|
Activity Number:
|
426
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #303694 |
|
Title:
|
Marginalized Frailty Models for Multivariate Data
|
|
Author(s):
|
Megan Othus*+ and Yi Li
|
|
Companies:
|
Harvard University/Dana Farber Cancer Institute and Harvard University/Dana Farber Cancer Institute
|
|
Address:
|
677 Huntington Ave. , Boston, MA, 02115,
|
|
Keywords:
|
Semiparametric normal transformation ; Proportional hazards model ; Clustered failure time data
|
|
Abstract:
|
This talk introduces a class of normal transformation models for clustered failure time data. The failure time outcomes are assumed to marginally follow a proportional hazards model, while the normally transformed variates allow a shared frailty. As a result, the model permits population-level interpretation of covariates in the proportional hazards model, but also directly models the correlation of the transformed failure times. The method allows for varying cluster sizes and we are able to predict the shared frailty of the transformed failure times. Predictions of the frailties allow us to evaluated the underlying cluster effects on subjects' survival. We propose a profile estimation procedure and derive asymptotic properties under this estimation scheme. We apply the method to a Children's Oncology Group study of acute lymphoblastic leukemia.
|
- 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 2009 program |
