2 - The Cross-National Epidemiology of Mental Disorders

Introduction

Successive advances have been made in our understanding of the epidemiology of mental disorders over the past several decades. These gradual advances in our understanding in turn rest on a foundation of incremental progress in various areas, including conceptualization of mental disorders, community survey methodology and data analysis, and global collaborative research networks. The World Mental Health (WMH) Surveys Initiative, comprising the largest and most sophisticated coordinated set of community surveys on mental disorders to date, exemplifies the progress that has been made and provides a state-of-the-art portrayal of the epidemiology of mental disorders around the world. In this background chapter, we provide some relevant historical and theoretical context in order to outline the scope and value of the WMH surveys, and to address some key criticisms of the work.

Operationalization of Mental Disorders

Mental disorders have been recognized and described by clinicians for millennia (Kendler 2009). Hippocrates and many other physicians of the ancient world provided detailed descriptions of depression and alcohol dependence; conditions which we now include under the rubric of ‘common mental disorders’. In the nineteenth century, Kraepelin and others working in mental hospital settings delineated psychotic disorders such as schizophrenia and bipolar disorder: entities that are now often termed ‘serious mental disorders’. Freud, Janet, Westphal, and others working in outpatient settings described a range of anxiety and impulse-control disorders. Mental disorders were classified and categorized in early editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM), and of the International Statistical Classification of Diseases (ICD).

At the same time, by the middle of the twentieth century, a number of problems in psychiatric classifications were readily apparent (Spitzer et al. 1980). A particularly important issue that came to light was disagreement between clinicians in different countries about the diagnosis of schizophrenia; in the United Kingdom there was a comparatively high threshold for diagnosis of this condition, while in the United States there was a much lower threshold (Cooper et al. 1969). Critics reported that healthy individuals who feigned mental illness could readily be admitted to a psychiatric hospital (Rosenhan 1973). In order to address these problems, researchers began to develop diagnostic criteria to operationalize psychiatric disorders; for example, the Feighner criteria were developed by Robins and colleagues and the Research Diagnostic Criteria were developed by Spitzer and colleagues (Kendler et al. 2010).

This work in turn provided the basis for the third edition of the DSM, published by the American Psychiatric Association in 1980. This volume, led by Spitzer, provided diagnostic criteria for a range of mental disorders and so enabled clinicians across the United States to more reliably determine the presence of diagnostic entities. The manual was translated into a range of languages and within a few years became the dominant nosology for psychiatric research across the globe. The diagnostic criteria were incorporated into clinician-administered diagnostic interviews, they formed the basis for inclusion criteria into clinical trials of pharmacotherapy and psychotherapy for mental disorders, and they provided a foundation for community and clinical surveys of mental disorders in a range of different settings.

Subsequent decades saw the publication of DSM-IV and DSM-5, as well as further revisions of the ICD. Whereas DSM-III broke new ground by providing a novel approach to the operationalization of mental disorders, DSM-IV and DSM-5 made incremental progress, with changes to the diagnostic criteria based on accumulating research (Frances et al. 1989). Thus, for example, high thresholds for the introduction of new disorders in DSM-IV and DSM-5 were set, and the focus was on evidence-based refinement of the diagnostic criteria of existing mental disorders. DSM is produced by the American Psychiatric Association, and recent revisions have emphasized the importance of diagnostic validity (that is, going beyond the earlier emphasis on reliability of operational criteria to focus on evidence supporting the relevant construct) (Kupfer & Regier 2011). ICD is produced by the World Health Organization (WHO), and recent efforts at revision have emphasized the importance of clinical utility and of global applicability (Reed & Ayuso-Mateos 2011). Nevertheless, DSM-IV and the research version of ICD-10 have a good deal of overlap, and there are ongoing attempts to harmonize DSM-5 and ICD-11 (Regier et al. 2013).

Community Surveys Methodology

A range of advances in the methodology of community surveys have been crucial for contemporary epidemiology of mental disorders. Once diagnostic criteria had been developed for psychiatric disorders, an immediate question was whether they could feasibly be administered in the community, for example, by master's level clinicians (Weissman et al. 1978). A key advance was the development of a fully structured lay interview based on the Feighner criteria, the Research Diagnostic Criteria, and the DSM-III diagnostic criteria: the Diagnostic Interview Schedule (DIS) (Robins et al. 1984). Work on the DIS was supported by the National Institute of Mental Health Center for Epidemiological Studies, which funded the Epidemiological Catchment Area study in the United States; the first large multi-site community survey of mental disorders to be based on the new approach to the operationalization of mental disorders. Comparison of data from lay-administered interviews with independent clinician-administrated interviews indicated acceptable reliability and validity.

