Local Food System Approaches to Address Food and Nutrition Security among Low-Income Populations: A Systematic Review

Food and nutrition insecurity disproportionately impact low-income households in the United States, contributing to higher rates of chronic diseases among this population. Addressing this challenge is complex because of various factors affecting the availability and accessibility of nutritious food. Short value chain (SVC) models, informally known as local food systems, offer a systemic approach that aims to optimize resources and align values throughout and beyond the food supply chain. Although specific SVC interventions, such as farmers markets, have been studied individually, a comprehensive review of SVC models was pursued to evaluate their relative impact on food security, fruit and vegetable intake, diet quality, health-related markers, and barriers and facilitators to participation among low-income households. Our systematic literature search identified 37 articles representing 34 studies from 2000–2020. Quantitative, qualitative, and mixed-method studies revealed that farmers market interventions had been evaluated more extensively than other SVC models (i.e., produce prescription programs, community-supported agriculture, mobile markets, food hubs, farm stands, and farm-to-school). Fruit and vegetable intake was the most measured outcome; other outcomes were less explored or not measured at all. Qualitative insights highlighted common barriers to SVC use, such as lack of program awareness, limited accessibility, and cultural incongruence, whereas facilitators included health-promoting environments, community cohesion, financial incentives, and high-quality produce. Social marketing and dynamic nutrition education appeared to yield positive program outcomes. Financial incentives were used in many studies, warranting further investigation into optimal amounts across varying environmental contexts. SVC models are increasingly germane to national goals across the agriculture, social, and health care sectors. This review advances the understanding of key knowledge gaps related to their implementation and impact; it emphasizes the need for research to analyze SVC potential comprehensively across the rural-urban continuum and among diverse communities through long-term studies of measurable health impact and mixed-method studies investigating implementation best practices. This trial was registered at PROSPERO as CRD42020206532.


Introduction
In 2021, >10% of households in the United States were considered food insecure, meaning that they lacked access to sufficient food for an active, healthy life [1].Low-income households experience greater rates of food insecurity: 32% of households with incomes below the federal poverty line experienced food insecurity in 2021 [1].Low-income and food-insecure adults experience higher rates of chronic diseases, such as type 2 diabetes [2][3][4], heart disease [3,5], mental health conditions, such as depression [6,7], and lower quality of life (QoL) [8,9].These disparities have been attributed, in part, to poor diet quality.The diets of low-income and food-insecure populations are notably low in fruits and vegetables (FVs) and are of significantly less nutritional quality than the diets of those with higher socioeconomic status [10,11].It is well-established that insufficient resources-financial and otherwise-are a critical barrier to healthy food access and intake; these resources include lack of transportation [12], high housing and food costs [13,14], being under or unemployed [13], and having low assets [15].
Social safety-net programs, such as the Supplemental Nutrition Assistance Program (SNAP), aim to address food insecurity among low-income individuals.Although food security status is positively impacted by program participation, some analyses have found that poorer diet quality persists, remaining lower among participants relative to their higher-income nonparticipant counterparts [16].This and other factors have driven greater investment in promoting diet quality among federal food assistance beneficiaries.For example, the Gus Schumacher Nutrition Incentive Program (GusNIP) awards grant funding to nonprofit organizations and government agencies for projects that incentivize FV purchases by SNAP participants [17], and the 2022 White House Conference on Hunger, Nutrition, and Health resulted in explicit calls for "food is medicine" interventions designed to treat or prevent diet-related health conditions via healthy food "prescriptions" [18].
Parallel to the increased focus on nutrition incentives and "food is medicine" interventions, "nutrition security"-a concept that embodies goals related to food security, diet quality, and health equity-has emerged as a necessary national target.The formal definition of nutrition security, as defined by the USDA, is "having consistent access, availability, and affordability of food and beverages that promote well-being and prevent (and if needed, treat) disease, particularly among racial/ethnic minority populations, lower income populations, and rural and remote populations" [19].Several federal agencies and efforts have committed to moving beyond food security and dedicating resources to combat nutrition insecurity: the National Strategy drafted after the 2022 White House Conference on Hunger, Nutrition, and Health includes multiple pillars focused on nutrition security [18]; the USDA announced a 4-pillar strategic approach to tackle nutrition security [20]; and the 2020-2030 Strategic Plan for NIH Nutrition Research underscores the significance of nutrition security [21].
