Abstract WMP47: Low Sensitivity Of Widely Used Emergency Dispatch Algorithm For Thrombectomy Patients - Implications For Mobile Stroke Units

Introduction: Widely used emergency dispatch algorithms such as the Advanced Medical Priority Dispatch System (AMPDS) have limited diagnostic accuracy for prehospital diagnosis of stroke. With advent of mobile stroke units (MSU), this inaccuracy prevents optimal dispatch to patients who may benefit. Expedited endovascular thrombectomy (EVT) is a major contributor to net benefit of MSUs. We assessed the accuracy of AMPDS for recognizing stroke in thrombectomy patients in the Australian state of Victoria. Methods: We included consecutive patients accepted for EVT (direct and secondary transfer) to The Royal Melbourne Hospital from 2007-2021 in whom linked AMPDS dispatch codes could be obtained from Ambulance Victoria. The primary outcome was the proportion of cases dispatched as stroke vs non-stroke with subgroup analyses of the effect of baseline clinical severity, metropolitan vs rural dispatch and time to thrombectomy. Chi square and Mann Whitney tests were used as appropriate. Results: A total of n=618 patients were included with baseline NIHSS 16 (IQR 10-20). Of these, only 62% (95% CI 58-66) were initially dispatched as suspected stroke, with the most common non-stroke diagnoses being “Unconscious/Fainting” (19.2%) and “Falls” (6.9%). Those with a higher baseline severity (NIHSS ≥10) were less likely to be classified as stroke than those with lower severity (59% vs 76%, p<0.001), while no difference was found between metropolitan and rural patients (p=0.066). Overall, no significant time differences were found between stroke and non-stroke dispatches for ambulance dispatch to arterial access (median 208 vs 216 min, p=0.593) or hospital arrival to arterial access (median 42 vs 42 min, p=0.851). However, only 32 patients were treated on the MSU, which commenced operation November 2017. Conclusions: Almost 40% of thrombectomy patients did not receive an initial AMPDS dispatch of suspected stroke and those with higher baseline severity were more likely to be misclassified. Although time to thrombectomy was not significantly different between stroke vs non-stroke dispatches, MSU treatment was under-represented. Our findings have implications for emergency medical services and particularly mobile stroke units which rely on accurate stroke dispatch.