Net Reclassification Index Statistics Do Not Help Assess New Risk Models

When evaluating a new risk factor for disease (eg, a measurement from imaging studies), many investigators examine its value above and beyond existing biomarkers and risk factors. They compare the performance of an "old" risk model using established predictors and a "new" risk mo...

Full description

Saved in:
Bibliographic Details
Published inRadiology Vol. 306; no. 3; p. e222343
Main Author Kerr, Kathleen F.
Format Journal Article
LanguageEnglish
Published United States Radiological Society of North America 01.03.2023
Subjects
Biomarkers
Humans
Models, Statistical
Reviews and
Risk Assessment - methods
Risk Factors
Online AccessGet full text
ISSN0033-8419
1527-1315
1527-1315
DOI10.1148/radiol.222343

Cover

More Information
Summary:When evaluating a new risk factor for disease (eg, a measurement from imaging studies), many investigators examine its value above and beyond existing biomarkers and risk factors. They compare the performance of an "old" risk model using established predictors and a "new" risk model that adds the new factor. Net reclassification index (NRI) statistics are a family of metrics for comparing two risk models. NRI statistics became popular in some medical fields and have appeared in high-impact journals. This article reviews NRI statistics and describes several issues with them. Problems include unacceptable statistical behavior, incorrect statistical inferences, and lack of interpretability. NRI statistics are unhelpful (at best) and misleading (at worst).
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
ObjectType-Review-3
content type line 23
ISSN:0033-8419
1527-1315
1527-1315
DOI:10.1148/radiol.222343