton, N.J.: The RobertWood Johnson Founda- tion, 1987. Johnson, J. Deaths of unattended children spur day care billsin Congress. The New York Times, Nov. 25, 1987, p. B-9. Jones, E.F.; Forrest, J.D.; Goldman, N.; Henshaw, S.;Lincoln, R.; Rosoff, J.; Westoff, C.; und Wulf, D. Teenage pregnancy in developedcountries: Deter- minants and policy implications. Family Planning Perspectives17:53-63, 1985. Kamerman, S.S., and Kahn, A.J. Income transfers, work and theeconomic well being of families with children. International SocialSecurity Review 3:345-382, 1982. Kash, D.E., and Ballard, S. Academic and applied policystudies: A comparison. American Behavioral Scientist 30:597-611, 1987. Kenney, A.M. School-based clinics: A national conference. FamilyPlanning Perspectives 18:44-46, 1986. Kerner, O. Report of the National Advisory Commission onCivil Disorders. New York: E.P. Dutton, 1968. Klarman, H.E., and Guzick, D. In: Selby, P., ed. Influenza Virus, Vaccines and Strategy. New York: Academic Press, 1976, pp.255-268. Klaus, M.H., and Kennell, J.H.Maternal-Infant Bonding.St. Louis: C.V. Mosby, 1976. Klein, R. NHS reorganization: The politics of the second best. Lancetii:418- 420, 1972. I r, PUBLIC POLICY 153

Knox, E.G. Theoretical aspects of -ubellavaccination strategies. International Journal of Infectious Diseases 7:194-197, 1984. Koplan, J.P., and White, G.C. An update onthe benefits and costa of measles and rubella immunization. In: Gruenberg,E.M.; Lewis, C.; and Goldston, S.E., eds. Vaccinating Against Brain SyndromeR:The Campaign Against Measles and Rubella. New York: Oxford UniversityPress, 1986. Krasnegor, N. Chief, Human Learning andBehavior Branch, NICIU) Personal Communication, 1987. Lamb, M.E., and Hwang, C.P.Maternal attachment and mother-neonate bonding: A critical review. In: Lamb, M.E.,and Brown, A.L., eds. Advances in Developmental Psychology. Vol. 2.Hillsdale, N.J.: L Erlbaum, 1982. Lazar, I.; Darlington, R.; Murray, H.; Royce, J.;and Snipper, A. Lasting Effects of Early Education. Monographs of the Societyfor Research in Child Develop- ment 47(1-2, Serial No. 194), 1982. Lozoff, B., and Brittenham, G.Infant care:Cache or carry.Journal of Pediatrics 95:478-483, 1979. McCartney, K., and Galanopoulos, A.Child care and attachment: A new frontier the second time around. AmericanJournal of Orthopsychiahy 58:16- 24, 1988. McCord, J. A thirty-year folio' i-up of treatment effects.American Psychologist 33:284-289, 1978. McCormick, M.C. The contribution of low birthweight to infant mortality and childhood morbidity. New England Journal ofMedicine 312:82-90, 1985. McKeown, T. The Role of Medicine: Dream,Mirage or Nemesis? London: Nuffield Provincial Trust, 1976. Murphy, F.D., and Ebert, RH. Report of thePresident's Biomedical Research Panel. MEW Publication No. (0S)76-500.Washington, D.C.:U.S. Govt. Print. Off., 1976. National Research Council. Genetic Screening: Programs,Principles and Re- search. Washington, D.C.: National Academy of Sciences,1975, p. 126. National Research Council. Risking the Future:Adolescent Sexuality, Pregnan- cy and Childbearing.Washington, D.C.: National Academy Press, 1987. Nkowane, B.M.; Bart, S.W.; Orenstein, W.A.; and Baltier, M. Measlesoutbreak in a vaccinated school population: Epidemiology,chains of transmission and the role of vaccine failure. American Journal of PublicHealth 77:434-438, 1987. Office of Program Planning and Evaluation.NIH Data Book. Washington, D.C.: U.S. Department of Health and Human Services,1986. 1 (1. 154 PREVENTION OF MENTAL DISORDERS

Office of Technology Assessment.Cost Effectiveness of InfluenzaVaccination. Washington, D.C.: U.S. Govt. Print.Off., 1981. Phillips, D.; McCartney, K.; Scarr-Salapatek, S.; and Howes, C.Selective review of infant day care research: Acause for concern! Zero to Three: Bulletin of the National Center for ClinicalInfant Programs. Washington,D.C.: Supt. of Docs., U.S. Govt. Print. Off., 1987.

Powers, E., and Witmer, H. AnExperiment in the Prevention ofDelinquency. New York: Columbia UniversityPress, 1951.

Regier, D.A. Director of ClinicalResearch, NIMH PersonalCommunication, 1987.

Richardson, S.A. The relationof severe malnutrition ininfancy to the intelligence of school children withdifferent life histories.Pediatric Research 10:57-61, 1976. Richmond, J.B., and Kotelchuck, M. Coordination and developmentof strategies and policy for publichealth promotion in the UnitedStates.In: Holland, W.W.; Detels, R.;and Knox, E.G., eds. Textbookof Public Health. Vol. 2. London: Oxford Press,1986, pp. 195-206. Roberts, S.R. Types andtreatment of pellagra.Journal of the American Medical Association 75:21-26, 1920. Rolrrtson, L.S. Crash involvement of teenage drivers when drivereducation is eliminated from high school.American Journal of PublicHealth 70:599-603, 1980.

Robertson, L.S., and Zador,P.L. Drive education and fatalcrash involvement of teenaged drivers. American Journal of Public Health 68:959-965,1978. Rogers, D.E.; Blendun, R.J.;and Moloney, M.B.A. Whoneeds Medicaid? New England Journal of Medicine307:13-18, 1982. Rowley, P.T., and Huntzinger, D.J. Newborn sickle cellscreening:Benefits and burdens realized. AmericanJournal of Diseases of Children 1983. 137:341-345,

Russell, L.B. Is Prevention Betterthan Cure? Washington, D.C.: Institution, 1986. The Brookings

Schape, E.; Moskowitz, J,M.;Malvin, J.H.; and Schaeffer,G.A. Evaluation of seven school-based preventionprograms: A fina! report on the NAPA project. International Journal of theAddictions 21:1081-1112, 1986. Schinke, S.P.; Gilchrist, L.D.;and Snow, W.H. Skills cigarette smoking intervention to prevent among adolescents. American Journalof Public Health 75:665-667, 1985.

14 PUBLIC POLICY 155

Shannon, J.A. The National Institutes of Health: Some critical years, 1955- 1957. Science 237:865-868, 1987. Sherman, M., atid Shafritz, DA Hepatitis-B v4rus and hepatocepular carcinoma: Molecular biology and mechanistic considerations. Seminars on Liver Diseases 4:98-112, 1984. Singer, H.D. Mental and nervous disorders associated with pellagra. Archives of Internal Medicine 15:121-146, 1915. Sydenstricker, V.P. The history of pellagra, its recognition as a disorder of nutrition and its conquest. American Journal of Clinical Nutrition 6:409-414, 1958. Telch, M.J.; Killen, J.D.; McAlister, A.L.; Perry, C.; and Maecoby, N. Long term follow-up of a pilot project on smoking prevention with adolescents. Journal of Behavioral Medicine 5:1-8, 1982. Teuber, H.L.; and Powers, E. Evaluating therapy in a delinquency pre-ention program. Proceedings of the Association for Research in Nervous and Mental Diseases 31:138-147, 1953. U.S. Bureau of the Census. Current population reports. Series P-23, No. 150. Population Profile of the United States: 1948-1985. Washington, D.C.: U.S. Govt. Print. Off., 1987. Warner, KE. Smoking and health implications of a change in the federal cigarette excise tax. Journal of the American Medical Association 255:1028- 1032, 1986. Westinghouse Learning Corporation. The Impact of Head Start: An Evaluation of Head Start on Children's Affective and Cognitive Development. Vols. I and II. Athens, Ohio: Ohio University Press, 1969. Willems, J.S., and Sanders, C.R. Cost-effectivenees and cost-benefit analysis of vaccines. Journal of infectious Diseases 144:486-493, 1981. Wilson, W.J. The Truly Disadvantaged: The Inner City, the Underclass and Public Policy. Chicago: University of Chicago Press, 1987. Winslow, C.E.A.The Life of Hermann M. Biggs.Philadelphia: Lea and Febiger, 1929. Wyngaarden, J.B. The National Institutes of Health in its centennial year. Science 237:869-874, 1987. Zabin, L.S.; Hirsch, M.B.; Smith, E.A.; Street, R.; and Hardy, J.B. Evaluation of a pregnancy prevention program for urban teen-agers. Family Planning Perspectives 18:119-126, 1986. Zigler, E., and Gordon, E., eds. Day Care: Scientific and Social Policy Issues. Boston: Auburn, 1982.

1 CHAPTER 5

Psychiatric Disorder in Parents as aRisk Factor for Children

Michael Rutter, C.B.E., M.D., F.R.S. Honorary Director MRC Child Psychiaty Unit Department of Child and Adolescent Psychiatry Institute of Psychiatry De Crespigny Park, Denmark Hill London Statistical Associations

More than 60 years have gone by since Janet (1925) drewattention to the importance of parental mental disorder as a psychiatric risk factorfor the children and outlined the possible mechanisms that might be involved.Sub- sequent research has amply documented the reality of risk, asdemonstrated by the consistent statistical associations between psychiatric disorder in parents and in their children (see Earls 1987; Feldman et al. 1987; Ilutter1966, 1987a; Rutter and Quinton 1984).Such associations have been demonstrated in numerous epidemiological studies of the generalpopulation, in case-control comparisons of the parents of children with a psychiatricdisorder, and in case-control comparisons of the children of parents with a mentaldisorder. The main risk is for persistent psychiatric disorders in the childrenrather than transient situational stress responses (Rutter and Quinton 1984).Moreover, recent evidence (Quinton et al. in press) has shown thatthis risk extends into adult life.

This paper was prepared for the Project Prevention initiative of theAmerican Academy of Child and Adolescent Psychiatry and is based in part on an earlier review "Parental mental disorder as a psychiatric risk factor," in Hales,R.E., and Frances, A.J., eds., American Psychiatric Association's Annual Review,Vol. 6, Washington, D.C.: American Psychiatric Association, 1987. pp. 647-663. Correspondence should be addressed to Dr. Michael Rutter, Honorary Director, MRC Child Psychiatry Unit, Department of Child and Adolescent Psychiatry, Institute of Psychiatry, De Crespigny Park, Denmark Hill, London SE6 8AF, England. Telephone: (01)703-7349. 158 PREVENTION OF MENTAL DISORDERS

Of course, correlations do not prove causation, and it could be that the problems of rearing a mentally disturbed child led to emotional disorder in the parent rather than the other way around. Doubtless that occurs; certainly thcza is evidence that the rearing of biologically handicapped children is associated with an increased rate of parental distress and family tension (Breslau et al. 1982; Byrne and Cunningham 1985; Cooke et al. 1982; Sabbeth and Leventhal 1984). Nevertheless, that does not seem likely to be the usual explanation if only because in many, if not most, cases the parent's disorder antedated the child's.Moreover, Richman et al.'s (1982) prospective epidemiological, longitudinal study showed that maternal depression when the children were 3 years old (and free of psychiatric disorder) predicted the development of child disorder during the subsequent 5 years. It may be concluded that the risks to the children associated with parental psychiatric disorders are real. Nevertheless, that conclusion does not necessarily mean that the risk derives from the parental illness per se.After all, child psychiatric disorder is also associated with chronic physical illness in the parents (Rutter 1966), parental death (Garmezy 1983), and parental criminality (Witter and Giller 1983)as well as with family adversities of various kinds (Rutter and Ma4ge 1976). It could be that the risks stem from the psychosocial stressors associated with parental illness rather than from the illness as such. Moreover, insofar as the risks are a function of parental illness, they may be genetically or environmentally determined. The environmental risks may stem from physical damage to the fetur (as from drugs or perinatal complications) or to the child (as from head injuries or other accidents de' 'ving from lack of adequate parental supervision), as well as from psychosocial factors. If effective prevention or intervention is to follow identification of the risk factor, it is important that the relevant risk mechanism or processes be identified.These constitute the main focus of this review.Obviously, the mechanisms involved may vary according to the type of parental disorder that constitutes the risk variable or the type of child disorder that ensues; that possibility will be borne in mind in considering the relevant empirical findings. Genetic Mechanisms

Much of the research on the risks to children of parents with mental disorders has been based on the premise that the risk is likely to be genetically determined. That premise derives from the empirical demonstration that genetic factors play a significant role in the determination of schizophrenia (Gottesman and Shields 1976), major affective disorders (Gershon et al. 1982; Weissman et al. 1984a), antisocial personality disorders and criminality (Crowe 1983), and some varieties of alcoholism (Bohman et al. 1981; Cloninger et al. 1981; Goodwin 1985).It also appears that genetic factors play a significant role inmany psychiatric disorders arising in childhoodthough thisarea has been less studied (McGuffin and Gottesman 1985; Rutter et al. in pressa; Vandenburg et P'17 DISORDERS IN PARENTS 159 al. 1986). Although there is an important heritable component to many types of mAjor mental illness, genetic factors seem much less important in the emotional disorders that make up most of adult psychiatric outpatient practice (Torgersen 1983).Yet these disorders, when they occur in parents, are associated with a substantially increased psychiatric risk for the children (Rutter 1966; Rutter and Quinton 1984). Even when the parental condition is genetically determined in part, however, it does not follow that the risk to the children is genetically mediated. This is because (a) with all the adult mental disorders there is a major nongenetic component; (b) the continuity between mental disorders in children and adult life is far from complete (Rutter 1984' and, even when there is continuity, the genetic component may be greater for disorders that persist into adulthood than for those confined to the childhood years (Rutter and Gil ler 1983); and (c) parental mental disorder is frequently accompanied by n*or environmental disturbance (Coyne et al. 1987; Feldman et al. 1987; Jacob and Seilhamer 1987). Thus, parental symptoms may directly impinge on or involve the children in some way (Radke-Yarrow et al. 1988; Rutter 1966); the parental illness may

interfere with parenting functions (Bettes 1988; Cohn et al. in press; Field et al. in press; Rodnick and Goldstein 1974; Susman et al. 1985; Weissman and Paykel 1974) or impair parent-child relationships and interactions (Cox et al. 1987; Davenport et al. 1984; Feldman et al. 1987; Zahn-Waxler et al in press). Moreover, the parent's illness may result in such family disruption that the children have to go into foster care (Rice et al. 1971), or it may be accomp, nied by marked marital discord and disharmony (Birtchnell and Kennard 1983a and b; Got lib and Hooley 1988; Rutter and Quinton 1984). This discord is associated with increased conflict over child rearing, greater segregation in decisionmaking, reduced affection, and altered patterns of dominance (Hinchcliffe et al. 1975; Kreitman et al. 1971). Specificity of Effects It is necessary to consider how to test the hypothesis that the risk to the thildren is genetically mediated. A genetic transmission of disorders would be suggested if the risks to the children were relatively confined to certain specific types of parental disorder. However, the evidence is clear that this is not the case.Raised rates of psychiatric disorder in children have been found, for example, for parental schizophrenia (Watt et al. 1984), depression (Beards lee et al. 1983; Cytryn et al. 1986; Keller et al. 1986; Weissman et al. 1984b, 1987); alcoholism (Earls et al. 1988; Nylander 1960; Rydelius 1981; Steinhausen et al. 1984), and personality disorder (Quinton et al in press; Rutter and Quinton 1984). A second test of the genetic hypothesis is to determine if the disorders in the children tended to be of the same type as those in the parents or, at least, showed a degree of specificity in relation to the parental diagnosis. Even that relationship does not seem to be generally the case. On the whole, there are only rather weak associations between the form of disorders in parents and children. There 1 160 PREVENTION OF MENTAL DISORDERS may be some limited specificity, however.First, although the children of schizophrenic parents may show a variety of psychiatric problems, some abnormalities both are particularly associated with this parental diagnosis and appear to constitute the childhood precursors of adult schizophrenia (Nuechter- lein 1986; Rutter 1984; Watt et al. 1984). The key features comprise (a) abnormalities in interpersonal relationshipsshown by odd, unpredictable behavior, social isolation, and rejection by peerstogether (in males) with solitary antisocial behavior in the home; (b) neurodevelopmental immaturities in the form of clumsiness, visuospatial difficulties, and verbal impairment; and (c) attention deficits characterized by poor signal-noise discrimination. Second, there is some tendency for parental personality disorder to be associated with conduct disturbance in the sons (Rutter and Quinton 1984; Stewart et al. 1980) and for parental alcoholism to be linked with both alcoholism and antisocial disorders in the male offspring (Rydelius 1981; Stein- hausen et al. 1984), but perhaps especially with disorders that combine both emotional and conduct disturbance (Earls et al. 1988). Third, and less certainly, major depression in the parents may be particularly likely to lead to depression in the children (Cytryn et al. 1986; Decina et al. 1983; Keller et al. 1986; Weissman et al. 1984b). Standardized psychiatric assessments have shown that about half the psychiatric disorders in the children of seriously depressed parents are depressive in form. The uncertainty stems from (a) the observation that some half the disorders in the children are not depressive in type and that the increase in nondepressive disorders may be as great as that for major affective conditions (Hammen et al. 1987; Weissman et al. 1987); (b) the lack of evidence concerning whether the rate of childhood depression in the offspring of parents with depression is higher than that found with other forms of parental mental disorder (most comparisons have been with normal controls); (c) the negative findings in some studies (Gers:2on et al. 1985) together with the high frequency of mixed symptomatology even when the longitudinal coutse suggests an affective diagnosis (Akiskal et al. 1985); and (d) the limited knowledge regarding the extent to which childhood depression is synonymous with the major depressive disorders of the adult life (Rutter et al. 1986). Nevertheless, there is some suggestion that depressive disorders in the children may be particularly linked with serious parental depression (Rutter and Quinton 1984). This linkage is probably most evident in the case of bipolar affective disorders with a prepubertal onset (Strober et al. 1988). From the very limited data so far available, it may be inferred that most psychiatric disorders in the children of depressed parents are not primarily genetically mediated, but that genetic mechanisms are likely to be much more influential in the association between bipolar affective disorders in parents and their children. Environmental Effects A third test of the genetic hypothesis is provided by studies that determine whether the association between disorders in parents and in their childrencan be accounted for by environmental variables. Rutter and Quinton (1984), in 1 C DISORDERS IN PARENTS 161 their study of a heterogeneous group of mentally ill parents, used a rangeof well-tested discriminating measures of the family environment.Their results showed that ths risk to the children was largely a function of the familydiscord and hostility associated with the parental mental disorder. Indeed,the risk to the children of mentally ill parents showed no significant increase over thelevel in the general population once the family adversity variables had beentaken into account. Similarly, Feldman et al. (1987) found that the psychiatricrisk for the children of mentally ill parents, as also for children of mentally healthy parents, derived in large part from the associated familyadversities and especially from discordant mother-child relationships. Children not living with their ill parents showed no increase in psych:stric disorder, providedthey were being reared in an alternative family setting and not in an institution. It seems that the risk to the children is largely environmentally mediated.However, there appear to be some important exceptions to the general finding.Emery et al. (1982) found that whereas discord constituted the main factor involvedin the conduct disturbances seen in the children of parents with depression or personality disorder, it did not account for the increased rate of disorders in the children of schizophrenics. Similarly, Folstein et al. (1983) found that discord accounted for conduct disturbances in the children of parents with Huntington's disease, but not for depression in the offspring. Keller et al. (1986) found that parental discord increased the risk of disorder in the children of depreesed parents, but that the severity and chronicityof maternal depression also did so (it is not clear whether this was so after controlling for discord; moreover, the risk to the children was not significantly affected by depression in the father). Rutter and Quinton's (1984) findings also are conpatible with some risk for childhooddepression that is additional to that associated with discord. Fendrich et ai. (in press) studied the role of a range of family risk factors as noes, ible mediators for the increased risk for psychiatric disorders in the offspring, Family discord, divorce, and lack of cohesion were important risk factors for conduct disorder irrespective of the presence or absence of parental depression. However, no family risk factor was associated with anxiety disorder; and for depression in the offspring, both family risk factors and parental depression seemed to act separately as mediators at risk. Radke-Yarrow et al. (1985) found insecure attachmeat to be more frequent in the children of severely depressed mothers than in controls (see also Gaensbauer et al. 1984). However, what was most characteristic was a type of insecurity associated with both resistance and avoidance, together with abnormal affect or stereotyped rnaladaptive behavior. Because this is the pattern that may also be associated with child abuse, it seems unlikely thatit is specific to parental depression. Nevertheless, it may represent a more pathological variety of insecure attachment. Radke-Yarrow et al. (1985) found that the abnormal pattern of attachment was significantly associated with the severity and chronicity of mental depression; it was unaffected by whether the father was depressed, but it was associated with maternal negative expressed emotions to the child and with the absence of a father in the household. The lack of effect 162 PREVENTION OF MENTAL DISORDERS

in this study and in the study of paternal depressionby Keller et al. (1986), together with the effect of discord and of negative expressedemotions, suggeets that genetic factors do not constitutea sufficient explanation. Rutter and Quinton (1984) found that parental personalitydisorder (of both antisocial and other types)was powerfully associated with disorder in the children. However, multivariate analyses showedthat this was more a consequence of the children's exposure to hostile or aggressive behaviorthan of the parental diagnosis per se. Nevertheless, although theeffect fell short of statistical significance, there wassome suggestion tlat personality disorder put the children at an additional psychiatric risk beyondthat accounted for byexposure to hostile behavior. The same suggestionstems from the finding that,even after controlling for disrupted parenting in childhood,parental deviance (much of which involved criminalityor personality disorders) predicted the development of personality disorder in adult life ininstitution-reared children (Quinton and Rutter 1988; Rutter et al. inpress b). Adopted-Away Children

Probably the strongest test of the genetic hypothesisis provided by determining rates of disorder in the children ofmentally ill parents whoare adopted in infancy and brought up by nonill parentsto whom they are not biologically related. This strategy has been employed withschizophrenia: Several studies from Heston (1966) to Tienari et al. (1985)have found an increased rate of schizophrenia in the adopted-away children.It has also been shown that the attentional deficits found in the fostered childrenof schizophrenics who have been brought up away from the ill parentsare similar to those found in remitted adult schizophrenics (Asarnow andMacCrimmon 1978; Asarnow et al. 1977). The results provide strong evidence ofa genetic mode of transmission. The genetic hypothesis has also been supportedfor other disorders; thus it has been found that there ib an increased rate of thesame disorder in adopted-away offspring of parents with alcoholism(Goodwin 1985), antisocial problems (Crowe 1983), and major affective disorders(Cadoret 1978; Mendlowicz and Rainer 1977). In addition, Stewart and deBlois (1983) found that the associations between antisocial behavior infathers and sons were greater whenthe fathers were in the home, althoughsome association was found when fathers were absent. It may be concluded that the adoptee-studyfindings confirm that genetic factors play an important role inmany chronic psychiatric disorders in adult life. However, the investigationsundertaken to date have provided rather limited evidence concerning the role ofgenetic influences in the psychiatric risks in childhood for the offspring of mentallyill parents. Gene-Environment Correlations andInteractions Although a great deal of attention has beenpaid to the heri ability of specific psychiatric conditions, heredity isnot the only way in which genetic factorscan play a role in the psychiatric riskexperienced by the children ofparents with some form of psychiatric disorder.Gene-environment correlations and 1 71 DISORDERS IN PARENTS 163 interactions may be influential through their effects on environmental risk mechanisms (Kendler and Eaves 1986; Plomin 1986; Pogue-Geile and Rose 1987). There is very little evidence showing the operation of such mechanisms, but a few findings suggest that they do occur. Two processes require special mention. First, it is possible that to some extent people create their own environments, so that genes partially shape the environment (Scarr and McCartney 1983). Thus, McGuffin et al. (1988) have shown that psychosocial stressors load in families. Several studies have shown that individuals who show deviant behavior in childhood are more likely to lead disrupted lives in adult life. For example, Kandel and Davies (1986) found that depressed and nondepressed adolescents differed in their pwaern of social relationships in adult life; and Caspi et al. (1989) found that the adult careers of explosive children tended to be characterized by disorganization and instability. Similarly, using retrospective data from the Epidemiologic Catchment Area (ECA) Study, Robins (1986) found that adverse life experiences in adulthood were linked with previous psychopathology in childhood. None of the studies has tested the hypothesis that the links over time were genetically mediated, but it is possible that genetic factors played a part. This possibility was also suggested by Cado-et et al.'s (in press) finding that although parental alcoholism and antisocial_iorder had no direct effect on depression in the adult offspring, there wasa strong indirect effect through the link with antisocial personality disorders in thegrown children who had been adopted in infancy. Such disorders secondarily increased the risk of depression in adult life. Second, it may be that genetic factors increase people's vulnerability to environmental hazards. This is suggested by the finding that environmeigal risks are greatest in individuals who are also genetically vulnerable. Several studies have shown that the risk of antisocial behavior is greatest when there is criminality in both the biological and the adoptive parents; and the increase in risk is greater than expected on the basis of a simple additive effect (Cadoret 1985; Cadoret et al. 1983).Similarly, Cadoret et al. (in press) showed that parental antisocial or alcohol disorder was associated withan increased risk of adult depression in the offspring following late adoption,an increase not found in the absence of the parental disorder. There is also some suggestion of similar mechanisms in connection with parental schizophrenia.In particular, Parnas et al. (1985) found that institutionalization in infancy predisposed to schizophrenia in the offspring of schizophrenic mothers, but not in controls; this finding is consistent with gene-environment interactions. More direct evidence comes from studies of the rearing environment of adopted-away children with direct study of the adoptive home environment. The Colorado Adoption Study (Plomin and De Fries 1985) provides just such evidence, but so far not with respect to psychiatricoutcomes. The only psychiatric investigation is that undertaken by Tienari et al. (1985; in press) of the adoptive families of children born to schizophrenic parents. The data from analyses on the first 91 cases (half the sample) showed that the 172 164 PREVENTION OF MENTAL DISORDERS

psychiatric risk was increasedwhen the adoptive family environmentwas disturbed, suggestinga gene-environment interaction. However, thiseffect was evident for borderline and characterdisorders rather than for schizophrenic psychoses. The evidence so faron gene-environment correlations andinteractions is extremely scanty, butwe may conctuee that there area variety of mechanisms by which a genetic predispositionmay increase environmental links for child psychiatric disorder. The extentto which this happens in practiceremains uncertain. Clinical Implications

The clearest conclusions apply toparental schizophrenia: It is there is a definite, genetically apparent that mediated, increased risk ofschizophrenia in the children. In addition, the increasedrisk for a broaderrange of schizophrenia spectrum disorders probably includesa genetic component (Kendler and Gruen- berg 1984); but the criteria forsuch disorders remain quiteunclear, and the genetic link with schizophrenia is still somewhat uncertain(Torgersen 1984). Moreover, there isan incrnased risk for other types of psychiatricdisorders that probably is not geneticallymediated, The clinical pictureof social oddity, neurodevelopmental abnormalities,and attention deficits ismost likely to represent a precursor of schizophrenicpsychosis; but the criteriaare not sufficiently clear cut to warrant a definite diagnosis at that stage. Thereis no advantage in creatingan expectation of genetic predestinationwhen there is good evidence that environmentalfactors also play an importantrole in etiology. Brown et al. (1962) showed that overinvolvement with relatives whoexpress high levels of criticism (highEE) is associated withan increased risk of relapse in adult schizophrenia; this finding was replicated byLeff and Vaughn (Leff and Vaughn 1981; Vaughn andLeff 1976; Vaughn et al,1984). Leff and Vaughn (1981) also found that this effectwas additional to that obtained by appropriate medication; anda controlled therapeutic trial (Leffet al. 1982, 1985) demonstrated that social interventionwith high EE families that fall in EE significantly reduced resulted in a the relapse rate. Goldsteinand his colleagues (Doane et al. 1981; Goldstein 1985) found that high EE andpoor communication in the families of disturbedadolescents predicteda worse outcome (including the development of schizophrenicspectrum disorders). Itmay be inferred that when the children of schizophrenic parents develop psychiatricdisorder, they aro most likely to be helped by therapeuticinterventions that reduce highly critical parental overinvolvementand that improve harmonious communication. parent-child It seems probable that genetic factors playsome contributory role in the development of conduct disorders(possibly especially in those chronic) in the children of that are most parents with personality disorders,and of' depression in the children of depressed parents. The genetic evidence,however, is inconclusive and as yet pointsto no particular mode ofprevention or intervention. 173 DISORDERS IN PARENTS 165

that arises fromconsideration of gene-environment Perhaps the main point vul- interactions is that agenetic predisposition maywell create a greater that warrant attentionin nerability to environmentaladversities, and it is they preventive and therapeuticinterventions. of alcoholic parentsprobably results from both Alcoholism in the offspring needed would br imilarto genetic and environmentalfactors. The treatment However, the reality of tfurisk means that required foralcoholism generally. alcoholic parents need tobe alerted to thedangers for that clinicians treating with heavy drinking or the offspring and to takeseriously the risks associated with drinking to relievestress. Effects on theFetus

Alcohol &see of alcohol in thefirst trimester of pregnancyact It is known that high cranio-facial as a teratogen,causing mental retardationand a characteristic malformation that has come tobe known as the "fetalalcohol syndrome" (Porter Steinhausen and Spohr1986). It seems that in thefirst 10 weeks et al. 1984; in midpregnancythere alcohol is cytotoxic, causing adeficiency in brain growth; and delay of neural cellmigration and develop- is a transient disorganization (CNS) neurotransmitter ment and interferencewith central nervous system production leading toneuroendocrine abnormalities.It remains uncertain the fetus or whetherthere whether there is a thresholdfor this alcohol effect on at alcohol levels too/ow to is a continuum of effectswith subclinical damage syndrome; the balance ofevidence suggests a continuumeffect. cause the full evidence Moreover, it is unclearwhether the effect is specificto alcohol; some (Hingson et al. 1982).Because suggests that marijuana mayact in a similar way postnatal environmen- parental alcoholism tends tobe associated with so many has proved difficult toseparate fetal frompostnatal effects tal disturbances, it However, Aronson when examining psychiatricdisturbances in the children. that the offspring ofalcoholic mothers had IQ scoresthat et al. (1985) found foster parents were below thoseof controls even whenthey had been reared by physical stigmata of fetalalcohol syndrome. Thechildren and did not show the inattention. Steinhausen et of alcoholic parents alsoshowed hyperactivity and small, incomplete sample,found that the behaviorof al. (1984), in a study of a somewhat during the children with fetal alcoholeffects tended to improve there was not severe mentalretardation; but hyper- preschool years, provided 8-1/2 years was not activity tended to persist.The psychiatric outcome at the extent of morphologicaldamage as clinically significantly associated with Streissguth et al. (1984), in a assessed. Much of the bestdata are provided by followup of some 500 infants,half of whom were born detailed and systematic Alcohol-related to heavy drinkersand half to light andinfrequent drinkers. behavior effects were stil,,,rident at 4 years of age, afterstatistical adjustment most evident for reaction for possible confoundingvariables. The effect was I 7 166 PREVENTION OF MENTALDISORDERS

time, response tendency, and attention, suggestingsome impairment in the central processing ofinformation. The findings rather than a threshold suggested a dose-response effect, but clinicalabnormalities heaviest drinking levels were found only at the (more than 59g of absolute alcoholper day). We lack knowledge on how beet to treat thesealcohol-related attentional deficits and associatedbehavioral disturbances. crucial need is to Clearly, however, themost prevent the initial fetaldamage (althoughsome of the behavioral sequelaeare likely to be related to education on the damages postnatal influences).Public of heavy drinking isa priority. It should be noted, however, thata reduction in drinking, or even abstinence from alcohol,once a woman realizes that she ispregnant will not constitute because the most effective prevention serious damageoccurs during the first 4 weeks, just after the firstmissed menstrual before and period. People needto appreciate that the consumption of alcoholshould be kept lowwhenever considered likely. pregnancy is plannedor Opioids

It is known that opioidspass the placental barrier or morphine addicts will be and that offspring ofheroin born opioid dependent(Jeremy and Bernstein Withdrawal after birthleads to an abstinence 1984). and rejections of syndrome with markedirritability overtures of comfort.&cause parentsare likely to consider such behavioras "difficult," there have been fears that it wouldlead to impaired parent-infant relationships.Very little evidence that risk. Jeremy is available(.n the extent of and Bernstein foundlittle differenceat 4 months between methadone-exposed andcomparison infants intheir patterns of both, poor maternalcommunication interaction (in and worse coordination). was associated with greaterinfant tension The mothers whoabused drugs showedworse interaction and communicationwith their infants nal functioning than did controls,but poor mater- was related more to lack ofcurrent emotional drug use perse. The findings, resources than to so far as they go, are consistentwith a risk to the children of being rearedby a mother who suggest that the main is addictsd to opioids;but the results risk does not stemfrom the infants' drug syndrome in the neonatalperiod. withdrawal Pregnancy Complications There has been muchdiscussion of the tions in the etiology possible role ofpregnancy complica- of schizophrenia(Walker and Emory derived from the 1983). The interest findings chat schizophreniais associated with slightly raisedincidence of obstetric a history of a complications; that suchcomplications are more likely in schizophrenictwins than in their and that discordant monozygoticcotwins; pregnancy complicationsmay lead to neurodevelopmental if the complicationscause neural damage, such disabilities as by intraventricularhemor- rhage (Stewart 1983).It was thoughtthat if a schizophrenic likely to have abnormal woman were more pregnancies, this mightconstitute part of the for the psychiatricrisk for the offspring. reason It is now clear thatthey are notmore 175 DISORDERS IN PARENTS 167 likely to have abnormal pregnancies, but it seems that the children born to schizophrenic mothers may be more vulnerable to damage from obstetric complications (Walker and Emory 1983). There is some suggestion that the increased ventricular size found in some cases of schizophrenia maybe associated with obstetric complications (Schulsinger et al. 1984), as well as being more common in schizophrenic individuals without apositive family history (Murray et al. 1985). Nevertheless, it remains uncertain whether pregnancycomplica- tions play other than a minor contributory role in the transmission of schizophrenia from parent to child. Family Environmental Effects

Exposure to Specific Parental Symptoms

Rutter's (1966) early study of child psychiatric disorders rum ...with parental mental illness suggested that the risk to the children was greater when parental symptoms directly impinged on or involved the children.Direct symptom impact occurred when parental delusions incorporated the children in some way, when children were forced to participate in parental ritualsand compulsions, or when the parental illness led to marked restrictions in the children's social activities.Many of these examples concerned severe and somewhat unusual types of parental disorders.Rutter and Quinton (1984) conducted a 4-year prospective study of a representative sample of the families of newly referred psychiatric patients who were parents of children under 15 years of age. The results showed that such direct involvement of children in parental symptoms was not common and did not account for much of the risk to the children. Neither the children's exposure to psychiatric symptoms nor their exposure to parental affective symptoms was significantly associated with psychiatric disorder in the children, once account had been taken of other factors. However, their exposure to hostile or aggressive behavior by the parent was very strongly associated with an increased risk of psychiatricdisorder, irrespective of the parental diagnosis. Similarly, Billings and Moos (1986), in a 1-year followup study, found that the children of depressed parents continued to show increase in maladjustment even when the parental symptoms remitted. The children's maladjustment correlated with the quality of family emotional resources, as well as with the severity of the depression in both parents. Family Discord Marital discord constituted one important source of hostile behavior in the Rutter and Quinton study, and it showed a strong association with an increased psychiatric risk for the children. Sons tended to develop disturbances earlier than daughters in the presence of family discord; but if the discord persisted, the girls suffered in the long run, although not to quite the same extent as boys (Rutter and Quinton 1984).Similar findings on the risks associated with family discord derive from all other studies that have included systematic C 168 PREVENTION OF MENTAL DISORDERS

discriminating measures of parent-child and marital relationships (Cox et al. 1987; Feldman et al. 1987; Keller et al. 1986; Radke-Yarrow et al. 1985; Richman et aL 1982). Most of the research has concerned the effects of parental depression, but the major psychiatric risks to the children associated with family discord and parental hostility or critical overinvolvement have been found with personality disorder (Rutter and Quinton 1984) and Huntington's disease (Folstein et al. 1983), as well as with general population samples of parents who do not suffer from any mental illness (Emery 1982; Rutter 1982; Rutter and Giller 1983). The effect is greater when discord is associated with other family adversities, but it is not necessary that such adversities include parental mental disorder.The consist-rly and pervasiveness of the risk associated with serious persistent family discord that is found in quite disparate populations testifies to its importance as a psychiatric risk factor. Probably the risks are greatest when the discord results in parental criticism or hostility that is directly focused on one or more of the children, but the risk is still evident when the tension and quarrelingare mainly between the two parents. Discord plays a crucial role as a mediatcr of the psychiatric risks associated with parental mental disorder. This role isa consequence both of the strength of the discord effect and of the frequency with which mental discord is associated with marital discord. The nature of the connections between mentaldisorder and marital discord are complex; there is evidence of causal influencesin both directions. In addition, both may be determined in part by familystresses and adversities outside the marriage. Nevertheless, however the discord arises,it serves as an important risk mechanism for the children. Family Breakup Parental mental disorder, especially when it is associated withother psychosocial hazards, not infrequently leads to temporaryor permanent family breakup (Rice et al. 1971). Thus, of children admitted to institutionalcare or to family foster care because of parenting difficultiesor child neglect or abuse, a high proportion have mentally ill parents (Quinton and Rutter 1988). Even short-term admissions to residentialcare are associated with a substantially increased psychiatric risk for the children (Wolkind and Rutter 1973).We lack studies in which there has been systematic psychiatricassessment of the parents in families from which the childrengo into foster or institutional care. Nevertheless, the evidence that exists shows that interms of psychiatric sequelae, both in childhood (Roy 1983) and in early adult life (Quintonand Rutter 1988), the main risks derive from the adverse experiencesrather than from the parental mental illness perse. Probably the main adversity is not the child's separation from parentsas such, but rather the multiple stresses with which it is associated, and the family discord that precededand succeeded the child's admission into fostercare (St. Claire and Osborn 1987; Wolkind and Rutter 1985). DISORDERS IN PARENTS 169

Brown et al. (1986) showed thatwith respect to effects onvulnerability to adult depression, parental losecreated a risk factor only if it led topoor-quality parental care that was lackingin warmth and affection.Nevertheless, the insecurities associated with goingin and out of foster care maywell add to the risks.Clinicians face real dilemmas:Children who have been admitted to residential nurseries do poorly on returnto their biological parents(Hodges and Tizard 1989a and b)but thelong-term outcome is also bad forthose who remain in institutional care (Quintonand Rutter 1988). The consequences are better if either there is a stablefoster family (Roy 1983) or ifharmony is restored however, is all in the biological family (Rutter1971). Foster care breakdown, too common; and the restorationof good relationships in thefamily is not easily brought about. Physical Risks A few researchers have noted anincreased rate of accidents in thechildren of depressed mothers (Brown andDavidson 1978), and it is known that severe head injuries create a significantand substantial psychiatric risk(Rutter et al. 1983). Though this mechanism (i.e.,brain damage from head injuriesresulting from inadequate parental supervision) maybe important in individual children, the infrequency of severe head injury meansthat it is not a relevant factor in other than a small minority of cases.Nevertheless, it is striking that Weissman et al. (198) found that thechildren of depressed parents had anincreased rate of a wide range of medical problems.It is not known what role theseproblems played in the risk for psychiatricdisorder. Parenting Although it has been appreciated for sometime that mental disorders including schizophrenia (Rodnick andGoldstein 1974) and depression (Belle 1982; Rutter in press b; Tronickand Field 1986; Weissman andPaykel 1974) may significantlydisorganize, distort, and impair parenting,only recently have there been systematic, observationalstudies of the parenting of mentallyill mothers (fathers have yet to be studied).Radke-Yarrow et al. (1985) found that severe (but not minor)maternal depression, especially of thebipolar variety, was associatedwith a marked increase in an unusual(and presumably psychopathological) ,..ariety ofresistant/avoidant insecure attachment. Depressed mothers were rated as moredisorganized, unhappy, tense, inconsistent, and ineffective with theirchildren (Davenport et al. 1984). Cox etal. (1987) found that depressed mothers in thecommunity differed from controls inbeing less likely to respond to their 2-year-oldchildren's overtures, less facilitative of social interactions, less adept in responding totheir children's cues, and more likely to respond with control when theirchildren were distressed. Stein et al. (submitted for publication) reported similarfindings from their longitudinal study of 49 women with postpartumdepression and 49 individually matched nondepressed controls.Observations of mother-child interactionwhen the children were 19 months old showed that, comparedwith controls, the depressed women interacted less withtheir children and were lessfacilitating; their Is 170 PREVENTION OF MENTAL DISORDERS

children showed less affectivesharing, were muchmore likely to show marked distress during a planned brief departurefrom their parent'sroma, and were lees likely to show initial sociability with a stranger. Similar butreduced effects were seen in the subgroup of families in whichthe mothers had been depressed postnatally but were no longerdepressed. The findings suggest thattha altered parenting quality is nota direct consequence of current depression.Other investigations, too, have shown alteredpatterns of mother-infant interaction associated with maternal depression(Bettes 1988; Cohn et al. inpress; Field 1984; Field et al. inpress; Livingood et al. 1983; Lyons-Ruthet al. 1986). The evidence is consistent inshowing significant impairmentsin parenting associated with maternal depreseion,hit all the studies havenoted marked individual differences, withsome depressed womer._leafing well. In addition, they have consistently shownthat the pareming differencestended to remain (albeit at a somewhat reduced level)after remission of the depression. Hawton et al. (1985) found a substantially increased risk of childabuse in mothers who attempted suicide.Suspected or actual abusewas more frequent than in young, nonsuicidal,working-class mothers with home. preschool children at However, the parental suicidalattempt did not constitutea useful alerting mechanism for possiblechild abuse; as in mostcases, the abuse pr, ceded the suicide attempt. Less is known about parentingdifferences associated with other parental mental disorder. forms of Sameroff et al. (1982) found that,compared with deurotic dpression, maternalschizophrenia tended to be associatedwith rather fewer impairments inp& renting ard fewer emotional disturbancesin infants. Ntislund et al, (1984a and h) fon that the 1-year-old offspringof women with schizophrenia or cycloid psychosisdiffered from controls in showinga lack of fear of strangers. They alsonoted an increase ofanxious attachment in the offspring of schizophrenics, but not in the children of mothers withother types of mental illness. Thesame stet('showed that mother-infant each instance were interactions in mote negative when the motherwas psychotic (McNeil et al. 1985; Naslund et al. 1985; Persson-Blranow et al, 1984).As previously noted, :gdke-Yarrow et al. (1985) showeda marked increase in insecure in the children of depressed attachment mothers. Although the evidenceis contradictoryon whether thereare effects on parenting thatare specific to parental diagnosis, it seems probable that there are not. Most, if not all, forms ofserious mental disorder may be associatedwith difficulties inor distortions of parenting. They are, in turn, associated with abnormalities in the dyadic relationshipbetween mother and child; this associationprobably plays a role inleading to an increased psyAiatric link(although direct evidencecn this issue is lacking). Clinical Implications

Parental mental disorder isfrequently associated withwidespread disturbances in many aspects of family interact(on and of parenting;moreover, it is clear that these familydisturbances constituteone of the main risk mechanisms 1 7 DISORDERS IN PARENTS 171 by which parental mental disorder leads to psychiatric problems in the children. se associated psychosocial disturbances often continue well after acute arental symptoms abate (Bothwell and Weissman 1977).Hence, it is no surprise that there are no close connections between the ebb and flow of parental symptoms and the course of disorder in the children (Hobbs 1982; Rutter and Quinton 1984) or that the psychiatric risk to the children continues well after remission of the parental disorder, although the risk is greater if it persists (Billings and Moos 1986; Rutter and Quinton 1984). The most obvious implication is that clinicians treating parents with a mental disorder need to assess the extent to which the disorder is associated with family discord, with negative feelings to any of the children, and :. ith impairments of parenting. The main risk to the children occurs when these family features are affected; the risk is much less if they are not. Clearly, it is highly desirable to intervene therapeutically to improve patterns of family relationships; but. this has not proved easy to accomplish in practice. Rounsaville et al. (1979) found that depressed women with serious marital difficulties had a poor outcome compared with those who were single or in supportive relationships. They also found that psychotherapy effected little improvement in the marriage although it was effective in enhancing other aspects of social functioning.Conjoint marittherapy might well be more effective, but not all husbands are willing to be engaged in treatment; nevertheless, this seems likely to be the preferred approach. Equal attention needs to hos paid to the difficulties in parenting associated with mental disorder. Most of the parenting differences involve subtle aspects of parent-child interaction--tather than gross neglect or abusealthough occasionally this occurs (Hawton et al. 1985); nevertheless, the findingson insecurity of attac ment indicate that there are significant consequencee for the children.Of course, insecure attachments occur commonly in the general population; they are associated with later difficulties in peer relationships (Wolkind and Rutter 1995). Radke-Yarrow et al.'s (1985) findings, however, suggest that the pattern of insecurity includes abnormal features not ordinarily seen in the absence of parental pathology. Sometimes there is a tendency to assume that very young children are not aware of family tensions and disputes and hence are relatively protected from family discord.Tin, evidence firmly contradicts thio sanguine view. Not only does discord frequently interfere with parenting, but also it is clear that toddlers are quick to pick up negative feelings (Zahn-Waxler et al. 1984). Cummings et al. (1985) showed that 2-year-olds typically responded with distress to angry verbal exchanges between adults; and the children subsequently showed increases in aggression between peet3. Young infants may also be vulnerable to cognitive ill effects associated with maternal depression during the early years of their lives. One study (Cogill et al. 1986) showed cognitive deficits that seemed to be specifically associated with (relatively mild) maternal depression during the first year of motherhood. The finding awaits replication. 172 PREVENTION OF MENTAL DISORDERS

Parents who are depressed or suffering from some other form of mental disorder need to be helped to reduce intrafamilial conflicts, to prevent such conflicts from impinging on the children, and to improve their functioning as parents. In addition to direct efforts to bring about improved parenting, there are likely also to be benefits from improving the parents' availability and use of emotional supports. For example, often it may be helpful to encourage the healthy parent to take a greater role in looking after the children. One specific issue that arises with puerperal psychosis is whether the mentally ill women should be encouraged to continue to look after their newborn infants or whether separation might be safer. We lack data on the factors that should be taken into account in making that decision. Mother-and-baby units in psychiatric hospitals have been available for many years (Rutter 1966); often the results for both mothers and babies seem satisfactory, and there may be advantages in babies remaining with their mothers while they are hospitalized (Grunebaum et al. 1975). Yet it is not clear that babies do not suffer from exposure to their severely ill mothers. The matter urgently needs further study. Meanwhile, we must to seek to ensure that the care of infants is as good as it can be when mothers are mentally ill, whether the neonates are jointly admitted with the mothers or remain at home with the fathers or other family members. Individual Differences in Children's Responses

Finally, it is necessary to pay attention to the important universal observation that children differ markedly in their response to parental mental illness. Ordinarily fewer than half succumb to any form of psychiatric disorder, many come through the experience without psychological damage, and some even seem to gain strength from having coped successfully with stress and adversity (El-Guebaly and Offord 1980; Rutter 1985). To some extent, this individual variation is a function of the characteristics and context of the parental mental illness. The children are less likely to be affected adversely if the parental disorder is mild; of short duration; is unassociated with family discord, conflict, and disorganization; is unac lompanied by impaired parenting; and does not result in family breakup.lowever, there are also other featuree that are associated with resilience or vulnerability. During the 1970s, there was much rhetoric about "invulnerable" children who are supposedly unaffected by bad environments (Anthony 1974). On the basis of mainly anecdotal observations, but with rome quantitative data on families with a psychotic parent (Lander et al. 1978; Worland et al. 1984), Anthony argued that invulnerability involved lack of involvement and of identification with the sick parent and an ability to sustain independence and psychological separation from the illness. The suggestions are in keeping with the observations of Others, but knowledge is lacking on the importance of these variables when considered in conjunction with other risk and protective factors.

1 DISORDERS IN PARENTS 173

Garmezy (1971, 1974) drew attention to the relatednotion of "competence" shown by children at psychiatric risk throughbirth to, and rearing by, a schizophrenic parent. His emphasis differed from Anthony's inits concern with social coping skills ratLer than with intrapsychicmechanisms, but it shared the focus on personal qualities. His own research(Garmezy et al. 1984) has been important in further elucidating firnature of these qualities; but it also has led to a broadening of the range of protective factors toinclude a warm, emotionally supportive family milieu and the presence of anextended support system, as well as personality dispositions (Garmezy 1985).Mrsten and Garmezy (1985) pointed to the need F r preventive interventions to include steps topromote all three types of preventionthrough teaching socialproblem-solving skills, reducing family discord, and bringing support agencies intoplay. Feldman et al. (1987) emphasized a somewhat similar set of strategiesbut noted the difficulties engendeted by many families' poor use of resourcesand the need to take account of opportunities outside, as well asinside, the family.They suggested the possible gains from children's exposure towell-adjusted individuals and from their participation in rewarding activitiesoutside the con- fines of the home. Sex of Child Many studies have found that boys tend to be more susceptible toill effects following exposure to family discord or disruption (Rutter 1982;&slow and Hayes 1986); this sex difference is likely to influence children's responsesto parental mental disorder ',;!.thvugh we lack adequate data on whetherboys do indeed differ from girls in this respect). Rutter and Quinton (1984)found little difference between boys and girls in their overall likelihood ofdeveloping disorder in association with parental mental disorder. However, boystended to develop disorders earlier following family discord, and girls weresomewhat mote likely to develop disorder in theabsence of discord. In childhood, there was some suggestion that children of the same sex c ;the ill parent were most at risk.In that vein, Radke-Yarrow et al. (1988) observed that depressed mothers tended to seek comfort from their daughters in a way thatseemed to draw them into their own depressed state,Similarly, Hinde and Stevenson- Hinde (1987) noted that negative mood in mothers was associated with an increase in the cuddling of daughters but not of sons. It is probable that several different mechanisms are involvedin sex differences in response to discord (Rutter in presbb). There may be a tendency for parents to be more likely to quarrel in front of their sons thantheir daughters (Hetherington et al. 19E2). Moreover, boys are more liable to respond to conflict with aggression and girls with distress (Cummings et al, 1985). The boys' mode of response may lead to parents' being less tolerant of their difficulties andhence more likely to respond negatively with scapegoating orother forms of focused criticism. 174 PREVENTION OF MENTAL DISORDERS

Aggression in boys is more likely to be met with parental punitivenessor backing away; and both types of responses are liable to increase the likelihood of escalation (Dunn and Kendrick 1982; Maccoby and Jack lin 1983; Patterson 1982). When a family breaks up, sons are much more likely that, daughters to be placed in some form of institutional care (Packman 1986)a placement that increases their psychiatric risk (Quinton and Rutter 1988; Walker et al. 1981). Temperamental Factors Children's temperamental features have been shown to be associated with differences in their response to a variety of stress situations.Rutter and Quinton (1984) showed that children of mentally ill parents with high-risk temperamental attributes (defined in terms of the constellation of negative mood, low regularity, low malleability, and low fastidiousness)were twice as likely as temperamentally easy children to develop an emotionalor behavioral disturbance.More detailed analyses suggested that the children's characteristics put them at increased risk because they elicited different parental behavior. Children with temperamental risk factorswere only slightly more likely to come from discordant homes but, within such homes, theywere more than twice as likely as other children to be the target of parental hostility and criticism. Similarly, Lee and Bates (1985) found that toddlers withdifficult temperaments were more likely to elicit coercive responses from their mothers. The implication is that the children's temperamental features in part determined the likelihood that they would be drawn intoa maladaptive pattern of parental interaction, a pattern that in turn predisposed to the developmentof psychiatric disturbance in the child. Protective Features As with other stress situations, a good relationship withone or more parents has been found to be a protective factor (Rutter 1971, inpress a). The presence in the :rome of an emotk wily supportive, mentally healthy otherparent seems generally beneficial for both the children and the ill parent. Littleis known about whether good relationships outside the homecan serve a similar protective function. Presumably much dependson their closeness and intimacy and hence on the extent to which they can be used supportively by the children. Schreiber (1985) found that the quality (but not quantity) ofpeer relationships was associated with levels of disorder in 9- to 12-year-old boys in both divorced and nondivorced families. Moreover, Pelligrini et al. (1986),in a study of 23 children of patients with bipolar affective disorder and 23 childrenof normal controls, found that offspring without psychiatric probl-ms tendedto have more adequate personal supports than those with psychiatric disorder.Affective disorder in the children was much less likely if they hada best friend as supporter Although good relationships are onlyone of several possible mitigating factors (El-(uebaly and Offord 1980; Rutter 1985, 1987a), others havebeen littk, studied in relation to the effects of mental illness in mothersor fathers. Bleuler I '3 DISORDERS IN PARENTS 175

of schizophrenic patients, com- (1978), discussing hisstudy of the children could be health enhancingif they were both manage. mented that the stresses in able and of a kind thatgives rise to rewardingtasks. The suggestion is from other situationsthat resilience is characterized keeping with the evidence (Rutter 1985; 1987a); by some sort of action todeal with the strws situation of self-esteem, a belief inone's ovr, such self-efficacy isassaciated with a sense ability to cope with life'schallenges, and a repertoireof social problem-solving cognitive set seems likelyto be fostered by both approaches. This positive and secure, stable,affeetional relationshipsand prior experiences of success achievement. accounting for Clearly, protective mechanisms arelikely to be important in children's responses to parentalmental disorders, as individual differences in (Rutter 1987b, in press a)but they are with otherpsychosocial adversities empirical data on their effects arelacking. Clinical Implications individual differences in Three main implications stemfrom the findings on children's responses to parentalmental disorders.First, it is important to how children in appreciate that such differencesexist, and to focus clinically on responding, rather than to assume the families of adultpatients are actually that they are generally atrisk. Second, resilience or vulnerabilityresides in patterns ofpatient-child inter- constitutional qualities of action and transaction, and notjust in some inherent Children put themselves at children (althougl doubtlessthey play a role). increased or reduced riskby the ways in whichthey react, ways that in turn influence how parents respond tothem. There is a need tohelp children develop of good strategies successful coping strategies,recognizing that there is a range and that what works bestwith one child may notwork as well with another. Third, there needs to be a concernwith the family as a whole,together with patient as an individual. As its kin and peer-grouplinks, and not just with the (1985) noted, the families ofpsychiatric patients should Ku ipers and Bebbington problems are real and be seen as a positive endirreplaceable resource. Their need to be comidered as acentral concern of clinicians;but their positive qualities also can make asignificant difference to the courseof the patient's disorder, as well as to its impact onthe children. Conclusions

constitutes a There is good evidence thatpeychiatric disorder in parents significant psychiatric riskfactor for children. Themechanisms involved in- (usually as a contributoryfactor to vulnerability, elude genetic transmission such); damage to rather than as direct inheritanceof psychiatric disorders as 176 PREVENTION OF MENTALDISORDERS

the fetus (especially from high maternal alcoholconsumption); anda variety of family environmentaleffects (discord and impaired important). The evidence parenting are particularly concerning these mechanismshas implications for preventive policies and practice;but, so far, dataare lacking on the extent to which preventiveinterventions are in factprotective.

References

Akiskal, H.S.; Downs, J.;Jordan, P.; Wabon, S.; Affective disorders in Daygherty, D.; and Pruitt,D.B. referred children andyounger siblings of manic- depressives. Archives ofGeneral Psychiatry42:996-1003, 1985. Anthony, E.J. The syndrome of thepsychologically invulnerable Anthony, E.J., and child.In: Koupernik, C., eds. TheChild and His Family:Children at Psychiatric Risk. Vol. 3.New York: Wiley,1974. Aronson, M.; Kyllerman, M.; Sabel, K.G.; Sandin,B.; and Olegard, R.Children of alcoholic mothers:Developmental, perceptual teristics and compared and behaviouralcharac- to matched controls.Acta Paediatrica 74:27-35, 1985. Scandinavica

Asarnow, R.F., andMacCrimmon, D.J. Residual cally remitted schizophrenia: performance deficit inchroni- A marker ofschizophrenia? Journalof Abnormal Psychology 87:597-608,1978,

Asarnow, R,F.; Steffy,R.A.; MacCrimmon,D.J.; and Cleghorn, tional aesessment of J.M. An atten- foster children atrisk from schizophrenia.Journal of Abnormal Psychology86:267-275, 1977. Beardslee, W.R.; Bemporad, J.; Keller, M.B.; andKlerman, G.L. Childrenof parents with majoraffective disorder:A review. Psychiatry 140;832-852,1983. American Journal of

Belle, D., ed. Livesin Stress: Womenand Depression. Sage, 1982. Beverly Hills, Calif.:

Bettes, B.A. Maternaldepression and motherese: Temporal andinteractional features. ChildDevelopment 59:1089-1096,1988. Billings, A.G., andMoos, R.H. Children olpacents with unipolardepression: A controlled 1-year follow-up.Journal of Abnormal 1986. Child Psychology14:149-166,

Birtchnell, J., andKennard, J. Does inaritalmaladjustment leadto mental illness? Social Psychiatry18:79-88, 1983a. DISORDERS IN PARENTS 177

Birtchnell, J., and Kennard,J. Marriage and mentalillness. British Journal of Psychiatry 142:193-198,1983b. Bleuler, M. The SchizophrenicDisorders:Long-term Patient and Family Studies. New Haven: YaleUniversity Press, 1978. and Clonincer, R. Maternalinheritame of alcohol Bohman, N.; Sirardsson, S.; Archives of General abuse:Cross-fostering analysis of adopted women. Psychiatry 38:965-969, 1981. after an acute Bothwell, S., and Weissman,M.M. Social impairnent four years depressive episode. AmericanJournal of Orthopsychiatry47:231-237, 1977. Breslau, N.; Staruch, K.S.;and Mortimer, E.A.Psychological distress in of Children mothers of disabled children.American Journal of Diseases 136:682-686, 1982. Brown, G.W., and Davidson, S.Social class, psychiatric disorderof mother, and accidents to children. Lanceti:378-381, 1978. Brown, G.W.; Harris, T.O.; andBifulco, A. Long-term effects ofearly loss of parent. In: Rutter, M.;Izard, C.; and Read, P., eds.Depression in Young People: Developmental and ClinicalPerspectives. New York: GuilfordPress, 1986. pp. 251-296. Brown, G.W.; Monck, E.M.;Carstairs, G.M.; and Wing, J.K.Influence of family life on the course of st'.zophrenic illness. British Journal ofPreventive and Social Medicine 16:55-68,1962. Byrne, E.A., and Cunningham,C.C.The effects of mentallyhandicapped children on familiesA conceptualizedreview. Journal of ChildPsychology and PE whiatry 26:847-864,1985. Cadoret, R. Evidence for geneticinheritance of primary affoctive disorderin adoptees. American Journal ofPsychiatry 136:463-466, 1978. Cadoret, R.J. Genes, environmentand their interaction in thedevelopment of psychopathology. In: Sakai, T., and Tsuboi,T., eds. Genetic Aspects of Human Behavior. Tokyo: Igaku-Shoin,1985. pp. 166-176. Cadoret, R.J. ; Cain, C.A.; andCro-.7e, R. Evidence for gene-environmentinter- action in the development ofadolescent antisocial behavior.Behavior Genetics 13:301-310, 1983. Cadoret, R.J.; Troughton, E., Moreno,L.; and Whitters, A. Early lifepsychoso- cial events and adult affective symptoms.In: Robins, L.N., and Rutter, M.,eds. Straight and Devious Pathwaysfrom Childhood to Adulthood.New York: Cambridge University Press, in press. 178 PREVENTION OF MENTALDISORDERS

Caspi, A.; Bem, B.J.; and Elder, G.H.Continuities andconsequences of interactional styles across the lifespan. Journal of Personality 57:1-32, 1989.

Cloninger, C.R.; Bohman, M.; andSigvardsson, S. Inheritance of alcoholabuse: Cross-fostering analysis of adoptedmen. Archives of General PsychiaPy 38:861- 868, 1981.

Cogill, S,R; Caplan, H.L.;Alexandra, H.; Robson, K.M.; and&mar, R. Impact of maternal postnatal depression on cognitive development ofyoung chikiren. British Medical Journal 292:1165-1167,1986.

Cohn, J.F.; Campbell, S.B.; andMatias, R. Face-to-faceinteractions of postpartum depressed and nondepressed mother-infant pairs at two months..Develop- mental Psychology, inpress.

Cooke, K.; Bradshaw, J.; Glendinning,C.; Parker, G.; and Badwin,S. "1970 Cohort: 10 Year Follow UpStudy: Interim Report."Working paper of the Social Policy Unit, University of York.York, England: Universityof York, 1982. Cox, AD.; Puckering, C.;Pound, A.; and Mills, M.The impact of maternal depression in young children. Journal of ChildPsychology and Psychiatiy 28:917-928, 1987.

Coyne, J.C.; Kahn, J.; andGotlib, I.H. Depression.In: Jacob, T., ed. Family Interaction and Psychology:Theories, Methods andFindings. New York: Plenum Press, 1987.pp. 509-533.

Crowe, R.R. Antisocial personalitydisorder. In: Tarter, RE.,ed. The Child at Psychiatric Risk. Oxford: OxfordUniversity Press, 1983.pp. 214-244. Cummings, E.M.; Lanott, R.J.;and Zahn-Waxler, C.Influence of conflict between adults on the emotions and aggression ofyoung children. Developmen- tal Psychology 21:495-507,1985. Cytryn, McKnew, D.W.; Zahn-Waxler,C.; and Gershon, E.S.Developmental issues in risk research: Theoffspring of effectively illparents. In: Rutter, Izard, C.; and Read, P., eds. Depression in Young People:Developmental and Clinical Perspectives. NewYork: Guilford Pres&1986. pp. 163-188. Davenport, YB.; Zahn-Waxler,C.; Adland, M.C.; andMayfield, A. Early child- rearing practices in families witha manic-depressive parent. AmericanJournal of Psychiatry 141:230-235,1984. Decina, P.; Kastembaum,C.J.; Farber, S.; Kron, L.; and Fieve, R.R. Clinical and Gargan, M.; Sackeim, H.A.; psychological assessment ofchildren of bipolar probands. American Journalof Psychiatyy 140:548-553,1983. Doane, JA; Goldstein, M.J.;and Rodnick, E.H. Parental style and the development patterns of affective of schizophreniaspectrum disorders. Family Process 20:337-349, 1981.

1 ci 7 DISORDERS IN PARENTS 179

Dunn, J., end Kendrick, C. Siblings: Love, Envy and Understanding. London: Grant McIntyre, 1982. Earls, F. On the familial transmission of child psychiatric disorder. Journal of Child Psychology and Psychiatry 28:791-802, 3987. Earls, F.; Reich, W.; Jung, K.; and Cloninger, C.R Psychopathology in children of alcoholic and antisocial parents. Akoholisni: Clinical and Experimental Research 12:481-487, 1988. El-Guebaiy, N., and Offord, D.R. The competent offspring of psychiatrically ill parents. Part I: A literature review. Canadian Journal of Psychiatry 25:457- ,463, 1980. Emery, RE. Interparental conflict and the children of discord and divorce. Prychologicca Bulletin 9'2:310-330, 1982. Emery, R.E.; Weintraub, S.; and Neale, J.M. Effects of marital discord on the school behavior of children with schizophrenic affective disorders and normal parents. journal of Abnormal Psychology 10:216-228, 1982. Feldman, RA.; Stiffman, InR.; and Jung, K.G. Children at Risk: In the Web of Parental Mental Illness, New Brunswick: Rutgers University Press, 1987. Fendrich, M.; Warner, V.; and Weissman, M.M. Family risk factors, parental 4epression and childhood psychopathology.Developmental Psychology, in press. Field, 'En Early interactions between infants and their postpartum depressed mothers. Infant Behavior and Development 7:517-522, 1984. Field, T.; Healy, B.; Goldstein, S.; and Guthertz, M. Behavior state matching and synchrony in mother-infant interactions of nondepressed vs. "depreesed" dyaas. Developmental Psychology, in press, Folstein, S ,E.; Franz, M.L.; Jensen, B.A.; Chase, G.A.; and Folstein, M.F. Conduct disorder and affective disorder among the offspring of patients with Huntington's disease. Psychological Medicine 13:46-52, 1983. Gaensbauer, W.; Harrison, R.J.; Cytryn, L; and McKnew, D.H. Social and affective develo9ment in infants with a manic-depressive parent. American Journal of Pvchiatry 141:223-229, 1984. Garmezy, N.Vulnerability research anl the issue of primary prevention. American Joumal of Orthopsychiatry 41:101-116, 1971. Garmez.,N.The study of competence in children at risk for severe psychopath ',logy.In: Anthony, E.J., and Koupernik, C., eds, The Child and His Family; Children at Psychiatric Risk. Vol. 3. New York: Wiley, 1974. pp. 77-97. 180 PREVENTION OF MENTAL DISORDERS

Garmezy, N. Stressors of childhood.In: Garmezy, N., and Rutter, M,, eds, Stress, Coping and Development in Children. New York: McGraw-Hill, 1983. pp. 43-84. Garmezy, N. Stress resistant children: The search for protective factors. In: Stevenson, J., ed. Recent Research in Developmental Psychopatholog y. A book supplement to the Journal of Child Psychokgy and Psychiatry, No. 4. Oxford: Pergamon Press, 1985. Garmezy, N.; Masten, A.S.; and Tellegen, A. The study of stress and competence in children: A building block for developmental psychopathology.Child Development 55:97-111, 1984. Gershon, E.S.; Hamovit, J.; Guroff, J.; Dibble, E.; Leckman, J.F.; Sceeny, W Targum, S.P.; Nurnberger, J.1.; Goldin, L.R; and Bunney, W.E. A family study of schizoaffective, bipolar I, bipolar II, unipolar and normal controls. Archives of General Psychiatry 39:1157-1167, 1982. Gershon, B.S.; McKnew, D.; Cytryn, L.; Hamovit, J.; Schreiber, J.; Hibbs, E.; and Pellegrini, D. Diagnoses in school-age children of bipolar affective disorder patients and normal controls. Journal of Afftctive Disorders 8:283-291, 1986. Goldstein, M.J. Family factors that antecede the onset of schizophrenia and related disorders: The results of a fifteen year prospective longitudinal study. Acta Psychiatrica Scandinavica 71:7-18, 1985. Goodwin, D.W. Alcoholism and genetics: The sins of the fathers. Archives of General Psychiatry 42:171-174, 1985. Gotlib, L., and Hooley, J.M. Depression and marital distress: Current status and future directions. In: Duck, S., ed. Handbook of Personal Relationships. Chichester, England: Wiley, 1988. pp. 543-570. Gottesman, LI., and Shields, J. A critical review of recent adoption, twin and family studies of schizophrenia:Behavioral genetics perspective. Schizophrenia Bulletin 2:360-400, 1976. Grunebaum, H.; Weiss, J.L.; Cohler, B.J.; Hartman, C.R; and Gallant, D.H. Mentally Ill Mothers and Their Children. Chicago: University of Chicago Press, 1975. Hammen, C.; Gordon, D.; Burge, D.; Adrian, C.; Jaenicke, C.; and Hiroto, D. Maternal affective disorders, illness, and stress:Risk for children's psychopathology. American Journal of Psychiatry 144:736-741, 1987. Hawton, K ; Roberts, J.; and Goodwin, G. The risk of child abuseamong mothers who attempt suicide. British Journal of Psychiatry 146:486-489, 1986. Heston, LI,Psychiatric disorders in fo3ter home reared children of schizophrenic mothers. Britis4 Journal of Psychiatry 112:819-825, 1966. DISORDERS IN PARENTS 181

parents and Cox, M.; and Cox, R.Effects of divorce on Hetherington, E.M.; Families. Hillsdale,N.J.: L children. la: Lamb,M.E., ed. Non-Traditional Erlbaum Associates,1982. pp. 233-288. A study of the D.; Roberts, F.J.;and Vaughan, P.W. Hinchcliffe, M.; Hooper, their spouses. BritishJournal of interaction betweendepressed patients and Psychiatry 126:164-172,1975. approach Stevenson-Hinde, J.Implications of a relationship Hinde, R.A., and Journal E:221-235, for the study ofgender differencr,s.Infant Mental Health 1987. S.; Oppen- Day, N.; Dooling, E.;Kayne, H.; More lock, Hingson, R.; Alpert, J.J.; drinking and marijuana use heimer, E.; and Zuckerman,B. Effects of maternal Pediatrics 70:539-546,1982. on fetalgrowth and development. disorder in parents andchildren. Hobbs, P. The relativetiming of psychiatric British Journal ofPsychiatry 140:37-43,1982. of ex-institutional Hodgss, J., andTizard, B. IQ andbehavioural adjustment adolescents. Journal ofChild Psychology andPsychiatry 30:53-75, 1989a. B. Social and familyrelationships of ex-institutional Hodges, J., and Tizard, 1989b. adolescents. Journal ofChild Psychology andPsychiatry 30:77-97, In: Jacob, Seilhamer, R.A. Alcoholismand family interaction. Jacob, T., and Methods and Find- T., ed. FamilyInteraction andPsychopathology: Theories, ings. New York:Plenum Press, 1987. pp.535-580. Paul, E.C.) London:Allen & Janet, P. PsychologicalHealing. Vol. 1 (transl.: Unwin, 1925. Bernstein, V.J. Dyads atrisk: Methadonemaintained women Jeremy, R.J., and 55:1141-1154, 1984. and their four-month-oldinfants. Child Development adolescent depressive Kandel, D.B., a IdDavies, M.Adult sequelae of symptoms. Arch, vesof General Psychiatry43:255-262, 1986. H.; and W.R.; Dorer, D.J.;Lavori, P.W.; Samuelson, Keller, M.B.; Be; irdslee, illness on Klerman, G.R. Impact ofsevarity and chronicityof parental affective in children,Archives of General adaptive functioningand psychopathology Psychiatry 43:930-937,1986. and Eaves, L.J.Models for the jointeffect of genotype Kendler, K.S., and Journal of Psychiatry environment on liability topsychiatric illness. American 143:279-289, 1986.

( 182 PREVENTION OF MENTALDISORDERS

Kendler, KS., and Gruenberg, A.M. Anindependent analysisof the Danish adoption study ofschizophrenia. VI. The disorders as defined by relationship betweenpsychiatric DSM-III in the relativesand adoptees. Generul Psychiatry41:555-564, 1984. Archives of Kreitman, N.; Collins,C.; Nelson, B.;and Troop, J. interaction.IV. Neurosis and marital Manifest psychologicalinteractions. Psychiatry 119:243-262,1971. British Journal of Kuipers, L., and Bebbington, P. Relativesas a resource in the functional illneas. British management of Journal of Psychiatry147:465-470, 1986. Lander, H.S.; Anthony, E.J.; Kass, L.;Franklin, L.; and Bass,L. A vulnerability to risk ofparental psychosis. measure of and Chiland, C., eds. In: Anthony, E.J.;Koupernik, C.; The Child and HisFamily: Vulnemble New York: Wiley,1978. pp. 325-333. Children. Vol. 4. Lee, C.I., and Bates, J.E.Mother-child interactionat age two perceived difficulttemperament. Child years and Development 56:1314-1325,1985. Leff, J.P.; Kuipers, L.; Berkowitz, R.;Eberlein-Vries, R.; and controlled trial of socialintervention in the Sturgeon, D. A families of schizophrenicpatients. British Journal ofPsychiatry 141:121-134,1982. Leff, J,P.; Kuipers, L.; Berkowitz, R.;and Sturgeon, D. A intervention in thefamilies of schizophrenic controlled trial of social patients: Twoyear follow-up. British Journal ofPsychiatry 146:594-600,1985. Leff, J,P., and Vaughn, C.The role ofmaintenance therapy expressed emotion inrelapse of schizophrenia: and relatives A two-yearfollow-up. British Journal of Psychiatry139:102-104, 1981. Livingood, A.B.; Daen, P.; and Smith, B.D.The depressed of stimulation forher infant, Journal mother asa source of Clinical Psychology39:569-375, 1983. Lyons-Ruth, K.; Zoll,D.; Connell, D.; mother and her and Grunebaum,H.U. The depressed one-year old infant:Environmental context, interaction andattachment and infant mother-infant development. In:Tronick, E., and T., eds. TheDepressed Mother and Field, Bass, 1986. Her Young Child,San Francisco:Jossey- Maccoby, E.E., and Jacklin, C.N. The"person" characteristics the familyas environment. In: of children and Magnusson, D., andAllen, V., eds. Development: An Intvactional Perspective, Human pp. 75-91. New York: AcademicPress, 1983. Masten, A.S., and Garmezy, N. Risk,vulnerability and developmental psychopathology. protective factorsin In: Lahey, B.B.,and Kazdin, AE., vances in Clinical ChildPsyci ?logy. Vol. eds. Ad- pp. 1-52. 8. New York:Plenum Press, 1985.

10 1 183 DISORDERSIN PARENTS abnormal influences onnormal and I. Genetic Adolescent McGuffin, P.,and Gottesman, L., eds.Child and Rutter, M.,and Hersov, Scientific,1985. development.In: Oxford:Blackwell Approaches.2d ed, Psychiatry:Modern 17-33. pp. TheCamberwellcollaborative andBebbington, P. of depressed McGuffin, P.;Katz, R.; adversity inthe relatives study. IV.Depression and 1988. depression of Psychiatry152:775-782, probands.British Journal Offspring of women I.; andKakj, L. and Ntislund, B.;Persson-Blennow,interaction atthree-and-a-half McNeil, T.F.; Mother-infant 1985. with nonorganicpsychosis: 71:551-558, Acta PsychiatricaScandinavica six monthsof age. transmis- studysupportinggenetic and Rainer,J.D. Adoption 1977. Mendlowicz, J., illness. Nature268:326-329, sion inmanic-depressive Towards anaetiological clas- and Reveley,A.M. Lewis, S.W.; 1985. Murray, RM.; Lancet1:1023-1026, sification ofschizophrenia. L.; andMalmquist-Larsson, I.; McNeil,T.; Kaij, the Persson-Blennow, Infantattachment to Niislund, B.; psychosis: 1984a. of womenwithnonorganic 69:231-241, A. Offspring PsychiatricaScandinavica mother at oneyear of age.Acta andMalmquist-Larsson, I.; McNeil,T.; Kaij, L.; during the Niislund, B.;Personn-Blennow, psychosis: Fearof strangers with nonorganic 1984b. A. Offspringof women Scandinavica69:435-444, of life. ActaPsychiatrica first year Kaij, L.Offspring of women I.; McNeil,T.F.; and weeks of Naslund, B.;Persson-Blennow, interaction atthree and six psychosis:Mother-infent 1985. with nonorganic Scandinavica71:441-450, age.Acta Psychiatrica Journal of of adultschizophrenia. K.H.Childhood precursors Nuechterlein, 27:133-144,1986. Psychology andPsychiatry Child PaediatricaScandinavica alcoholicfathers. Acta Nylander, I.Children of 1960. 49(Suppl.):121, Scientific, 1986. Care? Oxford:Blackwell Packman, J .Who Needs rearing and Schulsinger, H.Institutional Teasdale, T.W.;and mothers: Aprospectivehigh Parnas, J.; of schizophrenic diagnostic outcomein children 42:762-769, 1985. GeneralPsychiatry risk study.Archives of Eugene, Ore.:Castalia, 1982. Coercive FamilyProcess. Patterson, G.R. Gershon, Cytryn, L.;McKnew, D.H.; Nackman, P.; in childrenof Pelligrini, D.;Cosisky, S.; sociaI resources Cammuso, K.Personal and subjects. E.; Hamovit,J.; and children ofnormal control bipolar affectivedisorders and patients with 143:856-861,1986, American Journalof Psychiatry 184 PREVENTION OF MENTAL DISORDERS

Persson-Blennow, I.; Naslund, B.; McNeil,T.; Kaij, L.; and Malmquist-Larsson, A. Offspring ofwomen with nonorganic psychosis; Mother-infantinteraction at three days of age. Acta PsychiatricaScandinavica 70;149-159, 1984. Plomin, R. Development, Geneticsand Psychology. Hillsdale,N.J.: L Erlbaum Associates, 1986.

Plomin, R., and De Fries, J.C.Origins of Individual Differencesin Infancy: The Colorado Adoption Project. Orlando,Fla.: Academic Press, 1985. Pogue.Geile, M.F., and Rose,R.J.Psychopathology: A behavior-genetic perspective.In: Jacob, T., ed.Family Interaction andPsychopathology: Theories, Methods and .Findings,New York: Plenum Press, 1987.pp. 629-650. Porter, R.; O'Conner, M.; andWhelan, J., eds. Mechanismsof Alcohol Damage in Utero. Ciba Symposium105. London: Pitman, 1984. Quinton, D., and Rutter, M.Parenting Breakdown: Makingand Breaking of Inter-Generational Cycles. Aldershot,Hants, England; Gower, 1988. Quinton, D.; Rutter, M.; andGulliver, L.The social functioning inearly adulthood of the children of psychiatric patients. In: Robins,L.N., and Rutter, M., eds. Straight and Devious Pathways From Childhoodto Adulthood. New York: Cambridge UniversityPress, in press.

Radke-Yarrow, M.; Cummings,E.S.; Kuczynski, L.; andChapman, M. Patterns of attachment in two- and three-year olds in normal familiesend families with parental depressiovi. ChildDevelopment 56;884-893, 1985. Radke-Yarrow, M.; Richters, J.;and Wilson, W.E. Childdevelopment in a nety-ork of relationships. In;Hinde, R.A., and Stevenson-Hinde,J., eds. &la- tions Within Families: MutualInfluences. Oxford; Oxford 1988. University Press,

Rice, E.P.; Ekdahl, M.C.;and Miller, L.Children of Mentally Ill Problems in Child Care. New Parents: York: Behavioral Publications,1971. Richman, N.; Stevenson,J.; and Graham, P.J.Preschool to School: A Be- havioural Study, London:Academic Press, 1982. Robins, L.N. Theconsequences of conduct disorder in girls. Block, J.; and Radke-Yarrow, In: 'Olweus, D.; M., eds. Development ofAntisocial and Prosocial Behavior: Research, Theoriesand Issues. New York; pp. 385-414. Academic Press, 1986. Rodnick, E.H., and Goldstein, M.J. Premorbid adjustmentand the recovery of mothering function inacute schizophrenicwomen.Journal of Abnormal Psychology 83;622-628, 1974.

1P3 DISORDERS IN PARENTS 185

M.M.; Prusoff, BA;and Herceg-Baron,RL. Rounsaville, B.J.; Weissman, Comprehensive Marital disputes andtreatment outcomein depressed women. Psychiatry 20:483-490,1979. Paper Enough? SubstituteCare and Socialization." Roy; P. "Is Continuity Psychiatry Sec- presented at SpringScientific Meeting, Childand Adolescent March 1983. tion, Royal College ofPsychiatrists, London, Study. Sick Parents: AnEnvironmental and Psychiatric Rutter, M. Children of Oxford Univer- Institute of PsychiatryMaudsley MonographNo. 16. London: BUY Press, 1966. the children. separation:Psychological effects on Rutter, M. Parent-child 1971. Journal of Child Psychologyand Psychiatry 12:233-260, approaches to the studyof develop- Rutter, M.Epidemiological-longitudinal Collins, W.A., 3d. TheConcept of Development.The Minnesota ment. In: L Erlbaum Associates, Symposia on Child Psychology.Vol. 15. Hillsdale, N.J.: 1982. pp. 105-144. of and development:1. Childhood antecedents Rutter, M. Psychopathology Zealand Journal ofPsyd:iatry adult psychiatric disorder.Austrialian and New 18:225-234, 1984. Protective factors andresistance Rutter, M. Resilience inthe face of adversi4y. Psychiatry 147:598-611, 1985. to psychiatric disorder.British Journal of psychiatric risk factor. In:Hales, Rutter, M. Parentalmental disorder as a eds. American PsychiatricAssociation's Annual Review. RE., and Frances, AJ., Association, 1987a. pp.647- Vol. 6. Washington,D.C.: American Psychiatric 663. mechanisms. AmericanJour- Rutter, M. Psychosocialresilience and protective nal of Orthopsychiatry57:316-331, 1987b. mechanisms. In: Rolf, J.; Rutter, M. Psychosocialreelience and protective Nuechterlein, K.; and Weintraub,S., eds. Risk and Masten, A.; Cicchetti, D.; New York: Protective Factors inthe Development ofPsychopathology. Cambridge University Press,in press a. considerations in the effects Rutter, M. Commentary:Some focus and process Developmental Psychology, in pressb. on children ofparental depression. 0.; and Shaffer, D.Head injury. In: Rutter,M., ed. Rutter, M.; Chadwick, Livingstone, 1983. Developmental Neuropsychiatry.New York: Churchill pp. 83-111. Trends and Perspectives. Rutter, M., and Giller,H. Juvenile Delinquency: Harmondsworth, Middlesex,England: Penguin, 1983.

n 186 PREVENTION OFMENTAL DISORDERS

Rutter, M.; Izard, C.; and Read, P.,eds. Depression mental and Clinical in Young People:Develop- Perspectives. New York:Guilford Prs,1986. Rutter, M.; MacDonald, H.; Le Couteur,A.; Harrington, A. Geneticfactors in child R.; Bolton, P.;Bailey, nal of Child psychiatric disorders.II. Empiricalfindings. Jour- Psychology andPsychiatry, inpress a.

Rutter, M., andMadge, N. Cycles of Disadvantage:A Review of London: HeinemannEducational, 1976. Research. Rutter, M., andQuinton, D. Parental psychiatric disorder:Effects on children. Psychological Medicine14:853-880, 1984. Rutter, M.; Quinton,D.; and Hill, J. Adult outcomeof institution-reared children: Malesand femalescompared. Straight and In: Robins, L.N.,and Rutter, M.,eds. Devious PathwaysFromChildhood to Adulthood. Cambridge UnhersityPress, inpress b. New York: Rydelius, PA Children of alcoholicfathers. Acta (Suppl. 286),1981. Paediatrica Scandinavica Sabbeth, B.F., and Leventhal, J.M.Marital adjustment illness: A critiqueof the literature. to chronic childhood Pediatrics 73:762-768,1984. Sameroff, A.J.;Seifer, R.; and Zax,M. Early emotional disorder. development of childrenat risk for Monographs of theSociety for Research ment 47(No. 199),1982. in Child Develop- Scarr, S., andMcCartney, K How theory of people make theirown environments: A genotype-environmentaleffects. Child Development 54:424-435,1983. Schreiber, M.D.'Peer Social Support and Coping inChildren of Divorcedand Nondivorced Families."Paper presentedat the Annual American PsychologicalAssociation, Los Convention of the Angeles, Calif., Aug.1985. Schulsinger, F.;Parnas, J.; Peterson, Mednick, S.A.; E.T.; Schulsinger,H.; Teasdale, Moller, L.; andSilverton, L. T.W.; offspring of Cerebral ventricularsize in the schizophrenic mothers:A preliminary Psychiatry 41:602-606,1984. study. Archivesof General St. Claire, L., and Osborn, A.F.The ability and been "in-care"or separated from behaviour of childrenwho have their parents.Early Child Care 28(3, wholeissue), 1987. Development and Stein, A.; Gath, D.H.; Bucher, J,;Bond, A.; Day, relationship between A.; and Cooper,B.J. The postnatal depressionand mother-child mitted forpublication. interaction. Sub- DISORDERS IN PARENTS 187

In: Lahey, B.B., Steinhausen, H.-C., andSpohr, H.-L, Fetalalcohol syndrome. Clinical Child Psychology.Vol. 9. New aad Kazdin, A.E.,eds. Advances in York: Plenum Press,1986. pp. 217-243. Psychopathology in theoffspring Steinhausen, H.-C.; Gobel,D.; and Neetler, V. American Academy ofChild Psychiatry of alcoholic parents.Journal of the 23:465-471, 1984. perinatal hazards. In:Rutter, M., ed.Developmental Stewart, A. Severe Livingstone, 1983. pp.15-31. Neuropsychiatry. NewYork: Churchill in aggressiveand Stewart, M., and deBlois, S.Father-son resemblances Psychiatry 142:78-84;143:310-311, antisocial behaviour.British Journal of 1983. disorder in parents C.S.; and Cummings,C. Psychiatric Stewart, MA; de Blois, disorder, Journal ofChild Psychol- of hyperactive boysand those with conduct ogy andPsychiatry 21:283-292,1980. Sandman, B.M.; Kirchner,G.L.; Streissguth, A.P.; Martin,D.C.; Barr, H,M,; Intra-uterine alcohol andnicotine exposure:Attention and and Darby, B.L, 20:533-541, reaction time in4-year-old children.Developmental Psychology 1984. C.; Danforth, H.; andFreeman, Strobe, M.; Morrell, W.;Burroughs, J.; Lampet, of bipolar I disorderin adolescence:Early-onset symptoms R. A family study lithium resistance. Journalof Affective linked to increasedfamilial loading and Disorders 15:255-268,1988. P.K; Lannotti, R.J.;Hollembeck, B.E.; andZahn- Susman, E.J.; Trickett, abusive and normalmalee. Waxier, C. Child rearingpatterns in depressed, American Journal ofOrthopsychiatry 55:237-251,1985. Moring, J.; Kaleva, M.;Wahlberg, Tienari, P.; Lahti, I.;Sorri, A.; Naarala, M.; offspring of schizophrenicsand controls: K -E.; and Wynne,L.C. Adopted away and family study ofschizophrenia.In: Robins, L.N., The Finnish adoptive From Childhood toAdulthood. Rutter, M., eds. Straightand Devious Pathways NEW York: CambridgeUniversity Press, in press. Moring, J. Naarala, M.; Wahlberg,K -E.; Pohjola, J.; and Tienari, P.; Sorri, A.; Acta Interaction of geneticand psychosocialfactors in schizophrenia. Psychiatrica Scandinavica71(Suppl. No. 319), 1985. sampling variation uponthe Torgersen, S. Genetics ofneurosis: The effects of Psychiatry 142:126-132,1983. twin concordance ratio.British Journal of nosological aspects ofschizotypal and borderline Torgersen, S. Genetic and 41:546-554, 1984. personality disorders.Archives of General Psychiatry 188 PREVENTION OFMENTAL DISORDERS

Tronick, E., and Field, F., eds. MaternalDepression and Infant San Francisco:Jossey-Bass, 1986. Disturbance. Vandenberg, S.G.;Singer, S.M.; andPau ls, D.L. The Disorders in Adultsand Children. Heredity of Behavior New York: PlenumMedical, 1986. Vaughn, C.E., andLoft J.P. The influence of familyand social factors course of psychiatric illness.British Journal on the of Psychiatry129:125-137, 1976. Vaughn, C.E.;Snyder, KS.;Jones, S.; Freeman, Family factors in W.B.; and Falloon,I.R.H. schizophrenic relapses:A replicationin Calikrnia of research on expressedemotion. Archives British 1984. of General Psychiatry41:1169-1177,

Walker, E., andEmory, E. Infantsat risk for schizophrenic parents. psychopathologyOftlpring of Child Development54:1269-1285, 1983. Walker, E.F.; Cudeck, R; Mednick,SA; and Schulsinger, absence and F. Effects ofparental institutionalization ofdevelopment of clinical risk children.Acta Psychiatrica symptoms in high- Scandinavica 63:95-109,1981. Watt, N.; Anthony, E.J.; Wynne, L.C.;and Rolf, J.E.,eds. Children Schizophrenia: Alinigitudinal Perspective. at Risk for Press, 1984. New York: CambridgeUniversity

Weissman, M.M.;Gammon, G.D.;John, K.; Prusoff, B.A; and Merikangas, KR;Warner, V.; Sholomskas, D.Childmn of depressed psychopathology and parents: Increased early onset ofmajor depression. Psychiatry 44:847-853,1987. Archives of General Weissman, M.M.;Gersion, E.S.; Kidd, Dibble, E.; K.K; Prusoff,B.A; Leckman, Hamovit, J.;Thompson, W.D.; J.F.; Psychiatric disorders Paula, D.L.; andGuroff, J.J. in the relativesof probands Archives of General with affectivedisol ders. Psychiatry 41:13-21,1984a. Weissman, M.M.;John, K.; Merikangas, P.; Gammon, KR; Prusoff,B.A.; Wickramaratne, G.D.; Angold, A.;and Warner, V. children: Generalhealth, social and Depressed parentsand their of Diseases of psychiatric problems.American Journal Children 140:801-805,1986. Weissman, M.M., and Paykel, KS.The Depressed Relationships. Chicago: Woman: A Stz1 -ly of Socia2 University of ChicagoPress, 1974. Weissman, M.M.; Prusoff, BA.;Gammon, G.D.; J.F.; and Kidd,KK. Pochopathology Merikangas, K.R.;Leckman, and normal of the children(ages 6-18) of parents. Journalof the American depressed 23:78-84, 1984b. Academy of ChildPsychiatry

17 DISORDERS IN PARENTS 189

Wolkind, S., and Rutter, M.Children who have been "In-Care." An epidemiological study. Journal of Chikl Psychology and Psychiatry 14:97-105, 1973. Wolkind, S., and Rutter, M. Separation, loss and family relationships.In: Rutter, M., and Hersov, L., eds. Child and Adolescent Psychiatry:Modern Approaches. 2d ed. Orford: Blackwell Scientific, 1985. pp. 84-57. Worland, J.; Janes, C.L.; Anthony, E.J.; McGinnis, M.; and Cass, L. St. Louis Risk Reses-ch Project: Comprehensive progress report of experimental studies. In: Watt, N.F.; Anthony, E.J.; Wynne, L.C.; and Rolf, J.E., eds. Children at Risk for Schizophrenia: A limgitudinal Perspective. New York:Cambridge University Press, 1984. pp. 105-147. Zahn-Waxler, C.; Kochansha, G.; Krupnic, J.; and McKnew, D. Patterns of guilt in children of depressed and well mothers. Developmental Psychology, in press. Zahn-Waxler, M.; McKnew, D.H.; Cummings, M.; Davenport, Y.B.; and Radke- Yarrow, M. Problem behaviors and peer interactions of young children with a manic-depressive parent. American Journal of Psychiatry 141:236-240, 1984, Zaslow, M.J., and Hayes, C.D.Sex differences in children's response to psychosocial stress: Towards a cross-context analysis. In: Lamb, M.E.; Brown, A.; and Rogoff, B., eds.Advances in Developmental Psychology.Vol. 4. Hillsdale, N.J.: L Erlbaum Associates, 1986. pp. 285-337.

