This communication examines the effect of a lengthy patient history interview in modifying a diagnosis made after a thorough examination of the patient's present mental state. The diagnoses given to 396 patients after a mental state examination were compared with those given after a subsequent patient history interview. Changes across major diagnostic categories occurred for only 15% of the sample; most of these changes occurred when there was uncertainty present in the mental state diagnosis. There was little change in the degree of confidence with which the mental state and patient history diagnoses were expressed, and little change in the number of subsidiary diagnoses given. The diagnostic distribution of the entire sample after the patient history interview was remarkably similar to that after the mental state examination.