The implementation and effectiveness of outlet‐level healthy food and beverage accreditation schemes: A systematic review

Summary Healthy food outlet accreditation schemes represent an avenue for incentivizing food retailers to promote healthy eating patterns by improving the healthiness of food environments. This systematic review aimed to (i) assess the impact of food outlet‐level accreditation schemes on outlet practices and customer purchases and (ii) identify barriers and enablers to scheme implementation. Peer‐reviewed and grey literature were systematically searched. Eligible studies related to outlet‐level food and beverage accreditation schemes across any food retail setting. Findings were narratively synthesized by retailer type according to (i) scheme characteristics (governance, targeted products, support, and monitoring); (ii) scheme outcomes (rate of uptake, proportion of certified retailers, impact on purchasing, customer perspectives, and retailer perspectives); and (iii) barriers and enablers to implementation. From 21,943 records screened, 48 were included, covering 26 schemes. Most (18) targeted restaurants or convenience stores. Average uptake was 65% of all outlets approached to participate. Implementation of accreditation schemes was associated with healthier customer purchases in convenience stores, schools, and hospitals, but evidence from restaurants was mixed. Enablers of scheme implementation included support for implementation and maintenance, flexible scheme criteria, and motivated retail staff. Healthy food outlet accreditation schemes represent a promising mechanism for engaging retailers to improve the healthiness of food retail environments.

within the consumer nutrition environment that seek to improve population eating patterns are potentially powerful from a public health perspective. Multiple systematic reviews have found that interventions that change food and beverage environments so that merchandising and marketing of foods and beverages favors healthy options can lead to healthier purchasing and consumption. 10,11 Healthy food and beverage interventions have been previously characterized as modifying one or more of the "4Ps" of food environment merchandising: the available "products," the presence and use of "promotions" to advertise those products, the "prices" at which those products are sold, and the "place" or positioning within the food outlet. 11,12 Other research that has explored elements likely to improve the healthiness of food environments has expanded on the 4Ps framework by identifying additional intervention targets in food retail settings. These additional intervention targets, supplementing the 4Ps to make up the 7Ps, include the "people" (or employees) who sell products, the "processes" by which products are delivered to the consumer (including, for example, the standard side dishes or condiments served with a meal), and "partnerships" between retailers and other stakeholders. 13 Engaging food and beverage retailers to intervene across the 7Ps and change the consumer nutrition environment at food retail outlets remains a challenge. [14][15][16] In particular, previous studies have identified that retailers frequently express concerns that healthy food and beverage interventions will compromise business commercial viability and that appropriate healthier product alternatives are not readily available to stock. 17,18 Additional challenges previously identified by retailers include lack of perceived consumer demand for healthy food, confusion in what constitutes a "healthy" food offering, and fear of profit loss. 16 Food outlet-level accreditation schemes (hereafter referred to as "schemes") represent one strategy for engaging with retailers to improve the healthiness of food and beverage outlets. 10 Using predefined criteria 10 to assess organizational practice(s), such schemes may increase the healthiness of consumer purchases within food outlet settings by changing, among other food environment characteristics, the relative availability, placement, promotion, and price of healthier options. [19][20][21][22] However, we are only aware of one systematic review that has included an analysis of the impact of such schemes. The previous review, conducted in 2017, focused on the impact of accreditation schemes on practices to promote healthier ready-to-eat meals, finding increases in healthier catering practices and availability of healthier options. 10 Both included studies were of a weak study design. No systematic reviews have comprehensively examined the impact of schemes on a full range of outcomes of interest to retailers and policymakers, including changes to outlet practices; impacts on consumer purchasing behavior; retailer and customer awareness; understanding, satisfaction, and support of schemes; and barriers and enablers to successful scheme uptake and maintenance. A greater understanding of the implementation and impact of schemes has the potential to lead to improved scheme design that could result in greater uptake of these initiatives by retailers and policymakers and increase the effectiveness of these schemes to improve population health and nutrition outcomes.
The aim of this systematic review was to assess the impact of nutrition-related food retail outlet-level accreditation schemes on food retail outlet practices and customer purchasing behavior. We also aimed to identify the reported barriers and enablers to scheme implementation, including scheme uptake (the proportion of retailers signing up for a scheme) and certification (the proportion of retailers meeting scheme requirements).

| Search strategy
The selection, analysis, and reporting of the results for this study were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 23 (see Table S1 for the completed checklist). An initial scoping review was conducted to identify key studies, which were used to inform the final search strategy. The search protocol was registered online with PROSPERO on April 3, 2021 (CRD42021240769).

