*Article* **Supermarkets in Cyberspace: A Conceptual Framework to Capture the Influence of Online Food Retail Environments on Consumer Behavior**

**Neha Khandpur 1,2,\*, Laura Y. Zatz 2, Sara N. Bleich 2, Lindsey Smith Taillie 3, Jennifer A. Orr 4, Eric B. Rimm <sup>2</sup> and Alyssa J. Moran <sup>5</sup>**


Received: 27 October 2020; Accepted: 16 November 2020; Published: 20 November 2020

**Abstract:** The rapid increase in online shopping and the extension of online food purchase and delivery services to federal nutrition program participants highlight the need for a conceptual framework capturing the influence of online food retail environments on consumer behaviors. This study aims to develop such a conceptual framework. To achieve this, mixed methods were used, including: (1) a literature review and development of an initial framework; (2) key informant interviews; (3) pilot testing and refinement of the draft framework; and (4) a group discussion with experts to establish content validity. The resulting framework captures both consumer- and retailer-level influences across the entire shopping journey, as well as the broader social, community, and policy context. It identifies important factors such as consumer demographic characteristics, preferences, past behaviors, and retailer policies and practices. The framework also emphasizes the dynamic nature of personalized marketing by retailers and customizable website content, and captures equity and transparency in retailer policies and practices. The framework draws from multiple disciplines, providing a foundation for understanding the impact of online food retail on dietary behaviors. It can be utilized to inform public health interventions, retailer practices, and governmental policies for creating healthy and equitable online food retail environments.

**Keywords:** online food retail; conceptual framework; consumer behavior; food choices; online shopping; retailer policies

#### **1. Introduction**

Online food retail is an increasingly popular means of acquiring food and is expected to grow rapidly over the coming decade. In 2017, online food retail represented \$13 billion in sales [1] and was projected to increase to \$100 billion [2], reaching 70% of U.S. shoppers by 2025 [3]. The 2019 coronavirus (COVID-19) pandemic has accelerated this growth and served as a catalyst for retailers to increase investments in their online food retail infrastructure and services. In April 2020, U.S. shoppers spent \$5.3 billion on online food purchases, an increase of 37% from the previous month [4,5]. Based on

this recent surge, revised growth projections estimate a nine-fold increase in online grocery purchases between 2017 and 2023 [6].

Online food retail has also emerged as an important avenue to improve food access. In 2018, the United States Department of Agriculture (USDA) selected 10 retailers in 9 states for a two-year Online Purchasing Pilot ("online Electronic Benefits Transfer (EBT)") to test the use of Supplemental Nutrition Assistance Program (SNAP) benefits as payment for online food purchases [7]. The affordability, access, and delivery inequities, as well as the disproportionate food security challenges faced by low-income communities have been brought to the fore during the pandemic [8], prompting the USDA to expand online EBT both geographically and across a wider range of retailers. At the time of writing, at least 40 states had been approved to participate in online EBT, with several others in the planning stage [9]. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) is also considering how to offer online food purchasing options to its participants [10].

These shifts in consumer food purchase behaviors, the increasing investment in online infrastructure, and the expansion of online food purchasing to participants in federal programs, highlight the need for an assessment of the online food retail environment. A substantial and growing body of evidence captures the influence of in-store food marketing on consumer purchases [11,12]. In comparison, little is known about the relationship between the online food retail environment and consumer food choices. Some evidence supports the influence of social, contextual, and demographic factors on consumers' use of online platforms for food purchases [13–15]. Retailer-level factors like credibility, product freshness, product quality and price have also been shown to predict online purchasing [16]. However, there is a dearth of information about factors influencing the spectrum of online consumer behaviors which is important for understanding how online food retailers influence consumer food purchases and subsequent dietary intake and health outcomes.

What is currently lacking is an integrated framework capturing both consumer- and retailer-level factors and their interaction that influence consumer behaviors within online environments. The few existing frameworks focus on specific consumer determinants like their attitudes, privacy concerns, social influences, facilitating conditions, hedonic motivations, and perceived risk or satisfaction with the online experience [17–19]. Retailer influence is incompletely captured or mentioned briefly [20]. The models do not illustrate the interactive and dynamic nature of online food retail platforms or the active and responsive role that consumers play in shaping their food purchase experience. Existing frameworks are also explicitly geared towards retaining and expanding the consumer base or maximizing profits [21]; they are not designed to study the effects these platforms have on dietary behaviors or health. Perhaps most revealing is the concentration of this literature in the fields of marketing, retail, decision sciences, and informatics; studies are largely absent in the public health domain.

In the absence of a comprehensive conceptual framework that looks at consumer grocery purchase behaviors it becomes impossible to systematically study the effect of food retail environments on food choices. Such a framework is crucial for informing public health interventions, guidelines, retailer practices, and governmental policies to create healthy and equitable online food retail environments. To address this gap in the literature, the present study aims to develop and refine a conceptual framework capturing factors that influence consumer food purchase behaviors within online food retail environments. For the purposes of this study, online food retail environments were described as websites providing click-and-collect (i.e., order online and pick up at the store) or food delivery services. 'Retailers' include e-commerce platforms hosted by the retailer themselves or by a third-party vendor (e.g., Instacart).

#### **2. Materials and Methods**

The development and refinement of the conceptual framework was guided by the approach suggested by Jabareen, 2009 [22], and involved extensive reading and categorizing of data; identifying concepts; deconstructing, categorizing, synthesizing, and integrating concepts; and validating the final framework. To achieve this, mixed methods were used, including:


The study methodology was reviewed and determined to be non-human subjects research by the Institutional Review Board at the Johns Hopkins Bloomberg School of Public Health.

#### *2.1. Literature Review*

A scoping review was undertaken between April and June of 2019 to identify peer-reviewed and grey literature, in English, on the attributes, preferences, and shopping behaviors of consumers that make purchases online, consumer interaction with and acceptance of technology, online and in-store food marketing and merchandising, and design of online retail environments. Databases were searched from January 2009 to May 2019 (ProQuest, PubMed, and Thomson ONE) using combinations of the keywords *food, grocer\*, supermarket\*, retail\*, shop\*, store\*, purchas\*, buy\*, online, ecommerce (or e-commerce), internet,* and *web*. Search results were supplemented with health agency reports, trade publications, and industry reports from 2015 to 2019. Reference lists from peer-reviewed publications were also scanned. While peer-reviewed literature was not restricted by geographic location, only US-focused analyst reports and trade publications were included to ensure contextual relevance. Paper titles and abstracts were screened. A total of 136 industry reports and 97 peer-reviewed papers informed the development of the draft framework.

