Twelve-month prevalence, severity, and treatment of common mental disorders in communities in Japan: preliminary finding from the World Mental Health Japan Survey 2002–2003


Dr Norito Kawakami, Okayama University Graduate School of Medicine and Dentistry, Hygiene and Preventive Medicine, 2-5-1 Shikata-cho, Okayama 700-8558, Japan. Email:


Abstract  To estimate the prevalence, severity, and treatment of Diagnostic and Statistical Manual of Mental Disorders (4th edn; DSM-IV) mental disorders in community populations in Japan, face-to-face household surveys were conducted in four community populations in Japan. A total of 1663 community adults responded (overall response rate, 56%). The DSM-IV disorders, severity, and treatment were assessed with the World Mental Health version of the World Health Organization (WHO) Composite International Diagnostic Interview (WMH-CIDI), a fully structured lay-administered psychiatric diagnostic interview. The prevalence of any WMH-CIDI/DSM-IV disorder in the prior year was 8.8%, of which 17% of cases were severe and 47% were moderate. Among specific disorders, major depression (2.9%), specific phobia (2.7%), and alcohol abuse/dependence (2.0%) were the most prevalent. Although disorder severity was correlated with probability of treatment, only 19% of the serious or moderate cases received medical treatment in the 12 months before the interview. Older and not currently married individuals had a greater risk of having more severe DSM-IV disorders if they had experienced any within the previous 12 months. Those who had completed high school or some college were more likely to seek medical treatment than those who had completed college. The study confirmed that the prevalence of DSM-IV mental disorders was equal to that observed in Asian countries but lower than that in Western countries. The percentage of those receiving medical treatment was low even for those who suffered severe or moderate disorders. Possible strategies are discussed.


Cross-national psychiatric epidemiology has consistently revealed a lower prevalence of mood and anxiety disorders among East-Asian countries, such as South Korea,1 Taiwan2 and China (Hong Kong),3 compared with Western countries.2–15 A similar pattern was observed for alcohol abuse/dependence,4–6 with an exception for South Korea, in which a high prevalence was reported.1

In a community-based survey in Japan with an original diagnostic instrument, Kitamura et al. reported a high lifetime prevalence of Diagnostic and Statistical Manual of Mental Disorders (3rd edn, revised; DSM-III-R) major depression (14%) but modest lifetime prevalences for other mood and anxiety disorders (1–2%).16 Another community-based survey of mental disorders was conducted using the University of Michigan version of the Composite International Diagnostic Interview (CIDI)6 in an urban population in Japan. The findings showed that the prevalences of mood and anxiety disorders and alcohol abuse/dependence were intermediate between those in East Asia (Taiwan and Hong Kong) and those in the USA and Europe.2,3,15 For medical treatment rates among those who suffered mental disorders, Kitamura et al. reported that only 10% of those who received a diagnosis of mental disorders visited a doctor for medical treatment.16 This rate was far lower than those reported in previous surveys in the USA.13,17 The other survey did not report the medical treatment rate.15 In addition, one dimension that has been lacking in previous psychiatric epidemiologic surveys in relation to the assessment of unmet needs is the severity of mental disorders.18 Many mental disorders are mild and do not require treatment. No previous study in Japan has considered the severity of mental disorders in the epidemiology of mental disorders and medical treatment. Needs and unmet needs relevant to mental disorders are still not clear for the Japanese population.

The World Health Organization (WHO) established the World Mental Health (WMH) Survey Consortium in 1998 to address unmet needs considering the severity of the disorders across developed and developing countries.19 The WMH collaborators expanded the CIDI to include detailed questions about disorder severity, impairment, and treatment and then carried out a coordinated series of WMH-CIDI surveys in 28 countries around the world, including Japan. The first paper from the cross-national collaborative study reported 12-month prevalence, severity, impairment, treatment, and sociodemographic correlates in 14 countries, indicating a large difference in these indicators among countries, particularly, between Western and Asian countries.18 The survey revealed that the medical treatment rate was still lower among people who suffered mental disorders with a severe or moderate impairment, almost in every country.

