Community Engaged Research Strategies Used in Food Retail Interventions: A Scoping Review

Abstract

Background

Researchers across the United States often leverage community engagement (CE) as a strategy in interventions aiming to alter the retail food environment (RFE), especially in areas serving racially segregated neighborhoods with low incomes. However, little is known about the full breadth, intensity, and approaches used to engage communities in RFE intervention work.

Objective

The purpose of this scoping review is to identify what CE research approaches have been applied by researchers in the RFE intervention literature and how they vary by type of retail settings, phase of intervention, year of intervention, and key domains of equity.

Methods

Following the JBI (formerly known as Joanna Briggs Institute) and the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) extension for Scoping Review guidelines, any study published in academic journals and English that discussed activities or strategies for CE in RFEs, irrespective of the type of study, was included. PubMed, CINAHL, and ProQuest were searched for reports published from inception until August 2023. CE research strategies were extracted and classified following the Centers for Disease Control and Prevention (CDC) continuum of community engagement framework, including outreach (lowest CE), consult/involve, collaboration, and shared leadership (highest CE). CE research strategies were then examined for their variation across RFE setting, intervention phase, intervention year, and key equity domains related to healthy food retail (eg, affordability).

Results

A total of 98 RFE interventions reported in 104 reports were included in this review, and most were implemented in either supermarkets (21%), corner stores (20%), or multiple RFE settings (21%). All interventions employed CE research strategies of outreach (n = 98), whereas approximately half employed strategies of shared leadership (n = 52). Exploring CE research strategies by RFE settings and intervention phase, this review found stronger forms of CE in less traditional RFE settings, including mobile markets, and among interventions that used CE research strategies across all phases of the intervention study. RFE interventions that implemented the highest forms of CE research strategies (ie, shared leadership) were also those that addressed all key equity domains.

Conclusion

The findings of this review reveal that the form of CE in RFE interventions varied widely, with more domains of equity addressed when higher forms of CE were used. Insights from this review suggest that future research should prioritize assessing the effectiveness of shared leadership CE strategies on achieving and sustaining nutrition-related health equity outcomes for communities.

UNHEALTHY DIETS CONTRIBUTE TO THE DEVELOPment of multiple chronic conditions such as diabetes, obesity, and cardiovascular disease.^1-3 These chronic disease conditions disproportionately burden certain communities and groups, including families with low incomes, communities of color, indigenous people, and racially segregated neighborhoods.^4,5 Diet-related health disparities are partially attributable to social and structural barriers in the retail food environment (RFE),⁴ such as structural racism and economic inequality, which can impact the types of stores and products made available. The RFE is typically operationalized as the main place (eg, grocery store, corner store, farmers market) where people purchase their food within a community.⁶ The RFE is shaped by a complex network of physical, social, economic, cultural, and political factors that interact at the consumer, store, and community levels.⁷ These factors determine the accessibility, availability, and affordability of food in any community and influence consumers’ dietary choices and nutritional status, as well as other population health outcomes.⁸ Thus, for many interventions focused on community nutrition, the RFE serves as a key venue for implementing strategies.

Intervention research around addressing the RFE to promote population health has had an exponential growth in the past 2 decades. A variety of different strategies have been implemented, such as introducing healthy foods at existing stores, introducing new food stores, raising nutrition knowledge of consumers or community residents via product labels, and leveraging the 4 Ps of marketing (ie, product, price, place, and promotion) to promote healthy food availability and purchasing.^9-12 An integral part of many of these interventions is to shape dietary behavior and promote better nutrition-related health outcomes of communities by attracting consumer^9,11 and/or community engagement (CE).^9,12 Directly engaging communities enables researchers to understand local contexts that can support more effective implementation and sustainability of intervention strategies.¹³ Such engagement is particularly critical for interventions focusing on historically and intentionally excluded communities, because evidence demonstrates that more relevant and effective interventions can be developed by forming supportive relationships, enhancing community capacity, and involving the community in decision-making.¹³ Hence, comprehensive and participatory approaches in RFE interventions are essential to reduce diet-related health inequities.

Meaningful CE that centers diversity, equity, inclusion, and ownership¹⁴ can be complex and challenging, but also very rewarding. According to the Centers for Disease Control and Prevention (CDC), CE can be defined as a mechanism that involves people, partnerships, and coalitions who serve as catalysts for mobilizing community resources to influence systems toward improving the health and well-being of the community.¹⁴ Often guided by principles of the Community-Based Participatory Research model (CBPR), the emphasis of CE is a collaborative process that requires understanding the strengths of each partner and involving them in all aspects of the intervention or program.¹⁵ CE has the potential to transform the power dynamics between partners through the adoption of community inclusiveness in decision-making.¹⁶

CBPR has been widely implemented in public health interventions, including those focusing on the RFE.¹² Although numerous individual publications on strategies to engage communities in RFE changes have been described,^10-12,17 little is understood about the entire breadth, intensity, and methodological approaches used for CE in the RFE intervention literature and how well these studies address equity. Existing systematic or scoping reviews that focus on RFE interventions have explored the impact on diet and health,^18,19 healthy food consumption,²⁰ and obesity-related outcomes.²¹ Although these reviews help to summarize the wider RFE literature on effective approaches for health,^18-21 most of these reviews do not focus on the specific CE research approaches used. In addition, only 1 prior review was identified that examined the use of CE strategies in RFE interventions,²² yet they only considered 1 CE research approach (eg, shared decision-making) and not the complete spectrum of potential strategies.

