|
Activity Number:
|
66
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306465 |
|
Title:
|
Assessment of Diagnostic Tests in the Presence of Verification Bias Using Multiple Imputation and Resampling Methods
|
|
Author(s):
|
Michael P. McDermott*+ and Hua He
|
|
Companies:
|
University of Rochester Medical Center and University of Rochester Medical Center
|
|
Address:
|
Department of Biostatistics and Computational Biology, Rochester, NY, 14642,
|
|
Keywords:
|
bootstrap ; missing data ; multiple imputation ; sensitivity ; specificity ; verification bias
|
|
Abstract:
|
Sensitivity and specificity are common measures of the accuracy of a diagnostic test. The usual estimators of these quantities are unbiased if data on the diagnostic test result and the true disease status are obtained from all subjects in an appropriate sample. In some studies, verification of the true disease status is performed only for a subset of subjects, possibly depending on the result of the diagnostic test and other characteristics of the subjects. Estimators of sensitivity and specificity based on this subset of subjects are typically biased. We view this as a missing data problem and apply commonly-used techniques in the survey sampling literature, multiple imputation and bootstrap resampling, to derive estimators and corresponding confidence intervals that are corrected for this verification bias. Comparisons are made between these and existing bias-correction methods.
|