Electronic databases (Embase, EBSCO Medline, EBSCO Global
Health, EBSCO Business Source Complete, and ERIC) and grey literature (Google Advanced Search, first 100 results) were systematically searched to identify studies that related to retail outlet-level healthy food and beverage accreditation schemes. The four hedge terms ("food outlets," "accreditation schemes," "nutrition," and "outcomes of interest") were combined with the operator "AND," and within each hedge, specific search terms were combined with the operator "OR." The included EBSCO Medline-specific search terms are shown in Table 1. Boolean search operators were adjusted for each database searched (see Table S2 for full search strategies for each database). Backward searching of reference lists of included articles and forward searching of articles that have cited included articles were undertaken to optimize the search process. A research librarian was consulted to develop this search strategy.

| Inclusion criteria
To be included, studies must have reported on the adoption of, or compliance with, nutrition-related food outlet-level schemes, or the impact of participation on outcomes of interest (Table 2). Accreditation schemes were defined as interventions or programs, including awards, accreditation, or other recognition, based on an assessment of organizational practice(s) using predefined criteria. 10 To be included, the outlet, organization, or the scheme had to be related to food provision or food retail, and the scheme had to include an element related to food or nutrition, including nutrition schemes, alcohol provision policies, or food sustainability policies. We included observational and experimental study designs.

| Study selection process
Following the database searches, article titles, keywords, and abstracts were imported into COVIDENCE for removal of duplicates and subsequent screening by two independent authors (OH and TBR).
Articles deemed to be potentially relevant based on title and abstract content had their full texts screened against the inclusion/exclusion criteria by two independent authors (OH and JA). Any disagreements were resolved through discussions with a third author (MB). For the grey literature search, the first 10 pages of search results (100 results) were screened by two authors (OH and MB) to identify potentially relevant records.

| Data extraction
Two authors independently extracted data from included studies using a standard template in Microsoft Excel (for each included study, any two of OH, SS, JA, MB, and TBR completed data extraction). This information was cross-checked, and any disagreements were resolved through discussion with a third author. Extracted data included bibliographic data, study design, study funding, scheme criteria required to receive accreditation, value proposition to retailers, governance (engagement and recruitment, enforcement, assessment, and monitoring processes), implementation (uptake, certification, provision of support and resources, responsibility for implementation, and monitoring), outcomes (effect on customer purchasing or eating patterns, business outcomes [customer perspectives, retailer perspectives, and commercial viability], 24 cost-effectiveness, and process outcomes), program costs/resources, scheme duration, and sustainability. Scheme uptake was defined as the proportion of retailers who committed to participate in the scheme relative to the number of retailers that were T A B L E 1 Search terms for EBSCO Medline. invited to participate. Scheme certification was defined as the proportion of retailers who received the scheme award or accreditation relative to the total number of retailers who committed to participate. The aspects of the food environment that were targeted by schemes were defined according to the 7Ps: "product," "promotion," "price," "place," "processes," "people," and "partnerships." [11][12][13]25 2.5 | Quality appraisal The quality of included quantitative, qualitative, and mixed methods studies was determined using the Mixed Methods Appraisal Tool Although the original MMAT does not provide cut-offs for high-, medium-, and low-quality studies, we applied previously applied cutoffs 26 to identify studies that scored ≥80% as high-quality studies, studies that scored 50%-80% as medium-quality studies, and studies that score ≤50% as low-quality studies.

