|
Activity Number:
|
10
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #305271 |
|
Title:
|
Bayesian Semiparametric Methods for Joint Modeling of Longitudinal and Survival Data
|
|
Author(s):
|
Adam Branscum*+ and Timothy Hanson and Wesley O. Johnson
|
|
Companies:
|
University of Kentucky and University of Minnesota and University of California, Irvine
|
|
Address:
|
College of Public Health, Lexington, KY, 40536,
|
|
Keywords:
|
Polya trees ; Bayesian nonparametrics
|
|
Abstract:
|
Development of statistical models for the analysis of survival data with time-dependent covariates is an active area of research. The observed values of time-dependent covariates are often available at discrete time points. To circumvent potential bias inherent in last-value-carried-forward approaches, current methodologies propose modeling longitudinal covariate processes jointly with survival data. Recent approaches have focused primarily on modeling longitudinal processes using mixed effects methods, while the survival component is characterized by a Cox proportional hazards regression. In this talk, we will review briefly Bayesian approaches to joint survival/longitudinal modeling and present a novel Bayesian semiparametric approach, which is illustrated using data relating blood chemistry measurements to survival time of kidney dialysis patients.
|