Js CHAPTER 6

Risks forMaladjustmentAssociated with ChronicIllness inChildhood

I. Barry Pless, M.D. National Health Scientist Professor of Pediatrics McGill University Montreal

Terence M. Nolan, M.D.,Ph.D. Senior Lecturer Melbourne University Department of Pediatrics Royal Children's Hospital Melbourne which chronic illness is a Before beginning anexamination of the extent to eesential that the concept ofrisk be risk indicator forpsychiatric disorder, it is primitive sense any factorthat has a causal relation to a clarified. In the most The key, however, is disease may be viewed as arisk factor for that disease. provide no real proof ei the question of causality;statistical associations alone Many other criteria mustbe met, the most importantof such a relationship. Further, a which is evidence that the causepreceded the outcome of interest. distinction should be madebetween factors that aresubject to change through intervention and thcee that arenot. For example,cholesterol is a determinant, changed; whereas or risk factor,for heart diseasebecause blood lipids can be heart disease, is not viewedin the same sex, which mayalso be associated with (In epidemiologic parlance, sexwould be referred to light for obvious reasons. between causal as an "effectmodifier.") There are, then,important differences when seen from theviewpoint of the clinicianseeking factors, particularly for intervention is to guidance for prevention.In most cases the best strategy determinant factor. Unfortunately,because this is not find ways to modify the identification of groups at often possible, an alternativeto be considered is the increased "risk." It is inthis sense that the role ofchronic illness is examined in this paper.

for the Project Preventioninitiative of the American This paper was prepared by a National Academy of Child and AdolescentPsychiatry. Dr. Phase is supported Dr. Nolan acknowl- Health Scientist Award fromHealth and Welfare, Canada. edges the support of the RoyalChildren's Hospital ResearchFoundation. I P 9 192 PREVENTION OF MENTALDISORDERS

If a preventive or therapeutic interventionis based not underlying factor, but on modifying the rather on theuse of strategies aimedat related factors, the initialdeterminantserves another role. The interventions is enhanced effectiveness of most whea theyare targeted at a subgroupof the population for whom the riskis greatest. Whether under these the interventionis more efficacious circumstances is debatable.But there is little focusing limitedresources on those at highest doubt that by risk, rather thanadopting a more universal approach,the approach islikely to be economical. more efficient andmore This preamble isof particular relevance in the case of chronicillness viewed as a determinant of psychiatricdisorder in childhood. believe that the illness There is littlereason to itself can be modifiedor prevented in most The mkjority of chronicdisorders of childhood instarv;es. are permanent; and towhatever extent medicaltreatments may beeffective, it is virtually applied, regardlessof the extent to certain they will be disorder. which they affect therisk of psychiatric

Notwithstanding thisviewpoint, there isevidence that interventionsmay alter the course of some psychological disease over andabove what achieved byconventional medical can be stress management approaches. Forexample, counseling and may reduce the frequencyof asthmatic attacks children (Perrinet al. 1988, personal in some commun.), and similarapproaches to modify the numberof episodes of appear Despite these diabetic ketoacidosis(White et al. 1984). results, at present suchinterventions are by mainstream of medicalcare for these children. no means part of the rule. Nonetheless, They are theexception, not the if the severity ofmedical symptoms any approach, reactive can be reduced through psychological symptomsare likely to be lessened. studies, however,suggest that certain Some selves accentuate modalities of medicalcare may them- risk for psychiatricdifficulties. For with whichrepeat visits for checkups example, the frequer..cy are requestedmay be related to increasing subsequent psychologicalproblems. Despite these disclaimers,in the majority of "targeting," as described instances the applicationof later, seems metappropriate when chronic viewed as a risk factor.That is, the illness is of the population presence of such an illnessidentifies a sector that hasa greater probability of experiencing variousforms of maladjustmentor psychiatric disorder.It is presumed arises as a resultof the experience that thisexcess risk of having tocope with the underlying biological disorder.The advantagein so identifying better target this population isthen to preventive or therapeuticinterventions. The main challengesfor investigators from the basic working in this fieldshifted longago issue of establishingthe risks associated the task of identifying with chronic illness,to subgroups of thechronically ill who risk. In thecourse of such studies are at still greater some light has been shedon the mechanisms responsible for the underlying linkages. CHRONIC CHILDHOOD ILLNESS 193

Much of the effort toward the ascertainment of subgroup risks is driven not only by this basic stretegy, but further by a paradox that characterizes all work in this domain. On the one hand, virtually all the chronic illnesses of childhood are, unlike chronic illnesses in adulthood, relatively infrequent. Theprevalence of such conditions as diabetes, rheumatoid arthritis, or epilepsy is in the order of 1 to 5 per 1,000 in the general population. There are one or two exceptions to this statement, the most important being asthma. On the other hand, in contrast to the rarity of individual conditions, the chronic disorders of ch ildhood, taken as a whole, have a prevalence of nearly 10 percent. Thus, if it were established that the specific diagnosis or category of illness is of little relevance in the augmentation of riekin contrast to the common feature of chronicity aloneintervention strategies would have to include as much as 10 percent of the general population. This is not an attractive situation; accordingly, as has been stated, much effort has been devoted to trying to further specify the elements of risk most strongly associated with childhood illness. This chapter is divided into three main parts. The first reviews studies to demonstrate the strength and nature of the association between illness and maladjustment broadly defined. These serve as a basis for establishing the concept of risk. The approach in this section is essenUally epidemiologic; the strength of evidence is evaluated in relation to the nature of the research design employed. After considering other attributes of the study, this section includes questions of sampling, sample size, the choice of measures, and the manner in which they are administered. The second section examines the literature dealing with interventions that have been evaluated in a reasonably objective and scientifically acceptable fashion. Also included are a few whose evaluation has been lees rigorous but that appear to shed some light on how the underlying problem may be best approached. The third section attempts to extract insights into the fundamental mechanisms responsible for the association. These insights not only help to understand why the phenomenon of risk exists, but also provide guidance for prevention and treatment. A final caveat:The perspective in this review is that of pediatric epidemiologists. It is not, therefore, a clinical document, and it is certainly not written from the perspective of a child and adolescent psychiatiist and the qualified mental health practitioner. Our values and criteria for efficacy undoubtedly differ from those of child psychiatrists and other clinicians. We tend to view anecdotal or case reports as most useful in generating hypotheses. Our bias is to assume that such studies can rarely, if ever, serve as a basis for testing hypotheses.Conversely, it is obvious that for many child and adokscent psychiatrists and other mental health professionals, epidemiological data may often appear crude and simplistic. Such data fail to take account of, or capture adequately, the dynamics of a complex reactive system. It should be noted,

0 rs 194 PREVENTION OF MENTAL DISORDERS

however, that one of us has collaborated witha child psychiatrist in an earlier critical review of the literature in this area (Plees and Pinkerton 1975),and the conclusions reached did not differ greatly from those described in thispaper. Evidence in Support of Chronic Illness as a Risk Factor

In a major review published in 1986, Nolan and Flees presenteda critical appraisal of the research pertaining to the emotional correlates andconsequences of birth defects. This section provides an overview of the findings and conclusions of that review. The review began by making the point that birthdefects are, in most respects, reasonable analogues forany chronic disabling condition of childhood. It argued that clear distinctions between such termsas emotional disorders, psychosocial maladjustment, and behavioral disordersare difficult and unimportant for a basic description of overall risk. More refined attemptsat quantifying risks for specific types of maladjustment have not been attempteduntil recently (Breslau 1985; Cadman et al. 1987). For the mostpart, the measures used have operationally defined the construct of maladjustmentbeing examined. The term, therefore, is intended to include assessments of behavioralpathology, self- concept, cy self-esteem, as well as psychological disturbanceas observed by clinicians. Of much greater concern than the diversityof measures is the nature of the research design used by investigators to makea case for a causal relationship one in which emotionai problems are viewed asa consaquence of a presumaMy preexisting physical disability. Accordingly, thereview was organized with the most primitive and scientifically least acceptable designsfollowed by those that, progressively, permit the strongest inferencesabout causality. In the first category of design, 17case studies were cited, none of which, by definition, include any comparisongroups (table 1). They cover a spectrum of disorders from diabeteR to cleft palate, includestudies of as few as 20 children to nearly 300 children, and study subjectsranging in age from infancy to adulthood. Although, as stated,many different measures were used, including one that employed only projective tests, onlya few had adequate norms against which the findings could be compared.It is therefore perhaps not surprising that virtually all these studies reportedhigh proportions of maladjustment. Obviously, any conclusions aboutcausality based on such findings must be viewed with skepticism.

2n2 CHRONIC CHILDHOOD ILLNESS 195

The secoLd level of design was deecribed as "case series with controls." In this grtitip 19 studies were summarized (table 2).In each, a control or comparison g-:oup was identified, usually consisting of healthy children, often siblinr In some, the controls were matched individually with cases; in most, the strategy of matching by group was chosen, rendering the two comparable by age, sex, or social status. Agen, the range of disorders was diverse, as were the measures employed. Many were parent rating scales, but also included are teacher rating scales and a few self-assessment procedures such as the Cooperstnith Self Esteem Inventory, the Manifest Anxiety Scale, and the Piers/Harris Self-Concept measure (Piers and Harris 1969). The most popular age range in this group of studies was from 6 to 12 years. Generallythe samples involved fewer than 50 subjects and an equivalent number of controls. The results of these studies present a more balanced picture. Notwithstand- ing issues of statistical power, 12 of the 19 showed statistically significant results in support of the hypothesis that those with chronic disorders are at greater risk for emotional disturbances. In the remainder, the results either were inconclusive or failed to show any significant differences. Also included (table 3) is a summary of three community-based, large-scale, prevalence surveys (Isle of Wight, Rochester, and Genesee County). In these, between 1,700 and 3,000 children were studied, and in each about 10 percent had a chronic physical disorder (CPD). The relative risk of ma/adjustment from these generally sound but nonetheless cross-sectional designs was between 1.3 and 7.4.Several were just above 2.0, suggesting that those with chronic disorders have, on average, twice the risk of maladjustment when compared with healthy controls. More recently, Cadman et al. (1987) published the results of the Ontario Child Health Study, in which 3,294 randomly selected children were surveyed.Their findings confirmed the relationship between chronic illness, psychiatric disorder, and social adjustment problems, showing at _mer- all relative risk of 2.2. We slso examined eight cohort or prospective studies (table 4). Short of an impossible randomized controlled trial, these offer the most conclusive evidence bearing on the causal direction of the association. All but two were large-scale surveys; and, in each, substantial numbers of children were identified with disorders prior to the time when their emotional status was assessed. Although . Ainiestudies the cases were identified retrospectively, the time relationships were always clear. The results from these studies are highly consistent. They report relative risks ranging from 10.9 (in the case of Densen et al.'s (1970) study of those with physical illnesses later rejected by the draft for emotional reasons), to 1.3 in the National Survey in the United Kingdom using parent reports of behavioral symptoms as the measure of disturbance (Plass and Roghmann 1971). Several investigators controlled possible confounding through the use of multivariate statistical analyses, and this lends added credence to their findings.

2 n 3 Table 1. Case Series: Uncontrolled or Norm-ReferencedTests Only

Sample Ago Outcome Reference Disorder sim (Yr) maseurr(s) Findinp 1975Dormer Spina bifida 63 13-19 Clinicil asessisment 66% Depession 1976Boyle et al. Cyatic fibrosis V 13-30 50% Severe isolation Cliniad asemenumt. Rohrechs- 46% Poor or fair daily coping 1976Tavonnina et al. abistes Da-we-Person, Thematic apert.ptlon 78%lsolation or hostility 144 6-19 Piers -Harris Self-Concept Cystic fibrosis AI Ns* a nem tg (LOD butcleaf< =Mgc Jimior Eysenck Personality Inventory All sample DOM Hearing impairment tiesring, higher 'corm but not significant Missouri Childran's Picture Series Sample a norm on emission and activity Heanng a:worming, more aggressive Diabetes more aggressive, active Asthma All ample a norm on alienation Psychological Screening Inventory Diabetes and Poring a norm on alienation, 19'77 Simonds Diabetes defensiveness 1977 40 6-18 Clinics) assesameut Anon ot Diabvtes 292 7.5% Serious psychiatric disorders 10/9Sullivan 0-Adult Clinical assessment and questionnaire Diabetes 106 51.7% Emotional maladjustment 1979 12-16 Diabetic Adjustment Scale MacLean and BeckerHearing impairment 20 7% Maladjustment 1979O'Malley et at 13-20 Clinical *moment Matipancies 113 5-36 5% Maladjuetment Rutter and Graham Interview 36% Mild maladjustment Clinical asessement (combined 23% Moderato-mien maladjustment 1980Grey et al. Diabetes adjustment rating) 20 7-13 Rodgers' Parent Interview 1981 Bywater Cystic fibrosis Cooperemith Self-Esteem Inventory 27 12-16 56% Moderate-eel/ere maladjustment Clinical assessment 63% 'Meaty" in previous year 1981 Helier et al. Rut er B2 Behavice Scale (Teaches-) Cleft lip/palate 96 18-Adult 7% Behavior problem at school Stn. (Awed interview based on school 33% Marginal or inadequate psychosucial and work achievement, pest rind flinctioning 1983OT ougherty et al. Transposition of greet present family life 31 i 9 Behavior Rating Sails (Esaminier)* arteries "Behavioral iifficulties were prominent ina Achenbach Syn.ptom Checklist sub-group of these children." (Parent)* See notes at :he end of the table Table 1. Case Series: Uncontrolled or Norm-Referenced Tests Only (continued)

Sample AP Outo3me Raffinate Disorder size (Yr) msesure(s) Findinp

1963 Smith et al. Cystic fibrosis 12.18 Tennessee Self.Concept Scale Group moan at 90th cantile on general maladjustment scale 1963 Harper Duchene)* muscular 4 4 17 Minnesota Multiphasic Personality 60% Depressive feelings dystrophy Inventory 47% Maladjustment Orthopedic impairment 48% Maladjustment 1964 Cowen et al. Cystic fibrosis 176 16.Adult Cornell Medical Index 19% Moderate-severe emotional disturbance (males) 43% Moderate-severe emotional disturbance (females) 1984 Team Visual impairment 23 x 11.5 Achenbach CBCL Total score = normal range Introversion score = Sample Norm (11 c 0.05) 1985 Richman e al. Cleft lip/palate 36 14-17 Behavior Problem Checklist 47% Poor adjustment (Quay-Peterson)

No data given. Note CBCLChild Behavior Checklist

205 Table 2. Case Series With Controls

Reference Disorders Cases C.:eitrols (yr) Control ealectir Outcome measuns(s) Findings

1974 McAnarney et al. Juvwill.chxnnk 42 42 6-17 Group match on age, sex, SES Cooperemith Solf-Estaem Inventory arthritis Ficsminfias 64% Owls vs Children's Manifest Anxiety Scale 40% controls (NS), emotional California Test of Personality health not exersllent) Teacher rating 33% Cases va 9% tents* (NS), kr« adjustment 1976 Kumar et al. Sickle cell disease 29 29 12-18 Group match on age, ethnicity General Anxiety Scale for Controls< 0.006) lime anxious Children (Samson) Piers-Harris Self-Concept Scala Controls (p < 0.001) higher self-ooncept California Test of Personality No difference in overell achustmant 1977 Steinhausen et al. Diabetes 56 61 8-18 Group match on age, sex, SES Child ren's Penonality No difference questionnaire Hamburger Neurotiziamus Extravassionashala Fur Kinder Und Jugendliche 1977 Gayston et al. Cystic fibrosis 33 31 5-13 Oldest sibling Piersaarrie Sell-Corcept Scale No difference Missouri Children's Picture (Matched analysis not done) Series Holtzman Inkblot Test 1978 Simonds and Cleft lip/palate 40 40 6-18 Group match on age, sex, SES tfeimburgor Clinical amassment No difference Behavior questionnaire (Parent) 1980 Gath et al. Dis_betes 76 70 5-16 Next ckild on class list of Teacher questionnaire, Rutter 52 20% Cams vs 13% controls, aum sex Behavior Scale (Teacher) behavioral deviancy at school (NS) 28% Cases, emotional diaturbance Clinical assessment No comparable figure given for controls (clinical) See notes at the end of the table. 206 Table 2. Case Series With Controls (continued)

Age Reference Disorders Cases Controls (yr) Control selection Outcome nreasursts) Findings

1980 From et al. Familial 43 23 2-19 Siblings Connor's Parent Questionnaire No difference overall hyperlipo- Clinical assessment (Matched analysis not done) proteinemia Male cams significantly more impulsivefityperectrve Female cases Ognificantlyhighsr on perfectionism 1980 Meijer Hemophilia 20 20 Not Frequency match on age, sex, Manifest Affect Rating Scale Cases significantly more hostile, stated SES Mother-Child Questionnaire defiant 1981 Steinhausen Cystic fibrosis as as 5-18 Frequency match on age, sex, Clinical assessment (Mother) 58% Cases vs 31% controls, and Schindler SES definite psychiatric problems (slight to marked) 1981 Dreier et aL Cystic fibrosis 108 122 3-13 Siblings of CF patients, Louisville Behavior Checklist Egrentigling 19% Cases vs Other respiratory unmatched healthy children School Behavior Checklist 5%4% minds, maladjustment illnesess Achenbach CBCL 21112111EXILISE No difference Came signi&antly higher CBCL Total

1982 Gordon et al. Constitutionvl 23 23 6-12 Matched (individual) on IQ Piers-Harris SW-Concept Scale Behavior Problem &ores short stature (WISC-R), SW, age, sax

1982 Iwwis and Khaw Cyatie fibrosis 57 27 7-12 Frequency match on age, sex, Quay-Petarson Behavior Cease significantly mom Asthma SES Problem Checklist maladjusted (frequency of Piers-Harris Self-Concept Scale behavior problems) No difference in ealf-oncept 1982 Steinhaumn and Inflammatory 17 17 7-18 Individually matched on age, Graham and Rutter Parent 60% GSM vs 18% controls, tai Kies bowel disease sax. SES Interview Psychiatric disorder Rutter Children's Behavior (CBQ) Cases significently higher CBQ Clinical assessment subscores for *optimal discolors 1982 Lavigne at aL Diabetes 41 43 6-16 Individually matched on age, Achenbach CBCL Male cases significantly more sex intenalidng and mternaliting IIYMPWIM Group mean 0.8-0.9 SD above controls Fetnalall no difference (Matched analysis not dons) flee note at the end of Oa table. 207 Table 2. Case Series With Controlcontinued)

Ago Reference Disorders Cases Controls (yr) Control selection Outcome measureit0 Findings

1983 Stelnhausen et al. Cystic fibrosis 72 36 4.14 Frequency math on age, sex, Adapted Gramm and Rutter Asthma No direct comparison between sibling rank, SES Parent Interview cams and controls Multiple regression used to examine predictors of outowne measures for CF, estluna, and 1984 Hoare control groups reperately Seizure disorder 113 113 5.14 Individual match by class, Rutter Children's Behavior Cb.roeicen1es,y 48% Cases vs RP) sax (Teacher) 13% controls, disturbed* Rutter Children's Behavior tbitigilgsgr 45% Cases vs (Parent) 10% controls, distburbee ginigisuosmc 17% Cares vs 3% controls, disturbed* lizajablear 17% Cams vs 7% controls, disturbed INS) 1985 (Matchird analysis not done) Thompson Diabetes 119 Not Not No matching Missouri Children's Behavior Mean subscale scores not Cystic fibrosis stated stated Healthy controls from Checklist different &cm controls, and pediatrician's offke, lees than psychiatric group perhistric controls from community guidance clinic 1985 Breslau Cystic fibrosis 304 360 3-18 Random sample from Cleveland Myelodysplesia Psychiatric Screening Inventory 27% Cases vs 11% mntrols, arefl (Maher) Multiple physical severe psychiatric impairment disorders Between diseare groups, no Cerebral palsy differences co parent conflict, regressive anxiety 8W:scales; overall, higher than controls (Analysis of covariance cusnifing il1C01216 and maternal 1985 Cowen et al. education) Cystic fibrosis 41 31 2.5 Healthy day care children Preschool Behavior gumtionnaire No significant difference (lAomatching) (modified Rutter CR) overall, but CF significantly higher on hcetile-aggreas subscale Note NS = not statistically significant, SES= socioeconomic status; CF *Statistically significant. cystic fibrosis, WISC-R = Wechsler Intelligence Sallee for Children (Revised), CBCLChild Behavior Checklist. Ob Table 3. Prevalence Surveys

Age Sample Reference Survey sit. Sampling procedure (Yr) size Outcome measure(s) Findings

1971 Plass and Roghmann Isle of Wight, Entire population surveyed 9-11 3,271 Psychiatric essesament (Rutter and 17% with CPD, 7% in United Kingdom Graham after screening by standard- others (RR 2.4) ized parent and teacher questionnaires)

1971 Plass and Roghmann Monroe County, Systematic sample ci house. 6-16 1,756 California Test of Personality 23%40% with CPD, 13%46% Rochester, N.Y. holds random sample of Coopersmith &dilate.= Inventory in others (RR 2.3) children from these households Children's Manifest Anxiety Scale Cowan Trachaes Behavitx Scale Kearsley Behavior Symptom Questionnaire

1981 'Nailer et al. Genesee County, Cluster sample dank.= ban 0.17 3,072 Parent questions abovt behavior Michigan random sample of households Four questions to parent from Rochester 3.5%.20.6%* with CPD, 2.8% in survey, about behavior, social, whole amp* behavior problem learnin& and school problems (RR 1.34.4) 3.8-36.3%° with CPD, 1.8% in whole sample, social problem (RR 2.1-19.6)

Not*. CPD v chronic physical disorder, RRrelative risk. *Rates vary with degree of functional impairment, type of C PD.

209 Table 4. Cohort Studies

Sample else Ars Reference Disorder Cohort Cohort Disorder(yr) Data Source Outcome messure(s) Findinp

1970 Num et el. Physical problems Retrospective stratified 3,611 851 0-18 School records Draft rejection on sample a males from lisissitsa mat Physical New York City public "mental" grounds disorders, 7.6%, no elementary schools problems, 0.7% (RR 10.9) 1971 Plass and Chronk physical National Survey (19461 (Univeziate analysis only) 4,649 537 0-15 Repeated interviewsBehavioral symptom Roghmann dinorders sample of United Kingdom Eitssninssisztinhatme and examinations questionnaire') (UK) birth cohort !imams) Physical (parents, teachers. disorders 25%, others 17% children) (RR 1.51 Teacher rated nervous/ aggressive: CP) 39%, others 31% (RR 1.3) Child retrort (neurotic). CPD 14%, others 11% (RR 1.3) 1978 Peckham And Asthma Netional Child Development (Univariate analysis only) 13,509 1,644 7 & 11 Parents, teachers Rutter Home Butler Study I 1958) sample of Significantly higher scores birth cohort Behavior Scale on parent scale in (Parent) children with asthma Bristol Social Dose.responee effect Adjustment Guide No difference on Bristol (Teacher) after controlling for sex, sccial class (hfultivariate analysis controlling for sex, 1981 Ahnsjo Diabetes 111,lent cases of cliabete: social class) 64 A) =lin* 4-17 Psychuitrists Clinical an..ontrols asseesment No difference detected at CAM/ Rorschach Test Individual matching of 30 time 1 or time 2 controls on sex, age, Diabetes showed increase in parent occupation, family aggression from 201 to c haracteristica 30%,controls decrease Time 1 measurement within from 16% to 7% (2 < 0.05) 5 mo of diagnosis; time 2, 3 years later

See note at the end of the table. 1 0 Table 4. Cohort Studien (continued)

Sample sire Age Outcome mammas) Findinp Reference Disorder Cohort Cohort Disorder(yr) Data Source

CPI Score average is 27% Fran 1% mad= seraph of 144 13-22 Interviewers, subjectsStructured 1984 Orr et al. Chronic physical eases, 15% former cases disorders Monne CouMy households Interview in 1968 California Two measurements Psychological at 8.year interval Inventory Hospital admission lir% Cases vs 6.8% non-cases 1984 Britten ot al. Epilepsy National Surly (1948) 5,362 46 0-26 Hospital records sample of UK birth and subjects or other treatment had outcome (RR 2.7) (Univariate analysis) cohort for psychiatric, emotional problems age 15 to 26 years

Interview Schedule Psychiatric disorder Eangdmoi: Prevalence 1985 Kovacs et al Diabetes Newly diagnosed 74 8-13 Retrospective ascerulnment for Children or at least four of prodiabetk of outoime before notable symptoms psychosocial difficulty, diagnosis or signs of 18% Ilagysrgerie: 14% distress psychiatric disorder nalfagramlimt at 24 wk. 36%; at 9 mo 93% recovered Prevalenot aladjustrnent Heller et al. Cleft hp/palate Subjects recruited from 140 4-13 Structured interwew Achenbach Child 1985 1, 24% at Heart disease clinic lists (parent, child) Behavior Checklist 33% at ti time 2 Hearing defects Tko missauremixts trt llear Children's Self- interval Report Psychiatric (Univariate analysis) Rating Scale

Note. CPDchromic physical diaorder, RR = relative risk. 211 204 PREVENTION OF MENTAL DISORDERS

The conclusion based on this extensive review ofresearch published over the past 20 years is that the postulated causal relationship doesexist. Although we concluded that the risk of this outcomewas not "exceedingly large, it is of sufficient magnitude to be of concern to clinicians. In broadterms, it may safoly be assumed that at least twice asmany children with these disorders have a higir probability of experiencing what must be regardedas a secondary handicap of potentially great importance in hisor her development" (Nolan and Pless 1986). Interventivn

Although until recently evidence supporting thecausal link between chronic disorder in childhood and psychiatric illness has beensomewhat equivocal, the clinical suspicion that suchan association exists has been present fora long time. Indeed, as is often the case, exhortationsfor greater responsiveness and sensitivity to the psychological needs of thesechildren had been issued by experienced clinicians well in advance ofany supportive scientific evidence. And, while better evidencewas accumulatinggenerally since the late 1950s case reports and accounts of studies of doubtful scientific meritcontinued to appear. The intuitive case for intervention isa case for therapeutic treatment and not for prevention. The distinction between thesetwo forms of intervention is important.Though results of either would undoubtedlyshed light on the underlying cenceptual modelor mechanisms whereby illness increases risk for maladjustment, the answers providedare likely to be very different, as are the clinical implications. In a broad sense much of the exi, ting researchdemonstrating associations between illness and psychiatric problemsimplicitly draws ona preventive model. As stated, the argumentgoes that if it can be shown that chronic illness is a risk factor, and perhaps thatsome illnesses or some attributes in combination with certain illnessespose still greater risks, these children could be targeted for some form of intervention.Precisely what drat intervention would be is difficult to say, because at thepresent time the evidence is incomplete. At best, with a few possible exceptions,the commonly held view supports the notion that any and all chronic ili:leases increase risk to some extent. Furthermore, as has been stated, the actual extent of this risk, in relativeterms (RR, relative risk), is rarely more than double that inthe Oneral population (Po, prevalence). The "population attributable riskproportion," that is, the proportion of psychiatric illness affecting childrenor adolescents that would be removed if all chronic physical disorderswere to vanish, is, in fact, only about 9 percent. (Assuming the prevalence of CPD (Po) is 10percent and the RR of psychiatric disorder is 2.0. Calculated from the formula:APt=[(RR-1)/(RR(1/Po)-11]. See Rothman, 1986, pp. 38-39.) CHRONIC CHILDHOOD ILLNESS 205

This percentage, from a public policy point of view, is relatively small. However, another way to appreciate the extent of the problem is through a parameter known as the etiologic fraction. This is the proportion of psychiatric disorder among children with CPD that is attributable to CPD itself. Under the same assumptions as above, this would be 50 percent. It is not without foundation, then, that strong arguments have been put forward to support various nonspecific, or global, preventive interventions, despite the caveats articulated by Rutter (1982, 1987). The evidence in favor of their effectiveness, however, is extremely tentative. The most tenable model is one built on the assumption that the manner in which illness acts tothe detriment of mental health is to increase stress, both for the child and family. Because, as we have stated, the root cause of O. a sttess cannot be modified in any direct or significant fashion, the most popular approach h s been to try to enhance coping, primarily through counseling parents and fostering social support networks. This somewhat generic tactic has been well described in a conceptual model put forward by Cassel (Kaplan and Cassel 1975): In the most general terms, the theory that has guided these studies (those reportc4 in the monograph referenced), has been that susceptibility to a wide variety of diseases and disorders (including somatic as well as emotional and behavioral disorders) is influenced by a combiaation of exposure to psychosocial stressful situations and the protection afforded against these situations by adequate social supports (p. 2). Most attempts to implement these ideas and other related strategies have come in either of two forms. The first is the development of what might be viewed as an "artificial network" usually consisting of parent groups or (e.g., for adolescents), patient groups, to provide mutual support and assistance. None of these, so far as we are aware, have been properly evaluated. It is likely that even if one or more were found to be efficacious, the extent to which the findings could be generalized would be severely limited because of self-selection. It must be assumed that those most likely to benefit from this type of process would choose it, whereas others, perhaps in equal or greater need, would not. The other alternative in this model is to provide an individual, professional or nonprofessional, whose main task is to offer nonspecific support, usually to the parents, but occasionally to the child. It is believed that through this type of support a general reduction in stress will follow and some degree of protection is offered. There is a small amount of evidence that this may actually be the case. Some such evidence is found in the study by Pless and Satterwhite (1975), in which nonprofessional counselors were used, and in the extensive work in which a pediatric ambulatory care team provided more comprehensive, home- based care for children with chronic illnesses than is customary (Stein and Jessop 1984a).A comprehensive summary of the results of various such interventions is shown in table 5.

2 1 3 Table 5. Intervention Studies

Assignment Reference Disorier (yr) Sample aim moosdure Outcome manur() Intervention Remit

1973 McCraw and Diabetes 7-16 33 Treatment Frequency matching Coopetandth fiell-Fateem Inventory 'Novi& 26 Control kir op, ma, race, 3-wk camp; no spacific Siwalik:ant improvement Chlklmn's Maniket Anxiety Scala description of actMties SES, &ration of in seltestmen ansesure diabetes- Controls for Innala campers only, chosen from variety improvemants in moms of physician for ell subjects regard- lam of trestmentkontrol statue 1975 Plass and Wide ranp 6-15 56 '11.eatmentRandom after California Tea of Personality Satterwhite of chronic 'Artily Causally hymn% 60% 71mitment subjects 42 Control stratification os Cooperemith Self-Istasan Inventaty Fhrtical 1-yr taminsd nonprolassictel 41% controls, improved Family Function Children's Manifest Anxiety Scale disorder' Wes mature woman acted as mythologic status family couneelkes 1984 Stein and Iseitop Wide reap 0-11 106 Treatment Random after Personal Adjustment and Rol. Pediatric Roane COM Prqgrein of chronic 104 Control etratifisd on Significsntly better Skills Scale (PARS 11) 'Nem cam by pediatrician, PhYlical Judpd Ability to impmeniect in adjust- Functional Status Ifeacunt nurse pectitioner, with dleordets Cops, and Overall mat at 6 mo in tawnier* access to social worker, Duiden huts' pony (g0.04), with psychiatrists. etc. trend in same direction Pretest, and posttest at 1 yr ( g0.06) imams et 6 ani 12 mo (g37.33) No cliffirence in functional statue nassettre at either time

214 207 CHRONIC CHILDHOODILLNESS

that have been The difficulty with themain therapeuticintervention studies have bet-,a developed usingmodels in which themediating reported is that they manifestations focused factor was, as described, someform of stress or other might include negativechanges ;frimarily on the child.The latter, for example, the preventive model,few interven- in self-concept.However, particularly in in such a way that theprincipal focus is indeedthe tions have been designed the Rochester child rather than a parent,most often themother. For example, in Satterwhite 1975) and in theMontreal Social Workerstudy study (Flees and kept, the proportion ofcontacts (Nolan et al. 1987), inwhich these figures were child were only 12 percentand 7 percent, respectively. devoted primarily to the the Thus, very few of theinterventions have beendirected exclusively at they appear to havebeen successful, it is notknown children. And, where self-concept, or by reducing whether they succeeded byimproving the child's child, or both.In the Montreal study,there were no stress in parent or of the detectable changes in anyof the subdomainsof perceived competence child. however, is that fewintervention programs havebeen The overriding reality, the properly designed andadequately evaluated.Perhaps paradoxically, have represent attempts toprovide rather globaland majority of those that to which a moredirect generally preventive formsof intervention. The extent be successful is not clear.It would be reasonable therapeutic approach would peychotherapy" would not be very to assume that theresults of "conventional chronic illnesses who haveemotional problems from different for children with It is, results that have been foundfor children who arephysically healthy. be more difficult to offer interesting to speculate,however, whether it would illnesses, or less difficult. effective psychotherapy forchildren with chronic generally be more success- On the one hand, therapyfor these children might ful because there is aspecific issue that can bereadily identified--adaptation other hand, at a purelyanecdotal level, the personal to the illness. On the psychiatrists are reluctant experience of one of the authorssuggests that child because they feel insecurein to become involved withthese childrenperhaps problems that coexist. Otherobservations support this the face of the medical research in this domain; view: the dearth of childpsychiatrists who have done research grants studyingthis the few who serve asprincipal investigators on issue, especially in relation toprevention of psychiatricdisorder (Rutter 1982); professional associations and, indeed, the compositionof scientific meetings of point is word-. makingbecause if it wen more widely of child psychiatrists. The in this high- acknowledged that child psychiatry,in fact, is underrepresented readily be taken to remedythe situation. Apart from risk arena, steps could methods to assess speculation, there is no directcontrolled evidence of propt.r the efficacy of psychiatrictherapy under thesecircumstances. this is the extensive 1.)3clyof literature arising from One possible exception to therapy model in the the work of Minuchin (1970),who has adopted a family with chronic illnesses.It is difficult, however, to refitment of many children furthermore, there evaluate these reports. Theycomprise mostly case histories; 2,r 208 PREVENTION OF MENTALDISORDERS

is virtually no way to be certain about theextent to which self-selection for whatever results accounts might be reported.Another is the work the most part, dealswith approaches of Reiss, which for to family therapy, butonly rarely in the context of chronic illnessin childhood (Reiss 1986; Reiss and Oliveri1980). In addition, many of the studies reviewedfailed to describe with of precision thelevel of expertise of any degree the outcomes the therapist. Thisundoubtedly influences observed. In general,it may be safely child and adolescent assumed thata qualified psychiatrist, clinical childpsychologist, or other health professionalwill achieve better mental results thana nonprofessional attempting to use thesame approaches. And, regardless of discipline,the more perienced the therapist,presumably, the better ex- realm of psychiatry the results. Evenwithin the alone, however, itis virtually impossible most published studies to discern from the therapist's levelof experience.It should also be noted that severalreports suggest that children tend to pediatricians and otherswho treat underdiagnose maladjustment,whatever the context et al. 1986). Thismay serve to underestimate (Goldman interventions because the effectivenessof psychiatric children identifiedby physicians forreferral to psychiatristsmay be selected ina "biased" manner, for those whoare the most seriously ill example, by choosing kind. or most resistant to therapyof a less skilled

In summary, theevidence clearlysupports the view that childhood isa risk factor for significant a chronic illness in evidence indicates maladjustment. Furthermore,some that many of theemotional disorders childhood or earlyadolescence are likely manifest during to persist into adulthoodor may arise in adulthood aftera lag period (Pless et al. difficult to be certain 1989). Beyond this,however, it is which subgroupsamong the chronically ill, if greatest risk. Oileimportant exception any, are at that involve the is the case of childrenwith disorders ceritral nervoussystem (CNS), to be rapid developmentof new therapeutic discussed later. Withthe tion, which techniques, suchas organ transplanta- are associated with theuse of immunosuppressant radiotherapy that drugs or may damage thenervous system, the number tions may wellincrease in the future. of such condi- Equally, many of thenew technologies may have an indirect effecton maladjustment because and psychological of the powerfulethical issues associatedwith theiruse. In addition, it isreasonable tosuppose that other discrete pain, conditionsthat result in elements suchas brief admissions significant disruptionsin routine,or frequent to hospital,may identify such subgroups. explained later, thereis little evidence However, as is risk. Hence that suchgroups are truly at increased recommendations forintervention, particularlyat the preventive level, must necessarilylook to community to individual clinical mental health strategiesas opposed approaches. It isobviously not reasonable every child witl 4chronic to suggest that illness receivepsychiatric intervention (Nolan et al. 1987).Instead, what is from the outset strongly recommendedis that the medical care provided for thesechildren, usually by to the probabilitythat an emotional subspecialists, be carefullyattuned problem may arise.When thisoccurs, 216 CHRONIC CHILDHOOD ILLNESS 209 prompt referral must beconsidered. The problem for childpsychiatrists is to decide whether such referrals wouldbe considered "appropriate," and,if so, what priority would be assigned tothem in situations in which psychiatric resources are limited. It must also be noted that othertherapists, especially psyzhologists, have achieved what appear to be impressiveresults.Focused brief intervention techniques, including educational strategies,behavior modification, and relaxation therapy, have been employed.However, many of the results include relatively few subjects and/or lack propercontrols. These cannot be viewed as examples of primary prevention; but to theextent that they succeed, however, they undoubtedly contributs, to secondaryprevention. Hence they deserve further consideration, particularly with aview to determining how effective (as opposed to efficacious) they may be, and towhat extent they represent generally applicable forms of intervention. Risk Indicators forPsychosocial Dysfunction

Understanding the distribution and determinantsof any disorder and, par- .icularly, planning therapeutic and preventiveinterventions require knowledge of the prevalence and potency of causativeagents and of the factors (effect modifiers) that change the relationsh'n betweenthe host and these agents (Cassel 1976). Rutter articulated sevem. keyproblems that must be considered when attempting to prevent psychosocialdisorder (Rutter 1982). First, planners must ensure that an efficaciousintervention reaches its target population. Second, they must understandthat short-term improvements may not necessarily leed to enduring Jrlong-term benefits.Third, they must recognize the need to start interventions early inchildhood, even though critical periods of development probably do not exist.This is because patterns of failure, once established, tend to persist.Fourth, planners must explore the cost-benefit aspect, which should take account of anydisadvantages, or side effects, of an intervention. Most fundamentally, however,Rutter pointed out that the first and crucial step is bridging the gap betweenthe identification of a damaging factor and knowing how to eliminate or reduceits effect. The following section reviews the evidencefor indicators of risk tor psychosocial dysfunction in chronic illness. Broadlyspeaking, candidates can be conceptualized as personal or environmental (table6).Personal factors include demographic and biologic characteristics, togetherwith individual susceptibility.Environmental factors can be thought of asbeing in the social or medical environment. 210 PREVENTION OF MENTAL DISORDERS

Table 8. Candidate RiskIndicators for Psychosocial Dysfunction in Childhood ChronicIllness

Factor Indicator

1.Demographic Factors Age, sex, socioeconomicstatus 2.Biologic Factors Type of disorder, braininvolvement, severity, visibility, predictability,age at onset, duration of illness 3.Individual Susceptibility Locus of control, personality 4.Social EnvironmentalFactors Family functioning, parents' psychologic state 5. Medical EnvironmentalFactorsContinuity of care,access to care

Demographic Factors The risk of emotional problems in the generalpopulation increases withage (Rutter et al. 1970), andthe same is true for children 1972). Boys are at greater with CPD (Pless et al. risk than girls (Rutter 1982),E;ut6pecific interactions with chronic illnesses havebecn recorded only ina few situations, notably girls with hearing impairmentand boys with congenital 1985). heart disease (Helleret al. Social class may also play an important role.Data from the 1970 birth cohort indicate that British families with disabledchildren were significantly more likely to be living in suboptimalhousing circumstances (Cooke 1984). and Lawton Others have shownthat there isa prominent association between economic stress and psychosocialoutcomes (Stein and Jessop 1986; Reissman 1980). However, Stein and with few possible exceptions,the illnesses themselves showno clear social class gradients;hence, true interaction modification is unlikely. or effect Biologic Factors

The notion that riskof psychosocial disorderis linked to specific entities has not been disease supported by the bulk ofevidence from either adult (Cassileth et al. 1984) studies or to childhood studies (Breslau1985; Heller et al. 1985; Ness and Pinkerton1975; Stein and Jessop suggestions that children 1982). There have beensome with sensory impairmentsare at more risk than children with other medicaldisorders, but the evidence al. 1975; Pleas 1984). is scanty (Haggertyet

2 1 8 211 ILLNESS CHRONICCHILDHOOD

CNS involvement(especially abundant evidence,however, that problems There is important predictorof emotional in CPD is an 1970; Steinhausenand mental retardation) Marshall 1985;Rutter et al. (Breslau 1985;Breslau and of the Isleof Wightprevalence Rutter andcoworkers' analysis CPD that Wefers 1976). first to focut, on adistinction between a apparently the of a substantialexcess survey was that does not.Their finding involves thebrain and one brain-affected group wasinter- emotional problemsin the recently, in theprevalence of "organic," effect onbehavior. More preted as apossible direct, or with spinabifida, cerebralpalsy, studied acohort of children Screening Inven- Breslau (1985) fibrosis. Usingthe Psychiatric multiple handicaps,and cystic brain-affectedchildren were found thatone-third of revealed that the tory ofLangner, she multivariate analysis Furthermore, her mental retar- severely disturbed. children were notaccounted for by high isolatiu scoresin these this distressinglyhigh risk for crude ordinalscale). Although for priority dation (using a children clearlyearmarks them maladjustment inbrain-affected direct or secondaryresult of brain this effect is a distinct from targetedservices, whether Furthermore, thereremains the far fromresolved. arising from dysfunction remains excess risk maybe artifactual, possibility that someof the apparent devised for thegeneral For example,questionnaires methodological factors. likely to contain manyitems that emotionaldisturbance are affirm, population to mess involvement willfeel forced to teachers of thosewith CNS is not what parents or of these"symptoms" in such cases despite the factthat the meaning the attributionthat would be intended, and maybe quitedifferent from CPD. was is healthy or onewith another inferred in the caseof a child who types of CPD,Mess risk over andabove other Notwithstanding thisexcess approach to thestudy and advocated theso-called"noncategorical" the and others of chronicillness, emphasizing pwchosocial consequences and Pinkerton management of experience forchildren (Pless commonality of thechronic illness enthusiasm has notbeen sustained Jessop 1982).Moreover, disease (Moos 1975; Stein and paychosomatic causalrelationship for for the notionof a specific 1979). disor- severity and therisk of psychosocial The relationshipbetween disease conflicting evidence 1984). Recentstudies provide der is also notresolved (Pleas lack of astandardized of the problemrelates to the about thisassociation. Much particularly onethat is validand health statusinstrument, et al. 1986; severity index or (Eisen et al.1979; Newacheck reliable acrossdisease categories (Pless andSatterwhite 1975), a In the MonroeCounty survey Williams 1979). of interferencewith dailyactivities severity based onparent reports (Pleas and Graham measure of used in the Isleof Wight survey was used,much like the one Satterwhite (1975)stated: 1970). Pless and between (was there] adirect relationship Only in one halfof the measures In most and the frequencyof maladjustment. the severity ofthe disability maladjustment being more (was] curvilinear, of the othersthe relationship nondisabled groups,and less in frequent in theseverely disabledand the of disability (p.88). those withintermediate levels !, 212 PREVENTION OF MENTALDISORDERS This same phenomenonwas noted in arthritis (McAnarney a sample of children et al. 1974).By way of with chronic important predictor contrast, severitywas the most of psychopathologyin cystic fibrosis patients withasthma,even after controlling patients, but notamong events (Steinhausen for familyfunctioning and and Schindler1981; Steinhausen life McNichol et al.(1973) studied et al. 1983).Similarly, reported that a cohort of children "behavioraldisturbances aged 7 to 14years and significant level occurredmore often and at only in thesmallgroup of children a statistically asthma." withsevere and continuing Conversely, Harper (1983) foundno evidence of degree ofimpairment in a linear relationshipbetween adolescents withmuscular dystrophy thopedic problemsand scoreson the Minnesota and otheror- tory. Gath etal. (1980) Multiphasic noted thatpoor diabetic control Personality Inven- psychiatric disorderand reading was directly relatedto to address the retardation, but,again, th is cohort issue of thetemporal study failed In anothercohort study, sequence and itsrelationship to Heller and herassociates (1985) causality. congenityl heartdisease, cleft assessed childrenwith occasions, 1 lip and palate,and hearing year apart. Theyfound that impairmenton two severity (on over this shortperiod of time, a simple 3-pointordinal scale) disease persistence andonset of was directly relatedto both the Checklist (CBCL) maladjustment,as assessed bythe Child (Achenbach andEdelbrock 1983). Behavior Data from thehomecare study of Stein cial disorder and Jessopindicated that was not related totraditional psychos°. hospitalized, beddays), but medical morbiditymeasures (days impairment was related toschool absence on their ownmeasure of functional and functional d). Cadmanet al. (1987) status (Stein reporteda similar relationship. and Jessop1984b, In the Montreal social workstudy, the status general healthsubscale of the measure (FSII) byStein andJeesop functional as "Eat well?" "Cut was used. Thisincluded such down on thingshe/she usually items children?" Theresponse scale asks does?" and'Play withother 2 weeks and for thefrequency of t whether thebehavior is behavior in thepast findings emerged attributed tothe illness. from examiningthe relationship Two interesting as rated by thisscale and risk between severityof disability was an overall of behaviordisorderon the CBCL. nonlinearrelationship, beet First, there Chiklren whohad no modeled with disability hadonly a 13-percent a quadraticterm. whereas childrenwithsevere disability risk ofmaladjustment, a 50-percent risk. (scores greaterthan 4 When thisrelationship on the FSII) had interaction between was examinedfurther, however, sex and functionalstatus became an relationshipwas almost completely apparent (figure1). The on boys alone. In explained bythe apparent fact, boys witha chronic illness effect of severity 4.8 timesmore likely than and any levelof disability boys witha chronic illness were maladjusted(95-percent and nodisability to be whereas for confidenceintervalon RR is 2.7, girls the relativerisk was only 8.8; p< .001), 1.2 (0.6, 2.4;p = .62).

11.1111111...111111111.11.11- CHRONIC CHILDHOOD ILLNESS 213

Behavior Disorder and Disability 70

40 ( Males

30 Females

None Mild Moderate Severe Disability Figure L Relationship between behavior disorder, gender, anddisability.

One possible clue to understanding the apparentlyconflicting evidence in relation to disease severity comes from a consideration of thedisorder's "visibility," In a prevalence survey of 2,454 randomly selected adult applicants for disability benefits, Zahn (1973) found that paysicalcharacteristics that clearly indicated the presence of sickness or disability were associated with better interpersonal relations (i.e., self-assessment of family, pe^r, andother relationships). Furthermore, in a study of young adult survivors of end-stage renal disease, Beck et al. (1986) showed that visibility (Cushingoid appearance, obesity, scars, orthopedic aids, short stature) was inversely correlated with identity stability (on a self-image scale) and social maturity (on the Vineland Social Maturity Scale). Finally, Jessop and Stein (1985) analyzed results from the 209 children who were participants in their randomizedtrial of home care (Stein and Jessop 1984a,c) and found that on a variety of measures of psychosocial function, it was children with "nnrmal" appearance who had the worst outcomes. The mothers 214 PREVENTION OF MENTAL DISORDERS

of these children were also less satisfied withcare and had more psychiatric symptoms, whereas the child's condition reportedly hadgreater impact on the family compared with those who appeared "abnormal." Pless (1984) has argued that the degree of visibility ofa diseaseand the likelihood that this forces the child to accept himselfas a "disabled person" may be the force behind this process. The ambiguity produced bya "marginal" state, or personal indecision about incapeJity,was first elaborated by Wright (1960). A related phenomenon is the impact that theunpredictability of a disease process might haveon psychosocial function. In the analysis of the home care intervention cited previously (Jessop and Stein 1985), itwas found that mothers of children with conditions where itwas necessary to watch for, or expect change, perceived a more negative impact of theillness on the family and had more psychiatric symptoms themselves. Unfortunately, appropriate empirical investigationsare lacking concerning the role that age at onset of disease plays in modifyingthe effect of psychosocial disorder. The same applies to duration of illness,particularly after controlling for the effects of actual age andage at onset. One exception is the results from the social worker intervention study (Nolanet al. 1987). These show thatage of onset and illness duration, after controllingfor age, are not predictive of behavior disorder as measured by the CBCL. Individual Susceptibility The chat Aerization of personalityor temperament characteristics that may modify the risk of emotional problems in theface of illness-imposed stress is rendered difficult for severalreasons. The most complex is the problem of having a pool of subjects on whom the pertinentmeasures were made before the onset of the physical disorder. In addition, the suitabilityof moet measures currently available deserves carefulconsideration.Studies such as that of Perrin and Shapiro (1985) identify differencesbetween healthy and diseased populations on such characteristicsas health locus of control. They interpret their observations--that beliefs in the controlof their health by chance and by powerful others is significantly stronger inchildren with chronic illnesses and their parents thanamong healthy children and their parentsas representing the effects of the chronic disorderor its management. Any conclusion about the direction of this putative cause-offectrelationship seems premature, forreasons already stated. Evidence from large-scaleprospective studies is needed before an unbiased assessment of the role of individual predispositioncan be made. Another problem relates to clearly definingand accurately measuring the pertinent predictors of behavior. Ina recent investigation that focused on social and personal competence in childrenand adolescents with orthopedicand seizure disorders, Perrin and her coworkers(1987) regressed healthresources inventory (HRI) scoreson scores from an adapted middle-childhoodtempera- ment questionnaire and other predictors.They found that up to 42 percent of the explained variancewas attributable to temperament scores, especially with 2r'') CHRONIC CHILDHOOD ILLNESS 215 teacher reporta on the HRI. However, the total R squaresfor their models were in the range of .44 to .76. This raises serious questionsabout whether the same construct was being measured in both the dependent andindependent variables. Social Environmental Factors The social environment, or "psychological situation" (Barker etal. 1978), is a concept that emerged from the work of socialpsychologists Kurt Lewin and Egon Brunswick (Moos 1979). It specifies the influences that parents,teachers, and other children have on a child's behavior. Accepting the importauceof the family microenvironment has led to many attempts to measure familyfunctioning, and this has resulted in the development of several self-report instruments.These have been used, almost without exception, in cross-sectionaldesigns (Friedrich 1979; Kovacs et al. 1985; Lewis and Khaw 1982; McNichol et al.1973; Plass et al. 1972; Pless and Satterwhite 1973; Sabbeth 1984;Steinhausen et al. 1983). Although there is no entirely satisfactory measure of family function(Walker and Crocker 1987), there does seem to be abundant evidence thatfamily dysfunction is associated with emotional problems in these children. But again, the direction of the relationship remains uncertain. Even studies with prospective data that antedate the onset of CPD have not provided all the necessary temporal evidence to dissect the risk that family dysfunction confers for emotional problems. This applies, too, to the association of maternal reychological distress with the presence of childhoodCPD. This is another well-documented association based mainly on cross-sectional studiee (Breslau et al. 1982; Burden 1980; Friedrich 1979; Gayton et al. 1977; Tewand Laurence 1973). British cohort data show, however, that familiee with disabled children generally fail to receive as much support from relativee, friends,and neighbors (Cooke and Lawton 1984). Rutter (1987) and Quinton and Rutter (1984) made much the same points with respect to the adverse effects of certain elements in the family environment such as parental symptoms, family discord, franily breakup, and impaired parenting (e.g., due to the presenceof schizophrenia or depression). Although the evidence in each case appears persuasive, the levels of risk are not always quantified, and it is not known how the effects are modified, if at all, in the presence of a chronic illness in the child. Medical Environmental Factors In chronic illness, the medical environment is conceived of as a subset of the wider social environment. The impact that components of the health care system have on child psychosocial function has been explored mainly for acute illness, especially in relation to preparedness for surgery (Skipper and Leonard 1968). The importance of patient and parent education with respect to psychosocial function in chronic illness has been emphasized (Van Vechten et al. 1977), but this has not yet been subjected to thorough empirical investigation (Nolan et al. 1986). Education and other techniques to reduce uncertainty in chronic illness, or to facilitate adaptation to it, are obvious and important candidates for future research. Satisfaction with care Law been studied in chronic illness, `23 216 PREVENTION OF MENTAL DISORDERS

and it appears that seeing the same doctor is associated with improved satisfaction with specialty care (Breslau 1982; Breslau and Mortimer 1981). Although little solid evidence as yet exists, it has been suggested (Plesset al. 1978) that when patterns of medical care prevail in which responsibilities for the care of these children are dividedamong specialists and primary care physicians, certain aspects of care will be duplicated while othersmay be neglected. In the latter, more potentially serious, situation, each party believes the other is "carrying the ball" when, in fact, neither is.To the extent that, certain elements of care, such as counseling,are truly effective in modifying the impact of the illness on the child or family, this could wellprove to be an important determinant of risk. Even more optimally organized care may result in sending mixedmessages of psychologic significance to the child. In the usualcase, children with chronic conditions are followed by hospital-based specialty clinics. The patternin these is to request regular followup visits regardless of the medical needs of the child. Whether these are prompted by simple curiosity, by research needs,or by financial incentives is unimportant. What is important,at least in theory, is that the child, while being reassured that all is well and encouragedto lead a normal life, is at the same time being constantly reminded ofhis or her abnormality or deviant status. When such routinesare accompanied by restrictions or precautions that are not essential to the maintenance ofmedical well-being, or by recommendations that special education be provided(which, despite the mainstreaming movement, often remains segregated),the underlying problem is accentuated. Conclusion

This review shows that chronic physical illnesses of childhood clearlyincrease the risk of mental illness. Much uncertainty still remainsabout the extent of risk and about the degree to which other features of thechild, the illness, the family, or, indeed, the medical care system influence this risk.In addition, it is far from ciear what strategies of preventionare the most promising or what the role of clinical psychiatrists might be in reaching this goal. Tothe extent to which the concept of community mental health is stillalive and well, there may well be a more promising role for interventions basedon some of the principles enunciated by Caplan and some of theprograms described by contributors to the volume edited by Cowen et al. (1967). Conversely, it is possiblethat the keys to any substantial improvementsare held almost exclusively by the pediatricians and, in particular, by pediatric subspecialists.If indeed, the solution requires cooperation between these often rivalfactionsfor example, in some forced marriage often referred toas "behavioral pediatrics"the prognosis in the present climate is far from promising. CHRONIC CHILDHOOD ILLNESS 217

References

Achenbach, T.M., and Edelbrock, C. Manual for the Child Behavior Checklist and Revised Child Behavior Profile. Vermont: Queen City Printers, 1983. Ahnsjo, S.; Humble, K.; Larsson, Y.; Stettergen-Carlsson, G.; and Sterky, G. Personality changes and social adjustment during the first three years of diabetes in children. Acta Paediatrica Scandinavica 70:321, 1981. Amir, S.; Galatzer, A.; Frish, M.; and Laron, Z. A sociological survey of 296 juvenile diabetics. Pediatric and Adolescent Endocrinology 3:82, 1977. Barker, RG., and Associates. Habitats, Environments and Human Behavior: Studies in Ecological Psychology and Eco-Behavioral Science. San Francisco: Jossey-Bass, 1978. Beck, A.L.; Nethercut, G.E.; Crittenden, M.R; and Hewins, J.Visibility of handicap, self-concept, and :wild maturity among young adult survivors of end-stage renal disease. Journal of Developmental and Behavioral Pediatrics 7:93-96, 1986. Boyle, Lit.; di Sant'Agnese, PA; Sack, S.; Millican, F.; and Kulcyzcki, L.L. Emotional adjustment of adolescents and young adults with cy3tic fibrosis. Journal of Pediatrics 88:318, 1976. Breslau, N. Continuity reexamined: Differential impact on satisfaction with medical care for disabled and normal children. Medical Care 20:347-360, 1982. Breslau, N. Psychiatric disorder in children with physical disabilities. Journal of the American Academy of Child Psychiatry 24:87-94, 1985. Breslau, N., and Marshall, I.A.Psychological disturbance in children with physical disabilities: Continuity and change in a 5-year follow-up. Journal of Abnormal Child Psychology 13:199-216, 1985. Breslau, N., and Mortimer, E.A. Seeing the same doctor: Determinants of satisfaction with specialty care for disabled children. Medical Care 19:741-757, 1981. Breslau, N.; Staruch, K.S.; and Mortimer, E.A.Psychological distress in mothers of disabled children.American Journal of Diseases of Children 136:682-686, 1982. Britten, N.; Wadsworth, M.E.J.; and Fenwick, P.B.C. Stigma in patients with early epilepsy: A national longitudinal study. Journal of Epidemiology and Community Health 38:291, 1984. 218 PREVENTION OF MENTAL DISORDERS

Burden, RL. Measuring the effects of stress on the mothers of handicapped infants: Must depression always follow? Child Care, Health and Devekpment 6:111-225, 1980. Bywater, E.M. Adolescents with cystic fibrosis: Psychosocial adjustment. Ar- chives of Disease in Chiklhood 56:538, 1981. Cadman, D.; Boyle, M.; Szatmari, P.; and Offord, D.R. Chronic illness, disability, and mental and social well-being Findings of the Ontario Child Health Study. Pediatrics 79:805-813, 1987. Cassel, J. A contribution of the environment to host resistance. American Journal of Epidemiology 104:107-223, 1976. Cassileth, B.R.; Lusk, E.J.; Strouse, T.B.; Miller, D.S.; Brown, L.L; Cross, P.A.; and Tenaglia, AN. Psychosocial status in chronic illness: A comparative analysis of six diagnostic groups. New England Journal of Medicine 311:506- 511, 1984. Cooke, K, and Lawton, D. Informal support for the careers of disabled children. Child Care, Health and Development 10:67-79, 1984. Cowen, L.; Cory, M.; Keenan, N.; Simmons, R.; Arendt, E.; and Levison, H. Family adaptation and psychosocial adjustment to cystic fibrosis in thepreschool child. Social Science and Medicine 20553, 1985. Cowen, L.; Corey, M.; Simmons, R.; Keenan, N.; Robertson, J.; and Levison, H. Growing older with cystic fibrosis:Psychologic adjustment of patients more than 16 years old. Psychosomatic Medkine 46:363, 1984. Cowen, E.L.; Gardner, E.k; and Zax, M., eds. Emergent Approaches to Mental Health Problems. New York: Meredith Publishing, 1967. Denson, P.M.; Ullman, D.B.; Jones, E.W.; and Vandow, J. Childhood characteristics as indicators of adult health status. Public Health Reports 85:981, 1970. Dorner, S. The relationship of physical handicap to stress in families withan adoleseent with spina bifida. Developmental Medicine and Child Neurology 17:765, 1975. Drotar, D.; Doershuk, C.F.; Stern, R.C.; Bont, T.F.; Boyer, W.; and Matthews, L. Psychosocial functioning of children with cystic fibrosis. Pediatrion 67:338, 1981. Eisen, M.; Ware, J.E. ; Donald, C.A.; and Brook, R.H. Measuring components of children's health status. Medical Care 17:902-921, 1979. Friedrich, W.N.Predictors of coping behavior of inothers of handicapped children. Journal of Consulting and Clinical Psychology 47:1140-1141, 1979. 2; CHRONIC CHILDHOOD ILLNESS 219

Froese, AP.; Rose, V.; and Allen, D.M. Emotional implications of primary familial hyperlipoproteinemia in childhood and adoleecence. Pediatrics 65:469, 1980. Gath, A.; Smith, M.A.; and Baum, J.D. Emotional, behavioural, and educational disorders in diabetic children. Archives of Disease in Children 55:371-375, 1980. Gayton, W.F.; Friedman, S.B.; Tavormina, J.F.; and Tucker, F. Children with cystic fibrosis: I. Psychological test findings of patients, siblings, and parents. Pediatrics 59:888-894, 1977. Goldman, R.; Voulalas, D.S.; and Stein, RE.K "Do Parents or Doctors Know if Children Have Behaviorvl Problems?" Paper presented at the meeting of the American Pediatric Association, Washington, D.C., May 1986. Gordon, M.; Crouthamel, C.; Post, E.M.; and Richman, R.A.Psychosocial aspects of constitutional short stature: Social competence, behavior problems, self-esteem, and family functioning. Journal of Pediatrics 101:477, 1982. Grey, M.J.; Genet, M.; and Tamborlane, W.V.Psychological adjustment of latency age diabetics: Determinants and relationship to control. Pediatrics 65:69, 1980. Haggerty, RJ.; Roghmann, KJ.; and Plass, 1.B., eds. Child Health and the Community. New York: Wiley, 1975. pp. 78-94. Harper, D. Personality correlates and degree of impairment in male adolescents with progressive and non-progressive physical disorders. Journal of Clinical Psychology 39:859-867, 1983. Heller, A; Rafman, S.; Zvagulis, I.; and Pless, I.B. Birth defects and psychosocial adjustment. American Journal of Diseases of Children 139:257-263, 1985. Heller, A.; Tidmarsh, W.; and Pless, I.B. The psychological functioning of young adults born with cleft lip or palate. Pediatrics 20:459, 1981. Hoare, P. The development of psychiatric disorder among school children with epilepsy. Developmental and Medical Child Neurology 26:3, 1984. Jessop, D.J., and Stein, R.E.K. Uncertainty and its relation to the psychological and social correlateb of chronic illness in children. Social Science and Medicine 20:993-999, 1986. Kaplan, B.H., and Cassel, J.C. eds. Family Health: An Epidemiological Ap- proach. Chapel Hill: University of North Carolina Press, 1975. pp. 1-5. Kovacs, M.; Feinberg, T.L.; Paulauskas, S.; Finkelstein, R.; Pollock, M.; and Crouse-Novack, M. Initial coping responses and psychosocial characteristics of children with insulin-dependent diabetes mellitus.Journal of Pediatrics 106:827-834, 1985. 20,-7 220 PREVENTION OF MENTAL DISORDERS

Kumar, S.; Powars, D.; Allen, J.; and Haywood, L.J. Anxiety,self-concept, and personal and social adjustments in children with sickle cell anemia. Journalof Pediatrics 88:859, 1976.

Lavigne, J.V.; Traisman, H.S.; Marr, T.J.; and Chasnoff, I.J. Parentalpercep- tions of the psychological adjustment of children with diabetesand their si- blinp. Diabetes Care 5:420, 1982. Lewis, B. L., and Khaw, K Family functioningos a mediating variable affecting psychosocial adjustment of children with cystic fibrosis. Journalof Pediatrics 101:636-640, 1982.

MacLean, G., and Becker, S. Studies of the psychological adjustmentof the hearing-impaired. Canadian Journal of Psychiatry 24:744, 1979. McAnarney, E.R; Pleas, I.B.; Satterwhite, B.; and Friedman,S.B. Psychological problems of children with chronic juvenile arthritis.Pediatrics 53:523-528, 1974.

McCraw, RK, and Travis, L.B. Psychological effects ofa special summer camp on juvenile diabetics. Diabetes 2:275, 1973. McNicol, K; Williams, H.E.; Allan, J.; and McAndrew, I.Spectrum of asthma in children. III. Psychological and social components. BritishMedical Journal 4:16-20, 1973.

Meijer, A. Psychiatric problems of heinophilic boys and theirfamilies. Interna, tional Journal of Psychiatric Medicine 10:163, 1980. Minuchin, S. The use of an ecological framework in thetreatment of a child. In: Anthony, E.J., and Koupernik, C., eds. The InternationalYearbook of Child Psychiatry: The Child in His Family. New York: Wiley,1970. Moos, RH. Social-ecological perspectiveson health.In: Stone, G.C.; Cohen, F.; Adler, N.E.; and Associates, eds. Health PsychologyAHandbook. San Francisco: Jossey-Bass, 1979.

Newacheck, P.W.; Halfon, N.; and Budetti, P.O. Prevalenceof activity-limiting chronic conditions among children basedon household interviews. Journal of Chronic Diseases 39:63-71, 1986,

Nolan, T.; Desmond, K; Herlich, R; and Hardy, S.Knowledge of cystic fibrosis in patients and their parents. Pediatrics 77:229-235,1986. Nolan, T., and Pless, I.B.Emotional correlates andconsequences of birth defects. Journal of Pediatrics 198:201-216,1986.

Nolan, T.; Zvagulis, I.; and Ness, I.B. Controlledtrial of social work in childhood chronic illnes8. Lancet 411-415, August 22, 1987. CHRONIC CHILDHOODILLNESS 221

O'Dougherty, M.; Wright, F.S.;Garmezy, N.; Lowenson, R.B.;and Torres, F. heart defects. Later competence andadaptation in infants who survive severe Child Development 64:1129, 1983, O'Malley, J.E.; Koodier, G.;Foster, D.; and Slavin, L.Psychiatric sequelae 49:608, 1979. surviving childhood cancer.American Journal of Orthopsychiatry Orr, D.P.; Weller, S.C.; Satterwhite,B.; and Pleas, LB. Psychosocialimplica- tions of chronic illness inadolescence. Journal of Pediatrics104:152, 1984. Peckham, C., and Butler, N. A nationalstudy of asthma in childhood.Journal of Epidemiology andCommunity Health 32:79, 1978. Perrin, E., and Shapiro, E.Health locus of control beliefs ofhealthy children, children with a chronicphysical illness, and theirmothers.Journal of Pediatrics 107:627-633, 1985. Perrin, E.C.; Ramsey, B.K.; andSandler, H.M. Competent kids;Children and adolescents with chronic illness.Child Care, Health and Development13:13-32, 1987. Piers, E., and Harris, D. ThePier8-Harri.9 Children's Self-ConceptScale. Nash- ville, Tenn.: Counsellor Recordingsand Tests, 1969. Pleas, I.B.Clinical assessment:Physical and psychological functioning, Pediatric Clinics of North Ameriaz31:33-45, 1984. Pleas, I.B.; Crippe, HA.; Davies,J.M.C.; and Wadsworth, M.E.J.Chronic Illness in Childhood and Psychologicaland Social Circumstances inAdolescence and Early Life, Dev. Med. ChildHealth (In press), 1989. Pless, I.B., and Graham, P. Epidemiologyof physical disorder. In: Rutter,M.; Tizard, J.; and Whitmore, K., eds.Education, Health and Behavior.London: Longman, 1970. Pless, I.B,, and Pinkerton, P. 1975.Chronic Childhood DisorderPromoting Patterns of Adjustment. London: HenryKimpton, 1975. Pless, I.B., and Roghmann, K.J.Chronic illness and its consequences:Obser- vations based on three epidemiologic surveys.Journal of Pediatrics 79:351, 1971. Pless, LB.; Roghmann, K.; and Haggerty,R.J. Chronic illness, family functioning, and psychological adjustment:A model for the allocation ofpreventive health services. International Journalof Epidemiology 1:271-277, 1972. Pless, LB., and Satterwhite, B.B. A measureof family functioning and its implications. Social Science and Medicine7:613-621, 1973. 2 222 PREVENTION OF MENTALDISORDERS

Plass, I.B., and Satterwhite, B. Chronic illness. In:Haggerty, R.J.; Roghmann, KJ.; and Pleas, I.B., eds.Child Health and the 1975. pp. 78-94. Community. New York:Wiley,

Plass, LB.; Satterwhite,B.; and Van Vechten,D. Division, duplication neglect: Patterns of and care for children with chronicdisorders. Child Care, Health and Development 4:9,1978. Quinton, D., and Rutter, M. Parents with childrenin care.I.Current circumstances and parenting.Journal of Child Psychology 229, 1984. and Psychiatry 25:211- Reiss, 1). Family process, chronic illness and leath.On the weakness ofstrong bonds. Archives ofGeneral Psychiatry 43:795-804,1986. Reiss, D., and Oliveri, M.E. Family pat-adigmsand family coping: A for linking the family's proposal intrinsic adaptive capacitiesto its responses tostress. Family Relations29:431-444, 1980. Richman, LC.; Holmes, C.S.; and Eliason, M.J.Adolescents with cleft lipand palate and behaviorrelated to personality 22:93, 1985. adjustment. Cleft PalateJournal Rothman, K.J. Modern Epidemiology. Boston:Little, Brown & Co.,1986. Rutter, M. Prevention of children's psychosocialdisorders: Myth and substance. Pediatrics 70:883-894,1982. Rutter, M. Psychosocial resilience and protectivemechanisms. AmericanJour- nal of Orthopsychiatry57:316-331, 1987. Rutter, M.; Tizard, J.; and Whitmore, K.Education, health andbehaviour: Psychological and medicalstudy of childhood 1970. development. New YorkWiley, Sabbeth, B. Understanding the impact of chronicchildhood illnesson families. Pediatric Clinics of NorthAmerica 31:47-57, 1984. Simonds, J.F. Psychiatric status of diabetic childrenin good andpoor control. Pediatric and AdolescentEndocrinology 3:8, 1977. Simonds, J.F., andHeimburger, R.E. Psychiatric lip-palate matched with evaluation of youth withcleft a control group. Cleft PalateJournal 15:193, 1978. Skipper, J.K., andLeonard, R.C. Children,stress and hospitalization: experiment. Journal of A field Health and SocialBehcwior 9:275-287, 1968. Smith, MA; Treadwell,M.; aid O'Grady, L. change, and clinical Psychosocial functioning,life status in adolescentswith cystic fibrosis. Adolescent Health Care4:230, 1983. Journal of

9 ') CHRONIC CHILDHOODILLNESS 223

Stein, R.E.K., and Jessop,D.J. A noncategorical approach tochronic childhood illness. Public Health Reports97:354-362, 1982. Stein, R.E.K, and Jeesop,D.J. Does pediatric home caremake a differencr) for children with chronic illness?Findings from the pediatricambulatory care treatment study. Pediatrics73:845-853, 1984a. Stein, R.E.K, and Jessop, D.J.Assessing the Functional Statusof Children. Monitoring Child Health in theUnited States: Selected Issuesand Policies. Cambridge, Mass.: HarvardUniversity School of HealthPolicy Research and Education, 1984b. Stein, R.E.K, and Jessop, D.J.Evaluation of a Home Care Unit as anAm- bulatory I.C.U.National Technical InformationService publication. Springfield, Va.: U.S. Dept. ofCommerce, 1984c. Stein, R.E.K., and Jessop, D.J.Relationship between health statusand psychological adjustment amongchildren with chronic conditions.Pediatrics 73:169-174, 1984d. Stein, R.E.K., and Jessop, D.J."Long-term Mental Health Effects of aPediatric Home Cam Program." Paperpresented at the annual meeting ofthe Am- bulatory Pediatric Association,Washington, D.C., 1986. Stein, R.E.K, and Reissman,C.K. The developuent of animpact-on-family scale: Preliminary findings.Medical Care 18:465-472, 1980. Steinhausen, H.C.; Borner, S.; andKoepp, P.The personality of juvenile diabetics. Pediatric and AdolescentEndocrinology 3:1, 1977. Steinhausen, H.C., and Kies, H.Comparative studies of ulcerativecolitis a..d Crohn's disease in children andadolescents. Journal of Child Psychologyand Psychiatry 23:33, 1982. Steinhausen, H.C., and Schindler, H.P.Psychosocial adaptation in childrenand adolescents with cystic fibrosis.Journal of Developmental andBehavioral Pediatrics 2:74-77, 1981. Steinhausen, H.C.,; Schindler, H.P.;and Stephan, H.Correlates of psychopathology in sick children: An empiricalmodel. Journal of the Atnerican Academy of Child Psychiatry 22:559-564,1983. Steinhausen, H.C., and Wefers, D.Intelligence structure and personalityin various types of physicalhandicaps in childhood andadolescence. Neuropadiatrie 7:313-321, 1976. Sullivan, B.J. Adjustment in diabeticadolescent girls.I. Development of the diabetic adjustment scale. PsychosomaticMedicine 41:119, 1979.

2 3 1 224 PREVENT/ON OFMENTAL DISORDERS

Tavormina, J.B.; Kastner,L.S.; Slater, P.M.; children: A psychologically and Watt, S.L. Chronicallyill and emotionallydeviant population? Journa/of Abnormal Child Psychology4:99, 1976. Teare, J.F. Behavioral adjustment of childrenattending a residential the blind.Journal of Developmental school for and Behavioral Pediatrics5:287, 1984. Tew, B., and Laurence,KM. Mothers,brothers, and sisters spina bifida. of patients with Developmental and MedicalChild Neurology 1973. 15(Suppl. 29):69-76, Thompson, R.J. Delineation of child behaviorproblems: A basis for and intervention.Journal of Developmental assessment 1985. and Behavioral Pediatrics6:37,

Van Vechten, D.;Satterwhite, B.; and Pless,I.B. Health education for parents of physically literature handicapped caildrenAnupdate.American Journal of Diseases of Children131:311-315, 1977. Walker, D.K., andCrocker, R.W. Measuring family systemsoutcomes.In: Weiss, P., and Jacobq,F., eds.Evaluating Family Aldine, 1987. Programs.New York: Walker, D.K.; Gortmaker, S.L., and Weitzman,M. Psychosocial Problems Chronic Illness and Among Children inGenesee County. School of Public Health,1981. Boston: Harvard White, K; Kolman, M.L.; Wexler, P.;Polin, G.; and Winter,R.J. diabetes and unstablefamilies: A psychosocial Unstable with recurrent evaluation of diabeticchildren kstoacidosis.Pediatrics73:749-755, 1984. Williams, R.G.A.Theories and measurement in disability.Journal of Epidemiology and CommunityHealth33:32-47, 1979. Wright, B.A.Physical Disability--A per and Row, 1960. Psychological Approach.New York: Har-

Zahn, M.A. Incapacity,impotence and invisible impairment: The effectson interpersonal relations.Journal of Health and 1973. Social Behavior14:115-123, CHAPTER 7

Prevention ofPsychiatric Morbidity in Children AfterDisaster

Robert S. Pynoos, M.D., M.P.H. Director Program in Przvention Interventionin Trauma, Violence, ahd Sudden Bereavement University of California, Los Angeles Neuropsychiatric Institute and Hospital Center for the Health Sciences Los Angeles

Kathi Nader, M.S.W. Director of Evaluation Program in Prevention Intervention inTrauma, Violence, and Sudden Bereavement University of California, Los Angeles Neuropsychiatric Institute and Hospital Center for the Health Sciences Los Angeles Introduction

Disasters are ubiquitous, affecting all 50States, from major urban areas to remote rural communities. Sincethe inception of the 1974 Disaster ReliefAct (Public Law 93-288), 798 majorFederal disasters and 3,092 Federal emergencies, which are slightly less severe,have been declared in the United States(D. Dannels, Federal Emergency ManagementAgency, personal communication, Sept. 21, 1987). The Federal Governmenthas denied 38 percent of the additional requests from States for disasterassistance since 1974 (Rubinet al. 1986). The cost in human life is significant. Morethan 8,000 deaths per year in the United States are attributed to natural andmanmade disasters; estimates suggest there are 50 injuries for everydeath (Logue et al. 1981b). Even when loss of life is minimal, destruction to homesand businesses often causes serious economic losses.

This paper was prepared for the ProjectPrevention initiative of the American Academy of Child and Adolescent Psychiatry. 233 226 PREVENTION OF MENTALDISORDERS

No epidemiological datawdst on deathor morbidity of children exposed major disasters. A disaster, to however, is one of the fewlife stresses for which early access to affectedchildren and their familiesis authorized health measure. Section as a public 413 of the DisasterRelief Act mandatesthat the National Institute of MentalHealth provide mental and local agencies in the health assistance toStates disaster area "to helppreclude possible damaging physical or psychologicaleffects" (U.S. Government, provide an 1976). Thus, disastets uncommon opportunity for mentalhealth professionals preventive strategies that to employ will (a) decrease theincidence of psychiatricmorbidity and (b) foster theabilities of large numbers adversity. of children to adaptto Understanding the impact of disaster on childrenrequires multiple professional perspectives:medical, psychiatric, (Shore et al. 1986; Cohen sociological, ecological,and economic in press). Individualstudies of the effects ofdisaster on children are summarizedin the appendix. limitations in most of Garmezy (1986) foundmajor the studieson children and disasters.While a phenomenological constructof posttraumatic has slowly emerged stress disorder (PTSD)in children (Anthony 1986), only themost recent studies have the standardizedinstruments and applied used in adult studies systematic methodologiesthat have been examining disaster reactions,the role of process of coping and exposure, and the recovery. Furthermore, thesestudies indicate the levelof adult stress thatchildren are exposedto after a disaster, influence child morbidity. a factor that may This paper explores issues related to theeffects of disasteron children, including psychiatricmorbidity, mediating impact on child factors, child intrinsicfactors, the development, and methodsof prevention. PsychiatricMorbidity

Although populationsexposed to disaster and Adams 1984), undergo substantialstress (Adams controvemy persists aboutthe extent, nature, postdisaster psychiatric and causes of morbidity. Untilrecently, methodological of postdisasterstudies accounted for limitations much of thecontroversy (Logue et al. 1981a).Researchers examineddifferent mental health varied in their rateof onset and chronicity. phenomena, which Standard PTSD criteria(American Psychiatric Association1980), standardized tematic research psychiatric instruments,and sys- methodology havepermitted investigators longstanding issues. to clarifysome Primacy of Exposure Degree ofexposure is the key variable stress reactions. In the in determining initialposttraumatic Buffalo Creek Damflood study (Gleseret al. 1981), researchers usinga stressor scale found that was significantly associated subsequent psychiatricmorbidity with the degreeof exposure. Shoreet al. (1986) 234 MORBIDITY AFTER DISASTER 227 demonstrated that the onset of psychiatric disorders after the Mt. St. Helens volcanic eruption followed a dose-response exposure pattera. Considerations of exposure and personal impact have focused on proximity to the impact zone, life threat, physical injury, witnessing of injury or death, injury or death of a significant other, property damage, and financial loss. Onsite exposure is the factor most specifically correlated with the characteristic PTSD symptoms (Wilkinson 1985). Two years after the Beverly Hills Supper Club fire, Green et al. (1985a) found degree of exposure to be the principal factor explaining chronicity of psychiatric impairment. Some investigators have found that specific experiences, especially witnessing the grotesque or hearing cries of distress, are associated with psychiatric morbidity (Green et al. 1985b). Rescue workers, ,aspecially body handlers, not directly exposed to the actual disaster but exposed to the mutilatior and death of others have also exhibited increased morbidity (Jones 1985). Even without a clear model of PTSD in children, early studies suggested the primacy of exposure in children's disaster reactions (Bloch et al. 1956; Glaser et al. 1981; Burke et al. 1982; see appendix, pp. i and iii). Investigators found that immediate life threat, presence in the impact zone, severe injury, and death or injury of family members were factors significantly associated with postdisaster emotional disturbance. Children did not seem to be as affected by residential property damage as adults, although individual children reported distress at the loss of personal possessions. Pynoos et al. (1987a) syntematically confirmed that the degree of exposure to life threat in school-age children had a dose- response relationship to severity of PTSD that continued to be present at the 14-month followup. Perimeter of Danger In assessing the need for services, a major consideration is that the psychiatric impact of a disaster is correlated with ,he proximity to the disaster zone. Three systematic studies of adults found a perimeter of danger or impact beyond .which there was no measurable change in psychiatric morbidity (Doh- renwend et al. 1981; Green et al. 1983; Shore et al. 1986). The perimeter of impact may be difficult to determine in technological disasters, especially those involving "invisible contaminants" (Berezofsky 1987), because the danger zone and measurable life or health threat may not be well established, For children exposed to life threat, a perimeter was also demonstrated beyond which there was no appreciable increase in PTSD without the presence of mediating factors (Pynoos et al. 1987a). These findings, however, require replication before they can be generalized for children. Types of Psychiatric Morbidity Systematic studies after disasters have found an increase in some psychiatric disorders but not in others. Shore et al. (1986) demonstrated a dose-response onset of three disorders following exposure to the Mt. St. Helens volcanic eruption, cited in order of frequency: (a) PTSD, (b) single-episode depression,

1. 228 PREVENTION OF MENTAL DISORDERS and (c) generalized anxiety disorder. A significant rate of comorbidity existed; however, no change in the rate of onset of other disorderswas determined. Increased morbidity could not be attributed to mediating factors suchas age, education, income, employment, or state of physical health. The researchers concluded that a natural disaster can give rise toa variety of highly significant stress-response disorders in individuals. In a study of Cambodian adolescent refugees exposedto massive trauma, Kinzie et al. (1986; appendix, p. xvi) reported the onset of disorderssimilar to those found by Shore et al. Fifty percent of the childrenwere diagnosed with PTSD, 12 percent with mckjor depressive disorder, 37 percent withintermittent depressive disorder, and 18 percent with generalized anxiety disorder. Kinzie et al. also found a significant rate of comorbidity. Pynoos and Nader's (1987) 14-month followup study of children exposedto a sniper incident suggests a distribution of psychiatric disorders similarto that reported by Kinzie et al. In this school-age sample, separationanxiety disorders were more frequent than generalized anxiety disorders. Clinical studies of children exposed tosevere traumas have also documented a high frequency of PTSD symptoms. For example, Ter: (1979; appendix,p. xiii) reported that 5 to 15 months after being kidnapped and facinglife threat, an entire group of 25 children had developed moderate tosevere PTSD. More systematic data collection enabled investigators to correlatedifferent exposure factors with the relative frequency ofspecific forms of morbidity.The experience of life threat and witnessing injury and death is highlycorrelated with the onset of PTSD; loss of a significant other is correlated withthe onset of a single depressive episode or adjustment reaction;worry about or sudden separation from a significant other is correlated with persistent anxietyregarding the safety of significant others (Pynoos et al. 1987a; appendix,p. xvii). While these factors may sometimes operate independently, at other timesthere may be interplay among them. Grief, for example,may increase PTSD, while life threat may increase the risk of a depressive episode (Pynoos and Nader 1987). Multiple bhavioral measures are needed to providerA more adequate assessment of the impact of trauma on children and adolescenis. Adams and Adams (1984) concluded that postdigasterstress reactions are serious and relatively enduring. They demonstrated tLat adult stress reactions in the Mt. St. Helens community increased afterexposure to the disaster. Examination revealed a substantial increase in stress-induceddisorders, mental illness, alcohol abuse, family stress, aggression withrelated adjustment problems, and violence. The number of domestic violencecases reported to police increased 45.6 percent. Although Adams and Adams (1984) notedan increase in juvenile arrests in the acute postdisaster period, they did not obtaindata on overt, recordable childhood behavior comparable to their findings foradults. Some studies have found an increase in somatic complaints,especially in school-age children, reflected in greater demandson school and camp nursing services (Pynoos et al. 9 0) 229 MORBIDITY AFTERDISASTER

is no destructionof 1987a; Kliman 1976, pp.325-335).Even when there increase dramaticallyfor several weeks(Pynoos property, schoc 1ah ,mces may several years in the easeof and may remainslightly increased for et al. 1987a) (1979) found that 33percent of a group exposed children(McFarlane 1987). Terr the first showed deteriorationin schoolperformance during of children Global AssessmentScale, Kinzie et al. postdisaster year. Usingthe Childhood have a significant that severelytraumatized children (1986) demonstrated Well-controlled data onschool performance degree of functionalimpairment. are needed. Prevalence ofMorbidity variations in a poychiatric morbidityvaries greatly with The prevalence of of exposure, the community,the type and severity disaster's overall impact on St. Helens community of the populationaffected. In the Mt. and the percentage appeared in 20 percentof the women after the eruption,psychiatric disorders appeared in up to 75 the men (Shore etal. 1986). Disorders and 11 percent of Creek Dam disaster(Gleser et al. 1981). percent of thoseexposed to the Buffalo prevalence data forchildren have beencollected after mAjor No comparable exposed children havebeen reported disasters. After traumas,however, severely posttraumatic stress symptoms(Pynoos et al. to exhibit highprevalence rates of significantly higher rateof PTSD in childrenthan in 1987a). After finding a 918-924) proposed adults 1 year following severeburns, Andreasen (1985, pp. than adults. that children were morevulnerable to PTSD Course of Morbidity delayed onset of more systematicstudies, the concept of a With the completion have not found sufficientevidence disorder is losing favorbecause investigators indicates that early responsespredict later for it. Instead, mostof the evidence the adult literature,Figley (1986, pp.xvii-xxix) symptomatology. Summarizing to the traumatic concluded that "the disorder wasdetectable soon after exposure and delayed reactions arerare." Shore et al.(1986) event or catastrophe that the rate of onset washighest verified this statementwhen they found dropped sharply insubsequent years. Theyalso found during the first year and tended to persist butdepression that 3 years after thediaaster, PTSD symptoms had abated. Like theadult studies, more recentstudies and anxiety disorders found no apparent delayin the onset of school-age andadolescent children have and that an early onsetis strongly predictiveof later course of PTSD symptoms Pynoos and Nader 1987;Pynoos (Terr 1979; McFarlane1987, appendix, p. vi; preschool children is difficultbecause of the et al. 1987a).Judging the onset in self-reports and limits inherent in theavailable instrumentsand children's cognitive maturity. 230 PREVENTION OFMENTAL DISORDERS

MediatingFactors SituationalVariables Appraisal of Threat

Because of the uniquedangers they face, may experience additional certLin segmentsof the population stress aftera disaster.For example, women showeda significant level of pregnant anxiety after theThree Mile Island reactor accident becausethey perceived nuclear 1980). danger to theirunborn children(Bromet Very young children may gainpartial protectionfrom the traumatic because they donot understand the impact threat, however, extent of the danger.Their appraisal depends partiallyon the accompanying of actions and attitude,especially in situations adult's or sibling's warnings (Freud of potential danger,such as air raid and Burlingham1943; Mercier and many horrifying and Despert 1943).Still, in catastrophic situations,adults cannot appear unPlarmed. In fact,children be expected to put in grAter can be confused, disturbed,and potentially jeopardy by adultswho minimize Nader 1986). the obvious threat(Pynoos and Human Accountability The peychological aftereffects ofmanmade disasters than those ofnatural disasters. may be more chronic traumatic stress Several investicatomhave noted that reactions aremore persistent after post- beings are perceived an event for whichhuman to be responsible(Frederick 1980). PTSD in childrenhave been reported The highestrates of et al. 1986; Pynoos after acts of violence(Torr 1979; Kinzie et al. 1987a).Grinker and Spiegel phenomenon in adultspartly to the (1943) attributedthis debilitating effectof prolonged revenge fantasies,an effect also observed or unexpressed in children byPynoos and Pth(1986). Separation froma Significant Other Worry about thesafety of a significant additional other duringa disaster may be source of extremestress for children an have been notedto exhibit intense and their families.Mothers their children, concern with the whereabouts which is thenassociated and safety of (McFarlane 1987). .ch continuedanxiety afterwards Adolescents worriedabout a greater post, younger sibling havereported ter distress andmore somatic complaints (Dohrenwena J. 1981). For all than theirpeers family members,worry about a sipificant other maypers.st, leadingto chronic whereabouts preoccupation aboutthe person's or safety, emotionaldetachment, and functioning (Pynoosand Nader 1988b). impairment of daily Several investiptorshave found that separation fromparents or siblings immediately afterthe disastercan further exacerbate the stressreaction, 2 3 S MORBIDITY AFTER DISASTER 231 especially for the preschool and school-agechild (McFarlane and Raphael 1984; Pynoos and Nader 1988b). For example,the duration of symptoms increased measurably in children who were sent away tostay with relatives after the Australia Ash Wednesday fire (McFarlane1987). Guilt Guilt, an associated feature of PTSD, is not wellexamined in children because of major developmental difficulties inassessing guilt in children of varying ages. However, preliminary evidence from a study ofviolent events indicates that when children report guilt, the severity of theirposttraumatic stress reactions increase independent of exposure (Pynoos etal. 1987a). School-age children reported "feeling bad" at being unable to provideaid, being safe when others were harmed, or believingtheir actions endangered others (Pynoos andNader 1938b). Particular experiencesfor example,hearing a wounded person's cry for help or watching someone bleed todeathmay create an intense level of empathic arousal in children that remainsundiminished if no effective intervention occurs (Hoffman 1979). Whetherguilt is an at-risk indicator in the triage of children following catastrophic or violenteventg is an area for future investigation. Multiple Adversities Multiple adversities that reeult from an event such as adisaeter may have more than an additiveeffect in increasing psychiatric morbidity (Rutter1985). One disaster alone can lead to the experience of severelife threat, death of a significant other, loss of residence and relocation,involuntary unemployment of a parent, and change in the family'sfinancial state.This effect may be responsible for the high rate of comorbidity inseverely affected children and adults after major disasters and massive traumas. Intervening Variables Recovery of the Community Disasters disrupt community cohesion and function(Erickson 1976). Some evidence suggests that the inefficiency OL postdisasterorganizational efforts directly and independently influences the persistence ofpsychological problems (Quarantelli 1985, pp. 173-215). On the other hand, theinfluence of the social group can be positive andreparative (Quarentelli 1985). Fear of recurrence affects everyone in the communityregardless of their degree of exposure to the event (Pynoos et al. 1987a).This widespread fear is sometimes fueled by myth, rumor, lack of information, andmisinterpretation. Ethological studies have demonstrated how the transmissionof alarm signals can be contagious in groups(Anthony 1986). Washburn and Hamburg (1965, pp. 1-13) demonstrated how asingle violent incident can influence the degree of fearful behavior in a primate group for manymonths. 2 232 PREVENTION OF MENTAL DISORDERS

Cultural Factors

Cultural differencesmay influence postdisaster behavior. Kinzie et al.(1986) found no increase in alcoholor other drug use or delinquencyamong adokscent Cambodians, even though they hadendured massive trauma andfamily loss. This finding contrasts with the increaseddelinquency and alcohol and other drug use generally found among bereaved adolescentgroups (Krupnick 1984, pp. 99-141). No differenceswere found between black and Hispanicschool-age children exposed to a sniper attackin a Los Angeles neighborhood(Pynoos et al. 1987a). Additional studiesare needed to assess the relevance of culturalfactors among different subpopulations experiencing differenttypes of disasters. Family

Parental and family functioningis a major mediating factor inchildren's psychiatric morbidity (Rutter 1985).Parental distress, parental disagreement about appropriate action during thediaaster, and change in parentingfunction after the disaster influence children'sreactions and recovery (Blochet al. 1956; Handford et al. 1986). The persistenceof symptoms in children has beenfound to be significantly associated withfour parental responses: parents'excessive dependence on children forsupport (Silber et al. 1958),overprotectiveness (McFarlane 1987), a prohibitive attitudetoward temporary regressivebehavior or toward open expression and communicationabout the euperience (Blochet al. 1956), and preexisting parentalpsychopathology (Bloch et al. 1956). Family members often experiencesimilar exposure toa disaster and share the loss of a family member or property damage. Gleser et al. (1981)found high symptom correlation betweenparents and children; McFarlaneet al. (1987), between mothers and theirschool-age children. When children'sexposure exceeds that of other familymembers, their symptomaticbehavior may disrupt normal family functioning (Terr1979). Every study that has interviewedchildren regarding their PTSDand grief reactions has found significantdiscordance between parent andchild reports (Bloch et al. 1956; McFarlane et al. 1987; Weissman 1987). Childreninitially experience core symptoms quiteprivately, and parents andteachers may not notice behavioral changes.When their distress is not fullyappreciated, children may not receive adequate emotionalsupport.They may becomemore withdrawn or isolated (Kinzieet al 1986) or may face disturbing demands to act unaffected. parental Kinzie et al. (1986) documentedthe powerful influence that ships have in constructively family relation- mediating the effects of massivepsychic trauma. Cambodian adolescent refugeeswho lived with family members after relocating had or relatives a significantly lower rate of psychiatricmorbidity than those who lived in adoptivesettings.

