Incorporation of age into patient early warning scores significantly improves mortality prediction

Abstract

Background Age is a critical factor for the assessment of patients attended by emergency medical services (EMSs). However, how age modifies early warning scores’ (EWSs) predictive ability should be unveiled. Aim To determine how age influences the performance of EWS [National Early Warning Score 2 (NEWS2), VitalPAC-Early Warning Score (ViEWS), Rapid Acute Physiology Score (RAPS) and modified Rapid Emergency Medicine Score (mREMS)] to predict 2-day mortality. The secondary objective was to determine the performance of EWSs at different age ranges. Design A prospective, observational study performed between November 2019 and July 2023. Methods A multicenter, ambulance-based study, considering 38 basic life support units and six advanced life support units referring to four tertiary care hospitals. Eligible patients were adults recruited from among all phone requests for emergency assistance who were later evacuated to emergency departments. The primary outcome was 2-day in-hospital mortality (includes all-cause mortality). The main measures were demographical and vital signs needed for EWS calculation. Results and discussion A total of 8028 participants fulfilled the inclusion criteria, with 7654 survivors and 374 non-survivors. Among age ranges, the 2-day mortality was 2.8% for the ≤44 years, 3.3% for the 45–64 years, 4.1% for the 65–74 years and 6.7% for the ≥75-year age group. The inclusion of age significantly improved the Area Under the Curve (AUC) in all the scores (P = 0.006 for non-age-adjusted mREMS, P = 0.001 for NEWS2, P = 0.002 for ViEWS, P = 0.028 for RAPS, all compared with their counterparts with age). Conclusion Our results demonstrated that the incorporation of age into the EWS improved the performance of the scores. These results will allow the EMS to improve patient management and resource optimization by including an easy-to-obtain variable.