[Skip to Navigation]
Sign In
Table 1.  Characteristics of SuperSNAP Participants
Characteristics of SuperSNAP Participants
Table 2.  Differences in Preintervention Purchases Between 436 SuperSNAP Participants and 42 716 Comparison Group Participants
Differences in Preintervention Purchases Between 436 SuperSNAP Participants and 42 716 Comparison Group Participants
Table 3.  Changes in Purchases Associated With SuperSNAP Participation
Changes in Purchases Associated With SuperSNAP Participation
1.
Mokdad  AH, Ballestros  K, Echko  M,  et al; US Burden of Disease Collaborators.  The state of US health, 1990-2016: burden of diseases, injuries, and risk factors among US states.   JAMA. 2018;319(14):1444-1472. doi:10.1001/jama.2018.0158 PubMedGoogle Scholar
2.
Sotos-Prieto  M, Bhupathiraju  SN, Mattei  J,  et al.  Association of changes in diet quality with total and cause-specific mortality.   N Engl J Med. 2017;377(2):143-153. doi:10.1056/NEJMoa1613502 PubMedGoogle ScholarCrossref
3.
Sotos-Prieto  M, Bhupathiraju  SN, Mattei  J,  et al.  Changes in diet quality scores and risk of cardiovascular disease among US men and women.   Circulation. 2015;132(23):2212-2219. doi:10.1161/CIRCULATIONAHA.115.017158 PubMedGoogle ScholarCrossref
4.
Sugiyama  T, Shapiro  MF.  The growing socioeconomic disparity in dietary quality: mind the gap.   JAMA Intern Med. 2014;174(10):1595-1596. doi:10.1001/jamainternmed.2014.3048 PubMedGoogle ScholarCrossref
5.
Wang  DD, Leung  CW, Li  Y,  et al.  Trends in dietary quality among adults in the United States, 1999 through 2010.   JAMA Intern Med. 2014;174(10):1587-1595. doi:10.1001/jamainternmed.2014.3422 PubMedGoogle ScholarCrossref
6.
Wang  PY, Fang  JC, Gao  ZH, Zhang  C, Xie  SY.  Higher intake of fruits, vegetables or their fiber reduces the risk of type 2 diabetes: a meta-analysis.   J Diabetes Investig. 2016;7(1):56-69. doi:10.1111/jdi.12376 PubMedGoogle ScholarCrossref
7.
Alissa  EM, Ferns  GA.  Dietary fruits and vegetables and cardiovascular diseases risk.   Crit Rev Food Sci Nutr. 2017;57(9):1950-1962. doi:10.1080/10408398.2015.1040487PubMedGoogle Scholar
8.
Lelong  H, Blacher  J, Baudry  J,  et al.  Individual and combined effects of dietary factors on risk of incident hypertension: prospective analysis from the NutriNet-Santé cohort.   Hypertension. 2017;70(4):712-720. doi:10.1161/HYPERTENSIONAHA.117.09622PubMedGoogle ScholarCrossref
9.
Micha  R, Shulkin  ML, Peñalvo  JL,  et al.  Etiologic effects and optimal intakes of foods and nutrients for risk of cardiovascular diseases and diabetes: systematic reviews and meta-analyses from the Nutrition and Chronic Diseases Expert Group (NutriCoDE).   PLoS One. 2017;12(4):e0175149. doi:10.1371/journal.pone.0175149 PubMedGoogle Scholar
10.
He  FJ, Nowson  CA, Lucas  M, MacGregor  GA.  Increased consumption of fruit and vegetables is related to a reduced risk of coronary heart disease: meta-analysis of cohort studies.   J Hum Hypertens. 2007;21(9):717-728. doi:10.1038/sj.jhh.1002212 PubMedGoogle ScholarCrossref
11.
Dietary Guidelines Advisory Committee. Scientific report of the 2020 Dietary Guidelines Advisory Committee: advisory report to the Secretary of Agriculture and the Secretary of Health and Human Services. US Dept of Agriculture, Economic Research Service; 2020. Accessed July 6, 2021. https://www.dietaryguidelines.gov/sites/default/files/2020-07/ScientificReport_of_the_2020DietaryGuidelinesAdvisoryCommittee_first-print.pdf
12.
Poti  JM, Mendez  MA, Ng  SW, Popkin  BM.  Is the degree of food processing and convenience linked with the nutritional quality of foods purchased by US households?   Am J Clin Nutr. 2015;101(6):1251-1262. doi:10.3945/ajcn.114.100925 PubMedGoogle ScholarCrossref
13.
Malik  VS, Popkin  BM, Bray  GA, Després  JP, Hu  FB.  Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk.   Circulation. 2010;121(11):1356-1364. doi:10.1161/CIRCULATIONAHA.109.876185 PubMedGoogle ScholarCrossref
14.
Coleman-Jensen  A, Rabbitt  MP, Gregory  CA, Singh  A. Household food security in the United States in 2019. Accessed November 3, 2020. https://www.ers.usda.gov/publications/pub-details/?pubid=99281
15.
Feeding America. The impact of the coronavirus on local food insecurity. Accessed May 22, 2020. https://www.feedingamerica.org/sites/default/files/2020-05/Brief_Local%20Impact_5.19.2020.pdf
16.
Morales  ME, Berkowitz  SA.  The relationship between food insecurity, dietary patterns, and obesity.   Curr Nutr Rep. 2016;5(1):54-60. doi:10.1007/s13668-016-0153-y PubMedGoogle ScholarCrossref
17.
Seligman  HK, Schillinger  D.  Hunger and socioeconomic disparities in chronic disease.   N Engl J Med. 2010;363(1):6-9. doi:10.1056/NEJMp1000072 PubMedGoogle ScholarCrossref
18.
Seligman  HK, Berkowitz  SA.  Aligning programs and policies to support food security and public health goals in the United States.   Annu Rev Public Health. 2019;40:319-337. doi:10.1146/annurev-publhealth-040218-044132 PubMedGoogle ScholarCrossref
19.
Food and Nutrition Service, US Department of Agriculture. Supplemental Nutrition Assistance Program (SNAP). Accessed February 19, 2020. https://www.fns.usda.gov/snap/supplemental-nutrition-assistance-program
20.
Canning  P, Stacy  B. The Supplemental Nutrition Assistance Program (SNAP) and the economy: new estimates of the SNAP multiplier. Accessed July 3, 2020. https://www.ers.usda.gov/publications/pub-details/?pubid=93528
21.
Gundersen  C.  Food insecurity is an ongoing national concern.   Adv Nutr. 2013;4(1):36-41. doi:10.3945/an.112.003244 PubMedGoogle ScholarCrossref
22.
Carlson S. More adequate SNAP benefits would help millions of participants better afford food. Center on Budget and Policy Priorities. Accessed July 3, 2020. https://www.cbpp.org/research/food-assistance/more-adequate-snap-benefits-would-help-millions-of-participants-better
23.
Mozaffarian  D, Liu  J, Sy  S,  et al.  Cost-effectiveness of financial incentives and disincentives for improving food purchases and health through the US Supplemental Nutrition Assistance Program (SNAP): a microsimulation study.   PLoS Med. 2018;15(10):e1002661. doi:10.1371/journal.pmed.1002661 PubMedGoogle Scholar
24.
Olsho  LE, Klerman  JA, Wilde  PE, Bartlett  S.  Financial incentives increase fruit and vegetable intake among Supplemental Nutrition Assistance Program participants: a randomized controlled trial of the USDA Healthy Incentives Pilot.   Am J Clin Nutr. 2016;104(2):423-435. doi:10.3945/ajcn.115.129320 PubMedGoogle ScholarCrossref
25.
Berkowitz  SA, O’Neill  J, Sayer  E,  et al. Health center–based community-supported agriculture: an RCT. Am J Prev Med. 2019;57(6 suppl 1):S55-S64. doi:10.1016/j.amepre.2019.07.015
26.
Yoder  AD, Proaño  GV, Handu  D.  Retail nutrition programs and outcomes: an evidence analysis center scoping review.   J Acad Nutr Diet. Published online November 20, 2020. doi:10.1016/j.jand.2020.08.080PubMedGoogle Scholar
27.
Reinvestment Partners. SuperSNAP. Accessed November 4, 2020. https://reinvestmentpartners.org/what-we-do/produce-prescriptions/supersnap-1.html
28.
National Institute of Food and Agriculture, US Department of Agriculture. Gus Schumacher Nutrition Incentive Program. Accessed November 3, 2020. https://nifa.usda.gov/program/gus-schumacher-nutrition-incentive-grant-program
29.
Healthy Food America. Healthy food pricing incentives: a systematic review of current evidence. Accessed November 4, 2020. https://www.healthyfoodamerica.org/healthy_food_pricing_incentives_a_systematic_review_of_current_evidence
30.
Hager  ER, Quigg  AM, Black  MM,  et al.  Development and validity of a 2-item screen to identify families at risk for food insecurity.   Pediatrics. 2010;126(1):e26-e32. doi:10.1542/peds.2009-3146 PubMedGoogle ScholarCrossref
31.
Lacko  A, Ng  SW, Popkin  B.  Urban vs. rural socioeconomic differences in the nutritional quality of household packaged food purchases by store type.   Int J Environ Res Public Health. 2020;17(20):7637. doi:10.3390/ijerph17207637 PubMedGoogle ScholarCrossref
32.
Ng  SW, Hollingsworth  BA, Busey  EA, Wandell  JL, Miles  DR, Poti  JM.  Federal nutrition program revisions impact low-income households’ food purchases.   Am J Prev Med. 2018;54(3):403-412. doi:10.1016/j.amepre.2017.12.003 PubMedGoogle ScholarCrossref
33.