Research suggests that food-insecure households often sacrifice food quality and variety in favor of quantity (e.g., consuming low-cost, energy-dense, and nutrient-poor foods) [22].An emphasis on nutrition security warrants improvements in access to nutritious foods and adequate health services to prevent and treat disease, shifting away from a more calorie-centric focus to one that considers the nutritional content of foods [23].Contrary to colloquial narratives, there is some evidence that food-insecure households prefer more healthful foods (i.e., fruits, vegetables, and high-quality proteins) when given a choice, suggesting that lack of resources, not knowledge or desire for well-being, may be a key factor standing in the way of improved diet quality [24,25].Given the interconnectedness of diet quality, food insecurity, and chronic disease, the shift to nutrition security holds promise for enhancing innovation in clinical practice and public policy while also advancing health equity.
Addressing food and nutrition insecurity within the United States remains a challenge given the complexity of determinants impacting the food supply, particularly access to nutritious foods.Such intersecting determinants necessitate a systems approach that leverages resources and aligns values across the food supply chain.Short value chain (SVC) models of healthy food access-informally known as local food systems-fit this vision.A food value chain is "a business model in which producers and buyers of agricultural products form strategic alliances with partners along the supply chain to enhance financial returns through product differentiation that advances social or environmental values" and embodies values of "transparency, strategic collaboration, and dedication to authenticity" [26].Although traditional food supply chains may reflect some of the same operational activities as value chains, these models are unique because of their emphasis on shared missions and operational values.These missions may encompass healthy food access, farm viability, and environmental stewardship [26].
SVC models, such as farmers markets (FMs) and communitysupported agriculture (CSA), show promise for influencing key dietary and health outcomes among low-income consumers.FMs, for example, can be a source of healthy food products to SNAP and Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) recipients via the use of the Electronic Benefit Transaction system and incentive program vouchers, respectively.Patronage of FM is associated with increased food security status and increased FV consumption among SNAP participants [27].CSA participation has resulted in increased vegetable intake [28], decreased frequency of doctor's visits and expenditures at pharmacies, and improved healthy eating behaviors (e.g., eating salads and preparing dinner at home) [29].
Currently, a systematic review of the literature on all types of SVC models and their various impacts and implementation challenges does not exist.A 2016 systematic review of FM use among low-income consumers found there to be limited use of shared methods and metrics across included studies, limiting the broader understanding of factors that influence FM use [30].A 2020 scoping review synthesized factors that may impact effectiveness of FV incentive programs for current SNAP participants, although most included studies were conducted at FM [31].An additional scoping review broadly examined interventions targeting SNAP beneficiaries and their reported impact on diet and nutrition-related outcomes [32].These reviews, although valuable, occupied a narrower scope, either focusing on a particular low-income group of interest (e.g., SNAP participants) or on 1 type of intervention (e.g., FM-based programs).Given the inherent overlap between and increasingly common integration of SVC models in single interventions, an encompassing review of all SVC models is likely to be helpful in informing future research, practice, and policy.
The purpose of this systematic review was to evaluate whether participation in SVC models of healthy food access influenced food security status, FV intake, total diet quality, and health-related markers and outcomes among low-income households in the United States.The authors also aimed to understand barriers to or facilitators of participant engagement with SVC models in the United States.Given the novelty of the nutrition security construct in the United States, for which no validated measures existed until 2022 [33] and none are yet widely accepted, this review focused on measures of diet quality and food security (among other outcomes) for the quantitative synthesis and interrogated the health equity potential of each model based on qualitative assessments of their accessibility.

