Psychotic Experiences in the General Population: A Cross-National Analysis Based on 31 261 Respondents From 18 Countries

Importance  Community-based surveys find that many otherwise healthy individuals report histories of hallucinations and delusions. To date, most studies have focused on the overall lifetime prevalence of any of these psychotic experiences (PEs), which might mask important features related to the types and frequencies of PEs.

Objective  To explore detailed epidemiologic information about PEs in a large multinational sample.

Design, Setting, and Participants  We obtained data from the World Health Organization World Mental Health Surveys, a coordinated set of community epidemiologic surveys of the prevalence and correlates of mental disorders in representative household samples from 18 countries throughout the world, from 2001 through 2009. Respondents included 31 261 adults (18 years and older) who were asked about lifetime and 12-month prevalence and frequency of 6 types of PEs (2 hallucinatory experiences and 4 delusional experiences). We analyzed the data from March 2014 through January 2015.

Main Outcomes and Measures  Prevalence, frequency, and correlates of PEs.

Results  Mean lifetime prevalence (SE) of ever having a PE was 5.8% (0.2%), with hallucinatory experiences (5.2% [0.2%]) much more common than delusional experiences (1.3% [0.1%]). More than two-thirds (72.0%) of respondents with lifetime PEs reported experiencing only 1 type. Psychotic experiences were typically infrequent, with 32.2% of respondents with lifetime PEs reporting only 1 occurrence and 31.8% reporting only 2 to 5 occurrences. We found a significant relationship between having more than 1 type of PE and having more frequent PE episodes (Cochran-Armitage z = −10.0; P < .001). Lifetime prevalence estimates (SEs) were significantly higher among respondents in middle- and high-income countries than among those in low-income countries (7.2% [0.4%], 6.8% [0.3%], and 3.2% [0.3%], respectively; χ²₂ range, 7.1-58.2; P < .001 for each) and among women than among men (6.6% [0.2%] vs 5.0% [0.3%]; χ²₁ = 16.0; P < .001). We found significant associations with lifetime prevalence of PEs in the multivariate model among nonmarried compared with married respondents (χ²₂ = 23.2; P < .001) and among respondents who were not employed (χ²₄ = 10.6; P < .001) and who had low family incomes (χ²₃ = 16.9; P < .001).

Conclusions and Relevance  The epidemiologic features of PEs are more nuanced than previously thought. Research is needed that focuses on similarities and differences in the predictors of the onset, course, and consequences of distinct PEs.

In recent years, interest in the epidemiologic landscape of hallucinations and delusions has grown.^1,2 These psychotic experiences (PEs) are reported by a sizeable minority of the general population. A recent meta-analysis based on 61 studies² reported that the median lifetime prevalence of PE was 7.2%. Because this prevalence is substantially higher than the lifetime morbid risk for psychotic disorders, such as schizophrenia (median morbid risk, 0.7%),³ the field of psychiatric epidemiology has been forced to rethink how PEs fit into the epidemiologic landscape of psychotic disorders. The terms used to describe these experiences have also evolved over time. Although psychoticlike experiences has been used, we will use the general term psychotic experiences to encompass hallucinatory experiences (HEs) and delusional experiences (DEs).²

Early work on the epidemiologic features of PEs⁴ focused on these experiences as risk indicators for later conversion to full psychosis. This type of research has an appealing logic because many of the risk factors associated with PEs are also associated with schizophrenia and psychosis.⁵Quiz Ref ID Additional evidence has accumulated that PEs are also associated with the subsequent onset of a wide array of common mental disorders, including anxiety, mood, and substance use disorders,^6-8 and with an increased risk for suicidal ideation and intent.^9-11 Thus, awareness is growing that the presence of PEs may reflect a vulnerability to a wide range of adverse mental health outcomes (in addition to psychotic disorders).^4,12-16 These findings and the concern that antipsychotics may be inappropriately used to treat individuals with isolated PEs may have influenced the decision to exclude attenuated psychosis syndrome in recently revised diagnostic criteria.¹⁷

As the empirical data have accumulated, systematic reviews^2,15,18,19 have pooled prevalence estimates and applied meta-regression techniques to explore the sociodemographic correlates of PEs. These reviews provide valuable clues to the nature of PEs but also highlight important gaps in the literature. Four of these gaps are of special importance for the present study.