2 u 233 MORBIDITY AFTERDISASTER

Family Bereavement in disaster mortalityin the United States,dis- Despite a decreasing trend evidence asters still account forsignificant loss of life (Logue etal. 1981a). Some unexpected death of a parent orsibling is associated with indicates that sudden, morbidity in both risk of pathological or persistentgrief and psychiatric a higher 1984). As Rutter (1985) children (Pynoos and Nader1987) and adults (Lundlin number of changes in thechild's life, which can result noted, lose r. sts in motion a thus increase the risk of in chronically unsatisfactoryfamily circumstances and psychiatric morbidity. Scheel-age children appear tobe confused, frightened,and disturbed by their (Pynoos et al. 1987b). Becausethey may not be ade- normal grief reactions often overlooked by other quately understood,ehildren's grief reactions are preoccupied with their ownmourning (Bowlby 1980). family members, who are others and seek While adults commonly sharetheir grief experiences with reassurance, manychildren report that they do notreceive sufficient emotional support from family, teachers,and friends (Pynoos andNader in press). Influence on Peer Relationships Temporary dislocation and permanentrelocation of residence caninterrupt social support, eopecially for peer friendships,which are an important source of adolescents. In addition,posttraumatic irritability,inhibition, or aggression Posttraumatic stress reactionsin can strainsibling and peer relationships. children, especially those whohave not reachedadolescence, may include a enjoyment of normal activities, atendency to stay reduction of interest in and from inside more or nearer to protectiveadults, and feelings of estrangement isolation and, others (Pynoos and Nader 1988a).These symptoms can result in consequently, in disruption ofthe social, cognitive, andemotional developmental tasks accomplishedthrough play and interpersonalinteraction. Influence of the SchoolMilieu Garmezy (1986) commented onthe importance of theteacher's role as an stressed children, Teachersthemselves may be external support figure for distress as they do to severely affected, andchildren may respond to their parental distress. For example,children were shown to reflecttheir teachers' emotional responses to newsof President Kennedy'sassassination (Kliman 1968). Because children spend muchof their week in school,they may be just as likely to experience a disasterthere as at home. Howthe school reacts in the immediate situation and in theaftermath and reccvery periods maysubstan- disaster in tially affect children's recovery.The anxiety precipitated by a students and teachers can resultin general changes in ,aissroombehavior and milieu of disruption of the educational process.In addition, the postdisaster schools may vary in toleratingpostdisaster reactions depending onthe attitude of the principal and otheradministrators, 2 4 1 284 PREVENTION OFMENTAL DISORDERS

Child IntrinsicFactors Some childrenseem relatively resilient to a stressor, whileothers respond more significantly thantheir degree ofexposure explains. psychopathology, especially Preexistii.b anxiety and depressivedisorders, and or trauma have beenassociated with previous loss more severe and prolongedpostdisaster symptoms (Lacey 1972;Pynoos and Nader1987). The disaster child of a previoustraumatic event and may remind the Nader in press). renew his or her reactions(Pynoos and Predisposition toArousal Behavior After a traumaticevent, children frequently exhibitsleep disturbances, hypervigilance, andexaggerated startle Newman 1976, appendix, reactions (Pynoos andNader 1988a; p. ii; Burke et al. 1982;Davis 1983). be more likelythan adults to Children may stage 4 sleep phenomena experience neurophysiologicalchanges, suchas (Fisher et al. 1973).Some evidence persistence of increasedstates of arousal indicates that helps to reinforceother PTSD symptoms (Krameret al. 1984,pp. 81-95; Kolb 1987); disturbance has been for example, sleep correlated withdifficulties in attention performance (Pynoos and academic et al. 1987a). Children'spropensity to arousal may vary according togenetic, constitutional, behavior and environmentalfactors. Coping Whereas effective coping reducesdistress, maladaptive such as druguse, may exacerbate copingresponses, (Silver and Wortman distress or becomeproblems themselves 1980, pp. 279-341).There is no acceptable childhood coping(Garmezy 1986). taxonomy of Influenced by theirphase of development and prior experience,childrenvary widely in their event and their attempts to interpretthe symptoms, to regulatetheir emotions, and ing, information,and assistance. to search formean- Like adults, children prominently manifestavoidant behaviors associated with specifictraumatic remindeis. and anxiety the auditory and Some appearto overgeneralize visual aspects ofa traumatic reminder, specific stimuli (Bloch especially incident- et al. 1956); othersdisplay accurate tion.The way in whichchildren cognitive discrimina- renewed anxiety process these remindersand manage the may significantly affecttheirrecovery. Childrenvary in their acceptance of their postdisasterreactions. They interpret theirreactions asan indication that may and may feel that something iswrong with them other childrenare not similarly affected. caregivers, they Along with their may unrealisticallyexpect their These expectationscan intensify distress recovery time to be shorter. (Silver and Wortman and prevent seekingneeded support 1980; Kaltreideret al. 1979).

242 236 MORBIDITY AFTERDISASTER

factors influencing more con- Further studies are necessaryto investigate the child behaviors, such asincreased academic motivation, structive postdisaster (Pynoos et a). 1987a). increased courage, or moreempathic responses to others varying influence on aspectsof child These behaviors, inturn, may have development. Age reactions among age groupsafter Several studies havereported differences in Carey-Trefzer (1949) observedthat younger children were catastrophic events. These children had reflect adults' reactionsto war conditions. more likely to endangered. The older thechild, neurotic reactions onlyif they were personally destruction aroused anxiety.Similarly, Gleser et al.(1981) the more the sight of flood exhibited more severe found that school-agechildren exposed to a Psychiatric EvaluationForm psychiatric impairmentthan preschool children: and depression scoresincreased with age;belligerence (PEF) overall severity 16 teenagers, and anxiety washigher for the oldest group, ages was higher for accident, Handford et al.(1986) found to 20. After the ThreeMile Island nuclear than 8 did not appear torecognize the potentialdanger that childm younger combined fantasy and reality, resulting from the accident.Children older than 8 sometimes embellishingthe danger. school-age children after Findings are dissimilarfor studies of elementary winter storm, Burke etal. (1982) compared 436-year- disasters. After a severe Questionnaire 7-year-olds,Connor's Parent and Teacher old children to 21 In contrast, there were antisocial scale scores werehigher for the 6-year-olds. found in children ages 5 to12 after a sniper attack(Pynoos et al. no differences Cambodian 1987a). Kinzie et al.(1986) found norelationship between the experience, age, or sex,and the presence of adiagnosis in adolescence as measured by the ChildhoodGlobal Assessment Scale. Impact on ChildDevelopment children's levels Although there has been somediscussion of the influence of initial reactions andassimilation of traumatic events(Eth of maturity on their available and Pynoos 1985, pp.36-52; Cohen 1986), fewsystematic data are disaster or other traumaticexperiences on child concerning the effects of differential outcome of sub- development. Studies areneeded to examine the stress at different ages. stantial numbers ofchildren who have experienced a evidence for an interactionbetween trauma and Although there is no conclusive influence a number developmental stage, posttraumaticstress phenomena may affect the developmental process,including cognitive of characteristics that outlook, and impulse functioning, initiative,personality style, self-esteem, reexperiencing phenomer a canaffect cognitive functioning control. Intrusive, symbolic traumatic by altering attention eithertoward or away from concrete or reminders. Without resultingin a phobic disorder,traumatic avoidant behavior

3 236 PREVENTION OF MENTALDISORDERS

can lead to inhibitionsor altered interests. Children's imaginativeplay can become constrictedand leo enjoyablewith the repetition themes in play (Terr1979). of disaster-related Several researchershave reported very young children (Terr prominent personalitychanges even in 1979; Gislason andCa111982, appendix,p. viii). These changes range fromreduced impulsecontrol to increased attraction to danger inhibition, from to a debilitatingsense of fear, from emotional to exhibitionism.Researchers have described withdrawal pany the onset of adult PTSD changes in self-imagethat accom- (Kaltreider et al. 1979).How children'ssense of self-efficacy, self-confidence,or self-esteem is influenced Without adequate remains unknown. predisaster data, it isdifficult to and whether thereis an actual assess the degree of change discontinuity in personalitydevelopment or merely an exagge-stionof preexistingtraits. Silver and Wortman(1980) concluded unnoticed,one of the most devastatingthat, although thechange maygo effects of traumaon adults is the tendency for themto permanentlychange their views trauma studies have of the world.Childhood consistently founda marked change in the future,including a sense of orientation toward altered attitudes foreshortened future,negative expectations,and toward marriage,having children, Pynoos and Eth and career (Terr1979; 1984). Childrenmay anticipate that the will affect theiradult behavior. experitmce ofa trauma A study ofconcentrationcamp survivors demonstrated theinfluence of t"Alma on later parental pp. 295-313). behavior (Danieli1985,

PreventionInterventionStrategies Disasters present opportunities to implementthe preventive principles in outreachintenention strategies psychiatry focus on strengthening (Lystad 1984). Thesestrategies individual and familycoping capacitiesas well as decreasing adverseinfluences onrecovery. They include adaptation of resilient fostering the continued children and assistingthose withsevere stress reactions. Preventive interventions may be implementedbefore, during, disaster andmay focus on the individual(s), or after a ment.Degrde of the agent (event),or the environ- exposure and personalimpact primarily variability ofresponse. Interventions determine the should take specifically gearedto these varyingeffects precedence over lessfocused solutions.

24 4 237 MORBIDITY AFTERDISASTER

Before a Disaster Exposure property destruction,and community If exposure to lifethreat, injury, loss, goal risk factor for psychiatricmorbidity, then the first disruption is the primary formulation be to minimize exposure.Prevention requires of prevention should policyfor example,improved build- and implementationof instrumentai social damage. reduce earthquakeinjuries, loss, and property ing standards to of public policy. Psychiatry can assume animportant role in this area predisaster training iscommonly conducted inschools In the United States, the training are where there is a highrisk of' naturaldisaster. The goals of the storage of appropriatehome emergency severalfold:(a) to instruct in students with certain typesof natural disasters(for supplies; (b) to familiarize simulates an example, in LosAngeles Countyschools, a traveling van teach methods of physicalself-protection during and earthquake); and (c) to that children Although these programstypically acknowledge after a disaster. them for the broad rangeof will be scared, theymake little effort to prepare emotional reactions. Inother countriee, schoolshave commonly experienced with a wide spectrumof implemented intervention programsto enhance coping including disasters (Avalon1979; Klingman 1978).No potential situations inoculation efforts havebeen evaluation studies ofpreparedness or stress reported. is important to prepareparents and teachersto act Before a disaster, it By despite feeling personallyoverwhelmed by a crisis. decisively and effectively control of the situation,adults responding systematicallyand appearing to be in will help to reducechildren's stress (Klingrnan1978). During a Disaster Evacuation schools, medical teams,and In helping to prepareevacuation protocols for address bothadaptation-enhancing the media, mentalhealth professionals can factors during a disaster.Early warning and prompt and stress-producing Creek Dam flood evacuation are recommended.Evidence from the Buffalo evacuated before thedestruction sufferedsignificantly less suggested that those Familiarity with psychiatric morbidity andchronic functionalimpairment. understanding of evacuationprocedures at evacuation plans,including parents' be important in allayinganxiety during and after their children's schools, can parents, Inopportune or abrupt separationof young children from a disaster. distress. siblings, and other trustedadults carries its ownrisk for postdisaster

240 238 PREVENTION OFMENTAL DISORDERS

Emergency MedicalRelief Emergenty medicalrelief is intended to ensure prompt,coordinated large numbers ofinjuredpersons. Inefficiency care of of injured children can add to the medicalmorbidity and substantiallyincrease their immediate stress. Thismay occur, for example, psychological when casualtiesare evacuated to various hospitals throughoutthe region. Theauthors have aftermath of twoschool emergencies, provided consultationin the which an insecticide one in which a bombexploded andone in pollutant contaminatedclassrooms. In each or sick childrenwere transported by case, injured hospitals. No records ambulance or helicopterto several outlying indicated whichchildrenwere sent to which Consequently, parentscould not locate hospital. one case for hours, in their children forextended periods,in another formore than a day. After children commonlyfear repeated a traumatic incident, nificant others. disaster or dangerto themselvesor to sig- Prolonged separationcan increase and intensify preoccupations and theconcomitant strass. these fearful the prolonged After the schoolbombing incident, separation andworry became focal children's subsequentanxiety. points of parents'and Secondary Exposure

Children needprotection from lated, anti dead. unnecessary exposure to theinjured, muti- Involving childrenand adolescents compounds their in disasterrescue work exposure, introducingsecondary risk of After the disaster,inappropriate media psychiatric morbidity. coverage, such as exhibitingcorpsee or mutilated bodies,may also havea harmful effect. News Media

The news mediahave the capacity to play a greaterrole in informing educating the publicabout the mental and coping strategies, Lealth aspectsof disaster parenting, andavailable services. reactions, effort., have notbeen consistent Too often,however, media with good mentalhealth practices. After the MexicoCity earthquake, outreach a group of psychiatristsorganized a major program in collaborationwith the media exists for (Palacios et al.1986). Need a library of accessibleand well-evaluated including materials multimedia designed forchildren. Studies messages, best methodsof public education. ete needed to evaluatethe Aftera Disaster Psychological FirstAid Psychological first aid provides promptrelief from direct victims andfor those awaiting acute distress bothfor Clinical studies word on thecondition ofa family member. have recommendedfirst aid techniques developmentalstages (Pynoos and applicable tospecific For example, Nader 1988b;Farberow and from preslhooj,to seicOhd grade, Gprdon 1981). z4 b psychological firstaid would MORBIDITY AFTERDISASTER 239

clarifications, consistentcaregiving, and help in include repeated concrete children might be cautioned verbalizing fears, feelings,and complaints. Older posttraumatic tendencytoward impulsive behaviorand about an increased Systematic risk-taking (Tem 1979; Ethand Pynoos 1985, appendix, p.xv). research is needed on the useof age-appropriatetechniques and on early homes, disaster intervention at key sites, such astemporary relocation centers, relief offices, and schools. In working withfamilies and children atrisk for psychiatric morbidity, prevention goals include(a) ameliorating traumaticstress reactions and with child development and facilitating grief work; (b)preventing interferences and (c) promoting competencein effectively the resulting maladjustments; requires adapting to the crisis situation.Successful preventive intervention who are identifiably at risk,treatment of populationsunder- access to children prevention of the onset of going normative reactionsto extreme stress, and disorders or 'eduction of theirduration and progression. Family The family is the key t;ettingin which feelings ofvulnerability can be restored. The goals for familywork are (a) to mitigated and a sense of security establish a sense of give children the experienc..4of being supported, (b) to physical security, (c) to validaterather than dismiss children'saffective re- by sponses, and (d) toassist children in dealingwith traumatic reminders accepting their renewedanxiety and providinghelpful reassurances. Family intervention members often need support,guidance, and sometimes therapeutic effectively help their children. to reduce their ownlevels of stress before they can disaster responses, They need information aboutthe wide range of children's the effect of traumaticreminders, the presence ofarousal behavior, realistic expectations regarding recovery,and the need to encourage opencommunica- tion with their children. Much remains unknown.Research is particularly neededin the following parenting ability following traumaand loss; (b) managing areas: (a) enhancing and arousal states; children's new, incident-specificfears, regressive behaviors and (c) facilitating children'sgrief work. Anthony (1986) has noted thatafter a disaster, individuals feelsignificantly less secure and morevulnerable, and exhibit increasedattachment behavior. In children this often manifests as acontinued fear of recurrence, newincident- afraid of specific specific fears, and regressivebehavior. Children may beceme places, concrete items, andhuman behaviors that specificallyremind them of the incident. They may alsoexhibit more generalized fear ofbeing vulnerable; for instance, they may becomeafraid of strangers, beinginthe dark, being alone, being in their own roam or inthe bathroom, or going tobed. Regressive behaviors commonly observed inchildren following a disaster includeenuresis, abandonment of previously learnedskills, and increased dependency,including 4"I 240 PREVENTION OF MENTALDISORDERS

clinging, a need to staynear home, and asking to sleep with al. 1982; Newman 1976). parents (Burke et School-based MentalHealth Intervention Some school districts haveinstituted psychologicalcrisis teams that immediately to emergencies can respond or community disasters. Usingclassroom consult- ations, an entire schoolcan be screened and children varying levels of risk. prioritized to.cording to Therapeutic consultationsshould follow earlycase fmdings. The school is the optimalsite because (a) it is parents and (b) the stigmatization most convenient to childrenand that accompaniesuse of mental health facilities is obviated. Theschool is an ideal locus and children in to involve parents, teachers, preventively oriented,trauma responseprograms. Crabbs (1981), Blom (1986), andPynoos and Nader (1988b) for such programs. Elements have provided guidelines ofsuch a program includeconsultation with school administrators, training ofteachers, and educationof parents and children. Identifying and addressingthe mostcommon rumors, misconceptions, has helped to minimize and fears anxiety in all members ofthe school communityand to limit interference witheveryday activities. Classroom drawing exercises and engaging children,over time, in symbolic reconstruction in play haveproved to be effectivemethods of initial intervention (Pynoos and Nader 1988b). The usefulness of theseclassroom interventionsin reducing children's fearof recurrencewas demonstrated in the aftermath Italian earthquake (Galante of an and Foa 1986). Redramatizingthe earthquake's destructive force and thenreconstructing their village important two-step in play proved to bean process in reducing children'scognitive preoccupations (Galante and Foa 1986,appendix, p. v).Special procedures reintegrate hospitalized are needed to or severely traumatized childreninto the classroom,to deal with bereavedchildren, and to monitor school behavior andperformance. Intervention withIndividual Children Children who havehad severe exposure to life threat and havewitnessed injury or deathmay require direct forms of individual serious psychiatric intervention to avertmore morbidity. While onlypreliminary investigationsinto the appropriate therapeutictechniques for children adult PTSD have found are available, investigators of that the optimal timefor intervention is in theacute period, when theintrusive phenomena are most apparent and theassociated affect is most available(Kaltreider et al. 1979). reminders are most easily Incident-specific traumatic identifiable during theacute phase after the disaster. Clinically, school-age children and adolescents haveparticipated in thesame kind of acute debriefingthat has been and Eth 1986; Frederick a hallmark of adult trauma work(Pynoos 1985, pp. 71-99). Thegoal is to assist children thoroughly exploring in their subjectiveexperience and to helpthem understand the meaning oftheirresponses. The consultation observing-ego and reality-testing bolsters the children's functions, thereby dispelh,,gcognitive confusions MORBIDIT Y AFTER DISASTER 241 and encouraging active coping(Caplan 1981). Children are therebyassisted in identifying traumatic remindersthat elicit psychophysiologicreactions, intrusive imagery, and intenseaffective responses. The specificaim is to increase children's sense of being able toanticipate or, at least, manage their recurrence. Managing what are usuallyunavoidable daily reminders maybe the key to enhancing the children's sense ofmastery of a disaster experience.Enabling children to share these traumaticreminders with their parents increasesthe likelihood that they will receiveessential parental support andunderstanding. Perhaps the best gauge of the effectiveneesof acute individual intervention may to be children's improved capacity toparticipate in bolving problems secondary the crisis, for example, to prevent unnecessaryseparations from siblings or to arrange contacts ;;th peers. Posttraumatic stress reactions areexpectable and understandable psychological phenomena that resultfrom traumatic exposure. By usingtheir authority, mental health professionals canlegitimize the children's feelings and reactions and assist them inmaintaining their self-esteem. Children canalso be prepared to anticipate and copewith the transient return ofunresolved feelings over time. Although trauma debriefing andconsultation have proven helpful, many severely exposed children will require moreextended therapeutic interventions. There is an immediate need forinvestigation of brief, focal psychotherapywith traumatized children (Pynoos and Eth1984). Adult studies provide evidence that such therapy can be effective inPTSD treatment (Kaltreider et al. 1979). These methods need to be modified toreflect developmental considerationsand to be tested for efficacy. Modifying arotuPl behavior nay be animportant aspect of an overall treatment plan. Laboratory and preliminaryclinical data in adults indicate that this neurophysiological response can be attenuatedby pharmacological intervention (Kolb 1987). A pilot investigation ofpersistent arousal behavior in children exposed to gunfire incidents is beingconducted at the University of California, Los Angeles (UCLA). Preliminarydata suggest that abnormalities inacoustic startle response and stage 4 sleepphenomena are present and can bearrested by the use of clonidine. In a study of acutePTSD in children, arousal behavior, especially sleep disturbance, wasemaciated with interference with attention and learning (Pynoos et al. 1987a).Early alleviation of this symptom may decrease chronicity and reduce functionalimpairment in severely exposed children. Conclusion

A growing knowledge base is currentlyemerging from solid research studies of the psychiatric impact of disasters.These newer studies are leading to more data-based strategies of intervention forexposed populations. The study of children's reactions is an area of particular concern,requiring more systematic investigation. 2413 242 PREVENTION OF MENTAL DISORDERS

The involvement of child psychiatrists in thisarea provides a new dimension for psychiatrists concerned with school mentalhealth, preventive andcommunity psychiatry, child development,and child psychopathology. Given the high frequency of disasters, it is importantfor child psychiatrists to beaware of their roles and functions in the planning anddelivery of appropriate preventive mental health services.

25o MORBIDITY AFTER DISASTER 243

Appendix Childhood n'auma Studies

Table 2 briefly summarizes studiesof traumatized children between1945 and 1987 and is provided to assistthe reader in comparing and contrastingthe studies, their methods, and results.Please refer to table 1 for the key tothe terminology and setup of table 2. 244 PREVENTION OF MENTAL DISORDERS

Table L Key to Table 2

Column Term Meaning

Author Author(s) and Date of the Study Subjects Number of Subjects Subjects n The number of subjects included in the study Study Characteristics of the study Study Group size > 10 Whether the size of the studied group or comparison group was greater than 10 Study Time period The length of time after the event that the child(ren) was (were) studied or questioned Study Control group Whether a nonexposed comparison group with characteristics similar to the study group was studied simultaneously Study Selection The method of selecting subjects Study Questioned Who was questioned Study Instrument The instrument used or the method of study (PEF = Psychiatric Evalua- tion Form: CGAS= Childhood Global Assessment Scale) Findings Results Found Findings Parent/child Conclusions about parental influence on child(in)'s reactions Findings Exposure The effects of varying proximities and circumstances on child(ren)'s reactions Findings Symptoms Symptoms observed or reported for the child(ren)

25 2 Table 2. Childhood Trauma Studies

Hospitalizations Author Subjects Study Findings Postoperative sequelae; cases reviewed Symptoms: 25% with emotional se- Levy n = 124 0-10 or more years after the operation quelae similar to those of adults with com- 1945 ages 0-14 bat neurosis; highest number of symptoms Group size > 10: no in children younger than 3. Time period: varied Symptoms and age of prevalence: Night Control group: no terrors, ages 1-2; negativistic reactions Selection: from previous patients' (defiance, dependence, tantrums), 4 and records older; fears, 0-8. Questioned: mothers, retrospective Categorized emotional sequelae: condi- Instrument: none; 2- to 4-hour interview tioned fear (dependency and regressive latent fear; phobias; anxiety states; hostile reactions; obsessions; hysteria). Natural Disasters Author Subjects Study Findings Tornado in Mississippi; parents (and 3 Exposure: Proximity to the disaster zone Bloch et al. n = 185 families) interviewed and loss of injury to family members were most 1956 preeictive of anxiety, symptom formation, or Group size > 10: unclear intensification of pathological character traits. :1, Time period: 1 week after Parent/child: Suppression of discussion Control group: not el ac-rly delineated may have intensified symptoms for child. Selection: families selected from those Symptoms: increase in dependency, cling- who had filled out a questionnaire dis- ing, need to stay near the home, asking to sleep tributed through one of the community with parents, regression (enuresis, abandon- schools ment of previously learned skills, night terrors Questioned: 88 parents, 3 families in which experience was relived, tornado Instrument: none; 3/4- to 2-hour games, general irritability and sensitivity to unstructured interviews noise, phobic, and avoidance symptoms). Table 2. ChildhoodTrauma Studies(continued) Natural Disasrs continued Author Subjects Study Findings Blaufarb and n = 300 A California earthquake; Levin parents and Parent/child: Children takecues from 1972 children seen in 1 to 2group meetings to relieve stress parents. Symptoms: Common fearswere sleeping alone, being alone ina room, clinging Group size > 10: unclear (ages 3-6), scattered incidence ofregression in toilet and eating habits. Time period: froma few days to 5 weeks after Control group:no Selection: respondentsto an offer to provide services Questioned: parents andchildren Instrument: none; crisis-orientedgroups

Lacey 1972 n = 66 Aberfan, Wales, mining tip complex gave Parent/child: Children reflected elementary way and roared down mountain,killing 116 parents. children and 28 adults; Symptoms:Sleeping difficulties,ner- children children andvousness, lack of friends, school parents interviewed afterparent referral avoidance, reduction in play, instability,enuresis, wor- Group size >10: unclear sening of existing symptoms.linging, fear Time period: within 4years after of being alone, fearof damness and of Control group:no weather, angerover presence of tip mines, Selection: referred dueto parentalconcerninitial poor school performance,games of Questioned: parents andchildren burying in the sand,anxiety (children with Instrument: none; clinicalinterviews previous anxiety creatingsituations in their backgroundswere most affected). Table 2. Childhood Trauma Studies (continued) Natural Disasters (continued) Author Subjects Study Findings Newman 1976n = 224 Buffalo Creek slag dam flood; clinical obser- Exposure: l lisaster effects attributed to vations developmentslevel, perceptions of their Group size > 10: unclear family's responses, and direct exposure to Time period: unclear the disaster. Control group: no Symptoms: Nightmares, screaming in Selection: survivor plaintiffs reenonse to contact with water, instability, enunssis, a clear and enduring creativity, Questione& families and children Instrument: none; clinical interviews and hopefulness in contrast to their parents.

Glaser et al. n = 207 Buffalo Creek slag dam broke and flooded; Parent/child: Children's reactions were 1981 PEF administered. highly correlated with those of their Group size > 10: yes parents. Time period: 16 months-2 years after Exposure:Adults' reactions were as- Control group: no sociated with proximity to the flood zone, Selection: litigants loss, and reactions of other family mem- Questioned: families bers. Instrument: PEF; informal family Symptoms: Reduced interest in school interviews correlated with decreased grades. Grades also correlated with overall severity (PER Bedwetters (6% of the children) scored high on belligerence, anxiety, and overall severity.Children with posttraurna obesity scored higher on depression and overall severity. Bereavement intensified 255 reactions moi e than did property damage. Table 2. ChildhoodTrauma Studies (continued) Natural Dismters (continued) Author Subjects Study Finding Burke et al. n = 81 Severe winter storm and 1982 flood; Conner's Exposure: Closer proximityto the flood Parent and TeacherQuestionnaire administered area and economic status affectedscores. Symptoms: Children with specialneeds Group size > 10: unclear were at risk for increased aggressivecon- Time period: 4-5 months after duct scores. All children showedimproved Control group: unclear; proximityto school conductscores. Boys' anxiety scores flood vs. damage not clearlyspecified increased and girls' decreased. Selection: Head Start children Questioned: parents and teachers Instrument: Connor's Parent andTeacher Questionaaires

2 5 t ; 1) Table 2. Childhood Trauma Studies (continued)

Natural Disasters (continued) Author Subjects Study Findings Handford et al.n = 35 Nuclear plant accident at Three Mile Island Exposure: No effect of distance of 1988 nuclear power plant residence from Three Mile Island. Parent/child:Parente reported fewer Group size > 10: no (4 psych) symptoms for children than children Time period: 1-1/2 years atter reported for themselves. Children's reac- Control group: yes tions not related to parents' reactions un- Selection: unknown less the two parents responded with Questioned: parents and children different intensities. Instrument: for children: structured Symptoms: Continued anxiety. interviewe, questionnaire devised for thisChildren expressed having fun, fear, or event, Quay and Peterson Behavior Prob-anger related to the evacuation. Contradic- lem Checklist, Children's Manifest tions about the presence of symptoms Anxiety Scale (Casteneda et al. 1056), (revealed symptoms before or after denying Kinetic Family drawings; for parents: them). Symptoms not seen in a majority of SCL-90-R, profile of mood states, MMPI children: lack of memory, games related to the incident, avoidance of thoughts about the incident, dreams, denial of concern, pride in their reactions, rationalization, fatalism, preoccupation, fears about the future (no significant effect of age, sex, or IQ). Reactions less severe than thcee reported for Chowchilla kidnap victims.

2 5 7 Table 2. Childhood Trauma Studies(continued) Natural Disasters continued) Author Subjects Study Findings Age:No significant difference in symptoms by age. Children (mean 9years, 11 months), undifferentiated fears; adolescents (14 years, 1-2 months), fears for self and family; adolescents (15years, 5 months), concerns for self, family, friends, and flituie offspring. Misconceptions and exaggerations seen in children 8-12. No recognition of danger in childrenyounger than 8 years, 5 months. Only children older than 13 seemed to understand theproper- ties of radiation. Course: Compared to a different study of same incident, parents' reactions ata few months after did not persist; children with emotional disorders were at risk; children reported continued anxiety.

25S G 1 .

Table 2. Childhood Trauma Studies (continued)

Natural Disasters (continued) Author Subjects Study Findings Severe earthquake in Italy; Rutter's Be- Exposure:The amount of damage, Galante and n = 300 havioral Questionnaire administered bydestruction, and death to a village were not Foa 1986 grades 1-4 teachers always predictive of risk for neurotic or antisocial disturbances. A significant cor- Group size > 10: yes relation between family deaths and at-risk Time period: 6 months after scores occurred for only one village studied. Cortrol group: no Symptoms: Intrusive imagery, omens, Selection: 1-4 graders in six villages, fears of recurrence, drawings full of menac- then from Calibritto because it had ing and threatening figures, fears of im- the most at-risk children pending doom, retelling, regression, Questioned: teachers themes of death in play, anger over delayed Instrument: Rutter's Children's Behavior assistance during the disaster, anniversary Questionnaire for completion by teachers,fears. then group sessions of 4 children Table 2, Childhood TraumaStudies (continued) Natural Disasters continued Author Subjects Study Findings McFarlane a = 808 Australian bushfire; Rutter's Parent et al. 1987 and Exposure: Morbidity rates lower for the ages 5-12 Teacher Questionnaire administered exposed group at 2 months after the dis- Group size > 10; yes aster; no significant differences reported by Time period: 2, 8, and 26 months after teachers at 8 and 26 months after; parents Control group: yes reported greater symptoms for affected Questioned: teachers and Farents children at and 26 months.Affected Instrument; Rutter's questionnaires children were rated as more obedient,perhaps explaining the teachers' reports.

McFarlane Symptoms; Delayed onset of symptom; n = 808 Australian bushfire; Rutter's parent andpreoccupation with the Pre: night-mares; et al. 1987 ages 5-12 teacher questionnaires administered underachievement in school and increased As above (McFarlane et al.); after 26absences; Rutter itemsangry, phobic, months parents were questioned aboutneurotic, somatic, distractable, restless. events after the fire for their children and Exposure: Mothers' intrusive imagery, about their own responses to the fire changed parenting, and separation in the days right after the firewere more predictive of ch ildren's poettraumatic phenomena than direct expoeure to the disaster. Symptom persistence;Early disturbance predictive of later disturbance. In- tervening adverse life eventswere predictors of children's continuing preoccupations with the disaster. Irritabil- ity of mothers predictive of children's 2() posttraumatic symptoms at 8 months after; mother's preoccupation with andfear of future fire at 26 months. Table 2. Childhood Trauma Studies (continued)

Animal Attacks Author Subjects Study Findings MacLean 1977 n = 1 Leopard attack on father and son Symptoms: Traumatic neurosis preschooler Group size > 10: no (single case study) reenactments of the event or taking boy and Time period: 5 months after father to hospital; avoidance of emotions, Control group: no aggression, and regression. Included were Selection: father's referral fear and terror; not wanting to be separated Questioned: child and parent from parents; clinging; concern about his Instrument: clinical interview father's whereabouts; concerns about being "eaten up"and accidents to family members in play; anger; fear of retaliation; ambivalence; defenses against emotions identifying with the aggressive animal, father, mother, or therapist; a pressure to activity. Pattern of love, anger, fear, hate. Course: Decrease in sy.nptoms over the course of therapy. MacLean 1980ri = 1 4 yeate aftera leopard attack (as above, Syrontoms: Stress situation of the class- MacLean) room resulted in "regression uuder stress.' Response to "harsh school teacher," related to the original leopard attack.Fear of water and of drowning, world seen Et:, dangerous place with dangerous surprisee, hospital and wicked witch play, preoccupation with death (there was also an intervening threatened loss with mother's pregnancy and hospitalilation). 2 f; 1 Table 2. Childhood TraumaStudies (continued) Animal Attacks (continued Author Subjects Study Findings Gislason and n = 3 Moderate to severe dog bites (3 children) Call 1982 Symptoms: All three childrenwere more Group size > 10: no (singlecase studies) afraid of strange dogs and preoccupiedwith Time period: 20, 40, and 4 monthsafter biting in general. They exhiliited clingingto Control group: no their mothers, lowered self-esteem,revenge Selection: attorney's referral fantasies, and nightmares. They becameun- Questioned: child and parent usually caution% inhilited, andimpaired in Instrument: clinical interview their capacityfor enjoyment, enthusiasm,and pleasum Each child maintaineda vivid recol- lection of the dog-bite incident. Fewregreesive trends and no reenactment behaviorswere observed, although childrensometimes showed a preoccupation with dogsin their thinking and playing. The psychiatristsob- served constriction in thought andfantasy and adverse effect on personality.

Violence Author Subjects Study Findings Bergen 1958 n = 1 Schizophrenic lather stabs motherof Symptoms: Repetitive play and 4-year-old reenactment (murder game, painting handsred), Group size > 10: no (singlecase study) repetitive retelling, guilt,separation Time period: unknown anxiety, fears, fear the dead, thinkyou Control group: no don't love them, fear of attackand injury, Selection: pediatrician referral loneliness, denial, aggressiveness followed Questioned: child by hand washing, impulsive behaviorthat Instrument: clinical interviews/treatmentwas dangerous to herself, fears ofown aggression, anxiety, oedipal issues. Table 2. Childhood Trauma Studies(continued) Violence (continued)

Author Subjects Study Findings Ruben 1974 n = 1 Comparison of a strain and a shock trauma Symptoms:Became painfully shy, a prelatency girl, age 6, nipped herdesire to become somebody special, sibling brother's penis, drawing blood, andwasissues reenacted in later life, masochistic forced to perform fellatio character traits, infantile narcissism, and Group size > 10: no (single case study) need for a mother, sadness, dislike of Time period: age 13-1/2 mother, lack of friends, horgility, guilt. Control group: no Phobias associated with a threat bya Selection: parent referral babysitter. Questioned: teenager Instrument: clinical interviews/statement Burgess 1975 n = 9 families Reactions of families of homicide victims Symptoms: Acute griefphrase ego-oriented Group size > 10: no (single case studies) thought about lcse of family member, horror Time period: varied over the manner of death, thoughts about how Control group: no the victim must have felt, desire to know the Selection: self-referral facts, outrage, anger, aggression, and the desire Questioned: families to physically do something about the crime; Instrument: clinical interviews/treatment amplicating factors-the public learns through the media, identifying the body, whetheror not to open the casket, numbnees and confusion, insomnia, sleep pattern disturbances, headaches, chest pain, palpitations, and gastrointestinal upsets. Reactionsover time gist work guilt ( If only ..." ), dreams and

2 fl 3 Table 2. Childhood Trauma Studies(continued)

Study Findings nightmares (some wish fulfillment dreams), phobic reactions, identification with the deceased, possible role change, dealing with the court process, undermining of faith in the world, helplessness, blame, stigmatization.

Preschoolers whose fathers murdered their Symptoms:Disavowal of the death, mothers (2 pairs of siblings ages 11 months nactment in play, headache, delayed and 2, 2-314, and 4-1/2 years) grieving, repeated retelling of event, thumb Group size > 10: no (single case studies) sucking, "balkiness," telling tales, crying in Time period: 10 days after and 2years 2-year-olds, withdrawal after visits with after murderer's parents, enuresis, anxiety at Control group: no separation, guilt, fears of vampires, anxiety Selection: adult referral with discussing event, anxiety about Questioned: children sibling's whereabouts, loyalty conflicts. Instrument: clinical interviews/treatment Other factors: Stability of placement important to the recovery of the children. Need for swift resolution of custody issues. A familiar environment is preferable; however, it is important that surrogate parents be able to handle reality and maintain openness to discussing the event, which requires understanding of age variables, lack of ii vestment in the murd^r, andac- curate information about what i):.-.Ippened. 2 E Table 2. Childhood Trauma Studies (continued)

Violence (continued) Author Subjects Study Findings Preschool siblings yips 2 years, 3 months Pruett 1979 n = 2 Symptoms:Symptoms of massive and 3 years, 7 months) who witnessed thetrauma.Regression (including toilet murder of their mother by their father;habits), terror, awakening distressed look- father attempted suicide ing for mother, pressured questioning about the way the parents looked, lack of understanding about death, attempts to comfort grieving grandparents, fear, fear of death, nightmares, denial and avoidance. Age:Age-appropriate concerns about death and the preservation of bodily in- tegrity. Murderous fantasies, wishes, and fears that are age appropriate. Treatment:Helping grandparents to answer questions, help child differentiate self and own fate from that of mother, lend- ing suppat to ego strength. Information and clarification should be kept at minimum to avoid inhibition of defenses and thwart adequate repression of the trauma. Provide answers to pressured inquiries. Course: Decrease in symptoms over the course of therapy.

215 Table 2. Childhood Trauma Studies(continued) Violence (continued)

Author Subjects Study Findings Terr 1979 n = 20 Children kidnapped from their schoolbus; Symptoms: Omen formation; hallucina- traveled for 11 hours in darkness of boardedtions; fear of repeat; disturbances of cogni- van without food or drink; buried for 16tion (perception, timesense, thought); hours in truck trai;er until they dug them-repeated traumatic dreams; posttraumatic selves out play; reenactment of attitudes, fears,or Group size > 10: not by age actions that occurred during the kidnap- Time period: 5-13 months after ping; fears; psychophysiologicaloccurren- Control group: no ces; anxiety; personality changes; absence Selection: offer of assistance of flashbacks, amnesia,or haziness; denial Questioned: child alternating with intrusive repetitive Instrument: clinical interviews phenomena or ego dysfunction.

211; Table 2. Childhood Trauma Studies (continued)

Violence (continued) Author Subjects Study Findings Eth and n = 50 Child witnesses to violence Symptoms: PTSD symptoms; sleep dis- Pynoos 1985 Group size >10: no (single case studies) turbances, night terrors, and somnam- Time period: unknown bulism; startle reactions to traumatic Control group: no reminders; intrusive imagery and as- Selection: referral sociated affect; denial-in-fantasy, inhibi- Questioned: parent and child tion of spontaneous thought, fixation to the Instrument: clinical interviews/treatmenttrauma, preoccupations with future harm used to limit traumatic anxiety; cognitive reappraisals of the event; traumatic helplessness; identification with third parties intervening and development of newcareer interests; challenge to trust in adult restraint and to own impulse control; un- characteristic aggressive, reckless, or self- destructivebehavior;unconscious reenactment; revenge fantasies or dreams; confusion; age-related difficulties in processing the event.

2 Table 2. Childhood TraumaStudies (continued) Violence (continued Author Subjects Study Findings Zeanah and n = 1 4-year-old child witnessed her father Burke 1984 History.Previous abuse and removal strangle her mother from home. Group size > 10: no (singlecase study) Symptoms: PTSD symptoms; hyperac- Time period: 1 month after tivity, sleep disturbances, fear of toileting Control group: no alone, enuresis, and traumatic play and Selection: referral reenactment; aggressive behavior; iden- Questioned: child tification with the victim; sleep distur- Instrument: clinical interviews/treatmentbance; need for reassurance about what happened; fear of aggression; anxiety about aggressive impulses; identification with the protector; anxiety; sense of danger in thr external world. Treatment: Included work with foster parents, PTSD, mourning, identification of aggression.

2j Table 2. Childhood Trauma Studies (continued)

Irit2lee continued) Author Subjects Study Findings Cambodian adolescent children who lived Parent/child: A strong relatiOnship be- Kinzie et al. n = 46 through 4 years of severe concentration 1986 tween living situation and psychiatric diag- camp-like experience. nosis: children who lived with one or more Group size > 10: no nuclear family members were less likely to Time period: 4 years after have a diagnosis than children living in Control group: of 6 children foster homes. Selection: Cambodian refugees, local area Symptoms: PTSD symptoms; night- Questioned: guardian and child mares, recurring dremns, intrusive mental Instrument: CGAS states of shame over being alive, and sometimes seemed pressured to tell horrible events of their pasts. Loss of energy and interest, avoiding memories of Cambodia and avoiding discussion of traumatic events; startle reactions and guilt; helplessness and hopelessness; pessimism and brooding; headaches and concerns about health. Fossibly because of their Cambodian values, there was no social acting-out behavior, truancy, other disruptive school behaviors, or drug use. Age/sex: No relationship between the Cambodian experience, age, or sex and the presence of a diagnosis (CGAS). 2 0 Table 2. Childhood Trauma Studies(continued) Violence (continued)

Author Subjecta Study Findings Terr 1983 n = 26 4 years after a kidnapping freina schoolbun Parent/child:ftelaVonship of family Geoup size > 10: no pathology to lack of community bonding; Time period: 4-5 years after recent family problems and individual vul- Control group: no nerabilities were related to severity. Selection: original subjects A Symptems:Occasional recurrences of Questioned: child aumatic anxiety; not wanting people to know Instrument: clinical interviews about the victimiartion; fears; SUIV1394011or conecious avoklance of thoughts about the kkl- napping; absence of &mof external reality; memory c4' the trauma but forptfalnees shcut symptoms; only four children with related prcblems w echool perfcemance; emoaonal die tancing through me d metaphor, physicalBen- salem when recallingthe .sent ; displacement of affect; mispereeptkns; pereeptualover- generalizations resulted in startle reactions, suspicions, and physical discomfort; continued time distortion; omens; Rxeshortened &WM; pa:simian; nightmares; posttraumatic play, reenactments or eapeated physkal sensations. Over time: more evidentshame, thoughtsup- pression, denial and repression of symptems, unlinking of memories from affect, memories of misperceptions, sense of foreshortened future, death dreams, and the dangerous nature,con- 270 tagion and repetition of pcsttlaumatie play and reenactment; disappear--some fears, exact repetitions in dreams. Table 2. Childhood Trauma Studies (continued)

Violence (continued) Author Subjects Study Findings Los Angeles children exposed to a sniper Exposure: PTSD symptoms and Pynoos ei al. n = 159 attack on their elementary schoolseverity level increased as the degree of 1987 playground exposure increased. Group size >10: yes Symptoms: PTSD symptoms:More Tim" period: 1 month after than 70% reported identification of the C. ..rol group: yes event as an extreme stressor, getting upset Selection: random when thinking about the event, fear of Qucetioned: children recurrence of the event. Fewer than one- Instrument: PTSD Reaction Index th ird reported experiencing loss of interest in significant activities, difficulty paying attention, interference with learning, estrangement and interpersonal distance, or guilt.Significant differences between exposure levels for all but two items, fear of recurrence and guilt.Symptoms always present in severe reaction:interpersonal distance, reduced interest in activities, difficulty paying attention, sleep disturbance, intrusive imagery, intrusive thoughts, and emotional avoidance. Children with mild or no reactions almost never reported a reduced interest in activities, interpersonal distance, or guilt. Increased knowledge of the victim significantly related to increased symptomatology.Previous trauma and worry about a sibling not statistically sig- 271 nificant. Previous trauma elicited renewed Table 2. Childhood TraumaStudies (continued)

Author Subject, Study Findings thoughts and images rf the previousevent. Children in all expos.regroups identified worry about a sibling's safety assource of extreme stress and continued to have specific anxieties about it. Age, sex, and life events: Igo significant difference for age, sex, or previous life e, exposure level. 265 MORBIDITY AFTERDISASTER

References

American Psychologist Adams, P.R., and Adams,G.R. Mt. St. Heiens' ashfall, 39:252-260, 1984. Statistical Manual ofMentcd American PsychiatricAssociation. Diagnostic and Disorders. 3d ed., revised.Washington, D.C.: AmericanPsychiatric Associa- tion, 1980. and Dadock, Andreasen, N.C. Posttraumatic stressdisorder. In: Kaplan, Hi., Vol. 1. Baltimore: Wil- B.J., eds. ComprehensiveTextbook of Psychiatry / IV. liams & Wilkins, 1985. Anthony, E.J. The response tooverwhelming stress: Someintroductory comments. Journal of theAmerican Academy of Childand Adolescent Psychiatry 25:299-305, 1986. Death Education Avalon, 0. Community orientedpreparation for emergency. 3:227-244, 1979. Berezofsky, R.E. Post-TraumaticStress Disorder and theTechnological Dis aster. Washington, D.C.:The Bureau of NationalAffairs, Inc., 1987. Bergen, M. Effect of severe trauma on a4-year-old child. PsychoanalyticStudy of the Child 13:407-429,1958. Blaufarb, H., and Levin, J.Crisis intervention in anearthquake, Social Work 17:16-19, 1972, Bloch, D.; Silber, E.; and Perry,S. Some factors in theemotional reaction of children to disaster. AmericanJournal of Psychiatry 113:416-422,1956. Journal of Blom, G.E. A schooldisasterIntervention and research aspects. the American Academy ofChild Psychiatry 25(3):336-345,1986, Bowlby, J . Attachment andLoss. Vol, III, New York:Basic Books, 1980. Promet, E. Three Mile Island:Mental Health Findings.National Institute of Mental Health Contract Report,Pub, No. (SM) 287-79-0048.Washington, D.C.: U.S. Govt. Print. Oft, 1980. Burke, J,D., Jr.; Borus, J.F.;Burnes, B.; Millstein, KH.; andBeasley, M.D. Changes in children's behavior after anatural disaster. AmericanJournal of Psychiatry 139:1010-1014, 1982. American Journal of Ortho- Burgess, A.W.Family reaction to homicide. psychiatry 45(3):391-398, 1975. Caplan, G. Mastery of stress:Psychological aspects. AmericanJournal of Psychiatry 138:413-420, 1981, 273 266 PREVENTION OF MENTALDISORDERS

Carey-Trefzer, C.J. The resultsof a clinical study of war-damagedchildren who attended the child guidanceclinic, The Hospital for SickChildren, Great Ormond Street, London. Journalof Mental ScienceXCV:535-559, 1949. Castaneda, A.; McCandlees, B.R;and Palermo, D.S. Thechildren's form of the Manifest Anxiety Scale. ChildDevelopment 27(3):317-326, 1956. Cohen, RE. Developmental phases of children's reactionsfollowing natural disasters. Journal of Emergencyand Disaster Medkine1-4:89-95, 1986, Cohen, RE. Intervention programs for children. In: Lystad, M., ed.Human Response to Mass Emergencies.New York: Brunner/Mazel,in press. Crabbe, M. School mental health services followingan environmental disaster. Journal of School Health51:165-167, 1981. Danieli, Y. The treatment and prevention of len-termeffects and intergenera- tional transmission ofvictimization: A lesson fromholocaust survivors and their children. In: Figley,C., ed. Trauma and ItsWake. New York: Brunner/ Mazel, 1985.

Davis, M. Neurobiologicalperspectives on anxiety. letin 19:457-465, 1983. Psychopharmacology Bul-

Dohrenwend, B.P.; Dohrenwend,B.S.; Warheit, G.J.;Bartlett, G.S.; Goldsteen, R.L.; Goldsteen, K; and Martin, J.L. Stress in thecommunity: A report to the President's Commission on the accident at Three Mile Island.Annals of the New York Academy of Science365:159-174, 1981. Erickson, KT. Loss of communality at Buffalo Creek.American Journal of Psychiatry 138:302-305,1976. Eth, S., and Pynoce, R. Developmentalperspectives on psychictrauma in children. In: Figley, C.,2d. Trauma and Its Wake. 1985. New York: Brunner/Mazel,

Farberow, N.L., andGordon, N.S. Manual forChild Health Workers in Disasters. DIMS Pub. No. Major (ADM)81-1070. Washington,D.C.: National In- stitute f Mental Health,1981. Figley, C.R. Introduction. In: Figley, C., ed.Trauma and Its Wake. Vol.2. New York: Brunner/Mazel,1986. Fisher, C.; Byrne, J.; Edwards, A; and Kahn,E. A psychophysiologicalstudy of nightmares andnight terror.Journal of Nervous 157:75-98, 1973. and Mental Diseases Frederick, C.J. Effects of natural versus human-inducedviolence upon victims. Evaluation and ChangeSpecial Issue: 71-75, 1980. 274 267 MORBIDITY AFTERDISASTER

situatir as. In: Eth,S., Frederick, C.J. Childrentraumatimx1 by catastrophic Disorder in Children.Wailington, and Pynoos, R., eds.Post-Traumatic Stress D.C.: AmericanPsychiatric Press, 1985. York: Medical War Freud, A, andBurlingham, D. Warand Children. New Books, 1943. study of psychic traumaand Galante, R., and Foa,D. An epidemiological for children after anatural disaster.Journal of the treatment effectiveness Psychiatry 25:3357-3363,1986. American Academy ofChild and Adolescent Critique and commentary.Journal Garmezy, N. Childrenunder severe stress: Adolescent Psychiatry25:384-392, 1986. of the AmericanAcademy of Child and Dog bite in infancy:Trauma L .personality I.L., and Call, J. Adolescent development. Journal ofthe AmericanAcademy of Cnild and Psychiatry 21:203-207,1982. Psychosockd Effects of B.L.; and Winget,C. Prolonged Glaser, G.C.; Green, 1981. Disaster: A Study ofBuffalo Creek. NewYork: Academic Press, following M.C.; and Glesser, G.C.Long-term impairment Green, B.L.; Grace, and Clinical Psychol- the Beverly F1Supper Club fire.Journal of Consulting ogy 53:672-67e,1985a. Titchener, J.L.; andlAndy, J.G. Levels Green, B.L. Grace,M.C.; Lindy, J.D.; following a civilian disaster:The Beverly Hills Supper of functional impairmlnt 1983. Club fire. Journal ofConsulting and ClinicalPsychology 50:573-580, J.D.; and Grecs, M.0Post-traumatic stressdisovier: Green, B.L.; Lindy, Disorders 173:406-411,1985b. Toward DSM IV. Journalof Nervous and Mental Spiegel, J.P. War neurooesin North Africa:The Tunisian Grinker, RR., and Sept. 1943. New York:Josiah Campaign. The AirSurgeon, Army Air For Macy, Jr. Foundation. H-mphrey, F.J.; Bagnato,S.; Handford, M.A.; Mayes,S.D.; Mattison, R.E, ; J.D. Child and parentreaction to the 'fhreeMile Island Bizler, E.O.; and Kales, Academy of Child andAdolescent nuclear accident.Journal of the American Psychiatry 25:346-356,1986. feeling and behavior.American Hoffman, M. Developmentof moral thought, Psychologist 34:958-966,1979. :iisaster victims: Theemotional effects of recoveringand Jones, D R. Secondary 142:303-307, identifying human remains.Anwri Journal of Psychiafty 1985.

n ;.) 268 PREVENTION OF MENTALDISORDERS

Kaltreider, N.B.; Wallace, A.; and Horowitz,M,J. A field studyof the stress response ayndrome:Young women after Americc hysterectomy.Journal of the :Mica: Association242:1499-1503, 1979. Kinzie, J.D.; Sack,W.H.; Angell, R.H.; effects of massive Manson, S.; and Rath,B. The psychiatric trauma on Cambodianchildren: I. The children. the American Academyof Child and Journal of Adolescent Psychiatty25:370-376, 1986. laiLlan, A. The Corningflood project: disaster. In: Parad, Psychological first aidfollowing a natural H.J.; Resnik, H.L.P.;and Parad, L.G., eds. Disaster Management.Bowie, Md.: The Emergency and Charles Press Publishers,Inc., 1976. Kliman, G. PsychologicitlEmergencies of Childhood. Stratton, 1968. New York; Grune& Klingman, A. Children in stress: Anticipatoryguidance in the framework the educationalsystem. Personnel and of Guidance Journal57:22-26, 1978. Kolb, LC. Aneuropsychological hypothesis disorders. American explaining posttraumaticstress Journal of Psychiatry144:989-995, 1987. Kramer, M.; Sclyjen,L.S.; and Kinney,L. The dream disturbed Vietv-n experience in dream- veterans. In: Van derKolk, B., ed. Disorder: Psychological Post-Traumatic Stress and BiologicalSequelae. Washington, Psychiatric Press, 1984. D.C.: American K.rupnick, J.L.Bereavement during chikaiood and adolescence.In: Osterweis, M.; Solomon, F.;and Green, M.,eds. Bereavement and Care. Washington, Reactions, Consequences, D.C.: NationalAcademy Press, 1984. Lacey, G. Observationson Al- ofan. Journal of 260, 1972. Psychosomatic Research16:257- Levy, D.M. Psychic trauma of operationsin children. Diseases of Children69:1-25, 1945. American Journalof Logue, J.N.; Hansen, H.; and Struening,E. Some indications health effects ofa natural disaster. Public of long-term Health Reports96:67-80, 198'a. Logue, J.N.; Melick,M.E.! and Hansen, the epidemiology H. Research issuesand directions in of heakh eftedsof disasters. 162, 1981b. Epidemiological Reviews3:140- Lundlin, T. Morbidity following suddenand unexpected Journel of Psychiatry144:84-88, 1984. bereavement. British Lystad, M. Children'sissponso to disaster: Family Journal of Family implications. International Psychiatry 5:41-60,1984. MacLean, G. Psychictrauma and traumatic year old boy. Canadian neuross: Play therapywith a four Psychiatric AssociationJournal 22:71-75,1977. 269 MORBIDITY AITERDISASTER

Canadian Journal of MacLean, G. Addendum to a caseof traumatic neurosis. Psychiatry 25:506-508,1980. phenomena in a longitudinalstudy of children McFarlane, AC. Posttraumatic Academy of Child and following a natural disaster.Journal of the American Adolescent Psychiatry26:764-769, 1987. effects of a fire. McFarlane, AC., andRaphael, B.Ash Wednesday: The Australian and New ZealandJournal of Psychiatry18:341-351, 1984. effects of the war onFrench Mercier, M.H., and Despert,J.L. Psychological children. PsychosomaticMedicine 5:266-272, 1943. observations at BuffaloCreek. Newman, C.F. Childrenof disaster: Clinical American Journal ofPsychiatry 133:312-316, 1976. Cuevas, P.; Ayala, J.;and Palacios, A; Cueli, J.;Camacho, J.; Cleriga, R.; effect of masscommunication in the MexicoCity Cossoff, L. The traumatic International Jour- earthquake: Crisis interventionand preventive measures. nal of Psycho-Analysis13:279-294, 1986. witnessed their mother's Pruett, KR. Home treatmentfor two infants who murder. Journal of theAmerican Academy ofChild Psychiatry 18:647-657, 1979. Social Pynoos, R.S., and Eth, S.The child as witness tohomicide. Journal of Issues 40:87-108, 1984. of Pynoos, R.S., and Eth, S.Witness to violence: Thechild interview. Journal Psychiatry 25:306-319, 1986. the American Academy ofChild and Adolescent C.; Nader, K.; Arroyo,W.; Steinberg, A.; Eth,S.; Nunez, Pynoos, R.S.; Frederick, posttraumatic stress inschool-age F.; and Fairbanks, L.Life threat and children. Archives of GeneralPsychiatry 44:1057-1063,1987a. Later Accuracy of Pynoos, R.S., and Nader,K."Proximity to Violence and Psychiatric Associa- Recall." Symposium on theMaking of Memory, American tion Annual Meeting,Washington, 73 0., May 1986. K. 'Preliminary Resultsof a Follow-up Study of Pynoos, R.S., and Nader, presented at the American Childhood Bereavementand PTSD." Paper Psychiatric Association,Washington, D.C., October1987. sexual assaults of their Pynoos, R.S., and Nader, K.Children who witness the mothers. Journal of theAmerican Academy of Childand Adolescent Psychiatry 27(5), 1988a. Psychological first aid andtreatment approach to Pynoos, R.S., and Nader, K. Journal of children exposed to communityviolence: Research implications. Traumatic Stress 1(4):445-473,1988b. 277 270 PREVENTION OF MENTALDISORDERS

Pynoos, R.S., and Nader,X Children exposed in press. to violence. PsychiatricAnr als,

Pynoos, R.S.; Nader, K;Frederick, C.; Gonda, L.; in school age children and S tuber, M. Griefreactions followirg a sniper attackat school. lsrcwl Journalof Psychiatry and RelatedSciences 24:53-63, 1987b. Quarantelli, E.L. Anassessment of conflictingnews on mental health: The consequences of traumatic events. In: Figley, C., ed.Trauma and Its Wake. New York: Brunner/Mazel,1985. Ruben, M. Trauma in the lig: .t of clinicalexperience. The Psychoanalytic of the Child 29:369-387,If /4. Study Rubin, C.B.; Yezer, Z.M.; Hussain, Q.; andWebb, A. SummaryofMajor Natural Disaster Incic:enta in theUnited States, 1965-85. D.C.: Natural Hazards Special Pub. 17.Washington, Research & ApplicationsInformation Center, George Washington University,1986. Rutter, M. Resiliencein the face of adversity: to paychiatric disorder. Protective factors andresistance British Journal ofPsychiatry 147:598-611,1985. Schetlw, D. Preschoolers' responses to murder of theirmothers by their fathers: A study of fourcases. Bulletin American 6:45-57, 1978. Academy of Psychiatryand Law Shore, J.; Tatum, E.; and Vollmer, W.Psychiatric reactionsto disaster: The Mt. St. Helensexperience. America,4 Journal of Psychiatry143:590-595, 1986. Silber, E.; Perry, S.;and Bloch, D. Patterns of parent-childinteraction in disaster. Psychiatry21:159-167, 1958. a Silver, R.L., andWortman, C.B.Coping vAth undesirable Garber, J., and life events.In: Seligman, M.E.P., eds.Human Helplessneds. Academic Press, 1980. New York: Terr, L. Childrenof Chowchilla: Study of psychic trauma.Psychoanalytic Study of the Child34:547-623, 1979. Terr, L. Chowchilla revisited: The effectsof psychic trauma school bus kidnapping. four years aftera American Journal ofPsychiatry 140:1542-1550,1983. U.S. Government.Federal Register. Washington, D.C.: Officeof Federal Na- tional Archives andRecord Service, Nov.6, 1976. Washburn, S., and Hamburg, D. The studyof primate behavior. I., ed. PirnateBehavior. New York: In: De Vore, Holt, Rinehart andWinston, 1965. Weissman, M.M. Assessing psychiatricdisorders in children: between mothers' andchildren's reports. Discrepancies 753, 1987. Archives of GeneralPsychiatry 44:747- 278 1 MORBIDITY AFTER DISASTER 271

Wilkinson, C.B. Aftermath of a disaster:The collapse of the Hyatt Regency Hotel skywalks. American Journal ofPsychiatry 140:1134-1139, 1985. Zeanah, C.H., and Burke, G.S.A young child who witnessed hermother's murder:Therapeutic and legal considerations.American Journal of Psychotherapy 38:132-145, 1984.

2 7 ;) CHAPTER 8

Conduct Disorder: Risk Factorsand Prevention

D. R. Offord, M.D. Department of Psychiatry,McMaster University Hamilton, Ontario Introduction

The essential feature ofconduct disorder, according tothe Diagnostic and Statistical Manual of MentalDisorders (DSM III) (APA1980), is "a repetitive basic rights of others or and persistent pattern ofconduct in which either the mkjor age-appropriate societal normsor rules areviolated. The conduct is more serious than the ordinarymischief and pranks of childrenand adolescents" (APA subtypes based 1980, p. 45). DSM III dividesconduct disorder into four specific absence of adequate socialbonds and the presence orabsence on the presence or this method of identifying of aggressive antisocialbehavior. The validity of subcategories of conduct disorderis controversial (Rutter1978; APA 1980). conduct-disordered children are aheterogenous There is no disagreement that should be group; thedisagreement centers on howchildren with this disorder subclassified (Offord and Waters1983, pp. 650-682). conduct disorder is the number,type, duration, Another diagnostic issue in for a frequency or intensity, andseverity of antisocial symptoms necessary diagnosis. It is clear that mostchildren display some antisocialbehavior at some time (Gold 1966; Offordet al. 1986a). The thresholdat which the persistence of conduct disorder is and severity of antisocialsymptoms warrant a diagnosis Slight changes in the to a considerable extentarbitrary (Offord and Boyle 1986). threshold can, for instance,significantly alter prevalence ratesof conduct disorder in a cross-sectionalcommunity survey (Boyle et. al.1987a, b). have been caught The term "delinquent children"usually refers to those who evidence suggests by the police and processedthrough the court system. The represent a subgroup ofseriously conduct-disordered that delinquent children and the courts are children. The youth whobecome involved with the police among those withthe most numerous and severeantisocial behaviors (West and Farrington 1973, 1977).

This paper was prepared forthe Project Prevention initiativeof the American Academy of Child and AdolescentPsychiatry. 274 PREVENTION OF MENTALDISORDERS

Two further points aboutdiagnosis should be made. proportion, even First, a substantial a majority, of children with conductdisorder will have other psychiatric diagnoses (Offord and Waters 1983; Offordet al. 1986a). Themost common overlapping diagnosis isattention-deficit disorder with Second, the vast majority hyperactivity. of children with conductdisorder will have been diagnosed on the basis ofone source (e.g., parent basis of two or teacher) and seldomon the sources (e.g., parent and te er) (Rutter et al. 1970; 1986a). Thus, the behaviors Offord et al. contributing Lo a diagnosisof conduct disorderare usually reported with thenecessary severity in one situation home) but seldom (e.g., schoolor across situations (e.g., school andhome). Conduct-disordered childrenconstitute a heavy burden have been noted, for of suffering. They instance, to comprise thelargest singlegroup of emotionally disturbed children treatedor untreated (Rutter et al. Children and adolescents 1970; Robins 1974). with the disorder haveimpaired functioning inmany areas of their lives, suchas school and peer relationships 1983). In addition, they (Offord and Waters have a poor adultprognosis; half of them have psychosocial disturbances serious in adulthood (Robins1970; Rutter and Giller 1983). Conduct-disordered boys andgirls have equallypoor prognoses for overall adult psychiatric diagnoses. As adults, the males havemore externalizing disorders (e.g., antisocial personality,alcohol and drug abuse) females have than females, andthe more internalizing disorders(e.g., affective and (Robins 1986, anxiety disorders) pp. 385-414). Finally, this conditionimposes a heavy burdenon society because of thepersonnel andmoney involved in diagnosis, and the judicial treatment, process. Thus, because of themagnitude and seriousness problem, conduct disorder of the deserves intensive researchaimed at discovering effective interventionstrategies. The prevention ofconduct disorder isattractive for several 1987). First, by the reasons (Offord time conduct-disorderedchildren (and their diagnosed and treatment families) are is initiated, theyusually have suffereda good deal. Second, the treatmentof established cases has been disappointing(Offord and Reitsma-Street 1983; Rutterand Giller 1983). ing evidence of positive Few studies have shownconvinc- long-term effects oftreatment. Indeed, evidencesuggests that sometreatments may have had adverse effects, possiblythrough mechanisms of labeling andstigmatization of the affected 1979, pp. 627-684). child (Robins 1974, Third, even if itwere effective, treatmeut difficult and extremely would be very expensive to deliver to theaffected children andtheir families.For instance, theevidence is consistent facilities serve that existingtreatment a minority of children withpsychiatric disorders, conduct disorder, and those including who are seen in suchfacilities are notnecessarily the ones most in needof treatment (Rutteret al. 1970; Langner (,fford et al. 1987).For these et al. 1974; reasons, the search foreffective prevention programs for this condition isextremely appealing. This paper centers on the prevention of conductdisorder, that is,a reduction in the number ofnew cases of the condition. The range of topics and data, must paper, because it coversa wide be selective. Itconcentrates on major worksin 281 CONDUCT DISORDER 275 the field and does not professto be exhaustive in its coverageof the various of delinquency, using thelegal definition, is not extensive areas. Its coverage review of some because ths focus is conductdisorder. The paper begins with a general points about riskfactors and then presents,in a critical way, the evidence on risk factors forconduct disorder. It examinesthe significance of reported primary prevention programsfor conduct disorder andconcludes with suggestions for future research. Risk Factors

To qualify as a risk factor for adisorder, a variable must meetthree criteria (Department of ClinicalEpidemiology and Biostatistics 1981): 1. The presence of thevariable must be associatedwith an increased probability of disorder; 2. The occurrence of thevariable must antedate the onset ofdisorder; and 3. There must be evidencethat the variable plays a causalrole in disorder. Criterion 1 defines a correlate of, or amarker for, a disorder. Criterion2 is essential if the correlate is to qualify as arisk factor. In the absence ofcriterion 2, a correlate is not arisk faci:or but may be a part ofthe disorder or a consequence of it.Criterion 3 ensures that the observedrelationship between the variable and disorder trulyis a causal one. Evidence for thiscould include, for instance, that a changein ' severity of the variable is associatedwith a change in the incidence of disorder, orthat the mechanisms whereby the variable exerts its causative influence ondisorder are known. It also mustbe shown that the observedrelationship between the variableand disorder is not the result of another riskfactor. Table 1 pro-des data on potentialrisk factors. It rates the variablesaccord- ing to the three qualifyingcriteria for risk factors and indicateswhother the variable or its mechanisms of action arepotentially modifiable. In addition, table 1 provides data on so-calledprotective factors, which also mustfulfill the three criteria outlined above.For criterion 1, however, thevariable must be associated with a decreased ratherthan an increased probability ofdisorder. Individual Characteristics Sex Conduct disorder is more common inboys than girls. In the Isle ofWight (IOW) study (Rutter et al. 1970), forinstance, the rate of conduct disorderin 10 and 11 year olds was 6.0 percent forboys and 1.6 percent for girls, resultingin a boy:girl ratioof 3.8:1. Similarly, in the Ontari,Child Health Study (OCHS), a province-widecommunity prevalence survey, tlrates for boys and girls, ages 4 to 11, were 6.6 percent and1.8 percent respectively, for a boy:girlratio of 3.6:1 2 c 276 PREVENTION OF MENTALDISORDERS

(Offord et al. 1986a). In 12to 16 year olds in the OCHS,however, the rates in adolescent boys and girlswere 10.4 percent and 4.1 percent, resulting in a reduced boy:girl respectively, ratio of 2.5:1. This reduced ratiois largely a result of the later onset of conductdisorder in girls, the 9 to 10, and in girls it average age of onset in boys is is 12 to 13 (Offord andWaters 1983).The male preponderance of conduct disorderapplies whether the data parent, or teacher reports are based on self, or on official delinquency statistics(Rutter and Giller 1983; Offord et al. 1986a).The finding applies stronglyto school-age children and adolescents, but it is not as firmly established forpreschool children (Richman et al. 1982).

Several hypothesized mechanismsexplain the male preponderancein conduct disorder. The twomajor ones are the boys' and Jacklin 1980a, greater aggressivity (Maccoby pp. 92-100, 1980b) and the boys'apparently greater vulnerability to family upsetor marital discord (Rutter and Giller although the risk factor, 1983). Thus, sex, is not modifiable, the mechanismsby which itmay produce increased rates ofconduct disorderare potentially modifiable. Race

Much of the workon the relationship betweenrace and antisocial behavior has focused on teenage youth. Evidence from severalstudies comparing black and white youths in theUnited States indicate thatblacks have prevalence of antisocial behavior an increased and delinquency, especiallyof crimes involving violence againstpersons (Berger and Simon 1974; Hindelang et al. 1979; Elliott and Ageton 1978; Hindelang 1980). This black-whitedifference probably cannot be accounted for bypolice or court biasesor by social class variations. It is unclear, however, to what extent the differencesare explicable by other variables suchas area of reeidence, family circumstances, or livingconditions (Rutter and Giller 1983), Inthe Uaited Kingdom, findingson race and antisocial behavior show that the delinquency rate for Asianshas consistently beenequal to or sower than that of thewhite population. In blacks is substantially contrast, the arrest ratefor higher than that for whites,especially for violent crime (Rutter and Giller 1983), In a study of elementaryschool-age children, British Intitian children data indicate that West ars more disruptive in school thanothers, but this disturbance t..)ea not extend to emotional probiemsor to the honiesetting (Rutter et The behavioral disturbance al. 1974). in school may be partlya reaction to the unusually high rate of edueAtional retardation among thesechildren (Yule et al. 1975)and partly because they tendto be enrolled in schoolswith characteristics with a high frequency of associated behavior problems (Rutteret al. 1975b). The psychosocial correlates of thee('behavior problems children with were similar to those found for white one important difference: Conductdisorders were especially common in West Indian girls, comparedwith their relative infrequency girls. in white

2 C3 CONDUCT DISORDER 277

to have higher ratesof North American Indianchildren have been reported including antisocial behaviorand delinquency, fromthe emotional disturbance, al. 1981; Beiser early school yearsthrough adolescence(Beiser 1981; Green et These studies of NorthAmerican Indians, however,have and Atteave 1982). grounds. First, fewstudies have been criticized on threemajor methodological population denominator onwhich to been able to identifyclearly a well-defined prevalence of mental healthproblems (Manson andShore base eetimates of the who have a 1981). Second, thetask of distinguishingthose Indian children those who do not is extremelydifficult because culture mental disorder from (Beiser 1981; Green etal. defines whether a given setof behaviors is "deviant" when investigators haveused 1981). The thirdmethodological pitfall has arisen of data, such as recordsof mental health facilities. relatively inexpensive sources potential biases that can This procedure leads tothe introduction of important estimates (Chambers andWood- result in inaccurate andmisleading prevalence ward 1980; Meketon1983). variations in the frequencyof conduct disorder,antisocial In summary, some different racial groups. behavior, and delinquencyhave been reported for aside, it is likely thatmuch of this relationship Methodological considerations by family for by location of residenceand living conditions and can be accounted and conduct disorder. and school circumstancesthat are related to both race Body Build (Epps and Parnell 1962; Several studies ofinstitutionalized delinquents reported that delinquentboys Glueck and Glueck 1966;Gibbens 1963) have nondelinquents to have ameeomorphic or were almosttwice as likely as It is likely, however,that this finding appliesonly to the muscular build. 1983). It is probable institutionalized delinquentpopulation (Rutter and Giller that deknquents with amesomorphic build, becauseof their tougher and likely to be incarcerated.For instance, West and stros '>;er look, are more found no associa- Farefington's study (1973) of 411working-class boys in London ratio or strength of grip as tiou between delinquencyand either height-weight The British NationalSurvey of Health and measured by a dynamometer. study, found that Development (Wadsworth 1979), abirth cohort longitudinal delinquents tended to reachpuberty later thannondelinquents and to be lighter. These results werestronger for serious delin- somewhat smaller and mesomorphs. Thus, there is quents and opposite ofwhat would be expected for between body build anddelinquency inde- no firmevidence of a relationship pendent of institutionalization. Chronic Physical HealthProblems population surveys of childrenassociates chronic ill A consistent finding in problems (Rutter et al. health with pa-, ehiatricdisorders and social adjustment Walker et al. 1981; Cadman 1970; Plass and Setterwhite1975; Eisen et al, 1980; OCHS (Cadman et al. 1987),for inutance, childrenwith both et al. 1987). In the than threefold risk chronic illness and associateddisability were at a greater 278 PREVENTION OF MENTALDISORDERS

for psychiatric disordersand at considerablerisk for social adjustment problems. Children withchronic medical conditionsbut no disability considerably less risk; about were at a twofold increase in psychiatricdisorders, but little increased riskfor social adjustmentproblems, was observed. communitysurveys did not These indicate that the risk forconduct disorderwas higher than for other types ofpsychiatric disorder. The reported an delinquency literaturehas excess of physical illness, accidents,and disability quents compared with among delin- nondelinquents (Rutter andGil ler 1983); but thishas not been found in allstudies (Reitsma-Streetet al. 1986). present, is slight and This excess, when is associated withparental problems andadverse childhood experiences.In the OCHS, however, health problems the data indicate thatchronic increase the risk for psychiatricproblems, including conduct disorder, independentof the effects of other sociodemographic, parental, variables measured inthe and family domains(Offord et al. 1986b).The identification of the causalprocesses mediating the relationship chronic health problems between in children and increasedrates of conduct disorder needs further work.These causal processes may include variablessuch as low self-esteem, poorpeer relationship°, and al. 1987). poor school performance (Cadmanet Brain Damage

Children with braindamage are at increasedrisk for psychiatric the IOW study, forinstance, the rate of disorder. In higher in youngsters psychiatric disorderwas five times with cerebral palsy,epilepsy, or some other the brain stem (Rutter disorder above 1977). It is clear froma number of studies (Rutter 1981; Brown et al. 1977, 1981) that brain damageputs children at risk disorder in generalrather than for for psychiatric a specific type of disturbance.Thus, in the case of conduct disorder anddelinquency, there is brain damage other no specific association with than that present forpsychiatric disturbance The mechanisms by in general. which brain damageleads to increasedrates of conduct disorder are not clear.OCHS data suggest that the association of earlydevelop- mental problems, !ncludingbrain damage, with least to some extent, child psychiatricdisorder is, at independent of sociodemographic,family, and parental variables (Offord etal. 198(3b).The effect of the mediated, for instance, brain damage couldbe through resultingcognitive disabilityor abnormal temperament (Offord andWaters 1983). Physiological Characteristics A number of studieshave shown an association betweenantisocial behavior and autonomicreactivity (Rutter and Giller 1983). Comparedwith nondelin- quent boys, delinquentboys have been reported (Davies and Miliphant to have lower restingpulse rates 1971b; Wadsworth1976), to have reduced pulse rates . times of probablestress (Davies and 1977), and to have lower Miliphant 1971a; Westand Farrington skin conductancereactivity witha longer recovery time (Borkovec 1970; Siddleet al. 1973, 1976). There of the biological is also the suggestionthat one mechanisms involvedin a geneticpredisposition toward 2 .(,) 5 CONDUCT DISORDER 279 criminal behavior may be the responsivenessof the autonomic nervous system (ANS) (Mednick et al. 1986, pp. 83-50). Forinstance, some data suggest that adult criminal offenders exhibit lowerarousal and slower recovery of ANS than nonoffenders (Mednick and Volavka 1980, pp.86-158) and that children of criminal offenders show patterns of ANSresponsiveness that would be expected from their parents (Mednick 1977, pp. 1-8).There is also preliminary evidence that ANS responsiveness, measuredprospectively, can predict antisocial behavior in children and adolescents (Loeb andMednick 1977, pp. 245-254; Clark 1982). These autonomic features support abiological basis for the reported reduced anxiety and impaired passive avoidancelearning following punishment in antisocial persons (Rutter and Giller1983).While several studies provide supporting data for this hypothesis inadults, there is a paucity of data for children and adolescents. One study (Daviesand Miliphant 1974) of 11 to 16 year olds with "refractory"behavior difficulties supports the hypothesisof impaired passive avoidance learning. It shouldbe emphasized that the data from this area of investigation are limitedand, in many cases, preliminary. Personality and Temperament Considerable work has been done on the possibilitythat conduct disorder and delinquency are associated with preexistingenduring patterns of behavior. These patterns are at times included underthe term "temperament" and at other times under "dimensions of personality." This area of research has major methodologicalproblems. First, it is difficult to find measures of temperament orpersonality that accurately assess enduring patterns of behavior (Offord and Waters 1983;Rutter and Giller 1983). Most of the work on temperament has employed parentalpereeptions, which have only a modest agreement withexternal observers (Bates 1980). Second, it ishard to separate msaaures of abnormal temperamentfrom the beginning stages of disorder itself.This relates to the previously raised diagnosticissue of the difficulty in arriving at a threshold for conduct disorderalong the continuum of increasing frequency and severity of antisocial symptoms. Taking into account these methodological problems, thelimited data suggest that measures of difficult temperament in infancy do notsignificantly predict childhood behavior problems (Thomas et al. 1968).By age 3 or 4, however, some prediction is possible.In West and Farrington's study (Farrington1986), "trouliesomeness," 8B rated by peers and teachers, at ages 8 to10 was a strong predictor, independent of the effects of other variables, ofconvictions at 10 to 13. Similarly, other data (Graham et al. 1973) revealthat elementary school-age children v,ith difficult temperaments, as measured byparental perceptions, were far more likely than the normaltemperament group to show evidence of psychiatric disorder 1 year later.In these latter two studies, it is not clew- whether temperamental or personality characteristics arebeing measured or whether the instruments are recording the beginning stages ofdisorder. 2 .0 280 PREVENTION OF MENTAL DISORDERS

In the delinquency literature focusingon this topic, most investigations have revealed differences in personality dimensionsbetween delinquent andnon- delinquent groups (Rutter and Gil ler 1983).This literature has two added methodological problems: Many of the studies have includedonly incarceiated individuals, and the control sampleswere usually inadequately matched on key variables. Rutter and Gil ler (1983) concluded thatthere is probably a minor tendency for high measures on Eysenck's neuroticismscale (Eysenck 1977) to be associated with an increased risk of antisocialbehavior or delinquency in older adolescents and young adults butnot in younger children. Finally, there is considerable overlap betweenconduct disorder and hyperactivity (attention-deficit disorder with hyperactivity).Indeed, there is some doubt about whether they should be consideredseparate disorders (Offord and Waters 1983; Offord et al. 1986a). Thereare sparse data on differences between conduct-disordered children with and withouthyperactivity, although one study (Offord et al. 1979) of boyson probation showed that those who were definitely hyperactive tended to be more antisocial and hadpoorer school performance and lower birth weights. To what extent conduct-disorderedchildren or delinquents have preexisting patterns of reactivitycommon in hyperactive children is unknown. It has been reported, however, that10- and 11-year-old children who were diagnosed as having pervasive hyperactivity (identifiedas hyperactive in the home and in the school) hadfour times the rate of behavior disturbance 4 years later than childrenoriginally identified as hyperactive in otAy one setting (Schachar et al. 1981). The mechanisms by which children withabnormal temperament develop conduct disorder at an increased rateare beginning to be understood. Data indicate, for instance, that children of mentallyill parents and with difficult temperaments are more than twiceas likely as children with easy temperaments to be the target of parentalanger and criticism and to develop emotional and behavioral disturbance (Rutter andQuinton 1984). Childrenwere much more likely to be scapegoated in families with marked parentaldiscord. Genetic Factors

The available twin and adoptionstudies suggest a genetic traip-nissionof antisocial behavior and criminalityat the adult level (Offord and Waters 1983; Mednick et al. 1986). The data for childrenare much less convincing. In fact, a study of adolescent adopted-awayoffspring of criminalor alcoholic parents (Bohman 1978) could not correlate thesocial maladjustment of biological criminal parents and their adopted-awayoffspring. The evidence indicates that genetic factors do not playan important etiologic role in conduct disorder overall. Genetic factors, however,may be a significant causal variable in certaincases. They appear to havea stronger effect on a subgroup of criminal adultswith persistent antisocial behavior, which probablybegan in childhood in the majority of cases (Robins 1986). It isnot at all clear what is inherited. Mednicket al. (1986) argue for an altered responsivenessof the ANS, but another possibilityis certain temperamental traits (Goldsn and Gottesman 1981). CONDUCT DISORDER 281

IQ, Learning Disorders,and EducationalRetardation IQ and conduct There are data suggestiag anassociation between low et al. 1970), for instance, disorder and delinquency.In the IOW study (Rutter slightly below average IQin conduct disorder wassignificantly associated with boys, but not in girls. InWest and Farrington'sstudy (West and Farrington with delinquency, especiallypersistent delin- 1973), low IQ was associated behavior" vari- quency. Thisassociation was independentof the "troublesome able measured several yearsearlier. of conduct-disorderedchildren or delinquents is The educational retardation instance, in more striking thantheir IQ deficits (Offordand Waters 1983). For al. 1970) of 10- and11-year-old children, one-thirdof the IOW study (Butter et disorder, and one-third children severely retarded inreading showed conduct children were at least 28months retarded in their reading of conduct-disordered maintained even after (after IQ was partialled out).These associations were controlling for family size andsocial class. school perform- A mkior issue is the natureof the relationship between poor lisorders and conduct disorder ordelinquency. This relation- ance or learning Gil ler ship might take threemajor forms (Offord andWaters 1983; Rutter and 1983). develop poor school performance orlearning disorders First, children could failure as a consequenceof conduct disorder. Althoughit is clear that school and follow conduct disorder,there is little evidence to suggest can accompany significantly below IQ level) that learning disorders(where achievement is develop as a consequence ofconduct disorder. school perform The second possibility is thatthe learning disorders and poor the onset of conduct disorderand am etiologically linkedto it. A ance antedate and Mann 1972) have number of workers (Rutter etal. 1970; Frease 1972; Gold suggested this possibility. Theypostulate a causal chain inwhich early school provoke the child to failure leads to feelings of lowself-esteem, which in turn engage in antisocialbehavior in an effort to raisehis or her self-esteem and to Several lines of evidence gain a feeling of accomplishmentand confidence. support this approach,especially data that showthat improving academic performance reduces antisocialbehavior (Ayllon andRoberts 1974). The third possibility is thatthe learning disorder, or poorschool performance, factors and that one is and conduct disorder arisefrom common or overlapping not etiologically linked tothe other. For example,Sturge and Offord (Sturge that within a relatively poor 1972; Offord 1982a, pp.129-151) argued strongly urban population, the educationalretardation and the antisocialbehavior arise that the educational from common or coexistingadverse family influences and retardation itself is not causallyrelated to the antisocialbehavior. The common factors need not be limited topsychosocial ones but couldinclude temperamental characteristics (Rutter and Giller1983). 2 282 PREVENTION OF MENTAL DISORDERS

Studies that prove the relativeimportance of the second and thirdpos- sibilities as etiologic mechanismsin conduct disorder have notbeen carried out (Offord and Waters 1983; Rutterand Giller 1983). Basedon current data, it appears likely that both of these posaibilitiesare involved as etiologic factors. Psychosocial Factors Family Factors Parental Deviance

Parents with severe psych iatricimpairment, especially criminality,are found much more commonly in thef"milies of antisocial childrenor delinquents than in families of age-matched controls (Glueck and Glueck 1950; Lewisand Balla 1976; Offord 1982a; Farringtonand West 1981). It is likely thatin most cases the parent's criminalitywas preeent before the onset of conduct that it is independently related disorder and to conduct disorder and delinquency(Farrington and Weet 1981). The causal mechanisms involved in therelationship between parental deviance and conductdisorder are not primarilygenetic, but almost certainly includb factors that can both accompany parental devianceand be important in the genesis ofconduct disorder (Rutter andGiller 1983). These factois include, for example, the modeling by the child ofaggressive and violent parental behavior; poor parenting practices resulting inpoor supervision; and parental criticism and hostility,marital discord, and relianceon welfare. The relative importance of thesevariables in producing conductdisorder is not known (Offord and Waters 1983). Poor Parenting Poor parental behavior, especially harsh discipline,rejeciing attitudes, and pc, )r supervision, has been shown to antedate serious antisocialbehavior and delinquency and to have a significant, independent associationwith these outcomes (Farrington and West 1981).In addition, detailed analysesof family interactions indicate thatdisruptions in parental associated with higher monitoring and disciplineare rates of antisocial behavior(Patterson 1982, 1986, pp. 235-261). When the motherand the problem childbecome involved in extended coercive exchanges,the problem child'stroublesome behavior escalates and spreads frow involving the parent to involving thesiblings. Although theie interactional patterns betweenmother and child evidence that the:- antedato are bidirectional, there is and are usually linkedto the child'saggressive behavior. For instance, datareveal that parents' coercivebehavior is associated with an increased probabilityof the maintenance of child. Intervention aggressive behavior in the programs aimed at reducing the frequencyof these coercive interchanges have resultedin a reduction of children's (Patterson 1982). aggressive behavior CONDUCT DISORDER 283

