A0523
Title: Sensitivity analysis for causal estimation with time-varying unmeasured confounding
Authors: Kuan Liu - University of Toronto (Canada) [presenting]
Yushu Zou - University of Toronto (Canada)
Liangyuan Hu - Rutgers University (United States)
Abstract: Causal inference relies on the untestable assumption of no unmeasured confounding to ensure the causal parameter of interest is identifiable. Sensitivity analysis quantifies unmeasured confounding's impact on causal estimates. Among sensitivity analysis methods proposed in the literature, the latent confounder approach is favored for its intuitive interpretation via the use of bias parameters to specify the relationship between the observed and unobserved variables, and the sensitivity function approach directly characterizes the net causal effect of the unmeasured confounding without explicitly introducing latent variables to the causal models. These two sensitivity analysis approaches are developed and extended, namely the Bayesian sensitivity analysis with latent confounding variables and the frequentist sensitivity function approach from a prior study, for the estimation of time-varying treatment effects with longitudinal observational data subjected to time-varying unmeasured confounding and time-varying impact of the unmeasured confounding. The performance of these methods is investigated in a series of simulation studies and is applied to multi-center pediatric disease registry data to provide practical guidance on implementation.
