MENTAL HEALTH HAS become one of the major

Transcription

1 Psychiatry and Clinical Neurosciences 2011; 65: doi: /j x Regular Article Screening performance of K6/K10 and other screening instruments for mood and anxiety disorders in Japanpcn_ Keiko Sakurai, MPH, 1 * Akihiro Nishi, MD, MPH, 2,5,6 Kyoko Kondo, MD, PhD, 3 Kosuke Yanagida, MD 4 and Norito Kawakami, MD, DMSc 1 1 Department of Mental Health, School of Public Health, 2 Department of Public Health, Graduate School of Medicine, University of Tokyo, Tokyo, 3 Yanagawa Clinic, Junpukai and 4 You Clinic, Okayama, Japan, 5 Department of Society, Human Development, and Health, Harvard School of Public Health, and 6 Institute for Quantitative Social Science, Harvard University, Massachusetts, USA Aims: This study aimed to establish the screening performance and optimal cut-off points for the Japanese version of Kessler (K)6, K10 and the Depression and Suicide Screen (DSS). Methods: A self-report questionnaire including K6, K10 and DSS, as well as the Center for Epidemiologic Studies Depression Scale (CES-D), was administered to a random sample of community residents in Japan (non-cases, n = 147) and psychiatric outpatients diagnosed with mood or anxiety disorders according to DSM-IV (cases, n = 17). A receiver operator characteristics (ROC) curve was drawn to estimate the area under the curve (AUC), the sensitivity, and specificity with the optimal cut-off points for K6, K10, and DSS, which were then compared with those of CES-D. The community sample was also asked to rate each measure on a scale from very easy to very hard to use. (0.95), but DSS showed a significantly smaller AUC (0.89) than CES-D (P < 0.05). The optimal cut-off points were estimated as 4/5 for K6, 9/10 for K10, and 1/2 for DSS. The sensitivity of these three scales was similar, but the specificity was lower for DSS than for the other two. K6, K10 and DSS were rated as being very easy or easy to use significantly more than CES-D (P < 0.01). Conclusion: The screening performance of the Japanese versions of K6 and K10 was comparable with that of CES-D, and better than that of DDS. K6/K10, particularly K6, might have an advantage, even over the CES-D, because of its similar screening performance and better acceptability. Key words: depression and suicide screen, K6/10, mood/anxiety disorder, scale, screening performance. Results: K6 and K10 showed a high AUC ( ), which was comparable to that of CES-D MENTAL HEALTH HAS become one of the major global burdens of disease, accounting for 14% of all diseases. 1 In Japan, Kawakami and his colleagues reported that the 12-month prevalence of mood disorders was 3.1% (major depressive disorder: 2.9%) *Correspondence: Keiko Sakurai, MPH, Department of Mental Health, School of Public Health, Graduate School of Medicine, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo , Japan. sakura07-tky@umin.ac.jp Received 27 March 2010; revised 21 February 2011; accepted 17 May and anxiety disorders 4.8% (general anxiety disorder: 1.2%). 2 Many people suffer from major depression. Major depressive disorder is one of the dominant reasons for committing suicide in Japan (27.6%). 3 Mood/anxiety disorder is also one of the biggest concerns that are linked to reduced social productivity and individual satisfaction in daily life. 4 6 To minimize the social and individual influence of mood and anxiety disorders, including suicide, it becomes essential to detect mental illnesses in their early stages. 7 9 Many self-administered questionnaires have been developed for screening mood and 434

2 Psychiatry and Clinical Neurosciences 2011; 65: Screening performance of K6/K anxiety disorders in the general population (e.g. the 22-item Langner Scale, the General Health Questionnaire-12), or for screening depression in particular (e.g. the Self-rating Depression Scale [SDS], Hospital Anxiety and Depression Scale [HADS], Center for Epidemiologic Studies Depression Scale [CES-D]) Short screening questionnaires, such as the K6 and K10 (hereafter referred as K6/K10), which contain six or ten items of the same set of questions on depression and anxiety, respectively, have been increasingly used in community settings. 15 K6 and K10 were also used in national surveys in the USA, Canada, Australia, and Japan. 16 Using item response theory (IRT), self-administered or intervieweradministered K6/K10 questionnaires were developed to detect general psychological distress. 