First, the use of pooling in systematic reviews of PEs has encouraged the use of coarse dichotomous measures (eg, presence or absence of lifetime prevalence) to harmonize the wide array of scales and diagnostic instruments used to assess PEs.² This dichotomy has reduced the subtlety of the associations examined in these reviews. Second, the studies included in the systematic reviews have varied in many key design elements. As noted by Linscott and van Os,² substantial heterogeneity in the data has hampered analyses related to the relationship between PEs and sociodemographic variables. Third, most of the community studies of PE prevalence and correlates have been performed in high-income countries. A major exception is the World Health Survey, which included 4 brief PE questions in surveys of 52 nations.²⁰ However, the World Health Survey assessment of PEs had several limitations (eg, it lacked information on the frequency of PE occurrence, and questions about DEs were not asked in a fashion that excluded experiences related to alcohol, illicit drugs, or sleep). Finally, to allow pooling of data from different studies, some reviews have collapsed different variables across orthogonal axes. For example, Kaymaz et al¹⁵ compiled composite variables related to weak and strong PEs that in theory could be built from data related to (1) the count of different types of PEs, (2) the frequency of occurrence, (3) associated distress, (4) comorbidity, and/or (5) certainty (eg, confidence in the psychotic nature of the experience).

Leading commentators have repeatedly called for more fine-grained analyses of PEs to guide the field.^1,21 The present report provides initial results of analyses designed to address the above limitations by examining data collected in the World Health Organization World Mental Health (WMH) surveys, a series of population-based surveys administered in many countries using consistent instruments and field procedures designed to facilitate pooled cross-national analyses of the prevalence and correlates of mental disorders. These data provide an unprecedented opportunity to explore the epidemiologic landscape of PEs.

The WMH surveys are a coordinated set of community epidemiologic surveys administered in probability samples of the household population in countries throughout the world (http://www.hcp.med.harvard.edu/wmh/). Data were collected from 2001 through 2009. Eighteen of the 26 completed WMH surveys administered the Composite International Diagnostic Interview (CIDI) psychosis module. These 18 countries are distributed across North and South America (Colombia, Mexico, Peru, Brazil [São Paulo], and the United States), Africa (Nigeria), the Middle East (Iraq and Lebanon), Asia (People’s Republic of China [Shenzhen]), the South Pacific (New Zealand), and Europe (Belgium, France, Germany, Italy, the Netherlands, Portugal, Romania, and Spain). All 18 surveys were based on multistage, clustered-area probability, household sampling designs (Table 1). The weighted mean response rate across all 18 countries was 72.1%. Most surveys were based on nationally representative sample frames, but a few excluded rural areas (ie, Colombia and Mexico) or focused on particular regions (ie, Nigeria and China) or cities (Brazil). Participating sites were grouped into 3 country-level income strata according to World Bank criteria^22,23 as low- and lower middle–income countries (Colombia, Iraq, Nigeria, China, and Peru), upper middle–income countries (Brazil, Lebanon, Mexico, and Romania), and high-income countries (the European countries, New Zealand, and the United States). The age ranges reported herein include 18 years and older except in 3 countries (Mexico, Colombia, and Peru) in which 65 years was the upper age limit.

In keeping with previous studies of PEs,^10,12,24-29 we made the a priori decision to exclude individuals who had PEs with positive findings of a screen for possible schizophrenia and/or psychosis and manic depression and/or mania (ie, respondents who reported schizophrenia, psychosis, or manic depression and/or mania in response to the question, “What did the doctor say was causing this/these experiences?” and those who ever took any antipsychotics for these symptoms). This process resulted in the exclusion of 140 respondents (0.4% of all respondents), leaving 31 261 respondents for this study (Table 1).

All WMH surveys were conducted face to face in the homes of respondents by trained lay interviewers. Written informed consent was obtained before beginning the interviews in all countries. Procedures for obtaining informed consent and protecting individuals (ethical approval) were approved and monitored for compliance by the institutional review boards of the collaborating organizations in each country.³⁰ Full details of these procedures are described elsewhere.^31,32

All WMH interviews had 2 parts. Part 1 was administered to all respondents and contained assessments related to core mental disorders. Part 2 included additional information relevant to a wide range of survey aims, including assessment of PEs. All respondents who met criteria for any part 1 DSM-IV mental disorder and a probability sample of other respondents were administered part 2. Respondents to part 2 were weighted by the inverse of their probability of selection for part 2 to adjust for differential sampling. Within the different sites, questions related to PEs were administered to all respondents or a random sample of those administered part 2. Analyses in this study were based on the weighted part 2 subsample of respondents administered the CIDI psychosis module. Additional weights were used to adjust for differential probabilities of selection within households and for nonresponse and to match the samples to sociodemographic distributions in the population.

