Joint Modeling of Longitudinal Health Predictors and Cross-Sectional Health Outcomes via Mean and Variance Trajectories

• Activity Number: 243
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Bei Jiang*+ and Michael R. Elliott and Mary D. Sammel and Naisyin Wang
• Companies: University of Michigan and University of Michigan and University of Pennsylvania and University of Michigan
• Keywords: Growth mixture models ; variability ; joint modelling ; longitudinal trajectories ; cross-section outcome

Abstract

Growth mixture models (GMMs) can be used to model the heterogeneity in the longitudinal trajectories that cannot be fully explained by measured covariates in the linear mixed effects models when assuming a common within-subject variance parameter. We can further allow subject-level variability in GMMs to differ individuals by individuals by assuming the within-subject variances follow some distribution. As Carroll (2003) noted, "systematic dependence of variability on known factors" may be "fundamental to the proper solution of scientific problems" in certain settings, heterogeneity in the within-subject variances may be important in explaining disease risks. We develop a method that simultaneously examines the association between the heterogeneities in both the mean growth profile classes and the within-subject variances and the cross-sectional binary health outcomes. We consider an application to predict severe hot flashes using the hormone levels collected over time for women in menopausal transition from Penn Ovarian Aging Study.