Psychological Medicine

Original Articles

Detecting common mental disorders in primary care in India: a comparison of five screening questionnaires

V. Patela1a2 c1, R. Arayaa3, N. Chowdharya2, M. Kinga4, B. Kirkwooda1, S. Nayaka3, G. Simona5 and H. A. Weissa1

a1 London School of Hygiene and Tropical Medicine, London, UK

a2 Sangath, Alto-Porvorim, Goa, India

a3 University of Bristol, Bristol, UK

a4 Royal Free and University College Medical School, London, UK

a5 Center for Health Studies, Group Health Cooperative, Seattle, USA


Background Screening of patients for common mental disorders (CMDs) is needed in primary-care management programmes. This study aimed to compare the screening properties of five widely used questionnaires.

Method Adult attenders in five primary-care settings in India were recruited through systematic sampling. Four questionnaires were administered, in pairs, in random order to participants: the General Health Questionnaire (GHQ, 12 items); the Primary Health Questionnaire (PHQ, nine items); the Kessler Psychological Distress Scale (K10, 10 items), and from which we could extract the score of the shorter 6-item K6; and the Self-Reporting Questionnaire (SRQ, 20 items). All participants were interviewed with a structured lay diagnostic interview, the Revised Clinical Interview Schedule (CIS-R).

Results Complete data were available for 598 participants (participation rate 99.3%). All five questionnaires showed moderate to high discriminating ability; the GHQ and SRQ showed the best results. All five showed moderate to high degrees of correlation with one another, the poorest being between the two shortest questionnaires, K6 and PHQ. All five had relatively good internal consistency. However, the positive predictive value (PPV) of the questionnaires compared with the diagnostic interview ranged from 51% to 77% at the optimal cut-off scores.

Conclusions There is little difference in the ability of these questionnaires to identify cases accurately, but none showed high PPVs without a considerable compromise on sensitivity. Hence, the choice of an optimum cut-off score that yields the best balance between sensitivity and PPV may need to be tailored to individual settings, with a higher cut-off being recommended in resource-limited primary-care settings.

(Received March 15 2007)

(Revised August 28 2007)

(Accepted October 10 2007)

(Online publication November 30 2007)