Predictors of Hearing-Aid Outcomes.

This was an observational study with no control group and no randomisation, so it can show associations but not that any factor caused better hearing-aid outcomes. The sample was 68 participants with symmetrical sensorineural hearing loss who met strict audiometric criteria, so it may not apply to people with conductive or mixed loss, asymmetric loss, or milder hearing loss. The best aided speech-in-noise model accounted for 43% of the spread, and the best self-report models accounted for 0%, 21%, and 8%, so much of the outcome was unexplained. Hearing-aid settings were adjusted after user complaints and requests, so final amplification and compression varied across participants. All hearing aids used slow-acting compression, so the results may not apply to fast-acting compression or individually optimized compression settings. Some measures were missing for some participants, including binaural masking level difference for 54 of the 68 participants. The study did not measure initial motivation to wear hearing aids, which may affect self-reported benefit. Speech-in-noise was measured with a time-reversed two-talker masker in Spanish, so everyday listening situations and other languages may behave differently. Participants had about 10 dB more hearing loss at low frequencies than participants in a similar hearing-aid study, so the sample was not exactly typical.

The supplied text does not report funding or author conflicts. Participants received a hearing aid free of charge, and the hearing aids were Starkey devices that were precommercial at the start and later became commercial X-Series devices.

Do not read the 30% figure as proof that unaided quiet speech scores can accurately predict an individual patient's aided performance. The study was observational, and the best aided speech-in-noise model accounted for 43% of the spread in scores.