Identification and management of patient deterioration—Comparing the afferent limb of early warning systems

It is mandated that all Australian hospitals employ a means of detecting patient deterioration, and in doing so, most have chosen to adopt Early Warning Scores (EWS). Over the past three decades, many variations of EWSs have been developed and tested in larger regional/metropolitan hospitals. However, there is a paucity of evidence as to which EWS is most effective in predicting deterioration events in small, poorly resourced regional/rural hospitals. The aim of this study was to inform small, poorly resourced regional/rural hospitals on the selection of the best EWS or class of EWS, to augment patient safety within their context. This multi-phase retrospective case-controlled study compared the efficiency of 12 existing EWSs using patient data from two small regional private hospitals (Phase 1). Outcomes from Phase 1 informed Phase 2 and the development of a new EWS for use in poorly resourced regional private hospitals. The new EWS was then validated using two independent patient cohorts from small, poorly resourced regional/rural public hospitals (n=7) and large, well-resourced public regional/metropolitan hospitals (n=6). Results showed that in small regional private hospitals, the aggregated weighted EWS, called Compass, was most effective in identifying deteriorating patients with an Area under the Receiver Operator Characteristic Curve (AUROC) of 0.747 (CI 0.73-0.76). However, Compass had a low sensitivity of 0.44, meaning that less than 50% of the deteriorating patients achieved an emergency call score. Given the suboptimal efficiency of the 12 EWSs tested, a new, more efficient EWS was developed. The first step in developing a new EWS involved determining the ability of the vital signs in predicting patient deterioration. While vital signs are good indicators of patient deterioration, no single vital sign was found to predict patient outcomes strongly. A rapid heart rate and the need for supplementary oxygen were identified as the best indicators of an impending clinical deterioration event in this patient cohort. Based on these findings, a new combination EWS, called MOD-6 was then developed. This was achieved by adding a single trigger component to Compass, extending the existing vital sign trigger threshold ranges and incorporating a graduated weighted scale for supplementary oxygen use. The new combination EWS, MOD-6, was 20% more effective than Compass when used in poorly resourced regional private hospitals. However, when validating the new MOD-6 using two independent patient cohorts from the public sector, the MOD-6 EWS was no more effective than existing EWSs. This study has produced evidence that EWSs are used differently in large, well-resourced regional/metropolitan hospitals compared to poorly resourced regional/rural hospitals. In the well-resourced hospitals, there is evidence that the EWSs are being used to identify clinical deterioration events, reactively using the EWS scores to guide the escalation of patient care. At the small, poorly resourced hospitals, the EWSs appear to be used proactively to predict patient deterioration and prevent adverse patient events. At these small regional/rural hospitals, staff appear to respond early to changes in the patients’ vital signs, transferring patients out to a better resource facility before triggering an emergency threshold score on the EWS. In conclusion, an aggregated weighted EWS, such as Compass, or a combination EWS such as the new MOD-6 EWS should be considered for implementation in the small, poorly resourced private hospitals. The best indicators of patients requiring transfer out to a higher level of care in this patient cohort are the use of supplementary oxygen and those with tachycardia. Findings from this research will inform nursing practice in small, poorly resourced regional/rural hospitals and positively contribute to patient safety.