
Medicare maps show a Southern dementia diagnosis belt that tracks cardiometabolic risk more than age
Medicare claims show diagnosed dementia concentrated in the South and in high-cardiometabolic counties, including places that are not simply the oldest. Researchers link the pattern to diabetes, vascular disease, and uneven diagnosis—not age alone—while stressing that the maps count recorded cases, not proven causation.
Documented: Diagnosed dementia among Medicare beneficiaries is not evenly spread across the United States, and the pattern does not simply follow where the population is oldest. NORC at the University of Chicago’s Dementia DataHub, built from Medicare claims, reported that in 2020 the highest state rates in its “highly likely” dementia category included Puerto Rico (10.6%), Florida (8.2%), Texas (8.1%), Connecticut (8.0%), and New York (7.9%). Axios, citing the same hub, confirmed Florida’s 8.2% state rate and noted that the system counts diagnosed cases only. The hub itself has stressed that a large share of people with dementia never receive a diagnosis code, so these figures are not true prevalence.
Documented: A December 2025 paper in Alzheimer’s & Dementia by Yale neurologist Adam de Havenon and colleagues modeled 2020 Dementia Data Hub figures covering about 61.5 million Medicare beneficiaries across 3,076 counties. Counties in the top tertile averaged 1,986 diagnosed cases per 100,000 residents; 43.8% of those high-burden counties were in the South, versus 9.4% of Western counties. A model that included higher diabetes prevalence, higher uninsured rates, greater fast-food access, higher White population share, smaller household size, smoking, and elevation predicted top-tertile status with an area under the curve of 0.84. The authors describe association and prediction, not causation. Fast-food access was one of several retained predictors; the paper does not show that proximity to restaurants causes dementia.
Documented: That Southern concentration overlaps a longer-established geography. CDC analyses of the Stroke Belt have long shown elevated stroke death rates concentrated in the South. A separate Barrow Neurological Institute analysis of roughly 20.8 million Medicare beneficiaries, published in Alzheimer’s & Dementia and summarized by the institute in 2025, estimated that eight cardiometabolic conditions together accounted for about 37% of incident dementia cases nationally, with the largest county-level fractions in the Southeast. Hypertension (9.6%), ischemic heart disease (6.7%), and heart failure (5.7%) were the largest individual shares. Researchers said regional differences were partly consistent with higher obesity, inactivity, and poorer diet in Southern states. A 15% proportional reduction in those risks was estimated to cut incident cases by about 6.3%—a modeled counterfactual, not an observed intervention result.
Documented: Diagnosis itself is geographically uneven. A 2024 analysis led by Julie Bynum, covered by CNN, found that people in ZIP codes with the lowest “diagnostic intensity” were about 28% less likely to receive a timely dementia diagnosis than the national average, while those in the highest-intensity regions were about 36% more likely. Traditional research has also placed a larger share of Alzheimer’s and related dementias in the Southeast. De Havenon has told reporters that New York and California’s relatively high recorded rates may partly reflect greater access to clinicians, while sparse specialist care in some rural states can leave cases uncounted. That claim is consistent with the diagnostic-intensity literature; it is not a completed correction of the maps.
Documented, at the global level: The 2024 Lancet Commission estimated that about 45% of dementia cases worldwide are potentially preventable or delayable by addressing 14 modifiable risks across the life course, including less education, hearing loss, hypertension, smoking, obesity, depression, inactivity, diabetes, excess alcohol, traumatic brain injury, air pollution, social isolation, high LDL cholesterol, and untreated vision loss. That is a population attributable fraction from observational evidence, not proof that any single U.S. county’s excess is fully reversible.
Claimed: Popular accounts, including a Daily Mail feature drawing on de Havenon’s comments, describe a “dementia belt” in which seven of 11 traditional Stroke Belt states also rank among the 15 states with the highest recorded dementia rates, and in which New England states with very old populations show comparatively low recorded rates. Those state rankings are plausible given the NORC and county papers, but the precise “seven of 11” cut depends on which year’s Medicare definition and which list of Stroke Belt states is used. Birth in a high-stroke-mortality state has also been linked, in earlier cohort work, to higher later dementia risk even after people move—suggesting early-life conditions can travel with people—but that is a different study design from the current Medicare maps.
Speculated: Framing the maps as evidence of a specific failure of environmental regulation overreaches the published models. In the de Havenon county analysis, air quality was not among the retained predictors; diabetes, insurance coverage, smoking, and fast-food access were. Whether fast-food density is a marker of poverty, diet, car-dependent land use, or something else is not established. Whether closing the diagnostic gap would shrink or widen the apparent Southern excess is also unknown: better detection raises recorded rates even if underlying disease does not change.
What the record supports is narrower and still consequential. Recorded dementia clusters where vascular and metabolic disease already cluster, and where the chance of being coded as having dementia also varies. Aging explains the national rise; it does not explain why neighboring counties and regions diverge this sharply.
Older adults and their families in high-burden Southern and high-cardiometabolic counties face more recorded dementia alongside the vascular diseases already concentrated there, while uneven diagnosis means some communities are undercounted and others look worse because care is easier to reach.
Sources (8)
- [1]Unraveling disparities in county-level dementia diagnosis prevalence across the United States (de Havenon et al., Alzheimer's & Dementia)(https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/alz.70954)
- [2]NORC Launches County-Level Dementia Data Portal(https://www.norc.org/research/library/norc-launches-county-level-dementia-data-portal.html)
- [3]Dementia DataHub (NORC)(https://www.dementiadatahub.org/content/dementiadatahub/us/en.html)
- [4]Florida leads nation in dementia among Medicare recipients, research finds (Axios)(https://www.axios.com/local/tampa-bay/2024/09/11/dementia-florida-medicare-research-hillsborough)
- [5]Your ZIP code may determine your dementia diagnosis, study finds (CNN)(https://www.cnn.com/2024/08/26/health/dementia-zip-code-wellness)
- [6]Barrow Study Examines Geographic Ties Between Dementia, Cardiometabolic Disease in the US(https://www.barrowneuro.org/about/news-and-articles/research/barrow-study-examines-geographic-ties-between-dementia-cardiometabolic-disease-in-the-us/)
- [7]The Lancet: Nearly half of dementia cases could be prevented or delayed by tackling 14 risk factors(https://www.eurekalert.org/news-releases/1052982)
- [8]Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission(https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)01296-0/abstract)