Automated and machine learning approaches in diagnostic hearing assessment: a scoping review

Abstract

Hearing loss affects one in five people worldwide and is estimated to affect one in four by 2050. Treatment relies on accurate diagnosis of hearing loss, but this first step is out of reach for more than 80% of those affected. Increasingly automated digital approaches are being developed for self-administered hearing assessment without professionals’ direct involvement. This scoping review provides an overview of automated approaches, based on 56 reports from 2012 until June 2021, adding to the 29 published prior to 2012. Twenty-seven automated approaches were identified. An increasing number of approaches report similar accuracy as manual hearing assessments. Machine learning approaches are more efficient and personal digital devices make assessments more affordable and accessible. Validity can be enhanced by quality surveillance including noise monitoring and detecting inconclusive results. Employed within identified limitations, automated hearing assessments can support task-shifting, self-care, and clinical care pathways.

Using Machine Learning and the National Health and Nutrition Examination Survey to Classify Individuals With Hearing Loss

“…One such test, the Home Hearing Test, has been shown to produce reliable results in the home ( 4 , 5 ). For a more thorough review of automated and in-home audiometric testing, please see Pragt et al ( 6 ).…”