|
Activity Number:
|
156
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #307436 |
|
Title:
|
A Marginal Model for Multistate Panel Data under Heterogeneity
|
|
Author(s):
|
Wei-Ting Hwang*+
|
|
Companies:
|
University of Pennsylvania School of Medicine
|
|
Address:
|
628 Blockley Hall, Philadelphia, PA, 19104,
|
|
Keywords:
|
panel data ; frailty ; Markov process ; estimating equation
|
|
Abstract:
|
The goal of this work is to estimate disease progression rates from panel data with multiple stages when the population is heterogeneous because of unobserved variables. Frailty is introduced to describe the heterogeneity in the population. We developed a marginal model based on generalized estimating equations to obtain the parameter estimates. Comparison to conditional model based on the likelihood function will be discussed. The results and methods are applied to data on Alzheimer's Disease (AD) research where repeated measurements on subjects' cognitive functions were collected over time through a Clinical Dementia Rating (CDR) scale. This scale has five ordinal stages which measure the functional impairment in areas such as memory, orientation and language.
|
