|
Activity Number:
|
469
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308348 |
|
Title:
|
Martingale Residuals for Nested Case-Control Data
|
|
Author(s):
|
Ornulf Borgan*+ and Bryan Langholz
|
|
Companies:
|
University of Oslo and University of Southern California
|
|
Address:
|
Department of Mathematics, Oslo, N-0316, Norway
|
|
Keywords:
|
Cohort sampling ; Counter-matching ; Counting processes ; Cox's regression model ; Regression diagnostic plots ; Relative risk regression
|
|
Abstract:
|
Standard use of Cox's regression model and other relative risk regression models for censored survival data requires collection of covariate information on all individuals under study even when only a small fraction of them die or get diseased. For such situations nested case-control sampling offers a useful alternative. For cohort data, martingale residuals provide a useful tool for assessing the fit of a model. We introduce a natural extension of the cohort methods to nested case-control data, and show that plots of martingale residual processes provide a useful tool for checking model-fit. A formal goodness-of-fit test to go along with the plots is also presented. The methods are illustrated using data on lung cancer deaths in a cohort of uranium miners.
|