Berkowitz  SA, Delahanty  LM, Terranova  J,  et al.  Medically tailored meal delivery for diabetes patients with food insecurity: a randomized cross-over trial.   J Gen Intern Med. 2019;34(3):396-404. doi:10.1007/s11606-018-4716-z PubMedGoogle ScholarCrossref
34.
Thomas  LE, Li  F, Pencina  MJ.  Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial.   JAMA. 2020;323(23):2417-2418. doi:10.1001/jama.2020.7819 PubMedGoogle ScholarCrossref
35.
Li  F, Morgan  KL, Zaslavsky  AM.  Balancing covariates via propensity score weighting.   J Am Stat Assoc. 2018;113(521):390-400. doi:10.1080/01621459.2016.1260466 PubMedGoogle ScholarCrossref
36.
Gittelsohn  J, Trude  ACB, Kim  H.  Pricing strategies to encourage availability, purchase, and consumption of healthy foods and beverages: a systematic review.   Prev Chronic Dis. 2017;14:E107. doi:10.5888/pcd14.170213 PubMedGoogle Scholar
37.
Engel  K, Ruder  EH.  Fruit and vegetable incentive programs for Supplemental Nutrition Assistance Program (SNAP) participants: a scoping review of program structure.   Nutrients. 2020;12(6):1676. doi:10.3390/nu12061676 PubMedGoogle ScholarCrossref
38.
Moran  A, Thorndike  A, Franckle  R,  et al.  Financial incentives increase purchases of fruit and vegetables among lower-income households with children.   Health Aff (Millwood). 2019;38(9):1557-1566. doi:10.1377/hlthaff.2018.05420 PubMedGoogle ScholarCrossref
39.
Seguin  RA, Morgan  EH, Hanson  KL,  et al.  Farm Fresh Foods for Healthy Kids (F3HK): an innovative community supported agriculture intervention to prevent childhood obesity in low-income families and strengthen local agricultural economies.   BMC Public Health. 2017;17(1):306. doi:10.1186/s12889-017-4202-2 PubMedGoogle ScholarCrossref
40.
Mancino  L, Guthrie  J, Just  DR.  Overview: exploring ways to encourage healthier food purchases by low-income consumers—lessons from behavioral economics and marketing.   Food Policy. 2018;79:297-299. doi:10.1016/j.foodpol.2018.03.007 Google ScholarCrossref
41.
Audet AJ, Zezza MA. How behavioral economics can advance the design of effective clinician incentive programs. Health Affairs blog. Accessed December 4, 2020. https://www.healthaffairs.org/do/10.1377/hblog20150915.050556/full/
42.
Food and Nutrition Service, US Department of Agriculture. FINI Grant Program. Accessed July 6, 2021. https://www.fns.usda.gov/snap/FINI-Grant-Program
43.
Berkowitz  SA, Edwards  ST, Polsky  D.  Cash benchmarking for integrated health care and human services interventions: finding the value added.   Health Aff (Millwood). 2020;39(4):582-586. doi:10.1377/hlthaff.2019.01579 PubMedGoogle ScholarCrossref
44.
Gundersen  C, Ziliak  JP.  Food insecurity and health outcomes.   Health Aff (Millwood). 2015;34(11):1830-1839. doi:10.1377/hlthaff.2015.0645PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    1 Comment for this article
    EXPAND ALL
    Error in Food Purchasing Data in Study of Fruit and Vegetable Subsidy Program Among US Individuals With Low Income
    Seth A. Berkowitz, MD, MPH | Division of General Medicine and Clinical Epidemiology, Department of Medicine, University of North Carolina at Chapel Hill School of Medicine
    I write on behalf of my co-authors to report errors in the Original Investigation, “Association of a Fruit and Vegetable Subsidy Program With Food Purchases by Individuals With Low Income in the US,”1 published in JAMA Network Open on August 11, 2021. All study authors have approved this message. This longitudinal cohort study used data from transaction records of a large supermarket chain in North Carolina from October 2019 to April 2020. We assessed monthly spending on all fruits, vegetables, legumes, and nuts (primary outcome); spending on less healthy food categories; and spending on sugar-sweetened beverages as 1 category of less healthy foods among participants who used the Supplemental Nutrition Assistance Program (SNAP) combined with an additional monthly benefit to purchase healthy foods (SuperSNAP) and SNAP beneficiaries who did not use SuperSNAP but shopped in the same stores. We found that SuperSNAP participation was associated with increased monthly purchases of fruits, vegetables, nuts, and legumes. We recently discovered an error in the food purchasing data we received from the supermarket chain. These data underlie the analyses reported in this paper. These are point of sale data that record purchases made by customers. Specifically, we discovered that in some instances when individuals used more than 1 payment method in the same transaction (eg, both cash and credit card), called a “split-tender” transaction, there was duplication of the purchase record. For example, if a person purchased a single can of beans in their shopping trip but paid with 2 different tender types, sometimes a record was generated such that the purchase of 2 cans of beans was reflected—1 paid with cash and 1 with credit card. Overall, this inflated the number of purchases and total spending. We discovered this error when we were further using this data for additional analyses. Discovery of this error prompted us to review the entire article, data, and analyses for any other errors. We have now corrected the data and re-run all analyses. We confirm that all errors identified are discussed herein, and we are not aware of any additional errors. Fortunately, this issue with the data applies proportionally to both the intervention and comparison group and throughout the entire time period in our study, so there is little qualitative change in the findings. The error does change several descriptive statistics and point estimates and CIs of analyses, but these changes are modest (eg, the primary outcome is that the intervention group spent approximately $24 dollars more per month on fruits and vegetables, compared with $30 as previously reported). The direction and statistical significance of study outcomes are not changed, and there are no changes in the interpretations or conclusions reported. In addition, there are 2 other changes. First, the number of comparison group participants increased from 33246 to 42716. This has to do with the construction of the cohort. As described in the Supplement, to remove implausibly high purchase values associated with a shopper ID number (eg, when a store employee lets a shopper use a store-based shopper ID number if they have forgotten their own discount card), we removed outlier shopper ID values at a cut-off level corresponding to the top 1% of purchases. In the original article, this level was $1452, and this has decreased to $1008 in the corrected dataset. The removal of duplicates also changed the distribution of high spending, removing bunching at the top. Taken together, these changes result in more control participants meeting eligibility criteria. Second, 3 outcomes related to out-of-pocket expenditures (ie, expenditures that were not related to SNAP or WIC funds) were removed because they cannot be estimated accurately. Given the issue with split-tender transactions, we cannot be sure if the purchases should be attributed to program benefits or out of pocket spending (the data error produces records for both). To address these errors, corrections are needed in the Abstract, Key Points, Methods and Results sections, Tables 2 and 3, and the Supplement. We have requested that the journal publish a Correction notice and correct the article. We deeply regret the error and apologize that we did not notice it earlier. Fortunately, the error does not affect the original study conclusions that participation in SuperSNAP was associated with meaningful increases in healthy food purchasing. REFERENCE 1. Berkowitz SA, Curran N, Hoeffler S, Henderson R, Price A, Ng SW. Association of a Fruit and Vegetable Subsidy Program With Food Purchases by Individuals With Low Income in the US. JAMA Netw Open. 2021;4(8):e2120377. doi:10.1001/jamanetworkopen.2021.20377
    CONFLICT OF INTEREST: Conflict of Interest Disclosures: Dr Berkowitz reported receiving personal fees from the Aspen Institute outside the submitted work and receiving grants from the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health during the conduct of the study.
    READ MORE
    Original Investigation
    Public Health
    August 11, 2021

    Association of a Fruit and Vegetable Subsidy Program With Food Purchases by Individuals With Low Income in the US

    Author Affiliations
    • 1Division of General Medicine and Clinical Epidemiology, Department of Medicine, University of North Carolina at Chapel Hill School of Medicine
    • 2Center for Health Equity Research, Department of Social Medicine, University of North Carolina at Chapel Hill School of Medicine
    • 3Reinvestment Partners, Durham, North Carolina
    • 4Carolina Population Center, University of North Carolina at Chapel Hill
    • 5Department of Family Medicine and Community Health, Duke University School of Medicine, Durham, North Carolina
    • 6Department of Nutrition, Gillings School of Global Public Health, University of North Carolina at Chapel Hill
    JAMA Netw Open. 2021;4(8):e2120377. doi:10.1001/jamanetworkopen.2021.20377
    Key Points

    Question  Is a program that provides additional funds to Supplemental Nutrition Assistance Program (SNAP) beneficiaries for purchase of fruits and vegetables (SuperSNAP) associated with changes in food purchasing?

    Findings  In this cohort study of 667 SuperSNAP participants and 42 716 SNAP beneficiaries not enrolled in SuperSNAP, participation in SuperSNAP was associated with a significant increase in purchases of fruits and vegetables of $24.34 per month.

    Meaning  The findings suggest that a program that incentivizes the purchase of healthier foods for SNAP beneficiaries may be associated with improved healthfulness of food purchases.

    Abstract

    Importance  Food insecurity is associated with a less healthy diet. The Supplemental Nutrition Assistance Program (SNAP) is associated with reduced food insecurity, but benefit levels may be insufficient for beneficiaries to afford healthy foods.

    Objective  To evaluate whether participation in SuperSNAP, a program that provides an additional $40 per month for the purchase of fruits and vegetables with no added sugar, sodium, or fat to SNAP beneficiaries, is associated with changes in food purchasing.

    Design, Setting, and Participants  This longitudinal cohort study used data from transaction records of a large supermarket chain with approximately 500 stores located across North Carolina from October 2019 to April 2020. Participants were SNAP beneficiaries.

    Exposure  SuperSNAP participation.

    Main Outcomes and Measures  Monthly spending on all fruits, vegetables, legumes, and nuts (primary outcome); spending on less healthy food categories; and spending on sugar-sweetened beverages as 1 category of less healthy foods. Monthly data on purchases by SNAP beneficiaries before and during SuperSNAP participation were compared with data from SNAP beneficiaries not enrolled in the program who shopped at the same stores. Overlap weighting (a propensity score–based method) was used to account for confounding, and linear mixed-effects models were fitted with random effects to account for repeated measures and clustering by store.

