Misclassification by psychiatric screening questionnaires.

Two psychiatric screening instruments were applied to a sample of attenders (n = 875) in three primary medical care clinics in the city of Sao Paulo. The two questionnaires were validated in a subsample (n = 260) against the Clinical Interview Schedule [8]. The effect of sociodemographic variables on misclassification (by both questionnaires) was investigated using Relative Operating Characteristic (ROC) analysis and linear logistic modelling. Both these techniques proved to be useful in this regard, and it is recommended that they be more widely used in screening studies. Sex and educational level exerted important influences on misclassification. In general, men were more likely than women to be classified as false negative, while poorly educated respondents were more likely to be classified as false positive. These results are interpreted in an illness-behavior context.