key: cord-1029390-88vj8czi
authors: Choi, Shinae L.; Men, Fei
title: Food Insecurity Associated with Higher COVID-19 Infection in Households with Older Adults
date: 2021-09-11
journal: Public Health
DOI: 10.1016/j.puhe.2021.09.002
sha: ccf49abba313589207c7a6d73beb280c91ad651d
doc_id: 1029390
cord_uid: 88vj8czi

Objectives As a well-documented social determinant of health, food insecurity may be associated with COVID infection in households with older adults. We examined whether older adults were vulnerable to COVID infection during the early pandemic if they were food insecure versus food secure. Study design Cross-sectional study using a nationally representative population-based survey of US older adults. Methods We used a random subsample of Health and Retirement Study (HRS) drawn in June 2020 (N = 3,212). We compared the odds of reporting COVID infection in a household, COVID infection and mortality among acquaintances, and respondent’s perceived fair or poor health across household food insecurity status resulted from financial or nonfinancial barriers. Baseline health and socioeconomic circumstances were adjusted in the models. Results Results showed higher COVID infection rate among food-insecure households than among their food-secure counterparts during the pandemic. Food insecurity due to nonfinancial obstacles was associated with greater likelihood of COVID infection both within respondents’ households (adjusted odds ratio [aOR] = 1.73, 95% confidence interval [CI]: 1.03–2.90) and among their acquaintances (aOR = 1.32, 95% CI: 1.05–1.65). Food insecurity caused by both nonfinancial and financial constraints was associated with twice the likelihood of knowing someone who died from COVID than their food-secure counterparts (aOR = 2.14, 95% CI: 1.27–3.61). Conclusions Food insecurity driven by nonfinancial constraints played an important role in the ongoing pandemic among US older adults. Policies addressing COVID need to recognize the vulnerability of food-insecure older adults beyond lack of monetary resources.

The COVID-19 (COVID hereafter) pandemic has caused great suffering around the globe, especially among older adults. Older adults are at particularly high risk of COVID infection, hospitalization, and mortality due to their weakened immunity and other health vulnerabilities. 1, 2 Poor nutrition and existing chronic conditions could further magnify the effect of COVID. 2, 3 Food insecuritythe limited access to adequate food due to lack of money and other resourcesis a serious public health issue in the developed countries, [4] [5] [6] affecting one in nine US households and 7.2 percent of households with elderly. 4 The mass unemployment and mobility restrictions due to the COVID have exacerbated food insecurity. [7] [8] [9] [10] The share of US households experiencing food insecurity was estimated to increase to one in seven in 2020. 10 Among seniors 60 years and older, food insufficiencya measure of severe food insecurityincreased by 75 percent during the early stage of the pandemic. 11 Food insecurity has been associated with a number of negative health outcomes among older adults, [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] including chronic conditions, infectious diseases, and health care utilization. 13, 15, 17, 19, 21, 22 The food insecurity-health association has been robust and strong, above and beyond the mediation of income and assets. 16 Food insecurity may heighten risk of COVID infection through weakened immune system 2, 23, 24 and increased viral exposure. 8 Food-insecure adults still in the labour force may disproportionately work in crowded environments with little to no social distancing. 8, 25 Nonfinancial constraints like closure of or restricted access to food stores may hamper access to food, forcing older adults to visit crowded stores and food charities, use public transit, and increase their exposure to virus. Moreover, given the correlation between strength of social networks and COVID spread, 26 it is likely that older adults from food-insecure households J o u r n a l P r e -p r o o f FOOD INSECURITY AND COVID- 19 INFECTION 4 know other similarly disadvantaged people vulnerable to COVID and, 27 as a consequence, have more acquaintances infected by and dead from COVID than their food-secure counterparts.

According to the social ecological model 28 and theories on social determinants of health, 29 socioeconomically disadvantaged households should heighten the chance of contracting COVID and knowing others infected by the virus due to neighbourhood adversities, health vulnerability, and economic instability among other household and community factors. 30 A few studies have associated aggregate measures of income with COVID infection and mortality rate in the US; [31] [32] [33] however, they may have masked important inter-household differences in COVID infection. Income and food insecurity are related yet fundamentally different measures of economic well-being. 34 Compared to income, food insecurity more accurately captures experience of material hardships and appears to be a much more powerful predictor of health. 16 Past studies suggest food insecurity as an independent predictor of adverse health outcomes even among lower-income older adults. 18, 35 The official measure of food insecurity developed by the US Department of Agriculture (USDA) stressed lack of money as its underlying cause. 4 While financial constraints play a major role in determining food insecurity, the USDA measure may have overlooked common nonfinancial constraints among older adults such as health-related mobility issues. 36 This issue may be especially salient during the early stage of the pandemic, 7 when normal operation of stores and transport systems was disrupted by lockdowns and other restrictive measures. Living in food deserts with few retailers of nutritious food could further hamper food access. 37 Indeed, perceived long distance to grocery store has been associated with food insecurity among older adults, 38 suggesting the need to look beyond income when studying their food insecurity.

