Eliminating hearing loss would in theory prevent 7.0% of dementia cases worldwide, level with high LDL cholesterol (6.9%) at the top of 14 modifiable risk factors (Lancet Commission on Dementia Prevention, Intervention, and Care, 2024 report). The association scales with severity: a pooled analysis of 50 cohort studies and 1,548,754 participants put the hazard ratio for dementia at 1.35, with 16% more risk for every 10 dB of hearing lost (Ageing Research Reviews, Yu et al. 2024). The ACHIEVE randomized trial found no overall cognitive benefit from a hearing intervention across 977 older adults, but a 48% reduction in three-year cognitive change among the 238 participants at higher baseline risk (The Lancet, Lin et al. 2023). Meanwhile, 83% of the 401.4 million people who need a hearing aid do not use one (International Journal of Audiology, Orji et al. 2020). We aggregated data from the Lancet Commission, the ACHIEVE and CHOICE randomized trials, the World Health Organization, the Global Burden of Disease Study, the Alzheimer’s Association, NIDCD, and cohort studies in the JAMA journals; this roundup sits alongside our data on over-the-counter hearing aid adoption and hidden hearing loss.
TL;DR
- Hearing loss carries a weighted population attributable fraction of 7.0% for dementia, out of 45.3% for all 14 modifiable risk factors (Lancet Commission 2024 report).
- With hearing measured by audiometry, 32.0% (95% CI 11.0% to 46.5%) of incident dementia in a US cohort was attributable to hearing loss (JAMA Otolaryngology, Ishak et al. 2025).
- Adult-onset hearing loss is associated with a dementia hazard ratio of 1.35 (95% CI 1.26-1.45) across 50 cohort studies (Ageing Research Reviews, Yu et al. 2024).
- Each 10 dB of worse hearing is associated with a 16% higher dementia risk (Ageing Research Reviews, Yu et al. 2024).
- Hearing aid users had a 19% lower hazard of long-term cognitive decline, hazard ratio 0.81 (JAMA Neurology, Yeo et al. 2023).
- ACHIEVE found no overall effect on 3-year cognition (difference 0.002, p=0.96) and a 48% reduction in its higher-risk ARIC subgroup (The Lancet, Lin et al. 2023).
- The same hearing intervention cut mean falls over 3 years by 27%, from 1.98 to 1.45 (The Lancet Public Health, Goman et al. 2025).
- In the CHOICE trial, 3.09% of the hearing aid group and 4.74% of controls progressed to dementia-level impairment in 24 months, P = .23 (JAMA Neurology, Chen et al. 2026).
- 57 million people were living with dementia in 2021, with nearly 10 million new cases a year (WHO Dementia Fact Sheet 2026).
- 1.57 billion people had hearing loss in 2019, projected at 2.45 billion by 2050 (The Lancet, GBD 2019 Hearing Loss Collaborators 2021).
- 83% of the 401.4 million people who need a hearing aid do not use one (International Journal of Audiology, Orji et al. 2020).
- US health and long-term care costs for dementia are projected at $409 billion in 2026 (Alzheimer’s Association, 2026 Alzheimer’s Disease Facts and Figures).
1. The Attributable Share: How Much Dementia Traces Back to Hearing Loss
A population attributable fraction (PAF) answers a narrow question: if a risk factor vanished, what share of cases would vanish with it? The answer depends heavily on the relative risk and the prevalence assigned to the factor, and the 2024 Commission changed both. It cut the relative risk from 1.9 to 1.4 and raised the assumed prevalence of hearing loss from 31.7% to 59.0%, so the headline figure moved from 8.2% to 7.0% while the logic underneath it shifted: a less common, stronger risk factor became a very common, weaker one.
The bigger disagreement is about measurement. When hearing is tested with audiometry instead of a questionnaire, the attributable share of incident dementia in a US cohort rises to 32.0%, and self-reported hearing loss showed no association with dementia at all in the same analysis (JAMA Otolaryngology, Ishak et al. 2025). In that cohort 66.1% of participants had audiometric hearing loss while only 37.2% reported it, which is why the choice of test matters as much as the choice of model. Objective measurement is its own field; our roundup of auditory brainstem response testing covers the clinical end of it.
| Metric | Value | Source |
|---|---|---|
| Weighted PAF of dementia for hearing loss (midlife risk factor) | 7.0% | Lancet Commission on Dementia, 2024 report (Livingston et al.) |
| Inputs behind that figure | Relative risk 1.4 (95% CI 1.0-1.9), prevalence 59.0%, unweighted PAF 19.1% | Lancet Commission on Dementia, 2024 report (Livingston et al.) |
| Weighted PAF for all 14 modifiable risk factors combined | 45.3% | Lancet Commission on Dementia, 2024 report (Livingston et al.) |
| Hearing loss weighted PAF in the previous report | 8.2% (relative risk 1.9, prevalence 31.7%) | Lancet Commission on Dementia, 2020 report (Livingston et al.) |
| PAF of incident dementia from any audiometric hearing loss, ARIC-NCS (2,946 adults aged 66 to 90) | 32.0% (95% CI 11.0% to 46.5%) | JAMA Otolaryngology 2025 (Ishak et al.) |
| PAF by severity in the same cohort | Mild 16.2%; moderate or greater 16.6% | JAMA Otolaryngology 2025 (Ishak et al.) |
| PAF of prevalent dementia from moderate or greater hearing loss, NHATS 2021 (2,470 adults) | 16.9% (95% CI 4.1% to 28.7%) | Journals of Gerontology Series A 2023 (Smith et al.) |
| Same PAF split by sex | Men 40.5%; women 3.2% (95% CI -12.7% to 17.9%) | Journals of Gerontology Series A 2023 (Smith et al.) |
Context note: a PAF is a model output, not a count of patients, and every figure in this table assumes the association is causal. The 2024 report is available from The Lancet, and the cohort-level estimate from JAMA Otolaryngology. The wide confidence intervals in the US cohorts, and the gap between men and women in the NHATS analysis, show how unstable these estimates remain.
2. Relative Risk by Degree of Hearing Loss
The dose-response pattern is the strongest argument that the link is more than coincidence. Across studies that measured hearing in decibels, risk climbs with each step of loss, and every 10 dB of worse hearing is associated with a 16% higher dementia risk (Ageing Research Reviews, Yu et al. 2024).
How steep the slope looks depends on the study. The famous three-step gradient, with hazard ratios of 1.89, 3.00 and 4.94, comes from 639 people and 58 dementia cases, with only 6 participants in the severe group. Larger and newer cohorts land on a flatter curve: 1.49 for mild and 1.68 for moderate-to-severe loss over a median of 8 years in ARIC, and a hazard ratio of just 1.07 (95% CI 1.04-1.11) in a Danish registry cohort of 573,088 adults (JAMA Otolaryngology, Cantuaria et al. 2024). In this set of studies the smallest sample produced the largest effect and the largest sample the smallest, which is a reason to read the 4.94 with care.
| Metric | Value | Source |
|---|---|---|
| Incident dementia with adult-onset hearing loss (50 cohort studies, 1,548,754 participants) | HR 1.35 (95% CI 1.26-1.45) | Ageing Research Reviews 2024 (Yu et al.) |
| Dementia risk per 10 dB worsening of hearing | 16% higher (95% CI 1.07-1.27) | Ageing Research Reviews 2024 (Yu et al.) |
| Alzheimer’s disease dementia with hearing loss | HR 1.56 (95% CI 1.30-1.87) | Ageing Research Reviews 2024 (Yu et al.) |
| Commission’s own meta-analysis of six higher-quality studies (n = 666,370) | HR 1.37 (95% CI 1.00-1.87), I2 = 80% | Lancet Commission on Dementia, 2024 report (Livingston et al.) |
| Mild hearing loss, ARIC-NCS (3,602 adults, median 8 years) | HR 1.49 (95% CI 1.16, 1.92) | Frontiers in Epidemiology 2026 (Deal et al.) |
| Moderate-to-severe hearing loss, same cohort | HR 1.68 (95% CI 1.24, 2.28) | Frontiers in Epidemiology 2026 (Deal et al.) |
| Dementia prevalence by hearing level, US adults (normal, mild, moderate to severe) | 6.19%, 8.93%, 16.52% | JAMA 2023 (Huang et al., NHATS 2021) |
| Hazard ratios for mild, moderate, severe hearing loss (639 adults, 58 cases) | 1.89, 3.00, 4.94 | Archives of Neurology 2011 (Lin et al.) |
Outlier note: the 4.94 for severe hearing loss carries a 95% CI of 1.09 to 22.40. Most recent available data for that three-level audiometric gradient: Archives of Neurology, Lin et al. 2011. The pooled estimates are in Ageing Research Reviews, the newest cohort in Frontiers in Epidemiology, and the US prevalence gradient in JAMA. In the ARIC analysis the association did not differ between Black and White participants (P-interaction = 0.92).
