Prediction of delayed psychiatric disorders in breast cancer patients by using a screening procedure was investigated. Two questionnaires, the Psychological Distress Inventory and the Hospital Anxiety and Depression Scale, were administered before and during chemotherapy, and at the first follow-up visit. A psychiatric diagnosis was assigned to 50 of the 132 patients (38%) evaluated at follow-up. Including a set of clinical and demographic variables in a logistic regression, increasing age (P=0.001) and psychiatric history (P<0.001) were associated with psychiatric morbidity at follow-up. The accuracy of the two questionnaires in predicting delayed psychiatric disorders increased from the evaluation before chemotherapy to the evaluation during chemotherapy. The most accurate prediction was observed for the concurrent evaluation at follow-up. The accuracy of three predictive models developed for each evaluation point by including age, psychiatric history and psychological distress measured with each of the two questionnaires was not significantly better than that observed using only the questionnaires' scores as predictors.