*Article* **A Model Depicting the Retail Food Environment and Customer Interactions: Components, Outcomes, and Future Directions**

**Megan R. Winkler 1, Shannon N. Zenk 2, Barbara Baquero 3, Elizabeth Anderson Steeves 4, Sheila E. Fleischhacker 5, Joel Gittelsohn 6, Lucia A Leone <sup>7</sup> and Elizabeth F. Racine 8,\***


Received: 7 September 2020; Accepted: 16 October 2020; Published: 19 October 2020

**Abstract:** The retail food environment (RFE) has important implications for dietary intake and health, and dramatic changes in RFEs have been observed over the past few decades and years. Prior conceptual models of the RFE and its relationships with health and behavior have played an important role in guiding research; yet, the convergence of RFE changes and scientific advances in the field suggest the time is ripe to revisit this conceptualization. In this paper, we propose the Retail Food Environment and Customer Interaction Model to convey the evolving variety of factors and relationships that convene to influence food choice at the point of purchase. The model details specific components of the RFE, including business approaches, actors, sources, and the customer retail experience; describes individual, interpersonal, and household characteristics that affect customer purchasing; highlights the macro-level contexts (e.g., communities and nations) in which the RFE and customers behave; and addresses the wide-ranging outcomes produced by RFEs and customers, including: population health, food security, food justice, environmental sustainability, and business sustainability. We believe the proposed conceptualization helps to (1) provide broad implications for future research and (2) further highlight the need for transdisciplinary collaborations to ultimately improve a range of critical population outcomes.

**Keywords:** grocery store; restaurant; environment; retail; food purchasing behavior; dietary intake

#### **1. Introduction**

Dramatic changes in the retail food environment (RFE) are evident over the past few decades, and even the past few years [1,2]. The number of traditional supermarkets are declining, while alternative grocery formats such as discount and convenience focused grocers are proliferating [1]. Food is increasingly found everywhere, across stores and businesses that are not traditionally considered "food" outlets [3,4]. Exponential growth in the number of dollar stores, pharmacies, and their grocery offerings exemplifies both of these trends [1,2,5]. Due in part to technological advances, online grocery shopping with delivery or curbside pick-up may be the wave of the future, further accelerated by consumer and federal responses (e.g., expanding online shopping options for US Department of Agriculture Supplemental Nutrition Assistance Program (SNAP) participants) to the coronavirus pandemic [6]. Still, prior to the pandemic, the majority of the American food dollar went to food prepared away from home [7]. Prepared food delivery has surged, with digital ordering and third-party delivery services helping to fuel its rise [8,9]. These changes partially reflect a growing consumer demand for convenience due to time scarcity [10–12], but also the decisions of a variety of other food actors including outlet owners, suppliers, and manufacturers to compete for customers through facilitating convenience. The RFE—including these recent trends—has implications for health, but also for other outcomes such as community and economic development.

Over the past 15 years, conceptual models of the RFE have played an important role in guiding research and intervention efforts, and thus have advanced the field. In 2005, the Model of Community Nutrition Environments by Glanz and colleagues identified several key components of the RFE, such as the "consumer" and "community" nutrition environments, which facilitated communication in the field [13]. The ecological framework depicting multilevel, interacting influences on what people eat by Story and colleagues positioned retail food sources as a key aspect of the physical environment [14]. In her book, Morland expanded on the pathways by which the RFE affects obesity and personal factors that moderate these relationships [15]. Yet, the recent convergence of changes in the RFE and advances in the field suggest the time is ripe to revisit how we conceptualize the RFE. Previous models tend to miss important components of the current and emerging environment, such as the wide varieties of retail food sources, involved actors, and business models, focus solely on diet and/or health as the outcomes of interest, and underemphasize the broader context that influences and interacts with the RFE to affect a diverse range of population outcomes.

The proposed model in this paper was prepared by The Healthy Food Retail Working Group leadership team. The Healthy Food Retail Working Group is a US collaboration of over 150 researchers and stakeholders jointly supported by Healthy Eating Research, a national program of the Robert Wood Johnson Foundation, and the Nutrition and Obesity Policy Research and Evaluation Network (NOPREN), which is supported by a cooperative agreement from the Centers for Disease Control and Prevention's Division of Nutrition, Physical Activity, and Obesity. The Healthy Food Retail Working Group holds bimonthly webinars on retail food topics and convenes smaller sub-groups to explore topics in further depth and develop collaborative research, practice, or policy projects.

In March 2019, the working group leadership met at the annual NOPREN meeting and strategized on research needs and future directions including a conceptual model to guide research. This process began as a brainstorming activity and a review of the previously published RFE conceptual models. We agreed that there were elements of the RFE missing from previous conceptualizations. To address this, we began meeting throughout the next year, and with feedback from the wider membership, developed a conceptual model to reflect RFE evolutions and its complexity, as well as what has been learned about the RFE over the past 20 years of public health research. Our focus was on developing a model that captured the chronic, ongoing processes, and outcomes of the RFE, and much of our efforts preceded the recent COVID-19 pandemic and historic protests against police brutality across the U.S. While we believe some model components and outcomes are highlighted by the COVID-19 pandemic and the movement for racial justice, there are others we do not address (e.g., state-mandated restaurant closures). As a compliment, Leone and colleagues (see Special Issue "Retail Strategies to Support Healthy Eating" https://www.mdpi.com/1660-4601/17/20/7397) offer ways that the proposed model could be used to inform research directions during significant disruptions, such as pandemics.

The aim of this paper is to propose an updated conceptual model of the RFE and its relationships with customer behavior that produce a host of significant population outcomes. Below, we present an overview of the conceptual model and our underlying assumptions and motivations. We then describe and justify each of the model components. Last, we discuss how the model can be used to direct broad future directions in observational, intervention, and policy research to understand and modify the interactions between customers and the RFE with the intention of improving societal outcomes.

#### **2. Overview and Motivation for the Retail Food Environment and Customer Interaction Model**

As an overview, the Retail Food Environment and Customer Interaction Model (Figure 1) breaks the RFE down into business models, actors, and sources and their influence on the customer retail experience (e.g., food availability, promotion, quality). Our model depicts reciprocal relationships and influence between the RFE and customers, including their individual, interpersonal, and household characteristics that affect sales/purchases. The model highlights the multilevel context in which the RFE and customers operate and expands the population outcomes produced by RFEs and customers that should be considered moving forward: health, food security, food justice, environmental sustainability, and business sustainability. See Table 1 for component definitions.

Several underlying assumptions motivated the proposed model components and relationships. First, we took a highly-inclusive posture to address the multifactorial nature of the RFE in the US and its wide-ranging, discipline-crossing implications for society. However, we recognize as predominantly public health scholars that our focus remains on health and thus describe much of the model from that evidence base. We also conceptualized the model's diverse and multidimensional components as a complex dynamic system. This is represented not only in the reciprocal relationship between RFEs and customers, but also by the inclusion of multilevel contexts that can affect RFEs, customers, and their interactions. Finally, we speculated that an important driver of the RFE evolution has been the supply and demand for convenience and highlighted this in several model components. Time scarcity [10–12], growing mental fatigue and stress [16], and changing social norms [17] around daily food preparation may all contribute to customers' increasing demand for highly-accessible, limited-preparation products [18]. This demand has often been met by RFEs

providing an abundance of ultra-processed, highly palatable, calorically-dense products through an ever-growing accessibility [19–21]. Yet, these patterns are juxtaposed by others that suggest that large swaths of the US are devoid of a variety of convenient foods and sources [22–24]. Thus, even an important driver, such as convenience, must be considered in a larger system of relationships and factors in order to understand why diverse outcomes can be produced.



#### **3. Retail Food Environment**

A key focus of our efforts to advance prior conceptualizations was to more comprehensively identify specific components of the RFE. We define the RFE as the environment where all food and beverages are purchased by consumers, including foodservice operations such as restaurants. We also recognize that the RFE is part of a larger food system, including agriculture, farming, and food production. However, in our model, we focus on the retail components most immediate to where food is sourced and purchased by customers, including: Retail Food Sources, Retail Food Actors, Retail Food Business Models, and the Retail Food Customer Experience. While differentiating the various components of the food environment is helpful, we acknowledge that overlap can and does exist among these components.

#### *3.1. Retail Food Sources*

Retail food sources (e.g., stores, restaurants, websites) are settings where people can purchase food and beverages, and are a well-known, well-studied concept in food environment research. Most investigations have studied these sources by examining the geographic-related aspects, such as number of, proximity to, and density of food outlets (i.e., the physical locations whose primary business is to sell food, such as restaurants and stores) [13]. Using these measures, research has aimed to characterize community food environments and examine their associations with community residents' diet and health related outcomes [25–28]. For example, prior evidence suggests positive relationships between convenience store availability and obesity among children [25] and between relative availability of unhealthy (e.g., fast food, convenience stores) to healthy (e.g., supermarkets, farmers' markets) sources with adult obesity [28].

Yet, such conceptualizations of retail food sources have insufficiently addressed the full and evolving range of settings and modalities where food and beverages can be purchased. Business responses to address customer convenience (i.e., reduce customer time and effort in food preparation and acquisition) have likely driven a growth in retail food sources in the US [29] and contributed to an ever-increasing ubiquity of ready-to-eat foods and beverages available for purchase. Thus, our conceptualization (Figure 2) aims to capture a more complete range of retail food sources that have evolved and classifies them across two dimensions of customer convenience: food preparation and accessibility.

**Figure 2.** Common and Emerging Retail Food Sources across Two Dimensions of Customer Convenience: Accessibility and Degree of Required Preparation. Accessibility involves the ability for customers to obtain food products from a retail source from their immediate location (e.g., food can be delivered to their location or customers are required to travel to source). Degree of required preparation captures the typical proportion of products offered by the source that is prepared: ready-to-eat versus unprepared.

The first convenience dimension—food preparation—demonstrates the variation across sources in the typical proportion of products offered that are prepared: ready-to-eat versus unprepared. As shown in Figure 2, there is an apparent imbalance in the types of sources that primarily offer products that eliminate at-home food preparation versus those that offer unprepared versions. Some sources, such as fast food, restaurants, and food trucks, only offer ready-to-eat products. However, ready-to-eat foods are also staples in gas-marts and convenience stores through offerings of pre-packaged foods and increasingly grab-and-go delis and hot prepared food [30]. Even grocery stores and supermarkets are part of this prepared food trend [7,31], though continue to offer a greater percentage of products that require some (e.g., frozen pizza) or complete (e.g., eggs) at-home preparation. These offerings stand in contrast to other sources, such as farmer's markets and meal kit deliveries, which continue to sell a majority of products that require some degree of preparation (e.g., cut, chop, and sauté fresh vegetables).

Sources have also evolved to address customer convenience through the dimension of accessibility. We view accessibility as the ability of customers to purchase products from their immediate location (e.g., home, work, school). Changes in accessibility were first observed through the staggering spread of brick-and-mortar food sources that narrowed customers' travel distances to venues. For instance, evidence suggests that the density of fast food chains and restaurants near US homes and workplaces significantly increased between 1971 and 2008—in some cases doubling [32]. While these changes contributed to today's approximately 200,000 fast food venues [33] and 153,000 convenience stores/gas-marts [34], accessibility has also recently evolved to no longer require people to travel to and visit brick-and-mortar locations. Such immediate accessibility has in some respects been around for decades through vending machines, worksite cafeterias, and pizza delivery. However, accessibility

in recent years seems to be exponentially expanding. Ready-to-eat packaged foods (e.g., candy) are offered in non-food outlets and checkout aisles (e.g., barber shops, home improvement stores, clothing stores) [3,4]; sit-down and fast food restaurants regularly offer options for delivery, often via third-party online applications and platforms [9]; and even sources that primarily sell products requiring preparation are now delivering (e.g., meal kit deliveries, online grocery delivery). Moving forward, we need a better understanding of the impacts of these increasing forms of accessibility and prepared food products offered by retail sources. Future research can investigate how some modalities might be used to improve the ubiquity of healthier ready-to-eat options as well as disentangle for whom these convenience dimensions are more or less available.