3. Hearing Aid Use and Cognitive Decline in Observational Data
Observational studies agree on direction and disagree on size. Pooled across 8 cohort studies with 126,903 participants followed for 2 to 25 years, hearing aid users had a 19% lower hazard of long-term cognitive decline than people with uncorrected hearing loss (JAMA Neurology, Yeo et al. 2023). That is a comparison between people who chose, could afford and kept using a device and people who did not, so it cannot separate the device from the person wearing it.
Two details show how fragile the picture is. The Framingham result that drew headlines in 2025, a hazard ratio of 0.39 for hearing aid users under 70, rests on 72 hearing aid users in that age group out of 2,953 participants. And one widely cited study on the question, a UK Biobank analysis published in The Lancet Public Health in April 2023, was retracted in December 2023 and is excluded from this roundup. How much benefit the devices deliver in real listening conditions is a separate question, covered in our data on directional microphones and signal-to-noise ratio.
| Metric | Value | Source |
|---|---|---|
| Long-term cognitive decline, hearing aid users vs uncorrected hearing loss (8 studies, 126,903 participants) | HR 0.81 (95% CI 0.76-0.87) | JAMA Neurology 2023 (Yeo et al.) |
| Short-term cognitive test scores after hearing aid use (11 studies, 568 participants) | 3% improvement (ratio of means 1.03, 95% CI 1.02-1.04) | JAMA Neurology 2023 (Yeo et al.) |
| Dementia among hearing aid users, pooled from four studies | HR 0.83 (95% CI 0.77-0.90) | Lancet Commission on Dementia, 2024 report (Livingston et al.) |
| Dementia risk vs no hearing loss, Denmark: without hearing aids and with hearing aids | HR 1.20 (95% CI 1.13-1.27) and HR 1.06 (95% CI 1.01-1.10) | JAMA Otolaryngology 2024 (Cantuaria et al.) |
| Dementia prevalence with moderate to severe hearing loss: hearing aid users vs non-users | 11.46% vs 21.53% (prevalence ratio 0.68, 95% CI 0.47-1.00) | JAMA 2023 (Huang et al., NHATS 2021) |
| Incident dementia, hearing aid users younger than 70 at hearing test (Framingham) | HR 0.39 (95% CI 0.17-0.89), P = .03 | JAMA Neurology 2025 (Francis et al.) |
| Incident dementia, hearing aid users aged 70 or older (Framingham) | HR 0.93 (95% CI 0.63-1.36), P = .69 | JAMA Neurology 2025 (Francis et al.) |
Context note: in ARIC, hearing aid use among people with hearing loss was not significantly associated with dementia (HR 0.87, 95% CI 0.68, 1.12), and only 8% of Black participants with hearing loss used hearing aids against 33% of White participants (Frontiers in Epidemiology, Deal et al. 2026). Primary sources: the meta-analysis in JAMA Neurology, the Danish cohort in JAMA Otolaryngology, the Framingham research letter, and the retracted UK Biobank paper.
4. The ACHIEVE Trial: A Null Result and a 48% Subgroup
ACHIEVE is the reason this topic moved from epidemiology to trial evidence, and its headline number is routinely quoted without its context. The primary analysis was negative: three-year change in global cognition was -0.200 standard deviation units with the hearing intervention and -0.202 with health education, a difference of 0.002 (95% CI -0.077 to 0.081; p=0.96).
