Development and External Validation of a Machine Learning Model for Prediction of Potential Transfer to the PICU.

Unrecognized clinical deterioration during illness requiring hospitalization is associated with high risk of mortality and long-term morbidity among children. Our objective was to develop and externally validate machine learning algorithms using electronic health records for identifying ICU transfer within 12 hours indicative of a child's condition. Observational cohort study. Two urban, tertiary-care, academic hospitals (sites 1 and 2). Pediatric inpatients (age <18 yr). None. Our primary outcome was direct ward to ICU transfer. Using age, vital signs, and laboratory results, we derived logistic regression with regularization, restricted cubic spline regression, random forest, and gradient boosted machine learning models. Among 50,830 admissions at site 1 and 88,970 admissions at site 2, 1,993 (3.92%) and 2,317 (2.60%) experienced the primary outcome, respectively. Site 1 data were split longitudinally into derivation (2009-2017) and validation (2018-2019), whereas site 2 constituted the external test cohort. Across both sites, the gradient boosted machine was the most accurate model and outperformed a modified version of the Bedside Pediatric Early Warning Score that only used physiologic variables in terms of discrimination ( C -statistic site 1: 0.84 vs 0.71, p < 0.001; site 2: 0.80 vs 0.74, p < 0.001), sensitivity, specificity, and number needed to alert. We developed and externally validated a novel machine learning model that identifies ICU transfers in hospitalized children more accurately than current tools. Our model enables early detection of children at risk for deterioration, thereby creating opportunities for intervention and improvement in outcomes.

Informations de copyright

Copyright © 2022 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.

Déclaration de conflit d'intérêts

Dr. Mayampurath received funding from the National Heart, Lung, and Blood Institutes (NHLBI; K01HL148390); he received support for article research from the National Institutes of Health (NIH). Dr. Mayampurath has performed consulting services for Litmus Health, Austin, TX, outside of submitted work. Ms. Hegermiller received funding from FabFitFun, Indus Consulting Inc., University of Chicago, and Northwell Health, outside of submitted work. Dr. Edelson received funding from the NIH/University of Wisconsin-Madison (R01 GM123193), the NIH/Idaho State University (R01 GM137083), the U.S. Health and Human Services/AgileMD (ASPR-BARDA 21-00592) (Principal Investigator), the Department of Defense/University of Wisconsin-Madison, and EarlySense; she disclosed that she owns equity in Agile MD stock and serves as the President and Co-Founder of AgileMD, and that she owns Intellectual Property Rights (ARCD.P0535US.P2); she also disclosed that AgileMD acquired the exclusive licensing rights to electronic Cardiac Arrest Risk and Triage when it acquired QuantHC. She received research support and honoraria from Philips Healthcare (Andover, MA). Dr. Churpek’s institution received funding from the National Institute of General Medical Sciences (R01 GM123193-05), the Department of Defense (PRMRP-W81XWH-21-1-0009), the NHLBI (R01- HL157262), the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK126933-A1), the National Institute on Aging (R21 AG068720-01), the National Institute on Drug Abuse (R01 DA051464-01), and an EarlySense Research Grant; he disclosed that he has a patent-pending (ARCD. P0535US.P2) and that he is a consultant with AgileMD. The remaining authors have disclosed that they do not have any potential conflicts of interest.