A Bayesian Net Benefit Approach to Cost-effectiveness Analysis in Health Technology Assessment

The economic literature on cost-effectiveness analysis in the context of decisions by health technology assessment agencies assumes as the quantity of interest a linear combination of the mean of the sampling distribution of the effectiveness and the cost. We argue that this is not always reasonable. Our reasons for this assertion are that (i) treatments are compared on the basis of mean values, and for some useful models the mean of the distribution of the cost, which is conditional on the available data, does not exist, and (ii) even for models for which the mean does exist, it might not constitute an accurate reflection of the distribution. This paper presents a general Bayesian cost-effectiveness analysis of a single treatment, where the quantity of interest is the distribution, conditional on the data, of the net benefit. This approach permits a natural extension to several treatments, which enables us to make a statistical comparison. Illustrations with treatment comparisons for real and simulated data are given.