Hearing intervention did not slow cognitive decline across the full ACHIEVE trial

The bottom line

ACHIEVE did not find less three-year cognitive decline in its full randomized cohort. Global cognition changed by -0.200 standard-deviation units with hearing treatment and -0.202 with health education, a difference of 0.002 (95% CI -0.077 to 0.081; p=0.96). (Primary trial)

A prespecified analysis found that effects differed between the ARIC and de-novo recruitment groups. That signal is worth following, but the overall primary result remains null and the trial did not test whether hearing treatment prevents dementia.

Why this matters

Age-related hearing loss is common and observational studies associate it with cognitive decline. A randomized trial can test whether treating hearing loss changes cognition rather than merely tracking two conditions that share age, vascular, social or other causes. It also asks a narrower question than whether hearing care improves communication or daily life.

What researchers did

Four US sites randomized adults without substantial cognitive impairment to audiological counseling plus hearing devices or to individual health-education sessions. Participants were assessed every six months for three years. The primary endpoint combined a comprehensive battery into one standardized global-cognition score.

The cohort intentionally included two populations: 238 older ARIC participants with more baseline cognitive-risk factors and 739 generally healthier volunteers recruited from local communities. The study was unmasked, although cognition was assessed using standardized procedures.

What they found

The combined primary analysis was null. A prespecified interaction by recruitment source was statistically significant (p=0.010), and the ARIC subgroup declined less with hearing treatment while the de-novo group did not. In a 2025 secondary model, participants in the top predicted-risk quartile had 61.6% slower cognitive decline with hearing treatment (95% CI 33.7%–94.1%). That model-derived result was not the trial’s primary endpoint.

How strong is the evidence?

The randomized overall comparison is the strongest result and directly argues against a universal cognitive benefit over three years. The subgroup interaction was prespecified, making it more credible than a purely post-hoc split, but ARIC and de-novo participants differed in several ways. The later risk analysis helps explain heterogeneity; it does not turn the full trial positive.

What this study does not show

ACHIEVE does not show that hearing treatment prevents mild cognitive impairment or dementia, reverses brain aging, extends healthspan, or lengthens life. It does not establish which baseline risk profile would reliably predict cognitive benefit in routine care. A null cognition result also does not mean that clinically appropriate hearing care lacks communication, participation or quality-of-life value.

Safety and conflicts

No significant adverse events were attributed to either intervention. Real-world benefit can still depend on fitting, comfort, adherence, support and access. NIH funded the trial; Sonova/Phonak supplied hearing technology in kind. Several authors disclosed consulting, advisory, honoraria, research or equipment relationships with hearing-technology companies and other organizations.

What happens next

Longer follow-up should test diagnosed cognitive impairment and dementia, not only change in a composite test score. Replication should prospectively define higher-risk groups and examine whether social engagement, reduced listening effort, vascular risk or another pathway explains any difference.