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

ICU Admission Prediction for Patients With Kawasaki Disease or MIS-C Using Machine Learning.

JiWon Woo1, Rebecca Mosier1, Rishima Mukherjee1

  • 1Department of Biomedical Engineering, Johns Hopkins University Whiting School of Engineering, Baltimore, Maryland, USA.

JACC. Advances
|March 27, 2025
PubMed

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

Using Mixed-Methods Research to Address Stagnant Operative Mortality Rates in Congenital Heart Surgery.

Critical care explorations·2026
Same author

Pneumococcal meningitis among hospitalised children after introduction of pneumococcal conjugate vaccine in India: a sentinel hospital surveillance (2019-2022).

The Lancet regional health. Southeast Asia·2026
Same author

Rare type I collagen variants in early-onset bicuspid aortic valve disease: Overlap with Ehlers-Danlos syndrome and osteogenesis imperfecta.

HGG advances·2026
Same author

Importance of Fostering International Collaboration for Optimal Outcomes of Kawasaki Disease Worldwide: A Science Advisory From the American Heart Association.

Journal of the American Heart Association·2026
Same author

Kawasaki Disease-Related Mortality in Arab Countries: A Multinational Physician Survey from the Kawasaki Disease Arab Initiative (Kawarabi).

Pediatric cardiology·2026
Same author

Infections of Selected Nonvalvular Cardiovascular Devices: A Science Advisory From the American Heart Association.

Journal of the American Heart Association·2026
Summary
This summary is machine-generated.

Machine learning models can predict intensive care unit (ICU) admission for children with Multisystem Inflammatory Syndrome in Children (MIS-C) or Kawasaki Disease (KD). Early prediction aids clinical decisions and may prevent severe deterioration.

Area of Science:

  • Pediatric critical care medicine
  • Computational biology
  • Health informatics

Background:

  • Multisystem inflammatory syndrome in children (MIS-C) and Kawasaki disease (KD) can lead to rapid, unexpected clinical deterioration.
  • Patients may require intensive care unit (ICU) admission due to disease severity.

Purpose of the Study:

  • To develop a machine learning (ML) model for predicting future ICU admission in pediatric patients diagnosed with KD or MIS-C.
  • To enhance clinical decision-making by providing early warnings for potential critical care needs.

Main Methods:

  • Utilized data from 2,539 pediatric patients with MIS-C or KD from the International Kawasaki Disease Registry.
  • Developed and compared snapshot and time-series ML models (logistic regression, XGBoost, random forest) using clinical features.
Keywords:
Kawasaki diseaseclinical deteriorationmachine learningmultisystem inflammatory syndrome in children (MIS-C)risk prediction

Related Experiment Videos

  • Incorporated engineered time-series features to improve predictive accuracy.

    • Main Results:

    • ML models accurately predicted ICU admission within 48 hours.
    • The time-series window-XGBoost model achieved an AUROC of 0.92 and an area under the precision-recall curve of 0.86.
    • Key predictors included high ferritin, anticoagulant/heparin use, high C-reactive protein, and low platelet count.

    Conclusions:

    • Machine learning algorithms demonstrate high efficacy in predicting ICU admission for pediatric MIS-C and KD patients.
    • These predictive models can assist physicians in proactively implementing supportive care measures.
    • Early intervention guided by ML predictions may mitigate the risk of severe clinical deterioration.