NHATS is the National Health and Aging Trends Study, a long-running survey of more than 8,000 Medicare beneficiaries age 65 and older, run by Johns Hopkins researchers Vicki Freedman and Jennifer Wolff with the University of Michigan's Institute for Social Research, alongside its companion study of family caregivers, NSOC (the National Study of Caregiving). The headline numbers coming out of this pair of surveys describe a caregiving system under growing strain, but the way NHATS counts dementia cases means the real picture is almost certainly worse, not better, than the published figures suggest.
That distinction matters for anyone reading dementia statistics with a personal stake — someone noticing memory changes in themselves or a parent. A survey built to be cautious about labeling someone as having dementia will undercount people in the early stages of the disease, which is a different problem than a survey that exaggerates risk. Understanding how NHATS draws that line helps explain why its numbers should be read as a floor, not a ceiling.
Medical information disclaimer: This article is for general educational purposes only and does not provide medical advice, diagnosis, or treatment. Always consult a physician or other qualified health professional about symptoms, medications, tests, or treatment decisions.
Table of Contents
- What NHATS and NSOC Actually Measure
- The Caregiving Numbers Behind the Headline
- Why the Dementia Count Likely Runs Low, Not High
- What Today's Numbers Leave Out Entirely
- Where the Data Is Headed in 2026
- How to Read the Next Dementia Statistic You See
- Frequently Asked Questions
What NHATS and NSOC Actually Measure
NHATS follows the same panel of Medicare beneficiaries over time, tracking their health, daily functioning, and cognitive status as they age. NSOC interviews the unpaid family members and friends who help those same older adults, capturing what caregiving actually looks like week to week, according to Johns Hopkins Bloomberg School of Public Health.
Because the two studies are linked, researchers can connect an older adult's cognitive status directly to the experience of the person caring for them. That link is what made the February 2025 Health Affairs findings possible: a clear picture of how caregiving demands have shifted over more than a decade, not just a snapshot of any single year.
The Caregiving Numbers Behind the Headline
Comparing 2011 to 2022, the number of unpaid family caregivers supporting older adults rose 32%, climbing from 18.2 million to 24.1 million people, according to the Johns Hopkins Bloomberg School of Public Health press release on the Health Affairs study. That is a measure of breadth: more households pulled into caregiving than before. The intensity of that work grew even faster for one group in particular.
Among caregivers specifically helping someone with dementia, average weekly hours climbed nearly 50%, from 21.4 hours in 2011 to 31.0 hours in 2022. Low-income caregivers faced the greatest risk of negative consequences from that load, the same Johns Hopkins release reports. Read together, more people are caregiving, and the ones caring for someone with dementia are doing substantially more hours of it than caregivers did a decade earlier.
Why the Dementia Count Likely Runs Low, Not High
The place to slow down is how NHATS decides who has dementia at all. The survey sorts respondents into "probable," "possible," or "no dementia" categories using a proxy-administered screening tool called AD8, combined with any self-reported or doctor-reported diagnosis on file. Researchers describe this cutoff as conservative, meaning people with mild cognitive impairment or early-stage dementia are often classified as having no dementia, according to the analysis published on NCBI/PMC. That design choice has a real-world echo in the diagnosis records themselves.
A claims-linked analysis of NHATS participants who died between 2011 and 2020 found that only 67.9% had a formal Alzheimer's or related-dementia diagnosis on record; 32.1% died without ever being formally diagnosed, according to data reported in Geriatrics journal. That pattern points to underdiagnosis as the dominant source of error in this population — not survey respondents being counted as having dementia when they do not. For someone noticing memory changes in themselves or a family member, this means a "no dementia" classification in population data is not the same as a clean medical clearance. It reflects where a conservative screening threshold happened to fall, not a diagnosis.
What Today's Numbers Leave Out Entirely
A second gap sits outside the survey's screening method altogether: who gets interviewed in the first place. An issue brief built on NHATS data from the HHS Office of the Assistant Secretary for Planning and Evaluation explicitly excludes nursing-home residents from its dementia-prevalence and caregiver estimates, a limitation the agency states directly, according to ASPE's report. Dementia prevalence tends to be highest among people in institutional care, so leaving that group out understates the total burden across the older adult population.
There is also a timing gap worth knowing about. NHATS and NSOC public-use files release without an embargo, but the more detailed restricted files — the ones linking survey answers to Medicare claims or other administrative records — require a Data Use Agreement that can take roughly a year to approve, according to SGIM's resource page on the studies. That means the most detailed, claims-verified dementia statistics often reach the public a year or more after the underlying data was actually collected. A caregiving statistic circulating in 2026 may describe conditions from 2023 or 2024, not the present moment.
Where the Data Is Headed in 2026
The agencies behind these studies are aware of these gaps and are actively building toward closing them. NIA's current limited-competition funding notice for NHATS/NSOC lists expanded cognitive and dementia measures, richer dementia-caregiving content, and improved data on residential and nursing-home care as stated priorities for continued data collection, per NIH Guide notice RFA-AG-24-037. Eligibility for that funding is restricted to investigators who already have access to the existing sample.
A specific project already underway reflects that direction. Johns Hopkins and University of Michigan researchers Judith Kasper and Vicki Freedman are enhancing NSOC under NIA funding to directly compare caregivers of people with and without Alzheimer's or related dementias, adding a module on caregiver involvement in medical care along with caregiver time-diary data, according to the Johns Hopkins project page. Both efforts point toward finer-grained measurement, not a correction suggesting current numbers overstate the problem.
How to Read the Next Dementia Statistic You See
Given how the measurement works, a few habits keep a reader from over- or under-reacting to any single number drawn from NHATS or NSOC:.
- Treat a "no dementia" classification in population statistics as a screening result, not a diagnosis — it misses many people with mild cognitive impairment or early-stage dementia.
- Assume published dementia and caregiver figures describe community-dwelling older adults unless a source says otherwise, since nursing-home residents are typically left out.
- Check the data year behind any statistic; a figure can lag the headline by a year or more when it draws on claims-linked restricted files rather than the public-use release.
- Read caregiver-hours statistics as a sign of growing intensity, not just growing headcount — the 2011-to-2022 comparison shows both moving at once, with dementia caregivers' hours rising fastest.
- Expect future NHATS/NSOC releases to carry more detailed cognitive measures and caregiving content as NIA-funded enhancements reach the data, rather than a scaling back of the dementia categories now in use.
Frequently Asked Questions
Does a "no dementia" result in NHATS mean a person's memory is fine?
Not necessarily. NHATS uses a conservative screening method, so people with mild cognitive impairment or early dementia are often placed in the "no dementia" category rather than flagged.
Are nursing-home residents included in NHATS dementia statistics?
Generally not. An ASPE issue brief built on NHATS data states that its dementia-prevalence and caregiver estimates exclude nursing-home residents, which understates total burden among institutionalized older adults.
Why do some dementia caregiving statistics seem out of date?
Restricted NHATS/NSOC files linking to Medicare claims require a Data Use Agreement that can take about a year to approve, so the most detailed statistics often lag actual data collection by a year or more.
You Might Also Like
- The Ideal Dementia Care Map Discussed by NIA in 2026: Questions Adult Children Can Ask Local Care Programs
- The $100 Million PROTECT-Cog Alzheimer’s Study Announced in 2026: Research Participation Questions for Adults With Memory Concerns
- The Ideal Dementia Care Map Discussed by NIA in 2026: What the Report Means for Adult Children
