Comparing the effectiveness of early warning scores in predicting adult patient outcomes in the emergency department

Objective:
This study compared the effectiveness of the Modified Early Warning Score, the National Early Warning Score, the VitalPAC Early Warning Score (ViEWS), the Rapid Acute Physiology Score, the Rapid Emergency Medicine Score (REMS), and the Worthing Physiological Scoring system for predicting hospitalization, intensive care unit (ICU) admission and in-hospital mortality.
Method:
Our study was performed on all the patients admitted to the emergency department, except those who were pregnant, dead on arrival or self-discharge patients between 1 January 2019 and 31 March 2019. The outcomes of the study were hospitalization, ICU admission, and in-hospital mortality. Receiver operating characteristics (ROC) curve analysis was performed to evaluate and compare the six scores.
Results:
REMS was found to have the greatest efficacy for the prediction of hospitalization. The area under the ROC curve (AUROC) was 0.705. ViEWS was found to have the greatest efficacy for prediction of ICU admission and in-hospital mortality (include the 2-day and 7-day mortality). AUROC was 0.727 in ICU admission, 0.918 in 2-day mortality, 0.813 in 7-day mortality, and 0.775 in in-hospital mortality.
Conclusion:
The efficiency of ViEWS was better than other scores for predicting the patients’ severity. For predicting hospitalization, REMS could be the best predictor because of the age variable.

Key words: Emergency departments; Early Warning Score; Outcome; Mortality