To fill these gaps, this scoping review aimed to identify what CE research approaches have been applied by researchers in RFE intervention literature. Specifically, this review examined the form and extent of CE in RFE interventions and how CE varied across RFE settings, intervention design phase (eg, planning, implementation), and intervention year. Furthermore, the review aimed to examine whether consideration of equity as defined by Kumanyika (2019)²³ and Mui et al (2021)²⁴ were included in the RFE intervention design; namely, if interventions increased healthy food availability and options, affordability, or accessibility for historically and intentionally excluded communities (eg, low-income, Indigenous).

METHODS

This scoping review was performed using guidance from the JBI scoping review²⁵ and the Preferred Reporting Items for Systematic Reviews and Meta-analyses extension for Scoping Reviews (PRISMA-ScR).²⁶ An initial protocol was drafted and shared with all co-authors for feedback. The protocol was registered under Open Science Framework registration and is available at https://osf.io/q6jtz, and all authors agreed on iterative protocol development as needed.

Search Strategy

The search strategy was created in collaboration with the University of Illinois College of Medicine Rockford’s research librarian, with test searches run in December 2022. The first database searches were executed on January 6, 2023, and they were rerun on August 18, 2023, to capture any new publication(s) since the first search. PubMed, CINAHL, as well as 5 databases in ProQuest (APA PsycInfo, PAIS Index, ProQuest Dissertations & Theses A&I, ProQuest Dissertations & Theses Global, and Sociological Abstracts) were searched. The search included terms related to the concepts of “retail food environment (RFE)” and “community engagement (CE).” The search process was iterative, and additional terms were added as identified throughout the process to create the final search strategy (Supplementary Figure 1). The team also searched for additional citations of extracted reports.

Eligibility Criteria

Following the JBI recommendations,²⁵ inclusion criteria were defined around context, concept, and population of interest. The context was retail food stores in the United States, given the unique, longstanding systematic injustices in neighborhood resource allocation, including food access, that historically marginalized communities in the United States continue to experience.²⁷ A broad definition of retail food stores, including “any establishment where food is offered to the consumer and intended for off-premises consumption”²⁸ was used. This encompassed a wide range of settings, such as supermarkets, grocery stores, corner stores, gas marts, dollar stores, ethnic grocers, community food stores, farmers markets, mobile markets, and restaurants. The concept was RFE interventions implementing environmental change strategies to promote healthy eating environments in communities (eg, increasing food outlets or adapting the environment of existing food outlets to offer healthy food). Any participant who was involved or participated in RFE interventions was our population of interest. The measured outcome was activities used to engage communities in any RFE intervention.

Following the inclusion/exclusion criteria, reports written in English were considered. While ProQuest Dissertations & Theses A&I and ProQuest Dissertations databases were searched, our final inclusion/exclusion criteria required reports to be published in academic journals. Any reports that discussed the activities or strategies to engage customers or communities in RFE interventions, regardless of the type of manuscript (such as protocol, formative research, evaluation, or program implementation), were included. Some researchers published intervention protocol papers, and others discussed their activities in an evaluation or formative research paper. The purpose was not to focus on the type of report or manuscript but to understand the intervention discussed in the report. Reports in which an RFE intervention was not conducted were excluded. Reports targeting non-RFE settings, such as childcare facilities, schools, hospitals, and worksite cafeterias, were also excluded unless there was an explicit collaboration between these settings and retail food stores to implement intervention activities.

Study Selection

Covidence, a web-based platform, was used to manage the review.²⁹ All identified records fulfilling the inclusion/exclusion criteria were uploaded to Covidence. Two co-authors independently screened the title and abstract. Two co-authors who did not participate in the title and abstract screening resolved the conflicts. Additionally, 3 authors performed the full-text screening with a minimum of 2 authors screening each article independently, and any reason for excluding the article was recorded. All disagreements throughout the process were resolved, and a consensus was achieved through discussions.

Data Extraction

A template was created, tested, and uploaded to Covidence²⁹ to extract all relevant information. Data extracted from each article included authors, intervention year, population of focus (race and ethnicity, income), intervention site (urban/rural/mixed), CE activities, and the focus of the intervention (affordability, availability, accessibility). When the intervention year was not available, the publication year was used as the intervention year. Extraction was divided between 6 reviewers, with each reviewer extracting all relevant information for their assigned report. Extracted data was then verified by a second reviewer. When there was more than 1 publication/report from the same intervention (eg, both a protocol and intervention evaluation), relevant extracted information was merged in the extraction form, with the more detailed information maintained.

