Abstract
Background
Upstream socioeconomic circumstances including food insecurity and food desert are important drivers of community-level health disparities in cardiovascular mortality let alone traditional risk factors. The study assessed the association between differences in food environment quality and cardiovascular mortality in US adults.
Design
Retrospective analysis of the association between cardiovascular mortality among US adults aged 45 and above and food environment quality, measured as the food environment index (FEI), in 2615 US counties. FEI was measured by equal weights of food insecurity (limited access to a reliable food source) and food desert (limited access to healthy food), ranging from 0 (worst) to 10 (best). Age-adjusted cardiovascular mortality rates per 100,000 adults aged 45 and above in the calendar year 2017–2019. County-level association between CVD mortality rate and FEI was modeled using generalized linear regression. Data were weighted using county population.
Result
Median CVD deaths per 100,000 population were 645.4 (IQR 561.5, 747.0) among adults aged 45 years and above across US counties in 2017–2019. About 12.8% (IQR 10.7%, 15.1%) of residents were food insecure and 6.3% (IQR 3.6%, 9.9%) were living in food desert areas. Comparing counties by FEI quartiles, the CVD mortality rate was higher in the least healthy FE counties (704.3 vs 598.6 deaths per 100,000 population) compared to the healthiest FE counties. One unit increase in FEI was associated with − 12.95 CVD deaths/100,000 population. In the subgroup analysis of counties with higher income inequality, the healthiest food environment was associated with 46.4 lower CVD deaths/100,000 population than the least healthy food environment. One unit increase in FEI in counties with higher income inequality was associated with a fivefold decrease in CVD mortality difference in African American counties (− 18.4 deaths/100,000 population) when compared to non-African American counties (− 3.63 deaths/100,000 population).
Conclusion
In this retrospective multi-county study in the USA, a higher food environment index was significantly associated with lower cardiovascular mortality.
Supplementary Information
The online version contains supplementary material available at 10.1007/s11606-023-08335-9.
KEY WORDS: cardiovascular disease, food environment, food insecurity, food desert, income inequality, social determinants of health
INTRODUCTION
Broadly defined, food insecurity is “a lack of consistent access to enough food for an active, healthy life.”1 In 2019, about 10.6% of US households were food insecure.2 The COVID-19 pandemic exacerbated shortcomings of the country’s struggle with food security in the context of a volatile supply chain, financial insecurity, and employment uncertainty. To a large extent, food insecurity is more prevalent in low-income families who also face housing affordability, disadvantages resulting from structural racism, chronic or acute health problems, and healthcare access barriers.3 While individual-level food preferences are influenced by these fundamental social determinants and food environment (FE), living in areas devoid of healthy food options also known as “food desert” additionally impairs healthy eating behaviors, leading to poor health outcomes.4,5
Compromised dietary intake due to food deserts and food insecurity may lead to adverse metabolic effects and increase cardiovascular disease (CVD) risk.6,7 A growing body of literature already indicates that food insecurity is associated with increased CVD risk and related factors, including hypertension, diabetes, and high cholesterol.8–11 Additionally, food insecurity may increase CVD risk by activating the stress response, triggering harmful coping behaviors, medication nonadherence, and reduced ability to manage chronic conditions.6,12,13
Food insecurity (FI) and food deserts (FD) have been increasingly recognized as important environmental contributors to individual health as well as a potential target for community-level interventions to improve CVD health outcomes.14,15 However, existing literature on CVD outcome variation has historically focused on individual-level differences in behavior and healthcare factors, rather than environmental factors.16–18 Moreover, past studies are limited by using nominal indicators of food insecurity rather than a comprehensive measure of the food environment (FI and FD).9,19 Thus, these studies may not have detected important differences in health outcomes driven by incremental differences in food insecurity. Recent research has demonstrated the significant contribution of community-level risk factors to CVD outcomes.20,21 For instance, nearly 75% of the variation in CVD mortality across US counties is explained by county-level risk factors, such as median household income and other socioeconomic factors.20 Despite this growing evidence, the majority of interventions to reduce CVD mortality take individual-level or healthcare system factors into account, and fail to acknowledge community-level social determinants of health.22,23
Therefore, in this study, we aim to address two questions. First, how does the county-level food environment, encompassing both food insecurity and food deserts, vary by demographics, socioeconomic, and healthcare characteristics at the county level? Second, is CVD mortality associated with a poor food environment, after accounting for county-level demographics, socioeconomic and environmental factors, healthcare characteristics, and behavioral risk factors?
METHOD
CVD Age-Adjusted Mortality Rate
The study used county-level CVD age-adjusted mortality rates (AAMRs) among adults aged 45 years and above from the publicly available CDC Wide-Ranging Online Data for Epidemiologic Research (CDC-WONDER) database for 2017–2019. The data is based on death certificate records collected by the National Center for Health Statistics.24,25 County-level CVD mortality rates were defined as the number of deaths per 100,000 individuals with diagnosed diseases of the circulatory system with International Classification of Diseases, Tenth Revision (ICD-10) codes I00–I78. The age adjustment was based on the 2000 US Census standard population. We included counties with at least 10 reported deaths due to CVD (n = 3139). Of 3139 counties, 524 counties were dropped because of missing data across all the selected county features. The final sample included 2615 counties.
County-Level Food Environment Index (FEI)
The county-level food environment is measured using a county-level food environment index (FEI) that accounts for both food insecurity and food desert derived from the County Health Rankings compilation (CHR).26 The index ranges from a scale of 0 (worst) to 10 (best) and equally weights these two measures. Food insecurity was measured as the percentage of the population who did not have access to a reliable source of food during the past year. Limited access to healthy food referred to as a “food desert” estimates the percentage of the population with income less than or equal to 200 percent of the federal poverty threshold for the family and live at least 1 mile away from a grocery store in nonrural or 10 miles away in rural areas. Food insecurity estimates were derived from the 2017 Map the Meal Gap (MMG) study conducted by Feeding America,27 which uses US Census American Community Survey (ACS) and Current Population Survey (CPS) data, and Bureau of Labor Statistics data. Food desert estimates were based on the data from the US Department of Agriculture Food Environment Atlas, 2015. We further classified county-level FEI into quartiles (1st—healthiest food environment, FEI 7.5–10.0; 4th- least healthy food environment, FEI 0–2.5) within each state.28
Other County Characteristics
We included county-level demographics, socioeconomic and environmental, access to healthcare, and behavioral variables that were assessed contemporaneously before 2017. In some specifications, we included the GINI coefficient as a comparative structural measure of financial disparity within each county since the FD measure of FEI is primarily focused on lower-income populations. GINI coefficient represents income equality based on the median income, calculated by taking the ratio of income at the 80th percentile to income at the 20th percentile for each county. A higher coefficient indicates lower equality between the top and bottom ends of the income spectrum for a county. The GINI coefficient was obtained from the County Health Ranking (CHR) 2021 database.29 Conceptual framework for the study, additional variables, and data sources are outlined in eMethod 1 and eTable 1 of Supplement.
