|
Activity Number:
|
360
- Contributed Poster Presentations: Section on Bayesian Statistical Science
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
|
Abstract #313138
|
|
|
Title:
|
Causal Inference in Observational Studies with a Post-Treatment Variable Using Bayesian Inference
|
|
Author(s):
|
Li He* and Yu-Bo Wang and William Bridges
|
|
Companies:
|
and Clemson University and Clemson University
|
|
Keywords:
|
Propensity score;
Principal Stratification;
Polya-gamma distribution;
Bayesian logistic inference
|
|
Abstract:
|
Propensity score methods are commonly used to correct for the effects of confounding variables on the estimated treatment effect in observational studies. Principal stratification is a statistical technique used to correct for the effect of post-treatment variables on the estimated treatment effects in randomized clinical trials. When estimating the treatment effect in observational study with post-treatment variables present, it is critical to adjusting for both the confounding variables and the post-treatment variables. There is little literature discussing approaches to combining propensity scores methods and principal stratification methods. We propose a Bayesian inference framework which considers joint modeling of the propensity score with general principal stratification. The proposed framework will be applied to several simulation cases to evaluate the versatility of the method. Moreover, since the propensity score is easily extended to non-binary cases, we also consider extending the proposed framework to multiple treatment groups.
|
Authors who are presenting talks have a * after their name.
