ABSTRACT NUMBER: ESOC2026A1435 PRIME-IVT: FEASIBILITY AND EARLY EFFECTIVENESS OF A NATIONWIDE REGISTRY-EMBEDDED STEPPED-WEDGE CLUSTER RANDOMIZED TRIAL OF AN AUDIT-AND-FEEDBACK INTERVENTION TO IMPROVE INTRAVENOUS THROMBOLYSIS USE

Abstract Background and aims Intravenous thrombolysis (IVT) use in South Korea has shown a declining trend with significant interhospital variation. We conducted a nationwide feasibility stepped-wedge cluster randomized trial of a registry-embedded audit-and-feedback intervention to optimize IVT delivery. Methods Twenty hospitals in the CRCS-K-NIH registry were randomized into four clusters that sequentially transitioned to the intervention phase (April–August 2025). The intervention comprised a preintervention meeting, monthly center-level email reports, and case-level text-message feedback to clinicians regarding eligible but untreated patients. Feasibility outcomes included reporting timeliness, feedback delivery, clinician response rates, and recruitment. The primary effectiveness endpoint was IVT use among eligible patients. Results We enrolled 932 eligible patients (intervention, n = 465; control, n = 467). The program demonstrated high feasibility: all reports were delivered on schedule (0-day delay), and > 80% of individualized feedback messages were successfully delivered. The clinician response rate was 69.5%, and recruitment reached 77.6% of the planned. IVT was administered in 41.5% in the intervention phase compared to 34.7% in the control phase (absolute difference, 6.8 percentage points; matched OR 1.43; 95% CI, 0.92–2.22; P = 0.152). Centers with high clinician engagement (≥60% response rate) showed a stronger effect (1.77; 1.00–3.12; P = 0.067), suggesting a potential engagement–response gradient. Conclusions A nationwide registry-embedded audit-and-feedback program is feasible, achieving excellent reporting timeliness and feedback delivery. Although the overall increase on IVT use did not reach statistical significance in this feasibility trial, the larger effect observed in highly engaged centers supports the need for a fully powered trial with enhanced strategies to maximize clinician engagement. Conflict of interest Eung-Joon Lee: Nothing to disclose. Figure 1 - belongs to Conclusions