Bayesian Methods for Biostatistics introduces doctoral students to the principles and practice of probabilistic inference when uncertainty is quantified with prior distributions rather than fixed-frequency assumptions. After revisiting Bayes’ theorem in continuous form, the course moves quickly to modern computational tools (Markov-chain Monte Carlo, Hamiltonian Monte Carlo, and variational approximations) that make complex hierarchical models tractable for large public-health datasets. Students learn how to elicit and critique priors, build Bayesian generalized linear and mixed-effects models, evaluate convergence diagnostics, and compare models with posterior predictive checks, WAIC, and Bayes factors. Emphasis is placed on interpreting full posterior distributions for parameters and public-health decision metrics, communicating probabilistic results to policy stakeholders, and integrating Bayesian workflows with earlier frequentist training. Case studies span vaccine-effectiveness monitoring, small-area disease mapping, adaptive clinical-trial design, and real-time epidemic forecasting, giving students hands-on experience with state-of-the-art software (R) and reproducible reporting practices.
None