Choosing an instrument to present the results of the economic assessment of innovation? Measuring Health benefit of an innovative treatment

Abstract

Measuring the Health benefits of an innovative treatment requires the strict indications and precautions. The user’s guide for the technical report of an ex-ante evaluation that is currently developing by HAS recommends presenting the results with the cost-effectiveness acceptability curve for one comparator and with the health benefits and the cost-effectiveness acceptability frontier in the case of multiple treatment alternatives. When the health economic evaluations are conducted using individual patient data in the context of a stochastic model, the uncertainty related to sampling fluctuations and variability of the measurement can be quantified by confidence intervals. The theorem of Fieller and the nonparametric methods of resampling are preferred tools to calculate the confidence interval of the incremental cost-effectiveness ratio (ICER). However, a decision maker cannot ignore the problems of cost-effectiveness evaluation when the bootstrap simulations are scattered in all four quadrants of the cost-effectiveness plane. The questions remain when uncertainty is explored in the context of a decision analysis model. A Bayesian approach allows associating of each parameter of the model to the specific probability distribution and providing numerous simulations. In both, stochastic and decision approaches, the cost-effectiveness acceptability curve and the net health or monetary benefits can overcome the difficulties of calculating the confidence interval of ICER. The objective of this article is to conduct an explicit description of these tools.