Regression models for accuracy estimation

Niels Henrik Bruun
Additional contact information
Niels Henrik Bruun: Aalborg University Hospital

Biostatistics and Epidemiology Virtual Symposium 2026 from Stata Users Group

Abstract: I present regression-based methods for estimating and comparing diagnostic accuracy measures while addressing the STARD 2015 requirements. Key metrics include sensitivity, specificity, AUC, PPV, NPV, and accuracy. True-positive and false-positive rates, independent of prevalence, are estimated using OLS regression with robust variance. The derived measures, PPV, NPV, and accuracy, are computed from prevalence, sensitivity, and specificity using nonlinear formulas. For single-modality analysis, sensitivity and specificity are obtained by regressing test outcomes on the "true" values, such as those obtained from pathology. For multimodality studies on the same subjects, data are stacked with a modality indicator, and mixed-effects models with random intercepts are used to account for correlation. A new confreg command combines regression and nonlinear estimation to estimate accuracy metrics under dependency structures. These methods provide a flexible framework for robust comparisons of diagnostic performance across instruments.

References: Add references at CitEc
Citations:

Downloads: (external link)
http://repec.org/biep2026/Bio26_Bruun.pdf

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:boc:biep26:02

Access Statistics for this paper

More papers in Biostatistics and Epidemiology Virtual Symposium 2026 from Stata Users Group Contact information at EDIRC.
Bibliographic data for series maintained by Christopher F Baum ().