|
Activity Number:
|
439
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #305446 |
|
Title:
|
Bayesian Analysis of Doubly Censored Data with Covariates with Application to AIDS and Dementia
|
|
Author(s):
|
Binbing Yu*+
|
|
Companies:
|
National Institute on Aging
|
|
Address:
|
7201 Wisconsin Avenue, Suite 3C309, Bethesda, MD, 20892,
|
|
Keywords:
|
AIDS ; dementia ; doubly-censored ; MCMC
|
|
Abstract:
|
Dementia is one of the leading causes of death in older people. Knowledge of life expectancy after dementia diagnosis may help planning for future care. Different study designs and definitions of diagnosis age result in different strategies of estimating survival after dementia onset. The studies based on prevalent cases are subject to length bias. Studies that use incidence dementia avoid this bias, but introduce other difficulties such as how to decide the exact onset time. Conventional definitions like midpoint imputation leads to biased estimates of survival time. In addition, the variances are under estimated because the variation of inaccurate onset times is ignored. A Bayesian method is used to account for the interval-censored dementia onset time. The method is applied to data from a population-based dementia study and a study of AIDS incubation time.
|
