Background: Dispatch algorithms used by emergency 000 call-takers to recognise stroke have limited diagnostic accuracy, but it is currently unknown whether recognition may be better in patients requiring thrombectomy due to more severe deficits.
Aims: We aimed to evaluate the accuracy of 000 call-taker dispatch codes for recognising thrombectomy patients as having a stroke.
Methods: We linked 000 call-taker dispatch codes to metropolitan and rural patients receiving thrombectomy (direct presenting and secondary transfer) at The Royal Melbourne Hospital from 2007-2021. The primary outcome was the proportion of cases dispatched as stroke versus non-stroke, with secondary analyses of differences in time to thrombectomy and recognition accuracy according to baseline clinical severity. Chi-square and Mann-Whitney tests were used where appropriate.
Results: A total of 618 patients were successfully linked, of whom 382 (61.8%) were recognised by the 000 call-taker as having a stroke. This rose to 552 (89.3%) post paramedic arrival and assessment. Of the non-stroke dispatches, the most common were “Unconsciousness/Fainting” (117; 49.6%) and “Falls” (40; 16.9%). Patients with higher baseline severity (NIHSS ≥ 10) were less likely to receive stroke dispatch than those with a lower severity (59% vs 76%, p<0.001). No statistically significant time differences were found between stroke and non-stroke dispatches for time from dispatch to thrombectomy (median 208 vs. 216 minutes, p=0.593) or time from hospital arrival to thrombectomy (median 42 vs. 42 minutes p=0.851).
Conclusion: Nearly 40% of thrombectomy patients were not recognised as having a stroke by the 000 call-taker, with 10% still unrecognised after paramedic assessment. Patients with higher baseline severity were less likely to receive a stroke dispatch despite theoretically having more obvious symptoms. Time to thrombectomy; however, was not significantly longer for non-stroke dispatches.