Statistical Analysis
Study variables are reported with median and interquartile ranges. We also compared the distribution of county characteristics by food environment index quartiles (1st—healthiest food environment, 7.5–10.0; 4th—least healthy food environment, 0–2.5). The Kruskal–Wallis test was used to compare the characteristics between the food environment quartiles. We used generalized linear models (GLM) sequentially to estimate the association between Food Environment and CVD mortality. Akaike information criterion (AIC) and the Bayesian Information Criterion (BIC) were used to evaluate the performance of the regression models. Regression modeling and additional method are outlined in the Supplement.
To further explore the heterogeneity of the income inequality and food environment on CVD mortality, we used the same regression model but conducted a subgroup analysis by GINI coefficient quartiles to compare counties with lower income inequalities (Q1) to higher income inequalities (Q4). While this approach controls for heterogeneity within counties by income, the distribution of people in a county may have a strong influence on the phenomenon of income inequality. We, therefore, included an interaction term between the food environment index and counties with a higher proportion of African Americans. Data has been weighted using county population in statistical analyses. All analyses were conducted using STATA v17.
RESULT
A total of 2615 US counties were included in the analysis. The largest concentration of counties with the least healthy food environment and increased CVD mortality were clustered across the southeast and northeast central regions of the USA (Fig. 1).
Figure 1.
Bivariate association between food environment index and cardiovascular disease mortality in the USA, 2017–2019.
Between 2017 and 2019, the median CVD deaths per 100,000 population were 645.4 (IQR 561.5, 747.0) among adults aged 45 years and above (eTable 2). Among counties, a median of 2.0% (IQR 0.6%, 9.1%) residents were African Americans aged 45 years and above, 58.3% (IQR 50.2%, 66.5%) with some college education, and had a median income inequality ratio of 40.4 (IQR 40.0, 40.9). About 10.5% (IQR 8.6%, 12.8%) residents were burdened by higher housing costs, 12.8% (IQR 10.7%, 15.1%) were food insecure, 6.3% (IQR 3.6%, 9.9%) were living in food desert areas, and 13.2% (IQR 9.6%, 17.1%) were benefiting from food stamp/SNAP program. By healthcare features, a median of 10.1% (IQR 7.1%, 13.6%) residents were uninsured, while 16.5% (IQR 12.4%, 21.4%) were insured by Medicaid or other public insurance. A median of 50.3 (IQR 33.7, 72.3) PCPs, 55.8 NPs, 23.6 PAs, and 11 healthcare centers were available per 100,000 population. The median annual screening rate of routine glycated tests for diabetes was 87.5% (IQR 84.3%, 89.9%) among Medicare enrollees with a history of diabetes.
Comparing counties by FEI quartiles, the CVD mortality rate was higher in the least healthy FE counties (704.3 vs 598.6 deaths/ 100,000 population) compared to the healthiest FE counties (Table 1). By demographic characteristics, the proportion of African Americans was greater in the least healthy FE (3.9% vs.1.5%) than in the healthiest FE, whereas the Asian and Hispanic population distribution was similar across the FEI quartiles. Of FEI quartiles, median household income was greater ($59,938) in the healthiest FE counties while income inequality (GINI coefficient) was greater in the least healthy FE counties. Interestingly, housing cost burden, overcrowding, and air quality remained comparable across FE quartiles of counties. Among healthcare characteristics, counties with the least healthy FE had a larger population with Medicaid coverage (20.4% vs 13.1%) and a greater number of NPs (65.3% vs. 49.5%), PAs (25.9% vs. 20.9%), and healthcare centers (16.9 vs 7.9 per 100,000 population) than the healthiest FE counties. The proportion of PCPs (51.9% vs 48.9%) was marginally higher in the least healthy FE counties than healthiest FE counties. Whereas, diabetes (10.6% vs 12.6%), obesity (32.1% vs. 34.4%), and smoking (15.6% vs. 19.0%) rates were lower in healthiest FE counties than in the least healthy FE counties.
Table 1.
Distribution of Characteristics by Food Environment Index Quartiles
| Factor | FEI quartile 1 Healthiest food environment (N = 749) Median (IQR) |
FEI quartile 2 (N = 664) Median (IQR) |
FEI quartile 3 (N = 614) Median (IQR) |
FEI quartile 4 Least healthy food environment (N = 588) Median (IQR) |
p value |
|---|---|---|---|---|---|
| CVD Mortality rate per 100,000 population aged ≥ 45 years | 598.6 (524.3, 687.2) | 641.4 (563.7, 741.7) | 664.2 (572.6, 751.6) | 704.3 (611.1, 817.8) | < 0.001 |
| Demographic Composition | |||||
| Population | 37,349.0 (17,209.0, 111,876.0) | 33,830.0 (16,378.0, 69,087.0) | 33,304.0 (15,902.0, 90,382.0) | 27,776.0 (13,141.0, 77,423.5) | < 0.001 |
| % Aged ≥ 45 years | 45.0 (40.8, 48.1) | 46.1 (42.3, 49.3) | 45.2 (41.5, 48.9) | 45.0 (41.0, 48.8) | < 0.001 |
| % Female aged ≥ 45 years | 46.6 (42.2, 49.7) | 47.9 (44.0, 51.0) | 47.1 (43.1, 50.8) | 47.3 (43.0, 50.9) | < 0.001 |
| % African American, NH aged ≥ 45 years | 1.5 (0.5, 4.8) | 1.8 (0.5, 7.8) | 2.1 (0.5, 10.2) | 3.9 (0.7, 25.8) | < 0.001 |
| % Asian, NH aged ≥ 45 years | 0.6 (0.4, 1.5) | 0.5 (0.4, 1.1) | 0.6 (0.4, 1.2) | 0.6 (0.4, 1.1) | 0.002 |
| % Hispanic aged ≥ 45 years | 0.8 (0.6, 1.1) | 0.8 (0.6, 1.0) | 0.9 (0.6, 1.2) | 0.9 (0.7, 1.2) | < 0.001 |
| % Disability | 13.5 (11.0, 16.3) | 15.4 (13.1, 18.1) | 16.0 (13.5, 19.1) | 17.1 (14.5, 20.3) | < 0.001 |
| % Rural population | 55.6 (26.9, 80.9) | 54.7 (34.5, 75.4) | 51.1 (28.7, 72.7) | 52.0 (29.4, 75.5) | 0.19 |
| Socioeconomic and Environmental Features | |||||
| % Some College-level education | 60.8 (52.0, 69.0) | 58.7 (50.4, 66.7) | 58.2 (50.5, 66.0) | 55.6 (46.9, 63.3) | < 0.001 |
| % Unemployed | 3.4 (2.9, 4.1) | 3.8 (3.1, 4.5) | 4.1 (3.4, 5.0) | 4.7 (3.8, 5.7) | < 0.001 |
| Median household income ($) | 59,938.0 (50,941.0, 70,649.0) | 51,985.0 (46,447.0, 58,956.0) | 49,457.0 (43,079.0, 55,378.0) | 45,072.0 (39,124.0, 50,234.0) | < 0.001 |
| GINI coefficient | 40.1 (30.8, 40.5) | 40.4 (40.0, 40.7) | 40.5 (40.2, 40.9) | 40.8 (40.3, 50.3) | < 0.001 |
| % Household owned | 75.2 (70.6, 78.9) | 73.0 (68.4, 76.3) | 70.6 (65.7, 75.1) | 69.0 (62.4, 73.6) | < 0.001 |
| % Housing with cost burden | 10.5 (8.7, 12.5) | 10.5 (8.6, 12.8) | 10.6 (8.3, 13.0) | 10.4 (8.6, 13.0) | 0.97 |
| % Housing with overcrowding | 1.9 (1.2, 2.9) | 1.8 (1.2, 2.8) | 1.9 (1.4, 2.7) | 2.0 (1.3, 2.8) | 0.23 |
| % Housing with poor condition | 0.9 (0.6, 1.2) | 1.0 (0.7, 1.4) | 1.0 (0.7, 1.5) | 1.0 (0.7, 1.5) | < 0.001 |
| % Housing vacant | 12.1 (7.8, 18.9) | 15.3 (10.5, 22.0) | 16.0 (11.6, 22.2) | 17.6 (12.9, 24.2) | < 0.001 |
| Crime rate per 100,000 population | 173.3 (102.7, 274.5) | 201.4 (129.5, 311.8) | 240.4 (134.4, 381.8) | 275.7 (159.8, 441.4) | < 0.001 |
| % Food insecurity | 10.6 (9.0, 12.4) | 12.4 (10.9, 14.1) | 14.0 (12.0, 16.2) | 15.8 (13.2, 19.0) | < 0.001 |
| % Food desert | 3.4 (1.8, 5.3) | 5.5 (3.6, 7.8) | 7.8 (5.2, 10.5) | 11.6 (8.1, 15.1) | < 0.001 |
| % Food Stamp/SNAP | 10.0 (7.0, 13.2) | 12.8 (10.1, 15.9) | 14.3 (10.8, 17.9) | 17.2 (13.3, 21.5) | < 0.001 |
| Air quality, PM2.5 concentration (μg/m3) | 9.5 (8.2, 10.5) | 9.4 (8.0, 10.6) | 9.4 (7.9, 10.3) | 9.6 (8.3, 10.4) | 0.23 |
| Social Association rate per 100,000 population | 10.2 (7.9, 13.4) | 11.7 (8.7, 14.8) | 11.5 (8.8, 14.3) | 11.3 (8.6, 14.2) | < 0.001 |
| % Population with exercise opportunity | 70.4 (53.6, 84.0) | 68.0 (53.2, 79.9) | 68.8 (53.0, 81.2) | 65.5 (49.6, 79.9) | 0.003 |
| Healthcare Access features | |||||
| % Uninsured | 9.8 (6.7, 13.1) | 10.1 (7.0, 13.5) | 10.4 (7.3, 14.2) | 10.2 (7.5, 13.9) | 0.012 |
| % Medicaid/other public insurance | 13.1 (9.6, 17.4) | 16.1 (12.7, 19.9) | 17.2 (14.1, 21.9) | 20.4 (16.6, 25.8) | < 0.001 |
| PCP rate per 100,000 population | 48.9 (32.8, 72.9) | 48.8 (34.0, 70.4) | 52.2 (33.8, 75.3) | 51.9 (34.7, 70.5) | 0.43 |
| NP rate per 100,000 population | 49.5 (31.5, 70.5) | 52.8 (36.2, 77.6) | 61.0 (40.5, 88.5) | 65.3 (43.7, 96.7) | < 0.001 |
| PA rate per 100,000 population | 20.9 (8.7, 36.0) | 22.2 (9.4, 39.3) | 26.4 (11.7, 44.5) | 25.9 (10.6, 46.3) | < 0.001 |
| MHP rate per 100,000 population | 99.7 (45.7, 185.1) | 105.7 (51.1, 205.5) | 136.6 (65.3, 237.2) | 146.2 (61.0, 255.1) | < 0.001 |
| Healthcare centers per 100,000 population | 7.9 (2.0, 17.4) | 9.8 (4.2, 21.1) | 12.2 (4.6, 26.3) | 16.9 (6.4, 31.9) | < 0.001 |
| % HbA1c test among Medicare enrollees with diabetes | 88.0 (85.1, 90.4) | 87.7 (84.5, 89.8) | 87.2 (84.1, 90.0) | 86.7 (83.5, 89.2) | < 0.001 |
| % Annual ambulatory care visit among Medicare enrollees | 83.4 (78.8, 86.4) | 83.4 (78.4, 86.6) | 83.6 (79.5, 86.2) | 83.2 (78.9, 86.1) | 0.78 |
| Behavioral features | |||||
| % Diabetes | 10.6 (8.7, 12.9) | 11.7 (9.4, 14.2) | 12.0 (9.6, 15.0) | 12.6 (10.4, 15.5) | < 0.001 |
| % Obesity | 32.1 (28.0, 35.2) | 32.9 (29.6, 36.1) | 33.3 (30.0, 36.7) | 34.4 (30.7, 38.0) | < 0.001 |
| % Smoking | 15.6 (14.0, 18.1) | 16.8 (15.0, 19.2) | 17.5 (15.5, 19.9) | 19.0 (16.4, 21.4) | < 0.001 |
| % Hypertension | 36.6 (32.9, 40.4) | 36.7 (33.5, 40.6) | 37.2 (33.8, 41.5) | 36.8 (33.5, 40.5) | 0.069 |
| % Physically Inactive | 25.7 (21.5, 29.4) | 26.8 (23.6, 30.8) | 27.8 (24.0, 31.7) | 28.8 (25.2, 32.5) | < 0.001 |
| % Excessive drinking | 18.4 (16.2, 20.1) | 17.7 (15.7, 19.7) | 17.4 (15.2, 19.3) | 16.7 (14.2, 19.0) | < 0.001 |
FEI, food environment index; CVD, cardiovascular disease; PCP, primary care provider; NP, nurse practitioner; PA, physician assistant; MHP, mental health provider
Table 2 shows the association between FEI and CVD mortality alongside model performance. County-level factors collectively explained 58.5% variation in CVD mortality (Table 2; model 4). In the bivariate association, each unit increase in FEI accounted for − 67.26 deaths (95% CI: − 73.66, − 60.85) per 100,000 population. By features, the largest unadjusted mortality differences in CVD deaths per 100,000 population were associated with the proportion of African Americans (44.52 deaths/100,000 population), college education (− 50.88 deaths/100,000 population), income inequality (34.85 deaths/100,000 population), percentage of the population with exercise opportunity (− 36.93 deaths/100,000 population), a higher rate of PCPs (− 25.35 deaths/100,000 population), percentage of diabetes, obesity and smoking rates.
Table 2.
Association between County-Level Characteristics and Cardiovascular Deaths per 100,000 Population, USA 2017–2019
| Bivariate | Model-1 | Model-2 | Model-3 | Model-4 | |
|---|---|---|---|---|---|
| County features | |||||
| Food Environment Index (FEI) |
− 67.26*** (− 73.66 to − 60.85) |
− 53.79*** (− 61.69 to − 45.89) |
− 25.52*** (− 33.37 to − 17.67) |
− 22.04*** (− 29.89, − 14.18) |
− 12.93*** (− 20.53 to − 5.33) |
| Demographic composition | |||||
| Age, 45 years and above |
26.28*** (20.27, 32.28) |
35.14*** (23.86, 46.42) |
13.95** (4.20, 23.71) |
16.27** (6.57, 25.96) |
10.34* (1.03, 19.65) |
| Age X Age |
− 0.33*** (− 0.41, − 0.26) |
− 0.48*** (− 0.62, − 0.34) |
− 0.24*** (− 0.35, − 0.13) |
− 0.26*** (− 0.37, − 0.14) |
− 0.16** (− 0.27, − 0.05) |
| % Female, 45 years and above |
− 6.49* (− 12.58, − 0.40) |
11.01*** (5.69, 16.33) |
19.20*** (13.60, 24.80) |
18.38*** (12.79, 23.98) |
12.48*** (6.97, 18.00) |
| % African American, NH, 45 years and above |
44.52*** (38.93, 50.11) |
21.42*** (14.30, 28.54) |
− 5.16 (− 12.55, 2.22) |
− 8.24* (− 16.12, − 0.35) |
− 8.62* (− 16.51, − 0.73) |
| % Asian, NH, 45 years and above |
− 29.65*** (− 35.15, − 24.16) |
− 20.38*** (− 25.49, − 15.28) |
− 10.25*** (− 15.23, − 5.28) |
− 9.98*** (− 14.34, − 5.62) |
0.06 (− 4.36, 4.47) |
| % Hispanic, 45 years and above |
12.41*** (6.62, 18.19) |
− 9.58** (− 15.59, − 3.57) |
− 36.28*** (− 42.30, − 30.25) |
− 35.99*** (− 42.88, − 29.11) |
− 20.68*** (− 28.16, − 13.20) |
| % Urban |
26.19*** (20.01, 32.36) |
26.59*** (19.76, 33.42) |
5.02 (− 3.81, 13.85) |
8.42* (0.08, 16.76) |
9.32* (1.49, 17.16) |
| Socioeconomic and environmental features | |||||
| % Some college-level education |
− 50.88*** (− 54.22, − 47.54) |
− 61.60*** (− 68.48, − 54.71) |
− 53.95*** (− 61.35, − 46.55) |
− 28.47*** (− 36.41, − 20.53) |
|
| GINI Coefficient |
34.85*** (30.49, 39.21) |
17.65*** (11.13, 24.18) |
21.74*** (15.08, 28.39) |
18.49*** (12.30, 24.68) |
|
| % Household owned |
− 6.53** (− 10.59, − 2.47) |
8.19* (0.37, 16.01) |
1.47 (− 6.34, 9.27) |
5.96 (− 1.29, 13.22) |
|
| % Housing with cost burden |
6.93*** (3.26, 10.61) |
2.61 (− 1.48, 6.71) |
4.04* (0.07, 8.01) |
3.19 (− 0.77, 7.15) |
|
| Crime rate per 100,000 population |
25.82*** (21.92, 29.72) |
15.22*** (9.13, 21.31) |
14.64*** (9.04, 20.25) |
9.64*** (4.19, 15.09) |
|
| Social association rate per 100,000 population |
− 13.08*** (− 18.44, − 7.73) |
− 3.25 (− 8.34, 1.85) |
− 2.89 (− 9.58, 3.80) |
1.05 (− 5.19, 7.29) |
|
| % Population with exercise opportunity |
− 36.93*** (− 40.65, − 33.20) |
− 10.55** (− 18.23, − 2.87) |
− 9.60** (− 16.15, − 3.06) |
− 6.40* (− 12.68, − 0.11) |
|
| Healthcare access features | |||||
| % Uninsured |
21.79*** (18.16, 25.42) |
0.07 (− 6.65, 6.80) |
7.41* (1.01, 13.81) |
||
| PCP rate per 100,000 population |
− 25.35*** (− 31.60, − 19.10) |
− 16.00*** (− 22.61, − 9.39) |
− 12.05*** (− 17.69, − 6.40) |
||
| % Hb1Ac test among Medicare enrollees with diabetes |
− 5.36 (− 11.64, 0.92) |
0.16 (− 5.27, 5.59) |
− 0.04 (− 5.36, 5.28) |
||
| % Annual ambulatory care visit among Medicare enrollees |
31.68*** (27.00, 36.36) |
10.69*** (6.60, 14.79) |
8.09*** (4.25, 11.93) |
||
| Behavioral features | |||||
| % Diabetes |
49.60*** (45.68, 53.52) |
11.57*** (5.79, 17.36) |
|||
| % Obesity |
48.76*** (45.09, 52.43) |
− 20.07* (− 36.36, − 3.78) |
|||
| % Current smoking |
62.72*** (59.47, 65.97) |
30.98*** (23.38, 38.58) |
|||
| N | 2615 | 2615 | 2615 | 2615 | 2615 |
| Adj R2 | 0.303 | 0.431 | 0.456 | 0.585 | |
| AIC | 39,121.9 | 39,120.8 | 38,950.5 | 38,948.6 | |
| BIC | 39,180.6 | 39,226.4 | 39,114.6 | 39,089.3 | |
FEI, food environment index; CVD, cardiovascular disease; PCP, primary care provider; AIC, Akaike information criteria; BIC, Bayesian information criteria
*** p < 0.001
** p < 0.01
* p < 0.05
Examining the mortality differences by models, each unit increase in FEI (moving from worst to best) was associated with 53.79 lower CVD deaths per 100,000 population when controlled only for demographics (Table 2; model 1). After adding socioeconomic and environmental factors in model 2, the CVD mortality further attenuated to 25.52 deaths per 100,000 population. Adding healthcare access features (model 3) marginally improved the model with mortality further declining by less than 3 deaths per 100,000 population. In fully adjusted model 4, each unit increase in FEI was associated with 12.95 (95% CI: − 20.53, − 5.33) lower CVD deaths per 100,000 population.
Table 3 shows estimates of the CVD mortality rate by levels of income inequality and FEI. The healthiest food environment (FEI Q1) was associated with on average 55.7 lower CVD deaths/100,000 population when compared to the least healthy food environment (FEI Q4) in counties with lower income inequalities (Table 3, column 1). Restricting the sample to counties with higher income inequality, counties with the healthiest food environment reported 46.4 lower CVD deaths/100,000 population compared to the least healthy food environment counties (Table 3, column 2). To illustrate the interaction, regression plots of the predicted probability of CVD mortality as a function of FEI and income inequality (GINI coefficient) were generated (Fig. 2A and 2B). Consistent with modeling results, panel A shows a declining mortality rate with increasing FEI and decreasing income inequality, with the lowest mortality rate at the highest FEI and lowest income inequality. While comparing counties with lower income inequality (Q1) vs higher income inequality (Q2–Q4), CVD mortality declined linearly with increased FEI (panel B).
Table 3.
Estimation of the CVD Mortality Rate per 100,000 Population for Counties with Levels of Income Inequalities and Available Food Environment
| Lower Income Inequality (GINI Q1) (n = 653) |
Higher Income Inequality (GINI Q4) (n = 637) |
Lower Income Inequality (GINI Q1) (n = 653) |
Higher Income Inequality (GINI Q4) (n = 637) |
|
|---|---|---|---|---|
| β (95% CI) | β (95% CI) | β (95% CI) | β (95% CI) | |
| FEI Quartiles | ||||
| Least healthy food environment (Q4) | Ref | Ref | ||
| Average healthy food environment (Q3) |
− 14.66 (− 47.25, 17.93) |
− 26.80* (− 51.17, − 2.42) |
||
| Moderate healthy food environment (Q2) |
− 15.11 (− 45.09, 14.87) |
− 30.39* (− 59.64, − 1.13) |
||
| Healthiest food environment (Q1) |
− 55.73*** (− 87.41, − 24.04) |
− 46.39* (− 85.71, − 7.06) |
||
|
Food Environment Index (Continuous variable) |
− 18.99** (− 31.86, − 6.13) |
− 3.63* (− 5.44, − 1.83) |
||
|
FEI in African American-dominant counties## (Interaction term) |
30.52 (− 16.73, 77.77) |
− 14.80* (− 36.62, − 7.02) |
||
This table only reports the key estimated coefficients for the association between CVD mortality and food environments by county income inequality (GINI coefficient), where GINI quartile 1 represents counties with lower income inequality, and GINI quartile 4 represents counties with higher income inequalities. Columns 1 and 2 report the estimated association between CVD mortality and food environment expressed as a categorical variable by FEI quartiles. Columns 3 and 4 report the estimated differentials in the association between CVD mortality and food environment in black/non-black counties## within respected GINI quartiles. The model specification includes the same estimated control variables shown in table-3, and model-4 (excluding interactions) but is not reported for the purpose of simple presentation of the results
##African American-dominant counties are considered based on a binary variable, where at least 50% of the county population is African American
*** p < 0.001
** p < 0.01
* p < 0.05
Figure 2.
Panel A: Predicted probability of CVD mortality by food environment index and income inequality. Panel B: Predicted probability of CVD mortality by food environment index and quartiles of income inequality.
Furthermore, counties were binarized by whether African Americans represented 50% or more of the county population. Restricting the sample to counties with lower income inequalities, one unit increase in the food environment index was associated with 19 lower CVD deaths/100,000 population in counties where African Americans represented less than 50% of the population (Table 3, column 3). A similar increase in FEI showed a non-significant association between CVD mortality in counties with higher proportions of African Americans. In counties with higher income inequalities, one unit increase in the FEI was associated with 3.63 lower CVD deaths and 18.4 lower CVD deaths (14.80 + 3.63) in counties with low and high proportions of African American respectively (Table 3, column 4).
DISCUSSION
A growing body of literature establishes a strong association between food insecurity and CVD-related outcomes. We add to this literature by examining a measure of the food environment that includes both food insecurity and food desert status and also includes rich county-level covariates spanning socioeconomic, environmental, healthcare access, and behavioral factors. We found that moving from the least healthy to the healthiest food environment was associated with improved CVD mortality rates. Greater variation in CVD mortality was evident with demographic, socioeconomic, and environmental features, but not with a broader set of healthcare access and behavioral features across different levels of the food environment.
After adjusting for socioeconomic and environmental features, the mortality advantage of a better food environment was attenuated. Nonetheless, the mortality rate with a percentage of African Americans before socioeconomic, environmental, healthcare, and behavioral adjustments were reversed. Furthermore, the advantage with the percentage of the Hispanic population greatly improved after consideration of all other covariates. While previous studies have shown widened rural–urban mortality over the last three decades, with a comparatively greater rise in rural CVD mortality, these studies were limited in terms of adjusting for a broader set of socioeconomic, environmental, healthcare, and behavioral risk factors.30,31 In contrast, our finding suggests that mortality patterns associated with these demographic compositions (race and urbanization) were masked by socioeconomic, environmental, healthcare access, and behavioral factors.
Social and economic deprivation may limit access to healthy food resources. Evidence shows that communities with a higher percentage of low-income residents typically have fewer full-service supermarkets,4 and chronic exposure to stress from living in these deprived environments can lead to adverse CVD outcomes.32,33 However, our study finds that in counties with higher income inequality, improvement in the food environment significantly decreased the CVD mortality rate when compared to counties with lower income inequality. Novel findings of our study suggest a potential relationship between food environment and CVD mortality, which may also be improved by addressing these socioeconomic disparities.15,34,35
This study also finds that a higher food environment index may be associated with a lower CVD mortality rate and such a relationship was more pronounced in counties with larger African Americans with higher income inequalities. These findings have implications for addressing diet-related disparities and linked comorbidities, which are persistent among African Americans and low-income communities.36,37 Increasing access to full-service supermarkets in underserved areas has been shown to improve dietary quality and reduce health disparities. Studies have found that interventions promoting supermarket development positively impact dietary behaviors, including increased consumption of fruits and vegetables.38 Furthermore, programs that provide financial incentives to attract grocery stores and increase healthy food options in low-income areas have demonstrated positive impacts.39 Community-based initiatives, such as community gardens and farmers’ markets have also shown promising results in improving dietary behaviors and reducing chronic disease risks in underserved communities.40 While income inequalities pose challenges to improving health in the African American population, evidence supports the potential effectiveness of interventions targeting food environments. Strategies such as supermarket development, healthy food financing initiatives, income supports, and community-based approaches may contribute to better health outcomes, including reduced cardiovascular disease risk, even in the presence of income disparities. Addressing both income inequalities and food environments is crucial for promoting health equity and reducing health disparities among African Americans.
Healthcare access factors accounted for limited variation in CVD mortality across counties. More PCPs in a county were associated with a relatively large reduction in CVD mortality by 16 deaths per 100,000 population, potentially pointing to the role that PCPs playing managing chronic conditions and continuity of care. At the same time, Medicaid and other insurance coverage showed limited effects in the current analysis, despite prior studies showing positive impacts of health insurance on chronic conditions.40–42 Preventive services were also associated with small mortality advantages. This is notable because of the comparative prevalence of behavioral risk factors (diabetes, obesity, smoking) across counties by different levels of FE. If the mortality advantage is indeed attributable to demographic dispersion, and socioeconomic and environmental factors in the healthiest FE counites, then this has implications for how we invest in reducing socioeconomic differences relative to healthcare services.
Our study has several strengths. A major strength of the study is the utilization of the validated food environment index. The USDA’s Economic Research Service (ERS) has conducted extensive research on the FEI, including the development and validation of the index. The ERS developed the FEI to measure the availability and affordability of healthy foods in neighborhoods across the USA. The data sources used in the FEI, including the Census Bureau's American Community Survey, the Bureau of Labor Statistics’ Consumer Price Index, and the USDA's Food Availability Data System are vetted by ERS.1 Furthermore, CDC also has been using FEI as part of its efforts to monitor and improve the food environment in the USA.43 Furthermore, this study examined CVD mortality and its association with the food environment while controlling for a broad set of county-level socioeconomic and environmental factors. Combined the study features accounted for more than 50% of county CVD mortality variation. Finally, our modeling approach also accounted for the potential correlation between county-level factors within each state.
Although our study used a large number of covariates, these are not causal estimates and there may still be additional confounding variables, such as food insufficiency—the inability to get enough food to meet one's basic nutritional needs—that have a negative impact on cardiovascular health. We cannot completely rule out the possibility that food shortage serves as a proxy for other elements harming cardiovascular health. Furthermore, the study considered CVD mortality among adults aged at least 45 years; however, deaths occurring before age 45 are also often amenable to food environments along with socioeconomic and behavioral interventions. Although we included more than three-fourths of US counties our results are not necessarily generalizable to other counties or different calendar years.
CONCLUSION
In summary, CVD mortality is associated with food insecurity and food desert status. The racial disparity has additional mortality disadvantages in the presence of income inequality and a poor food environment. Although traditional risk factors were similarly prevalent across different levels of the food environment, healthcare access factors, except for an increasing number of PCPs, were of marginal importance in CVD mortality advantage after controlling for insurance access and income. This suggests that in addition to traditional targets, policymakers should test interventions that create healthier food environments and address social determinants of health.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements:
During the realization of this work, Tarang Parekh was a summer research fellow, recipient of a Doctoral Research Scholarship, and the High Impact Grant supported by the Office of the Provost, George Mason University.
Author Contributions
Dr. Parekh had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis.
Concept and design: Parekh.
Acquisition, analysis, or interpretation of data: Parekh.
Drafting of the manuscript: Parekh.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Parekh.
Administrative, technical, or material support: Parekh, Xue, Cheskin, Cuellar.
Supervision: Cuellar.
Data Availability
Publicly accessible data were used in the study. The supplement includes information on the linkages and data sources.
Declarations:
Conflicts of Interest:
The authors declare that they do not have a conflict of interest.
Financial Disclosure:
No financial disclosures were reported by the authors of this paper.
Disclaimers:
The study is a continuation of the work presented in the Ph.D. dissertation of Tarang Parekh.
Footnotes
Key Points
Question
Are differences in county-level food environment associated with differences in cardiovascular mortality among US adults?
Findings
This retrospective analysis of 2017–2019 data from 2615 counties, finds that for each unit increase in FEI, the county-level CVD mortality was lower by − 12.95 deaths/100,000 population. One unit increase in FEI in counties with higher income inequality was associated with a fivefold decrease in CVD mortality differences in African American countries compared to non–African American counties.
Meaning
In US counties between 2017 and 2019, better food environment quality measured by the index was associated with improved cardiovascular mortality among US adults.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.USDA ERS - Definitions of Food Security. Accessed 21 Dec 2021. https://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-us/definitions-of-food-security.aspx.
- 2.Coleman-Jensen A, Rabbitt MP, Gregory CA, Singh A. Household Food Security in the United States in 2019. Accessed 21 Dec 2021. http://www.ers.usda.gov/publications/pub-details/?pubid=99281.
- 3.What Is Food Insecurity in America? Hunger and Health. Accessed 21 Dec 2021. https://hungerandhealth.feedingamerica.org/understand-food-insecurity/.
- 4.Zenk SN, Schulz AJ, Israel BA, James SA, Bao S, Wilson ML. Neighborhood Racial Composition, Neighborhood Poverty, and the Spatial Accessibility of Supermarkets in Metropolitan Detroit. Am J Public Health. 2005;95(4):660–667. doi: 10.2105/AJPH.2004.042150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Li J, Kim C. Exploring relationships of grocery shopping patterns and healthy food accessibility in residential neighborhoods and activity space. Appl Geogr. 2020;116:102169. doi: 10.1016/j.apgeog.2020.102169. [DOI] [Google Scholar]
- 6.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]
- 7.Seligman HK, Schillinger D. Hunger and Socioeconomic Disparities in Chronic Disease. N Engl J Med. 2010;363(1):6–9. doi: 10.1056/NEJMp1000072. [DOI] [PubMed] [Google Scholar]
- 8.Franklin B, Jones A, Love D, Puckett S, Macklin J, White-Means S. Exploring Mediators of Food Insecurity and Obesity: A Review of Recent Literature. J Community Health. 2012;37(1):253–264. doi: 10.1007/s10900-011-9420-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Berkowitz SA, Berkowitz TSZ, Meigs JB, Wexler DJ. Trends in food insecurity for adults with cardiometabolic disease in the United States: 2005–2012. PLoS ONE. 2017;12(6):e0179172. doi: 10.1371/journal.pone.0179172. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Holben DH, Taylor CA. Food Insecurity and Its Association With Central Obesity and Other Markers of Metabolic Syndrome Among Persons Aged 12 to 18 Years in the United States. J Osteopath Med. 2015;115(9):536–543. doi: 10.7556/jaoa.2015.111. [DOI] [PubMed] [Google Scholar]
- 11.Parekh T, Xue H, Cheskin LJ, Cuellar AE. Food insecurity and housing instability as determinants of cardiovascular health outcomes: A systematic review. Nutr Metab Cardiovasc Dis. 2022;32(7):1590–1608. doi: 10.1016/j.numecd.2022.03.025. [DOI] [PubMed] [Google Scholar]
- 12.Men F, Gundersen C, Urquia ML, Tarasuk V. Prescription medication nonadherence associated with food insecurity: a population-based cross-sectional study. CMAJ Open. 2019;7(3):E590–E597. doi: 10.9778/cmajo.20190075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Palakshappa D, Ip EH, Berkowitz SA, et al. Pathways by Which Food Insecurity Is Associated With Atherosclerotic Cardiovascular Disease Risk. J Am Heart Assoc. 2021;10(22):e021901. doi: 10.1161/JAHA.121.021901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Beaulac J, Kristjansson E, Cummins S. A systematic review of food deserts, 1966–2007. Prev Chronic Dis. 2009;6(3):A105. [PMC free article] [PubMed] [Google Scholar]
- 15.Kelli HM, Hammadah M, Ahmed H, et al. Association Between Living in Food Deserts and Cardiovascular Risk. Circ Cardiovasc Qual Outcomes. 2017;10(9):e003532. doi: 10.1161/CIRCOUTCOMES.116.003532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Charkhchi P, FazeliDehkordy S, Carlos RC. Housing and Food Insecurity, Care Access, and Health Status Among the Chronically Ill: An Analysis of the Behavioral Risk Factor Surveillance System. J Gen Intern Med. 2018;33(5):644–650. doi: 10.1007/s11606-017-4255-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Garcia SP. Incremental Health Care Costs Associated With Food Insecurity and Chronic Conditions Among Older Adults. Prev Chronic Dis. 2018;15. doi:10.5888/pcd15.180058. [DOI] [PMC free article] [PubMed]
- 18.Liu Y, Eicher-Miller HA. Food Insecurity and Cardiovascular Disease Risk. Curr Atheroscler Rep. 2021;23(6):24. doi: 10.1007/s11883-021-00923-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Stupplebeen DA. Housing and Food Insecurity and Chronic Disease Among Three Racial Groups in Hawai’i. Prev Chronic Dis. 2019;16:E13. doi: 10.5888/pcd16.180311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Patel SA, Ali MK, Narayan KMV, Mehta NK. County-Level Variation in Cardiovascular Disease Mortality in the United States in 2009–2013: Comparative Assessment of Contributing Factors. Am J Epidemiol. 2016;184(12):933–942. doi: 10.1093/aje/kww081. [DOI] [PubMed] [Google Scholar]
- 21.Roth GA, Dwyer-Lindgren L, Bertozzi-Villa A, et al. Trends and Patterns of Geographic Variation in Cardiovascular Mortality Among US Counties, 1980–2014. JAMA. 2017;317(19):1976–1992. doi: 10.1001/jama.2017.4150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Thornton RLJ, Glover CM, Cené CW, Glik DC, Henderson JA, Williams DR. Evaluating Strategies For Reducing Health Disparities By Addressing The Social Determinants Of Health. Health Aff (Millwood). 2016;35(8):1416–1423. doi: 10.1377/hlthaff.2015.1357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Frieden TR. A Framework for Public Health Action: The Health Impact Pyramid. Am J Public Health. 2010;100(4):590–595. doi: 10.2105/AJPH.2009.185652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.CDC WONDER. Accessed 22 Dec 2021. https://wonder.cdc.gov/WelcomeA.html.
- 25.Centers for Disease Control and Prevention. Interactive Atlas of Heart Disease and Stroke. http://nccd.cdc.gov/DHDSPAtlas. Accessed on 12 Dec 2021.
- 26.Food environment index. County Health Rankings & Roadmaps. Accessed 22 Dec 2021. https://www.countyhealthrankings.org/explore-health-rankings/measures-data-sources/county-health-rankings-model/health-factors/health-behaviors/diet-exercise/food-environment-index.
- 27.Food Insecurity Report Briefs | Feeding America. Accessed 22 Dec 2021. https://www.feedingamerica.org/research/map-the-meal-gap/overall-executive-summary.
- 28.McCullough JM, Leider JP. Associations Between County Wealth, Health and Social Services Spending, and Health Outcomes. Am J Prev Med. 2017;53(5):592–598. doi: 10.1016/j.amepre.2017.05.005. [DOI] [PubMed] [Google Scholar]
- 29.Explore Health Rankings | 2021 Measures. County Health Rankings & Roadmaps. Accessed 24 Dec 2021. https://www.countyhealthrankings.org/explore-health-rankings/measures-data-sources/2021-measures.
- 30.Cross SH, Mehra MR, Bhatt DL, et al. Rural-Urban Differences in Cardiovascular Mortality in the US, 1999–2017. JAMA. 2020;323(18):1852–1854. doi: 10.1001/jama.2020.2047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Khan SU, Javed Z, Lone AN, et al. Social Vulnerability and Premature Cardiovascular Mortality Among US Counties, 2014 to 2018. Circulation. 2021;144(16):1272–1279. doi: 10.1161/CIRCULATIONAHA.121.054516. [DOI] [PubMed] [Google Scholar]
- 32.Palakshappa D, Ip EH, Berkowitz SA, et al. Pathways by Which Food Insecurity Is Associated With Atherosclerotic Cardiovascular Disease Risk. J Am Heart Assoc. 2021;10(22):e021901. doi: 10.1161/JAHA.121.021901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Holben DH, Taylor CA. Food Insecurity and Its Association With Central Obesity and Other Markers of Metabolic Syndrome Among Persons Aged 12 to 18 Years in the United States. J Osteopath Med. 2015;115(9):536–543. doi: 10.7556/jaoa.2015.111. [DOI] [PubMed] [Google Scholar]
- 34.Kelli HM, Kim JH, SammanTahhan A, et al. Living in Food Deserts and Adverse Cardiovascular Outcomes in Patients With Cardiovascular Disease. J Am Heart Assoc. 2019;8(4):e010694. doi: 10.1161/JAHA.118.010694. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Liu Y, Eicher-Miller HA. Food Insecurity and Cardiovascular Disease Risk. Curr Atheroscler Rep. 2021;23(6):24. doi: 10.1007/s11883-021-00923-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Noonan AS, Velasco-Mondragon HE, Wagner FA. Improving the health of African Americans in the USA: an overdue opportunity for social justice. Public Health Rev. 2016;37(1):12. doi: 10.1186/s40985-016-0025-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Satia JA. Diet-Related Disparities: Understanding the Problem and Accelerating Solutions. J Am Diet Assoc. 2009;109(4):610–615. doi: 10.1016/j.jada.2008.12.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Dubowitz T, Ghosh-Dastidar M, Cohen DA, et al. Diet And Perceptions Change With Supermarket Introduction In A Food Desert, But Not Because Of Supermarket Use. Health Aff Proj Hope. 2015;34(11):1858–1868. doi: 10.1377/hlthaff.2015.0667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Snelling AM, Yamamoto JJ, Belazis LB, Seltzer GR, McClave RL, Watts E. Incentivizing Supplemental Nutrition Assistance Program Purchases with Fresh Produce in Corner Stores to Reduce Food Inequity in Underserved Areas of Washington DC. Health Equity. 2020;4(1):386–393. doi: 10.1089/heq.2020.0028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.McCormack LA, Laska MN, Larson NI, Story M. Review of the nutritional implications of farmers’ markets and community gardens: a call for evaluation and research efforts. J Am Diet Assoc. 2010;110(3):399–408. doi: 10.1016/j.jada.2009.11.023. [DOI] [PubMed] [Google Scholar]
- 41.Hsu J, Price M, Huang J, et al. Unintended consequences of caps on Medicare drug benefits. N Engl J Med. 2006;354(22):2349–2359. doi: 10.1056/NEJMsa054436. [DOI] [PubMed] [Google Scholar]
- 42.Goldman DP, Joyce GF, Escarce JJ, et al. Pharmacy benefits and the use of drugs by the chronically ill. JAMA. 2004;291(19):2344–2350. doi: 10.1001/jama.291.19.2344. [DOI] [PubMed] [Google Scholar]
- 43.Healthy Food Environments. Centers for Disease Control and Prevention. Published April 20, 2023. Accessed 12 May 2023. https://www.cdc.gov/nutrition/healthy-food-environments/index.html.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Publicly accessible data were used in the study. The supplement includes information on the linkages and data sources.