Large US Study Finds Hearing Loss Raises Dementia Risk, With Loneliness Explaining Only a Small Part

A prospective analysis of more than 317,000 adults in the NIH All of Us cohort found that hearing loss was linked to a 38 percent higher risk of dementia, and that loneliness accounted for only about 2.5 percent of that connection.

For years, researchers have debated why hearing loss so consistently shows up as one of the strongest modifiable risk factors for dementia. One popular theory holds that hearing loss isolates people socially, loneliness sets in, and that loneliness in turn erodes cognitive health. If that were the main pathway, tackling loneliness might matter as much as treating the hearing itself.

A new study published in BMJ Public Health puts that theory to a direct test in one of the largest and most diverse cohorts available in the United States. The answer it returns is striking: loneliness matters for cognition in its own right, but it explains only a small slice of the link between hearing loss and dementia.

Background: Why the Researchers Looked at This

Hearing loss and loneliness have each been tied to cognitive decline in earlier research, but the two tend to travel together. People who struggle to follow conversations often withdraw from social situations, and social withdrawal is a well-known ingredient in loneliness. That overlap has made it hard to say whether hearing loss harms cognition directly or mostly by way of the isolation it can create.

The distinction is not academic. If loneliness carried most of the effect, public health efforts might focus on social programs. If hearing loss acts through other routes, such as reduced stimulation of the auditory system or the extra mental effort of straining to hear, then treating the hearing itself becomes the more urgent target. Mediation analysis, the statistical approach used here, is designed to split an association into the part that flows through an intermediate factor and the part that does not.

The researchers, based at the University of Arizona, turned to the National Institutes of Health All of Us research program, a large national dataset that pairs electronic health records with participant surveys and deliberately recruits people from groups underrepresented in medical research.

How the Study Was Done

The team analyzed records from 317,108 adults followed for an average of 11.8 years. Hearing loss was identified from diagnoses documented in each participant's electronic health record, and 50,849 people, about 16 percent of the cohort, had such a diagnosis. Loneliness was measured in a subset of 130,266 participants who completed the UCLA Loneliness Scale, an eight-item questionnaire scored as an average.

The researchers used Cox proportional hazards models, a standard method for studying how exposures relate to the timing of health events, to estimate the risk of two outcomes recorded in the health records: new dementia diagnoses and new diagnoses of impaired cognition. The models adjusted for age, sex, race, education, income, smoking, and alcohol use.

They then ran a four-way decomposition mediation analysis, which partitions the total effect of hearing loss on cognition into a direct part and an indirect part flowing through loneliness, while allowing the two exposures to interact. Sensitivity checks examined racial and ethnic subgroups, accounted for death as a competing risk, and tested alternative ways of modeling hearing loss over time.

What the Researchers Found

Over the follow-up period the cohort recorded 5,663 new dementia diagnoses and 21,117 new diagnoses of impaired cognition. Participants with documented hearing loss had a 38 percent higher risk of developing dementia and a 67 percent higher risk of impaired cognition than those without, after adjustment for demographic and lifestyle factors.

Loneliness carried its own independent risk. Each one-point increase on the loneliness scale was associated with a 41 percent higher risk of dementia and a 48 percent higher risk of impaired cognition.

The central question, though, was how much of the hearing loss effect ran through loneliness. The answer was surprisingly little. The mediation analysis estimated that loneliness accounted for roughly 2.5 percent of the association between hearing loss and dementia, and about 2 percent of the association with impaired cognition. In other words, more than 97 percent of the link between hearing loss and cognitive outcomes appears to operate through pathways other than loneliness.

Subgroup analyses suggested the hearing loss and dementia association was stronger among Black and African American participants, and the association with impaired cognition was stronger among Asian and Black and African American participants. Sensitivity analyses largely confirmed the main results.

What It Means for People with Hearing Loss

The authors conclude that addressing hearing loss directly may yield broader cognitive benefits than targeting loneliness alone. Social connection still matters, and the study found loneliness to be a meaningful independent risk factor. But for people with hearing loss, the findings suggest that the hearing itself deserves attention, not just its social consequences.

That aligns with a growing body of evidence, including the widely cited ACHIEVE randomized trial, pointing toward hearing rehabilitation as a promising avenue for protecting cognition in older adults at elevated risk. This study cannot prove that treating hearing loss prevents dementia, but it strengthens the case that the connection between the ear and the brain is direct enough to warrant action rather than waiting.

If Hearing Deserves Direct Attention, Access Matters

Because this study points to hearing loss itself, rather than loneliness, as the main driver of cognitive risk, it adds weight to the argument that hearing should be treated early rather than left unaddressed. FDA-OTC hearing aids were created to lower the practical barriers that keep many adults with mild-to-moderate hearing loss from doing exactly that.

Panda Quantum is one example of this newer generation: a receiver-in-canal device with 16-channel WDRC processing and adaptive noise reduction aimed at clear speech in noisy environments, the everyday listening situations where conversation is hardest to follow. An optional companion app can run an in-ear hearing test and personalize the sound to the user's hearing profile, though the device works fully without it. It runs 20 hours per charge, with a case that carries three additional full charges for 80 hours total, and comes with a 5-year warranty and a 45-day trial that starts when the product arrives.

One caveat worth keeping in mind: OTC hearing aids are approved for mild-to-moderate hearing loss. People with severe or profound loss still benefit most from a clinical fitting with an audiologist.

Limitations of This Research

This is an observational study, so it can show association but not prove cause. Hearing loss, dementia, and impaired cognition were identified from electronic health records, which capture only diagnoses that were made and documented; undiagnosed hearing loss or early cognitive change would be missed. Loneliness was measured at a single point in a subset of participants who chose to complete the survey, which may not represent the full cohort.

The mediation estimates also depend on the assumption that all important shared causes of loneliness and cognition were measured. Unmeasured factors, such as depression or physical health, could shift the size of the mediated effect in either direction.

Where This Leaves Us

The takeaway from one of the largest analyses of its kind is refreshingly clear: hearing loss and loneliness are both worth taking seriously for long-term brain health, but they are largely separate problems. Supporting social connection helps, yet it is no substitute for addressing hearing loss itself, and that is a message both clinicians and families of older adults can act on today.

Yang S, Sbarra DA, Wu J, Klimentidis YC. Hearing loss and loneliness in the development of impaired cognition and dementia in the US All of Us cohort: prospective analyses using electronic health records. BMJ Public Health. 2026;4(3):e004524. Retrieved from PubMed. https://doi.org/10.1136/bmjph-2025-004524