Activities to engage communities were extracted, and the CDC continuum of community engagement framework was followed to classify the level of CE in RFE interventions.¹⁴ Category levels included outreach, consult/involve, collaborate, and shared ownership/decision making (Figure 2 for definitions). The review also evaluated whether the focus of the interventions addressed key equity domains of healthy eating in food retail. The selection of equity domains was guided by the Getting to Equity in Obesity Prevention framework by Kumanyika²³ and by key dimensions of an equitable regional food system identified by Mui and colleagues (2021).²⁴ Domains included increased healthy food availability at existing locations (eg, adding produce or additional varieties of low-sodium canned vegetables at existing establishments, and so forth); increased healthy food availability by adding new sites; increased affordability of healthy food items (eg, reducing prices for healthy food through incentives or vouchers); and increased accessibility of healthy options at food retail sites (eg, mobile produce markets or providing transportation). Information was also extracted about whether the intervention was intended to address RFEs in a historically and intentionally excluded community (eg, conducted in neighborhoods with low healthy food access or in retail settings that accept food assistance benefits) or population (eg, low-income, rural, Indigenous, or communities of color). Data regarding geography was extracted based on the authors’ description of the intervention location (urban/rural/mixed/not specified); when it was not identified but the city or county name was provided, the United States Department of Agriculture (USDA) Economic Research Service 2023 (ERS) Rural-Urban Continuum³⁰ codes were followed. If no specific location was mentioned in the report, it was categorized as “not specified.” A quality appraisal of included reports was not performed, given that the focus of this review was to identify the different CE research approaches used in RFE interventions, as opposed to how well the intervention was designed to measure an effect.

Figure 2.

Description of categories in the continuum of community engagement.

Data Synthesis

Extracted data were synthesized in narrative and tabular form, guided by the research questions. To measure the extent of CE, each study was assessed for CE research activities that met the four CDC CE categories of outreach, consult/involve, collaborate, and shared ownership/decision making (Fig 2) and assigned a score of 1, 2, 3, or 4, respectively, when such activities were present. Two reviewers independently reviewed the scores and provided a consensus. To compare CE score across types of RFE settings, intervention phase (ie, the intervention phase communities were engaged), and intervention year, a mean score with minimum and maximum (min/max) for each setting, phase, and year was calculated. As a result, even a study with shared decision-making in a single intervention phase, such as planning, would have received a score of 4. RFE settings were defined as either corner stores, supermarkets, farmers markets, mobile markets, restaurants, others, or multiple if the intervention targeted more than 1 type of setting. The intervention design phase was categorized based on when CE activities occurred, including planning, implementation, evaluation, or any combination of phases. The 4 key equity domains for healthy food retail and population of focus were also evaluated across CE scores. Equity domains were examined individually as well as whether “All” or “>1 domain” were addressed in the RFE intervention.

RESULTS

The literature search resulted in 13 619 references. After removing 8980 duplicates, the title and abstract screening were conducted on 4639 records against the inclusion criteria. A total of 310 full-text reports were subsequently assessed for eligibility, and of these, 104 reports or articles were included in the review. There were 98^10,31-127 unique RFE studies/interventions (Fig 3) evaluated for their CE research strategies, after combining reports (n = 6)^128-133 on the same interventions (Fig 4). Going forward, the 98 unique studies will be referred to as interventions in this manuscript.

Figure 3.

Characteristics of 98 studies included in the scoping review on community engaged research strategies in food retail interventions.

Figure 4.

Flow diagram of the literature search and screening results for the scoping review of community engagement in retail food environment interventions.^{26 a}Reasons for exclusion at the title and abstract screening step were not recorded.

Description of Included Studies

Most of the 98 interventions (Table 1) were conducted in corner stores (20%; n = 20),^31-50 supermarkets (21%, n = 21),^51-71 or multiple settings (21%, n = 21)^10,72-91 such as different combinations of wholesalers, corner stores, community settings, carryout, worksites, and schools. The remaining interventions were conducted at farmers markets (15%, n = 15),^92-106 mobile markets (10%, n = 10),^107-116 restaurants (4%, n = 4),^117-120 or another RFE setting (7%, n = 7), such as food hubs, food co-ops, vending machines, or general grocery store (ie, did not specify corner store or supermarket).^121-127 Included interventions were conducted between 1989 and 2023, with nearly two-thirds (n = 62/98) conducted after 2010 (Table 1). Approximately 62% (n = 61)^{10,31-34,37-46,48,50,55-59,61-67,70-72,75,76,78,79,83,84,86,89,94-96,98,102,103,105,107-111,113,114,117,118,120,123-125,127} of interventions were conducted in urban settings, 22% (n = 22)^{47,49,51-54,60,73,74,77,85,87,90,91,97,101,104,106,112,116,119,126} in a “rural setting,” 8% (n = 7)^{36,69,82,88,92,93,121} “mixed,” 1% (n = 1)³⁵ suburban, and 6% (n = 7)^{68,80,81,99,100,115,122} “did not specify” the geography type (Fig 3).

