Power and sample size calculations for stochastic cost-effectiveness analysis

Abstract

As the data for economic analyses are increasingly collected prospectively alongside clinical trials, many commentators have highlighted that the sample sizes in such trials should be based on the requirements for the economic analysis as well as those for the clinical evaluation. However, issues associated with sample size calculations for economic analysis have yet to receive the rigorous attention given to sample size calculation for clinical evaluation. In particular, no sample size formula for cost-effectiveness analysis is available for analysts hoping either to calculate the required sample size at the design stage of a study or to calculate the power a given size of clinical trial will generate for cost-effectiveness analysis. Building on the recent literature for calculating confidence intervals for cost-effectiveness ratios, the authors explore possible techniques for deriving a sample size formula for cost-effectiveness analysis based on simple combination of the confidence limits on costs and effects.