The existing body of literature points to the plausible variation in the COVID infection risk by older adults' food insecurity status, with potential distinction between food insecurity driven by financial scarcity and nonfinancial barriers. Relative to the food-secure, food-insecure older adults may also be more likely to know someone with COVID due to their stronger social connections with the vulnerable population. We capitalize on a recent nationally representative population-based health survey to evaluate the associations of household food insecurity with COVID infection and health status among U.S. older adults.

Data are from a random 25 percent subsample of the Health and Retirement Study (HRS) participants who completed a COVID-19 module introduced in June 2020. 39 The HRS is a nationally representative longitudinal survey of U.S. adults ages 51 and older and their spouses of any age. Detailed survey design can be found in Sonnega and colleagues. 40 The sample included 3,266 persons. Following the convention of past geriatric studies, [41] [42] [43] we excluded 54 respondents younger than 50 to reach our analytic sample with 3,212 persons aged 50 and older.

COVID infection in respondent's household. Self-reported COVID diagnosis of respondent and respondent's coresidents were assessed with two questions: "Have you had or do you now have COVID-19, the disease caused by the novel coronavirus?"; "Has anyone in your household other than you been diagnosed with COVID-19?" Given the high likelihood of intrahousehold COVID spread, 44-46 a single dichotomous outcome was created to indicate COVID infection of anyone in respondent's household.

J o u r n a l P r e -p r o o f COVID infection and death of respondent's acquaintances. We created two binary outcomes for COVID infection and deaths from COVID of respondent's acquaintances, respectively, based on the questions "Has anyone else you know been diagnosed with COVID-19?" and "Has anyone you know died from COVID-19?" (yes; no).

Low self-rated health. Self-rated health status was assessed using a single item, "Would you say your health is excellent, very good, good, fair, or poor?" We dichotomized this variable, where "poor or fair" was coded 1, and "good, very good, or excellent" was coded 0. Collapsing "poor" and "fair" categories has been a standard approach in studying self-rated health. [47] [48] [49] [50] 

Food insecurity during the COVID pandemic. The survey evaluated whether a respondent: didn't have enough money to buy food; and had trouble buying food even though had money (yes; no). The first item stresses the financial constraints, while the second emphasizes the nonfinancial aspects. We created four mutually exclusive categories: food-secure (reference); finance-driven food insecurity only; nonfinance-driven food insecurity only; and both finance-and nonfinance-driven food insecurity. Also, we created a dummy variable indicating food insecurity status: food-secure (reference); food-insecure (i.e. either finance-or nonfinancedriven food insecurity).

We controlled for factors that may confound the association between food insecurity and COVID infection. 51, 52 Demographic and socioeconomic controls include age, sex, race/ethnicity, marital status, number of household members, living arrangement, education, employment status, total household income, liquid wealth, and non-liquid wealth. Health covariates included depressive symptoms (range: 0 to 8) assessed using a subset of items from the Center for J o u r n a l P r e -p r o o f Epidemiologic Studies Depression scale 53 and number of chronic health conditions (high blood pressure, diabetes, cancer, lung disease, heart problems, stroke, psychiatric problems, or arthritis, range: 0 to 8).

We first compared the four food insecurity categories on all variables included in the analysis, using one-way analysis of variance (ANOVA) tests. We estimated multivariate logistic for health covariates due to collinearity. Additionally, we conducted several sensitivity analyses to assess the robustness of the results. Previous literature suggests that household income and liquid wealth are the common drivers of food insecurity. 34 Thus, we experimented with models without those two in the sensitivity tests.

