Moving beyond AUC: decision curve analysis for quantifying net benefit of risk prediction models
Mohsen Sadatsafavi, Amin Adibi, Milo A. Puhan, Andrea S. Gershon, Shawn D. Aaron, Don D. Sin
Abstract
<b>Statistical constructs such as ROC/AUC do not answer the critical question of how much clinical utility a risk prediction model confers. This paper overviews decision curve analysis, a novel method for quantifying net benefit of a risk prediction model.</b>https://bit.ly/3h1rraX
Topics & Concepts
MedicineReceiver operating characteristicRisk modelArea under the curveStatisticsMachine learningEconometricsRisk analysis (engineering)Internal medicineComputer scienceMathematicsClinical practice guidelines implementationMedical Coding and Health InformationMeta-analysis and systematic reviews