Can early warning scores identify deteriorating patients in pre-hospital settings? A systematic review

Abstract

Objective: To evaluate the effectiveness and predictive accuracy of early warning scores (EWS) to predict deteriorating patients in pre-hospital settings.

Methods: Systematic review. Seven databases searched to August 2017. Study quality was assessed using QUADAS-2. A narrative synthesis is presented.

Eligibility: Studies that evaluated EWS predictive accuracy or that compared outcomes in populations that did or did not use EWS, in any pre-hospital setting were eligible for inclusion. EWS were included if they aggregated three or more physiological parameters.

Results: Seventeen studies (157,878 participants) of predictive accuracy were included (16 in ambulance service and 1 in nursing home). AUCs ranged from 0.50 (CI not reported) to 0.89 (95%CI 0.82, 0.96). AUCs were generally higher (>0.80) for prediction of mortality within short time frames or for combination outcomes that included mortality and ICU admission. Few patients with low scores died at any time point. Patients with high scores were at risk of deterioration. Results were less clear for intermediate thresholds (≥4 or 5). Five studies were judged at low or unclear risk of bias, all others were judged at high risk of bias.

Conclusions: Very low and high EWS are able to discriminate between patients who are not likely and those who are likely to deteriorate in the pre-hospital setting. No study compared outcomes pre- and post-implementation of EWS so there is no evidence on whether patient outcomes differ between pre-hospital settings that do and do not use EWS. Further studies are required to address this question and to evaluate EWS in pre-hospital settings.

Keywords: Critical care; Deteriorating patients; Early warning score; Pre hospital setting; Track and trigger system.