Quantifying a Learning Curve for Video Head Impulse Test: Pitfalls and Pearls.

The video head impulse test (vHIT) is nowadays a fast and objective method to measure vestibular function. However, its usability is controversial and often considered as a test performed by experts only. We sought to study the learning curve of novices and to document all possible mistakes and pitfalls in the process of learning. In a prospective cohort observational study, we included 10 novices. We tested their ability to perform correctly horizontal head impulses recorded with vHIT. We assessed vHITs in 10 sessions with 20 impulses per session giving a video instruction after the first session (S1) and individual feedback from an expert for session 2 (S2) up to session 10 (S10). We compared VOR gain, the HIT acceptance rate by the device algorithm, mean head velocity, acceleration, excursion, and overshoot between sessions. A satisfying number of accepted HITs (80%) was reached after an experience of 160 vHITs. Mean head velocity between sessions was always in accepted limits. Head acceleration was too low at the beginning (S1) but improved significantly after the video instruction ( = 0.001). Mean head excursion and overshoot showed a significant improvement after 200 head impulses ( < 0.001 each). We showed that novices can learn to perform head impulses invHIT very fast provided that they receive instructions and feedback from an experienced examiner. Video instructions alone were not sufficient. The most common pitfall was a low head acceleration.