| Data synthesis
Results were narratively synthesized because of anticipated heterogeneity of populations and outcomes. 27 Schemes were grouped by the target food outlet type: (i) restaurants, cafés, and bars; (ii) convenience and corner stores; (iii) schools and childcare centers; (iv) hospitals and healthcare settings; (v) other general workplaces (henceforth referred to as "workplaces"), or (vi) multiple different settings targeted with the same broad accreditation criteria.
Scheme findings were summarized according to (i) scheme characteristics, including food environment targets, scheme governance, support offered for scheme implementation, and scheme monitoring and compliance; (ii) scheme outcomes, including uptake, certification, impact on purchasing, customer perspectives, and retailer perspectives; and (iii) barriers and enablers to scheme implementation. Data synthesis included the vote-counting method following Cochrane advice 28 to summarize effect estimates for scheme outcomes of interest as in (ii) above. Schemes were counted as having an overall "positive" or "negative"/"neutral" impact. Schemes were considered as having a "positive" impact if the study supported the award scheme: increased uptake of healthy retail practices, increased the healthiness of customer purchases, was supported by customers or retailers, or was associated with commercial outcomes favorable to retailers.
Outcomes were also classified as "positive" if over 50% of the counted variable were deemed as having been an improvement to food environments, or customer or retailer behaviors or perspectives.
Barriers and enablers to scheme implementation were extracted if they were explicitly identified by the original authors as barriers and enablers in the results sections of included studies. The authors of the current review then inductively coded each barrier and enabler according to similar constructs (e.g., lack of retailer time for implementation). These barriers and enablers were then grouped into themes.

| Study characteristics
The date of publication for included studies ranged from 2004 to 2021 (Table S4). Of the 48 included studies, one was a randomized controlled trial, one was a nonrandomized quantitative experimental study, 22 were descriptive observational quantitative studies, seven were observational qualitative studies, and 18 were observational mixed methods studies. As determined by the MMAT, 26 18 of the included studies were high quality, 15 were medium quality, and 15 were of low quality. Details of the MMAT scoring for each study are found in Table S5.

| Accreditation scheme characteristics
The 48 included studies covered 26 different schemes (Table S6). Of these 26 schemes, 14 were based in the USA, five each were based in Canada and the UK, and two were based in Australia. Nine schemes targeted restaurants, nine targeted convenience and corner stores, three targeted schools and childcare settings, one targeted hospitals, one targeted workplaces, and four targeted multiple retailer types.

| Accreditation scheme governance
Of the 26 identified schemes, 22 had their governance mechanism, or their overarching managing body, described. 19 29,30 Accreditation scheme food classification criteria were applied in various ways but were most commonly used to encourage product or menu reformulation to meet a set standard and designate which products should be relatively more or less available, and which products should be promoted.

| Accreditation scheme environmental changes
All components of the food environment (as characterized by the 7Ps [11][12][13]25 ) were targeted by at least one identified scheme. All identified schemes targeted some aspects of "product." Two common examples of this included changing the available food products to make menus healthier and offering smaller portion sizes. For example, the US ¡Por Vida! Initiative encouraged restaurants to alter menu items to meet a range of nutrient criteria, 21 whereas US "Shape Up Somerville: Eat Smart, Play Hard" included requirements for provision of smaller portion sizes in restaurants. 20 Seventeen schemes encouraged food retail outlets to implement "promotions." This was commonly the use of posters, table tents, and other promotional materials to promote healthy eating and/or the scheme itself. Eleven schemes attempted to change the "people" aspect of food environments, most commonly by training outlet staff in the preparation and upselling of healthier foods and beverages. Ten schemes encouraged food retail outlets to change their "processes." This took the form of fundamental shifts in the way that food was offered at outlets, without changing the food that was actually available. For example, under the "Healthier Catering Commitment", UK fast food outlets were encouraged to no longer add salt to menu items (instead allowing customers to do so themselves). 19 Seven schemes targeted the "place" aspect of food environments, often by changing store layouts to make healthier items more accessible or prominent. For example, "Choose Health LA Restaurants" required that drinking water be easily accessible in restaurants. 21 Six schemes leveraged "partnerships" to promote healthier customer choices. This included partnering outlets with local councils to provide support with promoting healthy eating and partnering outlets with other healthpromoting businesses. For example, corner stores that signed up for the "Healthy HotSpot" initiative received support in the form of community outreach and assistance with engagement with local institutions. 42 Finally, one scheme encouraged food retail outlets to change their "prices" to make healthier foods and beverages relatively more affordable.

