Food Insecurity, Memory, and Dementia Among US Adults Aged 50 Years and Older

Haobing Qian; Aayush Khadka; Suzanna M Martinez; Sonali Singh; Willa D Brenowitz; Adina Zeki Al Hazzouri; Tanisha G Hill-Jarrett; M Maria Glymour; Anusha M Vable

doi:10.1001/jamanetworkopen.2023.44186

. 2023 Nov 21;6(11):e2344186. doi: 10.1001/jamanetworkopen.2023.44186

Food Insecurity, Memory, and Dementia Among US Adults Aged 50 Years and Older

Haobing Qian ¹, Aayush Khadka ^1,^✉, Suzanna M Martinez ², Sonali Singh ², Willa D Brenowitz ³, Adina Zeki Al Hazzouri ⁴, Tanisha G Hill-Jarrett ⁵, M Maria Glymour ⁶, Anusha M Vable ¹

¹Department of Family and Community Medicine, University of California, San Francisco

²Department of Epidemiology and Biostatistics, University of California, San Francisco

³Kaiser Permanente Center for Health Research, Portland, Oregon

⁴Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, New York

⁵Global Brain Health Institute, University of California, San Francisco

⁶Department of Epidemiology, Boston University, Boston, Massachusetts

Accepted for Publication: October 10, 2023.

Published: November 21, 2023. doi:10.1001/jamanetworkopen.2023.44186

^✉

Corresponding Author: Aayush Khadka, PhD, Department of Family Community Medicine, University of California, San Francisco, 1001 Potrero Ave, #2211, San Francisco, CA 94110 (aayush.khadka@ucsf.edu).

Author Contributions: Drs Qian and Khadka had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Qian, Martinez, Zeki Al Hazzouri, Glymour, Vable.

Acquisition, analysis, or interpretation of data: Qian, Khadka, Martinez, Singh, Brenowitz, Hill-Jarrett, Vable.

Drafting of the manuscript: Qian, Khadka, Martinez, Singh.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Qian, Khadka, Glymour.

Obtained funding: Martinez, Vable.

Administrative, technical, or material support: Qian, Singh.

Supervision: Qian, Martinez, Vable.

Conflict of Interest Disclosures: Dr Brenowitz reported grants from the National Institute on Aging (NIA) outside the submitted work. Dr Hill-Jarrett reported receiving Neuropsychologist consulting fees for Cogstate. Dr Glymour reported receiving grants from the National Institutes of Health NIA during the conduct of the study. No other disclosures were reported.

Funding/Support: The study was supported by NIA grants RF1AG079202 (Martinez, Vable) and R01AG074351 (Vable).

Role of the Funder/Sponsor: The NIA had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

^✉

Corresponding author.

PMCID: PMC10663972 PMID: 37988079

Key Points

Question

Is food insecurity among older adults associated with higher subsequent dementia risk and memory decline?

Findings

In this cohort study of 7012 older adults, food insecurity was associated with an increased estimated dementia risk. Food insecurity was also associated with lower memory scores and faster memory decline.

Meaning

The findings of this study suggest that food insecurity among older adults is associated with worse cognitive performance and higher dementia risk.

Abstract

Importance

Despite existing federal programs to increase access to food, food insecurity is common among US older adults. Food insecurity may affect Alzheimer disease and Alzheimer disease–related dementias via multiple mechanisms, yet there is almost no quantitative research evaluating this association.

Objective

To examine whether food insecurity in older adults is associated with later-life cognitive outcomes.

Design, Setting, and Participants

This cohort study of US residents aged 50 years and older from the US Health and Retirement Study was restricted to respondents with food insecurity data in 2013 and cognitive outcome data between calendar years 2014 and 2018. Analyses were conducted from June 1 to September 22, 2023.

Exposure

Food insecurity status in 2013 was assessed using the validated US Department of Agriculture 6-item Household Food Security Module. Respondents were classified as being food secure, low food secure, and very low food secure.

Main Outcomes and Measures

Outcomes were dementia probability and memory score (standardized to 1998 units), estimated biennially between 2014 and 2018 using a previously validated algorithm. Generalized estimation equations were fit for dementia risk and linear mixed-effects models for memory score, taking selective attrition into account through inverse probability of censoring weights.

Results

The sample consisted of 7012 participants (18 356 person-waves); mean (SD) age was 67.7 (10.0) years, 4131 (58.9%) were women, 1136 (16.2%) were non-Hispanic Black, 4849 (69.2%) were non-Hispanic White, and mean (SD) duration of schooling was 13.0 (3.0) years. Compared with food-secure older adults, experiencing low food security was associated with higher odds of dementia (odds ratio, 1.38; 95% CI, 1.15-1.67) as was experiencing very low food security (odds ratio, 1.37; 95% CI, 1.11-1.59). Low and very low food security was also associated with lower memory levels and faster age-related memory decline.

Conclusions and Relevance

In this cohort study of older US residents, food insecurity was associated with increased dementia risk, poorer memory function, and faster memory decline. Future studies are needed to examine whether addressing food insecurity may benefit brain health.

This cohort study examines cognitive outcomes in older individuals with food insecurity.

