Effect of Hormonal Contraception on the HIV Acquisition Using Structural Models

• Activity Number: 84
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Pai-Lien Chen*+
• Companies: Family Health International
• Keywords: Marginal structural model ; structural nested model ; Cox proportional hazards model ; observational study

Abstract

Marginal structural models (MSMs) using inverse probability weighting and structural nested models (SNMs) using g-estimation have been recognized as preferred methods that account for the confounding and intermediate role of time-dependent risk factors and reduce selection bias that may be present with observational study designs. In this paper, we apply MSMs and SNMs to a large multi-center cohort study that investigates the effect of hormonal contraception (HC) on the HIV acquisition. We compare the results between conventional Cox proportional hazards models, the structural Cox proportional hazards models, accelerated failure time models, and examine the sensitivity of confounding of these analyses.