Background. The General Health Questionnaire (GHQ) has been validated in different languages and cultures and in diverse settings. However, the validity of the 12-item version, increasingly used for screening for psychiatric morbidity in primary care, has not been established among ethnic Indians living in the United Kingdom.

Methods. The GHQ-12 was used to screen for psychiatric morbidity in a study of patterns of consultation and explanatory models of mental illness in a general practice in West London. All individuals who scored 2 or more and an equal number of individuals who scored 0 or 1 were interviewed using the Revised Clinical Interview Schedule (CIS-R) to confirm psychiatric morbidity. Hindi versions of the both these instruments were also employed. Thresholds of GHQ were compared against the standard of the CIS-R using the recommended threshold of 12 and above as indicating caseness. A receiver operator characteristic curve was drawn to obtain the best threshold value for screening.

Results. The optimal threshold for screening as assessed by receiver operator characteristic analysis was 2/3. This threshold had a sensitivity of 96·7% and a specificity of 90%.

Conclusions. The sensitivity and specificity of the 12-item General Health Questionnaire among women of ethnic Indian origin living in the United Kingdom is high. It can be employed as a screening instrument to identify individuals with psychiatric morbidity in this population.