| Accreditation scheme implementation support
Of the 26 identified schemes, 24 were reported as offering outlets support to implement the scheme. 19

| Accreditation scheme certification
Accreditation scheme certification refers to the proportion of participating schemes that achieved the scheme requirements to be certified.
Thirty-five included studies reported on scheme impact on the healthiness of food outlets ( Figure 2 and Fourteen schemes reported on the proportion of retailers who achieved scheme certification (nine schemes were reported on in high-quality studies). The mean certification rate among these schemes was 64% (range 6%-100%). The mean proportion of certified retailers reported in high-quality studies was 54% (range 6%-100%). 20,44,45,49,50,54,56,67,69 Schemes that provided multiple levels of certification (such as bronze, silver, and gold certification levels) commonly had a higher proportion of certified retailers as businesses could aim for a lower level and still be considered "certified." For example, 24 private hospitals joined the Healthy Hospital Food Initiative. Nine (38%) reached "gold" accreditation (achieving nutrition standards in all four domains), seven (29%) reached "silver" accreditation (achieving nutrition standards in two or three domains), F I G U R E 2 Accreditation scheme impact counting for included studies. three (13%) reached "bronze" accreditation (achieving nutrition standards in one domain), and five (21%) did not implement any standards. 50 Therefore, the Healthy Hospital Food Initiative had an overall proportion of certified retailers of 79%, although just 38% of hospitals reached full ("gold") certification.
Another observed driver of the proportion of certified retailers was the number of retail outlets recruited for participation. Schemes with fewer participating outlets tended to provide greater levels of support and had a higher proportion of certified retailers. Waupaca Eating Smart recruited seven restaurants and two supermarkets to participate, and all outlets implemented some Waupaca Eating Smart activities. 52 In comparison, 2989 schools were recruited to participate in the Food For Life Partnership, and 192 achieved any certification. 56

| Accreditation scheme impact on eating patterns and purchasing
Eighteen included studies reported on scheme impact on customer eating patterns and purchasing ( Figure 2 and The impact of scheme participation on customer nutrient intakes was reported by a high-quality study for just one scheme: Start Right-Eat Right. 48 This scheme was associated with improvements in both food and nutrient intakes among children attending participating centers.

| Accreditation scheme retailer perspectives and commercial outcomes
Seventeen included studies reported on scheme impact on retailer perspectives or commercial outcomes ( Figure 2 and

| Accreditation scheme customer awareness and perspectives
Twelve included studies reported on scheme impact on customer awareness and perspectives ( Figure 2 and Table S8) where measured, was consistently high (4/4 schemes). 29,30,35,39,51 Conversely, customer awareness and understanding of schemes was usually low (4/5 schemes), and this was associated with low scheme impact on the healthiness of customer purchases. 34,36,41,57,58 Conversely, Project FIT saw increases in customer understanding of the scheme and associated increases in purchasing of healthier foods. 69

| Barriers and enablers to accreditation scheme implementation
Barriers and enablers to implementation were reported for 20 schemes (Table 3). In general, elements that were associated with scheme uptake, certification, and impact on customer purchases explicitly considered the perspectives, value to, and support of the retailer.
T A B L E 3 Summary of barriers and enablers of accreditation scheme implementation.

