Special Theme – Mental Health Psychiatric epidemiology: selected recent advances and future directions Ronald C. Kessler1

Reviewed in this article are selected recent advances and future challenges for psychiatric epidemiology. Major advances in descriptive psychiatric epidemiology in recent years include the development of reliable and valid fully structured diagnostic interviews, the implementation of parallel cross-national surveys of the prevalences and correlates of mental disorders, and the initiation of research in clinical epidemiology. Remaining challenges include the refinement of diagnostic categories and criteria, recognition and evaluation of systematic underreporting bias in surveys of mental disorders, creation and use of accurate assessment tools for studying disorders of children, adolescents, the elderly, and people in less developed countries, and setting up systems to carry out small area estimations for needs assessment and programme planning. Advances in analytical and experimental epidemiology have been more modest. A major challenge is for psychiatric epidemiologists to increase the relevance of their analytical research to their colleagues in preventative psychiatry as well as to social policy analysts. Another challenge is to develop interventions aimed at increasing the proportion of people with mental disorders who receive treatment. Despite encouraging advances, much work still needs to be conducted before psychiatric epidemiology can realize its potential to improve the mental health of populations.

Keywords: psychiatry; epidemiology, trends; mental disorders, epidemiology; psychopathology; review literature.

Voir page 471 le re´sume´ en franc¸ais. En la pa´ gina 472 figura un resumen en espan˜ol.

Introduction psychiatric epidemiologists study risk, they tend to focus on broad nonspecific risk markers, such as Epidemiology is concerned with understanding and gender and social class, rather than on modifiable risk controlling disease epidemics by investigating em- factors, hence limiting the possibilities for interven- pirically the associations between variation in tion. However, this situation is changing as descrip- exposure to disease-causing agents external to the tive issues are being resolved, more analytical individual, variation in the resistance of individuals questions are being addressed, and preventive exposed to the disease-causing agents, and variation interventions are being implemented. in resistance resources in the environments of exposed individuals (1). These investigations are initially carried out by examining natural variations. Descriptive psychiatric epidemiology Hypotheses based on these analyses are then, usually, tested provisionally in naturalistic quasi-experimental Adult community epidemiological surveys situations with matching or statistical controls used to Descriptive psychiatric epidemiology has gone approximate the conditions of an experiment. If the through an unprecedented period of growth over hypotheses stand up to these preliminary tests, they the past twenty years. Starting with the Epidemiolo- are evaluated in interventions aimed at preventing the gic Catchment Area (ECA) study in the USA (2), large onset or altering the course of the disorders. surveys of adult mental disorders in the general Psychiatric epidemiology traditionally lags population have been carried out in numerous behind other branches of epidemiology because of countries throughout the world. An important difficulties encountered in conceptualizing and innovation of the ECA was the use of a fully measuring mental disorders. As a result, much structured research diagnostic interview known as contemporary psychiatric epidemiology continues the Diagnostic Interview Schedule (DIS, 4). Method- to be descriptive, focusing on the estimation of ological studies demonstrated that the DIS yields disorder prevalences and subtypes (2) at a time when reliable and valid diagnoses (5), a result that was very other branches of epidemiology are making progress important in promoting the ECA–DIS methodology in documenting risk factors and developing pre- in subsequent general population surveys. ventive interventions (3). To the extent that The first expansion of the ECA–DIS metho- dology was carried out by WHO in collaboration with the US Alcohol, Drug, and Mental Health Admin- 1 Correspondence should be addressed to Dr Kessler, Department of Health Care Policy, Harvard Medical School, 180 Longwood Avenue, istration to include International Statistical Classifica- Boston, MA, USA (email: [email protected]). tion of Diseases (ICD) criteria for research and to Ref. No. 00-0543 produce versions of the instrument in many different

464 # World Health Organization 2000 Bulletin of the World Health Organization, 2000, 78 (4) Psychiatric epidemiology: selected recent advances and future directions languages. The resulting instrument, the Composite than those obtained in blind clinician reinterviews International Diagnostic Interview (CIDI, 6), first (20), concerns about the high prevalence estimates became available in 1990. WHO technical support have focused largely on the underlying validity of the led to an unprecedented number of major epidemio- ICD and DSM systems. Clinical significance criteria logic surveys using the CIDI in countries as diverse as were added to nearly half the diagnoses in the DSM- Brazil (7), Canada (8), Germany (9), Mexico (10), the IV system in response to these concerns to address Netherlands (11), and Turkey (12). the perceived problem that the previous diagnostic In 1997, WHO created the International criteria led to overdiagnosis of disorder among Consortium in Psychiatric Epidemiology (ICPE) to people whose symptoms were clinically insignificant. coordinate the comparative analysis of these data However, this has led to even more controversy (13). ICPE also provides technical assistance to regarding whether the inclusion of these new criteria researchers planning new CIDI surveys. The WHO is legitimate (21). This controversy illustrates the World Mental Health 2000 (WMH2000) initiative has point made at the very beginning of this article: grown out of these technical assistance activities. problems in conceptualizing and measuring mental WMH2000 will coordinate general population CIDI illness have hampered progress in psychiatric surveys in 20 countries in the year 2000, distributed epidemiology much more than in other branches of globally in North America (Canada, USA), Latin epidemiology. America (Brazil, Colombia, Mexico, Peru), Europe Irrespective of the ultimate resolution of these (Belgium, France, Germany, Italy, the Netherlands, conceptualization and measurement issues, concerns Spain, the Ukraine), the Middle East (Israel), Africa about high prevalence have led to a new interest in the (South Africa), Asia (China, India, Japan) and the assessment of severity and impairment in psychiatric Pacific (Indonesia, New Zealand). epidemiological surveys as well as a view that Several important results have consistently dimensional assessment of mental disorders and emerged from the DIS and CIDI surveys. global assessments of ‘‘case-level psychiatric mor- . Mental disorders are among the most prevalent bidity’’ are more useful than detailed evaluations of classes of chronic diseases in the general popula- many separate ICD or DSM disorders. The new tion, with lifetime-to-date prevalences often close WMH2000 surveys include structured versions of to 50% of the population and with 12-month standard disorder-specific, dimensional clinical se- prevalences typically in the 15–25% range (2). verity measures and assessments of the functional . Mental disorders typically have much earlier ages impairments and disabilities associated with current of onset than other chronic diseases. Anxiety mental disorders with the aim of developing disorders have median ages of onset in the early to dimensional and global severity measures. Impor- late teens in most of these surveys, while mood tantly, the WMH2000 surveys will also carry out and substance use disorders have median ages of identical assessments of the functional impairments onset in the early to mid twenties (14). and disabilities associated with a representative . Mental disorders are among the most impairing of sample of physical disorders in order to provide all chronic diseases (15). comparative information. . Respondents with the most severe and disabling mental disorders in these surveys usually meet Adult clinical epidemiological surveys lifetime criteria for a number of different ICD and The techniques developed in the ECA study to carry Diagnostic and Statistical Manual of Mental out fully structured psychiatric diagnostic interviews Disorders (DSM) syndromes (16). have more recently been extended to primary care . Only a minority of the respondents in these settings. The first of these studies was the Medical surveys who meet criteria for a mental disorder Outcomes Study (MOS), an investigation of a series report that they received treatment in the of chronic diseases, including depression, that preceding year (17). The measures of disorder documented that depression is associated with levels severity included in the surveys are consistently of functional impairment in a wide range of life associated with probability of service use, like- domains that are comparable to, if not greater than, lihood that service use occurs in the specialty those found among patients with physical disorders sector, and intensity of treatment, documenting such as hypertension, diabetes, and arthritis (22). that there is some rationality both in help-seeking Subsequent clinical epidemiological surveys have and in the allocation of treatment resources. attempted to evaluate the impairments associated However, the surveys also show that only a with a broader range of mental disorders in primary minority of patients describe a course of therapy care samples. The largest and most influential of that is even minimally adequate in terms of these has been the WHO Primary Care Collaborative currently available treatment guidelines (18). Study (23). These surveys documented that mental disorders are highly prevalent among people who The high rates of disorder found in these surveys seek help from family physicians, that these disorders have led some commentators to raise questions are associated with substantial impairment in role about the plausibility of the prevalence estimates (19). functioning, and that most of these disorders go As clinical reappraisal studies clearly show that the undetected by primary care physicians. A series of prevalence estimates in CIDI surveys are not higher

