Test-retest reliability of remote home-based audiometry in differing ambient noise conditions

Home-based remote audiometry has been emerging due to the increasing accessibility of mobile technology and the need for healthcare solutions that are available worldwide. However, the challenges presented by uncontrolled conditions, such as noisy environments, could compromise the reliability of the hearing assessment.In this study, we evaluate the reliability of the Jacoti Hearing Center (JHC) smartphone application in differing ambient noise environments. Hearing test data were synchronized from the JHC application to the Jacoti earCloud database (JEC). We collected, de-identified, and analyzed real-world, home-based audiometric test data spanning from 2015 to 2023, extracted from the JEC database. A set of exclusion criteria was defined to perform data cleaning, ensuring the removal of incomplete and unreliable data, as well as, data from users who had completed a large number of tests. The final dataset comprised 9,421 test and retest threshold pairs from 1,115 users. Hearing tests conducted under relatively quiet and noisy conditions were categorized based on threshold-to-noise ratio.The test-retest reliability of the JHC application demonstrated an average absolute difference of 4.7 dB within the range from 20 to 75 dB, ranging from 3.7 dB to 6.2 dB across frequencies. A strong positive correlation of 0.85 was found between test and retest thresholds. Moreover, the pure tone average differences were within 5 dB for 84.6% of the audiograms. No clinically significant effects of ambient noise were observed on thresholds determined between 20 and 75 dB HL.Our results demonstrate that the JHC application can provide reliable audiometric data for hearing loss, even in non-ideal acoustic conditions. This highlights the potential of home-based audiometric assessment, reinforcing the idea that, with continuous noise monitoring and noise-aware control of the testing procedure, remote audiometry can be reliable.