#### *3.2. Retail Food Actors*

Retail food actors are the people that work in the RFE whom, at various steps in the process typically towards the middle and ends of the food supply chain, determine the foods and beverages available at a source (e.g., managers/owners, suppliers/distributors, merchandising managers, and sales representatives). The retail food actors interact to determine which items are feasible to sell, store, and transport while maintaining quality and minimizing waste. For instance, when source managers or restaurant owners plan to sell a new item, they identify potential suppliers and understand the space, cost, and shelf life requirements necessary to sell the product in a safe and profitable way. Food manufacturer sales representatives are another example, who work with store managers to promote products and marketing strategies, such as in-store displays [35].

Each actor has their own specialty and focus. A sales representative's focus is often to develop relationships with retail outlets that will provide environments for food products to reach consumers and cultivate demand. A distributor's focus may be to develop a supply chain that efficiently moves food from warehouses to stores and restaurants. Alternatively, a store manager or restaurant owner's focus may be to provide an array of items that customers demand in an efficient and pleasant environment [36,37]. The varying foci and goals of these actors have often resulted in an efficient system that provides an abundance of convenient, non-perishable, manufactured food and beverage items, as these are often more logistically and financially appealing to manage [38–40].

Relative to other RFE components in the model, very little literature in public health nutrition has investigated the impact of these actors on the RFE, though there is a growing base of research examining the role of store managers [36,41–43]. Such research is important as these actors develop reciprocal and deterministic processes that influence the current RFE (e.g., informal and formal product contract agreements [44,45], managers requesting products from distributors based on customer demand and what they can maintain due to resources and infrastructure) [46,47]. A better understanding of how the retail food source is influenced by the goals, foci, and decisions of these actors may be necessary to develop more effective policies and sustainable interventions to improve population outcomes.

#### *3.3. Retail Food Business Models*

Another RFE component that requires additional research is the business models used across each retail food source. Business models direct a source's operations, financing, target customer base, and mission. Understanding the business model of a source, particularly products sold and services provided, helps to understand their priorities. For example, sources offering culturally-tailored products might be demonstrating a priority to address the needs and preferences of a specific ethnic community [48–50], while sources offering products with specific values, such as locally-sourced, or dietary requirements, such as gluten-free [51], may be targeting and prioritizing other customer groups. Services provided (e.g., fast food versus "dining experience") can also indicate a source's targeted customer base (e.g., income/class, available time, cooking abilities/preferences). Products sold might also reflect a source's priorities to generate additional revenue streams, such as stores that participate in federal assistance nutrition programs [52] to expand their customer base, as well as how much local demand is valued over operational convenience (e.g., product variation versus the same products at all locations) [53].

Business outcomes, including revenue and profits, are often the ultimate goal for many sources. Such goals are at times a necessity, given that some source types (e.g., grocery wholesalers and stores) struggle with low profit margins [54]. Yet, some sources may have additional goals beyond profit. For example, institutional foodservice companies might be profit driven, but they contract with community-based institutions, such as colleges, workplaces, prisons, or hospitals [55]. This partnership creates a mix of profit motive and community benefit where the institution's goals, such as for healthy eating and/or locally-sourced products, influences the foods that the foodservice company provides.

Ownership is another indicator of the business model, and a range of ownership types with diverse goals exist across the RFE. The majority of foods and beverages purchased in US are sold by publicly-traded corporations [56–58], such as Walmart, Kroger, McDonalds, Sysco, and Starbucks. However, there are a number of large-chain food retailers that are privately-owned such as Chick-fil-A, Publix, Meijer, and Subway. Such privately-owned chains, while not always held to produce profits for shareholders, continue to dominate RFE spaces (in terms of profits, reach, etc.) over the private and independently-owned source with only one or two locations. Other examples of ownership models include food cooperatives (co-ops) and community-owned businesses. Co-ops involve groups of people that use membership fees to collectively operate a food retailer. Some co-ops are not-for-profit companies, allowing more flexibility to operate the co-op in a manner aligning with the co-op's mission or changing member needs. Community-owned business food retailers are often for-profit businesses that are financed, owned [59], and operated collectively by community members (e.g., Baldwin Market in Florida), and differing from co-ops often raise more capital and investments to allow "capital-intensive enterprises to start at scale [59]".

The past several decades have brought an important RFE transformation from small independent ownership to large chain often corporate/franchise ownership [60–62]. In some cases, entire groups of sources may be corporately-owned, such as fast food. In other cases, ownership at sources, such as grocery stores and supermarkets, remains relatively diverse; though, these also show growing declines in the presence (number and market share) of independent ownership [60]. With these shifts in centralizing ownership to fewer hands, much remains to be investigated and understood about how these different ownership types and business models contribute to the RFE [63,64].

#### *3.4. Customer Retail Experience*

Together, retail actors, business models, and retail sources combine and lead to the final component of the RFE: the customer retail experience. This component consists of the characteristics of food and beverage products for sale and the broader environment that people encounter when making their purchases. Referred to by Glanz and colleagues as the "consumer nutrition environment" [13], these features were mainly conceptualized as occurring within a physical location. Yet, given increasing shifts to online purchasing, customers are now also experiencing retail food spaces through webpages and mobile applications.

The traditional marketing mix of product, price, place, promotion, and people remains a helpful way to classify the customer retail experience [65]. In comparison to research on retail food sources, fewer studies in the field have examined how features of the customer retail experience within those sources relate to purchasing, consumption, or health outcomes [26,27]. This work is important as studies examining links with sources often rely on classifying entire source types as either healthy (e.g., supermarkets) or unhealthy (e.g., fast food); yet, this can neglect the variation in product mixes (e.g., supermarkets offer plenty of unhealthy products), placement, and other marketing features within a source that influence customer purchasing [66–68].

Of the limited evidence examining features of the customer retail experience, many have studied food product availability or prices. Both the absolute and relative availability and prices of healthful and unhealthy foods, as well as availability of culturally-appropriate products [49,50], may be relevant for consumers' purchasing decisions [69–71]. Often, unhealthy products are more available [72–75] and less expensive than healthful products [76]. Product quality and variety (i.e., number of options), such as for produce or milk options, also influence purchasing decisions [77–81] and can vary across source type and neighborhood [82–84].