15 Because of their brevity (six and ten questions for K6 and K10, respectively, to be completed within 2 3 min), K6 and K10 have a great advantage over other well-known scales. Reliability and validity of the original versions have been repeatedly evaluated in the USA and Australia; 17,18 the K6/K10 scales were reported to detect major depression and dysthymia most effectively according to the DSM-IV. Kessler recommended using the cut-off score of 12/13 as an indicator of severe mental illness. K6/K10 scales were translated into Japanese 19 and the interview version was validated in the general population in Japan. 19 The screening performance of K6/K10 scales in detecting 30-day DSM-IV mood and anxiety disorders was excellent with values as high as 0.94 for K6 and 0.94 for K10 in the areas under receiver operater characteristic curves (AUC). 19 This is the only study to investigate the screening performance of the Japanese versions of K6/K10. It is not clear what was the best cut-off score for K6/K10 in Japan, or if K6/K10 had a better performance compared with other scales of depression and anxiety, such as CES-D. The other depression scale, which has also been frequently used, is the Depression and Suicide Screen (DSS), 20 a five-item scale based on the SDS for the early detection of depression and suicidal ideation in Japan. 12 The scale has been widely used in Japan as part of the National Suicide Prevention Program Manual. 21 DSS had AUC values as high as 0.77 for predicting depression cases identified by the Geriatric Depression Scale Short-form in a sample of elderly. 20 However, the scale has not been validated against a psychiatrist s diagnosis of depressive disorder and the screening performance has not been examined in non-elderly populations. The aim of this study was to evaluate the screening performance of K6/K10 and DSS for mood and anxiety disorders, and to compare it with CES-D, a widely used standard instrument, which has been well validated for screening depression 14,22,23 and used in large-scale population-based studies in Japan. 24 We compared AUC, sensitivity, and specificity among these scales, estimated based on the comparison between psychiatric outpatients with DSM-IV mood and anxiety disorders and a community sample. In addition, previous studies of a self-administered questionnaire for mental illness detection in Japan did not investigate acceptability of these scales in a general population. 19 We also compared the degree of acceptability (easiness to complete a questionnaire) of these scales judged by the community sample. METHODS Subjects The study subjects were drawn from psychiatric outpatients with DSM-IV mood and anxiety disorders and a community sample. The data collection was conducted between January and February Psychiatric outpatients were selected from two psychiatric outpatient clinics, based on the availability of trained psychiatrists who cooperated for our study, and patient profiles of the clinics with a greater proportion of mood/anxiety disorders, in Okayama City, Japan. All patients who met the diagnosis of DSM-IV for mood (major depressive disorder, dysthymia, and bipolar I II disorders) or anxiety disorders (panic disorder, specific phobia, social phobia, agoraphobia without panic, and PTSD) (n = 18, 10 men and eight women, and six people aged years, nine people aged years and three people aged 60 and older) were invited and all agreed to participate in the study. The subjects were asked to complete a self-administered questionnaire consisting of screening scales for depression. Data from one woman was excluded from the analysis because of a missing response on one K10 item. The current diagnoses of the remaining 17 patients were obtained according to the DSM-IV criteria from psychiatrists in charge of patient treatment, who were fully experienced in clinical practice and also well-trained in the use of DSM-IV criteria for more than 10 years. The diag-