#### *2.2. Key Informant Interviews*

We interviewed seven experts in grocery merchandizing and marketing, e-commerce and online retail, behavioral psychology, public policy, computer science and data privacy and digital marketing. Experts were identified through a combination of known contacts, through their published work, or were referred to by other experts during interviews. The interviews were conducted in person or via teleconference by a member of the research team and lasted 45–60 min.

The interview began with an overview of the objectives of the study. Experts were presented with the draft framework and asked: (i) for their feedback on the extent to which it captured their understanding of factors influencing consumer food choices when shopping online; (ii) to identify constructs that could be improved or simplified; and (iii) for additional constructs that should be included. Follow-up questions were tailored to each key informant's area of expertise. For example, an expert in computer science and engineering was asked an additional question about when and how personal data are collected from consumers along the path to purchase. Experts in food retail marketing were asked how personal data are used to adapt online marketing practices to specific consumers. Insights were requested on how the key domains influenced one another. Suggestions were incorporated, and the framework was refined after each interview.

#### *2.3. Pilot Testing*

Study researchers (A.J.M., N.K.) independently tested the internal consistency of the conceptual framework by using it to guide a mock shopping exercise. This was done to identify additional concepts that may have been missed during the literature review and the key informant interview, but not with the aim of formally testing the framework. An online shopping account was created at two U.S. online food retailers. Researchers navigated each store's website, browsed through their departments, added three grocery items to the shopping cart, and proceeded to the checkout. The applicability of the conceptual framework to the experience of a consumer searching for and selecting food was discussed in detail. Revisions to the framework were incorporated as necessary.

#### *2.4. Expert Discussion*

A teleconference discussion was convened with six members of the Healthy Food Retail Working Group to determine the content validity of the conceptual framework. This group is a collaborative effort of the Robert Wood Johnson Foundation's Healthy Eating Research program and the Centers for Disease Control and Prevention's Nutrition and Obesity Policy Research and Evaluation Network. Members include researchers and technical experts working on healthy food retail and related areas.

To engage fully with the framework, members of the Working Group were asked to undertake a mock shopping exercise, similar to the one conducted by the study authors, two weeks prior to the call. A Qualtrics form was created to guide the sequence of product searching and selection activities and to record feedback. After selecting an online food retailer from a list of 21 options, members navigated to the grocery department homepage, the breakfast cereal department page, and the product pages of two specific brands of bread and canned fish. They documented marketing strategies, customizable features, tools, options, and multimedia content on each of these webpages. Members also recorded ease of navigation and site policies. Their feedback from this exercise was discussed in the teleconference, during which the purpose of the research was clarified, and a draft of the conceptual framework (including what the concepts represent and how they relate to each other) was presented. Members were asked whether the current framework captured their understanding of the range of factors influencing consumer behavior in the online store and for ideas for improvement. Notes from the discussion were recorded, and relevant revisions were incorporated into the framework.

#### **3. Results**

#### *3.1. Evolution of the Conceptual Framework*

Insights from the literature review informed the development of a draft framework that captured the influence of consumer characteristics, online food marketing, and retailer and website characteristics on online grocery shopping intention. Feedback from the key informant interviews added further detail by delineating the different stages of the online shopping process, the sequence of online consumer behaviors, the key role of personalized marketing, relationships between online retailers and manufacturers, disclosure of sponsored content, factors influencing site design, and the use of personal information in customizing the platform.

The pilot test provided additional insights into retailer characteristics related to membership and loyalty programs, privacy and data use, and order payment, fulfillment, and collection. This exercise also differentiated the static versus the dynamic elements of the online platform (i.e., attributes of the site that are consistent for all consumers versus those that can be personalized by the retailer or customized by the consumer), differences in site functionality as viewed by an anonymous shopper versus a registered shopper, and marketing strategies employed at checkout.

Healthy Food Retail Working Group members acknowledged the detail and clarity of the current version of the framework and agreed that it captured relevant constructs that influence consumer food choice in online environments. Their feedback focused on the applicability of the conceptual framework to specific subgroups of consumers (e.g., SNAP and WIC beneficiaries), retailers (e.g., small stores), and online formats (e.g., mobile applications). These ideas were discussed, and relevant constructs in the framework were added or made more salient.

The final framework combined elements of the Technology Acceptance Model, consumer behavior and decision-making frameworks, brick-and-mortar marketing categories (product, price, placement, and promotion) [11,23], key themes from the data privacy and e-commerce literature, and constructs described in existing food environment measures (e.g., Nutrition Environment Measures in Stores) [18,20,24–30].

#### *3.2. Description of the Conceptual Framework*

#### 3.2.1. Path to Purchase

Central to the conceptual framework is the sequence of consumer behaviors involved in online grocery shopping. This Path to Purchase consists of four stages—Pre-Shop, Online Shopping, Pick-Up or Delivery, and Post-Shop—encompassing six behaviors. Under Pre-Shop behaviors, the consumer selects an online retailer. He/she then searches for or discovers, selects, and purchases food (Online Shopping). The consumer receives the order (Pick-Up or Delivery) and prepares and consumes the food or/and discards it (Post-Shop). The quality of the consumer's experience engaging in each of these behaviors determines their overall satisfaction with the retailer and likelihood of shopping again.

Consumer behaviors are influenced by consumer- and retailer-level attributes, presented above and below the Path to Purchase in Figure 1. Attributes presented in solid outline are not likely to change over the shopping journey (static domains). Attributes presented in dashed outline are likely to change over the duration of a single shopping visit (dynamic domains). The cross-cutting themes of Equity and Transparency influence retailer-level factors, while Social, Community, and Policy Context influences both retailer- and consumer-level factors.

**Figure 1.** Consumer behavior in the online food retail environment.

#### 3.2.2. Consumer-Level Attributes

Consumer Characteristics, Preferences, and Past Behaviors: this domain encompasses consumers' Technology Acceptance, Individual and Household Demographics, Food-Related Preferences and Behaviors, and Attitudes and Beliefs (detailed in Table 1). These attributes may influence decisionmaking at each stage of food selection and purchase. For instance, consumers with positive experiences with e-commerce and online food retail may be more likely to select an online retailer. Individual and demographic characteristics such as age, education, income, household composition (e.g., having young children) and location also influence the likelihood of shopping online, the platform selected, the foods purchased, and their preparation and consumption.

#### 3.2.3. Retailer-Level Attributes

Retailer Policies and Practices: they include static attributes of online retailers that may influence retailer selection, food receipt, and food preparation and consumption. During retailer selection, Site Access policies include information required to purchase groceries online, retailer incentives offered to first-time users (e.g., free trials), membership requirements prior to shopping, and delivery service areas (e.g., zip codes served). Privacy and Data Sharing policies include the terms and conditions that govern the collection, tracking, storage, use, and sharing of personal information by the retailer. Policies on Inventory Management guide the availability of products and brands, pricing strategy, and accuracy of inventory tracking, while policies on Collection and Payment determine the payment options accepted by the retailer (e.g., SNAP), the integration of loyalty and reward programs, fees for delivery and restocking, and delivery options (e.g., availability of click and collect or delivery).

