Semiparametric Efficient Estimator for Incomplete Longitudinal Binary Data with Application to Smoking Trends

• Activity Number: 179
• Type: Topic Contributed
• Date/Time: Monday, August 4, 2008 : 2:00 PM to 3:50 PM
• Sponsor: Section on Health Policy Statistics
• Author(s): Jamie Perin*+ and John Preisser and Paul Rathouz
• Companies: The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill and The University of Chicago
• Keywords: Dropout ; Generalized Estimating Equations ; Longitudinal studies ; Missing at Random

Abstract

Generalized estimating equations, while commonly used in the analysis of incomplete longitudinal data, are invalid for data that is not missing completely at random. Robins, Rotnitzky and Zhao (1995) introduced a class of weighted estimating equations that are valid under dropouts missing at random, including an easy-to-implement, but inefficient member. Robins and Rotnitzky (1995) gave the more complex semi-parametric efficient estimator in this class and applied it to continuous data. A specific form of the estimator is developed for binary data and used to assess the efficiency of the simpler estimator in a simulation study. Both are applied in the estimation of cigarette smoking trends in the United States from a cohort of 5,078 young adults. This extends an earlier analysis by Preisser et. al. (2000) who analyzed smoking trends with the inefficient weighted estimating equations.