A Nonparametric Test for Comparing ROC Curves from a Paired Design

• Activity Number: 374
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Medical Devices and Diagnostics
• Author(s): María Carmen Pardo* and Alba María Franco-Pereira
• Companies: Complutense University of Madrid and Complutense University of Madrid
• Keywords: receiver operating characteristic ; summary measure ; two-sample nonparametric test ; diagnostic marker

Abstract

When measurements on two diagnostic markers are available, a question of interest is which marker best discriminates between healthy and disease subjects. The most commonly approach to comparing ROC curves is to test the equality of a summary measure from the ROC curves, such as the area under the curve. In this work, we propose a nonparametric test to test equality of two ROC curves which has higher power than the conventional DeLong test and other known nonparametric test.