Why this works

16 modifiable risk factors, each backed by at least one published study. The attributable-risk figures come from the 2024 Lancet Commission and the studies it cites. Every claim on this page links to its source: 20 studies, each listed with its title, journal, year and DOI.

How to read the evidence grades

Commission report
A standing panel of researchers reviews every study on a factor and weighs the evidence together. The 2024 Lancet Commission stands behind most of the numbers on this page.
Meta-analysis
One analysis that pools the results of many studies, so no single odd result can carry the conclusion.
Randomised trial
People were assigned by chance to either the intervention or a control group. The strongest test of cause and effect we have.
Cohort study
A large group of people followed for years or decades. Shows a real-world link, though not proof of cause by itself.

diet

Unhealthy diet— of population risk

Diets low in greens, fish and olive oil and high in processed food are associated with faster cognitive decline.

Meta-analysisMorris MC et al., Alzheimer's & Dementia 2015Barnes LL et al., New England Journal of Medicine 2023

education

Less education1.7% of population risk

Fewer years of formal education roughly doubles the dementia rate in pooled studies, likely through reduced cognitive reserve.

Meta-analysisMeng X et al., PLoS One 2012Livingston G et al., The Lancet 2024

environment

Air pollution0.7% of population risk

Fine-particle pollution is linked to higher dementia risk, including at levels common in US cities.

Meta-analysisWilker EH et al., BMJ 2023Livingston G et al., The Lancet 2024

injury

Head injury1.0% of population risk

A single head injury raises dementia risk by about 24 percent. Five or more raise it nearly threefold.

Cohort studyFann JR et al., The Lancet Psychiatry 2018Livingston G et al., The Lancet 2024

metabolic

High LDL cholesterol7.0% of population risk

The 2024 Lancet Commission added high LDL cholesterol as the second-largest modifiable factor, accounting for 7% of population dementia risk.

Commission reportLivingston G et al., The Lancet 2024
Diabetes or high blood sugar1.1% of population risk

Type 2 diabetes raises dementia risk by a quarter to nearly double across pooled studies, through vascular damage and insulin resistance in the brain.

Meta-analysisXue M et al., Ageing Research Reviews 2019Livingston G et al., The Lancet 2024
Obesity0.7% of population risk

Obesity at age 50 nearly doubles dementia risk. Weight loss in the years just before diagnosis is a sign of the disease, not protection.

Cohort studySingh-Manoux A et al., Alzheimer's & Dementia 2018Livingston G et al., The Lancet 2024

movement

Physical inactivity1.7% of population risk

Physically active people have about 20 percent less dementia risk in pooled studies. Walking counts.

Commission reportLivingston G et al., The Lancet 2024Iso-Markku P et al., British Journal of Sports Medicine 2022

sensory

Hearing loss8.2% of population risk

Untreated midlife hearing loss is the single largest modifiable risk factor identified by the 2024 Lancet Commission.

Commission reportLivingston G et al., The Lancet 2024Lin FR et al., The Lancet 2023
Vision loss2.0% of population risk

Uncorrected vision impairment was linked to a 78% higher dementia rate in a UK cohort of 117,000 older adults.

Cohort studyZhu Z et al., American Journal of Ophthalmology 2022Livingston G et al., The Lancet 2024

sleep

Sleep apnea— of population risk

Untreated obstructive sleep apnea increases dementia risk by about 26% through oxygen deprivation and sleep fragmentation.

Meta-analysisLeng Y et al., JAMA Neurology 2017Sabia S et al., Nature Communications 2021

social

Depression1.9% of population risk

Late-life depression raises dementia risk by about 85 percent, and vascular dementia risk by two and a half times.

Meta-analysisDiniz BS et al., British Journal of Psychiatry 2013Livingston G et al., The Lancet 2024
Social isolation1.6% of population risk

Loneliness and lack of social contact increase dementia risk by about 50%.

Meta-analysisKuiper JS et al., Ageing Research Reviews 2015Livingston G et al., The Lancet 2024

vascular

High blood pressure1.9% of population risk

Midlife hypertension damages small blood vessels in the brain, raising dementia risk. Intensive control reduces it.

Commission reportLivingston G et al., The Lancet 2024Williamson JD et al., JAMA 2019
Smoking1.4% of population risk

Current smoking raises dementia risk by about 27 percent and Alzheimer's risk by about 80 percent. Quitting at any age reduces risk.

Commission reportLivingston G et al., The Lancet 2024Anstey KJ et al., American Journal of Epidemiology 2007
Excessive alcohol0.5% of population risk

Drinking 14 or more UK units a week (about 7 US drinks) was linked to shrinkage of the brain's memory centre. Even light drinking showed no protection.

Cohort studyTopiwala A et al., BMJ 2017Livingston G et al., The Lancet 2024

What we leave out

You will not find these in the app, and the reason is the evidence. Next to hearing, sleep and vascular health, these have not held up.

Brain-training games
The one trial that showed a benefit (FINGER) combined brain training with diet, exercise and vascular monitoring. Games alone have not shown they prevent cognitive decline.
Supplements
B vitamins, ginkgo, vitamin E and fish-oil capsules have each been tested in randomised trials. None held up.
Coconut oil
It reached Facebook before it reached the journals. There is no trial evidence behind it, and its saturated fat points the wrong way against the diet evidence above.

When something new passes a real trial, it gets a page here with the DOI, same as everything else.

This is research information, not medical advice. Nothing here replaces your parent's doctor.