What You Can Do to Help
Adults who reached middle age with none of three common vascular problems — high blood pressure, diabetes, smoking — lived on average thirty years without dementia. Those carrying all three lived 17.5.10 Twelve and a half years, and every one of those three can be treated. The 2024 Lancet Commission puts the whole preventable share at around 45% across fourteen risk factors.1 That is a population figure and not a promise to anyone. What follows is what the evidence supports doing, where it is strong, where it is thin, and what to ask in the room.
Researchers followed about 12,000 American adults, average age 56, for a median of 26 years, and reported the result in a unit that needs no translation: dementia-free survival years. Adults who reached midlife with none of high blood pressure, diabetes or current smoking averaged close to 30 dementia-free years. Adults carrying all three averaged 17.5.10
Twelve and a half years is the gap, and all three of those conditions are treatable by ordinary medicine that already exists. This is an observational study, so it shows a pattern rather than proving cause — but two randomised trials now support the blood pressure half of it directly, and they are in the next section.
Type 2 diabetes on its own may roughly double the risk of dementia.11 Nothing else on this page moves a number that far.
The best-evidenced item on this page, and the only one where a randomised trial has now reduced dementia itself.
The large trial that worked. Nearly 34,000 adults aged 40 and over with uncontrolled blood pressure were randomly assigned either to a target below 130/80 or to usual care. At four years the treated group had significantly less dementia of any cause. Followed to seven years, intensive control produced a 15% reduction in dementia risk.12
The earlier trial, stated honestly. SPRINT MIND randomised more than nine thousand adults to a tighter or standard target and produced less mild cognitive impairment, and less of the combined outcome of MCI or dementia — but it did not significantly reduce dementia on its own.4 For two years that was the honest ceiling of what could be claimed. It is no longer the whole picture, because a trial four times larger has since reached the harder endpoint.
And the guidelines have moved. Citing both trials, the American Heart Association and the American College of Cardiology now recommend that adults with hypertension lower systolic blood pressure below 130 mm Hg specifically to help prevent mild cognitive impairment and dementia.13 A recommendation naming dementia as the reason is a different thing from a hopeful association.
The one most people would never connect to memory — and the joint largest of the fourteen.2
If following conversation in a noisy room has become work, that is worth acting on. The trial that tested hearing aids for this did not meet its primary endpoint: across everyone in the study, hearing aids made no measurable difference to cognitive decline. But in a pre-specified group of 238 people who were older and already at higher risk, decline was about 48% slower over three years.3 That is a subgroup result, not a settled one. It is also the reason we ask.
The most active people in the research have roughly a seventh less dementia than the least active.
Pooled across 21 studies and 26 cohorts, the relative risk in the most active compared with the least is 0.86.5 That is an association drawn from observing people rather than from randomising them, so some of it may run the other way — early changes can make people less active. But it is consistent, it is large, and movement is free.
Loneliness is associated with a 31% higher risk of dementia.
Across more than 600,000 people in 21 cohorts, people who reported feeling lonely had substantially higher rates of later dementia.6 Again this is association: withdrawing from people can be an early sign rather than a cause. It is on this page because it is one of the fourteen,1 because it is measurable, and because it is one of the few things on the list a family can do something about this week.
Two of the largest levers on the list, and both are already treated routinely by medicine you can get this month.
Type 2 diabetes may roughly double the risk of dementia11 — and it is one of the three conditions in the thirty-years-versus-seventeen finding at the top of this page. High LDL cholesterol in midlife carries 7% of the population risk, the joint largest share of the fourteen alongside hearing.1
Neither is something this platform can measure — they need a blood test and a clinician. We ask about them so that the report you carry into the room already says whether they are in the picture, and so nobody spends the appointment establishing what you already knew.
The most treatable thing on the page, and the most often missed.
Persistent pain is associated with about 9.2% faster memory decline.7 And in a randomised trial in nursing homes, simply treating pain properly reduced agitation by around 17% — behaviour that would otherwise have been read as the dementia getting worse.8 That is why questions about pain sit inside a memory review.
Depression is treatable and can look exactly like early dementia. It is also on the list of fourteen.1
The other direction matters too: in 181,093 veterans, post-traumatic stress disorder carried an adjusted hazard ratio of 1.77 for later dementia — close to double.9 That is why this platform opens with 27 questions about mood, anxiety, bipolar symptoms and trauma rather than going straight to memory.
Most of this page is single risk factors, because that is how the research is usually done. But the strongest signal of the last decade came from testing several changes together, which is closer to how a person actually lives.
Finland, 2015. About 1,300 people aged 60 to 77, all at raised risk of dementia, were randomly assigned to two years of structured diet, exercise, cognitive training and vascular monitoring, or to general health advice. Both groups improved. The structured group improved 25% more.14 It was the first randomised trial to show that changing how someone lives can protect thinking.
