In WP 3 we will develop methods to predict patients’ responses to targeted treatments based on genetic features or other biomarkers such that subgroups of patients for which the benefit risk balance of a treatment is positive can be identified and confirmed. While it has been shown that study designs and analysis methods based on multiple testing procedures and/or adaptive designs may be more efficient compared to standard approaches many of the proposed procedures are ad-hoc methods based on single biomarkers and a systematic approach to derive optimized enrichment designs and analysis methods is lacking. Furthermore, the literature focuses on efficacy and insufficient consideration is given to the establishment of positive benefit risk and the closely related problem of the estimation of treatment effects (on safety and efficacy) in targeted subgroups. Finally, current approaches address maximization of statistical power but do not take the prevalence of the identified subgroups into account.
We will derive optimised study designs based on one or more biomarkers, incorporating information from surrogate and safety endpoints to specify and confirm subgroups controlling frequentist error rates. This work will cover trials with different objectives, such as the identification of any subgroup, all subgroups or the maximal total population, where the treatment has a positive benefit risk balance. The frequentist analysis approach will allow to assess the level of evidence such trials can provide in terms of current standards.
Besides optimizing the probability of success we will use decision theoretic approaches developed in WP2 to develop optimized methods for the identification of subgroups where the benefit-risk balance is positive. Especially, accounting for the prevalence of subgroups, as well as the estimated treatment effect and safety profile in different subgroups will allow incorporation of health economic aspects into the optimization of the identification and confirmation of subgroups. Also for the decision theoretic approach we will develop methods that allow incorporation of information from surrogate endpoints.
Furthermore, we will develop optimized adaptive enrichment clinical trial designs allowing for subgroup selection in an interim analysis. We develop methods that exploit information from one or more biomarkers and surrogate endpoints as well as safety outcomes for the selection of a subpopulation for the remainder of the trial. For the derivation of the subgroup selection rule we will apply frequentist and Bayesian decision theoretic methods.
Section for Medical Statistics (IMS)
CENTER FOR MEDICAL STATISTICS; INFORMATICS, AND INTELLIGENT SYSTEMS (CeMSIIS)
Medizinische Universität Wien
A - 1090 Wien
Links: Medical Univeristy of Vienna