Speech Audiometry at Home: Automated Listening Tests via Smart Speakers With Normal-Hearing and Hearing-Impaired Listeners.
Jasper Ooster, Melanie Krueger, Jörg-Hendrik Bach and 3 others
PMID 33272109WHAT IT FOUND
An Amazon Echo skill accurately detected hearing loss in most users, with an AUC of 0.95 for clinical thresholds.
However, it overestimated hearing thresholds by up to 2.2 dB in normal-hearing listeners and showed higher error rates in older adults.
Key findings
01The smart speaker application showed high classification performance for detecting a deviating clinical speech reception threshold, with an AUC of 0.95.
02There was a significant bias between the smart speaker and clinical reference measurements, averaging 1.40 dB, with the smart speaker yielding higher (worse) SRT values.
03Automatic speech recognition errors were significantly higher in elderly listeners compared to young normal-hearing listeners, which contributed to reduced measurement reliability.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study was conducted in a laboratory simulator, not actual homes, so real-world background noise and room acoustics were not fully captured. A human supervisor was present during the measurements, which may have influenced participant behavior or provided unintended cues. Participants were naive users of smart speakers; experienced users might interact differently with the device, potentially altering error rates. The decision thresholds derived are valid only for trained listeners, as untrained users showed different error patterns. The bias between smart speaker and clinical results varied significantly across hearing loss groups, complicating direct comparison.
Declared interests
The authors declare no conflicts of interest. The research was supported by non-U.S. government funding.
The easy way to misread this
Do not use the smart speaker SRT values as a direct substitute for clinical diagnostic thresholds. The test systematically overestimates hearing thresholds (shows worse hearing) by up to 2.2 dB in normal-hearing listeners and has higher variability in older adults, so it should be used for screening and referral rather than definitive diagnosis.