OBJECTIVES: The overall aim of this investigation was to determine the best use of post-transplant immunosuppression therapies, in terms of clinical and economic outcomes, for the liver transplant population at the University of Colorado Hospital (UCH), regardless of expected reimbursement. METHODS: Patients were sequentially assigned to either tacrolimus (FK) or emulsified cyclosporine (CYA) with or without mycophenolate mofetil (MMF) yielding four treatment arms: MMF + CYA (60), MMF + FK (51), CSA (39), and FK (41). All patients were followed for one year post-transplant. Using intent to treat or crossover analysis outcomes were similar between FK- and CYA- treated groups, allowing comparison of MMF patients (111) to no-MMF patients (80). Costs (included: hospital, clinic, emergency department, outpatient immunosuppression; excluded: physician fees) were obtained using relative value units (RVUs, microcosting), and wholesale acquisition prices. Ratios of cost-to-charges (RCCs) for cost-estimate method comparisons were obtained from a Medicare cost report. RESULTS: All treatment arms had similar severity and chronology of rejection episodes. In addition, there were no significant differences between treatment arms when actual costs (RVUs), or other cost-estimation methods were used. Cost-center specific ratio of costs to charges (RCCs) performed better than global hospital RCCs, but both were significantly different from, and underestimated, RVU costs. Pharmacy appeared to be the reason why cost-center specific RCCs did not perform well. There were no significant differences between cost-center specific RCCs and RVU costs when pharmacy RCCs were adjusted. CONCLUSIONS: There is not a preferred immunosuppressive treatment substitute (MMF + FK, MMF + CYA, FK, CYA) for the post-liver transplant population at UCH. Traditional cost-estimate methods do not accurately approximate costs. A pharmacy RCC adjustment may be required for disease states that use significant inpatient pharmacy resources. Cost-center specific RCC method cost-estimates may be improved with this methodology.