Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 357 - Cross-Cutting Research in Causal Inference and Survival Analysis
Type: Invited
Date/Time: Thursday, August 12, 2021 : 12:00 PM to 1:50 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #316805
Title: Aiding Causal Inference Through Parameter Expansion of the Cox Model to Allow Effects to Vary with Follow-Up
Author(s): James O'Malley* and Todd MacKenzie and Pablo Martinez Camblor
Companies: Dartmouth College and Dartmouth College and Geisel School of Medicine at Dartmouth
Keywords: Causal inference; Cox model; Instrumental variable; Non-proportional hazards; Parameter expansion; Survival time

I discuss the extension and estimation of the Cox regression model for time-to-event data when the treatment variable is exposed to unmeasured confounders and its effect varies over follow-up. An intuitive explanation will be given of how the treatment effects are identified, including of how the assumed treatment selection mechanism is used to help account for unmeasured confounders. This is used to motivate an estimation procedure based on the expression for the survival-time probabilities implied by the model. Limitations of the procedure will also be discussed in relation to alternative approaches. The procedure is applied to data sets in vascular surgery and cardiology from studies for which the time until death is of interest.

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

Back to the full JSM 2021 program