The accuracy of Johannesburg-based ambulance personnel in identifying stroke

Background. Stroke is a potentially life-threatening, time-dependent event that requires urgent management to reduce morbidity and mortality. It has been suggested that earlier recognition by ambulance personnel and transport to stroke centres may significantly reduce treatment delays. For this reason it is vitally important that ambulance personnel are able to accurately diagnose stroke. Methods. A series of vignettes were created that included images, video and audio displaying either signs or symptoms of stroke or those of another condition. Ambulance personnel were asked to review each vignette and state whether the patient described was suffering from a stroke or not. Further investigation was sought by requesting each individual to motivate their answer, mentioning upon what their diagnosis was based.  Results. A total of 40 basic life support (BLS) and intermediate life support (ILS) personnel from different sites diagnosed 280 vignettes. BLS personnel were able to diagnose stroke with a sensitivity of 85.3% and a specificity of 89.9% (positive predictive value (PPV) 86.7%, negative predictive value (NPV) 88.8%), while ILS achieved a sensitivity of 98.2% and specificity of 94.0% (PPV 91.7%, NPV 98.8%). The combined sensitivity and specificity were 91.5% and 92.0%, respectively (PPV 89.2%, NPV 93.8%). In order to aid their diagnosis, only 5% of BLS and 18.34% of ILS utilised validated stroke screening tools. Conclusion. Despite not using validated screening tools, the ambulance personnel sampled in this study were able to identify stroke with high accuracy. Further studies should be considered to identify how these diagnoses were reached in order to identify training needs.