Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

ROC curve estimation when covariates affect the verification process.

C Rodenberg1, X H Zhou

  • 1Procter and Gamble, Inc., Cincinnati, Ohio, USA.

Biometrics
|December 29, 2000
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Discovery of an I^{π}=10^{+} Isomer in ^{150}Yb: Nature of the Longest 10^{+} Isomeric Chain.

Physical review letters·2026
Same author

Isomer Depletion of ^{93m}Mo Triggered by Inelastic Nuclear Scattering Rather than Nuclear Excitation by Electron Capture.

Physical review letters·2026
Same author

[Diagnostic value of symptomatic condylar angle for limited mouth opening in patients with temporomandibular disorders based on MRI].

Zhonghua yi xue za zhi·2026
Same author

[Selection and identification of S100B-specific nanobodies for melanoma biomarker].

Zhonghua zhong liu za zhi [Chinese journal of oncology]·2025
Same author

Z=14 Magicity Revealed by the Mass of the Proton Dripline Nucleus ^{22}Si.

Physical review letters·2025
Same author

[Risk factors of coronal imbalance following posterior lumbar short-level fixation and fusion for degenerative lumbar scoliosis].

Zhonghua yi xue za zhi·2025

This study introduces a method to correct verification bias in medical test accuracy estimation using receiver operating characteristic (ROC) curves. The approach adjusts for patient selection factors, improving diagnostic test evaluation.

Area of Science:

  • Biostatistics
  • Medical Informatics
  • Diagnostic Accuracy Studies

Background:

  • Receiver operating characteristic (ROC) curves are standard for evaluating medical diagnostic test accuracy.
  • Verification bias can distort ROC curve estimations when patient selection for verification is non-random.
  • Existing methods may not adequately address bias when verification depends on patient characteristics (covariates).

Purpose of the Study:

  • To develop and validate a statistical method for correcting verification bias in ROC curve analysis.
  • To account for covariates that influence both test performance and the likelihood of verification.
  • To provide adjusted maximum likelihood estimates for ROC curves in the presence of covariate-dependent verification.

Main Methods:

  • Application of the Expectation-Maximization (EM) algorithm within ordinal regression models.

Related Experiment Videos

  • Derivation of maximum likelihood (ML) estimates for ROC curves as a function of covariates.
  • Utilizing the observed information matrix for asymptotic variance estimation under the missing-at-random assumption.

    • Main Results:

    • The proposed method provides adjusted ROC curve estimates that account for verification bias influenced by covariates.
    • Asymptotic variance estimates are derived, enabling robust statistical inference.
    • The method is demonstrated on a dementia screening study, showing its practical applicability.

    Conclusions:

    • The EM algorithm applied to ordinal regression effectively corrects for verification bias in ROC analysis.
    • This approach enhances the reliability of diagnostic test accuracy estimates when verification is covariate-dependent.
    • The method offers a robust framework for analyzing data from two-phase diagnostic studies.