Hearing Loss and Dementia: What the Research Actually Says

A family sits down to Sunday dinner with eight people at the table. Their father answers a question nobody asked, laughs a beat after everyone else, then goes quiet for the rest of the meal. On the drive home somebody says the word they have been avoiding for a year.

He may well be fine. That table is one of the hardest listening environments a human ear encounters: overlapping voices, hard surfaces, no visual cue about who is speaking next. A person with moderate hearing loss can look exactly like a person with early dementia in that room and be neither.

Which is the first reason this topic matters. The second is that the research connecting hearing loss to dementia is real, widely reported, and widely reported wrong.

What the association actually is

Hearing loss shows up on the major list. The Lancet standing Commission on dementia prevention, in its 2024 report, named 14 modifiable risk factors that together account for roughly 45 percent of dementia cases worldwide. Hearing loss is one of them, carried forward from earlier reports, alongside things like hypertension, smoking, physical inactivity, diabetes, social isolation and depression. The 2024 update added high LDL cholesterol in midlife and uncorrected vision loss.

That is a serious finding from serious people. It is also a statement about risk across populations, not a promise to any individual, and "modifiable risk factor" is not the same claim as "treating this prevents dementia."

The difference between those two sentences is where almost all the confusion lives.

The trial everyone cites, and what it found

The study behind most of the headlines is ACHIEVE, published in The Lancet in 2023. It is worth knowing what it actually did, because the version that reached the public is not quite the version that happened.

ACHIEVE randomized 977 adults aged 70 to 84, all with untreated hearing loss and none with substantial cognitive impairment, to either a hearing intervention or a health education control, and followed cognition for three years. Those 977 came from two different groups: 238 participants drawn from the long-running Atherosclerosis Risk in Communities study, who were at higher risk of cognitive decline, and 739 healthy volunteers newly recruited from the community.

The primary result, across the whole study population, was null. No significant difference in cognitive change between the group that got hearing treatment and the group that did not.

In the higher-risk ARIC subgroup, hearing intervention reduced three-year cognitive change by 48 percent.

Both of those are true, and the second one is what got printed.

One point in fairness to the researchers, since most of this section is about how their work got mangled. The difference between the two cohorts came out of a sensitivity analysis they had specified in advance, not one they went looking for after the primary result disappointed. That is the difference between a real signal worth following and a fishing trip, and it is why the trial's own conclusion is that hearing treatment might help people already at higher risk.

Why the popular version is wrong twice

You will see the claim stated as "hearing aids cut dementia risk by nearly half." That sentence misreports the finding in two separate ways.

Wrong population. The 48 percent came from a subgroup of 238 people already at elevated risk, not from the trial as a whole. In the healthy community volunteers who made up three quarters of the study, the intervention did not produce a significant effect over three years.

Wrong outcome. ACHIEVE measured the rate of cognitive change over three years. It did not count dementia diagnoses. Slowing measured cognitive decline in a higher-risk group is a genuinely encouraging result. It is not the same finding as preventing dementia, and three years is a short window for a condition that develops over decades.

We are being this careful for a practical reason, not an academic one. A family that buys hearing aids believing they are buying dementia prevention has been set up for disappointment, and disappointed people put the devices in a drawer. The honest case for treating hearing loss is strong enough that it does not need the overclaim.

The honest case

Strip the dementia headline out entirely and the argument still stands.

A person who cannot follow conversation withdraws from it. They stop going to the things where conversation is the point, which is most things. The isolation that follows is itself on the Lancet Commission's list of risk factors, and long before it is a dementia question it is a quality of life question, a safety question, and a question of whether someone is still a participant in their own family.

There is also the diagnostic problem we opened with. Untreated hearing loss can imitate cognitive impairment convincingly enough that families, and sometimes clinicians, misread it. A hearing evaluation is a cheap way to remove one large confound before anyone concludes anything about memory.

And the fair reading of ACHIEVE is that for people already at higher risk of cognitive decline, treating hearing loss may slow that decline. That is a reasonable thing to act on. It is just not the same sentence as the headline.

How common this is

The National Institute on Deafness and Other Communication Disorders puts numbers on it. Twenty two percent of people aged 65 to 74 have disabling hearing loss, and among those 75 and older it is 55 percent. Disabling is the operative word there: the threshold is a loss of 35 decibels or more in the better ear, roughly the level at which someone would generally benefit from hearing aids. Milder loss, the kind that ruins a dinner table without ever registering as a diagnosis, is more common still.

So for most families reading this, the question is not whether a parent will face this. It is whether it gets addressed or quietly absorbed.

The money, which is where this usually stalls

Here is the part that surprises families, every time.

Original Medicare does not cover hearing aids. Medicare's own coverage page puts it in a single line: "Original Medicare doesn't cover hearing aids or exams for fitting hearing aids." Both halves of that sentence cost money, and the exam is the half families do not see coming.

What Part B does cover is diagnostic hearing and balance exams, when a doctor or other provider orders them to determine whether you need medical treatment. After the Part B deductible you pay 20 percent of the approved amount, plus a copayment in a hospital outpatient setting.

There is one provision worth knowing that almost nobody does. You can see an audiologist once every 12 months without an order from a provider for non-acute hearing conditions, meaning hearing loss that has developed over many years. That covers most age-related hearing loss. A parent does not need to get a referral first to have their hearing properly evaluated once a year.

Some Medicare Advantage plans include hearing benefits that Original Medicare does not. Whether a given plan does, and what it actually pays, varies enough that it has to be checked plan by plan.

Since October 17, 2022, there is also a fourth route. The FDA established a category of over-the-counter hearing aids for adults 18 and older with perceived mild to moderate hearing loss, sold in stores and online without a prescription, an exam, or an audiologist. They are explicitly not intended for severe or profound hearing loss. For a parent who has refused to see anyone about it for a decade, an OTC device is sometimes the thing that gets tried first, and it beats nothing.

The part nobody plans for

We will say the unglamorous thing, because it decides whether any of the above matters.

Hearing aids only work when they are worn, and worn devices need small daily competence. Batteries or charging. Insertion. Cleaning. Noticing when one has stopped working rather than concluding that it is not helping. Not losing something the size of a bean.

For a person who also has arthritis, low vision, or memory trouble, that maintenance is not automatic, and a device that lives in a nightstand drawer is not treatment. This is worth thinking about at the point of purchase rather than six months later: who is going to help with the upkeep, and does the device chosen suit the hands and eyes of the person using it. Rechargeable models exist partly for this reason.

For families in Troy weighing what a parent needs day to day, hearing sits alongside vision as one of the things that quietly determines whether someone participates or drifts. We wrote separately about the vision side of that, in what low vision actually costs a person's independence.

The short version

Hearing loss is on the Lancet Commission's list of 14 modifiable dementia risk factors, and those 14 together account for about 45 percent of cases worldwide. That is real.

The famous 48 percent figure is not what most people think. It comes from a higher-risk subgroup of 238 people inside a 977-person trial whose overall primary result was null, and it measured the rate of cognitive change over three years rather than dementia diagnoses.

Treat the hearing anyway. Untreated hearing loss imitates cognitive decline, drives the isolation that is itself a risk factor, and takes people out of their own families. Original Medicare pays none of the cost of the aids, but it covers diagnostic exams when ordered, and a yearly audiologist visit needs no referral for ordinary age-related hearing loss.

And decide who is helping with the batteries before you buy anything.