|
Activity Number:
|
252
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Stat. Sciences*
|
| Abstract - #301133 |
|
Title:
|
Combining Multiple Kaplan-Meier Estimators: A Bayesian Model
|
|
Author(s):
|
Peter Bouman*+ and Xiao-Li Meng and Vanja Dukic+
|
|
Affiliation(s):
|
Harvard University and Harvard University and University of Chicago
|
|
Address:
|
Science Center 610, 1 Oxford Street, Cambridge, Massachusetts, 02138, 5841 South Maryland Avenue, MC2007, Chicago, Illinois, 60637,
|
|
Keywords:
|
Meta-analysis ; Survival Analysis ; Bayesian Statistics ; Kaplan-Meier estimator
|
|
Abstract:
|
The Kaplan-Meier product-limit estimator $KM(t)$ is a popular method of estimating a non-parametric survival curve $S(t)$ from (possibly censored) follow-up data from a given study group. If we want to make inferences about a population survival curve from multiple, heterogeneous study-specific $KM_{i}(t)$, though, the nonparametric nature of $S(t)$ makes it difficult to combine multiple estimators. We propose a simple, flexible Bayesian model for survival curve heterogeneity which estimates overall survival probabilities at specific time horizons of interest. We apply the model to a meta-analysis of protease-inhibitor survival data, producing estimates of study heterogeneity and the overall survival distribution.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |
