|
Activity Number:
|
19
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306068 |
|
Title:
|
Evaluation of a Noninvasive Diagnostic Device Using Weighted Least Squares Approach
|
|
Author(s):
|
Zhen Huang*+ and Huiman Barnhart
|
|
Companies:
|
Duke Clinical Research Institute and Duke University
|
|
Address:
|
P.O. Box 17969, Durham, NC, 27715,
|
|
Keywords:
|
agreement ; correlated kappa ; correlated sensitivity/specificity ; weighted least squares
|
|
Abstract:
|
In CINEMA study, the performance of a noninvasive multi-row detector subsecond computed tomography (MRD-SUCT) is evaluated by comparison with the commonly accepted invasive diagnostic method coronary catheterization. The kappa coefficient and sensitivity and specificity are used as measures of agreement and reliability. Because 28 coronary segments per patient are available for evaluation, there are 28 estimated kappa coefficients, as well as subunit sensitivities and specificities. We use the weighted least squares approach to take into account the correlation between these estimated statistics and share our experience in testing the equality of these parameters and related issues.
|