Marital Discord Marital disharmony has been consistentlyassociated with conduct disorder (Rutter et al. 1970; Offord and Boyle 1986).The important element in the often-reported relationship between brokenhomes and antisocial behavior and delinquency clearly is not the broken home itselfbut the marital discord that precedes the break (Rutter and Gil ler 1983).Further, there is evidence that the association between marital discord or poorfamily functioning and conduct disorder is still present after controlling for theeffects of sociodemographic and parental variables (Rutter and Gil ler1983; Offord and Boyle 1987). The data suggest that marital discord oftenprecedes a child's behavior disturbance and that a le-owning of the discord is accompaniedby an improvement in the child's behavior (Rutter and Gil ler 1983). The processesinvolved in this relationship are probably similar to thosehypothesized for parental deviance (Offord and Waters 1983; Rutter and Gil ler 1983) and also mayinclude streases and adversities outside the marriage (Rutter 1987). Asnoted, the relative importance of these hypothesized causal mechanismsis unknown (Offord and Waters 1983). Large Family Size An association between large family size (usuallyfour or more children) and conduct disorder and delinquency has been reported forboys (Rutter et al. 1970; Farrington and West 1981, pp. 138-145) hut not forgirls (Jones et al. 1980). In the OCHS, however, large family size was not significantlyrelated to psychiatric disorder (Offord and Boyle 1986). West andFarrington's data (Farrington and West 1981) indicate that large family size is related toantisocial behavior and delinquency, independent of sociodemographic and parentalfactors. The mechanisms underlying this association are notclear but include at least three possibilities. The first is that in a large family,material and educational resources, especially among alreadydisadvantaged families, are stretched beyond the breaking point (Offord 1982a; Rutter and Giller1983). This contention is supported by the finding that the relationshipbetween large family size and delinquency is most marked in poor ordisorganized families (West and Farrington 1973; Wadsworth 1979). A second possibilityis that among boys, there can be a potentiation of antisocial behavior(Jones et aL 1980) or a contagion effect (Robins et al. 1975). It has been found,for insta nce, that the level of antisocial behavior among boys in a family wasassociated with the number of brothers in the family but not with the number ofsisters (Jones et al. 1980). In fact, with the number of brothers in the familyheld constant, the greater the number of sisters, the lower the level ofantisocial behavior among the brothers. The sisters appeared to suppress antisocialbehavior among their brothers. A third possibility is that the relationship betweenfamily size and antisocial behavior is cause6 by the confounding effects ofeducational retardation (Rutter and Giller 1983). This latter variable is a riskfactor for conduct disorder and delinquency and is associated with large family size.Under this

2 ) 284 PREVENTION OF MENTAL DISORDERS

explanation, the educational retardation,rather than the large familysize, is the important causal variable. Nodata are available 4- allowranking of these proposed mechanisms in order ofimportance. Community and SocioeconomicFactors Area and Socioeconomic Class Conduct disorder is muchmore common in certain comminitiee.Rutter et al. (1975a) noted, for instance,that both conduct and nearoticdisorders were twice as common ina poor inner-city area thanon the IOW. Similarly, Langner et al. (1970, pp. 185-202) found that in midtown Manhattan,the frequency of psychiatric impairmentamong children increased markedly indoscending levels of the social class ladder.He reported that thepercentages of children with impairment for high-, middle-,and low4ncomegroups were 8, 12, and 21, respectively. In the OCHS (Offordet al.1986a), therewas a significant relationship between conduct disorderand severalmeasures of socioeconomic disadvantage, including welfarestatus, subsidized housing,low income, unemployment, and overcrowding.The strongest relationshipwith conduct disorder among these variablesoccurred in thecase of welfare stet us. The delinquency literature alsosuggests at least a modestassociation between low social class and delinquency, which becomes stronger when seriousand violent crime is at issue (Rutter andGiller 1983). Why are inner cities and low socioeconomic class associatedwith conduct disorder and delinquency? The reasons are not completely understood, butthey point primarily toward factorsreflected in the under-the-roofculture of the child's home (Robins 1979). It appears that children across social classesbecome psychiatrically disturbed for similarreasons. For instance, in both the IOW and the innw-London borough (Rutter et al. 1970, 19754),nsychiatrically disturbed children lived in families withworse marriages, were more likelyto have psychiatrically disturbedparents, and came from largerfamilies than other children. All these factors were more common in the inner-Londonborough. A major reason, then, for the increased prevalence of psychiatricdisturbance in poor areas is the excess of factors thatresult in a disturbed culture or family. A under-the-roof consequence of this is that children withoutthese correlates are at low risk for developing psychiatricdisturbance regardless of live in poor or middle-class whether they areas. Another corollary of the centralimportance of the under-the-roof cultureof the home in producing that the association between psychiatric disorder is poverty and child psychiatricdisturbance is not independent of parental andfamily factors. Most studies(Robins 1979; Rutter and Giller 1983; Wordet al. 1986b), but not all support this contention. (Farrington and Weet 1981), Peer Groups There is evidence that thecharacteristics ofgroups in which childron participate may have importantbearing on their behavior pp. 263-284). More specifically, evidence (Maccoby 1986, at both the preschool(Patterson et al. I, 4, CONDUCT DISORDER 285

1967) and school-age levels (Westand Farrington 1973) showsthat the peer group may play a partin the acquisition andmaintenance of aggressive behavior. Some reports supportthe idea that if antisocial youthabandon their antisocial peer group, their levelof antisocial behavior diminishes(Knight and West 1975; Osborn and West1980). The data are sparse andsuggestive only, and thus the importance of peer groupsin the etiology of conduct disorderis not well established. Schools There is evidence that the quality ofLigh schools can have dramatic effects on students'academic achievement and behaviorduring and after their school careers; these effects appearto be independent of thestudents' intake characteristics, both academic and behavioral(Rutter et al. 1979; Rutter 1983). The participating schools varied greatly on anumber of measures of students' success, includingbehavior in the classroom, rates ofattendance, level of delinquency, examination success,college entrance, and employmentduring and after leaving school. The factorsfostering students' success were numerous but did not include severalvariables commonly considered important,for example, resources, size of school, sizeof classroom, and amount ofpunishment. The positive factors includedbalanced intake of students withdiffering scholastic ability, emphasis on academicachievement, reasonable discipline, effective techniques of classroom management,pleasant working conditions for the student, ample opportunities for studentparticipation and responsibilities, and excellent relationships between theadministration and the staff.Although certain high schools appear to increasethe risk of antisocial behavior, it is not clear to what extent a "bad" school canbe changed into a "good" one. Television Evidence suggests that televisionviewing has an impact on childrens' attitudes and behaviors, and that in a certain group,probably small, it can play a significant role inpromoting aggressive behavior (Rutterand Giller 1983; Eron and Huesmann 1986, pp. 285-314).Specific factors p,re associated with increased television viewing in children. Forinstance, the viewing of violence on television can lead to heightenedaggressiveness, which in turn appears to increase viewing of television violence.In addition, children who are less popular and do poorly in school tend to watch moretelevision, thus viewing more violence (Eron and Huesmann 1986).This chain suggests that the effects of television violence will be most marked onchildren who already are at increased risk for antisocial behavior. Physical Environment Some data suggest that the physical designfeatures of buildings and the amount of surveillance may affectthe amount of antisocial behaviorin a community (Newman 1973; Rutter andGiller 1983).Results indicate that reduced vandalism is associated with both greatersurveillance and a physical design that provides geographic areas forwhich the residents feel responsible

21ti2 286 PREVENTION OF MENTAL DISORDERS

(Newman 1973; Wilson 1978). Inthese studies, the residents' characteristics (e.g., number of one-parent families, numberof children) weremore important than surveillanceor design variables in predicting vandalism.However, situational and physical enrichmentfactors appear to havean independent effect in the production of antisocialbehavior, but their strength is notknown. Further, it is not clear whether modifyingthese factors reduces antisocialbehavior or simply displaces it toa different location or to a different type ofbehavior. Labeling

In labeling theory (Scheff 1966;Lemert 1967), society'srew tion to an actual or primary behavioral deviation, rather thancorrecting it, provokesa stronger and more enduring secondary deviation.Evidence suggests that beingap- prehended by police escalatesantisocial behavior in youth withconduct disorder (Gold 1970; Farrington 1977; Farrington et al. 1978). The mechanismsinvolved are not well understood, but the apprehensionitself appears to lead to increased antiauthority attitudee and perhapsto a change toward more-deviantpeer group. To what extent labeling is involvedin producing conduct disorder unknown. is Protective Factors

The term "protective factors"refers to variables thatimprove a person's response to an environment that places thatperson at high risk for disorder (Rutter 1985). This interestgrew out of the observation thata significant proportion of children from high-riskbackgrounds appeared to haverelatively good psychosocial outcomes.Those in this good outcomegroup originally we -e referred to as "invulnerable children" (Anthony 1974);more recently, they hale been called "resilient children"(Masten and Garmezy 1985,pp. 1-52). The number of children fromhigh-risk backgrounds whoare free of adjustment problems is quite smallonce measurement difficultiesare taken into account and a wide range ofpsychiatric outcomes is considered(West and Farrington 1973; Weet 1982). The workon protective factors is preliminary, anda number of potential candidatesare now discussed (Rutter and Giller 1988). Characteristics of the SocialGroup Some data suggeet thata voluntary change in peergroup may prevent antisocial behavior or delinquency.In West and Farrington's(1973) study, persistent recidivists (thosewho had court convictions inlate adolescence) more likely than temporary recidivists were to continue going around withan all- male group (Knight and West1975; Osborn and West 1980).Causal inferences are made uncertain by the difficulty inmatching the family backgrounds recidivists end nonrecidivist3 of the and the impossibility of knowingwhether the peer group change preceded or followed thechange in the level of antisocialbehavior. Additional data suggestthat the association between and low delinquency good parental supervision rates in high-risk environmentsmay be mediated in part by effective control of peer group activities (Rutter andGiller 1983). 2 9 3 CONDUCT DISORDER 287

Other evidence supporting the possible protective function of a change in peer group comes from Rutter's data(Rutter et al. 1979) indicating that the differences among the rates of antisocialbehavior of students in various high schools may have been dependent on the characteristicsof the groups of students at intake. These peer group characteristics included amount of behavioral deviancy, intellectual level, and ethnic background.Similarly, it may be that the reduction in antisocial activities following school dropout is partly due to a change in peer group (Elliott and Voss 1974; Bachman et al. 1978). Finally, there is information to suggest that changes in living situations can reduce delinquency, due in part, perhaps, to a change for the better in the peer group (West 1982). The data on this issue, while consistent, are sparse and focus primarily on adolescents who already have serious antisocial behavior. The data currently cannot point to a clear causal connection between the presence of the hypothesized protective factor and the reduction in the incidence of conduct disorder. Employment There is a consistent association between unemployment and delinquency (Wootton 1959). The question at issue is whether holding a job reduces the risk for deviancy in persons at increased risk for antisocial behavior. Longitudinal studies focusing on this issue are almost nonexistent (Rutter and Giller 1983). However, Bachman's longitudinal study (Bachman et al. 1978) of male adolescents in the United States did show a weak association between employment and self-reported antisocial behavior, controlling for previous level of antisocial symptoms. Measurement and matching difficulties in the study make the validity of the results uncertain. In any case, this hypothesized protective factor could not be of major importance in conduct disorder because of the age restrictions on the diagnosis. Change of Circumstances Do improvements in social circumstances lead to a reduction in antisocial behavior? There is evidence that among children who were separated from their parents early, those who subsequently lived in discord-free homes had fewer conduct disorders than those who lived in homes with disharmony (Rutter 1971). Similarly, it has been shown that the diminishment of the disturbed behavior of children of divorcing parents is related to the extent to which the divorce improves family relationships (Heatherington et al. 1978; Wallerstein and Kelly 1980). Again, these findings apply primarily to children who already have a good deal of disturbance.It has not oeen shown directly that changes in circumstances prevent new cases of conduct disorder among children whose original environment would place them at increased risk for this condition. 288 PREVENTION OF MENTAL DISORDERS

Good Relationship

It has been shown that children from discordant homeswho have a good relationship with one parentare less likely to develop conduct disorders than those from the same circumstances who donot have this good parental relation- (kip (Rutter and Gil ler 1983). The results ofseveral other studies suggest that a good relationship with an adult inside or outside the familymay reduce deviant behavior in children from high-risk families(Rutter and Giller 1983; Reitsma- Street et al. 1985). Compensatory Good Experiences The extent to which good experiences outsidethe home can reduce antisocial behavior in children from high-risk familiesis not known. The literature suggests good (Rutter et al. 1979), good nonacadamic skilldevelopment programs (Jones and Offord 1989), school competence (Rutteret al. 1976c; Stiffman et al. 1986; Rae-Grantet al. 1986), and activity participation and competence (Rae-Grant et al. 1986; Stillmanet al. 1986) as candidates for protective factors. In each of thesecases, the data are encouraging but not conclusive. In the OCHS) for instance,among children ages 4 to 11, being a good student and participating in sport andnonsport activities had an independent effect on lowering the prevalence of psychiatricdisorder even after controlling for level of risk (Rae-Grant et al. 1986). In the12- to 16-year-old group, the findinp were confirmed only for the variableof being a good student. Similarly, among children whose parents were emotionally ill, activityparticipation and competence made important independent contributionsto the prediction of ch ild behavior problems (Stiffman et al. 1986). In a recently completed demonstrationproject, children living in a publicly supported housing complex whowere exposed to a nonschool skill development program showed lower levels of antisocial behavior,as indicated by community measures such as vandalism and police calls, than children livingin a comparable housing complex who didnot receive the intervention (Jones and Offord 1989). In addition, the potential savingfrom this reduction in antisocial behavior in the community greatlyexceeded the cost of theprogram. Hypothesized mechanisms accountingfor this beneficial effect include improved self-esteem and better timeuse among the children, as well as increased opportunities for them to identify withcompetent non-antisocial adults. A major issue with all these data is theextent to which the compensatory experience is known to antedate the onset ofthe disorder. Coping Mechanisms

Coping has been conceptualizedto include the individual's attempts bothto directly alter the threatening conditionsand to change the appraisal of themto avoid feeling threatened (Rutter andGiller 1983). The mostcommon school-age child-focused programs aimed at improvingcoping are affective education and social problem-solving training .(Durlak1986). The aim of affective education CONDUCT DISORDER 289 is to improve children's emotional and social adjustmentby increasing their awareness and acceptance of the ways inwhich feelings, attitudes, and value*, influence interpersonal behavior (Medway and Smith 1978;Baskin and Hess 1980). It is hoped that gains in these areas will result in improvementsin social and emotional adjustment. The results of outcome studieshave been disappointing; the improvements in adjustment have been confined,in general, to self-report instruments with no extension to more objective measures(Word 1987). Social problem solving contends that interpersonal cognitiveproblem-solving competence allows the growing child to relate to peers and adults inhealthy and satisfying ways (Spivack and Shure 1974; Spivack et al. 1976). Thecontention is that this should promote good adjustment and prevent psychosocialdifficul- ties, including conduct disorder. Although children can be taught toimprove cognitive problem-solving skills, it has not been shown consistmtly thatim- provement in these skills results in improved emotional or socialadjustment (Durlak 1985; Offord 1987). In summary, the improved coping techniques that have been taughthave not been shown to improve the childrens psychosocial adjustment. Possibilities for Prevention

The role of epidemiology in prevention is to identify risk factors and then to postulate causal chains leading to disorder (Robins 1978).Subsequently, programs aimed at breaking the causalchain should be launched. Then it should be determined, through rigorous evaluation, whether the incidence of disorder is reduced as a result of the intervention. The formation of causal chains involving risk and protective factors is a complex process. Some factors act in an additive fashion. The effect of the etiologic factors together is equal to the sum of their effects in isolation. Brain damage and an adverse environment are examples of additive factors in producing psychiatric disturbance in the child (Rutter 1977).In other cases, the presence of one risk factor may potentiate thenegative effect of another. For instance, the effect of low birth weight on intelligence is most marked in children who come from particularly disadvantaged social circumstances (Sameroff and Chandler 1975, pp. 187-244).Another mode of interaction among etiologic factors is transactional, in wh ich one factor increases the likelihood of experiencing another. For example, in families where scapegoating occurs, the child with a difficult temperament is most likely to be the onewho is scapegoated (Graham et al. 1973). Finally, there is evidence that in some cases factors can interact in a nonadditive or interactive fashion (Rutter andGiller 1983). Here the sum of the effects of the joint factors is different from the effects of the factors considered individuaily. A further complication in this area is that these interactions among factors vary with the child's stage of developmentand can be understood only in a limited way by cross-sectional studies (Rutter 1985). 290 PREVENTION OF MENTAL DISORDERS

As shown in table 1, fewer than half thecorrelates qualify without reservation as risk or protective factors according to the criteria ofbeing associated withan increased or decreased prevalence ofconduct disorder, antedating theonset of the condition, and havinga causal relationship with tho disorder. None of the potential protective factors fulfillsall three criteria. Of thoso variablesthat qualify or may qualifyas risk factors, some (sex, race, and genetic factors,for example) clearly are not modifiable.These variables, along with correlatesthat have not been shown to be riskfactors, cannot be candidates for changein primary preventionprograms but can act as markers indkatingpopulations with an increased prevalence ofconduct disorder. As mentioned earlier,however, although a risk factor such as sex is not modifiable,the mechanisms by which male sex results inan increased rate of conduct disorder maybeamenable tu intervention. One other pointshould be kept in mind when choosingpar- ticular risk factorsas critical variables in a primary preventionproject. Some risk factors, such as brain damage,are so rare that even f their frequency could be lessened, this reduction wouldnot be expocted to preventa sizable number of conduct disorders. The attributablerisk (Lilienfeld 1976) associatedwith brain damage is small.

Some of the variables thatappear to qualify as risk fat:tors and showsome promise of being able to be modifiedthemselves or through theirmechanisms include sex, race, chronic physicalillness, personality and temperament,educational retardation, parental deviance,poor parenting, marital discord, large family size, socioeconomic nlass,schools, television, and the physicalenviron- ment and situational effects (table 1).There is little direct evidencethat reducing the frequencyor severity of these variables actually reducesthe incidence of conduct disorder.Preliminary evidence for this existsfor marital discord, poor parenting, andeducational retardation (Rutter andGiller 1983; Berrueta-Clement et al. 1984; Offord1987), but few data of thistype are available for other risk factots.

Recent reviews of the primaryprevention of conduct disorder(Joffe and Offord 1987) and emotional andbehavioral disorders (Offord 1987)reveal encouraging leads but few well-establishedfacts.The relevant literature reviewed in the remainder of this paper is organized accordingto the three types of programs directed at specifictarget groups: milestone, high-risk,and communitywide. Milestoneprograms focus on children at a givenage or developmental level.High-risk programs restrict themselvesto certain groups of children thought to be at increasedrisk for conduct disorder.Communitywide programs focus, not on particular children, buton a population of childrenfor example, those living ina specified geographic areaor attending a particular school. The most promisingprograms are those baaed on the risk factors and mechanisms of causation for whichthe most solid data exist (seetabk 1). Table 1. Ratings of Potential Risk Factors of Conduct Disorderon Selected Variables

Variable or Its Causal Mechanisms of Action Strength of Relationship Antedates Relationsh ip Are Potentially Variables with Disorier Disorder with Disorder Modifiable

(A) Individual Characterist

(i) Sex Boy:girl ratio 2-4:1 Yes Yes Yes

(ii) Race In some cases 1.5-2:1 Yes Probably not Yes (iii) Body Build Mesomorphy Yes No Not applicable 2:1 in institutionalized delinquents (iv) Chronic Physical 2-3:1 Yes, probably Yee Yes Illness in majority of cases (v) Brain Damage 2-5:1 Yea, usually Yes Yes (vi) Physiological Not firmly established Yes, at least Unknown Unknown Characteristics in some cues (vii) Personality & Not firmly established Yes Yes Yes Temperament TO) le 1. Ratings of Potential RiskFactors of Conduct Disorderon Selected Variables (continued)

Variable or Its Causal Strength of Relationship Mechanisms of Action Antedates Relationship Ale Potentially Variable3 with Disorder Disorder with Disorder Modifiable

(viii) Genetic Factors Not strong Yes Yes Unknown (ix) IQ, Learning Reading retardation Probably in Yes Yes Disorders & 2-3:1 some cases Educational Retardation

(B) Psychosocial Factoys (1) ram& F adroit% (a) Parental Parental criminality Yes Yes Yes Deviance 2:1 (b) Poor Parenting 2:1 Yes Yee Yes (e) Marital 1.5 - 2:1 Yes Yes Discord Yes (d) Large Family 1.0 - 1.5:1 Yes, usually Size Yes Yes

(ii) Community aud Socioeconomic isAmt (a) Area and 2:1 Yes Mixed data Socioeconomic Yes Class 299 Table I. Ratings of Potential Risk Factors of Conduct Disoryler on Selected Variables (continued)

Variable or Its Causal Mechanisms of Action Strength of Relat ionship Antedates Relationship Are Potentially Variables with Disorder Disorder with Disorder Modifiable

(b) Peer Groupe Unknown Unknown Unknown Yes (c) Schools 2:1 Yes Yee Yes (d) Television Unknowl Yes Prolo...bly Yes (e) Physical Unknown Yes Probably Yes Environment G at Situational Effects Ct (1) Labeling Unknown Unknown Unknown Yee El (C) Protective Factors (1) Characterintics Unknown Unknown Unknown Yes of Social Group C (ii) Employment Unknown Unknown Unknown Yee t*. (iii) Change in Social Unknown Unknown Unknown Yes k Circumstances (iv) One Good Unknown Yee, p vbably Unknown Yee Relationship in some cases tv, Compensating 2-3:1 Unknown Yes Yes Good aperiencee (vi) Coping Unknown Yes Unknown Yes Mechanisms

300 294 PREVEMION OF MENTAL DISORDERS

Milestone Programs Interventions aimed at reducing mental retardation (Ramey et al. 1984), educational retardation in the early school years (Gray and Klaus 1970), and brain damage in the perinatal period (Magrab et al. 1984, pp. 43-73) would, if effective, be expected to reduce the incidence of conduct disorder. The Perry Preschool Project (Berrueta-Clement et al. 1984), a well-designed preschool educational program, provides some evidence of a teduction in delinquency and self-report antisocial symptoms in the experimental group, compared with the control group, up through age 19. The results, though somewhat inconsistent, are promising enough to warrant attempts at replication. Preschool programs that result in improved behavior and school prformance in the early school grades might be considered successful in preventing conduct disorder because behavioral and scholastic difficulties in the early school years are themselves indicators of increased risk for conduct disorder. Based on the data in hand, first-rate preschool programs are a promising primary prevention intervention for conduct disorder. Clearly, further randomized trials tire needed to test the effectiveness of various types of preschool programs in reducing childhood educational and behavioral morbidity. The development of effective programs to improve parenting skills and to reduce marital discord would be expected to reduce the incidence of conduct disorder in children and adolescents. Evidence indicates that the programs should focus on reducing parental criticism and anger toward the child (Far- rington and West 1981; Rutter and Quir.ton 1984) and on lessening the amount of coercive interaction between the parent and child (Patterson 1982, 1986). Any promising programs in this area should be evaluated, preferably through a randomized controlled trial. Work should continue on evaluating the effectiveness of affective education and social problem-solving training programs for school-age children. As noted earlier, thare is no conclusive evidence that these programs reduce levels of antisocial behavior (Offord 1987). Finally, a reduction in motor vehicle accidents and unwanted births could be expected to decrease incidence of conduct disorder. A major sequelae of accidents is head injury, which in turn is associated with an increased frequency of psychiatric disorder, including conduct disorder (Offord and Waters 1983). Children who are unwanted at birth are at increased risk for psychiatric disorder (Rutter 1982). Unfortunately, no firm evidence of effective interventions in these areas currently exists (Health and Welfare Canada 1980; Rutter 1982; Roberts et al. 1984, pp. 173-199).In addition, the number of cases of conduct disorder attributable to these risk factors is small. High-risk Programs All the programs mentioned in the previous Jection might be more efficient if they were launched in populations at high risk for conduct disorder. For instance, preschool programs could focus on poor, urban children, or parenting

3 1 CONDUCT DISORDER 295

In programs couldbe directed at parents withmental illness or criminality. either case, the target condition,conduct disorder, will have amuch higher frequency than in the generalpopulation. Other high-risk groups includecLcidren of divorce and childrenwith chronic medical illness. Children ofdivorce appear to be at increasedrisk for emotional Wallerstein and Kelly and behavioral disorders(Heatherington et al. 1978; 1980). Several intervention programsfor these children have beendescribed, but few have hem evaluated(Robson and Rae-Grant 1987).One of the most promising programs is the DivorceAdjustment Project (Stolberg andGarrison and 1985), which provides a 12-sessionpsychoeducational program for children a 12-week support groupfor parents.Increased self-esteem and prosocial behavior among the children andimproved adjuatment among the parentshave been reported. Designweaknesses in this project makeit imperative that replications of this or similar programsbe carried out (Offord 1987). For children with chronic medicalillness, three randomizedcontrolled trials are aimed atreducing the prevalence of emotionaland behavioral problems (Cadman et al. 1986). All three reportpositive results; however, one studyhad serious methodological flaws,the results of the second were notstatistically significant, and the data from thethhd were preliminary. Obviously,there is a need forfurther work. One other study deservesmention. In a randomized controlledtrial, client- centered family intervention infamilies with a delinquent childreduced by half the subsequent court contacts ofsiblings (Klein and Alexander1977). Evidence indicates that the interventionreduced deviant behavior in a group atincreased risk for conduct disorder, namelythe siblings of delinquents. Finally, it should be kept inmind that while the high-risk strategyin primary prevention has the potential advantageof greater efficiency, it alsoincludes the possibility of incorrect labeling andstigmatization (Afford 1987). The potential benefits and harms of such anapproach should be monitored in anyevaluation procedure. Communitywide Programs An advantage of this type of programis that individual children are not singled out, and thus labeling andstigmatization are avoided (Offord 1982b). Th two major focuses of these arethe school and the community atlarge. The school is a promising settingfor primary prevention programsfor conduct disorder. The balancing of children withdifferent ability levels in each school and superior classroom organizationand teacher practices appear tobenefit students in both the behavioral and theacademic domains (Rutter et al. 1970; Offord 1987). Improved schools,characterized by many of the factors present in the most successful secondaryschools in Rutter's study, would almost certainly reduce the incidence ofconduct disorder. Data are needed in two areas. First, it mustbe determined at the secondaryschool level whether an 310 2 296 PREVENTION OF MENTAL DISORDERS

intervention strategycan turn a "bad" school into a "good"one. Second, information is Leeded about whether thereare differences among elementary schools similar to those in secondary schooht.That is, is there evidence thatspecific elementary schools havea greater beneficial effect on academicperformance and deviant behavior, independentof the students' intake characteristics?If so, the mechanisms involved in producing this effectshould be identified. Third, interventions based on these mechanismsshould be attempted to determine whether a relatively ineffective elementaryschool can be turned intoa relatively effective one. At the community level, furtherwork needs to be doneon testing the effectiveness of various physical designsof buildings and other strategiesaimed at diminishing the opportunitiesfor antisocial behavior. Differentsurveillance patterns could be evaluated and thebenefits of grouping economicallydisad- vantaged families versus interspersing them among middle-classfamilies could be compared.

Evidence is neededon the extent to which programs in thecommunity, but not directly involving the family,can reduce antisocial behavior in childrenat increased risk, ltappears that schools can have this effect. Asnoted earlier, data indicate thata nonschool skill developmentprogram can reduce community rates of antisocial behavioramong children and adolescents living ina public housing complex (Jones andOfford 1989). Moreprograms of this nature need to be launched and evaluated.One advantage of theseprograms is that they require little involvementof the parents other than providingpermission for their children's participation.A program dependenton the involvement of the parents of conduct-disorderedchildren may fail becauseof inadequate parental cooperation.

Finally, further work needsto be done on the effectson antisocial behavior of placing different types ofchildren together ingroups. For instance, some evidence indicates that thepresence of healthy girls can havea suppreesion effect en the level of antisocialbehavior among boys (Joneset al. 1980; Offord and Waters 1983). It may be that groupingantisocial boyo withno female presence results in an escalation of the boys'deviant behavior.Different grouping strategies involving childrenof both sexes and with differingdegrees and types of behavior problemsshould be implemented andevaluated in a variety of settings, includingfrom community schools,recreation groups, and institutional settings.

Conclusion

In summary, while there isa paucity of proven, effectiveprograms for the prevention of conduct disorder, many leads are worthy of pursuit.The most promising include enriched preschoolprograms; interventions aimed at improving parenting skills or decreasing marital discord; programs targetedon specific high-risk groups, suchas children with chronic health problemsor the offspring 303 CONDUCT DISORDER 297 of divorcirg parents; strategieswith the goal of improving schools;community interventions aimed at enrichingthe lives of and providing competenciesto children at increased risk for conductdisorder; and differing the compositionof groups of childrenbased on, for instance, the sex ratio andthe relative number of children with deviant behaviorpatterns. Research is urgently needed in threerelated areas.First, work should continue on learning more preciselywhich correlates actually qualify asrisk or protective factors according to thethree criteria in table 1. Second, morework needs to be done on the nature of causalchains linking risk and protectivefactors with disorders. Third, primary prevention programsaimed at reducing important risk factors (or their harmfulsequelae) should be launched andrigorously evaluated.It should be kept in mind that theprevention program need not directly attack the risk or protectivefactor but may interrupt the causal chain anywhere between the occurrencn of therisk or protective factor and the onset of the disorder. More demonstration projects areneeded. They will have to belaunched in the face of incomplete knowledge aboutetiology and causal chains. Ajudgment will have to be made that thehypothesized causal chain underlying theproject is likely both to be true and toaccount etiologically for a sizablesubgroup of conduct disorders. Finally, it willhave to be realized that primary prevention research in this area is difficult, not onlybecause of the methodological problems of treatment research, but also foranother reason: The incidence of conduct disorder will be relatively low in communitypopulations and, perhaps, even in high-risk groups. As a result, samplesizes must be extremely large to show, in a reliable manner,clinically important differences in outcomesattributable to the interventions. Thus, preventionstudies in this area, if properly carried out, will usually be extensive, time-consuming,and costly. The potential payoff, however, is worth this effort and cost. Acimowledgments

The author thanks the following for theirhelpful criticisms of earlier drafts of this paper: Prof. M. Rutter, Mr. M.Boyle, and Drs. J. Fleming, N. Price- Munn, and P. Szatmari. 298 PREVENTION OF MENTAL DISORDERS

References

American Psychiatric Association. Diagnosticand Statistical Manual ofMental Disorders. 3d ed. Washington, D.C.:American Psychiatric Association,1980. Anthony, E.J. The syndrome of thepsychologically invulnerable child.In: Anthony, E.J., and Koupernik, C., eds.The Child and His Family:Children at Psychiatric Risk. Vol. 3. New YorkWiley, 1974. Ayllon, T., and Roberts, M.D.Eliminating discipline problems bystrengthening academic performance. Journal ofApplied Behavioural Analyees7:71-76, 1974. Bachman, J.G.; O'Malley, P.M.; andJohnston, J. Adolescence to adulthood Change and stability in the lives ofyoung men. Youth in Transition. Vol. VI. Ann Arbor, Mich.: Institute forSocial Research, University ofMichigan Press, 1978.

Baskin, E.J., and Hess, R.D.Does affective education work? Ariew of seven programs. Journal of School Psychology 18:40-50,1980. Bates, J.E. The concept of difficulttemperament. Merrill Palmer Quarterly 26:299-319, 1980.

Beiser, M. Mental health of AmericanIndian and Alaska Nativechildren. White Cloud Journal 2:37-47, 1981.

Beiser, M., and Atteave, C.L.Mental disorderamong Native American children: Rates and risk periods for enteringtreatment. American Jour. al ofPsychiatry 139:193-198, 1982.

Berger, A.S., and Simon, W. Blackfamilies and the Moynihanreport: A research evaluation. Social Problems22:145-161, 1974.

Berrueta-Clement, J.R.; Schweinhart, L.J.; Barnett, W.S.; Epstein,A.S.; and Weikart, D.P. Changed Lives:The Effects of the Perry PreschoolProgram on Youths Through Age 19. Ypsilanti,Mich.: High Scope, 1984. Bohman, M. Some genetic aspects of alcoholism and criminality.Archives of General Psychiatry 35:269-276,1978.

Borkovec, T.D. Autonomicreactivity V sensory stimulationin psychopathic, neurotic and normal juveniledelinquents. Journal ofConsulting and Clinical Psychology 35:217-222, 1970.

Boyle, M.H.; Byles, J.k; Offord,D.R.; Crawford, J.W.; Szatmari,P.; Rae-Grant, NI; Links, P.S.; Cadman,D.T.; and Stenerson, P.E.Ontario Child Health Study: Measurement of disorder.Submitted for publication, 1987a.

3Or' CONDUCT DISORDER 299

Boyle, M.H.; Offord, D.R; Hofmann, H.F.; Catlin, G.P.; By les, J.A.; Cadman, D.T.; Crawford, J.W.; Links, P.S.; Rae-Grant, N.I.; and Szatmari, P. Ontario Child Health Study: Methodology. Archives of General Psychiatry, 1987b. Brown, G.; Chadwick, 0.; Shaffer, D.; Rutter, M.; and Traub, M. A prospective study of children with head injuries: III. Psychiatric sequelae. Psychological Medicine 11:63-78, 1981. Cadman, D.; Boyle, M.; Szatmari, P.; and Offord, D.R. Chronic illness, disability and social well-being: Findings of tho Ontario Child Health Study. Pediatrics, 1987. Jadman, D.; Rosenbaum, P.; and Pettingill, P. Prevention of emotional behavioral and family problems of children with chronic medical illness. Journal of Preventive Psychiatry 3:167-190, 1986. Chambers, L.W., and Woodward, CA Guide to Health Needs Assessment. Ottawa: Canadian Public Health Association, 1980. Clark, F. Relationship of electrodermal activity at age 3 to aggression at age 9: A study of physiologic substrate of temperament. Los Angeles: University of Southern California Library, 1982. Davies, J.G.V., and Miliphant, R. Autonomic responses of male adolescents exhibiting refractory behaviour in school. Journal of Child Psychology and Psychiatty 12:115-238, 1971a. Davies, J.G.V., and Miliphant, R. Refractory behaviour at school in normal adolescent males in relation to psychopathy and early experience. Journal of Child Psychology and Psychiatry 12:35-42, 197 lb. Davies, J.G.V., and Miliphant, R. Refractory behaviour in school and avoidance learning. Journal of Child Psychology and Psychiatry 15:23-32, 1974. Department of Clinical Epidemiology and Biostatistics, McMaster University Health Sciences Centre. How to read clinical journals:IV. To determine etiology and causation. Journal of the Canadian Medical Association 124:985- 990, 1981. Durlak, J. Primary prevention of school maladjustment. Journal of Consulting and Clinical Psychology 53:623-630, 1985. Eisen M.; Donald, C.A.; Ware, J.E.; and Brook, R.H. Conceptualization and Measurement of Health for Children in the Health Insurance Study. R-2313- HEW. Santa Monica: The Rand Corporation, 1980. Elliott, D.S., and Ageton, S.S. Reconciling race and class differences in self- reported and official estimates of delinquency. American Sociological Review 45:95-110, 1980.

3 0 fp 300 PREVENTION OF MENTAL DISORDERS

Elliott D.S., and Voss, H.L. Delinquencyand Dropout. London: Lexington Books, 1974.

Epps, P., and Parnell, R.W. Physiqueand temperament ofwomen delinquents compared with women undergraduates.British Journal of Medical Psychology 25:249-255, 1952.

Eron, L.D., and Huesmann, L.R. Therole of television in the developmentof prosocial and antisocial behaviour. In:Olweus, D.; Block, J.; and Radke-Yar- row, M., eds. Development of Antisocial and PrvsocialBehavior. New York: Academic Press, 1986.

Eysenck, H.J. Crime and Personality.London: Paladin, 1977. Farrington, D.P. The effects of publiclabelling. British Journal of Criminology 17:112-125, 1977.

Farrington, D.P. Stepping stonesto adult criminal careers. In: Olweus,D.; Block, J.; and Yarrow, MR, eds.Dz.velopment of Antisocial and Prosocial Behavior. New York: Acadero&Press, 11036.

Farrington, D.P.; Osborn, S.; and West,D.J. The persistence of labellingeffects. British Journal of Criminology18:277-284, 1978.

Farrington, D.P., and West, D.J.The Cambridge study in delinquentdevelop- ment (United Kingdom). In: Mednick,S.A., and Baert, A.B., eds. Prospective Longitudinal Research: An EmpiricalBasis for the Primary Preventionof Psychosocial Disorders. New York.:Oxford University Press, 1981. Freese, D.E. The schools, self-conceptand juvenile delinquency. BritishJour- nal of Criminology 12:133-146,1972.

Gibbens, T.C.N. Psychiatric Studiesof Borstal Lads. London: OxfordUniver- sity Press, 1963.

Glueck, S., and Glueck, E.Unravelling Juvenile Delinquency.Cambridge: Harvard University Press, 1950.

Glueck, S., and Glueck, E. Physiqueand Delinquency. New York:Harper and Row, 1956.

Gold, M. Undetected delinquentbehavior. Journal of Researchin Crime and Delinquency 3:27-46, 1966.

Gold, M. Delinquent Behaviorin an American City. Belmont,Mass.: Brooks Cole, 1970.

Gold, M., and Mann, D.Delinquency as a defense. AmericanJournal of Orthopsychiatry 42:463-479, 1972. 307 CONDUCT DISORDER 301

Goldsmith, H.H., and Gottesman,LI. Origins of variation inbehavioral style: A longitudinal study oftemperament in young twins.Child Development 52:91-103, 1981. Graham, P.; Rutter, M.; andGeorge, S.Temperamental characteristics of behavior disorders in children.American Journal ofOrthopsychiatry 43:328- 339, 1973. Gray, S.W., and Klaus, RA Theearly training project: Aseventh-year report. Child Development 41:909-924,1970. Green, B.E.; Sack, W.H.; andPambrun, A A review of childpsychiatric epidemiology with special referenceto American Indian andAlaska Native children. White Cloud Journal2:22-36, 1981. Report of a Task Health and Welfare Canada."Periodic Health Examination." Supply and Force to the Conference of DeputyMinisters of Health. Minister of Services, Canada, 1980. Heatherington, E.M.; Cox, M.; andCox, R. The aftermath of divorce.In: Stevens, J.H., and Matthews, M.,eds. Mother-Child, Father-ChildRelations. Washington, D.C.: NationalAssociation for the Education ofYoung Children, 1978. Hindelang, M.Race and involvement incommon-law personal crimes. American Sociological Review 43:93-109,1978. Hindelang, M.J.; Hirschi, T.; andWeis, J.G. Correlates of delinquency:The illusion of discrepancy betweenstif-report and official measures.American Sociological Review 44:995-1014, 1979. Joffe, R.T., and Offord, D.R. Theprimary prevention of antisocialbehavior. Journal of Preventive Psychiatry,in press, 1987. Jones, M.B., and Offord, D.R. Preventionof antisocial behavior in poorchildren by non-school skill development.Submitted for publication, 1989. Jones, M.B.; Offord, D.R.; andAbrams, N. Brothers, sisters andantisocial behavior. British Journal ofPsychiatry 136:139-145, 1980. Klein, N.G., and Alexander, J.F.Impact of family systemsintervention on recidivism and sibling delinquency:A model of primary preventionand program evaluations,Journal of Consulting and ClinicalPsychology 45:469-474, 1971. Knight, B.J., and West, D.J. Temporaryand continuing delinquency.British Journal of Criminology 15:43.50,1975. 3 302 PREVENTION OF MENTALDISORDERS

Langner, T.S.; Gersten, J.C.;Greene, E.L.; Eisenberg, McCarthy, ED. Treatment J.G.; Herson, J.H.; and of psychological disordersamong urban children. Journal of Consulting andClinical Psychology42:170-179, 1974. Langner, T.S.; Herson, J.H.;Greene, E.L.; Jameson, Children of the city: J.D.; and Goff, J.A. Affluence, poverty and mentalhealth. In: Allen, V.L.,ed. Psychological Factors inPoverty. Chicago: Markham,1970. Lemert, E.M. HumanDeviance, Social Problemsand Social Control, Englewood Cliffs, N.J.:Prentice-Hall, 1967. Lewis, D.O., and Balla, D.A.Delinquency al ,d Psychopatl Grune and Stratton, 1976. dogy. New York:

Lilienfield, AM. Foundationsof Epidemiology. New Press, 1976. York: Oxford University Loeb, J., and Mednick,SA A prospective study of predictors ofcriminality: Electrodermalresponse patterns. In: Mednick, S.A., eds. Biosocial Bases of and Christiansen, KO., Criminal Behavior. NewYork: Gardner Ness,1977. Maccoby, E.E. Socialgroupings in childhood: Their relationship toprosocial and antisocial behaviorin boys and girls.In: Radke-Yarrow, M., eds. Olweus, D.; Block, J.;and Development of Antisocialand Prosocial Behavior: Research, Theories andIssues. New York: AcademicPress, 1986. Maccoby, E.E., and Jacklin, C.N. Psychologicalsex differences. In: Rutter, M., ed. Scientific Foundationsof Developmental 1980a. Psychiatly. London:Heinemann, Maccoby, E.E., and Jacklin, C.N. Sex differencesin aggression: A rejoinderand reprise. Child Development51:964-980, 1980b. Magrab, P.R.; Soetek, AM.; and Powell, B.A.Prevention in the prenatal In: Roberts, M.D., and period. Peterson, L., eds. Preventionof Problems in Childhood: Psychological Research and Applications. New York:Wiley, 1984. Manson, S., and Shore,J.H.Psychiatric epidemiologic American Indians and research among Alaska Natives: Somemethodological issues. White Cloud Journal 2:48-56,1981. Masten, A.S., and Garniezy, N. Risk, vulnerabilityand protective factorsin developmental psychopathology.In: Lahey, B.B., and vances in Clinical Child Psychology. Kazdin, AE., eds. Ad- Vol. 8. New York:Plenum Press, 1985. Mednick, S.k A biosocialtheory of the learning Mednick, S.A., and of law-abiding behavior.In: Christiansen, KO., eds.Biosocial Bases of CriminalBe- havior. New York:Gardner Press, 1977.

309 CONDUCT DISORDER 303

Mednick, S.A.; Moffitt, T.;Gabrielle, W., Jr.; and Hutchings,G. Genetic factors in criminal behavior: Areview. In: Olweus, D.; Block,J.; and Radke-Yarrow, Research Theories M., eds. Development ofAntisocial and Prosocicd Behavior: and Issues. New York: AcademicPress, 1986. Mednick, SA, and Volavka, J.Biology and crime. In: Morris, N.,and Tonry, M., eds. Crime and Justice:An Annual Review of Research.Vol. II. Chicago: University of Chicago Press, 1980. Medway, F.J., and Smith, R. C., Jr.An examination of contemporaryelementary school affective education programs.Psychokgy of the Schools 15:260-269, 1978. Meketon, M.J.Indian mental health: Anorientation. American Journal of Orthopsychiatry 53:110-115, 1983. Newman, O. Defensible Space.London: Architectural Press, 1973. Offord, D.R. Family backgrounds ofmale and female delinquents. In:Gunn, J., and Farrington, D.P., eds.Delinquency and the Criminal JusticeSystem. New York: Wiley, 1982a. Offord, D.R. Primary prevention:Aspects of program design andevaluation. Journal of the American Academy ofChild Psychiatry 21:225-230, 1982b. Offord, D.R. Prevention ofbehavioral and emotional disorders ofchildren. Journal of Child Psychology andPsychiatry 28:9-19, 1987. Offord, D.R.; Alder, R.J.; and Boyle,M.H. Prevalence and sociodemographic correlates of conduct disorder. AmericanJournal of Social Psychiatry 6:272- 278, 1986a. Offord, D.R., and Boyle, M.H.Problems in setting up and executinglarge-scale psychiatric epidemiological studies.Psychiatric Developments 3:257-272,1986. Offord, D.R., and Boyle, M.H."Correlates of Conduct Disorder in aCommunity Prevalence Survey." Working paper,1987. P.S.; Cadman, Offord, D..; Boyle, M.H.;Szatmari, P.; Rae-Grant, N.1.; Links, D.T.; Bylee, J.A.; Crawford, J.W.;Munroe Blum, H.; Byrne, C.; Thomas,H.; and Woodward, C.A "Ontario Child HealthStudy: Correlates of Disorder." Work- ing paper, 19866. Offord, D.R; Boyle, M.H.; Szatmari,P.; Rae-Grant, N.1.; Links, P.S.;Cadman, D.T.; Byles, J.A.; Crawford, J.W.;Munroe Blum, H.; Byrne, C.; andWoodward, C.A. Ontario Child Health Study:Prevalence of disorder and rates of service utilization. Archives of General Psychiatry,in press, 1987. Offord, D.R., and Reitsma-Street, M.Problems of studying antisocial behavior. Psychiatric Developments 2:207-224, 1983. bO4 PREVENTION OF MENTALDISORDERS

Offord, D.R.; Sullivan, K.; andAllen, N. Delinquency andhyperactivity. Jour- nal of Nervous and MentalDisease 167:734-741, 1979.

Offord, D.R., and Waters, B.G.Socialization and its failure.In: Levine, M.D.; Carey, W.B.; Crocker, A.C.; and Gross, R.T., eds.Developmental-Behavioral Pediatrics. Philadelphia: Saunders,1983.

Osborn, S.G., and West, D.J.Do young delinquentsreally reform? Journal of Adolescence 3:99-114, 1980.

Patterson, G.R. Coercive FamilyProcess. Eugene, Ore.: Caztalia Company, 1982. Publishing

Patterson, G.R. The contributionof siblings to trainingfor fighting: A micrceo- cial analysis. In: Olweus,D.; Block, J.; and Radke-Yarrow,M., eds. Develop- ment ofAntisocial and ProsocialBehavior: Research, Theoriesand Issues. New York: Academic Press, 1986.

Patterson, G.R.; Littman, R.A.;and Bricker, W.Assertive behavior in children: A step towarda theory of aggression. Monograph of the Society for Researchin Child Development 32(6),1967. Pless, I.B., and Satterwhite, B. Chronic illness. In:Haggerty, R.J.; Roghmann, KJ.; and Pless, I.B., eds.Child Health and the Community. 1975. New York% Wiley,

Rae-Grant, N.I.; Thomas, H.;Boyle, M.H.; and Offord, in Reducing the Prevalence D.R 'Protective Factors of Common Behavioraland Emotional Disordersin Childhood." Submitted forpublication, 1986.

Ramey, C.T.; Yeates, K.O.;and Short, E.J. Theplasticity of intellectual development: Insights from preventiveintervention. 1925, 1984. Child Development 55:1913- Reitema-Street, M.; Offord, D.R.; and Finch, T. Pairsof same-sexed siblings discordant for antisocialbehaviour. British Journal 1985. of Psychiatry 146:415-423, Richman, N.; Stevenson, J.; and Graham, P.J.Pre-school to School: ABe- havioural Study. London:Academic Press, 1982. Roberts, M.C.; Elkins, P.D.;and Royal, G.P. Psychological prevention of accid9nts and applications to the illnesses. In: Roberts,M.C., and Peterson, L., eds. Prevention of Problems in Childhood: PsychologicalReseamh and Applications. New York: Wiley, 1984.

Robins, L.N. The adultdevelopment of the antisocialchild. Psychiatry 6:420-434, 1970. Seminars in CONDUCT DISORDER 305

Robins, L.N. Antisocial behavior disturbances ofchildhood: Prevalence, prognosis and prospects. In: Anthony, E J., andKoupernik, C., eds. The Child in Hi8 Family: Children at Psychiatric Risk. Vol. 3.New York: Wiley, 1974. Robins, LN. Psychiatric epidemiology. Archives ofGeneral Psychiatry 35:697- 702, 1978. Robins, L.N. Longitudinal methods in the study of normaland pathological development. In: Kisker, K.P.; Meyer, J.E.; Muller, C.;and Stromgren, E., eds. Psychiatrie der Gegenwart, Band, I. Grundlagen undMethoden der Psychiatrie. Heidelberg: Springer-Verlag, 1979. Robins, L.N. The consequences of conduct disorder ingirls.In: Olweus, D., Block, J., and Radke-Yarrow, M., eds. Development ofAntisocial and Prosocictl Behavior: Research Theories and Issues. New York:Academic Press, 1986. Robins, L.N.; West, P.A.; and Herjanic, B.L. Arrests anddelinquency in two generations: A study of black urban families and theirchildren. Journal of Child Psychology and Psychiatry 16:125-140, 1975. Robson, B.E., and Rae-Grant, Q. Divorce: A marker for prevention.Journal of Preventive Psychiatry 3(3):241-250, 1987. Rutter, M. Parent-child separation: Psychological effects onchildren. Journal of Child Psychology and Psychiatry 12:233-260, 1971. Rutter, M. Brain damage syndromes it, childhood:Concepts and findings. Journal of Child Psychology and Psychiatry 18:1-21, 1977. Rutter, M. Liagnostic validity in child psychiatry.Advances in Biological Psychiatry 2:2-22, 1978. Rutter, M. Psychological sequelae of brain damage inchildhood. American Journal of Psychiatry 138:1533-1544, 1981. Rutter, M. Prevention of children's psychosocial disorders:Myth and substance. Pediatrics 70:883-894, 1982. Rutter, M. School effects on pupil progress:Research findings and policy implications. Child Development 54:1-29, 1983. Rutter, M. Resilience in the face of adversity: Protective factors andresistance to psychiatric disorder. British Journal of Psychiatry147:598-611, 1985. Putter, M. 'Psychiatric Disorder in Parents as a Risk Factor forChildren." Piepared for Prevehtion Project, American Academy of Child Psychiatry, 1987. Rutter, M.; Cox, A.; Tupling, C.; Berger, M.; and Yule, W. Attainmentand adjustment in two geographical areas: 1. The prevalence of psychiatric disorder. British Journal of Psychiatry 126:493-509, 1975a.

3 ! 2 306 PREVENTION OF MENTAL DISORDERS

Rutter, M., and Gil ler, H. Juvenile Delinquency:Trends and Prospectives. New York: Penguin Books, 1983.

Rutter, M.; Maughan, B.; Mortimore,P.; Ouston, J.; and Smith, A.Fifteen Thousand Hours:Secondary Schools and TheirEffects on Children. Cambridge: Harvard University Press,1979.

Rutter, M., and Quinton, D. Parentalpsychiatric disorder: Effectson children. Psychological Medicine 14:853-880, 1984.

Rutter, M.; Tizzard, J.; and Whitmore,K Education, Health andBehavior. London: Longntan, 1970

Rutter, M.; Yule, W.; Berger, M.;Yule, B.; Morton, J.; and Bagley, C.Children of West Indian immigrants. I.Rates of behavioral deviance and ofpsychiatric disorder. Journal of Child Psychologyand Psychiatry 15:241-262, 1974. Rutter, M.; Yule, B.; Morton, J.;and Bagley, C.Children of West Indian immigrants. III. Home circumstancesand family patterns. Journalof Child Psychology and Psychiatry 16:105-123,1975b.

Rutter, M.; Yule, B.; Quinton, D.;Rowlands, O.; Yule, W.; and Berger,M. Attainment and adjustment intwo geographical areas.III.Some factors accounting for area differences.Br. 'Fh Journal of Psychiatry126:520-533, 1975c.

Sameroff, A.J., and Chandler,M.J.Reproductive risk and continuumof caretaking casualty.In: Horowitz, F.D., ed. Reviewof Child Development Research. Vol. 4. Chicago: Universityof Chicago Press, 1975, Schachar, R.; Rutter, M.; and Smith,A. The characteristics ofsituationally and pervasively hyperactive children:Implications for syndrome definition.Jour- nal of Child Psychology andPsychiatry 22:375-392, 1981. Scheff; T. Being Mentally Ill. Chicago:Aldine, 1966.

Siddle, DAT.; Mednick, S.A.;Nichol, A.R.; and Foggitt, R.H.Skin conductance recovery in antisocial adolescents.British Journal of Social andClinical Psychology 15:425-428, 1976.

Siddle, DAT.; Nichol, A.R.; andFoggitt, R.H. Habituation andover-extinction of the GSR component of theorienting response in antisocialadolescents. British Journal of Social andClinical Psychology 12:301-308,1973. Spivack, G.; Platt, J,J.; and Shure, M.B. The Problem-Solving Approachto Adjustment. San Franciso:Jossey-Bass, 1976. Spivack, G., and Shure, M.B. Social Adjustment ofYoung Children:A Cognitive Approach to Solving Real-LifeProblems. San Francisco:Jossey-Bass, 1974.

3 ' :3 307 CONDUCT DISORDER

risk model for Stiffman, A.R.; Jang,KG.; and Feldman,RA. A multivariate Journal of Orthopsychiatry56:204- childhood behaviorproblems. Anterican 211, 1986. primary prevention program Stolberg, A.L., andGarrison, KM. Evaluating a Community Psychology13:111-124, for children of divorce.American Journal of 1985. Submitted in "Reading Retardationand Antisocial Behaviour." Sturge, C. degree, University ofLondon, partial fulfillment ofthe requirements of M.Phil. 1972. Temperament and BehaviorDisorders Thomas, A.; Chess, S.;and Birch, H.G. in Children. NewYork: New YorkUniversity Press, 1968. early emotional deprivation. Wadsworth, M.E.J.Delinquency, pulse rates and British Journal ofCriminology 16:245-256,1976. Adolescence and Crime. Oxford: Wadsworth, M. Roots ofDelinquency: Infancy, Martin Robertson, 1979. illness and psychosocial Walker, D.; Gortmaker, S;and Weitzman, M. Chronic ptr5lerns among childrenin Genesee County.Boston: Harvard University School of Public HealthPublications, 1981. Children and J.S., and Kelly, J.B.Surviving the Break-Up: How Parente Cope with Divorce.New York: Basic Books,1980. London: West, D.J.Delinquency:Its Roots, Careers andProspects. Heinemann, 1982. West, D.J., and Farrington,D.P.Who Becomes Delinquent?London: Heinemann, 1973. London: West, D.J., and Farrington,D.P.The Delinquent Way ofLife. Heinemann, 1977. housing estates. In: Wilson, S. Vandalism and"defensible space" on London R, ed. TacklingVendalism. Home OfficeResearch Study, No. 47, London: H.M.S.O., 1978. London: Allen & Unwin, Wootton, B. Social Scienceand Social Pathology, 1959. Rutter, M.; and Yule, B.Children of West Indian im- Yule, W.; Berger, M.; of migrants: II.Intellectual performance andreading attainment. Journal Child Psychology and Psychiatry16:1-17, 1975.

3 4 CHAPTER 9 and Drug PreventionofAlcohol CriticalReview ofRisk Abuse: A Strategies FactorsandPrevention

Ph.D. Karol L. Kumpfer, Institute Psychiatry andSocial Research Department of City University ofUtah, Salt Lake Introduction complicated anddifficult and drugabuse is a The preventionof alcohol are many,and different of use andabuse of chemicals undertaking. The causes will firezdiscuss biological affect differentyouth. This paper vulnerability to clusters of factors thought to increasea child's and environmentalrisk factors preventionstrategies com- dependency andthen review do not developing chemical in this paperis that youth underlying assumption abuse, without monly usedtoday. An such as alcoholand drug develop disablingaddictive disorders, Youth rarelybecome emotional, orcognitive precursors. bullet" of peer prior behavioral, alcohol or di iigs(as if hit bythe "magic rapidly dependent on stressors arecombined withbiological ovarwhelming life academic pressure) unless youth manifestbehavioral and Often, high-risk Hence, preventionprograms vulnerabilitiec kindergarten andfirst grade. problems as early as the child's lifeand may takethe dependency shouldbegin early in in vulnerable for chemical remediation servicesfor risk factors form of specialtreatment or reviewed at theend of this paper of the preventionstrategies disorders children. Many prevalence of manydifferent childhood would help todecreass the besides alcoholand drug abuse. Importance forMedicalPractitioners child andadolescent practitioners such aspediatricians and Because medical to see childrenat high risk likely in theirprofessional practice signs psychiatrists are will discuss theearly warning and other drugabuse, this paper inheritance or the for alcohol is increasingevidence for of risk factors.Because there

initiative of theAmerican prepared for theProject Prevention This paper was Psychiatry. Academy of Childand Adolescent 3 1 5 310 PRKVENTION OF MENTALDISORDERS

"life-style disease"concept of alcoholism should be and drugahuse, medical knowledgeable ofthese research practitioners vulnerability studies so thatthey can concept to high-risk,using, or abusing explain the youth and theirfamilies. Risk FactorAssessments Althoughcomprehensive risk been developed factor assessmentinstruments have and evaluated,many prevention not yet discussing theneed for such specialists in thefield are screeninginstruments (DuPont instruments couldidentify children in press).Such attentioncan be paid to who are athigh riskso that a iitional are not specific reducing theirrisk factors. for alcoholand drug Many of theserisK factors adolescent and abuse, but adul problems(i.e., conduct are predictive ofmany other teenage suicide, disordersandjuvenile teenagepregnancy, emotional delinquency, problems). Thisfinding underscores problems, andlearning tification, referral, the importanceof intensifying and servicesystems for our iden- children manifestingproblems. Researchers andpractitioners other Aychiatric are increasinglydiscovering that, disturbances, thesingle most likemany psychiatricor medical disorder predictive earlyrisk factorfor a tion of the is positivefamily history disorder. Hence,the single or parental manifesta- chemically best predictorthat a youth dependent ishaving family might become or who have members whoare chemically manifesteda vulnerabilityto chemical dependent ing familyhistory and dependency. Hence,gather- medical specialists vulnerability datais a valuable who care forhigh-risk youth. prevention activityfor Need toTarget High-risk Children andAdolescents Given the small amountof prevention should be theprimary target resources available,high-risk children capita of preventiveefforts. In 1986, was spent annuallyon the prevention only 77 centsper problem annually of chemical costa the Nationmore than $850 dependency, yetthis DeMarsh 1986a).Considering the per person (Kumpferand demand-side small amount preventionstrategies, this " fundingavailable for targeting high-risk author hasrepeatedly youth whoaccount for most recommended and drug abuse. of the economiccosts of alcohol This public policy strategyhas recently programs for high-riskyouth as been implementedwith prevention Rehabilitation and mandated inthe Comprehensive Treatment Actof 1986. Drug Abuse munities, and This actmandates schools, preventionagencies todevelop com- youth presumedto be at risk: strategies fornine categoriesof 1. Children of alcohol andother drugabusers; 2. Victimsof physical, sexual, orpsychological abuse; 3. Schooldropouts; 4. Pregnantteenagers; 3 ALCOHOL AND DRUG ABUSE 311

5.Economically disadvantaged youth; 6.Delinquent youth; 7. Youth with mental healthproblems;

8.Suicidal youth; and 9.Disabled youth. A new prevention agency, theOffice for Substance AbusePrevention (OSAP), has been created in the Alcohol,Drug Abuse, and Mental HealthAdministration (ADAMHA) to coordinate and disseminatealcohol and ether drug abuse prevention strategies. High-risk youth arethe nukior targets of the $24million in newly awarded community demonstration/evaluationprojects, which include a high percentage of funding for minorityyouth. It is hoped that these projectswill to ad- produce well-documented andwell-evaluated high-risk youth projects vance theknowledge of the most effective ways toprevent alcohol and other drug abuse in high-risk youths.As one of the national evaluators forthe OSAP demonstration grant program, theauthor helped to review all of the130 new prevention programs. One ofthe mtkjor contributions of thisdemonstration program may be thedevelopment of new evaluation and programmaterials adapted for ethnic youth and families. Children with chemically dependent parents orrelatives are possibly at the highest risk of these mentioned groups.The majority of clients in treatmentfor alcohol and drug abuse (the averagein a large number of studies isfrom 60 to 70 percent) have had chemicallydependent parents or relatives(Cotton 1979; Templer et al. 1974; Goodwin 1971).Because a significant amount of important current research, including theauthor's (see Kumpfer 1987b for acomplete review), has been conducted on therisk factors of these children, this paperwill address risk factors found in childrenof chemically dependent parents aswell as other high-riskpopulations. Risk Factor Research Issues Risk factor research in the field ofalcohol and other drug abuse iscurrently in its infancy. Although many riskfactors have been proposed in theliterature, a number ofproblems have limited the discovery ofmajor etiuiogical variables of alcohol and drug abuse. Thesefactors include 1. The lack of empirically derivedcausal models; 2. The lack, until recently, ofstatistical analysis procedures sufficient totest causal models; 3. The specificity of risk factorsfor different types of drugs anddifferent types of youth; 4. The lhck of specificity of somerisk factors for drug use or abuaerather than other behavioral and emotionalproblems found in psychiatric syndromes related to drug abuse; and 3 I 7 812 PREVENTION OF MENTAL DISORDERS

5. The lack of longitudinal studiesneeded to create sufficient databases for etiological analysis.

Each of these deterrentsto etiological research hinders thedevelopment of a simple list of risk factors for alcoholand drug abuse that could be risk assessment included in a or should be addressed in prevention interventions.Each of these methodological andconceptual problems in riskfactor research is discussed in more detail below. Lack of Causal Models

A mAjor problem has beenthe lack of strong, empiricallytested theoretical models of etiology. Previous etiological researchwas flawed by an emphasison clinical-intuitive data (Nardi 1981)and contained majormethodological problems (i.e., lack of controlgroups, small sample sizes, unrepresentative samples, nonstandardized datacollection techniques, unreliableor invalid instrumentation, and lack of triangulationof data sources) (Olson 1983). Lack of Predictive StatisticalAnalysis Methods Another problem has been thelack of statistical analysis could test these theoretical procedures that models. Many risk factorsdiscussed in the alcohol and drug abuse literature have been discovered in researchstudies relyingon correlational statistical proceduresthat do not simultaneouslyconsider the relationship of a number of variablesto drug abuse. Hence, theimportance of the fact(or the relationship of the variables iscurrently unknown. Additionally, correlations showonly a strength of associationand not which factor predicts the other. Forinstance, zero-order correlationsas noted in the Kandel et al. (1986) review ofrisk factors for illicit druguse and delinquency from adolescence toyoung adulthood were not predictive oflater use when other factors were combined intoa path analysis model. Evenmore surprising is that risk variables with values too low to meet statisticalsignificance were actually the most predictive variables (highest beta weights) in thepath analysis model. Specifically, closeness toparents has a nonsignificantcorrelation (r=.026) with marijuana use, and maternalauthoritarian decisionmakinghas a nonsig- nificant correlation (r=.043)with other illegal druguse in young adult females, yet the beta weights for thepath analysis (beta= .106 and .108, respectively) demonstrated these variables to be the most significant predictivevariables for drug use. Variables withvery significant correlations (p< .001), such as prior use of marijuana, frequency of alcoholor cigarette use, and peer activity index, were not significant predictive factors of drug use. Hence,pure correlations are not very useful in determiningdirection of effect and causation. of multiple regression, With the advent path analysis modelingprocedures and event history analysis, advances within the field of alcohol anddrug abuse etiology should advance rapidly. ALCOHOL AND DRUG ABUSE 313

Specificity of Risk Factors Recent longitudinal studies(Kaudel et al. 1986; Baumrind 1985)also make it clearer that risk factors aredifferent or specific for each of thefollowing: 73-99); 1.Different drugs (Huba andBentler 1982; Handel et al. 1978, pp. 2. Different levels of drug use(Robins and Pryzbeck 1985); 3. Different types of youth(i.e., ages, age cohorts (Robins etal. 1986); 4. Ethnicity or race (Johnston etal. 1986), and gender (Kandel etal. 1978); and 5. Possibly different communityenvironments (Baumrind 1985). Battjes and Jones (1985) in theirconcluding chapter to their National Institute on Drug Abuse (NIDA)monograph, Etiology of Drug Abuse:Implica- tions for Prvuention, observe that"drug use is not a unitaryphenomenon" and involves a wide variety of substances,legal and illegal, having a varietyof pharmacological effects. A number ofimportant risk factoni have beeniden- tified for adolescence (particularlyearlier stages of drug use, such ascigarette and alcohol use), yet relatively littleattention has been focused on infancy,early childhood, and preadolescent riskfactors. According to the author'stheoretical model (to be discussed later), theserisk factors are diffennt at each stageof development. In addition, drug use patternsin America change over time and are different in differentregions of the country (for example,Berkeley, Califor- nia, compared with New YorkState), so that age cohorts are affecteddifferen- tially. For instance, before the late1960s, drug use was considered atypicaland deviant, became normative untilabout 1979, and then began todecrease (Johnston et al. 1986). Hence, drug usein youth who attbnded high school or college during the "prodrug" years willlikely have different etiological roots. Risk Factors for Use or Abuse A number of authors (Hawkins etal. 1985; Baumrind 1985; Murrayand Perry 1985; Robins and Pryzbeck1985) stress that the etiology ofdrug abuse is different from the etiology of drug use.Research has shifted from the study of drug abuse in clinical populations tothe study of drug initiation and use in general population epidemiology surveys.Much of the emphasis in the past few years has been onrisk factors for initiation in youth.However, the mAjority of youth who experiment with cigarettes,alcohol, or drugs will not become regular users. Johnston etal. (1986) calculated the discontinuationrates of high school seniors who had ever used a drug,but did not use it again in the yearbefore the survey.These rates are between 49 and 63 percentfor inhalante, methaqualone, sedatives, barbiturates, heroin,nitrites, and tranquilizers and 39 to 42 percent for other opiates, LSD,PCP, stimulants, and hallucinogens. Drugswith the lowest noncontinuation rates, whichshould be of most concern for parents and adolescent psychiatrists, arealcol -A (7 percent), cigarettes (16 percent), cocaine (24 percent), and marijuana (25 percent).In the New York Longitudinal 3 9 314 PREVENTION OF MENTALDISORDERS

Study, only 25 percent of 15 to 16 year olds in 1971 whoexperimented with illegal drugswere still using illegal drugs atage 23 (Kandel and Yamaguchi 1985). Because most of thiscountry's social, medical,economic, and cultural costs are caused by drug abuse,rather than druguse, this author believes that the emphasis should bereturned to the study of riskfactors for drug abuse. Lack of Specificity of RiskFactors for Drug Abuse The specificity of risk factors associated with drugor alcohol abuse is of increasing interest. Some ofthe biomedical riskfactors in alcoholism, in fact, are proving not to be specific to alcoholismalone. Much publicity hes the discovery of P300 surrounded amplitude decrements andlonger latencies in childrenof alcoholics (Begleiter et al.1984, Porjesz and Begleiter1985, pp. 138-182), but this biological marker is notspecific for alcoholismand also occurs in schizophrenia and epilepsy. persons with Another barrier to discovering risk factors specific to drugabuse or alcoh olism is the high probabilitythat the moresevere forms of both are related psychiatric syndrome that to a includes antisocial personality,Briquet's Syndrome (sometimes called the St. Louisian Triad, the DevianceSyndrome, Antisocial Personality), and chemicaldependency. All of these been found more prevalent psychiatric conditions have in the same families(Robins 1986) andeven in children adopted from thesetypes of families (Goodwin1985). The association between delinquent activitiesand illicit druguse has been established in repeated studies (Elliott and Huizinga 1984;Jessor et al. 1980; O'Donnellet al. 1976). Considerable study(Kandel et al. 1986) has whether drug been conducted todetermine use causee crime or whethercrime causes druguse, when it is most likely that theyare covariates for most youth syndrome. manifesting this psychiatric Lack of LongitudinalData Needed to DeriveRisk Factors In their chapter on "Implications of EtiologicalResearch for Prevention Interventions and Future Research," Battjes and Jones(1985) repeatedly made a strong case for the necessity oflongitudinal studies in the alcohol and drug abuse. understanding of Several longitudinal studieshave focusedon correlates of mental health problemsand juvenile delinquency and Boyle 1979; Elliott in adolescents (Brunswick et al. 1982; Jessor andJessor 1977) ur cL ildren and Brown 1982; Baumrind kKellam 1985), but few completedlongitudinal studies have focused primarilyon alcohol and drug abuse. Several longitudinalresearch projects designed specificallyto study drug abuse public high school study are Kandel's New York State begun in 1971 (Handel1975; Kandel et al.1976a); Johnston's early followupof a national male cohort 24 in 1974 (Johnston from sophomoreyear to age 1973); and Monitoringthe Future, Johnston'snational study of high schoolseniors (Johnston et al. 1971, seven longitudinal 1982, 1983, 1984, 1985,1986). In studies (Kandel, Jessorand Jessor, Brunswick, Johnston, Smith, Kaplan,and O'Donnell)were continued with funding to specifically on the focus consequences of drug use on the functioningof young adults. 3 .1: ALCOHOL AND DRUG ABUSE 315

using causal mod- Some of these longitudinalstudies (specifically Kande]) are eling procedures that willalso produce data relevant forrisk factor research. Risk Factors for Alcohol andOther Drug Use in Youth

Risk factor research has evolvedthrough several different phasesin its short history. The first phase of studiesfocused on youth currentlyusing and correlates of their use bycross-sectional studies of different groups.This research focused primarily ondemographic variables such as race,gender, socioeconomic status, and education.The next phase of etiologicalstudies examined risk factors for drug use orabuse in several different domainsof psychosocial environment, such ascommunity, peer, school, andfamily do- alcoholism field, has mains. More recent etiologicalresearch, particularly in the focused on biological and inheritedrisk factors.Each phase of risk factor research is discussed in moredetail below. Phase One: DemographicRisk Factors The first atempts to determinerisk factors for alcohol andother drug abuse focused almost exclusively ondemographic factors such as race,socioeconomic status, and sex (Fraser 1987b).Many of these factors, such asminority status, male gender, and low socioeconomicstatus, have strong correlations tochemical dependency. A number of preventioninterventions are currently targeted at members of these demographic groupssimply because the total group is athigh risk. However, this level ofunderstanding of the etiology of alcoholand other drug abuse does not offer theprevention specialist much help idetermining why members of these groups ofyouth are at high risk or what canbe done to reduce these risk factors. Certainlychanging a person's gender or race inorder to eliminate these riskfactors is not an option. Hence,demographic risk factor research does little to informprevention interventions, althoughit may help providers to decide target services.These demographic riskfactorsgender, ethnicity, age, geographic location,and projectsand theirrelationship to chemical dependency are discussedbelow. Gender Many general incidence andprevalence surveys show that males are more likely than females to become dependent onalcohol or drugs (except for nicotine and other stimulants).According to the U.S. NationalDrinking Practices Survey (Clark and Midanik 1982),about 6 percent of women and about 10 percent of men in the UnitedStates have alcohol problems.Lifetime prevalence of alcohol disorder is about 24percent for men and 6 percentfor women, al. according to the EpidemiologicCatchment Area (ECA) study (Robins et 1986). Daily alcohol use is reported tobe about 3 percent for highschool women and 7 percentfor high school men (Johnston etal. 1986). 3 2.1 316 PREVENTION OF MENTALDISORDERS

The National High SchoolSenior Survey (Johnstonet al. 1986) consistently reports higher rates of use and abuse of alcohol, marijuana,and other drugs in males than in females.Only in theuse of stimulants (diet pills) do exceed males in annual females prevalence rates (16.4vs. 14.9 percent), whileyoung females smoke cigarettesabout as oftenas young males. The rate ce cigarettesmoking by females has currently 12 percent of high increased over theyears; school senior girls and12.3 percent of highschool senior boys smokea half-pack or more daily, and higher for high school ever-used rateswere slightly senior girls than boys (69.7vs. 67.4 percent). Since 1980, cigarette smoking has consistently been higheramong females than males in conege (approximately12 vs. 8 percent). High school senior males show considerablyhigher dailyuse rates for marijuana than females (6.9vs. 2.8 percent). These daily are higher for young adults (18 marijuana use rates to 25 years old). in 1982,27 percent ofyoung adult males and 15percent of young adultfemales reported daily (Miller et al. 1983). marijuana use Monthly use rates of highschool senior malesare also proportionately higher thanthose of females fora number of illegal drugs such as heroin (0.3 vs. 0.1 percent), LSD(2.4 vs. 8.0 percent), hallucinogens (3.4 PCP (2.4 vs. 7.0 percent), vs. 1.4 percent), inhalants (2.8vs. 1.7 percent), and amyl and butyl nitrites (3.0vs. 4.0 pement). Differences in male and femalemonthly use rates are not as large forusers of cocaine (7.7 vs. 5.6 percent), such as tranquilizers prescription drugs (2.2 os. 1.9 percent),other opiates (2.6 sedatives (3.0 vs. 2.0 percent), vs. 1.8 percent), and barbiturates(2.4 vs. 1.6 percent). Binge drinking is considerably higher in highschool boys than ingirls, with 45 percent of males andonly 28 percent of drinks in a females reportinguse of five or more row in the prior 2 weeks. Thisfigure for boys increased to 57 percent in college to from 45 males and is the onlyalcohol statistic thatincreased from 1982 to 1985 (52to 57 percent); for college 37 to 34 percent. females this ratedropped from Ethnicity and Race Ethnic and racial differences in prevalencerates are difficult to because theyare often confounded by determine socioeconomic status andliving conditions. Although white youthcompared with nonwhite youth have reportedconsistent- ly higher levels ofuse of marijuana, cocaine, and prescribed and unprescribed), other illegal drugs(both these differencesare now only several percentage points (Miller et al. 1983). The ECA studyby Robins et al. (1986) overall differences repo7ted that in lifetime alcoholdisorders betweenraces are small. They found lifetime alcoholismrates for all ages t be 23 percent for '.lieck percent for whitemen, 24 men, and 28 percent for othermen (predominantly Hispanic), and 4 percent for whitewomen, 6 percent for black other wcmen. women, and 5 percent for Probably the moststriking result of theECA study is the significant race-by-ageinteraction. Although lifetimealcoholism ratesare substantially higher inyoung adult whites than blacks 3 9 (29 vs. 13 percent),they ALCOHOL AND DRUG ABUSE 317 are substantially lower inelderly whites than blacks (13 vs, 24 percent). The meaning of this diametric switch in alcohol disorder ratesbetween whites and blacks is open to speculation. Robins hypothesized acultural integration and economic opportunity theory. These statistics suggestthat blacks are not more genetically vulnerable than whites to alcoholism and thatsocioeconomic level and cultural norms can change alcohol use. Urban American Indian youth appear particularlyvulnerable to early and heavy alcohol and drug use (Miller 1981). A comparisonof white, Hispanic, black, American Indians, and racially mixed youthfound that the latter two groups experimented with drugsearlier (Jackson et al. 1981). Among whites, Vaillant (1983) found, in a longitudinal study ofalcohol use, that Americans of northern European descent manifesthigher rates of alcoholism than Americans of southern European descent. Age Young adults (18 to 29 years old) use alcohol and drugs more than anyother age cohort. One exception tothis is the earlier noted reversed pattern with blacks. Young adult blacks use significantly less alcohol thantheir parents or grandparents. Robins et al. (1986) speculated that this may haveoccurred because the parents were "second generation" blacks who moved tothe city and had few opportunities for good housing and jobs, whereasthe young adults now are "third generation" blackswith better community connections and opportunities. Socioeconomic Status The major indicators of socioeconomic status are discussedbelow. Income There is little evidence that socioeconomic level per se influencesdrug use. It is possible that low income may contribute to stress,which could be a psychosocial precursor of drug abuse. Chemical dependency, however, appears to cut across all income categories and is not related to residency in ghetto areas. The authorfound in her researth of illegal drug abusers a U-shaped distribution of incomein drug abusers in treatment, just the opposite of distribution in the area's generalpopulation (Kumpfer 1987a). Many drug abusers had very low incomes and othershad very high incomes, possibly because ofillegal drug sales.Because of this confounding factor of drug-related income, income level may be a poorindicator of socioeconomic status. Employment Controlled drug 1156 appears to be possible without jeopardizing many types of employment; however, the medical, social, and behavioral consequencesof drug and alcohol abuse leave many abusers without satisfactory employment

0 318 PREVENTION OF MENTAL DISORDERS

(Aps ler 1979).The final stages of alcohol dependencyare detrimental to economic self-sufficiency, particularlyin a supervised, conventional jobwith much responsibility. Users in the middleand upper income levels oftenhave sufficient incomes to purchaseexpensive drugs, anda high proportion of habitual drug usersappear to be able to work while addicted to illegaldrugs (Bale 1979). When their income doesnot cover the cost of the drugs,users may consider street crimeor rehabilitation. Loss of employment is often citedas a life stressor thatmay help to precipitate heavy drinkingor drug use. Many rehabilitation centers consider steadyemployment to be a protective factor against relapse (Friedman 1980,pp. 171-188). Education

Education is probablya better indicator of socioeconomic statusbecause income and employment levelsare likely to reflect theconsequences of alcohol and drug problems. The ECA study(Robins et al. 1986) suggestedthat higher educational attainment is correlatedwith lower lifetime alcohol disorderrates. These differencesare for both males and females and for differentage cohorts, although they were notas dramatic for the elderly, possibly becausefewer attended college or completed highschool. The largest differencesare noted for 30- to 59-year-old men, who have a high 36 percent rate oflifetime alcohol disorder, compared with only 20percent for college graduates. Johnston et al. (1986) reported littledifference betweenyoung adults who do and do not attend college in annualprevalence of use of alcohol (bothabout 92 to 95 percent with full-time collegestudents slightly higher),marijuana (both about 45 percent),or any illegal drug other than marijuana at about 28 percent). For (primarily cocaine, more frequent drug usage, noncollegersapondents had significantly higher rates thancollege students. However, collegestudents had significantly lower rates ofregular daily use ofa half pack or more of cigarettes, but higher rates of periodicheavy drinking (fiveor more drinks in a row) compared with noncollege-agemates. Regional Differences

in the United States, overallalcohol and drug consumptionrates are lowest in the South and north CentralStaten and highest in theNortheast and West. Alcohol consumption for the generalpopulation is highest in theRocky Moun- tain region, the West, and theNortheast. The highestconsumption ratesare in Washington, DC. In youth,the rate of occasional heavydrinking is lower in the South and West than in theNortheast and north CentralStates (Johnston et al. 1986). Regular cigarette smoking in high school seniorsis lowest in the West (8 percent) and South (10percent), but higher in the northCentral States and Northeast (17 percent).Cocaine use has the largestregional differences with nearly three timesgreater annual prevalence inthe Northeast (20.8 percent) and West (19.7 percent) compared with the South (7.5percent) and north Central States (8.2 percent).Use of centralnervous system depressants 3°4 ALCOHOL AND DRUG ABUSE 319

tranquilizers, and heroin) islowest in the West (barbiturates, methaqualone, the West have the and highest in the Northeast.The north Central States and highest rates of stimulant use. Population Density illegal drug use in The largest metropolitan arephave the highest levels of high school seniors accordingto the most recentJohnston et al. (1986) survey. found for cocaine, for whichthe usage rate in the The greatest differences are high (19 percent) as in the larger metropolitan areasis more than twice as nonmetropolitan areas (9 percent).Marijuana annual usage is alsohigh, with 44 percent of urban seniorsusing compared with 37 percentof nonmetropolitan youth. Summary of DemographicRisk Factors classification for As mentioLed earlier, onedisadvantage of using group specify which targeting prevention interventions,is that this method does not youths in the group are athigh risk Hence, many youthsin the group will be false positives and thereforeidentified as high risk and putthrough intensive unnecessarily. In addition, this groupclassification prevention interventions high-risk or how to does little to clarify the reasonswhy the group is considered intervene to lower the risk. Phase Two: PsychosocialEnvironmental Risk Factors For these reasons, preventionresearchers and specialistshave turned to experimental research studies,rather than general population surveysto determine which groups are athighest risk and the etiology of thisrisk More recent papers on risks foralcohol and other drug abusehave considered risk and protective factors by thefollowing psychosocial environmentaldomains: community, school, and family.Reviews of these risk factorshave been extensively covered by a number ofresearchers such as Kumpfer andDeMarsh (1985); Kumpfer (1987b); Hawkins et al.(1985, 1986, 1987); Fraser andHawkins (1984); Huba and Bentler (1980); Jessorand Jessor (1977); Johnston etal. (1982); and Murray and Perry (1985). A Conceptual Framework factors is A conceptual framework fororganizing these risk and protective presented in figure 1. After a completereview of the literature for allrisk factors plotted on the found to empirically correlate withdrug abuse, these factors were Public Health Service (PHS) Model ofPrevention, which consists of a triadof host, agent, and environmentcluster variables. Other similargroupings of factors have been created usingthe PHS prevention model(Schinke and Gilchrist 1986; Nathan 1983), but noneof these focuses on theinteraction of the three cluster variables. Theagent/environment and agent/host riskfactors (often considered community riskfactors) and the more commonpsychosocial host/environment factor are discussed in moredetail below. 3r" 320 PREVENTION OF MENTALDISORDERS

AGENT Availabiaty 0Type of substance Costs Sid deficits * Strength Low Saresteern Waive propeities Behavior problems Blocking potential Prodrug Navies Health effects Low academic mativatbn Side effects Psychobgical distutances ReInfoiting properties Lack of pest refusal skitis Perceived psychological ReOrtion of proemial values/rah**, effects

PRESUBSTANCE ABUSE SYNDROME

ENVIRONMENT HOST

Family Environment Child s Cognons Cognitions Stressors Streuors Coping Resources/Skit Cooing Resources/Skills

Community ent Social/Peer Environment Cognitions Cognitons Stressors Stressors Coping Resources/Skins Copt% Resources/Skills

Figure L VASC (Value/Attitudes,Stressors, Coping Responses) theory of alcohol anddrug abuse,

Agent! Environment andAgent I Host Factors Agent (alcoholor drug) risk research has been alcohol preveution conducted primarily inthe arena, because data concerningprice and availability controlled for illegal drugs. are less Moskowitz (1986) hascompleted an excellentreview of risk factors andpreventionprograms for alcohol use. Agent by the cultural factors affected environment that have beenfound to correlate with are availability and cost (('oate alcohol use and Grossman 1985;Grossman et al. 1984). Unfortn.nately, the religiouscomposition of different to correlate highly with areas of the country tends cost and to some degreewith availability;therefore, 3 i) ALCOHOL AND DRUGABUSE 321

community "drinking sen- when Coate and Grossman(1985) introduced these in youth was eliminated. timent" controls, the effect ofprice on beer consumption reduced beer consumptionremained The effect of loweredminimum legal age on variable. Neither wine norbeer consumption appearsto be infkenivd a salient monopoly systems with few by availability measures(e.g., State-controlled whether drug and grocerystores can aell alcohol, outlets, laws concerning alcohol can be advertised on number of alcoholretail licensee, or whether billboards). increased taxes and price. Liquor consumption appearsto be more sensitive to studies that track changesin alcohol consumption Several quasi-experimental additional taxes, supports within States after the costof liquor is raised through beer consumption) is the finding that liquorconsumption (but not wine or sensitive to cost and decreasesin years following a taxincrease (Cook 1981, Cook and Tauchen 1982).Moskowitz (1986, p. 22)observed that pp. 255-285; communities it is such an beer may not be as pricesensitive because in many (comparable in price to softdrinks), "thus the effects on inexpensive commodity be trivial." consumption of any marginaldifferences in the price of beer may of different types ofliquor costs support the The differential sensitivity kr different drugs or previously mentioned pointthat risk factors are different alcohol. At some level of awareness,the youth (host) evaluatethese agent sanctions, health effects or factors (e.g., addictiveproperties, social and legal to determine whether risks, cost, and perceivedpsychological benefits or effects) drug, how much to use, and whento use it. Thiscomplicated to use a particular values and attitudes, assessment is also affectedby the hild or youth's current streesors, and copingskills, as shown in figure 1. Environment / Host Continuumof Domain Factors Psychosocial risk factors mostoften covered in riskfactor reviews am or- (Kumpfer and DeMarsh 1985)into a 3-by-4 matrix ganized by the author's model drug use, and coping of values/attitudes(cognitions), stressors influencing resources that protectagainst drug abuse by thefour major domains of family, in the model, each community, schools/peers, andindividual factors. As shown peers--influences the child, of these clusters offactorsfamily, community, and and the child influenceshis or her environment.Each of these four environmental domains is discussedin detail below, These clusters of environmentalvariables influence the childdifferentially during development. The familyis kpothesized toinfluence the child the most schools, religious early in life, with the community(television, cartoons, books, influence in early childhood.As the child be- institutions) becoming another increases until it is the most comes anadolescent, the influence of peers influential in some youthswho are alienated fromfamily and society. Throughout this developmental sequence,these environmental factomshape including values and at- the youth's (the host's)personality and behaviors, titudes, stressors, and coping resources.The combination of thesefactors is hypothesized to predispose a youth to use ornot use drugs. A vulnerableyouth 327 822 PREVENTION OFMENTAL DISORDERS

is hypothesizedto manifest earlyin life a "preeubstance (shown in the center abuse syndrome" of the triangle),consisting of eight more distinct in vulnerable major risks thatbecome youths as they becometeenagers: 1. Skill deficits; 2. Low self-esteem; 3. Behaviorproblems; 4. Prodrugattitudes; 5. Low academicmotivation; 6. Psychologicaldisturbance; 7. Lack ofpeer refusal skills; and 8. Rejection ofprosocial values undreligion. These are the major characteristicsthat have beenfound in studies to discriminateyouth who become a number of (Jessor and Jessor abusers from thosewho abstain 1977; Kandel et al.1986). A numberof these were determined in studiesthat compared characteristics delinquent and alcoholor other drug-abusing youth.Many of thesestudies used simple rather than multipleregression or multivariable correlational analyses recent longitudinalresearch that analyses. Unfortunately,more these characteristics employs causal modelshas found fewerof to be predictive ofadolescent drug abuse. Kandel et al. (1986)in a longitudinal followup ofa 1971 cohort of 15 year olds (N4004) in NewYork State found to 16 is predicted (22 that druguse in young adult males percent) in Kandel'smodel by adolescent of employmentand is protected cigarette use and lack by marriage.Risk for drug is predicted (10percent) by use in young women never being married andhaving an authoritarian mother. Psychologicalfactors such in adolescence as emotional disorderand dysphoricmood were most predictive ofadult crime hi cent delinquencywas not predictive. women, whereas adoles- CommunityEnvironment community Valuesand Attitudes Youth who live incommunities that more likely to condone theuse of alcc,hol and drugsare use them. Thepreviously mentioned Grossman (1985)suggests that research of Coateand religious composition a community's "drinkingsentiment" and are major determinantsof alcohol community has itsown informal social control consumption.Each system that generatesnormative influences pertainingto drinking anddrug use and associated with each(Maloff et al. 1979). appropriate behaviors Cultural recipesthat have of each subgroup normative propertiesare reinforced by members and internalizedby individuals. (1986), if there As discussed byMoskowitz are many %dietingmessages, as within 6 our complex society, ALCOHOL AND DRUG ABUSE 323 the cultural recipes may breakdown, resulting in diminishedinternalization and a greater need for compliancemechanisms. Formal control systems(laws, rules, enforcement, and punishments) arethen needed to regulate individuals' use or possessionof illegal substances. Formal controlsystems have not been shown to produce long-lasting decreasesin alcohol or other drug abuse (Polich and Bloom 1985) and are notconsidered effective prevention strategies(see Fracer 1987a for a complete reviewof supply-side prevention strategies). Social Bonding to Traditional Communities However, there are those who willabuse alcohol or other drugs in communities where the n m is abstinence ormoderate use.In this type of community, alcohol or other drugabusers are generally youth who are less bonded to society (Hirschi 1969). These youth arealienated from the dominant values of their community (Smith andFogg 1978, pp. 87-102); more rebellious against mocietal rules and authority(Kandel 1982; Goldstein and Sappington 1977); less attached to their parents(Jessor and Jessor 1977); less committed to their school (Elliott and Voss1974); less involved in activities of religious institutions (Schlegel and Sandborn1979); and less involved in recreational, sccial, and cultural activities (Kumpferand De Marsh 1986a, pp. 49-91). Community Dysfunction Communities differ in their degree of healthiness as measurtiaby the number of stressors and the number ofcommunity resources available to citizens and youth.Unfortunately, few studies have beenconducted on the impact of community wellness on alcohol or otherdrug abuse.Major changes in a community's stress levels, such as substantialincreases or decreases in unemployment or population (as in boom towns) oreconomic problems, could serve as naturally occurringexperiments that should be evaluated.Risk factors for crime that are likely to covary withalcohol or other drug abuse are high population density, rapid changes in neighborhoodpopulations (Sampson et al. 1981), and frequent residential moves (Farnsworth1984). Catalano et al. (1985) found that residential mobility is associatedwith higher rates of drug initiation and frequency of use. School and Peer Environment School and Peer Attitudes and Values As previously mentioned for communities, the normsof a youth's school and peers will affect theyouth's use of alcohol or drugs. It is notknown whether youth with increased vulnerability andmotivation to use drugs choose drug- using peers or whether peer use affects theyouth's use of drugs (Kandel and Yamaguchi 1985). The outcome is the same,however, in the chain of risk factors for abuse. One of the strongest predictors ofadolescent drug use is association with drug-using peers (Elliott et al. 1985;Kandel 1982). A youth's interpretation 324 PREVENTION OF MENTAL DISORDERS

of the school andpeer norm is influential in drug use. Several studies have found that perceived use of drugs by othersis a strong predictor of druguse (Robins and Ratcliff 1979; Jessor andJessor 1978). In young adulthood,peer use has also been found to correlate significantly with marijuana and illegal druguse (Kandel et al. 1986). However, other variables were more predictive of druguse when a causal modeling analysiswas conducted, namely adolescent cigaretteuse, lack of employment, and unmarried status for males and unmarried status and havingan authoritarian mother for females.

it is also unknown whether children,youth, or you ng adu Its have much choice in their associates. Somemay wish to associate with high status, traditional friends but are not acceptable to thatgroup.Studies suggest that early childhood social and emotional disturbancesare significant predictors of adolescent alcohol and drug abuse and delinquency(Kellam and Brown 1982; Lerner and Vicary 1984). Hence, it is possiblethat children with conduct disordersand emotional problems tend to associate withthose children low in the socialorder who by default will have themas friends. Hawkins et al. (1987, p. 94) stressed that "it is not known at what pointpeer associations become important in predicting delinquency and drug use."Longitudinal studies of the effectsof childhood peer associationson adolescent drug use are just beginning (Giordano et al. 1986; Coie and Dodge 1983).

