Longitudinal Model Shifts of Machine Learning-Based Clinical Risk Prediction Models: Evaluation Study of Multiple Use Cases Across Different Hospitals.

Journal: Journal of medical Internet research
PMID:

Abstract

BACKGROUND: In recent years, machine learning (ML)-based models have been widely used in clinical domains to predict clinical risk events. However, in production, the performances of such models heavily rely on changes in the system and data. The dynamic nature of the system environment, characterized by continuous changes, has significant implications for prediction models, leading to performance degradation and reduced clinical efficacy. Thus, monitoring model shifts and evaluating their impact on prediction models are of utmost importance.

Authors

  • Patricia Cabanillas Silva
    Dedalus Healthcare, Antwerp, Belgium.
  • Hong Sun
    Beijing Hospital, National Center of Gerontology, Institute of Geriatric Medicine, Chinese Academy of Medical Sciences, Beijing, China.
  • Mohamed Rezk
    Dedalus HealthCare, Antwerp, Belgium.
  • Diana M Roccaro-Waldmeyer
    Dedalus Healthcare, Antwerp, Belgium.
  • Janis Fliegenschmidt
    Institute of Anesthesiology and Pain Therapy, Heart and Diabetes Centre North Rhine-Westphalia, University Hospital of Ruhr-University Bochum, Bad Oeynhausen, Germany.
  • Nikolai Hulde
    Institute of Anesthesiology, Heart and Diabetes Centre NRW, Ruhr-University Bochum, Bad Oeynhausen, Germany. Electronic address: nhulde@hdz-nrw.de.
  • Vera von Dossow
    Institute of Anesthesiology, Heart and Diabetes Centre NRW, Ruhr-University Bochum, Bad Oeynhausen, Germany. Electronic address: vvondossow@hdz-nrw.de.
  • Laurent Meesseman
    Dedalus HealthCare, Roderveldlaan 2, 2600 Antwerp, Belgium.
  • Kristof Depraetere
    Dedalus HealthCare, Roderveldlaan 2, 2600 Antwerp, Belgium. Electronic address: kristof.depraetere@dedalus-group.com.
  • Joerg Stieg
    Dedalus HealthCare, Antwerp, Belgium.
  • Ralph Szymanowsky
    Dedalus HealthCare, Roderveldlaan 2, 2600 Antwerp, Belgium. Electronic address: ralph.szymanowsky@dedalus-group.com.
  • Fried-Michael Dahlweid
    Dedalus Healthcare, Antwerp, Belgium.