Methods
This review adhered to the PRISMA guidelines.The protocol was developed by 2 coauthors (HH and JAG) in consultation with a library sciences expert and feedback from 1 other author (AB).This protocol was registered with PROSPERO (CRD42020206532) and is available for review.

Search strategy
Three major topical domains-disparities, SVCs, and food-were used to develop the search.Each domain included a series of keywords and Medical Subject Heading (MeSH) terms.A search strategy was initially prepared for PubMed and adapted for each database by the research librarian.Database-specific indexing terms were included when applicable.Detailed information regarding search terms and search strings is provided in Appendix A. The search was not restricted to the outcomes of interest to afford the most comprehensive search possible.Articles with no relevant outcomes were excluded during the full-text screening process.
Literature published in the English language and in full-text from 2000 to 2020 was searched and accessed via the following electronic databases by 1 author (AB) for upload into Covidence: Agricola, Center for Agriculture and Biosciences International (CABI Abstracts), Cumulative Index of Nursing and Allied Health (CINAHL), Embase, Public Affairs Index, PubMed, Scopus, SocINDEX (a database of sociological scholarship, including social work), and Web of Science.Because of the expansive scope of the search and the extent of relevant literature and reports accessed via these databases, the decision was made not to search for other potential sources of gray literature.

Selection of articles
Articles were eligible for inclusion if they reported on !1 SVC model designed to the following: 1) minimize the disconnect between farms and consumers by reducing 1 or more "middle" portions of the traditional food supply chain and 2) leverage local or regional sources of healthy food (e.g., FVs).Such SVC models may include but are not limited to, FM, produce prescription (PRx), mobile market (MM), CSA, farm-to-school (FTS), farm stand (FSt), and food hub (FH) models.Included studies had to either evaluate the effects of these models on !1 diet-and health-related outcome (for quantitative studies) or explore barriers to and facilitators of engagement with such models (for qualitative studies).All studies had to focus on households within the United States considered low-income (i.e., 185% of the current Federal Poverty Level or as indicated by study authors).We relied on authors' definitions and descriptions to determine the use of SVC models, and there had to be explicit mention that the produce used in the interventions was sourced locally.If this was unclear, we attempted to contact the authors via email to clarify.If no clarification was received, the article was excluded from the review.
For the quantitative portion of this analysis, randomized controlled trials (RCTs), nonrandomized controlled studies (e.g., controlled prepost studies), and quasi-experimental studies designed to afford causal inference met our inclusion criteria.To enable a comparison between individuals and households, with some compared with no exposure to the eligible interventions, studies had to include a control or comparison group.Outcomes of interest were food security status (as measured by any version of the USDA Economic Research Service's food security survey [34], The Hunger Vital Sign (Children's HealthWatch team, Boston, MA, USA) screening tool [35], or the 1-item screening question included in Safe Environment for Every Kid screener [36]); FV intake (as measured by the National Cancer Institute Fruit and Vegetable Screener [37] or other validated measures); and total diet quality (as measured by the Healthy Eating Index [38] or other validated measures).Secondary outcomes included anthropometric measures (e.g., BMI (in kg/m 2 ), weight and waist circumference); biomarkers of health (e.g., blood pressure, total cholesterol, HDL cholesterol, cholesterol ratio, triglycerides, and fasting glucose); health outcomes (e.g., chronic disease diagnoses); and QoL indicators (as measured by WHO QoL-BREF (the abbreviated version of the WHO's 100-item QoL survey) [39], QoL10 [40], or other validated measures).
For the qualitative portion of our analysis, any study that collected data via focus groups or in-depth interviews was included (regardless of the overall study design).Extraction and analysis were focused on the synthesis of insights regarding participant barriers to and facilitators of intervention engagement.

Screening
Six reviewers (HH, AB, KIP, KG, JAG, and KKG) screened titles and abstracts for eligibility and inclusion using Covidence.Duplicates were removed.Two reviewers were required to screen each title and abstract independently.To settle discrepancies between 2 reviewers, a third reviewer (KG) was consulted to make a blinded, final determination.After the initial screening process, the full text of potentially eligible articles was obtained and screened in more detail.The Population, Intervention, Comparator, and Outcome screening guide is provided in Appendix B.

Data extraction
For included studies, relevant outcomes from quantitative studies or findings from qualitative studies were compiled into an Excel document.Data from all included studies were extracted by 9 team members (KG, KKG, KJ, KIP, ECK, EL, YL, KA-M, and HH) and checked for accuracy and completeness by 2 team members (KG and KKG).Extracted data included the name of the first author, year of publication, methodology (qualitative, quantitative, or mixed method), study design, study objective, type of SVC intervention, intervention components, geographic area (rural compared with urban), sample size, author criteria for (or definition of) low-income, and key quantitative and/or qualitative findings.

Risk of bias assessment
The risk of bias was assessed for each study.Quantitative studies were appraised using the National Institute of Heart, Lung, and Blood Quality Assessment Tools for Controlled Intervention Studies, Observational Cohort Studies, and Case-Control Studies [41].Quality rating of included studies was completed independently by 1 of 7 reviewers (KG, KKG, KA-M, KIP, KJ, ECK, EL, and YL).Secondary reviews-i.e., detailed checking and confirmation of each criterion and overall score-were completed by 3 reviewers (KG, KKG, and KA) such that all studies were reviewed by !2 team members.Disagreements were resolved via discussion between the primary and secondary reviewers, with a third author serving as the tiebreaker when necessary.Qualitative studies were appraised using The Standards for Reporting Qualitative Research (SRQR) devised by O'Brien et al., (2014) [42].

Data synthesis
Extracted data were reviewed iteratively by the author team to produce a comprehensive summary of included studies and their attributes in table form.Upon completion of the table (Table 1), data synthesis involved additional iterative, teambased reviews of the extracted content over a series of several intensive meetings to come to consensus on key insights across the following 6 domains: 1) the types of data and study designs used to evaluate SVCs to date, 2) the methodologic quality of studies to date, 3) intervention subtypes and key intervention characteristics, 4) the role of nutrition education (given the field's historic focus on education as a key strategy for individual behavior change), 5) whether the studied interventions demonstrated impact across the outcomes of focus, and 6) what we know to date about barriers to and facilitators of program engagement with the studied SVC models.For the reader's convenience, the results section was organized accordingly, and all insights were summarized narratively.For articles reporting qualitative data, the manuscripts were uploaded into NVivo (Lumivero) for coding of themes regarding barriers to and facilitators of SVC engagement by target participants.