Table 1.

Community engagement (CE) variation across retail food environment setting, intervention phase, and intervention year among 98 studies included in the scoping review on community-engaged research strategies in food retail interventions

	N (98)	Mean CE score (min-max)	CE Scorea % (n/N)
			1	2	3	4
Retail food environment settings
Corner store	2031-50	3.0 (1–4)	20% (4/20)	10% (2/20)	25% (5/20)	45% (9/20)
Farmers market	1592-106	3.1 (1–4)	7% (1/15)	27% (4/15)	13% (2/15)	53% (8/15)
Mobile market	10107-116	3.7 (3–4)	0	0	30% (3/10)	70% (7/10)
Multiple	2110,72-91	3.6 (1–4)	4% (1/21)	10% (2/21)	10% (2/21)	76% (16/21)
Otherb	7121-127	3.7 (2–4)	0	14% (1/7)	0	86% (6/7)
Restaurant	4117-120	2.5 (2–4)	0	75% (3/4)	0	25% (1/4)
Supermarket	2151-71	2.2 (1–4)	33% (7/21)	33% (7/21)	10% (2/21)	24% (5/21)
Intervention Phase of CE
Planning	260,63	3.0 (2–4)	0	50% (1/2)	0	50% (1/2)
Planning and implementation	510,42,55,62,110	2.8 (2–4)	0	40% (2/5)	40% (2/5)	20% (1/5)
Implementation	2337,39,41,54,57-59,61,65,67,68,70,78,82,86,91,92,106,108,113,117,119,124	2.0 (1–4)	39% (9/23)	35% (8/23)	17% (4/23)	9% (2/23)
Evaluation	536,43,66,99,112	2.0 (1–3)	40% (2/5)	20% (1/5)	40% (2/5)	0
Implementation and evaluation	832,38,51,56,69,95,96,107	1.9 (1–4)	25% (2/8)	50% (4/8)	12% (1/8)	12% (1/8)
All phases	5531,33-35,40,44-50,52,53,64,71-77,79-81,83-85,87-90,93,94,97,98,100-105,109,111,114-116,118,120-123,125-127	3.8 (2–4)	0	6% (3/55)	9% (5/55)	85% (47/55)
Intervention Year
1995 or earlier	555,58,73,77,82	3.0 (1–4)	20% (1/5)	0	40% (2/5)	40% (2/5)
1996–2000	368,91,123	2.3 (1–4)	33% (1/3)	33% (1/3)	0	33% (1/3)
2001–2005	452,78,81,105	3.3 (1–4)	25% (1/4)	0	0	75% (3/4)
2006–2010	2431,33,34,37-39,41,42,45,59,63-65,80,83,88,94-96,104,118,120,121,127	3.0 (1–4)	17% (4/24)	13% (3/24)	21% (5/24)	50% (12/24)
2011–2015	3910,35,36,40,44,46,48,50,54,56,57,60-62,66,67,72,74,75,79,84,85,87,90,93,97,100,101,103,107,108,112-114,116,117,122,125,126	3.2 (1–4)	5% (2/39)	23% (9/39)	18% (7/39)	54% (21/39)
2016–2020	2032,43,47,49,51,53,69,70,76,86,92,98,99,102,106,109,111,115,119,124	2.7 (1–4)	20% (4/20)	35% (7/20)	5% (1/20)	40% (8/20)
2021 or later	371,89,110	3.3 (2–4)	0	33% (1/3)	0	67% (2/3)

^aCommunity engagement score: 1 = Outreach; 2 = Outreach + Consult/Involve; 3 = Outreach + Consult/Involve + Collaborate; 4 = Outreach + Consult/Involve + Collaborate + Shared decision-making.

^bOther’ = Food hubs, food co-ops, vending machines, or general grocery store (ie, did not specify corner store or supermarket).

RFE interventions included in this review focused on different historically and intentionally excluded populations. More than 90% of interventions (n = 89)^{10,32-54,56-67,69,71,72,74-85,87-90,92-117,119-121,123-127} explicitly focused on improving RFEs for low-income neighborhoods, communities of color, Indigenous populations, or areas with low access to grocery stores or healthy food. In the following section, the findings are summarized as (1) forms of CE, (2) variation in CE by type of RFE settings, intervention phase, and intervention year, and (3) equity considerations in RFE interventions by CE score.

