The goal of this course is for students to gain good understanding of the Bayesian concepts involved in the design, analysis and reporting of clinical trials. Bayesian statistical paradigm use prior information or beliefs, along with the current data, to guide the search for parameter estimates. In the Bayesian paradigm probabilities are subjective beliefs. Prior information/beliefs are input as a distribution, and the data then helps refine that distribution. The choice of prior distributions, posterior updating, as well as dedicated computing techniques are introduced through simple examples. Bayesian approaches for design, monitoring, and analysis of randomized clinical trials are taught in this class. These approaches are contrasted with traditional (frequentist) approaches. The emphasis will be on concepts. Examples are case studies from the instructors’ work and from medical literature. R, OpenBUGS and SAS will be the main computing tools used.
Upon completing this course the students will

- Have a thorough understanding of the components of Bayesian paradigm.
- Carry out probability calculations involving posterior distributions from simple conjugate Bayesian problems – Beta-Bernoulli, Gamma-Poisson, Normal-Normal.
- Have a critical understanding of the similarities and differences between the Bayesian and traditional (frequentist) approach to design, analysis and interpretations of results of data arising from randomized clinical trials.
- Be familiar with mechanism of eliciting prior distributions.
- Be able to design, analyze and report a clinical trial using Bayesian methods.

### Prerequisites:

- Familiarity with basic concepts of clinical trials
- Working knowledge of statistics
- Working knowledge of R

Author
Thomas Zhou
Author
### Introduction

### Bayesian Methods Overview

### Bayesian Approach to Statistics – Discrete Case

### Bayesian Approach to Statistics – Continuous Case

### Estimation and Hypothesis Testing in Clinical Trials – Bayesian Approach

### Prior Elicitation

### Wrapping Up