**Table 1.** Description of consumer- and retailer-level domains and constructs that influence consumer behavior in the online food retail environment. EBT, Electronic Benefits Transfer, GMO, genetically modified organism.



**Table 1.** *Cont.*


**Table 1.** *Cont.*


**Table 1.** *Cont.*


**Table 1.** *Cont.*

The Pick-Up and Post-Shop behaviors are likely to be determined by Retailer Policies and Practices that govern Order Fulfillment, Order Delivery, and Returns and Order Cancellation. Policies on Order Fulfillment include the price and appropriateness of product substitutions during stock-outs and the quality of the food received. The availability of convenient time slots, delivery coordination (e.g., text message communication, online order tracking), and the availability of secure delivery options are examples of policies associated with Order Delivery. Policies determining the ease of cancelling incomplete or incorrect orders of items purchased online (e.g., store credit, vouchers, full refund) relate to Returns and Order Cancellation.

Personalized Marketing by Retailers: a consumer's search, selection and purchase of food are influenced by factors within the dynamic domain of Personalized Marketing. These marketing strategies are based on personal information provided by the consumer when registering with the online platform and from past purchases and browsing history. They may also be based on consumer data purchased by the retailer, shared by third parties, or automatically collected upon visiting the site (e.g., IP address, operating system). The consumer, either knowingly (by actively agreeing to) or unknowingly (by using the site, by signing-up for an account or a loyalty card that activates implicit agreements) or without full knowledge of the implications, allows retailers to use various sources of information to create a tailored experience.

Personalized Marketing maps onto the marketing mix of Product, Price, Placement, and Promotion ("the 4Ps"), but manifests differently in the online food retail environment than in brick-and-mortar stores [11]. Product-related personalized marketing includes Product Mix or the range and variety of products the consumer can view and explore. In the online space, this may include personalized storefronts created for the consumer which display products aligning with the consumer's revealed preferences (e.g., vegan, gluten-free). Product Price relates to different types of Discounts, Rewards, or Time-Limited Special Deals offered. These strategies can be personalized (e.g., the discounted products or discount amount is tailored) and often interact with Placement or Promotional strategies (e.g., recommended products may also be discounted). Under Placement, Cross-Promotion describes suggested complementary products, Search Result Ordering describes the default appearance of products (e.g., higher cost or sponsored products may appear before other options), and Recommendations captures the strategies used to increase exposure to featured, seasonal, or popular products on the site—all of which can be personalized to the consumer's preferences. Promotion refers to personalized marketing to increase consumer exposure to or visibility of sponsored products through Advertisements (e.g., title cards or banner advertisements), Branded Site Content, User Feedback (e.g., product ratings and reviews), links to Social Media, and other Point-of-Purchase Information (e.g., product labels). These strategies are frequently used in combination to influence product discovery, selection and purchase on the website homepage, search results page, product page, or at checkout.

Customization of the Website by the Consumer: this is the other dynamic domain that influences the search, selection, and purchase of food and includes Product Information Display, Site Navigation, and Shopping Tools. These attributes allow consumers to change what nutrition information they see (Product Information Display), filter products based on preferred attributes, save shopping lists, or request certain product comparisons (Shopping Tools). Combined with tools and tutorials to ease website navigation (Site Navigation), website customization features can increase the convenience of product searches, enhance consumer engagement with the product catalogue and improve the quality of the food purchase experience.

Other Food Marketing: this dynamic domain includes Promotional Strategies, Social Media Strategies, and Immersive Strategies, recognizing that exposure to marketing in other settings (brick-andmortar stores), through direct-to-consumer promotions, product endorsements, and sponsorships, and targeted marketing through social media platforms will affect food choices made online.

Sophistication of Website and Frequency of Use: technological progress in interface design, communications, and data security is likely to improve consumer trust and increase the volume and frequency of purchases made online. Advancements in personal data collection, advertising technology, purchase data analytics, and consumers' increased involvement in the co-creation of food retail platforms will allow retailers to better profile customers and more efficiently match them to products and promotions, increasing engagement. In this way, more sophisticated online platforms and frequent consumer visits will ensure greater personalization of retailer marketing strategies and a more customized website.

#### 3.2.4. Cross-Cutting Domains

Equity and Transparency are fundamental to retailer engagement with the consumer. Equity refers to the differential impact of retailer policies and practices on the food behaviors and privacy of underserved populations (e.g., individuals and communities of Color, low-income households, older adults, people with disabilities, households in rural areas). For example, a consumer's ability to utilize online services may be affected by the retail service area, availability of convenient delivery slots, or accepted payment methods. A retailer's targeted and personalized marketing strategies may trigger impulse purchases or increase the basket size, differentially impacting low-income consumers, especially if the promoted products are of inferior nutritional quality. Transparency in policies and practices captures the retailer's clear and upfront disclosure of data collection, storage and use of data, surveillance methods, marketing, sponsorships, etc., that may consciously or unconsciously influence the consumer's choices along the Path to Purchase. For instance, disclosure of fees and hidden costs and collection of personal data prior to checkout may affect a consumer's choice of retailer. Disclosure of product sponsorship at point of purchase may affect product selection.

Social, Community, and Policy Context: this conceptual framework is nested in the socio-ecological model [31]. Consumer and Retailer-Level attributes are likely influenced by the social context (e.g., social norms, endorsements from trusted members of society) and the structural factors (community, institutional, and policy contexts) in which they are embedded. The surge in online grocery purchases resulting from the physical distancing measures implemented in 2020 is an example of how contextual factors can affect consumers behaviors. Similarly, the acceptance of SNAP benefits for online food purchases would first require a favorable state-level policy context (e.g., states need approval for use of EBT test cards) [7], before retailer policies can be implemented. In this way, the provision of certain services by retailers is very much dependent on the broader political context.

#### **4. Discussion**

This study presents a conceptual framework that captures factors influencing consumer behavior within online food retail environments. It also details the methodology for framework development and refinement—a process that identified and integrated evidence across multiple fields of study. The conceptual framework captures both consumer- and retailer-level influences across the entire Path to Purchase as well as the broader social, community, and policy context. Important static attributes of retailer policies and practices and consumer characteristics, preferences, and past behaviors are captured. The framework also emphasizes the dynamic attributes of the online platform, including those of personalized marketing by retailers, customization of the website content, navigation by the consumer, and the two-way interaction between these domains that enables a variety of online food retail interfaces uniquely tailored to consumer preferences.

