A cost-effectiveness analysis to illustrate the impact of cost definitions on results, interpretations and comparability of pharmacoeconomic studies in the US.

BACKGROUND: There is a lack of a uniform proxy for defining direct medical costs in the US. This potentially important source of variation in modelling and other types of economic studies is often overlooked. The extent to which increased expenditures for an intervention can be offset by reductions in subsequent service costs can be directly related to the choice of cost definitions.
OBJECTIVES: To demonstrate how different cost definitions for direct medical costs can impact results and interpretations of a cost-effectiveness analysis.
METHODS: The IMS-CORE Diabetes Model was used to project the lifetime (35-year) cost effectiveness in the US of one pharmacological intervention 'medication A' compared with a second 'medication B' (both unspecified) for type 2 diabetes mellitus. The complications modelled included cardiovascular disease, renal disease, eye disease and neuropathy. The model had a Markov structure with Monte Carlo simulations. Utility values were derived from the published literature. Complication costs were obtained from a retrospective database study that extracted anonymous patient-level data from (primarily private payer) adjudicated medical and pharmaceutical claims. Costs for pharmacy services, outpatient services and inpatient hospitalizations were included. Cost definitions for complications included charged, allowed and paid amounts, and for medications included both wholesale acquisition cost (WAC) and average wholesale price (AWP). Costs were reported in year 2007 values.
RESULTS: The cost-effectiveness results differed according to the particular combination of cost definitions employed. The use of charges greatly increased costs for complications. When the analysis incorporated WAC medication prices with charged amounts for complication costs, the incremental cost-effectiveness ratio (ICER) for medication A versus medication B was $US6337 per QALY. When AWP prices were used with charged amounts, medication A became a dominant treatment strategy, i.e. lower costs with greater effectiveness than medication B. For both allowed and paid scenarios, there was a difference in the ICER of over $US10,300 per QALY when medication prices were defined by WAC versus AWP. Ratios of medication costs to cardiovascular complication costs ranged from under 0.45 to over 1.7, depending upon the combination of costing definitions.
CONCLUSIONS: Explicitly addressing the cost-definition issue can help provide meaningful cost-effectiveness data to payers for policy development and management of healthcare expenditures. It can also help move the pharmacoeconomics and outcomes research fields forward in terms of both methodology and practical application.