Subsequent work on a new diagnostic instrument was undertaken by the WHO in collaboration with the US Alcohol, Drug, and Mental Health Administration in order to include criteria from the ICD, and to translate the instrument into a range of different languages (Kessler 2000b). This first version of the Composite International Diagnostic Interview (CIDI) was made available in 1990, and included both DSM-III and ICD-10 criteria, with subsequent versions incorporating DSM-III-R and DSM-IV diagnostic criteria. The CIDI has a range of features designed to improve the reliability and validity of the data obtained from lay-administered interview, including efforts to increase respondent understanding, motivation, and ability to provide accurate survey responses (Kessler and Üstün 2004). Such features, and the concordance of the CIDI with clinical diagnosis, are discussed in more detail in the next chapter.

Prior to the DIS and CIDI, psychiatric epidemiology lagged behind work done in other areas of health. The DIS and CIDI allowed a new generation of psychiatric epidemiology research (Dohrenwend and Dohrenwend 1982), which employed operational definitions of mental disorders, and so provided rigorous data on their prevalence and correlates. A number of key lessons emerged from this new generation of studies, which include a range of surveys focusing on particular mental disorders, such as the National Vietnam Veterans Readjustment Study (Schlenger et al. 2007) and the National Epidemiological Survey on Alcohol and Related Conditions (Hasin & Grant 2015). First, depression, anxiety disorders, and substance-use disorders are highly prevalent conditions in the United States. Second, these disorders have an early age-of-onset, protracted course, and high comorbidity. Third, mental disorders are associated with significant impairment, as well as high costs to society. Finally, mental disorders are significantly under-diagnosed and under-treated.

The new generation of epidemiological research on mental disorders has also provided a series of lessons about particular mental disorders. Consider, for example, some key findings on anxiety disorders (Kessler et al. 2010). First, data on prevalence of different psychiatric disorders has emphasized that anxiety disorders are the most common class of psychiatric disorders, perhaps speaking to the adaptive value of fear responses (Stein & Nesse 2011). Second, data on onset and course have emphasized that many anxiety disorders precede mood and substance-use disorders, so emphasizing the potential clinical value of early recognition and robust treatment (Goodwin & Gorman 2002). Third, data on comorbidity have emphasized that depression and generalized anxiety disorder (GAD) have similarly high levels of comorbidity and of associated impairment; such work has contributed to the conceptualization of GAD as an independent diagnostic entity (Kessler 2000a). Fourth, work on the treatment gap in anxiety disorders has given impetus to the emergence of consumer advocacy organizations that work to promote awareness of these conditions (Stein et al. 2001). Analogous lessons have emerged for a range of psychiatric conditions.

Global Collaborative Research Networks

Soon after the development of the DIS, studies were undertaken with this instrument in a number of different countries (Weissman et al. 1994, 1996). The CIDI, based on a newer version of the DSM, and including ICD diagnostic criteria, was developed with the explicit aim of being translated into a range of different languages and being used in different regions of the globe, with modifications being made to ensure cross-cultural compatibility – including a range of work to address cross-national variations in the vocabulary of distress (see also Chapter 3). Comparison of data from across countries allows an initial approach to the question of how the prevalence and correlates of mental disorders differ across different contexts. However, in order to fully exploit the richness of data from different contexts and countries, international consortia that gather and pool such data are needed. Fortunately, such work has advanced in the past two decades.

Surveys using the DIS and DSM-III criteria in a limited number of countries suggested that although there were some differences in prevalence of mental disorders across different contexts, the conditions assessed (mood, anxiety, and substance-use disorders) were common, and individuals with these conditions had substantial unmet treatment needs (Weissman et al. 1994, 1996). Given that the WHO focuses on the ICD for its work on data gathering and policy formulation, this organization created a cross-national network of researchers using the CIDI in different countries, the WHO International Consortium in Psychiatric Epidemiology (ICPE) (2000). Data from the ICPE emphasized the high prevalence, early onset, substantial persistence, and high comorbidity of mental disorders in a range of different settings. In addition, however, ICPE investigators were able to pool epidemiological data to provide input on questions about risk of disorders; thus for example an ICPE analysis of predictors of onset and persistence of GAD found that history of specific phobia may be a risk marker for GAD onset (Kessler et al. 2002).