3 436 K. Sakurai et al. Psychiatry and Clinical Neurosciences 2011; 65: noses included major depression (n = 8), dysthymia (n = 4), generalized anxiety disorders (n = 4), and the comorbidity of panic disorder and dysthymia (n = 1). The survey of patients was unfortunately terminated with the limited number of patient subjects because fewer than expected patients visited the clinics and met the criteria during the survey period, and because a psychiatrist moved to another clinic and the survey could not continue beyond the predetermined period. For the community sample, a questionnaire was mailed to a sample of 500 community residents aged over 20 years who were randomly selected based on a voter registration list of the city. They were asked to complete the questionnaire and return it by mail. A total of 180 community residents (61 men, 108 women, and 11 with unknown sex, and 35 people aged years, 53 people aged years, 79 people aged 60 and older and 13 with unknown age) responded to the survey (response rate; 36%). An additional 33 respondents who had missing responses on K6/K10, CES-D, or DSS were excluded. Data from the remaining 147 respondents in the community sample were analyzed. The community sample was not assessed for psychiatric diagnosis but was regarded as non-cases because the 12-month prevalence of mood and anxiety disorders in the community was reported to be low (about 4%) in Japan. 2 The Ethics Committee for Epidemiological Study of the Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University reviewed and approved the study aims and procedures. Measures Psychiatric outpatients and community residents completed the self-administered questionnaire, which included relevant questions including demographic variables and four screening scales for mood and anxiety disorders: K6 scale, K10 scale, 15,17 DSS, 20 and CES-D. 22 K6/K10 scales K6 and K10 consisted of six and ten items, respectively, measured on a 5-point scale (0 4). The total score (ranging from 0 24 for K6 and 0 40 for K10) has been used as an indicator of severe mental disorders 18 or mood and anxiety disorders. 17 High scores indicate more severe mental disorders. K6 and K10 were translated into Japanese and showed its validity to DSM-IV mood and anxiety disorders diagnosed by a lay-interviewer-administered diagnostic interview in the community sample, 19 although the study did not report the best cut-off score to be used for the screening of mood and anxiety disorders. DSS The DSS consists of five dichotomous questions (yes/ no): Is your life pretty full?, Do you still enjoy doing the things you used to do?, Do you think it is too much trouble to do the things you used to do?, Do you feel that you are a useful person who is needed by others?, and Do you feel tired without any specific reason? with the total score ranging from 0 to 5. Using a cut-off of 1/2, DSS was reported to predict depression as assessed by the short-form Geriatric Depression Scale 25,26 with sensitivity of 70.5% and specificity of 72.9%. 20 DSS was included in the guidelines for prevention of depression in the community published by the Japan Ministry of Health, Labour, and Welfare as a convenient tool for depression. 27 CES-D The CES-D is a 20-item scale that has been evaluated for its reliability and validity in community surveys in the USA 14,22 and Japan. 23 A cut-off score of 15/16 has been used to screen for clinical depression in Japan. 23 Assessment of easiness to use The community respondents were also asked to rate the easiness of completing K6/K10, DSS, and CES-D on a four-point scale (very easy to answer, moderately easy to answer, slightly hard to answer, and very hard to answer). Demographic variables The questionnaire asked participants about their sex and age. Statistical analyses Cronbach s alpha coefficient was calculated for each of the four scales. Among the patients diagnosed with mood/anxiety disorder, the distributions of K6/K10 scores were obtained. The receiver operator characteristics (ROC) curves were drawn for each scale to

4 Psychiatry and Clinical Neurosciences 2011; 65: Screening performance of K6/K Table 1. Sex and age distributions of the community sample respondents and psychiatric outpatients with mood/anxiety disorders Community sample respondents (n = 147) Psychiatric outpatients with mood/anxiety disorders (n = 17) P-value* Sex Men 47 (32%) 10 P = Women 91 (62%) 7 Unknown 9 (6%) Age (years) (23%) 5 P = (32%) 9 60 or older 55 (37%) 3 Unknown 11 (7%) *We calculated P-value without unknown categories., no case. differentiate patients with mood or anxiety disorders (i.e. psychiatric outpatients) from the community sample with the calculation of AUC. The optimal cut-off points of K6/K10 for mood/anxiety disorder were calculated in two ways, using the Youden index to select an optimal cut-off point that would yield the largest sum of the sensitivity and specificity 28 and using the shortest distance to upper left corner method to select a cut-off point that would yield a position closest to the upper left corner on ROC, coordinates (0.1). 