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

A Post-Processing Fairness Mitigation Method for Medical Prediction Models.

Noga Stern, Inbal Livni Navon, Omer Reingold

    IEEE Journal of Biomedical and Health Informatics
    |March 31, 2026
    PubMed
    Summary
    This summary is machine-generated.

    Related Concept Videos

    Regression Toward the Mean01:52

    Regression Toward the Mean

    7.3K
    Regression toward the mean (“RTM”) is a phenomenon in which extremely high or low values—for example, and individual’s blood pressure at a particular moment—appear closer to a group’s average upon remeasuring. Although this statistical peculiarity is the result of random error and chance, it has been problematic across various medical, scientific, financial and psychological applications. In particular, RTM, if not taken into account, can interfere when...
    7.3K

    You might also read

    Related Articles

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

    Sort by
    Same author

    Embeddings of clinical codes enable knowledge-grounded AI in medicine.

    NPJ digital medicine·2026
    Same author

    Learning Normal Representations for Blood Biomarkers.

    ArXiv·2026
    Same author

    Effect of cumulative blood pressure exposure on long-term cardiovascular outcomes in the community, a nationwide cohort study.

    The American journal of medicine·2026
    Same author

    Minkowski-Space Modeling of Hyperbolic Lenses.

    Physical review letters·2026
    Same author

    Neural tracking at theta predicts drumming-induced altered states of consciousness.

    Scientific reports·2026
    Same author

    West Nile Virus-Neutralizing Plasma for West Nile Virus Disease.

    NEJM evidence·2026

    Resource allocation in medicine faces fairness challenges. This study introduces a post-processing algorithm that improves fairness in prediction models for limited resources, offering practical usability advantages.

    Area of Science:

    • Medical informatics
    • Health equity
    • Machine learning in healthcare

    Background:

    • Resource allocation in medicine is a critical challenge.
    • Prediction models aid decisions but can exhibit bias and unfairness.
    • Ensuring fairness in healthcare AI is essential for equitable patient outcomes.

    Purpose of the Study:

    • To introduce and evaluate a novel post-processing algorithm for enhancing fairness in medical resource allocation models.
    • To compare the performance and usability of this post-processing algorithm against modified pre-processing and in-processing algorithms.
    • To assess the impact of fairness interventions on predictive performance and resource allocation constraints.

    Main Methods:

    • Developed in-hospital mortality predictors using ICU data from two datasets.

    Related Experiment Videos

  • Applied a novel post-processing algorithm to enforce equal opportunity under resource constraints.
  • Compared the post-processing algorithm with modified pre-processing and in-processing algorithms.
  • Evaluated algorithms based on fairness metrics (equal opportunity), predictive performance (sensitivity, positive predictive value), and usability.

    • Main Results:

    • All three algorithms significantly improved the "equal opportunity" metric, reducing sensitivity value span by an average of 52%.
    • No single algorithm consistently outperformed the others in fairness or predictive performance.
    • Enforcing fairness led to a modest average decrease in predictive performance (4% sensitivity, 3% positive predictive value).
    • The post-processing algorithm preserved numerical risk predictions and did not require retraining for intervention group size adjustments, offering usability benefits.

    Conclusions:

    • All evaluated methods enhance fairness in medical prediction models, with comparable overall predictive performance, though sometimes reduced.
    • The developed post-processing algorithm provides practical usability advantages over pre-processing and in-processing alternatives.
    • Fairness interventions are crucial for equitable resource allocation, and the choice of method impacts usability and implementation.