Other features, including placement and promotion, have been less studied, although industry practices provide indirect evidence that these, too, are important for creating a customer retail experience that translates into sales. For instance, food/beverage manufacturers spend an estimated USD 50 billion per year, or 70% of their marketing budget, on in-store trade-promotion fees [35]. Such fees can guarantee certain product placement (e.g., checkout aisles) and/or promotion through cooperative advertising (e.g., store circulars) and discount campaigns (e.g., "2 for 1"). These practices also occur in online shopping spaces, such as pop-up advertisements, notifications, and cart "reminders" [85]. Promotion also occurs at the packaging level, as significant efforts have been made by manufacturers to attract customers (e.g., children's cereal boxes [86]) and by public health to inform customers of a product's nutritional composition and quality (e.g., nutrition label reform [87], front-of-package, and traffic-light labeling [88–90]). Even newer features of shelf promotion, such as undershelf lighting in the candy aisle, signals that these features will continue to evolve as the competition for customer attention and thus sales endure among companies and product categories [91].

The final feature—people—also affects customers' decisions on where to shop and the food and beverage products to which they have access. Despite limited literature, studies indicate that negative social interactions influence people's shopping locations and can range from inefficient, unenthusiastic service to forms of discrimination and stigma [92–95]. For instance, Black Americans have described employees watching, following, or treating them with less respect and experienced this behavior while shopping in predominantly White neighborhoods or in stores owned by individuals of a race/ethnicity different than their own [94,96–99]. Research also highlights that some customers frequent sources that they trust and especially those with which they have a built relationship [100,101].

#### **4. Retail Sales and Customer Purchasing**

The conceptual model involves two sides—an RFE side that presents key components that are most immediate to where food is sourced and purchased by customers and a customer side that presents the many aspects relevant to individual variation in customer purchasing and dietary intake (see Section 5). The two sides connect at the point of a transaction or where a product is sold by the retailer and equivalently purchased by the customer (Figure 1). In comparison to prior models, we view the relationship between the two sides as reciprocal: actions of RFEs influence customer behaviors and customer behaviors influence RFE actions. Moreover, aspects of both the RFE and individual customers can interact to lead to a customer purchase.

#### **5. The Customer: Individual Dietary Intake, Individual Characteristics, and Household Characteristics**

The right-side of the model (Figure 1) represents the customer and the multidimensional characteristics that influence decisions about personal dietary intake and food and beverage purchasing. The relationship between dietary intake and purchasing is bidirectional, and we propose that a wide range of individual, interpersonal, and household characteristics influence individual purchasing and ultimately dietary intake (Table 2) [14]. At the individual level are the intrapersonal factors that influence dietary intake and purchasing behaviors. Previous models and a large body of evidence indicate that factors such as attitudes, knowledge, food preferences, socio-demographic characteristics, lifestyle behaviors (e.g., smoking), stress, and cultural norms influence these behaviors. For example, individuals who have less education and/or poor employment consistently report lower dietary quality [102,103], which may be due to limited time or financial resources. Individuals' food knowledge and attitudes are also important, as greater nutrition knowledge has been associated with better dietary

quality and may reflect a better nutrition label literacy and ability to overcome food marketing tactics to make healthier purchases [104].

**Table 2.** Examples of individual, interpersonal, and household characteristics relevant to food and beverage purchasing and dietary intake behavior.


Individuals are embedded in households and other interpersonal contexts, and much evidence suggests that characteristics from these contexts, such as SNAP status, income, social support, social norms, shopping behaviors, and food preparation skills and decision-making, are also related to dietary intake and purchasing. For example, sources and timing of food benefits (e.g., SNAP benefit schedule) shape the number of food shopping trips and their food baskets, as diets tend to be healthier around the weeks that SNAP household benefits are redeemed versus other times of the month [105,106]. Household income is also consistently related to purchases with higher household incomes purchasing healthier foods and beverages and having greater access (e.g., fruits and vegetables) compared to lower income households [107]. In households of immigrant families, the level of acculturation of the head of household influences what food is purchased [108,109]. Additionally, household members' work schedules and transportation options are related to shopping trip frequency and foods purchased and may contribute to customers' increasing need for convenience [110].

#### **6. Community, State, Tribal, National, and Global Contexts**

As shown in Figure 1, the RFE, as well as customers do not operate in a vacuum. Instead, macro-level factors, including economic, social, media, built environment, policy, and others, influence the RFE, customers, and their relationships. A growing body of evidence examines factors at this macro-level, and we group these factors under two contextual levels: community context, including neighborhoods and city/local jurisdictions, and the broader state, tribal, national, and global context. Table 3 provides examples of relevant factors in each context. These factors may directly affect the RFE and customers, as well as modify RFE–dietary intake relationships.


**Table 3.** Macro-level contexts with example factors that influence the retail food environment; customer purchasing, dietary intake, and individual and household characteristics; and their relationships.

Note. SNAP, US Department of Agriculture Supplemental Nutrition Assistance Program; WIC, US Department of Agriculture Special Supplemental Nutrition Program for Women, Infants, and Children.

Under the community context, zoning codes and commercial real estate professionals can directly influence where food sources are located; transportation systems can affect the food sources that consumers can reach; local social norms might ultimately influence the food options available; and tax policies such as municipal sugar-sweetened beverage taxes affect the prices that consumers face [111]. Under the broader contexts, numerous factors from the state, tribal, and federal levels influence food retail and customer behavior such as the following examples. First, stocking requirements for participating retailers in SNAP and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) affect the products they carry, and these are particularly impactful in smaller sources, such as dollar stores, that may not otherwise carry as many healthier items [43,112]. Second, minimum wage laws affect the financial resources that consumers have to purchase foods. Third, industry advertising of food products may affect food choices and thus dietary intake. Finally, healthy food purchasing incentive programs and nutrition education programs (e.g., within SNAP) increase financial access and education about healthy foods and beverages. As suggested by the embedding of the community context within the broader context in Figure 1, factors at the state, tribal, national, and global levels can influence the community context as well. For example, preemptive laws can prevent state and local governments from enacting policies that influence the RFE, such as taxes and labeling [113].