An objective of the present paper was to describe the 12-month prevalence, severity, and treatment of mental disorders and their demographic correlates based on data specific to Japan that were collected between 2002 and 2003 (WMH Japan 2002–03 Survey) as a part of the WMH surveys.18 Based on the findings, the specific needs, unmet needs, and possible treatment options are presented and discussed in relationship to a proposal for a plan to provide mental health care in Japan.


Survey populations and subjects

Four community populations in Japan were selected as study sites in 2002–03. The sites included two urban cities (Okayama, population 660 000; and Nagasaki, population 450 000) and two rural municipalities (Kushikino, population 25 000) and Fukiage, population 8500, in Kagoshima prefecture). These sites were selected in consideration of geographic variation and the availability of site investigators. Mainly due to the latter factor, all survey sites were located on the west coast of Japan for the 2002–03 WMH Japan surveys. The proportion of those aged ≥65 years old ranged from 17% (Okayama) to 36% (Fukiage); the proportion of those who engaged in agricultural or fishery occupations to the total employed population ranged from 2% to 3% (Nagasaki and Okayama) to 16% (Fukiage). A random sample was selected from residents aged ≥20 years old in each survey site, based on a voter registration list or a resident registry. After a letter of invitation was sent, trained interviewers contacted the subjects and interviewed those who agreed to participate in the survey using the standardized instrument. We excluded subjects who did not meet eligible criteria: those who had died, moved, or were institutionalized. A completed interview was defined as one in which, at least, the pharmacoepidemiology (PH) section of the instrument had been completed. A total of 1664 interviews were obtained. One respondent from the Okayama site was eliminated from the analysis presented in the previous WMH collaboration paper18 because of a coding error. Thus, we eliminated this respondent from further analysis to maintain consistency with the previous study.18 The response rate was calculated as the number of completed interviews divided by the number of eligible subjects (excluding ineligible subjects who were deceased, had moved, or had been institutionalized). The overall response rate was 56% (Table 1). Unfortunately, the response rate at the Nagasaki site was very low (26%) because a different survey method had been used, while the response rates were fairly high (66–81%) at the other three sites. Excluding data from the Nagasaki site did not greatly alter the findings. Thus we included the data in the analysis, together with data from other three sites.

Table 1.   Survey site details: WHO WMH Japan 2002–2003 survey
DispositionOkayamaNagasakiKagoshima PrefectureTotal
  • WHO WMH, World Health Organization World Mental Health.

  • Ineligible subjects include those who were deceased, had moved, or had been institutionalized. Ineligible subjects also include a small number (n = 15) of those who had impaired cognitive functions in the Okayama site.

  • Response rate = (no. completed interviews)/(no. total initial sample–no. ineligible).

Total initial sample16071008001005871002301003224100
Completed interview925 57.6208 26.0354 60.3177 77.01664 51.6
Incomplete interview6  0.43  0.4  9  0.3
No contact80  5.0296 37.0  376 11.7
Refused397 24.7280 35.0185 31.5 40 17.4902 28.0
Ineligible199 12.413  1.6 48  8.2 13  5.7273  8.5
Response rate  65.7  26.4  65.7  81.6  56.4

An internal sampling strategy was used in all surveys to reduce respondent burden by dividing the interview into two parts. Part I included the diagnostic assessment, and part II included information about correlates of a disorder. All respondents completed part I. All part I respondents who met the criteria for any mental disorder and a probability subsample of approximately 10% of other respondents were then given part II (n = 477). The interviews for the respondents who were not selected into part II were terminated after part I. The part I sample was used to estimate the 12-month prevalence, severity, and treatment for most disorders; the part II sample, that is, a subsample of the part I sample, was used to estimate the 12-month prevalence of some disorders and to examine the association between demographic variables and the 12-month prevalence, severity and treatment. The part II respondents were weighted by the inverse of their probability of selection to adjust for the differential sampling of cases and non-cases. In addition to this part II weight, all samples were weighted to adjust for differential probabilities of selection and post-stratified to match the population distributions on the cross-classification for sex and age,20 for which the non-response weight in a given group for sex and age was the inverse of the response rate in this category (Appendix I). The unweighted and weighted distributions of the subjects in part I and part II are shown in Appendix II.

