Assessing the Role of Medical Caption Technology to Support Physician-Patient Communication for Patients With Hearing Loss: Mixed Methods Pilot Study.

Only 10 participants were analysed, and the study was too small to detect differences between subgroups. The visits were scripted mock scenarios with medical students wearing masks, not real clinical conversations. Most participants already used captioning or other hearing assistive technology, so people trying captions for the first time may react differently. The sample preferred spoken English and did not include people who primarily use American Sign Language. The captioning system was not optimised for medical vocabulary, and participants worried it could misrecognise complex terms or medication names. Word error rate treats all transcription errors equally, so it may not show the risk of a wrong medication name or diagnosis. One participant's data were lost because of equipment failure.

The authors declared no conflicts of interest.

Do not read this as proof that live captions improve real clinical care. It was a small simulated pilot of usability and acceptability, and accuracy for complex medical terms was not established.