This conceptual framework makes an important contribution to our understanding of the burgeoning field of online food retail. It serves as a foundation for a deeper study into the influences on consumer food purchase behaviors within online platforms and the interactions between them. To our knowledge, this is the first time that the relatively under-studied domains of personalization and customization of the online food retail environment or the domains of equity and transparency or the social, community, and the policy context, have been considered in any framework. Further investigation is certainly warranted. Future studies could use the framework to compare brick-and-mortar retailers and their online platforms to identify the convergence and divergence of consumer behaviors and retailer responses within these settings. The conceptual framework itself could be empirically tested to support its validity and better establish a hierarchy between attributes. Previous work has used structural equation modelling techniques to examine the hypothesized relationships between constructs in a proposed model and identify possible causal relationships between them [32]. While online retail platforms seem largely similar across contexts, a multi-country investigation of the framework's domains and their interactions would help confirm its applicability to different settings. Testing the framework among a diverse group of consumers to ensure that it adequately captures their selection and purchase experiences is also warranted.

The conceptual framework provides a foundation for understanding how a lack of transparency within online retail platforms could impact health equity. Indeed, if an equity lens is not applied in the development and implementation of retailer policies, online platforms may unintentionally widen disparities in healthy food access, affordability, and diet quality for vulnerable groups [33]. For instance, lack of transparency around membership fees, costs associated with platform access or delivery, accepted forms of payment, or inappropriate product substitutions for SNAP beneficiaries, may be more detrimental to a household with scarce resources to spend on food than to a household with greater financial resources. Concerns have also been raised about the surveillance practices, data collection, use, storage, sharing, and privacy measures of online retailers [34]. Current practices leverage digital tools to capture sensitive data on consumer purchases, location, and preferences or require people to share personal information to avail of savings [34]. This may facilitate targeted marketing on the basis of race, ethnicity, or socioeconomic status. Although targeted marketing is not harmful in and of itself, it may exacerbate existing disparities in diet-related chronic diseases if used to disproportionately advertise unhealthy products.

The framework may help to study the effect of predatory marketing tactics, similar to those employed across other digital media where communities of color are targeted with the least nutritious products [35]. Indeed, several forms of discriminatory and disparate advertising to vulnerable groups have been identified [34], exposure to which is likely to encourage unhealthy purchases. More research to identify and address these equity gaps is crucial to safeguarding the sub-groups that are disproportionately affected by adverse health outcomes and those that stand to benefit most from policy, systems, and environmental interventions that promote healthy eating.

Finally, the conceptual framework could be used to inform and evaluate public health interventions aimed at improving consumer food choices in the online food environment. On the policy and practice front, the framework could inform: (i) recommendations and standards for best practices related to online food marketing; (ii) specific guidance for online retailers to ensure policy transparency, equitable access, and assurance of privacy; (iii) tailored educational content for consumers unfamiliar with online grocery retail; and (iv) personalized nutrition education and communication. For example, nutrition interventions delivered via the online retail platform could offer personalized healthy shopping lists that draw on information about consumer preferences and budget constraints, offer personalized healthy recipes or meal solutions, or develop a personalized labeling campaign that makes specific nutritional attributes of a product more salient at the point of purchase. SNAP-Ed—SNAP's voluntary nutrition education program—could partner with online retailers to allow participants to interact with a registered dietitian in real time while food shopping. Local WIC agencies could work with online retailers to create a WIC-friendly web interface with WIC-eligible products, label products as being part of WIC food packages, and allow only WIC-approved substitutions in case of stock-outs.

This study does have its limitations. Despite a comprehensive approach to development, the resulting conceptual framework may not have captured all the elements of the online food retail environment. The descriptions of the constructs in the framework serve as examples and are not exhaustive. The literature reviewed was almost entirely from the US and Europe. Pilot testing and content validity were established for the US context. Therefore, it is possible that the framework may not adequately capture the online environments in other contexts and would need further testing to gauge applicability in different settings, including its applicability to food purchases made via mobile applications. Advances in technology will result in new and innovative features that will need to be incorporated into the evolving conceptual framework.

This study has several strengths. It leveraged multiple methods in the development and refinement of the framework, including key informant interviews, comprehensive systematic literature reviews, mock shopping exercises and group discussions, improving its validity. The framework development drew from multiple disciplines and benefited greatly from the insights of experts across different fields, allowing for an in-depth understanding of the factors influencing consumer purchases and underscoring the need for the public health community to collaborate with scientists and policymakers from non-traditional public health disciplines to map the influence of the online food environment. Finally, applying a public health perspective to the development of the framework expanded its utility in informing future interventions in this field.

#### **5. Conclusions**

This paper integrated multiple perspectives across a wide range of fields to develop a framework capturing both consumer- and retailer-level factors influencing consumer purchases in the online food retail environment, as well as the broader social, community, and policy context. It identifies important static factors and emphasizes the dynamic nature of personalized marketing by retailers and customizable website content. Equity and transparency in retailer policies and practices are also captured. Researchers, retailers, advocates and policymakers are encouraged to utilize the framework to guide the development and evaluation of interventions, policies, and practices in the online food retail space.

**Author Contributions:** Conceptualization—N.K., A.J.M.; methodological design—N.K., A.J.M., L.Y.Z.; data collection— L.Y.Z., J.A.O., N.K., A.J.M.; analysis—L.Y.Z., J.A.O., N.K., A.J.M.; interpretation—N.K., L.Y.Z., S.N.B., L.S.T., E.B.R., A.J.M.; original draft preparation—N.K., A.J.M.; review and editing—N.K., L.Y.Z., S.N.B., L.S.T., E.B.R., A.J.M.; project administration—A.J.M.; funding acquisition—A.J.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by Healthy Eating Research, a national program of the Robert Wood Johnson Foundation.

**Acknowledgments:** The authors would like to acknowledge the contributions of members of the Healthy Food Retail Working Group (Lucia Leone, Diana Grigsby-Toussaint, Pasquale Rummo, Jared McGuirt, Alice Ammerman, Betsy Anderson Steeves, Stephanie Jilcott Pitts) in the development of the conceptual framework.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*

## *Article* **Developing a National Research Agenda to Support Healthy Food Retail**

**Amelie A. Hecht 1,\*, Megan M. Lott 2, Kirsten Arm 2, Mary T. Story 2, Emily Snyder 3,**†**, Margo G. Wootan 3,**‡ **and Alyssa J. Moran <sup>1</sup>**


‡ Affilicated to the Center for Science in the Public Interest at the time of the study.