Training and field procedures

The fieldwork for the WMH Japan Survey was carried out by a survey center at each survey site in connection with the headquarters at the National Institute of Mental Health (NIMH), Japan. The centers include Okayama University Graduate School of Medicine and Dentistry, Ijuin Public Health Center, Kagoshima Prefecture, and Nagasaki University Graduate School of Biomedical Sciences. The development of the instrument, training of the interviewers, and preparation of data analysis were supported by the technical support center at Okayama University Graduate School of Medicine and Dentistry (NK).

Before the survey, interviewers received a 5-day standardized instrument-specific training. The training included didactic sessions of general interview skills and reviews of the instrument sections, mock interviews and role-playing exercises. Two official trainers (NK, NI) and other assistants who were previously trained in the instrument provided 5-day training to the interviewers at each survey site.

At the Okayama site, an invitation letter was sent to each subject and then an interviewer visited the homes of the subjects to seek permission to participate in the survey. In the two Kagosima prefecture sites, community volunteers first contacted the subjects in their homes to recruit them into the survey. If the subject agreed, the interviewer conducted a face-to-face interview in the home or at the survey center if the participant preferred. At the Nagasaki site, an invitation letter was sent to each subject, and an interviewer conducted the face-to-face interview with those who replied positively. When an invitation letter was mailed twice and no response was received within a month, no further effort was made to contact the individuals. Written  consent  was  obtained  from  each  respondent at each site. The Research Ethics Committees of Okayama University (for the Okayama site), Japan NIMH (for the Kagoshima site), and Nagasaki University (for the Nagasaki site) approved the recruitment, consent, and field procedures.

Survey instrument

The survey used the Japanese computer-assisted personal interview (CAPI) version of the WMH Initiative version of the WHO Composite International Diagnostic Interview (WMH-CIDI),21 a fully structured diagnostic interview, to assess disorders and treatment. The original English version of WMH-CIDI was translated into Japanese by a team under the supervision of the investigators (NK, NI, TF). Key questions of the final draft of the Japanese version were translated back into English and sent to the WMH Coordinating Center at the Harvard Medical School for a review to check for cross-national consistency. Methodological evidence collected in the WHO-CIDI Field Trials and later clinical calibration studies showed that all disorders were assessed with acceptable reliability and validity both in the original CIDI22 and in the original version of the WMH-CIDI. A pilot study using the Japanese version of WMH-CIDI with a small number of clinical patients showed good concordance between clinical diagnosis and WMH-CIDI diagnosis of major depression and alcohol abuse/dependence.23

Twelve-month diagnosis, severity and treatment

Disorders assessed in the survey were (i) anxiety disorders [agoraphobia, generalized anxiety disorder (GAD), panic disorder, post-traumatic stress disorder (PTSD), social phobia, specific phobia], (ii) mood disorders (bipolar I and II disorders, dysthymia, major depressive disorder), (iii) disorders that share a feature of problems with impulse control [intermittent explosive disorder (IED)], and (iv) substance disorders (alcohol and drug abuse and dependence). Disorders were assessed using the definitions and criteria of the Diagnostic and Statistical Manual of Mental Disorders (4th edn, DSM-IV). The CIDI organic exclusion rules were imposed on all diagnoses.

The WMH-CIDI/DSM-IV disorders were classified as serious, moderate, or mild. Serious disorders were defined as either those meeting the criteria for bipolar I disorder or substance dependence with a physiological dependence syndrome; a suicide attempt in conjunction with any other WMH-CIDI/DSM-IV disorder; reporting at least two areas of role functioning with severe role impairment due to a mental disorder in the disorder-specific Sheehan Disability Scales (SDS);24 or reporting overall functional impairment at a level consistent with a Global Assessment of Functioning (GAF)25 of ≤50 in conjunction with any other WMH-CIDI/DSM-IV disorder.18 Respondents not classified as having a serious disorder were classified as moderate if interference was rated at least moderate in any SDS domain or if the respondent had substance dependence without a physiological dependence syndrome. All other disorders were classified as mild.

