Abstract
Supermarket-based interventions are one approach to improving the local food environment and reducing obesity and chronic disease in low-income populations. We implemented a multicomponent intervention that aimed to reduce environmental barriers to healthy food purchasing in a supermarket in Southwest Baltimore. The intervention, Eat Right-Live Well! used: shelf labels and in-store displays promoting healthy foods, sales and promotions on healthy foods, in-store taste tests, increasing healthy food products, community outreach events to promote the intervention, and employee training. We evaluated program implementation through store environment, taste test session, and community event evaluation forms as well as an Employee Impact Questionnaire. The stocking, labeling, and advertising of promoted foods were implemented with high and moderate fidelity. Taste test sessions were implemented with moderate reach and low dose. Community outreach events were implemented with high reach and dose. Supermarket employee training had no significant impact on employees’ knowledge, self-efficacy, or behavioral intention for helping customers with healthy purchasing or related topics of nutrition and food safety. In summary, components of this intervention to promote healthy eating were implemented with varying success within a large supermarket. Greater participation from management and employees could improve implementation.
Introduction
In Baltimore City, African Americans suffer from higher rates of diet-related chronic diseases, including obesity, diabetes and hypertension, than other ethnic groups (Baltimore City Health Department, 2010). Environmental barriers to purchasing healthy foods may contribute to these disparities (French, Story & Jeffery, 2001). Forty-three percent of African American neighborhoods have low availability of healthy foods compared to 4% of White neighborhoods (Franco, Diez Roux, Glass, Caballero, & Brancati, 2008). Recognition of unequal access has led to strategies that focus on altering the food environment in underserved neighborhoods (Escaron, Meinen, Nitzke, & Martinez-Donate, 2013).
Supermarket interventions that have implemented one or more strategy, including point-of-purchase information; price manipulations; increased availability, variety, and convenience of healthy options; advertising; and researcher–management collaborations, have been effectively implemented; however, their impact on healthy food purchasing has been mixed (Baquero, Linnan, Laraia, & Ayala, 2014; Escaron et al., 2013; Fisher & Strogatz, 1999; Foster et al., 2014; Glanz & Mullis, 1988; Glanz & Yaroch, 2004; Paradis et al., 1995; Patterson et al., 1992; Rodgers et al., 1994). This study advances research on supermarket interventions because it is one of the first multicomponent interventions in a large, urban supermarket that was initiated and supported by the store owner, which included additional strategic community outreach and employee training components. Combining these elements, we implemented the Eat Right-Live Well! (ERLW) campaign, with the primary goal of increasing sales of promoted healthy foods.
Formative research for ERLW in a low-income, largely African American Baltimore neighborhood suggested that purchasing was influenced by in-store labeling, product placement, and promotion of inexpensive processed foods (Zachary, Palmer, Beckham, & Surkan, 2013). Participants in the study reported difficulty finding affordable healthy foods in the supermarket and obtaining information about healthy products from store employees. As a result of these and other factors, shoppers often purchased relatively cost-effective but calorie-dense foods, often higher in fat, sodium, and carbohydrates (Zachary et al., 2013).
Through the present process evaluation of ERLW, we aim to understand: (1) how well the ERLW intervention was implemented in terms of fidelity, reach, and dose; (2) how barriers and facilitators affected the intervention; and (3) whether the intervention was sustainable.
Method
Setting
The ERLW intervention supermarket was located in Southwest Baltimore, a predominantly African American community where average life expectancy is 6 years shorter than Baltimore overall (Ames et al., 2011). Median household income is $27,752 and approximately 30% of residents live below the federal poverty level (Baltimore Neighborhood Indicators Alliance, 2013). The community has several small corner stores but only two supermarkets. The study supermarket is a 40,000-square-foot full-service retail grocery store with a large selection of discounted food items.
The Eat Right-Live Well! Supermarket Intervention
The intervention was initiated by the supermarket owner who expressed concern over his customers’ purchasing habits and sought to improve them through collaboration with a university-based research team. The owner, store management, and research team participated in planning meetings that were held over the course of a year.
Based on previous research showing strong support for multicomponent interventions and our formative research, ERLW aimed to increase the availability and recognition of healthy foods, while reducing costs through: (1) stocking more healthy foods, (2) identifying healthy foods with shelf labels and in-store displays, (3) sales and promotions of healthy foods in the store circular, (4) in-store taste tests and healthy recipe card distribution, (5) community outreach events to promote healthy eating, and (6) store employee training.