    Results  The study included 667 SuperSNAP participants and 42 716 SNAP beneficiaries who did not use SuperSNAP but shopped in the same stores; 436 SuperSNAP participants had preintervention data and were included in the main analysis. SuperSNAP participation was associated with increased monthly purchases of fruits, vegetables, nuts, and legumes ($24.34; 95% CI, $23.96-$24.72; P < .001; 227.49 oz; 95% CI, 223.68-231.31 oz; P < .001). Only a small increase in spending on less healthy food categories compared with the SNAP beneficiaries who did not use SuperSNAP ($2.09; 95% CI, $1.45-$2.74; P < .001) was observed. As total spending increased, the proportion of total food and beverage spending on less healthy foods significantly decreased (difference, 4.25%; 95% CI, 4.02%-4.48%; P < .001). Monthly spending on sugar-sweetened beverages decreased (difference, $1.81; 95% CI, $1.39-$2.24; P < .001).

    Conclusions and Relevance  In this cohort study, participation in SuperSNAP was associated with meaningful increases in healthy food purchasing. Subsequent studies should investigate whether healthy food incentive programs improve health outcomes.

    Introduction

    Greater fruit and vegetable consumption and less sodium, sugar, and saturated fat consumption are associated with longer life and reduced risk of poor health.1-13 In 2019, more than 10% of US households were food insecure, meaning that they had uncertain access to enough food for an active, healthy life; this percentage has increased during the COVID-19 pandemic.14,15 Food insecurity is associated with the consumption of less healthy foods that are cheaper on a per-calorie basis than healthy foods.16-18 The Supplemental Nutrition Assistance Program (SNAP) is the largest program in the US to address food insecurity.19 Although SNAP is associated with reduced food insecurity, current SNAP benefit amounts are often insufficient to allow purchase of the recommended amounts of fruits and vegetables.20-22

    Healthy food incentive programs, which provide additional funds to facilitate the purchase of healthier foods, have emerged as a key intervention to reduce food insecurity and improve diet.23-26 In this study, we evaluated SuperSNAP,27 a program funded by a US Department of Agriculture grant (formerly Food Insecurity Nutrition Incentive and now known as the Gus Schumacher Nutrition Incentive Program),28 that was designed to improve diet quality for SNAP beneficiaries.

    Because the primary mechanism whereby SuperSNAP may improve health is through changes in food purchasing, evaluating whether changes in food purchasing occur would be informative. Proof of this concept would support further investigation into whether programs such as SuperSNAP have effects on diet quality and health. A recent report29 identified the following recommendations for food subsidy programs: electronic issuance, longer duration (>24 weeks) with repeated incentives, coverage of a broad selection of healthy foods (eg, including frozen and canned fruits and vegetables rather than only fresh produce), and redemption in stores (vs solely in farmers’ markets). Because SuperSNAP has several of these features, we used this program to evaluate whether providing incentives to purchase healthy foods is associated with changes in food purchasing. We hypothesized that SNAP beneficiaries who participated in SuperSNAP, compared with SNAP beneficiaries who did not participate in the program, would have increased fruit and vegetable purchasing.

    Methods
    Study Design, Setting, and Participants

    This retrospective cohort study was performed from October 2019 to April 2020 with a dynamic cohort, meaning that participants could enter and exit the cohort at different times. The primary source of data was the transaction records of a large supermarket chain with approximately 500 stores located across North Carolina. The University of North Carolina institutional review board did not consider this study of deidentified data to be human participant research; therefore, the requirement for informed consent was waived. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    The unit of analysis was the monthly transactions linked to a particular shopper identification (ID) number. The shopper ID number is a unique identifier available to all shoppers that can be used to obtain store discounts on purchases. The ID numbers were used to track SuperSNAP benefit use, and all SuperSNAP benefits were linked to a unique shopper ID number. For this study, we included all transaction data in a given month in which a shopper ID number was linked to a purchase that used any SNAP benefits during that month. SNAP benefits only needed to be used for 1 transaction (not every transaction) in a given month to be included in this study.

    SuperSNAP

    SuperSNAP provided $40 per month in additional funds to SNAP beneficiaries that could be spent on fresh, frozen, or canned fruits and vegetables with no added sugar, fat, or salt. SuperSNAP benefits could be redeemed at any location of a single supermarket chain. Eligible foods were programmed into the supermarket point-of-sale system, and when eligible foods were purchased, the cost of those foods was deducted from the available SuperSNAP balance at the point of purchase. SuperSNAP participants were adults (age, ≥18 years) who were current SNAP beneficiaries and identified by staff at their primary care clinic as likely to benefit from the program because of their health status (eg, owing to diabetes or obesity). Participants were enrolled by clinic staff at 9 clinics—primarily federally qualified health care centers—throughout North Carolina. Participants completed a brief enrollment survey with clinic staff; the survey included questions about household composition, food security (using the Hunger Vital Sign 2-item screener),30 and proportion of food shopping done at the supermarket chain. Access to the program expired if participants did not spend any SuperSNAP benefits for 2 consecutive months, and unused benefits at the end of each month rolled over for 2 months.

    Data Collection and Categorization of Spending

    The transaction data included every item bought during each shopping episode and the payment method. Eligible fruit and vegetable purchases made through SuperSNAP were identified based on a unique offer code applied at checkout. We were thus able to determine the source of food spending for each transaction (eg, SNAP, SuperSNAP, or out of pocket). All transactions within a given month were aggregated to produce monthly data for each shopper ID.

    On the basis of other ongoing work using large, national data sets on household purchases and nutritional label data at the barcode level,31,32 we were also able to categorize every purchased item as a nonfood item (eg, toilet paper, magazine, or tobacco) or a food item. Among food items, we further examined 3 specific categories as being particularly relevant for health outcomes (eTable 1 in the Supplement): (1) fruits, vegetables, legumes, and nuts; (2) less healthy foods including processed meats and processed seafood, desserts, sweet snacks, salty snacks, candy, chocolate, gum, sweeteners, and toppings; and (3) sugar-sweetened beverages.

    Outcomes

    The primary outcome for this study was total expenditure, in US dollars, from any spending source on all fruits, vegetables, legumes, and nuts. Secondary outcomes were spending on fruits, vegetables, nuts, and legumes in ounces and as a proportion of total spending on food and beverages. Because food incentives, even if restricted to healthier foods, may allow for resources to be spent on less healthy foods, thus negating the potential impact of a healthy food incentive, we included these secondary outcomes examining spending on less healthy food categories. Furthermore, because prior research has suggested that consumption of sugar-sweetened beverages, which is associated with poor health, may decrease when healthier foods are subsidized,25,33 we examined sugar-sweetened beverages as a subcategory of less healthy foods.