Descriptive statistics for the sample characteristics by food insecurity category are presented in Table 1 . We contrasted the COVID infection rates, and the health, demographic, and socioeconomic characteristics of the four food insecurity categories by conducting ANOVA and J o u r n a l P r e -p r o o f Tukey's post-hoc tests; the right-hand column denoted statistically significant contrasts between specific pairs of food insecurity categories (Table 1) . During the early months of the COVID pandemic, 3.2 percent of study participants reported that they or other members in the household were diagnosed with COVID, with significantly different rates across the food insecurity categories. The results showed higher COVID infection rate among food-insecure households, particularly nonfinance-driven food-insecure households than among their food-secure counterparts. Nearly two in five participants reported that anyone else they know were diagnosed with COVID, with similar rates across the four food insecurity categories. One-fifth reported that anyone they know died from COVID. Older adults with both financial and nonfinancial obstacles towards food security were twice as likely as food-secure persons to report that anyone they know died from COVID (39.2 vs. 19.0%). The COVID death rate was not different between food-secure people's acquaintances and acquaintances of people experiencing either financedriven or nonfinance-driven food insecurity. Compared to all other food insecurity groups, both finance-and nonfinance-driven food-insecure older adults reported significantly higher COVID death rates of their acquaintances during the pandemic.

[ Table 1 about here] 1.23 (p < 0.05, 95% CI = 1.00, 1.52, Panel C) times that of food-secure households, respectively.

The odds ratios attenuated slightly and were no longer statistically significant after adjusting for health, demographic, and socioeconomic factors.

In Specification II, we used the four-category food insecurity variable. The association attenuated yet persisted in the adjusted model (OR = 2.14, p < 0.01, 95% CI = 1.27, 3.61, Panel C). Finance-driven food insecurity alone was not related to COVID outcomes, with or without confounders adjustment.

[ Table 2 about here]

Panel D in Table 2 presents logistic regression predicting perceived health status (poor/fair vs. good/very good/excellent) among older adults. Compared with food-secure J o u r n a l P r e -p r o o f persons, food-insecure persons were more likely to report poor or fair health status (OR = 1.97, p < 0.001, 95% CI = 1.65, 2.35) in Specification I. This association still remained significant after adjusting for demographic and socioeconomic characteristics (OR = 1.51, p < 0.001, 95% CI = 1.24, 1.84). In Specification II, food-insecure persons were more likely to report poor or fair health status than food-secure persons regardless of the cause of food insecurity. Food insecurity due to both financial and nonfinancial obstacles was associated with the worst health status. The magnitudes of these significant associations were attenuated yet persisted at the p < 0.001 level after covariates adjustment, with the exception of finance-driven food-insecurity.

We reran the adjusted model without accounting for total household income and liquid wealth, which yielded virtually the same results as those from the main analyses (Table 3) .

[ Table 3 about here]

Using a nationally representative sample of U.S. older adults, we found that food insecurity due to nonfinancial obstacles was associated with poorer health and greater likelihood of COVID infection both within respondents' households and among their acquaintances. Food insecurity caused by both nonfinancial barriers and financial constraints was associated with twice the likelihood of knowing someone who died from COVID. Our study is among the first illustrating the differential COVID infection risk across households of different food insecurity status while contributing to the growing body of literature that links food insecurity to poor health outcomes among older adults.

That food-insecure older adults showed 50% higher odds of reporting poor or fair health compared to their food-secure counterparts is in line with prior evidence. 12 However, only food insecurity driven by nonfinancial factors was relevant to COVID infection while food insecurity J o u r n a l P r e -p r o o f due to financial and nonfinancial causesbut not either cause alonewas associated with COVID mortality among respondents' acquaintances. These findings suggest that food insecurity is a correlate of COVID and health in general among older adults insofar as nonfinancial factors are concerned. One possible explanation is that nonfinance-driven food insecurity indicates mobility issues caused by disability or chronic conditions. People of poor health are more likely to contract COVID and experience disease symptoms than those who are in better health. 3 There is also emerging evidence that comorbidity such as hypertension and diabetes are risk factors for death from COVID. 54 Food insecurity and the consequent inadequate nutrition, heightened inflammation, and psychological stress could only further inflate the risk of COVID infection. 2, 21, 23, 24 Another possibility is that policies restricting mobility and interrupting business may have forced many to shop for food in more crowded environments (e.g., food pantries and grocery stores with reduced hours and increased demand), 55 increasing their risk of viral infection.

The lack of association between finance-driven food insecurity and COVID infection was somewhat unexpected and stands in contrast to the prior findings connecting poor health to food insecurity presumably caused by financial constraints. 21, 22 This might be because access to COVID testing remained limited during early pandemic in neighbourhoods with higher social vulnerability, 56 where food insecurity due to financial constraints may be particularly prevalent.

Lack of internet access and technological competency may pose additional barriers to COVID testing among financially deprived older adultsmore so compared to their better off counterparts, further exacerbating undercount of COVID cases among the former. Stigma about the virus that disproportionately affect the lower-resource food-insecure people could also lead to underreporting of COVID cases. Indeed, COVID was associated with substantial stigma during the early pandemic; 57 those with greater financial constraints may seek to minimize such stigma J o u r n a l P r e -p r o o f through underreport. Moreover, the food insecurity questions used in our study differed significantly from the USDA questionnaire used in previous research, which would have led to the result discrepancies. More detailed comparison of the two questionnaires is warranted to better understand the findings we observed.

Food-insecure persons due to both financial and nonfinancial obstacles are significantly more likely than food-secure persons to report that anyone else they know died from COVID during the pandemic. One explanation is that food-insecure people with low financial resources and nonfinancial obstacles such as health issues are the ones with the greatest socioeconomic disadvantages and therefore may know other disadvantaged people with substantial exposure to COVID and least resource to cope with its consequences. 27 Economically disadvantaged people are more likely to live in overcrowded housing with minimal social distancing and hold publicfacing front-line jobs that do not provide opportunities to work from home and put them at high risk of infection. 58 In addition, chronic conditions may be more prevalent among acquaintances of food-insecure adults, which, along with social disadvantages, can further increase exposure to and mortality from the COVID. 3, 54, 58 Income supplement has been associated with reduced food insecurity. In the US, receipt of unemployment insurance was associated with large reduction in food insecurity during the pandemic. 59 Enrollment in Supplemental Nutrition Assistance Program (SNAP) has been shown to attenuate the association between very low food security and poor physical health among older adults, though the stigma attached to SNAP participation may offset its benefit on mental health. 60 Food security barriers seemingly unrelated to household finance such as disability and mobility restrictions may also be partially solved by social capital (e.g., food sharing from family and friends) and adequate income (e.g., paying for food delivery). 61, 62

Even prior to the COVID pandemic, researchers have raised concerns regarding local food environments for disadvantaged residents, who disproportionately relied on nearby stores with more limited food items. 63 Low-income households may visit several food stores to get access to the most affordable food, often traveling long distances to acquire food. 63 Although many innovative strategies have been adopted to mitigate the devastating impact of pandemic on food insecurity for children from low-income households (e.g., delivering food to families by using school busses driving along their regular routes, 64 grab-and-go food pick-up for multiple days' worth of meals at one time), 65 only few strategies have been implemented to fulfil daily nutritional needs and alleviate food insecurity challenges for older adults. It is important to ensure that the nutritional needs of vulnerable population are met in order to reduce exacerbating disparities in health outcomes, 65 particularly older adults with functional and cognitive impairments. Policies addressing COVID need to recognize the vulnerability of food-insecure older adults beyond lack of monetary resources.

We've seen significant growth of the online grocery delivery sector during the pandemic, 66 yet it is prominent that both a lack of wireless broadband access or devices at home and unfamiliarity with computer training have been barriers among older adults. 67, 68 Ensuring resource (e.g., Internet access, devices) and digital literacy training to bridge and close the digital divide among older adults should be also a policy priority along with the continued availability of essential food to vulnerable communities.

Several limitations exist in the current study. We examined self-reported health outcomes (i.e., COVID infection and perceived health status), which could suffer from underreport due to stigma and unawareness. It is possible that the respondents' acquaintances may have died from causes other than COVID. With regard to the COVID infection and death of respondents' acquaintances variables, the study questions did not specify whether the acquaintances were friend, relatives, neighbors, or a wider range of social networks, which warrant greater distinction in future investigation.

Additionally, nonfinance-driven food insecurity was measured by a single question without information on why food-insecure adults were unable to buy food even though they had money. Possible reasonswhich we could not discerninclude respondent's mobility constraints such as functional impairments and transportation limitations and social distancing policies such as the lockdown and the grocery stores' altered hours. We acknowledge that the differences between "financial" and "nonfinancial" barriers are not always clear-cut and that certain nonfinancial barriers could in theory be attenuated through financial means (e.g., spending more to order food delivery if allowed). However, we had to restrict our definitions of "financial" and "nonfinancial" barriers to the subjective interpretation of our respondents due to data limitations. We could not specify geographical residence of respondents due to data limitations. Thus, we acknowledge that a fuller understanding of geographic effects (e.g., food deserts) is necessary in the future research. Our sample consists of US households with older adults and therefore does not generalize to other countries or age groups.

Our results indicate that nonfinance-driven food insecurity during the pandemic is significantly associated with poorer health and higher likelihood of COVID infection both within respondents' households and among their acquaintances even after adjusting for potential confounders. Food insecurity caused by both nonfinancial barriers and financial constraints was associated with twice the likelihood of knowing someone who died from the COVID. Our J o u r n a l P r e -p r o o f Exponentiated betas (odds ratios) and confidence intervals are presented; Covariates included health conditions, demographic, and socioeconomic characteristics for Panels A, B, and C. Panel D adjusted for demographic and socioeconomic characteristics but did not adjust for health covariates; Statistical significance denoted as *** p < .001; ** p < .01; * p < .05.

J o u r n a l P r e -p r o o f Exponentiated betas (odds ratios) and confidence intervals are presented; Model 1 did not adjust for total household income; Model 2 did not adjust for total household income and liquid wealth; Statistical significance denoted as *** p < .001; ** p < .01; * p < .05.

Race, ethnicity, and age trends in persons who died from COVID-19-United States

Nutrition, immunity and COVID-19

COVID-19 and Older Adults: What We Know

Household food security in the United States in 2019: United States Department of Agriculture

Household food insecurity in Canada

Changing probability of experiencing food insecurity by socioeconomic and demographic groups during the COVID-19 pandemic in the UK. medRxiv

The early food insecurity impacts of COVID-19

Food insecurity amid the COVID-19 pandemic: Food charity, government assistance and employment

Food insecurity and COVID-19: Disparities in early effects for US adults

The impact of coronavirus on food insecurity in 2020 & 2021. 2021. J o u r n a l P r e

Food hardship during the Covid-19 pandemic and Great Recession

Food insecurity and health outcomes

Conceptual framework for understanding the bidirectional links between food insecurity and HIV/AIDS

Food insecurity: A key social determinant of health for older adults

Food insecurity and health care utilization among older adults in the United States

Food insecurity, chronic pain, and use of prescription opioids. SSM-Population Health

Association between household food insecurity and mortality in Canada: A population-based retrospective cohort study

Severe food insecurity associated with mortality among lower-income Canadian adults approaching eligibility for public pensions: A population cohort study

Food insecurity is associated with higher health care use and costs among Canadian adults: Study examines the association of food

The ecology of human development: Experiments by nature and design

Impact of social determinants of health on the emerging COVID-19 pandemic in the United States

Nutrition interventions in women in low-income groups in the UK. Proceedings of the Nutrition Society

Inequality and the coronavirus: Socioeconomic covariates of behavioral responses and viral outcomes across US counties: National Bureau of Economic Research2020

Mobility network models of COVID-19 explain inequities and inform reopening

Coronavirus infections and deaths by poverty status: The effects of social distancing

The economics of food insecurity in the United States

Food insecurity is associated with chronic disease among low-income NHANES participants

Food insecurity, comorbidity, and mobility limitations among older US adults: Findings from the Health and Retirement Study and Health Care and Nutrition Study

Disparities and access to healthy food in the United States: A review of food deserts literature

Elderly perception of distance to the grocery store as a reason for feeling food insecurity-Can food policy limit this?

2020 HRS COVID-19 Project data description and usage: The Institute for Social Research

Cohort profile: The health and retirement study (HRS)

Association of social deprivation with cognitive status and decline in older adults

Inequalities in elevated depressive symptoms in middle-aged and older adults by rural childhood residence: The important role of education

The influence of multimorbidity on leading causes of death in older adults with cognitive impairment

Assessing the effects of metropolitanwide quarantine on the spread of COVID-19 in public space and households

High household transmission of SARS-CoV-2 in the United States: Living density, viral load, and disproportionate impact on communities of color. medRxiv

Transmission of SARS-COV-2 infections in households-Tennessee and Wisconsin

The relationship between frailty and diabetes: An investigation of self-rated health, depression symptoms and quality of life in the Study of Health Aging and Retirement in Europe

Decline in health for older adults: five-year change in 13 key measures of standardized health

Changes in health, functional performance and activity predict changes in self-rated health: A 10-year follow-up study in older people

Do welfare regimes moderate cumulative dis/advantages over the life course? Cross-national evidence from longitudinal SHARE data

Predictors of households at risk for food insecurity in the United States during the COVID-19 pandemic

Predictors of food insecurity among older adults in the United States

Government capital, intimate and community social capital, and food security status in older adults with different income levels

Urban food environments and residents' shopping behaviors

Schools race to feed students amid coronavirus closures

Feeding low-income children during the COVID-19 pandemic

Food supply chains during the COVID-19 pandemic

Impact of the digital divide in the age of COVID-19

Digital health equity as a necessity in the 21st century cures act era