The intervention was implemented between April and December of 2012. The 475 promoted food items chosen for increased stocking, labeling, and in-store advertisement were selected by a registered dietician (RD) on the research team using U.S. Food and Drug Administration (FDA) and Institute of Medicine (IOM) guidelines to select foods (FDA, 2009; Wartella, Lichtenstein, Yaktine, & Nathan, 2011). Approximately 50 new items were selected as healthy alternatives to commonly purchased items, based on the supermarket’s sales data and findings from our formative research. For example, the popularity of traditional canned vegetables led to stocking, labeling, and advertisement of canned vegetables with no added salt. Previously stocked healthy foods were selected for labeling. Promoted food items fell into six categories: (1) condiments and spreads (49 items); (2) fruit and vegetable products (93 items); (3) dairy products (63 items); (4) soda, snacks, and desserts (141 items); (5) grains, such as cereals and pastas and bread (80 items); and (6) miscellaneous items, which included canned tuna, soup, and frozen dinners (49 items). Some items in each category corresponded to the same product or food sold in different sizes (e.g., small and large boxes of the same brand of cereal).
The promoted products were then assigned shelf labels based on their healthy attributes. Label content was developed based on findings from the formative research on the main dietary concerns of the supermarket customers (i.e., “low sodium” for cardiovascular disease and “healthier sugar level” for diabetes). In combination with those findings, the RD determined the final labels based on FDA and IOM nutritional labeling guidelines: “low fat” and “low sodium” based on FDA standards, “healthier sugar level” based on IOM criteria, “100% juice” based on the product’s ingredient list, and “better choice,” which was used for healthier alternatives that did not fall into one of the above label categories (e.g., whole-wheat pasta and canned peaches in water; FDA, 2009; Wartella et al., 2011).
In July of the intervention period, ERLW in-store advertisements were displayed throughout the store. These promotional messages were developed by the RD and a graphic designer and were used to draw attention to ERLW, publicize the promoted products (e.g., skim milk vs. whole milk), and educate customers on the ERLW shelf labeling scheme. In total, 19 advertisements were placed throughout the store, ranging from small posters on freezer cases to large banners hanging from the ceiling.
Taste tests were conducted by trained graduate and undergraduate students, a community leader, and one of the study investigators who worked with the in-store manager to schedule taste test sessions. Sessions lasted for approximately 2 hours. Taste tests were based on simple recipes and ready-to-eat foods that were selected from the 475 promoted products. Taste tests started in June, but evaluation of the sessions did not begin until August. For the taste tests, shoppers sampled the promoted foods and received materials about the intervention. Facilitators explained the labeling scheme and invited shoppers to participate in community outreach events.
Community outreach events were implemented starting in July to increase the visibility of ERLW in the neighborhood. Initially, events were organized by reaching out to community partners from the formative research phase. Additional community partners were directly approached, referrals made, and events scheduled with each of them. Events targeted adults in the local area surrounding the intervention supermarket, as well as children, since the formative research indicated that children influence adults purchasing behaviors (Wingert, Zachary, Fox, Gittelsohn, & Surkan, 2014).
Employee training included five modules—orientation to the intervention, general nutritional concepts, customer service to encourage healthy purchasing, food safety, and background in food marketing and messaging—and were conducted over 10 training sessions between April and June. Graduate students and faculty from the Johns Hopkins Bloomberg School of Public Health led the sessions. Each training session was approximately 1-hour long and reached most of the 102 supermarket employees from diverse positions throughout the store. A small subset of these attendees who expressed interest in further engagement with the intervention were asked to participate in an employee advisory board that met regularly with the research team to discuss the continued implementation as well as barriers and successes of ERLW.
Instruments
Store Environment Evaluation
Fidelity was evaluated by capturing the availability of promoted food items, proper placement of ERLW shelf labels, and placement of in-store advertisements. Evaluation occurred bimonthly: once during the 2-week period when Supplemental Nutrition Assistant Program (SNAP) benefits were distributed and commonly used, which resulted in an increased number of customers in the store, and once when benefits were not distributed.
Taste Test Evaluation
Interventionists assessed reach and dose by measuring the number of visitors who participated in each taste test session and the number of recipe cards and shelf label guides distributed during each session. Participant reactions to taste test items were also recorded as positive, negative, or neutral and were then totaled at the end of each session.