Bulletin of the World Health Organization, 2000, 78 (4) 465 Special Theme – Mental Health

innovative programmes has been developed based by a good deal of emotional turbulence, much of on these surveys to help primary care doctors detect which resolves in adulthood. This means that and treat mental disorders (24). legitimate questions can be raised about the clinical One of the most important implications of significance of child and adolescent symptoms that these clinical epidemiological surveys is that un- are not currently causing a serious impairment, unless treated comorbid mental disorders might complicate a strong case can be made that these symptoms are the treatment and management of physical disorders. associated with risk of future clinically significant For example, Roose & Glassman documented that disorders. The complication is that long-term long- comorbid depression is a powerful predictor of early itudinal data are needed to confirm current diag- mortality among survivors of first heart attacks (25). noses. In comparison, it is much easier to reach Based on this finding, new interventions have been agreement on the classification of serious emotional developed to screen for and treat depression among disorders among youth in that the level of agreement cardiac patients. A number of related, but as-yet- among informants is much greater when the child’s unpublished, clinical epidemiological research initia- symptoms are severe (30). tives are currently investigating the effects of comorbid mental disorders on the onset, course, and management of other physical disorders. Challenges for descriptive psychiatric Preliminary studies suggest that at least some of epidemiology these investigations are likely to yield important practical results (26). Systematic underreporting The methodological advances surrounding develop- Child and adolescent community ment of the DIS and other fully structured measures epidemiological surveys of disorder and severity have addressed many of the The adult epidemiological survey finding that mental measurement problems that previously limited disorders have early ages of onset has promoted progress in psychiatric epidemiology, while empirical interest in the mental health of children and data produced by the new generation of surveys adolescents. Risk factor research is a good deal more initiated by the ECA study have stimulated healthy advanced in studies of child and adolescent psycho- debate about deeper conceptual issues regarding the pathology than in studies of adults, but progress in validity of the ICD and DSM classification systems. the descriptive epidemiology of child and adolescent However, formidable challenges in this area still mental disorders has been hampered by a more remain since mental disorders are highly stigmatized severe version of the same measurement difficulties conditions that many people want to keep private. that plague adult studies. Two reasons hinder studies Because of their embarrassment or fear of discrimi- of younger respondents: nation, their mental disorders can only be defined on – Childhood disorders are much less ‘‘crystallized’’ the basis of clusters of symptoms they report. than adult disorders, posing special challenges for In the case of patients seeking professional their assessment; and treatment, there is reason to believe that self-reports – It is impossible to carry out direct interviews with will be fairly complete and honest. However, this is not young children, making it necessary to rely on the case in epidemiological surveys. It is little wonder, parents and teachers as informants. then, that concerns have been raised that under- reporting is a very serious problem in surveys of this Even when children are old enough to be inter- sort (31). Consistent with this concern, methodo- viewed, questions arise about their ability to under- logical studies have shown that reports about mental stand all questions (27), making it useful to retain disorders, substance-use problems, and other embar- informant reports from parents and teachers. How- rassing issues, such as abortion, criminal behavior, and ever, there is a problem in this; parent, teacher, and homosexuality, are extremely sensitive to subtle child reports often diverge (28). This creates variations in context and mode of questioning (32). problems in knowing how to combine the different An important implication is that the preva- reports into overall prevalence estimates (29). lences of emotional problems reported in epidemio- Concerns exist that prevalence estimates are un- logical surveys should generally be considered lower- realistically high when diagnostic algorithms combine bound estimates rather than accurate reflections of parent, teacher, and child reports using an ‘‘or’’rule, the true prevalences in the population. This is true while diagnostic algorithms that require agreement even when interviews are carried out by clinicians, among raters yield prevalence estimates that are seen since methodological research has shown that some as being unrealistically low. respondents are less liable to disclose embarrassing As in surveys of adult disorders, these concerns information when they are aware that their inter- seem related more to fundamental uncertainties viewer is a mental health professional (33). This can about current diagnostic criteria than to any objective bias estimates of correlates if there is systematic data suggesting that the estimates are in error. variation in willingness to disclose symptoms as a However, the situation is more complicated in that function of a putative risk factor. Such a differential the childhood and adolescent years are characterized willingness hypothesis, for example, has been proposed as a plausible explanation for the widely

