A Bayesian approach to economic analyses of clinical trials: the case of stenting versus balloon angioplasty.

New results about the costs and effects of a new therapy may be weighted with prior information. As such, classical confidence intervals surrounding the costs and effects of a therapy may not reflect the real uncertainties. Bayesian techniques may improve this by formalizing the way that prior information is taken into account in assessing the new evidence. Costs and effects can be analysed separately, but also, when considering the balance between costs and effects, they can be analysed simultaneously. Here, an example is given using data from two trials that compared costs and effectiveness of stent implantation with balloon angioplasty. The Bayesian results make it clear that different prior distributions may lead to different decisions, and it is concluded that even Bayesian analysis may not always reflect the process of capturing the remaining uncertainties.