The instrument used in the WMH surveys was the World Health Organization CIDI,³³ a validated, fully structured diagnostic interview designed to assess the prevalence and correlates of a wide range of mental disorders according to the definitions and criteria of the diagnostic systems of the DSM-IV³⁴ and International Statistical Classification of Diseases, 10th Revision.

Quiz Ref IDThe CIDI psychosis module included questions about the following 6 PE types: 2 related to HEs (visual and auditory hallucinations) and 4 related to DEs (2 bizarre delusional items [thought insertion and/or withdrawal and mind control and/or passivity] and 2 paranoid delusional items [ideas of reference and plot to harm and/or follow]) (eTables 1 and 2 in the Supplement). For example, respondents were asked if they ever experienced PEs (eg, “Have you ever heard any voices that other people said did not exist?”). This question was followed by a probe question to determine whether the reported PEs ever occurred when the person was not “dreaming or half-asleep or under the influence of alcohol or drugs.” Only responses of the latter type are considered herein. The sequence of these follow-up probe types differed slightly between the first 6 WMH surveys, which were administered in Europe (Belgium, France, Germany, Italy, the Netherlands, and Spain), and those in the remaining 12 countries (eTables 1 and 2 in the Supplement).

Respondents who reported PEs were then asked about the presence of the PEs in the past 12 months and the frequency and/or occurrences of the PEs in their lifetime. In this study, we present prevalence estimates for any PE, any HE (with or without associated DEs), any DE (with or without associated HEs), pure HEs (without DEs), and pure DEs (without HEs). In addition, we will present the following 2 key PE-related metrics: (1) count of types of PEs (henceforth referred to as PE type metric) and (2) frequency of occurrence of PE episodes (henceforth referred to as PE frequency metric). Respondents may have had more than 1 hallucination and/or delusion type associated with a single episode of PE. For the PE frequency metric, reported frequency of lifetime PE episodes was divided into the following 5 categories: 1 time only, 2 to 5 times, 6 to 10 times, 11 to 100 times, and 101 times or greater. This 5-category scheme was collapsed into 1 to 10 vs 11 or more times in the analyses of sociodemographic correlates of PE frequency among respondents with lifetime PEs.

The sociodemographic factors considered herein include sex, age, number of years of education, employment history, marital status, family income, and nativity (ie, born inside the country of assessment). For the bivariate and multivariate analyses, the sociodemographic variables were stratified into broad categories based on methods described elsewhere.³⁰

Data were analyzed from March 2014 through January 2015. Weighted prevalence estimates were calculated for the various PE types and related metrics. Odds ratios and design-corrected 95% CIs are reported. Because the WMH survey data featured geographic clustering and weighting, SEs of parameter estimates were generated using the design-based Taylor series linearization method³⁵ implemented in a commercially available macro (SAS, version 9.4; SAS Institute, Inc). Multivariate significance was evaluated using Wald χ² tests based on design-corrected coefficient variance-covariance matrices. The association between the PE type metric and the PE frequency metric was evaluated using the Cochran-Armitage test.³⁶ Statistical significance was evaluated consistently using 2-tailed .05-level tests.

Table 2 presents country-specific lifetime PE prevalence estimates. Quiz Ref IDLifetime prevalence (SE) of at least 1 PE was reported by 5.8% (0.2%) of the 31 261 respondents. Lifetime prevalence of any HE was 5.2% (0.2%); of any DE, 1.3% (0.1%). The median (interquartile range) of lifetime PEs, HEs, and DEs were 5.5% (2.8%-7.5%), 4.4% (1.8%-6.5%), and 1.3% (0.9%-1.8%), respectively (eFigures 1-3 in the Supplement present the cumulative distribution of PE, HE, and DE estimates, respectively). Twelve-month prevalence (SE) of any PE was 2.0% (0.1%), whereas the median (interquartile range) was 1.4% (1.0%-2.8%).

Lifetime prevalence (SE) of PEs was significantly higher among women than men (6.6% [0.2%] vs 5.0% [0.3%]; χ²₁ = 16.0; P < .001). Similar sex differences were found for prevalence of HEs (5.9% [0.2%] vs 4.3% [0.3%]; χ²₁ = 19.4; P < .001) but not DEs (1.4% [0.1%] vs 1.3% [0.1%]; χ²₁ = 0.3; P = .61). The significant sex difference was also found for respondents with pure HEs (5.2% [0.2%] vs 3.7% [0.3%]; χ²₁ = 19.3; P < .001) but not pure DEs (0.7% [0.1%] vs 0.7% [0.1%]; χ²₁ = 0.1; P = .80).