466 Bulletin of the World Health Organization, 2000, 78 (4) Psychiatric epidemiology: selected recent advances and future directions observed finding that women report higher rates of . A number of ongoing data collection systems that anxiety and depression than men (34). make use of these screening measures have been Grappling with the problem of systematic developed and implemented to screen individuals underreporting is a major challenge for the future of in need of treatment and chart aggregate trends in psychiatric epidemiology. One way of doing this is to the prevalence of unmet need. Systems of this sort build on the work of survey methodologists, who areavailableaspartofperiodichealth-risk have developed a number of strategies to increase the appraisal surveys and require management screen- accuracy of responses to embarrassing questions. It is ing surveys carried out by employers and managed also possible to study variations in responses as a health care organizations. The expense is kept to a function of question sensitivity in split ballot minimum by using one of several low-cost data experiments built into epidemiological surveys that collection methods that include: paper and pencil manipulate wording, anonymity, mode, or other self-administration (typically in mail surveys) aspects of the question answering situation in an coupled with optical scanning of responses; effort to investigate sensitivity of responses to computerized self-administration (typically in a these manipulations. Finally, it is possible to include doctor’s office); and interactive voice response standard psychometric measures of social desirabil- administration using a digitized voice to ask ity, yay-saying , or lying into epidemiological surveys questions over the telephone and a telephone and to use responses to these measures to investigate touch-tone keypad to enter responses. the possibility that risk factor associations are biased . Statistical methods are being developed to make because of their associations with these measures. All small area estimations of disorder prevalences and of these strategies need to be explored in future unmet need for services from large-scale popula- research. tion surveys (35). These methods are designed to blend the direct small area data collected in Small-area estimation ongoing screening surveys with more in-depth Descriptive epidemiological studies are often used by periodic data collected in large-scale epidemiolo- public health agencies to estimate the magnitude of gical surveys. Until now there has been no attempt untreated disorders and to study barriers to receiving to develop integrated systems to coordinate the treatment for purposes of planning future changes in collection and integration of these two types of outreach and treatment activities. However, these data, although such proposals have been planning activities are usually carried out much more made (36). The development of systems along frequently (typically on an annual basis) than these lines represents an important challenge for epidemiological surveys (typically no more than once the future of descriptive psychiatric epidemiology. a decade). Furthermore, planning decisions are usually made at a much lower level of geographical Developmental psychopathology aggregation (typically towns, health districts, or There is great interest in refining the definitions of states) than the epidemiologic surveys (typically child and adolescent mental disorders in develop- national). It is not feasible to carry out expensive mental terms. Refinement based on knowledge of the general population epidemiological surveys more behavioural, cognitive, and emotional problems frequently or at the levels of geographical aggregation associated with risk of long-term mental health where health resource allocation decisions are made. problems would represent a great advancement. An Therefore, some other approach is needed to important challenge for child psychiatric epidemiol- increase the usefulness of such surveys for resource ogists is to develop strategies that can help shortcut allocation planning purposes. As discussed below, a this process of discovery until the time when good deal of work along three lines is currently in definitive long-term prospective studies have been progress. carried out. As outlined below, there are three . A number of short fully structured measures of options available for doing this. psychopathology have been developed to screen . Collaborative comparative retrospective or follow- for clinically significant mental disorders (23). back studies can be developed to capitalize on the These instruments can be self-administered in less number of large child mental health studies carried than ten minutes and yield fairly accurate assess- out two decades or more ago. What little is known ments of overall psychopathology (i.e., the like- from such follow-back studies is that retrospective lihood that the respondent has any clinically reports are surprisingly accurate under some significant psychopathology) as well as useful circumstances (37). provisional information about differential diag- . Since some retrospective reports will likely be nosis. These characteristics make such screening accurate, retrospective case–control studies could instruments much more feasible to use in ongoing be carried out to compare demographically local data collections than the more comprehen- matched young adults with mental disorders to sive interviewer-administered instruments such as others without such disorders in an effort to DIS and CIDI that are typically used in obtain provisional information about modifiable epidemiological surveys. determinants of these disorders. Evaluations of the accuracy of recall information in retroactive

Bulletin of the World Health Organization, 2000, 78 (4) 467 Special Theme – Mental Health

studies by using baseline reports as a gold standard considered. Such difficulties pose special challenges could provide information about the bounds of for assessing the existence of psychopathology and the accuracy of these retrospective reports. especially for differential diagnosis. Furthermore, hypotheses based on retrospective case–control studies could be confirmed in later prospective studies and experimental interven- Analytical and experimental tions. psychiatric epidemiology . The third option is to launch a series of parallel short-term longitudinal and retrospective cohort Modifiable determinants of illness onset sequential studies that attempt to cover the full age Analytical epidemiology uses nonexperimental data range from early childhood to adulthood in the to generate, refine, and provisionally test causal period of only a few years. In addition to obtaining hypotheses (38). In contrast, experimental epide- overlapping prospective data, respondents in each miology tests hypotheses by evaluating the effects of cohort-specific sample (a combination of parents, interventions on the prevention or amelioration of teachers, and focal respondents that would vary disease outcomes. As noted at the beginning of this depending on the age range of the cohort) would article, analytical and experimental psychiatric epide- be asked retrospectively to recall information miology are much less well developed than most from the final waves of data collection that was other branches of epidemiology. This is partly reported contemporaneously in the baseline attributable to conceptual and measurement pro- assessments. This would allow researchers to blems, but is also true because of the nature of the evaluate the accuracy of short-term recalled causal mechanisms involved in the onset of mental reports and to use this information to evaluate disorders. These causes relate much more strongly to the accuracy of retrospective answers obtained broad measures of environmental adversity than to from baseline respondents in contiguous cohorts. the comparatively narrow and easily modifiable risk factors (e.g., diet, exercise, smoking) that increase the In all of these approaches, it is important that future risk of chronic physical illnesses such as cancer and studies are initially neutral regarding categorical heart disease. diagnostic distinctions. Dimensional measures of Important work that focuses on modifiable risk symptoms and syndromes should be favored so that factors for particular disorders is emerging in researchers will be able to discover whether thresh- analytical psychiatric epidemiology. Examples in- olds of the sort implied by current diagnostic systems clude work linking: obstetrical complications to the can be justified empirically. risk of childhood-onset schizophrenia (39), exposure to famine during childhood to antisocial personality Other special population studies disorder (40), and early-life exposure to lead to Although not highlighted above, there are also a Alzheimer disease (41). Nevertheless, despite these number of important challenges in carrying out examples, the greater complexity of environmental epidemiological studies in other special populations. etiological agents in studies of psychiatric rather than Studies of the elderly are complex due to difficulties physical disorders has led many psychiatric epide- in determining whether organic exclusion rules miologists to focus much of their analytical efforts on should be invoked to disqualify certain mental broad nonspecific risk factors. There has been a disorder diagnoses. In studies of the elderly there special interest in exposure to stressful life experi- are the additional problems of assessing both the ences, including various types of childhood adversity cognitive capacity of the person being interviewed and adult stressors, which have been consistently and the accuracy of their retrospective recall of age of linked to a wide range of child, adolescent, and adult onset and other aspects of illness course. Sample mental disorders. There has also been a great deal of selection bias is another special challenge in general interest in stress-buffering factors such as social population studies of the elderly due to associations support and active coping. of psychopathology with early mortality and institu- Since it is difficult to devise interventions that tionalization. can prevent exposure to stress, most experimental Epidemiological studies in less developed interventions aimed at preventing mental disorders countries also pose special challenges. Many such have been designed to increase access to stress- countries lack a tradition of free speech and of buffering resources either in total populations or in conducting public opinion surveys, leading to high-risk population segments. There are quite a few unwillingness to disclose information about mental promising interventions of this sort (42). Unfortu- disorders. This unwillingness usually does not take nately, however, psychiatric epidemiologists have the form of high rates of refusal to participate in generally not played central roles in these interven- epidemiological surveys, but rather of high rates of tions for the reasons outlined below. agreement accompanied by implausibly low reported . Most psychiatric epidemiologists are more inter- rates of disorders. There are also complexities ested in descriptive and broad-gauge analytical involving local interpretations that make it difficult epidemiology than in the detailed analytical to know which illness experiences are being investigations required to target and shape pre- ventive interventions. For example, while a great

