Early Warning Score Accuracy for Older Adults in the ED

A retrospective prognostic study conducted in Rome, Italy, has found that five different early warning scores (EWSs) provide acceptable short-term accuracy in predicting clinical deterioration for patients aged 80 years or older visiting an emergency department (ED).

The study focused on identifying clinical deterioration, which was defined as either death or admission to an intensive care unit (ICU) within 24 hours of the patient’s arrival at the emergency department.

Study Scope and Methodology

The research included all consecutive nontrauma ED visits by patients aged 80 years or older between January 2015 and December 2024 at a large urban teaching hospital in Rome.

Researchers analyzed a total of 50,645 patients, with a median age of 85 years and an interquartile range of 82 to 88 years. Of these participants, 54.6% were female.

Out of the total patient population, 1,233 individuals, or 2.4%, experienced the primary outcome of clinical deterioration within the 24-hour window.

Performance of Early Warning Scores

The study evaluated five specific scoring systems based on physiologic parameters recorded at the time of admission:

• National Early Warning Score (NEWS)
• National Early Warning Score 2 (NEWS2)
• Modified Early Warning Score (MEWS)
• Rapid Emergency Medicine Score (REMS)
• International Early Warning Score (IEWS)

All five scores demonstrated fair discrimination, with the area under the receiver operating characteristic curve (AUROC) ranging from 0.747 for MEWS to 0.782 for NEWS.

While NEWS achieved the highest AUROC, the Rapid Emergency Medicine Score (REMS) showed the best calibration, recording a Brier score of 0.0220.

The findings suggest that REMS provided the most consistent performance specifically for patients aged 94 years or older, indicating its potential utility for targeted risk stratification in the oldest patient populations.

Clinical Context for Older Adults

The increasing use of emergency departments by older adults has created a critical need for the early and accurate identification of patients at risk of rapid decline.

Prior to this study, data regarding the performance of EWSs in identifying clinical deterioration among acutely ill adults aged 80 years or older were considered scarce.

Early warning scores are widely utilized in emergency settings to support triage and risk stratification, helping clinicians determine which patients require more intensive monitoring or immediate intervention.

The research utilized Shapley additive explanations (SHAP) values to analyze the comparative contributions of different variables to the scores’ predictive power.