Moreover, community and broader contexts may modify effects of the RFE on sales/purchasing and in turn individual dietary intake. For instance, the effect of retail food outlet availability on where people purchase foods may depend on safety of the surrounding community environment, as shoppers have reported avoiding stores or certain shopping times due to unsafe neighborhood conditions, including drug sales, violence, and harassment [93,94]. Because consumption of high-fat, sugary foods and beverages can alleviate stress, exposure to such community stressors may also increase the effect of household availability of these food types on individual consumption choices. With regard to the broader state, tribal, national, or federal context, for example, the impact of in-store food/beverage marketing to children on purchasing may vary depending on regulations for television food advertising to children. That is, it is possible that being exposed to both in-store marketing and television food advertising will have a stronger impact on purchasing than when one is only exposed to one or the other. Thus, our model recognizes a wide variety of factors that may influence the RFEs, customers, and their sales/purchasing interface as well as alter these relationships.

#### **7. Population Outcomes**

We posit that the dynamics and interactions between RFEs, individuals and households, and their larger contexts can produce a host of population outcomes. Scholars have previously articulated the importance of examining the multiple outcomes produced by national and global food systems [114,115]. In this conceptualization, we offer five for consideration: health; food security; environmental sustainability; business sustainability; and food sovereignty, equity, and justice.

Population health is the outcome most familiar to RFE researchers from the discipline of public health. It aims to uncover the ways this system contributes to diet-related non-communicable diseases, such as obesity, type 2 diabetes, and cardiovascular disease [13,14]. Those interested in improving this outcome often offer RFE modifications that help "make the healthy choice, the easy choice," such as offering and widely promoting products consistent with national dietary guidelines [116].

Food insecurity is another outcome, and one that at the time of this writing has dramatically risen as a result of the economic implications from executive orders required to curb the spread of COVID19 (e.g., miles of cars waiting at food pantries [117]) as well as damage to RFE locations that accompanied peaceful protests for racial justice [118]. In usual times, food insecurity is likely to occur when federal nutrition assistance is not accepted at all sources, price structures lead high-fat and high-sugar products to be most affordable (i.e., lowest-cost dietary option), and fresh and healthier options are not equally available across communities. As such, healthy food and beverages are not affordable or accessible to all groups, and this most often impacts the economically and socially disadvantaged [14,116].

Outcomes of the system not only relate to people but to the environment. Food waste is one example, as more than 400 pounds of approximate waste per person was observed at the US retailer and consumer levels in 2010 [119]. Other examples relate to the agriculture and transportation practices required for the types of products sold and purchased. Many suggest that the majority of available products are produced and commonly transported in a way that leads to environmental degradation, as they require methods that can diminish soil fertility, emit greenhouse gases, deplete freshwater resources, and/or neglect biodiversity [114,115].

The system also contributes to business and economic outcomes, which reflect the "health" of the source's business performance in the US market economy [120]. Here, goals of generating sales, profits, and competitiveness are key and for some businesses may be the primary motivators for decision-making [120–123]. Food retailers and companies often aim to achieve such goals by interrogating consumer "choice" and the predictors of which retailers will be shopped and which products purchased (e.g., price strategies, product mix, store layout) [122,124]. Of the outcomes identified, this outcome has arguably been the best performing in recent decades, as US supermarket and fast food industries experienced an estimated annual revenue in 2019 of USD 682 billion and 293 billion [125,126], respectively. However, the COVID-19 pandemic is likely to change this success

for some industries, such as restaurants, which observed a 51% drop in food-away-from-home monthly expenditures in March 2020 compared to March 2019 [127].

Finally, there are also significant outcomes of the system characterized through the lens of food sovereignty, equity, and justice. In this perspective, inequalities in power are central, and the rights of individuals and communities to define, produce, and sell their own food are emphasized [116,128]. To achieve such outcomes requires addressing the socio-structural barriers (e.g., economic inequality, racism, sexism) that have historically-marginalized, inequitably targeted, and resource-deprived certain groups and populations [129]. It also demands the development of sustained opportunities for communities to create the RFE that best serves their needs and interests (e.g., supporting tribal food sovereignty and Black-owned businesses).

Articulating these five outcomes is necessary to not only highlight the multiple outcomes produced and that need to be considered in future research but the challenges and opportunities that also lie ahead. For instance, when we focus on a single RFE goal and ignore that other outcomes are produced, we create solutions that may address our goal but simultaneously produce harm in other areas. Such consequences may be unintentional or well-known (e.g., promotion of unhealthy, processed foods which increase profits, but are associated with non-communicable disease [130–132]). Yet, moving forward it may be important to reframe these varying outcomes from inevitable systemic trade-offs to sites of opportunity. Diverse groups working to improve the RFE could identify ways to work at cross-purposes, achieve goals for multiple outcome areas, and potentially do so with greater efficiency and less duplication and resources. Working together will also push discipline-centric change agents to consider the feasibility and sustainability of their proposed solutions and may help spur the creation of more worthwhile and effective transformations. While collaboration and attention to multiple outcomes will be easy for some, other groups may require support or even accountability measures to help cultivate "common ground" (e.g., reframing from businesses profitability to sustainability), and many have already been calling for and provided specific strategies to do so [115,129,131,133].

#### **8. Future Directions**

The Retail Food Environment and Customer Interaction Model attempts to capture key RFE and customer components in the US that converge to shape food and beverage purchases with diverse societal outcomes. Expanding upon previous frameworks [13–15], we believe this updated model highlights: (1) the multifactorial nature of the RFE; (2) the wide-ranging and discipline-crossing outcomes produced for society; (3) the reciprocal and dynamic relationships between RFEs and customers as well as with factors from multilevel contexts creating a complex system; and (4) the importance supply and demand for convenience has and continues to play in shaping the US RFE. As such, the model adds important information that can guide future research on the broader RFE context for dietary intake and help to inform public health interventions and policies aimed at improving RFE settings.

