STEEP: Statistically efficient methods for precision medicine trials

Basket trials are an innovative approach to precision oncology that can enrol patients across different cancer types, for example, colorectal cancer and breast cancer, who share common genetic profiles. I have developed several advanced methods for the design and analysis of basket trials. These methods significantly improve the efficiency of basket trials by borrowing of information between patient subgroups achieved by modelling the similarity of the subgroup-specific treatment effects. Flexible basket designs with added efficiency (e.g., allowing patient allocation) are desirable yet their development lags behind. In this project, I shall investigate how basket trials can be designed to allow (a) multiplicity correction, (b) sequential decision making, given the accumulating trial data or new information from outside the trial. Specifically, I shall develop statistical methods that
1) control the risk of incorrectly claiming efficacy when there is none,
2) allocate more patients to receive a better-suited treatment during the trial,
3) supplement the basket trial (especially for rare cancers) with historical controls or real-world evidence,
4) enable adding new patient subgroups or new control treatments to an ongoing basket trial.