It is hypothesized that variations in the processes and structures of the selection of patients and the conduct of the surgical procedure may influence risk-adjusted outcome in patients undergoing cardiac surgery. For this reason, the results of the pilot phase of this Veterans Affairs cooperative study were reviewed to determine the variation in the operative practices at six pilot institutions. There were large variations in the percentage of elective, urgent, and emergent cases at each institution, ranging from 58% to 96% elective, 3% to 31% urgent, and 1% to 8% emergent. There was also a tenfold increase in the preoperative use of intra-aortic balloon pump for control of unstable angina, varying from 0.8% to 10.6%. Five of the six centers had accredited thoracic surgical residency programs. There was large variation in the preoperative participation of attending surgeons, from 100% participation at three centers to less than 5% in one. The operation was performed by the attending surgeon in 28% of cases, but this varied from 0% to 100%, depending on the hospital. Cold cardioplegia was used almost uniformly (99%) for myocardial protection; the use of retrograde cardioplegia varied from 2% to 89% among hospitals, and the use of blood cardioplegia ranged from 0% to 100%, with an average of 54% of cases. The use of myocardial temperature monitoring varied between hospitals, from 25% to 99%. The use of the cell saving devices to scavenge shed blood varied from 5% to 99%, and the frequency of the use of banked blood varied from 25% to 65%, depending on the institution. The internal mammary artery was used for 67% patients undergoing coronary artery bypass graft, with a variation between hospitals of 39% to 83%. One hospital used a single cross-clamping technique for the performance of proximal anastomoses in 95% of cases, as opposed to all other hospitals, who used this technique in less than 10% of cases. Aortic venting varied from 58% to 98% and left ventricular venting from 1% to 38%. The use of porcine valves varied approximately 15% in three hospitals to 30% to 40% in the other three hospitals. There were tremendous variations in the duration of operative procedure (5.2-7.3 hours), actual operating time (4.0-5.6 hours), total cardiopulmonary bypass duration (102-146 minutes), and ischemic time (50-87 minutes). The use of inotropic support varied from 41% to 91% between hospitals.(ABSTRACT TRUNCATED AT 400 WORDS)