Introducing Bayesian Thinking and Applications to Health Science Researchers

• Activity Number: 453
• Type: Roundtables
• Date/Time: Wednesday, August 7, 2013 : 7:00 AM to 8:15 AM
• Sponsor: Section on Teaching of Statistics in the Health Sciences
• Author(s): J. Jack Lee*+
• Companies: Univ of Texas-M D Anderson Cancer Center
• Keywords: Bayesian methods ; computation software ; graphs and videos ; teaching statistics to non-statisticians

Abstract

Although the frequentist framework has dominated statistical inference, the Bayesian approach offers many advantages as it is more intuitive, more directly addresses the question of interest, and is uniquely suited to medical applications. For example, Bayes theorem is widely applied in evaluating the accuracy of disease diagnosis. Hierarchical models synthesize information across many groups. Prior information is easily incorporated for assessing treatment effect. The Bayesian learn-as-we-go approach is ideally suitable for adaptive designs. Utility can be constructed to evaluate the optimal treatment choice balancing efficacy and toxicity. A big challenge is how to communicate the Bayesian thinking to researchers. In addition, there are relatively few tools to learn and to implement Bayesian methods. I will share my experience in conveying the Bayesian thinking to nonstatisticians via graphs and videos. The features of the frequentist and Bayesian approaches are compared and contrasted. I will also show simple tools for Bayesian computation using WINBUGS, R, BRUGS, SAS, and Excel, etc. Participants are encouraged to share their experiences and successful examples.