Optimal Design for Diagnostic Accuracy Studies When the Biomarker Is Subject to Measurement Error

• Activity Number: 42
• Type: Contributed
• Date/Time: Sunday, July 29, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Matthew Thomas White*+ and Sharon Xie
• Companies: University of Pennsylvania and University of Pennsylvania
• Keywords: Biomarker ; Diagnostic test ; Measurement error ; Reliability sample ; Study design

Abstract

Biomarkers have become increasingly important in recent years for their ability to effectively distinguish diseased from healthy individuals, potentially avoiding unnecessary and invasive diagnostic testing. Given their importance, it is critical to design and analyze studies to produce reliable estimates of diagnostic performance. Biomarkers are often obtained with measurement error, which may cause the biomarker to appear ineffective if not taken into account in the analysis. We develop design strategies for studying the effectiveness of an error-prone biomarker in differentiating diseased from healthy individuals and focus on the area under the receiver operating characteristic curve (AUC) as the primary measure of effectiveness. Using an internal reliability sample within the diseased and healthy groups, we develop optimal study design strategies that minimize the variance of the estimated AUC. We develop optimal allocations of the number of subjects, the size of the reliability sample, and the number of replicate observations per subject in the reliability sample under constraints where 1) the total number of observations is fixed or 2) the total cost of the study is fixed.