009 The ambulance clinical triage-for acute stroke treatment (ACT-FAST) algorithmic pre-hospital triage tool for endovascular thrombectomy: ongoing paramedic validation. Issue 6 (24th May 2018)

• Record Type: Journal Article
• Authors: Zhao, Henry; Pesavento, Lauren; Rodrigues, Edrich; Salvaris, Patrick; Smith, Karen; Bernard, Stephen; Stephenson, Michael; Parsons, Mark; Yassi, Nawaf; Davis, Stephen M; Campbell, Bruce CV
• Is Part Of: Journal of neurology, neurosurgery and psychiatry. Volume 89:Issue 6(2018)
• Page Start: A5
• Page End: A5
• Publication Date: 2018-05-24
• Subjects: Neurology -- Periodicals; Nervous system -- Surgery -- Periodicals; Psychiatry -- Periodicals; 616.8
• DOI: 10.1136/jnnp-2018-ANZAN.9
• Languages: English
• ISSNs: 0022-3050

Abstract

Abstract : Introduction: The ambulance clinical triage-for acute stroke treatment (ACT-FAST) algorithm is a severity based 3-step paramedic triage tool for pre-hospital recognition of large vessel occlusion (LVO), designed to improve specificity and paramedic assessment reliability compared to existing triage scales. ACT-FAST sequentially assesses 1. Unilateral arm fall to stretcher <10 s; 2a. Severe language disturbance (right arm weak), or 2b. Severe gaze deviation/hemi-neglect assessed by shoulder tap (left arm weak); 3. Clinical eligibility questions. We present the results of the ongoing Ambulance Victoria paramedic validation study. Methods: Ambulance Victoria paramedics assessed ACT-FAST in all suspected stroke patients pre-hospital in metropolitan Melbourne, Australia, and in the Royal Melbourne Hospital Emergency Department since July 2017. Algorithm results were validated against a comparator of ICA/M1 occlusion on CT-angiography with NIHSS ≥6 (Class 1 indications for endovascular thrombectomy). Results: Data were available from n=119 assessments (ED n=68, pre-hospital n=51). Patient diagnoses were LVO n=20 (15.6%), non-LVO infarcts n=45 (38.5%), ICH n=10 (8.3%) and no stroke on imaging n=44 (37.6%). ACT-FAST showed 85% sensitivity, 88.9% specificity, 60.7% (72% excluding ICH) positive predictive value and 96.7% negative predictive value for LVO. Of 10 false-positives, 4 received thrombectomy for non-Class 1 indications (basilar/M2 occlusions/cervical dissection), 3 were ICH, and 1 was tumour. Three false-negatives were LVO with milder syndromes. Discussion: The ongoing ACT-FAST algorithm validation study shows high accuracy for clinical recognition of LVO. The streamlined algorithmic approach with just two examination items provides a more practical option for implementation in large emergency service networks. Accurate pre-hospital recognition of LVO will allow bypass to endovascular centres and early activation of neuro-intervention services to expedite endovascular thrombectomy.