A revision of the CIDI in the late 1990s provided the impetus for the WHO to establish the WMH Surveys Initiative, and to encourage countries around the world to use the revised CIDI to undertake nationally representative surveys. Grants from a number of agencies, including the National Institutes of Health in the United States allowed the creation of a core infrastructure for the Initiative, so leveraging expertise and allowing cost-saving for participating countries. Many low and middle-income countries, which had not previously participated in psychiatric epidemiological research and which lacked both financial and human resources to undertake such work, were now able to conduct high-quality and large-scale community surveys, and to build capacity in psychiatric epidemiology. The WMH Surveys Initiative now comprises over 30 countries, and includes more than 150,000 respondents (see Chapter 3).

The WMH surveys have brought enormous scope and value to psychiatric epidemiology, providing, for the first time, rigorous data on the prevalence and correlates of a range of mental disorders in a wide range of low, middle, and high-income countries. Notably, coincident with the collection, analysis, and publication of this new body of data, we have witnessed the emergence of the new field of global mental health (Patel 2012; Becker & Kleinman 2013). Indeed, just as epidemiology can be conceptualized as a key pillar of public health, so work such as that of the WMH surveys and the Global Burden of Disease study (GBD Collaborators, Murray et al. 2015) has provided a foundation for global mental health, which in turn has emphasized the high disability due to mental disorders, and the large treatment gap associated with these conditions. The WMH Surveys Initiative has provided a key evidence base for the clarion call of global mental health for ‘no health without mental health’ (Prince et al. 2007).

Criticisms of Contemporary Psychiatric Nosology and Epidemiology

The operationalization of mental disorders provided by DSM and ICD, advances in survey methodology outlined earlier, and the development of global collaborative research networks such as the WMH surveys all lay the foundation for remaining chapters of this volume to describe the cross-national epidemiology of a wide range of mood, anxiety, substance-use, and disruptive behaviour/impulse-control disorders. At the same time, it is relevant to consider a number of the key criticisms of these foundations, and so of the work described in this volume. We consider in turn, some related criticisms of contemporary psychiatric nosology, of the approach of the WMH surveys, and of the employment of the CIDI across different regions of the world.

Criticisms of current psychiatric nosology include those emphasizing the reification of mental disorders, artefactual comorbidity, and vested interests. Some argue that psychiatrists now use checklists to make diagnoses; they are unaware of symptoms not listed in DSM, and they fail to appreciate that diagnostic criteria are merely an hypothesis about how a disorder most commonly presents – for any particular patient current diagnostic constructs may not serve well (Hyman 2010; Kendler 2016). Patients often have more than one DSM diagnosis, which suggests that some of this comorbidity is artefactual in nature (Maj 2005). Neuroscientifically minded critics have argued that some of these flaws reflect the fact that diagnostic operationalizations are not sufficiently based on biological discoveries; neurocircuitry-based constructs, it has been suggested, may ultimately provide a more valid and useful foundation for assessment and treatment (Insel et al. 2010). More sceptical critics have argued that more disorders translate into increased sales, and that psychiatry has colluded with the pharmaceutical industry to inappropriately apply Western diagnostic constructs in non-Western contexts, ignoring the issue of fundamental differences in the structure of distress and psychopathology across the globe, and instead medicalizing non-pathological distress and deviance (Moynihan et al. 2002; Stein 2015).

Certainly, the diagnostic operationalization embedded in contemporary psychiatric nosology reflects the status of our current knowledge of mental disorders. At the same time, considerable effort has been made to establish and to improve the diagnostic validity and clinical utility of modern diagnostic constructs (Reed & Ayuso-Mateos 2011; Regier et al. 2013). Such efforts avoid reification by focusing on the evidence in support of contrasting diagnostic conceptualizations and operationalizations, fully accepting that symptoms lie on dimensions and that disorders have fuzzy boundaries with non-disorder and with one another. Conflicts of interest are rigorously addressed, and participation of a broad range of stakeholders is vigorously encouraged (Stein & Phillips 2013). While new biologically based conceptual frameworks such as the Research Domain Criteria (RDoC) may well be useful in informing future research (Cuthbert & Insel 2010), it is by no means clear that superior diagnostic constructs for clinical work will easily be found (Stein 2014). Indeed it is unlikely that the next editions of DSM and ICD will represent a paradigm shift; it is much more likely that they will represent iterative improvement (Kendler & First 2010).