Significant differences were found across the 3 country-level income strata in lifetime prevalence of any PE, any HE, and any DE. In each comparison, the prevalence estimates were significantly higher among respondents in middle- and high-income countries than among those in low-income countries (χ²₂ range, 7.1-58.2; P < .001 for each) (Table 2).

Table 3 shows the lifetime prevalence estimates (SE) of individual PE types and counts of different PE types. Quiz Ref IDThe most common PE type overall was visual hallucinations (3.8% [0.2%]), followed by auditory hallucinations (2.5% [0.1%]). Prevalence estimates of individual DE types were low (0.3%-0.7%). Among those with any lifetime PE, 72.0% (representing 4.2% of the total sample) reported only 1 PE type; 21.1% (representing 1.2% of the total sample), exactly 2 types; and 6.8% (representing 0.4% of the total sample), 3 or more types.

Psychotic experiences were typically infrequent, with 32.2% of the respondents with lifetime PEs reporting only 1 episode (Table 4). An additional 31.8% of respondents with lifetime PEs experienced only 2 to 5 PE episodes. Thus, for nearly two-thirds of respondents (64.0%) with lifetime PEs, these experiences occurred only 1 to 5 times in their lives. Quiz Ref IDAn additional 10.0% of respondents with lifetime PEs reported 6 to 10 lifetime episodes; 20.0%, 11 to 100 episodes; and 6.0%, 101 or more episodes. The relationship between the PE type metric and the PE frequency metric is best displayed in Table 5. Those respondents with more PE types are disproportionately more likely to have more PE episodes (Cochran-Armitage z, −10.0; P < .001).

eTable 3 in the Supplement shows the association of sociodemographic variables with lifetime PEs, HEs, and DEs in bivariate and multivariate models. The following sociodemographic variables were associated with increased odds ratios for PEs, HEs, and DEs in both models: (1) being a homemaker or classified as having other employment (eg, looking for work or disabled) vs employed (for PEs, multivariate model, χ²₄ = 10.6; P < .001); (2) being nonmarried (ie, never married or separated, widowed, or divorced) vs married (for PEs, multivariate model, χ²₂ = 23.2; P < .001); and (3) lower vs high household income level (for PEs, multivariate model, χ²₃ = 16.9; P < .001). In addition to these findings, several sociodemographic variables were associated with only 1 type of PE. Young respondents (aged 18-29 years) were significantly more likely to have DEs (compared with those older than 60 years), whereas age was unrelated to HEs (and overall PEs). Although female sex was associated with an increased prevalence of PEs (in both models), this finding was driven by an increased risk for HEs (but not DEs). Low educational level, in comparison, was associated with an increased risk for DEs (but not HEs). Unexpectedly, those born outside the country (ie, migrants) were significantly less likely than the native born to report HEs (but not DEs) in the bivariate and multivariate models.

Among the factors that influence the PE type metric (in those who had experienced PEs) in the multivariate model, the 3 younger strata (ie, spanning 18-59 years) were significantly more likely to have more than 1 PE type (compared with those aged ≥60 years) (eTable 4 in the Supplement). None of the other sociodemographic characteristics was associated with the PE type metric. Among the correlates of the PE frequency metric, student status was significantly associated with a lower frequency of PE occurrence. None of the other sociodemographic variables was associated with PE frequency (eTable 5 in the Supplement).

Based on cross-national samples from 18 countries, we found that 5.8% of respondents reported having 1 or more PEs at least once in their lifetime and 2.0% in the previous year. These overall estimates are broadly consistent with those in the previous literature.² Our study contributes important new information regarding the count of PE types and frequency of PEs that go beyond the issues considered in previous community-based studies of PEs. Perhaps the most striking finding is that PEs are infrequent for most of the individuals who experience them, with 32.2% reporting 1 PE episode in their life and 64.0% reporting no more than 5 lifetime occurrences. In the general population, those with 2 or more types of PEs are also significantly more likely to have more PE episodes. For example, of those who reported 3 or more PE types, nearly one-quarter (24.5%) reported 101 or more occurrences.

