|
Activity Number:
|
421
- Contributed Poster Presentations: Health Policy Statistics Section
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Health Policy Statistics Section
|
|
Abstract #322747
|
|
|
Title:
|
A Bayesian Approach for Instrumental Variable Analysis
|
|
Author(s):
|
Samrachana Adhikari* and Sharon-Lise Normand
|
|
Companies:
|
Harvard Medical School and Harvard Medical School
|
|
Keywords:
|
treatment effect heterogeneity ;
latent variables analysis ;
health policy research
|
|
Abstract:
|
For making a causal inference in the presence of unmeasured confounders instrumental variable (IV) analysis plays a crucial role. Valid instruments that satisfy assumptions about relationships between the instruments, a treatment assignment and an outcome are necessary for the IV analysis. Most of the existing estimation methods utilize a method of moments approach within a structural models framework. While these models do not make any distributional assumptions, they rely on either constant treatment effect or monotonicity constraints for identification. We explore likelihood-based estimators in the IV analysis and the necessary assumptions to make a valid inference based on such estimators. We focus on a Bayesian approach to make a posterior inference on the estimates of causal effects, with particular emphasis when there is treatment effect heterogeneity. We utilize simulations to characterize operating characteristics of various estimators. A novel application to determine the effect of radial artery access on bleeding outcomes compared to femoral artery access for patients undergoing cardiovascular procedures demonstrates the utility of the Bayesian approach. (R01-GM111339)
|
Authors who are presenting talks have a * after their name.