29 Sensitivity and specificity based on the optimal cut-off points were compared across the scales. A similar series of analyses was also conducted separately for mood disorders (n = 12) and the community sample. The same analyses were not conducted separately for anxiety disorders because of the small number of cases with anxiety disorders (n = 6). The proportion of the respondents who rated the scales very easy or easy was compared between K6/K10 and CES-D and between DSS and CES-D (Wilcoxon rank sum test). All analyses were conducted using SPSS version 14 for Windows. RESULTS The community sample respondents consisted of a larger proportion of women than men, and of a similar age distribution with the psychiatric outpatients with mood/anxiety disorders (Table 1). Psychiatric outpatients with mood/anxiety disorders had significantly greater average scores on K6, K10, DSS, and CES-D compared to the community respondents (all P < 0.001) (Table 2). Cronbach s alpha coefficients for these scales were 0.91 for K10, 0.85 for K6, 0.77 for DSS, and 0.86 for CES-D. Among the patients diagnosed with mood disorder, the mean (standard deviation) scores were 12.8 (4.3) for K6, 20.9 (5.9) for K10, 3.9 (1.2) for DSS, and 32.8 (8.9) for CES-D. CES-D showed the greatest AUC for screening mood/anxiety disorders, followed by K6 and K10 (Fig. 1a). The AUC for DSS was the lowest. For mood/anxiety disorders, the AUC was 0.93 (95%CI, ) for K6, 0.94 (95%CI, ) for K10, 0.87 (95%CI, ) for DSS, and 0.95 (95%CI, ) for CES-D. DSS had a significantly smaller AUC compared to CES-D (Z = 2.23, P = 0.027); the AUC for K6 or K10 was not significantly different from that of CES-D (P > 0.05). The pattern was similar when it was applied to screening only mood disorders (Fig. 1b). The AUC for mood disorder only was 0.94 (95%CI, ) for K6, 0.94 (95%CI, ) for K10, 0.90 (95%CI, ) for DSS, and 0.95 (95%CI, ) for CES-D. The AUC for K6, K10, or DSS was not significantly different from that of CES-D (P > 0.05). Table 2. Mean scores of each scale among the community sample and outpatients with mood/anxiety disorder Community sample respondents (n = 147) Psychiatric outpatients with mood/anxiety disorder (n = 17) CES-D Mean SD Mean SD P-value K P < K P < DSS P < CES-D P < CES-D, Center for Epidemiologic Studies Depression Scale; DSS, Depression and Suicide Screen; K, Kessler.

5 438 K. Sakurai et al. Psychiatry and Clinical Neurosciences 2011; 65: (a) Table 3. Sensitivity and specificity for screening mood/ anxiety disorder by K6, K10, DSS, and CES-D among community samples and outpatients with mood/anxiety disorder* Optimal cutoff-point Sensitivity (%) Specificity (%) Prevalence of screened cases in the community sample (%) Sensitivity K6 4/ K6 12/ K10* 9/ DSS 1/ CES-D 15/ (b) 1 - Specificity *Based on 17 cases with mood/anxiety disorder (outpatients) and 147 non-cases (a community sample). The optimal cut-off points for K6 and K10 that maximize the sum of sensitivity and specificity based on our data. The cut-off points for K6, DSS, and CES-D based on Kessler et al. 4, Fujisawa et al. 20, and Shima et al. 23, respectively. CES-D, Center for Epidemiologic Studies Depression Scale; DSS, Depression and Suicide Screen; K, Kessler. Sensitivity 1 - Specificity Figure 1. (a) The receiver operater characteristic (ROC) curve for ( ) Kessler (K)6, ( ) K10, ( ) Depression and Suicide Screen (DSS) and ( ) Center for Epidemiologic Studies Depression Scale (CES-D) differentiating outpatients having mood or anxiety disorders from the community sample (outpatients [n = 17], community samples [n = 147]). (b) The ROC curve for ( ) K6, ( ) K10, ( ) DSS and ( ) CES-D scale differentiating outpatients having mood disorders from the community sample (outpatients [n = 13], community samples [n = 147]). The Youden index and the shortest distance to upper left corner method provided the same optimal cut-off point for each scale. The optimal cut-off point was 4/5 on K6, 9/10 on K10, with sensitivity of 100% and moderate-to-high specificity (Table 3). The specificity for K6/K10 was slightly greater than