Developing and externally validating machine learning models to forecast short-term risk of ventilator-associated pneumonia
Machine learning models trained on routinely collected ICU data with careful labeling can anticipate VAP onset up to a week in advance with strong predictive performance. Model performance generalized to data from an entirely different hospital system despite differences in practice and labeling patterns, but did not perform well when there was poor feature overlap. Future work should focus on real-time prospective evaluation.
Authors
- Alexander V. Misharin
- Alec Peltekian
- Mengjia Kang
- Ankit Agrawal (ORCID: https://orcid.org/0009-0006-1700-2397)
- Zewei Liao
- S. Ehrmann
- Nikolay S. Markov
- G.R. Scott Budinger
- Elsa Tavernier
- Benjamin D. Singer
- Luke V. Rasmussen
- Thomas Stoeger
- Richard G. Wunderink
- Wan-Ting Liao
- Rebecca K. Clepp
- Karolina Senkow
- Catherine A. Gao
Institutions
- Université de Tours (FR)
- Northwestern University (US)
- Inserm (FR)
- University of Chicago (US)
- Centre Hospitalier Universitaire de Tours (FR)
- Clinical Investigation Center Plurithematic Tours (FR)
- Lung Institute (US)
Publication Details
- Journal
- medRxiv
- Published
- 2026-01-30
- DOI
- https://doi.org/10.64898/2026.01.28.26344858
- Primary Topic
- Nosocomial Infections in ICU
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Science Foundation
- American Heart Association
- American Thoracic Society
- Gilead Sciences
- Northwestern University
- Princeton Center for Complex Materials
- National Institutes of Health
- National Heart, Lung, and Blood Institute
- National Institute of Allergy and Infectious Diseases