|
Activity Number:
|
154
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #309095 |
|
Title:
|
Inference in Clinical Trial Using ROC Curve Regression
|
|
Author(s):
|
Lin Zhang*+ and Jack Tubbs
|
|
Companies:
|
Baylor University and Baylor University
|
|
Address:
|
Department of Statistical Science , Waco, TX, 76798,
|
|
Keywords:
|
nonparametric ; clinical trial ; interaction effect ; ROC curves regression ; bootstrap
|
|
Abstract:
|
This paper considers a method proposed by Dodd and Pepe where the area under the ROC curve (AUC) provides a measure of the treatment effect that allows for the incorporation of strata variables, covariates, and their possible interaction in the usual logistic regression model. Dodd and Pepe proposed using bootstrapped standard errors as the independence assumption of the logistic model was violated when using a function of the usual Mann-Whitney statistic as the response variable. In this paper we consider the 'exact' estimates for the variance of the estimate of the AUC as a function of strata variables and discrete covariates. The confidence interval for the AUC based on this method is compared with the bootstrap procedure.
|
