Bayesian Precision Medicine

This short course will include topics chosen from my 2024 Chapman and Hall book. I will begin with examples and concepts related to precision medicine, which uses individual patient covariates to choose treatment or doses. The first topic is a toy example of problems that may arise when comparing immunotherapy to prayer for treating a severe disease, followed by a brief review of beliefs and concepts that underly the religion of causality. As time permits, I will present four Bayesian clinical trial designs that aim to identify optimal subgroup-specific doses or treatments, and two data analyses. The designs include a phase I-II trial to optimize natural killer cell dose when treating advanced hematologic malignancies, a general design for randomized phase II treatment screening and selection, a randomized design to evaluate nutritional prehabilitation for reducing post operative morbidity, and a generalized phase I-II design to identify optimal subgroup-specific CD70 NK cell doses. The data analyses apply Bayesian nonparametric regression models to identify optimal covariate-specific treatments, the first to optimize the dose of intravenous busulfan in a preparative regimen for stem cell transplantation, and the second to choose a targeted therapy for advanced breast cancer.