Twelve-month treatment was assessed by asking respondents if they ever saw any of a list of professionals either as an outpatient or inpatient for problems with emotions, nerves, mental health, or use of alcohol or drugs. Professionals were classified into the following categories: (i) psychiatrist or mental health specialist, (ii) general medical (other MD or nurse), and (iii) human services (religious provider, social worker or counselor in a non-mental health setting) and complementary and alternative (Internet group, self-help group, or alternative provider). Further, health-care service was defined as psychiatrist, mental health specialist, or general medical. The subjects who had used any of these services in the previous 12 months were placed in a category labeled ‘any treatment’, and those who did not seek treatment were categorized as ‘no treatment.

Patient variables

The demographic characteristics used in the analysis were age, gender, education, marital status, and income. Income was dichotomized based on the median income per family member. Demographic characteristics of the study sample after sampling weighting were: for gender and age (Appendix II); for education, none or some primary, 30.9%; completed secondary, 34.1%; some post-secondary, 18.8%; and college graduate, 16.5%; for marital status, married, 62.8%; and not married, 37.2% (previously married, 15.1%; and never married, 22.1%); for income, below average, 50%; and above average, 50%.

Analysis methods

Data are reported on prevalence, severity, and associations of severity with treatment. Simple cross-tabulations were used to calculate prevalence and severity. Logistic regression analysis was used to study sociodemographic correlates. Standard errors of descriptive statistics were estimated using the Taylor series method implemented in the sudaan software system (Research Triangle Park, NC, USA) to adjust for the weighting of cases.26 The logistic regression coefficients were transformed to odds ratios (OR) and are reported here with design-adjusted 95% confidence intervals (CI). Multivariate tests were based on Wald χ2 tests computed from design-adjusted coefficient variance–covariance matrices. Statistical significance was based on two-sided design-based tests evaluated at the 0.05 level of significance.


Twelve-month prevalence and disorder severity

Approximately 9% of community residents had experienced any WMH/CIDI DSM-IV disorder in the previous 12 months (Table 2); 5% had experienced any anxiety disorder; 3%, any mood disorder; 1%, an impulse-control disorder (i.e. IED); and 1.7%, any substance use disorder. Among the specific disorders assessed in the survey, the 12-month prevalence was greater for major depressive disorder (2.9%), specific phobia (2.7%), and then GAD (1.2%).

Table 2.   Prevalence of 12-month WMH-CIDI/DSM-IV disorders and disorder severity
Disorder12-month prevalenceSeverity of the disorder
  • CI, confidence interval; GAD, generalized anxiety disorder; IED, intermittent explosive disorder; NA, 95% confidence intervals could not be calculated; PTSD, post-traumatic stress disorder; WMH-CIDI/DSM-IV, World Mental Health version of the World Health Organization Composite International Diagnostic Interview/Diagnostic and Statistical Manual of Mental Disorders 4th edn.

  • Part II sample.

  • Part II sample. No adjustment was made for the fact that one or more disorders in the category were not assessed for all part II respondents.

  • –, no cases.