    Statistical Analysis

    We sought to analyze whether participation in SuperSNAP was associated with changes in food purchasing. We analyzed data on monthly purchases by SNAP beneficiaries before and during SuperSNAP participation compared with data from SNAP beneficiaries who never enrolled in the program and shopped at the same stores (ie, 2 groups were observed during 2 periods). The 2 periods were demarcated by an index date, which divided the study into periods before and after the index date. For those who participated in SuperSNAP, the index date was defined as the month in which SuperSNAP participation began. For the comparison group of SNAP beneficiaries who never enrolled in SuperSNAP, the index date was selected at random from a distribution that matched the distribution of SuperSNAP enrollment months. To help account for confounding, we used a propensity score–based analytic strategy called overlap weighting,34,35 which assigns an overlap weight equal to the estimated probability (ie, the propensity score) of being in the group in which the participant did not actually belong. Overlap weighting has the property of exactly balancing all covariates used to estimate the propensity score (when estimated with a logistic regression model).34 To calculate the propensity score, we used purchase record data organized by shopper ID number. Only data from before the index date were used to estimate the propensity score. To enhance internal validity, we excluded shopper ID numbers that did not have observations before the index date because we would not be able to estimate a propensity score for those shopper ID numbers. To estimate the propensity score, we fit logistic regression models that estimated the probability of SuperSNAP participation based on the variables listed in the eAppendix in the Supplement, encompassing absolute and relative measures of expenditures across different categories, absolute measures on volume or weight purchased across different categories, measures of payment source, and time measures.

    To evaluate whether there were changes in purchasing associated with SuperSNAP participation, we fit linear mixed models for the aforementioned outcomes. The unit of analysis was the shopper-month. The models included terms for SuperSNAP participation (1 or 0), time in relation to index date (0, before index date; 1, on or after index date), and a SuperSNAP × time interaction term to estimate whether a given outcome differed between those who participated in SuperSNAP and those who did not after the index date. To account for seasonality in shopping and secular trends (eg, the COVID-19 pandemic), we adjusted for the month and year of observation and the duration of follow-up time. Because follow-up time was a variable after the index date, it was not balanced by overlap weighting. Because the store at which an individual shopped may have provided information about geographic characteristics and socioeconomic status and because those who shopped at the same stores may have been more similar than those who shopped at different stores, we used a random intercept term for store ID (the store that received a plurality of expenditures for a given shopper ID in a given month). To account for repeated measurements within shopper ID numbers, we also included a random intercept term for shopper ID number.

    Although the overlap weighting approach used enhanced internal validity, this approach could also affect generalizability of the findings because SuperSNAP participants without pre-SuperSNAP data were excluded. Therefore, we conducted additional analyses that included all shopper ID numbers for shoppers who used SuperSNAP even if there were no pre-SuperSNAP data. For these analyses, we used the same modeling approach as described above but without weighting.

    Analyses were conducted using SAS, version 9.4 (SAS Institute Inc). A 2-sided P < .05 was considered to indicate statistical significance. The overlap-weighted analysis with the fruit, vegetable, nut, and legume spending outcome was considered primary, and other analyses were considered secondary.

    Results

    Overall, 824 shoppers were enrolled in SuperSNAP from October 2019 to April 2020. Of these, 667 used SuperSNAP benefits (81.0% SuperSNAP redemption rate) at some time during this period, and 436 had preintervention data and were thus included in the main analysis. The comparison group included 42 716 SNAP beneficiaries who never enrolled in SuperSNAP. The 667 shoppers who had used SuperSNAP benefits were included in the sensitivity analyses. Table 1 shows characteristics of those who were enrolled in SuperSNAP. The mean (SD) household size was 2.3 (1.7; range, 1-12), and 341 (91.7%) participants reported having experienced food insecurity. SuperSNAP participants who had pre-SuperSNAP data available were similar to those who did not have these data available.

    Table 2 shows purchasing characteristics before the index date. Before weighting, mean (SD) monthly total spending on food and beverages was higher in the comparison group than in the SuperSNAP group ($259.77 [$219.48] vs $224.57 [$192.60]; P = .004), but mean (SD) monthly spending on fruits, vegetables, nuts, and legumes ($25.21 [$26.23] vs $25.32 [$28.09]; P = .95) and mean (SD) monthly spending on sugar-sweetened beverages ($26.34 [$33.39] vs $25.75 [$33.21]; P = .77) were similar. In the weighted comparisons, all characteristics were exactly balanced. There was a small difference in the mean (SD) duration of follow-up between SuperSNAP participants and the comparison group (3.0 [1.5] months vs 2.7 [0.1] months; P < .001); thus, regression analyses adjusted for duration of follow-up.

    Intraclass correlation statistics showed that 93.2% of the variation in spending on fruits, vegetables, nuts, and legumes was explained at the level of the shopper and 4.5% was explained at the level of the store. Table 3 shows differences in outcomes between SuperSNAP participants and shoppers in the comparison group during the period after the index date. The total purchases increased by $25.10 (95% CI, $22.79-$27.41; P < .001) in the SuperSNAP group compared with the group that never used the program.

    When categories of purchases were examined, spending on fruits, vegetables, nuts, and legumes increased by $24.34 (95% CI, $23.96-$24.72; P < .001) in the SuperSNAP group (complete model results are given in eTable 2 in the Supplement). The proportion of total food and beverage spending on fruits, vegetables, nuts, and legumes increased by 14.25% (95% CI, 14.04%-14.47%; P < .001) in the SuperSNAP group. This amounted to a monthly increase of 227.49 oz (95% CI, 223.68-231.31 oz; P < .001). In a mean household size of 2.24 individuals in the analytic sample, this represented 3.4 oz, or approximately 1 serving per person per day.11

    Spending on less healthy food categories increased minimally in the SuperSNAP group compared with the group that never used the program (difference, $2.09; 95% CI, $1.45-$2.74; P < .001). As total spending increased, proportion of total food and beverage spending on less healthy food decreased by 4.25% (95% CI, 4.02%-4.48%; P < .001).

    In the analysis of sugar-sweetened beverages, total spending on food and beverages decreased by $1.81 (95% CI, $1.39-$2.24; P < .001) and spending as a proportion of total spending on food and beverages decreased by 3.83% (95% CI, 3.61%-4.04%; P < .001).

    Results of sensitivity analyses in which the entire sample was examined were similar to results of the main analyses (eTable 3 in the Supplement). In the sensitivity analyses, spending on fruits, vegetables, nuts, and legumes was greater among SuperSNAP participants compared with those who never used the program (difference, $21.98; 95% CI, $20.04-$23.93; P < .001). Spending on less healthy foods was not substantially different, with a mean difference of −$0.64 (95% CI, −$4.54 to $3.26; P = .75). Spending on sugar-sweetened beverages was estimated to have decreased by $3.95 (95% CI, $1.74-$6.16; P = .001) or 4.5% (95% CI, 3.3%-5.7%; P < .001).