Community Events Evaluation
Reach and dose at community events included documentation of the number of participants at each outreach event, the educational messages delivered, and the featured promoted foods. Interventionists recorded descriptions of the participants and the organization hosting the event. An interventionist evaluated the event as it occurred.
Employee Impact Questionnaire
The Employee Impact Questionnaire (EIQ) assessed self-reported knowledge, self-efficacy, and behavioral intention regarding healthy purchasing among the supermarket employees. These constructs are related to social cognitive theory and the theory of planned behavior and were adapted from a similar questionnaire used in food environment interventions (Mead, Gittelsohn, Roache, Corriveau, & Sharma, 2013; Suratkar et al., 2010). Knowledge and self-efficacy are derived from behavioral capability and self-efficacy constructs from social cognitive theory, while behavioral intention is based on the same construct from theory of planned behavior (Ajzen, 1991; Bandura, 1986). The questionnaire from which the EIQ was adapted exhibited moderate reliability and face validity (Mead, Gittelsohn, De Roose, & Sharma, 2010; Suratkar et al., 2010).
To promote the intervention and assist shoppers in making healthy purchases, the employee training focused on several concepts including an overview of ERLW, nutrition, food safety, and customer service. The EIQ consisted of 42 items, 14 for each construct: knowledge, self-efficacy, and behavioral intention. The knowledge scale used multiple-choice items with four possible answers and covered topics such as low-sodium foods, cross-contamination in food preparation, and appropriate responses to customer complaints. The self-efficacy scale used four-level Likert-type items ranging from very difficult to very easy on how easy or difficult employees found, for example, explaining the health benefits of certain foods, dealing with expired foods, and providing answers to customers. The behavioral intention scale used three-level Likert-type items from not willing to very willing and included questions focusing on how willing employees would be to undertake behaviors such as demonstrating proper serving sizes, washing their hands, and helping upset customers. Questionnaires were conducted prior to the intervention in March, and in July and August during the intervention but after the training sessions.
Analysis
Fidelity for stocking, labeling, and advertisements was calculated. For stocking, fidelity was calculated on a monthly basis by dividing the number of correctly stocked promoted food items documented at the two bimonthly evaluations by the total possible number of items assessed in these two evaluations. Fidelity for labeling and advertising was calculated in a similar manner. Fidelity standards were based on previous process evaluations of healthy eating interventions in low-income, urban communities in Baltimore; low fidelity was considered 0% to 49%, moderate 50% to 74%, and high 75% to 100% (Gittelsohn et al., 2010).
Taste test session reach and dose were calculated as counts and means. The total number of sessions per month was calculated as counts while numbers of participants, recipe cards, and label guides distributed were calculated as means per session. Reach was based on the average number of participants per number of sessions in a month; low reach was defined as 0 to 24 participants, moderate was 25 to 49, and high was 50 and over. Dose was based on an average of two educational giveaways (e.g., label reading guide, recipe card) provided per visitor; low dose was zero or one giveaway and high dose was two or more giveaways.
Reach for community outreach events was calculated as the mean number of participants per month. Descriptions of event participants and activities were noted to calculate dose. Because there were no prior existing defined cut-offs for these parameters in past similar research, we set standards for reach at community events. Low, moderate, and high reach was set at zero to 49, 50 to 99, ≥100 participants, respectively. Low dose was defined as zero or one educational message per session and high dose was two or more messages.
To analyze the EIQ, categories representing “low,” “medium,” and “high” were created for each of the knowledge, self-efficacy, and behavioral intention scales. For knowledge, scores were based on the number of correct answers at baseline (range 0-14). For self-efficacy, responses were scored from one for very difficult to four for very easy (range 14-56). For behavioral intention, answers were scored from one for not willing to three for very willing (range 14-42). To represent the “low,” “medium,” and “high” categories for each of the separate scales, natural tertiles were based on the distribution of responses for each. A one-tailed paired t test compared pre- and post-training summary scores for employees who were evaluated at both time points.