The 48% figure belongs to 238 of the 977 participants. They were recruited from the long-running Atherosclerosis Risk in Communities (ARIC) study, and their control group declined by -0.402 units over three years against -0.151 among the healthier volunteers recruited through advertising. A hearing intervention can only slow a decline that is happening, which is the explanation the authors offer for the split, and the test for a difference between the two cohorts was significant (p=0.010). A later analysis that ranked all participants by predicted risk found the same shape: decline was 61.6% slower in the top quartile.
| Metric | Value | Source |
|---|---|---|
| Screened and randomized (adults aged 70 to 84, four US sites) | 3,004 screened; 977 randomized (490 intervention, 487 control) | The Lancet 2023 (Lin et al., ACHIEVE) |
| Primary outcome: 3-year change in global cognition, intervention vs control | -0.200 vs -0.202; difference 0.002 (95% CI -0.077 to 0.081), p=0.96 | The Lancet 2023 (Lin et al., ACHIEVE) |
| ARIC cohort (n = 238) | Difference 0.191 (95% CI 0.022 to 0.360), p=0.027; 48% reduction in 3-year cognitive change | The Lancet 2023 (Lin et al., ACHIEVE) |
| De novo community cohort (n = 739) | Difference -0.061 (95% CI -0.151 to 0.028), p=0.18 | The Lancet 2023 (Lin et al., ACHIEVE) |
| Test for different effects between the two cohorts | p=0.010 | The Lancet 2023 (Lin et al., ACHIEVE) |
| Adherence and crossover at year 3 | Intervention group wore hearing aids a mean 7.2 hours per day (SD 5.2); 16.5% of controls (76 of 462) obtained hearing aids on their own | The Lancet 2023 (Lin et al., ACHIEVE) |
| Slowing of 3-year decline in the top quartile of predicted risk | 61.6% (95% CI 33.7% to 94.1%) | Alzheimer’s & Dementia 2025 (Pike et al.) |
Context note: participants averaged 76.8 years of age with a mean better-ear pure-tone average of 39.4 dB (SD 6.9), so the results describe that range of hearing loss and not more severe loss. The full paper is open access at PubMed Central, and the risk-stratified reanalysis is in Alzheimer’s & Dementia. Most recent available data for the primary cognitive endpoint: The Lancet, Lin et al. 2023.
5. After ACHIEVE: Falls, Loneliness, and the CHOICE Trial
The secondary analyses turned out clearer than the cognitive result. Over three years the hearing intervention group averaged 1.45 falls against 1.98 in the control group, a 27% reduction (The Lancet Public Health, Goman et al. 2025). Social networks shrank in both arms, by less in the intervention arm, and the authors themselves caution that the difference in network size may not be clinically meaningful.
The newest randomized evidence comes from Shanghai. CHOICE enrolled people who already had mild cognitive impairment plus moderate to severe hearing loss, the population in which an effect on dementia should be easiest to detect, and its primary endpoint still missed significance. Its secondary endpoint, return to a normal Clinical Dementia Rating score, favored hearing aids by a wide margin. Two large trials now share the same pattern: a null primary result with signals around it.
| Metric | Value | Source |
|---|---|---|
| Mean falls over 3 years, hearing intervention vs control (ACHIEVE) | 1.45 vs 1.98; difference -0.54 (95% CI -0.77 to -0.31); 27% reduction | The Lancet Public Health 2025 (Goman et al.) |
| Social network size over 3 years, intervention vs control (ACHIEVE) | 22.6 to 21.3 people vs 22.3 to 19.8 people; difference 1.05 (95% CI 0.01 to 2.09) | JAMA Internal Medicine 2025 (Reed et al.) |
| Loneliness score difference over 3 years (ACHIEVE) | -0.94 (95% CI -1.78 to -0.11) | JAMA Internal Medicine 2025 (Reed et al.) |
| CHOICE enrollment (adults 60 or older with mild cognitive impairment, Shanghai) | 21,908 screened; 703 randomized (353 hearing aids, 350 education) | JAMA Neurology 2026 (Chen et al.) |
| CHOICE primary outcome: conversion to dementia-level impairment at 24 months | 3.09% vs 4.74%; relative risk 0.59 (95% CI 0.25 to 1.38), P = .23 | JAMA Neurology 2026 (Chen et al.) |
| CHOICE secondary outcome: cognitive improvement (Clinical Dementia Rating of 0) | 15.34% vs 2.36%; relative risk 5.94 (95% CI 2.46 to 14.37) | JAMA Neurology 2026 (Chen et al.) |
| CHOICE participants completing the trial | 610 (86.77%) | JAMA Neurology 2026 (Chen et al.) |
Sources: the falls analysis in The Lancet Public Health, the social isolation analysis in JAMA Internal Medicine, and the CHOICE trial in JAMA Neurology, published online in September 2026.