The United States, 2025. US POINTER ran the same idea with about 2,100 Americans aged 60 to 79 at elevated risk. After two years the structured, higher-intensity programme beat the self-guided one on global cognition — the finding behind the "cognitive scores equivalent to people up to two years younger" described further down this page.15 A trial across eleven Latin American countries reported similar results in 2026, using a culturally adapted version of the same programme.16 The approach has now been taken up in more than seventy countries.19
What honesty requires us to add. The benefits in these trials are described by their own investigators as fairly modest, and whether they translate into fewer people getting dementia is not yet known. Some of the improvement in every group, including the controls, is the practice effect of sitting the same cognitive test repeatedly. And one large trial of a ten-year weight-loss programme in people with diabetes found the picture depended on where someone started: cognitive impairment was reduced among those who were overweight at baseline, unchanged among those who were obese, and increased among those with the greatest obesity — a result its own authors call hard to explain, and which does not appear to track how much weight anyone lost.17 The investigator's own summary is the fairest thing anyone has said about it: losing weight is good for diabetes and good for the heart, and following what the POINTER programme asked of people is probably good for the brain.
The practical version. Nothing in these trials required anything exotic. Blood pressure watched and treated. Blood sugar and cholesterol watched and treated. Not smoking. Moving most days. Sleeping. Eating reasonably. Keeping weight in a sane range. The American Heart Association bundles those eight as Life's Essential 8, and a 2026 review concluded that following them may protect the brain as well as the heart.18 It is the least surprising list in medicine, and it is the one with a randomised trial behind it.
Everyone has heard “exercise and eat well.” These are the five that surprise people — each one tied here to the strongest source behind it rather than to the headline that carried it. Where the evidence is a trial, it says so. Where it is observational, it says that too.
20%
Get the shingles vaccine. A natural experiment in Wales — eligibility set by exact date of birth, so people born weeks apart differed only in whether they could get the shot — found vaccination cut new dementia diagnoses by 3.5 percentage points over seven years, a 20% relative reduction. The effect was stronger in women.
Eyting et al., Nature 2025 · regression discontinuity, not a correlation2 of 14
Treat hearing loss, and check your eyes. Hearing loss and vision loss are both on the Lancet Commission's list of fourteen modifiable risk factors. They are among the easiest to act on and the least likely to be raised in an appointment about memory.
Lancet Commission 2024, via Facts and Figures 2026, page 194×/wk
Move — and lift, not just walk. U.S. POINTER's exercise prescription was 30–35 minutes of moderate-to-intense aerobic activity four times a week, plus strength and flexibility work twice a week. Participants in the structured programme finished with cognitive scores equivalent to people up to two years younger.
U.S. POINTER, via Facts and Figures 2026, pages 99–10013 yrs
Mind the vascular set in midlife. High blood pressure, diabetes and smoking, measured between ages 45 and 65: people with none of the three went about 30 years before dementia appeared; people with all three, about 17. High cholesterol is on the Lancet list alongside them.
ARIC cohort, 12,409 adults, 26 years · Neurology Open Access 2026Air
Reduce air-pollution exposure. Also one of the fourteen — and the one nobody lists when asked to name a dementia risk factor. It is a vascular exposure as much as a respiratory one.
Lancet Commission 2024, via Facts and Figures 2026, page 19—
And one we will not oversell. “Good stress versus bad stress,” novelty-seeking, specific brain-training regimens: the mechanisms are plausible and the trial evidence is thin. Cognitive and social engagement is on the Lancet list; a particular puzzle is not.
Stated as a limit, because a health page that oversells one item devalues the other fourA note on where these lists come from. The popular round-ups — including the World Economic Forum's widely shared “Dementia is rising fast — 5 surprising ways to protect your brain” — are summaries of the same underlying work: the Lancet Commission's modifiable risk factors, the U.S. POINTER trial, and the 2025 shingles-vaccine natural experiment. This page cites the underlying work rather than the summary, so the strength of each claim can be judged on its own.
Midlife is where the modifiable risk sits. In the same ARIC cohort — 12,409 adults measured between ages 45 and 65 and followed an average of 26 years — people with none of high blood pressure, diabetes or smoking went about 30 years before dementia appeared; people with all three went about 17. A gap of nearly 13 dementia-free years (Neurology Open Access, 2026).
The report notes that subjective cognitive decline “is often an indicator of other, treatable conditions, such as sleep apnea, depression or anxiety” (page 29). This is precisely why the sitting carries both instruments rather than a cognitive one alone: mood, sleep and worry are measured alongside observed thinking, so a clinician can see which one moved, and when.
Facts and Figures 2026, Overview, page 19
The fourteen: lower education, vision loss, high cholesterol, head injury, physical inactivity, smoking, excessive alcohol consumption, hypertension, obesity, diabetes, hearing loss, depression, infrequent social contact and air pollution. In the United States, a study of more than 375,000 participants estimated that nearly 37% of dementia cases were associated with eight modifiable factors (page 19).