Results
Fidelity, Reach, and Dose of Implementation
Stocking and Labeling of Promoted Foods
Stocking of promoted foods achieved high fidelity with an average of 88% of promoted items on store shelves (Table 1). Fidelity ranged from 66% in December to 93% in June. Staple items such as bread, milk, and cereal were more consistently stocked compared with condiments and spices. Labeling of promoted foods achieved moderate fidelity (Table 1). Throughout the intervention, labeling fidelity ranged from 49% to 81% with an overall mean of 71%.
Fidelity of Stocking, Labeling, and Advertising of Promoted Foods
In-Store Advertising
ERLW advertisements achieved high fidelity with an average of 75% signage found in the correct location throughout the store (Table 1). Fidelity ranged from 26% in December to 97% in August. Though implemented with high fidelity, advertisements exhibited greater variability from month to month compared to shelf labels. Every month except December reported a high level of fidelity of ERLW advertisements. Store renovations likely led to low fidelity in December as many advertisements were attached to refrigerator cases and shelf end caps. Both ERLW advertisements and the labels frequently competed with other special promotions.
Taste Test Sessions
On average, eight taste test sessions occurred monthly with an average attendance of 30 participants (Table 2), achieving moderate reach. An average of 9 label guides and 12 recipe cards were distributed to the 30 participants per taste test session, achieving low dose. Sessions had consistent moderate participation throughout the intervention. Taste test sessions were successful in demonstrating the labeling system via label guides, sharing recipes, and incorporating weekly promoted items from the store circular. These interactions often engaged customers in conversations about healthy eating behaviors and diet-related diseases.
Taste Test Session Reach and Dose Delivered
We compared the popularity of different sampled items by asking participants if they liked the item, disliked it, or were neutral. These responses were tracked on the taste test evaluation. Assessment and modification of taste test items and recipe cards occurred roughly every month. New recipe cards allowed us to feature seasonal ingredients and adapt recipes to holidays such as the Fourth of July and events like back-to-school. The RD ensured that recipes adhered to dietary guidelines. The frequency of taste tests varied by month based on events, the store’s needs, and ERLW interventionists’ schedules.
Community Outreach Events
Community outreach events averaged 106 participants per month, achieving high reach (Table 3). Outreach was broad, with events reaching elementary and high school students, young parents, families, church groups, seniors, and those in substance abuse recovery programs. Events included an introduction to ERLW, an interactive explanation of the shelf labeling system, and at least one additional educational message, meeting the standard set for high dose. Additional activities used educational messages to understand nutrition facts labels, shop seasonally as a way to improve food quality and decrease costs, use ERLW-promoted products in everyday meals, estimate sugar and fat content of typical unhealthy foods, and stimulate discussion of the change in serving sizes over time.
Community Outreach Events Reach and Dose Delivered
Events provided other opportunities to interact with store customers and the wider community. Surveys conducted midway through implementation showed low visibility of ERLW materials in the store; as a result, the interventionists re-engaged community partners who had participated in the formative research phase, along with others in the community. The objective of the outreach events was to introduce and provide information about the ERLW intervention and how customers could use the labeling system, recipes, and taste tests to make healthy affordable food choices.
Employee Training
Results of the employee training are presented in Table 4. Some respondents did not complete all sections of the EIQ; therefore sample sizes are different for each of the three scales. There were no significant increases in employees’ knowledge, self-efficacy, or behavioral intention for helping customers with healthy purchasing, food safety, nutrition, or customer service observed after employee trainings.
Employee Training Results
Implementation Barriers and Facilitators
Store Renovations
The unanticipated renovation of the store in November and December presented obstacles for intervention fidelity. While expansion of the produce section was an asset for ERLW’s healthy eating campaign—and had been prompted by formative research findings that people would buy more produce if this department were more visible and visually appealing—the renovations were disruptive, and limited the availability of store management and employees to meet with the research team regularly during this phase.
Store Owner, Management, and Employee Issues
The store owner and chief operating officer played pivotal roles in all phases of intervention design and implementation. Their support translated into significant financial and in-kind contributions during the intervention including the purchase of ERLW T-shirts and hats, payment of employees’ wages during training sessions, donation of foods used in taste tests, and space in the weekly circular to advertise discounts on ERLW-promoted healthy items. The store owner was committed to the ERLW intervention, but the same support was not received from the store manager given his dual loyalties in being asked to help implement the intervention and being responsible for maintaining a profit for the store. This presented day-to-day challenges in maintaining the promotional table in the front of the store, keeping the ERLW advertisements visible, and coordinating in-store events, which constrained reach. 1
High rates of turnover had implications for employee participation in the program. Store employees hired after April did not receive any formal orientation to ERLW and missed the multiday employee training at the beginning of the implementation period. These circumstances posed challenges for maintaining consistent support of ERLW from employees and for developing relationships between the interventionists and employees.
