This is a cross-sectional community study conducted to assess the one-year prevalence of psychotropic use in the city of São Paulo. A representative stratified sample of the city was drawn from three sub-districts selected on the basis of their health indicators (Ramos & Goihman, 1989). The probability of a psychiatric disorder was estimated by means of the Adult Psychiatric Morbidity Questionnaire (QMPA) developed by Santana (1982). The total sample comprised 1742 subjects: 11·7% of males and 24·6% of females were probable cases in the QMPA, at the cut-off point 7/8. The overall psychotropic consumption was 101·6 persons/1000 inhabitants. The rate of psychotropic use was higher for females (142·3 persons/1000 inhabitants) than males (50·0 persons/1000), a difference statistically significant (χ2 = 18·0, 1 df, P < 0·001). The highest rate of consumption was for tranquillizers (80·4/1000 inhabitants) and the general physician was found to be the leading prescriber (46·9%), being followed by cardiologists (15·3%). A log-linear model was constructed to study the combined effect of sociodemographic factors on the probability of being a tranquillizer user. Women were found to take more tranquillizers than men, consumption increased with age, and the positives in the QMPA were more likely to be users than were the negatives. The higher the family income per capita the higher the risk of being a tranquillizer user. These findings applied regardless of the sub-district, marital status, and migration status of the subjects. These results are discussed in the light of the alternative possible interventions by general practitioners.