Form of CE Research Activities in RFE Interventions

Following the CDC CE continuum framework,¹⁴ all included RFE interventions (n = 98)^10,31-127 incorporated CE activities of outreach, 87% (n = 85)^{10,31-37,40,42,44-50,52-56,59,60,62-64,66,67,69,71-77,79-98,100-127} included activities of consultation or involvement, two-thirds included collaboration activities (n = 66),^{31,33-37,42,44-50,52-55,63,64,71-77,79-90,93,94,96,97,100-105,107-116,120-123,125-127} and approximately half demonstrated activities of shared decision-making and ownership (n = 52)^{33-35,44,46-50,52,54,63,64,71-73,75-77,79-81,83-90,93,94,100-105,107,109-111,114-116,120-123,125-127} (Fig 5).

Figure 5.

Community engagement activities and level of involvement across retail food environment interventions among 98 studies included in the scoping review on community engaged research strategies in food retail interventions.

Outreach.

All interventions had some type of outreach, focusing on either store customers or the broader community to inform residents about the intervention. Thirteen interventions used only in store outreach activities to inform customers or store employees^{38,39,41,43,51,57,58,61,65,68,70,78,99} (Fig 5). The most common in-store outreach activities included store promotion of the intervention (eg, point of sale signs, shelf labels, or nutrition education messages) and incentives to purchase healthy food items. The remaining interventions (n = 85)^{10,31-37,40,42,44-50,52-56,59,60,62-64,66,67,69,71-77,79-98,100-127} had outreach activities to engage wider communities, which included sharing information through media, social media, newsletters, marketing or social marketing campaigns, and other promotional material for the study (eg, recipe card distribution, informational postcards at community sites, yard signs, and billboard signs) (Fig 5).

Consult/Involve.

Eighty-five^{10,31-37,40,42,44-50,52-56,59,60,62-64,66,67,69,71-77,79-98,100-127} interventions (87%) implemented CE research activities to consult or involve communities—where residents had the opportunity to provide feedback and the flow of information was bidirectional. Numerous strategies were used, including community group discussions, focus groups, interviews, listening sessions, feedback on intervention material, surveys, interactive education sessions, interactive demonstrations, taste tests in the community, formative research, and involving community residents through community organizations. For instance, Dannefer et al (2012)³⁷ distributed consumer request cards among community organizations for residents to request specific healthy food to be carried in a store. Gans et al (2018)¹⁰⁹ conducted focus groups and discussions with residents to inform intervention development in which residents provided their input regarding the location of the intervention, the content of promotional materials, the type of produce to be sold, and nutrition education topics.¹⁰⁹

Collaboration.

Sixty-six interventions (67%)^{31,33-37,42,44-50,52-55,63,64,71-77,79-90,93,94,96,97,100-105,107-116,120-123,125-127} involved collaborative CE research activities with multilevel collaboration through different community organizations to build partnerships and trust in the community; however, the community’s role in final decision-making was not evident in the published description of the CE activities. Collaboration activities included building coalitions in the community, connecting farmers with retailers, having community members as advocates, and involving community organizations in intervention. For instance, Gittelsohn et al (2014)⁷² conducted an intervention called B’More Healthy Communities for Kids, which was a collaboration between wholesalers, recreation centers, corner stores, youth, and families. Youth delivered interactive nutrition sessions in the corner stores and were featured on program promotional materials.⁷² As another example, Baker et al (2006)¹²³ involved the community through discussions around the intersections between faith and health and identified the need to create retail food infrastructure (eg, supermarket or community-run produce market) rather than just information on nutrition and physical activity to address obesity.

Shared Decision-Making/Shared Ownership.

At the highest level of community engagement, just over half of interventions (n = 52)^{33-35,44,46-50,52,54,63,64,71-73,75-77,79-81,83-90,93,94,100-105,107,109-111,114-116,120-123,125-127} included activities or strategies in which community residents, representatives, or community-based organizations were involved in a way that influenced decision-making regarding the intervention. Community residents engaged widely across interventions from youth, faith-based advocates, lay church members, business leaders, growers, grocers, academics, and community leaders, but all were part of the project decision-making. Some of the CE research activities included having community advisory boards, involving community residents as owners, store managers, or employees, and developing community advocacy groups to influence local RFE policy. For instance, Meister and Zapien (2005)⁸¹ trained members of community coalitions (also known as special action groups) to be policy advocates and develop a policy agenda for their communities. In this intervention, RFE was a part of the larger community initiative. Special action groups performed different activities, such as making presentations at the city council, participating in long-range planning committees, organizing an annual community forum, and cooking demonstrations at grocery stores to promote healthy food. As another example, Healthy Kids Healthy Communities was part of a national effort implementing policy, system, and environmental changes to address obesity.⁸⁰ In this project, community residents from 49 cities across the United States were involved in assuming leadership roles and participating in the government decision-making process to promote the availability of affordable, healthy food at different venues, including RFE. In another study, called Tribal Health and Resilience in Vulnerable Environments (THRIVE), tribal leaders and local community residents played a critical role and participated in all aspects of intervention, from formative research to implementation in the Chickasaw and Choctaw Nations. A partnership agreement was formed at the beginning of the intervention between academic researchers and tribal nations, and all aspects of the partnership were reviewed by tribal institutional review boards.⁴⁷

