EUVIMEDEUVIMED entdecken
VergleichsstudieOpen Access · CC BY 4.0

Machine Learning Versus Traditional Statistical Models for Emergency Department Volume Forecasting

Machine Learning versus traditionelle statistische Modelle für die Prognose des Notaufnahme-Patientenvolumens

Kian Huang, Enola Okonkwo, Haris Choudhary, Rahul Mhaskar, Nathan Lapaix, Michael Lozano, Reginald Saint-Hilaire

Western Journal of Emergency Medicine · 14.10.2026 · 27 (6)

DOI: 10.5811/westjem.62266 ↗

Auf einen Blick

SARIMA, XGBoost und LGBM prognostizierten das tägliche Patientenvolumen 7 Tage im Voraus ähnlich gut. Alle drei lagen im Mittel bei etwa 9 bis 10 Patienten pro Tag neben dem Ist-Wert.

Fragestellung

Welche der fünf Prognosemethoden sagt das Notaufnahme-Patientenvolumen 7 Tage im Voraus am genauesten voraus?

Methode

Retrospektive Studie mit 3.202 aufeinanderfolgenden Tagen täglicher Besuchsdaten aus einer akademischen freistehenden Notaufnahme; Vergleich von 5 Vorhersagemodellen über 453 rollierende Wochenfenster.

Ergebnisse

SARIMA, XGBoost und LGBM hatten die niedrigsten MAE: 9,60 (95 %-KI 9,29–9,91), 9,61 (95 %-KI 9,31–9,91) und 9,64 (95 %-KI 9,33–9,94) Patienten/Tag. ARIMA: 10,54; LSTM: 12,01. XGBoost war besser als ARIMA und LSTM (DM −6,09 bzw. −6,96; P < .001).

Schlussfolgerung der Autor:innen

SARIMA, XGBoost und LGBM schnitten vergleichbar gut ab und lieferten laut Autor:innen zuverlässige, operational nutzbare Prognosen. SARIMA als traditionelle Methode war ebenso gut wie die neuen Machine-Learning-Modelle.

Kernaussagen

  • SARIMA, XGBoost und LGBM waren die besten Modelle
  • MAE etwa 9,6 Patienten/Tag
  • SARIMA war einfach und nicht schlechter als die ML-Modelle

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

Kurz gefasst

Vergleichsstudie in der Zeitschrift Western Journal of Emergency Medicine, online erschienen am 14.10.2026. Verfasst von Kian Huang und 6 weiteren Autor:innen. Beteiligte Einrichtung: University of South Florida, Morsani College of Medicine, Tampa, Florida. Hauptthemen laut MEDLINE: Emergency Service, Hospital, Machine Learning, Models, Statistical. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Emergency department (ED) crowding harms patient care and strains hospital operations. Forecasting daily patient arrivals can help ED medical directors, department chairs, and hospital administrators plan staffing and bed capacity in advance, rather than reacting after a surge has already occurred. This study compared five computer-based forecasting methods, including three machine learning models (eXtreme gradient boosting [XGBoost], light gradient boosting machine [LGBM], and long short-term memory [LSTM]) and two traditional statistical models (autoregressive integrated moving average [ARIMA] and seasonal ARIMA [SARIMA]), to determine which most accurately predicts ED patient volume seven days in advance, using eight years of daily visit data from an academic freestanding ED.

Methods

This retrospective study analyzed 3,202 consecutive days of daily patient volume (March 27, 2017, to December 31, 2025) from an academic freestanding ED with approximately 31,102 annual visits. Each model used the preceding 365 days of data to forecast the subsequent 7 days, repeated across the study period to reflect the weekly cycle by which most EDs plan staffing. Our primary measure of accuracy was mean absolute error (MAE), defined as the average number of patients per day by which a forecast missed the actual volume. Secondary measures included root mean squared error (RMSE), mean absolute percentage error (MAPE), the coefficient of determination (R², a score from 0 to 1 reflecting how well predictions tracked actual day-to-day volume), and the percentage of forecasts accurate within 5, 10, 15, and 20 patients, thresholds selected for their relevance to staffing decisions. The best-performing model on initial metrics, XGBoost, served as the reference model for statistical comparison using the Diebold-Mariano (DM) test, which determines whether an observed difference in accuracy between two models is statistically meaningful or attributable to chance.