Four meetings were held with the employee advisory board between July and November consisting of 5 to 12 employees with diverse roles in the store (e.g., cashiers, deli worker, bagger, and administrative personnel). Meetings were intended to occur monthly, but their frequency and attendance waned due to scheduling conflicts. This limited our ability to obtain employee feedback on the intervention and respond to their concerns.
Discussion
To our knowledge, this is the first process evaluation of a comprehensive, multicomponent environmental intervention in a large supermarket in an underserved African American community that was also initiated and supported by the store owner. The intervention components varied in the success of their implementation.
There was high fidelity for stocking promoted items but only moderate fidelity for the labeling of these items. In-store advertisements of promoted foods also achieved high fidelity. Store renovations interfered with the implementation of these components. The months with the lowest number of labels were November and December, which coincided with the staggered renovations of the store’s sections (frozen food, dairy, cereal, etc.). It is possible that store traffic from shoppers and employees also increased due to holiday food purchases and more frequent stocking during these months. The constant re-stocking of items in a store with high customer volume meant that labels were frequently detached or misplaced. We observed no significant difference in fidelity between weeks in which SNAP benefits were distributed and weeks in which they were not. The release of SNAP benefits is state-dependent so these findings may be context-specific.
Despite the renovations, stocking and labeling of healthy foods were implemented with similar fidelity compared to other interventions (Gittelsohn et al., 2010; Novotny et al., 2011; Rosecrans et al., 2008). Intervention advertisements were implemented with high fidelity but with competition from other nonintervention marketing on shelves and display cases. This challenge in implementation has been noted with interventions in small and large stores across various types of advertisements (Curran et al., 2005; Foster et al., 2014; Gittelsohn et al., 2010; Rosecrans et al., 2008).
Taste tests had moderate reach and low dose but were effective in engaging customers. This experience contrasts other interventions where higher dose has been achieved in small and large stores (Gittelsohn et al., 2006; Gittelsohn et al., 2010; Novotny et al., 2011; Rosecrans et al., 2008). Community events were well attended with participants responding positively to messages, achieving both high reach and dose; however, a large event in December may have inflated the mean number of participants at community events per month. Few studies have implemented community events outside of the intervention stores and ERLW appears to be unique in its high reach and dose (Rosecrans et al., 2008).
Employee training showed no significant effects, though our sample size was small. There have been few interventions that incorporate employee trainings similar to ERLW. A study conducted in the Netherlands trained employees to give advice on healthy eating; however, customer surveys indicated no significant changes in advice when comparing the intervention and control groups (Van Assema et al., 2006). Another study in ethnic food stores in North Carolina showed effective implementation of employee trainings, however, did not measure changes in employees’ knowledge, or behavioral antecedents (Baquero et al., 2014).
The store renovation that occurred in the last 2 months of the intervention appeared to have a negative effect by limiting the store’s capacity to implement several components (stocking, labeling, and taste tests of promoted items). Other studies have reported similar external barriers to stocking items, such as problems with ordering systems and budget constraints that prevented ordering promoted foods (Curran et al., 2005; Mead et al., 2013; Rosecrans et al., 2008).
A distinguishing characteristic of ERLW was that it was initiated and supported by the store owner. Few other food store interventions have collaborated with store ownership and management like ERLW. In the Healthy Foods North intervention, stores helped decide which foods to promote, although motivation for the intervention came from local health staff and not the stores themselves (Mead et al., 2013). Foster et al. (2014) were able to collaborate with study supermarkets in Philadelphia such that the intervention was implemented by store staff. The store owner in our study was supportive in donating food and paying employees to assist with in-store taste tests. This contrasts with other store interventions where store owners were reimbursed for these activities (Curran et al., 2005; Gittelsohn et al., 2010).