Our findings provide an empirical foundation on which to investigate factors that influence the persistence of PEs.¹ When viewed within the context of the gap between 12-month and lifetime PE estimates from the present study (ie, 2.0% vs 5.8%), we can infer that most individuals do not have persistent PEs (mindful that lifetime prevalence estimates for mental health disorders are often biased downward owing to underreporting).³⁷ Linscott and van Os² estimated that of those individuals who report any PEs, approximately 80% would have had transient experiences. This estimate is consistent with our empirical finding that about 64.0% of individuals with PEs report only 1 to 5 lifetime occurrences.

Based on the set of PEs examined, our study confirms that HEs were more common than DEs (5.2% vs 1.3%), and this general pattern was consistent across the 3 country-level income strata. Lifetime prevalence of PEs was lower in the low-income to lower middle–income countries (3.2%) compared with the upper middle–income and high-income countries (7.2% and 6.8%, respectively). Although we cannot compare our results directly with the single previous cross-national study of PEs²⁰ owing to differences in how the PEs were assessed, both studies (optimized for consistent design and PE assessment) provided insights in variation between sites.

One of the strengths of cross-national studies, such as the WMH Survey, is that they are able to identify risk factors that exist consistently across countries despite site-specific cultural factors. We found an increased prevalence of HEs and DEs associated with being unmarried or not employed and having a lower household income level. However, certain demographic features were differentially associated with HEs but not DEs and vice versa. For example, women had a significantly higher prevalence of HEs but not DEs. We found a significant relationship between being younger and having DEs but not HEs. Unexpectedly, migrants in our study were significantly less likely to report lifetime HEs (compared with native-born respondents). These novel findings provide important points of distinction between the epidemiologic features of psychotic disorders and PEs.^5,38

Although several sociodemographic variables were significantly associated with the lifetime prevalence of PEs, these features were not associated with the PE type metric or the PE frequency metric. We speculate that comorbid psychiatric illness (eg, depression and anxiety disorders) and other risk factors known to be associated with PEs and mental disorders (eg, family history, substance use, and trauma exposure) may contribute to these PE-related metrics. The comprehensive nature of the WMH survey will allow us to explore these hypotheses in future analyses.

Although our study has many strengths (eg, large sample size, range of countries, uniform methods for data collection, and innovative analysis of PE-related metrics), several limitations are notable. In keeping with other population-based surveys, we relied on trained lay interviewers to administer the questionnaire. Although we excluded those individuals who had screen-positive findings for possible psychotic disorders, we did not have access to valid measures of clinical psychotic disorders. Lifetime prevalence estimates are prone to underreporting.³⁷ We only assessed 4 types of DEs, and these probes may have been insensitive to culture-specific delusional beliefs.¹⁶

We have provided, to our knowledge, the most comprehensive description of the epidemiologic landscape of PEs published to date. Although the lifetime prevalence of PEs is 5.8%, these events are typically rare. For nearly one-third of individuals who have PEs (ie, 32.2%), these events were solitary (ie, 1 event only). In the general population, a small subgroup of individuals has multiple types of PEs and experiences these types of PEs more frequently. The research community needs to leverage this fine-grained information to better determine how PEs reflect risk status. Our study highlights the subtle and variegated nature of the epidemiologic features of PEs and provides a solid foundation on which to explore the bidirectional relationship between PEs and mental health disorders.

Submitted for Publication: February 10, 2015; final revision received and accepted March 24, 2015.

Corresponding Author: John J. McGrath, PhD, MD, Queensland Brain Institute, University of Queensland, St Lucia, Queensland, Australia 4072 (j.mcgrath@uq.edu.au).

Published Online: May 27, 2015. doi:10.1001/jamapsychiatry.2015.0575.

Author Contributions: Dr Kessler had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: McGrath, Al-Hamzawi, Alonso, Haro, Kovess-Masfety, Lepine, Ochoa, Viana, Kessler.

Acquisition, analysis, or interpretation of data: Saha, Alonso, Bromet, Bruffaerts, Caldas-de-Almeida, Chiu, de Jonge, Fayyad, Florescu, Gureje, Haro, Hu, Kovess-Masfety, Lepine, Lim, Mora, Navarro-Mateu, Ochoa, Sampson, Scott, Viana, Kessler.

Drafting of the manuscript: McGrath, Saha, Al-Hamzawi, Kessler.

Critical revision of the manuscript for important intellectual content: McGrath, Saha, Alonso, Bromet, Bruffaerts, Caldas-de-Almeida, Chiu, de Jonge, Fayyad, Florescu, Gureje, Haro, Hu, Kovess-Masfety, Lepine, Lim, Mora, Navarro-Mateu, Ochoa, Sampson, Scott, Viana, Kessler.