468 Bulletin of the World Health Organization, 2000, 78 (4) Psychiatric epidemiology: selected recent advances and future directions

many epidemiological studies have been carried comorbid psychopathic syndromes that often in- out on the stress-buffering effects of social clude a combination of panic, generalized anxiety, support, only a few of these studies have refined depression, phobia, and substance abuse (16). These their evaluations of social support to investigate syndromes differ substantially in their ages of onset. what should be included in preventive interven- Anxiety, oppositional-defiant, and attention-deficit tions aimed at providing support to socially problems typically are the component syndromes isolated people at risk of mental disorder (43, 44). with the earliest ages of onset. Assuming that this . Community psychologists and other human cumulation of disorders is of causal significance — a service professionals who have taken the lead in hypothesis urgently in need of evaluation — efforts most mental health preventive interventions come to intervene among children and adolescents to from clinical backgrounds and rely on their clinical prevent the cumulation of multiple psychopatholo- experience to design and implement their pro- gical syndromes hold out great promise for reducing grammes. Such individuals mistakenly see little the prevalence of serious mental disorders (46). The need for input from psychiatric epidemiology. analytical investigation of patterns and determinants Psychiatric epidemiologists must expend much of this cumulation of syndromes is a critically more effort to reach out to their preventionist understudied area in developmental psychiatric colleagues in order to make this error clear. A epidemiology. major challenge for the future is for psychiatric epidemiologists to become integrated into these Genetic epidemiology efforts. In a paper on the future of psychiatric epidemiology, published in 1992, Lee Robins suggested that the There is another class of intervention programmes, greatest hope for breakthroughs in our under- much larger than the ones discussed above, which standing of the etiology of mental disorders would also represent an opportunity for epidemiological come from genetic epidemiology (47). There is no collaboration. These are the many government indication that this promise has begun to be fulfilled entitlement programmes such as public assistance in the intervening years. Linkage studies have been for the unemployed, social security for retired people, unable to identify a single specific gene or gene and aid to single mothers with dependent children marker for any major mental disorder after more that exist in most developed countries. These than ten years of active research. Once such markers programmes are, of course, much more than mental are identified, integration of psychiatric epidemiol- health preventive interventions, but they have ogy with population genetics will be valuable in a enormous implications for mental health. Although number of ways (48), but, it is not clear when this these programmes would profit from the input of will occur. psychiatric epidemiologists, this is seldom the case. Psychiatric epidemiologists have evinced great For the most part, these programmes have been interest in behavioral genetics, despite the absence of designed by economists and implemented by social gene markers for major mental disorders. Much of workers and other human service professionals who this interest focuses on twin and extended twin- generally see little reason to be concerned with mental family studies. Such studies use structural equation health issues. modelling techniques to partition variances and covariances into genetic and environmental compo- Modifiable determinants of illness course nents (49). Although convincing data have been It was noted earlier that most mental disorders have presented in these studies that common mental early ages of onset, that there are difficulties inherent disorders are heritable (50), behavioural genetic in defining an onset of the noncrystallized symptom research has been disappointing in not advancing clusters that characterize many children and adoles- far beyond this basic finding. A good part of the cents with emotional problems, and that these reason for this disappointment stems from a symptom clusters have high base rates in the general misunderstanding on the part of psychiatric epide- population of young people. The idea of focusing miologists not familiar with the methods of beha- prevention efforts on ‘‘primary’’ prevention of vioural genetics about exactly what these studies are mental disorders — that is, the prevention of first capable of showing. onset — is therefore probably unrealistic. A much In this regard, it is important to appreciate that more feasible approach is to develop early universal, the significant heritability of variables that are probably school-based, interventions aimed at pro- thought to be risk factors for mental disorders, such moting stress-buffering resources among children as IQ and personality, means that there will be that are then supplemented with a series of more induced associations between genes and mental targeted secondary preventive interventions aimed at disorders even if there are no direct genetic effects preventing crystallization of nascent emotional on mental disorders over and above the mediating problems (45). effects of environmental pathways. A critical im- One observation of potentially great impor- plication of this is that documentation of heritability tance in this regard is that the vast majority of adults does not mean that environmental interventions will with serious mental disorders experience a series of be ineffective. This being the case, psychiatric

Bulletin of the World Health Organization, 2000, 78 (4) 469 Special Theme – Mental Health

