In a five-year prospective follow-up study of 115 psychiatrically hospitalized patients, interinterview reliability for depressive symptoms was 94%. Specific interrater diagnostic reliability was 80%, using explicit diagnostic criteria. Disagreement was solely on degree of certainty within a specific diagnosis, not between different diagnoses. Interrater reliability was 100% for presence or absence of affective disorder.
Follow-up was blind. Of 52 patients initially diagnosed by explicit criteria as having primary affective disorder, 43 were followed up. Blind diagnostic agreement was 86%. An additional 9% not concordant by blind diagnosis fell within the usual clinical concept of primary affective disorder. Only two patients (5%) had a course incompatible with the natural history of primary affective disorder. To our knowledge this is the first blind, prospective validation of criteria for the diagnosis of primary affective disorder.