Objective: Evaluate the impact of multidisciplinary process improvements and FAST-ED score stratification on neurointerventional team activation, resource optimization and decreasing neurointerventional thrombectomy times.

Background: Timely intervention in acute ischemic stroke (AIS) is essential for improving outcomes. Activating neurointerventional resources in non-large vessel occlusion (LVO) cases can create inefficiencies. The FAST-ED score, a validated prehospital tool for predicting LVOs, offers a way to streamline team mobilization and reduce Door to Puncture times.

Methods: FAST-ED scores were added to initial stroke code pages. Viz.ai’s AI platform converted pre-hospital alerts into “pager cards.” Scores ≥5 prompted core stroke team members to expand the card to include neurointerventional radiology (NIR). If CTA ruled out LVO, the team was stood down via the same card. Countermeasures included:

• Revised Policies, EMS Stroke Code process, incorporated FAST-ED and Viz.ai-LVO pathways
• Targeted education for MICNs, telecom, ED, and neurology staff
• Mock stroke codes at two campuses
• Multi-source data collection
• FAST-ED score integration into EPIC for streamlined abstraction

Results: FAST-ED score stratification enabled targeted activation of neurointerventional teams, avoiding unnecessary mobilization in 122 true negative cases (FAST-ED≤4 AND no LVO). The collective countermeasures led to a 37.61% reduction in median door-to-puncture time (113 to 71 minutes, resulting in a statistically significant, p<0.0001. FAST-ED score performance metrics were: True Negatives (TN) = 122, True Positives (TP) = 5, False Positive (FP) =6, False Negatives (FN) = 1, yielding a sensitivity of 83.33% and specificity of 95.31%.

Conclusion: Integrating FAST-ED scoring into EMS and hospital workflows improved procedural efficiency and resource utilization. Avoiding neurointerventional activation for FAST-ED scores ≤4 reduced unnecessary team mobilization. This may have  contributed to the 37.61% decrease in door-to-puncture times. This model demonstrates a replicable framework for stroke systems to enhance operational efficiency through data-driven triage and activation protocols.