|
Activity Number:
|
417
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #309705 |
|
Title:
|
A Semiparametric Empirical Likelihood Method for ROC Curve Regression of Ordinal Diagnostic Test in the Presence of Verification Bias
|
|
Author(s):
|
Xiaofei Wang*+ and Haibo Zhou
|
|
Companies:
|
Duke University and The University of North Carolina at Chapel Hill
|
|
Address:
|
Medical Center, Room 8027, Hock Plaza, Durham, NC, 27717,
|
|
Keywords:
|
Empirical Likelihood ; Ordinal Diagnostic Test ; Receiver Operating Characteristic (ROC) Curve ; Semiparametric Method ; Verification bias
|
|
Abstract:
|
In medical studies to assess the accuracy of diagnostic tests, definitive disease assessment is too invasive or expensive to be ascertained on all the study subjects. It is often more ethical or cost effective to ascertain the true disease status with a higher rate in study subjects where the diagnostic test is suggestive of disease. Without taking the study design into account, standard statistical methods can yield biased estimates for diagnostic accuracy. It is known as verification bias in the literature. To accommodate the biased sampling scheme, we develop a semiparametric empirical likelihood method to estimate the parameters of the covariate-specific ROC regression model for ordinal diagnostic test. We consider biased sampling schemes in which the set of selected subjects include either a simple random sample, or a test-result-dependent sample or both.
|
