Abstract Background and aims Stroke is a major cause of disability and mortality worldwide, yet data on stroke workflow efficiency and treatment delays remain limited. This study evaluated acute stroke workflow performance within a regional stroke network, benchmarked time metrics against international standards, and identified predictors of treatment delay Methods This retrospective cohort study included 1,794 patients with stroke from a regional stroke registry (2021–2025). Data collected comprised demographics, stroke type, stroke severity (NIHSS), treatment rates for intravenous thrombolysis and mechanical thrombectomy, and workflow metrics (Door-to-CT, Door-to-Needle, and Door-to-Groin times). Comparisons were made by shift (day vs. night) and presentation day (weekday vs. weekend). Time trends were projected using autoregressive integrated moving average (ARIMA) models. Machine learning models identified predictors of delayed thrombolysis (Door-to-Needle 60 minutes) Results The mean age was 56.3 ± 14.9 years, with 67% male patients. Ischemic stroke accounted for 81% of cases (median NIHSS = 6). Intravenous thrombolysis was administered in 11.4% and mechanical thrombectomy in 4.0% of patients. Median Door-to-CT time improved from 34 minutes in 2021 to 25 minutes in 2025, while Door-to-Needle time decreased from 95 to 60.5 minutes. No significant differences were observed between shifts or weekdays and weekends. ARIMA projections indicated continued improvement in workflow times through 2027. Key predictors of delayed thrombolysis included shift timing, stroke severity, and day of arrival (AUC-ROC = 0.81). Conclusions Acute stroke workflow efficiency improved substantially over time, particularly for imaging and thrombolysis. Persistent variability in Door-to-Groin processes highlights the need for further optimization of endovascular coordination and interfacility transfer pathways. Conflict of interest Najod Alsabaan. Nothing to disclose
Alsabaan et al. (2026) studied this question.