Competence of final year otolaryngology residents with the bedside head impulse test

BACKGROUND: The bedside head impulse test (bHIT) is a clinical method of assessing the vestibulo-ocular reflex (VOR). It is a critical component of the bedside assessment of dizzy patients, and can help differentiate acute stroke from vestibular neuritis. However, there is evidence showing the bHIT is often not performed in appropriate clinical settings or is performed poorly. To date, there have been no studies evaluating the bHIT competence of graduating physicians. METHODS: 23 final year Otolaryngology –Head &Neck Surgery (OTL-HNS) residents in Canada were evaluated on the use of bHIT using a written multiple-choice examination, interpretation of bHIT videos, and performance of a bHIT. Ratings of subject bHIT performance were completed by two expert examiners (DT, DL) using the previously published Ottawa Clinic Assessment Tool (OCAT). RESULTS: Using a cut-off of an OCAT score of 4 or greater, only 22% (rater DT) and 39% (rater DL) of residents were found able to perform the bHIT independently. Inter-rater reliability was fair (0.51, interclass correlation). The mean scores were 65% (14.1% standard deviation) on the video interpretation and 71% (20.2% standard deviation) on the multiple-choice questions. The scores on multiple choice examination did not correlate with bHIT ratings (Pearson r = 0.07) but there was fair correlation between video interpretation and bHIT ratings (Pearson r = 0.45). CONCLUSION: Final year OTL-HNS residents in Canada are not adequately trained in performing the bHIT, though low interrater reliability may limit the evaluation of this bedside skill. Multiple choice examinations do not reflect bHIT skill. These findings have implications for development of competency-based curricula and evaluations in Canada in critical physical exam skills. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s40463-019-0326-y) contains supplementary material, which is available to authorized users.