the one for CES-D. DSS also showed high sensitivity but its specificity was moderate (60%), which was the lowest among the four scales. Prevalence of the positives was similar for K6 (with the cut-off of 4/5), K10, and CES-D, ranging from 25% to 30%. DSS yielded a higher prevalence. The community respondents rated K6/K10 and DSS as easier to answer compared to CES-D. The proportion of the respondents who rated the scales very easy or easy to complete was 87.8% for K6/K10, 85% for DSS, and 51.3% for CES-D. There was a significant difference between K6/K10 and CES-D and between DSS and CES-D (P < 0.001, Wilcoxon rank sum test). DISCUSSION The present study demonstrated that K6 and K10 have high screening performance for mood and anxiety disorders, equal to that of CES-D. The screening performance (AUC) of DSS was lower than that

6 Psychiatry and Clinical Neurosciences 2011; 65: Screening performance of K6/K of K6, K10, and CES-D, and significantly different from CES-D. The internal consistency reliability of K6 and K10 was sufficiently high. The present study showed high screening performance (AUC) of K6 and K10 for mood and anxiety disorders ( ), with AUC values similar to those reported in a previous study in Japan 19 and even greater than those reported in the USA. 18 As reported previously, almost no difference in AUC was observed between K6 and K The AUC for K6 and K10 were quite similar to that of CES-D. With the optimal cut-off points, the sensitivity was 100% for K6 and K10. The specificities of the two scales were similar, only slightly greater for K10 than for K6. The sensitivity and specificity of K6, K10, and CES-D were quite comparable, while the specificity of K10 was even greater than that of CES-D. The screening performance of K6 and K10 is considered acceptable for the use in the community 30 and similar to that of CES-D. When we compared the respondents burden to complete these scales, the community residents reported that K6 and K10 were easier to complete than CES-D. This is probably because K6/K10 consists of fewer items than CES-D but also because, unlike CES-D, K6/K10 does not have reverse scored items, which might be difficult for respondents to answer. The K6 and K10, particularly K6, have the advantage of brevity over the CES-D as they consist of only six or 10 multiple-choice questions that perform very similarly. In this study, the optimal cut-off point was 4/5 and 9/10 for K6 and K10 based on the Youden index 28 or the shortest distance to upper left corner method both assuming that the sensitivity and specificity are equally important. These cut-off points are close to the K6 and K10 scores for which the stratum-specific likelihood ratios started to increase. 19 A previous study stated that the Youden index or the shortest distance to upper left corner method are the most suitable methods to determine an optimal cut-off point because they are least dependent on population prevalence. 31 The cut-off point of 4/5 for K6 and 9/10 for K10 may be used in the screening of mood/anxiety disorders in the general population. In addition, these cut-off points for K6 and K10 seem to yield estimates of depression or psychological distress prevalence similar to estimates based on CES-D, as the prevalence in the community sample is quite close. K6 and K10 with these cut-off points could also be used in epidemiological studies on depression or psychological distress, as the CES-D and General Health Questionnaire-12 have been used for this purpose in many studies. However, the optimal cut-off score varies depending on the purpose of the screening, prevalence in a target population, cost-effectiveness, and other factors. To screen or monitor severe mental illness, a previous study proposed a cut-off point of 12/13 for K6, 18 which was determined by the prevalence-matching method 31 for selecting a cut-off that would produce a prevalence of positives similar to the prevalence of a target disorder when favoring specificity over sensitivity. In the present study, the specificity was high for this cut-off point, while the sensitivity substantially decreased. The cut-off point of 12/13 for K6 may be useful if a screening program targets severe mental illness, has limited resources for the secondary screening (such as manpower of health care professionals), and thus expects a high postprobability of the disorders in the positives. DSS also showed high internal consistency reliability, as reported in a previous study. 