epidemiologists are increasingly coming to realize Challenges for analytical and that behavioural genetic methods are actually much more useful in studying environmental than genetic experimental epidemiology influences. Workers in other areas of epidemiology Integration with prevention efforts had this insight a good deal earlier than those in and social policy analysis psychiatric epidemiology (51). Nevertheless, analy- A current challenge is bridging the gap that currently tical methods designed to control genetic fixed exists between analytical epidemiology and preven- effects in order to identify environmental effects tion research. Psychiatric epidemiologists also need more clearly, such as the discordant MZ twin design to become involved in more large-scale social policy (which is a comparison, using a matched pairs design research interventions. For example, recent federal of monozygotic twins who differ on an outcome in welfare reform legislation in the USA has led to a order to study the effects of environmental influ- series of state-level natural experiments moving ences while controlling for genetic influences), welfare mothers into the labour force. Early evalua- remain very underused in psychiatric epidemiology. tions of these experiments by economists clearly show that the previously neglected high rates of Barriers to help-seeking mental disorders found among welfare recipients are Retrospective epidemiological studies of speed of major impediments to successful transitions into the initial treatment contact show that it often takes many labour force (55). This observation had stimulated years for people with anxiety, mood, or substance- debate about the importance of providing mental use disorders to seek professional help after first health services as a central part of welfare-to-work onset of their disorder (52). Furthermore, more transition programmes. Unfortunately, this debate recent studies of service use show that only a minority has had no empirical or conceptual input from of people with a recent mental disorder obtain psychiatric epidemiologists. It is critical that psychia- treatment (17). These are disturbing results, espe- tric epidemiology becomes more central to this and cially in the light of clear evidence that treatments for other emerging social welfare and entitlement most common mental disorders are both safe and programme reform debates and interventions. effective. Studies of the determinants of help-seeking in The importance of secondary prevention the USA show that financial barriers are important Most theorizing and research on mental health impediments to treatment and that treatment rates prevention continues to place either a universal or increase substantially when these barriers are high-risk focus on primary prevention. Focused removed (53). At the same time, a recent compara- secondary preventions overlain on basic universal tive study of help-seeking in the USA and Canada interventions make much more sense for two found that the same low proportion of people with reasons. First, it is increasingly recognised that the mental disorders seek treatment in the two countries prodromes of many mental disorders start at such an even though Canadians enjoy free access to mental early age that it is very difficult to envision a broad- health treatment while people in the USA do not based programme that could prevent their occur- (54). An investigation of reasons for not seeking rence. Second, the complexities of interventions treatment carried out in this comparative study needed to prevent the progression of mental found that the typical mentally ill person not in disorders from early manifestations to more serious treatment reported a number of reasons for not and chronic cases are so great that it is necessary to seeking help, including perceived lack of efficacy of focus delivery of these interventions in high-risk treatment, believing that the problem will eventually segments of the population. Analytical psychiatric go away by itself, and the feeling that he/she wants epidemiologists need to reorient their research in to handle the problem himself/herself, without ways that will facilitate the development and testing outside help. of hypotheses focused on secondary preventions of These and related findings in other epidemio- this type. logical studies of the help-seeking process strongly One very important and currently neglected suggest that misunderstandings about the nature of research paradigm for this purpose is the naturalistic mental illness and perceived stigma continue to longitudinal study of the determinants of illness interfere with the help-seeking process. Public course in cohorts of current patients. There have education campaigns have been launched in some been several very influential large-scale longitudinal countries to address these problems, but these efforts naturalistic studies investigating illness course in are too recent to have developed a solid knowledge representative adult patient samples (56). Never- base regarding the effectiveness of communication theless, these have been clinical rather than epide- messages or channels or to have tested emerging miological studies. Analytical investigations of the hypotheses about other effective outreach possibi- predictors of these outcomes are needed to support lities. However, this is likely to be an area of the development of principled adult secondary considerable growth over the next decade. preventive interventions.

470 Bulletin of the World Health Organization, 2000, 78 (4) Psychiatric epidemiology: selected recent advances and future directions

Understanding the determinants ment as the end result of a modifiable process of of help-seeking sorting through a hierarchy of coping strategies in The problem of unmet need for treatment is much which treatment ranks rather low on the preference more severe for mental health disorders than for hierarchy of many people. disorders in most other areas of medicine. Standard conceptual models for studying the help-seeking process highlight the importance of health beliefs, Conclusions including perceived need for treatment, perceived efficacy of treatment, barriers to seeking treatment, A number of encouraging advances have occurred in and facilitating factors (57). These models have been psychiatric epidemiology over the past twenty years. useful in understanding and modifying the help- However, uncertainty regarding diagnostic categories seeking process in many different areas of medicine. and criteria and underreporting due to respondent However, in the case of mental illness, such models reluctance to admit symptoms continue to be major could usefully be extended in several ways. For sources of difficulty. Additional problems exist in example, it is important to appreciate that during the studies of special populations, including the young, initial stages of a mental disorder the signs and the elderly, and people in less developed countries. symptoms are often quite nonspecific, making it Innovative methods of minimizing and evaluating difficult for suffers to realize that they are in need of the effects of measurement error and especially of help (58). Another potentially useful way of extend- systematic underreporting are needed to advance the ing standard conceptual models would begin with the aims of analytic epidemiology. Psychiatric epidemiol- realization that the range of culturally available and ogists need to move beyond their current fixation on acceptable strategies for coping with emotional description and analysis of broad-gauged risk problems is much more diverse than for physical markers to study modifiable intervention targets disorders. For example, alternative and complemen- and to develop collaborations with their colleagues tary medicine, the use of informal social support involved in preventing mental illness as well as with networks, other problem-focused strategies aimed at social policy analysts, who are currently at the forefront of developing, implementing, and evaluat- resolving the presumed situational determinants of n the emotional problems (e.g., life style change), and ing interventions. cognitive strategies aimed at redefining the situational determinants of the distress so that they lose their Acknowledgements sting (e.g. cognitive reappraisal, displacement), are all Preparation of this article was supported by US ways of coping with mental illness. Given the stigma Public Health Service Grant K05 MH00507 and by a associated with mental illness, it is reasonable to grant from the US National Institute of Drug Abuse assume that most people will work their way through (R01 DA11121). The author appreciates the helpful many, if not all, of these strategies before seeking help comments of Evelyn Bromet, Kathleen Merikangas, from a mental health professional. Insight into the Bedirhan U¨ stu¨n, and Uli Wittchen on an earlier help-seeking process might be increased if epide- version of the paper. miological studies conceptualized professional treat-