Results
The search identified a total of 24,001 potentially relevant studies that were imported into Covidence for screening and review.After 10,138 duplicates were removed, the authors screened 13,863 titles and abstracts.We identified 512 potentially eligible studies for full-text review; a large number of articles progressed to full-text review as abstracts often did not include the detail necessary to discern whether an SVC model, with explicit local sourcing of food, was employed.Following full-text review, 37 articles representing 34 distinct studies were identified for inclusion.Common reasons for exclusions were wrong outcomes/focus (n ¼ 130), lack of full text (i.e., abstract only) (n ¼ 124), and wrong study design (n ¼ 90).The PRISMA flow diagram (Figure 1) illustrates the selection process of articles for systematic review.
The intensity of nutrition education varied widely in frequency, length, and spread of sessions.The lowest frequency was 1 lesson at the start of the intervention [43,50,53], and the highest frequency was 22 lessons spread across the duration of the intervention [76].Other reported frequencies included 3 [68,71], 4 [78], 6 [67], 9 [48,69], and 10 lessons [73].The length of the education lessons varied by format.Online lessons were short (10-20 min) compared with in-person lessons, which were often reported to be 1 h in duration.Lessons were spread across the duration of the intervention and/or farmer's market season and were often offered weekly, biweekly, or monthly.Five studies leveraged existing curricula and dietary guideline resources from federal health agencies and professional associations, including the USDA Dietary and Physical Activity Guidelines for Americans [69,78], the national standards for Diabetes Self-Management Education [50], the Dietary Approaches to Stop Hypertension and Diabetes Prevention Program [73], and the Cook Smart, Eat Smart curriculum [48].Four studies explicitly described the use of theoretical frameworks to guide the development of a study-specific curriculum [45,53,69,73], such as Social Cognitive Theory [53,69] and Adult Learning Theory [73].
For both online synchronous and in-person curricula, active learning components and cooking demonstrations were common.Field-based learning was employed in 2 studies and included tours of grocery stores, farms, and FMs guided by health educators [48,69].Two studies described using a "tailored" nutrition education curriculum; Stotz et al., (2019) [73] tailored the curriculum to the cultural preferences, social needs, and educational needs of their target population, whereas the trial from which White et al., (2018) [69] drew their qualitative sample tailored the curriculum to the CSA season and availability of produce [81].
Two studies described offering their nutrition education sessions to families [69,78], and 1 study focused on children as the primary target [76].Three of the 14 studies offered nutrition education lessons and materials in both the English and Spanish languages [50,52,53].Reported educational materials include lesson handouts, produce information (e.g., purchasing, storing, and preparation tips), and recipe cards.
Of the 12 articles that measured FV intake, 7 found SVC intervention participation to significantly increase FV intake [43][44][45][46]49,51,52], 4 found no effect [48,50,53,54], and 1 found a negative effect (i.e., the comparison group had a significant increase in FV intake when compared with the intervention group) [47].Among the studies that found a positive impact, improvements in FV intake were characterized differently depending on the methods and measures employed.Anderson et al., (2004) [43] used a structural equation model with a latent variable representing 3 measures of FV intake and found a regression coefficient of 0.33.Johnson et al., (2004) [51] used questions from the Behavioral Risk Factor Surveillance System and observed a 1.04-serving improvement in FV intake.Gans et al., (2018) [52] and Leone et al., (2018) [45] both used the National Cancer Institute's FV screener and observed a 0.44 cup and 0.31 cup improvement, respectively.Herman et al., (2008) [46] and Berkowitz et al., (2019) [49] both employed 24-h recalls, but the former characterized their impact in terms of servings/1000 kcal consumed-observing an increase of 1.4 compared with controls-and the latter calculated Healthy Eating Index sub-scores, finding increased scores for total vegetables (þ0.5), total fruit (þ1.0), and whole fruit (þ0.7) relative to controls.Finally, Kropp et al., (2018) [44] used plate waste data to study an FTS intervention and estimated a 0.06-serving improvement in vegetable intake.
Findings were generally null or mixed for the other outcomes.Four articles included anthropometric measures as an outcome, although none of them found an effect [47,49,50,73].Two of the 4 articles that measured total diet quality found the SVC intervention to have no effect [47,73], whereas 2 found a significant increase in total diet quality among intervention participants [49,76].Three articles assessed biomarkers of health (i.e., blood pressure [49,73], hemoglobin A1c [50,73], fasting blood glucose [73], and lipid panel [73]), but only 1 found an effect: a study in which the intervention group had a significant decrease in diastolic blood pressure [49].Both articles that measured food security status found the interventions to have no effect [49,73].
The most common barriers noted across all intervention types were insufficient program awareness, poor logistical access or convenience, and issues related to cultural incongruence.Participants reported lacking the knowledge necessary to fully utilize the SVC program outside the parameters of the study (for example, lack of clarity regarding outlet location, hours of operation, and available food assistance programming options).Further, participants reported poor logistical access or inconvenience, with specific concerns regarding the cost of produce, transportation limitations, and locations and/or hours that interfered with long working hours or busy family schedules.For interventions involving FMs and CSAs, participants reported dissatisfaction with the limited variety and reliability of produce available, especially relative to supermarkets.Spoilage of fresh produce was also a concern, often cited in tandem with tight food budgets.Many studies examined the utility of financial incentives, but logistical issues with voucher distribution and redemption were frequently reported.
Issues related to cultural incongruence were reported across 18 studies.This was expressed in several distinct ways.Most often, participants reported that fresh FV was either not routinely consumed or not a part of their traditional cultural foods.Among studies examining FM-based programs, experiences of bias (e.g., racial) were common and deterred regular participation.Experiences ranged from perceived bias against the presence of young children and language barriers to stigma associated with the use of food assistance programs and came from both vendors and other shoppers.
Common facilitators of SVC engagement included the healthpromoting environment of SVC markets, feelings of community cohesion, financial incentives, and FV quality.The healthpromoting environment was the most encompassing facilitator and involved opportunities for nutrition education (e.g., preparation and preservation techniques, recipes, and cooking skills) and social interaction, ultimately enhancing participants' desire to eat more healthfully to prevent and manage chronic disease.
Community cohesion was identified as a facilitator distinct from the health-promoting environment, given that it was driven not by what the participants gained from the program but by how it enabled them to support their network and community.For example, participants appreciated the opportunity to interact and exchange information with FM vendors and CSA farmers and to support the local food economy.Others reported enjoying sharing nutrition education and excess FV with family and friends.Studies wherein health professionals facilitated the intervention, such as PRx models, reported that participants appreciated the collaboration between community resources.
Financial incentives were a commonly reported facilitator when available, although participants reported a resurgence of cost as a barrier as soon as the intervention concluded.Existing food assistance programs were more often discussed along with barriers, such as lack of awareness (e.g., how to use them at nontraditional markets), insufficient voucher amounts, and stigma related to program use.
Finally, the quality of FV available through SVC outlets was a key facilitator.Indeed, studies reported instances where participants were willing to put in extra time or effort to overcome barriers related to SVC engagement because of the high perceived quality of the available FV, especially relative to FV options available at local supermarkets or convenience stores.
Definitions of each barrier and facilitator, along with strategies to consider for enhancing future engagement, are outlined in Table 2.The applicability of each barrier and facilitator across studies is summarized in Table 3.Although strategies were not directly solicited from participants across all studies, several were reported based on their emergence during data collection.

