A 'league table' of contingent valuation results for pharmaceutical interventions: a hard pill to swallow?

Pharmaceutical expenditure represents a large percentage of total healthcare expenditure, and has thus received much attention within the economic evaluation literature. However, although the number of contingent valuation (CV) studies measuring willingness to pay (WTP) in healthcare has increased, little is known about the relative magnitude of values elicited across different interventions, diseases or countries, or the methodological comparability of these values. We address this gap by seeking to establish if it is feasible to use elicited WTP values in resource allocation, illustrated by attempting to compile a 'league table' of WTP values for pharmaceutical interventions. A review database was compiled for CV studies in healthcare published from January 1985 to December 2005. Of 210 studies identified, 40 considered pharmaceutical interventions. Values are presented as mean or median WTP values, adjusted where necessary to pound and $US for 2004/5. Lack of reporting in some instances of either the mean or median, together with heterogenous methods and infrequent reporting of costs, made 'league table' construction difficult. This raises questions about the use of existing studies for resource allocation decisions, despite the fact that most studies were seemingly undertaken for policy objectives. However, four interventions had more than one study, making it possible to compare the values elicited. The values elicited across studies were fairly consistent for two interventions (anti-hypertensive therapy and tumour necrosis factor [TNF]-alpha blockade for rheumatoid arthritis), whereas WTP values for insulin and post-operative emesis therapy were very divergent. No single methodological difference seemed to explain this pattern; however, the more methodological differences between studies the greater the likelihood of divergent values. A checklist, or minimum reporting set of information, is the first step towards improving the consistency of methods, and therefore values, published. In the longer term, a move towards the use of a reference case akin to that used for cost-utility studies would seem important if such studies are to be used for comparative purposes and thereby be relevant to resource allocation decision making.