Received: 28 August 2020; Accepted: 2 November 2020; Published: 4 November 2020

**Abstract:** The food retail environment is an important driver of dietary choices. This article presents a national agenda for research in food retail, with the goal of identifying policies and corporate practices that effectively promote healthy food and beverage purchases and decrease unhealthy purchases. The research agenda was developed through a multi-step process that included (1) convening a scientific advisory committee; (2) commissioned research; (3) in-person expert convening; (4) thematic analysis of meeting notes and refining research questions; (5) follow-up survey of convening participants; and (6) refining the final research agenda. Public health researchers, advocates, food and beverage retailers, and funders participated in the agenda setting process. A total of 37 research questions grouped into ten priority areas emerged. Five priority areas focus on understanding the current food retail environment and consumer behavior and five focus on assessing implementation and effectiveness of interventions and policies to attain healthier retail. Priority topics include how frequency, duration, and impact of retailer promotion practices differ by community characteristics and how to leverage federal nutrition assistance programs to support healthy eating. To improve feasibility, researchers should explore partnerships with retailers and advocacy groups, identify novel data sources, and use a variety of study designs. This agenda can serve as a guide for researchers, food retailers, funders, government agencies, and advocacy organizations.

**Keywords:** food and beverage; grocery retail; supermarket; marketing; policy; research agenda; healthy food retail; food environment

#### **1. Introduction**

The food retail environment is an important driver of dietary choices in the U.S. Components of the food retail environment, including access to food retail, availability and price of healthy products in stores, and presence of in-store marketing, all play a role in shaping dietary patterns [1,2]. Food and beverage manufacturers spend billions of dollars annually to ensure retailers stock, prominently place, and promote their products [3]. Unhealthy products are promoted more often than healthy products, and evidence suggests that promotion of unhealthy products shapes consumer purchasing more than promotion of healthy products [4,5].

Current dietary patterns, which, compared to the *2015–2020 Dietary Guidelines for Americans,* are low in fruits and vegetables, whole grains, and lean protein, and high in sodium, added sugars, and saturated fat, put many Americans at elevated risk of chronic health conditions, including obesity and diabetes [6,7]. Low-income and racial/ethnic communities, who experience greater prevalence of diet-related chronic diseases, may also be more likely to be targeted by marketing of unhealthy foods and beverages [8–10]. For example, in-store marketing of unhealthy beverages has been shown to increase at the time of month when Supplemental Nutrition Assistance Program (SNAP) benefits are distributed, particularly in neighborhoods with high SNAP participation [10].

As national attention toward health disparities and diet-related chronic diseases has increased in recent years, researchers, advocates, and policymakers have recognized the need to improve the food retail environment. In 2010, the Robert Wood Johnson Foundation (RWJF) and The Food Trust convened researchers, public health advocates, food retailers, manufacturers, and marketing professionals to discuss strategies to promote healthy retail, with a particular focus on children in low-income communities. The report that followed, *Harnessing the Power of Supermarkets to Help Reverse Childhood Obesity*, recommended marketing tactics to promote healthier purchases that jointly benefited consumers, retailers, and manufacturers [11]. In the intervening years, progress has been made toward identifying retail practices that undermine healthy eating and designing interventions that promote healthy eating in the retail food environment. At the same time, the retail food landscape has evolved: grocery store chains have consolidated, dollar stores have gained market share, and some consumers have shifted their purchases online. Research to fill remaining and emerging gaps in the food retail literature is needed.

This article outlines a national research agenda to support healthy food retail developed by Healthy Eating Research (HER; a national program of RWJF), the Center for Science in the Public Interest (CSPI), The Food Trust, and other researchers. This is the first national research agenda focused on healthy food retail. Research agendas have been developed to guide work on a variety of other public health topics [12,13]. Agenda-setting helps to identify important gaps in knowledge and to build consensus and support to fill those gaps among funders, advocates, and researchers. This agenda describes key areas for research to better understand current food retail practices and consumer behaviors and potential retail strategies to promote healthy eating while addressing racial and income disparities in diet quality and related disease. Research in these domains can inform policy strategies and corporate practices to improve the food retail environment and promote health equity. This article describes the collaborative and iterative methods used to develop the research agenda and the results generated at each step of the process. It then presents a final set of research questions in a comprehensive research agenda, key considerations for how to conduct that research, and ways in which the research agenda can be used to advance the field and public health.

#### **2. Methods**

The research agenda was developed through an iterative process between October 2019 and July 2020 that included the following steps: (1) convening a scientific advisory committee; (2) commissioning five systematic literature reviews and one original research project on food retail practices and interventions; (3) in-person convening of expert stakeholders; (4) thematic analysis of meeting notes and refining research questions; (5) follow-up survey of convening participants; and (6) developing the final research agenda (Figure 1). The scientific advisory committee provided input at each stage of the process. This agenda-setting process was based on methods used by Duffy et al. [12].

**Figure 1.** Flow chart depicting the process of developing the national healthy retail research agenda.

The research agenda was developed with an emphasis on health equity and the demographic groups that are at highest risk for poor health, especially nutrition and weight-related health disparities. These priority populations, identified by HER, include Non-Hispanic Black, Hispanic/Latinx, Asian American, Native Hawaiian/Pacific Islander, American Indian/Alaska Native, and rural children and their families [14].

#### *2.1. Convening a Scientific Advisory Committee*

A scientific advisory committee was formed and included seven researchers from government, academia, and nonprofit organizations, representing a variety of substantive areas related to psychology, nutrition, health behavior, anthropology, and public policy. The committee was selected based on prior work in the field, leadership in related working groups and professional organizations, and experience working with HER's priority populations. The committee provided input on topics for commissioned research, the in-person convening agenda and guest list, and content of the follow-up survey and final research agenda. Committee members also took notes and guided small group discussions at the in-person convening.

#### *2.2. Commissioned Research*

Five literature reviews and one original research project were commissioned for the in-person convening and were conducted by experts in the field. (Five of these papers are published jointly with this special issue.) These works aimed to provide an overview of previous research on key topics and guide convening discussion. Commissioned papers were organized into three themes: (1) retailer and manufacturer marketing practices, (2) consumer food purchasing trends by race/ethnicity, socioeconomic status, and geographic location, and (3) effectiveness of government- and researcher-led retail interventions to increase healthy food access and purchases. The original research paper used Nielsen Homescan Consumer Panel data from 2008–2018 to assess how packaged food purchases differ by store type and consumer demographics (urban vs. rural, high vs. low income).

#### *2.3. In-Person Convening of Expert Stakeholders*

The goals of the in-person convening were to (1) summarize previous research on healthy food retail, (2) identify gaps in the literature, (3) generate and prioritize questions for future research, (4) highlight best practices for research collaboration with the food industry, and (5) facilitate relationships between retailers and researchers to implement and evaluate healthy retail interventions. The full-day event was held in Washington, DC on 29 January 2020 and was organized by staff from HER, CSPI, and The Food Trust and the scientific advisory committee. Forty-six expert stakeholders from academia, government, advocacy, and the food industry participated.

