Evaluating the Effects of Time-Varying Exposures in Observational Studies Using Cox Models with Stabilized Inverse Probability of Treatment Weights

• Activity Number: 403
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Stanley Xu*+ and Susan Shetterly and Michael Ho and Thomas Tsai and David Magid
• Companies: Kaiser Colorado and Kaiser Colorado and Denver VA Medical Center and Denver VA Medical Center and Kaiser Colorado
• Keywords: Drug-Eluting Stent ; Clopidogrel ; Time-varying treatment ; Cox model ; Stabilized inverse probability of treatment weights

Abstract

Stabilized inverse probability of treatment weights (S-IPTW) have been used to compare two groups in the presence of confounders in observational studies. The objective of this paper is to develop a stratified Cox model with S-IPTW adjustment to examine the effect of a treatment which varies over time in an observational study. S-IPTWs are calculated on outcome event days and are used in a stratified Cox model to estimate the effect of an anti-platelet agent, clopidogrel, on subsequent bleeding, myocardial infarction, and death after a Drug-Eluting Stent was implanted into a coronary artery. Clopidogrel use may change over time based on patients' behavior and physicians' recommendations. We also compared the results to those from a Cox model for counting processes adjusting for all covariates used in creating S-IPTWs. We demonstrate that 1) results from the stratified Cox model without S-IPTW adjustment and the Cox model for counting processes without covariate adjustment are identical in analyzing the clopidogrel data; and 2) effects of clopidogrel on outcomes are larger in the stratified Cox model with S-IPTW adjustment compared to those from the Cox model for counting processes.