Statistical analysis: McGrath, Saha, Chiu, Lim, Sampson, Kessler.

Obtained funding: McGrath, Alonso, Bruffaerts, Caldas-de-Almeida, Gureje, Kovess-Masfety, Lepine, Mora.

Administrative, technical, or material support: McGrath, Bromet, Caldas-de-Almeida, de Jonge, Gureje, Hu, Lepine, Ochoa.

Study supervision: Fayyad, Navarro-Mateu, Ochoa, Sampson.

Conflict of Interest Disclosures: During the past 3 years, Dr Kessler has been a consultant for Hoffman-La Roche, Inc, Johnson & Johnson Wellness and Prevention, and Sanofi and has served on advisory boards for Mensante Corporation, Plus One Health Management, Lake Nona Institute, and US Preventive Medicine. No other disclosures were reported.

Funding/Support: Each World Mental Health (WMH) country obtained funding for its own survey. The São Paulo Megacity Mental Health Survey is supported by Thematic Project Grant 03/00 204-3 from the State of São Paulo Research Foundation; the Shenzhen Mental Health Survey, by the Shenzhen Bureau of Health and the Shenzhen Bureau of Science, Technology, and Information; the Colombian National Study of Mental Health, by the Ministry of Social Protection and the Saldarriaga Concha Foundation; the European Study of the Epidemiology of Mental Disorders project, by contracts QLG5−1999−01042 and 2004123 from the European Commission, the Fondo de Investigación Sanitaria (the Piedmont Region [Italy]), grant FIS 00/0028 from the Instituto de Salud Carlos III (Spain), grant SAF 2000−158-CE from the Ministerio de Ciencia y Tecnología (Spain), grants Centro de Investigación Biomédica en Red CB06/02/0046 and Redes Temáticas de Investigación Cooperativa en Salud RD06/0011 REM-TAP from the Departament de Salut, Generalitat de Catalunya, Spain, Instituto de Salud Carlos III, other local agencies, and an unrestricted educational grant from GlaxoSmithKline; the Iraq Mental Health Survey (IMHS), by funding from the Japanese and European funds through United Nations Development Group Iraq Trust Fund; the Lebanese National Mental Health Survey, by the Lebanese Ministry of Public Health, the World Health Organization (WHO) (Lebanon), anonymous private donations to the Institute for Development, Research, Advocacy and Applied Care, Lebanon, and unrestricted grants from Janssen Cilag, Eli Lilly and Company, GlaxoSmithKline, Roche, and Novartis; the Mexican National Comorbidity Survey (MNCS), by grant INPRFMDIES 4280 from The National Institute of Psychiatry Ramon de la Fuente, grant CONACyT-G30544-H from the National Council on Science and Technology, and supplemental support from the Pan American Health Organization; the New Zealand Mental Health Survey, by the New Zealand Ministry of Health, Alcohol Advisory Council, and the Health Research Council; the Nigerian Survey of Mental Health and Well-being, by the WHO (Geneva), the WHO (Nigeria), and the Federal Ministry of Health, Abuja; the Peruvian World Mental Health Study, by the National Institute of Health of the Ministry of Health of Peru; and the US National Comorbidity Survey Replication, by grant U01-MH60220 from the National Institute of Mental Health, the National Institute on Drug Abuse, the Substance Abuse and Mental Health Services Administration, grant 044708 from the Robert Wood Johnson Foundation, and the John W. Alden Trust. Additional support was from John Cade Fellowship APP1056929 from the National Health and Medical Research Council (Dr McGrath).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views and opinions expressed in this report are those of the authors and should not be construed to represent the views or policies of any of the sponsoring organizations, agencies, or the WHO.

Additional Contributions: The surveys discussed in this article were administered in conjunction with the WHO WMH Survey Initiative. The WMH staff assisted with instrumentation, fieldwork, and data analysis, and these activities were supported by grant R01MH070884 from the US National Institute of Mental Health, the John D. and Catherine T. MacArthur Foundation, the Pfizer Foundation, grants R13-MH066849, R01-MH069864, and R01 DA016558 from the US Public Health Service, Fogarty International Research Collaboration award R01-TW006481 from the Fogarty International Center, the Pan American Health Organization, Eli Lilly and Company, Ortho-McNeil Pharmaceutical, Inc, GlaxoSmithKline, and Bristol-Myers Squibb.