Variation in CE by Type of RFE Setting, Intervention Phase, and Intervention Year

The CE score (range, 1–4) was used to compare the variation in the extent of CE research strategies across RFE setting, intervention phase, and year. Table 1 presents both the mean CE score (min/max) for each category and the proportion of studies across CE score. Exploring CE scores across RFE settings, the highest scores were observed for “mobile market” (mean = 3.7), “other” retail settings (mean = 3.7), and “multiple” settings (mean = 3.6). Settings with the lowest CE scores included supermarkets (mean = 2.2) and restaurants (mean = 2.5), with most interventions in these settings^{51,56-62,65-70,117-119} having a CE score of 2 or lower (supermarket = 14 of 21 studies; restaurants = 3 of 4 studies). These interventions were focused more on promotion of the intervention, affordability (incentive program), and the community role in program implementation and did not explicitly discuss community partnerships or their role in program decision-making. Although all settings had at least one example of an intervention incorporating shared leadership/ownership activities (ie, a CE score of 4), it was evident that interventions implemented in “mobile market,” “multiple,” and “other” settings were most consistently able to incorporate the highest levels of CE, as 70%,^{107,109-111,114-116} 76%,^{72,73,75-77,79-81,83-90} and 86%^{121-123,125-127} of interventions, respectively, received a score of 4. Interventions in these diverse RFE settings used comprehensive approaches to make system-level changes, empower youth to be community advocates, promote nontraditional business models, and link farmers with traditional RFE; thus, they required community inclusiveness from multiple organizations and enabled shared decision-making in the process.

Examining the CE score by intervention phase (Table 1), the highest score occurred when communities were engaged throughout (mean = 3.8). A total of 55^{31,33-35,40,44-50,52,53,64,71-77,79-81,83-85,87-90,93,94,97,98,100-105,109,111,114-116,118,120-123,125-127}, interventions integrated CE approaches throughout the process, with 85% (n = 47)^{33-35,44,46-50,52,64,71-73,75-77,79-81,83-85,87-90,93,94,100-105,109,111,115,116,120-123,125-127} demonstrating shared decision-making and ownership with the community. These interventions included CE activities in which community members, directly or through stakeholders representing community members, had the opportunity to take a lead in the project and guide the program design, implementation, and evaluation. In comparison, lower scores for CE were observed when communities were involved only in the implementation (mean = 2), evaluation (mean = 2), or both phases (mean = 1.9). Across these 36 interventions,^{32,36-39,41,43,51,54,56-59,61,65-70,78,82,86,91,92,95,96,99,106-108,112,113,117,119,124} communities were involved more to execute rather than decide program activities. Finally, examining CE activities across intervention years, there were no clear changes over time, with a persistent proportion of studies only implementing CE activities of outreach and consult/involve.

Considerations of Equity in the RFE Intervention

This review also examined whether interventions demonstrated various considerations for equity and how these varied across CE score (Table 2). Equity domains that were assessed included whether the RFE intervention involved trying to (1) increase the availability of healthy food at an existing store, (2) increase the availability of healthy food by introducing a new site, (3) increase the affordability of healthy food through financial support such as coupons, or (4) increase the accessibility of healthy food. Overall, interventions demonstrated a strong commitment to improving RFEs for historically and intentionally excluded communities, as 91% (n = 89) ^{10,32-54,56-67,69,71,72,74-85,87-90,92-117,119-121,123-127} of interventions had this explicit focus. Across the 4 domains of equity, interventions overall had a greater focus on increasing availability of healthy food at existing sites (74%)^{10,31-62,65,67-70,72-74,76-88,90-92,95,97-99,104,117-122,124-127} and improving its affordability (56%)^{32,33,35,37,38,40,43,44,46,48,53,56,59,61,64-69,71,76,78,79,86,87,92-111,113-116,119,123,124,126,127} rather than adding new community locations of stores selling healthy food (32%)^{48,63,64,66,71,75,76,79,87,93,94,96,100-105,107-116,123,126,127} or addressing concerns of healthy food accessibility (31%),^{48,50,64,66,67,76,87,93,94,96-98,100-105,107-116,123,125} such as improving transportation. Examining these domains across CE scores, it was observed that as CE research activities improved (ie, CE scores increased), interventions displayed a greater focus on addressing accessibility and making new sites available. In addition, although most interventions considered more than 1 equity domain (58%),^{32,33,35,37,38,40,43,44,46,48,50,53,56,59,61,64-69,71,76,78,79,86,87,92-105,107-116,119,123-127} only interventions that incorporated communities in shared decision-making and ownership addressed all 4 domains. One consistent way researchers involved communities in shared decision-making was by developing or engaging with an existing local food policy council (or board), which occurred in 2^76,104 of the 4 interventions^48,76,87,104 that addressed all equity domains. For instance, Janda et al (2021)⁷⁶ discussed working with the County Food Policy Board to promote geographic and economic access to healthy food through farm stands, mobile markets, corner stores, and subsidized prices.⁷⁶

Table 2.