20 Similarly to K6 and K10, the community sample rated DSS as a less burdensome instrument compared to CES-D. However, DSS showed moderate AUC, which was significantly smaller than that of CES-D but nonsignificantly smaller than that of K6 and K10. The specificity of DSS was particularly lower than that of the other scales. As a result, two out of five community respondents screened positive on DSS. The specificity was even lower than the one reported for DSS in a previous study. 20 This is probably because the previous study used cases defined by a different screening questionnaire. Another possibility is that the current respondents were adults and not the elderly, which was the case in the previous study. It might be beneficial to amend DSS to increase the screening performance, specifically the specificity. This study has several limitations. First, in this survey, both outpatient and healthy controls were selected from different settings, while the study was conducted in the same area, Okayama city. However, the patients selected from the clinics may not be representative of all the patients in this population; the community sample may be biased because of non-response to the survey. These two samples may differ in some demographic and behavioral variables, such as social class and help-seeking styles. It is most desirable to survey a random sample of the whole population with a structured clinical interview to identify cases with disorders. Second, since the community sample was expected to include a certain percentage of those who suffered from mood and

7 440 K. Sakurai et al. Psychiatry and Clinical Neurosciences 2011; 65: anxiety disorders, 2 this might result in underestimation of the specificity, thus also AUC, of each scale. The specificity and AUC estimated in the study should represent a minimum value. If people with current mood and anxiety disorders were more or less likely to respond to the survey compared to other people, the response rate also might affect the findings. Third, the present study used outpatients diagnosed with DSM-IV mood or anxiety disorders and not individuals with mood or anxiety disorders selected from a community. As the clinical sample probably included only definite cases, not cases meeting the diagnostic threshold, the sensitivity and thus AUC were likely to be overestimated. Fourth, the clinical diagnosis in this study, made by a welltrained medical doctor with more than 10 years of experience, is not gold standard unless a clinical psychiatric interview schedule with a longitudinal assessment is used. However, clinician diagnosis according to DSM-IV has often been used in research or international survey on the screening performance of depression scales. 32 Also, one of the intended purposes of DSM-IV is for screening cases in study, not only clinical usage. It has appropriate reliability when a doctor with sufficient skills uses it. Fifth, the clinical sample was small. The estimated screening performance (AUC, sensitivity, and specificity) could be statistically unstable. Sixth, there may be a sex difference in the screening performance of K6/K10 and DSS. The present study could not investigate the screening performance by sex. Despite these limitations, the present study supports high performance, equal to that of CES-D, of the Japanese versions of K6 and K10 in screening for mood and anxiety disorders with the optimal cut-off points of 4/5 for K6 and 9/10 for K10. However, K6 and K10 were more acceptable to community residents than CES-D. K6 and K10 might be useful in screening for mood and anxiety disorders in the community population in Japan compared to DSS. DSS might be useful in the community elderly population if it were amended in order to increase the specificity. ACKNOWLEDGMENTS The present study was conducted initially supported by a Grant for Research on Psychiatric and Neurological Diseases and Mental Health (H16-KOKORO-011) from the Japan Ministry of Health, Labour, and Welfare. The preparation of the manuscript was partly supported by the Nakajima Foundation. REFERENCES 1. Murray CJL, Lopez AD. The Global Burden of Disease and Injury Series, Volume 1: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to Harvard University Press, Cambridge, MA, Kawakami N, Takeshima T, Ono Y et al. Twelve-month prevalence, severity, and treatment of common mental disorders in communities in Japan: preliminary finding from the World Mental Health Japan Survey Psychiatry Clin. Neurosci. 