Implementation of a nurse-initiated in-hospital stroke response system: association with early imaging and response standardization in non-specialist wards.

In-hospital stroke is associated with delayed evaluation and poorer outcomes than community-onset stroke. We implemented the Musashino In-Hospital Stroke Action Team (MISAT), a nurse-initiated system that summons stroke-specialist nurses when stroke is suspected, and evaluated its association with early imaging and response standardization in non-specialist wards. This single-center retrospective observational study analyzed 80 consecutive MISAT activations for suspected in-hospital stroke in October 2023-December 2025. A pre-implementation cohort served as reference. The primary outcome was time from symptom recognition to brain imaging. Secondary analyses assessed process time metrics and phase-based trends in non-specialist wards. Of 80 MISAT activations, 52 (65.0%) were cerebrovascular events and 28 (35.0%) were stroke mimics. The median [interquartile range] recognition-to-imaging time was significantly shorter post-implementation than in the pre-implementation reference cohort (44.0 [30.0-68.3] min vs. 72.0 [48.0-122.5] min; p = 0.002). Recognition-to-activation and activation-to-imaging times were 12.0 [5.0-28.0] min and 28.5 [20.5-38.0] min. In non-specialist wards, recognition-to-activation time improved across phases from 54.0 [21.5-88.0] min to 8.0 [3.0-22.0] min and 10.0 [5.0-17.0] min (Kruskal-Wallis p = 0.0039), with reduced variability (Fligner-Killeen p = 0.0020). Recognition-to-imaging time also decreased across phases, with reduced variability. The MISAT system facilitated earlier evaluation of suspected in-hospital stroke. In non-specialist wards, improved recognition-to-activation time and reduced variability indicate potential educational and standardization benefits. This nurse-initiated system may be a practical stroke-specific complement to existing emergency response systems.