|
Activity Number:
|
160
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #305538 |
|
Title:
|
Different Schemes of Verification Bias in Evaluating Medical Tests
|
|
Author(s):
|
Marina V. Kondratovich*+
|
|
Companies:
|
FDA
|
|
Address:
|
HFZ-550, 1350 Piccard Dr.,, Rockville, MD, 20850,
|
|
Keywords:
|
sensitivity ; specificity ; verification bias ; diagnostic accuracy
|
|
Abstract:
|
An issue of verification bias presents in a lot of submissions to the FDA. If it is not addressed properly by an appropriate study design or an appropriate type of statistical analysis, it can create problems for unbiased evaluation of medical tests. We present different schemes of verification bias for a comparison of two medical tests (Old and New tests) : 1) a random sample of the subjects with negative results via both tests has gold standard results; 2) no subjects with negative results via both tests have gold standard results; we describe two schemes: one for a paired study design (pre-market studies) and another for an unpaired study design (post-market studies); 3) only subjects with positive results by the Old Test have results of the gold standard. The results of investigation of conditions for the unbiased comparison of the two tests are presented along with examples.
|
- 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 |