6. Global Prevalence: Two Conditions Growing on the Same Curve
Both conditions are driven by the same demographic fact: 62.1% of people with hearing loss are older than 50 (The Lancet, GBD 2019 Hearing Loss Collaborators 2021). Age-standardised dementia prevalence is projected to stay essentially flat through 2050, so the growth in cases is population ageing and population growth, not rising individual risk. That is also why a common exposure such as hearing loss weighs so heavily in the attributable-risk tables.
The overlap is easiest to see where hearing is measured directly. In the United States, 65.3% of adults aged 71 and older have hearing loss, and the Alzheimer’s Association counts 7.4 million Americans aged 65 and older living with Alzheimer’s dementia in 2026.
| Metric | Value | Source |
|---|---|---|
| People living with dementia worldwide | 57 million (2021), over 60% in low- and middle-income countries | WHO Dementia Fact Sheet 2026 |
| New dementia cases per year | Nearly 10 million | WHO Dementia Fact Sheet 2026 |
| Dementia cases, 2019 estimate and 2050 forecast | 57.4 million to 152.8 million | The Lancet Public Health 2022 (GBD 2019 Dementia Forecasting Collaborators) |
| Alternative projection | 78 million in 2030; 139 million in 2050 | Alzheimer’s Disease International, Dementia Statistics |
| People with hearing loss worldwide, 2019 | 1.57 billion (20.3% of the population) | The Lancet 2021 (GBD 2019 Hearing Loss Collaborators) |
| Projected hearing loss, 2050 | 2.45 billion, a 56.1% increase from 2019 | The Lancet 2021 (GBD 2019 Hearing Loss Collaborators) |
| Disabling hearing loss requiring rehabilitation | 430 million people; over 25% of people older than 60 | WHO Deafness and Hearing Loss Fact Sheet 2026 |
| US adults aged 71 and older with hearing loss | 65.3% (21.5 million), reaching 96.2% at age 90 and older | JAMA Network Open 2023 (Reed et al., NHATS 2021) |
Context note: the two 2050 dementia projections differ by 13.8 million (152.8 million minus 139 million) because they come from different models; each is cited here as its own publisher reports it. Underlying sources include the WHO dementia fact sheet, the WHO deafness and hearing loss fact sheet, Alzheimer’s Disease International, and the Alzheimer’s Association. Most recent available data for the global hearing loss count: GBD 2019, published in The Lancet in 2021.
7. The Hearing Aid Uptake Gap and What It Costs
The attributable-risk figures describe a theoretical maximum, and uptake is nowhere near it. Globally, 83% of people who need a hearing aid do not use one, rising to 90% in the African region (International Journal of Audiology, Orji et al. 2020). In the United States, where nationally representative audiometric data exist, 29.2% of adults aged 71 and older with hearing loss use a device.
Cost explains part of the gap. Among US Medicare beneficiaries who could not get hearing care, 75% cited cost (KFF 2021), and WHO puts the annual global cost of unaddressed hearing loss at almost US$ 1 trillion against a required investment of less than US$ 1.40 per person per year. Set next to a US dementia care bill of $409 billion, the spending involved in hearing care is small, though no source in this roundup claims that closing the gap would reduce that bill by a stated amount.
| Metric | Value | Source |
|---|---|---|
| People worldwide who need a hearing aid, and share not using one | 401.4 million; 83% (90% in the African region) | International Journal of Audiology 2020 (Orji et al.) |
| US adults aged 71 and older with hearing loss who use hearing aids | 29.2% (about 6.4 million people) | JAMA Network Open 2023 (Reed et al., NHATS 2021) |
| US adults who could benefit and have ever used hearing aids | Fewer than 1 in 3 (30%) aged 70 and older; about 16% aged 20 to 69 | NIDCD Quick Statistics About Hearing 2024 |
| Average out-of-pocket spending among Medicare beneficiaries using hearing services | $914 in 2018; top 10% spent $3,600 or more | KFF, Dental, Hearing, and Vision Costs and Coverage Among Medicare Beneficiaries 2021 |
| Average annual dollar limit on hearing aids in Medicare Advantage plans | $960 (range $66 to $4,000) | KFF, Dental, Hearing, and Vision Costs and Coverage Among Medicare Beneficiaries 2021 |
| Extra total health care cost over 10 years with untreated hearing loss | $22,434 (95% CI $18,219-$26,648), 46% higher | JAMA Otolaryngology 2019 (Reed et al.) |
| Annual global cost of unaddressed hearing loss | Almost US$ 1 trillion | WHO Deafness and Hearing Loss Fact Sheet 2026 |
| US health and long-term care costs for Alzheimer’s and other dementias | $409 billion projected for 2026 | Alzheimer’s Association, 2026 Alzheimer’s Disease Facts and Figures |
Context note: WHO estimates a return of nearly US$ 16 for every US dollar invested in ear and hearing care over 10 years, and the global cost of dementia at US$ 1.3 trillion in 2019. Sources: the global needs analysis in the International Journal of Audiology, NIDCD Quick Statistics, the KFF issue brief, and the cost study in JAMA Otolaryngology. Most recent available data for Medicare out-of-pocket hearing spending: KFF 2021, using 2018 survey data.
