The bedside head impulse test (bHIT) is used to differentiate vestibular neuritis (VN) from posterior circulation stroke (PCS) in patients presenting with acute vestibular syndrome (AVS). If assessed by neuro-otological experts, diagnostic accuracy is high. We report on its diagnostic accuracy when applied by nonexperts during routine clinical practice in the emergency department (ED), its impact on patient management, and the potential diagnostic yield of the video-oculography-supported head impulse test (vHIT). Medical chart review of 38 AVS patients presenting to our university medical center's ED, assessed by neurology residents. We collected bHIT results (abnormal/peripheral or normal/central) and whether patients were admitted to the stroke unit or general neurological ward. Final diagnosis (VN, n=24; PCS, n=14) was determined by clinical course, magnetic resonance imaging, and vHIT. The bHIT's accuracy was only 58%. Its sensitivity for VN was high (88%), but due to many false-abnormal bHITs in PCS (36%), the specificity was low (64%). The vHIT yielded excellent specificity (100%) and moderate sensitivity (67%). The decision on the patient's further care was almost arbitrary and independent from the bHIT: 58% of VN and 57% of PCS patients were admitted to the stroke unit. The bHIT, applied by nonexperts during routine practice in the ED, has low accuracy, is too often mistaken as abnormal/peripheral, and is not consistently used for patients' in-hospital triage. As false-abnormal bHITs can lead to misdiagnosis/mistreatment of stroke patients, we recommend that bHIT applied by nonexperts should be reassessed by a neuro-otological expert or preferably quantitative vHIT in the ED.