302 Use of a Modified Early Warning Score to Predict Early Clinical Deterioration in Admitted Emergency Department Patients

Abstract

Early Warning Scores (EWS) have been validated as a means of identifying patients at risk of physiologic deterioration and who may benefit from early transfer to an intensive care unit (ICU). Delays in the ICU transfer of these patients (“step-ups”) are associated with increased morbidity and mortality in both admitted and emergency department (ED) patients. Several studies have also found that EWS systems can be used in the ED to predict disposition, cardiac arrest, and even mortality. The goal of this study was to determine if an EWS, based on the last set of vital signs prior to admission, could identify patients at risk for deterioration within the first 48-hours of inpatient care. This retrospective study examined patients presenting to two urban, academic, tertiary care hospitals over a six-month period. All ED patients who required inpatient admission were included for review. The last set of vital signs prior to departure from the ED, as well as data regarding length of stay, admitting service, admitting diagnosis, and mortality was collected. Using this data, a modified EWS was then calculated using the previously validated scoring system. Data from patients who underwent an escalation of care within 48 hours of hospital admission was compared to patients who remained in the ward. Subjects were screened for differences in baseline characteristics, transfer vital signs, EWS score, and ultimate patient outcomes. Data sets were available for 2,240 patients, with 119 of these patients experiencing an ICU step-up. There were no statistically significant differences in age, sex, or race between patients admitted to the ward, ICU, or those who required ICU step-up. The median hospital length-of-stay of ICU step-up patients (172.9 hours, IQR: 92.5-330) mirrored that of patients admitted directly to the ICU (137.8, 79.2-227.1), and was nearly double that of patients who remained on the ward (92.9, 54.1-151.8, p < 0.0001). Furthermore, mortality in ICU step-up patients (8.5%) was significantly higher than patients who remained on the ward (1.4%, p < 0.0001) and approached that of the ICU population (12.3%). The average EWS was significantly higher for step-up patients (1.93 ± 1.72) compared to ward patients (1.10 ± 1.35, p < 0.0001). Additionally, using a previously validated screening cut-off score of 4, roughly 18% of patients who decompensated would have triggered an alert at time of transfer, as compared to only 6% of patients who remained stable. A pre-defined subgroup analysis revealed a significant difference in the transfer EWS score for step-up patients being admitted to Pulmonary Medicine (2.37 ± 2.10 vs. 1.26 ± 1.62, p = 0.002) and those whose admitting diagnosis was sepsis/infection (2.32 ± 1.89 vs. 1.46 ± 1.61, p = 0.02) compared to similar patients who remained on the ward. Patients requiring ICU step-up had significantly increased length of stay and mortality compared to ward patients, consistent with previous studies. Despite only minor differences in individual vital signs, the average EWS for step-up patients was significantly higher than ward patients, and was even more notable in certain patient subgroups. These results support the use of an electronic EWS at time of transfer to screen for patients at risk of early clinical deterioration.