The encompassing nature of our model has broad implications for future research and can guide numerous research questions. However, here for the sake of brevity, we focus our comments on three important gaps that we identified throughout model development. First, additional research is necessary to investigate the role and influence of certain understudied RFE components: retail actors, business models, and the customer retail experience. A better understanding of these components is required to develop effective interventions and partnerships that are more likely to improve outcomes. Second, there is a need, especially in public health, to broaden our awareness of outcomes beyond health in an attempt to anticipate the wide array outcomes that a single change to the RFE and customer interaction can generate. Finally, while literature examining why convenience is an important driver of behavior exists in the disciplines of psychology, behavioral economics, and cognitive science, there remain relatively less investigation and understanding of nutrition and public health. Uncovering what convenience means to customers and how best to capitalize on it to improve health and other population outcomes are important directions moving forward.

Given the complexity, dynamics, and reciprocal processes of the Retail Food Environment and Customer Interaction Model, we also suggest a need for more sophisticated research methods and transdisciplinary partnerships. Two recommended research approaches are systems science and multilevel, multicomponent (MLMC) interventions [134–136]. Systems science involves methodological approaches, often computational models, that aim to understand the impacts produced from complex interrelated mechanisms and relationships among multiple factors [134,135]. Except for a few exceptions [137–140], relatively little work has studied the RFE using such methods, and incorporating these approaches could help to not only identify solutions that improve multiple outcomes but identify those to avoid to circumvent unexpected consequences. MLMC interventions are large, complex, multidimensional interventions that often require significant coordination, stakeholder buy-in, and resources; yet, their utility also lies in identifying which individual and/or set of components most effectively improves outcomes [141,142]. Both the model's complexity and these research approaches suggest that transdisciplinary, collaborative leadership will be required. Bringing together stakeholders from many disciplines, such as agriculture, business, public policy, regional/urban planning, nutrition, social sciences, and public health, could help to build more and stronger transdisciplinary projects that are better positioned to effectively improve the RFE for a variety of societal outcomes.

#### **9. Conclusions**

This paper provides a model depicting the interactions of the RFE and consumer behavior while also highlighting some of the outcomes of this system as witnessed in the US. We view the Retail Food Environment and Customer Interaction Model as a "living" conceptualization and hope that it inspires many additional, more refined versions. We encourage research utilizing this model to help us better understand why food sources operate in certain locations, how food sources decide which foods to carry, and why customers choose to purchase certain foods. Then using this insight, transdisciplinary efforts should work to develop solutions that modify the RFE-customer relationship in ways that ultimately improve a range of population outcomes.

**Author Contributions:** Conceptualization, M.R.W., S.N.Z., B.B., E.A.S., S.E.F., J.G., L.A.L., and E.F.R.; methodology, B.B., E.F.R., M.R.W., and S.N.Z.; writing—original draft preparation and writing, B.B., E.F.R., M.R.W., and S.N.Z.; review and editing, E.A.S., S.E.F., J.G., and L.A.L. All authors have read and agreed to the published version of the manuscript and contributed substantially to the work reported.

**Funding:** All of the authors are on the leadership team of the Healthy Food Retail Working Group, jointly supported by Healthy Eating Research (HER), a national program of the Robert Wood Johnson Foundation (RWJF), and the Nutrition and Obesity Policy Research and Evaluation Network (NOPREN). NOPREN is supported by Cooperative Agreement No. 5U48DP00498–05 from the Centers for Disease Control and Prevention (CDC), Prevention Research Centers Program. All authors receive a stipend from HER for their leadership role with the working group. Support for MRW's effort was provided by the National Heart, Lung, and Blood Institute (NHLBI), grant number K99HL144824 (Principal Investigator: MRW). Publication fees were supported by Healthy Eating Research, a national program of the Robert Wood Johnson Foundation. The content is solely the responsibility of the authors and does not necessarily represent the official views of the HER, RWJF, NOPREN, CDC, or NHLBI.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **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*

## *Communication* **Healthy Food Retail during the COVID-19 Pandemic: Challenges and Future Directions**

#### **Lucia A. Leone 1,\*, Sheila Fleischhacker 2, Betsy Anderson-Steeves 3, Kaitlyn Harper 4, Megan Winkler 5, Elizabeth Racine 6, Barbara Baquero <sup>7</sup> and Joel Gittelsohn <sup>4</sup>**


Received: 7 September 2020; Accepted: 8 October 2020; Published: 11 October 2020

**Abstract:** Disparities in dietary behaviors have been directly linked to the food environment, including access to retail food outlets. The Coronavirus Disease of 2019 (COVID-19) pandemic has led to major changes in the distribution, sale, purchase, preparation, and consumption of food in the United States (US). This paper reflects on those changes and provides recommendations for research to understand the impact of the pandemic on the retail food environment (RFE) and consumer behavior. Using the Retail Food Environment and Customer Interaction Model, we describe the impact of COVID-19 in four key areas: (1) community, state, tribal, and federal policy; (2) retail actors, business models, and sources; (3) customer experiences; and (4) dietary intake. We discuss how previously existing vulnerabilities and inequalities based on race, ethnicity, class, and geographic location were worsened by the pandemic. We recommend approaches for building a more just and equitable RFE, including understanding the impacts of changing shopping behaviors and adaptations to federal nutrition assistance as well as how small food business can be made more sustainable. By better understanding the RFE adaptations that have characterized the COVID-19 pandemic, we hope to gain greater insight into how our food system can become more resilient in the future.

