A Seminonparametric Approach to Joint Modeling of a Primary Binary Outcome and Longitudinal Data Measured at Discrete Informative Observation Times

• Activity Number: 82
• Type: Contributed
• Date/Time: Sunday, July 31, 2011 : 4:00 PM to 5:50 PM
• Sponsor: ENAR
• Author(s): Song Yan*+
• Companies: North Carolina State University

Abstract

A common joint modeling approach is proposed to use subject-specific random effects in a linear mixed model for longitudinal data as predictors in a model (e.g., logistic model) for the final binary outcome. In many applications, the observation times of longitudinal data may be informative. We propose to introduce a third model in the joint model for the informative observation times, and relax the normality distribution assumption of random effects using the semi-nonparametric (SNP) approach of Gallant and Nychka (1987). An EM algorithm is developed for parameter estimation. Extensive simulation designed to evaluate the proposed method indicates that ignoring either informative observation times or distribution assumption of the random effects would lead to invalid and/or inefficient inference. Applying our new approach to the New York University In Vitro Fertilization(IVF) data reveals some interesting findings the traditional approach failed to discover.