Reliability and diagnostic accuracy of home video recording in differentiating sleep-related hypermotor epilepsy from disorders of arousal.

Abstract

Objective: This study evaluates the interobserver reliability (IR) and diagnostic accuracy of home video in distinguishing sleep-related hypermotor epilepsy (SHE) from disorders of arousal (DoA) among neurologists with varying levels of expertise. Additionally, it assesses the learning effect in nonexpert raters and extrapolates key diagnostic elements and confounding factors.

Methods: We selected 24 home videos capturing sleep-related events recognized as typical from 12 patients with confirmed SHE and 12 with DoA. During a full-day video session, a panel of 18 experts in epileptology or in sleep medicine and 12 general neurologists and residents were asked to classify each video as "SHE," "DoA," or "unknown," based on the semiology of the event. Baseline IR among all raters and among neurologists before and after experts' discussion was calculated using kappa statistics. The learning effect for nonexperts was assessed using McNemar test by comparing performance before and after training. A thematic analysis of raters' comments was conducted to extrapolate discriminatory features for diagnosis and to identify confounding factors.

Results: The overall agreement was 75%, with a "moderate" IR (kappa = .56). Raw agreement among experts was 79%, with "substantial" IR (kappa = .62). Among general neurologists, raw agreement improved from 70% to 82% after training, corresponding to "moderate" IR (kappa = .47) and "substantial" IR (kappa = .65), respectively. Disagreement was higher for DoA cases. The thematic analysis identified abrupt onset and brief duration of episodes, ambulatory behavior, and irregular breathing as the main confounders.

Significance: Home video is a reliable tool for distinguishing between SHE and DoA. IR is substantial when assessed by experts and among nonexperts after appropriate educational training. The use of home video as a diagnostic tool is feasible for clinicians with limited access to specialized diagnostic facilities, as well as for sleep and epilepsy specialists, aiding in the diagnostic process and in the follow-up assessment.