Equity considerations in interventions by community engagement (CE) score among 98 studies included in the scoping review on community-engaged research strategies in food retail interventions

Score of CEb	Domains of Equitya						Historically and intentionally excluded communities considerations
	Addressed availability of healthy food at existing sites	Addressed availability of healthy food by adding new site	Addressed affordability of healthy food	Addressed accessibility of healthy food	>1 domain of access considered	All 4 domains considered
Overall (N=98)	74% (72/98)	32% (31/98)	56% (55/98)	31% (30/98)	58% (57/98)	4% (4/98)	91% (89/98)
1 (N=13) 38,39,41,43,51,57,58,61,65,68,70,78,99	100% (13/13)	0	54% (7/13)	0	54% (7/13)	0	85% (11/13)
2 (N=19) 10,32,40,56,59,60,62,66,67,69,91,92,95,98,106,117-119,124	90% (17/19)	5% (1/19)	68% (13/19)	16% (3/19)	63% (12/19)	0	89% (17/19)
3 (N=14) 31,36,37,42,45,53,55,74,82,96,97,108,112,113	71% (10/14)	29% (4/14)	43% (6/14)	36% (5/14)	50% (7/14)	0	86% (12/14)
4 (N=52) 33-35,44,46-50,52,54,63,64,71-73,75-77,79-81,83-90,93,94,100-105,107,109-111,114-116,120-123,125-127	62% (32/52)	50% (26/52)	56% (29/52)	42% (22/52)	60% (31/52)	8% (4/52)	94% (49/52)

^aEquity domains were informed by Kumanyika’s ‘Getting to Equity in Obesity Prevention’ framework²³ and key dimensions of an equitable regional food system identified by Mui and colleagues (2021).²⁴

^bCommunity Engagement Score: 1=Outreach; 2=Outreach + Consult/Involve; 3= Outreach + Consult/Involve + Collaborate; 4= Outreach + Consult/Involve + Collaborate + Shared decision making.

DISCUSSION

This scoping review examined the breadth and extent of CE in RFE interventions by exploring the range of strategies employed and how these varied across RFE settings, intervention phase, intervention year, and key equity domains for healthy food retail. The findings revealed that CE research strategies of outreach were used across all RFE interventions, whereas enabling communities to participate in decision-making (ie, the highest level of CE) occurred in only half. In addition, the strength of CE varied across RFE settings, intervention phase, and domains of equity. Interventions conducted in diverse RFE settings, such as “mobile markets,” “multiple,” and “other,” and interventions involving communities throughout the intervention design (ie, from design through evaluation) had the highest level of CE (ie, shared decision-making) based on CE activities described. Similarly, interventions addressing all 4 domains of equity were also those that employed the highest forms of CE. Surprisingly, there was very little change in the strength of CE research strategies used in RFE interventions over time.

The review also found that the phase and purpose of employing CE are critical regardless of the RFE setting. Any intervention (1) recognizing the importance of community ownership, (2) involving the community throughout the process from understanding their needs to program implementation, and (3) implementing comprehensive strategies rather than solely focusing on store promotion demonstrated the highest forms of CE. Such examples demonstrate that considering RFE as a part of a system (Policy, Systems, and Environmental [PSE] model-based interventions) rather than as a stand-alone determinant and ensuring community residents have the ability to influence decisions may be an important strategy for researchers to explore to develop more effective RFE interventions. These findings are consistent with the literature highlighting the importance of PSE change strategies to address complex health behaviors and to have a sustainable impact.¹³⁴ PSE demands collaboration across sectors that are complex and labor-intensive.¹³⁵ Partners might have variations of invested interests, and forming a common goal is critical. Once the partnership is established, researchers have emphasized additional challenges such as maintaining a partnership, shared roles and governance, transparency, and accountability.¹³⁶ Such challenges may be especially salient when researchers work in the RFE, where it can be difficult to navigate the interests and desires of the retail partner with goals to engage community members authentically. Yet, even though some RFE settings appear to be more amenable to using stronger forms of CE research strategies than others, findings of this review also suggest that a significant portion of RFE interventions in the United States have been engaging in this important work and have been able to do so across all settings. Thus, even though supermarket- and restaurant-focused RFEs had lower CE scores on average, some intervention studies were still able to have high CE scores in these settings.^{52,54,63,64,71,120}

