Evaluating the Diagnostic Accuracy of the Head-Impulse Test: A Scoping Review.

IMPORTANCE

Vestibular symptoms rank among the most common complaints in medicine worldwide. Underlying disorders manifested by these symptoms are generally associated with an impairment of the vestibular-ocular reflex and can be assessed with different diagnostic procedures. In recent years, an increasing number of diagnostic test accuracy studies comparing various head-impulse test (HIT) methods with other diagnostic procedures have been published but not systematically reviewed.

OBJECTIVE

To conduct a scoping review and describe key characteristics of the growing number of diagnostic studies in patients presenting with vestibular symptoms.

EVIDENCE REVIEW: In April 2017, published studies were identified through searches of 4 bibliographic databases: Medline, Science Citation Index Expanded, the Cochrane Library, and ScienceDirect. Studies were included if they provided diagnostic accuracy data (sensitivity and specificity) for any HIT method with reference to any other vestibular test or clinical diagnosis in patients with vestibular symptoms. Study key characteristics were extracted, and the current literature was described narratively. All analysis took place between June 2017 and July 2018.

FINDINGS: We identified a total of 27 diagnostic studies (including 3821 participants). There were disagreements between diagnostic test accuracy data both within and between studies when different HIT methods were compared with other diagnostic procedures. The proportion of correctly identified people having the disease (sensitivity) ranged between 0% and 100% (median, 41%), whereas the proportion of correctly identified people without the disease (specificity) was higher and ranged between 56% and 100% (median, 94%).

CONCLUSIONS AND RELEVANCE: Based on the studies included in this review, sensitivity, specificity, and, more importantly, the risk of misdiagnosis and associated undertreatment or overtreatment cannot be reliably estimated by HIT methods for patients with vestibular symptoms. We recommend that further diagnostic studies consider (1) multiple possible underlying causes of vestibular symptoms and multiple test thresholds, (2) a representative sample of patients with and without the disease, and (3) reporting guidelines for diagnostic test accuracy studies.