Introduction

The number of US residents aged 65 years and older living with Alzheimer disease and Alzheimer disease–related dementias (AD/ADRD) is expected to increase from 5.8 million in 2020 to 14 million by 2060.¹ Similarly, food insecurity, defined as a lack of consistent access to enough food for a healthy, active lifestyle, is persistent, and the prevalence in households with elderly individuals increased from 5.3% (2001) to 7.1% (2021).^2,3,4 Among adults between the ages of 50 and 59 years, the prevalence of food insecurity is estimated to be higher at 9.4% in 2021.⁵ Older adults living with food insecurity are more likely to have lower nutrient intakes and experience poorer health outcomes, such as cardiovascular and metabolic diseases, increased stress and depression, and increased dementia risk.^6,7,8,9,10

The Lifecourse Health Development Framework, which explains how health trajectories develop over a lifetime, informs several plausible mechanisms by which food insecurity impacts dementia risk (Figure 1).¹¹ First, food insecurity arises due to financial constraints, which limit access to healthful foods and contribute to a lower quality, quantity, and variety diet.^12,13 Next, food insecurity may lead to poor nutrition, trigger stress pathways, or increase the likelihood of poor cardiometabolic health and mental illness. Ultimately, these factors, including food insecurity, may increase the risk of cognitive decline.^{9,14,15,16,17,18}

Figure 1. — Multiple pathways through which food insecurity may plausibly impact later-life cognitive outcomes and risk of dementia.

Few studies have rigorously investigated food security in terms of its association with AD/ADRD. Earlier studies on this topic have been conducted primarily on cross-sectional data with small and selected subpopulations or have used inconsistent measures of food insecurity and later-life AD/ADRD risk.¹⁹ A recent systematic review²⁰ identified only 1 longitudinal study examining food insecurity’s association with subsequent cognitive decline in older adults, with an association over a 2-year follow-up, consistent with several cross-sectional studies that also documented associations between food insecurity and cognitive decline.^{21,22,23,24,25}

As food insecurity may be modifiable through existing government programs (eg, the Supplemental Nutrition Assistance Program [SNAP]) and there are limited existing treatment options for dementia, it is important to evaluate whether food insecurity is associated with increased dementia risk.^26,27 We rigorously evaluated the association between food insecurity in later life and cognitive health, including dementia risk and age-related memory decline, in a large and diverse sample of US older adults. Our work builds on prior literature by using longitudinal data, validated measures of food insecurity and dementia risk, and robust adjustment for life course socioeconomic variables.

Methods

Sample

This cohort study used data from the Health and Retirement Study (HRS), a nationally representative, biennially fielded longitudinal survey of noninstitutionalized individuals aged 50 years and older and their spouses.²⁸ The HRS and its various substudies collect rich data on life course demographic characteristics, health, labor market, and socioeconomic status. This study was waived from institutional board review because it did not involve human participation by the Human Research Protection Program at the University of California, San Francisco. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

We used food security status data from the 2013 Health Care and Nutrition Study (HCNS), an HRS substudy regarding health care access, food purchases, food consumption, and nutrition (N = 8073).²⁹ We combined HCNS data with outcome and covariate data from the broader HRS survey. Our study period was from calendar years 2013 to 2018.

Among HCNS participants, we excluded those younger than 50 years (n = 176, typically spouses of age-eligible participants) or those not included in the HRS cross-wave respondent tracker data set (n = 2). We also excluded participants who did not provide complete data on food security status (n = 501) or had missing cognition data for the entire study period (n = 382).

Exposure

Food security status was assessed using the validated US Department of Agriculture (USDA) 6-Item Short Form US Household Food Security Survey Module.³⁰ Respondents were asked a series of questions related to food purchases and consumption over the past 12 months to determine food security status, such as whether they were able to afford the food they needed or whether they ate less due to financial constraints (full questionnaire presented in eTable 1 in Supplement 1). Using USDA guidelines, we summed the number of affirmative answers to the 6 questions (score range, 0-6), where higher raw scores indicate higher food insecurity levels. We then categorized the total score into 3 levels in accordance with the USDA guidelines: food secure (high and marginal statuses, scores: 0-1), low food secure (scores: 2-4), and very low food secure (scores: 5-6).

Outcomes

The outcomes were algorithmically defined and previously validated dementia probability and memory scores measured in the 2014, 2016, and 2018 waves of the HRS.³¹ Both measures were developed specifically in the HRS by training prediction models on the Aging, Demographics, and Memory Study (ADAMS) substudy. The ADAMS substudy focused on HRS respondents aged 70 years and older in 2000 and 2002 and had participants go through a full neuropsychological battery and dementia diagnosis.³²

Dementia probability is the risk of a respondent being demented at the time of the survey. It is a composite measure that combines responses to various cognitive questions in the HRS as well as demographic information.^31,33 Dementia probability values range from 0 to 1, and higher values reflect higher dementia risk. Compared with the standard clinical diagnosis of dementia, the dementia probability algorithm achieved a C statistic of 0.67, indicating acceptable performance.

Memory score is a measure of a respondent’s memory-related cognitive function at the time of the survey. It is also a composite measure that combines responses from various cognitive questions and demographic information. Higher values of memory score reflect better memory performance. We standardized memory score in our sample to the 1998 memory score distribution in the HRS for ease of interpretation.

Covariates

We adjusted for the following potential confounders: age at food security assessment (in 2013), sex (male and female), self-reported race and ethnicity as part of the HRS surveys (non-Hispanic Black, Hispanic, non-Hispanic White, and Other), educational level (in years, linear and quadratic terms), birthplace (by census regions), married or not married status (in 2012, the year preceding the food security assessment), age at each interview wave (centered at 70 years, linear and quadratic terms), and body mass index (self-reported, 2012). Race was included as a proxy for racialized experiences that structurally minoritized groups face, which is associated with both the exposure and outcome. The Other race category included non-Hispanic categories of Alaska Native, American Indian, Asian, Native Hawaiian, Pacific Islander, and any other self-specified race.

