Providing systematic guidance in pharmacoeconomic guidelines for analysing costs

Despite the considerable amount of work that has gone into the standardisation of costs, a recent survey indicated that there is still a wide variation in methods used in costing studies in the published literature. Part of the answer to the lack of standardisation may lie in pharmacoeconomic guidelines themselves.

The purpose of this paper is to categorise the current recommendations for the analysis of costs within an economic evaluation, and to use these categories to assess variations in guidelines.

Based on identified costing methodology, we developed a detailed categorisation. We then surveyed existing guidelines to determine similarities and differences in the guidelines.

Our framework had five categories: general design items; resource identification and classification; resource measurement; resource valuation; and overall reporting. We noted differences in perspective, some of which were due to methodological uncertainty and differences in the purposes of the studies covered by the guidelines. There was very little guidance in how to put the basic cost measure into operation. In addition, the preferred concept for resource utilisation was that of resource use under actual (rather than experimental) conditions, and there was very little guidance as to how to achieve this measure.

There are still large variations, as well as gaps, in guidelines in the costing area. This is due, in part, to wide-ranging objectives for studies, methodological uncertainty and a neglect of the fundamentals of costing. We recommend that the contents of the basic measure of 'average cost' should be explained in more detail, with recurring, capital and allocated overhead costs being reported separately. A target capacity level (e.g. 80%) should be reported, even if in a sensitivity analysis. More guidance should be provided regarding the translation of utilisation under experimental care to that under routine care. For resource valuation, adjustments to the 'average cost' measure should be made explicit. There should be more guidance on how to project system-wide costs from the costs of one hospital. The manner in which overall results are reported should be addressed, including the identification of how costs should be categorised.