Summary: Hearing Loss & Dementia Risk by the Numbers
| Metric | Value | Source |
|---|---|---|
| Weighted PAF of dementia: hearing loss, and all 14 modifiable risk factors | 7.0% and 45.3% | Lancet Commission on Dementia, 2024 report |
| PAF of incident dementia from audiometric hearing loss, US cohort | 32.0% (95% CI 11.0% to 46.5%) | JAMA Otolaryngology 2025 (Ishak et al.) |
| Dementia hazard ratio with adult-onset hearing loss | 1.35 (95% CI 1.26-1.45) | Ageing Research Reviews 2024 (Yu et al.) |
| Dementia risk per 10 dB worsening | 16% higher | Ageing Research Reviews 2024 (Yu et al.) |
| Dementia hazard ratio, moderate-to-severe hearing loss, ARIC | 1.68 (95% CI 1.24, 2.28) | Frontiers in Epidemiology 2026 (Deal et al.) |
| Dementia prevalence, normal hearing vs moderate to severe loss | 6.19% vs 16.52% | JAMA 2023 (Huang et al.) |
| Long-term cognitive decline, hearing aid users | HR 0.81 (95% CI 0.76-0.87) | JAMA Neurology 2023 (Yeo et al.) |
| Incident dementia, hearing aid users under 70 | HR 0.39 (95% CI 0.17-0.89) | JAMA Neurology 2025 (Francis et al.) |
| ACHIEVE 3-year cognitive change: all 977 participants, and higher-risk ARIC subgroup (n = 238) | Difference 0.002, p=0.96 overall; 48% reduction in the subgroup | The Lancet 2023 (Lin et al.) |
| Mean falls over 3 years with hearing intervention | 27% reduction (1.45 vs 1.98) | The Lancet Public Health 2025 (Goman et al.) |
| CHOICE conversion to dementia-level impairment at 24 months | 3.09% vs 4.74%, P = .23 | JAMA Neurology 2026 (Chen et al.) |
| People living with dementia worldwide | 57 million (2021) | WHO Dementia Fact Sheet 2026 |
| Dementia cases forecast for 2050 | 152.8 million | The Lancet Public Health 2022 (GBD 2019) |
| People with hearing loss worldwide | 1.57 billion (2019); 2.45 billion by 2050 | The Lancet 2021 (GBD 2019) |
| People needing a hearing aid who do not use one | 83% of 401.4 million | International Journal of Audiology 2020 (Orji et al.) |
| US adults 71 and older with hearing loss who use hearing aids | 29.2% | JAMA Network Open 2023 (Reed et al.) |
| Average Medicare out-of-pocket hearing spending | $914 (2018) | KFF 2021 |
| US dementia health and long-term care costs | $409 billion (2026) | Alzheimer’s Association 2026 |
Methodology and Sources
Every figure above was read in the primary document, or in a named intermediary’s report and traced back to the primary paper, during research for this article. We prioritized randomized trials, systematic reviews, nationally representative cohorts and official statistical releases, and excluded commercial blogs and retracted work. Nothing here is medical advice.
- Lancet Commission on dementia prevention, intervention, and care: 2024 report (Livingston et al., The Lancet 2024): PAF table, six-study meta-analysis, pooled hearing aid estimate.
- Lancet Commission 2020 report (Livingston et al., The Lancet 2020): previous hearing loss PAF and inputs.
- ACHIEVE randomized controlled trial (Lin et al., The Lancet 2023): enrollment, primary and cohort-specific outcomes.
