Subgroup Cluster Based Bayesian Adaptive Design for Precision Medicine
The Subgroup Cluster Based Bayesian Adaptive Design (SCUBA) is a powerful solution for phase 2 subgroup enrichment trials (Biometrics; Guo et al., 2017b). SCUBA aims to find the optimal subgroup that benefits from the treatment.
In addition, for an umbrella or platform trial where different treatments or doses are compared, SCUBA is able to 1) identify the optimal subgroup (if any) for each arm; 2) enrich the patient allocation to their desirable treatments by adaptive randomization. Despite of the feature of patients enrichment, SCUBA can also be a powerful tool of subgroup analysis for the clinical data, and then subgroup-based probability of success can be calculated, which is useful to steer the sponsor’s decision for future drug development.
- SCUBA is flexible in detecting the subgroups and the biomarkers can be either categorical or continuous.
- SCUBA has a large power in detecting the optimal subgroup, compared to existing Bayesian methods.
- SCUBA is able to detect different high-response subgroups for different treatment arms.