    Discussion

    In this longitudinal cohort study, participation in SuperSNAP was associated with increased purchasing of fruits and vegetables. Propensity score–based analyses that incorporated data before and after SuperSNAP enrollment revealed that spending on fruits and vegetables was approximately $24 per month greater after SuperSNAP enrollment, with no meaningful increase in spending on less healthy foods and decreased spending on sugar-sweetened beverages. Sensitivity analyses that included the entire sample of SuperSNAP users showed similar findings.

    This study makes 3 notable contributions to the literature on fruit and vegetable subsidies. First, prior studies24,26,36-39 have found that incentives may be associated with meaningful increases in purchasing of fruits and vegetables, but these studies typically examined programs based in smaller local food stores and farmers’ markets or that used paper-based enrollment and vouchers. The present study used data from a large supermarket chain and a streamlined online enrollment process and electronic issuance, all of which increase the potential for high-volume scalability. Second, in the present study, increases in healthy food purchasing occurred in a setting with a variety of less healthy and healthy food options, suggesting that programs such as SuperSNAP may be associated with improved health behaviors. Third, we found no meaningful increases in unhealthy food purchases (and decreased spending on sugar-sweetened beverages). Enrollment in SuperSNAP through clinics may have encouraged mental accounting of the $40 incentive, helping to restrict changes to purchases of healthier foods.40,41 Alternatively, if demand for healthier food among participants were greater than baseline resources, additional resources might be preferentially directed to meet this demand.

    The findings of this study suggest directions for future work. Examination of how food incentives and subsidies may change diet quality, health outcomes, health care utilization, and health care cost is needed. Policies, such as the Gus Schumacher Nutrition Incentive Program, may play a key role in this research. Previous programs42 did not have a great emphasis on evaluation of potential health benefits, but examining whether these subsidy programs are associated with health benefits will be important ongoing work. Other future directions of research include examining potential program benefits for all members of the household (not just the enrolled individual, because food is shared within households), establishing the optimal amount and duration of these benefits, and examining possible heterogeneous treatment effects by social circumstances and clinical characteristics. In addition, how associations found in the study may apply to other samples and to SNAP beneficiaries overall should be evaluated.

    Strengths and Limitations

    This study has strengths. This study used objective food purchasing data and compared SuperSNAP participants with a large number of SNAP beneficiaries throughout North Carolina. The consistency of results between 2 different analytic approaches, one of which emphasized internal validity and the other generalizability, suggests that the results observed in this study are likely to reflect the experience of SuperSNAP participants more broadly.

    This study also has limitations. We had access only to shopper ID–level purchasing data; thus, we were not able to assess how changes in purchasing translated to changes in diet quality for individuals (if they did at all). However, we believe that a change in purchasing is a first step in the pathway to change in diet quality and improved health. Second, follow-up was relatively brief, and whether changes observed were sustained is not known. Third, how the results might have been affected by food purchases made outside the participating supermarket chain is unclear. Because we only had transaction data and not demographic, social, or clinical data, the possibility of unmeasured confounding cannot be ruled out. However, to confound the association between SuperSNAP participation and food purchases, the unmeasured factors would have had to affect food purchasing. Therefore, the preintervention data used for the propensity score analysis should help account for unmeasured factors because the associations, if any, of time-invariant demographic, social, and clinical characteristics with food purchasing would be reflected in food purchases before the index date. Furthermore, SuperSNAP participants were enrolled by clinicians for the purpose of receiving a benefit intended to improve diet-related chronic disease. Thus, the participants likely represent a unique subset of the overall population of SNAP beneficiaries who may be most likely to benefit from a program such as SuperSNAP. Whether the results observed in the group selected for SuperSNAP participation could be generalized to the broader population of SNAP beneficiaries is unclear. Reasons the results may not be generalizable include the characteristics of those enrolled and the setting of clinic-based enrollment, which may have worked synergistically with the cash value of the benefit provided to influence purchases of healthy food. Additional studies should explore these issues, potentially by using cash-benchmarked study designs that can isolate the impact of clinic-based approaches from the cash value of the benefit offered.43

    Conclusions

    In this cohort study, SuperSNAP, which provides a fruit and vegetable subsidy to SNAP beneficiaries, was associated with increased purchasing of fruits and vegetables, a minimal increase in purchasing of less healthy foods, and decreased purchasing of sugar-sweetened beverages. Future studies should examine the potential effects of subsidies for the purchase of fruits and vegetable on diet quality, food security, health, and health care utilization. Given the high prevalence of food insecurity and its association with poor health outcomes,44 fruit and vegetable subsidies may be associated with important public health benefits in the US.

    Back to top
    Article Information

    Accepted for Publication: June 2, 2021.

    Published: August 11, 2021. doi:10.1001/jamanetworkopen.2021.20377

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Berkowitz SA et al. JAMA Network Open.

    Correction: This article was corrected on December 13, 2021, to fix data errors in the Abstract, Key Points, Methods and Results sections, Tables 2 and 3, and the Supplement.

    Corresponding Author: Seth A. Berkowitz, MD, MPH, Division of General Medicine and Clinical Epidemiology, Department of Medicine, University of North Carolina at Chapel Hill School of Medicine, 5034 Old Clinic Building, CB# 7110, Chapel Hill, NC 27599 (seth_berkowitz@med.unc.edu).

    Author Contributions: Dr Berkowitz had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Berkowitz, Price, Ng.

    Acquisition, analysis, or interpretation of data: Berkowitz, Curran, Hoeffler, Henderson, Ng.

    Drafting of the manuscript: Berkowitz, Price, Ng.

    Critical revision of the manuscript for important intellectual content: Curran, Hoeffler, Henderson, Price, Ng.

    Statistical analysis: Berkowitz, Henderson, Price, Ng.

    Obtained funding: Berkowitz, Hoeffler, Ng.

    Administrative, technical, or material support: Curran, Hoeffler, Ng.

    Supervision: Berkowitz, Curran.

    Conflict of Interest Disclosures: Dr Berkowitz reported receiving personal fees from the Aspen Institute outside the submitted work and receiving grants from the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health during the conduct of the study. Mr Curran reported receiving grants from the Rockefeller Foundation and Blue Cross Blue Shield of North Carolina Foundation outside the submitted work. No other disclosures were reported.

    Funding/Support: This work was supported by grant UL1TR002489 from the North Carolina Translational and Clinical Sciences Institute and the Center for Health Equity Research, School of Medicine; grant K23DK109200 from the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (Dr Berkowitz); and grant P2C HD050924 from the Carolina Population Center (Dr Ng). SuperSNAP is a project of Reinvestment Partners funded by the US Department of Agriculture’s National Institute of Food and Agriculture, but the US Department of Agriculture did not provide funding for this study.

