Optimal treatments in cost-effectiveness analysis in the presence of covariates: Improving patient subgroup definition

In the presence of covariates, the cost-effectiveness analysis of medical treatments shows that the optimal treatment varies across the patient population subgroups, and hence to accurately define the subgroups is a crucial step in the analysis. A patient subgroup definition using only influential covariates within the potential set of patients covariates established by the expert has recently been proposed, and the influential covariates were chosen from the univariate distributions of the effectiveness and the cost, conditional on the effectiveness. In this paper, we argue that the Bayesian variable selection procedure should be developed using the bivariate distribution of the cost and the effectiveness, which is not the usual practice. This new approach, provides results with wider applicability and more understandable without a significative increase in the complexity of the procedure. For real and simulated data sets, optimal treatments for subgroups are found, and compared with that from previous methods.

► A novel approach to patient subgroup definition in a bivariate costeffectiveness model is considered.
► Using objective Bayesian tools, subgroup optimal treatments are derived.
► Illustrations show that the proposed approach is easily implementable.