EUVIMEDOpdag EUVIMED
KohortenstudieOpen access

Explainable Machine Learning for Predicting Hospital Admission at Emergency Department Triage: Retrospective Cohort Study

Erklärbares Machine Learning zur Vorhersage einer Krankenhausaufnahme bei der Triage in der Notaufnahme: retrospektive Kohortenstudie

Kalash Dave, Antony Robert, W Alton Russell, David L Buckeridge

JMIR Medical Informatics · 2026-11-10 · 14 · e98273-e98273

DOI: 10.2196/98273 ↗

Auf einen Blick

Routinedaten aus der Triage konnten eine Krankenhausaufnahme zuverlässig vorhersagen. Das XGBoost-Modell erzielte die beste Leistung.

Fragestellung

Lassen sich Krankenhausaufnahmen am Ende der Triage mit Machine-Learning-Modellen auf Basis von Routinedaten vorhersagen?

Methode

Retrospektive Kohortenstudie zu erwachsenen Notaufnahmebesuchen zwischen Januar 2015 und Dezember 2022 in 2 akademischen Krankenhäusern; Split in Training, Validierung und Test; Vergleich mehrerer Klassifikatoren; SHAP-Analyse.

Ergebnisse

526.145 Besuche; 91.363 Aufnahmen (17,4 %). XGBoost: AUC-ROC 0,86 (95 %-KI 0,858–0,864), Sensitivität 0,67, Spezifität 0,86; LR: AUC 0,84. Wichtige Prädiktoren: Alter, CTAS, Ankunftsmodus, frühere Hospitalisierungen.

Schlussfolgerung der Autor:innen

In urbanen tertiären Versorgungseinrichtungen können strukturierte Triage-Daten eine Krankenhausaufnahme effektiv vorhersagen. Modelle wie XGBoost erscheinen als praktikables, erklärbares Werkzeug für frühe Entscheidungen.

Kernaussagen

  • 526.145 Besuche, 91.363 Aufnahmen (17,4 %)
  • XGBoost mit AUC-ROC 0,86 (95 %-KI 0,858–0,864)
  • SHAP hob Alter, CTAS, Ankunftsmodus und frühere Hospitalisierungen hervor

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

At a glance

Kohortenstudie in JMIR Medical Informatics, published online on 2026-11-10. Written by Kalash Dave and 3 co-authors. Institution: Department of Medicine, McGill University, 845 Sherbrooke Street West, Montreal, QC, H3A 0G4, Canada, 1 4384624410. Major MEDLINE topics: Triage, Emergency Service, Hospital, Machine Learning, Patient Admission. 51 references. The full text is freely available in PubMed Central.

Auf einen BlickAbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Background

Emergency departments (EDs) are vital access points for urgent care, yet persistent overcrowding in health care systems worldwide compromises service quality and efficiency. Urban tertiary care centers are increasingly strained by surging patient numbers and scarce inpatient beds. Prior admission prediction studies mostly rely on limited triage features and discrimination metrics, neglecting operational factors, calibration, and explainability. Identifying likely admissions early could enable proactive resource allocation and shorter wait times.

Objective

This study aimed to develop and evaluate machine learning models for predicting hospital admission at the end of triage using routinely collected structured data while addressing prior limitations regarding feature selection, model calibration, and explainability.

Methods

We conducted a retrospective cohort study of adult ED visits between January 2015 and December 2022 at 2 academic hospitals in the McGill University Health Centre. We excluded visits with dispositions other than admission or discharge to ensure a binary outcome. Data were split into training (70%), validation (10%), and testing (20%) sets using stratified random sampling. We trained classifiers, including logistic regression (LR), decision trees, random forests, artificial neural networks, and gradient boosting algorithms (extreme gradient boosting [XGBoost], light gradient boosting machine, and categorical boosting), using demographics, vital signs, chief complaints, comorbidities, and historical hospitalization frequency available at triage. Sensitivity analyses tested exclusion criteria and temporal data shifts, while subgroup analyses examined differences by age, sex, and site. Performance was evaluated using area under the receiver operating characteristic curve (AUC-ROC), sensitivity, specificity, and calibration. We also performed a decision threshold analysis to assess operational flexibility and used Shapley additive explanations (SHAP) values to interpret feature contributions.

Results

The final cohort included 526,145 visits (mean age 50.9, SD 21.1 years; 272,767/526,145, 51.8% female), with 91,363 patients (17.4%) admitted. The XGBoost model achieved the highest performance, with an AUC-ROC of 0.86 (95% CI 0.858‐0.864), a sensitivity of 0.67, and a specificity of 0.86, outperforming the LR baseline (AUC 0.84). Key predictors identified by SHAP values included age, Canadian Triage and Acuity Scale scores, arrival mode, and historical hospitalization frequency. Subgroup analyses confirmed strong performance across age, sex, and sites. Sensitivity analyses evaluating alternative outcome dispositions and chronological data splits confirmed overall model stability. Threshold analysis showed that adjusting the decision threshold to the 0.70 to 0.75 range improved the positive predictive value and did not disproportionately reclassify older or higher-acuity patients.

Conclusions

In urban tertiary care settings, structured triage data available at triage can effectively anticipate hospital admissions. By incorporating broad feature sets and operational variables, models such as XGBoost offer a feasible, explainable tool for early decision-making. Such tools could improve coordination between emergency and inpatient services and ease overcrowding. Future work should prioritize prospective validation and real-time clinical workflow integration.

Subjects

MeSH terms (MEDLINE)

TriageEmergency Service, HospitalMachine LearningPatient AdmissionHospitalizationHumansRetrospective StudiesFemaleMaleMiddle AgedPredictive Learning ModelsBoosting Machine Learning AlgorithmsEmergency Room VisitsClassification AlgorithmsAdultLogistic ModelsRandom ForestROC CurvePrediction AlgorithmsAgedCohort Studies

Author keywords

bed managementclinical decision support systemsdigital healthemergency department crowdingexplainable AIlength of staymachine learningpatient flowpredictive analyticstriagewaiting time

Authors and institutions

  • Department of Medicine, McGill University, 845 Sherbrooke Street West, Montreal, QC, H3A 0G4, Canada, 1 4384624410.
  • Department of Emergency Medicine, McGill University Health Centre, Montreal, QC, Canada.
  • School of Population and Global Health, McGill University, Montreal, QC, Canada.

Cite

APA

Dave, K., Robert, A., Russell, W. A., & Buckeridge, D. L. (2026). Explainable Machine Learning for Predicting Hospital Admission at Emergency Department Triage: Retrospective Cohort Study. JMIR Medical Informatics, 14, e98273-e98273. https://doi.org/10.2196/98273

Vancouver

Dave K, Robert A, Russell WA, Buckeridge DL. Explainable Machine Learning for Predicting Hospital Admission at Emergency Department Triage: Retrospective Cohort Study. JMIR Medical Informatics. 2026;14:e98273-e98273. doi:10.2196/98273

BibTeX
@article{Dave2026,
  author = {Dave, Kalash and Robert, Antony and Russell, W Alton and Buckeridge, David L},
  title = {Explainable Machine Learning for Predicting Hospital Admission at Emergency Department Triage: Retrospective Cohort Study},
  journal = {JMIR Medical Informatics},
  year = {2026},
  volume = {14},
  pages = {e98273--e98273},
  doi = {10.2196/98273}
}
Start for free →