TS2017 Conference Pages Banner.jpg
Home Registration Hotel Schedule Education Program

The "Why, What, and How" of Bayesian Approaches in the Analysis of Translational Data

Track: Regulatory Science and Translational Methods

Session Number: 1015
Date: Fri, Apr 21st, 2017
Time: 10:30 AM - 11:30 AM

Description:

Although new design and analysis methods for translational science are increasingly developed in a Bayesian framework, these methods have not been readily adopted in practice. In part, this is perhaps a result of the language of the Bayesian framework and that the tools used for estimation are different than non-Bayesian approaches. Consequently, translational scientists have limited experience with interpreting study results using Bayesian analyses and evaluating whether an analysis is rigorous. The purpose of this session is to introduce the language of the Bayesian framework and give the audience the ability to interpret and evaluate Bayesian analyses. The first half of the session covers the nomenclature of Bayesian approaches through three examples relevant to clinical and translational researchers. The second half of the session compares results produced from a traditional analysis and an analysis conducted in a Bayesian approach using MCMC (the major computational engine of Bayesian analysis) for a logistic regression model. The outcome of attending this session will be an increased awareness and understanding of Bayesian methodologies and an ability to interpret and evaluate analyses conducted in a Bayesian framework.
Session Type: Lecture

Discipline: None
Session Type: Lecture

Discipline: None

Speakers

Moderator/Chair
Matthew J Mulvahill, MS
Research Instructor
University of Colorado at Denver
Speaker 1
Nichole Carlson, PhD
University of Colorado at Denver
Speaker 2
Edward John Bedrick, PhD
Professor
Dept of Epi and Biostatistics