Anxiety disorders
 Panic disorder0.5(0.0–1.1) 24.8 (0.0–58.3)69.1(32.8–100.0) 6.1 (0.0–18.6)
 GAD1.2(0.6–1.8) 29.2 (8.0–50.4)70.8(49.6–92.0)NA
 Specific phobia2.7(1.5–3.9) 10.9 (0.0–24)46.9(27.9–65.9)42.2(26.1–58.3)
 Social phobia0.8(0.2–1.4) 31.6 (0.0–66.7)68.4(33.3–100.0)NA
 Agoraphobia without panic0.3(0.1–0.5) 33.7 (0.0–79.6)66.3(20.4–100.0)NA
 PTSD0.4(0.0–0.8) 65.6(37.4–93.8) 5.9 (0.0–16.3)28.5 (0.0–59.7)
 Any anxiety disorder4.8(3.0–6.6) 15.4 (4.8–26.0)59(46.1–71.9)25.6(14.8–36.4)
Mood disorders
 Major depressive disorder2.9(2.1–3.7) 22.1(10.7–33.5)60.4(44.9–75.9)17.5 (5.3–29.7)
 Dysthymia0.7(0.3–1.1) 41.2 (0.0–83.3)43.7 (0.0–93.1)15.1 (0.0–32.7)
 Bipolar I-II disorders0.1(0.0–0.3)100NANANA
 Any mood disorder3.1(2.1–4.1) 25.8(13.6–38.0)56.4(39.5–73.3)17.8 (6.2–29.4)
Impulse-control disorders
 IED1.0(0.4–1.6) 15.3 (0.0–40.6)33 (0.0–71.4)51.7(20.1–83.3)
Substance use disorders
 Alcohol abuse or dependence1.6(0.2–3.0) 26.7 (1.2–52.2)10.6 (0.0–24.7)62.7(30.6–94.8)
 Alcohol dependence0.4(0.0–0.8) 81.7(44.1–100.0)18.3 (0.0–55.9)NA
 Drug abuse or dependence0.1(0.0–0.3)100.0NANANA
 Drug dependence0.1(0.0–0.3)100.0NANANA
 Any substance use disorder1.7(0.3–3.1) 29.7 (3.2–56.2)10.2 (0.0–23.7)60.1(27.8–92.4)
Any disorder
 Any8.8(6.4–11.2) 16.7 (9.8–23.6)46.6(35.6–57.6)36.7(26.3–47.1)
Total sample  1.5 (0.7–2.3) 4.1 (2.7–5.5) 3.2 (1.8–4.6)

Among those who had experienced a disorder in the previous 12 months, one of six (1.5% of the total population) had experienced a severe disorder, and approximately one half (4.1% of the total population) had experienced a moderate disorder in the previous 12 months. More severe cases were in the categories of PTSD, bipolar I–II disorders (as defined), alcohol dependence, and drug abuse and dependence.

The 12-month prevalence of any DSM-IV disorder was significantly different among the four survey sites after controlling for gender and age groups (data not shown, d.f. = 3, P = 0.038): The estimated OR of having any DSM-IV disorder in the previous 12 months were 1.63 (95%CI, 1.00–2.66) in Nagasaki-city, 0.92 (95%CI, 0.56–1.50) in Kushikino, and 0.47 (95%CI, 0.20–1.10) in Fukiage, compared with Okayama as a reference. Similar tendencies were observed for any mood disorder and any anxiety disorder, although the difference was not significant for any of the four disorder categories (d.f. = 3, P > 0.05).

Association of 12-month disorder severity with treatment

A severe disorder and a moderate disorder were combined into one category because <30 cases had a serious disorder. Those who had a severe/moderate disorder were more likely to receive health care and medical treatment (P < 0.05), particularly among the mental health specialties (Table 3). However, only 19% of those who had a severe/moderate disorder sought medical treatment.

Table 3.   Association of 12-month WMH-CIDI/DSM-IV disorder severity with treatment
TreatmentAny disorderNo disorderTotal respondents
Serious or moderateMild
  • CAM, complementary and alternative; WMH-CIDI/DSM-IV, World Mental Health version of the World Health Organization Composite International Diagnostic Interview/Diagnostic and Statistical Manual of Mental Disorders 4th edn.

  • The categories were combined because fewer than 30 individuals had a serious disorder. Significant differences in a proportion for any treatment, mental health specialty, health care, absence of health care, and proportion of no treatment between the two severity categories of a disorder (d.f. = 1, all P < 0.05).