Re´ sume´ Epide´ miologie psychiatrique : progre`sre´ cents et perspectives Exposant dans leurs grandes lignes les enjeux futurs de CIM et du DSM; enfin, que seulement une minorite´ des l’e´pide´miologie psychiatrique, cet article fait le point des personnes qui remplissent les crite`res de´finissant un avance´es re´ centes. Les progre`s importants enregistre´s trouble mental de´ clarent avoir e´te´ soigne´es pendant re´cemment dans le domaine de l’e´pide´miologie psychia- l’anne´e e´coule´e. Il convient aussi de noter que trique descriptive sont la mise au point d’entretiens l’e´pide´miologie psychiatrique descriptive est a` l’origine diagnostiques entie`rement structure´s, fiables et per- de travaux de recherche en e´pide´miologie clinique. Les tinents, et l’utilisation de ces entretiens dans des enqueˆtese´ pide´miologiques cliniques conduites dans des enqueˆtes transnationales paralle`les sur la pre´valence et services cliniques ou de soins primaires sont utilise´es les variables corre´le´es des troubles mentaux. Ces pour e´ valuer les handicaps comparatifs et le mode enqueˆtes montrent re´gulie`rement que les troubles d’utilisation des services pour un large e´ventail de mentaux figurent parmi les maladies chroniques les plus troubles mentaux. re´pandues dans la population ge´ne´rale ; que les troubles Il subsiste ne´anmoins plusieurs proble`mes a`re´soudre mentaux apparaissent en principe a`unaˆ ge beaucoup dans ce domaine. plus pre´coce que les autres maladies chroniques ; que les . La question de l’affinage des cate´ gories et des crite`res troubles mentaux figurent parmi les maladies chroniques diagnostiques continue de se poser depuis que les les plus incapacitantes ; que les personnes qui souffrent proble`mes rencontre´s pour conceptualiser et mesurer des troubles mentaux les plus graves et les plus la morbidite´ ont remis en question la validite´ incapacitants remplissent, a` vie, les crite`res correspon- fondamentale des syste`mes de la CIM et du DSM. dant a` un certain nombre de syndromes diffe´rents de la

Bulletin of the World Health Organization, 2000, 78 (4) 471 Special Theme – Mental Health

. Une e´valuation des biais syste´matiques de sous- a` l’origine de l’apparition des troubles mentaux sont notification dans les enqueˆtes sur les troubles beaucoup plus fortement lie´s a` une adversite´ ambiante mentaux s’impose. Compte tenu du fait que les marque´e qu’a` des facteurs de risque aise´ment modifia- personnes interroge´es dans les enqueˆtes e´pide´mio- bles tels que l’alimentation et le tabagisme. Toutefois, les logiques a` grande e´chelle n’ont gue`re de raisons de psychopathies concomitantes constituent un domaine de re´pondre honneˆtement aux questions personnelles et recherche important et prometteur. La re´ussite des embarrassantes et que la notification des troubles interventions destine´es a`pre´venir les syndromes mentaux est sensible aux moindres variations du psychopathologiques multiples chez les enfants et les contexte et de la fac¸on dont les questions sont adolescents devrait permettre de re´duire la pre´valence de pose´es, les taux de pre´valence des proble`mes affectifs troubles mentaux graves. Plutoˆ t que sur l’apparition des qui ressortent de ces enqueˆtes doivent ge´ne´ralement troubles il conviendrait de faire des recherches sur eˆtre conside´re´s comme des sous-estimations. l’e´volution des troubles, autre domaine de recherche . Des instruments d’e´valuation pre´cis permettant important, mais ne´glige´ jusqu’ici. A l’avenir, les d’e´tudier les troubles de l’enfance et de l’adolescence e´pide´miologistes psychiatriques devront s’attacher a` et la psychopathologie du de´veloppement doivent accroıˆtre l’utilite´ de leurs recherches analytiques pour eˆtre cre´e´s et utilise´s. Leur mise au point a e´te´ entrave´e leurs colle`gues qui participent aux e´tudes sur la par l’impossibilite´ de conduire des entretiens directs pre´vention ou analysent les politiques sociales et qui avec de jeunes enfants et parce que les troubles de sont en premie`re ligne pour l’e´laboration, la mise en l’enfance sont beaucoup moins cristallise´s que les œuvre et l’e´valuation des interventions. Il faudra en outre troubles de l’adulte. parvenir a` mieux comprendre le processus qui conduit a` . Des syste`mes permettant de proce´der a`des solliciter une aide et e´laborer des interventions destine´es estimations dans des zones limite´esetd’e´valuer les a` accroıˆtre la proportion des malades mentaux qui besoins, de planifier les programmes et de pre´voir be´ne´ficient d’un traitement. l’allocation des ressources doivent eˆtre mis en place. Malgre´ des progre` s encourageants, un important Les progre`s re´alise´s dans le domaine de l’e´pide´- travail reste a` faire avant que l’e´ pide´miologie psychia- miologie analytique et expe´ rimentale ont e´te´ plus trique ne donne toute la mesure de sa capacite´a` modestes. Cela tient en partie au fait que les me´ canismes ame´ liorer la sante´ mentale des populations.

Resumen Epidemiologı´a psiquia´ trica: algunos avances recientes y futuras orientaciones En este artı´culo se revisan algunos avances recientes en el servicios correspondientes a un amplio espectro de contexto de los retos futuros que tiene ante sı´ la trastornos mentales. epidemiologı´a psiquia´ trica. Entre los importantes avances Sin embargo, en este terreno sigue habiendo recientes de la epidemiologı´a psiquia´ trica descriptiva cabe varios retos: citar el desarrollo de entrevistas diagno´ sticas fiables y . El perfeccionamiento de las categorı´as y los criterios va´ lidas plenamente estructuradas y la utilizacio´ n de esas diagno´ sticos sigue siendo una cuestio´ n pendiente, ya entrevistas en encuestas transnacionales paralelas de la que los problemas que plantean la conceptualizacio´n prevalencia de los trastornos mentales y los factores con y la medicio´ n de las enfermedades han desencade- ellos relacionados. Entre los resultados sistema´ ticamente nado una pole´ mica sobre la validez de los sistemas de notificados en esos estudios cabe citar los siguientes: los la CIE y del DSM. trastornos mentales son uno de los grupos de dolencias . Es necesario evaluar el sesgo sistema´tico de ma´ s prevalentes entre las enfermedades cro´ nicas que subnotificacio´ n de que adolecen los estudios sobre afectan a la poblacio´ n general; los trastornos mentales los trastornos mentales. Considerando que en los aparecen normalmente a edades mucho ma´ s tempranas estudios epidemiolo´ gicos realizados a gran escala los que otras enfermedades cro´ nicas; los trastornos mentales encuestados esta´ n poco motivados para informar con son una de las enfermedades cro´ nicas ma´ s incapacitantes; sinceridad sobre algunas cuestiones personales y los individuos que presentan los trastornos mentales ma´s embarazosas, y que los resultados de los informes graves y discapacitantes suelen satisfacer durante toda la sobre enfermedades mentales son sensibles a vida los criterios con que se definen varios sı´ndromes en la pequen˜ as variaciones del contexto y de la manera CIE y en el Manual Diagno´ stico y Estadı´stico de Trastornos de formular las preguntas, las prevalencias de Mentales (DSM); y so´ lo una pequen˜ a parte de quienes se problemas emocionales halladas en dichos estudios ajustan a los criterios que caracterizan un trastorno mental deberı´an considerarse en general como estimaciones sen˜ alan que han recibido tratamiento durante el an˜o del lı´mite inferior para la poblacio´n. precedente. Tambie´n cabe destacar el impacto de la . Es necesario crear y aplicar instrumentos precisos de epidemiologı´a psiquia´ trica descriptiva en el inicio de evaluacio´ n para estudiar los trastornos y la psicopa- investigaciones de epidemiologı´a clı´nica. Diversos estu- tologı´a del desarrollo de los nin˜ os y los adolescentes. dios de epidemiologı´a clı´nica realizados en entornos de El desarrollo de esos instrumentos se ha visto atencio´ n clı´nica y primaria esta´ n siendo utilizados para dificultado por la imposibilidad de entrevistar comparar el efecto incapacitante y las pautas de uso de directamente a nin˜ os de corta edad y por el hecho