Discussion
To our knowledge, this is the first systematic review to examine the impact of participant experiences with SVC models of healthy food access in the United States.The dual objectives, drawing on both quantitative and qualitative studies, afford a robust review from which we generate nuanced insights regarding the burgeoning scholarship on local food system models for advancing food and nutrition security and health equity.This review found mixed efficacy of SVC models, with improved FV intake being the most consistently demonstrated impact.SVC interventions vary widely in design, although FMs are more commonly studied than other intervention types.Despite such model variety, we found there to be a common set of barriers to and facilitators of participant engagement across model types.

Quantitative findings
Among quantitative studies, FV intake was the most frequently measured outcome and 1 for which findings were generally promising.Other quantitative outcomes were sparsely measured or not measured at all.Even so, improving FV intake is a key, proximal mediator of longer-term health impacts; FVs encompass a wide array of foods that provide dietary fiber, vitamins, and minerals and are a source of phytochemicals that have numerous protective mechanisms [82].Evidence indicates that higher intake of FV is associated with reduced risk of heart disease [83], stroke [84], lower mortality [85], and has a positive impact on mental health status in adults [86].The impacts observed in the reviewed studies-namely those for which intake improved by !1 serving (or a half cup)-were clinically meaningful, which has far-reaching implications given that low-income, food-insecure households are at a heightened risk of chronic disease and mental health conditions [5,7,8,87].
Scholars and practitioners alike tend to be concerned about the long-term sustainment of impacts measured over relatively short-term studies.In the case of FV intake, follow-up studies have been done to assuage this concern.Marshall et al., (2020) [88] conducted a 2-y follow-up on a school-based intervention that increased child intake of FV and found a sustained and significant increase in participant intake compared with baseline.Neville et al., (2015) [89] also conducted an 18-mo follow-up of an RCT in older adults and observed long-term positive changes in FV intake.If we want to move beyond understanding the proximal impacts of such interventions, although, and discern for which models the proximal impacts translate into more distal impacts on chronic disease morbidity, studies of greater duration than was generally observed in this review will be necessary.
Anthropometric measures, total diet quality, health biomarkers, and food security status were assessed less frequently.This was a surprise, particularly the lack of food security data, as many studies cited this as a motivating concern and rationale for targeting low-income households.The dearth of such outcomes may be related to the burden incurred by both researchers and participants to collect it (with potential trade-offs for study retention) or the lack of changes to such outcomes in pretrial pilot studies (though this would be expected if pilots were relatively short in duration and underpowered, as is often the case).None of the included studies reported QoL or health outcomes.This may also be related to relatively short study durations or the participant burden associated with rigorous interventions, which may plausibly harm QoL [90].This is something for researchers to consider, given the practical importance of QoL for participant well-being and the key role of health-related QoL data in cost-utility analyses comparing interventions toward evidence-based resource allocation.