Related criticisms have been made of both the WMH surveys and of global mental health; emphasizing that psychiatric epidemiology and public mental health are flawed insofar as they rely on DSM and ICD operationalizations (Summerfield 2012). On the one hand, some critics have argued that given that we already know that mental disorders are highly prevalent and under-diagnosed and under-treated, the WMH surveys add little that is new (Weich & Araya 2004). On the other hand, others argue that WMH and global mental health have inflated the prevalence of mental disorders and their associated disability; the argument is that policy-makers should not be misled by such data, and should be careful not to conflate psychological distress with medical disorder (Summerfield 2012). Those who are less sceptical of contemporary psychiatric nosology and epidemiology still have a range of other criticisms. These include that the CIDI takes insufficient account of cultural idioms of distress, that the reliance on recall biases age-of-onset and lifetime prevalence estimations, and that the cross-sectional design of the WMH surveys makes it difficult for causal factors underlying mental disorders to be delineated.

While it is important to be aware of the limitations of any scientific research and consequent efforts at advocacy, the sophistication of the WMH surveys and of efforts in global mental health should not be underestimated. Although it is true that WMH surveys have relied on DSM and ICD diagnostic conceptualizations and operationalizations, they have also made important contributions to psychiatric nosology by rigorously comparing different diagnostic constructs (Ruscio et al. 2007; Lee et al. 2009; Karam et al. 2010; Stein et al. 2010, 2013; McLaughlin et al. 2015). Relatedly, WMH surveys have documented how the symptoms of common mental disorders lie on dimensions, and have contributed to psychiatric nosology by evaluating the impact of choosing different cut-points and thresholds for determining the presence of disorders. Finally, the WMH surveys have expended significant energy on addressing cross-national variations in the vocabulary of distress, in optimizing recall of symptoms, and in improving analyses of risk and resilience factors (see Chapter 3). Similarly, the discipline of global mental health has emphasized the importance for public health of appreciating the range that spans serious mental disorders, psychological distress, and well-being, and of addressing the specific social and cultural context within which psychiatric symptoms emerge.

Indeed, we would argue that in a number of respects there has been significant convergence between nosological advances, clinical neuroscience, psychiatric epidemiology, and global mental health; this convergence has helped address critiques of contemporary psychiatric theory, and has contributed to strengthening clinical practice and mental health policy. Thus, for example, the question of how to define a case has long been debated in each of these arenas. The WMH surveys have made an important contribution to this issue by measuring symptom severity of specific mental disorders, by paying particular attention to the question of how best to delineate subthreshold from clinical symptomatology, and by providing an evidence base for proposals to improve the operationalization of diagnostic entities in DSM and ICD. Relatedly, an emphasis on the dimensional nature of psychiatric symptomology has been a key issue in psychiatric nosology, clinical neuroscience, psychiatric epidemiology, and global mental health. Thus DSM-5 introduced new chapters on related psychiatric conditions, emphasized that psychiatric symptoms lie on dimensions of severity and are present across disorders, and provided a range of trans-diagnostic symptom measures. RDoC has emphasized that symptoms lie on dimensions and cut across different disorders; and WMH analyses have investigated different cut-points and thresholds for determining mental disorder.

Conclusion

The subsequent chapters of this volume provide data on the cross-national descriptive epidemiology of a wide range of mental disorders. Certainly our epidemiological data are only as good as our current study instruments, as the memory of survey respondents permits, and as their willingness to participate fully in the interview process allows. Still, notwithstanding the limitations of current diagnostic operationalizations, survey methodology, and global research collaborative networks, these data provide the most comprehensive available perspective on the prevalence and distribution of mental disorders around the world. Taken together, they provide a compelling picture of the burden of mental disorders and of the treatment gap; information that is crucial for those working to improve mental health policies and services. At the same time, the data are an important bridge across the classical divide between psychiatric epidemiology and clinical practice, asking and answering a range of important questions about the prevalence and correlates of mental disorders, and providing clinicians with key facts and figures about the most important conditions seen in everyday work. We believe that the wealth of analyses provided in this volume comprise an important step in the evolution of epidemiological research on mental disorders. Our hope is that they also provide a foundation from which to evaluate future policy and clinical interventions aimed at increasing access to and effectiveness of mental disorder treatments.