Diagnostic cluster methodology groups patients having similar medical conditions according to their International Classification of Diseases, 9th Revision codes. Episodes of care related to the diagnostic cluster can then be tracked from the claims data to determine the total charges associated with patient management. A retrospective claims analysis using an episode registry database was conducted to determine the 1-year (July 1, 1995, to June 30, 1996) covered charge for statin therapy, the overall cost of treating related cardiovascular (CV) disease, and the cost impact of coadministration of drugs that potentially compete for hepatic metabolism. The three statin treatment groups (lovastatin, pravastatin, and simvastatin) were similar with respect to age, gender, mean number of prescription refills, rate of refill compliance, and prevalence of the coadministration of potentially interacting agents. Before adjustment for severity of illness, there were no significant differences between groups in prescription drugs/services (statin Rx/Svc) or total CV charges. After adjustment for severity of illness, the pravastatin group had the lowest statin Rx/Svc and total CV charges. Within the group with the greatest severity of illness, statin Rx/Svc charges were significantly lower with pravastatin than with lovastatin and simvastatin. The statin Rx/Svc charges were not significantly different between lovastatin and simvastatin. Coadministration of a potentially interacting agent significantly increased both the statin Rx/Svc and total CV charges within the simvastatin-treated group but did not significantly influence costs in the lovastatin- or pravastatin-treated groups. The estimates of direct costs derived from this analysis are consistent with findings in the published literature and demonstrate that pravastatin has cost advantages compared with lovastatin and simvastatin. Diagnostic cluster methodology also generated valuable information regarding drug surveillance and the health care cost impact of potential drug-drug interactions with selected statins.