Qualitative findings
The aggregation of qualitative insights revealed insufficient program awareness, poor logistical access or convenience, and issues related to cultural incongruence as common barriers to participant engagement across intervention types.Ubiquitous facilitators of engagement included the health-promoting environment of SVC outlets, feelings of community cohesion, financial incentives, and FV quality.
Although identifying barriers, participants also offered strategies-often unsolicited-to enhance sustained participation in future programs.This reflects a broad interest in and commitment to improving SVC models among low-income individuals.Participants perceived seasonal FV nutrition education (i.e., preparation and storage methods) and community connection to be facilitators with unique applicability to SVC models and may be key points of focus for program administrators.Despite apparent interest in the studied models, this review revealed how a common set of barriers has persisted throughout the 20-y review period.Given the ubiquitous and persistent nature of these barriers, further research on such barriers and facilitators may be less impactful than efforts to understand and test implementation solutions.Of note, there was substantial interdependence between barriers and facilitators; the decision by low-income households to use an SVC model is multifaceted, weighing economic, logistical, and sociocultural factors beyond individual control.This suggests that systems-level interventions may be more effective than singular or isolated approaches.Addressing barriers in the long term requires sustained cross-

Multimodal interventions
Financial and self-efficacy barriers create distinct and welldocumented challenges to achieving a nutritious diet [91][92][93].This review included 15 multimodal interventions that aimed to mitigate these 2 barriers simultaneously.Modalities for enhancing financial access to FV included the provision of FM coupons or free or discounted produce.Modalities to support diet-related self-efficacy included nutrition lessons, cooking skill workshops and demonstrations, and educator-guided tours of FM and grocery stores.
Evidence suggests that multimodal interventions have a greater likelihood of affecting health behavior change compared with unimodal interventions [94,95].Multimodal interventions are particularly recommended for interventions targeting household-level changes (as is common for SVC interventions) and for managing common and complex health conditions, such as obesity, diabetes, and cancer [94][95][96][97].In this review, 8 of the 15 multimodal interventions reported on our quantitative outcomes of interest [43,45,48,50,52,53,66,76]. Of those, only 4 showed significant positive changes-for the outcomes of diet quality [76] and FV intake [43,45,52].Examination of the modalities used in these interventions revealed 2 characteristics that seem key to facilitating successful outcomes: social marketing and intensive nutrition education.
Social marketing strategies appear to be a poignant mechanism for increasing awareness of SVC programs, encouraging engagement, and promoting a sense of community.Sharpe et al., (2020) [47] concluded that improving spatial access to healthy foods alone was ineffective in improving diet quality among disadvantaged communities living in USDA-defined Low-Income Low-Access areas and suggested a multifaceted approach focused on barriers experienced by the target community.Incorporating promotional activities in intervention design can be particularly advantageous, as a lack of awareness about the existence and operations of SVC programs (including location, hours, and acceptance of Electronic Benefit Transaction) was reported as a barrier in 11 of the 25 qualitative studies in this review [55,[57][58][59]61,62,66,70,74,75,79].Gibson et al., (2014) [76] engaged parents in a 6-mo FTS intervention via monthly in-school FM displays with free seasonal produce and printed recipes.Leone et al., (2018) [45] and Gans et al., (2018) [52] leveraged reduced-price mobile fresh market models and consistently marketed throughout the duration of the interventions (6 and 12 mo, respectively).Their strategies included visually attractive newsletters delivered regularly via mail and email (weekly and monthly) with market information and invitations to join intervention-related community events (e.g., cooking demonstrations, taste-testing, and prize raffles, respectively).
A second mechanism for driving positive outcomes appeared to be the implementation of a dynamic nutrition education curriculum that offered frequent lessons, promoted both knowledge and skills, incorporated field-based learning activities, and, most importantly, was tailored to produce seasonality.Although a general lack of nutrition knowledge and cooking skills has been reported as a barrier to increasing FV intake [57,65,73], unfamiliarity with FM and CSA produce items were specifically mentioned as a barrier to engagement in numerous studies [10,62,64,69,77].In Gibson et al., (2014) [76], weekly nutrition lessons included farmers serving as guest speakers, cooking classes, a field trip to the grocery store, and school-based gardening.In Leone et al., (2018) [45] and Gans et al., (2018) [52], all intervention elements were focused on in-season produce and included content on key nutrients, health benefits, relevant recipes, and tips for selection, storage, and time-efficient and budget-friendly approaches for preparation and integration in the diet.