The Alzheimer’s Association’s own 10 Healthy Habits for Your Brain covers the same ground (page 98), and the U.S. POINTER trial — the first large randomized U.S. trial to show a lifestyle intervention protecting cognitive function — built one of its four domains around exactly this: “Health Monitoring: regular check-ins on blood pressure, weight and lab results” (page 100). Participants in the structured programme finished with cognitive scores equivalent to people up to two years younger (page 99).
Everything on this page is about what can be changed. These are the four questionnaires the review actually asks, which is how any of it gets noticed in the first place. Each is one published instrument, reproduced unaltered and scored against its own published thresholds — no blends, and no boundary invented here. Two are answered by the person; two by someone who knows them well.
The Short IQCODE — 16 questions. Everyday things, each rated against how they
were ten years ago: remembering a conversation, handling money, working a familiar machine.
Not “can they do it” but “has it changed” — which is what people
closest to someone notice first, often years before a clinic test does.
Jorm AF. A short form of the IQCODE: development and cross-validation.
Psychol Med. 1994;24(1):145–153. Community accuracy from Quinn TJ et al.,
Cochrane Database Syst Rev. 2021;CD010079 — sensitivity 0.80, specificity 0.84.
No copyright; never billed for.
The IADL-C short form — 11 questions. Activities that
need planning and judgment rather than strength: a budget, a sequence of errands, an insurance
form. It asks the thing the others do not — whether lists, reminders and routines are
being used to keep managing. Someone slipping and coping looks nothing like someone slipping
and not coping, and only this instrument can tell them apart.
Schmitter-Edgecombe M, Parsey C, Lamb R. Arch Clin Neuropsychol.
2014;29(8):776–792. Cut-offs from Rahman S et al., same journal, 2025;40(6):1101–1111.
Washington State University — licence requested, not yet granted.
The M3 Checklist — 27 questions. Depression, anxiety, bipolar spectrum and
post-traumatic stress on a single page. It is first in the sitting for the reason this page keeps
returning to: depression in an older adult is regularly mistaken for dementia, and unlike
dementia it is treatable. One question asks about thoughts of suicide, and a positive answer
brings up the 988 line immediately, whatever else the sitting says.
Gaynes BN, DeVeaugh-Geiss J, Weir S, et al. Feasibility and diagnostic validity
of the M-3 checklist. Ann Fam Med. 2010;8(2):160–169. n = 647, against the MINI.
Depression sensitivity 0.84, specificity 0.80; any disorder 0.83 and 0.76.
M3 Information, LLC.
The DVPRS — 5 questions. Pain right now, then how much it
interfered with activity, sleep, mood and stress over the past day. Every number carries a plain
description — 6 is “hard to ignore, avoid usual activities” — which
matters on a form filled in by someone whose words are going.
Polomano RC, Galloway KT, Kent ML, et al. Pain Medicine.
2016;17(8):1505–1519, N = 307. Defense & Veterans Center for Integrative Pain Management,
Uniformed Services University of the Health Sciences. A US federal work, free to use as is,
without alteration.
All four run the same way: a higher score is worse. A score falling is not the mirror image of a score rising. Mood and pain genuinely improve, and a fall on those two most likely means what it looks like. The IQCODE asks about ten years ago, so a lower score means the person describing them has changed their account — not that the person recovered. The report says which every time.
Only one of the four publishes a threshold for calling a change real. The M3 does: 2–3 points. The IQCODE, the IADL-C and the DVPRS publish none, so movement on those three is reported as direction and size and nothing more. Nobody has established what a meaningful change is on them, and this platform does not fill that in.
Full reference lists, the psychometrics for each instrument and the rights position on each are set out on the evidence page.
Everything above this happens over months. These take ten minutes, and cost nothing.
Who picks up matters more than the number. The Helpline is staffed in part by master’s-level dementia specialists, and what they give is a care consultation — your situation worked through, not a leaflet.
Good reasons to call
Other ways in
And if you would rather act than worry
10 Healthy Habits for Your Brain, and the free Brain Health Habit Builder — check the habits you have now and build a plan from there.
The ten are: challenge your mind · stay in school · get moving · protect your head · be smoke-free · control your blood pressure · manage diabetes · eat right · maintain a healthy weight · sleep well.
Four of those ten are things this review already asks you about — blood pressure, diabetes, moving and sleep — and each one has its own section further up this page. That is the join: this page tells you what is changeable, and the Habit Builder is somewhere to put it.
alz.org/…/10-healthy-habits-for-your-brain · alz.org/…/brain-health-habit-builder
Taken from alz.org on 21 September 2026 — the Helpline, chapter search, TrialMatch and brain health pages. Hours and services change; the Helpline number is the one to trust if anything here has gone stale.
M3 Information is not affiliated with, endorsed by, or sponsored by the Alzheimer’s Association, and no money moves in either direction. Nothing you enter into this review is shared with them, or with anyone else, unless you send it yourself.
If you are worried about someone right now, that is a reason to call their clinician rather than to wait for a quarterly review.