Scheme characteristics Enablers Barriers
Scheme criteria Flexible delivery allowing retailers to select which scheme criteria they wished to meet 19,20,32,64 Eligibility criteria to join exclude some businesses 19,64 Tiered scheme with multiple levels of criteria 19,64 Short initiative timeframe 42 Feasible, culturally acceptable, and tailored delivery to businesses 19,64 Scheme not worthwhile given retailer food options 70 Some businesses only have small changes to make to meet criteria 19,64 Resource and time intensity of delivery 19,32,58,63,64,68 Incorporating pilot scheme learnings to ensure compatibility with business practices 47 Requirements for staff training and support 47,53 Low resource and space requirements 53 Retailer recruitment Convenience of applying to participate 70,72 Misunderstandings about how to qualify for the program 72 Existing scheme easy to pick up 19,64,70 Slow approval process 42 Several key and overarching enablers of scheme implementation were identified. Pilot schemes, supported by research, were noted as useful for ensuring compatibility with businesses and facilitating implementation. 47 Such compatibility included ensuring that scheme criteria aligned with what was feasible for retailers, developing a convenient scheme recruitment strategy, ensuring strong communication and engagement from governing bodies with retailers, and providing ample support with implementation. Retailers were more likely to participate in a scheme that they perceived as easy to pick up 72 and where success was publicly recognized and led to a competitive advantage. 19 Barriers to scheme implementation were also identified. These included promotional activities or materials being unsuitable for display in some businesses, 20,33,40 poor publicizing of a scheme, 33 a slow process for the ordering and delivering of scheme equipment and materials, 42 and low levels of retailer time and availability, 20,37,38,46,47,72 resources, 37,38,40 or technical skills 32 to implement an accreditation scheme. Further, some schemes were relatively time or resource intensive to implement. 19,32,58,64,68 The requirement to provide healthy products was also seen as a barrier, as some retailers reported difficulties in sourcing these, 29  Key factors associated with scheme uptake and implementation identified included support provided for scheme implementation and maintenance, flexibility for retailers in meeting scheme criteria, and motivation of retailers and staff. Average uptake across the seven schemes reporting uptake was 65% (range 43%-88%). Customer purchases of targeted healthier items increased in 10 of the 14 schemes, which reported on scheme impact on the healthiness of customer purchasing. Only the Start Right-Eat Right early childhood education scheme was evaluated for impact on nutritional intake (rather than using purchasing as a proxy measure for healthier consumption). 48 With heterogeneity in outcomes, it was difficult to assess the overall magnitude of scheme impact on purchasing behavior.
A meta-analysis 74  Multicomponent retail food environment interventions, and interventions that make larger changes to the food environment (targeting multiple of the 7Ps, rather than just "products"), have been more consistently associated with favorable impacts on customer purchasing behaviour. 10,11 Schemes that target multiple aspects of the retail environment may therefore be more effective at improving the healthiness of customer purchases. Tiered schemes, with multiple levels to achieve, were associated with increased scheme uptake, 64 healthy changes to the food environment, 19,39,40 and increases in customer purchases of healthier items. 40 There was also evidence that retailers might "stall" at lower scheme levels and fail to make further health-promoting changes. 40 The evidence synthesized here suggests that such schemes should incorporate greater support and incentives for retailers to make further changes to achieve higher levels of certification. In the case of the Healthy HotSpot Initiative, retailers were awarded additional funding in response to further changes to the store environment. 42,43 Across the included schemes, customer support for schemes was high, although awareness and understanding were generally low. Low scheme awareness was associated with a low impact on customer purchases. Behavior economic theory suggests that consumer awareness of an intervention is not required for changes in consumer behavior. 75 Indicators of retailer satisfaction with schemes were also high, and qualitative research echoed previous research findings that retailers' ability to contribute to customer and community wellbeing was an important motivator. 24 It is likely that at least some of the favorable retailer feedback is related to a participation bias, as retailers more supportive of healthy food environments are more likely to adopt such schemes. As evaluations did not typically compare participating and nonparticipating retailers, it is not known which scheme formats have the strongest recruitment potential, nor how to expand the reach of schemes to a greater diversity of retailers (and communities).
The higher proportion of convenience store schemes reporting increases to the healthiness of customer purchases (83%) compared with restaurant schemes (33%) may be related to the small number of included restaurant schemes, or perhaps variation in the accreditation criteria used in different settings. As convenience stores typically sell packaged nonperishable food, these outlets may be able to offer direct substitutes for less healthy alternatives more easily than restaurants, which may require more substantive changes in food storage, cooking, and ordering practices. Several schemes targeting restaurants noted the complexity and time intensity of full nutritional analyses of menus by registered dietitians. 29,30,50 Schemes using simple foodbased criteria (e.g., restrictions on deep-fried foods and encouraging fresh fruit and vegetables) may therefore be more easily understood by retailers and less costly to monitor but may also only be appropriate for certain retailer types, such as those offering substantive quantities of fresh fruits and vegetables. We are aware of only one previous systematic review that examined the findings of two award schemes 10 but was unable to draw general conclusions about this type of retail intervention or factors that are likely to be associated with scheme impacts. More work is needed to further understand the characteristics of impactful schemes in these settings and whether scheme characteristics including accreditation criteria or implementation support may need to differ by setting.
In the current review, the most common enablers reported to increase uptake and implementation of schemes included a tiered approach to scheme participation with multiple levels of achievement, public recognition of certification, and provision and sharing of resources including support provided by people skilled in nutrition science to implement changes, franchise or retailer executives' approval to participate, high retailer motivation, and collaborative efforts and partnerships drawing on shared expertise. The most common barriers reported included difficulties in sourcing healthy products; lack of retailer or franchise engagement; and cost, resourcing, and time intensity of implementation. These enablers and barriers were largely similar to those identified in a recent review of reviews of factors influencing implementation of healthy food retail interventions, 16 which found key influences including "Retailer knowledge, skills and preferences regarding healthy food (and interventions)," "Organisational Support (Control and Ownership over Food Store Supplies)," "Resources (Staff, Time, Capital)," and "Establishing Partnerships" with a range of stakeholders. Further reviews of factors affecting implementation of healthy food retail interventions have also emphasized difficulties in maintaining a constant supply of healthy alternatives at an affordable price. 76,77 The current review did not make direct comparisons between the characteristics or the barriers and enablers of accreditation schemes compared with other healthy food retail initiatives. It is likely that the "offer" to the retailer, including public recognition or certification, is particularly important in the context of accreditation schemes.