Results

Among 453 rolling weekly forecast windows, SARIMA, XGBoost, and LGBM demonstrated the lowest absolute errors and best forecasting of patient volume, with MAE values of 9.60 (95% CI, 9.29–9.91), 9.61 (95% CI, 9.31–9.91), and 9.64 (95% CI, 9.33–9.94) patients/day, respectively; these three models therefore missed the actual daily patient count by approximately 9 to 10 patients, on average. Models ARIMA and LSTM were less accurate, with mean MAE values of 10.54 (95% CI, 10.21–10.87) and 12.01 (95% CI, 11.63–12.39) patients per day. Relative to the XGBoost reference (MAPE 13.82%), ARIMA (MAPE 14.93%) and LSTM (MAPE 17.27%) showed a marked decrease in predictive accuracy. For operational planning, the top-performing models (SARIMA, XGBoost, and LGBM) correctly predicted daily volume within a margin of 20 patients in 90.3% to 90.6% of instances, while LSTM performed less well and reached this threshold in only 83.3% of forecasts. Statistical testing confirmed these differences were unlikely due to chance, with XGBoost significantly outperforming ARIMA (DM = −6.09, P < .001) and LSTM (DM = −6.96, P < .001).

Conclusion

Models SARIMA, XGBoost, and LGBM performed comparably well, and all three would provide reliable, operationally useful forecasts for ED staffing and resource allocation decisions. A traditional statistical method, SARIMA, which is simpler and less resource-intensive to implement, performed as well as the newer machine learning methods evaluated in this study. This finding suggests that hospitals do not necessarily need to invest in more complex artificial intelligence tools to obtain accurate ED volume forecasts.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Emergency Service, HospitalMachine LearningModels, StatisticalHumansRetrospective StudiesForecastingBoosting Machine Learning AlgorithmsPredictive Learning ModelsLong Short Term MemoryEmergency Room VisitsCrowdingPrediction Algorithms

Autor:innen und Institutionen

  • University of South Florida, Morsani College of Medicine, Tampa, Florida.
  • University of South Florida, Morsani College of Medicine, Tampa General Hospital, Department of Emergency Medicine, Tampa, Florida.
  • University of South Florida, Morsani College of Medicine, Division of Evidence-based Medicine and Outcomes Research, Tampa, Florida.

Zitieren

APA

Huang, K., Okonkwo, E., Choudhary, H., Mhaskar, R., Lapaix, N., Lozano, M., & Saint-Hilaire, R. (2026). Machine Learning Versus Traditional Statistical Models for Emergency Department Volume Forecasting. Western Journal of Emergency Medicine, 27(6). https://doi.org/10.5811/westjem.62266

Vancouver

Huang K, Okonkwo E, Choudhary H, Mhaskar R, Lapaix N, Lozano M, et al. Machine Learning Versus Traditional Statistical Models for Emergency Department Volume Forecasting. Western Journal of Emergency Medicine. 2026;27(6). doi:10.5811/westjem.62266

BibTeX
@article{Huang2026,
  author = {Huang, Kian and Okonkwo, Enola and Choudhary, Haris and Mhaskar, Rahul and Lapaix, Nathan and Lozano, Michael and Saint-Hilaire, Reginald},
  title = {Machine Learning Versus Traditional Statistical Models for Emergency Department Volume Forecasting},
  journal = {Western Journal of Emergency Medicine},
  year = {2026},
  volume = {27},
  number = {6},
  doi = {10.5811/westjem.62266}
}
Kostenlos starten →