Eur Arch Psychiatry Clin Neurosci (2004) 254 : 36 – 42 DOI 10.1007/s00406-004-0452-y ■ Abstract Background Common mental disorders (CMD) are critical issues in community studies. The 12- item version Chinese Health Questionnaire (CHQ-12) has not been sufficiently explored for use in Chinese communities around the world. This paper evaluates CMD among ethnic Chinese living in Brazil, by examin- ing its psychometric properties and factorial structure of CHQ-12. Method For 211 community subjects who completed the CHQ-12, we used the ROC curve analysis to calculate its internal consistency and criterion valid- ity, adopting the interview SCAN as criterion. Principal component analysis was performed. Results Conside- rable evidence was found for the internal consistency (Cronbach’s α) and item-total correlation of CHQ-12. The best cut-off point was 2/3, which allowed calculating criterion validity: 75 %, sensitivity, 71 % specificity, 55 % PPV, 86% NPV, 28% misclassification rate and area under the ROC curve 0.728 (SE: 0.102, 95 % CI: 0.528–0.927). The 3-dimensional structure of CHQ-12 could be described as somatic, depressive and worrying factor, accounting for 47.8% of the total variance. Con- clusions When applied to the Brazilian Chinese community, the CHQ-12 presented comparable psy- chopathologic pattern, psychometric properties and factorial structure to that of Taiwan. ■ Key words mental health · screening · reliability and validity · Chinese · Brazil Introduction The nature and dimensional structure of common men- tal disorders (CMD) among culturally diverse commu- nity samples has recently attracted much interest of in- ternational cross-cultural epidemiology (Jacob et al. 1998; Wernecke et al. 2000). The symptom profile of CMD among community respondents can allow the evaluation of the psychopathological differences be- tween Chinese and Western individuals, a critical issue for research. Since its introduction in 1970, the General Health Questionnaire (GHQ) has been shown to be a useful screening method aimed to detect non-psychotic men- tal disorders in general clinical settings (Goldberg and Blackwell 1970). The GHQ has been translated into many languages and regarded as applicable in many countries, with good validity indicators (Goldberg et al. 1997, 1998). Cheng and Williams (1986) translated and recon- structed the 60-item English version of the GHQ after discriminatory analysis, adapting it to the Chinese cul- ture. Renamed as the Chinese Health Questionnaire (CHQ), its applicability to detect CMD in overseas sam- ples has been reported in Chinese communities of the United Kingdom (Pan and Goldberg 1990) and New Zealand (Cheung and Spears 1993; Abbot et al. 1999). Al- though Cheng et al. (1990) explored the psychometric properties and factorial structure of CHQ-30, there is no study on its factorial structure or on the concurrent val- idity of CHQ-12. This paper aims to report the sympto- matic pattern of a Chinese community sample of Brazil from the results of CHQ-12,by examining its psychome- tric properties, factorial structure and concurrent valid- ity. ORIGINAL PAPER Yuan-Pang Wang · Clarice Gorenstein · Laura H.Andrade Patterns of psychopathological manifestations among ethnic Chinese living in Brazil Received: 20 May 2003 / Accepted: 20 August 2003 EAPCN 452 Dr. Yuan-Pang Wang () · C. Gorenstein · L. H. Andrade Instituto de Psiquiatria Hospital das Clínicas Faculdade de Medicina Universidade de São Paulo R. Dr. Ovídio Pires de Campos s/n. Caixa Postal 8091 CEP 05403-010, São Paulo, SP, Brazil Tel.: +5511/5087-9474 Fax: +5511/5087-9475 E-Mail: gnap_inbox@hotmail.com Y.-P. Wang Faculdade de Medicina de Santo Amaro UNISA, Brazil