Schools with norms that condonethe 'Ise of marijuana and other drugshave been found to have both higheroverall use rates and also higherusage by better adjusted youth (Baumrind 1985). Aswith communitynorms, it is likely that the formal control system of schoolpolicies will reflect the schoolnorms toward alcohol and druguse. Little research has been conductedon the impact of these school policies. School Stressors and Coping Resources Youth who have increased stress in school in terms of increasedacademic and behavioral problems (Heijanicet al. 1977, pp. 445-455; Rimmer 1982); placement in special classrooms(Holmberg 1985); increased rejectionby school peers (Coie and Kuperschmidt 1983; Kumpferand DeMarsh 1986a); and cognitive deficite and lack of academicmotivation (Smith and Fogg 1978)are more likely to become abusers of alcoholand other drugs. At-risk youthwho are less bonded and have low commitmentto school are more likely touse alcohol and drugs and to become delinquent(Johnston et al. 1986). To dealwith the stress of school, they tend touse inappropriate coping responses, suchas skipping classes, being truant, droppingout of school, and using alcoholor drugs. Family Environment Factors

Family members, because theyinfluence the child's early psychosocialen- vironment, have a tremendousimpact on the child's vulnerabilityor proneness to alcohol or druguse. As mentioned at the beginning ofthis paper, both 3C 0 ALCOHOL AND DRUG ABUSE 325 parental use of alcohol and drugs andother stressors on the child, such as parental dysfunction, create increasedrisk in the child. The author has just finished a complete review of earlychildhood psychosocial risk factors for a NIDA research monograph (Kumpfer1987b). Some of the family factors covered in that review are summarized below. Parental and sibling alcoholism(Cotton 1979; Goodwin 1985) and use of illicit drugs (Smart and Fejer 1972; Thorne andDeBlassie 1985) significantly increase the youth's vulnerability to becoming analcohol or drug abuser. A number of mediating variables have been hypothesized toexplain this relationship, such as inherited geneticvulnerability (to be discussed in more detail inthe next major section), identification and modelingof parental or sibling drug use, accessibility to drugs and alcohol, and lack offamily sanctions for use. Family Attitudes and Values Early family environment is consideredth a primary determinant of attitudes, life stressors, and coping skills that willeventually influence the youth's need for and choice of drugs, including alcohol(Kumpfer and DeMarsh 1986a). Looking more specifically at adolescentalcohol and drug use, positive family relationships, involvement, and attachment appear todiscourage youth's initiation into drug use. Kandel (1980) foundthat parental influence varies with the stages of drug use that she identified.Parental role modeling of alcohol use is poeitively associated with adolescent use ofalcohol, whereas the quality of the family relationship is inversely related to the useof illicit drugs other than marijuana. According to Kandel, three parentalfactors help to predict initiation into drug use: parent drug-using behaviors,parental attitudes about drugs, and parent-child interactions. The latter factor ischaracterized by lack of closeness, lack of maternal involvement in activitieswith children, lack of or inconsistent parental discipline, and low parental educationalaspirations for the children. Family Stressors and Coping Resources Research seems to be consistent regarding theeffects of quality and consistency of family management, familycommunication, and family cohesiveness on alcohol and drugabuse. Stanton (1979) and Ziegler-Driscoll (1979)suggested that familial stressors on the child include apattern of overinvolvement by one parent and distance or permissiveness by theother. Similarly, families with children who abuse drugs are described by Kaufmanand Kaufman (1979) as families in which fathers are "disengaged" andmothers are "enmeshed." Early childhood family stressors that a health care specialistwould want to monitor generally include parental peychiatricdysfunction, family dysfunction, and parenting dysfunction (each discussed in moredetail below).

1.Parental psychiatric dysfunction. To the degree thatthe parents suffer from chemical dependency, affective disorders(Kumpfer and DeMarsh 1986a), antisocial personality (Robins 1966;Booz-Allen and Hamilton 1974), or other psychiatric syndromes, the child willlikely not receive the 331 326 PREVENTION OF MENTAL DISORDERS

kind and amount of care and attentionthat all children need.Since psychiatric disorders often run in families,the child could actually bea special needs child born to parents whoare least well equipped to deal with such a child. 2. Family dysfunction. A few recent studiesof the dynamics of families in which one ol the parents is chemically dependentsuggest increasAd family problems, such as increased familystress and family coin; st and decreased family cohesion (Kumpfer andDe Marsh 1985); decreased family organization, home managementskills, and family rituals (Bennett and Wolin 1985); increased family socialisolation (Kumpfer and De Marsh 1985); increased marital distress and conflict(Jacob and Seilhamer 1986; O'Farrell and Birchr...1985); andfrequent family moves (Vaillant and Milofsky 1982; Catalano et al. 1985; Kaplanet al. 1984). Parental marital discord has been found to be stronglyassociated with use of heroin and other illegal drugs (Simcha-Faganand Gersten 1986). Becauseso few studies of family dynamics have beenconducted (fewer than 10 in thepast 12 years), and because thisappears to be a promising direction for risk-factor research, simdies of thistype should be replicatod with direct observations of the family interactionscomparing general population families with families that abuse alcoholor other drugs to determine the extent of these differences in family interaction. 3. Parenting dysfunction. Parentingis learned primarily from one'sown parents. Two major longitudinal studieshave demonstrated thatcross- generational deficiencies in parentingand discipline practices have been demonstrated to correlate with antisocialoffspring who are at higher risk for alcohol or drug abuse (Elder et al.1983, pp. 93-118; Huesmann et al. 1983). A number of studies have foundthat alcohol or drug abusers experienced parental disciplinethat was slack, inconsistent,or authoritarian (Baumrind 1983; Kandelet al. 1986; Sowder and Burt 1978a, b). Disruptions in family managementare a major mediating variable in children's future dysfunctionalbehavior (Patterson 1982). Variablesassociated with antisocial problems includehouseholds that are disorganized andhave poorly defined rules and thosethat have inconsistent, ineffectivefamily management techniques. Patterson's (1986)causal modeling datasuggest that failure "by parents to effectively dealwith garden variety, coercive behaviorsets into motion coercive interactionsequences that are the basis for training in aggression." The author's researchfound that parents who abusealcohol or other drugs spent less time withtheir children and spent less timepositively reinforcing their children for goodbehaviors (Kumpfer and DeMarsh1985). Problem behaviors have been foundto increase a child's risk of alcoholand drug abuse (Jessor and Jessor 1978;O'Donnell and Clayton 1979,pp. 63-110). ALCOHOL AND DRUGABUSE 327

Child Personality andBehavior Factors Child's Attitudes andValues of any drug or alcohol A number of studieshave found that initiation into use favorable to its use (Kandel etal. 1978; Smith and Fogg is preceded by attitudes attitudes and values 1978). According to theauthor's theoretical model, these froir the youth's community,school and peers, and family.These are learned and identification of the cognitions are mediated bythe degree of social bonding influences on their socializatim.Unfortunately, youth with those primary those who have youth who are most likely tobecome alcohol or drug abusers are and a strong need the least identification withothers, the least social bonding, 1976). These youth tend tobe more normless (Paton for independence (Jessor et al. 1977), and and Handel 1978), have ahigh tolerance for deviance (Brooks resist traditional authority(Goldstein and Sappington1977). Basic Personality Traits for an alcohol For many years, researchershave searched with little success (1986), in a sequen- and other drug abusingpersonality. Labouvie and McGee tial longitudinal analysisof the Rutgers Health andHuman Development Project data of 882 adolescents(12 to 21 years old), foundthat light, moderate, heavy users and heavy users differed inbasic personality traits. Early onset, avoidance on the scored lower on achievement,cognitive structure, and harm Jackson Personality ResearchForm (1968) and higher onautonomy, exhibition, impulsivity, and play than laterand lighter users. These "ligh-riskyouth move rapidly to multiple drug useby age 15 and cocaine useby age 21. Because of positive family or schoolimpact, light users tended tohave exactly the opposite personality traits and to limittheir use to alcohol even at age21. These data suggest thatyouth at high risk for chemicaldependency have different personality structures tobegin with and that their basicpersonalities drug or alcohol do not change significantly overtime (3 years) with increased use. These results,combined with information fromother longitudinal studies of youth, support thenotion that youth prone toheavy chemical dependency are already different by the timethey reach the age of alcohol ordrug use. child or Youth's Stressors andCoping Resources Children and youth from lesssupportive and functional families, communities, and schools have morestreesors to cope with in theirlives. Some invulnerable children from theseunfavorable environments learneffective coping skills, such as soliciting peer,adult, and religious support andfocusing competencies, good wellness practices,and other skills. High-risk on developing behaviors to help youth are more likely to usealcohol and drugs and antisocial them reduce stress, Earlyinitiation of drug use predictssubsequent misuse of drugs, greater frequency of' use,and probability of involvementin deviant activities such as crime andselling drugs (Kandel 1982; Brunswickand Boyle 1979). 333 328 PREVENTION OF MENTALDISORDERS

Early antisocial behaviors,including acting out, overinvolsment in socially disturbing behaviors, impatience,impulsivity, and acting defiantand negative (Spivak 1983), have been found tocorrelate positively with adolescentdrug use. Aggressive and thrill-seeking behaviorshave also been found in youthat high risk for alcohol and other drug abuse. Kellam and Brown(1982) found, ina longitudinal study of black children,that the combination ofaggression and shyness as early as the first gradewas predictive of adolescent alcohol anddrug abuse.Patterson (1986) said thatmore than a dozen longitudinal studies (Olweus 1979, 1980) haddemonstrated that aggressionis as stable as intelligence in children. His most recent research supports thedevelopment and maintenance of aggression by faultydiscipline and parentingpractices. Dis- rupted family management skillslead to the development of anticocial behavior in the children. These sarressive, coercive, and noncompliantbehaviors then increase the child's risk of academic failure and rejectionby normalpeers. Patterson hypothesizes thatpoor social and academicsuccess, combined with noncompliant behaviors, then leadsto parental rejection and lowself-esteem. Phase Three: BiologicalVulnerabilities The last alcohol and drug abuserisk factors to be discussedare biological characteristics due to inheritance,accident, ilhess, or in uterofactors. Substan- tial research in the alcoholismfield has dealt withinherited risk factors for alcoholism, but less has been done until recently fordrug abuse.Several researchers are beginning to developvulnerability models thatcombine biological and psychosocial variablee (Kumpfer and DeMarsh 1984,1986a; Hill et al. 1985; Huba and Bender 1982). These models also includevariables suchas life streesors, which become triggeringmechanisms for sustained alcoholand drug abuse episodes when theyoverwhelm coping capacities.The Kumpfer and DeMarsh VASC (Values/Attitudes,Stressors, Coping Resources) phasizes primarily psychosocial theory em- factors, whereas the Hill et al.model emphasizes biological factors. A combined "BiopsychosocialVulnerability Model" (figure proposed by Kumpfer (1987b). 2) has been This model includesbiological factors suchas genetics, the prenatalenvironment, and physiological/cognitive they would interact with disorders as environmental factors, suchas family, community, and social environments, toinfluence the child's vulnerabilityto alcohol and drug abuse. The justification for the impoe-nce of addingbiological variables to the model is summarized brieflybelow and can be found in (19876, more detail in Kumpfer A groN, g body of literature suggests thatchildren of alcohol anddrug abusers may differ significantlyfrom other children in their cal, neurophysical, genetic, biochemi- nb,.ropsychological, and physicalmakeup. Because of these differences, theyare hypotheeized to bemore prone to developing alcohol drug abuse problems, thus and perpetuating a cycle. It currentlyappears that theee children are likely todisplay a large number ofpossible biological alcohol and drug abuse. Not markers for all children of alcohol anddrug abusers will develop

.4 329 ALCOHOL AND DRUGABUSE

vulnerability model of these biomedicalpnbler.-,s. The author's some or all largest number ofrisk factors are more hypothesices thatchildren with the (Kumpfer and De Marsh1984). As researchin likely to abusealcohol or drugs to determinewhich risk factorscluster this field progresses,it may be possible together and whether some aremore salientthan others.

Infant-Child- Youth-Adult Physiological *Alcohol or Drug In Utero Temperament Abuse Genetic Variables Variables Variables Cognitions Coping Resources

Community/School Social/Peer Family Environment Environment Environment Values Values Values Stressors Stressors Stressors Coping Resources Coping Resources Coping Resources

Figure 2. A biopsychosoclalvulnerability model.

Genetic Factors inpatient treatment More than half ofthe alcohol anddrug abusers in drug abuse (Goodwin1985; Templer programs have afamily history of alcohol or also noted that thegeneral average for et al. 1974;Cotton 1979). Goodwin for all types ofchildren of alcoholicsis four to five vulnerability to alcoholism fathers may have up times the risk for thegeneral population.Sons of alcoholic probability of becoming anakoholic than sons ofnonal- to nine times greater Cloninger et al. 1981). coholic fathers (Bohmanet al. 1981; early family studies,sibling and half-s iblingstudies Research stemming from Svikis 1986), and twin studies (Kaij1960; Pickens and (Schuckit et aL 1972), Cadoret et al. 1985) adoptive studies (Bohmanet al. 1981;Cloninger et al. 1981; susceptibility to chemicaldependency in children withbiologi- suggests higher Recently, researchershave begun cal parents who werealcohol or drug abusers. biochemical, neurochemical,tempera- to search forbiomedical markers (i.e., markers are ment variables thatexplain the geneticlink). These biomedical discussed below. Reaction to Alcoholand Drugs that children ofalcohol and drug abusers maydiffer This research suggests chemicals (Schuc kit andRayses significantly in their reactionto and tolerance of found to have decreasedsubjective 1979). Adult children ofalcoholics have been 3 5 330 PREVENTION OFMENTAL DISORDERS

feelings of intoxication, decreased psychomotorimpairment (Alpert and unpublished data),increased relaxation Schuckit creased slow alpha (Schuckit andBernstein 1981), in- (Vogel et al. 1979;Pollock et al. 1983), and synchronizingeffect on brain waves" increased "normalizing autonomic (Propping et al. 1981),and increased nervous system reactivity(Kissen et al. 1959). TemperamentVulnerabilities Some researchers (Tarter et al. 1985)have speculated temperament differences that basic biological may predispose a childto alcohol or drug Children of alcoholicparents have been found abuse. be more activeand to have more often than otherchildren to decreased attentionspans, decreased emotional homeostasis, increasedemotional lability, increased gregariousness,and decreased socialinhibition (Tartar1985). NeurologicalVulnerabilities Children of alcoholics have shown additionalneurological differences excessively highfrequency EEG (Gabriella such as alpha activity et al. 1982); deficiencyin slowwave (Propping et al. 1981),reduced P300 amplitude visual evoked potential and latency of the (Porjesz and Begleiter1985), and decreased (Schuckit and Bernstein1981). sleep time NeurochemicalVulnerabilities Several researchersare looking for neurochemical cohol and drug differences betweenal- abusers andnonusers or children of ter levels of serotoninand dopamine alcoholics. Neurotransmit- and other drugs appear to be affected bythe use of alcohol (Kent et al. in press);hence some researchers that lower baselevels ofsome essential have speculated youth to self-medicate neurotransmittersmay predispose a with alcoholor illegal drugs (Goodwin Melchior 1977,pp. 373-430) or carbohydrates 1985; Myers and (Wurtman andWurtman 1986). Several researchershave foundsuppression of alcohol (Murphy et al. 1985)and in humans consumption in animals (Naranjo et al. 1984)with the monoamine reuptakeinhibitors that have use of as Zimelidine and serotonin a high specificity forserotonin, such uptake inhibitors(fluoxetine and norepinephrine uptakeinhibitors (desipramine). fluvoxamine) and NeuropsychologicalVulnerabilities Because academicproblems are oftenrelated to later some researchers have alcohol or drug abuse, begun to look forcognitive dysfunctions account for theseproblems. Children that could help of alcoholics, whohave been found at higher risk foralcoholism (Goodwin to be 1985) and drugabuse (Johnstonet al. 1986), have beenfound in severalstudies to have and overall IQ decreased verbalperformance (Gabriella andMednick 1983), (Herjanic et al. 1979), delayed mentaldevelopment increased left-handednessor ambidextrousness Feldstein and Harburg1982; Nasrallah ;Lee- et al. 1983), decreasedabstraction and 3 '.'f; 381 ALCOHOL ANDDRUG ABUSE

1983), anddecreased ability capability (Nolland Zucker (Goodwin personal problem-solving the MinnesotaCard Sort Test to shiftcognitive sets on communication). Vulnerabilities In Utero noted in children and behavioralrisk factors Many of thebiological, cognitive, features of fetalalcohol syndrome and drugabusers are major mothers. of alcoho! ice (FAE), andinfants born todrug-addicted (FM), fetalalcohol effect studies of childrenof alcoholics previouslymentioned biomedical limit the effectof Most of the alcoholic fathersin order to conducted withonly sons of drug-abusing were If morechildren ofalcohol or FM on theresearch results. might be even moreapparent. used, thebiomedical differences complica- mothers were include low birthweight, neonatal Associated symptomsof FAS/FAE disabilities, andEEG attention-deficitdisorders, learning 1982-1983). It tions, hyperactivity, Finnegan 1976;Stimmel et al. abnormalities (Abel1981, 1982; by the mother,father, or the use ofalcohol or drugs may bediscovered that that can bepassed on tooffspring. grandparents may causegenetic changes EtiologicalFactors Summary of contribute to negative biologicalfactors can This researchsuggests that When thisincreased risk is alcohol or otherdrug abuse. that are not vulnerability to school, orcommunity factors interpersonal family, likely to become combined with development, thechild is more supportive of thechild's positive factors withpsychoso- This interactionof biological an alcohol ordrug abuser. in the fieldofjuveniledelinquency also beenstudied by experts researchers in that cial factors has drug abuse).A number of (often a precursorof alcohol or family is amodest gical factor ofcriminality in the field concludethat the "biol factors" (Pattersonand in comparisonwith other family Cadoret et predictor ofdelinquency studies ((loningeret al. 1981; Stouthamer-Loeber1966). Biological between biologicalfactors demonstrated apositive relationship discussed by al. 1983) have chronic, hardcoredelinquency as particularly for large. and delinquency, although the effectsizes are not very Rutter andGiller (1983), ImplicationsforPrevention proposition that somechildren cited in this papersupports the The research drug abusersbecause of biological vulnerable tobecoming alcohol or and are more The degre4 andtype ofinherited biochemical and psychosocialrisk factors. child.In addition,the vulnerabihty willdiffer 4.br each raised in a neuropsychological damage sustainedby the child extent and typeof psychosocial will also vary. chemically dependent ordysfunrtional parent household with a effectiveness,prevention believes that forthese reasonsand cost This author to high-riskpopulations.These should beprimarily targeted specific needs. programs and tailored tothe participants' programsshould be flexible 337 832 PREVENTION OF MENTALDISORDERS

An idealstrategy for to start with developingtailoredprevention an extensivebiological and interventionswould be DuPont inpress for several psychosocial risk suggested lists assessment (see design specificprevention of riskfactors). One appears to be interventions foreach risk could V-n very costly andto involve factor. At firstglance, this directly associatedwith the agenciee andspecialties not child prevention ofchemical generally psychiatrists andother child dependency juvenile justice development (pediatricians, specialists).However, specialists,special educators, that themost effective a number ofresearch studies preventionprograms are have found (Kumpfer 1987a.,Moskowitz 1986) generally themost intensive populations. and tailoredto the specific needs ofhigh-risk Some prevention specialists(Hawkins et efforts shouldtarget high-risk al. 1987)believe that neighborhoods, prevention than high-riskindividuals schools,or communities high-risk because of theconcern for labeling rather or involving youthand families individualsas if theyturn out tobe false in interventionsthat they do children into positives. Thisis a possible not need a programwithout way to get the prevention specificallylabeling them high-risk field providesprevention high risk.Currently, general schoolor public services tomany false positives programs, namelyall school in population, inexpensive, children andthe general that have obtrusive, andoversimplistie occasionallybeen foundto increase preventionprograms 1988).Involving children alcohol anddrug tions that and familiesfrom the use (Hansen they donot need and general populationin interven- ineffective, and thatmay have potentiallyunethical. negative effectsis costly, Some healthprofessionals vention and may well be afraidto provide preventionservices to needed earlyinter- example, abusedor neglected children withan identified children, children, problem (for learning-disordered emotionallyor behaviorally because of children, anddepressed disturbed some nebulousconcern for labeling. or suicidalchildren) often discussed (Introductory the Rosenthal"expectancy" psychologyclasses never beenreplicated.) If effect, althoughthis research children will, this ill-foundedfear of labeling has in manycases, be denied is not countered, Neither the ch empirically these ild nor theparents need proven beneficialservices. or drug abuse." be told thechild is "at The newprevention high risk foralcohol target high-riskor vulnerable programs createdwith OSAP label them. children andyouth, but funding will Youthare selectedfor these in generalthey donot members ofhigh-risk special services abusers, groups of youth,such because theyare children livingin public as children ofalcoholor drug alreadymanifesting housingor high-density problems(behavioral, housing,children gateway drugusers. academic,or emotional), generally These high-riskchild and advertised tothe teachers, demonstrationprograms are services thatwill help communitymembers, and children inany of these parents asextra adjusted,more skillful,and groups to be more successfulin life. happier, better To date,many alcohol and providing drug abuseprevention education inschools. These programs havefocusedon educationalapproaches 3 :1 are designed to ALCOHOL AND DRUG ABUSE 333 provide information to the general studentbody. They are a good beginning, but it should be obvious from the reviewof risk factors that more intensive interventions are needed to affect high-riskyouth. The next section will review the type of prevention programs thathave been implemented and evaluated. In general, the rarity of adequate evaluationsof these programs makes a critical review difficult. Alcohol and Drug Use Prevention-centeredApproaches Within the PHS model of prevention, three majorapproachesschool-based prevention interventions, community-basedprevention interventions, and family-focused prevention interventionshave beenpromoted by the Federal Government and the States. All three of thesemekjor approaches are reviewed in depth in the author's coedited book,Childhood and Chemical Abuse: Preven- tion and Intervention (Ezekoye et al. 1986),and are reviewed below. School-based Prevention Interventions One third of all prevention programs areconducted in the schools. School- based programs are the primary method for accessingyouth for the prevention of alcohol and drug abuse. Many readers mayremember the lessons on tobacco, drugs, and alcohol that they received in their schoolhealth classes. Today's programs are muchimproved and do not rely on the scare tactics promotedby Harry J. Anslinger and the Bureau of Narcotics,because they were shown to be counterproductive or ineffective (Bukoski 1979; Wepner1979). A recent study demonstrated that knowledge retention in students isbetter in a low-fear rather than high-fear appeal and better with a crediblecommunicator (Williams et al. 1985). As the "back to basics" educational revolution grows,this route of easy access to a large number ofyouth is declining(Adler and Raphael 1983). The percentage of high school seniorsreporting having received any drug education information in schools has declined from 79 percent in 1976 to68 percent in 1982a surprising result given the increase in funding forprevention in recent years. Of these students, only 20 percent reported receiving aspecial course on alcohol and drug abuse; 59 percent of these studentsreported that the course was valuable (Johnston et al. 1985). A wide variety of school-based programs arebeing implemented and tested in the Nation. These approaches have been categorizedby Bukoski (1986) into five domains: incognitive, affective/interpersonal,behavioral, env;ronmental, and therapeutic. These will be discussed brieflyunder the headings of cognitive, affective, alternative, behavioral, community, andfamily-focused programs. Knowledge about Alcohol and DrugsCol titive Programs Programs influencing the cognitive domain focus onincreasing the students' knowledge about (a) pharmacological effects of alcohol anddrugs; (b) health and social causes and consequences of abuse; (c)school and community attitudes and norms and legal sanctions; and (d) generalhealth education.These 3 ,1 9 334 PREVENTION OF MENTAL DISORDERS

programs often consist of films and didactic instruction by classroomteachers or health educators.Occasionally, law enforcement officialsor physicians present to the classes legal realitiesor health consequences.Reality-oriented assemblies presented by ex-addicts and drugparaphernalia displays have been found to be ineffective but oi..^asionallyare still used. A teacher who passes around a joint to a class should not besurprised to get two back. These educational programs are generally effectivein increasing students' knowledge about alcohol and drugs, but whetherthey have any impacton decreasing or delaying the onset of alcoholor drug use is not known, because most educational programs do not include evaluation of behavioral objectives(Moskowitz 1983). Proponents of these programs face thefollowing issues: 1.The Knowkdge Attitudes l IntentionsBehavior Theory. This theory, the primary theory of change underlyingthese programs,assumes that a change in knowledge will affect attitudes,intentions to use, and eventually behavioraluse but has never been empirically demonstrated.A number of studies have providedlittle empirical support for this theory (Wallack and Barrows 1981; Goodstadt1981). Some researchers have been able to demonstrate knowledgeretention without changes inattitudes, intentions,or behavior (Schaps et al. 1982; Moskowitzet al. 1984b; Williams et al. 1985); otherresearchers have been able to demonstrate attitude changes butno change in intentions or behavior (Schlegel and Norris 1980). In addition,some researchers have found changes in intentions andopposite changes in behavior (Kumpferand DeMarsh 1986a). The author's study foundyouth in one condition (the Family Skills Training groups) thatsignificantly increased their intentions to use drugs (like theirdrug-using parents), but those whowere users significantly decreased theiruse. A number of studies found that intentions to use drugs accountedfor little of the variance in subsequent use after controlling for prior use (Bentler andSpeckart 1979; Huba et al. 1981). 2.Decreased exposure to educationalprograms by high-risk youth. Many high-risk children, including childrenof alcohol or drug abusers, donot regularly attend school andare likely to "skip out" on an alcoholor drug lecture or film. In addition, thesechildren drop out of school earlierand are lost to the pretest or followup evaluationof outcomes because of frequent family moves. 3.Quality control. When teachers,students, or trainers withmany different attitudes toward alcohol and druguse are implementing theprograms, quality control can be difficultto achieve. Process evaluationsare rarely conducted to determine whetherthe trainers are implementingthe program as specified. 4. Weak programs. Most ofthese programsare not intensive or enduring enough to have much impactunless their goalsare very simple. One

(). ALCOHOL AND DRUGABUSE 335

study has found thatlonger programming maynot be necessary if the simply presents the goals are to increaseknowledge and the curriculum facts. Schlegel et al.(1984) found that studentsexposed to a three-session posttest and 8-month facts curriculum hadlower alcohol consumption at followup than students whohad the facts sessions plusfour more sessions of values clarificationand decisionmaking skills. ineffective be- 5. Inappropriate programs.Cume educational programs are developmentally or culturallyappropriate. Prevention cause they are not and fit programs shouldbe designed to match thechild's cognitive level the child's cultural andethnic traditions. have had Although few of theseknowledge-oriented prevention programs mewurable impact on alcohol ordrug abuse, some reeearchersstill support this approach and say that moreresearch is needed. It is theauthor's opinion that the short-term negative consequencesof alcohol or providing information about experimentation with these drug abuse may help to deterlow-risk youth from substances. This effect willhelp to reduce stress on anumber of families that the highly worry that theirchildren will become drugabusers. Knowledge about addictive qualitiee of somedrugs (crack or deeignerdrugs) may be more important, since even low-riskyouth could becomeaddicted to these drugs through occasional use. The author believes thateducational proib-ams couldalso be used to help high-risk youth understandthe behavior ofadultstheir parents, grandparents, relatives, orfriendswhen they drink toomuch.Low-risk children can also be taught tobe supportive and helpful tochildren lees fortunate than themselves who areliving in homes with chemicallydependent parents. Teachers could betrained to recognize the signsof a child of chemically dependent parents and couldthen refer the child for help orprovide extra understanding and support for thechild at school (Ackerman1983). Affective and InterpersonalEducation Programs Affective and interpersonalprevention activities attemptto indirectly abuse by increasing thechild's self-concept, under- prevent alcohol or drug of the standing of feelings andinterpersonal relationships, and awareness communication and decisionmaking processes.A wide variety of programs are grouped under this approach.Some programs involveyouth in discussion groups or peercounseling sessions aboutfeelings, decisionmaking, problem- solving, communications, or valuesclarification.Occasionally, professional consultation is provided toindividuals or groups. Evaluations have found little supportfor the effectiveness ofthis approach attitudes, or delaying the in decreasing adolescents'intentions to use, changing Goodstadt and Sheppard 1983; onset of use (Huba et al.1980; Goodstadt 1980; Moskowitz et al. 1984a, b). TheNapa Project, which was basedprimarily on eighth grade this model, demonstrated noeffects on seventh grade males or males or females. A positiveeffect on use was found foralcohol and marijuana, 341 336 PREVENTION OF MENTALDISORDERS

but not for cigarettes among seventh grade girls at posttestand not at a 1-year followup (Schap@ et al.1982; Moskowitz et al. 1984a,b). Another combined values decisionmaking program that with social competencyskills training ina 12-sessionprogram found a positive effecton tobacco use at posttest, but negative effect a on alcohol use and no effecton marijuana use (Gersick et al. 1985). Johnson et al.(1986) reported negative by a decisionmaking effects on alcohol andmarijuana and social competencycurriculum compared withtheir social influences,peer-resistance trainingprogram. that both It is not clear, however, programs were implemented with thesame degree of enthusiasm. These studies show conflicting resultson different substances, but overall these affective education programs have not been provedvery effective given limited programmingfunding and time. The these programs author believes that thegoals of are laudable, but theyare too far removed from the goal reduced alcohol and drug of abuse. More intensiveprogramming for high-risk youth whoare actually found to be deficientin these skills productive. may be more Alternative Programs

This type of approachto prevention is basedon the theoretical assumption that providing youthwith "alternative highs"or skill- and competency-building activities will !educe alcohol and drug abuse.Schaps et al. (1981) found review 'of 122 prevention in a evaluation studies that only12 dealt with alternative progremarning. The effectiveness of this approachwas equivocal. Five of the programs had positive outcomes, butseven reported no program impact. et al. (1984) found that Stein the Channel One nationwideprogram, in which many high-risk youthwere involved in community recreation and businessprojects, failed to prevent theshort-term increases in of some substances frequeacy of drunkennessand use expected in high-riskyouth. However, it ispossible that these programswere actually effective in reducing alcohol and drugs and high-risk youth'suse of that this result wouldhave been shown if ing youth were compared the participat- with similar high-riskyouth in controlgroups. When evaluating preventionprograms with youth already drug problems, manifesting behavior and one must remember that thesetroubled youth often rapidly escalating base have a rate of problem behaviorsand that any decreasein this rapid increase of druguse is progress. However, in bring a number of alternative ; rograms that gateway drug-using youthtogeti.er, one must alsoconsider the contagion effect.An experimentalevaluation of three of the sites showed slight Channel One positive findings forimproved democraticproblem solving and participation inalternatives, but increaseduse of inhalants, hallucinogens, and alcohol (Hu et al.1982). An important discoveryby Swisher and Hu of the contradictory (1983) that helps to explainsome findings for this approachis that some alternative promote decreased activities use and others promote increaseduse, depending on the social environment andpeople associated with ment, sports, social, each type of activity.Entertain- extracurricular, and vocationalactivities are associated 3 9 ALCOHOL AND DRUG ABUSE 337 with increased use of alcohol or drugs; academicand religious activitiee and active hobbies are associated with decreased use. Murray and Perry (1986) are in the procees of evaluating analternatives program that is based on thefunctional relevance of drugs for youth at different ages.They have found that transition marking and socialacceptance are important functions of alcohol and drugs among youngeryouth, whereas stress reduction appears more important among older youth.All the youth used alcohol and drugs to enhance personal eaergy, for recreation,and for relief from boredom or loneliness. Murray and Perry are evaluatingtheir school-based program, Amaz ing Alternatives,which helps youth to identify health-enhancing alternative activities for each functio:, aerved for themby alcohol or drug use. Behavioral Prevention Programs These programs attempt to behaviorally train students toresist peer pres- suree to use tobacco,alcohol, and marijuana through social learning, reciprocal determinism, and efficacy theory (Bandura 1977, 1986). Thecurrently popular social competency apprcaches to prevention includethree different strategies: (1) the "social influences" approach as found in socialinoculation and peerresistance social skills training strategies promotedinitially by Evans et al. (1978, 1981), (2) modeling or training health-promotingbehaviors, and (3) the broader "life/social skills" approaches promoted by Botvin andEng (1980, 1982). Social influences prevention programs have been well supportedfiscally and have been researched by many well-known preventionspecialists, for example, McAlister and his Stanford associates (1979, 1980); Perry and herMinnesota associates (1980a, b; 1983a, b); Tell and associates in Oslo(1984); Vartianen and associates in North Karelia, Finland (1983); Schinke,Gilchrist, and their Seattle associates (1983a, b; 1986); Pentz and her Tennessee orUniversity of Southern California (USC) associates (1982, 1983); Johnson and hisMinnesota or USC associates (1984,1985); Flay and his Waterloo or USC associates (lle2, 1983a, b, c, 1985); Fisher and his associates in Australia (1983);Dielman and his M. ,thigan group (1984a, b); and Biglan, Severson, and colleagues inOregon (1986). These programs appear to have some effectiveness indelaying onset of tobacco use in junior high school studentc, although Moskowitz(1986) pointed out that the "pattern of effects is inconsistent acrossstudies even of the same program. Some studies find reductionsin new smoking; others find reductions in experimental or more regular smoking." The reason for thedelayed onset is unknown bece.-le of the many mediating variables in these programs. Some researchers have speculated that if real changes are occurring in these programs, the changes may be due toreinforcement of existing school norms against tobacco (Kumpfer et al. 1986), to recent changes in the social climate against smoking cigarettes, or to the informal social control climate (Perry et al. 1980a, b; Rodin 1985, pp. 805-882; Mcekowitz 1983; Polich et al. 1984).

343 338 PREVENTION OF MENTAL DISORDERS

Botvin and McAlister (1982) reported that studentinterviews revealed a positive change in the entire social atmosphereregarding smoking following their life-skills antismoking intervention. Because of increasedawareness of tobacco's health risks, due to the antismoking campaign, beginning with the 1964U.S. Surgeon General's Report, many Americans have decreased theiruse of health-compromising mild stimulants such as tobacco. This changemay not have been as difficult as for other drugs becausemore acceptable mild stimulants are availablecaffeine, sugar, and zantheneto serve a similar function. In fact,the caffeinated soft drink market has increased significantlyin recent years, and sugar and choco- late consumptions are high. Hence, antismoking"say no" programs did not have to address the basic functions of nicotine. Braucht(1980) and Murray and Perry (1985) have stressed thatwe need to understand the different psychosocial functions played by alcohol and otherdrugs before effective prevention programs can be designed. The author believes thatin designing the best replacement strategy we also need tounderstand the biological functions that alcohol and drugs play. There is growing concernamong prevention researchers and practitioners about the high percentage of preventionresearch funding supporting this single prevention approach when few empirical datasupport ita effectiveness (Bell and Battjes 1985). A preliminary evaluationof the application of this approachto alcohol and marijuana preventionin the USC Project Smart suggestsdisap- pointing outcomes; thisprogram did not prevent significant numbers of children from using alcohol or marijuana (Johnsonet al. 1985). Recently, prevention practitioners and researchers expressedconcern about the real versus the statistical effectiveness of theseprograms, even with antismoking applications. For instance, because of the highattrition (25 percent to 64 percent) in these school-based programs, the often-cited50-percent reduction in youth who begin use can mean very few youth when percentagesare translated to actual numbers. Hence, because of their highcost and intrusiveness into schools' daily academic schedules, theseprograms do not appear to be cost effective. The highest risk youth are often missingfrom schools and do not receive thebenefits of the continuity of these school-basedprograms. Also, these programs rarely report what happens to high-riskor other youth who are already using tobacco, alcohol, or drugs.It is entirely possible that theseprograms are having a negative impact by increasing alcohol,tobacco, or drug use in youth whohave already tried them.

In a recent review of thisprevention approach, Flay (1987)concluded: "Overall, the findings from themost rigorous studies to datesuggest that the social influences approach to smokingprevention can be effectivesome of the time." However, this conclusionseems somewhat fragile, given the considerable differences between studies in thepatterns of reported results. Also,at least two plausible alternative interpretations of the reported effectsnamely,sen- sitizing effects of the pretest testing(or screening), and the HawthorneEffect remain because many controlgroup schools receh,, no specialprogram. In 3 4 4 ALCOHOL AND DRUG ABUSE 339 addition, Moskowitz (1986) pointed outthat "the effect sizes and significance levels reported in these studies arebiased and cannot be trusted asthe statistical analyses were notconducted on the units assigned toconditions, schools nr classrooms, but rather onstudents" (Big lan et al. 1985; Cook1985; Moskowitz 1983). So it appears that eventhe fourth generation studies suffer from =Or methodological weaknessesand that claimed positive effects ofthe "say no" skills-training programsneed to be evaluated more closely. This author does not believe thatsuch a simplistic approach can be very effective with high-risk youth, whoby this pc in their development will have multiple reasons to "say yes" to drugsand alcohol. These programs do not address the 'I'd 'n reasons whyvulnerable youth use drugs. A numberof researchers have also suggested that the peer pressurenotion as the primary determinant of alcohol and drug abuse isoversimplistic (Moskowitz 1983; Eiser and van der Pligt 1985; Sheppard etal. 1986). A young student representative of the Say No Clubs recentlyexplained at a Chicago conference that"peer pressure is more likeAdidas. If you are the only one withoutthem, you want them." This doesn't sound much likethe type of peer pressure promoted bythe social inoculation program providers orresearchers who present a picture of drug-using youth badgering nonusingyouth to 'just try it, you'll like it." A recent research study by Shope andDielman (1985) fourd that most fifth and sixth grade students already had sufficientrefusal skills (or saying-no) skills before the program and that a four-sessionsocial skills training curriculum did not decrease their alcohol consumption. Community-based Prevention Interventions Public Media This approach to prevention isNMof the =Or methods for providing prevention for adults. The primary techniquesused are mass media or public awareness campaigns.A coordinated media campaign couldinvolve one or all of the following: radio, television, newspaper,magazine, billboard, and poster consumer awarenessadvertisements; special face-to-face presentationsin workshops, classes, conferences, and campaigns;and supporting publications such as pamphlets, books, films, videotapes, and direct-mail flyers.This prevention strategy is one of the earliestused in this country. The goal of this approach is generally to provide increasedinformation about the health consequences of alcohol and drugabuse or use. Unfortunately, evaluationsof these public awareness programs, particularly thosethat employed scare tactics, have not been encouraging (Blum et al. 1976).More research suggests that information-only programs can be more effective ifthey communicate in a straightforward way the adverse effects of drugs, usecredible communicators, target behavioral changes needed to support nonuse,and influence the recipients' perceptions about the acceptability ofaohol and drug use or abuse (Durrell and Bukoeki 1983; Polich et ai. 1984). 340 PREVENTION OF MENTAL DISORDERS

A primary weakness of theseearly information-only preventionprograms was that they assumed a knowledge (attitude) behaviormodel. A considerable body of research suggests thatincreased knowledge has little effecton attitudes and that changes in attitudes seldomlead to behavior changes (McGuire1969). In fact, the morecommon scenario is that changes in behavior leadto changes in attitudes. Fishbein and Ajzen(1975) and other social psychologistshave discovered that attitude changecan influence behavior change if two conditions are satisfied: (1) attitude is specifically about thebehavior to be measured, and (2) other social and environmentalfactors support the attitude-behaviorchange relationship (Rockeach and Kliejunas1972; Wicker 1971; Ostram 1969). The conclusion about mediacampaigns is that they do provideneeded information and do affect thecommunity's socialnorms in the long run when combined with other communityprevention strategies. In addition,the public demand for credible informationabout alcohol and drugs isincreasing and should be satisfied by accurate andscientifically crediblemessages. Parent Groups

Increasingly, prevention plannersare advocating efforts to createa community climate ofnonuse of drugs (Durrell and Bukoski 1983).The current involvement of parents'groups in promoting community preventionof both drugs and alcohol is significantThese ix-rentalgroups generally focus on one or more of three goals: (1) changes in thehome to counter prodrugmessages, (2) changes in the youth'ssocial environment, and (3)community awareness campaigns. Actually, because of thespontaneous nature of these parentgroups, little is known about the actualactivities of these parentgroups and their effectiveness. NIDA recently fundeda descriptive study of these national parent groups in hope of better understanding theirimpact (Klitzner 1984). A national example of the usefulness ofparent groups in macro-approachesto prevention are Chemical People, the Cottage ProgramInternational's Family Friendship Circlee, and the National Federation of Parents for Drug Free Youth.The focus of theee macrocommunityprevention efforts ison creating a climate in which children are getting "don't do drugs"messages from respected adults andpeers in schools, media, and thecommunity at large. Community Groups Over the years, a number of public service clubs, religiousinstitutions, corporations, and privatenonprofit organizations have helpedin the fight against alcohol and drug abuse.Civic groups suchas the Junior League, the Lions Club, Service Clubs ofAmerica, and the Rotary havesponsored their own alcohol and drug prevention programs or worked with State and localofficials to enhance the effectivenessof State or nationalcampaigns. A number of burinesses have volunteeredto sponsor high-risk youth inthe Channel One progre_ms.

3,16 ALCOHOL AND DRUG ABUSE 841

Family-focused PreventionInterventions Importance of the Family in Alcoholand Drug Abuse Prevention Prevention specialists are beginning torecognize the valuable resource of parents and families in increasingthe effectiveness of alcohol anddrug abuse prevention programs for youth.The author's recent research withchildren of alcohol- or drug-abusing parents(DeMarsh and Kumpfer 1986; Kumpferand DeMarsh 1986a) supports the fact thatthe family is highly involved inthe genesis and maintenance ofchemical dependency and can be veryinfluential in alleviating risk factors in children. Prevention Programs for Children ofChemically Dependent Parents These results suggest cost-effective,family-based prevention interventions must be developed to preventdrug and alcohol abuse in childrenof alcohol and drug abusers. Few family-focusedalcohol and drug abuse prevention programs have been developed for high-riskchildren, even though the applicationof parent training and family skillstraining programs to other problemsin children has been highly effective (Pattersonet al. 1975; Miller 1975; Gordon 1970; Dinkmeyer and McKay 1976;Forehand and McMahon 1981; Guerney 1964; L'Abate 1977). One reason forthis effectiveness is that parents canbe trained to be effective change agentsand their effect will be enduringand powerful. In the past 10 years,NIDA has supported sevbAtl family-focused prevention programs, namely theauthor's Strengthening Families Program (DeMarsh and Kurnpfer 1986; Kumpferand DeMarsh 1986a) for children of drug abusers in treatment, Alvy'sConfident Parenting Program (1986) for parents of black youth, andSzapocznik'a Family Effectiveness Training(1983, 1985, 1987) for parents of high-risk Hispanicadolescents. The Kumpfer and DeMarshStrengthening Families Program Once empirical findings supported thecharacteristic differences between families and children who do and do notabuse drugs, the author developed and teeted three different types offanilroriented prevention programsa parent training program, a children's socialskills training program, and a familyskills training programto determine theireffectiveness in reducing the children's risk factors. Preliminary analyses of the pretestand posttest data suggest that all three programs were successful inreducing the risk factors in the children, although each program's effect depended onits intended goals. Hence, the behavioral parent training program wassucceseful in reducing the children's problem behaviors and improving the parents'ability to discipline the children; the family skills training programimproved the family relationships and some of the children's problem behaviors;and the children's social skills program improved the children's social skills.Only in the complete Strengthening Families Program, which combines all threeinterventions, was alcohol and drug use actually decreasedin the older children. A proposedlongitudinal study is needed to determine which of these changeswill be most successful in the long run in preventing alcohol anddrug abuse in the younger children.Another 342 PREVENTION OF MENTAL DISORDERS

important finding of this researchis that regardless of theparents' dys- functionality, most parents can be coached and assisted in develr lingmore effective parenting styles that willaffect risk factors in their childre... Parent Traii,g Program Results Following exposure to the parenttraining program, parents reportedhaving fewer problems handling school-agechildren and demonstratedincreased knowledge of child behavior management principles. This increasedknowledge and improved parent disciplineeffectiveness had direct impacton the behavior of the children, who were reported to scream less, have fewertemper tantrums, get angry less, improve their homebehaviors, and display fewerproblems than other children theirown age.Further, the childrenwere reported by their parents to be happier, to like schoolbetter, and to increase theiroutside activities. The children reporteda significant decrease in intention to smoke and to drink, but not touse drugs. Family Skills Training Program

This program combines threedifferent phasesa filial playtherapy phase called Child's Game, in whichparents are taught to enjoy thechild; a family communications training phase; andan effective discipline trai ning ph Parent's Game. Preliminary ase called outcome effectiveness evaluationsindicate improvements in three theoreticallyspecified areas thoughtto influence a child's risk status:(a) family functioning, (b)children's behavior problems, (c) children's expressiveness. and Family functioning seemedto improve on several dimensions ticipation in the family skills following par- training program, includingincreased family communication of problems,improved relationsamong siblings, improved ability to think of family-oriented activities, clarity of familyrules, and more social contacts by parents.

Likewise, improvements inchildren's behavior problemswere found. Parents reported that their children behaved lessimpulsively, weremore well behaved et home, and hadfewer problem behaviors ingeneral. Also, children self-reported improved relationswith peers. The family skills training component of the StrengtheningFamilies Program curriculum alsoappears to have affected theor to express themselves .,t to which children were able within the family context,verbally and otherwise. Children asked formore help with their homework, talked they felt sad, sought to people more when more attention from parents, andcried more. As in the parent training program, thechildren reported atposttest significant decreases in their intentions touse tobacco and alcohol, but not drugs. Obviously, the author believesthat this approach hasconsiderable promise for reducing alcohol and drugabuse in high-risk youth. a parent training program, The positive effects of a family skills trainingprogram, and a children's social skills training program need to be tested for durabilityin a followup study,

3 4 .,:1) ALCOHOL AND DRUGABUSE 343

populations (minorities, and the program effectivenesswith other high-risk conduct-disordered children) needsto be tested. Family Therapy prevention or early A number of clinicians areusing family therapy as a intervention strategy forhigh-risk children and youth.Klein et al. (1977) strategy for discovered that functional familytherapy is an effective prevention breakthrough in this area.Hence, younger siblingsof delinquents, a major family therapy is a preventionstrategy to the degree that itreduces risk factors dependent members of the family(the spouse and for nonusers or nonchemically for the children).Unfortunately, many publicand private funding sources alcohol and drug abuse prohibitpayment for anyone but theabuser. Family therapy can take manyforme.Gallant et al. (1970) have experimented with the use ofmultiple-couple groups. Steinglass(1975, pp. 259-299) briefly employed the free useof alcohol (while being videotaped) in the initial assessment stages of anexperimental inpatient program atNIAAA's Laboratory for Alcohol Research tohelp the staff and couplesbetter understand alcohol's role in family dynamics. InUtah, the Teen Alcohol and DrugSchools One havo successfully employedmultiple family therapy principles for years. unique feature of this programis switching childr.,n and parentsfor initial communication exercises.It often appears easier forpeople to practice new skills with different parents orchildren. therapist or To work successfully withchemically dependent families, the trainer needs thoroughunderstanding of the dynamics andtypical developmental stages of these families(Steinglass 1980; Kaufman 1980).The pioneering work of Wegscheider (1981)and Ackerman (1983) in familytherapy with alcoholics has promotedunderstanding in this field. Sincerecruitment is often a problem,Szapocznik et al. (1986) haveexperimented with one-person family therapy and found it effective.In addition, Szapocznik hasdeveloped culturally relevant family therapy for Hispanicfamilies of Cuban descent.Maldanado and his associates have successfullyused their family model as anearly intervention strategy for Hispanic (Spanishand Mexican descent) first-offenderyouth and as a preventionstrategy for siblings (Courtney1984; Kumpfer et al. 1985). Family Self-Help Groups Self-help groups, such as Al-Anon for spousesof alcoholics, and Alafam for families of alcoholics, areincreasing in popularity. Morethan 5,000 Alafam groups existthroughout the world, making themthe single largest prevention program involvingfamilies of chemically dependent persons.These groups closely parallel Alcoholics Anonymous.According to Ablon (1974), they teach the basic lesson that alcoholismis a disease, along with threebasic principles: loving detachment from the alcoholic,reestablishment of their own self-esteem and independence, and reliance on ahigher power. Through shared experiences, the groupsteach that many families have the sameproblems and that they are not alone.Some treatment clinics forchemical dependency also involve 349 344 PREVENTION OF MENTAL DISORDERS

spouses (para-alcoholies) or families ingroups. The author (Kumpfer 1975) specializes in prevention groups for wives of chemically dependentmen using a specially structured curriculum.Through increased understanding,these spouses and families not only increase theprobability that the alcoholicwill achieve and maintain sobriety,but also decrease theirown risk of becoming alcoholics.(For further informationon para-alcoholics and co-alcoholics Greenleaf 1981.) see Multicomponent PreventionPrograms All of the previously discussedprevention approaches involvea single strategy. More recently, preventionprogrammers and planners have become excited about the possibleinteractive and enhancingpotential of sustained, integrated, multicomponentcommunity programsor campaigns. School-based programs are limited in their influenceon youth because the majority ofan adolescent's time is spentoutside school (at home, watchingtelevision, in the community, and so forth). Inaddition, the highest risk youthare the least likely to be at school, and significant onset uf alcohol and drug abuseoccurs after high school graduation. Hence,an optimal preventionprogram would incorporate mass media, community organization, andfamilies as wellas schools. Evaluation Evaluations of multicomponentprevention programs have been and suggest that by combining encouraging a number of programsa synergistic effect may increase the effectiveness of all components. Flay andhis associates (19836, c) combined a school and mediasmoking-preventionprogram with written homework assignments to becompleted with parents. Thejunior high school youth in thisprogram were only half as likely to 2 months between start smoking during the pretest and posttest.Significant secondaryeffects were observed in parents involvedin the program, with 35 and 69 percent attempting percent quitting smoking to cut back or quit. Thisresearch did not test the individual components; hence,it was not possible todetermine the efficacy of the multicomponent approach.

Bien and Bry (1980)employed a dismantlingdesign to improve academic performance (one of the covariants of alcohol and drug abuse).They found that only the seventh graderswith all threecomponents (teacher conferences, student goal-settinggroups, and home notesor calls to parents) had significant improvements in grades andattendance overa no-intervention group. This research tends to support the increased effectivenesshypothesis of the multicomponent prevention strategy;however, it is possible thatthe most important component is the family involvementadded last. Health PrnmotionProgram Additio.adlsupport for the multicomponent communityprevention approach can be gleaned from the heart diseaseprevention and health promotion munity programs: the Oslo ..orn- Study (Hohne et al.1982), the Stanford Tiree 359 ALCOHOL AND DRUG ABUSE 345

Community Study (Maccoby 1976; Solomon 1982, pp. 308-321), the North Karelia Project (McAlister et al. 1980, 1982). The Stanford Three Community Study demonstrated that the combination of in-home or group prevention sessions with public media significantly reduced the risk of heart disease, increased knowledge (Meyer et al. 1980), and decreased smoking and number of hypertensives (Farquahar et al. 1981) when comparing intervention with no-intervention control communities. Intermediate results were found for the media-ontr communities at 1- and 3-year followups. Likewise, the intensive communhy prevention programming in the North Karelia Project in Finland suggest3 that multicornponent programs may have synergistic effects. More North Karelia residents who viewed the nationally televised smoking-cessation programs ceased smoking than persons in unorganized communities. A number of new community heart diaease prevention programs are currently under way in the United States: the Stanford Five Cities Study (Far- quahar 1978), the Minnesota Heart Health program (Leupker et al. 1982), the Pawtucket Heart Health Program (Elder et al. 1982), and the Lycoming Com- munity Health Improvement Program (Felix 1983).These community programs seek to be more integral to the communities by involving key community leaders and fewer outsiders to run programs. Unfortunately, none of these prevention programs include an evaluation design that can demonstrate the relative effectiveness of the individual program components. Several community-based drug prevention programs have been developed in San Francisco (Wa Hack and Barrows 1981); Charlotte, North Carolina (Kim 1981, 1982); Ventura, California (NIDA 1982a, b) and Seattle, Washington (Resnik 1982). According to Johnson (1983), it is "not clear that any meaningful evaluation of these projects will be forthcoming." Environmental Prevention Approaches Some prevention approaches, particularly in the prevention of alcoholism, alcohol abuse, and alcohol problems (alcohol-impaired driving, creshes, illness), are concerned with community environmental changes as mentioned earlier for informal and formal social controls. For instance, specialists in the alcohol field have advocated numerous environmental and regulatory approaches to prevent alcohol-related problems. Recommended measures include regulating the content of alcoholic beverage advertising (Mosher and Wal lack 1981), increasing the accuracy of portrayals of the consequences of alcohol use in the mass media (Wallack in press), increasing counteradvertising through industry funding (Wallack 1984), increasing excise taxes and price (Mosher 1982; Grossman et al. 1984), and decreasing availability by (a) increasing the minimum age for legal purchase (Wagenaar 1981, 1982; Vingilis and DeGenova 1984; Williams and Lillis 1985); (b) reducing the number of outlets selling alcoholic beverages for off-pomise consumption (MacDonald and Whitehead 1983; Hooper 1983); (c) eliminating alcoholic beverage sales from service stations; and (d) restricting sales at public events (Wittman 1985). 346 PREVENTION OF MENTAL DISORDERS

Moskowitz (1986) recently completed a review of these community environmental prevention approaches. Conclusions in this field are hindered by the lack of experimental designs, comparable control groups, lack of data on mediating variables, and the absence of process or implementation data (Judd and Kenny 1981); however, it appears that some public policy changes can have modest impact in reducing alcohol problems. Moskowitz (1986) stated that: "A substantial body of well-designed research indicates that increasing the minimum legal drinking age to 21 is an effective means of reducing alcohol-related automobile crashes, injuries and fatalities among the affected age group." Wagenaar (1983) estimated that 20 percent of all alcohol-related crashes and 13 percent of all fatal crashes involving young drivers can be prevented (Arnold 1985). The effect of increased cost of alcohol on decreased use is unknown because of the substantial overlap of differences in community values or informal social controls with the amount of excise tax on alcohol. It does appear that if increased taxes are added to liquor, then youth and adults will switch to lower priced beer. Beer is now nimilar in price to soft drinks, because from 1967 to 1983 the price of alcoholic beverages increased only half as much as nonalcoholic beverages (Cook and Tauchen 1984). Alcohol-impaired driving laws appear to have a short-term effect, depending on the enforcement of the laws and the degree of mgdia coverage concerning the crackdown. This policy approach is more costly to implement than thepreviously mentioned policy approaches of increased minimum age end increased taxation. The increased taxation approach has an additional benefit because the general State funds raised through this policy can be earmarked, as some States have done, for alcohol and drug abuse treatment and prevention programs. Another secondary or tertiary prevention strategy designed to reduce alcohol problems and alcohol-impaired driving is server intervention (Mosher 1983). O'Donnell (1985) estimated that about half of all alcohol-impaired driversare driving to or from licensed on-premise establishments. Server intervention involves developing new policies and training managers andservers to refuse service to intoxicated customers and to increase profits by promoting food and nonalcoholic drinks. Two evaluation studies ofserver interventions are in progress, but have no results to date (Saltz 1985). Most of these formal control prevention strategies are appropriate only for the prevention of abuse of a ' nal substance like alcohol. Theyare less applicable to the drug abuse prevention field except for over-the-counter and prescription drugs. Also, because these environmental and regulatory approaches tend to legislate personal choice, they often face public and private resistance (Bell and Levy 1984, pp. 775-785) and raise ethical and moral questions (Roffman 1982).

3 5 2 ALCOHOL AND DRUG ABUSE 347

Conclusion

Public Policy Issues Alcohol and drug abuse, dependency, and problems are asignificant drain on the Nation's health and economic well-being.Despite the significant drop in use of alcohol and drugs by youth in this Nation since1981, a substantial percentage of adults are using drugs. The commitmentin this country to the prevention of alcohol and drug abuse has been limited toverbal commitments rather than financial commitments. Although alcohol anddrug abuse in 1986 cost each person in this countryabout $850 (economic cost estimated from lostproduc- tivity, treatment costa, societal costa in accidents,fires), only 77 cents per person was spent on alcohol anddrug abuse prevention. Public concern appeared to change dramatically in fall1988, and Congress authorized several billion dollars for the prevention ofalcohol and drug abuse. Unfortunately most of this money will be spent on supplyreduction techniques that have demonstrated little effect on actual usepatterns. According to Johnston (personal communication), 200 times morefunding is going toward supply reduction as compared with demand-reductionstrategies of prevention. The Rand Corporation Report on Strategiesfor Controlling Adolescent Drug Use (Polich et al. 1984) detailed why the use ofsupply-reduction interventions is ineffective in decreasing drug use andprobably have pushed youth into using more harmful drugs. In addition, a very small amount of the newfunding will be targeted to high-risk children and youth as advocated in this paper.Only about $24 million of the total amount will be used to funddemonstration/evaluation projects for comprehensive prevention and treatment, targetedprevention projects, and early intervention projects for high-risk youth.The creation of OSAP, which will administer these new targeted preventionprojects, is a step in the right direction. It is hoped that many new and innovative programswill be created and evaluated for dissemination by this funding. Cost Benefit of Prevention Programs Cost-effectiveness or cost-benefit analyses of current alcohol anddrug abuse prevention approaches need to be conducted as advocatedby the AACAP's prevention project.Currently, it is difficult to determine cost effectiveness, because effectiveness for many prevention approaches is notwell established. However, gross estimates are possible and should be us.ed inadvance of establishing a prevention program to determihe whether the projectis worthwhile. Little is published on the program costs from which to estimate costeffectiveness. In 1977, NIDA published someinformation that could be used to determine gross estimates of cost effectiveness. NIDA estimated, atthat time, that to be cost effective, information or media campaigns needed toaffect 0.15 percent of the participants, education programs needed toaffect 6.25 percent of the 348 PREVENTION OF MENTAL DISORDERS

participants, alternative programs needed to affect8.53 percent of the participants, and interventionprograms needed to affect 4.48 percent of the participants. Research evaluations rarely publish their effectivenessrates and compare them with the costs of continueduse of alcohol or drugs or unexpected benefits. The 1986 estimated yearly cost to society ofa dysfunctional drug user is about $8,000, up from about $6,200 in 1982.Hence, this figure on estimated cost of dysfunctional use needs to be indexed eachyear. Cost-effectiveness evaluations are obviously the b )st indicator of whether the programwas really the most cost effective, because outcomesare compared across several programs with the same type of participants. However, cost-benefit analysesare possible with programs that have only one experimental group andone randomly assigned control group. In addition, this type of analysishas the added advantage of allowing the evaluator to add other spinoffor unexpected benefits to society. For example, the author has conducteda cost-benefit analysis of an alternative job skills building preventionprogram for high-risk youth. This program proved to be cost beneficial only when the value of theyouth's community homebuilding projects was considered.

Cost-effective programsare likely to include the following elements:community volunteers working with youth in communitysettings; strategies that target high-risk youth;messages that stress healthy lifestyles and focuson short-term healthconsequences; integration of enduring, coordinated, and pervasive strategies that addressmany of the environmental domains addressed in this paper. Recommended Directions In a presentation to the AACAP in 1986,the author discussed several promising directions for the prevention of alcoholand drug abuse (Kumpfer 1986). The most important pointsare summarized below: 1.Prevention interventions need to be tailoredto the intended audience, taking into account age,race, culture, gender, and life circumstances. 2.Prevention interventions need to be basedon the best known etiological factors contributing toor protecting youth from alcohol and drug abuse. 3. More etiological research is neededto track high-risk you th longitudinally from childhood into the drug-usingadolescent and early adult stages of their lives. 4.Increased prevention effectivenessresearch is needed for a broadrange of prevention interventions, ratherthan the current strategy of putting most funding into education-based approachesusing one mEkj or strategy "say no" skills training.

3 5 .1 ALCOHOL AND DRUG ABUSE 349

5.Cost-effectiveness and cost-benefit analyses (even if verycrude) should be conducted to determine whether development of a particularprevention strategy is likely to be worthwhile.This type of analysis may help prevention designers to consider ways to cut costs andincrease benefits.

6.Preferred prevention strategies coordinate local communityinvolvement, include messages that stress healthy lifestyles, targethigh-risk youth, and are enduring, naturalistic prevention programs. Role of Health Care Providers Medical staff members often come in contact withindividuals at high risk for alcohol and drug abuse because health problems, mentalhealth disorders, and accidents are often as3ociated with alcohol and drug abuse.Health care professionals are becoming more aware of the signs and symptomsof alcohol and drug abuse. Pediatricians and child or adolescent psychiatrists canbetter diagnose the child at risk for alcohol and drug abuse if theyunderstand the etiological risk factors and the four stages of drug use outlined byMacDonald (1984), former director of ADAMHA. Professional health care providerswho work with troubled youth could provide early intervention services if they wouldbe willing to determine risk for alcohol or drug abuse by using a briefinstrumunt that has been developed for the American Academy of Pediatricians(Klitzner and Schon- berg 1988; Petchers et al. 1988). In addition, doctors need to be sensitive to the possiblenegative consequences of overprescribing drugs to youth, such as (a) thatyouth may infer from their experience with prescription medics t ions that drugs can cure anypain, physical or mental, and (b) that drugs areoccasionally diverted into street use. One pediatric study suggests that youth who are prescribedpsychoactive medica- tions, even for valid medical reasons, are more likely to becomeabusers (Roush et al. 1980). One problem with this correlational studyis that it is difficult to attribute cause. Many youth look up to members of the medical profession as rolemodels in the communityBecause of their involvement with children and youth as caregivers and their high status (both of which increase identificationand role modeling), health care providers in their involvement with youthshould be sure to model nonuse of tobacco and illegal drugs and abstinence ormoderate use of alcohol. This is a position that all responsible adultsincludingparentswho are around children should take. Summary There are no simple solutions in the fight against alcohol anddrug abuse. Alcohol and drug abuse is due to long-term, complex causes that begin early in life. Early childhood risk factors put some children on a different developmental path that terminates in alcohol or drug abuse and other problem behaviors. Health care providers need to use their professional training and knowledge to attempt to help these children by early intervention and referral efforts. 350 PREVENTION OF MENTAL DISORDERS

References

Abel, E.L., ed. Fetal Akohol Syndrome I: An Annotated and Comprehensive Bibliography. Boca Raton, Fla.: CRC Press, Inc., 1981. Abel, E.L., ed. Fetal Akohol Syndrome II: Human Studies. Boca Raton,Fla.: CRC Press, Inc., 1982. Ablon, J. Al-Anon family groups: Impetus for change through thepresentation of alternatives. American Journal of Psychotherapy 28(1):30, 1974. Ackerman, R. Children of Akoholics. Holmes Beach, Fla.: Learning Publica- tions, 1983.

Adler, R., and Raphael, B. Children of alcoholics. Australian and NewZealand Journal of Psychiatry 17:3-8, 1983. A!pert and Schuckit. Unpublished data. Alvy, K.'Parenting Programs for Black Parents." Paper presentedat the Primary Prevention of Psychopathology Conference (Familieein transition: Primary prevention programs that work), National Institute of MentalHealth, University of Vermont, Burlington, Spring 1985. Apsler, R. People control the amounts of substances theyuse. Journal of Drug Issues 9:145-160, Spring 1979.

Arnold R. Effect of Raising the Legal Drinking Ageon Driver Involvement in Fatal Crashes: The Experience of Thirteen States. DOT Pub. No.(HS)806 902. Washington, D.C.: National Highway Traffic P. ifety Administration,Nov. 1985. Bale, R.N. The validity and reliability of self-reporteddata from heroin addicts: Mailed questionnaires compared with face-to-faceinterviews.The Interna, tiona:. Journal of the Addictions 14(7):993-1000,1979. Bandura, A. Social Learning Theory. Englewood Cliffs,N.J.: Prentice Hall, 1977.

Bandura, A. Social foundations of thought and action:A social cognitive theory. Englewood Cliffs, N.J.: Prentice-Hall, 1986.

Battjes, R.J., and Jones, C.L. Implications ofetiological research for preventive interventions and future research.In:Jones, CI., and Battjes, R.J., eds. Etiology of Drug Abuse:Implications for Prevention.NIDA Research Monograph 56, DBES Pub. No. (ADM)85-1335.Washington, D.C.: U.S. Govt. Print. Off., 1985. pp. 269-276.

Baumrind, D. "Why Adolescents Take ChancesAndWhy They Don't." Paper presented at the National Institute of Child Healthand Human Development, Betha;da, Md., Oct. 1983. AWOHOL AND DRUG ABUSE 351

Baumrind, D. Familial antecedentsof adolescent drug use: Adevelopmental perspective. In: Jones, C.L.,and Battjes, R.J., eds.Etiology of Drug Abuse: Implications for Prevention.NIDA Research Monograph56, DIMS Pub. No. (ADM)85-1335. Washington, D.C.:U.S. Govt. Print. Off., 1985. pp.13-44. Begleiter, H.; Porjesz, B.;Bihari, B.; and Kissen, B.Event-related brain potentials in boys at risk foralcoholism. Science 225:14934496,1984. Prevention Research: DeterringDrug Abuse Among Bell, C., and Battjes, R. DHHS Pub. No. Children and Adolescents.NIDA Research Monograph, (ADM)85-1335. Washington, D.C.:U.S. Govt. Print. Off., 1985. Bell, C.S., and Levy, S.M. Publicpolicy and smoking prevention:Implications for research. In: Matarazzo,J.D.; Weiss, S.M.; Herd, J.A.; andMiller, N.E., eds. Behavioral Health. NewYork: Wiley, 1984. pp. 775-785. "A Comparison of Childrenfrom Alcoholic and Bennett, L.A., and Wohn, S.J. Paper Non-alcoholic Families:Cognitive and PsychosocialFunctioning." presented at the National Council onAlcoholism Meeting, Washington,D.C., 1985. Bentler, P., and Speckart,G.Models of attitudesbehaviorrelations. Psychological Review 86:452-464, 1979. Bien, NA, and Bry, B.H. Anexperimentally designed comparisonof four intensities of school-basedprevention programs foradolescents with adjustment problems. Journal ofCommunity Psychology 8:110-116, 1980. Biglan, A.; Severson, H.H.; Ary,D.V.; Faller, C.; Thompson, R.;Nautel, C.; Lichtenstein, E.; and Weissman,W.W. Refusal skills trainingand the prevention ofadolescent cigarette smoking.In: Bell, C. and Battjee, R.,eds. Prevention Research: Deterring Drug AbuseAmong Children and Adolescents.DHHS Pub. No. (ADM)85-1344. Rockville,Md.: NIDA, 1985. pp. 170-195. Blum, R.; Blum, E.; and Garfield, E.Drug Education: Results andRecommen- dations. Lexington, Mass.: LexingtonBooks, 1976. Bohznan, M.; Sigvardsson, S.; andCloninger, R. Maternal inheritanceof alcohol Archives of General abuse:Cross-fostering analysis of adopted women. Psychiatty 38:965-969, 1981. Booz-Allen and Hamilton, Inc. An assessmentof the needs of and resourcesfor children of alcoholic parents.Rockville, Md.: NIAAA, 1974. Botvin, G.J., and Eng, A. Acomprehensive school-based smokingprevention program. Journal ofSchool Health 50: 209-213, 1980. Botvin, G.J., and Eng, A. Theefficacy of multicomponentapproach to the prevention of cigarette smoking.Preventive Medicine 11:199-211,1982. 357 852 PREVENTION OF MENTALDISORDERS

Botvin, G., and McAlister, A.Cigarette smokingamong children and adolescents: Causes and prevention.In: Arnold, C., ed. AnnualReview of Disease Prevention. New York: Springer,1982. Braucht, G. Psychosocialresearch on teenage drinking:Past and future. In: Searpitti, F., and Bateeman,S., eds. Drugs and the YouthCulture. Beverly Hills: Sage Press, 1980.

Brooks, L.S.; Lunkoff, I.F.; andWhiteman, M. Peer, family,and personality domains as related to adoleecents'drug behavior.Psychological Reports 41:1095-1102, 1977.

Brunswick A.F., and Boyle, J.M.Patterns of drug involvement: and secular influences Developmental on age at initiation. Youth and Society2:139-162, 1979. Bukoski, W. "Drug Abuse Prevention Evaluation: AMeta-Evaluation Process." Paper presented at the Annual Meeting of the American PublicHealth Associa- tion, New York, 1979. Bukoski, W.J. School-based substance abuse prevention:A review ofprogram research. In: Ezekoye, S; Kumpfer, K ; and Bukoski, W. J.,eds. Childhood and Chemical Abuse:Prevention and Intervention. 1986. pp. 95-115. New York: Haworth Press,

Cadoret, R.J.; Cain, C.A.;and Crowe, R.R.Evidence for gene-environment interaction in the development of adolescent antisocialbehavior. Behavior Genetics 13:301-310, 1983. Cadoret, R.J.; O'Gorman, T.W.; Troughton, E.; andHeywood, E. Alcoholism and antisocial personality: Interrelationships, genetic andenvironmental factors. Archives of General Psychiatry27:161-167.1, 1985. Catalano, R.F.; Hawkins,RIM.; White, H.; andPandina, R. Marijuana Use and "Predicting Delinquency in Two LongitudinalStudies."Paper presented at the AnnualMeeting of the AmericanSociety of Criminology, Diego, Nov. 1985. San

Clark, B., and Midanik, L.Alcohol use and alcohol Results of the 1979 national problems among U.S. adults: survey. In: Alcohol and HealthMonograph 1, Alcohol Consumption andRelated Problems. DHHSPub. No. (ADM)82-1190. Washington, D.C.: U.S. Govt.Print. Off., 1982. Cloninger, R.; Bohman, M.; and Sigvardsson, S.Inheritance of alcohol abuse. Archives of GeneralPsychiatry 38:861-868, 1981. Coate, D., and Grossman, M. "Effects of AlcoholicBeverage Prices and Legal Drinking Ages on Youth AlcoholUse: Results from the Second and Nutrition Examination National Health Survey." Unpublishedmanuacript. New York: National Bureau of EconomicResearch, 1986. ALCOHOL AND DRUG ABUSE 353

children's social status: Coie, J.D., and Dodge, K.A.Continuities and changes in Quarterly 29(3):261-282, 1983. A five-year longitudinalstudy. Merrill-Palmer Coie, J.D., andKuperschmidt,.J.B. A behavioral analysis ofemerging social status in boys' groups.Child Development 54:1400-1416,1983. Cook, P. The effect of liquor taxes ondrinking, cirrhosis and auto accidents.In: Moore, M., and Gerstein, D., eds.Akohol and Public Policy: Beyondthe Shadow of Prohibition. Washington,D.C.: National Academy Press,1981. Cook, P., and Tauchen, G. Theeffect of liquor taxes on heavydrinking. The Bell Journal of Economics13:379-390, 1982. Cook, P., and Tauchen, G. The effectof minimum drinking agelegislation on youthful auto fatalities, 1970-1977.Journal of Legal Studies 13, 1984. Cook, T. Priorities in research insmoking prevention. In: Bell, C.,and Battjes, Children and R., eds. Prevention Research:Deterring Drug Abuse Among Adolescents. DIMS Pub. No. (ADM)85-1334. Rockville, Md.: NIDA,1985. pp. 196-220. Cotton, N.S. The familial incidence ofalcoholism. Journal of Studies on il.cohol 40:89-116, 1979. Courtney, R.J. Functional familytherapy intervention for Hispanicjuvenile offenders and their parents. Projectconsultation on the Institute for Human Resources Development PreventionProgram.Salt Lake City: Utah State Division of Alcoholism and Dru 1J84. DeMarsh, J., and Kumpfer, K.L.Family-oriented interventions for the prevention of chemical dependency inchildren and adolescents.In:Ezekoye, S.; Kumpfer, K.; and Bukoski, W., eds.Childhood and Chemical Abuse: Prevention and Intervention. New York:Haworth Press, 1986. pp. 117-151. Dielman, T.E.; Horvath, W.J.; Leach,S.L.; and Lorenger, AL. Peer pressurein recruitment to smoking. Final reportto NIDA University of Michigan,1984a. Dielman, T.E.; Leach, S.L.; Lyons,AC.; Lorenger, AT.; Klos, D.M.; andHor- vath, W.J. Resisting pressures to smoke:Fifteen-month follow-up results of an elementary school based smokingprevention project. InternationalJournal of Health Education. 1984b. Dir "...meyer, D., and McKay, G.D.Systematic Training for Effective Parenting. Circle Pines, Minn.: AmericanGuidance Service, Inc., 1976. DuPont, R. Stopping drug abuse before itstarts: The future of drug abuse prevention. Rockville, Md.: Committee ofthe Institute for Behavior and Health, Inc., in press. 354 PREVENTION OF MENTAL DISORDERS

Durrell, J., and Bukoski, W. "ThePrevention of Substance Abuse:State-of-the- Art." Unpublishedpaper, 1983.

Eiser, R., and van der Pligt, J.Attitudinal and Social Factorsin Adolescent Smoking: In Search of PeerGroup Influence. Exeter,England: Department of Psychology, University of Exeter,1985.

Elder, G.H.; Caspi, A., andDowney, G. Problem behaviorin family relationships: A multigeneratiozalanalysis. In: Sorensen, A.; rod, L., eds. Human Weinert, F.; and Sher- Development: InterdisciplinwyPerspective. Hillsdale, N.J.: Erlbaum, 1983.

Elder, J.P.; Artz, L.M.; Carleton,R.A.; Knisley, P.M.; Peterson, P.E.; and Snow, R.C.K. G.S.; Rosenberg, "Establishing Cost-efficientMember-owned Heart Health Programs in OrganizationalSettings." Paper presentedat the Annual Convention of the Society ofBehavioral Medicine, Baltimore,Md., 1982. Elliott, D.S., and Huizinga, D. "The Ralationship BetweenDelinquent Behavior and ADM Problem Behaviors."Paper prepared for the of the Art Research ADAMIIA/OJJDP State Conference on Juvenile Offenderswith Serious Drug/Alcohol and Mental Health Problems,Bethesda, Md., 1984. Elliott, D.S.; Huizinga, D.; and Ageton, S.S. ExplainingDelinquency and Drug Use. Boulder, Colo.:Behavioral Research Institute,1982. Elliott, D.S.; Huizinga, D.; and Ageton, SS. ExplainingDelinquency and Drug Use. Beverly Hills,Calif.: Sage Publications,1985. Elliott, D.S., and Voss,H.L. Delinquency andDropout. Lexington, Mass.: Heath and Company, 1974. D.C. Evans, RI.; Rozelle, R.M.; Maxwell, S.E.; Raines, B.E.;Dill, C.A.; Guthrie, T.J.; Henderson, A.H.; and Hill,P.C.Social modeling filmsto deter smoking in adolescents: Results ofa three-year field investigation.Journal of Applied Psychology 66:399-414, 1981.

Evans, RI.; Rozelle, R.M.;Mittelmark, M.; Hansen, W.B.; J. Bane, A.; and Havis, Deterring the onset ofsmoking in children:Knowledge of immediate physiological effects and copingwith peerpreesure, media pressure, and parent modeling. Journal of Applied Social Psychology 8(2):126-135,1978. Ezekoye, S.; Kumpfer, K.;and Bukowski, W. Childhood Prevention and Intervention. and Chemical Abuse: New York: Haworth Press,1986. Farquahar, J.W. The community-based model of lifestyle intervention trials. American Journal of Epidemiology108:103-111, 1978. Farquahar, J.W.; Magnus,P.F.; and Maccoby, N. The and education in role of public information cigarette smoking controls.Canadian Journal of Public Health 72(6):412-420,1981. 3 ALCOHOL AND DRUG ABUSE 365

Farnsworth, M. Family structure, family attributes, and delinquency in a sample of low-income minority males and females.Journal of Youth and Adokscence 13(4):349-364, 1984. Felix, M.R.J. The Pennsylvania county health imnrovement program (CHIP). Unpublished project description, 1983. Finnegan, L.P. Clinical effects of pharmacologic agents on pregnancy, the fetus and the neonate. Annals of the New York Academy of Sciences 281:74-89, 1976. Fishbein, M., and Ajzen, I. Belief, Attitude, Intention and Behavior: An Intro- duction to Theory and Research. Reading, Mass.: Addison-Wesley, 1975. Fisher, D.A.; Armstrong, B.K.; and de Klerk, N.H. "A Randomized-controlled Trial of Education for Prevention of Smoking in 12-year-old Children." Paper presented at the Fifth World Conference on Smoking and Health, Winnipeg, Canada, July 1983. Flay, B.R. Psychosocial approaches to smoking prevention: A review of the fmdings. Health Psychology, 1987. Flay, B.R.; d'Avernas, J.R.; Best, J.A.; Kersell, M.W.; and Ryan, KB. Cigarette smoking: Why young people do it and ways of preventing it. In: McGrath, P., and Firestone, P., eds. Pediatric and Adolescent Behavioral Medicine. New York: Springer-Verlang, 1983a. Flay, B.R.; Hansen, W.B.; Johnson, C.A.; and Sobel, J.L.Involvement of Children in Motivating Smoking Parents to Quit Smoking with a Television Program." Paper presented at the Fifth World Conference on Smoking and Health, Winnipeg, Canada, and the 91st Annual Convention of the American Psychological Association, Anaheim, Calif., 1983b. Flay, B.R.; Johnson, CA.; Hansen, W.B.; Grossman, L.M.; Sobel, J.L.; and Collins, L.M. "Evaluation of a School-based, Family-oriented, Television- enhanced Smoking Prevention and Cessation Program: The Importance of Implementation Evaluation." Paper presented at the Joint Meeting of Evalua- tion Network and the Evaluation Research Society, Chicago, 1983c. Flay, B.R. ; Johnson, C.A.; Hansen, W.B.; Ulene, A; Grossman, L.M.; Alvarez, L.; Sobel, D.F.; Hochstein, G.; Sobel, J.L. Evaluation of a mass media enhanced smoking prevention and cessation program. In: Baggaley, J.P., and Tanega, P., eds. Experimental Research in Televised Instruction. Vol. 5.Montreal: Concordia University Press, 1982. Flay, B.R.; Ryan, K.B.; Best, J.A.; Brown, KS.; Kersell, M.W.; d'Avernas, J.R.; and Zanna, M.P. Are social psychological smoking prevention programs 3 ffec- five? The Waterloo Study. Journal of Behavioral Medicine 8(1):37-69, 1986. Forehand, R.L., and McMahon, R.J.Helping the Noncompliant Child. A Clinician's Guide to Parent Training. New York: The Guilford Press, 1981. 3 di 356 PREVENTION OF MENTAL DISORDERS

Fraser, M.W. "Social Responsibility and Substance Abuse." Paper presented at the Annual Program Meeting of the Council on Social Work Education, St. Louis, 1987a (March). Fraser, M.W. Reconsidering drug involvement among youth and adolescents: Implications for targeted prevention. Journal of Sociology and Social Welfare 14(4):21-46, 1987b. Fraser, M.W., and Hawkins, J.D. Social network analysis and drug misuse. Social Service Review 58(1):81-97, 1984. Friedman, L.N. Employing the ex-addict: An experiment in supported work. In: Lawson, I., ed. Quantitative Explorations in Orug Abuse Policy. New York: SP Medical and Scientific Books, 1980. Gabriella, W.F., and Mednick, SA Intellectual performance in children of alcoholics. The Journal of Nervous and Mental Disease 171(7):444-447, 1983. Gabriella, W.F.; Mednick, SA; Volavka, J.; Pollock, V.E.; Schulsinger, F.; and Itil, T.M.Electroencephalograms in children of alcoholic fathers. Psychophysiology 19:404-407, 1982. Gallant, D.M.; Rich, A; Bey, E.; and Terranova, L Group psychotherapy with married couples: A successful technique in New Orleans alcoholism clinic patieWs. Journal of Louisiana State Medical Society 122:41, 1970. Gersick, X; Grady, K; and Snow, D. "Social-cognitive Skill Development with Sixth Graders and Its Initial Impact on Substance Use."Unpublished manuscript. New Haven: Yale University, Department of Psychiatry, 1985. Giordano, P.C.; Cernkovich, S.A., and Pugh, M.D. Friendships and delinquency. American Journal of Sociology 91(5):1170-1202, 1986. Goldstein, J.W., and Sappington, J. Personality characteristics of students who become heavy drug users: An MA/11 study of an avant-garde. American Journal of Drug and Akohol Abuse 4:401-412, 1977. Goodstadt, M. Drug Education: A turn on or a turnoff.Journal of Drug Education 10:89-99, 1980. Goodstadt, M. Planning and evaluation of alcohol education programs. Journal of Alcohol and in ,ig Education. 26:1-10, 1981. Goodstadt, M., arid Sheppard, M.Three approaches to alcohol education. Journal of Studies on Alcohol 44(1):362-380, 1983. Goodwin, D. Is alcoholism hereditary? A review and critique. Archives of General Psychiatry 25:545-549, 1971. Goodwin, D.W.Alcoholism and genetics.Archives of General Psychiatry 42:171-174, 1986. 3 fi 2 ALCOHOL AND DRUG ABUSE 357

Gordon, T. Parent Effectiveness Training.New York: P.H. Wyden, 1970, Greenleaf, J. Co-alcoholic, Para-akoholic. LosAngeles: Greenleaf Associates, 1981. Grossman, M.; Coate, D.; and Arluck,G."Price Sensitivity of Alcoholic Beverages in the United States." Unpublishedmanuscript. New York: Nation- al Bureau of Economic Research, 1984. Guerney, B.G., Jr. Filial therapy: Descriptionand rationale. Journal of Con- sulting Psychology 28(4):304-310, 1964. Hansen, W.B. H'Ileory and implementationof the social influence model of primary prevention." Paper presented atthe National Prevention Network and NASADAD Prevention Research Conference,Kansas City, 1988. Hawkins, J.D.; Lishner, D.M.; andCatalano, R.F. Childhood predictors and the prevention of adolescent substance abuse.In: Jones, C.L., and. Battjes, R.J., eds. Etiology of Drug Abuse: Implicationsfor Prevention. NIDA Research Monograph 56, DIMS Pub. No. (ADM)85-1335.Washington, D.C.: U.S. Govt. Print. Off., 1985. pp. 75-126. Hawkins, J.D.; Lishner, D.M.; Jenson, J.M.;and Catalano, R.F. Delinquents and drugs:What the evidence suggests about preventionand treatment programming. In: Brown, B.S., and Mills,A.R., eds. Youth at High Risk for Substance Abuse. 1987. Herjanic, B.M.; Barredo, V.H.; Herjanic, M.;and Tomelleri, C.Children of heroin addicts. International Journal of theAddictions 14(7):636-641, 1979. Herjanic, B.M.; Herjanic, M.; Penick, E.C.; Tomelleri,C.J.; and Armbruster, R.B.S. Children of alcoholics. In: Seixas, F.,ed. Currents in Alcoholism. Vol. 2. New York: Grune and Stratton, 1977. Hill, S.Y.; Steinhauer, S.R.; and Zubin, J.Biological markers for alcoholism: A vulnerability model conceptualization. 34th AnnualNebraska Symposium on Motivation:Alcohol and Addictive Behavior.Lincoln, Neb.:University of Nebraska Press, 1985, Hirschi, T. Causes of Delinquency. Berkeley:University of California Press, 1969, Holmberg, M.B. Longitudinal studies of drug abusein a fifteen-year-old population, Acta Psychiatrica Scandinavica 16:129-136,1985, Holme, E.I.; Helgeland, A.; Hjermann, I.; and Leren,P. Socio-economic status as a coronary riskfactor:The Oslo study.Acta Med Scandinavka 660 (Suppl.):146-151, 1982. 363 358 PREVENTION OF MENTAL DISOKDERS

Hooper, F. ""rhe Relationship Between AlcoholControl Policies and Cirrhosis Mortality in the United States Counties,"Paper presented at the Annual Conference of American Public HealthAssociation, Dallas, 1983. Hu, T.; Swisher, J.; McDonnell, N.; and Stein,J. Cost-Effectiveness Evaluation: A Drug Abuse Prevention AlternativesProgram.Research Report for the Prevention Branch of the NIDA. Grant No.E07 DA02464-02, 1982. Huha, G.J., and Bentler, P.M.Developmental theory of druguse: Derivation and assessment ofa causal mode/ing approach. In: B&W*, P.B., and BrimD., eds. Life-Span Development andBehcwior. New York: Academic Nees, 1982.. pp. 147-203.

Huba, G.J.; Wingard, JA; and Bender, P.M. Longitudina/ analysis of the role of peer support, adult models, andpeer subcultures in beginning adolescent substance use: An application of setwisecanonical correlation methods Mul- tivariate Behavioral Research 15:259.279,1980.

Huba, G.J.; Wingard, J.; and Bender,P.M.Intentions to use drugsamong adolescents: A longitudinal analysis.International Journal of Addictions 16:331-339, 1981.

Hue:mann, L.R.; Eron, L.D.; Lefkowitz,M.N.; and Wakler, L.O. 'Tee Stability of Aggression Over Time and Generations."Paper presented at the Meeting of the Society for Research in ChildDevelopment, Detroit, 1.983. Jackson, N.; Carlisi, Greenway, C.; and lalenick, M. Ageof initial drug experimentation among white and nonwhiteethnics. The International Journal of the Addictions 16(8):1371-1386,1981. Jacob, T., and Seithamer, R.A.Alcoholism and family interaction.In: Jacob, T., ed. Family Interaction andPsychopathology: Theories, Methods,and Find- ings. New York: Plenum Pub., 1986.

Jeesor, R.. Predicting time of onaetof marijuana use: A developmentalstudy of high school youth. Journal of Consultingand Clinical Psychology 44:125434, 1976.

Jeesor, R.; Chase, J A; and Donovan,J.E. Psychosocial correlates ofmarijuana use and problem drinking in 41 national sampleof adolescents.American Journal of Public Health 70(6):604-613,1980.

Jessor, R.; Donovan, J.E.; andWidmer, K.Psychosocial Factors in Adolescent Alcohol and Drug Use: The 1978National Sample Study, and the1974-78 Panel Study. Boulder, Colo.: Instituteof Behavioral Science, 1980. Jessor, it., and Jeseor, S.L. ProblemBehavior and Psychosocial Development A Longitudinal Study. New York:Academic Pram, 1977.

3 ALCOHOL AND DRUG ABUSE 359

Jessor, R., and Jessor, S.L.Theory testing in longitudinalresearch on marijuana use.In:Kandel, D., ed. LongitudinalResearch on Drug Use, Washington, D.C. Hemisphere PubliahingCo., 1978. Johnson, C.A. Intervention in publicschools using informational, teacher,and peer strategies.In:Grabowski, J., and Glynn, T., eds.Interdisciplinary Analysis of Strategies in DrugAbuse Prevention. National Institute onDrug Abuse Research Monograph, inpreparation, Johnson, C.A.; Hansen, W.B.;Collins, L.M.; and Graham, J.W. FinalReport: The High School Anti-smokingProject. Report to National Institute onDrug Abuse. University of SouthernCalifornia, 1984. Johnson, C.A.; Hansen, W.B.; Flay, H.R.;Graham, J.; and Sobel, J. 'Prevention of Multiple Substance Abuse in Youth."Paper presented at National Prevention Network Conference, Nashville, 1985. Johnston, L.D. Drugs and AmericanYouth. Ann Arbor, Mich.: Universityof Michigan, Institute for Social Research,1978. Johnston, L.D.; Bachman, J.G.; andO'Malley, P.M. Student Drug Use in America, 1975-1981. DHHS Pub. No.(ADM)82-1221. Rockville, Md.: U.S. Department of Health and HumanServices, 1982. Johnston, L.D.; Bachman, J.C.; andO'Malley, P. Student Drug Use, Attitudes, and Beliefs, National Trends, 1976-1982.National Institute on Drug Abuse DHHS Pub. No. (ADM)83-1260.Washington, D.C.; U.S. Govt. Print. Off., 1983. Johnston, L.D., O'Malley, P.; andBachman, J.G. Drugs and AmericanHigh School Students, 1975-1983, NationalInstitute on Drug Abuse DHHS Pub. No. (ADM)85-1374. Washington, D.C.: U.S.Govt. Print. Ott, 1984. Johnston, L.D.; O'Malley, P.; andBachman, J.G. Use of Licit and IllicitDrugs by America's High School Students,1976-1984. DHHS Pub. No. (ADM)85-] 395. Washington, D.C.: U.S. Govt. Print. Off.,1985. Johnston, L.D.; O'Malley, P.M.; and Bachman,J.G. Drug Use Among American High School Students, College Students,and Other Young Adults: National Trends Through 1986. Ann Arbor, Mich.:University of Michigan, Institute for Social Research, 1986. Judd, C., and Kenny, D. Estimating the Effects ofSocial Interventions. London: Cambridge University Press, 1981. Kaij, L. Studies on the Etiology andSequels of Abuse of Alcohol. Sweden: University of Lund, Department of Psychiatry,1980. Kandel, D.13. Stages in adolescent involvementin drug use. Science 190:912- 914, 1975. 365 360 PREVENTION OF MENTAL DISORDERS

Kandel, D.B. Developmental stages in adolescentdrug involvement. In: Let- tieri, D.J.; Sayers, M.; and Pearson, H.W., eds.Theories on Drug Abuse. NIDA Research Monograph 30, DHHS Pub. No. 80-967.Washington, D.C.: U.S. Govt. Print. Off., 1980.pp. 120-127.

