1041

Abstract

Introduction: Interventions for severe sepsis are time sensitive. There are limited studies of electronic decision support tools for the identification of severe sepsis. The utility of a computerized sepsis alert as part of a comprehensive sepsis program is unclear. Hypothesis: To determine whether use of an electronic sepsis prompt to identify severe sepsis in patients admitted to an intensive care unit improves bundle compliance and outcomes. Methods: This was a single center cohort study at a community hospital. 484 consecutive patients who presented with criteria for severe sepsis or septic shock over a period of 18 months were identified by one of three ways – at presentation in the emergency department (ED), by the bedside clinician in the intensive care unit (ICU) or a third category (eICU) identified by an electronic sepsis prompt during tele-ICU surveillance (PHILIPS VISICU). The prompt alerted the intensivist in the tele-ICU in real time when systemic inflammatory response syndrome criteria and evidence of organ dysfunction were detected. Interventions were protocol driven in the form of a 6 hour sepsis bundle including fluids, antibiotics and early goal directed therapy (EGDT) for the patients who remained hypotensive after initial fluid challenge. Results: At baseline, the 3 groups were comparable in age; there were more males in the ED group. Their acute physiology and APACHE IV scores measured within the first 24 hours of ICU admission were comparable. The serum lactate levels were lower in the ICU group (3.5+/- 4.1) compared to the ED group (4.6+/- 3.6, p=0.028). More patients in the ICU and eICU groups underwent all the components of EGDT. There was better bundle compliance at 6 hours in the ICU (62.4%, p=0.001) & eICU (55.3%, p=0.031) groups compared to the ED (43.3%). Hospital mortality was lower in ED group as compared to the eICU group (20.3% vs. 30%, p = 0.044). ICU length of stay was higher in the eICU group (5.0 +/- 5.3) compared to the ED group (3.7 +/- 4.0, p = 0.04). Hospital length of stay was lowest in the ED group (8.4 +/- 6.9). Conclusions: An electronic sepsis prompt enables early identification and bundled intervention for patients with severe sepsis and can be incorporated as part of a sepsis program. This group of patients has worse outcomes compared to that identified in the emergency department despite better bundle compliance and similar severity of illness.