Patient-level, Temporal, and Dynamic Operational Indicators of Emergency Department Prolonged Length of Stay: A Retrospective Cohort Study.

Authors

Alexandre Vallée

Published in

Archives of academic emergency medicine. Volume 14. Issue 1. Pages e25. Epub Jul 02, 2026.

Abstract

Emergency department length of stay (ED-LOS) is a key indicator of crowding and care quality. This study aimed to identify patient-level, temporal, and real-time operational predictors of prolonged ED-LOS.
We conducted a retrospective cohort study using routinely collected data from all ED visits at Foch Hospital, France, between January and November 2025. Prolonged ED-LOS was defined as ED-LOS >8 hours. Data were split chronologically into a training period from January to September and a held-out test period from October to November. Three logistic regression models were evaluated: Model 1 included patient-level and temporal variables; Model 2 additionally included dynamic congestion indicators; and Model 3 further included early process delays.
Among 41,818 ED visits, 41,431 were included in the final analytic sample. Overall, 8,119 (19.6%) visits had ED-LOS >8 hours. In the test set, Model 1 showed good discrimination (area under the receiver operating characteristic curve (AUC): 0.752, 95% confidence interval (CI): 0.740 - 0.766), which improved modestly after adding dynamic congestion variables in Model 2 (AUC: 0.761, 95% CI: 0.749 - 0.775). Model 3 achieved the best performance (AUC: 0.804 (95% CI: 0.793 - 0.815); Brier score: 0.127 (95% CI: 0.122 - 0.132)). Older age, triage acuity level 3 (classification infirmière des malades aux urgences: CIMU ), weekend arrival, dynamic congestion at arrival, and early process delays were the main predictors of prolonged ED-LOS.
Based on the findings, older age, intermediate triage acuity, weekend arrival, dynamic congestion at arrival, and early process delays were the independent predictors of prolonged ED stay. The addition of dynamic congestion variables improved prediction beyond patient-level and temporal characteristics, while the strongest performance was achieved after incorporating early delays to triage and physician assessment.

PMID:
42524472
Bibliographic data and abstract were imported from PubMed on 29 Jul 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.