Abstract 33: Diagnostic Accuracy, Reperfusion Delay and Triage Scale Comparison: Real-World Paramedic Validation Results From the ACT-FAST Triage Algorithm for Thrombectomy
Background: Severity-based selection tools for large vessel occlusion (LVO) are limited by lack of validation in unselected prehospital stroke patients and concerns regarding delayed thrombolysis (tPA) and comprehensive stroke center (CSC) burdening. We examined these issues in a real-world validation of the two exam step (severe arm motor + speech or neglect) ACT-FAST LVO triage algorithm. Methods: The ACT-FAST statewide validation involved 15 metro and 17 rural hospitals in Victoria, Australia from Nov 2017-July 2019 with training of paramedics using an 8 min video. Prehospital paramedic assessments were correlated with hospital imaging to determine presence of LVO. Data were then examined for diagnostic accuracy, time saving for direct bypass to CSC using a validated Google maps model, rates and magnitude of delayed tPA in false-positive non-LVO infarcts, and extra CSC workload. Results: In 517 completed assessments, 58% involving non-EVT centers and including 114 (22%) LVO, ACT-FAST sensitivity was 81% (92/114) and specificity was 81% (325/403; 89% if ICH are not regarded as false-positive) for LVO. Figure compares to other LVO scales. Bypass to CSC was modelled to save median 71 min for analysis of 29 thrombectomy patients requiring inter-hospital transfer. Of 27 non-LVO infarcts with false positive ACT-FAST, only 4 (15%) received tPA at a non-CSC center, and bypass would have only added median 10 mins in these cases. The increase in CSC presentation using ACT-FAST triage was estimated to be 2-3.3 patients/week using estimated 7,200 suspected stroke cases/year across entire metro Victoria. Conclusion: In comprehensive real-world validation, the simple ACT-FAST algorithm detected LVO or ICH in almost 80% of positive assessments with highly favorable comparison to other scales. Prehospital bypass to CSC substantially reduces thrombectomy delay, and appears to strongly outweigh negatives of bypassing false positive cases on tPA delay and CSC overburdening.