Interpreting area under the receiver operating characteristic curve

The Lancet Digital Health, vol. 4, pp. e853–e855

Abstract

Clinicians often ask biostatisticians to provide a judgement on the quality of a prediction model. The prediction quality can be assessed with different performance criteria, with discriminative ability being a main issue: how well can we separate high-risk from low-risk patients? Discriminative ability is typically quantified by the area under the receiver operating characteristic curve (AUC) when we consider prediction of a binary event.

Authors 3

1. Anne de Hond corresponding Aachen

   Leiden University · KU Leuven · Stanford Medicine · Stanford University · Leiden University Medical Center

   Affiliation as printed

   Clinical AI Implementation and Research Lab, Leiden University Medical Centre, 2333 ZA Leiden, Netherlands; Department of Medicine (Biomedical Informatics), Stanford University, Stanford, CA, USA. Electronic address: a.a.h.de_hond@lumc.nl

   Clinical AI Implementation and Research Lab (AAHdH, EWS) and Department of Biomedical Data Sciences (EWS, BvC), Leiden University Medical Centre, 2333 ZA Leiden, Netherlands; Department of Medicine (Biomedical Informatics),

   KU Leuven, Leuven, Belgium

   Stanford University, Stanford, CA, USA (AAHdH); Department of Development & Regeneration,

2. Ewout Willem Steyerberg Aachen

   Leiden University · KU Leuven · Stanford University · Leiden University Medical Center

   Affiliation as printed

   Clinical AI Implementation and Research Lab, Leiden University Medical Centre, 2333 ZA Leiden, Netherlands; Department of Biomedical Data Sciences, Leiden University Medical Centre, 2333 ZA Leiden, Netherlands

   Clinical AI Implementation and Research Lab (AAHdH, EWS) and Department of Biomedical Data Sciences (EWS, BvC), Leiden University Medical Centre, 2333 ZA Leiden, Netherlands; Department of Medicine (Biomedical Informatics),

   KU Leuven, Leuven, Belgium

   Stanford University, Stanford, CA, USA (AAHdH); Department of Development & Regeneration,

3. Ben Van Calster Aachen

   Leiden University · KU Leuven · Stanford University · Leiden University Medical Center

   Affiliation as printed

   Department of Biomedical Data Sciences, Leiden University Medical Centre, 2333 ZA Leiden, Netherlands; Department of Development & Regeneration, KU Leuven, Leuven, Belgium

   Clinical AI Implementation and Research Lab (AAHdH, EWS) and Department of Biomedical Data Sciences (EWS, BvC), Leiden University Medical Centre, 2333 ZA Leiden, Netherlands; Department of Medicine (Biomedical Informatics),

   KU Leuven, Leuven, Belgium

   Stanford University, Stanford, CA, USA (AAHdH); Department of Development & Regeneration,