Any treatment22.7(10.7–34.7)12.0 (1.4–22.6) 6.2(3.1–9.3) 7.3 (4.4–10.2)
Mental health specialty17.5 (7.1–27.9) 6.0 (0.0–14.4) 1.3(0.0–2.7) 2.5(1.3–3.7)
General medical 6.6 (0.9–12.3) 5.1 (0.0–11.4) 3.5(1.7–5.3) 3.7(2.1–5.3)
Any health care18.8 (8.4–29.2) 4.8(2.8–6.8) 5.8(3.8–7.8) 5.8(3.8–7.8)
Non-health care (Human services or CAM) 8.6 (1.3–15.9) 0.9(0.0–2.9) 1.5(0.0–3.1) 1.9(0.5–3.3)
No treatment77.3(65.3–89.3)88.0(77.4–98.6)93.8(90.7–96.9)92.7(89.8–95.6)

Sociodemographic correlates of 12-month prevalence, severity, and health-care treatment

None of the sociodemographic predictors (i.e, sex, age, family income, marital status and education) was significantly associated with the 12-month prevalence of any disorder (P > 0.05, Table 4). The severity of any disorder experienced was significantly greater among older individuals and those who were not currently married (P < 0.05). The probability of seeking medical treatment was greater among those who had completed high school than among those who had received less education and those who had graduated from college (P = 0.001).

Table 4.   Predictors of WMH-CIDI/DSM-IV 12-month prevalence, severity, and health-care treatment (part II sample, n = 477)
 Any 12-month disorderSeverity/DisorderTreatment
OR95% CIOR95% CIOR95% CI
  • CI, confidence interval; OR, odds ratio; WMH-CIDI/DSM-IV, World Mental Health version of the World Health Organization Composite International Diagnostic Interview/Diagnostic and Statistical Manual of Mental Disorders 4th edn.

  • Serious or moderate (coded 1) vs mild (coded 0) among those who experienced any 12-month disorder.

  • Controlling for disorder severity.

  • *

    Significant at the 0.05 level, two-sided test.

  inline image = 0.2 inline image = 0.1 inline image = 0.1
Age (years)
  inline image = 1.7 inline image = 11.0* inline image = 0.7
Family income
  inline image = 0.4 inline image = 2.4 inline image = 0.1
Marital status
 Currently married1.001.001.00
 Never married/separated/widowed/divorced1.50(0.55–4.10)4.30(1.17–15.82)0.75(0.24–2.33)
  inline image = 0.7 inline image = 5.5* inline image = 0.3
 Middle school or less0.85(0.29–2.54)0.28(0.03–2.54)0.65(0.12–3.63)
 High school1.45(0.54–3.92)0.54(0.13–2.17)1.88(0.51–6.95)
 Some college1.88(0.60–5.93)0.53(0.11–2.50)1.05(0.35–3.21)
 College or higher1.001.001.0
  inline image = 3.1 inline image = 1.6 inline image = 20.4*

Sociodemographic correlates of WMH-CIDI/DSM-IV disorder types

Any mood disorders experienced in the previous 12 months were more prevalent in those who were never married (Table 5). Substance use disorders in the previous 12 months were more prevalent among men, those aged 35–49 years old, and among those who had higher-than-average income (P < 0.05). None of the sociodemographic predictors was significantly associated with 12-month prevalence of any anxiety disorder or impulse control disorder (P > 0.05).

Table 5.   Predictors of WMH-CIDI/DSM-IV disorder types (part II sample, n = 477)
 Any moodAny anxietyAny impulseAny substance
OR95% CIOR95% CIOR95% CIOR95% CI
  • CI, confidence interval; OR, odds ratio; WMH-CIDI/DSM-IV, World Mental Health version of the World Health Organization Composite International Diagnostic Interview/Diagnostic and Statistical Manual of Mental Disorders 4th edn.

  • *

    Significant at the 0.05 level, two-sided test.

  inline image = 1.3 inline image = 0.1 inline image = 0.6 inline image = 4.7*
Age (years)
  inline image = 3.5 inline image = 2.9 inline image = 1.3 inline image = 10.7*
Family income
 High1.001.00 1.01.00
  inline image = 0.0 inline image = 0.0 inline image = 0.2 inline image = 6.1*
Marital status
 Currently married1.00
 Never married/separated/widowed/divorced3.00(0.94–9.53)1.46(0.59–3.63)0.12(0.01–1.71)3.65(0.25–53.03)
  inline image = 3.9* inline image = 0.8 inline image = 2.8 inline image = 1.0
 Middle school or less0.25(0.03–2.43)1.00(0.26–3.88)1.52(0.04–52.69)10.20(0.37–278.16)
 High school1.04(0.24–4.54)1.90(0.62–5.82)3.45(0.13–92.09)1.65(0.0–38.93)
 Some college0.73(0.17–3.17)2.78(0.92–8.41)1.90(0.09–41.39)6.26(0.20–194.00)
 College or higher1.
  inline image = 2.8 inline image = 6.9 inline image = 0.9 inline image = 5.0


