Using Doubly Robust Estimator to Estimate an Average Treatment Effect in Observational Studies When Treatment Switching Exists

• Activity Number: 186
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Chunhao Tu*+ and Woon Yuen Koh
• Companies: University of New England and University of New England
• Keywords: Average Treatment Effect ; Propensity Score ; Treatment Switching ; Observational Studies

Abstract

In this paper, we propose an adjusted generalized doubly robust (AGDR) estimator for estimating average treatment effects (ATE) for binary treatments in observational studies when treatment switching exists. We evaluate our proposed estimator, in terms of bias, mean squared error (MSE), empirical standard error (ESE), and coverage probability (CP), with two existing inverse probability weighting estimators, IPW1 and IPW2. Simulation results show that in general, AGDR estimator outperforms IPW1 and IPW2 estimators in terms of the above evaluation criteria.