In advance of the meeting, participants were asked to read six brief reports with the preliminary findings from the five commissioned systematic reviews and one original research project. At the convening, academic researchers presented key findings from each of the commissioned projects. Presentations were grouped according to the three themes discussed in Section 2.2 (two presentations per theme). After each pair of presentations, scientific advisory committee members facilitated small group breakout discussions. In breakout groups, participants discussed findings from the presentations and research gaps related to the theme, including understudied populations. Participants were asked to brainstorm new research methods, data sources, and study designs to facilitate future evaluation.

Meeting organizers also facilitated a large group discussion during which participants were asked where they would recommend directing intervention research over the next ten years to have the greatest impact on population health and equity. Subsequently, a panel of industry representatives discussed best practices for researchers seeking to partner with retailers and food manufacturers on healthy retail research. Finally, in small groups, participants were asked prioritize research questions identified throughout the day that would help fill knowledge gaps.

After each small and large group discussion, participants were asked to write research questions that emerged on sticky notes. Sticky notes were placed on walls throughout the meeting room according to the theme. At the end of the convening, participants were asked to walk around the room and place dots next to the research questions they thought were most important for advancing health equity.

#### *2.4. Thematic Analysis of Notes and Refinement of Research Questions*

Notes taken by the scientific advisory committee at the convening and sticky notes generated by convening participants were thematically analyzed and grouped by three authors collaboratively (A.A.H., M.G.W., A.J.M.). The list of research questions was collated and refined by deleting duplicate questions, questions that were too vague or specific, and questions outside the scope of the research agenda. Cross-cutting considerations related to study design, setting, data sources, and partnerships raised during group discussions were also refined.

#### *2.5. Follow-Up Participant Survey*

An online follow-up survey was sent via email to convening participants in May 2020. The survey was developed by the authors with feedback from the scientific advisory committee. The survey was first entered into Qualtrics and tested for functionality and length. Respondents were asked to complete the survey within two weeks, during which time two reminder emails were sent.

A total of 40 research questions generated at the in-person convening were included in the follow-up survey. Survey respondents were asked to rank each research question on a scale from 1 (lowest) to 5 (highest) in terms of feasibility, equity, and importance (defined in Table 1). For each research question, average scores for each domain and composite scores were calculated using Microsoft Excel. Research questions that received low composite scores (<3) or low scores in all three domains (<3.5) were removed. This allowed research questions that received low scores in one domain but high scores in one or both of the remaining domains to be preserved (for example, a question that received a score of 2.0 for feasibility but a score of 3.7 for importance and 3.5 for health

equity would be preserved). Respondents were asked to list any missing research questions. Finally, respondents were provided a list of data sources for healthy retail research identified at the convening and provided an opportunity to list additional data sources.

**Table 1.** Definitions of domains used to rank healthy retail research questions in a follow-up survey sent to experts who previously participated in an in-person healthy retail research convening (n = 46).


#### *2.6. Developing a Final Research Agenda*

The final research agenda was developed based on findings from steps 2–5 (see Figure 1) and with critical input from the scientific advisory committee and select members of the Healthy Food Retail Working Group, which is supported by HER and the Centers for Disease Control and Prevention's (CDC) Nutrition and Obesity Policy Research and Evaluation Network (NOPREN). The final research questions (selected based on follow-up survey results) and the cross-cutting considerations for research were grouped into key themes.

#### **3. Results**

#### *3.1. Commissioned Research Findings*

Key findings from the commissioned research papers, including research gaps, are discussed briefly here; five of the commissioned papers are also published in this special issue.

Two commissioned systematic reviews focused on retailer and manufacturer marketing practices. The first identified four key strategies that food and beverage manufacturers use to influence retailer marketing practices, but called for further research to understand the role that financial incentives from manufacturers play in shaping the retail environment, including analyses using proprietary data from retailers and manufacturers [15]. The review also found evidence that retailer marketing strategies, including price discounts and prominent store placement, are associated with increased product sales, but concluded that other in-store promotional strategies, such as signs and displays, are understudied. A second commissioned paper assessed marketing-mix and choice-architecture (MMCA) strategies used to promote and sell sugar-sweetened beverages (SSBs) in U.S. food stores and found that SSBs were widely available and price reductions and promotions were used often to boost sales. The authors found that targeted MMCA strategies may be used to influence SSB purchases among at-risk consumers on the basis of income or race/ethnicity, for example, and that MMCA strategies may vary by retail format. They noted that most studies were not designed to capture such differences, representing a need for future investigation to inform practice and policy approaches to mitigate health disparities [16].

Two additional commissioned papers focused on differences in consumer shopping patterns by race/ethnicity, socioeconomic status, and geographic location (urban vs. rural). In one systematic review, the authors called for more research that examines how these three factors intersect to influence U.S. consumer food purchasing [17]. In particular, they found a small proportion of included studies examined purchasing at the intersection of two factors (race/ethnicity and socioeconomic status), and no studies examined purchasing at the intersection of all three factors or assessed geographic differences in purchasing. The other paper, an original research project using household packaged food purchase panel data from 2008–2018, identified heterogeneity in the type and nutritional quality of packaged foods and beverages purchased by urban versus rural households and low- versus

high-income households in different retail formats [18]. The authors called for research to examine why these differences exist—for example, why rural households tend to buy more packaged foods from mass merchandisers and dollar stores, which offer foods of poorer nutritional quality.

The final two commissioned systematic reviews examined the impact of retail interventions on consumers and retailers. One review synthesized 148 evaluations of governmental policies designed to increase healthy food purchases in supermarkets and found that sweetened beverage taxes, revisions to the Special Supplemental Nutrition Program for Women, Infants and Children (WIC) food packages, and financial incentives for fruits and vegetables were associated with improvements in dietary behaviors [19]. Providing financial incentives to supermarkets to open in underserved areas and increases in SNAP benefits were not associated with changes in diet quality but may improve food security. The authors called for more research to understand the effects of calorie labeling in supermarkets and online SNAP purchasing on consumer purchasing and consumption. The second paper reviewed 64 in-store marketing studies conducted between 2009–2019 and found that the majority of interventions identified at least one positive effect related to healthier food purchasing, consumption, or sales. Promotion was the most commonly studied marketing strategy for single-component interventions, while changing promotion, placement, and product together were the most common for multi-component interventions. The quality of research, however, precluded definitive conclusions, as fewer than 36% of studies used experimental designs. The review called for more research to understand what combinations of strategies work best by product category and retail format [20].

