Once diagnostic work-up and first therapy are completed in patients visiting the emergency department (ED), boarding them within the ED until an in-hospital bed became available is a common practice in busy hospitals. Whether this practice may harm the patients remains a debate. We sought to determine whether an ED boarding time longer than 4 h places the patients at increased risk of in-hospital death. This retrospective, propensity score-matched analysis and propensity score-based inverse probability weighting analysis was conducted in an adult ED in a single, academic, 1136-bed hospital in France. All patients hospitalized via the adult ED from January 1, 2013 to March 31, 2018 were included. Hospital mortality (primary outcome) and hospital length of stay (LOS) were assessed in (1) a matched cohort (1:1 matching of ED visits with or without ED boarding time longer than 4 h but similar propensity score to experience an ED boarding time longer than 4 h); and (2) the whole study cohort. Sensitivity analysis to unmeasured confounding and analyses in pre-specified cohorts of patients were conducted. Among 68,632 included ED visits, 17,271 (25.2%) had an ED boarding time longer than 4 h. Conditional logistic regression performed on a 10,581 pair-matched cohort, and generalized estimating equations with adjustment on confounders and stabilized propensity score-based inverse probability weighting applied on the whole cohort showed a significantly increased risk of hospital death in patients experiencing an ED boarding time longer than 4 h: odds ratio (OR) of 1.13 (95% confidence interval [95% CI] 1.05-1.22), P = 0.001; and OR of 1.12 (95% CI 1.03-1.22), P = 0.007, respectively. Sensitivity analyses showed that these findings might be robust to unmeasured confounding. Hospital LOS was significantly longer in patients exposed to ED boarding time longer than 4 h: median difference 2 days (95% CI 1-2) (P < 0.001) in matched analysis and mean difference 1.15 days (95% CI 1.02-1.28) (P < 0.001) in multivariable unmatched analysis. In this single-center propensity score-based cohort analysis, patients experiencing an ED boarding time longer than 4 h before being transferred to an in-patient bed were at increased risk of hospital death.
Keywords: Crowding; Data interpretation; Emergency service; Hospital; Hospital mortality; Propensity score; Statistical.