Operational Determinants of Emergency Department Length of Stay: A 10-Year Big Data Analysis

<title>Abstract</title> <p> <bold>Background</bold> Managing patient flow in the Emergency Department (ED) is a major operational challenge involving multiple interconnected clinical and system-level processes. Emergency Department Length of Stay (LOS-ED) is a widely used operational quality metric associated with patient outcomes, physician workload, and healthcare system efficiency. We aimed to identify operational factors associated with prolonged LOS-ED and to evaluate longitudinal trends in LOS-ED over a 10-year period. <bold>Methods</bold> We conducted a retrospective cohort study using electronic health record data from all ED visits to a large public medical center between 2012 and 2021. Operational parameters associated with LOS-ED, including consultation burden, imaging utilization, continuity of care, patient workload, and time-based workflow metrics, were evaluated using longitudinal and multivariable analyses. <bold>Results</bold> Overall, 917,778 ED visits were included in the analysis. Mean LOS-ED increased during the study period from 216 to 255 minutes and exceeded four hours from 2016 onward. During the same period, the proportion of patients undergoing CT imaging increased from 3.9% to 17.5%, while the proportion requiring at least one consultation increased from 5.1% to 20.5%. Patients requiring advanced imaging studies or multiple consultations demonstrated substantially prolonged LOS-ED. LOS-ED was significantly longer among patients undergoing ultrasonography or CT imaging compared with those who did not require advanced imaging (p&lt;0.001 for both). Consultation burden demonstrated a progressive association with prolonged LOS-ED. Random forest analysis identified consultation burden, advanced imaging utilization, continuity of care, ED workload, and time to first physician assessment as the operational variables most strongly associated with prolonged LOS-ED. <bold>Conclusions</bold> LOS-ED is associated with multiple interconnected operational factors extending beyond isolated time-based quality metrics. Consultation burden, imaging utilization, ED workload, and continuity of care were among the variables most strongly associated with prolonged LOS-ED. Improving ED performance may therefore require broader systems-based approaches targeting operational bottlenecks and workflow processes. </p>