Financial incentives
Financial incentives have been found to influence short-term dietary behavior change positively.When used as a catalyst for change rather than a reward, as is the case for the SVC models in this review, financial incentives can aid in the long-term maintenance of dietary behaviors [98].Seventeen interventions (19 studies) described using financial incentives as part of the studied SVC model [43,46,49,50,[53][54][55][56]59,60,[66][67][68][70][71][72]75,77,78].Of those, 8 were FM-based and affiliated with a federally-funded nutrition assistance program: 4 WIC Farmers Market Nutrition Program [43,46,53,54,66] and 4 SNAP-Double Up Food Bucks [55,59,60,70].PRx, CSA, and MM interventions also offered financial incentives.Incentives amounts ranged from $5 one-time vouchers [56] to $600 toward a multiyear CSA share [49].The incentive value sufficient to motivate behavior change remains a subject of debate, especially in underserved communities where the trade-off of precious time and scarce resources can deem small incentives futile.A systematic review of financial incentives for dietary behavior change estimates $40 as an optimal starting incentive for weight management programs [98].Six of the 19 studies reported quantitative impacts [43,46,49,50,53,54], with 3 demonstrating a significant increase in FV intake [43,46,49] and 2 of these being the aforementioned studies that offered the highest incentive amounts of all reviewed studies [46,49].The federally-funded GusNIP supports the implementation of high-reach, low-intensity community projects and low-reach, high-intensity projects for underserved communities, with intensity being inclusive of services, nutrition education, and incentive amounts.Budgets allocated toward direct incentives by GusNIP grantees increased from 68.5% to 74.7% in a span of 1 y [99,100], suggesting a greater realization of the role of incentives in achieving program objectives.Future research on incentive amounts, with consideration of community characteristics and environmental context, may yield useful guidance on optimal incentives across SVC intervention types.

Demographics and geographic location
Our systematic review findings highlight several shortcomings worth discussing.Each included article reported a higher percentage of female participants than males.In fact, 6 articles had only female participants, and 9 had >80% female participants.This can be partially explained by the number of studies that focused solely on WIC-related programs.However, having a higher proportion of female participants is not uncommon in nutrition, health, and food security-related research; as scholars work to understand mechanisms for enhancing population-wide health equity, this will be something to consider more purposefully during study planning and recruitment.
Additionally, most studies were focused on adults.Future studies need to consider the complex household dynamics and whether it would be prudent to enroll more than a single individual.Family-based multimodal interventions are favorable for the management and/or treatment of chronic health conditions in both adult and children populations compared with standardof-care interventions [94].
Geographic representation is another area for which researchers will need to be more intentional moving forward.Studies in this review were largely conducted in urban areas.Pillar 1 of the White House National Strategy on Hunger, Nutrition, and Health calls for special attention to rural health, given the persistent structural inequalities (e.g., transportation issues) and disparities in food access that they face [18].In 2021, 9 out of 10 counties with the highest food insecurity rates were rural [101].This represents a major public health problem, given that 46 million Americans live in rural areas [102].The USDA and NIH have also heightened their focus on rural health, signaling the importance and necessity of this in future research [103,104].

Methodologic considerations
Assessing the risk of bias for the 37 articles that met our inclusion criteria presented some challenges.The tool used to appraise qualitative studies, the SRQR, offers well-defined standards for reporting qualitative data; however, it was challenging to use this tool for pragmatic qualitative studies.For example, qualitative research undertaken with a practical or clinical orientation scored less favorably for criteria emphasizing explicit discussion of the research paradigm (e.g., postpositivist or constructivist) and elements related to researcher positionality (e.g., researcher characteristics and practices demonstrating reflexivity).Relying on the SRQR as a tool for assessing rigor thus required the research team to interrogate all other elements of the pragmatic studies more thoroughly to determine a reasonable risk of bias assessment rating.This suggests that there are opportunities to enhance the rigor and transparency with which qualitative inquiry is pursued, even when approached pragmatically or as 1 part of a mixed-method effort.An additional challenge arose when assessing the rigor of qualitative studies because of a wide variation in how facilitators and barriers to SVC participation were reported and framed (e.g., the degree to which themes were explicit).
The National Institute of Heart, Lung, and Blood Quality Assessment Tools worked well for evaluating all quantitative studies.A notable exception arose when trying to determine how to use the tools for quasi-experimental designs that did not fit the parameters of controlled intervention studies, observational cohort and cross-sectional studies, or case-control studies.Ultimately, the study in question was excluded from the final review because it did not meet inclusion criteria.As quasi-experimental designs and econometric analyses-such as the differences-indifferences design leveraged by Olsho et al., (2015) [105]gain traction in the health policy literature given their ability to estimate causal impact in the absence of RCTs (e.g., when such designs are unethical or impractical), multidisciplinary teams conducting systematic reviews need to consider the value of employing cross-disciplinary tools appropriate for more diverse study designs.
Very few of the studies that met inclusion criteria, in fact, were RCTs (n ¼ 3) or cluster-RCTs (n ¼ 2).It is unclear whether this was merely reflective of our inclusion criteria or if there are issues of feasibility when seeking to conduct an RCT to assess the effectiveness of SVC models.This may also connect to broader conversations about the questions and interventions for which RCTs are indeed appropriate and ethical to use, particularly if the target population includes structurally marginalized groups that may benefit from a minimal level of access to a particular SVC intervention.
The 1 qualitative study examined a MM intervention [109] and identified similar barriers to and facilitators of engagement as those studies included in the review (i.e., poor logistical access but helpful financial incentives).These barriers and facilitators were collected via a concept mapping activity with intervention participants; such an activity would not have met our methodologic inclusion criteria but is in line with emerging best practices for community-engaged data collection.Future qualitative reviews will want to keep this in mind.
The design of the quantitative studies-3 RCTs [106,107, 110], 1 process evaluation of an RCT [108], and 3 quasi-experimental designs [106][107][108]-included randomization more often than did the studies included in our review.This is an encouraging trend toward high-quality quantitative study designs, suggesting that this area continues to be a compelling avenue of research and that concerns regarding the applicability of randomly assigned study designs to SVC models may be for naught.The relative dearth of qualitative studies since 2020 is concerning; continuing to integrate the qualitative perspectives of study participants is core to optimizing the implementation of these models, especially as they get scaled to new settings and culturally distinct populations.