472 Bulletin of the World Health Organization, 2000, 78 (4) Psychiatric epidemiology: selected recent advances and future directions

de que los trastornos infantiles esta´ n mucho menos reducir la prevalencia de trastornos mentales graves. cristalizados que los trastornos del adulto. Adema´ s, el estudio de los determinantes de la evolucio´n . Es necesario crear sistemas que permitan llevar a cabo de los trastornos, y no tanto de su aparicio´ n, constituye estimaciones a pequen˜a escala a efectos de un campo de investigacio´ n interesante y sin embargo evaluacio´ n de las necesidades, planificacio´n de descuidado. Un futuro e importante reto para los programas y planificacio´n de la asignacio´n de epidemio´ logos psiquia´ tricos consistira´ en realizar inves- recursos. tigaciones analı´ticas que sean ma´ s pertinentes para Los progresos de la epidemiologı´a analı´tica y aquellos de sus colegas que participan en estudios de experimental han sido ma´ s moderados. Ello se debe en prevencio´noenelana´ lisis de polı´ticas sociales, en la parte al hecho de que los mecanismos causales primera lı´nea del desarrollo, la aplicacio´ n y la evaluacio´n implicados en la aparicio´ n de trastornos mentales esta´n de las intervenciones. Otro gran desafı´o consiste en relacionados mucho ma´ s con para´ metros generales de la profundizar en el conocimiento de los procesos de adversidad ambiental que con factores de riesgo bu´ squeda de ayuda y concebir intervenciones que concretos y fa´ cilmente modificables como son la permitan aumentar la proporcio´ n de personas con alimentacio´ n y el tabaquismo. Sin embargo, un a´ rea de trastornos mentales sometidas a tratamiento. investigacio´ n importante y potencialmente fructı´fera son Pese a los alentadores progresos realizados, los sı´ndromes psicopa´ ticos como´ rbidos. Las intervencio- queda mucho trabajo por hacer para que la epidemio- nes aplicadas eficazmente en nin˜ os y adolescentes para logı´a psiquia´ trica actualice su potencial de mejora de la prevenir el desarrollo de numerosos sı´ndromes psicopa- salud mental de las poblaciones. tolo´ gicos permiten albergar grandes esperanzas de

References 1. Susser M. Causal thinking in the health sciences: concepts 15. Kessler RC et al. The effects of chronic medical conditions and strategies of epidemiology. New York, Oxford University on work impairment. In: Rossi AS, ed. Caring and doing for others: Press, 1973. social responsibility in the domains of family, work, and 2. Robins LN, Regier DA. Psychiatric disorders in America: community. Chicago, University of Chicago Press, 2000, in press. the Epidemiologic Catchment Area Study. New York, The Free 16. Kessler RC et al. Lifetime and 12-month prevalence of Press, 1991. DSM-III-R psychiatric disorders in the United States: results from 3. Elwood JM, Little J, Elwood JH. Epidemiology and control the National Comorbidity Survey. Archives of General Psychiatry, of neural tube defects. New York, Oxford University Press, 1992. 1994, 51: 8–19. 4. Robins LN et al. National Institute of Mental Health Diagnostic 17. Alegria M et al. Comparing mental health service use data Interview Schedule: its history, characteristics and validity. across countries. In: Andrews G, ed. Unmet need in mental Archives of General Psychiatry, 1981, 38: 381–389. health service delivery. Cambridge, Cambridge University Press, 5. Helzer JE. A comparison of clinical and Diagnostic Interview 2000: 97–118. Schedule diagnoses: physician re-examination of lay-interviewed 18. Katz SJ et al. Medication management of depression in cases in the general population. Archives of General Psychiatry, the United States and Ontario. Journal of General Internal 1985, 42: 657–666. Medicine, 1998, 13: 77–85. 6. Composite International Diagnostic Interview (CIDI), 19. Regier DA et al. Limitations of diagnostic criteria and Version 1.0. Geneva, World Health Organization, 1990. assessment instruments for mental disorders: implications 7. Andrade L et al. Lifetime prevalence of mental disorders in for research and policy. Archives of General Psychiatry, 1998, a catchment area in Sa˜ o Paulo, Brazil. Paper presented at the 55: 109–115. Seventh Congress of the International Federation of Psychiatric 20. Kessler RC et al. Methodological studies of the Composite Epidemiology, Santiago, Chile, August 1996. International Diagnostic Interview (CIDI) in the US National 8. Offord DR et al. Mental health in Ontario: selected findings Comorbidity Survey. International Journal of Methods from the Mental Health Supplement to the Ontario Health Survey. in Psychiatric Research, 1998, 7: 33–55. Ontario, Queen’s Printer for Ontario, 1994. 21. Spitzer RL. Diagnosis and need for treatment are not the 9. Wittchen H-U et al. Lifetime and six-month prevalence of mental same. Archives of General Psychiatry, 1998, 55: 120. disorders in the Munich Follow-up Study. European Archives 22. Wells K. The functioning and well-being of depressed of Psychiatry and Clinical Neuroscience, 1992, 241: 247–258. patients: results from the Medical Outcomes Study. Journal of 10. Caraveo J, Martinez J, Rivera B. A model for epidemiological the American Medical Association, 1989, 262: 914–919. studies on mental health and psychiatric morbidity. Salud Mental, 23. Sartorius N, Ustun TB. Mental illness in primary care: 1998, 21: 48–57. an international study. New York, John Wiley & Sons, 1995. 11. Bijl RV et al. The Netherlands Mental Health Survey 24. Katon W et al. Collaborative management to achieve treatment and Incidence Study (NEMESIS): objectives and design. Social guidelines: impact on depression in primary care. Journal of Psychiatry Psychiatric Epidemiology, 1998, 33: 581–586. the American Medical Association, 1995, 273: 1026–1031. 12. Ky´ly´c¸C.Mental Health Profile of Turkey: main report. Ankara, 25. Roose SP, Glassman AH. Antidepressant choice in the Ministry of Health Publications, 1998. patient with cardiac disease: lessons from the Cardiac Arrhythmia 13. Kessler RC. The World Health Organization International Suppression Trial (CAST) studies. Journal of Clinical Psychiatry, Consortium in Psychiatric Epidemiology (ICPE): initial work 1994, 55 (suppl A): 83–87. and future directions — the NAPE lecture 1998. Acta Psychiatrica 26. Stoudemire A. Psychological factors affecting medical Scandinavica, 1999, 99: 2–9. conditions. Washington, DC, American Psychiatric Press, Inc., 14. International Consortium of Psychiatric Epidemiology. 1995. Cross-national comparisons of the prevalences and correlates 27. Breton JJ et al. Do children aged 9 through 11 years understand of mental disorders: results from the WHO International the DISC version 2.25 questions? Journal of the American Consortium of Psychiatric Epidemiology. Bulletin of the World Academy of Child and Adolescent Psychiatry, 1995, 34: 946–956. Health Organization, 2000, 78: xxx–xxx.

