Background Procedure-related costs are of increasing concern in selecting the appropriate procedure for the treatment of coronary artery disease (CAD).Methods and Results To determine what preoperative factors influence total postoperative hospital costs, data on 604 coronary artery bypass graft surgery (CABG) patients from 1990 to 1991 were analyzed.Professional fees were excluded. Hospital costs were computed by multiplying patient charges by the Medicare cost-to-charge ratio used in determining federal reimbursement. Median postoperative cost was $12 912 (range, $7100 to $259 546). Data were analyzed with a semiparametric regression model. Patients dying in the hospital were censored at time of death. There were significant differences among surgeons in costs but no significant differences in operative mortality. Significant risk factors for increased cost after adjusting for surgeon were: older age (P<.0001), lower left ventricular ejection fraction (P<.0001), prior CABG (P<.0001), female sex (P<.0049), no prior percutaneous transluminal coronary angioplasty (P<.0091), increased degree of CAD (P<.0102), black race (P<.0190), and diabetes (P<.032).Conclusions These results suggest that preoperative characteristics have important economic and medical implications.Surgeons should compare their management strategies on the basis of data analysis to determine the most effective practice with regard to mortality and cost. (Circulation. 1994;90(part 2): II-124-II-128.)