Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 155 - Recent Developments in Statistical Methods for Data with Informative Cluster Size
Type: Topic Contributed
Date/Time: Tuesday, August 4, 2020 : 10:00 AM to 11:50 AM
Sponsor: ENAR
Abstract #314022
Title: Accounting for Informative Observation Process in Transition Models of Binary Longitudinal Data: A Joint Model Approach
Author(s): Joe Bible and Paul S. Albert and Danping Liu*
Companies: Clemson University and National Cancer Institute and Eunice Kennedy Shriver National Institute of Child Health and Human Development
Keywords: Longitudinal data; Joint model; Informative cluster size; Informative observation process; Transition model
Abstract:

We motivate our methodology through the need to analyze the Consecutive Pregnancy Study (CPS) data, a longitudinal cohort study of women with multiple pregnancies in 23 Utah hospitals from 2003 to 2010. The interest is to study the recurrence of adverse pregnancy outcomes, e.g., preterm birth, with a transition model with random effects. The right censored observation window poses unique challenges to the analysis of the CPS data because of selection bias associated with over sampling of women who are predisposed to shorter gap times between pregnancies. There is also concern regarding an informative cluster size type problem where women who are prone to experiencing favorable pregnancy outcomes are subsequently more likely to endure subsequent pregnancies. We address these concerns with a Shared Random Effect Model (SREM) and demonstrate that it performs well at estimating the transition probabilities even when the gap time and continuation portions of the model are misspecified with respect to the observation process. Through simulation studies, we demonstrate the inadequacy of fitting the transition model alone, ignoring the observation process and informative cluster size.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program