Doubly Robust Testing and Estimation of Model-Adjusted Effect-Measure Modification with Complex Survey Data

• Activity Number: 360
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Babette Brumback*+ and Hao Zheng and Xiaomin Lu and Erin Bouldin and Michael Cannell and Elena Andresen
• Companies: University of Florida and SunTrust Bank and University of Florida and Epidemiology Department, University of Washington and UNT Health Science Center and Oregon Health and Science University
• Keywords: Causal inference ; Confounding ; Doubly-Robust Estimation ; Complex Survey Data ; Propensity score

Abstract

Recently, we extended doubly robust methodology to test and estimate model-adjusted effect-measure modification with complex survey data. Our purpose was to estimate the effects of disability status on cost barriers to health care within three age categories, and to test for differences. The doubly robust methodology requires a model for the probability of the exposure conditional on the confounders (an exposure model), as well as a model for the expectation of the outcome conditional on the exposure and the confounders (an outcome model). We developed two different doubly robust approaches, and we applied the methodology to data from the 2007 Florida Behavioral Risk Factor Surveillance System (BRFSS) Survey. We also used the doubly robust approaches to develop and apply goodness-of-fit tests for the exposure and outcome models. We compare the exposure modeling, outcome modeling, and doubly robust approaches in terms of a simulation study and the BRFSS example.