OBJECTIVES: To evaluate the extent to which case-mix-adjusted, hospital inpatient costs for pharmaceutical services changed from 2000 through 2002 in short-term, general, non-federal (STGNF) hospitals in the U.S., and to identify hospital characteristics that were significantly associated with such changes. METHODS: Using all-payer data from more than 2500 STGNF hospitals, discharge-level drug costs were estimated using the Medicare Cost Report information on revenue-center-specific cost-to-charge ratios applied to the corresponding hospital charges. Cost estimates were analyzed by all DRGs to identify those in which statistically significant (alpha = 0.05) upward or downward trends were detected. Pharmacy costs relative to other hospital costs were also analyzed, as was the variability across hospitals in case-mix-adjusted average pharmaceutical cost. Hospital characteristics were evaluated as predictors of case-mix-adjusted average pharmaceutical costs, and national estimates of total hospital pharmaceutical costs were estimated for selected DRG-defined clinical groups. RESULTS: Numerous DRGs exhibited significant upward or downward trends in average costs per case for pharmaceutical services from 2000 through 2002. Significant trends over time in the proportion of total hospital costs that were attributable to pharmaceutical services were also detected. Extensive variability across hospitals was detected in both the total costs that were due to pharmaceutical services, as well as the proportion of the total costs that were due to pharmaceutical services. Many DRGs exhibited substantial increases in the average dollar amount of costs for drug treatment from 2000 through 2002. CONCLUSIONS: Inpatient costs for pharmaceutical services have increased substantially in recent years relative to other hospital services areas, such as laboratory and diagnostic radiology. Further, there exists extensive variability in the intensity of pharmaceutical services that are provided, even after case-mix adjustment. Finally, numerous hospital characteristics, such as teaching status and size, are significantly correlated with the relative utilization of pharmaceutical services for the treatment of inpatients.