Direct and indirect costs of medical and of surgical treatment are presented for patients entered into the Birmingham portion of the Coronary Artery Surgery Study. For comparison, similar results are shown for the Birmingham portion of the national Cooperative Unstable Angina Study. In the Unstable Angina Study, mean inpatient costs at the end of 1 year in the study were $6867 for medical therapy, $10,574 for surgical therapy and $23,045 for those who failed medical therapy and required late surgery. A stepwise multiple regression analysis shows that the single best predictor of cost was the number of myocardial infarctions that the patient had while in the study. A discriminant-function analysis identified 85% of the medical patients who required late surgery. A significantly lower proportion of surgical than medical patients returned to work. Total inpatient costs for patients in the Coronary Artery Surgery Study (i.e, patients with stable angina) were $3432, $11,100 and $13,554 for medical, surgical and late surgical patients, respectively, for the first year in the study. There was no significant difference in the percentage of medical and surgical patients who were working at the end of 1 year. According to their own perceptions, the surgical group was in the best and the late surgical group in the worst health.