- Design
- Multicentre, open-label randomised trial (CHOICE)
- Population
- 703 adults ≥60 with MCI and moderate to severe hearing loss, Shanghai
- Primary outcome
- Conversion to dementia-level impairment (CDR ≥1) at 24 months
- Effect
- 3.1% v 4.7%, RR 0.59 (0.25–1.38); improvement to CDR 0: 15.3% v 2.4%
CHOICE, published in JAMA Neurology on 28 September, randomised 703 adults aged 60 and over in Shanghai with mild cognitive impairment (CDR 0.5) and moderate to severe hearing loss to hearing aids or to hearing-care education, open label, for 24 months.
Conversion to dementia-level impairment (CDR 1 or more) occurred in 3.1% with hearing aids and 4.7% with education (risk difference −2.1 percentage points, 95% CI −4.7 to 0.6; RR 0.59, 0.25 to 1.38). That was not significant, partly because conversion was much rarer than expected in both groups. A prespecified secondary outcome — return to normal cognition, CDR 0 — occurred in 15.3% with hearing aids against 2.4% with education. No study-related adverse events were reported.
The trial does not show that hearing aids prevent dementia, and the secondary result is vulnerable to the open-label design, since CDR ratings rely on interviews that better hearing itself makes easier. But hearing loss is common, correctable and harmful to quality of life on its own terms, and nothing here argues against treating it. Offer hearing assessment and aids on their own merits, and be honest that the effect on dementia risk remains unproven.
- Screen for hearing loss in every patient with mild cognitive impairment.
- Offer hearing aids for their own benefits, not as proven dementia prevention.
- Be cautious interpreting cognitive scores in patients who cannot hear the test well.
- Review adherence to hearing aids at follow-up; unused aids help no one.
Why it matters
Hearing loss is often cited as the leading modifiable dementia risk factor, and the first randomised test in MCI did not confirm prevention.
Don't overread it
The cognitive improvement was a secondary outcome in an open-label trial; it does not establish that hearing aids reverse MCI.
The statistics, in plain English
The relative risk of 0.59 looks like a 41% reduction, but the interval from 0.25 to 1.38 includes no effect and even harm, so the result is inconclusive. Conversion was uncommon in both groups, which left the trial underpowered. The secondary 'improvement' result is large but was not the main question and could be influenced by patients simply hearing the assessor better.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for neurology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free