Kandel, D.B. Epidemiological and psychosocialperspectives on adolescent drug use. Journal of American Academic Clinical Psychiatry 21(4):328-347982. Kandel, D.B.; Kessler, R; and Margulies,R. Antecedents of adolescents' initiation into stages of drug use: A developmentalanalysis. In: Kandel, D.B., ed. Longitudinal Research in Drug Use: EmpiricalFindings and Methodological Issues. Washington, D.C.: Hemisphere-Wiley,1978. Kandel, D.B.; Simcha-Fagan, O.; and Davis,M. Risk factors for delinquency and illicit drug use from adoleecence toyoung adulthood. Journal of Drug Issues 60(1):67-90, 1986.

Kandel, D.B., Single, E.; and Kessler, R.The epidemiology of drug useamong New York State high school students: Distribution,trends and change in rates of use. American Journal of Public Health66:43-53, 1976a.

Kandel, D.B.; Treiman, D.; Faust, R.; andSingle, E. Adolescent involvementin illicit drug use: A multiple classificationanalysis. Social Forces 55:438-458, 1976b.

Kandel, D.B., and Yamaguchi, K.Developmental patterns of theuse of legal, illegal, and medically prescribedpsychotropic drugs from adolescence toyoung adulthood. In: Jones, C.L., and Battjes,R.J., eds. Etiology of Drug Abuse: Implications for Prevention. NIDA RosearchMonograph 56, DHHS Pub. No. (ADM)85-1335. Washington, D.C.: U.S.Govt. Print. Oft'., 1985. pp. 193-235. Kaplan, H.; Martin, S.; and Robbins, CPathways to adolescent druguse: Self-derogation, peer influence, wealr,.mg of aocial controls, and early substance use. Journal of Health and SoaalBehavior 25:270-289, 1984. Kaufman, E., and Kaufman, P.N. FamilyTherapy of Drug and Alcohol Abuse, New York: Gardner Press, 1979.

Kaufman, M. Myth and reality in thefamily patterns and treatment. of substance abusers. American Journal of Drugand Alcohol Abuse 7(3&4):257-279, 1980.

Kellam, S.G., emd Brown, H. SocialAdaptational and Psychologicai Antece- dents of Adolescent Psychopathology Ten Years Later. Baltimore: JohnsHop- kins University, 1982.

Kent, T.A.; Campbell, J.R.; andGoodwin, D.W.Blo(>d platelet uptake of serotonin in chronic alcoholics. Lancet,in press. ALCOHOL AND DRUG ABUSE 361

Kim, S. An Evaluation of OmbudsmanPrimary Prevention Program on Stucknt Drug Abuse. Charlotte, N.C.:Charlotte Drug Education Center, Inc., 1981. Kim, S. Feeder area approach: An impactevaluation of a prevention project on student drug abuse. International Journal ofAddictions 17(2):305-313, 1982. Nissen, B.; Schenker, V.; and Schenker,A. The acute effects of ethylalcohol and chlorpromazine on certain physiologicalfunctions in alcoholics. Quarterly Journal of Studies on Akohol 20:480-492,1959. Klein, N.C.; Alexander, J.F.; and Parsons,B.V.Impact of family systems intervention on recidivism and siblingdelinquency: A model of primary prevention and program evaluation. Journal ofConsulting and Clinical Psychology 45:469-474, 1977. Klitzner, M. "Descriptive Study of National ParentGroups." A grant submitted to the NIDA by the Pacific Institute forResearch and Evaluation, 1984. Klitzner, M., and Schonberg, S.K. Concernsregarding the indirect assessment for drug abuse among adolescents.Developmental and Behavioral Pediatrics. Vol. 9, No. 1. Baltimore: Williams and Wilkins,1988. Kumpfer, K.L. Prevention Approaches toAdolescent Substance Use/Abuse. Paper presented at the American Academyof Child Psychiatry Institute on Substance Abuse and Adolescence, San Antonio,TIMM, October 1985. Kumpfer, K.L."Prevention Services for Children ofSubstance-Abusing Parents." NIDA final technical report, R18 DA02758-101/02 and DA U3888-01, 1987a. Kumpfer, K.L. Special populations: Etiology andprevention of vulnerability to chemical dependency in children of substanceabusers. In: Brown, B.S., and Mills, AR, eds. Youth at High Risk forSubstance Abuse. NIDA Monograph, DHHS Pub. No. (ADM)87-1537, 19876. pp. 1-72. Kumpfer, KL. TBO: Therapy by objectives. A newclinical program for promot. ing changes. Symposium on new psychotherapytechniques for women. Rocky Mountain Psychological Association, Salt LakeCity, Utah, May 1975. Kumpfer, KL., and DeMarsh, J. 'Prevention Services toChildren of Substance- abusing parents: Project Rationale, DeP,Liption andResearch Plan. Technical report submitted to National Institute on DrugAbuse, Rockville, Md., 1984. Kumpfer, K.L., and DeMarsh, .1, Family environmentaland genetic influences on children's future chemicaldependency. Journal of Children in Contemporary Society 18(1&2):49-91, Fall-Winter 1985.

3 7 362 PREVENTION OF MENTAL DISORDERS

Kumpfer, K.L., and DeMarsh, J. Familyenvironmental and genetic influences on children's future chemical dependency.In: Ezekoye, S.; Kumpfer, KL.; and Bukoski, W.J., eds. Childhood andChemical Abuse: Prevention andInterven- tion. New York: Haworth Pram,1986. pp. 49-91.

Kumpfer, K.L.; Moskowitz, J.; andKlitzner, M. Future issues andpromising directions in the prevention ofchemical dependency. Journal ofChildren in Contemporaiy Society 18(1&2):249-278,1985. L'Abate, L. Enrichment: StructuredInterventions with Couples, Families,and Groups. Washington, D,C.:University Press of America, 1977. Labouvie, E.W., and McGee, C.R.Relation of personality to alcoholand drug use in adolescence. Journal of Consultingand Clinical Psychology54:289-293, 1986.

Lee-Feldstein, A., and Harburg,E. Alcohol useamong right- and left-h persons in a small community. Journal ofStudies on Akohol 43:824-82P Lerner, J.V., and Vicary, J.R.Difficult temperament anddrug use: Analyses from the New York longitudinalstudy. Journal of Drug Education 1984. 14(1):1-8,

Leupker, R.V.; Blackburn, H.;Mittelmark, M.; Jacobs, D.R.; andCline, F.G. A Community Research andDemonstration Projecton the Prevention of Car- diovascular Diseases: TheMinnesota Heart Health Program."Paper presented at the Ninth World Congress ofCardiology, Moscow, 1982. Maccoby, N. The Stanford heartdisease preventionprogram. In: Newman, I.M., ed. Consumer Behavior in the Marketplace. Symposiumproceedirgs. Nebraska: Nebraska Centerfor Health Education, 1976.pp. 31-44. MacDonald, D.I. Drugs, Drinkingand Adolescents. Chicago:Year Book Medi- cal Publishers, 1984. MacDonald, S., and Whitehead, P. Availability of outletsand consumption of alcoholic beverages. Journalof Drug Issues (Fall):477-486,1983. Maloff, D.; Becker, H.; Fonaroff,A.; and Rodin, J. Informal their influenc3 social controls and on substance use. Journal of Drug Issues9:161-184, 1979. McAlister, A.L.; Petry, C.; andMaccoby, N. Adolescentsmoking: Onset and prevention. Pediatrics 63:650-658,1979.

McAlister, AL; Puska, P.;Kaskela, K.; Pallonen, V.;and Maccoby, N. Mass communication and community organization for public healtheducation. American Psychology 35(4):375-379,1980.

3 g ALCOHOL AND ,)RUGABUSE 363

J.; and Koskelia, A. McAlister, A.L.; Puska, P.;Salonen, J.T.; Tuomilehot, from the North Karelia Theory and action forhealth promotion: Illustrations project. American Journalof Public Health72(1):43-50, 1982. McAlister, A.L.; Maccoby,N.; and Farquahar, J.W. Meyer, A.J.; Nash, J.D.; Journal of Con- Skills training in cardiovascularhealth education campaign. sulting and ClinicalPsychology 48(2):129-142,1980. adolescents. In: Lettieri,D.J., Miller, J.D. Epidemiologyof drug use among Drug Abuse and theAmerican Adolescent. NIDA and Ludford, J.P., eds. Washington, D.C.: U.S. Research Monograph, DIMSPub. No. (ADM)83-1277. Govt. Print. Off., 1981. pp.25-38. Gardner-Keaton, H.; Harrell,A.V.; Wirtz, P.W.; Abel- Miller, J.D.; Cixin, LH.; Findings son, H.I.; andFishburne, P.M. NationcdSurvey on Drug Abuse: Main Abuse, 1983. 1982. Rockville, Md.:National Institute on Drug and Issues. Cham- Miller, W. Systematic ParentTraining: Procedures, Cases paign, Ill.: Research Press,1975. Mosher, J.F. Federal taxlaw and public healthpolicy: The case of alcohol- related tax expenditures.Journal of Public HealthPolicy 3(3):260-283, 1982. preventing drinking Mosher, J.F. Server intervention:A new approach for driving. Accident Analysisand Prevention 5:483-497,1983. Mosher, J.F., and Wallack,L.M. Government regulationof alcohol advertising: Protecting industry profits versuspromoting the publichealth. Journal of Public Health Poll* 2(4):333-353,1981. adolescent substance abusethrough drug educa- Moskowitz, J.M.4PrOyenting Preventing tion.In:Glynn, T.1.1 Leukefeld, C.G.;and Ludford, J.P., eds. Adolescent Drug Abuse:Intervention Strategies. NIDAResearch Monograph 47, (ADM)83-1280. Rockville,Md., 1983. pp. 233-249. Moskowitz, J.M. The primaryprevention of alcohol problems:A critical review Center, Berkeley, of the research literature.Monograph of Prevention Research Calif., 1986. pp. 1-63. Moskowitz, J.; Malvin, J.; andSchaeffer, G. The effects ofdrug education at follow-up. Journal of Alcoholand Drug Education 30:45-49,1984b. Moskowitz, J.; Malvin, J.;Schaeffer, G.; and Schaps,E. An experimental evaluation of drug education course.Journal of Drug Education14:9-22, 1984a. Evaluation of an Moskowitz, J.; Malvin, J.;Schaeffer, G.; and Schaps,E. Prevention, 1986. effective development teachertraining program. Journal of 3 364 PREVENTION OF MENTALDISORDERS

Murphy, J.M.; Waller, M.B.;Gatto, G.J, ; McBride, W.J.;Lumeng, L.; and Li, T.K. Monoamine uptakeinhibitors attenuate ethanolintake in alcohol-prefer- ring (P) rata. Alcohol2:349-352, 1985. Murray, D.M., and Perry,C.L.The prevention of adolescent Implications of etiological, drug abuse: developmental, behavioral,e.nd environmental models. In: Jones, C.L. andBattjes, RJ., eds. Etiologyof Drug Abuse: Implica- tions for Prevention. NIDA Research Monograph 56,DHHS Pub. No. (ADM)85- 1335. Washington, D.C.;U.S. Govt. Print. Off., 1985.pp. 236-256. Myers, RD., and Melchior, C.L. Alcohol and alcoholism:Role of serotonin. In: Essman, W.B., ed.Serotonin in Health andDisease. Spectrum, 1977. Vol. 2. New York: Naranjo, C.k; Sellers, E.M.; Roach, C.A.; Woodley,D.V.; Sanchez-Craig, M.; and Sykora, K.Zimelidine-induced varietionsin alcohol intake by depressed heavy drinkers. non- Clinical Pharmacology Therapy35:374-381, 1984. Nardi, P.M. Children of alcoholics: A role-theoreticalperspective. The Journal of Social Psychology115:237-245, 1981. Nasrallah, H.; Keelor, K;and McCalley-Whittiers, alcoholic males. Biological M. Laterality shiftin Psychiatry 18:1065-1067,1983. Nathan, P. Failures inprevention: Whywe can't prevent the devastating of alcoholism and drug effect abuse. American Psychologist38:459-467, 1983. National Institute on Drug Abuse. A chronologyof growth: How theparent movement evolved. PreventionResources 6(1):11-12, 1982a. National Instituteon Drug Abuse. Networking:Allies in prevention/parents and youth. PreventionResources 6(2):4-6, 1982b. Noll, R, and Zucker, R. "DevelopmentalFindings Froman Alcoholic Vul- nerability Study: ThePreschool Years." Paper ing of the American presented at the AnnualMeet- Psychological Awociation,Anaheim, Calif., 1983. O'Donnell, J.A., and Clayton, R,R. Determinants ofearly marijuanause. In: Beschaer, G.M., and Friedman,A.S,, eds. Youth Drug and Treatment. Abuse: Problems, Issues Lexington, Mass.: LexingtonBooks, 1979. O'Donnell, J.A.; Voss, ILL.;Clayton, RR.; and Room, A Nationwide Survey. R. Young Men andDrugs: NIDA Monograph 5,Washington, D.C.: U.S.Govern- ment Printing Office,1976. O'Donnell, M. Research on drinking locations of alcoholimpaired drivers: Implications for preventionpolicies. Journal of Public 1985. Health Policy 6:510-525, 3 o ALCOHOL AND DRUG ABUSE 365

O'Farrell, T.J., and Bircher, Cx.R."Marital Relationsb:ps of Alcoholic, Con- flicted, and Nonconflicted Couples."Paper presented at the American Psychological Association, Los Angeles, 1985. Olson, D. "Methodological Concerns with FamilyLife Skills Research." Paper presented at the National Institute on DrugMuse Technical Review: Family Life Skills Training and its Research,Rockville, Md., 1983. Olweus, D.Stability of aggressive reaction patterns in males:A review. Psychological Bulletin 86:852-875, 1979. Olweus, D. The consistency issue in personalitypsychology revisitedWith special reference to aggression. British Jo:trnal ofSocial and Clinical Psychol- ogy 19:377-390, 1980. Ostram, T.M. The relationship between the affective,attitudinal, and social correlates of drug use. International Jourral of Addictions15:869-881, 1969. Paton, S., and Kandel, D.B. Psychological factorsand adolescent illicit drug use: Ethnicity and sex differences.Adolescence 13:187-200, 1978. Patterson, G.E.; Reed, J.B.; Jones, R.R.; and Conger,R.E. A Social Learning Approach to Family Intervention. Eugene, Ore.: CastaliaPublishing Co., 1975. Patterson, G.R. The management and disruption offamilies. A Social Learn- ing Approach:Coercive Family Process.Vol. 3.Eugene, Ore.:Castalia Publiing Co., 1982. Patterson, G.R. Performance models for antisocial boys.Amer'.can Psychologist 41:432-444, 1986. Patterson, G.R., and Stouthamer-Loeber, M. The correlationof family management practices and delinquency. Child Development,1966. Pentz, M.A. "Social Skills Training: Al iventive Intervention forDrug Use in Adolescents." Paper presented at Meeting of the AmericanPsychological As- sociat'-n, Washington, D.C., 1982. Pentz, M.A. Prevention of adolescent substance abusethrough social skill development. In: Glynn, T.J.; Leukefeld, C.G.; and Ludford, J.P., eds.Prevent- ing Adolescent Drug Abuse:Intervention Strategies.NIDA Research Monograph 47, DIMS Pub. No. (ADM)83-1280. Washington, D.C.:U.S. Govt. Print. Off., 1983. Perry, C.; Killen, .1 ; Slinkard, L.A.; and McAlister, A.L. Peerteaching and smolOng prevention among junk) Ai students. Adolescence 9(58):277-281, 1980a.

3 71 368 PREVENTION OF MENTAL DISORDERS

Perry, C.; Killen, J.; Teich, M.; Slinkard, L.A.; andDanaher, B.G. Modifying smoking behavior of teenagers: A school-basedintervention. American Journal of ic Heelth 70(7):722-725, 1980b. Perry, C.L.; Murray, D.M.; Pechacek, T.F.; and Pirie,P.L. Community-based smoking prevention: The MinnesotaHeart Health Program. Unpublished manuscript, University of Minnesota. 1983a. Porry, Teich, M.J.; Killen, J.; Dass, R; and Maccoby,N. High school smoking prevention: The relative efficacyof varied treatments and instruc- tions. Adolescence 18(71), 1983b. Petchers, M.K.; Singer, Angelotta, J.W.; and Chow, J. Pevalidation and expansion of an adolescent substance abuse screeningmeasure. Developmental and Behaviorul Pediatrics. Vol. 9.Baltimore: Williams and Wilkins, 1988. Pickens, R, and Svikis, D. Use of the twinmethod in the study of vulnerability to drug abuse. In: Genetic and BiologicalMarkers in Drug Abuse and Al- coholism.NIDA Research Monograph 66, Pub.No. (SN)017-024-01291-2. Washington, D.C.: U.S. Govt. Print. Off.,1986. Polich, J., ead Bloom, F. "P300 Reflects theDegree of Cognitive Decline from the Resia, A Effects of Alcohol Consumptionin Normals and Ibdividuals at Risk for Alcoholism."Abstracts, IVth World Congress of BiologicalPsychiatry, Philadelphia, 1985. p. 436.

Polich, J.M.; Ellickson, P.L.; Reut!_,P.; andKahan, J.P. Stratcgies for Control- ling Adolescent Drug Use. Santa Monica,Calif.: Rand, 1984.

Pollock, V.E.; Volavka, J.; Goodwin,D.W.; Medrick, S.A.; Gabrielli, W.F.;Knop, J.; and Schulsinger, J.F "The EEG AfterAlcohol Administration in Menat Risk for Alcoholism."Abstracts, Nth World Congress ofBiological Psychiatry, Philadelphia, 1983. p. 436.

Porjesz, B.,and Begleiter, H. Human brainelectrophysiology and alcoholism. In: Tarter, RE., iind Van Thiel, D.H.,eds. Alcohol and the Brain: Chronic Effects. New York: Plenum Pub., 1985.

Propping, P.; Kruger, J.; and Mark,W. Genetic disposition to alcoholism: An EEG study in alccli lies and theirrelatives. Human Genetics 59:51-59, 1981. Resnick, H.S. "It starts with people- II. Drug Abuse Prevention Programs for the 19800." Unpublishedpaper. Lafayette, Calif.: Pyramid, 1982. Rimmer, J. The children of alcoholics:An explPratory study. Childrenand Youth Services Rildiew 4:365-373,1982, Robins, L.N. Deviant Children GrownUp. Baltimore: Williamsena 1966. 372 ALCOHOL AND DRUG ABUSE 367

Robins, LN.; Helzer, J.E.; andPryzbeck, T.R "Epidemiological Perspective on Alcoholism: ECA data." Presentation atthe American Psychopathological Association Annual Meeting, 1986. Robins, LN., and Pryzbeck, T.R..Age of onset of drug uso as a factor indrug use and otherdisorders. NIDA Research MonographNo. 56, U.S. DHHS Pub. No. 1415. Washington, D.C.:U.S. Govt. Print. Off., 1985. pp. 178-193. Robins, L.N., and Ratcliffe, KS. Riskfactors in the continuation ofchildhood antisocial behavior into adulthood.International Journal of Mental Health 7:76-116, 1979. Rockeach, M., and Kliejunas, P.Behavior as a function of attitude-toward- object and attitude-toward-situation.Journal of Personality and Social Psychology 22:194-201, 1972. Rodin, J. The application of social psychology.In: Lindzey, G., and Aronson, E., eds. The Handbook of Social Psychology.Vol. 2. New York: Random House, 1985. Roffman, R.A. Marijuana as Medicine.Seattle: Madrona, 1982. Roush, G.C.; Thompson, BA; and Berberian,R.M. Psychoactive medicinal and non-medicinal drug use among high schoolstudents. Pediatrics 66:709-715, 1980. Rutter, M., and Giller, H. Juvenile Delinquency:Trends and Perspectives. New York: Penguin Books, 1983. Saltz, R. Server intervention: Conceptualoverview and current developments. Alcohol, Drugs, and Driving: Abstracts andReviews 1(4):1-14, 1985. Preven- tion Research Center, Server InterventionProject, 2532 Durant Avenue, Berkeley, California 94704. Sampson, R.J., Castellano, T.C.; and Laub,J.H. Juvenile Criminal Behavior and Its Relation to Neighborhood Characteristics.Washington, D.C.: Office of Juvenile Justice and Delinquency Prevention,1981. Scher's, E.; DiBartolo, R.; Moskowitz, J.; Palley,C.; and Churgin, S. Primary prevention avaluation research: A review of127 impact studies. Journal of Drug Issues 11:17-43, 1981. Schap, E.; Moskowitz, J.; Concon, J .; and Malvin,J.Process and outcome evaluation of a drug education course. Journal of DrugEducation 12:353-364, 1982. Schinke, S.P., and Gilchrist, L.D. Preventing cigarettesmoking with youth. Journal of Primary Prevention, 1983a. 368 PREVENTION OF MENTAL DISORDERS

Schinke, S.P., and Gilchrist, L.D.Primary prevention of tobacco smoking. Journal of School Health 53(7):416-419, 1983b. Schinke, S.P., and Gilchrist, L.D. Preventingcigarette smoking with youth. Journal of Primary Prevention, 1986.

Schlegel, R.; Manske, S.; and Page, A. A guideddecision making program for elementary school students: A field acperimentin alcohol education.In: Miller, P., rad Nirenberg, T., eds. Preventionof Alcohol Abuse. New York: Plenum Pub., 1984.

Schlegel, R., and Norris. J. Effects of attitudechange on behavior for highly involving issues: The case of marijuana smoking.Addictive Behavior 5:113- 124, 1980.

Schlegel, R., and Sandborn, M. Religiousaffiliation and adolescent drinking. Journal of Studies on Akohol 40:693-703, 1979. Schuckit, M.A., and Bernstein, L.!.Sleep time and drinking history: A hypothesis. American Journal of Psychiatry138:528-530, 1981. Schuckit, M.A.; Goodwin, D.W.; Winokur, G.A study of alcoholism in half siblings. American Journal of Psychiatiy 128:1132-1136,1972. Schuckit, M.A., and Rayses, V. Ethanol ingestion:Differences in blood acetal- dehyde concentrations in relatives of alcoholicsand controls. Science 203:54-55, 1979. Sheppard, M.; Wright, D.; and Goodstadt,M. Peer pressure and drug useEx- ploding the myth. Adolescence 20:949-958,1985. Shope, J., and Diehnan, T."An Elementary School-based PeerResistance Program of Preventing Adolescent Alcohol Misuse."Paper presented at Illinois Alcoholism and Drug DependenceAssociation Annual Cc ierence Workshop. Ann Arbor, Mich.: University of Michigan,1985.

Simcha-Fagan, 0., and Gersten, J.C.Early precursors and concurrentcorre- lates of items of illicit druguse in adolescence. Journal of Drug Issues 60(1):7- 28, 1986.

Smart, R.G., and Fejer, D. Druguse among adolescents and their parents: Closing the generationgap in mood modification. Jcu.rnal of Abnormal Psychol- ogy 79:153-160, 1972.

Smith, G.M., and Fogg, C.P. Pnychologicalantecedents of teenage druguse. In: Simmons, R. G., ed. Research in Communityand Mental Health. Vol. 1. Green- wich, Conn.: JAI Press, 1978.

374 ALCOHOL AND DRUG ABUSE 869

Solomon, D.S. Health campaigns ontelevision. In: Pearl, D.; Bouthilet,I.; and Lazar, J., eds. Television andBehavior: Ten Years of ScientificPregress and Implications for the Eighties. Vol.2: Technical Reviews.Washington, D.C.: National Institute of Mental Health,1982. Sowder, B., and Burt, M. Childrenof Addicts and Non-addicts:A Comparative Investigation in Five Urban Sites.Report to NIDA. Bethesda,Md.: Burt Associates, Inc., 1978a. Sowder, B., and Burt, M. "Childrenof Addicts: A Population inNeed of Coordinated Comprehensive MentalHealth Services." Paper presentedat the American Association of PsychiatricServices for Children, Atlanta, 1978b. Spivak, G. High Risk Early BehaviorsIndicadng Vulnerability to Delinquency in the Community andSchool.National Institute of JuvenileJustice and Delinquency Prevention, Office ofJuvenileJustice and Delinquency Prevention, Law Enforcement AssistanceAdministration. Washington, D.C.:U.S. Govt. Print. Oft, 1983. Stanton, M.D. The client as familymember: Aspects of continuing treatment. In: Brown, B.S., ed. Addictsand Aftercare: CommunityIntegration of the Former.Drug Abuser. New York:Guilford Press, 1979. Stein, J.A.; Swisher, J.D.; Hu, T.; andMcDonnell, N. Cost-effectiveness evaluation of a Channel One program.Journal of Drug Education 14:251-269,1984. Steinglass, P. Family therapy inalcoholism. In: Kissin, B., and Begleiter,H., eds, Biology of Alcoholism, Vol. 5,Treatment and Rehabilitation of theChronic Alcoholic. New York: Plenum Pub.,1975. Steinglass, P. The alcoholic family athome: Patterns of interaction in dry, wet and transitional stages of alcoholism.Archives of General Psychiatly 24:401- 408, 1980. Stiuunel, B.; Goldberg, J.; Reisman,A.; Murphy R.J.; and Teets, K.Fetal outcome in narcotic-dependent. women:The importance of the type of raaternal narcotic used. American Journal ofDrug o.nd Alcohol Abuse 9(4):383-395, 1982-1983. Swisher, J.D., and Hu, T.W. Alternativee todrug abuse: Some are and some are not. In:Glynn, T.J.; Leukefeld, C.G.; and Ludford,J.P., eds. Preventing Adolescent Drug Abuse: InterventionStrategies. NIDA Research Monograph 47, DHHS Pub. No. (ADM)83-1280.Washington, D.C.: U.S. Govt. Print. Off., 1988. pp. 141-153. Szapocznik, J."Family Effectiveness Training."Unpublished final report. University of Miami Family GuidanceCenter, 1983.

3 7 5 370 PREVENTION OF MENTALDISORDERS

Szapocznik, J.; Kurtines, W.M.;Foote, F.H.; and Perez-Vidal,A. Versus One Person Family Corijoint Therapy: Some Evidence forEffectiveness of Con- ducting Family Therapy ThroughOne Person. Coral Gables,Fla.: University of Miami Spanish FamilyGuidance Center, 1986.

Szapocznik, J.; Santisteban, D.;Rio, A; Perez-Vidal, A.; andKurtines, W.M. Family effectiveness training(FET) for Hispanic families.In: Lefley, H.P., and Pederson, P.B., eds. Cross Cultural Training for Mental HealthProfessionals. Springfield, Ill.: Charles C. Thomas,1985.

Szapocznik, J.; Santisteban, D.;Rio, A; and Santisteban,D. Family effectiveness training: An intervention to prevent problem behaviors in Hispanicadoles- cents. Hispanic Journal ofBehavioral Sciences, Aug.1987.Miami, Fla.: Spanish Family Guidance Center. Tell, G.S.; Klepp, KJ.; Yeller, O.D.; and McAlister, A.Preventing the onset of cigarette smoking in Norwegianadolescents: The Oslo YouthStudy. Preventive Medicine 13:256-275, 1984.

Templer, D.I.; Russ, C.F.; andAyers, J. Essential alcohohsmand family history of alcoholism. QuarterlyJournal of Studieson Alcohol 35:655-657, 1974. Thorne, C.R., and DeBlassie, KK. Adolescent substanceabuse. Adolescence 20(78):335-347, 1985.

Vaillant, G.E. The NaturalHistory of Alcoholism.Cambridge: Harvaxd University Press, 1983. Vaillant, G.E., and Milofsky, E.S. The et;ology of alcoholism:A prospective viewpoint. American Psychologist37:494-563, 19. Vartainen, E.; Pallonen, U.;McAlister, A ; Koskela, K.; and two years of educational Puska, P. Effect of intervention in adolescent smoking(The North Karelia Youth Project). Bulletinof the World Health Organization 61(3):529-532,1983. Vingilis, E.R., and DeGenova, K Youth and the forbiddenfruit: Experiences with changes in legal drinkingage in North America. Journal of Justice 12:161-172, 1984 Criminal

Vogel, F.; Schalt, E.;Kruger, J.; Propping, P.;and Leh nert, K,F. The cephalogram (EEG) electroen- as a research tool in human behaviorgenetics: Psychologi- cal examinations inhealthy males with variousinherited EEG variants. Human Genetics 47:1-45,1979. Wagenaar, AC. Effects of an increase in thelegal minimum drinkingage. Journal of Public HealthPolicy, Inc. 2(3):206-223,1981. Wagenaar, A.C. Aggregate beer and wine consumption:Effects of changes in the minimum legal drinkingage and a mandatory beverage law in Michigan. Journal container deposit of Studies on Alcohol43:(5):469-487, 1982.

P+4: ALCOHOL AND DRUG ABUSE 371

Wagenaar, A. Alcohol, Young Drivers and Traffic Accidents: Effects of Mini- mum Age Laws. Lexington, Mass.: Lexington Books, 1983. Wallack, L.M. The prevention of alcohol-related problems: Recommendations for public policy initiatives. Berkeley: Prevention Research Center, 1984. Wallack., L.M. Mass media, youth i the prevention of substance abuse: Towards an integrated approach. Jou, , tal of Children in Contemporary Society, in press. Wallack, LAI., and Barrows, D.C. Preventing Alcohol Problems in California: Evaluation of the Three Year "Winners Program," California Department of Alcohol and Drug Programs. Contract A-0345-8. Berkeley, Calif.: Social Re- search Group, 1981. Wegscheider, S. Another Chance: Hope and Help for the Akoholic Family. Palo Alto, Calif.: Science and Behavior Books, 1981. Wepner, S. Which way drug education? Journal of Drug Education 9:93-103, 1979. Wicker, A. W. An examination of the "other variables" explanation of attitude- behavior inconsistency. Journal of Personality and Social Psychology 19:18-30, 1971. Williams, R; Ward, D.; and Gray, L. The persistence of experimentally induced cognitive change: A neglected dimension in the assessment of drug prevention programa. Journal of Drug Education 15:3342,1985. Williams, TT., and Lillis, R.P. "Changes in Akohol Consumption by Eighteen Year Olds Following an Increase in New York State's Purchase Age to Nineteen." Paper presented at the National Council on Alcoholism, National Alcoholism Forum, Washington, D.C., 1985. Wittman, F.D. Reducing Environmental Risk of Alcohol Problems, Berkeley, Calif.: Prevention Research Center, 198r.e Wurtman, R.J., and Wurtman, J.J. Carbohydrate craeing, obesity and brain serotonin. Appetite 86(7)Suppl.:99-103, 1986. G. Family research study at Eagleville hospital rehabilitation center. Family Process 16:175-189, 1979.

3 7 7 CHAPTER 10

Prevention Issuesin YouthSuicide

David Shaffer M.B., B.8,, F.R.C.P., F.R.C.Psych.

Irving Philips, Prafessw ofChild Psy New York Stcae PsychiatricInstitute College of Physicians andSurgeons of Colurnbia University Ann Garland, MA. Research Associate

Kathleen Bacon, Ph.D. Research Associate Introduction

A rational approach topreventing suicide is mostlikely if we can find out possibly unique, experiences orclinical states that what. ars the special, and made their have affected individuals an,that. together or in a combination suicides more likely. These canbe regarded as risk factors.In general, the featuree of a risk factor that are mostrelevant to preventive interventions are-- 1. That they account for ahigh proportion of cases of thedisorder; 2. That t'.).ey be modifiable atreasonable cost. classified es follows: By the usual convention,preventive interventions are the individual from developingthe 1.Primary, The intervention prevents condition (i.e., being suicidal)either because it prevents theindividual from being exposed to therisk factor or because the riskfactor is eradi- cated. 2. Secondary. The conditionis pment, the patient issuicidal, and the suicidal state can be identified byappropriate inquiry, but its effects are slight, The intervention will preventthe condition from developing to cause significantsuffering or dysfunction.

This paper was prepared forthe Project Preveetion initiativeof the American Academy of Child and AdolescentPsychiatry. 378 374 PREVENTION OF MENTAL DISORDERS

3.Tertiary.The condition is establishedand the subject is clearly symptomatic; he or shemay have made one or more suicide attempts. The interventionfor all intents andpurposes, treatmentwill shorten its course, reduce the likelihood ofrecurrence, or lessen its consequencesor complications (i.e., death). Suicide is given as an example of a condition that cannot beprevented because it is so rare (Rosen 1954). If suicide is rare, youth suicideis even more so;even in late adolescence the deathrate from suicide is only abouttwo-thirds of the adult rate (Shaffer andFisher 1981). Skepticism about preventing youthsuicide may be warranted, butnot because youth suicide is so rare. Approximatthy2,000 American youngstersages 10 to 19 killed themselves in1982 (14 percent of all deathsin that age group), far more than were affectedby the inherited metaboliccondition phenylketonuria, for whichroutine preventive screening isaccepted without question. The difference isthat the known risk factorsor early signs of suicide are sensitive (that is, they could accuratelydefine and identify most suicide victims before their death) but very nonspecific (thatis, applying them toa screen would identify an overwhelmingproportion of individuals who noticea significant risk for suicide).The factors thatare used to screen babies for phenylketonuria, however, are both specific and sensitive; few childrenwith the disease slip through thenet because of an apparently normaltest. The lark of specificity of the knownrisk factors for suicidemeans that preventive efforts directed to high-risk individuals are likely to be wasteful because few ofthose identified would havegone on to commit suicide withor without an intervention. Even with these caveats, however, it is time to considerpreventing suicide in young people. Ourknowledge about specific riskfactors is limited, not because we have reached thelimits of what well-plannedresearch with existing methods can tell us, but becausemuch possible research has out. There is every yet to be carried reason to expect that new findings willsoon improve our understanding ofmore specific risk factors and, inturn, our ability to both anticipate truly suicidal individualsand protect them. This paper outlines whatwe know about the risk factors for youth and examines the effectiveness suicide of preventive strategies untilnow. Because the literature on the subject islimited, this paper alsoexamines preventive strategies used among adultsthat might also be appliedto young people. What Are We Tryingto Prevent? All Self-DestructiveBehaviors? A common psychologicalfallacy is that intentcan be inferred from outcome. An application of thisprinciple is that all self-initiated, behaviors, such as drug ultimately noxious use, gambling, and aggression, thatinvite retaliation should be viewed along withsuicide as beingon a continuum of self-destructive behavior. For the purposes of this review, this idda isrejected for a number of reasons: (a) it has the potential for incorporating too ruany differentbehaviors 3 375 YOUTH SUICIDE

ignores the power ofimmediate or short-term to be useful; (b)it implausibly very obvi- example, althoughregularly taking heroin may reinforcement (for the addict is one ofpleasure or relief ously be harmful, itsimmediate effect on and (c) it is notbacked up by anycorroborating empirical from discomfort); behaviors as accidentrepeti- findings. Thus, suchpurportedly self-destructive For these often in individualswho commit suicide. tion are not found more only behavior that term "suicide" asused in this review covers reasons, the about one's death. centers on a conscious ordeclared wish to bring All SuicidalBehavior or OnlyCompleted Suicide? especially among theBritish, to classify It has beenpsychiatrists' practice, completed suicide and"parasuicide."Parasuicide is suicidal behavior into predominantly by youngfemales who take usually a nonlethalbehavior, shown poisonous substance(Stengel and Cook a nonlethaloverdose of a potentially does not wish to 1962). It is assumedthat the parasuicide 1958; Neuringer anxiety or distress inothers to bring about die, but uses abehavior that produces (i.e., manipulative some immediatechange in his orher circumstances behavior). implications for suicideprevention.If This classificationhas important and completions areseparate, albeit over- suicidal thoughts,threats, att,empts, and if only characterized by differenttypes of psychopathology, lapping, entities threaten or attemptsuicide intend to die, a very smallminority of those who threateners who suicide preventionshould be on attempters or then the focus of completers. On theother hand, ifideators, are judged tobe potential suicide they have acted and completers differ onlyin how effectively attempters, and efforts should bebroadly directed not in what theyintended, then prevention attempts to predict reducing any form ofsuicide morbidity, and and aimed at behavior will bemisplaced. We suicide potential on thebasis of current suicide will argue for this. entities and attempted suicidebeing distinct diagnostic The case for suicide characteristics of suicideattempters is based mainly onthe demographic Goldacre 1982), (Bergstrand and Otto1962; Morgan et al.1975; Hawton and attempters are youngfemales, whereas deathcertificate which show that most 1985) indicate that data (Shaffer andFisher 1981; Centersfor Disease Control Europe and the UnitedStates, suicide becomesincreasingly at least in Western in males. However,there is common withadvancing age and is more common ideators, attempters,and completersshare many char- abundant evidence that by sex and intent acteristics, and it ispossible that methodchoice is confounded to die. Method, and Outcome Possible Confoundingof Age, Sex, Sex and lethod committed with firearms, Although most suicidesin the United States are method. Ingestiondeaths are sizable sex differences canbe found for any given 3 o 376 PREVENTION OF MENTALDISORDERS

significantlymore common among females, and shooting andhanging are used more often by males (Centersfor Disease Control are frequently ineffective 1985). However,overdoses because they takeeffect relatively slowly time for secondthoughts and effective and afford usually is) that these treatment.It could be argued(and differences in outcomemust be known to those them, so that methodpreference reflects who choose who overdose have a difference in intentionality.Females no wish to die andat* therefore assumed less severepsychiatric problem; to suffer froma men, on the other hand,as evidenced by their more frequent use of highlylethal methods, because of are more likely to wantto die a more serious underlyingpsychiatric condition. It may be, however,that the sexes do not differ in their(generally ambivalent) motivation to die, thatchoice of methodis a sex-typed behavior, suicidal males andfemales respond to and that when ways, which, at least in an extreme affect, they doso in different North America andWestern Europe,are likely to lead to death in boysbut are unlikelyto do so in girls. This explanationwould be compatible in the suicide with the findingthat sex differences rate are not universaland vary by Organization 1974).For example, in country (World Health a report of conseeutiveyouth suicides in India (Sathyavathi1975) thereare no sex differences This could be in the teenage suiciderate. because even if thevictim has second methods are lesseffective in that thoughts, resuscitation country; becauseanother, more lethal, preferred method,jumping froma height, is female- country where the more readily available inthat backyard well isa ubiquitous feature, preferred ingeetantis an insecticide or because the available. (paraquat) for whichno treatment is Sex variationsin rate are unlikely to be a result ofreporting bias, because mon) aggressive case-findingmethods do not et al. 1974). materially alter theratio (Kennedy Age and Method Representative dataare rare, but in Oxford, England, a survey of all suicidesin i 4e County of Hawton and Goldacre(1982) found that what morecommon in the early than attempts weresome- in the laterteenage years, then about age 24 (Shafferand Fisher 1981). peaked at be interpreted This picture ofreciprocal trends carefully becauseattempts are must This finding,however, is not more common than completions. incompatible with thenotion that mature they becomemore astute in deciding as teenagers response to a problem, whether suicide isan appropriate and that if theydo decide methods ismore appropriate. on suicide, their choiceof Age-related methodpreference in youthis not well there is evidencethat younger children documented, although suicide attempts and teenagersare more likely to make with less lethaldrugs (most gesics) than adults, commonly over-the-counteranal- who most oftenuse more dangerous, obtained bylegitimate prescription psychoactive drugs from doctors(Morgan et al. 1975).This 3S1 YOUTH SUICIDE 377 difference could be explained as wellby greater ease of availability ofprescribed drugs for older age groups as bydifferences in intentionality. If intent is the same in the two groups,then the use of more effective drugs by older patients and less effectivedrugs by younger ones wouldbe expected to result in (a) more failed suicideinitiatives and thus a smallerproportion of completed suicides attributable tooverdose in the young and (b) moresuccessful overdose initiatives in olderindividuals. These resulta have beendocumented (Centers for Disease Control 1985). Sex Differences and Secular Trends Suicide completion rates have increasedduring the past three decades. If attempts and completions reflectthe same underlying phenomena, oneshould expect to see somewhat differentsex-related secular trends. Attempts, asthe more direct index ofsuicidal behaviors, should show anincreasing trend in both sexes. However, ifcompletion depends on the sex-linkedselection of more lethal methods, increase should be confined tomales. This seems to be what has happened; attempts have increasedduring the past three decades in both sexes (Weissman 1974; O'Brien 1977) whiledeaths have increased only amongmales. Conclusion These fmdings do not rule outsex-related differences in psychopathology among completers,because the relationship between sex,psychopathology, and intent may be quite complex. In ourNew York study of 175 consecutive teenage and child suicide completers, wehave found trends for depression tobe more common in the smallproportion of girls who have committedsuicide and in older teens of both sexes. It appearsthat many of the dead boys were notdepressed but suffered from poor impulsecontrol. They committed suicide soonafter an acute precipitation, at a timewhen their intentionality was certainlyhigh but could have been predicted to diminishhad they survived. At theee times they acted In a highly effective fashion(usually by hanging or shooting themselves), whereas we guess that many girls whofelt similarly would have acted on their suicidal impulses by taking what would turnout to be an ineffective overdose Similarities Between Suicidal Ideationand Behavior Paykel et al. (1974b), i n a study of ahousehold probability sample of adults, found that responses to a range of questionsabout suicidal ideation and behavior were stronglyinterrelatod in a hierarchical fashion. Notsurprisingly, almost all subjects who respondod positively to moreserious questions (about attempts) also replied positively t ) questionsabout less severe thoughts (whetherthey ever wanted to bedeal or ever thought about committingsuicide).More interesting, the age and symptom of profilesof individuals who responded positively to the items on ideation, butwho had not made a prior attempt, resembled those of the attempters and differedfrom those of the population who had neither suicidal ideation nor behavior.Pfeffer et al. (1984) reported 3imilar

3 S 378 PREVENTION OF MENTAL DISORDERS

findings in school children: Thosewho had thoughts of suicide andthose who had made a suicide attempt showedsimilar profiles of associatedsymptoms. However, this evidence for overlap (whichsupports the idea of suicideas a unitary phenomenon) needs to betempered, for ideation is muchmore common than suicidal behavioramong teenagers. Indeed, ideationmay be so common that one needs to question whetherit is at all abnormal. Shafferet al. (1987) found that 40 percent ofa population of about 5,000 teenagers ina semirural New Jersey county had entertainedsuicidal ideas, but only 5percent reported having made a suicide attempt. Onthe basis of this epidemiologicalevidence, the relationship between suicidalideation and behavior is at bestnonspecific, because clearly all suicideattempters and completere must firsthave been ideators. Similarities Between Attemptsand Deaths The close link between suicideattempts and completions has beenestab- lished in a number of differentways. Similar Diagnostic Profiles andCorrelatep Preliminary examination ofan ongoing study of 175 consecutiveteenage suicide completers and age-, race-, and sex-matched attempters (whoneeded hospital admission) in thegreater New York metropolitanarea (Shaffer et al. 1987) reveals no differences in diagnostic profile, previous attempthistory, and, most significantly, familial incidenceof suicide and suicideattempts between attempters and completers. Similarly,in a far smaller study (Brent 1987),there were no significant differences in diagnosisand prior attempt historyafter age and sex differenceswere taken into account. History of Attempts AmongCompleters Evidence from retrospective psychological autopsy studies(ShaiTer 1974; Kennedy et al. 1974; Robinset al. 1959; Dorpat and Ripley et al. 1970) indicate prior 1960; Barraclough attempt rates of between 30 and50 percent. Con- versely, followup studies ofteenage attempters (Otto 1972; Motto1984; Hawton and Goldacre 1982) show suicideratee 50 to 60 percent higher thi. '.n the general population. those found However, these data must betaken in perspective because is present, it is clear that although overlap a relatively small proportion of suicideattempters will go on to commit suicide. Estimates inthe followup studies 10 percent of boys who range from 9 to are admitted to a psychiatric inpatientunit following a severe attempt (Otto 1972; Motto 1984)to less than 1 percent of boyswho presented at anemergenc3 room after an overdose and who to a psychiatric hospital were not admitted (Hawton and Goldacre 1982).Similar proportions for girls range from 1 percentfor former inpatientsto about 0.1 percent for those wh- received only medicaloutpatient care. 3 ' 3 YOUTH SUICIDE 379

Predkting Death Among Attempters If attempters and completers are drawn from the same population, we should expect it to be difficult to pick out future completers from among the attempters. Very few followup studies of young suicide attempters have had access ic good baseline data. However, the findings from one of these studies (Motto 1984) are in line with this prediction. In a 5- to 15-year followup of teenagers admitted to a hospital after an attempt or with serious depression,although certain factors were proportionately more common in those who went on to commit suicide, the same factors were numerically many times more common in attempters who did not; that is, the base rate in noncompleters was high, and there were no pathognomonic predictors of later completion. Similarly, in a followup study of adult attempters' demographic characteristics, the extent to which the suicide attempt was judged to be serious (that ia, precautions taken to avoid discovery and medical seriousness) were not predictive of later suicide (Greer and Lee 1967). Suicide Potentiators Affect Deaths and Attempts If completed 3nd attempted suicide are closely related, one should expect that factors that potentiate or inhibit suicide have a similar effect on both phenomena. Regrettably, there are no readily available techniques for inhibit- ing suicide, but there is extensive evidence that attention and publicity given to a suicidal death will lead to an increase in suicides. In the only study examining the effect of media on both deaths and attempts, Gould and Shaffer (1986), studying the impact of television programs that dramatized the plight of suicidal teenagers, found that a majority of the programs were followed by an increase in suicide attempts. Conclusions and Implications for Saicide Prevention Given that there is no clearly effective way to differentiate the proportion of suicide attempters who will go on to complete from those who will never make another suicide attempt, and given that there is no robust evidence that one can differentiate between attempters and completers except after the event, a conservative approach seems justified. Therefore, any suicidal behavior should be seen as presaging completion. Primary prevention could prevent the initial occurrence of suicidal ideation or behavior; secondary prevention could prevent tionlethal suicidal behavior from progressing to death. We have adopted this approach in organizing this paper and have grouped as "primary preventive measures" interventions that

1.Alter the set toward suicide in unaffected individuals by direct interventions with normal, nondisturbed schoolchildren or in special services for survivors. 380 PREVENTION OF MENTAL DISORDERS

2. Facilitate early identification and treatmentof conditions known to predispose children and teenagers toward suicidal behaviorbefore suicide is contemplated. Secondary preventive measuresare those designed to reduce the potential for completion among those who have already threatenedor attempted suicide, by 1. Removing the means for committing suicide; 2. Providing emergency crisis interventionsat times of distress; and 3. Providing treatment after the crisis haspassed. Who Is at Risk? Strategies for Determining Risk Three possible strategies, two of whichare practicable, can be used to determine suicide risk factors in childrenand teenagers. The first possible, but impractical, strategy is to undertakea prospective longitudinal study of an unselected group of young people whoare about to enter the period of enhanced morbidity (e.g., 15-year-olds) and obtainbaseline information thatcan be referenced during the period of prospectivesurveillance. The second strategy is to undertakea control study comparing, retrospectively, the characteristics ofa representative group of subjects who havecom- pleted or attemptod suicide to thoseof a control group of other nonsuicidal disturbed children (if the assumptionis that mental disturbance isa p: 3req- uiE;te for suicide)or normals or both, and to identify the characteristicsthat differentiate the suicidal from the nonsuicidalgroup. It is probably easier to identify a representative sample of completedsuicides, which are subject to reporting requirements, than of attempters.Not all suicides may be officially reported, but there is no clear evidence thatunderreported or misreportedcases are unrepresentative of reported suicides (Shaffer andFisher 1981). Informa- tion must be obtained through "psychologicalautopsy," that is, retrospective history-taking from surviving informantswho knew the proband duringlife; however, this is an inherently incompleteprocess and the limits of its reliability are not known. Several uncontrolled psychologicalautopsy studies of childor adolescent suicide have been carriedout (Shaffer 1974; Sanborn et al. 1974; Jan-Tausch 1964). However, the onlycontrolled study reportedso far (Shafii 1985) consists ofa small sample and used acquaintance controls,practices that can be criticizedas being open to bias. We are currently undertakinga large population-based psychologicalautopsy study of completed suicides with a randomly selected normal controlgroup and a matched comparisongroup of suicide attempters. Partial datafrom this study are presented below. Obtaining a ropresentative sampleof attempted suicides isa considerably more challenging task. Most descriptions of attemptedsuicides are drawn from ,o- YOUTH SUICIDE 381 consecutive cases that have been seen and treated at a particular clinic. However, clinics vary in the type of cases they attract, and by no means all children or teenagers who self-report a suicide attempt will receive treatment for it (Shaffer and Caton 1984; Garfinkel et al. 1982). Little is known about the characteristics of attempters whose attempt is not brought to clinical attention, and the need for population-based research in this area is considerable. The third strategy involves a high-risk, prospective longitudinal study to determine the characteristics that distinguish these who will die from those who will not. Because the incidence of suicide is low (see below), a large sample is required even for high-risk individuals. Clearly, the longer the followup, the more complete the information, because the period of risk for subsequent death is not finite. The high-risk sample should be representative; thus in a followup of teenage attempters it would be desirable to follow up all suicide attempters within a given, defined geographical area. There have been at least three followup studies of adolescents who were initially hospitalized because of severe depression or a serious suicide attempt and on whom baseline information was obtained at the time of their original presentation (Motto 1984; Otto 1972; Garfinkel 1982). All of these studies nport relatively high subsequent death rates (about 10 percent for boys and 1 to 2 percent for girls). There has been one followup study of attempters (Goldacre and Hawton 1985) who took an overdose and who were not subsequently hospitalized but treated as outpatients. This study reports a subsequent suicide rate approAimately one-tenth as high as suggested by the inpatient followups. Descriptive Characteristics of Young Suicides Much of the family and diagnostic information in this section is drawn from the New York State/Columbia University Study, which, although still incomplete with findings subject to revision, is being quoted because it is by far the largest of the range of studies reported above. Demographic Characteristics

Age Very few children younger than 12 commit suicide, although many will threaten and some v., ill make suicide attempts. Suicide becomes increasingly common after puberty, and its incidence increases in each of the teenage years, to reach a peak (National Center for Health Statistics 1983) at age 23. Shaffer and Fisher (1981) review the demographics of suicide. Scx In the United States, teenage boys commit suicide nearly five times more often than girls, although ethnic data in the New York study suggest that the ratio is considerably less in Hispanics and somewhat less for blacks.

3 c 382 PREVENTION OF MENTAL DISORDERS

Ethnicity In general, suicide rates in whites are higher than in blacks. The difference varies in different parte of the United States, black suicides beingunder- represented most in the South and least in the North Central States. The incidence of suicide varies a good deal in different American Indiangroups (Shore 1975). Some have rates more than 20 times higher than the national Lverage; others approximate the Nation's rate as a whole. Geography The youth suicide rate in the United States is byno means uniform. Fates are highest in the western States and Alaska and lowest in the South and North Central and North Eastern States. The reason for these differences isnot clear. It may reflect the demographic szructure ofa region, the ready availability of firearms, or the region's genotic makeup. Secular Trends During the past 25 years, suicide has become lesscommon in middle-age and older groups but more common in theyoung. The increase has been greatest (threefold) among white males ages 15 to 24, withan increase noted nearly every year. The rate of increase for black males during this period has not beenas great. There has been only a very small increase in the suicide rate forgirls across all ethnic groups. In 1983, there were promising signs that the increasewas leveling off. That promise, however, was not fulfilled; in 1984, the suiciderate in white make reached a record high. Suicide rAtempts are more difficult to monitor than suicide deaths, butthere is evidence that attempts have also increased during this periodin both sexes (Weissman 1974; O'Brien 1977). The Suicide Methods Both boys and girls are most likely to commit suicidewith a firearm. The next most common methods are hanging, for boys, and jumpingfrom a height, for girls. Drug overdose, which is by far the mostcommon method in suicide attempts, is unusual in completed suicide. However, itcannot be assumed that young people who take an overdose are makingan empty gesture and do not wish to die. Most suicide victims have a mixture of feelings aboutwanting to die and wanting to live. Because an overdose is notan immediately lethal method for committing suicide, it gives the victim time to reconsider theact. Fortunately, most overdoses can now be successfully treated if the patientreaches a medical center in time. 3 YOUTH SUICIDE 383

Precipitants Many teenagers commit suicide soon (often withinhours) after finding out that they are in trouble, when they are afraidand uncertain about the consequences (Shaffer 1974). Other, less common, precipitants include rejectionand humiliations, for example, a dispute with a girlfriend or boyfriend,being "ragged" or teased, failing at schoul, and failing to get work. Very few cases occur with no obvious precipitant, appearingto have been planned many weeks or even months in advance. Such suicides are morelikely to take place on certain dates, for example, onthe anniversary of a friend's suicide or on or close to the birthday of the victim. Associated Mental Health Problems Most youth suicides are precipitated by a stress, forexample, getting into trouble, breaking up with a boyfriend or girlfriend, or havingproblems at school or arguments with parents.However, these common stresses of adolescence affect countlees teenagers every day who do not respondwith suicidal behavior. To explain suicide, we have to look beyond the stressor to somefeature of the individual's personality or to a coexisting mental illness. Some of the summary information on the teenager whohas committed suicide is based on provisional data from the New YorkState Psychiatric Institute Project (Shaffer et al. 1988). As this is still in progress,findings should be regarded as provisional. Studies in different age groups and in different countries aresurprisingly consistent in showing that 25 to 40 percent of suicides have made aprevious known suicide attempt. These rates were found by Shaffer both in hisBritish study of children younger than 15 (1974) and in the New York study.Similar rates have been reported among adult suicides inthe United States (Dorpat and Ripley 1960; Robins et al. 1959), in England (Barraclough et al.1970), and in Scotland (Kennedy 1972). There is no information about whether the proportion of suicides preceded by a known attempt varies with sex or ethnicity. Drug and alcohol use and a predisposition to intense methods and aggressive outbursts are common in suicide victims. These problems may occur alongwith periods of depression. Learning disorders also appear to be common inthis group. Uncomplicated depression without any associated behavior problems is lees common but is found among some girls and among someolder boys (Shafii 1985). Brent (1987), in his study of 27 teen suicidee, found that bipolar symptoms were common, but that depressive features werenearly always found in association with some other diagnosis. 3 r 384 PREVENTION OF MENTAL DISORDERS

A subgroup of teenage suicide victimsdid not appear toanyone to have a problem. However, .such teenagers wouldworry a great deal about getting things "just right." They becameexceesively and unnecessarily anxious before tests, not because of parentalpressure but because of their own anxiety about performing well. Theywere unreasonably distressed at times of change and dislocation, such as moving toa new home or changing to a new school. Only a small proportion ofteenage suicides occur among teenagerswith manic-depressive or schizophrenic psychosis,because these :xinditionsare relatively rare. However,among teenage patients suffering from psyrhosis, therate of suicide is extremely high. Consistent biochemical abnormalitieshave been found in the braintissue of adult suicides and suicide attempters.The abnormalities consist of low levels of 5-HIAA, the breakdown productof the neurotransmitter serotonin(Stanley 1984). This abnormality has alsobeen found in individuals with aggressiveor impulsive tendencies. Suicide and Imitation

An accumulation of evidencefrom a number of differentsources, taken together, indicates that imitationmay be an important facilitator of suicidal behavior in young people. This evidenceincludes the following: 1.Phillips (1984) demonstrated thatprominent display of thenews of a real suicide in newspapers leads toan increase in suicidal deathsmainly among young peopleduring a 1- to 2-week periodfollowing the news. 2. Gould and Shaffer's (1986)research showed that suicide completionand attempt rates inc-eased during the 2week, following fictional television shows dealing with adolescentsuicide. 3. Kreitman (1976) noted thatyoung attempters had many more close contacts with others who had madea suicide attempt than would be expected from the known prevalenceof suicide attempts in the community.

4. Documented examples of imitativesuicides have taken place withina few hours after a vulnerableteenager has seen a film, reada book, or seen a news story featuring suicide (Shaffer 1974). 5. The occurrence of suicideclusters is thought to dependon imitation. It appears that teenagers who die ina cluster have not usually known each other personally, but have readabout the others' deaths duringthe extensive localnewspaper coverage of each case. These findingsare clearly relevant to suicide preventionactivities, which often involve presenting thefacts about suicide toa child or teeage audience. A major challenge is howto present such findings, ifnecessary, without encouraging imitation.

3 t1 ,!J YOUTH SUICIDE 385

Past and Family History Salk (1985), in a study on a representative sample of youth suicide completers, noted an excess of obstetric complications in their obstetric histories. It is not clear whether the obstetric complications are directly related to the later suicide or whether they are an indirect index to some other predisposing factor, for example, maternal alcohol consumption or maternal ingestion of psychoactive drugs. This is an area where more research is needed. No evidenoe clearly shows that young suicides or suicide attempters have an excess of early loss experiences, but studies to establishthis are difficult to undertake and the possibility cannot be ruled out. A high proportion of suicide completers have a first- or second-degree relative who previously attempted or committed suicide. It is not clear whether this indicates a genetic predisposition or whether prior example facilitates a later suicidal response to stress. Imitation may be the important factor here, because studies into the acquaintances of suicide attempters suggest that attempters are more likely to have a close friend who has similarly made asuicide attempt. The association between suicide and aggressive behavior seems clear. Ag- gression in young people is, in turn, frequently associated with a history of exposure to familial aggression. However, there is no directevidence to indicate whether suicides or attempted suicide* are more likely to have experienced intrafamilial aggression in early childhood. Righ-risk Groups Suicide is uncommon; only 12 out of every 100,000 teenagers between ages 15 and 19 will commit suicide each year. This means that most preventive efforts directed to the general population ("universal" strategies) will be inefficient, reaching relatively few vulnerable teenagers. A more effective strategy would be to use available resources on a high-risk group, thus reaching many more potential cases. Preliminary findings from the New York study (Gould et al. in press) and Bayes' theorem (Fleiss 1981) have been used to estimate the probable incidence of suicide in various subgroups (see table 1). For theee estimates, the proportion of suicide completer. to normal controls was used with a particular risk factor, along with sex- and age-specific suicide rates from the general population (e.g., 15 out of 100,000 for 15- to 19-year-old white bays). The utility of these estimates lies primarily in their ranking rather than in any absolute number, because estimates are based only on univariate analyses. A review of the few followup studies of teenagers who have been treated psychiatrically for suicide attempts or depression (see table 1) shows suicide rates that are much higher than we would project from data on subjects in the New York study or from Goldacre and Hawton (1985) study of attempters who, for the moat part, received nonpsychiatric treatment.The most plausible explanation for the difference is selection factors, that is, more severely 140i1 886 PREVENTION OF MENTAL DISORDERS

Table 1. Risk Factors for Suicide inTeenarrs New York Suicide Study

Males

Suicide Rate/ 100,000/yr ApproximateAffected by Risk Factor Normals Suicides Odds Ratio Risk factor (N = 65) (N = 97)

(General Population) 14a Prior attempt 1.2% 21% 22.5 270 Major depression 2.0% 21% 8.6 100 Substance abuse 7.0% 37% 7.1 70 Antisocial behavior 17.0% 67% 4.4 40 FH of suicide 17.0% 41% 3.0 35

Females (N = 20) (N = 17)

(General Population) 3.5' Prior attempt 6% 33% 8.6 20 Major depression 2% 50% 49 80 Substance abuse 7% 5% 0.8 3 Antisocial behavior 12% 30% 3.2 8 FH of suicide 13% 33% 2.7 6 aN.C.H.S. Division ofMorta14unpub1ished data

Q n YOUTH SUICIDE 387 disturbed patients are admitted for psychiatric treatment; however, one cannot rule out the possibility that hospital admission worsens tile attempters' prognosis. Analyses such as these do not tell the whole story. Some groups that have very high suicide rates but account for only a small proportion of all suicides (e.g., schizophrenics) have been omitted; other groups may be difficult to identify and access. However, these data suggest that the greatest preventive impact would come from effective intervention directed at teenage boys who have made a previous suicide attempt or who are depressed. Attempters are poten- easy to identify because many are referred to emergency rooms for medical care after their attempt (Kennedy et al. 1974). Furthermore, because they have already demonstrated suicidal behavior, there would be no concern about the introduction of suicidal preoccupations to individuals who may not be suicidal. Evaluating Prevention Programs

There are two principal types of evaluation strategy 1. Auditing the usage patterns and costa of a service.This is clearly important, but may not reveal much about the program's efficacy. 2. Evaluating a program's efficacy.This will always require some comparison (a) among individuals before and after they have received the intervention, or (b) between individuals or communities who have received and those who have not received the intervention, or (c) both (a) (b). Before-and-after comparisons require that the outcome to be measured has been systematically recorded before the start of the intervention. This presents no problem when a before-and-after study has been planne i in advance, but sometimes it would be useful to be able to evaluate the efficacy of a program that came into existence without any plan. In these instances, recourse to old records may be useful. Such records may include the number of suicidal deaths recorded within a community, the number of visits to local emergency rooms for suicide attempts, and the number of suicide hotline calls. Unfortunately, the before-and-after method may be invalid if marked shifts in the suicide behavior rate over time aro due to a factor other than the introduction of the prevention program (such as a shift in the age pattern of a community or another, more specific, factor such as the availability of a method for committing suicide). To reasonably link the program to a change in suicide rates, it is also necessary to compare the target community to a matched control community where there was no prevention program but where any general effects would still be operating. When the outcome one wants to study was unrecorded before the start of the intervention and does not exist in any available records, it is necessary to make comparisons with control communities or subjects who were known not to have experienced the program. 3 9 2 388 PREVENTION OF MENTAL DISORDERS

The timing of followup is important An intervention program may have important short-term effects (e.g., it may induce teenagers who are currently tesubled to seek referral, or it may possibly iniuce suicidal preoccupations in other troubled youngsters) that would be best picked up by an evaluation carried out within days of the intervention. However, it may have longer term effects (e.g., by teething appropriate coping strategies, it may reduce stress responses when the young people encounter later problems) that would not be shown by an evaluation limited to the period immediately following the termination of theprogram. Finally, it is important to consider what is being evaluated. For example, the program's proponents may feel that improved knowledge about certain aspects of suicide is important to the program's success. They may devisea before-andafter test showing that youngsters who go through the program do indeed acquire new knowledge about suicide. However, their theory that knowledge is important may be incorrect, because the acquisition of knowledgemay have no effect on subsequent suicidal behavior (or it mayeven have the effect of increasing the unwanted behavior). That would become apparent only if the evaluators studied both the acquisition of knowledge and the effect of the program on suicide attempt or completion rates. Before-and-after comparisons with teenagers are often difficult to interpret. This is because teenagers' and children's ideas, perceptions, and levels of understanding change as pert of the process of normal development. Any change identified at a followup examination may bea consequence of this growth and development rather than of the intervention. This ismore likely to be a problem if the followup period is lengthy and ifone is assessing cognitive skills such as coping strategies,It is le% likely to be a problem if one is measuring straightforward knowledge, such as a list ofresources. To control for mature- tion overtime, it is necessary to examinea separate age- and IQ-matched control group that did not receive the intervention. Primary Preventive Interventions General Psychiatric Care Psychological autopsy studies show that most suicides had symptoms ofa psychiatric disorder at the time of their death. Other factors being equal,the introduction of mental health services intoa community should, by reducing the burden of mental illness, also reduce the suicide rate. This proposition hasbeen studied in the general population although not specifically for childrenor teens. Walk (1967) examined suicide rates in the British eounty ofSussex and found no effect on suicide rates after the introduction of a community psychiatric service. Neilson and Videbech (1973) similarly foundno effect on suicide rates after the introduction of a psychiatric serviceon the Danish island of Sameo (the rates included all ages, but teenagers accounted forvery few cases). However, neither study was controlled,so it is possible that the apparently stable rates were occurring at a time of a more general rate increase. More important, they

3,r)3 YOUTH SUICIDE 389 were undertaken before the widespread use of antidepressants and lithium and so do not reflect the impact of more recent antidepressant therapies. Restricting Access to Suicide Methods Because youth suicide is so often an impulsive act, it is reasonable to expect that limiting the availability of or access to any common method could reduce the teenage suicide rate. The so-called "British experience" is the most frequently cited example of how reducing access to the means of su icide can significantly reduce the suicide rate. In 1967, the mean carbon monoxide content of domestic gas in Great Britain was 12 percent. Self-asphyxiation with domestic cooking gas accounted for more than 40 percent of all British suicides (Hassell and Trethowan 1972; Kreitman 1976). By 1970, the carbon monoxide content was reduced to 2 percent. During this period, British suicide rates from carbon monoxide asphyxiation declined; by 1971 this method accounted for fewer than 10 percent of all suicides, and the overall suicide rate declined by 26 percent. Almost all of the reduction could be attributed to the decrease in deaths from domestic gas asphyxiation; there was no compensatory increase in suicidal deaths by other methods, although the incidence of attempts by overdose increased (Johns 1977). It is implausible to assume that individuals who would have committed suicide using domestic gas did not turn to different methods to achieve their suicidal goals. What may have happened was that the suicidal population, denied access to a universally available, nondeforming, nonviolent method, did not turn to other more violent (and more lethal) methods, but rather chose another readily available nonviolent method:self-poisoning.However, the impact was limited because during the same period belf-poisoning became progressively leas lethal. This may have been partly because of the substitution of tls less dangerous ben- zodiazepine drugs for the highly toxic barbiturates, and partly because of haproved resuscitation methods. The detoxification of domestic cooking gas also occurred in other European countries, such as the Netherlands, but it was not associated with any reduction in suicide rate. In these other countries, however, the base rate of self-asphyxiation from domestic gas was less than 7 percent compared to Britain's 40 percent. Prevention methods that do not require the active participation and educe- tion of the public have traditionally been the most effective, for example, changing the water supply in South London at the time of the great London cholera epidemics.Such methods may also have potential for preventing suicide. Clearly, the British case of reducing the toxic components of domestic gas is not directly relevant in the United States, where suicide by self-asphyxiation is rare and most likely to occur through inhaling the exhaust of an automobile engine. Mnst suicides in the United States are committed with firearms, and it has been suggested (Boyd and Moscicki 1986) that the increasing penetration of firearms into U.S. households is a leading reason for the increase in U.S. suicide rates. This reasoning is not entirely convincing, because 394 PREVENTION OF MENTAL DISORDERS youth suicide rates have similarly increased in many European countries where Orae-y.ss are less readily available. Furthermore, there seems to be an association between alcohol and other drug abuse and suicide by firearm (Shaffer et al. 1988), so that the increase in suicide deaths may reflect the increasing role of alcohol and drug use rather than increased firearm availability. Even if this were the case, however, the suicide rate, at least among alcohol and drug users, might decline if access to firearms were better controlled. Regardless of whether firearms are the cause of the perceived increase in the number of adolescent suicides, they are the most common method for committing suicide. This knowledge makes clear an opportunity for suicide prevention through appropriate education programs reminding professionals to counsel parents of vulnerable children either not to maintain firearms in the home or to make sure that firearms are maintained under the most secure conditions. Similar programs might be directed to the public with great advantage. School-based Programs Teenage suicides sometimes occur in clusters. A community that hadpre- viously experienced few suicides might suddenly be faced with four or five deaths within a few weeks. These outbreaks are often highly publicized andmay be attributed by the lay public as being a consequence of specific stressesor faults in the community (e.g., the presence of many recent migrants or an insensitive or uncaring school administration). These explanations are implausible, because the putative stresses will nearly always have been present for some time, and ifsuch general influences were at work, one might expect an elevated suicide rate to operate continuously rather than episodically. This model may "mad to demands for a community-based suicide prevention program closely tied to the educational system. Such programs abound; in 1976, approximately 400 such programs were in operation. The goals of most of these programs are:

1.Teaching pupils to recognize certain clinical features or warning signs thought to presage seicide. This is based on the reasonable assumption that potentially suicidal students are awe likely to discuss their feelings and intentions with other st4dents. In addition to this instruction (which often inccrporates a nonstigmatizing model of suicide, notas a mental illness, but as a response to stereotypical teenage stress), teenagersare given advice on some of the strategies that can be used to break suicidal intentions made to them in confidence. They may also be taught listening skills to promote trust and disclosure from potentially suicidal students. 2.Prov!ding students and teachers with source information, for example, how therapeutic services operate and how they can be acceesed. 3.Describing common stresses encountered by adolescents and the psychological development of the adolescent and its alleged relationship to suicidal crises. 391 YOUTH SUICIDE

coping strategies. These 4. Providing trainingin stress management or attempt to identifyand support studentswith drug or programs also and so forth. alcohol problems, failinggrades, parent problems, Description either small or large groups orboth. Programs intended Programs may involve social studies or be incorporated into aregular curriculum (e.g., for students may substitute for regularclasswork. physical education) orbe offered as a special suicide, a parent orteacher's Small groups focus onindividual reactions to confidentiality, how best tocommunicate with teenagers, difficulty in breaching and preeent information onteenage suicide, resources, etc. Urger meetings rarely put for- and difficulties.These presentations adolescent development instead such stresses as pressureto ward a psychiatricviewpoint, emphasizing mobility, changing valuesystems, and use ofdrugs and succeed, residential usually inoludes symptomsof an acute alcohol. The descriptionof warning signs depression and what areconsidered pathognomonicbehaviors, onset episode of marked giving away valuedpossessions, and showing a such as making a will, often includes a film, change in academicperformance.The presentation the clinical features of asuicide-pronti child, some- sometimes to demonstrate skills, but most often toexcite times t., doomonstratetherapeutic or listening interm or concern in theproblem. typically psychologists orsocial workers recruitedfrom Course leaders are suicide prevention or crisis local mental health centers.They are sometimes in grief counseling,but rarely in center volunteers whohave received training Some programs aredirected solely to teachers, adolescent psychopathology. techniques directly totheir who are then expected totransmit information or for teachers to becalled on to implement new students. It is not uncommon professionals, or single training sessionfrom mental health programs atter a educator programs even to rely onthe train:mg theyreceived indirectly during or throughvideo-taped training mullions. graders and older,although Student programs aregenerally offered for ninth schools offer programs tu youngerstudents. We have not some intermediate screening to exclude children encountered any programsthat use preliminary who might respondadversely to the programcontent. has a therapeutic pro- Among special programs wehave encountered, one Graduates of this programaddress gram forpreviously suicidal students. class. Others use peercounselors, who are usually classmates during a social Candidates are then taught picked by a panel of teachersand administrators. credit, usually by a guidancecounselor. They raay counseling techniques, for "the care company" then be given such names as'natural helpers,""buddies," or and are taught earlywarning signs and referraltechniques. distribute small, printedwallet cards with factsabout Many programs suicidal intention in a friend suicide, warning signs, stepsto take if one suspects numbers of hotlines in the area.The cards or familymember, and phone "4 o 392 PREVENTION OF MENTALDISORDERS

typically includea statement concerning theneed to override individual's request for a suicidal confidentiality. The cardsmay include the numbers of such communitygroups as drug and alcohol clinics, control centers. and social service audpoison Evaluation There has been only one systematic controlledevaluation ofan in-school program (Shaffer and Garland1987, eee below), number of informal although there havebeen a reports that have takeneither referrals rates as an index of or changing suicide efficacy. These dependentmeasures are not easily interpreted. Rees (1980),describing the San Mateo the introduction of County program,reported that the program ledto an increase inreferrals to a suicide prevention program.Because the program included elementsdesigned to increase effectivereferrals for suicideattempts and threats, a legitimate index of this may be success.However, a referralincrease would also been seen if theprogram had had the paradoxical have behavior. effect of increasingsuicide The 1985 annualreport of the Fairfax Center reported that County Youth SuicidePrevention in the year before itwas started, therewere 11 teenage suicides in the county,during its firstyear of operation there the second were 5, and during year only 3. This reportillustrates several problems: 1. The number ofteenage suicides is smallin most tive areas, resulting educational-administra- in insufficientopportunity to demonstratea program's efficacy (orthe reverse) by the program's examining deaths beforeand after introduction.This problem couldbe circumventedif (a) severalprograms with broadly similar together, or (b) characteristicsare pooled a more prevalent index ofsuicidality (e.g., suicide tempts) is usedas the dependent variable. at- 2.If a program isestablished after a suicide cluster, theimmediate preprogram baseline isunrepresentative. In these must be used, and because cases a long baseline teenage suicide rateshave been unstableover the past 3 decades,control communities should also be studied.However, because suicide ratesshow marked ethnic munities need to be differences, the controlcom- demographically matchedto the index community. A less complex approach to measuringthe effects of short-term changes in programs is to measure knowledge aboutresources, warning signs, and and changes inattitudes toward so forth, help-seeking in studentswho attend such programs. Although thesevariables do not provide whether a direct informationabout program has succeeded inits goal of preventing provide informationon whether the suicide, they will steps that are program has ackiressed theintermediate hypothesized to be thecrucial mechanismsfor the preventive intervention. Shafferand Garland (1987) assessed changes insuicide-related knowledge and help-seekingattitudes in ing 12 high schools. approv;mately 2,000 childrenattend- Approximately 1,000 ..re assessed before andafter 3 7 YOUTH SUICIDE 393

of school-based intervention.One thousand attending one of three models given. The controls attended differentschools at which no programs were students ranged in age from 13to 18, most being 14 to 17.Control schools were matchad on size, ethnic mixture,urban/rural situation, and rankin average reading scores. The matchworked weil and there werefew significant differen- populations before the program. ces in knowledge orattitudes between the two Before experiencing the programs,most students held viewsand had knew many of the knowledge that generally wouldbe considered sound. They likely warning signs and took the viewthat mental health professionals were views. However, about to be helpful. The programsdid little to change these 20 percent of the pupils inboth experimental and controlschools expressed views that would generallybe regarded as tares forintervention. They believed that suicide was areasonable course of actionunder certain circumstances, they would notreveal the suicidal confidenceof a friend, and they would not see a mental healthprofessional for help if they felttroubled. The program did little tochange these attitudes.Approximately 8 percent of each sample admitted to havingmade a suicide attempt. Whentheir responses to the programs were analyzedseparately, it was found thatthey responded, paradoxically, in a consistently morenegative fashion. Not only werethey lees likely to find the programinteresting or satisfactory, but alsothey were more likely to have been troubled bythe program; if anything,the programs appeared and beliefs in a generally negativedirection. Thus, at to change their attitudes who had not the second examinationthey were less likely than attempters attended an intervention to want toreveal suicidal preoccupations toothers and to believe that they couldbe helped by a mental healthprofessional, and they were more likely tobelieve that suicide was areasonable course of behavior. These results need to bereplicated, although they wereremarkably consistent between both programsand schools. They address only paperand pencil responses and provide nocertainty that more negativeattitudes would necessarily be translated into behaviors.Their failure to induce useful change may or may not havebeen a function of their quality;again, there seemed to be few differences between the programs inthoir impact, and although oneof the programs was seen asmuch more interesting than the othertwo, it did not differ from the others in the negligibleamount of change that itinduced. The failure to respond may finally havebeen a matter of intensity. The programscalled for relatively brief involvement, and it maybe that more significant changesrequire either a different teaching strategy(although a small group programdid no better than a large group one) or greaterlength of involvement. Regardless of these caveats, one cannot escapethe conclusion that the findingsdo little to support the value of generaleducation programs. Most students donot need them, and those that do may beadversely affected by them. be At a more general level,school-based suicide prevention programs can criticized for following a low-risk strategy;that is, very few of the teenagerswho receive the programs would belikely to attempt or commit suicide, 3 8 394 PREVENTION OF MENTALDISORDERS

Shaffer and Garland's finding thata significant proportion of teenagers take the viow that they would never seek the help of a mental health professionalis an appropriate focus for an intervention, butthere is a clear need to understand the reasons for this better beforedesigning an intervention thatwould correct it.We need to know to whatextent it is aconsequence of adverse past experiences. The Shaffer and Garlandstudy suggested that thismaybe a factor because the teens who had madea previous suicide attempt weremore likely to hold this view, and one would expect that they were similarlymore likely to have had direct experience withsuch professionals. Programs emphasizing generalcoping skills and providinga helpful perspective on common adolescentproblems may ormay not help the many teenagers with minor problems. Noneof the programs assessedin the Shaffer and Garland study employed thesestrategies, and other researchon the impact of such programs is not available.

One other finding from theShaffer and Garland studywas that relatively few students took the vieweitherbefore or afterexposure to a programthat suicide was a manifestation of mental illness. Programdesigners want to encourage self-initiated referrals; to do thisthey need to minimizeany stigmatizing correlates of suicide,for example, that it isa feature of mental illness. This is a matter ofsome concern because there is evidencethat it is such, and that knowledge might alsobe expected to reduceimitation. Imitation is inherently more likely if suicideis portrayed asa tragic, heroic, or romantic response to a nebulous stress suchas pressure imposed by an uncaring adult world or school; itmay be less likely if portrayedas a deviant act by someone with a mental disturbance. Most existingprograms emphasize the role of stress as a cause of psychiatric symptoms ingeneral and suicide in particular. AL hough the appropriatenessof school-based didacticprograms must be questioned, suchprograms undoubtedly offeran opportunity for case finding, and the school as a clinica base offers advantages thatshould not be ignored by the child psychiatrist. Secondary and TertiaryInterventions Secondary interventionattempts to preventan already established condition from causing additionalsuffering or impairment. Inthe suicide context, this might mean intervening witha suicide ideator to prevent himor her from attempting or completing euicideand is exemplified by thesuicide hotline. Tertiary interventions (whichare usually the same as "treatments") intended to shorten the are course of a condition, lessen thelikelihood of its recurrence, and reduce its noxiousconsequences or complications. Television in the Serviceof Suicide Prevention Because of television's broad reach and its proven (inadvertising) efficacy in inducing behaviors andmolding attitudes, it has public's been used to increase the awareness of prevention programs andto illustrate their efficacy. 3 ( YOUTH SUICIDE 895

Holding (1974, 1975), in Edinburgh,examined the impact of an 11-part weekly television series, "TheBefrienders." Each program centered onthe predicament of a suicidal individual who wasthen helped by the Samaritans. During the season when the programs wereshown, referrals to the Samaritans increased by 140 percent. However,there was no change in the numberof suicides or attempted suicides treatedby hospitals in the city, even though in previous years the number of suicideshad declined during that season(Holding and Barraclough 1975). This suggeststhat television was effective in increasing referrals to the suicide preventionservice, but that only nonsuicidalindividuals were attracted tothe service (which is implausible),that the service was ineffective, or, given the absence of the usualseasonal reduction in suicides, that any positive effect onincreasing referrals was counterbalancedby inducing suicidal behavior in others. This last possibility is not so farfetched.Gould and Shaffer (1986) examined the impact of four isolate televisionpresentations that dramatized the problem of youth suicide. The pl ograms weretargeted to a teenage audience, heralded with publicity in local high schools,and, to a varying degree, coordinated with community programs. The incidence ofcompleted teenage suicides in three states and the number of attemptedsuicides treated at six large hospitalsduring the 4 months before and after eachshowing were examined. Both suicides and attempted suicidesincreased significantly among teenagers during the 2 weeksfollowing three of the four programa. Afterthe fourth program, both suicides and attemptsdeclined but not to a statistically significant extent. Had the effect been confinedto an increase in attempts, it might have been attributable to moresensitive referral practices, resulting in more suicide attemptersbeing sent for evaluation. However, thesimultaneous increase in deaths suggests that the programsinduced suicidal behavior. The discrepant effect of one of the programsstudied by Gould and Shaffer, even though notstatistically significant, raises the possibilitythat special features in the dramatization of youth suicideeither may not cause unwanted effects or may serve to prevent suicidalbehavior. These features could lie in the presentation's context (for example, the innocuousfilm included names and telephone numbers of various clinicalservices, thus placing suicide in a pathological rather than a romantic context) orin its content (the same film emphasized the sequelae of the death ratherthan the plight or s4 ation of the victim). This is clearly an important area forfurther research, with significant implications for prevention. Hotiines and Crisis Services A rationale for crisis intervention has beenprovided by Schneidman and Farberow (1957) end Litman et al.(1965, pp.324-344). This can be summarized as follows:

1.Suicide is often associated with a critical stress event.

