A US study drawing on more than two decades of data from the Panel Study of Income Dynamics finds that older adults who experience food insecurity in later life have roughly twice the odds of screening positive for probable dementia, compared with those who have reliable access to food. The research, led by Cindy W. Leung of the Harvard T.H. Chan School of Public Health and the University of Michigan, has been published in the American Journal of Clinical Nutrition.
In the United States, food insecurity among households with older adults has nearly doubled over the past two decades.
Food insecurity is linked to a range of health problems, including poorer diet quality, cardiometabolic disease, psychological distress, and worse mental health. Growing evidence suggests it may also affect the brain, with food‑insecure older adults showing poorer cognitive functioning and a higher risk of dementia—a pattern that tends to persist after accounting for education, income, and wealth.
Researchers have increasingly asked whether timing matters. Studies in Burkina Faso, Malaysia, and China have found that older adults who recall hunger or food insecurity in childhood are more likely to show cognitive impairment or dementia decades later. Food insecurity during mid-life may also be important, given the relatively high levels of insecurity in that life stage.
The study, conducted by Leung and colleagues, examined whether food insecurity in mid-life and in later life was associated with probable dementia among U.S. adults, and whether prolonged or repeated exposure carried even greater risk.
Data came from the Panel Study of Income Dynamics (PSID), described by the authors as the world’s longest‑running nationally representative household panel survey. Begun in 1968, PSID follows its original participants with an oversample of low‑income families. Data were collected annually until 1997 and every two years thereafter.
The analysis began with 2,685 respondents who completed the core survey in 1999, aged between 43 and 59. After excluding those with missing information on key variables, the final sample included 2,051 participants; 56% were women, and about three‑quarters were non‑Hispanic White.
Food security was measured every two years with an 18‑item questionnaire from the US Department of Agriculture. It asked about experiences such as being unable to afford enough food in the past year. Assessments treated 1999, 2001 and 2003 as mid‑life, and 2015, 2017 and 2019 as later life. Food insecurity was defined as low or very low food security, and participants were categorised as experiencing it in neither period, in only one period, or in both.
Dementia was assessed in 2019 and 2021 using the Eight Item Interview to Differentiate Aging and Dementia, a screening questionnaire that probes changes in memory, judgment, orientation to time and daily functioning over the past year. A score of 2 or more out of 8 was classified as probable dementia, with either the participant or another household member able to respond.
Most participants (86%) were food secure in both periods. About 3% experienced food insecurity in mid-life only, 5% in later life only, and 5% in both periods. Those with any period of food insecurity tended to have lower incomes and less education and were more likely to be women or Black or Hispanic. About 20% of the sample screened positive for probable dementia.
After accounting for age, sex, race and ethnicity, education, income, marital status, employment, home ownership, and prior diabetes, hypertension and heart disease, those who experienced food insecurity in both mid-life and later life had 2.18 times the odds of probable dementia compared with those who were food secure in both periods. Those who experienced food insecurity in later life only had 1.85 times the odds. In models that did not account for changes in these factors over time, the estimates were higher, at 2.96 and 2.52, respectively.
In terms of predicted probabilities, the typical estimated chance of probable dementia was about 11% for adults who were food secure in both periods and about 40% for those who were food insecure in both. People who were food insecure only in mid-life did not have significantly higher odds of probable dementia, although this group was small, at 88 people, making the estimate imprecise.
The number of food‑insecure years in later life also appeared to matter. Participants who were food insecure in two or three of the later‑life surveys had about 2.6 to 3.2 times the odds of probable dementia, while repeated episodes in mid‑life were not significantly associated with it. The results were similar for men and women and held when the researchers broadened the definition of food insecurity to include marginal food security.
“Food insecurity is an important social determinant of health that can impact cognitive impairment and probable dementia. Future research should investigate the roles of income and nutrition policies to alleviate food insecurity and to reduce the burden of cognitive impairment and dementias before older adulthood,” the study authors concluded.
The authors note that food insecurity is reversible and is the focus of programmes such as food pantries, “Food is Medicine” initiatives that provide produce prescriptions or medically tailored meals, and the Supplemental Nutrition Assistance Programme (SNAP). They raised concern that SNAP work requirements extended to adults aged 55 to 64 under a 2025 law could discourage participation among older adults and lead to more food insecurity.
There are caveats to consider. The study was observational, so it cannot show that food insecurity causes dementia. Later‑life food insecurity was measured close in time to the dementia screening, and because no cognitive data were collected in mid-life, researchers could not rule out that early cognitive decline made it harder for some people to secure enough food, rather than the other way around. Unmeasured factors such as neighbourhood conditions could also contribute to both.
The dementia measure was a screening tool rather than a clinical diagnosis, and food security was not measured between 2005 and 2013, leaving a gap in the middle of the study period. People excluded because of missing data tended to be poorer and less healthy, which the authors suggest may have made the associations appear weaker than they are. The study also included only people who lived long enough to be screened at age 65 or older, and it did not adjust for mental health, which is linked to both food insecurity and cognitive decline.
The paper, “Associations between mid-life and later-life food insecurity on probable dementia in a national sample of United States adults,” was authored by Cindy W. Leung, Noura E. Insolera, Heejin Lee, Julia A. Wolfson, Claire T. McEvoy, Esther M. Friedman, Steven G. Heeringa, and Wei Hao.