Policy and research considerations
Pandemic-era media coverage and the recent White House Conference on Hunger, Nutrition, and Health generated robust public and private commitments to addressing food and nutrition insecurity in the United States.With the 2025 United States Dietary Guidelines committee at work and discussions commencing regarding the next Farm Bill, the country is at a critical juncture regarding we will address the complex interface of food supply chains, nutrition, and health.This review highlights critical opportunities to bolster our understanding of how SVC models may be leveraged to advance national goals at the interface of agricultural, economic, social, biological (nutrition), and health care systems.
Specifically, we contend that any federal investments in this domain should include sufficient resource allocation for robust, nationally coordinated evaluation (as has been facilitated for GusNIP grantees).The central goal would be to interrogate the potential of SVC models using consistent methods and measures for programs implemented across the rural-urban continuum and among diverse communities via longer-term studies focused on measurable health impacts.Studies focused on understanding and testing implementation strategies designed to overcome known (persistent) barriers and maximize impacts-for participants and across the value chain-will also be key.
For scholars and practitioners working together to design, implement, and study SVC models, we encourage consideration of how social marketing and engaging, self-efficacy-enhancing forms of nutrition education could be employed and tested, given our findings regarding their role in successful interventions to date.We also encourage the explicit study of intervention dose, including the value of any financial incentives; engagement and study of household units, not just individuals; and the inclusion of rural and remote contexts.Together, the insights from these more strategically designed studies can balance the need for evidencebased public health investments with ever-present concerns regarding resource limitations.

FIGURE 1 .
FIGURE 1. PRISMA flow diagram of the study identification and selection process for eligible articles.PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

TABLE 1
Characteristics of the short value chain interventions included in the systematic review

TABLE 1
Fair (continued on next page) K. Garrity et al.Advances in Nutrition 15 (2024) 100156

TABLE 1 (
continued ) Farmers market (FM) (n ¼ 18)food markets at which local farmers sell directly to customers (continued on next page) K. Garrity et al.Advances in Nutrition 15 (2024) 100156

TABLE 1 (
continued ) Farmers market (FM) (n ¼ 18)food markets at which local farmers sell directly to customers and from the market, and even pricing (i.e., pricing rounded to the dollar to limit the need for change); the desire to support local farmers Good (continued on next page) K. Garrity et al.Advances in Nutrition 15 (2024) 100156

TABLE 1 (
Produce prescription program (PRx) (n ¼ 7)a clinic-community collaboration in which a healthcare representative refers patients to receive free or discounted fruits and vegetables continued )Farmers market (FM) (n ¼ 18)food markets at which local farmers sell directly to customers

TABLE 1 (
continued ) Produce prescription program (PRx) (n ¼ 7)a clinic-community collaboration in which a healthcare representative refers patients to receive free or discounted fruits and vegetables

TABLE 1 (
continued ) Produce prescription program (PRx) (n ¼ 7)a clinic-community collaboration in which a healthcare representative refers patients to receive free or discounted fruits and vegetables Mobile market (MM) (n ¼ 5)produce is aggregated, typically in a single large vehicle, and transported directly to various neighborhoods for short-term sale

TABLE 1 (
continued ) Mobile market (MM) (n ¼ 5)produce is aggregated, typically in a single large vehicle, and transported directly to various neighborhoods for short-term sale

TABLE 1 (
continued ) Mobile market (MM) (n ¼ 5)produce is aggregated, typically in a single large vehicle, and transported directly to various neighborhoods for short-term sale Community-supported agriculture (CSA) (n ¼ 4)community members buy a share of a farmer's produce and receive portions of the harvest regularly throughout the growing season

TABLE 1 (
continued ) Community-supported agriculture (CSA) (n ¼ 4)community members buy a share of a farmer's produce and receive portions of the harvest regularly throughout the growing season

TABLE 1 (
continued ) Community-supported agriculture (CSA) (n ¼ 4)community members buy a share of a farmer's produce and receive portions of the harvest regularly throughout the growing season

TABLE 2
Emergent barrier and facilitator themes with reported strategies and considerations for supporting participant engagement in short value chain interventions

TABLE 3
Occurrences of barrier and facilitator themes within included qualitative studies

TABLE 3
partnerships and multimodal interventions that address the interplay between program access, awareness, cultural congruence, and financial incentives.
Abbreviations: CSA, Community-supported agriculture; FM, farmers market; FSt, farm stand; FTS, farm-to-school; MM, mobile market; PRx, produce prescription program; SVC, short value chain; WIC, special supplemental nutrition program for women, infants, and children.sector