Using Repeated Measures to Aid Cut Point Selection

• Activity Number: 489
• Type: Topic Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: International Chinese Statistical Association
• Author(s): Zhezhen Jin and Xinhua Liu*+
• Companies: Columbia University and Columbia University
• Keywords: Concordance ; biomarkers ; cut point

Abstract

In biomedical research and practice, quantitative biomarkers are often used for diagnostic or screening purposes, with an established cut point to aid binary classification. To select a cut point with measurements of the biomarker from the subjects in each of two groups with and without the disease or conditions, one may examine sensitivity and specificity for each possible cut point and use either parametric or non-parametric method to select the cut point optimizing a meaningful criterion. We consider several options in selecting the cut point maximizing concordance probability for binary classification and examine the performance of each approach regarding bias and variation through simulation studies.