Evaluating equity considerations, it was found that nearly all RFE interventions (91%), regardless of the degree of CE employed, explicitly focused on historically and intentionally excluded communities. However, the extent to which the domains of equity (availability at existing site, availability at new site, accessibility, and affordability) were addressed did vary by CE, with only those interventions using the highest forms of CE able to address all 4 domains. This suggests that although CE can take additional time and effort, and inherently, some nontraditional retail settings may be better partners for this work than others, greater CE can result in more comprehensive and equitable solutions for addressing RFEs. Moreover, it aligns with the principles of the Strategies to Repair Equity and Transform Community Health Initiative (STRETCH) framework that centers on system change to achieve health equity and support power-sharing in community partnerships through community-led approaches.¹³⁷

Despite evidence that greater forms of CE are linked with greater attention to equity, this review also found that there has been little to no improvement in researchers implementing the highest levels of CE—shared decision-making and ownership—across RFE interventions over time. The importance of incorporating CE in interventions and working at more authentic levels that share power has gained increasing attention in public health after the launch of the first CBPR-focused research initiative at the National Institute of Health in 1995.¹³⁸ This led us to believe that more RFE interventions would incorporate these strategies over time. However, this review found that the proportion of studies using this level of CE remained stagnant at approximately half. The idea of shared power in public health interventions is also being promoted under the term co-creation.¹³⁹ However, Morales-Garzón et al (2023)¹⁴⁰ found in their scoping review that studies discussing the “co-creation” methodology in equity-focused public health actions have a lack of community involvement in decision-making.¹⁴⁰ The recommendations for researchers, organizations, funders, and systems to make CE a requirement for prevention programs are consistent across different health domains.^141-143 Together, this suggests that more efforts are needed to deepen the knowledge and implementation of CE research strategies to attain shared ownership in public health interventions, including those within the RFE.

Strengths and Limitations

This review has several strengths and limitations. A key strength was using a broader definition of RFE that allowed authors to include less traditional retail settings, such as mobile markets, farmers markets, and vending machines, and compare the CE of these settings with traditional settings, such as supermarkets and grocery stores. In addition, this review performed an extensive database search led by a librarian and implemented the JBI²⁵ and PRISMA-ScR²⁶ guidelines to select studies and extract data. In terms of limitations, the authors did not include dissertation/thesis, gray, and non-peer-reviewed literature, which may have failed to provide a full synthesis of CE research strategies used in US RFE intervention research to date. Second, the inclusion/exclusion criteria were designed to review only RFE interventions in the United States, given the unique US context around historical and present-day racism and economic inequality, which continue to limit access to healthy food.²⁷ Future reviews should consider examining CE research approaches across a wider range of countries and assess how these compare with the US-based results. Furthermore, examining regional differences within countries, such as rural vs urban, could be useful. Finally, some interventions may have had limited details about CE activities implemented, and even with our rigorous database search, additional publications with relevant CE information about an intervention may have been inadvertently missed. Because CE research strategies could only be assessed based on activities discussed in the published articles identified, in which authors are often forced to make difficult choices about what study details to include because of limited journal word counts, this may have resulted in some inaccurate CE scores. Furthermore, information on who from the community was engaged—members that could directly benefit from the intervention or a leader representing community interests—could not be extracted because that detail was typically not apparent in publications. Together, this implies that there may be a need for both dissemination standards to ensure CE research strategies are adequately described in future intervention publications¹⁴⁴ and a shift in journal priorities toward publishing more CE-focused method papers.

CONCLUSION

Effective CE in RFE interventions is challenging but crucial. Documenting how and to what extent CE research strategies occurred is important to further understand what research gaps remain and explore the ways CE strategies could be leveraged for a more equitable intervention design. The findings of this review highlight that public health researchers who collaborated with communities throughout all phases of the project and aimed to address multiple domains of equity related to food retail were most likely to employ the highest levels of CE (ie, shared decision-making). As such, future RFE intervention research should assess whether shared leadership CE strategies—which strive to ensure community voice is incorporated in all decisions—are a more effective approach than other CE approaches in narrowing nutrition-related health equity outcomes for communities. Moreover, the limited information about CE research strategies described in some publications suggests that standardized dissemination procedures for intervention protocols could be beneficial to ensure a more complete and accurate understanding of the CE research activities employed.

Supplementary Material

Figure 1 is available at www.jandonline.org

Figure 1.

Search strategy and databases used in a scoping review of community-engaged research strategies in retail food environment interventions.

RESEARCH SNAPSHOT.

Research Question:

What are the community engagement (CE) research strategies used in retail food environment (RFE) interventions, and do they vary by type of retail settings, phase of intervention, year of intervention, and key domains of equity?

Key Findings:

Across the 98 included RFE interventions, all forms of CE research strategies, including outreach, consult/involve, collaboration, and shared leadership, were identified, with outreach used in all interventions and shared leadership in half. The form of CE research strategy varied across intervention characteristics (eg, intervention phase), and interventions using stronger CE forms also addressed more domains of equity.