4 0 (1 396 PREVENTION OF MENTAL DISORDERS

2.Suicide is usually contemplated withpsychological ambivalence. The surviving attempter oftenreports that wishes to die existedalong with wishes to be reLlcued and saved. 3. The wish to commit suicideas a solution to the problem arises in the context of mental disturbance. However,there is partial insight intothe unsatisfactory nature of' this solution(hence the victim's ambivalence) that can be identified anddealt with by those with specialtraining. Most crisis services centeron a telephone service, the hotline, thatoffers several practical advantages.It is available outside usualoffice hours; it offers the caller in crisisan opportunity for immediate support; itis anonymous, allowing callers to say shocking or embarrassing things, whichthey might find difficult in a face-to-face interview; and it gives those whoare concerned with issues of control andpower the freedom to hang up. There are several different modelsof crisis service, but Bridgeet al. (1977) suggested the following criteria,which many modelsmeet. A crisis service should have (a)an identifiable individual in the communitywho is retsponsible for it, (b) 24-hour telephoneor other emergency access, and (c) of its presence. advertisement Some services target special groups, such as college students; at leastone has a telephone at a repeated suicide siteon a bridge. Many of the calls received byhotlines deal with matters other In one study of than suicide. a hotline for teenagers (Slem and Cotler1973), the mostcommon reasons for calling were family and schooldifficulties, relationship and loneliness. Both adult problems, and teenage callers oftenreport feelia. unlovedor misunderstood. Most callsare made by females. Masturbatorycalls are common among the minority of male callers. Most centers are locally organized;one exception is the National Adolescent Suicide Hotline, which hasdeveloped from the NationalYouth Runaway Switchboard. Local centersusually concentrateon giving information about the location of appropriate services; but some, especially thosethat are part ofa multiservice agency, perform as more active case managers, makingappoint- ments with the appropriateservice and followingup if the appointment is not kept. Relatively few, the main exception being the Samarit,ans,try to offer direct therapy on the telephone;their "befriending"procees has been likened to Rogerian psychotherapy withan emphasis on acceptance, warmth,and confidentiality (Hirsch 1981).Most other centers will break believe that it will avert confidentiality if they a suicide. Many do not hesitateto call in police help, and although most willask a teenager's permission others will do before contacting parents, so without permission if the situationseems serious and if they believe that the parent willbe cooperative. .Most centers are staffed by volunteers supervisedby social workersor other mental health professionals.Some hotlinesteenlinesare teenagers who receive the started by same training as adult volunteers.however, teenage 4 1; 1 397 YOUTH SUICIDE

difficulty in dealing with sexcalls, volunteers are likely toexperience special and their hours ofavailability are usuallylimited. Effecfiveness of CrisisServices and Hotlines little research on theimpact of crisis servicesestablished There has been very Cotler 1973) of a widelyadver- specifically for teenagers;the study (Slem and the only systematicevaluation tised hotline for teenagersin suburban Detroit is identified. The followuprate was low of a teenage crisisservice that we have proportion of suicidal users wasunspecified. User satis- (58 percent) and the reported that 68 but morbidity wasnot. The authors faction was evaluated slightly older contacted had a goodexperience. In a study in a percent of callers of 3,000 college King (1977) determinedclient satisfaction in a survey age group, fewer of the malesfound the freshme.a. Most of thefemales but significantly one-third of males andone-fifth of femalesreported hotline helpful. However, Male suicidal users were that using the hotlinehad made their problem worse. females to be satisfiedwith their experienceusing significantly less likely than they had been able totalk a hotline.Both sexes had beenbetter satisfied when to a listener of theopposite sex. focused on whether theestablishment of a suicidehotline Most studies have studies (Litman and reduces the mort6lityfrom suicide. Early in a community Lester 1973) that reported afall in Farberow 1969; Ringel1968; Weiner 1969; service had beenintroduced either usedinappropriate the suicide rate after a shifts in a community,which controls or failed to accountfor the demographic major influence on itssuicide rate. Bridge etal. (1977), taking may exert a the introduction of asuicide crisis account of such factors,found no evidence that suicide rate. Anincidental finding ofthat study, which center affected the although the incidenceof examined North Carolinacommunitios, was that, there was a high proportionof older, white, married suicide was highest where in areas with low persons, suicidecenters had mostoften been established proportions of thr se groups. quoted as demonstratingthat Bagley's British study(1968) has been widely of suicide. Bagleycontrolled suicide prevention centersdo reduce the incidence in selecting controlcommunities, but laterresearch for demographic factors to replicate hisfindings using more elaboratematching techniques failed (Barraclough et al. 1977;Jennings et al. 1978). and race-specific suicide Most recently, Miller etal. (1964), examining age- with and without andbefore and after theintroduction of rates in U.S. counties in the suicide a suicideprevention center, found asmall but significant reduction 100,000) among young whitefemalesa group known tobe the rate (1.75 out of centersassociated with the presence most frequent user ofsuicide prevention of a service. Reasons for Low Impact ofHotlines why hotlines may not beeffective. Miller's There are a number of reasons hotlines most (1984) study, which did find aneffect among the group that uses 402. 398 PREVENTION OF MENTALDISORDERS

frequently, suggested by implication thathotlines might have on the suicide rate if they a greater impact could reach thegroups at greatest risk. In the low utilizationrates by teenagers in this regard, particular and by malesmore generally pose a particular challenge.Teenagers make only general suicide a small proportion of callsto prevention centers (7itman et al. 1965), andteenage attempters are significantly less likelyto know of the attempters (Greer and existence of hotlinesthan adult Anderson 1979).However, knowledgecan be increased with appropriatelytargeted advertising. The Detroit hotlinestudied by Slem and Cotler (1973)extensively advertised tion. Ninety-eight its services to thehigh school popula- percent of studentsrecognized its 5.6 percent of these, name; a not insignificant most female, had madeuse of the service. In King's of 3,000 freshmen(197'4 2 percent of study the students hadused the service for own problems and another1 percent had done their so to report problems inothers. Given that it ispossible to boostutilization with the advice givenacceptable or therapeutic? appropriate advertising,is commonly given The form in whichadvice is most may not be acceptable,because only callers comply with a small proportion of the recommendationsgiven to them reasons for this may be quite (Lester 1970). The complex and almostcertainly involve both manner and the content ofthe response given. the presumably uniform Within a singleservice, with a approach, compliancerates vary with different operators.The critical factors telephone in nspondervariation have notbeen fully explored, althoughthey seem to beunrelated to the characteristics volunteer's conversational or whether the responderrefers directly to euphemism (Slaikuet al. 1975). suicide oruses some Many callers, however,receive poor quality studies simulating information froma hotline. In real callers, Bleachand Claiborn (1974) Hodes (1975) foundthat operators and Apsler and tended to givestandardizedanswers using little judgmentto obtain a good fitwith the callers' noted that volunteers problems. Hirsch(1981) are less skilled thanprofessionals in eliciting history and in beingable to integrate relevant operators are likely information offered bya caller. Unskilled to be more dogmaticand hasty in their (1979) and McCarthyand Berman (1979) responses. Knowles found that untrainedvolunteerswere commonly overdirective,prematurely offering information. Furthermore, advice on the basisof inadequate experience isno substitute for training. Cohen (1982) foundthat only those Elkins and volunteers whoreceived preliminary ing improved theirperformance after 5 train- quality of information months at workTraining affectedthe provided and alsosuch qualitiesas empathy and warmth (Bleach and Claiborne1974; Genther 1974; France 1975; Kalafatet al. 1979). Although it istempting to dismiss hotlines asa well-meaning but therapeutic approach,it should be noted ineffective that theyare widely distributed Inthe United States andthey appear toserve a vulnerable and tion. King's studyshowed that only 8 underserved popula- percent of callerswere currently receiving other mentalhealth services, andthere is abundant rate in hotlinecallers is evidence that thesuicide many times that for thegeneral population. 4r3 YOUTH SUICIDE 399

system, a However, given the vulnerabilitiesof a volunteer-operated hotline more reliableapproach might be to limittheir function to that of case management and, within thoselimits, adopting the mostefficacious methods. Thc. be a model worth service in Cleveland describedby Sudak et al. (1977) may following, because it reportedexceptionally high compliance rates.Hotline operators routinely madeclinic appointments for callers(instead of relying on had been kept, the caller's own initiative),ascertained whether the appointment and if it had not, undertook afurther followup. particularly to develop It is also important tobroaden utilization patterns and troubled boys. Whereas ap- a self-referralmechanism 'that is acceptable to propriately directed advertising maybring in more teenagers ofboth sexes, the problem of attracting boys has not yetbeen examined; a particulareffort is likely to be needed for this. Finally, it is likely that usefuladvances in the technology ofhotline administration will require systematicperformance monitoring and assessment. This should include, but notbe confined to, systematicallynoting the reasons why calls were made and the ageand sex proportions of callersand building in a simple methodfor determining the rate ofcompliance with triage recommendations. Treating Suicide Attempters Although the literature is repletewith suggestions for theappropriate treatment of teenage suicide attemptors(Trautman and Shaffer 1984, pp.307- 323), we have found no satisfactorystudies that have systematicallyevaluated the effects of treating suicidaltbonagers (i.e., comparing outcome over a reasonable period of time withother treated or non-treated groups,the use of standard measures at the initiationof treatment and at followup,the use of random assignment to different treatment groups,or the use ofplacebo or dummy interventions). Most of thestudies referred to in this sectionhave been done on groups unselected for age;therefore, most relate to thetreatment of adult at tempters. The Problem of Noncompliance Whether or not psychiatric treatmentwill help them (the evidence onthe whole suggests that it willsee below),it seems that most suicideattempters who have been brought to amedical emergency room for treatmentof the medical effectA of their attemptand who were then offered apsychiatric appointment will not keep thisappointment! In a small study of 27 adoleecents,Litt et al. (1983) found that only33 percent of first appointments werekept, a similar proportion to that notedin the adult studies quoted below. Failure to attend was morecommon in adolescentswho had made a previous attempt. Chamedes and Yamamoto (1973)found in a study of adults thatonly 36 percent, and Paykel et al. (1974a)that only 44 percent, of referrals madein an 400 PREVENTION OF MENTALDISORDERS

emergency room were kept. Furthermore,many of those who kept their first few appointments would drop out of the psychiatrictreatment programprema- turely (Kogan 1957; Jacobsonet al. 1965). It is likely thatpoor compliance is not necessarily of pathology. Chamedes associated with low levels and Yamamoto (1973)found that a high proportionof referral failures will bereferred to another nrmtal the year after the health professional during attempt. No consistent clinicaldifferences or differences the medical severity of the in attempt were found in thecompliant andnoncom- pliant groups in Paykel'sor Chamedes' studies. The patient's experiencesat the time of the to play some part in ewergency rot,1. visit are likely determining later compliance.In a study of nearly emergency room cases, managed by 300 15 clinicians, Knesper(1982) founda relationship betweennonattendance andcharacteristics of clinicians. Some clinicians emergency room were able to persuade most of theirpatients to attend a later appointment, but otherspersuaded very few. Clinicians the notion of admission who introduced to a patient suddenly andwithout warning at theend of an examination hada very low rate of inpatient unit. success in making referrals toan Emergency room procedurescontribute to compliance recorded 37 percent rates. Kogan (1957) compliance for attemptersseen in an emergency room who were simply given aname and telephone number; made for the patient when an appointmentwas during triage therewas 82 percent compliance. Rogawski and Edmundson (1971),using a more stringentindex of compliance appointments kept), found (two that only 30 percentof those givena name and number kept theirappointment but that 66percent did so when the ment was made for them. appoint- Naturalistic Treatment Studies Will patients who keep their appointmentsbenefit from the treatmentthey receive? Naturalistictreatment studies rely treatment experiences. on the natural variations in clinic These studies typicallycompare the outcomes of patients who attendedfor a longer period withthese who attended periods or who dropped for shorter out of treatment. Thesestudies might also outcome of patients in compare the a new program with those whohad been treated ina preexisting program. Theresults of these studiesare difficult to interpret because treatment is rare4yadministered in in both type and quality a standardized way andmay vary with different practitioners.Nonattenders area poor comparison group becausethey may be eithermore disturbed than attenders (unable to keepappointments because oftheir disturbance) so distressed, they or less so (not being are less motivated to attend).In either case, differences between attenders andnonattenders would be treatment. Studies comparing unrelated to the effectsof old and newprograms may also be misleading because the introductionof a new service referral agent may rely on a different patternof or attract a different type ofpatient from those who original service. attended the

'T(,) 401 YOUTHSUICIDE

psychiatric treat- suicide attempterswho receive suggest that repetitions),suggesting Four studies (with respectto suicide ment have abetter prognosis studies wasconfined toteenagers. is ineffective.None of these that treatment aneffect: studies that appearto demonstrate The following are been suicide attempterswho had (1971) contrasted further appoint- 1. Greerand Bagley roomwithout a discharged from amedical emergency and who kept with thosewho hadbeen given ment due tostaff oversight, reattempt rate was attend apsychiatricclinic. The anappointment to had not beengiven anappointment; amongthose who who attended more significantly higher did less wellthan those attended only once attempt didnot predict those who seriousness ofthe initial than once.The medical reattempt. rates of 204suicide attempters. (1972) reported onthe repetition (thought 2. Kennedy of priorsuicide attempts into accountdifferent rates repetition rates werefound After taking repetition), strong predictorof further admitted for ashort period to be a 142 attempters significantly loweramong outpeklent care to be than among6 who hadreceived to a suicidecrisis unit who received noaftercare. only and 56 fewer suiciderepetitions higher attendancerates and in a 3. Welu(1977) found 63 patientswho were seen followup periodamong 57 cases seen in a 4-month outreach componentthan among programwith a strong program wasintroduced. before the new a sampleof Motto et al.(1981, pp.148-154) identified 4. Motto(1976) and (They hadeither made patients at highrisk for suicide. their admis- 3,005 hospitalized been judgeddepressed during previous suicideattempt or 862 declinedand the a discharge all wereoffered aftercare; sion.) After assigned eitherto receiveintermittent remainder wererandomly 5 years orto receive no intervals over was telephone contactat decreasing in theexperimental group The actualcontact rate between compliers further contact. diagnostic andother differences far fromcomplete, and During the first2 years,suicide were notdescribed. converged and noncompliers moncontacted group,but the rates twice as commonin the thereafter. followup periodand that had thelongest (6-year) However, aDanish study programsdescribed above the clinicalfeatures of the The that incorporatedmany of treatment program. found no effectof acomprehensive daytime hotline (Ettlinger 1975), healthprofessionals, included free acceesto mental requested, closeliaison program home visitsmade whenever suicide rate and walk-inclinics, frequent outreach for 1 year.The 6-year with otherhospitals, andproactive the death ratefor 681 admissions wascompared with for 670consecutive started. The newservice the servicehad been found in attemptersadmitted before but nodifferences were popular and wasused freely, appeared to be between the twogroups. suicide rates orsocial adjustment 406 402 PREVENTION OF MENTALDISORDERS

ExperimentalStudies The optimal design forassessing the assign similartypes of impact ofa treatment is reduces the cases to differenttreatment to randomly chance thata favorable conditions.This design in favor ofor against outcome is dueto some one of the referral biasworking are small,a failure of interventions being studied, randomizationmay occur. although, ifnumbers Chowdhuryet el. (1973) routine randomlyassigned adult outpatientcare or toan enhanced repeatattempters to telephoneaccess and walk-in service thatincluded failed to keep facilities.Patients emergency appointmeats. were visited athome when reattempt rates Thegroups did not they nor on anymeasure of mental differ fromeach otherin followupperiod although state at the theexperimental end ofa 6-month problems (forexample, group experienced benefits) than difficultieswith housing, fewer social the controls. employment,obtaining psychiatriccare was not One cannotconclude from in different helpful becauseit was this studythat amounts. received bybothgroups, albeit Gibbons (1980) to a and Gibbonset al. (1978) course of intensive, randomlyassigned 200 to routine but time-limited(3 months), attempters treatment (somecases were task-centeredcaseworkor generalpractitioner). followed bya psychiatrist, from the Cases judgedto have high others bya randomassignment. suicide intentwere excluded between thetwo No differencesin later groups. adjustmentwere found Lieberman andEckman (1981) tempters eitherto 32 hours randomlyassigneda small of behaviortherapy (social group of at- management, andcontingency skills training, psychotherapy. contracting)or to unlimited anxiety No differences insight-oriented repetition ofsuicide were found inthe two fewer attempts,although those groups withrespect to symptoms,were less in the behaviortherapy threats. preoccupiedwith suicidal group had ideation, andmade fewer Comments A problem with all thetreatment studies treat attemptedsuicide that havebeen reviewed (a) It is as a singlediagnostic is that they highly likely entity. Thiscauses two problems: to superficially that patientswith different similar suicidal psychiatricdisorders interventions. behavior willrespond leading Antidepressanttreatment differentlyto different suicidalyoungster and may be appropriate for the may reducesuicidal risk, for thedepressed patient withan antisocial but it islikely to Le of attempted personality less effective suicidemay be different disorder. (b)The natural which it is according to history associated.Thus Chowdhury the type of suiciderepetition and Kreitmai.'sdisturbance with rates inpatients with study showedhigh bance; byLitman's histories ofchronic are most likely observations,this is the personalitydistur- to goon to commit group of crisisservice clients suicide.Treatment and who controlgroups that 41 7 111111111111mmu"' 403 YOUTH SUICIDE

likely to produce regard to thesesubchassifications are areconstituted without misleading findings. Other concerns are: outcome measures. suicide repetitionand death as 1. Moststudies focus on behaviors, even in ahigh-risk group,and These are bothlow frequency therefore give theresearch little power. have been studiedin suicide currently usedefficacious interventions 2. Few particularly fewpsychopharmacological attempters.There have been studies. whose mental do not, for themost part,refer to adolescents, 3. The studies (Shaffer 1974) fromthose found in health correlates maybe different adults (Robins etal. 1959). showing any which wouldmitigate against Given theseproblems, all of show a notes that fiveof the eightstudies reviewed therapeutic effect, one not an area fortherapeutic from intervention.This is clearly beneficial effect areawhere further,better designed nihilism, althoughit is very much an research is calledfor. BiologicalPredictors ofSuicide repeatedly in thisreview, thebehavioral correlates As has beenpointed out decade, a numberof biological generally nonspecific.In the past along with of suicide are identified Stanleyand Mann 1987); correlates of suicidehave been of suicide would the hope that morespecific predictors this development came clue to themechanisms of These predictorsmight provide a regularization of the be identified. permit thebiochemical suicide vulnerabilityand thereby providing effectivesuicide prevention. suicide-prone individual, of low replicated most oftenis the presence The findingthat has been serotonin) in 5.HIAA (a metaboliteof theneurotransmitter concentrations of and victims.The finding fluid (c.s.f.) ofsuicide attempters the cerebrospinal et al. in 1976in a studybeing undertaken on was firstreported by Asberg have confirmedthis Many, but not all,subsequent studies depressed patients. reported in suicidalindividuals wi0 avariety observation, and ithas now been including borderlineand ag- diagnoses (Stanleyand Mann 1987), who have of primary al. 1982) andviolent prisoners gressive personalitytypes (Brown et have general- (Linnoila et al. 1983).Nonconfirmatory studies attempted suicide proportion of thesuicidal patientshad a bipolar ly been those inwhich a high in that suggested thatserotonergic dysfunction disorder.Goodwin (1986) The c.e.f. is arecipient of metabolites disorder -lay obscurethe relationship. reported high and spinal cord,but Stanley etal. (1985) 1 from bota brain autopsy specimens)and c.s.f. 5-HIAA. correlations betweenbrain (studied on have studiedlimited numbers ofsubjects, The investigationsof c.s.f. 5-HIAA is not the suicidal andnonsuicidal population wad the base rateof low levels in 1 4n8 404 PREVENTION OF MENTALDISORDERS yet known. However, it isclear that attempters. Asberg a range of levelsis found in et al. (1976) haveused this finding suicide the prognosisof those to determinewhether attempters who havea lower level of c.s.f. from those inthe normalrange Seventy-six 5-HIAA differs (adults) were followedup during a 1-year hospitalized suicideattempters 5-HIAA patients period. Twenty-onepercent of the low died during thefollowup period the initialattempt) compared (usually within6 months after elevated levels with only 2percent of those (a tenfoldincrease). Similar with normalor et al. (1987). findings havebeen reportedby Roy The usefulnessof c.s.f. 5-1{LAA agent of prevention as a predictor ofsuicide and, dependskni whether serotonin therefore,as an over time (i.e.,whether they metabolite levelsare stable abnormal state are an index ofa suicidal trait in which theyare depressed only or rather ofan evidence on thisis by no during periodsof illness). The means complete.Van Prang investigators in thisfield, found that (1977), one ofthe early low in about low levels in half thepatients after their depressed patientsremained have beenrelatively brief recovery. However,followup studies individuals have (Traskman-Bendzet al. 1984) and stable levelswhile those of show thatsome know thesignificance of these others fluctuate.We do notyet o1, two patients different patterns.Asberg et aL whose levelscontinued to decline (1987) reported went on to commitsuicide. after their firstattempt. Both If decliningor stable low levels or tertiary prevention are poor prognosticfeatures, then might be servedby routine secondary after a suicideattempt, with ongoing monitoringof patients terns. If abnormal special attentionto those with levels can beshown to be pathological pat- be presentsome film eller more persistent thanthat and to there recovery they representa trait marker), may be a potentialfor theusa of serotonin then long-term predictor metabolism markers of vulnerableindividuals. This as a that may bringa new level of is clearlyan exciting field specificity topreventive efforts. Conclusion The prediction of suicide hasoften been similar topredicting dismissedas a futile enterprise al. 1964), rare events fromcommon ones (Rosen with all theproblems of low 1964; Temocheet acceptable only specificity (high if interventionsare inexpensive false-positive rates) the case insuicide. or efficaciousor bothfar from This argumentmay apply to general directed toa general risk preventiveinterventions (i.e., population), butas descriptive those who commitsuicide increases, knowledge aboutthose so that we should so will our knowledge eventually be able of more specificrisk factors, risk. Thib, to defineindividuals who coupled with theknowledge that carry a very high of previously we expect to beforthcoming, untried psychiatricinterventions for a range that it is fartoo soon to suicideattempters, suggests drairny sortof closureon suicide prevention. 0 , YOUTH SUICIDE 405

human sensitivity, and A considerable amountof energy and goodwill, conventional suicide preventionactivities, which kindness have gone into the evidence that these have focused in a general way onteenagers. There is little be effective; what is needed nowis an educational effort have been or will ever attempts at prevention, to to improve the focusand sophistication of these harness the human energyand motivation into methodsthat will be effective high risk. There is littleresearch to among groupsthat are almost certainly at that go on, which isdifferent from saying thatthere is an abundance of research tells us not to go on. heferences

Crisis Ape ler, R, and Hodes,M. Evaluating hotlines withsimulated calls. Intervention 6:14-21, 1975. Asberg, M.; Nordstrom, P.; andTraskman-Bendz, L.Biological factors in suicide. In: Roy, A., ed. Suicide.Baltimore, Md.: Williams andWilkins, 1986, pp. 47-51. fluid studies Asberg, M.; Nordstrom, P; andTraskman-Bendz, L. Cerebrospinal of Sciences 487:243- in suicide: An overview.Annals of the New York Academy 244, 1987. and Ringsberger, Asberg, M.; Thoren, P.;Traskman-Bendz, L.; Bertilsson, L.; the affective disorder? V. Serotonin depression:A biochemical subgroup within Science 191:478-480, 1976. Bagley, C. The evaluationof a suicide preventionscheme by an ecological method. Social Science andMedicine 2:1-14, 1968. Barraclough, B.M.; Jennings,C.; and Moss, J.R. Suicideprevention by the Samaritans. Lancet, July 30, pp.237-238, 1977. BarrfAclough, B., and Shea, M.Suicide and Samaritan clients.Lancet 2:868- 870, 1970. Bergstrand, C.G., and Otto, U.uicidal attempts in adolescence andchildhood. Acta Paediatrica 51:17-26, 1962. Bleach, G., and Claiborn, W.L.Initial evaluation of hot-linetelephone crisis centers. CommunityMental Health Journal 10(4):387-394,1974. Boyd, J.H., and Moscicki, E.K.Firearms and youth suicide, AmericanJournal of Public Health 76(10):1240-1242,1986. for the Brent, D.A. `The PsychologicalAutopsy: Methodological Considerations of Mental Study of Adolescent Suicide."Paper presented at National Institute Health Workshop on Strategies forSuicide and Suicidal Behavior,Washington, D.C., 1987. 4 1 () 406 PREVENTION OF MENTALDISORDERS

Bridge, T.P.; Potkin, S.G.;Zung, W.W.K; and Soldo,B.J. Suicide prevention canters: Ecological study ofeffectiveness. The Journal ofNervous and Mental Disease 164:18-24, 197'1. Brown, G.L.; Ebert, RH.; Goyner, P.F.; Jimerson, D.C.;Klein, W.J.; Bunney, W.E.; and Goodwin, F.K. Aggression, suicide andserotonin: Relationshipsto CSF amine metabolites.American Journal of Psychiatry 1982. 139(6):741-746, June Centers for Disease Control. Suicide Surveillance.Atlanta, Ga.: U.S. Depart- ment of Health and HumanServices, 1985.

Chamedee, W.A., andYamamoto, M.D. Referralfailures: A one-year follow-up. American Journal ofPsychiatry 130(10):1157-1158,1973. Chowdhury, N.; Hicks, R.C.;and Kreitman, N.Evaluation of an after-care aervice for parakluicide("attempted suicide") patients. 81, 1973. Social Psychiatry 8:67- Dorpat, T.L., and Ripley, H.S. A study of suicidein the Seattlf.area. Com- prehensive Psychiatry1(6):349-359, 1960. Elkins, R.L., Jr., and Cohen,C.R. A comparison of the and job experiences effects of prejob training on nonprofessional telephone crisiscounselors. Suicide and life Threatening Behavior12(2):84-89, 1982. Ettlinger, R. Evaluation of suicide prevr-ition afterattempted suicide. Acta Psychiatrica Scandinavica260 (Supp.):5-135, 1975. Fleiss, J. StatisticalMethods for Risks and 1981. Proportions. New York:Wiley, France, K. Evaluation of lay volunteercrisis telephone workers. Journal of Community American Psychology 3(3):197-220,1975. Garfinkel, B.D.; Froese,A.; and Hood, J.Suicide attempts in adolescents. American children and Journal of Psychiatry139(10):1257-1261, Oct. 1982. Genthner, R Evaluating the functioning ofcommunity based hotlines.Profes- sional Psychology 5:409-414,1974.

Gibbons, J.S. Managementof self-poisoning. In: eds. The Suicidal Farmer, R., and Himch,S., Syndrome. London: CroonHelm, 1980. Gibbons, J.S.; Butler, J.; Urwin P.; and Gibbons, J.L. Evaluation ofa social work service for self-poisoningpatients. British Journal 1978. of Psychiatry 133:111118,

GoldPere, M., and Hawton, K. Reception ofself-poisoning and subsequentdeath in adolescents whotake overdoses. British 1985. Journal of Psychiatry146:395-398, YOUTH SUICIDE 407

aggression and depression:A theoretical frameworkfor Goodwin, F.K. Suicide, ofSuicidal future research. In:Mann, J.J., and Stanley,M., eds. Psychobiology Behavior. New York:New York Arademy ofSciences, 1986. D. The impact ofsuicide in television movies: Gould, M.S., and Shaffer, 315:690-694, 1986. Evidence of imitation. NewEngland Journal of Medicine pathways from childhood Gould, M.S.; Shaffer, D.; andDavies, M. Tritncated death. In: Robins, L.V.,and to adulthood: Attritionin follow-up studies due to Rutter, M.R., eds. Straightand Devious Pathways toAdulthood. Cambridge: Cambridge University Press,in press. 325 parasuicide Greer, S., and Anderson,M.Samaritan contact among patients. British Journalof Psychiatry 135:263-268,1979. in attempted suicide: Greer, S., and Bagley, C.Effect of psychiatric intervention A controlled study.British Medical Journal1:310-312, 1971. lethal attempted Greer, S., and Lee, H.A.Subsequent progress of potentially suicides. Acta PsychiatricaScandinavica 43:361-371, 1967. British Medical Hassell, C., and Trethowan,W.H. Suicide in Birmingham. Journal 1:717-718, 1972. M.Hospital admissions foradverse effects of Hawton, K., and Goldacre, in the Oxford medicinal agents (mainlyself-poisoning) among adolescents region. British Journalof Psychiatry 141:106-170,1982. Hirsch, S.A.Critique of volunteer-staffedsuicide prevention centres. Canadian Journal of Psychiatry26:406-410, 1981. British Journal of Holding, T.A The B.B.C."Befrienders" series and its effect. Psychiatry 124:470-472, 1974. Holding, T.A Suicide and"The Befrienders." Briti..Medical Journal 3:751- 753, 1975. in a sample of a Holding, T.A., and Barraclough,B.M. Psychiatric morbidity 127:133-143, London coroner's openverdicts.British Journal of Psychiatry 1975. Jacobson, G.F., et al. The scopeand practice of an early-accessbrief treatment psychiatric center. AmericanJournal of Psychiatry 121:11764182,1965. Jan-Tausch, J . S uicide ofChildren 1960-63: New JerseyPublic School Stu- dents. Trenton, N.J.: Stateof New Jersey Departmentof Education, 1964. Jennings, C.; Barraclough,B.M.; and Moss, J.R. Havethe Samaritans lowered the suicide rate? A controlledstudy. Psychological Medicine413-422, 1978.

412 408 PREVENTION OF MENTALDISORDERS

Johns, M.W. Self-poisoning with barbiturates inEngland and Walesduring 1959-1974. British MedicalJournal 2:1128-1130, 1977. Kalafat, J.; Boroto, D.R.;and France, K. and value orientation Relationships among experiencelevel and the performanceof paraprofessional selors. American Journal telephone coun- of Community Psychology5:167-179, 1979. Kennedy, P. Efficacyof a regional poisoning further suicidal behavior. treatment center in preventing British Medical Journal4:255-257, 1972. Kennedy, P.; Kreitman,N.; and Ovenstone, and parasuicide LM.K. The prevalenceof suicide "attempted suicide" inEdinburgh. Psychiatry 124:36-41,1974. British Journal of King, G.D. An evaluationof the effectiveness of a telephone counselingcenter. American Journal ofCommunity Psychology5:75-83, 1977. Knesper, D.J. A studyof referral failures for method of medical potentially suicidalpatients: A care evaluation.Hospital and Community 33(1):49-52, 1982. Psychiatry Kogan, L.S. The short-term case ina family agency. U. Resultsof a study. Social Casework38:296-302, 1957. Knowles, D. On the tendency for volunteerhelpers to give advice.Journal of Counseling Psychology26:352-354, 1979. Kreitman, N. Age andparasuicide "attempted 6:113-121, 1976. suicide" PsychologicalMedicine Kreitman, N. Suicidein Scotland in comparison with Englandand Wales. British Journal ofPsychiatry 120:83-87,1972. Kreitman, N. The coalgas story.United Kingdom British Journal of suicide rates, 1960-71. Preventive and SocialMedicine 3u.86-93, 1976. Lester, D. Steps toward the evaluation ofa suicide prevention center: Crisis Intervention2(2):42-45, 1970. Part one.

Lester, D. Preventionof suicide. Journalof the American 225:992, 1973. Medical Association

Lieberman, R, andEckman, T. Behavior therapy vs. insightoriented therapy for repeated suicideattempters. Archives of 1981. General Psychiatry38:1126-1130, Linnoila, M., et al. Low cerebrospinalfluid 5-hydroxyindoleacetic centration differentiates acid con- impulsive fromnonimpulsive violent Science 33:2609-2614,1983. behavior. Life ! 3 YOUTH SUICIDE 409

Litman, RE. Sigmund Freud on suicide. Essaysin Self-Destruction. New York: Science House, 1965. Litman, R, and Faberow, N.L. "Evaluatingthe Effectiveness of Suicide Preven- tion." Proceedings of the Fifth Internationtil Conferencefor Suicide Prevention, London, 1969, pp. 246-250. Litman, RE.; Faberow, N.L.; Shneidman, E.13.;Helig, S.M.; and Kramer, J.A. Suicide prevention telephone service. Journal of theAmerican Medical Associa- tion 192:107-111, 1965. Litt, I.F.; Cuskey, W.R; and Rudd, S.Emergency room evaluation of the adolescent who attempts suicide:Compliance with follow-up.Journal of Adolescent Health Care 4:106-108, 1983. McCarthy, B.W., and Berman, AL. A student operatedcrisis center. Personnel and Guidance 49(6):523-528, 1979. Miller, H.L.; Coombs, D.W.; Leeper, J.D.; and Barton,S.N. An analysis of the effects of suicide prevention facilities onsuicide rates in the United States, American Journal of Public Health 74:340-343, 1984. Morgan, H.G.; Burns-Cox, C.J.; Pocock, H.J.; andPottle, S. Deliberate self- harm: Clinical and socioeconomic characteristics of 368patients. British Jour- nal of Pay cV ary 127:564-574, 1975. Motto, J.A.Suicide prevention for high-risk persons who refuse treaiment. Suicide and Life Threatening Behavior 6:223-230, 1976. Motto, J.A. Suicide in male adolescents. In: Segdak, H.S.;Ford, A.B.; and Rushforth, N.B., eds. Suicide in the Young. Boston:John Wright PSG Inc., 1984. Motto, J.A; Heilbron, D.C.; Juster, R.P.; and Bostrom,AG. Communication as a suicide prevention program.Depression et Suicide 1981, 148-154. National Center for Health Statistics.Department of Health and Human Services, Mortality Statistics Branch (published and unpublisheddata), 1983. Neilson, J., and Videbech, T. Suicide frequency before andafter introduction of community psychiatry in a Danish island.British Journal of Psychiatry 123:395-399, 1973. Neuringer, G. Methodological problems in suicide research.Journal of Con- sulting Psychology 26:273-276, 1962. O'Brien, J.P.Increase in suicide attempts by drug ingestion: The Boston experience. Archives of General Psychiatry 34:1165-1169, 1977. Otto, 0. Suicidal acts by children and adolescents: A follow-up study.Acta Psychiatrica Scandinavica 233, 1972. ei 1 d 410 PREVENTION OF MENTAL DISORDERS

Paykel, E.S.; Hallowell, C.; Dressler,D.M.; Shapiro, D.L.; and Weissman,M.M. Treatthent of suicide attempters: Adescriptive study. Archives of General Psychiatry 31:487-491, 1974a.

Paykel, E.S.; Meyers, J.K.; Lindenthal,J.J.; and Tanner, J. Suicidal feelingsin general population: A prevalencestudy. British Journal of Psyvhiatry124:460- 469, 1974b.

Pferer, C.R.; Zuckerman, S.; Plutchik,; and Mizrucht, M.S. Suicidal behavior in normal school children: Acomparison with child psychiatricinpatients. Journal of the American Academyof Child and Adolescent Psychiatry21:564- 569, 1984.

Phillips, D. Teenage and adulttemporal fluctuations in suicideand auto fatalities. In: Sudak, H.S.; Ford,A.B.; Ibishforth, N.B., eds. Suicidein the Young. Boston: John Wright PSG,Inc., 1984. Ringel, E. Suicide preventionin Vienna. In: Resnick, H.L.P.,ed. Suicidal Behaviors: Diagnosis and Management.Boston: Little, Brown, 1968. Robins, E.; Gassner, S.; Keyes,J.; Wilkinson, RH., Jr.; andMurphy, G.E. The communication of suicide intent: A studyof 134 consecutivecases of successful (completed) suicide. American Journalof Psychiatry 115:724-733, 1959. Rogawski, A.B., and Edmundson,B. Factors affecting theoutcome of psychiatry interagency referral. American Journalof Psychiatry 127:925-934,1971. Rosen, A. Detection of suicidalpatients: An example ofsome limitations in the prediction of infrequent events.Journal of Consulting Psychology 1954. 18:397-403,

Ross, C.P. Mobilizing schools forsuicide prevention. Suicide andlife Threaten- ing Behavior 10:239-243,1980. Roy, A., et al. Reduced cerebrospinal fluid concentrations of homovanillicacid and homovanillic acid to5-hydroxyindoleacetic acid ratiosin depreeRed patients: Relationship to suicidality anddexamethasone nonsuppression.American Journal of Psychiatry, 1987. Rutter, M., and Giller, H. Juvenile Delinquency Trends andPerspectives. New York: Guilford Press, 1984.

Salk, L., et al. Relationshipof maternal and perinatalconditions to eventual adolescent suicide. Lancet 1:624-627,1985.

Sanborn, D.E.; Sanborn, C.J.;and Cimbolic, P. Twoyears of suicide: A study of adolescent suicide in NewHampshire. Child PsychiatryHuman Development 3:234-242, 1974. YOUTH SUICIDE 411

Sathyavathi, K Suicide among children inBangalore. Indian Journal of Pediatrica 42:149-157, 1975. Schneidman, E.S., and Farberow, N.L. Clues toSuicide. New York: Blakison, 1957. Shaffer, D.Suicide in childhood and early adolescence.Journal of Child Psychology and Psychiatry 15:275-291, 1974. Shaffer, D. Teen suicide research goals.In: Farberow, N.; Atman, S.; and Thorne, A., eds. Report of the NationalConfeence on Youth Suicide, June 19-20. Washington, D.C.: Office of Human DevelopmentServices, 1985, pp. 101-112. Shaffer, D., and Caton, C.L.M. "Runaway andHomeless Youth in New York City." A Report to the Ittelson Foundation, NewYork City, Jan. 1984. Shaffer, D., and Fisher, P. The epidemiology ofsuicide in children and young adolescents. Journal of the American Academy of ChildPsychiatry 20:545-565, 1981. Shaffer, D., and Garland, A. "An Evaluation of NewJersey Youth Suicide Prevention Programs." Draft report of preliminaryfindings. June 29, 1987. Shaffer, D.; Garland, A.; Gould, M.; Fisher, P.; andTrautman, P. Preventing teenage suicide: A critical review. Journal ofthe American Academy of Child and Adolescent Psychiatry 27(6):675-687, 1988. Shafii, M.; Carrigan, S.; Whittinghill, J.R.; andDerrick, A.Psychological autopsy of completed suicide in children and adolescents.American Journal of Psychiatry 142:1061-1064, 1985. Shore, J.H. American Indian suicideFact and fantasy.Psychiatry 38:86-91, 1975. Slaiku, KA; Tulkin, S.R; and Speer, D.C. Process and outcomein the evaluation of telephone counseling referrals.Journal of Consulting and Clinical Psychology 43:700-707, 1975. Slem, C.M., and Cotler, S. Crisis phone services: Evaluationof hotlir.a program. American Journal ofCommunity Psychology 1:219-227, 1973. Stanley, M.Cholinergic binding in the frontal cortex of suicide victims. American Journal of Psychiatry 141:11, 1984. Stanley, M., and Mann, J .J. Biological Factors Associated withSuicide. New York, 1987. Stanley, M., et al. Correlations between aminergic metabolitessimultareously obtained from human CSF and brain. Life Sciences 37:1279-1286,1985. 4 6 412 PREVENTION OFMENTAL DISORDERS

Stengel, E., and Cook, N.G.Attempted Suicide. Press, 1958. London: OxfordUniversity Sudak, H.S.; Sawyer, J.B.;Spring, G.K; and sumer; rates ina crisis center. Hospital Coakwell, C.M.High referral 1977. and CommunityPsychiatry 28:530-532,

Temoche, A.; Pugh,T.F.; and McMahon, former mental B. Suiciderates amongcurrent and institution patients.Journal of Nervous 138:124-130, 1964. and MentalDisorders Traskman-Bendz, L,et al. CSF monoamine metabolitesof depreesed during illness andafter recovery.Acta Psohiatrica patients 1984. Scandinavica69:333-342, Trautman, P.D., and Shaffer, D.Treatment of child attempters. In:Sudak, H.S.; Ford, and adolescentsuicide Young. Boston: A.B.; and Rushforth,N.B. Suicide in John Wright PSG,Inc., 1984. the Van Praag, H.M. Significanceof biochemical treatment, andprevention of parameters in thediagnosis, 12:101.131, 1977. depressive disorders.Biological Psychiatry Walk, D. Suicide and communitycare. British Journal 1391, 1967. of Psychiatry113:1381-

Werner, LW.The effectivenessof a suicide Hygiene 53:357-363,1969. preventionprogram. Mental Weissman, M.M. The epidemiologyof suicide Psychiatry 30:737-746,1974. attempts. Archivesof General Welu, T.C. Afollow-upprogram for suicide ness. Suicide and I.,ife attempters: Evaluationof effective- Threatening Behavior7:17-30, 1977. World Health Organization. "Suicideand Attempted Papers. Geneva:World Health Suicide." PublicHealth Organization, 1974,p. 58.

4 r 7 CHAPTER 11

Prevention of LearningDisorders

Archie A Silver, M.D. Prvfessor and Director Division of Child and AdolescentPsychiatry College of Medicine University of South Florida

Rosa A. Hagin, Ph.D. Professor Graduate School of Education Fordham University Lincoln Center, New York Introduction The prevention of learning disordersin childhood is confounded by thefact that learning failure is but a symptom, onemanifestation, stemming from a multitude of factors acting singly or, moreoften, in synergistic combination. As in any complex behavior,learning involves the interaction of biological, psychological, social, and educationalfactors. Biological factors include genetic variants, developmental lags, andimposed pathological events; psychological factors include cognition, attention,motivation, and emotional influences; socioecunomic-environmental factors involve culturaland economic status, the quality ai. I quantity of stimulation atearly ages, and the vicissitudes of life experiences throughout time; educationalfactors deal with the appropriateness and adequacy of educational exposure.For any child, these factors havevarying degrees of influence on learning, but inchildren with learning disorders all these influences must be considered because each mayplay a synergistic and reciprocal role in placing the child at risk for alearning disorder. It is reasonable to assume that thefirst step in the primary prevention of learning disorders is to understand thespecific contribution to learning of each of the risk factorsbiological,psychological, social, and educational- then to remove the causativefactor or avoid exposure to it.If, for example, certain specific fac.,rs, unique to cultural differenceand economic disadvantage, are actual causa, agents for learning disorders,then broad social planning and educatioaal change to modify these factors maybe needed. Such conclusions

This paper was prepared for the ProjectPrevention initiative of the American Academy of Child and Adolescent PsychitAry. 4 1 8 414 PREVENTION OF MENTALDISORDERS

may be drawn from work such as that of Brody and Axelrad (1970), Hessand Shipman (1968), and Caldwell andBradley (1978). If perinatalevents, such as neonatal hypoxia,are identified as potent contributorsto learning disorders, then primary prevention involvesthe isolation of theseevents and their modification by new obstetrical and perinatal techniques.(The importance of perinatal hypoxia in cognitivedevelopment and later learninghas been documented [Broman et aL 1975;Broman 1979; Mednick and Baert1981).) On the other hand, insome children with learning failure the cannot yet be clearly identified. causative agent These children haveno discernible history of traumatic physical events of theprenatal, perinatal,or postnatal periods. They need not have suffered socialor economic disadvantage; their centrai system is not marked by nervous identifiable focal or even diffusestructural defect; their emotional state and educationalexperiences are not significantly those of their peers who do different from learn; their intelligence is withinthe normal range; and they do not sufferfrom defects in visualor auditory acuity. Yet these children do have academicdifficulty. We classify thesechildren as suffering from specific developmentallearning disabilities. We these children enter postulate that before kindergarten the anlage for disabilityis already present in their centralnervous system and that theremay be a genetic predisposition to that disability. The list of identified risk factorsstemming fromsocial-cultural-economic disadvantage, from pathologicalevents of pregnancy, and possible genetic influence particularly from and from the vicissitudesof l ife's experiencenotonly is long, but also does notyet lend itself to primary able to eradicate the risk prevention. We are not yet factorsthat is, to eliminatepoverty and disadvantage, prematurity and hypoxia, genetic defects anddevelopmental lags, and inadequate or inappropriatestimulation. In the prevention of learningdisorders, we are thereforeleft with a fallback positionto secondarypreventionof identifying children have learning disorders who we predict will in the future and providingappropriate intervention before learning failure hasoccurred. Currently, there isa large group of children for whomsuch secondary prevention of learning disordersis available. Theseare children in whom the various primary risk factors, either biologicalor experiential, leavea mark on the function of thecentral nervous systemso that the psychoneurological processes involved in learningare impaired.For example, hypoxic birth, the inappropriate a traumatic or and/or inadequate stimulationof poverty or cultural deprivation,or a genetic predisposition to guage in all its ramifications, ah developmental lags in lan- limy be reflected in psychoneurologicaldys- inction, which makes learningto read, write, spell,or do arithmetic especially dgficu It. It is now possible to identify these peychoneurologicaldefects by the time the child is 5 years old, topredict with confidencewhen the child enters whether that child will fail kindergarten in learning, and to offera program of interventionso LEARNING DISORDERS 415 that learning failure does not occur.It is clear that children identified as vulnerable to learning failure are aheterogeneous group, composed of children who may have been exposed todifferent risk factors and their combinations. Thus, the identification process is not adiagnosis but a first stage to secondaty prevention. We identify these childrenwith a constellation of neuropsychological dysfunctions in "one or more of the basic processesin understanding or using written or spoken language," as havinglearning disabilities. There is a rich literature on the identificationof factors within the spectrum of social-cultural-economic disadvantagethat predispose children to learning failure (Caldwell and Bradley 1984;Tableman and Katzenmeyer 1985; Werner and Smith 1977, 1982; Murphy andMoriarty 1976; Sameroff and Chandler 1975; Hess and Shipman 1968; Deutsch 1973),and also on the effects of prenatal, perinatal, and neonatal events on learning(Mednick and Baert 1981; Siegel 1983). The review of that literature,however, is beyond the scope of this chapter. This paper will discuss methods for theidentification of 5-year-olds vulnerable to learning disability and willdescribe programs of successful intervention. Identification in Kindergarten

Scanning is not diagnosis.* Scanning can beexpected only to identify a heterogeneous group whose members have one thingin common: their fu nction in those parameters assessed by the scanninginstrument is immature when compared to their peers. Statistical Considerations The level of inaccurate predictions may beadjusted by statistical manipula- tion of the scanning battery. False positives maybe decreased by lowering the cutoff score, the score at or below which failure ispredicted and above which adequate learning is suggested.Lowering the cutoff score will put fewer childnen in the vulnerable category. At the same time,however, decreasing the number of false positives will increase the numberof false negatives. Con- versely, raising the cutoff score wil/ increase the number offalse positives and decrease the number of false negatives.

*In this paper, we will use the word "scanning"rather than "screening." Screening impliee dichotomizing; the sample falling through the sereenis the "abnormal," the "sick"; those not falling through are"normal." Scanning does not imply sickness or wellness. It does implyexamination of an entire sample (i.e., all children in kindergarten), viewingthe distribution of function on a continuum. It does not dichotomize, categorize, or label.It conveys what we are trying to do, looking over(scanning) entire school classes, finding children who are quantitatively, not qualitatively, differentfrom their peers. 420 416 PREVENTION OF MENTAL DISORDERS

Content Considerations

Unfortunately, there is currentlyno generally accepted theory of thenature of the learningprocesses. The functional precursors to ing, and arithmetic reading, spelling, writ- are not clear, and thus there isno general agreement on what functionsa predictive instrument should evaluate.Also, the skills quired of a beginning reathr re- may differ from those of a more advancedreader. It has been said (Farr 1969)that tests, which practicaljudgment indicatesare related to early academiclearning, are most predictiveof that learning. Our task in prediction, however, is to find the skills thatare needed before the child is exposed to academiclearning. Surveys of the literatureon predictive variables suchas age, sex, laterality, visual perception, auditory perception, and intersensoryintegration conclude that, although each singlevariable contributes to learning skill is central to reading (i.e., reading), noone and therefore predictive(DeHirsch 1971; Jansky and DeHirsch 1972). In contrast to sensory modality studies, Blank andBridger (1967) found symbolic mediation to bea necessary condition to solve problems temporally presented stimuli involving regardless of modality. Thedifficulty, theysay, is in applying relevantverbal labels to stimulieven in the same modality. Vellutino (1978) stronglysupported this argument. We should not be surprisedthat the single modality able to predict learning research has not been disability with any useful degreeof accuracy. Consider- ing reading disability itself,we know that within thisgroup of children, marked individual differencee indistribution of assets anddeficits may appear in all aspects of perceptual, associative,and emissive language all aspects of dysgnosias, function as well as in dyspraxias, and dysphasias(Silver and Hagin 1960, 1972). Recent multivariable studies (Rourke and Strang1983; Rourke 1985) of neuropsychological patterns inchildren with readingdisability substantiate this variability. Theimportance of these variationsis that in any correlational study of large numbers of children, individual differencesin distribution of function may well canceleach other out, and theinvestigator will dismiss insignificant some variables that as may have crucial importance forsome children in the sample. A theoreticalformulation is needed that most of the variations would encompass allor seen in learning disabilities. Thecontent of predictive instruments thus shouldnot be a collection of myth and larceny, but extraneous items put together by should spring fromconceptual unity of thelearning process. A number of such conceptualpositions have been advanced. Satz and his associates of theFlorida Longitudinal 1970, pp. 17-39; Satz Project (Satz and Sparrow et al. 1971; Satz and VanNostrand 1973) postulate that . . reading disabilities reflecta lag in the maturation of thebrain which differentially delays thoseskills whichare in primary ascendancy at different chronologicalages. ... Skills which develop... earlier during childhood (e.g., visual perceptual and cross modalsensory integration) are 1 LEARNING DISORDERS 417

more likely to bedelayed in younger children ...maturationally immature. ... Skills whichhave a slower rate of developmentduring childhood (language and formal operations) are morelikely to be delayed in older children developmentally immature(Satz et al. 1978, p. 319). Thus, this theory predicts thatchildren who are delayed in visualperception and cross-modality skills at ages5 and 6 will eventually fail inreading. The theory further predicts thatthese children will eventu ally catch up onthe earlier developing skills but will subsequentlylag in conceptual linguistic skills. Some evidence for the predictive valueof sensory-perceptual factors maybe seen in analysis ofthe early standardization batteryof 14 variables used in testing 497 white malekindergarten pupils in the publicschool system in Alachua County, Florida (Satz andFriel 1973), in which factor analysisyielded three major factors:(1) a sensory-perceptual motor-mnemonicfactor, (2) a verbal-conceptual factor, and (3) averbal-cultural factor.In a later paper (1978), Satz et al described fourfactors:(1) a sensory-perceptual-motor mnemonic factor, (2) teacherevaluations, (3) a conceptual-verbal factor,and (4) a motor factor.As the project developed, however,and additional samples were studied, the number of variables inthis battery decreased, and the current Florida Kindergarten ScreeningBattery consists of five subtests:(1) the Peabody Picture Vocabulary Test, (2) avisual recognition discrimination test(a visual matching to samples of geometricdesign) devised by Small (1969), (3) the Beery test of visual motor integration,(4) alphabet recitation, and (5) finger localization. This battery introduceslanguage tests that, although of greater predictive value than the originalstandardization battery, dilute the theoretical concept of a lag in sensory-perceptualfunctions. Further, the concept of developmental lag does not ensure that childrenwill outgrow that lag. Followup studies (Silver and Hagin 1964) have demonstratedthe persistence into young adulthood of the perceptual defects originallyfound in 10- to 12-year-old children with learningdisabilities. It appears that perceptualdysfunction does not necessarily mature spontaneously, andthat reading disability must be conridered along- term problem. The concept of maturational lag wasadvanced by Gesell and Thompson (1934). Gesell and Thompsonasked, "Does the infant present specificlags and accelerations among components ofhis behavior equipment?" Bender and Yarnell (1941) were the first to apply thi.: term"developmental lag" or "maturational lag" to children with readingdisabilities in whom classical neurological examination was normal. Bender andSilver emphasized as early as 1948 that many of the conditions seenin children, then considered "organic" or"brain damage," were better understood asmaturational delays rather than as structural defects of the central nervous system. The problem of what functions are delayedin children who will fail in reading is not solved.Studies of subgroups of neuropsychologicaldysfunction in children with learning disorders have sufferedfrom the perennial problem of disparate samples, definitions, boundaries,and methodology.Multivariate 422 418 PREVENTION OF MENTALDISORDERS

studics have proliferated;most found variationson the themes of language defect and graphomotor and visual-perceptual deficits, withvaried emphasison the specific deficits present.Doehring(1968)found that the concept basic subgroups of reading disabilities to is impairment in sequentialprocessing abilities; Petrauskas and Rourke(1979)found one group with most finger-gnosis and immediate difficulty in memory for visual sequences. Satzand Morris (1981) found onegroup with no impairment in their Is there a concept that neuropsychological tests. can unify the many and variedneuropsychological immaturities found in children withspecific learning disability? Position in space and orderin time are fundamentalfunctiona of the central nervous system. Schilder (1942) emphasizedthese concepts when he said, deal with the fundamental "We fact that human existenceexpresses itself in space and in time." In thedevelopment of perception, both change) and form in movement (i.e., temporal space are essential, and "there isa constant interplay or integration between motor andsensory features which cannevel be separated" (Bender1958).Lashley (1951) emphasizedthis interrelationship: temporal order "Spatial and appear to be almost interchangeablein cerebral action." The fundamental nature of space and time was further stated byLashley (p. 114): "Temporal integrationsare especially characteristic of human contribute as much behavior and as any single factor to man'sintelligence." Rozin (1976, p. 15) emphasized the importance of spatial and temporalcontexts in human memory: "Events occur in particularspatio-temporal contexts, andthey are stored in memory withrespect to these contexts,so that the past order, reality and coherence." Further,temporal sequencingstructures a series of hierarchies of organization; language involves the ordering ofletters and phonemes ina word, words ina sentence, sentences ina paragraph, and finally, the of logical thought and sequences memory. The body image inspace defines the space coordinates of the child'sperceptions. As they reviewed the variedclusters of neuropsychological with reading disabilities, deficits in children Silver and Hagin(1972)were impressed that these deficits could be consideredproblems with orientation in time, that spatial and in space and organization temporal organizationwere not age-appropriate in children with learningdisabilities. Theee problems tions of visual, visual-motor, may be seen in all combina- auditory, and body imageperception. Thus, there is unity in the diversedeficits in the functionsof the learning-disabled That unity is disorientation child. in space and disorganizationin time. Silver and Hagin suggeeted that, ifthese deficits could bedetected in the 5-year-old child enters kindergarten, as the intervention might take placein the kindergarten and early elementary gradee. In the visual discriminationand recall ofasym- metric figures, spatialorientation could be tested; and in auditory in auditory rotesequencing memory, temporal organizationmay be seen; in right-left discrimination, praxis, andfinger-gnosis,awareness of the body image in may be determined; visual-motor space function may combinevisual perceptual skill and body imageawareness (praxis). 419 LEARNING DISORDERS

intensive examination ofall These perceptualtasks were included in an grade of a school in theKips Bay area of NewYork City children in the first 1972). This examination was an during 1969-70 and1970-71 (Silver and Hagin study involvingneurological, psychiatric,psychological, intensive individual These examinations social, and educationalexaminations of 171 children. with only 1 percent falsepositives, the children who yielded data that detected, grade. would fail in reading 2 yearslater, at the end of second Jansky and DeHirsch(1972, A different conceptualapproach was used by Predictive Screening Index: p. 45) inthe development of their usefulness in clinical The choice of tests. wasbased upon their prognostic ..past research. The practice, as well as predictiveefficacy established in . development were considered:perceptuo-motor or- followbg aspects of receptive and expressive ganization, linguisticcompetence in both its symbols. The heavy emphasis aspects and readinessto cope with printed tests derived from theauthor's conviction thatthe ability to on linguistic importance in comprehend and use orallanguage was of overwhelming loarning to read. of remaining scanningtests used to predictreading The theoretical position probes that most failure is essentially apragmatic onethe use of or learning judgment says are related tolearning. closely resemble theskills that practical testa are frequently used.The Meeting SchoolScreening Two of these predictive subtests, five in each ofthree Tests (Hainsworth andSinquelend 1969) use 15 visual-perceptual-motor, and bodyimage and motor control. clusters: language, how the child takes in, processes, Each cluster is composedof tests to determine cluster includes serialcounting, and responds toinformation. The language and auditory discrimination),articulation, sentence phrases (nonsense syllables cluster includes repetition, and verbal reasoning.The visual-perceptual-motor visual recognition ofabstract forms, directions copying geometric designs, the body image- involving spatial orientation,draw-a-person, and writing name; includes hopping, clapping,directions, and control of hand motor control cluster develop this predic- movements. The Hainsworthshave worked for 15 years to tive instrument, with norms on2,500 children 2-1/2 to6-1/2 years old. A developmental age equivalentis available for eachcluster. 1-6 Screening Test forIdentification of Children with The Slingerland Grade 1985) Learning Disability is alsoin common use. Itsreview in Buros (Mitchell academic states, "It is a test ofvarious auditory, visual andmotor skills related to grade and early second grade; areas." Three forms areavailable: Form A for first early third grade; C forthird grade and early fourthgrade. B for second grade to copying printed Each form consists ofeight subtests:tests 1 and 2 involve material; test 3, recall andmatching of printed words,letters, and numbers, responding; test 4, immediate presented in brief exposurewith a delay before 5, delayed copying ofwords, phrases, letters, and matching of printed words; test dictation; tests 7 and 8, designs; test 6, writing groupsof letters and numbers to 424 420 PREVENTION OF MENTALDISORDERS

skill with initial phonics and delayedmatching of spoken C are not reallyearly predictive words. Forms Band instruments, but readinessbatteries. It may beseen, therefore, that the theory underlyingvarious predictive instruments is varied; inmost, the tests sample reading. If one examines skills presumedto resemble the name of eachsubtest, there redundancy and overlap appears to be much among the skills measuredby the tests. On look, hcwever, theactual manner in which a closer ly different the subtest is givenmaybe complete- among scanning testsso that, even though the the test may be the identifyingname of same, the functions measuredmay be varied. The scanning teststo date have largely failure in reading, been designed topredict successor rather than predictlearning failure in scanning batteries general. Three major for prediction (4' readingfailure are the Florida Screening Battery (Satz Kindergarten and Fletcher 1982),SEARCH (Silver and 1981), and thePredictive Screening Hagin 1976, meet, in greater Index (Jansky andDeHirsch 1972). All or lesser degree, thestatistical guidelines Psychological Association. of the American In spite of problems with false positivesand false negatives, criteriinmeasures used for validation problems with of predictions, anda general lack of consensus on the content ofa predictive battery, ments are available. well-studied scanninginstru- At this stage ofour knowledge, it is certainly predict in kindergartenwhich children will possible to however, must be later fail in reading.Prediction, followed by appropriateintervention. This will in the next sectionof this report. be considered

Outcomes ofEducationalInterventions Commitment to primary prevention of learningdisorders is reflected often in exhortationson its value than in more efficacy. This point data-based demonstrationsof its can be illustrated in theresults of computer intcjor data bases,ERIC, and Psychinfo. searches of two These searchesproduced 69 references dealing with earlyidentification and these refereAces prevention of learningdisorders. That deal more oftenwith preventionin the abstract data-based researchon specific intervention than with strategies can beseen in table 1. Most conventionaltextbooks provide little informationabout preventive interventions. Tworecent volumeson the psychoeducational young children (Paget andBracken 1983; asseosment of mention preventive Lichtenstein andIreton 1984) scarcely interventions; Achenbach's(1982) section his volumeon developmental on prevention in psychopathology isan admirable exception practice. Thecontent analysis to this reported in table 1shows that general sions of the needsand values of discus- prevention area common type ofpaper. However, any significantapplication of prevention disorders requires to the problem oflearning more than good intentions.Effective prevention judicious selectionof goals and specific requires formulation of principlesfor targeting, 425 LEARNING DISORDERS 421

Table 1. Content Analysis of Pmvention Literature Search

Content Number of References

Discussions of the value of prevention Generally favorable 11 Generally unfavorable 3

Methods for early identification Recommending 16 Opposing 3

Specific intervention projects No evaluation data provided 17 Descriptions of favorable outcomes 5 Data on outcomes provided Favorable outcome 7 Instructional objectives not met 1

References not relevant to prevention 6

timing, and guiding intervention strategies, It is with these specifics of preventive intervention that the data base is thin. Of the 38 specific projects reported, 17 provide extensive descriptions of methods for selection of participants and procedures but offer little data on outcomes of these interventions. Five papers provide descriptive and anecdotal data on outcomes; only eight provide evaluation data to document program effects. Ultimately, discussions of the fine points of rationale for an approach and even sophisticated statistical computations of "hit rate" of a screening instrument are only academic unless they are part of a total program that achieves measurable results in interventions with vulnerable children. Interventions for the prevention of learning disorders will be considered here from this pragmatic standpoint. The review has been organized in terms of (1) general consideration of facilitators and barriers to preventive intervention; (2) experimental programs in early education; (3) administrative arrangements to prevent learning disorders; (4) specific interventions to prevent learning disabilities; and (5) needs highlighted by this review.

426 422 PREVENTION OF MENTAL DISORDERS

Facilitators and Barriers to Preventive Interventions A comprehensive description of the need for preventionof learning disabilities through efforts directed at the preschoolyears was voiced in a position paper offered by the National Joint Committee on Learning Disabilities (NJCLD) (1085). The multidisciplinary membership ofthis committee represents some of the major organizations in the field, includingAssociation for Children and Adults with Learning Disabilities, American Speechand Hearing Association, Council for Learning Disabilities, divisionsof the Council for Exceptional Children concerned with communicationdisorders and learning disabilities, International Reading Association, NationalAssociation of School Psychologists, and Orton-Dyslexia Society. Thisreport emphasized that the preschool years are a critical period. during which essentialpteventive and intervention efforts are most effective. Recognizing that learninglisorders are a heterogeneous group of disorders of presumed neurological origin, thereport cautions that indiscriminate labeling is notwe! ranted, because normal development is characterized by broad ranges of individual andgroup differencee, as well as by variability in rates and patterns ofmaturation. Early identification procedures were defined to include examination of biological,genetic, perinatal, and adventitious risk indicators through systematicobservation, screening, and other (unspecified) procedures. The NJCLDpaper noted, however, that early identification was only one step and that the"identification programs not followed by assessment, intervention, and followup are futile." The NJCLD paper advised caution in the design of suchprograms by observing that "traditional readiness activitiesare often not sufficient to ensure later school success." Periodic evaluation of effectivenessis recommended so that interventions not only focus on ameliorating deficits thataffect current functioning, but also on developing abilities, skills, and knowledgenecessary for later linguistic, academic, and socialfunctioning. Eight areas of needwere recommended: 1. Systematic identificationprograms; 2. Assessment based on interdisciplinary approachesthat provide an integrated statement of current status and needs; 3. Validated models for early intervention availableto all preschool children with identified developmental deficits; 4.Qualified personnelnecessary to meet the needs of the identified children; 5. Assistance to families in participating fullyin all phases of identification and treatment; 6. Public information concerning issuesof child development and its disorders; 7. Response to the unique needs ofnon-English-speaking and limited- English -speaking families; 4 f"? 423 LEARNING DISORDERS

related to provision ofservices for 8. Systematicresearch to address issues preschool children withsuspected learning disorders. clear and enlightenedposition. If one agrees The NJCLD paper presents a desirable, there is that prevention oflearning disorders is atall possible and that is controversial. Itis significant that thenature little in the NJCLD report completely through thecall for of preventive interventionsis sidestepped Whether this joint statementsought to avoid supporting "validated models." of research in model or to disregardthe results of at least 20 years any existing wastefulness of the second early identification andprevention is not clear. The However, because ofvariations in viewpoints amongthe alternative is obvious. have been chosen in organizations involved,the neutral first alternative may for prevention fromNJCLD order to produce a unifiedstatement of support members. unattainable ideal, Although the NJCLDrecommendations seem to offer an All HandicappedChildren Act it should be rememberedthat the Education for 94-142, passed 10 yearsearlier, mandated a "free ap- of 1975, Public Law children between the agesof 3 propriate public education"for ah handicapped the ...inique educationalneeds and 21. Very specificprovisions de6igned to meet and their familiesobligate school districts toidentify of handicapped children served; to inform parentsfully all handicapped childrenwithin the age groups the law; to pruvidecomprehensive assessments atleast of their rights under individual educational plan(IEP) once every 3 years;and to develop annually an the written by a committeethat includes a teacher, aschool administrator, specialized personnel asneeded, to provide services as child's parent, and other carried out in the "least delineated in the IEP and to seethat the services were ensured contact with theeducational restrictive environment"(i.e., settings that mainstream to the extentappropriate bor a givenchild). needs The law requires the IEPto include an analysisof the child's current goals stating the levelsto be achieved duringthe and capabilities; annual accomplish the annual goals;evalua- current year andshort-term objectives to techniques and Vieirlocation; and related services tion criteria; instructional transport&tion, counseling,and (i.e., noninstructionalservices, such as special physical therapy, necessaryfor the child to profitfrom the IEP). 94-142 could be interpreted Although the modelrepresented in Public Law of learning disorders, itsimplemen- to include provisionsfor primary prevention in provision of servicesonly after children have tation has generally resulted result occurred because of experienced a severe degreeof learning failure. This Although Public Law economic considerations andtheoretical considerations. for early identificationand most of the Stateplans 94-142 recognized the need financial support for local implementationprovided funds for early screening, of the need for remedia- for intervention continuesto be based on documentation tion alter failure hasoccurred. deliberate. The This choice of remediation overprevention probably was not what existed in gap betweenthe services mandatedin Public Law 94-142 and 424 PREVENTION OF MENTALDISORDERS

most school districts was wide. Personnel neededto be recruited and trained. The sheer numbers ofchildren identified the services required as learning disabled and thecost of had not been anticipated.The emphasison remediation came because the need for services for learning-disordered childrenhad not been met for so long that muchhad to be done to provide with the law. even minimal compliance Public Law 94-142 was not drafted with theassumption that the Federal Government wouleassume all costs of educating gress considered that it handicapped children. Con- was a joint responsibility oflocal, State, and Federal governments to share thesecosts. Unfortunately, the tion of an enlightened grass-roots implementa- law resulted ina special education model thatrewards programs that serve children afterthey have failed. A second reason for the loss of the opportunityto support preventive under the Educationfor All Handicapped programs professional Children Act lies inthe lack of consensus about the definition oflearning disabilities.Federal and State legislatorscan hardly be expected to fund enabling legislation forservices to a group of handicappedchildren whose and often contradictory. characteristics are varied,confusing, The very people whopurport to provide theseservices cannot agree ona definition; indeed,some of them question the clear-cut etiological existence of a group (Beers and Beers 1980;Sprinthall 1984). The debateover definition does a disservice who would mount to the cause of prevention.People preventive projects inthe field of learningdisorders are faced with the task ofjustifying interventionsto prevent phenonvina their colleagues that some of argue do not exist in the firstplace. Thevague and variable definitions do a similardisservice to the development In the face of the of preventive appmaches. definitional disagreements,the efficacy of successful tions may be doubtedby the compulsive interven- definers who will, inthe light of their idiosyncratic definition,dismiss the childrenwho responded intervention as not having favorably to the been truly learning disabledto begin with,or at least to have avoided learningproblems because of compensatory processes mysterious maturationaland (Lindsay and Wedell1982; Wedell 1980). Despite differences in some aspects of the definitionsof learning disorders and the methods bywhich this determination is probably can be made for individuals,there a considerable amount ofconcordanceamong professionals ing the nature oflearning disabilities. cor..ern- Ultimately, the definitionis one of exclusion in which learningdisability is said to be between a child's present whena discrepancy potential and actualachievement exists sensory acuity, adequate despite normal intelligence, conventionaleducational opportunities, and appropriatemotivation for learning. Preventiveprograms are organized to reduce the incidenceof these disabilities and the emotionaland behavioral consequence of such learningfailure. The need for with the problem of preventiveprograms to deal learning disabilities isespecially significantbecause of the long-term effects ofschool failureon every aspect of personality. 4 425 LEARNINGDISORDERS share three successful preventiveprograms According toBarclay (1984), basic components: identify thechildren at risk; 1. Assessmenttechnology to 2. A set ofinterventionprocedures; the results. 3. Sufficienttime to assess prevention depends onassump- in fulfilling thepromise of identifying the A program's success technology provides ameans for tions that theassessment with conventionalclassroom otherwise destinedto fail to learn children who are identified aremodifiable.It further that theproblems thus child with instruction and matching the needsof an individual depends on theeffectiveness of require sufficionttime strategies. Filially,successful programs with fol- the intervention monitor individualprogress the interventionstrategies, to basis of evaluation to apply the overall programon the lowup evaluations,and to improve feedback. assessment, inter- share the basicelements of Although successfulprograms choice of objectives. various programsdiffer in their vention, andevaluation, the 1982) that isselected point in the causalchain (Offord This choicedepends on the of learningdisorders Intervention programsfor the prevention for intervention. approaches. Thefirst approachconsists of two intervention aspects usually take either and cognitivestrengths in broad otefforts to promotesocial competency thus immunizedwill be less the assumptionthat children of their of functioning, on they may encounterin the course vulnerable tolearning difficulties enhancement efforts have beendescribed as schooling.These approaches identifies precursorsof learning (Lorion et al.1984). Thesecond approach regarded as specific skillsand competencies disorders andseeks to target The review of preven- successful learningduring the school years. essential to terms of thesetwo approaches. tive programsis organized in Experiments inEarly Education 1960's are anexcellent early interventionefforts of the The experimental objectives. These programs,with their with enhancement example of programs preschool children,their clearly target populationof poor,disadvantaged culminating in the and their long-termfollowup research all :439fined interventions, for LongitudinalStudies (1983), meet collaborative reportof the Consortium of the Consortium criteria. While eacnof the programs of Barcleir's success the pooled analysesof data 'arch and evaluationin its own site, of conductew opportunity forlong-term followup across prog. msprovided a remarkable analytic group,the Foun- early interventions.The independent in 1975 the results of the remarkablesample organized dation for HumanSer '':es, describes as follows: curriculum, focused on intervention studythat had t specific Every early completed prior to1969, and had an children oflow-income families, was invited to jointhe Consor- original sample in excessof 100 subjects was 4 3 426 PREVENTION OF MENTALDISORDERS titan. The investigators of allbut nr.t ofthe 15 eligible this invitation.Thus this is studies accepted is essentially not a sampleof preschool the wholepopulation of programs but rather studies conducted large-scalepreschool in the UnitedStates during intervention pouled studyare generalizable the 1960's.Results of the review is in thesame sensea thorough generalizableitsummarizes the literature tium forLongitudinal bmt available Studies 1983,pp. 411-412) data. (Consor- This researchopportunitywas a fortunate tion of one, because it long-term effectsof preschool permitted theevalua- required 15years of longitudinal programs that wouldordinarily have evaluation studies study.Resulta ofmost of the were mixed; thefindings of individual Head Start(Westinghouse a major evaluationattempt of content and Report 1969)were unsatisfactory methodology. Incontrast, the both in termsof cHldren from marked increasein enrollments that all socioeconomicgroups in early of many parentscontinued to childhoodprograms suggested efficacy of early believe intheir value. educationprograms in enhancing The questionof the functioning ofchildren at risk the cognitiveand affective for learningfailure remained The participating unanswered. research sitesprovided longitudinalresearch to a valid substitutefor prospective their original answer this question. raw data to The membersagreed tosubmit common protocol an independentgroup for analysis, for collectingfollowup data to developa control subjects,and to collect from theirexpekimental and analysis. It thecommon followup was also expectedthat individual data forindependent pursue individual projects in theconsortium would research directionsin termsof their resources; reports ofthese studies unique objectivesand However, for are included in the thepurposes of this consortium volume(1983). data aresignificant discussion, theresults based because theydocument on the pooled interventionprojects, all with common outcomesfrom diverse the goal ofenhancing the development ofyoung children. cognitive andaffective The major research questionof the overalleffectiveness of cor.sortiumprojectwas to assess and metaanalysis early educationprograms through the of dataon intelligence, primary,secondary, ment orientation, educational schoolcompetence, and achievement,achieve- than 1,000subjects. Ao ouupational might beexpected, the attainment frommore the projects;recovery rates ranged amount of attritionvaried from 31 to100 percent among percent. However,when recovered with a medianof 79 on four background subjectswere compared variables(pretest IQ with lostsubjects socioeconomicrating, and score, mother'seducational attrition IQscore at age 6), level, was random andthat the researchersconcluded that program and controlsamples Standardizedtests and were equivalent. used to address measures of socialand educational the question adaptationwere conclusions of the impactof early were reached: education. Thefollowing 1.Developed abilities, as assessed fromBinet, WISC, and achievementtest scores, and WISC-R showedsignificant scores 1'3 ' differencesfavoring the 427 LEARNING DISORDERS

after completing the program.In programparticipants up to 4 yews both participants performedbetter than controls on grade three, program results for four projects mathematics and readingachievement tests when five, pooled resultsshowed prograrn were pooled.At grades four and significantly higher inmathematics, but not inreading. graduates to be had generally disap- After grade six, theexperimental/control differences followup, achievementtest scores inreading and peared. In the final within the 25th to mathematics for programparticipants were generally 30th percentile range. positive attitudestoward achieve- 2. Early educationalso appeared to foster rated for young women.Older program participants ment, particularly controls without pre- themselves higher on schoolperformance than did school experience. school records 3. Analysis ofschool competence onthe basis of continuing effects on specialeducation placement and showed substantial program found between program in-grade retention.Robust differences were the likelihood of meetingthe basic educa- participants and controls on place- tional requirements oftheir schools byavoiding special education ment or repetitionof a grade. differential of 15 percent wasfound between theexperimen- 4. A significant obtaining a high schooldiploma. tal and control groupsin the likelihood of levels, 5. Overall there were nodifferential effects in termsof subgroups (IQ benefiting more thanothers from pre- sex, familystructure) of children school training. aspirations, were 6. Higher levelvocational aspirations, butnot educational associated with preschoolattendance. of school competenceand 7.Indirect effects, throughthe intervening variables achievement orientation, were seenin employment status,educational participants. attainment, and educationalexpectations for older program "Independently Lazar, who headed theconsortium group inIthaca, concluded: major studies of earlyintervention with low incomechildren and collectively, the positive effects of these programs reported in this volumeclearly demonstrate throughout childhood andadolescent years" (1983, p.461). Campbell 1984) at theFrank Porter The AbecedarianProject (Ramey and Center of the University ofNorth Carolina is a more Graham Child Development Since 1972 this recent example ofthe enhancementapproach in prevention. conducted to determinewhether systematic earlyeduca- experiment has been sample of psychosocially tion can prevent retardedintellectual development in a who, in the absence ofapparent biologicaldysfunc- defined high-risk children at. delayed intellectualdevelopment, 'Me experiment tion, appear to have preschool experiences teteach tempted to mod;fyenvironment and to provide skills required for school success. 4 2 428 PREVENTION OF MENTALDISORDERS

Unlike many of the earlier experiments,the Abecedarian high-risk families program identified and enrolled theirchildren as infantsin order to optimal developmentthrough the preschool support years. A high-risk index(based on parental education,family income, andhistory of mental learning failure in retardation or school the family) identifiedfour cohorts of approximately half in the 28 children each,with experimentalgroup and half in thecontrol group. These children begancoming to daycare at age 6 weeks. prehensivo serviceswere provided with Intensive, coin- and 1/6 in the a staff/child ratio of 1/3in thenursery preschool.Educational intervention Games for the First was based on Learning Four Years (Sperlingand Lewis 1981) children andconventional for the daycare nursery school activitieswith a communications skills emphasisfor the preschoolchildren. To controlgroups for physical equate the experimentaland care, bottled formula,pediatric services, work serviceswere provided for all and social participants, controland experimental. Children were testedon appropriate individual cognitive scales twicea year. Statistically significantdifferences favoring appeared on the the educationalinterventiongroup Bayley atage 18 months,on the Stanford-Binet and 48 months, andon the McCarthy scales at ages 24, 36, effect was computed at 42 and 54 months.The treatment to be approximatelyone standard deviation through 4years, with control children at ages 2 six timesmore likely to the retardedranges. These results earn scores within project awaits data are impressive. The finalassessment of the from school followupwith these children. The BrooklineEarly Educational search effort that Project (BEEP)is anotherlong-term re- follows theenhancement approach. program has soughtanswers about the feasibility Begun in 1972, the education of public school-basedearly programs.Evaluation datacollected as the second grade havebeen presented youngsters reached closed in October by Pierson et al.(1984). Whenenrollment 1974, 285 familiesfrom BrookThie Boston had respondedto the program's and adjoiningareas of group. Diverse background invitation forparticipationas the pilot characteristicswere represented in nicity, primarylanguages, maternal terms of eth- Three program age and education, andfamily structure. components were providedin a special line-Boston boundary: center near theBrook- (1) parent educationand support, (2) ing, and (3)educationalprograms for children. diagnostic monitor- Families were randomly assignedto one of the three and intensity:(1) the most levels ofprogram cost expensive level(projected at $1,200 year) involvedfrequent home visits, per child per scheduled at least meetings, andlimited childcare, each once every 4 weeks;(2) moe4rate the same offerings intensity ($800)involved with less frequentappointments scheduled and (3) the leastexpensive cost level every 6 weeks; ($400) involvedno outreach through visits, meetings,or child care, but offered home information andsupport at the BEEP Center, only atthe request ofthe parents. provided in Diagnostic examinations cooperation witha local medical center, were examinations but thepurpose of these was research andmonitoring, rether than primarymedical care. 4 13 LEARNING DISORDERS 429

Beginning at age 2 years,weekly play groups were heldfor the children. At prekindergarten ages 3 and 4,BEEP participants wereoffered a daily morning program in whichthe curriculum was influencedby the High/Scope Program (Hohmann et a). 1979). Theemphasis in this program was on"structuring space and materials to aflbrd eachchild an opportunity to develop a senseof effective- social skills essential for ness, to explore concepts,and to develop mastery and competencies in school." Evaluation of the BEEP resultshas provided multiviewpoint(teacher, Results at the end of parent, independentobserver) and multitimepoint data. second grade reflect thechildren's functioning 3 yearsafter the termination of BEEP services. Attrition rate wasapproximately 10 percent per year,with 169 random- of the original 285 familieeavailable for follovrup. A comparison group, ly selected and representativein relevant backgroundvariables, was observed along with program participantsfor six 10-minute in second grade classrooms identity of intervals by independent observerswho were unaware of the group the children. Results of theseevaluations indicated that: 1. A significantly smallerpercentage of BEEP participants(14.2 percent) than comparison children (28.4percent) were rated by observers as "having difficulty" with classroomlearning behaviors such as working independently, following directions,resisting distractions, completing work, and getting along withother children. 2. Assessment of readinglevels showed that 19.3 percentof the BEEP participants and 32 percent of thecomparison children were not decoding and comprehending stories atthe high second grade level,their current grade placement. 3. When program cost levels wererelated to level of parents' education, a direct relationship between levelof services and what was regarded as adequate reading performance wasfound with well-educated families; few children of well-educated parents werenot reading at grade level. educated, high 4.In contrast, with childrenwhose parents were not well levels of outreach were thought tobe necessary. In the programlevel group in which parents wererequired to initiate contact for services,there were no significantdifferences in the adequacy of readingbetween BEEP children and the comparison group atthe second grade. The experimenter concluded onthis basis that "for families who are nothighly educated, greater outreach was requiredin order to produce lastingeffects." The assumption of a causalrelationship between the level of servicesto parents and reading achievement appearsto go beyond the available data.An alternative hypothesis might be thatthe children's lack of school progressis associated with their parents' !ack ofeducational skills, which in turn madeit difficult for them to initiate requests forBEEP services. More careful analysis of intervening variables or possiblypath analysis similar to that done in the pooled analyses of the Consortium forLongitudinal Studies (1983, pp. 454-455) 434 430 PREVENTION OF MENTALDISORDERS

appear to be necessary to support thecunclusi on of a directrelationship between services to parents and readingachievement. It also appears that themeasures of reading outcomesare gross in view of the other aspects of BEEPS experimental design and thequestions for which answers were sought. To regardgrade-level achievement in quate" appears to be reading as "ade- setting rather limitedgoals for the children ofwell- educated parents. Furthermore,the adequate/inadequate take into account the dichotomy does not range of achievement onecan expect by second grade in any normally distributedgroup of children. To summarize, earlyeducation programs directed psychosocial competencies toward enhancement of have proved effective inrealizing these objectives. In contrast to comparablecontrol youngsters,a high proportion of graduates avoid the need for program special education placementor repetition of grades in school andearn fewer scores within the retarded They reflect positive ranges on intelligence tests. attitudes toward education,have greater likelihood graduating from high school, of and demonstratemere productive behaviors in the classroom.Older program graduateereflect the indirect effecta educational and social of greater competency in improved vocationalexpectations and status. Theseprograms, however, seem less effective taining specific educational in producing andmain- skills in reading and,to a lesser extent, inmathematics in about 20percent of program participants. interesting in that it corresponds This percentage is to the usual incidenceestimates of specific learning disability. Oneis tempted to speculatethat this group is composed learning-disabled youngsters of who requiremore specific interventions for their special learning problemsthan enhancementprograms provide. Administrative Strategiesto Prevent LearningDisorders A wide variety ofadministrative arrangements been recommended and innovativeprograms has for the prevention oflearning disabilities. enhancement approaches Unlike the described in the previoussection, these modifications are short-lived and often fall victim to the winds ofeducational change before clear-cut evaluation oftheir effectivenesscan be accomplished. Preventive interventionprograms are particularly vulnerable because of their lackof statutory funding andconsequent dependence tal funds for survival. on administrative or experimen- These conditions limitthe number of preventiveinterventions suitable forreview for thepurposes of this paper. limitations, however, 14 Despite these programs provide sufficient datafor some analysis of the efficacy of thesemodels. Departure from thetraditional lock-step been proposed organization of school gradeshas as an educational provision forI andergartners whoseem des- tined to fail in conventionalfirst grade class has been proposed programs. The pre-find gradetransition by a number ofinvestigators (DeHirschet al. 1966), but efficacy datawure not provided. More recently, language, reading, and Zenski (1983) comparedthe mathematics achievementof children who hadbeen

4 `)e LEARNING DISORDERS 431

of children who had placed in a transition classbefore first grade with those enrolled in second grade. repeated first grade whenboth groups of children were found between the two groupsin any of the No significant differences were Zenski noted that achievement measures at theend of second grade. However, practice with the academicskills the transition classexperiences provided little evaluation was based andspeculated whether moredefinitive on which the between the experimen- results might have beenfound if more time had passed tal procedures and theevaluation. Kilby (1983) produced A retrospective evaluationof a junior first grade by results from those of Zenski'sresearch. This ex post factostudy very different grades four through compared academic achieVementand social adjustment in first grade following their eight for children whohad participated in a junior kindergarten year.The achievement andadjustment of the experimental than that of comparableclassmates. It was also children was more favorable and fewer found that there were fewerreferrals to learning disability programs grade repetitions in the primarygrades for program participants.These results instruction in the junior first led to the conclusion that"the intensive reading grade may have had apositive and long term effect." although mixed Expansion of kindergartenofferings has also been suggested, the impact of early results have been reported.Weissman (1985), in a study of intervention on specialeducation students' readiness formainstreaming, found grade programs, whether theyinvolved regular educa- that expanded pre-first from unsuc- tion or special education programs,did not discriminate successful It is disappointing that thesocioeconomic level cessftil mainstream adjustment. intervention, was of the child's parents, regardlessof the kind of educational most closely associated withoutcomes. In contrast, Anderson(1984) reported increasededucational effectiveness associated with the full-day (asopposed to the half-day)kindergarten. The full-day program resulted in anincrease of instructional timefrom 180 to 270 minutes. The programs in bothhalf-day and full-day kindergartensdrew on similar curriculum content,with the major differencesbetween the classrooms consisting of ir,creased engagedinstructional time in the full-day programs. Classrooms were matched in termsof children's ages, socioeconomiclevels, sex, and entry-level skills onkindergarten skills inventory.Teachers in the full-day services of volunteer mothers asclassrvom aides, whereas the program had the Results of the half-day kindergarten teachershad the services of paid aides. Stanford Early School AchievementTest administered at the end ofkindergar- kindergartners in skills, ten indicated significantlyhigher scores for the full-day knowledge, and understanding inreading and mathematics. Parentalsupport that indicated of the full-day program washigh and may have influenced ratings greater self-confidence,independence, and ability to playcooperatively. Another kind of administrativeprovision suggested to preventlearning failure is grade placement on thebasis of individual patterns ofdevelopment. This approach is based on thework on school readiness doneat the Gesell 4,1.6 432 PREVENTION OF MENTALDISORDERS

Institute in New Haven, Connecticut, by 14 and Ames(1964). Their Gesell Readiness Screening Test consistsof a series of clinical tasks tion, copy forma, interview (block construc- questions, writing ofnames and numbers, Incom- plete Man Test, and assessment of gross motor control).Although the Gesell test has been normed and widelyused, few validation data have been presented. beyond case studies One recent study (Woodet al. 1984) examinedthe predictive validity of the Geselltein a sample of 84 referral for special needs kindergartners by using evaluation during kindergartenas the criterion for failure. According to Woodet al., developmentalage scores on the Gesell test were significantly related to thecriterion, but the variations of false negatives and in the percentages false positives with onlyslight changes in the cutoffscores were great enough to raise questionsabout the basic statistical of the measure. Nevertheless, characteristics this measure has hadconsiderable use by school personnel in making decisionsfor the grade placement cases, use of the test has resulted in of children. Insome the decision topostpone school admission even though the child had reachedthe legalage for school entrance. May and Welch (1984)have examined theapplication of the developmental placement approachas an administrative provision disabilities. for preventing learning They chosea developmental age cutoffscore of 4.5 years as indicating an "unready child"who will not succeed in it stressful. The kindergarten and will find developmental age conceptmeintains that 50 percentof all school problems couldbe preventedor remedied by placement in developmental ages. Problems terms of later diagnosedas emotional disturbance, learning disability, minimalbrain dysfunction, and result from asking children underachievementare said to to perform at levels forwhich they are not developmentally ready (May andWelch 1984).This approach does curriculum but suggests that not proposea children take anotheryear to mature in order to handle the regular schoolofferings. This approach is incontrast to early interventionbased on a child's needs assumes that waiting a and year in a less demanding environmentwill make a child ready for success in theconventionalprogram. May and Welch tested outcomes of predictions made the on 222 children in the second throughsixth grades of a suburban New Yorkschool district by locating tal agee as kindergartners children whose developmen- placed them inone of three groups: (1) childrenof developmental age below4.5 years whose parentsaccepted the recommendation to "buy a year" (i.e.,to postpone kindergarten mental age entrance); (2) children ofdevelop- scores below 4.5 but whoseparents did not accept the tion to buy a recommenda- year and who, according to thetheory, were "overplaced"; (8) children whose developmental and age scores were greater than 4.5 andwho were placed according to theirchronological age. The investigators found no significantdifferencee among the three numbers of referrals for groups in special educationplacement, speech andlanguage services, remedial servicesin reading or mathematics, or counseling.Two children from each of thethree groups had been and a few recommended to repeata grade more of the buy-a-yeargroup (significant at .05 level) 4 7 had been LEARNING DISORDERS 433 referred to adaptive-mos= and resource-roomprograms. May andWelch concluded that these results did notshow greater difficulties for theoverplaced group and thatmaturation alone will not make achild ready for schooling. These studies suggest thatadministrative arrangements bythemselves do not hold much promise for theprevention of learning disabilities.The exceptions may be in modificationsthat increase engaged time (as inthe full-day kindergarten) to the extent that abetter match between the child'seducational needs and the educational programtakes place. Projects for the Preventionof Learning Disabilities The design of most successfulintervention projects is based on anunderlying rationale that focuses on the natureof the tasks to be learned andthe specific links in the causal chain that theintervention will address.As might be expected, rationales differ in terms ofthe investigator's professionalorientation and the age of the population tobe served. Variations are greatbecause of complexities in the nature of learningdisability. Skarda (1974) and her associatesviewed learning disability in terms ofdelays the focus of their in language development.Thus communication skills became model preventive program, the objectiveof which was to developlife-oriented language skills in language-delayedchildren. This group provided a 2-year intervent;ovi program for earlyoral language in Wisconsin publicschools. Complex case finding proceduresinvolved referrals, prekindergarten screening, individual nevis assessmentthrough parent interviews, observationof the children in naturalistic settings,and multidisciplinary team evaluations.In- terventions involved "structuringthe auditory environment, reinforcing essential behaviors, and fosteringhome-school communication."Curriculum emphasis was on art, physical movement,and music activities, with extensive use of audiovisualaids. Evaluations based on parent interviews,language tests, and case studies provided evidencethat language functions wereimproved. However, followup data on the relationshipof these gains in terms of later school achievement were not available. Weiss (1980) also designed a preventiveproject with language as a central focus.Inclass Reactive Language UNREAL),located in Colorado, served Hispanic children who had limited Englishproficiency (LEP). The objectives of this project were to increase languagedevelopment of LEP children and to prevent later language-relatedlearning problems. intervention consistedof "non-stigmatizing methods using inclass servicedelivery, thus redefining the role of the speech-languagepathologist." Evaluation using matched INREAL and control groups showed that theexperimental program effected highly significant iraprovement in languageskills. A followup study 3 years after the original project's conclusion showed thatINREAL participants needed fewer remedial services (remedial reading,speech-language servicee) and were less likely to be retained in grade than controlsubjects. A cost-effectiveness study showed that an original investment of$175 per pupil in INREAL produced 434 PREVENTION OF MENTALDISORDERS

savings of $1,283.76 to $3,073.16,which might have beenrequired for special services to program participants.

The sequencing and timingof educational activitiesis another preventive focus proposed by some investigators (Ainscow and Tweddle1979; Stott 1974; Lindsay and Wedell 1982; Gredler 1978). These writerswere critical of predictive measuree and diagnosticprocedures as offering little tional intervention. guidance for educe- They argued that learningdisabilities result from unproductive learning styles and strategies rather than inherentdisability. This educationally focusedintervention model woulduse classroom-based screening by the teacher, followedby the setting of appropriate the children identified objectives for as having problems at that time.According to Ainscow and Tweddle (1979), thereis no implicit assumptionthat these children will academically in 1 or 2 fail years; the foeus is on present skills,using current evidence of functioning. This approachrequiree considerable training they can task-analyze of teachers so that educational content intosmaller steps, constituting mGre detailed aequence of specific a educational objectives. Suchobjectives provide criteria for ongoingmouitoring of childten's objectives used to monitor progress. The sequence of the child's acquisition ofbasic skills alsoserves as a means of tracing the child'sperformance to establish modifying the teaching. a starting point for It is unfortunate that, despite Lindsay and Wedell's(1982) critical appraisal of most early identificationefforts on both substantive and they have not provided statistical grounds, evidence of the efficacyof the objectives approachto intervention. The assumptionthat all children regular instructional can be helpel to cope with the program if teachers monitor theirperformance appropriately may be questionable.At least forsome children, this approach would merely defer provision of services untilthe child has failed,thus making the matterone of remediation rather thanprevention. Although this approach has value in that it placesresponsibility for educational management withthe classroom teacher,it may require higher planning skills in sequencing level objectives thanmany teachers are able to offer, given the prevailingconstraints of time and class To a great extent, the enrollments in typicalschools. efficacy of the objectivesapproach to prevention would depend on the qualityof implementation of themodel. In turn, the implementation would depend quality of on the classroom teacher'sprofessional skills, educational resourcee, time,and level of motivation. The importance of thequality of implementation illustrated in the of program models is ckise of a preventiveprogram designed and validatedin an Illinois school district.Evaluation data presentedto the Joint Dissemination Review Panel of the U.S. Office of Education andthe National Instituteof Education met the standardsof educational impact, effectiveness that qualified replicability, and cost the project for membershipin the National Diffusion Network (1980). Patricket al. (1984) reported the highly successful disastrous replicationof this original pi oject.Their evaluation of thereplication in 68

) LEARNING DISORDERS 435 schools in a school district in another statefound no significant differences in achievement test scores for schools inwhich the program had operated and those in which it had not. In fact, a smallnegative correlation was found between the number of minutes of treatment in the programand achievement test scores. These strange results prompted theevaluators to study the replication process. They surveyed implementation in theschools in which the program operated and found that the extent and quality ofimplementation varied conrklerably. Only 78 percent of the teachers in theseschools reported knowing how to implement the experimental program; 63 percent wereable to implement the program at all; 48 percentused the one-to-one or small group structure,and 49 percent used the modality centers that werepart of the design. The evaluators concluded that they could not determine theeffectiveness of the fully implemented program from the data they elicitedin their study. Thus, even well- designed models may be ineffective in settingsthat do not ensure adequate quality of implementation. Preventive approaches that draw on theneuropisychological subskills basic to reading have engaged theattention of a number of investigators and clinicians. Serwer (1971) designed an experimentthat contrasted these intervention strategies with conventional classroomapproaches involving the direct teaching of skills as advocated by the skills-orientedapproaches. Sixty-two first graders identified as being at risk for laterlearning difficulties were assigned to one of two special classes or weredistributed through regular first grade classes. Experimental treatments consisted of(1) direct teaching of reading (using supplementary phonics and language experienceactivities); (2) indirect teaching (perceptual-motor training); (3) combined directand indirect teaching approaches; and (4) a control condition (classroominstruction using a basal reader approach). Phase I of the program involved groupinstruction within treatment conditions, and phase II involvedindividualized instruction with the same treatment approaches.Major findings were that low, but statistically significant, correlations between treatment method andposttest achievement existed (the indirec t and combined treatment groupsshowed better achievement than the direct and control groups). Serwercommented that results may have been affected by the group's limited age range and thelimited intervention time spent on the intervention conditions (30minutes per day). Although research designed to contrast differing intervention approaches isneeded, few studies like Serwer's are reported. This lack of definitivestudies results from a number of conditions, including difficulties in keepingexperimental treatments from being contaminated in the natural atmosphere of theclassroom, the reluctance to withhold intervention in order to preserve acontrol condition, and the lack of commitment of resources and funding for experimentalresearch in schools. These barriers to development of preventive interventions were overcomein a fortunate collaboration between alearning disorders unit in a medical school department of psychiatry and an urban school district in aninterdisciplinary program using the SEARCH &TEACH model (Silver et al. 1978). 440 436 PRKVENT1ON OF MENTAL DISORDERS

The first step in the preventiveprogram is to locate children vulnerable to learning failure. This is done bymeans nf the scanning test, SEARCH (Silver and Hagin 1976, 1981). Prediction-performancecomparisons show rates of 5 to 10 percent false negatives and 1 to 9percent positives. The second step in implementing the preventive model is to providediagnostic examinations for children identified by SEARCH. The thirdelement of the preventiveprogram is the provision of educational intervention.These activitiesare described in TEACH (Hagin et al. 1976),a prescriptive approach designed to build foundation skills necessary forprogress in reading and t!..language arts. Data from the end of second grade producedstriking contrasts between the intervention and c( ,trol groups; reading comprehensionscores for the intervention group were significantly differentfrom those of controls. These educational gains have also been associated withsigns of normal behavioral adjustment in the upper elementary grades (Silveret al. 1981) and a lowering of the incidence gnonpromotion from 12 to 17 percent downto 1 to 3 percent during the 12years in which the preventiveprogram operated (Hagin 1984). Conclusion

The state of the art of preventiveinterventions with learning disabilities is not as bleak as first impressions ofthe research literaturemay suggest. Although moregaps exist in practice than in theory, thereare lessons to be learned from work that has already beendone in the field; 1.Both legal and theoreticalfoundations for prevention of learningdisabilities alroady exist. 2. A variety of methods for theidentif, cation of vulnerable childrenhave been developed and researched.