**Keywords:** retail food environment; food purchasing; federal nutrition assistance; COVID-19; grocery stores; restaurants; dietary intake

#### **1. Introduction**

In the United States (US), substantial socioeconomic and racial disparities exist in dietary behaviors [1]. Limited access to fresh food, coupled with a greater prevalence of fast food outlets in lower-income and minority neighborhoods, is partially responsible for sub-optimal eating patterns among residents [2]. The Coronavirus Disease of 2019 (COVID-19) pandemic has placed unprecedented strain on the US food system and changed the way food is distributed, sold, obtained, prepared, and consumed [3,4]. In the early weeks of the pandemic, grocery retailers saw overwhelming demand paired with panic buying resulting in empty shelves [5]. Restaurants have been temporarily closed in many communities and some have even permanently closed as they were unable to weather the financial burden of the temporary closures and/or the required additional pandemic safeguards [6]. To reduce risk of exposure, many consumers shifted to online food shopping and opted for curbside pickup or home delivery over entering retail stores [7]. Changes in consumer purchasing, coupled with government-mandated business closures, also negatively impacted food growers and producers. Due to the lack of demand from restaurants, there were reports of farmers who found it more economical to plow under crops and cull their herds [8–12].

The virus has disproportionality impacted food access for groups that already had higher rates of food insecurity (see Table 1) [13]. COVID-19 has also further exacerbated existing disparities as coping strategies (e.g., bulk purchasing, online ordering, food delivery) are largely unavailable to those with already limited food access [11,14]. Food insecurity disproportionately affects communities who have been historically oppressed, most notably communities of color, due to policies and structures obstructing access to affordable foods [15]. Individuals in these communities often do not have equal access to resources and are more likely to have lost jobs during the COVID-19 crisis, leading to a further increased risk of food insecurity [16]. Before the pandemic, 21% of Non-Hispanic Black households experienced food insecurity [17]; currently, that proportion is estimated at 38% and will likely continue to rise the longer the pandemic persists and during the resulting economic recovery [18]. Many communities have also been impacted by uprisings against police brutality and structural racism that may have damaged, disrupted, or destroyed food retail outlets and other infrastructure, creating even more food access issues [19].

**Table 1.** Food insecurity rates before and during COVID-19 for select population groups.


Factors contributing to food insecurity during COVID-19:

• Structural inequities regarding race and class

• Job loss

• Holding a low-wage job(s)

• Limited savings/access to credit

Despite the many possible effects of the pandemic on components of the retail food environment (RFE), no literature exists that explores these effects in a systematic manner, and then uses these findings to suggest and prioritize next steps. To address this gap, we used the Retail Food Environment and Customer Interaction Model developed by Winkler and colleagues (also in this issue) to describe the impact of COVID-19 on the US RFE in four key areas outlined by the model: (1) community, state, tribal, and federal policy; (2) retail actors, business models, and retail sources; (3) customer experiences in retail setting; and (4) customer dietary intakes. This new model is the first attempt to describe the role of RFE on diet, and the unprecedented change in RFE due to the pandemic allowed an opportunity to both test the new model and to systematically structure our paper. For the purposes of the model and this paper, the RFE includes food stores (grocery, supermarket), food service (restaurants, institutional food), and emergency food (food pantries, food banks). In an effort to build a stronger, more sustainable food system for the future, we also identified research priorities and strategic programmatic directions related to the RFE in the pandemic context.

#### **2. Community, State, Tribal, and Federal Policy A**ff**ecting the Retail Food Environment**

During the pandemic, a variety of macrolevel factors at the community, state, tribal, federal, and global levels have influenced the RFE and customers' behaviors. Table 2 lists a range of COVID-19-relevant US federal government responses, including policies, programs, and operational guidelines related to food distribution and donations, household food handling and eating out, federal nutrition assistance, and federal nutrition education and promotion. The key US government departments and agencies charged with RFE-related pandemic responses include the Department of Agriculture (USDA), the Department of Commerce, the Department of Health and Human Services, Centers for Disease Control and Prevention (CDC), the Food and Drug Administration (FDA), and the Department of Homeland Security, specifically the Federal Emergency Management Agency (FEMA). Policies targeting retailers included new guidance from the CDC related to safe operations, a Paycheck Protection Program (PPP) which provided forgivable business loans, and rapid dissemination and utilization of existing laws such as those that protect organizations donating food [22]. On the consumer side, the CDC published guidelines on food safety and running essential errands including food shopping. The USDA provided relief to many people struggling to afford food through the Farmers to Families Food Box Program [23]. For this program, the USDA contracted with food distributers and other retail actors to distribute excess farm products (which normally would have gone to restaurants) through emergency food channels. A variety of existing federal nutrition education and promotion materials have been disseminated, modified, or created during the pandemic, particularly around food safety [24–26]. The USDA denied waivers from several states to use SNAP Education (SNAP-Ed) funding to pay for staff to perform work for other federal programs such as school meal distribution [27]. SNAP-Ed is an evidence-based program that works to promote healthy eating at the community, state, and tribal levels by using policies, systems, and environmental supports, providing direct nutrition education, and supporting social marketing campaigns [28].

**Table 2.** Selected US federal government COVID-19 initiatives targeting the retail food environment and customers \*.

#### **Food Distribution and Donations**

*Retail Food Establishments*


*Charitable Food Network*


#### *Home Delivery*


#### *Export Services*

• US Department of Commerce announced temporary reductions in or eliminations of costs of several of their export services, which provides relief to US businesses and economic development organizations during this pandemic and encourages the promotion of foreign direct investment and the export of "Made in the USA" foods and beverages around the world during this economic depression

*State, Tribal, and Local Governments*

• FEMA public assistance grants, which could be utilized to support emergency food distribution during this pandemic

#### **Household Food Handling and Eating Out**


#### **Federal Nutrition Assistance—Local Access and Purchasing**


#### **Federal Nutrition Education and Promotion**

• USDA developed, modified, or created a variety of federal nutrition education and promotion materials during the pandemic, particularly around food safety and eating on a budget

Note: CDC = Centers for Disease Control and Prevention; USDA = United States Department of Agriculture; FEMA = Federal Emergency Management Agency; FDA = Food and Drug Administration; WIC = Special Supplemental Nutrition Assistance Program for Women, Infants and Children; P-EBT = Pandemic Electronic Benefits Transfer. \* Additional tribal, state, and local laws, along with retailer policies and practices, impacted the retail food environment during this pandemic and several other national and international responses impacted the broader food system.

