We examined the effects of Resource-based Relative Value Scale (RBRVS)- and physician diagnosis-related groups (MDDRG)-based payment for anesthesiology services related to surgery by simulating these physician payment reform options. We merged Medicare Part A (hospital) and Part B (anesthesiology) payment data for 7,770 patients for the MDDRG analysis and examined 10,431 surgical procedures for the RBRVS analysis within 27 diagnosis-related groups (DRGs) during the second half of 1986 in 16 hospitals representing different geographic regions, bed size, and teaching status. Assuming budget neutrality (i.e., constant total expenditure for anesthesiology services) and using the proposed methodologies, we simulated RBRVS and MDDRG payments and compared them to current payments for anesthesiology services. Individual surgical procedures demonstrated a two- to more than four-fold variation in duration, accompanied by a similar variation in anesthesiology payments. Within DRGs, there was a three- to ten-fold variation in duration, and a two- to seven-fold variation in anesthesiology payments. Anesthesiology time was highly correlated with surgical time (r = 0.86-0.96). Compared to the current system, RBRVS and MDDRG systems were associated with systematic variations in payments, such that on average, on each case, anesthesiologists practicing in rural and nonteaching hospitals would gain, whereas those in urban or suburban and teaching facilities would lose. After adjusting for complexity of procedure, the distribution of payment gains and losses was a function of duration of surgery, which is not influenced by the anethesiologist. Longer cases of a given surgical procedure result in payment decreases. The results document the importance of retaining a time factor in the payment methodology for anesthesiology services to maintain equitable payment across practice settings--an objective of physician payment reform.