As previously noted, the store management presented challenges that resulted in inconsistent implementation of intervention components. This contrasts with other interventions where researchers experienced high engagement from store managers (Gittelsohn et al., 2006; Glanz & Mullis, 1988; Foster et al., 2014). Our challenges with management emphasize the importance of integrating intervention goals with managerial responsibilities. Our experiences with the manager may indicate that there was a conflict between the economic constraints of the store’s need to remain profitable while also successfully implementing the intervention. The high rate of supermarket employee turnover also had implications for the success of implementation. This appears to be a recurring challenge with food store interventions, as employee attrition has been noted both in the stores and in outside groups promoting healthy eating (Curran et al., 2005; Mead et al., 2013). Inconsistent implementation could also be attributed to fluctuation in customer traffic due to the SNAP release schedule.
Research indicates that interventions integrated with communities or organizations are more likely to be sustained (Shediac-Rizkallah & Bone, 1998). Evaluations of other integrated food store interventions support this, as several have experienced continued delivery of program components (Gittelsohn et al., 2006; Mead et al., 2013; Novotny et al., 2011). However, interventions implemented by outside groups with short duration have been less likely to be sustained (Rosecrans et al., 2008).
Despite challenges from the manager and employee attrition, exceptional support from the store owner led to sustainability of ERLW components. In July 2013, the owner hired a part-time dietitian to develop strategies for promoting healthier choices among both customers and employees. The dietitian has continued the taste tests and community events. In addition, a corollary project aimed at children, ERLW for Kids! was implemented during the summer of 2013. The dietitian has also continued the children’s programming due to its popularity. Furthermore, the store has retained use of ERLW shelf labels. In response to the difficulties encountered in maintaining these labels during the intervention, store management is currently integrating ERLW label content into the store’s labeling system.
A study limitation was that the intervention was implemented in just one store, meaning that any unique factors experienced at this store cannot be compared to other settings. A control store was used, however, to be able to examine the intervention’s effects on the sales of promoted items. The taste test and community event evaluation forms went through several modifications during the implementation process, as we simultaneously adapted the intervention based on feedback from customers, the employee advisory group, and store management. However, this also led to challenges with data aggregation throughout the intervention period. Research team meetings were used to iteratively redefine several aspects of program (e.g., designing and refining the recipe cards, modifying the taste tests, and engaging with community partners). These meetings also served as a forum for refinement of evaluation instruments.
The large number of intervention items limited the extent to which they could be promoted. Fewer items could have been more heavily promoted through taste tests or other store activities. However, the large number of promoted items gave shoppers many choices for healthy alternatives and increased our presence throughout the store.
Future interventions may benefit from working with the supermarket employees more directly through longer term training and more frequent employee advisory meetings to promote the integration of intervention components into existing store processes and improve sustainability. If integrated stocking and labeling are not possible, we recommend that future programs consider promoting a number of items commensurate with the availability of intervention staff; it was our experience that 475 products were too many to label, track, and promote.
For ERLW, store owner support was not enough to guarantee full implementation. In the future, it will be important for interventionists to get buy-in and/or accountability from the manager before implementation. This could include establishing a more regular schedule for meetings that will help sustain employee support and improve the intervention delivery. It could also include continued program engagement such as in-service training with store employees working directly with customers (e.g., cashiers, deli workers), and training modules for all new hires. Alternatively, depending on the context, future interventions may also benefit from more relationship building up front leading to realistic expectations of employee participation, as was the case in Foster et al. (2014). Greater communication from both the store owner and manager regarding what is feasible and sustainable may be necessary given the often low profit margins and employee time constraints in the supermarket industry.
To avoid the challenge of simultaneously completing evaluation forms and interacting with participants during the taste tests, we recommended that a separate evaluator be hired to improve the quality of intervention delivery and evaluation. The bimonthly evaluation of stocking and labeling of foods and in-store displays took approximately 4 hours. Future interventions should be prepared for the time and costs associated with the evaluation.
Conclusions
The ERLW intervention was implemented with varied success in a low-income neighborhood in Baltimore with limited availability of healthy foods. Results suggest that it is feasible for multicomponent interventions to increase availability and promote healthy foods in a large supermarket environment. Intervention strategies should focus on commonly purchased items, and behavioral and environmental changes should be based on community input. Buy-in from store executives, managers, and employees should also be cultivated to sustain intervention components. Future research in large stores is needed to determine ways to improve consistent availability and promotion of healthy foods while competing with existing in-store messaging.
Footnotes
Acknowledgements
We are grateful to the supermarket customers, staff, and management for their participation. We are also appreciative of the store owner for his flexibility and support in the intervention.
