Hearing-aid users with normal better-ear Digit Triplet Test SRT values: were tests performed unaided, and how should these cases be handled?

Hello everyone,

I am working with UK Biobank hearing data and would appreciate clarification on how to interpret a small subgroup of participants who report hearing-aid use but show apparently normal better-ear speech reception threshold (SRT) values on the Digit Triplet Test. I would like to know whether these cases could reflect valid testing conditions, a documentation/recording issue, or something that should be handled analytically during quality control.

What I have observed
In my sample, some participants who report using hearing aids show normal-looking better-ear SRT values, although most of them still have poorer SRT in the worse ear, have reported tinnitus or hearing difficulties.

My understanding from the documentation is that the hearing test should be performed without hearing aids. However, the pattern I am seeing made me wonder:

1. Were participants explicitly instructed to remove hearing aids before the Digit Triplet Test?
2. Was this checked by staff at the assessment centre, or was it based mainly on participant compliance?
3. Is there any variable, flag, procedural note, or other way to verify whether a participant may have completed the test while wearing hearing aids?
4. Have other researchers encountered this issue in hearing-related analyses?

For participants who report hearing-aid use but have normal better-ear SRT values, would you typically:

• exclude them,
• keep them,
• or treat them as a separate subgroup / sensitivity analysis?

I would be especially grateful for replies from anyone who has worked with the UKB hearing variables and had to make a similar decision.

Some summary stats for your reference:
Better-ear SRT was generally in the normal range (mean = -6.82 dB, SD = 0.818, range = -10.00 to -6.00 dB). Worse-ear SRT was more variable and often poorer (mean = -4.16 dB, SD = 2.668, range = -8.50 to 8.00 dB).

Thanks a lot in advance.