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

    Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the US Department of Agriculture.

    Additional Contributions: Tanya Wolfram, MPP, was an employee of Reinvestment Partners. She received no specific compensation for the project. Christy Wu Kui and Amy Lo (University of North Carolina) provided research assistance, and Leigh Ann Van Acker, BA (University of North Carolina), provided administrative assistance. Ms Kui received no compensation for the project. Ms Lo and Ms Van Acker were compensated for their work. We thank the participating supermarket chain for sharing the transaction data for this evaluation.

    References
    1.
    Mokdad  AH, Ballestros  K, Echko  M,  et al; US Burden of Disease Collaborators.  The state of US health, 1990-2016: burden of diseases, injuries, and risk factors among US states.   JAMA. 2018;319(14):1444-1472. doi:10.1001/jama.2018.0158 PubMedGoogle Scholar
    2.
    Sotos-Prieto  M, Bhupathiraju  SN, Mattei  J,  et al.  Association of changes in diet quality with total and cause-specific mortality.   N Engl J Med. 2017;377(2):143-153. doi:10.1056/NEJMoa1613502 PubMedGoogle ScholarCrossref
    3.
    Sotos-Prieto  M, Bhupathiraju  SN, Mattei  J,  et al.  Changes in diet quality scores and risk of cardiovascular disease among US men and women.   Circulation. 2015;132(23):2212-2219. doi:10.1161/CIRCULATIONAHA.115.017158 PubMedGoogle ScholarCrossref
    4.
    Sugiyama  T, Shapiro  MF.  The growing socioeconomic disparity in dietary quality: mind the gap.   JAMA Intern Med. 2014;174(10):1595-1596. doi:10.1001/jamainternmed.2014.3048 PubMedGoogle ScholarCrossref
    5.
    Wang  DD, Leung  CW, Li  Y,  et al.  Trends in dietary quality among adults in the United States, 1999 through 2010.   JAMA Intern Med. 2014;174(10):1587-1595. doi:10.1001/jamainternmed.2014.3422 PubMedGoogle ScholarCrossref
    6.
    Wang  PY, Fang  JC, Gao  ZH, Zhang  C, Xie  SY.  Higher intake of fruits, vegetables or their fiber reduces the risk of type 2 diabetes: a meta-analysis.   J Diabetes Investig. 2016;7(1):56-69. doi:10.1111/jdi.12376 PubMedGoogle ScholarCrossref
    7.
    Alissa  EM, Ferns  GA.  Dietary fruits and vegetables and cardiovascular diseases risk.   Crit Rev Food Sci Nutr. 2017;57(9):1950-1962. doi:10.1080/10408398.2015.1040487PubMedGoogle Scholar
    8.
    Lelong  H, Blacher  J, Baudry  J,  et al.  Individual and combined effects of dietary factors on risk of incident hypertension: prospective analysis from the NutriNet-Santé cohort.   Hypertension. 2017;70(4):712-720. doi:10.1161/HYPERTENSIONAHA.117.09622PubMedGoogle ScholarCrossref
    9.
    Micha  R, Shulkin  ML, Peñalvo  JL,  et al.  Etiologic effects and optimal intakes of foods and nutrients for risk of cardiovascular diseases and diabetes: systematic reviews and meta-analyses from the Nutrition and Chronic Diseases Expert Group (NutriCoDE).   PLoS One. 2017;12(4):e0175149. doi:10.1371/journal.pone.0175149 PubMedGoogle Scholar
    10.
    He  FJ, Nowson  CA, Lucas  M, MacGregor  GA.  Increased consumption of fruit and vegetables is related to a reduced risk of coronary heart disease: meta-analysis of cohort studies.   J Hum Hypertens. 2007;21(9):717-728. doi:10.1038/sj.jhh.1002212 PubMedGoogle ScholarCrossref
    11.
    Dietary Guidelines Advisory Committee. Scientific report of the 2020 Dietary Guidelines Advisory Committee: advisory report to the Secretary of Agriculture and the Secretary of Health and Human Services. US Dept of Agriculture, Economic Research Service; 2020. Accessed July 6, 2021. https://www.dietaryguidelines.gov/sites/default/files/2020-07/ScientificReport_of_the_2020DietaryGuidelinesAdvisoryCommittee_first-print.pdf
    12.
    Poti  JM, Mendez  MA, Ng  SW, Popkin  BM.  Is the degree of food processing and convenience linked with the nutritional quality of foods purchased by US households?   Am J Clin Nutr. 2015;101(6):1251-1262. doi:10.3945/ajcn.114.100925 PubMedGoogle ScholarCrossref
    13.
    Malik  VS, Popkin  BM, Bray  GA, Després  JP, Hu  FB.  Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk.   Circulation. 2010;121(11):1356-1364. doi:10.1161/CIRCULATIONAHA.109.876185 PubMedGoogle ScholarCrossref
    14.
    Coleman-Jensen  A, Rabbitt  MP, Gregory  CA, Singh  A. Household food security in the United States in 2019. Accessed November 3, 2020. https://www.ers.usda.gov/publications/pub-details/?pubid=99281
    15.
    Feeding America. The impact of the coronavirus on local food insecurity. Accessed May 22, 2020. https://www.feedingamerica.org/sites/default/files/2020-05/Brief_Local%20Impact_5.19.2020.pdf
    16.
    Morales  ME, Berkowitz  SA.  The relationship between food insecurity, dietary patterns, and obesity.   Curr Nutr Rep. 2016;5(1):54-60. doi:10.1007/s13668-016-0153-y PubMedGoogle ScholarCrossref
    17.
    Seligman  HK, Schillinger  D.  Hunger and socioeconomic disparities in chronic disease.   N Engl J Med. 2010;363(1):6-9. doi:10.1056/NEJMp1000072 PubMedGoogle ScholarCrossref
    18.
    Seligman  HK, Berkowitz  SA.  Aligning programs and policies to support food security and public health goals in the United States.   Annu Rev Public Health. 2019;40:319-337. doi:10.1146/annurev-publhealth-040218-044132 PubMedGoogle ScholarCrossref
    19.
    Food and Nutrition Service, US Department of Agriculture. Supplemental Nutrition Assistance Program (SNAP). Accessed February 19, 2020. https://www.fns.usda.gov/snap/supplemental-nutrition-assistance-program
    20.
    Canning  P, Stacy  B. The Supplemental Nutrition Assistance Program (SNAP) and the economy: new estimates of the SNAP multiplier. Accessed July 3, 2020. https://www.ers.usda.gov/publications/pub-details/?pubid=93528
    21.