The present study has confirmed that the prevalence of mental disorders was lower in Japan than that in Western countries.15,18,27,28 The 12-month prevalences of any DSM-IV disorder, four broad categories of mental disorders, and specific mental disorders were almost equal to those in WMH surveys in Beijing, and some European Union (EU) countries (Spain, Italy, and Germany), greater than those in Shanghai and Nigeria, and lower than those reported in the USA, Colombia, and most EU countries.6,12,14,18,29

The present survey added new information on the prevalence of PTSD, IED and drug abuse/dependence in Japan. The 12-month prevalence of PTSD in the present study was 0.4%, which is much lower than that recently reported in the USA.30 One of 100 respondents experienced IED in the previous 12 months, which is characterized by discrete episodes of failure to resist aggressive impulses, resulting in serious assaults or destruction of property. However, the estimated figure should be interpreted with caution because the diagnosis of IED should be made only after a thorough medical work-up.31 The 12-month prevalence of drug abuse/dependence was much lower than that previously reported in the USA and other Western countries.6,32 This is consistent with the fact that the Japanese government has a strong policy for controlling the use of illicit and other psychotropic drugs.33

The WMH surveys found that mild cases met diagnostic criteria but involved almost no impairment of functioning.18 However, 5.6% of the total population experienced a severe or moderate disorder in the past 12 months; 2.4% of the total population experienced a severe or moderate form of major depression in the previous 12 months. These severe or moderate cases should be a primary target of mental health care.

People were more likely to have sought medical treatment if they had experienced a severe or moderate disorder within the previous 12 months. However, only one of five who had a severe disorder sought medical treatment. The medical treatment rate was lower than that in the WMH surveys in most developed countries, such as the USA and Europe, while it was slightly greater than that in China, Columbia and Lebanon.18 It is somewhat unexpected that those who had completed high school or attended some college were significantly more likely to seek medical treatment than others. Individuals at each end of the educational spectrum may be reluctant to seek medical treatment for mental disorders. Those with the least education may lack knowledge of mental disorders and their treatments, and the more educated may face psychological barriers, such as fear of losing social status, a threat to job security, or humiliation, which might prevent them from seeking care.

As observed in previous studies in Japan and many other countries, women had a greater risk of mood disorder and anxiety disorders and a lower risk of substance  use  and  impulse-control  disorders,  although the  association  was  not  always  significant,  likely  due to a small number of the part-II sample respondents.4,6,15,16,18,34 Greater prevalence of any substance use disorder among younger groups is also in concordance with previous studies.4 Although not significant, the present study supports a previous observation that mood disorders were more prevalent among the middle-aged in Japan.15 A greater risk of suicide has been observed among men in the same age range in Japan.35 The most likely factor underlying the phenomenon is economic recession but others should be considered, such as rapid changes in family and society as well as the possible burden of caring for elderly relatives. The middle-aged population may be a primary important target for mental health care in Japan. In addition, older respondents had a significantly greater risk of having a severe form of a mental disorder after they became affected with one. This may be attributable to a greater risk of recurrence of mental disorders and comorbidity with a somatic disorder among older patients.36,37 A greater risk of mood and anxiety disorders among those not currently married is also in concordance with previous observations.29,34 However, in contrast to previous observations in other countries,18,27,29 income showed almost no association with mood or anxiety disorders; substance use disorder was rather more prevalent in the group with high income. First, this may be attributable to a possible selection bias in that people with low income who suffered from these disorders may be in poor living conditions and were not likely to participate in the study. Another possibility included confounding by employment status: if those who were currently employed (and earned some income) suffered from stress at work and were more likely to develop mental disorders, the deteriorating effect of low income could be diminished or even reversed. However, a previous study in Japan has reported an inconsistent association between income and depressive episode.27 Income may thus not be a strong correlate of common mental disorders in Japan, meaning that future research in Japan will be needed to address this point.