We also included flexible specifications of several baseline (2012) measures of income and wealth to account for confounding by socioeconomic status: total earnings; total wealth; total assets, including second home; poverty status (yes or no); mother’s and father’s years of education (in years, linear and quadratic terms); labor force status (yes or no); home ownership (categorical variable: own or buying, renting, or living rent-free with relative, employer, friends, or others), veteran status (yes or no); Social Security income; welfare benefits; veteran benefits; and SNAP benefits the household received. As these socioeconomic status measures may all be correlated, we examined the covariance matrix to evaluate collinearity and found that no confounders were correlated above 0.8. All income, wealth, and social benefits were inflation adjusted to 2018 US dollars. When covariate data were missing in 2012, we used the 2010 value instead. We also created missing indicators (1 = data missing, 0 = otherwise) for mother’s educational level and father’s educational level as missing parental education data may indicate that the respondent did not reside with their parents during childhood, which is a marker of social capital.^34,35 To minimize the possibility that those with lower cognition are less likely to apply for SNAP benefits, we additionally controlled for preexposure cognition, ie, dementia risk and memory score in 2012.³⁶

Statistical Analysis

To reduce selection bias due to missing data, we assumed the missingness mechanism was missing at random and used multiple imputation chained equation to impute missing covariates (558 [7.1%] individuals for dementia risk; 556 [7.0%] individuals for memory score) and outcome data (382 [5.4%] for the entire study period, 157 [2.2%] for baseline dementia risk, and 154 [2.2%] for baseline memory score).³⁷ We constructed 10 imputed data sets. In each imputed data set, we estimated inverse probability weights to address censoring due to death (details in the eMethods in Supplement 1).^38,39

In each imputed data set, we estimated the food insecurity and dementia risk association using generalized estimating equations specifying an independent correlation structure and applying robust SEs to account for repeated-outcome measures.^{40,41,42,43,44,45} Similarly, in each imputed data set, we estimated the food insecurity and memory association using a linear mixed-effects model. We modeled the food insecurity association with age-related memory decline, using an interaction between the exposure and age. We used Rubin rules to combine estimates across all imputed data sets.⁴⁶

After combining the estimates, we contextualized the food insecurity association with memory by dividing coefficients on the exposure variable by the coefficient on the age covariate in our model to translate our primary associations into years of excess cognitive aging. We plotted predicted memory values by age and food insecurity status to visualize the food insecurity and memory association.

To test whether our results were robust to different exposure specifications, we redefined food insecurity status using alternative USDA cut points: first, as a binary indicator for food secure and insecure status (food secure = high or marginal food security; food insecure = low and very low food security), and second, as a 4-level categorical variable for food secure status (high vs marginal vs low vs very low).

We also tested whether our results were robust to different outcome specifications. For dementia probability, we used 3 additional algorithmically defined measures of dementia risk developed in the HRS data by Gianattasio et al.⁴⁷ These were based on a modified Hurd algorithm, an algorithm developed using expert input to select covariates, and a machine learning–based least absolute shrinkage and selection operator–reduced logistic regression algorithm. Data on all 3 dementia risk measures were available for direct download from the HRS website. For memory score, we used immediate and delayed word recall as alternative outcomes. These scores were not algorithmically defined and were only available for direct respondents to the HRS.

Additionally, we refit our primary models by excluding SNAP benefits as a covariate. This is because SNAP benefits could potentially be endogenous with the food insecurity exposure, as there may be some temporal overlap between SNAP receipt status and the 1-year period during which food insecurity is measured. We also conducted the primary analyses in the complete case sample to assess our method of handling missing data. To acknowledge that adjusting for preexposure cognition may introduce regression-to-the-mean bias in change score analysis, we also performed sensitivity analyses without adjusting for the 2012 memory score.⁴⁸

Regression models were adjusted for age at baseline, age centered at 70 years (linear and quadratic terms), 2012 dementia risk, sex, race and ethnicity, years of education, mother’s education, father’s education (linear and quadratic terms), birthplace, marital status, self-reported body mass index, income and wealth (linear and quadratic terms), poverty status, labor force status, home ownership, amount received from food stamps, welfare benefits, veteran status, veteran benefits, and Social Security income. Regression model parameters were estimated after imputing the primary analytic sample 10 times to fill in missing values using inverse probability of censoring weights to account for potential differential attrition and combining the results from all imputed data sets using the Rubin rules.

All analyses were conducted in Stata MP, version 18 (StataCorp LLC). One of us (A.K.) reviewed the entire code as is recommended practice.⁴⁹ Analyses were conducted from June 1 to September 22, 2023. We report 2-sided 95% CIs. All statistical tests were 2-sided, with a significance threshold of P < .05.

Results

Our primary analytic sample included 7012 participants (18 356 person-waves; mean person-waves per respondent, 2.6). Respondents’ mean (SD) age was 67.7 (10.0) years, 4131 (58.9%) were women, 2881 (41.1%) were men, 1136 (16.2%) were non-Hispanic Black, 4849 (69.2%) were non-Hispanic White, 4357 (62.4%) were married in 2012, and mean (SD) duration of schooling was 13.0 (3.0) years (Table). Overall, 18.4% of the analytic sample were food insecure: 10.3% experienced low food security and 8.1% experienced very low food security. In the analytic sample, approximately 11% of individuals aged 65 years or older at baseline reported being food insecure. In contrast, approximately 28% of individuals younger than 65 years in our analytic sample reported being food insecure. Compared with those who were food secure (marginal or high), those with low and very low food security were younger, more likely to be women and non-Hispanic Black or Hispanic, had fewer years of schooling, lived in poverty, earned less, received greater welfare support, and were renters. Respondents experiencing low and very low food security were also less likely to be married.

Table. Baseline Characteristics by Food Security Status in the Primary Analytic Sample^a.