    Gundersen  C.  Food insecurity is an ongoing national concern.   Adv Nutr. 2013;4(1):36-41. doi:10.3945/an.112.003244 PubMedGoogle ScholarCrossref
    22.
    Carlson S. More adequate SNAP benefits would help millions of participants better afford food. Center on Budget and Policy Priorities. Accessed July 3, 2020. https://www.cbpp.org/research/food-assistance/more-adequate-snap-benefits-would-help-millions-of-participants-better
    23.
    Mozaffarian  D, Liu  J, Sy  S,  et al.  Cost-effectiveness of financial incentives and disincentives for improving food purchases and health through the US Supplemental Nutrition Assistance Program (SNAP): a microsimulation study.   PLoS Med. 2018;15(10):e1002661. doi:10.1371/journal.pmed.1002661 PubMedGoogle Scholar
    24.
    Olsho  LE, Klerman  JA, Wilde  PE, Bartlett  S.  Financial incentives increase fruit and vegetable intake among Supplemental Nutrition Assistance Program participants: a randomized controlled trial of the USDA Healthy Incentives Pilot.   Am J Clin Nutr. 2016;104(2):423-435. doi:10.3945/ajcn.115.129320 PubMedGoogle ScholarCrossref
    25.
    Berkowitz  SA, O’Neill  J, Sayer  E,  et al. Health center–based community-supported agriculture: an RCT. Am J Prev Med. 2019;57(6 suppl 1):S55-S64. doi:10.1016/j.amepre.2019.07.015
    26.
    Yoder  AD, Proaño  GV, Handu  D.  Retail nutrition programs and outcomes: an evidence analysis center scoping review.   J Acad Nutr Diet. Published online November 20, 2020. doi:10.1016/j.jand.2020.08.080PubMedGoogle Scholar
    27.
    Reinvestment Partners. SuperSNAP. Accessed November 4, 2020. https://reinvestmentpartners.org/what-we-do/produce-prescriptions/supersnap-1.html
    28.
    National Institute of Food and Agriculture, US Department of Agriculture. Gus Schumacher Nutrition Incentive Program. Accessed November 3, 2020. https://nifa.usda.gov/program/gus-schumacher-nutrition-incentive-grant-program
    29.
    Healthy Food America. Healthy food pricing incentives: a systematic review of current evidence. Accessed November 4, 2020. https://www.healthyfoodamerica.org/healthy_food_pricing_incentives_a_systematic_review_of_current_evidence
    30.
    Hager  ER, Quigg  AM, Black  MM,  et al.  Development and validity of a 2-item screen to identify families at risk for food insecurity.   Pediatrics. 2010;126(1):e26-e32. doi:10.1542/peds.2009-3146 PubMedGoogle ScholarCrossref
    31.
    Lacko  A, Ng  SW, Popkin  B.  Urban vs. rural socioeconomic differences in the nutritional quality of household packaged food purchases by store type.   Int J Environ Res Public Health. 2020;17(20):7637. doi:10.3390/ijerph17207637 PubMedGoogle ScholarCrossref
    32.
    Ng  SW, Hollingsworth  BA, Busey  EA, Wandell  JL, Miles  DR, Poti  JM.  Federal nutrition program revisions impact low-income households’ food purchases.   Am J Prev Med. 2018;54(3):403-412. doi:10.1016/j.amepre.2017.12.003 PubMedGoogle ScholarCrossref
    33.
    Berkowitz  SA, Delahanty  LM, Terranova  J,  et al.  Medically tailored meal delivery for diabetes patients with food insecurity: a randomized cross-over trial.   J Gen Intern Med. 2019;34(3):396-404. doi:10.1007/s11606-018-4716-z PubMedGoogle ScholarCrossref
    34.
    Thomas  LE, Li  F, Pencina  MJ.  Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial.   JAMA. 2020;323(23):2417-2418. doi:10.1001/jama.2020.7819 PubMedGoogle ScholarCrossref
    35.
    Li  F, Morgan  KL, Zaslavsky  AM.  Balancing covariates via propensity score weighting.   J Am Stat Assoc. 2018;113(521):390-400. doi:10.1080/01621459.2016.1260466 PubMedGoogle ScholarCrossref
    36.
    Gittelsohn  J, Trude  ACB, Kim  H.  Pricing strategies to encourage availability, purchase, and consumption of healthy foods and beverages: a systematic review.   Prev Chronic Dis. 2017;14:E107. doi:10.5888/pcd14.170213 PubMedGoogle Scholar
    37.
    Engel  K, Ruder  EH.  Fruit and vegetable incentive programs for Supplemental Nutrition Assistance Program (SNAP) participants: a scoping review of program structure.   Nutrients. 2020;12(6):1676. doi:10.3390/nu12061676 PubMedGoogle ScholarCrossref
    38.
    Moran  A, Thorndike  A, Franckle  R,  et al.  Financial incentives increase purchases of fruit and vegetables among lower-income households with children.   Health Aff (Millwood). 2019;38(9):1557-1566. doi:10.1377/hlthaff.2018.05420 PubMedGoogle ScholarCrossref
    39.
    Seguin  RA, Morgan  EH, Hanson  KL,  et al.  Farm Fresh Foods for Healthy Kids (F3HK): an innovative community supported agriculture intervention to prevent childhood obesity in low-income families and strengthen local agricultural economies.   BMC Public Health. 2017;17(1):306. doi:10.1186/s12889-017-4202-2 PubMedGoogle ScholarCrossref
    40.
    Mancino  L, Guthrie  J, Just  DR.  Overview: exploring ways to encourage healthier food purchases by low-income consumers—lessons from behavioral economics and marketing.   Food Policy. 2018;79:297-299. doi:10.1016/j.foodpol.2018.03.007 Google ScholarCrossref
    41.
    Audet AJ, Zezza MA. How behavioral economics can advance the design of effective clinician incentive programs. Health Affairs blog. Accessed December 4, 2020. https://www.healthaffairs.org/do/10.1377/hblog20150915.050556/full/
    42.
    Food and Nutrition Service, US Department of Agriculture. FINI Grant Program. Accessed July 6, 2021. https://www.fns.usda.gov/snap/FINI-Grant-Program
    43.
    Berkowitz  SA, Edwards  ST, Polsky  D.  Cash benchmarking for integrated health care and human services interventions: finding the value added.   Health Aff (Millwood). 2020;39(4):582-586. doi:10.1377/hlthaff.2019.01579 PubMedGoogle ScholarCrossref
    44.
    Gundersen  C, Ziliak  JP.  Food insecurity and health outcomes.   Health Aff (Millwood). 2015;34(11):1830-1839. doi:10.1377/hlthaff.2015.0645PubMedGoogle ScholarCrossref
    ×