#### *3.2. In-Person Convening Findings*

Research questions generated at the meeting (n = 147) were initially grouped according to the three meeting agenda themes (retailer and manufacturer marketing practices; consumer food purchasing trends; and effectiveness of retail interventions). (Figure 2) Forty-nine questions fell under the retailer and manufacturer marketing practices theme, 59 under the consumer food purchasing trends theme, and 39 under the effectiveness of retail interventions. These questions were refined and reorganized prior to inclusion in the follow-up survey. Two themes—retailer and manufacturer marketing practices and consumer food purchasing trends—were condensed due to overlap between research questions in these categories. In total, 40 questions representing two themes were included in the follow-up survey.

**Figure 2.** Flow chart depicting how research questions were generated and refined through the agenda-setting process.

#### *3.3. Follow-Up Survey Findings*

Twenty-six convening attendees completed the follow-up survey (response rate 57%). Three research questions were eliminated due to low scores: one question earned a low composite score (<3), and two questions earned low scores across all three domains (<3.5) (Table A1).

Research questions that received the highest composite scores focused on describing how frequency, duration, and impact of retailer promotion practices differed by community characteristics and how to leverage SNAP benefits to support healthy eating behaviors. (Figure 3) Research questions that received the highest scores for importance and equity focused on (1) evaluating the impact of retailer marketing practices on consumer health, (2) understanding the optimal retail design to promote healthy and reduce unhealthy purchases, and (3) evaluating the impact of healthy retail policies to address the social determinants of health. These questions, however, received lower scores for feasibility. Research questions that received the highest scores for feasibility focused on describing the current retail environment, including assessing the healthfulness of products currently available and promoted in stores, and describing the factors that influence consumer decision-making.

**Figure 3.** Research questions with the 10 highest composite scores from the follow-up survey. Numbers listed before questions represent ranking from 1–10 by composite score. Research questions were ranked on a 5-point Likert scale in terms of feasibility (y-axis) and health equity (x-axis). Ratings for importance are not displayed due to low variation (3.8–4.5) among the top ten questions. SNAP is the Supplemental Nutrition Assistance Program.

In the open-ended portion of the survey, several participants suggested additional research questions related to the COVID-19 pandemic. The research questions that participants were asked to rank were generated at the January convening, before widespread awareness of COVID-19 pandemic in the U.S., but the survey was conducted in May during the pandemic. A few participants indicated an interest in evaluating how COVID-19, generally, and the U.S. Department of Agriculture (USDA) waivers during the pandemic for SNAP and WIC statutory and regulatory requirements, specifically, affected food supply, retailer marketing, and consumer purchasing. Another participant called for research on how expansion of the SNAP Online Purchasing Pilot Program (a federal program to test the feasibility and impact of allowing online food retailers to accept SNAP benefits [21]) affects small and independent grocers.

Survey respondents identified several additional data sources for healthy retail research in the open-ended section portion of the survey. See Table 2 for a full list of data sources identified through the convening and follow-up survey.

**Table 2.** Data sources for healthy retail research identified at the in-person convening and through the follow-up survey.


#### *3.4. Research Agenda Findings*

Based on the information gathered in steps 2–5 (see Figure 1), a total of 37 research questions, grouped into ten key issue areas, emerged as priorities for future research (Table 3). Five of these issue areas focus on understanding the current food retail environment and consumer behavior and five focus on assessing implementation and effectiveness of interventions and policies to attain healthier retail.




**Table 3.** *Cont.*

Through small and large group discussions at the in-person convening, several cross-cutting considerations for future research emerged and were grouped into three themes: potential research partners, data sources, and study designs and settings (Table 4).


**Table 4.** Cross-cutting considerations for future research on healthy food retail discussed by in-person convening participants.

#### **4. Discussion**

This article is the first to present a national agenda for research to support healthy food retail, developed iteratively and collaboratively by experts in public health research, advocacy, and food retail and marketing. This research agenda builds on the 2011 *Harnessing the Power of Supermarkets to Help Reverse Childhood Obesity* report, which proposed in-store marketing strategies developed collaboratively by retailers, researchers, manufacturers, and marketing professionals to encourage the purchase of healthy products while maintaining or improving retailers' bottom lines [11]. This research agenda reflects advancements in research that have occurred in the intervening years and outlines key areas for future research.

Thirty-seven key research questions, grouped into ten overarching themes, were identified. Priority topics include how frequency, duration, and impact of retailer promotion practices differ by community characteristics and how to leverage federal nutrition assistance programs to support healthy eating. Many of the research questions that received the highest scores in the follow-up survey for importance or health equity received low scores for feasibility, underscoring the need to address barriers to evaluation. Identified strategies to address these barriers include partnerships with retailers, government agencies, business schools, advocacy organizations, and others to implement and evaluate pilot programs and policies, as well as exploration of new study designs and data sharing opportunities.

Of the ten key research themes that emerged, half centered around describing the current food retail environment and how environmental factors shape consumer behavior. Considering that an estimated three-quarters of purchase decisions are made while shopping, a nuanced understanding of marketing strategies used by manufacturers and retailers and how those strategies drive behavior can guide targeted interventions [23]. Additionally, most research to-date has focused on grocery stores, but changes in the food retail environment, including growth in online retail and proliferation of dollar-stores in low-income and rural areas, point to a need for research on nontraditional retail outlets [24–26].

The other five key research themes focused on evaluating interventions designed to improve the retail environment and access to nutritious food. The commissioned reviews highlighted evidence of retailer-, researcher- and government-initiated interventions that have led to increased healthy purchases, including fresh fruit and vegetable prescriptions, revisions to the WIC packages, and financial incentives for healthy purchases using SNAP [19,27,28]. Yet, additional research is warranted to evaluate these interventions at a larger scale, in other settings, and over longer periods of time. Evaluation of novel policies through natural experimentation at the state and local level is also needed. As one step toward facilitating such policy evaluation, federal agencies should provide states greater flexibility to innovate. For example, the USDA could approve state or local waiver applications to remove SSBs from eligible SNAP purchases [29]. Considering SNAP serves as an important source of revenue for many retailers, changes in SNAP and other federal nutrition assistance programs could shift the broader food landscape [30].

#### *4.1. Implications for Research and Practice*

The agenda-setting process centered around promoting health equity, and the research questions identified account for and aim to address health disparities. As researchers and practitioners pursue the policy, systems, and environmental change strategies identified in this agenda, the Equity-Oriented Obesity Prevention Framework developed by Kumanyika can serve as a guide to ensure equity issues continue to be prioritized [31]. Specifically, Kumanyika calls for designing and evaluating interventions using an explicit equity lens that acknowledges the realities of social inequities, incorporates a "people perspective", and prioritizes community engagement.