Characteristic	Overall (N = 7012)	Food security			P value
Characteristic	Overall (N = 7012)	High and marginal (n = 5719 [81.6%])	Low (n = 725 [10.3%])	Very low (n = 568 [8.1%])	P value
Outcome measures
Dementia risk in 2014, mean (SD)	0.1 (0.2)	0.1 (0.2)	0.1 (0.2)	0.1 (0.2)	.06
Standardized memory score in 2014, mean (SD)	−0.1 (0.9)	−0.1 (0.9)	−0.1 (0.8)	0.1 (0.7)	<.001
Covariates
Age in 2013, mean (SD), y	67.7 (10.0)	68.7 (10.0)	64.5 (9.4)	61.8 (7.6)	<.001
Sex, No. (%)
Female	4131 (58.9)	3316 (58.0)	456 (62.9)	359 (63.2)	<.001
Male	2881 (41.1)	2403 (42.0)	269 (37.1)	209 (36.8)	<.001
Years of education, mean (SD)	13.0 (3.0)	13.4 (2.7)	11.1 (3.6)	11.7 (3.3)	<.001
Race and ethnicity, No. (%)^b
Hispanic	795 (11.3)	471 (8.2)	185 (25.5)	139 (24.5)	<.001
Non-Hispanic Black	1136 (16.2)	759 (13.3)	215 (29.7)	162 (28.5)	.22
Non-Hispanic White	4849 (69.2)	4325 (75.6)	290 (40.0)	234 (41.2)	<.001
Other	232 (3.3)	164 (2.9)	35 (4.8)	33 (5.8)	.59
Southern birth, No. (%)	2314 (57.0)	1795 (51.0)	283 (88.4)	236 (78.2)	<.001
Married in 2012, No. (%)	4357 (62.4)	3764 (66.1)	360 (50.0)	233 (41.1)	<.001
Years of schooling, mean (SD)
Maternal	10.2 (3.6)	10.6 (3.3)	8.5 (4.1)	9.2 (4.1)	<.001
Paternal	10.0 (3.7)	10.3 (3.5)	8.6 (4.0)	9.0 (4.0)	<.001
In labor force in 2012 (yes vs no), No. (%)	3018 (43.3)	2451 (43.1)	320 (44.4)	247 (43.5)	<.001
Poverty status in 2012 (yes vs no), No. (%)	793 (11.4)	389 (6.8)	212 (29.4)	192 (33.8)	<.001
Veteran status in 2012 (yes vs no), No. (%)	793 (17.3)	389 (18.8)	212 (11.2)	192 (10.6)	<.001
Total wealth in 2012, mean (SD)	75 240.5 (113 795.5)	85 353.0 (122 996.2)	31 827.3 (31 452.8)	30 680.2 (35 757.1)	<.001
Earnings in 2012, mean (SD)	18 787.0 (42 603.3)	20 729.2 (45 797.5)	10 527.8 (19 475.4)	10 127.4 (25 651.4)	<.001
Total assets in 2012, mean (SD)	484 090.9 (1 136 305.7)	580 492.9 (1 235 835.9)	75 673.9 (272521.0)	52 346.4 (139 314.0)	<.001
Received food stamps since last interview wave (yes vs no), No. (%)	673 (9.6)	303 (5.3)	177 (24.5)	193 (34.0)	<.001
Amount of food stamps received among recipients, mean (SD)	123.8 (573.8)	61.8 (389.6)	324.1 (892.6)	480.9 (1128.0)	<.001
Amount of Social Security income received among recipients, mean (SD)	180.0 (1283.9)	87.5 (960.8)	517.2 (2081.3)	664.1 (2221.6)	<.001
Amount of veteran benefits received among recipients, mean (SD)	753.7 (5082.4)	846.1 (5477.3)	357.4 (2739.9)	345.6 (2750.4)	<.001
Amount of welfare benefits received among recipients, mean (SD)	19.0 (357.1)	3.9 (116.3)	66.2 (692.9)	108.0 (890.1)	.003
Self-reported BMI, mean (SD)	28.8 (6.0)	28.5 (5.8)	30.3 (6.3)	30.5 (7.3)	<.001
Residency status, No. (%)
Own home	5284 (75.4)	4635 (81.0)	401 (55.3)	248 (43.7)	<.001
Rent home	1183 (16.9)	682 (11.9)	247 (34.1)	254 (44.7)	<.001
Live rent free with others	202 (2.9)	142 (2.5)	34 (4.7)	26 (4.6)	.90
Other	21 (0.3)	17 (0.3)	3 (0.4)	1 (0.2)	.18

Open in a new tab

Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared).

^{^a}

The distribution presented is from the analytical sample before imputations.

^{^b}

All race and ethnicity designations were self-reported by Health and Retirement Study participants. Other races include the following non-Hispanic categories: Alaska Native, American Indian, Asian, Native Hawaiian, Pacific Islander, and any other self-specified race.

Dementia Risk Results

Compared with older adults with food security, those who experienced low food security had 1.38 times higher odds of dementia (95% CI, 1.15-1.67; in log odds: 0.33; 95% CI, 0.14-0.51), and those who experienced very low food security had 1.37 times higher odds of dementia (95% CI, 1.11-1.69; in log odds: 0.31; 95% CI, 0.10-0.52) (Figure 2; eTable 2 in Supplement 1). Translated to years of excess cognitive aging, point estimates showed that food insecurity is associated with increased dementia risk equivalent to approximately 1.3 excess years of aging.

Figure 2. — The analytical sample size is 7012 individuals, and 18 356 is the person-wave observations.

Memory Score Results

Compared with older adults with food security, those who experienced low and very low food security had worse memory levels at age 70 years (low β = −0.04; 95% CI, −0.08 to 0.00; very low β = −0.06; 95% CI, −0.1 to −0.01) (eTable 3 in Supplement 1). Translated to years of excess cognitive aging, low food insecurity was associated with decreased cognitive levels equivalent to approximately 0.7 years of excess aging per year, and very low food insecurity was associated with 1 year of excess aging per year. Older adults with food insecurity also had a marginally faster rate of age-related memory decline (low × age β = −0.005; 95% CI, −0.008 to −0.001; very low × age β = −0.009; 95% CI, −0.014 to −0.003) compared with food-secure individuals (eTable 3 in Supplement 1; Figure 3).

Figure 3. — Estimates were made based on the regression model fit in the primary analysis. The analytical sample size is 7012 individuals, and 18 356 is the person-wave observations.

Robustness Checks

Results were substantially similar (ie, point estimates and 95% CIs may vary, but conclusions are the same) under different exposure specifications (eTable 4 and eTable 5 in Supplement 1) and using different dementia algorithms (eTable 6 in Supplement 1) as well as immediate and delayed word recall scores (eTable 7 in Supplement 1). Estimates in models without adjusting for SNAP receipt (eTable 8 in Supplement 1), using complete case data (eTable 9 in Supplement 1), or without controlling for preexposure cognitive outcomes (eTable 10 in Supplement 1) were also quantitatively similar to the primary results.

Discussion

This cohort study evaluated the association between food insecurity among older adults and subsequent cognitive health using validated measures from the large, diverse longitudinal US HRS. We found that individuals who experienced low or very low food security had a higher probability of dementia, worse memory level, and faster rate of memory decline compared with those who were food secure. Memory decline diverged slightly faster among the very low food security group than the low food security group. Our primary results were robust to alternative specifications of the exposure, outcome, and analytic models.

In the current study, 11% of HRS participants aged 65 years and older experienced food insecurity, which is approximately 4% higher than reported among older adult participants (age ≥60 years) in the 2020 Current Population Survey.⁵⁰ The higher prevalence of food insecurity in our data was possibly a consequence of the Great Recession (2007-2009) in this population, which is consistent with examined trends in food insecurity.^51,52 Current population survey data showed that nearly 9% of older adults experienced food insecurity with the onset of the Great Recession, with lingering effects into 2014.⁵³ We also found that individuals experiencing food insecurity were, on average, younger and had lower educational attainment compared with those with food security. These findings suggest that food insecurity is more common among the same socioeconomically disadvantaged groups who are at high risk of dementia.^54,55,56,57 Together, these findings highlight the role that social determinants of health may play later in life.

While much of the literature regarding food insecurity focuses on early life factors, our study is among the first highlighting the outcomes associated with food insecurity later in life.^24,58,59,60 Recent findings by Lu and colleagues⁶¹ examined SNAP participation among SNAP-eligible adults—a group that is vulnerable to food insecurity—and found faster memory function decline among those experiencing food insecurity or not participating in SNAP compared with their counterparts. Our study examined food insecurity as the exposure, which builds on findings by Lu and colleagues given that SNAP participation among eligible older adults is underused compared with other segments of the population (48% compared with 86% overall in 2018).^36,62

Strengths and Limitations

Our study has strengths. Using a large and diverse data set, our study is in line with and adds evidence to the limited previous literature on food security and brain health by using both validated exposure and outcome data: the 6-item USDA Household Food Security Module and algorithmically defined, validated dementia probability and memory scores. Additionally, our outcome measures allowed us to incorporate information from both direct and proxy respondents in the HRS. Another strength is that food security was assessed before the outcome; therefore, we were able to examine exposure that temporally preceded the outcome.

Our study has limitations. A previous study using HRS data found that among SNAP-eligible adults, participants with reduced levels of cognitive function were less likely to participate in SNAP.³⁶ However, in an advance over earlier work, we adjusted for preexposure cognition, minimizing this potential bias. Second, residual confounding remains a possibility in our study. To mitigate residual confounding, we included variables that reflected childhood and adulthood socioeconomic status and included flexible specifications of these variables. Weight loss in subclinical AD could exacerbate the effects of food insecurity on cognitive outcomes, although we adjusted for preexposure body mass index in 2012. Another potential confounder is food insecurity status before 2013, which was not available using the 6-item questionnaire. Future studies could examine food insecurity over a longer period. In addition, clinical diagnosis of dementia was not directly observed.

Our cohort study has several policy implications. The findings suggest that food insecurity remains high among older adults. Additionally, given limited treatment options for dementia, reducing its risk by targeting modifiable risk factors, such as food security, may be necessary. Food security may be easier to modify than other potentially modifiable risk factors, such as education and exercise, as it may be addressable through existing federal programs, such as SNAP.⁶³ Increasing SNAP participation rates among older adults might constitute a viable population-level solution to reduce AD/ADRD disparities and improve brain health. The take-up rates of SNAP among eligible individuals varies substantially by state, ranging from 55% in Wyoming to an estimated 100% in a number of states, such as Oregon, Washington, and Illinois.^62,64 In general, one reason for lower uptake is high administrative burden in completing paperwork among eligible participants, which suggests that simplifying the application process might increase the take-up rate, especially for those with any cognitive impairment.⁶⁵

Conclusions

In this cohort study, we found that food insecurity in older adulthood was associated with increased dementia risk and faster memory decline. Our study contributes to a limited literature by capitalizing on a large and diverse sample, validated exposure and outcome measures, and longitudinal data to robustly evaluate these associations, providing evidence in support of the connection between food insecurity in older adulthood and subsequent brain health. Our findings highlight the need to improve food security in older adults and that doing so may protect individuals from cognitive decline and dementia. Bolstering SNAP by making it easier for older adults who are SNAP eligible to apply could potentially mitigate the negative association food insecurity has with brain health.

