ABSTRACT
Background and Aims: As a consequence of the COVID-19 pandemic prehospital video assessment (PHVA) of possible stroke was commenced by University College London Hospital and London Ambulance Service. We compared the accuracy of this assessment tool against conventional screening with FAST. Methods: We examined the accuracy of PHVA from May 2020 to April 2021, together with final diagnoses, timings for treatment provided and ambulance crew primary assessment (FAST status) wherever this was recorded. Results: Of 1300 patients triaged, 600 (46%) were sent to HASU, 570 (44%) were AE diverts and 65 (5%) were sent to their local TIA clinic. 40 patients (3%) were returned to their LAS pathway and missing data was noted for 25 patients (2%). From 426 patients where FAST status was recorded, diagnosis of likely stroke/TIA was correct in 89% (278) of PHVA versus 55% (173) of FAST assessed cases. Correct identification of non-stroke was correct in 84% (96) of PHVA versus 65% (75) of FAST assessed cases. A positive predictive value of 93% (PHVA) versus 59% (FAST) and negative predictive value 96% (PHVA) versus 84% (FAST) is noted. 46 thrombolysis and 12 thrombectomy cases with median treatment time before and during PHVA of 35 minutes (Interquartile range (IQR)-26) vs 33 (IQR-15) and 136 (IQR-87) vs 125 (IQR-49) minutes respectively, were noted. Conclusions: PHVA has been validated further, showing superior accuracy to the existing FAST assessment and may improve hyperacute treatment times and their variance.