2005; 59: Cabinet Office, Government of Japan. White paper on the prevention of suicide Jijigahosha 2009; (in Japanese). 4. Kessler RC, Merikangas KR, Berglund P, Eaton WW, Koretz DS, Walters EE. Mild disorders should not be eliminated from the DSM-V. Arch. Gen. Psychiatry 2003; 60: Kessler RC, Akiskal HS, Ames M et al. Prevalence and effects of mood disorders on work performance in a nationally representative sample of U.S. Workers. Am. J. Psychiatry 2006; 163: Sanderson K, Andrew G. Common mental disorders in the workforce: recent findings from descriptive and social epidemiology. Can. J. Psychiatry 2006; 51: Ono Y. Suicide prevention program for the elderly: the experience in Japan. Keio J. Med. 2004; 53: Sakamoto S, Tanaka E, Neichi K, Ono Y. Where is help sought for depression or suicidal ideation in an elderly population living in a rural area of Japan? Psychiatry Clin. Neurosci. 2004; 58: Chida F, Okayama A, Nishi N, Sakai A. Factor analysis of Zung Scale Scores in a Japanese general population. Psychiatry Clin. Neurosci. 2004; 58: Langner TS. A twenty-two item screening score of psychiatric symptoms indicating impairment. J. Health Hum. Behav. 1962; 3: Goldberg DP, Williams P. A User s Guide to the General Health Questionnaire. NFER-NELSON, Berkshire, NY, Zung WWK. A self-rating depression scale. Arch. Gen. Psychiatry 1965; 12: Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983; 67: Weissman MM, Sholomskas D, Pottenger M, Prusoff BA, Locke BZ. Assessing depressive symptoms in five psychiatric populations: a validation study. Am. J. Epidemiol. 1977; 106: Kessler RC, Andrews G, Colpe LJ et al. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol. Med. 2002; 32: Ministry of Health, Labour and Welfare, Japan. Comprehensive Survey of Living Condition of the People on Health and

8 Psychiatry and Clinical Neurosciences 2011; 65: Screening performance of K6/K Welfare Health Welfare Statistics Association, Tokyo, 2008 (in Japanese). 17. Furukawa TA, Kessler RC, Slade T, Andrews G. The performance of the K6 and K10 screening scales for psychological distress in the Australian national survey of mental health and well-being. Psychol. Med. 2003; 33: Kessler RC, Barker PR, Colpe LJ et al. Screening for serious mental illness in the general population. Arch. Gen. Psychiatry 2003; 60: Furukawa TA, Kawakami N, Saitoh M et al. The performance of the Japanese version of the K6 and K10 in the World Mental Health Survey Japan. Int. J. Methods Psychiatr. Res. 2008; 17: Fujisawa D, Tanaka E, Sakamoto S, Neichi K, Nakagawa A, Ono Y. The development of a brief screening instrument for depression and suicidal ideation for elderly: the Depression and Suicide Screen. Psychiatry Clin. Neurosci. 2005; 59: Ministry of Health, Labour and Welfare, Japan. The manual corresponding to depression (in Japanese). Available form URL: manual/utsumanual2.pdf (last accessed 14 June 2011). 22. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl. Psychol. Meas. 1977; 1: Shima S, Shikano T, Kitamura T, Asai M. New self-rating scales for depression. Clin. Psychiatry 1985; 27: Japan Ministry of Health, Labour, and Welfare. Report on trend of health and welfare survey (mental health) Yesavage JA. Geriatric depression scale. Psychopharmacol Bull. 1988; 24: Muraoka Y, Ikuji S, Ibara K. Physical psychological social background of depressive elderly among local residents. Jpn. J. Geriatr. Psychiatry 1996; 7: (in Japanese). 27. Ministry of Health, Labour and Welfare, Japan. The guideline for prevention and support of depression (revised version) (in Japanese). 28. Youden W. An index for rating diagnostic tests. Cancer 1950; 3: Holmes WC. A short, psychiatric, case-finding measure for HIV seropositive outpatients: performance characteristics of the 5-item mental health subscale of the SF-20 in a male, seropositive sample. Med. Care 1998; 36: Akobeng AK. Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatr. 2007; 96: Kelly MJ, Dunstan FD, Lloyd K, Fone DL. Evaluating cutpoints for the MHI-5 and MCS using the GHQ-12: a comparison of five different methods. BMC Psychiatry 2008; 19: Mezzich JE. International surveys on the use of ICD-10 and related diagnostic systems. Psychopathology 2002; 35: