In this prospective, 1-year study, 360 males admitted to an inpatient alcoholism treatment program were administered a DSM-III compatible structured interview and subtyped by co-occurring psychiatric disorder. Forty percent satisfied diagnostic criteria for alcohol dependence while 27% met criteria for alcohol dependence and one additional psychiatric syndrome. The dually diagnosed patients were divided into: alcohol dependence plus drug abuse, alcohol dependence plus antisocial personality and alcohol dependence plus depression. These subtypes were compared on multiple dimensions at intake and at 1-year follow-up. At follow-up, all groups showed significant improvement in drinking and psychosocial functioning. The results suggest that subtyping alcoholics by co-morbid psychiatric disorders may be a good postdictor of clinical history, but a poor predictor of drinking outcome.