Exact Bayesian Inference in Matched Pairs Designs

• Activity Number: 362
• Type: Contributed
• Date/Time: Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Yong Chen*+ and Sining Chen and Haitao Chu
• Companies: Johns Hopkins Bloomberg School of Public Health and Johns Hopkins University and The University of North Carolina at Chapel Hill
• Keywords: Matched Pairs Designs ; Exact Bayesian Inference

Abstract

Matched binary pairs arise frequently in biomedical research including clinical trials involving pre-/post comparison, and observational studies such as matched case-control and matched cohort studies. The relative risk, odds ratio and risk difference are commonly used as the measurements of the association between exposure and outcome in matched binary pairs. However, to our knowledge, the exact probability distributions of those measures have not been given in the literature. In this paper, we derive the exact posterior distributions of these measures for paired binary endpoints, which could be widely used in biomedical research. Furthermore, we discussed Bayesian inference based on these exact distributions, particularly when the sample size is small. The proposed method is applied to several real biomedical studies.