3. The values of early educationhave been demonstrated throughcareful analysis of extant data from the earlyexperimental projects of the 1960's. The enhanced social adjustmentfind increased educationalcompetence of the program participantswere reflected in lower rates of nonpromotion, referrals for special education, andnoncompletion of high school than comparison groups. However, theatypical learnersmay require special intervention projects addressed specificallyto learning disabilities. 4. A number of administrativeprovisions have been proposedas preventive interventions. However, results ofimplementation of these innovations have been mixed. The mostpromising are those that provideincreased amounts of quality instructional time. 5.Interventions have been designedwithvariety of program emphases: communication skills, sequencing ofeducational objectives, andneuro- psychological. Research aimedat comparing the relative meritsof each of these emphases will probablybe less useful andmay result in further 4 41 LEARNING DISORDERS 437

fragmentation of an alreadydivided field. A moreconstructive approach that individuals would relate assessment tointervention approaches so most suitable for could be matched moreappropriately with the programs their needs. training To best serve theprevention of learningdisorders, research and and diagnostic methods should be directed towardrefining early identification be closely related toeducational intervention inthe natural so that they can involved in such work setting of the school. Theinterdisciplinary collaboration would provide a rich sourceof data for the improvementof the education of young children. References

John Wiley, Achenbach, T.M. DevelopmentalPsychopathology. New York: 1982. Chichester, U.K.: Ainscow, M., and Tweddle, D.Preventing Classroom Failure. John Wiley, 1979. kindergarten. Paper Anderson, E. Increasing schooleffectiveness. The full day presented at Annual Meetingof American EducationalResearch Association, April 23-27, 1984. ERICDocument (R1E) ED 248036,1984. Barclay, J.R. Primary preventionand assessment, Personnel& Guidance Journal 62:475-478, 1984. disabilities: Facts Beers, C.S., and Beers, J.W.Early identification of learning and fallacies. ElementarySchool Journal 81:67-76,1980. in children Bender, L.A. Problems inconceptualization and communication eds. Psychopathology with developmental alexia.In: Hoch, P.H., and Zubin, J., of Communication. NewYork: Gruna and Stratton,1958. Bender, L., and Silver, A.Body image problems of thebrain-damaged child. Journal of Social Issues 4:84-89,1948. Bender, L., and Yarnell, H.An observation nursery.American Journal of Psychiatry 97:1158-1172, 1941. Blank, M., and Bridger, W.H.Perceptual abilities and conceptualdeficiencies in retarded readers. In:Zubin, J., and Jervis, G.A., eds.Psychopathology of Mental Development. New York:Grune & Stratton, 1967. Brody, S., and Axelrad, S. Anxietyand Ego Formation inInfancy. New York: International Universities Press,1970.

442 438 PREVENTION OF MENTALDISORDERS

Broman, S.H. Perinatal anoxiaand cognitive development In: in early childhood. Field, T.M., ed.Infants Born at Risk. NewYork, London: Spectrum Publication, SP Medical &Scientific Books, 1979. Broman, S.H.; Nichols, P.L.; and Kennedy, W.A.Preschool IQ: Prenataland Early Develo;.iontalCorrelates.Hillsdale, N.J.: sociates, 1975. Lawrence Erlbaum As- Caldwell, B., and Bradley, RH. Home Observationfor Measurement of the Environment Manual. LittleRock: University of Arkansas,1978. Caldwell, B.; Bradley, R.; and Staff. HomeObservation for Measurementof the Environment Little Rock:University of Arkansas,1984. Consortium for Longitudinal Studies. As the Twig IsBent: Lasting Eflictsof Preschool Programs. Hillsdale,N.J.: Erlbaum, 1983. DeHirsch, K Prediction inreading disability: A reviewof the literature. In: Hayes, A., and Silver, A.,eds. Report of the Reading Disability. Interdisciplinary Committeeon Washington, D.C.: Centerfor Applied Linguistics,1971. DeHirsch, K; Jansky, J.; and Langford, W .S PredictingReading Failure. New York: Harper and RONV,1966. Deutsch, C.P. Social classand child development. Ricciuti, H.N., eds. In: Caldwell, B.M., and Review of Child DevelopmentResearch. Vol. 3. Development and Social Policy. Child Chicago: University ofChicago Press, 1973. Doehring, D.G. Patterns of Impairmentin Specific ReadingDisability. Bloomington: IndianaUniversity Press, 1968. Farr, R. Reading, what can be measured? ERIC/CRIERReading Review Series. Newark, Del.: InternationalReading Association,1969. Gesell, A, and Thompson, H. Infant Behavior, ItsGenesis and Growth. York: McGraw-Hill, 1934, New Gredler, G. Learning disabilities and readingdisorders: A currentassessment. Psychology in the Schools15:226-238, 1978. Hagin, RA. Effectsof first grade promotionpractices of a prevention of learning program for the disabilities. Psychologyin the Schools 21:471-476,1984. Hagin, R.A.; Silver, A.A.; and Kreeger, H.TEACM Learning Tasksfor the Prevention of LearningDisability. New York: poration, 1976. Walker Educational BookCor- Hainsworth, P.K., and Sinquelend, M.L. MeetingStreet School Screening Providence, R.I.: Crippled Test. Children and Adults ofRhode Island, 1969. LEARNING DISORDERS 439

Hess, RD., and Shipman,V.C. Maternal attitudes towardthe school and the Passow, A.H., ed. De ?lop- role of the pupil: Somesocial class comparisons. In: ing hvgrams for trEducationally Disadvantaged.New York: Tea. -me College, Columbia t.iversity, 1968. Hohmann, M.; Bennett, B.;and Weikart, D.P. YoungChildren in Action: A Manual for Preschool Education.Ypsilanti, Mich.: High Scope Press,1979. Ilg, F.L., and Ames, LB.School Readiness: BehaviorTests Used at the Gesell Institute. New York: Harper& Row, 1964. Jansky, J., and DeHirsch, K.Preventing Reading FailurePrediction,Diag- nosis, Intervention. New York:Harper and Row, 1972. Kilby, G.A. An ex post factoevaluation of the junior first gradein Sioux Falls, South Dakota. DissertationAbstracts International 43:3271A,1983. Lashley, KS. The problem of serialorder in behavior. In: Jeffries,L.A., ed. Cerebral Mechanisms in Behavior.New York: Wiley, 1951. Lichtenstein, R., and Ireton, H.Preschool Screening. New York:Grune & Stratton, 1984. Lindsay, G.A., and Wedell, K. Theearly identification of educationally"at risk" children revisited. Journal ofLearning Disabilities 15:338-341,1982. Lorion, R.P.; Work, W.C.; andHightower, A.D. A schot !lasedmultilevel preventive intervention:Issues in program developmeat.Personnel and Guidance Journal 62:479-484, 1984. May, D.C., and Welch, E.Developmental placement: Does it preventlearning problems? Journal of LearningDisabilities 17:338-341, 1984. Mednick, S.A., and Baert, A.E., eds.Prospective Longitudinal Research. New York: Oxford Univerlity Press,1981. Mitchell, J.V., ed. The NinthMental Measurement Yearbook. TheBums In- stitute of Mental Measurements.Lincoln: University of Nebraska Press,1985. Murp'dy, L.B., and Moriarty, A.E.Vulnerabi. Coping & Growth. New Haven: Yale University Press, 1976. National Diffusion Network.Educational Programs that Work. San Francisco: Far West Regional Laboratory,1980. hational Joint Committee on LearningDisabilities. Learning disability and the preschool child. ERIC Document206544, 1985. Offord, D.R. Primary prevention:Aspects of program design andevaluation. Journal of the American Academy ofChild Psychiatry 21:225-230, 1982. 444 440 PREVENTION OF MENTAL DISORDERS

Paget, KD., and Bracken, B.A. ThePsychoeducational Assessment ofYoung Children. Now York: Grune &Stratton, 1983.

Patrick, A.H.; Kimball, G.H.; and Crawford, J. Evaluation of the earlypreven- tion of school failure program for kindergarten students. ERICDocument 244981, 1984.

Petrauskas, R., and Rourke, B.Identification of subgroups of retardedreaders: A neuropsychologicalmultivariate approach. Journalof Clinical- Neuropsychology 1:17-37, 1979.

Pierson, D.E.; Walker, DK; andTivnan, T. A school-basedprogram from infancy to kindergarten for childrenand their parents. Personnel& Guidance Journal 62:448-455, 1984.

Public Law 94-142, Educationfor All Handicapped ChildrenAct. (1975) Federal Register 42:121,a,5, December29, 1977, p. 65083. Ramey, C., and Campbell, F.A. Preventive education for high riskchildren: Cognitive consequences of theAbecedarian project. AmericanJournal of Mental Deficien,7 88:515-523, 1984. Rourke, B.P. ed. Neuropsychology of Learning Disabilities:Essentials of Sub- type Analysis. New York: GuilfordPress, 1985. Rourke, B.P., and Strang, J.D. Subtypes of reading and arithmeticdisabilities: A neuropsychological analysis.In:Rutter, M., ed.Developmental Neuro- psychiatry. New York: GuilfordPress, 1983. Rozin, P. The Psychological Approachof Human Memory, 3-46.In: Rosenzweig, M.R., and Bennett,E.L., eds. Neural Mechan;smsof Learning and Memory. Cambridge: MIT Press,1976.

Sameroff, A.J., and Chandler,M.J. Reproductive risk andthe continuum of caretaker causality.In: Horowitz, F.D., ed.Reinew of Child Development Research. Vol. 4. Chicago:University of Chicago Press,1975. Satz, P., and Fletcher, J.M. The Florida KindergartenScreening Battery. Odessa, Fla.: PsychologicalAssessment Resources, Inc.,1982. ;at; P., and Friel, J. Somepredictive antecedents ofspecific reading disability: A preliminary one-year follow-up, 1977-78. In: Satz,P., and ROSS, J.J., eds. The Disabled Learner: Early Detection and Intervention.Rotterdam: Rotter- dam University Press, 1973.

Sat; P., and Morris, R.Learning disability subtypes;A review. In: Pirozzolo, F.J., and Wittrock, KC., eds. Neuropsychological and CognitiveProcesses in Reading. New York: AcademicPress, 1981. 445 LEARNING DISORDERS 441

theory of specific develop. Satz, P.; Rarden, D.; andRoss, J. An evaluation of a ment dyslexia, ChildDevelopment 42:2009-2021,1971. dyslexia: A theoretical for- Satz, P., and Sparrow, S.Specific developmental D.J., and Satz, P., eds.Specific Reading Disability: mulation. In: Bakker, University Press, Advances in Theory andMethod. Rotterdam: Rotterdam 1970. Satz, P.; Taylor, H.G.; Friel, J.;and Fletcher, J.M. Somedevelopmental precur- In: Benton, A.E., and Pearl,D., sors of readingdisabilities: A 6 yet,r follow-up. York: Oxford Univer- eds. Dyslexia. An Appraisalof Current Knowledge. New sity Press, 1978. An evaluation of a Satz, P., and Van Nostrand,G.K. Developmental dyslexia: theory. In: Satz, P., and Ross,J.J., eds. The DisabledLearner: Early Detection and Intervention. Rotterdam:RrAterdam University Press,1973. Constructive Aspects. New Schilder, P. Mind: Perceptionand Tht ught in Their York Columbia UniversityPress, 1942. for children with specific Serwer, B.Experimental model school program learning disabilities. ERICDocument ED 087149, 1971. and Siegel, L.The prediction of possiblelearning disabilities in pre-term full-term children. In: Field, T.,and Sostek, A., eds. InfantsBorn at Risk. New York: Grune and Stratton,1983, pp. 295-315. Delineation of the Silver, A.A., and Hagin, R.A.Specific reading disability: syndrome and relationship tocerebral dominance. ComprehensivePsychiaby 1:126-136, 1960. Silver, A.A., and Hagin, R.A.Specific reading disability:Follow-up studies. American Journal of Orthopsychiatry34:95-102, 1964. Silver, A.A., and Hagin, R.A.Profile of a first grade: A basisfor preventive psychiatry. Journal of the AmericanAcademy of Child Psychiatry11:645-674, 1972. Silver, kA., and Hagin, R.A.SEARCH: A Scanning Instrumentfor the Preven- Book Corporation, tion of Learning Disability.New York: Walker Educational 1976, 1981. Silver, A.A.; Hagin, R.A.; andBeecher, R. Scanning diagnosisand intervention in the prevention ofreading disabilities.Journal of Learning Disabilities 11:437-449, 1978. secondary prevention Silver, a t.; Hagin, R. A.;and Beecher, R. A program for of learning disabilities:Results in academic achievementand in emotional adjustment. Journal of PreventivePsychiatry 1:77-87, 1981. 446 442 PREVENTION OF MENTALDISORDERS

Skarda, D. Preacademic program for children delayed in oralcommunication skills. ERIC Document096776, 1974. Small, N. Levels of Perceptual Functioning inChildren: A Developmental Study. Unpublished master's tnesis. Gainesville:University of Florida, 1969. Sperling, J., and Lewis, I. Learning Games for the FirstThree Years. New York: Walker Educational BookCorporation, 1981. Sprinthall, N.A Primaryprevention: A road paved and procrastinations. with a plethora of promises Personnel & GuidanceJournal 62:491-495, 1984. Stott, D.H. A preventiveprogram for the primary grades. Journal 74:299-308, 1974. Elementwy School

Tableman, B., andKatzenmeyer, M.Infant mental health Tableman, B., and Hess, services.In: R., eds. Prevention: TheMichigan Eaperience. New York: Haworth Press,1985. Vellutino, F.R. Toward an understanding of dyslexia:Psychological factors in specific reading disability.In: Benton, AL., and Appraisal of Current Pearl, D., etls. Dyslexia:An Knowledge. New York:Oxford University Press,1978. Wedell, K Early identification and compensatoryinteraction. In: Knights, It, and Bakker, D.J., eds.Treatment of Hyperactive Children: Current Research. and Learning Disordered Baltimore: UniversityPark Press, 1980. Weiss, RS. Efficacy of INREAL interventionfor preschool andkindergarten language handicapped andbilingual (Spanish) 204071, 1980. children. ERIC DocumentED Weissman, C.S. The impact of early intervention,PL 94-142, and otherfactors on mainstreaming. ERICDocument 245911, 1985. Werner, E.E., andSmith, RS. Kauai's Children Come of Age.Honolulu: University Press of Hawaii,1977. Werner, E.E., and Smith,R.S. Vulnerable but Hill, 1982. Invincible. New York:McGraw-

Westinghouse LearningCorporation. HThe Impactof Head Start: An tion of the Effects ofHead Start on Children's Evalua- ment." Executive Cognitive and AffectiveDevelop- Summary. Washington,D.C.: ClearingHouse for Federal Scientific and TechnicalInformation, 1969. Wood, C.; Powell, S.; and Knight, R.C.Predicting schoolreadiness: The validity of the developmentalage. journal of Learning Disabilities 17:8-11,1984. Zenski, J.P. A study ofthe effects ofa prerirst grade transition class with first graderetention on reading as compared 1983. achievement. ERICDocument 248459, 447 CHAPTER 12

Prevention of PsychiatricDisorders in Children and Adolescents: A Summary of Findings and Recommendations from Project Prevention

David Shaffer, M.D., M.B., B.S., F.R.C.P. F.R.C.Psych Professor of Psychiatry & Pediatrics Chief of the Division of Child Psychiatry Columbia University New York State Psychiatric Institute Introduction For a time in the 19th century, mental illness waswidely regarded as a form of inherited "degeneracy," preventable only througheugenics, Except for that period, generally humane and optimistic views aboutpreventing children's unhappiness and disturbed behavior have prevailed during the past100 years. In the early decades of this century, when psychiatricdisorder in children was predominantly viewed as a distortion of normal development,it would be prevented by psychological interventionsa liberal upbringingthat avoided the dangers of repression, child-rearing informed by theprinciples of learning theory, or, in a utopian situation, prophylacticpsychoanalysis. After World War II, more emphasis was placed on the social correlates ofdisturbed behavior; it was hoped that a happyand productive childhood would follow from prolonging and improving education, extending prenatal care to all,and improving access to medical mources and housing.These goals, which require no scientific justification, were probably oversold with respect to preventingmental illness. A period of disillusionment seems to have followed. Few preventiveintervention program focus solely on childand adolescent mental disorders, and one senses that many mental health professionals are now skepticalabout being able to prevent these disorders. Concerned about this development, the American Academy ofChild and Adolescent Psychiatry led Project Prevention, an initiative thatcommissioned

This paper was prepared for the Project Prevention initiativeof the American Academy of Child and Adolescent Psychiatry. 4 4 8 444 PREVENTION OF MENTAL DISORDERS

the series of reviews by experts thatmake up this volume. They havebeen written at a time when knowledge andbelief about child psychopathology isin flux. Few would now accept thatmental disturbancescan be explained solely as a result of stress imposed on the normaldevelopmental process (withan implication that there is universalpotential for disturbance); social-environmental causes are challenged by theconcept of invulnerability and by research that shows powerful interactions betweengenetic and environmental variables. It is a time when genericconcepts of disturbanceare no longer accepted, but when confidence in the diagnosticclassification provided by the Diagnosticand Statistical Manual of the AmericanPsychiatric Association (DSM) andInterna- tional Classification of Diseases(LCD) systems still must bewon. We wait impatiently for research findings thatwill point to the causal influenceof discrete modifiable risk factors(needed to formulatea rational prevention policy). The recent discovery ofdiscrete gene loci for bipolar affectivedisorder and possibly for schizophrenia hintat a future when wecan identify carrier status and undertake prenatal diagnosis(and, paradoxically, returnto the eugenic strategies ofa century ago), but many still mistrust themedical illness model of child and adolescentpsychopathology. Where do the experts standat this time of change? Are theyoptimistic or overwhelmed and disillusioned?Is there anythingwe can do to prevent psychiatric disturbance in children?This collection ofpapers provides an excellent view of how experts, in thelate years of this decade,are thinking about the problem. None suggest thatit is easy, but all point toways in which we could act to shorten, if notprevent, young people's distress andunhappiness. Review of the Papers

The papers in this volume havebeen grouped into three categories:those that are conceptual and methodological,those dealing with childrenexposed to known risk factors, and thosedealing with the preventionof specific deviant outcomes. Conceptual and MethodologicalPapers Sameroff and Fiese'spaper addresses "Conceptual Issues inPrevention." A number of basic preventionconcepts are challenged by Sameroffand Chandler's transactional model of behaviordisturbance.The model depictsa dynamic relationship between the individual'scharacteristics and ev.vironment,with continuing adjustments in eachbeing made in reaction to thechanging characteristics of the other. It followsthat a child's risk status,an important concept in prevention, is not static.It may be unexpectedlymodified by a largely unpredictable relationship betweenthe child's individual characteristicsand his or her environment. Lorion, Price, andEaton, commentingon this chapter in their own paper, also notethat the concept of diseaseonset that is essential to Caplan's primary, secondary,and tertiary categorization ofprevention activithis 4,) SUMMARY 445 does not lend itaelf to the transactional model.This is because the early features of disorder may generate either an increase indysfunction or its correction, depending on the environmental response. As do most of the authors in this volume, Sameroffand Fiese point out that true primary prevention is rarely possibleand that most preventive interventions will be provided to children after theirrisk status has been established. They classify useful interventions into:remediations that will restore the child to the expected state; redefinitions toidentify to parents the possibilities of normal caregiving; and reeducation, teaching parentshow to raise their com- promised child. They caution that because differentcombinations of risk factors will occur in different individuals, no single interventionis likely to suffice for all children, even those in a single classforexample, the offspring of psychiatrically ill parents. They note that most interventions arelikely to be prolonged and costly (a point also made by Eisenberg)and may need to be applied repeatedly at different times during thechild's development. Given their emphasis on the modifying role of the environment,it is not surprising that they stress the need to include the child's familyin any preventive intervention plan. Lorion, Price, and Eaton's paper, "The Prevention of Childand Adolescent Disorders: From Theory to Research," describes the firstthree of the four steps that are involved in the development of a successfulprevention program. The first is epidemiological or other research to identifymodifiable risk factors for the condition to be prevented. This is followed byresearch to find ways to prevent exposure to the risk factoro, either byeradicating them from the environment or by teaching effective avoidance. In the third stage, aprototype or model program is field tested todetermine its immediate consequences, to evaluate its cost, and to note the tendency of implementors todepart from the original model. The fourth stage, broad implementationof the prevention program, is discussed in the following chapterby two of the same authors. Lorion and his coauthors point out that most of the disorders referredto in this volume have a very low base rate and suggest that, forethical reasons, this calls for a "high-risk" rather than a "universal" approach (i.e., anintervention that is provided to all regardless of risk status).They argue (1) that arLy intervention that is powerful enough to change behavior toproduce a beneficial effect could also have an unwanted (and usually unforeseen)adverse effect, and (2) that because an overwhelming proportion of the populationis not at risk, even a small negative effect or.this large group could outweigh more powerful benefits to the much smaller population at risk. The authors join in the caution that primary prevention will notbe possible until more is known about causal mechanisms, but they stress thatinterven- tions that lower prevalence by reducing the duration of adisorder (secondary preventions) or reduce the complkations or recurrence of adisorder that is already manifest (tertiary prevention) are very worthwhile goals.

4 o. 446 PREVENTION OF MENTAL DISORDERS

In their chapter, "Prevention Programmingas Organizational Reinvention: From Research to Implementation," Priceand Lorion study the fourthstage of the process referred to above, introducingto the public at largea program whose efficacy has been shown only under thespecial conditions ofan experimental trial. Because thisprocess often fails, the authors set out to analyzewhy, in order "to assist readers to avoid thatfate for their preventive efforts." Program implementation will nearlyalways be carried out byan existing organizational structure, suchas a school board or a community mental health service. The organizationmay fail to implement the intervention appropriately because (1) it is inappropriatelypresented by its supportersor inventors; (2) the organization may not, fora variety of reasons, be ready to accept theintervention; or (3) even under optimalcircumstances, the program will haveto be adapted to meet local conditions, and,during the course of this adaptation,core or necessary elements may become distorted,damaged, or lost. The authors suggestways to avoid these problems. Program designerscan set specific short-term goals thatwill provide early feedback forthose implementing the program. Theycan involve the organizationas a fully participating partner. They shouldnot stop there, but should also reachto other powerful communitygroups to add their weight and influence. Theycan learn about the community's experiencewith similar interventions inorder to avoid repeating their mistakes.

Indicators of organizational readinessinclude the community'sprevailing attitudes and beliefs; whether thecondition to be prevented isseen as a problem; and whether pressure to dosomethin g about prevention hasdeveloped in influential sections of the community.The organizations thatare most likely to implement a program successfullyare those that already have a structure and affiliations suited to theintervention, rather than those thatneed to develop new partnerships and ways of working. Finally,effective implementation is more easily attained in flexible rather thanrigid organizations and in thosewith good communication between staff members rather thanin those that relyon bureaucratic dictate to produce change. Leon Eisenberg paintson a broad canvas and usesmany examples to illustrate his readable andscholarly paper, "Public Policy:Risk Factor or Remedy?"

He concludes, almost before hebegins, that with few exceptions,prevention offers better health at additionalcost and that to pretend otherwiseis to make a promise that cannot be fulfilled: "What is proffered on the flau..gi premiseof cost control may justas easily be denied when it becomes evidentthat it adds to expenses." On the other hand,the humane benefits ofprevention are incalculable, andeven the economic balance can be madeto favor prevention if indirect costs (the cost to the childand the parents)are considered in addition to direct costs (the cost to the healthsystem). However, the costs the name of prevention incurred in are not confined to dollars and cents, andhe providesa number of instances in whichpreventive interventions haveinadvertently led 451 SUMMARY 447 to increased incidence of the problem they havebeen designed to prevent (as among children exposed to certain typesof alcohol and other drug use curricula and those who participated in certain delinquency preventionprogram). Eisenberg, like Sameroff, questions the paradigms used to justifyand describe prevention. He argues that the vaccine model is notapplicable to noninfectious diseases and that the most dramatic example ofprevention through vaccination, the eradication of smallpox, was a unique instance even among infectious diseases and depended on thespecific characteristics of the condition.It has no parallels with other immunization programs orwith preventing noninfectious conditions. In the diseases that concern us,whether intoxications or failure to meet a child's psych osocial needs, single interventions, unlike vaccinations, do not confer permanent protection, and repeated ex- posures to the noxious agent, rathel thanprotecting against the disorder, merely add to the injury. Research is needed to understand how risk factors operate and to learnwhat type of intervention program works. In a detailed analy3is at the endof his paper, Eisenberg carefully documents therelative underfunding of psychiatric research in general and for children in particular.However, a quantity of research-based knowledge already available has not been implemented in public policy. Only a pedant would require more research to show that poor quality child care, unsupported parents, and young teenage mothers may hurt and harm. What is needed is not only more research but also, in Eisenberg's words, political resolve. Papers on Children of Known Risk Status Three papers review the outcomes and preventive interventions for children known to have been exposed to risk conditions. The conditions are having a parent with a psychiatric disorder, having a atonic physical illness, and experiencing a psychological trauma. Rutter's paper, "Psychiatric Disorder in Parents as a Risk Factor for Children," is especially important to Project Prevention because it addresses a readily identifiable risk situation in a familiar system that may be, in Price and Lotion's terms, "organizationally ready." The paper is closely argued and well documented, providing evidence for mechanisms and preventive interventions that are within the repertoire of activity by clinicians who work with children and families. Rutter identifies a number of specific features of a parent's psychiatric illness that could promote disturbance in the child.These include damage to the biological environment (e.g., fetal alcohol syndrome and the high incidence of birth injury in the offspring of schizophrenics); the high rablof marital discord when one or both parents have a psychiatric illness; direct involvement of the child in the symptoms of the parent, which occurs in some anxiety and psychotic states; interference with relationship formation that may occur between depressed and withdrawn mothers and their young children; and the 448 PREVENTION OF MENTAL DISORDERS consequences of discontinuous care that occur as a result of repeated hospitalizations. Preventive interventions include reducing alcohol intake duringpregnancy; counseling for disturbed parents with significant marital problems; counseling parents to reduce "expressed emotion" when they havea child who may be at risk for schizophrenii spectrum disorder; and direct treatment of childrento alter their perception and reaction to their parents' disturbed behavior. Pleas and Nolan's paper, "Risks for Maladjustment Associated with Chronic Illness in Childhood," examines the strength of the relationship between chronic physical illness and psychiatric disorder. The overall rate of psychiatric disorder in children with a chronic physical illness is twice that of the physically healthy population, and many of the children so affected willgo on to experience persistent psychiatric disorders in adulthood. Because chronic physical illness is common, the effect is subetantial; the authors estimate thatas much as 10 percent of psychiatric morbidity in children is attributable to associated physical illness. There is a considerable need to identifymore vulnerable subgroups. Pless and Nolan find no convincing evidence foran interaction between chronic physical illness and demographic factorsnor with the type of medical care provided. However, illnesses involving the centralnervous system seem to enhance risk. The authors also suggest thatan illness' visibility is a protective factor because it elio its social supports. There is clearly a need for substantially more research into the mechanisms of risk and the impact of preventive interventions in thisarea. Pynoos and Nader address "Prevention of Psychiatric Morbidity in Children After Disaster." Exposure to a psychologically traumatizingevent constitutes a significant risk factor for later psychiatric disturbance at all ages. In most cases it takes the form of what is sometimes a very persistent posttraumatic stress disorder (PTSD). Episodes of depression, particularly when thetrauma is associated with a significant loss, alsooccur along with other problems that the authors group as regressive behaviors. Onset is rarelydelayed, and most children and adolescents with sequelaeare symptomatic immediately after the trauma. The likelihood of significant psychiatric sequelae developingappears to be directly proportional to the degree of trauma experienced.This can be gauged by such indicators as proximity to the impactzone and the extent to which the disaster posed a direct threat to the child's life, caused physicalinjitiy, or resulted in a significant loss. Certain modifiable experiencesthat often follow a disaster may promote the persistence of symptoms. These include the child or adolescent witnessing injury or death in others, being discouraged from talking about the experience,or being separated from their parents or significant others. In the more comprehensive studiesthere is little evidence for an interactive effect with social demographic factors.

453 SUMMARY 449

preventive interventionsthat are models of Pynoos and Nader suggest some Children and how knowledge aboutrisk factors can beapplied practically. riskthose who sufferedsignificant exposure andwho adolescents at greatest readily identifiable and are symptomaticimmediately after the traumaare debriefing and/or othertherapies that may be can be referredfor psychological disaster's impact can be reimbursable with Federaldisaster relief funds. The siblings whenever minimized by not separatingchildren from their parents or possible, encouraging them totalk about the experience,and if possible not using them as rescuers so thatthey may avoid seeingdead or mutilated victims. Papers on PreventingSpecific Disorders disorder, alcohol and These four papers focus onspecific disordersconduct have been commonly a focus other drug use, suicide, andlearning disorderthat for past preventiveactivities. Prevention," lists many Offord's paper, "ConductDisorder Risk Factors and family history of criminality;ethnic status; low risk factors, including a low levels of child autonomic reactivity/arousal;harsh parenting coupled with large family size (especiallyfor boys); a reinforcing supervision; marital discord; supervision and peer group;and attending a schonlthat provides inadequate discipline, has a high delinquencyrate, and puts lowemphasis on achievement. likely that there is a These factors are not allindependent of each other; it is dose-response effect so that evenif several factors are presentfor a child, reduction of one can be expectedto produce somebenefit. Among modifiable improving school characteristics,providing adequate risk factors, Offord lists parents, and providing con- day care centers forchildren with lax or harsh tinuous and secure care forfoster children. There hasbeen little research on the effectiveness of theseinterventions. of Alcohol and Drug Abuse:A Critical Review Kumpfer's paper, "Prevention detailed guide to the of Risk Factors andPrevention Strategies," provides a literature (citing many internal papersand monographs that are notreadily substantial body of ex- available) in a field where therehas already been a alcohol and other drug perience with preventiveinterventions. Risk factors for degrees of drug use, cohorts, use are complex,differing for different substances, ages, and ethnicstatus. universal and that Ku mpfer takes the viewthat the potential for abuse is not specific characterisCcs ofalcohol and drug usingadolescents have determined alcoholics' sons their vulnerability. Thesecharacteristics include being male having an associated (who have a ninefold increasedrisk of becoming alcoholics); having a personality characterizedby low social antisocial personality disorder; and a lack of bonding, a tolerance for deviance, astrong need for independence, skills; starting to smoke ordrink at an early age; andholding peer refusal family or prodrug attitudes. Environmentalrisk factors include living in a community or attending a schoolthat condones alcohol ordrug use and consort- ing with peers who usealcohol or other drugs. Much ofthe biological research 454 450 PREVENTION OF MENTALDISORDERS

in this area has focused on the apparently greatertolerance for alcohol anddrugs and on neurological,neurochemical, and by the offspring of neuropsychological differencesshown alcoholics and other drugabusers. Many of these are likely to be interrelated, and risk factors Kumpfer points out thatadvances in statistical techniques have improvedour ability to test causal models. Didactic drug preventionprograms are common and tion that increased operate on theassump- knowledge can leadto a change in attitudesand ultimately behavior. Thatmay be the case when the information providedrelates closely to the target for changeand when other change, such as the social-environmental factorssupport current social climate withrespect to cigarette smoking. However, high-riskyoungsters often avoid didactic programs is variable and difficult programs, the quality of the to supervise, andmany programs are ficial, briefly delivered,and inappropriately super- the audience. designed for theage and culture of

Another preventionmodel, characterizedby increasing pressure, reinforcing taboos, and resistance topeer using the "justsay no" approach, is reviewed by Kumpfer, whoconcludes that evidence for its efficacy is flawed.Although it may encourage low-riskteenagers to delayor even avoid initiation, its high-risk youth is effect on completely unknown andcould even be negative. expresses concern that such Kumpier a high proportion of preventionfunds is being spent on this type ofprogram. Other program models to be effective include that have not yet beenshown improving interpersonalskills, increasing personal feelings and awareness of directing teenagersto alternative activities. evidence thatsomf of these programs There is may increase alcoholor other drug use. Kumpfer's own work with the alcoholicparents of young children them to havea number of parenting has shown by parent training deficiencies. Correctingthese deficiencies or by family therapymay constitute a but there isas yet no long-term research primary prevention, to show that theyresult in a diminution of use when thechildren reach theage of risk. Shaffer, Garland, and Bacon's paper, "PreventionIssues in Youth Suicide," identifies the known riskfactors for suicide, prior history of suicide including malesex, white ethnicity, behavior, family historyof suicide, generalized disorder, history of anxiety aggressive behavior,depression, and alcoholand other drug use. Other factors thatseem to act as facilitators for suicide includeexposure to examples of suicideamong friends, in the community, media. Access to or through the news a means for suicide has beenthought to be Shaffer's researchindicates that a risk factor, but gun ownership ratesare similar in the families of suicides andcontrols; gun ownership rather than to enhance seems likely to influencemethod choice risk itself in otherwisesuicidal teenagers. This paper describesthe difficulty of rare event() but suggests that devising preventiveinterventions for as knowledge of specific riskfactors is acquired it should prove possibleto develop focused The widely applied interventions for adolescentsat risk. and underresearched,school-based suicide postvention projectsare described. Research prevention and indicates that mostchildren are SUMMARY 451 already aware of the content of the programs, and there is some disturbing evidence suggesting that the programs may perturb a small group of children at risk. On the basis of known risk factors, this paper suggests that the most appropriate directions for suicide prevention are to provide effective screening and treatment of teenage suicide attempters, especially males and attempt repeaters; to limit vulnerable teenagers' access to firearms; to systematically and repeatedly educate news madia about the appropriate handling of suicide news; and to undertake further research into the biologicalcharacteristics of suicides and suicide attempters. Silver, Hagin, and Karlen's paper, 'Prevention of Learning Disorders," reviews the educational backwardness that cannot be explained by gross psychosocial deficiencies or by low IQ. Extensive research has failed to identify any single risk factor, combination of risk factors, or area of psychological dysfunction that is predictably associated with educational failure. There is insufficient knowledge to develop a primary prevention program, and, like other authors, they suggest that the field is best served by the early identification of handicap through screening tests that cover a range of psychological functions (and which aretherefore inappropriate as guides to prescriptive teaching) that empirically predict later reading delay. The authors provide a clear review of outcome studies on preventive interventions and group them into enhancement programs, which provide a broad range of educational and supportive servicw to preschool children fromdisad- vantaged backgrounds; administrative manipulations to provide an extra year of education or to defer school entry to permit "maturation"; and specific tutoring programs. The enhancement programs (Consortium for Longitudbal Studies and the Abecedarian and Brookline projects) have found an identifiable effect on attainment for the first 3 to 4 years of school, but this effect does not persist. However, the programs do seem to confer other substantial benefits many years later, inciuding higher self-esteem, better employment prospects, lower dropout and arrest rates, and much-reduced likelihood of entering special education class. The rt.ason for the apparent discrepancy between a temporary educational effect and a more enduring social benefit is not clew:. One possibility is that early enhancement reduces the student's chances of being assigned to a special education program and that the social benefits reflect a sparing from the negative effect of such programs. The authors report no benefits for deferring school entry to allow maturation.Finally, little research has been done to investigate the effect of prescriptive reading programs, and the authors conclude that there is no good substitute for an increase in quality instructional time.

4 5 6 452 PREVENTION OF MENTAL DISORDERS

Recommendations

These papers provided the American Academy of Child and Adolescent Psychiatry with a basis for an extensive set of recomnendations. In preparing these we have heeded Price and Lorion's caution that an organization must be ready to participate in a preventive intervention. That argument applies as much to mental health professionals, whether they are child or general psychiatrists or psychologists, as to any other organizational entity.The Academy has therefore chosen an array of interventions that, wh ile serving solid prevention goals, are generally feasible and use available skills thatcan be delivered in a familiar context. These recommendations can be summarized in terms of the following educational efforts: 1,Providing information to specialized parent groups, suchas parents with a psychiatric illness; the spouses of prisoners with children; parents of alcohol and other drug users; parents of chronically ill children; and parents of mentally retarded children. 2. Providing information to the general public on such topicsas How to access child mental health systems; Role of mental illness in teenage suicide; Magnitude of the problem of conduct disorder; Who is at risk for conduct disorder; Benefits of early referral for aggressive and disobedient behavior; Value of parental involvement as a preventivemeasure in conduct disorder; Who is at risk after psychic trauma; Value of psychological interventions after psychic trauma; Risk of mental health problems in children with head injury and epilepsy;

Associated mental illness with alcohol and drug use and how thatcan be helped; and Dangers of firearms in homes with teenagers at risk.

4e 7 SUMMARY 453

3.asInfluencing activities by other professional or administrative groups, such Accreditation/regulatory bodies to ensure inclusion of knowledge about risk factors and preventive interventions in professional curricula and examinations; Disaster services to prepare psychological preventive interventions; Reimbursement agencies to advocate reimbursement for preventive work; Pediatric organizations to educate pediatriciano about risk factors among certain groups of the chronically physically ill (specifically the neurologically impaired child); Foster care agencies, local health departmenta, and prison health care facilities to develop screening procedures to identify young children with conduct disorder; and Education boards to develop anticipatory suicide postvention activities and peer modification programs for conduct disorder and to identify, early, young school absentees and risk factors in physically handicapped children. 4.Prornotiag initiatives among local groups of mental health professionals. These initiatives would include introducing early detection programs to serve children of psythiatrically ill parents, children with chronic physical illness, young truants, learning- disabled children, and children in foster care; and enhancing linkages/networks with (a) adult psychiatrists treating psychiatrically ill parents, (b) pediatricians serving emergency rooms to develop screening and management services for suicidal teens, and (c) those who cam for chronically ill children, to identify those with psychiatric symptoms.

5.Modifying training programs for child and adolescent psychiatrists in training (child fellows). This would include Advocating a special public/preventive/community child psychiatry track; Encouraging ci, ild fellows to participate in community activities; Ensuring that seminars are held on prevention theory and practice, risk factors, and interventions; Providing training in suicide postvention techniques, school consultation skills, and the techniques of psychotherapy for victims of psychic trauma; 458 454 PREVENTION OF MENTAL DISORDERS

Teaching parent training techniques for families of young, aggressive children; and Promoting child psychiatric trainee attendance at high-risk clinics that follow up brain damaged children. 6. Providing postgraduate training for child and adolescent psychiatrists, including special workshops and instituteson risk factors and preventive interventions, how to give workshops to parents, how to make media presentations after a suicide outbreak or a disaster situation, and how to set up screening programs to detect children at risk. 7. Providing postgraduate training for pediatricians, focusingon the psychological tido for children with chronic physical illness and howto improve the care and assessment of suicide attempters inemergency rooms. 8.Providing postgraduate ttsining for general psychiatrists, including how to identify and manage or refer children of psychiatrically ill parenta, children of suicidal parents, and children of alcohol and drugabusing parents. 9.Research advocacy, specifically promoting longitudinal, epidemiologically based reaearch of risk factors, case control, and population-based studies among patient groups; research of causal mechanisms, intervention, and screening strategies; and support for demonstration projects. Finally, the Academy itself will undertakea number of projects including publishing materials suitable for parenta of childrenat riak; promoting academic meetings about risk factors andpreven tive interventions; launching appropriate public service advertisements; and encouraging collaborativework- ing with parent/patient advocacy groups.

Conclusions

Project Prevention offers an encyclopedicarray of facts, references, and valuable insights. In reviewing thepapers, certain important themes recur: 1.There is little support for a universal vulnerability model(except perhaps for PTSD following exposure to extreme trauma). As Kumpferstates in her chapter on alcohol and drug abuse, "An underlyingassumption...is that youth do not develop disabling addictive disorderswithout prior behavioral, emotional, or cognitive precursors." 2.It follows logically that most authors favor usinga high-risk strategy, rather than programs addressed to all individuals. Thisis partly because of the low efficiency of universal approaches (a compoundof their generally low power and their wasteful distributionon children who do not need 455 SUMMARY

prevention because of ethicalconsiderations; many them), and partly unwanted sideeffects, not been fullyinvestigated for interventions have children and teenswho do not audience is low-risk effect is and their predominant have most tolose if the program's need them andwho therefore harmful. of children atrisk. Because strategies requirethe identification there 3. High-risk who may haveonly slight symptoms, of a fear oflabeling children have to be reluctance to do this.This means that we is understandable and be sure thatthe benefits tactful in devisingidentification procedures from stigmatization. of an interventionoutweigh any risk limited, virtually of causes andmechanisms is extremely 4. Our knowledge implication is that weshould support ruling out primaryprevention. The longitudinal researchto improve ourknow ledge moreepidemiological and of mechanismsand risk factors. is generally notpossible, there are 5. Even thoughprimary prevention to providesecondary andtertiary prevention substantial opportunities the developmentof complications that will curtailsuffering and prevent in children andadolescents. subject to ellen ge proposed interventions are"fragile," that is, 6. Many of the their users. This meansthat P` and corruptionin the hands of features of a be paid toidentifying the core ornecessary attention has to apply those faithfully. program to ensurethat implementors and tertiarypreventions 7. The use ofhigh-risk strategiesand of secondary risk usually there is no one-on-onerelationship between and the fact that outcome serve toincrease the importance status and typeof psychiatric professionals in theprevention field.They are needed of mental health parenting to normal help parentsnotto teach normal in programs to the distortions that occurin families parents, but toanalyze and mitigate and psychiatrically ill parents orwith normal parents with stressed or to confirm riskstatus (as among disturbed children.They are needed and to screened positive on somerisk status survey) children who have need to be tailoredto the devise interventionsthat in most cases individual. professionals face inthe not to underestimatethe difficulty it is important it may affecttheir enthusiasmfor climate of today'spsychiatry and how vaccines for child Eisenberg may assure usthat there are no prevention. While and that weshould look atother and adolescentpsychiatric disorders unduly pessimistic. difficult not to feelthat his warnings are paradigms, it is conditions described inthis volume will Gene loci for atleast some of the and although it isclear that the geneticenvironment undoubtedly be found, highly likely thatit is to producedisorder, it seems alone Le insufficient technology for genereplacement When gene loci areknown, can the necessary. paradigm may yetapply. Until thattime there be far behind?The vaccination 4 6 1 456 PREVENTION OFMENTAL DISORDERS is a need toresist feeling overwhelmed andenfeebled by thousands ofchildren and ignorance. Countless features of adolescents whoare at risk or showing a disturbance remainuntreated and the early take this initiative,they will continue unrecognized. Unlesswe can in many instances, to suffer bothduring their during theyears to come. childhood and, INDEX

AACAP. See American Academy of Child Alcohol, Drug Abuse, and Mental Health and Adolescent Psychiatry Administration (ADAMII.A) Absosdarian Project reesarch ftinding source, 142, 145, 147 description, 427-28 Alcohol and drug abuse high-risk family identification, 428 common side effects, 11-12 Ablon, J. future considerations, 16-17 Al-Anon family groups, 343-44 prevention perspectives, 12-16 Academics' attitude to policymaking, problem perspectives, 7-12 137-38 risk factars ADAMHA. See Alcohol, Drug Abuse, and biological, 328-31 Mental Health Administration demographic, 315-19 Adams, G. R. psychosocial environmental, 319-28 postlisaster stress refttions, 228 summary, 449-50 Adams, P. R. side effects of prevention, example, 134 postdisaster stress reactions, 228 statistics, 8-9 Adaptive features of intervention, 117-18 Alcohol and drugs Adolescent parents, reeducation efforts, differencee in reaction to, 329-30 46-47 effect of alcohol on fetus, 165-66 Adolescent pregnancy suicide risk factor, 383 cultural code implications, 34 Alcoholism policymaking applications, 139 adopted-away children, 162 Adiescents, delinquency-prone, 76 alcohol-related behavior effects study, Ado,itable intervention, 58-69 165-66 Adopted-away children effect on fetus, 165-66 conduct disorder study, 280 gene environment correlationof risk, 166 genetic risk factor application, 162 ; Ade risk factor, 159, 160 schizophrenia application, 163-64 Alpha factor, research design, 82-83 Adult prognosis, conduct disorder, 274 Alternative programs, school-based, Affective and interpersonal education 336-37 programs, school-based, 335-36 American Academy of Child and Age Adolescent Psychiatry (AACAP) alcohol and drug abuse role, 317 Prevention Initiative, 3-4 and method, suicide application, 376-77 preventive intervenaon recommenda- postdisaster psychiatric morbidity role, tions, 452-56 235 Ames, L. B. Agency services, client use, 109 developmental placement study, 431-32 Agent/environment risk factor, 320-21 Anderson, E. Agent/host risk factor, 320-21 full-day kindergarten study, 431 Al-Anon, 343 Anthony, E. J. Alafam, 343 insecurity after disaster, 239-40 4 2 458 PREVENTION OF MENTAL DISORDERS

Anti-Drug Abuse Act of 1986 Biological transactions, 31-32 high-risk youth definition, 11 Biophysical vulnerability model, 329 Antisocial behavior. See Conduct disorder Blacks Archival sources, analysis, 108 alcohol and drug use statistics, 316-17 Area and oocioeconomic class, conduct conduct disorder statistics, 276 disorder application, 284 suicide rates, 382 Arthritis Blank, M. psychosocial dysftinction example, 212 learning disorders study, 416 Asians, conduct disorder application, 276 Body build, conduct disorder risk factor, Asthma 277 intervention an disease-altering factor, Bonding, 131-32 192 Brain damage psychosocial dysfunction example, 212 conduct disorder application, 278 Attempts and deaths, suicide emotional problems, 211 history of attempts among completers, 378 psychiatric disorder, increased risk, 278 Breslau, N. potentiators, 379 central nervous system involvement and predicting death among attempters, 379 psychosocial drifunction, 211 similar diagnostic profiles, 378 Bridge, T. P. Attributable risk, 71-72 crisis service model criteria, 396 Attrition, types, and sample size, 85-86 hotline effectiveness, 397 Bridger, W. H. Bagley, C. learning disorders ,itudy, 416 suicide attempters and followup British National Survey of Health and psychiatric treatment, 401 Development, 277 suicide prevention centers, 397 Brookline Early Educational Project Baits rates (BEEP) definitivn, 57 evaluation results, 429 predicting specific risk, 66 learning disorders study, 428-30 sample size applications, 83, 84 levels of program cost and intensity,428 BEEP. See Brookline Early Educational Broekowski, A. Project program monitoring principlee, 119-20 Before-and-after comparisons, prevention Bry, B. H. programa, 387-88 improvement of academic performance "The Befrienders," suicide prevention study, 344 television series, 395 Budget, research Behavioral prevention programs, school- based alcohol and drug abuse, 347-48 strategiee, 337 child mental health, 145-47 tobacco use, 337-39 Federal agencies, 143-47 Beverly Hills Supper Club fire, 227 guidelines, 143-44 Bien, N. A. priorities, 144 improvement of academic performance tax policy, 147 study, 344 Buffalo Creek Dam flood study, 226,229, Binge drinking, 316 237 4 (7 3 INDEX 459

Cambodian adolescent refugee'', concept of risk, 191 postdieaster psychiatric morbidity conduct disorder application, 277-78 eczample, 228, 232, 235 evidence in support, 194-204 Cambridge-Somerville Delinquency intervention, 204-9 Prevention Project, 133 maladjustment, definition, 194 Campbell, F. A. Nolan and Plena research, 194-204 learning disordem study, 427-28 overview, 193, 216 Caplan, G. research designs, description, 194-95 triad oi'prevention categories use, 69-62 subgroup, targeting, 192 Carbon monoxide, suicide prevention summary, 448 example, 389 Cigarettee Carnegie Council on Adolescent and alcohol tax, 147 Development high ochool smoking study, 316, 318 problem behaviors, attributes, 14 school-based prevention programs, report on prevalence rates for problem 337-39 behaviors, le Citizen survey, 109 Case-control study Clinical implications attributable risk result, 71-72 family interaction and parenting, 170-72 risk-factor identification strategy, 71 genetic mechanisms of risk, 164 transactional and medical paradigms, 72 response to parental mentalillness, typical, deecription, 71 individual differences, 176 Cannel, J. Clinical trials, definition, 58 stress model, 205 Coate, D. Caucasians alcohol use study, 320, 322 alcohol and drug use statistics, 316-17 Cognitive programa, school-based conduct disorder statistics, 376 focus, 333 suicide rate, 382 problems encountered, 334-35 CBCL. See Child Behavior Checklist Colorado Adoption Study, 163 Chamedes, W. A. Committee for Economic Development, noncompliant* of suicide attempters, 138-39 399, 400 Community-based prevention, alcohol Channel One programs, 336, 340 and drug abuse Chapter summaries, 444-51 community groups, 340 Child Behavior Checklist (CBCL), 212 parent groups, 340 Childhood Global Assessment Scale, 229, public media techniques, 339-40 235 Conununity forum, 110 Childhood trauma studies appendix, Community recovery, postdisaster 24844 psychiatric morbidity factor, 231 Children of known risk status papers, Community resources, analysis, 109 summary, 447-49 Comorbidity Chowdhury, N. desciiption, 10-12 random assignment study of suicide attempters, 402 postdisaster, 228 Compensatory good experiences, conduct Chronic illness risk factor disorder application, 288 chronic disorders, prevalence, 193 464 460 PREVENTION OF MEW'AL DISORDERS

Comprehensive Drug Rehabilitation and direct and indirect costs, 127 Treatment Act of 1986 influenza vaccine, 128 mandate to develop strategiee to lead poisoning, screening, 129 high-risk youth, 310-11 measles and rubella, 127-28 Conceptual and methodologicalpapers, summary, 444-47 pneumovax, 128 Conduct disc.:der m useful exercise, 129 attraction of prevention, 274 Coder, S. cross-generational effects, 31 hotline study, 397, 398 definition, 273 Coulee of disease evolution diagnostic issues, 273-74 accuracy of risk assessment, 66 genetic factor, 164 child and adolescent disorders Patterson model, 30-31 applicatt:..-s, 65-66 poor adult prognosis, 274 inevitability concern, 65 spectrum of disorders effect, 65 prevention possibilities, 289-90,294-96 risk factors Crisis services. See Hotlines CRS. See Congenital rubella criteria, 275, 291-93 syndrome C.s.f. 5-HIAA, suicide predictor, individual characteristics, 275-82 403-4 Cultural code protective factors, 286-89 psychosocial, 282-86 comparison with family code, 35 summary, 449 coping skills, 42 individual code effect, 38 Congenital rubella syndrome(CRS), 127-28 ingredients, 34 Consortium for Longitudinal Studies regulatory functions, 39 conclusions from standardizedtests, role of prevention, 34 426-27 statutes, 33, 36, 39 learning disorders, earlyintervention Cultural factors, postdisasterpsychiatric study, 425-27 morbidity role, 232 major research question, 426 Cultural sensitivity, 102 Constellation of behaviors, 30-31 Cystic fibrosis Coopersmith Self Esteem Inventory,195 psychosocial dysfunction example,212 Coping mechanisms alcohol and drug abuse application, Data collection 327-28 agency services, use of, 109 tenduct disorder application,288-89 archival sources, 108 parental coping skills, 42 characteristics, 110 postdisaster, 234-35 citizen survey, 109 school stress application, 324 community forum, 110 suicide application, 394 community resources, 109 Core technology, 117-18 key informant interviews, 110 Cornish test of insanity, 4 program monitoring, 118-20 Cost-benefit arugysie D'Aunno, T. A. alcohol and drug abuseprograms, 347-48 organizational rezdiness mrAsi,103 cost savings as justification for Dawn Report, 9 prevention, 128-29 Day care, 140-41 4f,r-;k) 461 INDEX

Disaster De Fries, J. C. preventive interventionstrategies Colorado Adoption Study,163 emergency medicalrelief, 238 De Hirsch, K. evacuation, preparing for,237 scanning test, learningdisorders, 419 minimizing, 237 Delinquency. See Conductdisorder exposure, family support, 239-4e, De Marsh, J. individual therapy, 240-41 Strengthening Families Program,341-43 of, 238 VASC theory, 319, 320, 328 news media, use psychological first aid, 238-39 Demographic risk factors school-based mental heaithteams, 240 alcohol and drug abuse secondary exposure, protectionborn, age, 317 238 gender, 315 primacy of exposure role race, 316 Beverly Hills Supper Clubfire example, socioeconomic, 317-19 217 psychosocial dysffinction, 210 Buffalo Creek Darn floodstudy, 226 suicide Mt. St. Helens volcaniceruption age, 381 example, 227 gender, 381-82 proximity factor, 227 geography, 382 psychiatric morbidity inchildren race, 382 childhood trauma studiesappendix, secular trends, 382 24844 Depression course of morbidity,229 genetic risk factor, 159,160, 161, 164 degree of exposure, 226-27 impaired parenting risk,169-70 Disaster Relief Act authorization,226 parental, and psychiatricdisorders in mediating factors, 230-36 children, 160 prevalence of morbidity, 229 parental symptoms, risk of exposureto, prevention interventionstrategies, 167, 171, 172 236-41 postdisaster psychiatricmorbidity, 228 proximity to disasterconsideration, 227 suicide risk factor, 377, 383,385 summary, 448 Development., postdisaster impact,235-36 types of morbidity,227-29 Developmental agenda, 39, 98 Distal outcome Develc mental lag, learningdisorders, definition, 64 417 organizational focusingapplication, Developmental placement, 431-32 111 Deykin, E. Y. Divoroe Acbustment Project,295 prevalence rates assessment,alcohol Domestic cooking gasdetoxification, 389 and drug abuse, 9 Driver education Diabetes Robertson study, 135-36 intervention as disease-alteringfactor, side effects of prevention,example, 192 135-36 psychosocial dysfunction example,212 Drug abuse. See Alcoholand drug abuse Direct costa, definition, 127 462 PREVENTION OF MENTALDISORDERS

Dryfoos, J. G. learning disabilitiesdefinition importance of schools inpreventive controversy, 424 intervention, 15 Educational interventions,learning report on prevalence ratesfor problem disorders behaviors, 10 administrative strategies,430-33 summary of epidemiological literature, 14 computer searches available,420 effective preventionrequirements, 420-21 Early childhood development experiments in early education,425-30 contextual influences,23-49 facilitators and barriers, early education, 422-25 experiments, 425-30 prevention projects, 433-36 Head Start Program model, 26, 138, 426 Embryolcgical develor.it, 31-32 identifying high-risk conditions,26 Emergencj med...cal reliefduring disasters, parental pathology effect,38 238 Perry Preschool Project,86, 294 Employment postdisaster impact, 235-36 alcohol and drug abuserole, 317-18 preschool years importancein learning conduct disorder application, disorders, 422 287 Environment/host risk factor,321-22 prevention programsimplications, 26-27 Environmental alcohol anddrug abuse risk factors, 27-28 prevention programs Sameroff sty y conclusions, 28 alcohol-impaired driving laws,346 transactional model, 28-32 alcohol server intervention,346 Early warning signs of suicide, 374, 391 increased cost of alcoholeffect, 346 ECA. See Epidemiologic Catchment Area Moskowitz review, 346 Eckman, T. reconmendedmeasures, 345 ramom assignment study of suicide Err 'ronmental factors,psychoiocial attempters, 402 dysfiinction Economic issues. See also Policymaking medical environment,215-16 cost-benefit analysis, 126-29 social environment,215 pellagra incidence,prevention Er mmental organizations, application, 125-26 33 Environmental risk factors research budget, 143-47, 347-48 alcohol, ef. Jetson fetus, 165-66 Education. St, also Schools family effects, 167-72 alcohol and drug abuse problems, 318 opioids, effecton fetus, 166 cultural code implications,34 pregnancy complications, 166-67 learning disorders, earlyintervention, 425-30 Environmentalscanning criteria for initial scanning, preventing failure, 138-39 107 data-collecting strategies,108 10 Education for All HandicappedChildren Act in-loop assessment,107-fl cost responsibility, 424 Tanizational context,107 free appropriate publiceducation eived risks and benefits,1 )7-8 mandate, 423 workshop planning,106-8 individual educational plan(IEP), 423 Environmental variablesof risk family eri-iironment,161 4 1: 7 463 INDEX

parenting, 169-70 Environmental variables(continued) physical Hake, 169 parental mality disorder, 162 Environtypt,, 32, 33, 42,47 Family factors alcohol and drugs Epidemiologic CatchmentArea (ECM etudy, 163, 315, 316,317, 318 attitudes and values, 325 diaruptions in family management,326 ERIC computer search,learning disordere, 420 parental use, 324-25 325-26 Etiologic fraction streesors and coping resources, chronic illness application,206 conduct disorder Etiological chains, elements,68 large family size, 283-84 Ettiinger, R. marital discord, 283 suicide attemptors endfollowup parental deviance, 282 psychiatric treatment, 401 poor parenting,?,82 Evacuation, preparing for,237 postdisaster importance,239-40 Experimental procedures, powerof Family-focused alcohol anddrug abuse alpha factor, 82-83 prevention followup, 86-88 children of chemicallydependent sample size, 83-84 parents, 341 Experimental stuiliee, suicideattempters, family therapy, 343 402 importance of family, 341 parent training programresults, 342 Fairfax County YouthSuicide Prevention self-help groups, 343-44 Center, 392 skills training program,334, 342-43 Family bereavement,poetdisaster Strengthening Families Program,341 233 psychiatric morbidity role, Family history Family breakup, riskmechanism, 167-68 suicide application, 385 Family code Family therapy model, 207-8 coping skills, 42 Farberrow, N. L. cultural code bridge, 40 suicide intervention, 395-96 definition, 35 Farrington, D. P. individual code effect, 38 conduct disorder study,277, 279, 281, myths, 37 288 operation characteristics,39-40 Fetal alcohol effect, 331 paradigms, 37-38 Fetal alcohol syndrome,166, 331 regulatory functions, 1: 41 Fetus, effects on rituals, 36 alcohol, 166-66, 331 stories, 36 drug abuse, 331 Family discord, riskmechanism, 167-68 opioids, 166 Family environment asrisk factor pregnancy complications,166-67 breakup, 168-69 Fidelity and adaptation clinical implications,170-72 adaptive characteristics,117-18 discord, 167-68 advantagee and disadvantages,116-17 exposure to specificparental symptoms, core technology,117-18 167 464 PREVENTION OF MENTALDISORDERS

Firearms increased vulnerabilityto environmental restricting access for suicideprevention, hazards, 163 389-90 schizophrenia application,183-64 suicide method, 382 Generative studies, 8'; Flay, B. B. Genetic factors smoking preventionpropam review, 338-39 alcohol and drug abuse,329 conduct disorder, 280 Florida KindergartenScreening Battery, 417 Genetic mechanisms ofrisk Florida LongitudinalProject, 416-17 adopted-away children, 162 Followup alcoholism, 159, 160 clinical implications, design of measurementprocedures, 87 164-65 documentation, 87-88 depression, 169, 160, 161,164 generative studies, 87 environmental variables,160-62 measurement over developmental gene-environment correlations,162-64 stages, 87 parental personality disorder,160, 162 Perry Preschool Projectexample, 86 role of genetic factors,158-59 suicide preventionprograms, 388 schizophrenia, 158, 159,160, 161, 164 validation studies, 87 Genotype, 31-32, 42 Friel, J. Geography sensory-perceptual factors,predictive suicide rates application,382 value, 417 Gesell ReadinessScreening Test, 432 Future considerations Gibbons, J. S. 1990 Health Considerationsfor the random assignment studyof suicide Nation findings, 16 attempters, 402 Year 2600 HealthObjectives for the Gleser, G. C. Nation, 16 effect on family ofexposure to disaster, 232 Garland, A. Goals and objectives,forming school-based suicideprevention collaborative, importance,114 program, 392-94 distinction, 113 Garmezy, N. elements, 113 importance of teachers to stressed prevention programobjective example, children, 233 113 limitations of studieson children and Goldacre, M. disaster, 226 suicide survey, 376 Gender Gordon, R. S. alcohol and drug abuserole, 315-16 prevention categorikm, 24,67-69 conduct disorder riskfactor, 275-76 public health preventioncategories and method, suicideapplication, 375-76 comparison, 68-69 Gene-environment correlationsof risk Gould, M. S. alcoholism application, 165 imitation and suicidestudy, 384 depression application, 163 suicide high-riskgroups study, 385, 386 environment partially shaped by genes, televisionprograms targeted at youth 163 suicide, 396 4 INDEX 465

Greer, S. identifying high-risk groups, 295 suicide attempters and followup preschool, 294-95 psychiatric treatment, 401 Holding, T. A. Grossman, M. suicide prevention television series, 395 alcohol use study, 320, 322 Hotlines Guilt, postdisaster psychiatricmorbidity advantages, 396 factor, 231 befriending process, 396 crisis intervention rationale, 395-96 Hagin, R. A. criteria, 396 implementation of learning dik rder effectiveness, 397 prevention programs, 435-n low impact, reasons, 397-99 neuropsychological deficits, reading staffmg, 396-97, 398 disability, 418 training, 398 SEARCH & TEACH program, 435-36 Human accountability, postdisaster Hainsworth, P. K. psychiatric morbidity factor, 230 learning disorder scanning test, 419 Hyperactivity Handicapped children conduct disorder overlap, 280 FAucation for All Handicapped Children parental alcoholism role, 165 Act, 423-24 remediation application, 44 identification and education, legal mandate, 423 individual educational plan (IEP), 423 Iatrogenic effects redefinition application, 44-45 in children, 69 Hard-to-reach populations and attrition from intervention, 66-67 85-86 minimizing, 09 Hawton, K. Ideation, suicidal suicide survey, 376 and behavior, 377-78 Head Start Program Paykel study, 377 early childhood development model, 26 IEP. See Individual educational plan Westinghouse Learning Corporation Ilg, F. L. study, 138, 426 developmental placement study, 431-32 Health care providers Imitation and suicide studies, 384 awareness of drug problemneeded, 11 Implementation success factors labeling, fear of, 332 interaction, 114-15 negative effects of overpreecribing drugs, local initiative, 116 349 personal factor, 116 role in alcohol and drug abuse practitioner involvement, 115 prevention, 349 Stolz study conclusions, 116 role models, 349 summary table, 116 Health promotion programs, 344-45 trial run, 115 Heart disease community prevention In-loop assessment, environmental programs, 344-46 scanning, 1U7 High-risk programs, conduct disorder Income, alcohol and drug abuse role, 317 preventzon India, suicide rates and methods, 376 Divorce Adjustment Project, 295 466 PREVENTION OF MENTAL DISORDERS

Indicated interventions Kindergarten identification of learning definition, 68, 74 disorders screening procedures, 74-75 content considerations, 416-20 Indirect coats, definition, 127 scanning, definition, 415 Individual code statistical considerations, 415 contribution to family interactions, 38 King, G. D. effect on other codes, 38 hotlines study, 398 Individual educational plan (IEP), 423 Kinzie, J. D. Influenza vaccine, cost-benefit analysis, Cambodian refugee study, 228, 232 128 Klein, Rudolf Insecure attachment, 161, 169, 171. See policymaking in a democracy, 137 also Separation anxiety Knesper, D. J. Institute of Medicine (I0M) noncompliance of suicide attempters, 400 Board on Mental Health and BehavioralKogan, L. S. Medicine, research fimding report, 144 noncompliance of suicide attempters, 400 Koplan, J. P. low-birth-weight infante incidence reduction, 139 measles and rubella vaccination cost- reeearch funding, 144, 145 benefit analysis, 127 IOM. See Institute of Medicine Kreitman, N. Isle of Wight study imitation and suicide study, 384 conduct disorder application, 275, 278, suicide repetition rates, 402-3 281, 284 Kumpfer, K. L. prevalence survey, 211 Strengthening Families Program, 341-43 Jansky, J. VASC theory, 319, 320, 328 screening test, learning disorders, 419 Jessop, D. J. Labeling home care study, 212, 213-14 conduct disorder application, 286 Johnston, L. D. health professionals' fear of, 332 National High School Senior and CollegeLarge family size, conduct disorder Survey, 316 application, 283-84 Lead poisoning, screening, 129 Kandel, D. B. Learning disorders drug use model, 322 administrative strategies for prevention, 430-33 Kennedy, P. biological factors, 413 suicide attempters and followup psychiatric treatment, 401 conduct disorder applic%tion, 281-82 Kerner Commission, 142 educational interventions, outcomes, 420-36, 421 Key informant interviews, 110 identification in kindergarten, 416-20 Kidnapping, postdisaster psychiatric morbidity example, 228 overview, 436-37 Ki lby, G. A. primary prevention, 414 junior first grade study, 431 psychological factois, 413 psychoneurological defects, 471 identification, 414-15 INDEX 467

Learning disorders (continued) intervening variables, 231-34 secondary prevention, 414-15 situational 'variables, 230-31 419 socioeconomic-environmental factors, Meeting School Screening Tests, disorder; 413 Mental illness. See Psychiatric suicide risk factor, 383 Psychiatric morbidity summary, 451 Methods, suicide, 382 Lieberman, R. Microregulations, 39 disorder random assignment study ofsuicide Milestone programs, conduct attempters, 402 prevention reduction, 294 Iltran, R. E. motor vehicle accident suicide intervention, 395-96 parenting skills, improving, 294 Litt, I. F. Perry Preschool example, 294 noncompliance of suicide attempters,399 problem-solving training, 294 Loeber, R. Miller, H. L. delinquency-prone adolescents study,76 suicide rates study, 397-98 multiple-gating screening procedure,75 Miniregulations, 39 Low-birth-weight infants, policymalting Minnesota Multiphasic Personality applications, 139 Inventory, 212 Minuchin, S. Macdonald, D. I. family therapy model, 207-8 health care praviders awarenessneeded,Monitoring programs 11 shaping by staff, 81 Macroregulations, 39, 98 staff attitude, 81-82 Mainstreaming readiness study, 431 volunteers, use of, 82 of Malnutrition Monitoring the Future, national study prevention paradigm, 131 high school seniors, 314 remediation application, 43-44 Montre.+1 Social Worker study, 207 Manifest Anxiety Scale, 195 Moskowitz, J. M. alcohol use, review of risk factorsand Marijuana prevention programs, 320, 321 effect on fetus, 166 cigarette smoking, 337, 339 high school use study, 316 324 community environmental prevention peer goup effect on use, programs review, 346 Project Smart, 338 Motor vehicle accidents Maturational lag, learning d'Asorders, reduction, conduct disorder application, 416-18 294 May, D. C. Motto, J. A. developmental placement study, 432-33 suicide attempters and followup Measles psychiatric treatment, 401 cost-benefit analysis, 127 Mt. St. Helens volcanic eruption,227, 228, health effects, 127 229 Mediating factors, postdisaster psychiatricMulticomponent alcohol and drug abuse morbidity role prevention programs, 344 child intrinsic, 234-36 Multiple adversities, postdisaster impact on development, 236-36 psychiatric morbidity factor, 231 4 72 468 PREVENTION OF MENTALDISORDERS

Multiple gating, screeningprocedure, 76 Naturalistic treatment studies,suicide Murrell, S. A. attempters, 400-1 environmental scanning criteria,107-8 Neurochemical vulnerabilities,alcohol Muscular dystrophy and drug abuse, 330 psychosocial dystlinction example,212 Neurological vulnerabilities,alcohol and Myths drug abuse, 330 definition, 37 . ropsychological vulnerabilities, developmental problems, 37 alcohol and drug abuse,330-31 regulatory function, 37 News media role duringdisaster, 238 NIAAA. See NationalInstitute on Alcohol Nader, K Abuse and Alcoholism NICHD. See NationalInstitute of Child separation anxiety afterdisaster, 228 Health and Development Napa Project 336-36 NIDA. See National Institutefor Drug National AdolescentSchool Health Abuse Survey, 8, 9 NIH. See National Institutesof Health National Adoleecent Suicide Hotline, 396 NIMH, See NationalInstitute of Mental National High SchoolSenior Survey, 7, 0, Health 316 1974 Disaster ReliefAct, 226 National Household Survey,8 1990 Health Considerationsfor the National Instituteon Alcohol Abuse ard Nation, 16 Alcoholism (NIAAA), 143,145 NJCLD. See National JointCommittee on National Institute for DrugAbuse (NIDA), Learning Disabilities 143 Nolan, T. National Institute of Child Health and research on emotional Development (NICHD) correlates and consequences of birth defects, 194-204 research fundingsource, 142, 145 Noncompliance studies, suicide National Institute ofMental Health WWI) attempters, 399-400 Noninfectious diseases, 132 research fundingsource, 143, 145, 146, 148 Nonnormative event, 48 Normative event, 48 National Institutes ofHealth (NIH) North American Indians Biomedical Research andDevelopment Index, 145 alcohol and druguse, 317 budget, 144, 145, 146, 147 conduct disorder application,277 history, 142-43 suicide rates, 382 North Karelia Project, National Joint Committeeon Learning 345 Disabilities (NJCLD) areas of need, 422-23 OCHS. See OntarioChild Health Study early identificationprocedures defined, Office for SuhstanceAbuse Prevention 422 (OSAP) importance of preechoolyears, 422 high-risk youth targeted,311 National Survey in theUnited Kingdom, Office of TechnologyAssessment 195 child memal healthservices study, 146 National Youth RunawaySwitchboard, Onset 396 definition, 62

4 t3 INDEX 469

developmental process example,131 Onset (continued) malnutrition example, 131 pathogen identification, 63 noninfectious diseases example,132 prevention categories applications,63-65 131 prevention issue, 63 vaccine model limitations, smallpox Ontario Child Health, Study(OCHS) World Health Organization vaccine example, 130-31 conduct disorder apphcation,275-76, 277, 278, 283, 284, 288 Parasuicide, 375 individual Ontario Child Health Study,195 Parental mental illness, responses Opioids competence qual;ties, 173 effect on fetus, 166 gender of child, 173-74 Organizational focusing protective factors, 174-75 defining goals, 111 resilience and vulnerability, 172 forming goals and objectives,113-14 temperamental factors, 174 prevention program objectiveexample, Parental psychiatric disorder 113 impact on, problem statement, 112-13 early childhood development, problem workup example, 112 38 environmental mechanisms, 165-72 proximal and distal outcomes,111 genetic mechanisms, 158-65 Organizational readiness individual differences in response, adaptiveness, characteristics, 104 172-75 assessing, examplee, 104-5 risk of exposure to symptoms,167 of staff, attitudes, beliefs, and practices statistical considerations, 157-58 104 summary, 447-48 awareness and acceptanceof prm, 103 Parenting D'Aunno model, 103 conduct disorder application, 282 existing structures and services,104 impaired, as risk factor, 169-70 disorder support from environment,103 skills improvement, conduct application, 294 Organizational einvention Pathogenesis, transactional peripective, and overview, 97-99 process description 62 Orientation Patoerson, G. R. entrepreneurial, 102 aggression and faulty discipline,328 to external relationships,101. alcohol and drug use by parents,326 goal, 101 antisocial behavior model, 30-31 participative, 102 Paykel, E. S. Abuse OSAP, See Office for Substarwe suicidal ideation study, 377 Prevention Peer groups alcohol and drug abuse correlation, Paradigms, family 323.24 definition, 37 conduct disorder application, 284,286 prevention implications, 37 mariju ma use, effect on, 324 typology, 37 post& a ter influence on peer Paradigms, prevention relationlhips, 233 bonding example, 131-32

A 470 PREVENTION OF MENTALDISORDERS

Pellagra incidence, prevention application, reeearch irrelevance to politicalprocess, 125-26 138 Perry Preschool Project, 86, 294 summary, 446-47 Personality tax policy applications, 140 alcohol and drug abuse role welfare policies, 141 attitudes and values, 327 Poliomyelitis basic trsits, 327 historical perspective, 1 stressors and coping responses, 327-28 Sister Kenny's work, 1 conduct disorder risk factor Politicians' attitude to policymaking, hyperactivity overlap, 280 137-38 predicting, from infancy, 279 Population density, alcohol anddrug research problem I, 279, 280 abuse, 319 Pertussis vaccine, side effects of Postdisaster morbidity prevention program, 134 coping responses, 234-35 Phillips, D. couree of, 229 imiiation and suicide study, 384 Poettraumatic stress disorder (PTSD), PBS. See Public Health Service 226-29 Physical environment, conductdisorder Precipitants, suicide, 383 application, 286-86 Preconceptual and prenatalcare Physical risk mechanism, 169 controversy in policymaking, 139 Piers/Harris Self-Conceptmeasure, 195 Predictive Screening Index, 419 Pleas, I. B. Pregnancy complications peychosocial dysfkinction and disease effect on fetus, 166-67 severity relationship, 211 Presubetance abuse syndrome,=Or research on emotional correlatesand risks, 322 consequences of birth defects, 194-204 Preventing specific disorderspapers, stress model, 205 summary, 449-51 Plomin, R. Prevention, conduct disorder Colorado Adoption Study, 163 communitywide programs, 295-96 Pneumovax, cost-benefit analysis,128 high-risk programs, 294-95 Policymaking identifying risk factors, 289,291-93 academics' attitude, 137-38 milestone programs, 294 alcohol and drug abuse, funding overview, 296-97 commitment, 347 types, definitions, 290 availability of health:we for the poor, Prevention categories 139-40 distinctions between, 24 day care issue, 140-41 overlapping, 64 in a democracy, 137 prir ary, 24, 60-61, 63-65, 873,379-80, Kerner Commission conclusions,142 388-94 pluralism of American society,13647 secondary, 24, 60, 63-65,373, 380, politicians' attitude, 137-38 394-403 preeonceptual and prenatalcare summary, 444-45 controversy, 139 tertiary, 24, 59-60, 374, 394-403 preventing educational failure,138-39 Prevention perspectives e integrating interventions,14-16 471 INDEX

Preventive intervention Prevention perspectives(continued) chronic illness public health, 12-13 artificial network of support,205 Prevention programs disorders abuse central nervous system alcohol and drug exception, 208 community-based, 339-41 conventional psychotherapyresults, environmental, 345-46 207 family-focused, 341-44 dearth of child psychiatristsin the field, health promotion,344-45 207 multicomponent, 344 directed at childrenexclusively, 207 ochool-based, 333 etiologic fraction, 205 before-and-after comparisons,387-88 individual support, 205,206 conduct disorder Minuchin's family therapymodel, conununitywide programs,295-96 207-8 high-riet programs,294-95 preventive model, 204 milestorw programs,294 recommendations, 208-9 modifiable variables,291-93 stress model, 205 consideration, 208 early intervention,26-27 therapist expertise evaluating, 387 disaster strategies, 236-41 goal, 26 focus, 64 health care providersrole, 349 general strategies, 41-47 high-risk childdemonstration programs, implementation 332 as continuous process,105 labeling concern, 332 critical resources, 101 learning disorders environmental scanningand initial communication skills focus,433-34 linkages, 106-8 cost-effectiveness study,433-34 facilitating factors, 114-16 design rationale, 433 fidelity versus adaptation,116-18 direct teaching ofskills, 435 monitoring, 118-20 quality of implementation,434-36 organizational focusing, 111-14 sequencing and timing ofeducational organizational readiness,103-5 activities, 434 role attributes of innovator,100-2 multidisciplinary approach,13 soft technology, 100 recommended directions,alcohol and steps, 100 drug abuse, 348-49 summary, 446 suicide team enterprisecharacteristic, 105 388-89 general psychiatric care, lawns, 1-2 to methods, 389-90 restrictingaccess outcomes, 59-66 school-based, 390-94 parental psychiatricdisorder, risk factor, television application,394-95 158 Prevention research recipients, selection planning, 73 base rate, 66 process, stages,99 Gordon's categories, 67-69 Preventive applications,definition, 77 476 472 PREVENTION OFMENTAL DISORDERS

Preventive intervention (continued) Prqject Prevention, 17 iatrogenic effects, 6647 Project Smart, 338 screening procedures, 74-75 Proof, concept of successffil programs, summary, 445 importance to prevention summary of strategies, 446 research, 89 pretest and posttest studies,limitations, timing 88 functions, description,76 transactional model,89-90 transactional model, 76 Proximal outcome Preventive model definition, 65 chronic illness application, 204 organizational focusing Preventive trials application, 111 risk level application,66 conduct requirements,78 Psychiatric disorder definition, 58 historical view, 443 demands of settings, 77 Psychiatric Evaluation focus,17 Form, 235 Psychiatric morbidityin children, mexker point questions,79 postdisaster types maturational factors, 78 Cambodian adolescentrefugees example, positiveresponse, 79 228 Price, R. H. comorbidity, 228 prevention-researchprocess, 99 depression, 228 Primary prevention description, 227-28 definition, 24 kidnapping example,228 onset issue, 61, 63 Mt. St. Helensvolcanic eruption prevalence reduction,61 example, 228 suicide, 373, 379, 388-94 posttraumatic stressdisorder, 227-29 Problem perspectives,alcohol and drug school absenceexample, 229 abuse, 7-10 separation anxiety,228 Problem-solving training Psychiatric research,public policy conduct disorderapplication, 288.89,294 Alcohol, Drug Abuse,and Mental Health Problem statement Administration, 142 elements, 112 budget, 143-48 workup example, 112 Federal Governmentas funding source, Process studies 142-43 definition, 79 Psychiatric ScreeningInventory of Langner, 211 program elements, identifyingeffective, 79-81 Psychiatry, child andadolescent historical perspective, Program elements,identifying effective 2 documentation of procedures, Psychinfo computersearch, learning 80 disorders, 420 optimal combination,80-81 Psychological autopsy, random assignment,80 380 Program monitoring Psychological crisisteams, 240 data sources, 119 Psychological first aid,238-39 data types, 118 Psychoneurological defects,identification, 414-15 implementation principles,119-20 4 77 INDEX 473

chronic illness Race Psychoeocial dysfunction, 316-17 risk factors alcohol and drug abuse role, biological, 210-15 conduct disorder role, 276-77 demographic, 210 suicide rates, 382 environmental, 215-16 Ramey, C. prevention problems, 209 learning disorders study,427-28 Rutter study, 209 Rand Corporation Report onStrategies Drug Use, summary table,210 for Controlling Adolescent 347 Psychosocial environmentalrisk factors, alcohol and drug abuse Re-ding disability maturational lag, 416-18 community environment, 322-23 neurowchological patterns, 416 family, 324-26 time personality and behavior, 327-28 position in space and order in ftmctions, 418 prevention model, 319-22 predictive tests, 419, 420 school and peers, 323-24 predictive disorder sensory-perceptual factors, Psychosocial risk factors, conduct value, 417 community and socioeconomic,284-86 Redefinition family, 282-84 handicapped children application,44-45 disorder PTSD. See Posttraumatic stress strategy definition, 46 Public health temperamental problem application,45 goal of prevention, 12 Reeducation prevention needs, 12 adolescent parents application,46-47 categories, 59-62 preventive intervention definition, 46 12-13 risk factors, key questions, Regional differences, alcoholand drug selected interventions, 68 abuse, 318-19 Public Health Service (PHS)Model of Regulatory processes Prevention and cultural code, 39, 40 agent/envirortment and agent/hostrisk and family code, 40-41 factors, 319-21 functions, 39 approaches to alcoholism and drug transactional model, 42-43 abuse, 333-46 environment/hoet risk factors, 321-22 Regulatory systems Public policy. See also Paradig no, preven- cultural code, 33, 34 tion; Policymaking; Psychiatric environtype, 32, 33 research, public policy family code, 35-38 public health programs, examples, genetic, 31-32 125-26 individual code, 38 Pynoos, R. S. prevention implications, 41 separation anxiety after disuster,228 preventive intervention strategies,41-47 Remediation Quinton, D. definition, 43 parental symptoms, risk of exposureto, hyperactivity application, 44 167-68 malnutrition application, 43-44 478, 474 PREVENTION OF MENTALDISORDERS

Research designs multiple problem behaviors,10 cohort studies, 195, 202-3 positive, 71 with controls, 195,198-200 psychoeochd dysilinction,209-15 prevalence eurveys, 195, 201 Sameroff study, conclusions, uncontrolled or norm-referenvd, 28, 41 194, 11011-7 screening instruments, 310 transactional model, 70, Research issues, riskfactors 71, 72 types and examples, 70 causal models, lack of, 312 Rituals drug abuse, lack ofspecificity of risk factors, 314 definition, 35 limiting factors, majorvariables, 311-12 developmental progression,35 Robertson, L. S. longitudinal data, lackof, 314-15 phases of researchdescribed, 315-31 driver education study,136 research agenda, 2 Robins, L. N. risk factor correlations,3, 4 Epidemiologic CatchmentArea (ECA) shift in emphasis, 3 Study, 163, 315, 316 specificity of risk factors, Rochester study, 207 313 Role attributes, statistical analysisprocedures, lack successful prevention 312 innovator we and abuse, 313 cultural sensitivity,102 Resilient children, 286 entrepreneurial orientation,102 goal orientation, 101 rtesource sensitivity,102 orientation to externalrelationships, lisk-correlated attrition, 85 101-2 Risk determination strategies, suicide participative orientation, adolescents, followup 102 studies, 381 resource sensitivity, 102 psychological autopsy, 380 Roes, C. P. types of studies, 380 school-based suicide Risk factor research. prevention See Research issues, program, 392 risk factors Rubella Risk factors. See also Alcohol and drug cost-benefit analysis, abuee; Conduct dieorder;Disaster, 127-28 psychiatric morbidityin children; side effects to preventionprogram, 133-34 Learning disorders;Suicide case-control study, 71-72 vaccine program, monitoringcouipliance, 127-28 chronic illness, 191-216 combining, for selected Rutgers Health andHuman Development interventions, 74 Project, 327 early childhood development, 27-28 Rutter, M. elements, 70 Isle of Wight prevalencesurvey, 211 environmental, 27, 160-62,165-75 parental symptoms, riskof exposure to, features relative toprevention, 373 167-68 genetic mechanisms,157-65 prevention myths, 23 high-risk childrenas primary target, 310-11 psychosocial dysfunctionstudy, 209 identifying, 69-70, 71 Sameroff, A. J. importance of earlywarning signs, 309 risk factors study,28, 41

4 71 INDEX 475

Sameroff, A.J.(continued) pootdisaster reactions application, 233 transactional model, 28-29 psychological crisis teams, 240 Sample size streesor and coping resources, 324 attrition, 85-86 suicide prevention programs determining, 83-84 description, 891-92 developmental continua, 84-85 evaluation, 392-94 Satterwhite, B. B. Fairfax County Youth Suicide Preven- psychosocial dysfunction and disease tion Center, 392 severity relationship, 211 goals, 390-91 stress model, 205 Shaffer and Garland study, 392-94 Sat; P. Screening procedures reading disability study, 416-17 intervention recipients, selection, 74-75 sensory-perceptual factors, predictive multiple-gating, 75 value, 417 suggestive evidence, 88 "Say no" programs, 338, 339 transactional model, 75 Schizophrenia SEARCH & TEACH learning disorders adopted-away children, 162 program, 435-36 competence in chit dren of schizophrenics, Secondary prevention 173 advantages, 63-64 gene-environment correlation, 163-64 definition, 24 genetic risk factor, 158, 159, 160, 161, 164 prevalence reduction, 60 high levels of criticism (high EE) suicide, 373, 380, 394-403 correlation, 164 Secular trends, suicide, 377, 382 impaired parenting risk, 169-70 Selected interventions pregnancy complications applicatIon, applications, 73-74 166-67 combining risk factors, 74 Schneidman, E. S. definition, 67-68, 74 suicide intervention, 395-96 exampleg, 73, 74 Schools goal, 68 absence, postdisaster, 229 public health examples, 68 alcohol and drug abuse prevention Self-destructive behavior affective and interpersonal education suicide application, 374-76 programs, 335-36 Separation anxiety alternative programs, 336-37 during disaster, 238 behavioral prevention programs, 337-39 postdisaster, 228, 230-31 cognitive programs, 333-36 Serwer, B. association of early problems and alcohol, direct teaching of skills program, 436 drug and mental disorders, 14-16 Shaffer, D. conduct disorder application, 265 imitation and suicide study, 884 conduct disorder prevention, 295-96 school-based suicide prevention environmental risk factor, 323-24 program, 392-94 importance in preventive intervention, television programs targeted at youth 15-16 suicide, 396 IVO 476 PREVENTION OF MENTALDISORDERS

Shore, J. Stolz, S. B. dose-response exposure to disaster fidelity and adaptationcontroversy, pattern, 227-28 116-17 psychiatric disorder prevalenceafter Mt, success factors, program implementation St. Helens eruption, 229 study, 115 Sickle cell anemia Stories side effects of prevention,example, 134-36 act and content distinction, 36 definition, 36 Side effects of prevention developmental component, 36 driver education example,135-36 regulatory function, 36 drug use, attempts to control,example, 134 Streissguth, A. P. pertussis vaccine example, alcohol-related behavior effectsstudy, 134 165-66 rubella example, 133-34 sickle cell anemia example, Strengthening Families Program,334, 342 134-35 Stress model suicide prevention example,135 Silver, A. A. chronic illness application,205, 200 implementation of learning Subgroup, targeting, 192 disorder Sudak, H. S. prevention programs, 435-36 hotline service in Cleveland, neuropsycholoecal deficits,reading 399 disability, 418 Suicidal behavior SEARCH & TEACHprogram, 435-36 classifications, 375 Sinqueland, M. L. Suicide learning disorder scanningtest, 419 age and method differences, :176-77 Sister Kenny attempts and deaths, similarities,378-79 poliomyelitis therapy, 1 biological risk factors,403-4 Slem, C. M. early signs, sensitivityof, 374 hotline study, 397, 398 early warning signs,374, 391 Slinger land Grade 1-6Screening Test family history, 385 for Identification ofChildren with gender and methoddifferences, 375-76 Learning Disability, 419 Hawton and Goldacresurvey, 376 Smallpox vaccination, 130 high-risk groups, 385, 136,387 Socioeconomic status hotlines and crisis services,395-99 alcohol and drug abuse role, 317-19 ideation and behavior,377-78 Soft technology, 100 ideators, attempters, andcompleters, Sparrow, S. differences, 375 reading disability study,416-17 imitation as facilitator,3b4 Sum of disorderseffect., 65 impaired parenting risk,170 Stanford Three CommunityStudy, 346 lack of specificity inrisk factors, 374 Statistical considerations,power of mental health problems,associated, experimental procedures,82-88 383-84 Statutes, 33, 35, 39 methodological difficulties,135 Stein, R. E. K. methods, 382-83 hcme care study, 212,213-14 parasuicide, 375 4 Ni precipitants, 383 INDEX 477

Suicide (continued) Transactional model prediction, overview, 404-6 archival records, 75-76 prevention programs, evaluating,387-88 biological transactions, 31-32 primary prevention, 379, 388-94 and child development, 28-32 reeearch, rarity, 374 concept of proof in preventive risk determination strategies, 380-81 89-90 defining purpose of reeearch, 73 screening difficulties, 136 secondary prevention, 380, 394-403 definition, 56 intervention recipient selection, 72 secular trends, 377, 382 prevention categories application', 59-62 self-destructive behavior, 374-75 prevention implications, 29, 31 side effects of prevention, example, 135 regulatory processes, 42-43 suicidal behavior, 376 risk factor researth application,72-73 summary, 450-51 risk souroes, 57, 70, 71 tertiary prevention, 394-403 screening procedure application, 75-76 treating attempters, 399-403 timing of intervention, 76 young suicides, descriptive characteristics, 381-82 Transactional paradigm. See Transaction- Suicide attempters, treating al model Treatment-correlated attrition, 85 experimental studies, 402 naturalistic treatment studies, 400-1 Tubercu.osis environmental and biological causes, 25 noncompliance problem, 399-400 treatment studies problems, 402-3 role of social factors, 26 Twin and adoption studies, conduct Summaries of chapters, 444-61 disorder, 280

Tax policy applications in policymaking, Universal interventions, definition and 140 examples, 67 Teen Alcohol and Drug Schools, 343 U.S. National Drinking Practices Survey, Television 315 "The Befrienders" suicide prevention series, 396 Vaccine model conduct disorder application, 285 bonding, 131-32 strtide prevention, 395-96 developmental process, 131 Temperamental problems, redefinition application, 46 limitations, 130-32 Temperamental vulnerabilities, alcohol malnutrition, 131 and drug abuse, 330 noninfectious diseases, 12 Terr, L. smallpox, 130 followup study of kidnapped children, Validation studies, 87 228 Value/Attitudes, Stressors, Coping Tertiary prevention Responses (VASC), theory of alcohol definition, 24, 69-60 and drug abuse, 319, 320 long-term consequences of disorder, 69 VASC. See Value/Attitudes, Stressors, suicide, 374, 394-403 Coping Responses 4 Q 2 478 PREVENTION OF MENTALDISORDERS

Weissman, C. S. Wood, C. readinees for mainstreaming study, 431 developmental placembt study,432 Welch, E. Workshop planning developmental placement study, 432-33 collecting data, 107 Welfare policies, 141 criteria for initial scanning,107 We lu, T. C. outcome, 107 suicide attempters and followup rationale, 106 psychiatric treatment, 401 wnrksheet, 108 West, D. J. World Health Organization(WHO) conduct disorder study, 277, 279,281, 283, 286 smallpox vaccination campaign,130-31 West Indians, conduct disorder application, 276 Yamamoto, M. D. Westinghouse Learning Corporation noncompliance of suicideattempters, 399, 400 Head Start Program study, 138,426 White, G. C. Year 2000 Health Objectivesfor the Nation, 16 measles and rubella vaccinationcost- benefit analysis, 127 Zenski, J. P. WHO. See World HealthOrganization Wilson, W. J. pre-first grade transition classstudy, 430-31 relationship between employmentand marriage, 141-42

* U.S. GOVERNMENT PRINTINGOFFICE : 1990- 277- 021 : QL 3

4 3 OSAP

Reprinted by the Office forSubstance Abuse Prevention and distributed by the National Clearinghouse for Alcoholand Drug Information P.O. Box 2345 Rockville, MD 20852 4 b41