Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 473 - Advances in Bayesian Methods and Mixture Modeling for Health Data
Type: Topic Contributed
Date/Time: Wednesday, August 10, 2022 : 2:00 PM to 3:50 PM
Sponsor: SSC (Statistical Society of Canada)
Abstract #322991
Title: Bayesian Causal Estimation with Time-Dependent Latent Confounders
Author(s): Kuan Liu*
Companies: University of Toronto
Keywords: Applied Bayesian; Causal inference; Longitudinal data; Latent variables

The strongly ignorable treatment assignment assumption (also known as no unmeasured confounding) requires a sufficiently large set of covariates being measured to ensure that subjects are exchangeable across the observed exposure given measured covariates. Although administrative data are rich in information, key confounders might not be captured. Several Bayesian sensitivity analyses for unmeasured confounding have been developed that use bias parameters to capture the effect of latent confounders on the outcome and exposure. However, there is a lack of considerations to handle time-dependent latent confounders. In this talk, I will present two parametric Bayesian causal approaches to tackle causal estimation when the causal structure features time-dependent latent confounders. We will discuss these methods under different simulated scenarios with varying strength of relations between the latent confounders and the observed variables.

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

Back to the full JSM 2022 program