Abstract:
|
Pharmaceutical companies continue to seek innovative ways to explore whether a drug under development is likely to be suitable for all or only an identifiable stratum of patients in the target population. Here, we describe a general framework, Patient Response Identifiers for Stratified Medicine (PRISM), for the discovery of potential predictors of drug response and associated subgroups. PRISM, which is available within the “StratifiedMedicine” R package, is highly flexible and can have many “configurations”, allowing the incorporation of complementary models or tools for a variety of outcomes and settings. To facilitate design planning for continuous and binary outcomes, one promising approach is to use model-based partitioning for subgroup identification while using the double-robust estimator for parameter estimation. Bootstrap resampling is then used for “honest” subgroup-level treatment effect estimates and inference. Simulation results, along with data from a Merck clinical trial are used to illustrate the utility of the proposed framework.
|