Abstract Number: Esoc2026a1434 Comparison Between Manual and Automated Measurement of Stroke Thrombolysis Workflow Metrics

Abstract Background and aims Audit of thrombolysis workflow for performance monitoring often requires manual review of health records. We investigated whether automated processing of routinely collected electronic health record (EHR) data could provide reliable workflow metrics in a comprehensive stroke centre. Methods A locally developed web-based electronic proforma (StrokeApp) is used for real-time clinical documentation. Manual metrics were derived by an auditor using EHR or paper records. Automated metrics were generated from StrokeApp and routine hospital management software. We compared manual and automated caseload and Door-to-Needle Times (DNT) by month between Jan 2023 and Sep 2025 using Bland-Altman analysis. Pre-specified Limits of Agreement (LoA) were +/-2 for thrombolysis cases and +/-5 mins for DNT. Results The population included 512 patients, mean Age=70.5 (SD=13.6), male Sex=203 (53%), median admission NIHSS=10 (IQR=6-18) and 115 (23%) received tenecteplase. Thrombolysis metrics were similar between methods (Manual DNT=45mins (IQR=35.5-57.25), n=512; Automated DNT=44mins (IQR=35-59), n=509). Total case numbers over 33 months disagreed by 3/512 (0.6%). Estimated bias for monthly caseload was -0.09 (95%CI=-0.34 to 0.52). LoA was -2.38 (95%CI=-3.12 to -1.64) to 2.56 (95%CI=1.82 to 3.30). The estimated bias for monthly DNT was -0.79 minutes (95%CI=-1.48 to -0.10). LoA were -4.73 (95%CI=-3.54 to -5.91) to 3.15 (95%CI=1.96 to 4.33). Conclusions Automated metric generation using routine EHR data performs acceptably for tracking monthly thrombolysis performance. Small discrepancies in case numbers are likely attributable to default date settings but caseload and LoA were acceptable for case recognition. Monthly DNT metrics were closely matched. Conflict of interest C Brown: nothing to disclose, M Abdou: nothing to disclose, K W Muir: Advisory Boards - Boheringer Ingelheim. Consultancy - Boheringer Ingelheim, Biogen, Hyperfine, Woolsey Pharma, Takeda. Lecture Fees - Boheringer Ingelheim, IschaemaView, Brainomix. Trial Support - Boheringer Ingelheim (tenecteplase for ATTEST2). Funding from NIHR, BHF, The Stroke Association, CSO. Figure 1 - belongs to Results