Dissertation/Thesis Abstract

Multivariate mixture models to describe longitudinal patterns of frailty in American seniors
by Connor, Jason Todd, Ph.D., Carnegie Mellon University, 2006, 175; 3275170
Abstract (Summary)

Large, longitudinal, multivariate population surveys are increasingly common. Many analytic methods inspect changing rates of individual outcomes but ignore heterogeneous subpopulations that likely exist. I introduce two analytical methods which extend group-based trajectory models to multivariate outcomes. The first of these methods, which I label the marginal developmental trajectory model, identifies latent patterns separately for each outcome, then uses a contingency table framework to identify common combinations of trajectories. The second method, which I label the joint developmental trajectory model, considers all outcomes simultaneously to identify combinations of longitudinal trajectories. I use group-based longitudinal finite mixture models (i.e., developmental trajectory models) to identify and describe latent subpopulations for the multivariate outcomes of interest. I apply these methods using data from the National Long Term Care Survey, which measures various disabilities in American elderly from 1982 to 2004. Both models of longitudinal pathways of disability clearly illustrate the various frailty patterns in latent subpopulations of American seniors. Within latent classes, the models clearly illustrate the age of onset and order in which disability patterns typically appear. Finally, I demonstrate latent frailty patterns with illustrative plots that show multivariate temporal patterns within latent class, explore demographics differences between subpopulations, and describe how disability is changing over time.

Indexing (document details)
School: Carnegie Mellon University
School Location: United States -- Pennsylvania
Source: DAI-B 68/07, Dissertation Abstracts International
Subjects: Gerontology, Statistics
Keywords: Disability, Longitudinal patterns, Trajectory models
Publication Number: 3275170
ISBN: 978-0-549-15546-1
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy