Food insecurity (FI) is a major global health problem, associated with several adverse outcomes including poor physical(Reference Gundersen and Ziliak1), mental(Reference Leung, Laraia and Feiner2) and cognitive health(Reference Qian, Khadka and Martinez3), as well as higher healthcare expenditures(Reference Sonik4), and affecting roughly 28 % of people worldwide between 2021 and 2024(5). In the USA, the average prevalence of FI across states and the District of Columbia (DC) was 12·2 % in 2021–2023(6). However, state FI rates vary, from 7·4 % in New Hampshire to 18·9 % in Arkansas(6), and have been relatively stable; for example, both before and after COVID-19, some states like Arkansas, Texas and Mississippi had the highest FI rates, while others like New Hampshire, Massachusetts and North Dakota had the lowest(7). Notably, risk factors for state FI rates are complex and may include determinants beyond geographical or population size. For instance, California and Texas are two of the largest states in terms of population and geographical size in the USA but widely differ in FI prevalence rates, thus suggesting other, stronger determinants of FI.
Identifying state-level determinants of FI, particularly modifiable sociopolitical factors, is crucial to inform FI mitigation strategies and may have implications for jurisdictions across levels (e.g. nationally and locally) globally. For instance, proximal indicators of FI like food access and income(Reference Odoms-Young, Brown and Agurs-Collins8) may be reflected by food environment(Reference Odoms-Young, Brown and Agurs-Collins8), eligibility to nutrition assistance programmes such as Supplemental Nutrition Assistance Program (SNAP)(Reference Keith-Jennings, Llobrera and Dean9), minimum wage(Reference Freije, Wallace and Chaparro10,Reference Winkler, Clohan and Komro11) and insurance coverage (which reduce out-of-pocket healthcare costs)(Reference Himmelstein12). Other FI determinants may be indirect, such as inequalities contributing to health disparities(Reference Odoms-Young and Bruce13) like the gender wage gap(Reference Feeley14), restricted reproductive healthcare access(Reference Finlay and Lee15) and non-discrimination policies for minoritised racial/ethnic and sexual/gender groups, as these factors can affect income, educational, employment and housing opportunities/access(Reference Odoms-Young and Bruce13,Reference Gibb, Shokoohi and Salway16) . Notably, FI disproportionately affects women, minoritised racial/ethnic and sexual/gender groups, households with children, and those who are unemployed, received less education and experience poverty(6,Reference Odoms-Young, Brown and Agurs-Collins8) . Moreover, these subpopulations may be intersecting, for instance, low-income and minoritised racial/ethnic groups are more likely to reside in neighbourhoods with limited grocery store access(Reference Odoms-Young, Brown and Agurs-Collins8), which is partially rooted in structural racism(Reference Li and Yuan17).
Additionally, substance use may exacerbate FI (e.g. by straining finances)(Reference Anema, Mehra, Weiser, Hall, Hall and Cockerell18). Thus, FI rates may be impacted by state-level policies and regulations related to tobacco(Reference Chaloupka19), alcohol(Reference Campbell, Hahn and Elder20) and cannabis(Reference Mennis, McKeon and Stahler21). Notably, non-medical cannabis legalisation may increase access to cannabis(Reference Mennis, McKeon and Stahler21) but may also address historical disproportionate incarceration of Black individuals(Reference Yang, Berg and Burris22), who consequently experience greater barriers to employment, housing and nutrition assistance programmes(Reference Wildeman and Wang23).
Existing research has primarily investigated factors proximally related to FI, while broader social and policy conditions that may shape state-level FI are understudied(Reference Bartfeld and Dunifon24,Reference Bartfeld and Men25) . Thus, building on previous literature(Reference Odoms-Young, Brown and Agurs-Collins8–Reference Freije, Wallace and Chaparro10,Reference Himmelstein12,Reference Carter, Dubois and Tremblay26,Reference Broussard27) and frameworks(Reference Bartfeld and Dunifon24,Reference Bartfeld and Men25) , this cross-sectional study explored select state-level sociopolitical indicators (both proximal FI determinants and factors potentially indirectly influencing FI via financial stability, equity and substance use) among states with significantly higher or lower FI prevalence than the US average(6). Findings may shed light on sociopolitical factors that differentiate states with low v. high FI prevalence and inform multilevel FI mitigation strategies in the USA and globally.
Methods
Food insecurity
State-level FI prevalence rates were obtained from the US Department of Agriculture (USDA), which combined data from 2021 to 2023 to provide reliable state/DC and national FI estimates(6). The USDA uses standard methodology to identify states with statistically significantly different FI prevalence relative to the US average (12·2 %) at the 90 % confidence level (t > 1·645), as recommended by the US Census Bureau(Reference Rabbitt, Reed-Jones and Hales28). Additional methodological details have been published previously(Reference Rabbitt, Reed-Jones and Hales28).
This study included states with significantly higher and lower FI prevalence relative to the US average(6); we considered DC as a ‘state’ in this study. This sample included twenty-five states (including DC): seven had higher FI prevalence (Arkansas [AR], Kentucky [KY], Louisiana [LA], Mississippi [MS], Oklahoma [OK], South Carolina [SC] and Texas [TX]) and eighteen had lower FI (California [CA], Colorado [CO], Hawai’i [HI], Iowa [IA], Massachusetts [MA], Maryland [MD], Minnesota [MN], North Dakota [ND], New Hampshire [NH], New Jersey [NJ], Pennsylvania [PA], Rhode Island [RI], South Dakota [SD], Virginia [VA], Vermont [VT], Washington [WA], Wisconsin [WI] and DC) than the US average. The remaining twenty-six states with FI prevalence similar to the US average were excluded.
Variables
One PhD-level author extracted data from robust national public data sources on sixteen state-level sociopolitical indicators across three categories: (1) proximal indicators (Food Environment Index (FEI) scores(29), SNAP participation rates(Reference Nchako30), minimum wage amount(31), percentage uninsured(32), Medicaid expansion(33) and child tax credit (CTC) programmes(34)); (2) inequality indicators (income inequality(35), gender wage gap(36), protective policies for sexual/gender minoritised groups(37), Racial Equity Index (REI) scores(38) and reproductive rights(39)); and (3) tobacco, alcohol and cannabis regulation indicators (100 % smoke-free laws(40), cigarette excise taxes(41), non-medical cannabis legalisation(42), cannabis possession-related expungements or pardons(43) and alcohol outlet density regulation(Reference Michel, Treffers and O’Malley44)). Table 1 provides data source, description and operationalisation of each sociopolitical indicator. Data from 2023 (or before) were extracted to coincide with FI prevalence estimates (2021–2023). For all indicators except percentage uninsured and income inequality, higher scores (for continuous) or presence of policies (for dichotomous) indicated intended positive public health impacts (i.e. aimed at improving food access, financial resources, equality and substance use-related regulations). For percentage uninsured and income inequality, lower scores indicated intended positive public health impacts.
Sociopolitical indicator source, description and operationalisation
Data analysis
We described continuous and dichotomous sociopolitical indicators for each state based on FI status (low v. high) and then conducted exploratory bivariate analyses (t tests or Fisher’s tests, as appropriate) characterising them in relation to state FI status. Significance level was set at P < 0·05. Given the exploratory nature of these tests, we did not adjust for confounders; findings should be interpreted as hypothesis-generating.
Results
Table 2 summarises each state’s FI rates and sociopolitical indicators. FI rates in the eighteen low-FI states ranged from 7·4 % in NH to 11·4 % in CA. Rates in the seven high-FI states ranged from 14·4 % in SC to 18·9 % in AR.
Food insecurity prevalence and relevant state-level sociopolitical indicators among 25 states (including DC) in the USA representing high (n 7) and low (n 18) food insecurity prevalence
FI, food insecurity; FEI, Food Environment Index; SNAP, Supplemental Nutrition Assistance Program; LGBTQ, Lesbian, Gay, Bisexual, Transgender and Queer.
*1 = yes, 0 = no.
†Higher = intended positive public health impact.
‡Lower = intended positive public health impact.
§0 = Negative, 1 = low, 2 = fair, 3 = medium, 4 = high policies.
Proximal indicators
In low-FI states, FEI scores (range: 0–10, higher = greater access) ranged from 7·7 in SD to 9·5 in NH; SNAP participation ranged from 66 % in CA and ND to 97 % in MA; minimum wage varied from $7·25 (federal minimum) in IA, NH, ND, PA and WI to $17·0 in DC; and percentage uninsured ranged from 2·6 % in MA to 8·3 % in SD. All states but WI (94·4 %, n 17/18) had Medicaid expansion programmes. CA, CO, DC, MD, MA, MN, NJ and VT (44·4 %, n 8/18) had state-level CTC programmes.
High-FI states varied in FEI scores (from 4·0 in MS to 6·8 in KY), SNAP participation (62 % in AR to 84 % in OK) and percentage uninsured (5·4 % in KY to 16·4 % in TX). Six (85·7 %, n 6/7) had the federal minimum wage ($7·25); AR was higher ($11·00). AR, KY, OK and LA (57·1 %, n 4/7) had Medicaid expansion. Only OK (14·3 %, n 1/7) had state-level CTC programmes.
Inequality indicators
In low-FI states, Gini coefficients (higher = greater inequality) varied from 0·44 in NH to 0·52 in DC, gender wage gaps ranged from $0·76 in NH and ND to $0·89 in RI and REI scores (range: 19·5–62·4, higher = greater equity) ranged from 31·1 in SD to 62·4 in VA. All but IA, ND, PA, SD, VA and WI (66·7 %, n 12/18) had strong protective policies for sexual/gender minority groups, and 61·1 % (n 11/18; CA, CO, DC, HI, MA, MD, MN, NJ, RI, VT and WA) had abortion protections.
High-FI states varied in the Gini coefficient (0·47 in AR, OK and SC to 0·50 in LA), gender wage gap (0·71 in LA to 0·82 in AR, KY, TX) and REI scores (19·0 in MS to 51·9 in OK). None had strong protective policies for sexual/gender minority groups or abortion.
Tobacco, alcohol and cannabis regulation indicators
For low-FI states, cigarette excise taxes varied from $0·44 in ND to $4·50 in DC. All but NH, PA and VA (83·3 %, n 15/18) had 100 % smoke-free laws. All but HI, IA, NH, ND, PA, SD and WI (61·1 %, n 11/18) had legalised non-medical cannabis; most (83·3 %, n 15/18) had provisions for expunging/sealing/pardoning prior cannabis-related convictions. All but HI, IA, ND, VA and VT (72·2 %, n 13/18) had alcohol outlet density regulations.
In high-FI states, cigarette excise taxes varied from $0·57 in SC to $2·03 in OK. Only two states (28·6 %, n 2/7) had 100 % smoke-free laws (AR and LA) or alcohol outlet density regulations (AR and KY). None had legalised non-medical cannabis nor provisions for expunging/sealing/pardoning prior cannabis-related convictions.
Exploratory analysis
For proximal indicators, low-FI (v. high-FI) states had greater FEI scores, SNAP participation, minimum wage and percentage insured (P’s < 0·05, Table 3). For inequality indicators, low-FI (v. high-FI) states had narrower gender wage gaps and greater REI scores. A greater proportion of low-FI (v. high-FI) states had more protective policies for sexual/gender minority populations and abortion rights (P’s < 0·05). For tobacco, alcohol and cannabis regulation indicators, low-FI (v. high-FI) states had greater cigarette taxes. A greater proportion of low-FI (v. high-FI) states had 100 % smoke-free laws, legalised non-medical cannabis and provisions for expunging/sealing/pardoning prior cannabis-related convictions (P’s < 0·05).
Sociopolitical indicators and characteristics across 25 states (including DC) in the US with low v. high food insecurity (FI)
Boldface italics indicates statistical significance (P < 0.05).
Discussion
Relative to high-FI states, low-FI states had more sociopolitical indicators with intended positive public health impacts. Notably, low-FI states had both proximal sociopolitical indicators aimed to improve food access and distal factors influencing inequalities and substance use, which may have had beneficial spillover effects on FI. Thus, findings suggest that FI mitigation efforts should apply a systems perspective(Reference Hong, Bangpan and Stansfield45,Reference Perrotte and Noorestani46) and consider population-level strategies to increase financial resources, reduce disparities and regulate substance use in addition to strengthening food access. These insights may have implications not only for the USA but globally, as well.
FI mitigation efforts should consider which factors (i.e. proximal, inequality or substance use-related regulations) amplify or reduce FI and the magnitudes of their effects, individually and in the context of other factors. These considerations may explain unexpected findings; for instance, low-FI states like SD or VA have some sociopolitical indicators with low scores for their intended positive public health impacts (such as for SNAP participation rates, wage gap and tobacco regulations). Relatedly, it is important to understand how multiple indicators within the same category may impact FI. For instance, although OK and LA have high SNAP participation rates among eligible participants, they have high FI. It is possible that, despite SNAP participation, poor food environments (which relate to income and healthy food proximity) are more powerful determinants of FI rates in these states. Such differences may also suggest potentially unmeasured or interacting indicators that might influence FI. Thus, future research investigating the magnitude of the impact of these intersecting indicators on FI is warranted.
Notably, findings suggest that non-food-related sociopolitical indicators (that are not traditionally targeted by FI-related policies) may be important in addressing FI, underscoring the increasingly urgent need to prioritise cross-sector policy initiatives. This research reinforces the importance of a Health in All Policies(Reference Raskind47) approach, an integrated policymaking framework that considers health implications across all sectors to improve population outcomes(48). Thus, as FI rates continue to rise, sustainable FI mitigation efforts are crucial and must prioritise coordinated action across multiple government agencies and sectors to advance relevant policy initiatives. These may include enhancing healthy food access, expanding SNAP budgets, minimum wages and Medicaid coverage, promoting health equity and strengthening public health protections like substance-related regulations(Reference Raskind47,Reference Gaines-Turner, Simmons and Chilton49) .
Some aspects of this study warrant careful consideration. Study limitations include an ecological analysis of a small subset (n 25) of US states. Thus, findings may not generalise to other states, levels (e.g. individual, local) or international settings. Furthermore, due to the small sample size, we relied on bivariate analyses and did not account for unmeasured heterogeneity across states or potential confounders, for example, other state-level sociopolitical factors (e.g. substance use recovery programme accessibility or SNAP implementation features). Additionally, this study’s cross-sectional, exploratory nature precludes causal inference. Finally, despite use of robust data sources, some sociopolitical indicators may not fully or accurately reflect each state’s complex policy landscapes. Considering these limitations, patterns identified between sociopolitical environment and FI must be interpreted with caution.
Conclusion
Low-FI states exhibited more sociopolitical indicators with intended positive public health impact, particularly related to food access, financial security, equity and substance use-related regulations. A sustainable, multilevel, holistic approach addressing multiple intersecting determinants of FI, beyond proximal factors alone, is essential. Future research utilising longitudinal designs and examining a broader set multilevel determinants is needed to better understand causal relationships and interactions among sociopolitical indicators of FI.
Data availability statement
All data are publicly available.
Acknowledgements
None.
Authorship
R.C.: Conceptualisation, Data curation, Formal analysis, Methodology, Writing – original draft and Writing – review and editing. K.F.R.: Data curation, Writing – original draft and Writing – review and editing. C.L.: Writing – original draft and Writing – review and editing. Y.T.Y.: Writing – review and editing. C.J.B.: Conceptualisation, Methodology, Writing – original draft and Writing – review and editing.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
There are no conflicts of interest.
Ethics of human subject participation
This study was categorised as non-human-subjects research.
Open access