Supplement 1.

eTable 1. Six-Item USDA Food Security Module, Assessed in HRS Respondents in 2013

eTable 2. Association of Food Insecurity With Dementia Risk in Primary Analytic Sample

eTable 3. Association of Food Insecurity With Memory Levels and Age-Related Decline in Primary Analytic Sample

eTable 4. Association of Food Insecurity With Dementia Risk Under Different Exposure Definitions

eTable 5. Association of Food Insecurity With Memory Score Among Respondents Under Different Exposure Definitions

eTable 6. Association of Food Insecurity With Dementia Risk Under Different Definitions of the Dementia Risk Outcome

eTable 7. Association of Food Insecurity With Word Recall in Primary Analytic Sample

eTable 8. Association of Food Insecurity With Cognitive Outcomes in Primary Analytic Sample Excluding SNAP Receipt From Control Variables

eTable 9. Association of Food Insecurity With Dementia Risk and Memory Score in Complete Case Sample

eTable 10. Association of Food Insecurity With Dementia Risk and Memory Score in Models Not Controlling for Cognitive Outcomes in 2012

eMethods. Construction of Censoring Weights

Click here for additional data file.^{(250.2KB, pdf)}

Supplement 2.

Data Sharing Statement

Click here for additional data file.^{(17.3KB, pdf)}

References

1.Matthews KA, Xu W, Gaglioti AH, et al. Racial and ethnic estimates of Alzheimer’s disease and related dementias in the United States (2015-2060) in adults aged ≥65 years. Alzheimers Dement. 2019;15(1):17-24. doi: 10.1016/j.jalz.2018.06.3063 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.USDA Economic Research Service. Definitions of food security . Accessed October 9, 2023. https://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-u-s/definitions-of-food-security/
3.Feeding America. Hunger and food insecurity: a definition of food insecurity. Accessed October 9, 2023. https://www.feedingamerica.org/hunger-in-america/food-insecurity
4.Coleman-Jensen A. Household Food Security in the United States in 2021. US Dept of Agriculture, Economic Research Service. 2022; ERR-309. [Google Scholar]
5.Gundersen C, Ziliak J. Hunger among adults age 50-59 in 2021: an annual report. April 26, 2023. Accessed September 20, 2023. https://www.feedingamerica.org/sites/default/files/2023-04/Hunger%20Among%20Adults%20Age%2050-59%20in%202021.pdf
6.Gundersen C, Ziliak J. The health consequences of hunger among adults age 50-59 in the United States: evidence from the 1999-2016 NHANES. August 2021. Accessed September 20, 2023. https://www.feedingamerica.org/sites/default/files/2021-08/2021%20-%20Health%20Consequences%20of%20Hunger%20Among%20Adults%20Age%2050-59%20in%201999-2016_0.pdf
7.Gucciardi E, Vogt JA, DeMelo M, Stewart DE. Exploration of the relationship between household food insecurity and diabetes in Canada. Diabetes Care. 2009;32(12):2218-2224. doi: 10.2337/dc09-0823 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Laraia BA. Food insecurity and chronic disease. Adv Nutr. 2013;4(2):203-212. doi: 10.3945/an.112.003277 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Pourmotabbed A, Moradi S, Babaei A, et al. Food insecurity and mental health: a systematic review and meta-analysis. Public Health Nutr. 2020;23(10):1778-1790. doi: 10.1017/S136898001900435X [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Pereira MHQ, Pereira MLAS, Campos GC, Molina MCB. Food insecurity and nutritional status among older adults: a systematic review. Nutr Rev. 2022;80(4):631-644. doi: 10.1093/nutrit/nuab044 [DOI] [PubMed] [Google Scholar]
11.Halfon N, Hochstein M. Life course health development: an integrated framework for developing health, policy, and research. Milbank Q. 2002;80(3):433-479, iii. doi: 10.1111/1468-0009.00019 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Gregório MJ, Rodrigues AM, Graça P, et al. Food insecurity is associated with low adherence to the Mediterranean diet and adverse health conditions in Portuguese adults. Front Public Health. 2018;6:38. doi: 10.3389/fpubh.2018.00038 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Hanson KL, Connor LM. Food insecurity and dietary quality in US adults and children: a systematic review. Am J Clin Nutr. 2014;100(2):684-692. doi: 10.3945/ajcn.114.084525 [DOI] [PubMed] [Google Scholar]
14.Gottesman RF, Albert MS, Alonso A, et al. Associations between midlife vascular risk factors and 25-year incident dementia in the Atherosclerosis Risk in Communities (ARIC) cohort. JAMA Neurol. 2017;74(10):1246-1254. doi: 10.1001/jamaneurol.2017.1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Schubert CR, Fischer ME, Pinto AA, et al. Brain aging in midlife: the Beaver Dam Offspring Study. J Am Geriatr Soc. 2019;67(8):1610-1616. doi: 10.1111/jgs.15886 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Jefferson AL. Cardiac output as a potential risk factor for abnormal brain aging. J Alzheimers Dis. 2010;20(3):813-821. doi: 10.3233/JAD-2010-100081 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Jagust W, Harvey D, Mungas D, Haan M. Central obesity and the aging brain. Arch Neurol. 2005;62(10):1545-1548. doi: 10.1001/archneur.62.10.1545 [DOI] [PubMed] [Google Scholar]
18.Jefferson AL, Himali JJ, Beiser AS, et al. Cardiac index is associated with brain aging: the Framingham Heart Study. Circulation. 2010;122(7):690-697. doi: 10.1161/CIRCULATIONAHA.109.905091 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.McMichael AJ, McGuinness B, Lee J, Minh HV, Woodside JV, McEvoy CT. Food insecurity and brain health in adults: a systematic review. Crit Rev Food Sci Nutr. 2022;62(31):8728-8743. doi: 10.1080/10408398.2021.1932721 [DOI] [PubMed] [Google Scholar]