This research agenda can serve as a resource for researchers writing grant applications, retailers seeking to conduct healthy retail pilots on their own or with researchers or advocates, funders drafting requests for proposals, and advocates engaging in organizational strategic planning. In particular, private foundations and federal agencies including the USDA, CDC, and the National Institutes of Health (NIH) should integrate the research themes outlined in this agenda into their strategic plans, ongoing initiatives, and funding priorities.

While federal agencies have made progress toward recognizing the importance of the food environment and healthy retail as a strategy to reduce disease and disparities, much work remains. For example, in the *National Nutrition Research Roadmap for 2016-2021*, the federal Interagency Committee on Human Nutrition Research, which includes representatives from USDA and the U.S. Department of Health and Human Services, identified research on food retail as an area of interest [32]. The CDC has acknowledged the importance of the food retail environment in multiple reports and, in 2015, published *Healthier Food Retail: An Action Guide for Public Health Practitioners* [33,34]. The CDC also promotes healthier retail among small, independent retailers through cooperative purchasing initiatives and communities of practice in the High Obesity Program and Racial and Ethnic Approaches to Community Health program. The NIH, between 1975 and 2018, funded more than 200 grants related to healthy food retail, and the *2020–2030 Strategic Plan for NIH Nutrition Research* recognized the important role of the food environment in shaping dietary behavior [35,36]. At the same time, the *Strategic Plan for NIH Obesity Research* only briefly mentions the food environment and does not mention retail [37]. Similarly, healthy retail is missing from the *USDA Science Blueprint* [38].

As federal departments and agencies use this research agenda to guide future funding priorities, coordination and harmonization across these entities are needed to ensure existing efforts are leveraged and amplified and that critical areas are not overlooked. Drawing on recent recommendations from Fleischhacker et al., creation of a new authority for cross-governmental coordination of nutrition research and policy, as well as strengthened authority, coordination, and investment for nutrition research within the NIH and USDA could help to catalyze new science and partnerships [39].

Research on healthy retail requires collaboration across sectors and disciplines, including relationship-building and data sharing between researchers and retailers. Research institutions and funders should provide financial and technical support to advance these efforts without expectation of immediate research deliverables. For example, to improve accessibility and affordability of data, foundations could serve as a conduit between researchers and industry, following the model of the RWJF Health Data for Action program [22]. Another potential model is the Johns Hopkins Bloomberg American Health Initiative, which provides funding to researchers engaged in consultancies and special projects that facilitate cross-sector partnerships [40].

Progress toward meeting the research goals outlined herein should be monitored. In five years, key stakeholders should be re-convened to discuss achievements and remaining gaps. In the intervening years, researchers, retailers, manufacturers, funders, and advocates should convene periodically to foster partnerships and data sharing.

#### *4.2. Strengths and Limitations*

This study has limitations. First, the list of attendees for the in-person convening was developed with the aim of bringing together groups across research and practice with a mutual interest in promoting health; thus, some interested parties such as manufacturers and trade associations may have been excluded, and the research questions and other ideas generated at the convening may be subject to bias. Additionally, 43 percent of meeting participants did not complete the follow-up survey; therefore, survey results may be impacted by self-selection bias. Finally, the food retail landscape is rapidly evolving, and this agenda reflects priorities identified at a specific period in time. For example, research questions were generated at an in-person convening in January 2020, before widespread awareness of COVID-19 in the U.S. The pandemic brought about important changes in how people in the U.S. purchase groceries and inspired new research questions (e.g., what are the impacts of increased online grocery purchasing; increased at-home food preparation; expansion of the SNAP Online Purchasing Pilot Program?) [21,41].

The methods used in this study, however, are strong. This study used a multi-step, iterative approach to develop the final research agenda. A range of stakeholders who represented diverse disciplines and organizations, including retailers, were engaged in this process. Finally, a focus on health equity was incorporated in every stage of the retail research agenda-setting process, increasing the likelihood that the research questions identified as part of this process will help address disparities in health.

#### **5. Conclusions**

The food retail environment presents an ideal setting for intervention to improve diet quality and reduce the prevalence of chronic disease and health disparities. The collaborative agenda-setting process, which included representatives from academic, government, advocacy, funding organizations, and industry, built consensus around key research gaps. The research questions identified through this process aim to inform policies and corporate practices that improve the food retail environment, and, ultimately, public health. This agenda can serve as a guide for researchers, funders, and advocates, ensuring that future work fills critical knowledge gaps, promotes equity, and advances policy and practice.

**Funding:** This research was supported by Healthy Eating Research (a national program of the Robert Wood Johnson Foundation), the Johns Hopkins University Center for a Livable Future, the Institute for Health and Social Policy at the Johns Hopkins Bloomberg School of Public Health, the Bloomberg American Health Initiative of the Johns Hopkins Bloomberg School of Public Health, and the Center for Science in the Public Interest.

**Author Contributions:** Conceptualization and methodology: A.A.H., M.M.L., K.A., M.T.S., M.G.W., A.J.M. Formal analysis and investigation: A.A.H., M.G.W., A.J.M. Writing—original draft preparation: A.A.H., E.S., M.G.W., A.J.M. Writing—reviewing and editing: all. Project administration and funding acquisition: A.J.M., M.M.L., K.A., M.T.S., M.G.W. All authors have read and agreed to the published version of the manuscript.

**Acknowledgments:** The authors would like to express their gratitude and appreciation to the attendees of the January 2020 Healthy Retail Research Convening, members of the Scientific Advisory Committee (Joel Gittelsohn, Karen Glanz, Lisa Harnack, Allison Karpyn, Anne Palmer, Kate Reddy, Christina Roberto, Shannon Zenk), and other expert reviewers (Betsy Anderson Steeves, Sheila Fleischhacker, Lucia Leone) who were instrumental in shaping this research agenda. The authors would also like to thank staff of HER, The Food Trust, and Center for Science in the Public Interest for their valuable guidance in planning and executing the research convening, especially Darya Minovi, Julia McCarthy, and Julia Koprak.

**Conflicts of Interest:** The authors (A.A.H., E.S., M.G.W., A.J.M.) declare no conflict of interest. One funder (H.E.R.) did contribute significantly to the design of the study and in the collection of the data at the convening; however, the affiliated authors (M.M.L., K.A., M.T.S.) did not play a role in the analysis or interpretation of the data.

#### **Appendix A**


**Table A1.** Follow-up survey participant (n = 26) mean rankings of research questions in terms of feasibility, equity, and importance.


#### **Table A1.** *Cont.*


**Table A1.** *Cont.*

\* Indicates question was eliminated due to low composite score or low score for equity or importance (<3).

geography (e.g., urban vs. rural)? 3.70 4.00 3.78 3.83

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*
