Bayesian and Profile -likelihood Change Point Methods for Modeling Cognitive Function over Time

• Activity Number: 403
• Type: Contributed
• Date/Time: Thursday, August 15, 2002 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Stat. Sciences*
• Author(s): Jun Ying*+ and Charles Hall and Lynn Kuo and Richard Lipton
• Affiliation(s): University of Connecticut and Albert Einstein College of Medicine and University of Connecticut and Albert Einstein College of Medicine
• Keywords: Bayesian analysis ; change points ; cognitive aging ; longitudinal data ; Markov Chain Monte Carlo ; mixed models

Abstract

Change point models are often used to model longitudinal data. Under the assumption of common change point among the individuals, either Bayesian or profile-likelihood methods may be used to estimate the change point. The Bayesian method, however, has the advantage over the profile-likelihood method in relaxing the common change point assumption and allows for individual change points in the model. We compare and contrast both the hierarchical Bayesian method and profile-likelihood method in analyzing longitudinal cognitive data from the Bronx Aging Study. Markov Chain Monte Carlo simulations are used to implement the Bayesian method. Caution must be taken regarding sensitivity to choice of prior distribution, identifiability, and goodness of fit. Analyses show that decline in memory precedes diagnosis of dementia by 7.5 to 8 years, and individual change points are not needed to model heterogeneity across the subjects.