Among 102 million adults aged 20 years or over in Japan (Japan Statistics Bureau, 2002), 5.7 million (5.6%) may have experienced a severe or moderate mental disorder in the previous 12 months. The majority of them did not seek medical treatment. A primary objective of a future mental health policy in Japan should be to increase medical treatment rates for those experiencing severe or moderate mental disorders. Impaired functioning in several life domains is anticipated in such cases. Awareness of these cases by family, friends, supervisors/coworkers, and neighbors, possibly enhanced through education/training and the dissemination of information, may be useful to facilitate medical treatment, in addition to the increased awareness of such cases. A fairly large portion of those who had experienced mental disorders already received medical treatment from non-psychiatrists. Training non-psychiatrist physicians to treat mental disorders and establishing a liaison between non-psychiatrists and psychiatrists might help bridge the gap.

The survey sites were selected from western Japan, and they did not include a metropolitan city with a population of >1 million. The lower response rate (57%) may also limit the interpretation of the findings, which may lead to either underestimation or overestimation of prevalence. In the Nagasaki site, the response rate was particularly low and the prevalence of mental disorders was several times greater than those at the other three sites, suggesting that those who had a disorder may have been more willing to participate at this site. If this is true, the lower response rate may inflate the overall prevalence. The response rate was lower among younger age groups, which may result in a distorted association between age and the prevalence of mental disorders.

The Japanese version of the WMH-CIDI was not fully validated against clinical diagnoses, although it was developed by an expert group and checked though an expert review and a back-translation procedure. The validity of the instrument should be checked against clinical diagnosis. The observed prevalences may have been over- or under-estimated in the present study because of the use of this instrument.13 The prevalence of substance use disorders and PTSD, which was estimated based on the part II sample (n = 477), might be unstable. For the relevant analyses based on these respondents, the number of respondents may have been insufficient to provide for the detection of a true association. The on-going WMH Japan survey will replicate and expand the present findings.


The World Mental Health Japan Survey 2002–2003 was carried out in conjunction with the WHO World Mental Health Survey Initiative (Chairs, Prof. Ronald C. Kessler, Harvard Medical School, and Dr Bedhan Ustun, WHO; We thank the coordinating staff members of WMH for their assistance with the instrumentation and their consultation on field procedures. The study was supported by grants from the Japan Ministry of Health, Labor, and Welfare. Thanks are also due to the staff members, Hisateru Tachimori, PhD, Yuko Miyake, PhD, at the National Institute of Mental Health, Japan, and other field coordinators in the WMH Japan 2002–2003 Survey.



Part I: weight creation

1. Post-stratification (PS) weight
  • • Purpose: To compensate for differences between the sample and population characteristics due to frame undercoverage, non-response and sampling variability.
  • • Five age groups and two gender groups were used in this case. This indicates that we have 10 subgroups within each of four survey sites.
Let {PSWi = The post-stratification weight}, where i = 1, . . . , 1663.
2. Normalize the PS weight

Part II: weight creation

1. The part II (PII) weight
  • • p1 and p2 calculated as empiric probability of selection into Part II sample for each subgroup. Empirical probabilities obtained using data weighted with Part I weight (NPSi).
2. Post-stratification (PS) weight
  • • Post-stratification procedure carried out on Part II sample (n = 477). See Part I description.
3. Normalize the PS weight
  • • Normalization procedure carried out on Part II sample (n = 477). See Part I description.


Table 6.  Sociodemographic distribution (%) of the survey sample in the WHO WMH Japan 2002–03 survey compared to population
 Part I unweightedPart II unweightedPart I weightedPart II weightedPopulation census
  • WHO WMH, World Health Organization World Mental Health.

  • Survey site, gender and age groups were used in the post-stratification of weight.

  • Average distributions based on population statistics of four survey sites in 2002.

Age (years)