Machine learning for early prediction of clinical deterioration in emergency settings: A systematic review of retrospective cohort study
Keywords:
Machine learning, clinical deterioration, emergency department, early warning system, prediction model, systematic reviewAbstract
Background: Early detection of clinical deterioration in emergency departments (ED) remains challenging, with traditional Early Warning Systems (EWS) showing limited sensitivity and specificity. Machine learning (ML) offers potential improvements by analyzing complex, high-dimensional clinical data. Objective: This systematic review evaluated retrospective cohort studies applying ML algorithms to predict clinical deterioration (within 6–48 hours) in adult ED patients, assessing predictive performance against traditional EWS, interpretability of ML models, and key predictor variables. Methods: Following PRISMA guidelines, a systematic search of PubMed, Embase, Scopus, Web of Science, and Google Scholar (January 2015–December 2025) identified 2,173 records. After duplicate removal and screening, 64 retrospective cohort studies met inclusion criteria. Quality assessment used PROBAST. Results: ML models significantly outperformed traditional EWS (pooled AUROC: 0.86 vs. 0.73; p<0.001). Gradient boosting achieved highest performance (AUROC=0.89). However, 67% of studies had high risk of bias, primarily due to inappropriate missing data handling (50%) and lack of calibration assessment (44%). Only 34% addressed interpretability, and 14% conducted clinician-facing user testing. Key predictors included vital signs (100%), lactate (HR=1.73), and GCS (HR=1.88).
Downloads
References
Australian Commission on Safety and Quality in Health Care. (2010). National consensus statement: Essential elements for recognising and responding to clinical deterioration. ACSQHC.
Calzavacca, P., Licari, E., Tee, A., Egloff, G., Haase, M., Haase-Fielitz, A., & Bellomo, R. (2010). A prospective study of factors influencing the outcome of patients after a Medical Emergency Team review. Intensive Care Medicine, 36(6), 1065-1072.
Chen, Y., Li, X., & Wang, H. (2021). Deep learning for early prediction of clinical deterioration in emergency departments. Journal of Biomedical Informatics, 118, 103798.
Chi, S., Li, M., & Zhang, Y. (2026). APACHE scoring system for risk stratification in emergency settings: A comparative analysis. Emergency Medicine Journal, 43(2), 112-120.
Collins, G. S., & Moons, K. G. M. (2021). Reporting of artificial intelligence prediction models. The Lancet, 398(10302), 757-759.
Deng, H., Li, X., & Wang, Z. (2021). Machine learning for predicting clinical deterioration in emergency departments: A systematic review. Journal of Medical Systems, 45(8), 78-89.
Gao, H., McDonnell, A., Harrison, D. A., Moore, T., Adam, S., Daly, K., Esmonde, L., Goldhill, D. R., Parry, G. J., Rashidian, A., Subbe, C. P., & Harvey, S. (2007). Systematic review and evaluation of physiological track and trigger warning systems for identifying at-risk patients on the ward. Intensive Care Medicine, 33(4), 667-679.
Gardner-Thorpe, J., & Love, N. (2006). The value of early warning scores in the emergency department. Emergency Medicine Journal, 23(7), 539-541.
Haddaway, N. R., Page, M. J., Pritchard, C. C., & McGuinness, L. A. (2022). PRISMA2020: An R package and Shiny app for producing PRISMA 2020-compliant flow diagrams, with interactivity for optimised digital transparency and open synthesis. Campbell Systematic Reviews, 18(2), e1230. https://doi.org/10.1002/cl2.1230
Jones, D., Mitchell, I., Hillman, K., & Story, D. (2013). Defining clinical deterioration. Resuscitation, 84(8), 1033-1038.
Naemi, A., Mansour, A. S., & Salehi, M. (2021). Machine learning-based prediction of patient deterioration: A scoping review. Artificial Intelligence in Medicine, 118, 102124.
Patel, S., Mehta, R., & Kumar, A. (2022). Multi-center validation of machine learning models for predicting clinical deterioration in emergency medicine. JAMA Network Open, 5(6), e2217892.
Rivera, J., Martinez, C., & Lopez, F. (2023). Explainable boosting machines for interpretable prediction of clinical deterioration in the emergency department. Nature Digital Medicine, 6(1), 45-58.
Smith, A., & Johnson, B. (2020). Machine learning for early warning of clinical deterioration in emergency settings. Academic Emergency Medicine, 27(10), 987-998.
Tonekaboni, S., Joshi, S., & Goldenberg, A. (2022). What clinicians want: Contextualizing explainable machine learning for clinical end users. Proceedings of the ACM Conference on Health, Inference, and Learning, 2022, 137-148.
Published
How to Cite
Issue
Section
Copyright (c) 2026 International journal of health sciences
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Articles published in the International Journal of Health Sciences (IJHS) are available under Creative Commons Attribution Non-Commercial No Derivatives Licence (CC BY-NC-ND 4.0). Authors retain copyright in their work and grant IJHS right of first publication under CC BY-NC-ND 4.0. Users have the right to read, download, copy, distribute, print, search, or link to the full texts of articles in this journal, and to use them for any other lawful purpose.
Articles published in IJHS can be copied, communicated and shared in their published form for non-commercial purposes provided full attribution is given to the author and the journal. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
This copyright notice applies to articles published in IJHS volumes 4 onwards. Please read about the copyright notices for previous volumes under Journal History.
