Emergency Department Triage Early Warning Score (TREWS) predicts in-hospital mortality in the emergency department

Abstract

Aim

The purpose is to assess the adequacy of the National Early Warning Score (NEWS) in the emergency department (ED) and the usefulness of the Triage in Emergency Department Early Warning Score (TREWS) that has been developed using the NEWS in the ED.

Methods

In this retrospective observational cohort study, we performed univariable and multivariable regression analyses with 81,520 consecutive ED patients to develop a new scoring system, the TREWS. The primary outcome was in-hospital mortality within 24 h, and secondary outcomes were in-hospital mortality within 48 h, 7 days, and 30 days. The prognostic properties of the TREWS were compared with those of the NEWS, Modified Early Warning Score (MEWS), and Rapid Emergency Medicine Score (REMS) using the area under the receiver operating characteristic curve (AUC) technique.

Results

The AUC of the TREWS for in-hospital mortality within 24 h was 0.906 (95% CI, 0.903–0.908), those of the NEWS, MEWS, and REMS were 0.878 (95% CI, 0.875–0.881), 0.857 (95% CI, 0.854–0.860), and 0.834 (95% CI, 0.831–0.837), respectively. Differences in the AUC between the TREWS and NEWS, the TREWS and MEWS, and the TREWS and REMS were 0.028 (95% CI, 0.022–0.033; p < .001), 0.049 (95% CI, 0.041–0.057; p < .001), and 0.072 (95% CI, 0.063–0.080; p < .001), respectively. The TREWS showed significantly superior performance in predicting secondary outcomes.

Conclusion

The TREWS predicts in-hospital mortality within 24 h, 48 h, 7 days, and 30 days better than the NEWS, MEWS, and REMS for patients arriving at the ED.