Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 85 - 20 Years of Principal Stratification: Where to Now?
Type: Invited
Date/Time: Monday, August 8, 2022 : 8:30 AM to 10:20 AM
Sponsor: Health Policy Statistics Section
Abstract #319196
Title: Bridging Theory and Practice in Principal Stratification
Author(s): Fan Li* and Bo Liu
Companies: Duke University and Duke University
Keywords: causal inference; principal stratification; programming; Stan; noncompliance

Post-treatment confounding is a common problem in causal inference, with special cases of noncompliance, truncation by death, surrogate endpoint. Principal stratification is a general framework for defining and estimating causal effects in the presence of post-treatment confounding. Despite its versatility, principal stratification is not accessible to the vast majority of applied researchers because its inherent latent structure, which requires complex inference tools and highly-customized programming. We fill this gap by developing a computing platform R package "PStrata" to automatize inference of principal stratification for the most common scenarios. PStrata supports two inference paradigms and accommodate different assumptions. For the Bayesian paradigm, the computing architecture combines R, C ++, Stan, where R provides the user-interface, Stan automatizes posterior sampling, and C++ bridges the two owing. For the frequentist paradigm, PStrata implements the weighting-based triply-robust estimator of Jiang et al (2020). PStrata can accommodate regular outcomes and time-to-event outcomes. We will illustrate PStrata using a high-profile clinical trial on aspirin use.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program