20.Na M, Dou N, Ji N, et al. Food insecurity and cognitive function in middle to older adulthood: a systematic review. Adv Nutr. 2020;11(3):667-676. doi: 10.1093/advances/nmz122 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Wong JC, Scott T, Wilde P, Li YG, Tucker KL, Gao X. Food insecurity is associated with subsequent cognitive decline in the Boston Puerto Rican Health study. J Nutr. 2016;146(9):1740-1745. doi: 10.3945/jn.115.228700 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Gao X, Scott T, Falcon LM, Wilde PE, Tucker KL. Food insecurity and cognitive function in Puerto Rican adults. Am J Clin Nutr. 2009;89(4):1197-1203. doi: 10.3945/ajcn.2008.26941 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Koyanagi A, Veronese N, Stubbs B, et al. Food insecurity is associated with mild cognitive impairment among middle-aged and older adults in South Africa: findings from a nationally representative survey. Nutrients. 2019;11(4):749. doi: 10.3390/nu11040749 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Onadja Y, Atchessi N, Soura BA, Rossier C, Zunzunegui MV. Gender differences in cognitive impairment and mobility disability in old age: a cross-sectional study in Ouagadougou, Burkina Faso. Arch Gerontol Geriatr. 2013;57(3):311-318. doi: 10.1016/j.archger.2013.06.007 [DOI] [PubMed] [Google Scholar]
25.Tong M, Tieu L, Lee CT, Ponath C, Guzman D, Kushel M. Factors associated with food insecurity among older homeless adults: results from the HOPE HOME study. J Public Health (Oxf). 2019;41(2):240-249. doi: 10.1093/pubmed/fdy063 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Smith TA, Gregory CA. Food insecurity in the United States: measurement, economic modeling, and food assistance effectiveness. Annu Rev Resour Econ. 2023;15(1):279-303. doi: 10.1146/annurev-resource-101422-090409 [DOI] [Google Scholar]
27.Ackley SF, Zimmerman SC, Brenowitz WD, et al. Effect of reductions in amyloid levels on cognitive change in randomized trials: instrumental variable meta-analysis. BMJ. 2021;372(156):n156. doi: 10.1136/bmj.n156 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Health and Retirement Study. Accessed October 9, 2023. https://hrs.isr.umich.edu/about
29.2013. Health Care and Nutrition Study (HCNS). Health and Retirement Study. Accessed October 9, 2023. https://hrsdata.isr.umich.edu/data-products/2013-health-care-and-nutrition-study-hcns
30.Economic Research Service . USDA. Household Food Security Survey Module. September 2012. Accessed October 20, 2023. https://www.ers.usda.gov/media/8271/hh2012.pdf
31.Wu Q, Tchetgen Tchetgen EJ, Osypuk TL, White K, Mujahid M, Maria Glymour M. Combining direct and proxy assessments to reduce attrition bias in a longitudinal study. Alzheimer Dis Assoc Disord. 2013;27(3):207-212. doi: 10.1097/WAD.0b013e31826cfe90 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.The Aging, Demographics, and Memory Study (ADAMS) Wave A. Health and Retirement Study. October 2009. Accessed October 9, 2023. https://hrsdata.isr.umich.edu/data-products/aging-demographics-and-memory-study-adams-wave
33.Jorm AF, Jacomb PA. The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): socio-demographic correlates, reliability, validity and some norms. Psychol Med. 1989;19(4):1015-1022. doi: 10.1017/S0033291700005742 [DOI] [PubMed] [Google Scholar]
34.Glymour MM, Avendaño M, Haas S, Berkman LF. Lifecourse social conditions and racial disparities in incidence of first stroke. Ann Epidemiol. 2008;18(12):904-912. doi: 10.1016/j.annepidem.2008.09.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Vable AM, Gilsanz P, Nguyen TT, Kawachi I, Glymour MM. Validation of a theoretically motivated approach to measuring childhood socioeconomic circumstances in the Health and Retirement Study. PLoS One. 2017;12(10):e0185898. doi: 10.1371/journal.pone.0185898 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Zuo D, Heflin CM. Cognitive impairment and supplemental nutrition assistance program take-up among the eligible older Americans. J Gerontol B Psychol Sci Soc Sci. 2023;78(1):99-110. doi: 10.1093/geronb/gbac111 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Andersen A, Rieckmann A. Using mi impute chained to fit ANCOVA models in randomized trials with censored dependent and independent variables. Stata J. 2016;16(3):650-661. doi: 10.1177/1536867X1601600306 [DOI] [Google Scholar]
38.Cole SR, Hernán MA, Margolick JB, Cohen MH, Robins JM. Marginal structural models for estimating the effect of highly active antiretroviral therapy initiation on CD4 cell count. Am J Epidemiol. 2005;162(5):471-478. doi: 10.1093/aje/kwi216 [DOI] [PubMed] [Google Scholar]
39.Robins JM, Hernán MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology. 2000;11(5):550-560. doi: 10.1097/00001648-200009000-00011 [DOI] [PubMed] [Google Scholar]
40.Ghisletta P, Spini D. An introduction to generalized estimating equations and an application to assess selectivity effects in a longitudinal study on very old individuals. J Educ Behav Stat. 2004;29(4):421-437. doi: 10.3102/10769986029004421 [DOI] [Google Scholar]
41.Hernan M, Robins J. Causal Inference: What If. Chapman & Hall/CRC Press; 2020. [Google Scholar]