- Cognitive benefits of hearing intervention vary by risk of cognitive decline (Pike et al., Alzheimer’s & Dementia 2025): risk-stratified ACHIEVE analysis.
- Effects of hearing intervention on falls (Goman et al., The Lancet Public Health 2025).
- Hearing Intervention, Social Isolation, and Loneliness (Reed et al., JAMA Internal Medicine 2025).
- CHOICE randomized clinical trial (Chen et al., JAMA Neurology 2026).
- Population Attributable Fraction of Incident Dementia Associated With Hearing Loss (Ishak et al., JAMA Otolaryngology 2025).
- Population attributable fraction of dementia from audiometric hearing loss (Smith et al., Journals of Gerontology Series A 2023).
- Adult-onset hearing loss and incident cognitive impairment and dementia (Yu et al., Ageing Research Reviews 2024).
- Hearing loss and incident dementia over 8 years in Black and White older adults (Deal et al., Frontiers in Epidemiology 2026).
- Hearing Loss and Dementia Prevalence in Older Adults in the US (Huang et al., JAMA 2023).
- Hearing Loss, Hearing Aid Use, and Risk of Dementia in Older Adults (Cantuaria et al., JAMA Otolaryngology 2024).
- Hearing loss and incident dementia (Lin et al., Archives of Neurology 2011).
- Association of Hearing Aids and Cochlear Implants With Cognitive Decline and Dementia (Yeo et al., JAMA Neurology 2023).
- Self-Reported Hearing Aid Use and Risk of Incident Dementia (Francis et al., JAMA Neurology 2025).
- Retracted: hearing aid use and dementia in the UK Biobank cohort (Jiang et al., The Lancet Public Health 2023): cited only to record the retraction.
- WHO dementia fact sheet (2026) and WHO deafness and hearing loss fact sheet (2026).
- Global prevalence of dementia in 2019 and forecasted prevalence in 2050 (GBD 2019 Dementia Forecasting Collaborators, The Lancet Public Health 2022).
- Hearing loss prevalence and years lived with disability, 1990-2019 (GBD 2019 Hearing Loss Collaborators, The Lancet 2021).
- Alzheimer’s Disease International dementia statistics and the Alzheimer’s Association 2026 Alzheimer’s Disease Facts and Figures.
- Prevalence of Hearing Loss and Hearing Aid Use Among US Medicare Beneficiaries Aged 71 Years and Older (Reed et al., JAMA Network Open 2023).
- NIDCD Quick Statistics About Hearing (page updated September 2024).
- Global and regional needs, unmet needs and access to hearing aids (Orji et al., International Journal of Audiology 2020).
- KFF: Dental, Hearing, and Vision Costs and Coverage Among Medicare Beneficiaries (2021).
- Trends in Health Care Costs and Utilization Associated With Untreated Hearing Loss Over 10 Years (Reed et al., JAMA Otolaryngology 2019).
- Data watch: the attributable share of dementia for hearing loss ranges from 7.0% (Lancet Commission 2024, global model) to 16.9% (NHATS, prevalent dementia) and 32.0% (ARIC-NCS, incident dementia); the estimates use different populations, hearing measures and outcome definitions and are not interchangeable, and all assume causality. Dementia totals conflict across publishers: WHO reports 57 million for 2021, the GBD forecast 57.4 million for 2019 and 152.8 million for 2050, and Alzheimer’s Disease International 139 million for 2050. WHO rounds the 2050 hearing loss projection to nearly 2.5 billion where GBD reports 2.45 billion. The ACHIEVE 48% figure is a prespecified sensitivity analysis in 238 participants inside a trial whose primary result was null, and the CHOICE primary endpoint was also not significant. The Framingham estimate rests on 72 hearing aid users under 70. Several inputs are more than three years old and are used as the most recent available data: Lin et al. 2011, Orji et al. 2020, the GBD 2019 papers (2021 and 2022), KFF 2021 (2018 data), Reed et al. 2019, and the 2023 papers from ACHIEVE, NHATS and JAMA Neurology. Newer data is expected from the ACHIEVE long-term follow-up study (NCT05532657) and the next Alzheimer’s Association annual report.
Last updated: October 11, 2026. We update this roundup quarterly, and the next refresh is expected when the Alzheimer’s Association publishes its 2027 Alzheimer’s Disease Facts and Figures report.