Bulletin of the World Health Organization, 2000, 78 (4) 473 Special Theme – Mental Health

28. Achenbach TM, McConaughy SH, Howell CT. Child/ 44. Harris T, Brown GW, Robinson R. Befriending as an adolescent behavioral and emotional problems: implications intervention for chronic depression among women in an inner city. of cross-informant correlations for situational specificity. 2: role of fresh-start experiences and baseline psychosocial Psychological Bulletin, 1987, 101: 213–232. factors in remission from depression. British Journal of Psychiatry, 29. Bird HR, Gould MS, Staghezza B. Aggregating data from 1999, 174: 225–232. multiple informants in child psychiatry epidemiological research. 45. Offord DR, Fleming JE. Child and adolescent psychiatry and Journal of the American Academy of Child and Adolescent public health. In Lewis M, ed. Child and adolescent psychiatry: Psychiatry, 1992, 31: 78–85. a comprehensive textbook, 2nd edit. Baltimore, Williams & 30. Costello EJ et al. The prevalence of serious emotional Wilkins, 1996: 1166–1177. disturbance: a re-analysis of community studies. Journal of 46. Kessler RC, Price RH. Primary prevention of secondary disorders: Child and Family Studies, 1998, 7: 411–432. a proposal and agenda. American Journal of Community 31. Lee RM, Renzetti CM. The problems of researching sensitive Psychology, 1993, 21: 607–633. topics. American Behavioral Scientist, 1990, 33: 510–528. 47. Robins LN. The future of psychiatric epidemiology. International 32. Turner CF et al. Adolescent sexual behavior, drug use, Journal of Methods in Psychiatric Research, 1992, 2: 1–3. and violence: increased reporting with computer survey 48. Risch N, Merikangas KR. The future of genetic studies technology. Science, 1998, 280: 867–873. of complex human diseases. Science, 1996, 273: 1516–1517. 33. Reissman CK. Interviewer effects in psychiatric epidemiology: 49. Neale MC, Cardon LR. Methodology for genetic studies of a study of medical and lay interviewers and their impact on twins and families. Dordrecht, Netherlands, Kluwer Academic reported symptoms. American Journal of Public Health, 1977, Press, 1992. 69: 485–491. 50. Kendler KS. The structure of genetic and environmental risk 34. Kessler RC. Gender differences in major depression: factors for six major psychiatric disorders in women. Archives epidemiologic findings. In: Frank E, ed. Gender and its effect of General Psychiatry, 1995, 52: 374–383. on psychopathology. Washington, DC, American Psychiatric Press, 51. Kuller LH. The role of population genetics in the study of 2000: 61–84. the epidemiology of cardiovascular risk factors. In Sing CF, 35. Schaible WL. Indirect estimators in US Federal Programs. Skolnick M, eds. Genetic analysis of common diseases: New York, Springer, 1996. applications to predictive factors in coronary disease. New York, 36. Kessler RC et al. Methodology for estimating the 12-month Liss, 1979. prevalence of serious mental illness. In: Manderscheid RW, 52. Olfson M et al. Psychiatric disorder onset and first treatment Henderson MJ, eds. Mental health, United States 1998. contact in the United States and Ontario. American Journal Washington, DC, US Government Printing Office, 1998: 99–109. of Psychiatry, 1998, 155, 1415–1422. 37. Brewin CR, Andrews B, Gotlib IH. Psychopathology and early 53. Frank RG, McGuire TG. A review of studies of the impact experience: a reappraisal of retrospective reports. Psychological of insurance on the demand and utilization of specialty mental Bulletin, 1993, 113: 82–98. health services. Health Services Research, 1986, 21: 241–265. 38. Mausner JS, Bahn AK. Epidemiology: an introduction. 54. Kessler RC et al. Differences in the use of psychiatric outpatient Philadelphia, PA, WB Saunders, 1984. services between the United States and Ontario. New England 39. Nicolson R et al. Obstetrical complications and childhood- Journal of Medicine, 1997, 336: 551–557. onset schizophrenia. American Journal of Psychiatry, 1999, 55. Friedlander D, Burtless G. Five years after: the long-term 156: 1650–1652. effects of welfare-to-work programs. New York, Russell Sage 40. Neugebauer R, Hoek HW, Susser E. Prenatal exposure Foundation, 1996. to wartime famine and development of antisocial personality 56. Coryell W et al. The time course of nonchronic major depressive disorder in early adulthood. Journal of the American Medical disorder: uniformity across episodes and samples. Archives Association, 1999, 282: 455–462. of General Psychiatry, 1994, 51: 405–410. 41. Prince M. Is chronic low-level lead exposure in early life 57. Janz NK, Becker MH. The health belief model: a decade later. an etiologic factor in Alzheimer’s disease? Epidemiology, 1998, Health Education Quarterly, 1984, 11: 1–47. 9: 618–621. 58. Skelton JA, Croyle RT. Mental representations in health 42. Mrazek PJ, Haggerty RJ, eds. Reducing risk for mental and illness. New York, Springer-Verlag, 1991. disorders: frontiers for preventive intervention research. Washington, DC, National Academy Press, 1994. 43. Harris T, Brown GW, Robinson R. Befriending as an intervention for chronic depression among women in an inner city. 1: randomised controlled trial. British Journal of Psychiatry, 1999, 174: 219–224.

474 Bulletin of the World Health Organization, 2000, 78 (4)