42.Cui J, Qian G. Selection of working correlation structure and best model in GEE analyses of longitudinal data. Commun Stat Simul Comput. 2007;36(5):987-996. doi: 10.1080/03610910701539617 [DOI] [Google Scholar]
43.Tibshirani BERJ. An Introduction to the Bootstrap. Chapman and Hall/CRC; 1994. [Google Scholar]
44.Robins JM, Rotnitzky A. Semiparametric efficiency in multivariate regression models with missing data. J Am Stat Assoc. 1995;90(429):122-129. doi: 10.1080/01621459.1995.10476494 [DOI] [Google Scholar]
45.Tchetgen Tchetgen EJ, Glymour MM, Weuve J, Robins J. Specifying the correlation structure in inverse-probability-weighting estimation for repeated measures. Epidemiology. 2012;23(4):644-646. doi: 10.1097/EDE.0b013e31825727b5 [DOI] [PubMed] [Google Scholar]
46.Rubin DB. Multiple Imputation for Nonresponse in Surveys. Wiley; 1987. doi: 10.1002/9780470316696 [DOI] [Google Scholar]
47.Gianattasio KZ, Ciarleglio A, Power MC. Development of algorithmic dementia ascertainment for racial/ethnic disparities research in the US Health and Retirement Study. Epidemiology. 2020;31(1):126-133. doi: 10.1097/EDE.0000000000001101 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Glymour MM, Weuve J, Berkman LF, Kawachi I, Robins JM. When is baseline adjustment useful in analyses of change? an example with education and cognitive change. Am J Epidemiol. 2005;162(3):267-278. doi: 10.1093/aje/kwi187 [DOI] [PubMed] [Google Scholar]
49.Vable AM, Diehl SF, Glymour MM. Code review as a simple trick to enhance reproducibility, accelerate learning, and improve the quality of your team’s research. Am J Epidemiol. 2021;190(10):2172-2177. doi: 10.1093/aje/kwab092 [DOI] [PubMed] [Google Scholar]
50.Ziliak J, Gundersen C. The state of senior hunger in America in 2020: an annual report. May 2022. Accessed September 20, 2023. https://www.feedingamerica.org/sites/default/files/2022-05/The%20State%20of%20Senior%20Hunger%20in%202020_Full%20Report%20w%20Cover.pdf
51.Ziliak JP, Gundersen C, Vaudin A. Introduction to senior hunger: special issue. Appl Econ Perspect Policy. 2023;45(1):221-233. doi: 10.1002/aepp.13347 [DOI] [Google Scholar]
52.Effects of the Great Recession on older Americans’ health and well-being . Population Reference Bureau. Accessed October 9, 2023. https://www.prb.org/resources/todays-research-aging-great-recession/
53.Coleman-Jensen A, Rabbitt MP, Gregory C, Singh A. Household Food Security in the United States in 2014. US Dept of Agriculture, Economic Research Service; 2015, ERR-194. [Google Scholar]
54.Martinez SM, Webb K, Frongillo EA, Ritchie LD. Food insecurity in California’s public university system: what are the risk factors? J Hunger Environ Nutr. 2018;13(1):1-18. doi: 10.1080/19320248.2017.1374901 [DOI] [Google Scholar]
55.Coleman-Jensen A, Rabbitt MP, Gregory CA, Singh A. Statistical supplement to household food security in the United States in 2020. USDA Miscellaneous. September 2, 2021. Accessed October 9, 2023. https://ideas.repec.org//p/ags/usdami/313486.html
56.Mayeda ER, Glymour MM, Quesenberry CP, Whitmer RA. Inequalities in dementia incidence between six racial and ethnic groups over 14 years. Alzheimers Dement. 2016;12(3):216-224. doi: 10.1016/j.jalz.2015.12.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Gilsanz P, Mayeda ER, Glymour MM, Quesenberry CP, Whitmer RA. Association between birth in a high stroke mortality state, race, and risk of dementia. JAMA Neurol. 2017;74(9):1056-1062. doi: 10.1001/jamaneurol.2017.1553 [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Barnes LL, Wilson RS, Everson-Rose SA, Hayward MD, Evans DA, Mendes de Leon CF. Effects of early-life adversity on cognitive decline in older African Americans and Whites. Neurology. 2012;79(24):2321-2327. doi: 10.1212/WNL.0b013e318278b607 [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Gundersen C, Ziliak JP. Food insecurity and health outcomes. Health Aff (Millwood). 2015;34(11):1830-1839. doi: 10.1377/hlthaff.2015.0645 [DOI] [PubMed] [Google Scholar]
60.Momtaz YA, Haron SA, Hamid TA, Ibrahim R, Masud J. Does food insufficiency in childhood contribute to dementia in later life? Clin Interv Aging. 2014;10:49-53. doi: 10.2147/CIA.S69220 [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Lu P, Kezios K, Lee J, Calonico S, Wimer C, Zeki Al Hazzouri A. Association between supplemental nutrition assistance program use and memory decline: findings from the Health and Retirement Study. Neurology. 2023;100(6):e595-e602. doi: 10.1212/WNL.0000000000201499 [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Lauffer S, Vigil A. Trends in Supplemental Nutrition Assistance Program Participation Rates: Fiscal Year 2016 to Fiscal Year 2018. US Dept of Agriculture; 2021. [Google Scholar]
63.Gundersen C, Seligman H. How can we fully realize SNAP’s health benefits? N Engl J Med. 2022;386(15):1389-1391. doi: 10.1056/NEJMp2200306 [DOI] [PubMed] [Google Scholar]
64.Mathematica. Reaching those in need: estimates of state supplemental nutrition assistance program participation rates in 2015. January 19, 2018. Accessed October 9, 2023. https://www.mathematica.org/publications/reaching-those-in-need-estimates-of-state-snap-participation-rates-in-2015
65.Kenney EL, Soto MJ, Fubini M, Carleton A, Lee M, Bleich SN. Simplification of Supplemental Nutrition Assistance Program recertification processes and association with uninterrupted access to benefits among participants with young children. JAMA Netw Open. 2022;5(9):e2230150. doi: 10.1001/jamanetworkopen.2022.30150 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials