Eighty-one patients with advanced breast cancer completed the Hospital Anxiety and Depression Scale (HADS) and Rotterdam Symptom Checklist (RSCL) to determine how well these questionnaires identified patients suffering from an anxiety state or depressive illness, compared with an independent interview by a psychiatrist who used the Clinical Interview Schedule. A threshold score was defined for each questionnaire which gave the optimal sensitivity and specificity. Seventy-five per cent of patients were correctly identified as suffering from an affective disorder by both the Rotterdam Symptom Checklist and by the Hospital Anxiety and Depression Scale. Twenty-one per cent of 'normal' patients were misclassified by the Rotterdam Checklist and 26% by the Hospital Anxiety and Depression Scale. When the HADs anxiety and depression subscales were analysed separately, the performance of the anxiety items was superior to that of the depression items. Both questionnaires were found to have good predictive value and could be used in patients with advanced cancer to help screen out those with an affective disorder.