Impact of emergency department length of stay on in-hospital mortality: a retrospective cohort study

Abstract

Background and importance Emergency Department (ED) workload may lead to ED crowding and increased ED length of stay (LOS). ED crowding has been shown to be associated with adverse events and increasing mortality. We hypothesised that ED-LOS is associated with mortality. Objective To study the relationship between ED-LOS and in-hospital mortality. Design Observational retrospective cohort study. Settings and participants From 1 January 2015 to 30 September 2018, all visits by patients aged 15 or older to one of the two ED at Toulouse University Hospital were screened. Patients admitted to the hospital after ED visits were included. Visits followed by ED discharge, in-ED death or transfer to ICU or another hospital were not included. Outcome measure and analysis The primary outcome was 30-day in-hospital mortality. ED-LOS was defined as time from ED registration to inpatient admission. ED-LOS was categorised according to quartiles [<303 min (Q1), between 303 and 433 minutes (Q2), between 434 and 612 minutes (Q3) and >612 min (Q4)]. A multivariable logistic regression tested the association between ED-LOS and in-hospital mortality. Main results A total of 49 913 patients were admitted to our hospital after ED visits and included in the study. ED-LOS was not independently associated with in-hospital mortality. Compared to ED-LOS < 303 min (Q1, reference), odd-ratios (OR) [95% CI] of in-hospital mortality for Q2, Q3, and Q4 were respectively 0.872 [0.747–1.017], 0.906 [0.777–1.056], and 1.137 [0.985–1.312]. Factors associated to in-hospital mortality were: aged over 75 years (OR [95% CI] = 4.3 [3.8–4.9]), Charlson Comorbidity Index score > 1 (OR [95% CI] = 1.3 [1.1–1.5], and 2.2 [1.9–2.5] for scores 2 and ≥ 3 respectively), high acuity at triage (OR [95% CI] = 3.9 [3.5–4.4]), ED visit at Hospital 1 (OR [95% CI] = 1.6 [1.4–1.7]), and illness diagnosis compared to trauma (OR [95% CI] = 2.1 [1.7–2.6]). Night-time arrival was associated with decreased in-hospital mortality (OR [95% CI] = 0.852 [0.767–0.947]). Conclusion In this retrospective cohort study, there was no independent association between ED-LOS before admission to general non-ICU wards and in-patient mortality.