Remote self-report and speech-in-noise measures predict clinical audiometric thresholds

Abstract

The COVID-19 pandemic highlighted the need for remote, but reliable hearing tests. Previous studies used remote testing but did not directly compare results in the same listeners with standard lab testing. Digits-in-noise (DIN) is a reliable speech-in-noise test that can be self-administered remotely. This study investigated the predictive validity of a self-administered DIN test and a commonly used self-report, the speech, spatial, and qualities of hearing (SSQ-12), for lab-based, supervised DIN and audiometry. Speech reception thresholds (SRTs) of 34 adults (18-64 y/o), 16 normal-hearing (NH) and 18 hearing-impaired (HI), were measured at home (remote-DIN) and in the lab (lab-DIN). All DIN testing used English digits 0-9, binaurally presented as triplets in different speech-shaped noise maskers (broadband, low-pass filtered at 2, 4, 8 kHz). Audiometry was administered during lab testing. An SSQ-12 e-version was completed by participants at home. As expected, NH listeners had significantly higher SSQ scores, and remote- and lab-DIN SRTs than HI listeners. All test versions of DIN were significantly correlated with pure-tone-average (PTA), with the 2-kHz filtered test the best predictor, explaining 50% of variance in PTA. SSQ also significantly predicted PTA. Overall, DIN-SRTs were better predictors of audiograms than the SSQ. Remote-DIN correlated significantly with lab-DIN, and there was no significant mean difference between remote- and lab-DIN. Test-retest reliability was measured for broadband remote-DIN. High, significant intraclass correlation coefficients indicated strong internal consistency of the remote-DIN. This study shows that remote SSQ-12 and DIN are valid screening tools for capturing important aspects of auditory function.