A rule-based mobile screening classifier for six ICVD-defined vestibular disorders: a pilot study
摘要
Vestibular disorders are common conditions requiring specialized evaluation. This pilot study aimed to assess the feasibility of a rule-based mobile screening classifier for six ICVD-defined vestibular disorders and to compare its top-two candidate outputs with clinical diagnoses made by an experienced otologist. In this prospective, single-center pilot evaluation, 86 participants (65 females, mean age 59.9 ± 15.7 years) presenting with vestibular symptoms at a specialized clinic completed a 47-item questionnaire using a tablet device and underwent a comprehensive clinical evaluation including history taking and otologic and neurologic examination by an experienced otologist. The performance of the mobile classifier was measured by comparing its top two candidate outputs against the expert reference diagnoses. The mobile classifier demonstrated an overall classification accuracy of 82.1%, with 51.2% of cases completely correct and 30.9% partially correct. For vascular vertigo specifically, the screening branch showed an accuracy of 82.7%. Participants completed the questionnaire in an average time of 7.24 minutes. These pilot findings support the feasibility of the rule-based mobile classifier as an initial screening aid for prevalent vestibular disorders. External multi-center validation in independent cohorts will be required before the tool can be regarded as a clinically validated screening instrument.