P139: The impact of a pancreatitis admission algorithm on emergency department length of stay in a tertiary care academic hospital

Abstract

Background: Emergency department (ED) overcrowding is associated with a broad spectrum of poor medical outcomes, including medical errors, mortality, higher rates of leaving without being seen, and reduced patient and physician satisfaction. The largest contributor to overcrowding is access block – the inability of admitted patients to access in-patient beds from the ED. One component to addressing access block involves streamlining the decision process to rapidly determine which hospital service will admit the patient. Aim Statement: As of Sep 2011, admission algorithms at our institution were supported and formalised. The pancreatitis algorithm clarified whether general surgery or internal medicine would admit ED patients with pancreatitis. We hypothesize that this prior uncertainty delayed the admission decision and prolonged ED length of stay (LOS) for patients with pancreatitis. Our project evaluates whether implementing a pancreatitis admission algorithm at our institution reduced ED time to disposition (TTD) and LOS. Measures & Design: A retrospective review was conducted in a tertiary care academic hospital in Montreal for all adult ED patients diagnosed with pancreatitis from Apr 2010 to Mar 2014. The data was used to plot separate run charts for ED TTD and LOS. Serial measurements of each outcome were used to monitor change and evaluate for special cause variation. The mean ED LOS and TTD before and after algorithm implementation were also compared using the Student's t test. Evaluation/Results: Over four years, a total of 365 ED patients were diagnosed with pancreatitis and 287 (79%) were admitted. The mean ED LOS for patients with pancreatitis decreased following the implementation of an admission algorithm (1616 vs. 1418 mins, p = 0.05). The mean ED TTD was also reduced (1171 vs. 899 mins, p = 0.0006). A non-random signal of change was suggested by a shift above the median prior to algorithm implementation and one below the median following. Discussion/Impact: This project demonstrates that in a busy tertiary care academic hospital, an admission algorithm helped reduce ED TTD and LOS for patients with pancreatitis. This proves especially valuable when considering the potential applicability of such algorithms to other disease processes, such as gastrointestinal bleeding and congestive heart failure, among others. Future studies demonstrating this external applicability, and the impact of such decision algorithms on physician decision fatigue and within non-academic institutions, proves warranted.