Adjustment for Measurement Error in Evaluating Diagnostic Markers by Using an Internal Reliability Sample

• Activity Number: 634
• Type: Contributed
• Date/Time: Thursday, August 4, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Matthew White*+ and Sharon Xie
• Companies: University of Pennsylvania and University of Pennsylvania
• Keywords: Measurement Error ; ROC Curves ; Biomarker ; Diagnostic Test ; Predictive Values ; Sensitivity

Abstract

Repeated measurements of a biomarker vary within a subject due to measurement error. Sources of measurement error include variability within an individual over time and variability in laboratory conditions. A naive approach ignores the error, biasing the sensitivity and specificity of the marker and giving the erroneous impression that the biomarker is not effective. Using an internal reliability sample of the biomarker, we propose bias-correction approaches for estimating a variety of diagnostic performance measures that include sensitivity, specificity, the receiver operating characteristic (ROC) curve, positive and negative predictive values, and likelihood ratios when the marker is subject to measurement error. We derive their asymptotic properties and confidence intervals. We then perform simulations to compare our approaches to naive approaches in estimates of these diagnostic performance measures. The proposed methods have broad biomedical applications (e.g., renal disease, Alzheimer's disease) and are illustrated using a biomarker study in Alzheimer's disease.