Hearing and gait: an association that did not survive its own adjustment

Design

Cross-sectional study with prespecified composite outcomes and Holm multiplicity correction

Population

172 cognitively unimpaired adults aged 50 to 80 (119 women), mean age 63.9 years

Primary outcome

Three composite gait measures from the instrumented Timed Up and Go against better-ear pure-tone average

Effect

Unadjusted Holm P at or below 0.015 for all three; age and sex adjusted, global duration Holm P = 0.050, turning 0.064, walking 0.131

One hundred and seventy-two cognitively unimpaired adults aged 50 to 80 from the DataCOG cohort had pure-tone audiometry and an instrumented Timed Up and Go test, with three prespecified composite gait outcomes - global duration, turning, and walking duration - and a Holm correction for multiple comparisons set in advance.

Unadjusted, worse better-ear hearing was associated with poorer performance on all three composites (Holm-adjusted P at or below 0.015). Adjusting for age and sex removed most of it: global duration sat exactly on the boundary at 0.050, turning at 0.064 and walking duration at 0.131 no longer met the threshold. Sex-stratified estimates were larger in the 53 men than in the 119 women, but neither group crossed the prespecified threshold (men 0.054, women 0.459 or higher).

The authors call this hypothesis-generating and they are right. The interest is not the finding but the method: prespecified composites, a stated multiplicity criterion, and a willingness to report that the effect dissolved. Read against the wider literature linking hearing loss to falls and mobility, this is a reminder that much of that signal may be age doing the work in both directions. Nothing here should change how you counsel a patient about hearing aids and falls - but nor does it license the confident claim that treating hearing loss improves mobility.

• Do not quote a hearing-gait association as established; this study's own adjustment removed it.
• Note that age confounds both hearing and gait, which is why unadjusted associations look strong.
• The sex difference is a subgroup impression, not a finding; neither sex met the prespecified threshold.
• Continue to assess falls risk in older patients with hearing loss on its own merits.
• Watch for the same pattern in other hearing-and-ageing papers that report unadjusted results only.

The statistics, in plain English

A Holm-adjusted P of exactly 0.050 is a boundary value and should be read as no clear signal rather than a positive result. The prespecified multiplicity criterion is what makes this trustworthy: with three outcomes and several models, at least one nominally significant result would be expected by chance, and correcting for that is the step most papers skip. With 172 participants and only 53 men, the sex-stratified analysis is underpowered by design and cannot support either a difference or its absence.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.