Work

Integrating the interpretable machine learning Score for Emergency Risk Prediction (SERP) with emergency department triage to better predict 30-day mortality

Year: 2026

Type: article

Abstract: OBJECTIVE: This study integrated a machine learning-based Score for Emergency Risk Prediction (SERP), developed using objective mortality endpoints, with the Patient Acuity Category Scale (PACS) and evaluated its predictive performance in clinical use. METHODS: This single-center, retrospective cohort study included all emergency department (ED) patients from a large tertiary hospital between January 1, 2018, and December 31, 2019. Using a reclassification framework, SERP was incorporated into PACS to derive two enhanced triage models. PACS+ model 1 down-triaged patients with low predicted 30-day mortality risk and up-triaged those with high risk. PACS+ model 2 up-triaged only high-risk patients, whereas lowrisk patients retained their original category. Predictive performance in the test cohort was assessed using the area under the receiver operating characteristic curve (AUC) and decision curve analysis. RESULTS: The derivation cohort included 97,188 ED visits, and the test cohort included 97,212 ED visits. In the derivation set, the mean age of the patients was 58.97±18.41 years, and 47,993 (49.4%) were female. Overall, 19.9%, 57.4%, 22.5%, and 0.2% of patients were triaged to PACS categories P1-P4, respectively. The 30-day mortality rate was 2.8% in the derivation set and 2.7% in the test cohort. For 30-day mortality prediction, PACS+ model 1 (AUC, 0.828; 95% confidence interval [CI], 0.820-0.836) and PACS+ model 2 (AUC, 0.812; 95% CI, 0.805-0.818) outperformed PACS (AUC, 0.722; 95% CI, 0.714-0.729). PACS+ model 1 consistently achieved greater net benefit across the range of clinical thresholds. CONCLUSION: Integrating machine learning-based SERP with PACS improved 30-day mortality prediction in ED triage.

Source: Clinical and Experimental Emergency Medicine

DOI: 10.15441/ceem.25.253

Authors: Yvonne Wong Qi Feng, Yohei Okada, Stephanie Fook-Chong, Yilin Ning, Kenneth Boon Kiat Tan, Marcus Eng Hock Ong

Institutions: Duke-NUS Medical School, Kyoto University, Singapore General Hospital

Language: English

Topic: Emergency and Acute Care Studies

Subfield: Emergency Medicine

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: diamond