Semiparametric Regression Modeling of Longitudinal Binary Outcomes with Outcome-Dependent Observation Times

• Activity Number: 194
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Kay See Tan*+ and Andrea Troxel and Stephen E. Kimmel and Kevin G. Volpp and Benjamin French
• Companies: Unniversity of Pennsylvania and Univ of Pennsylvania School of Medicine and University of Pennsylvania Perelman School of Medicine and University of Pennsylvania Perelman School of Medicine and University of Pennsylvania Perelman School of Medicine
• Keywords: Joint models ; observation-time process ; outcome process ; outcome-dependent follow-up ; informative observation

Abstract

Conventional longitudinal data analysis methods assume that outcomes are independent of the data-collection schedule. However, the independence assumption may be violated, for example, when adverse events trigger additional physician visits in between prescheduled follow-ups. Outcome-dependent observation times may introduce bias when estimating the effect of covariates on outcomes using a standard longitudinal regression model. Existing semi-parametric methods that accommodate informative observation times are limited to the analysis of continuous outcomes. We develop new methods for the analysis of binary outcomes, while retaining the flexibility of semi-parametric models. Our methods are based on counting process approaches and provide marginal inference. In simulations we evaluate the statistical properties of our proposed methods. Comparisons are made to 'naïve' GEE approaches that do not account for outcome-dependent observation times. We illustrate the utility of our proposed methods in an application to a clinical trial evaluating the effectiveness of adherence intervention among patients treated with warfarin.