| Implications for practice
This systematic review provides the first synthesis of evidence that food retail accreditation schemes may be effective in improving the healthiness of some consumer food environments. Although none of the studies included in this review examined changes in energy intake or weight outcomes, accreditation schemes are unlikely to have a significant impact on population weight when used alone. Addressing the obesogenic food environment is widely acknowledged to require changes throughout the food system. 78 Our findings suggest that accreditation schemes may be an effective mechanism of engaging commercial retailers in healthy food retail change, an otherwise hard-to-reach group. 16 This review provides a number of key lessons for those designing and supporting healthy food retail accreditation schemes. We have formulated recommendations for governing bodies to guide scheme development and implementation based on the barriers and enablers reported and key common elements of scheme criteria, design, implementation support, monitoring and evaluation, and governance (see Figure 3).

| Strengths and limitations of studies included in the review
As determined by the MMAT (26), 17 included studies were high quality, 13 included studies were medium quality, and 16 included studies were of low quality. Of the 16 low-quality studies, 10 were quantitative descriptive studies. These commonly scored low because of a lack of reporting on methodological characteristics. This highlights the F I G U R E 3 Recommendations for practitioners to support the implementation of healthy food retail accreditation schemes.
importance of ensuring that methodological approaches to quantitative evaluations of accreditation schemes are robustly described. Additionally, 10 of the 17 high-quality studies were quantitative descriptive studies.
Our review was restricted to articles in English but did not exclude studies based on scheme location. Nineteen of the 26 schemes identified were located in North America, and all schemes were located in English-speaking OECD-member countries. This may limit the generalizability of the results to high-income countries, which may have different retail environments and consumer expectations. 79  This review did not identify any cost-effectiveness evidence for food retail accreditation schemes. Future research is required to determine the value for money of these interventions from various perspectives.

| Methodological strengths and limitations of the review
This study is the first review of food retail accreditation schemes that captured outcomes needed in the design and execution of schemes by policy maker and retailers. The inclusion of both grey and peerreviewed literature and studies with a range of qualitative, quantitative, and mixed methods designs and a range of outcomes facilitated discussion of the holistic impacts and considerations for implementing such schemes. Heterogeneity in study design and scheme design increased the difficulty in synthesizing associations between scheme characteristics and outcomes. As the review did not include retail interventions other than accreditation schemes, we were unable to make direct comparisons with the effectiveness of different approaches to incentivizing and supporting retailers to make food environment changes. Finally, we did not explicitly focus on the process of scheme development, which may limit the application of these results to future scheme development.

| CONCLUSIONS
Nutrition-related food outlet-level accreditation schemes represent a promising mechanism for engaging food retailers to improve the healthiness of food retail environments. Accreditation schemes may offer different incentives and accountability mechanisms, although it is unclear if they are more effective than other kinds of healthy food retail interventions. Schemes appear to be influenced by many of the same barriers and enablers as other healthy food retail initiatives, emphasizing the need to address structural barriers to retailer changes including the supply of healthier food products. Further research is required on the impacts of accreditation scheme participation on the healthiness of customer purchases and population eating patterns.

ACKNOWLEDGMENTS
Open access publishing facilitated by Deakin University, as part of the Wiley -Deakin University agreement via the Council of Australian University Librarians.