Reliability of prehospital diagnosis of acute cerebrovascular accident.

Abstract

Early and correct suspicion of an acute cerebrovascular accident (CVA) is necessary to minimise the time to reperfusion treatment. Our aim was to evaluate the reliability of a prehospital diagnosis of stroke or transient ischaemic attack made by healthcare professionals referring patients directly to a neurological Emergency Department (ED). This retrospective analysis included all consecutive patients referred between 1 January and 31 December 2014 by ambulance physicians, paramedics or outpatient physicians to the neurological ED providing care for the 300-350,000 inhabitants of a highly urbanised area. We calculated sensitivity and positive predictive value (PPV) with 95% confidence intervals (95% CI) for each group of healthcare professionals, and compared the proportions of undetected CVAs. Of 802 patients referred with a prehospital diagnosis of CVA, 544 were confirmed. Additional 95 CVA cases were referred with a diagnosis other than CVA. The highest sensitivity for detection of any CVA was among ambulance physicians (96%; 95% CI 92-98%), followed by paramedics (85%; 95% CI 80-90%; p < 0.001) and outpatient physicians (74%; 95% CI 70-79%; p < 0.001). PPV for stroke was 83% (95% CI 77-87%) among ambulance physicians, 73% (95% CI 65-80%) among paramedics, and 56% (95% CI 47-64%) among outpatient physicians. Ambulance physicians are highly sensitive in diagnosing any CVA, and are correct in 8 out of 10 cases. The inferior performance of paramedics and outpatient physicians indicates the need for regular stroke training for paramedics and the implementation of two-way communication with the stroke team to identify potential candidates for reperfusion therapy before arrival at the ED.