Early in the pandemic, Congress made unprecedented short- and long-term changes to federal nutrition assistance [29]. A key change for the USDA Supplemental Nutrition Assistance Program (SNAP), which provides funding to supplement the food budgets of income-eligible individuals and families, was the expansion of online purchasing, which is now available in 45 states and the District of Columbia, impacting more than 90% of SNAP participants [30,31]. More work remains to expand the SNAP-authorized retailers beyond Walmart and Amazon and to ensure proper protections against predatory exposures to unhealthy food marketing [32–35]. Despite increasing access to online shopping, SNAP benefits have not yet been increased [32]. However, states could request waivers from the USDA to provide additional SNAP benefits through emergency allotments (up to USD 646 for a family of four) through Pandemic Electronic Benefits Transfer (P-EBT) for households with children who would normally receive free or reduced-price school meals (estimated USD 114 per child per month) [36]. The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) regularly provides supplemental foods, health care referrals, and nutrition education for income-eligible pregnant, breastfeeding, and non-breastfeeding postpartum women, and to infants and children up to age five who are found to be at nutritional risk. Online food payment is not currently possible with WIC benefits, though workarounds are available to order online with curbside payment and pickup [37,38]. A recent brief detailed key COVID-19 provisions to help optimize program impacts, including modernizing and streamlining WIC enrollment, extending eligibility for mothers and children, enhancing and expanding outreach, examining WIC food package flexibilities, scaling up nationwide best practices, and evaluating changes in breastfeeding practices during the pandemic [39].

Tribes, states, and local governments have played significant roles in shaping the RFE including creating food retail capacity and opening restrictions aimed at reducing the transmission of COVID-19. They have also increased technical assistance and communication regarding enrollment in new and existing federal nutrition assistance programs (e.g., P-EBT and Grab-n-Go meals), funding to support emergency feeding programs, and policies and resource allocations that support home delivery to vulnerable populations (e.g., state agencies covering delivery fees for SNAP online purchases) [40–42]. To offer Grab-n-Go meals to children during school closures, the USDA granted schools and other community sites flexibility to serve meals that do not meet the National School Lunch and School Breakfast programs' nutrition standards [43].

#### **3. Retail Sources**

COVID-19 has had multiple impacts on access to the places and the means by which people obtain food. Perhaps the most notable effects have come from restaurant closures. As about half of America's food dollars and a third of the food products produced in this country go to food service (food prepared away from home), including both restaurants and institutional food service (e.g., schools, hospital cafeterias), the closing of many of these venues significantly shifted both where Americans get their food and where food supplies are sent (or not sent). March 2020 spending on food away from home was 51% percent lower than it was in March of 2019 [44]. More cooking and eating at home has meant that other sources, like grocery stores and fresh food delivery services, have seen a surge in demand.

Since dollar stores and larger retailers like Target and Walmart sell groceries, they have been able to stay open when other retail outlets have been forced to close and may have seen increased sales as a result. Even as food sales have shifted from prepared food sources to retail food stores, shopping access has been limited by shorter store hours that allowed staff more time for cleaning or by designated shopping times for vulnerable community members (e.g., seniors, immunocompromised) [45]. Grocery workers protesting poor working conditions have also threatened to limit grocery store access [46].

During the initial onset of the pandemic, online ordering and sales, which allow for no or low-contact purchasing, surged with as many as 78.7% of consumers reporting having shopped online (compared with 39% pre-pandemic) [7]. However, regular online shopping rates have remained modest, with 33% of people reporting shopping online at least once a week compared to 27% in 2019 [47,48]. Prior to COVID-19, online grocery shopping rates were highest in the 30–44 age group with 28.3% reporting shopping for groceries online in 2019. This trend, mainly driven by families with children who desire convenience, has continued during COVID-19 [49]. Younger shoppers in general are more likely to embrace the technology needed to shop online than their older counterparts [50]; only 10% of baby boomers report that they will continue shopping for food online after the pandemic is over compared with 35–40% of younger shoppers [51].

Online shopping may also be disproportionately observed among wealthier, urban consumers as opposed to people who have limited income or live in rural communities, who may lack credit cards or reliable internet [52,53]. Federal nutrition assistance benefits are not accepted online at all (in the case of WIC) or only by select retailers (in the case of SNAP). On the business side, the switch to online sales may have left some small businesses behind. Larger supermarkets with existing online ordering capability have more easily adapted to the online order environment, while smaller stores without dedicated e-commerce platforms have been left scrambling to complete phone orders or create homespun website solutions [54].

For restaurants, many delivery services (e.g., Grubhub, Skip the Dishes) are expensive and can cut into already reduced margins [55]. Consumers may assume that restaurants that do not use these services are closed or they may merely overlook them. Local food purveyors such as farmers and mobile markets that directly sell produce and other foods to consumers have also struggled to adapt to online sales and home delivery as they may have limited staff for deliveries or are unable to take orders from customers with limited internet access [56,57]. Regardless of the food source (restaurant, grocery delivery, mobile market), delivery options may be limited, or totally unavailable to many rural consumers due to the extensive distances needed to travel to provide delivery in those areas, further exacerbating disparities in accessing food in rural communities [47].

Finally, food banks and food pantries have seen a huge surge as consumers who lost their jobs have turned to food aid. Emergency food is also an important source for those not eligible for federal nutrition assistance or those who fear the chilling effect of the Public Charge Rule, which threatens the legal status of immigrants who accept certain forms of government assistance [58,59]. At the same time, food banks and pantries were impacted by disruptions in the food supply as well as decreases in donations and volunteers, who tend to be older and more vulnerable to COVID-19.

#### **4. Retail Actors**

Important retail actors affected by the pandemic include the owners and managers of food retail and food distribution businesses that control what food is available where. The most notable pandemic challenges for food distributors has been that supply chains and products developed for food service and restaurants need to be adjusted to get food to grocery stores and other retail outlets (i.e., packaging sizes are different for stores vs. restaurants) [3]. For many distributors, switching from food service to food retail sales requires flexible packing lines and transportation channels as well as diverse distribution relationships which most of them do not currently have [3]. Without the ability to quickly readjust distribution channels, the shift in demand has led to temporary decreases in availability of many foods at retail outlets (see additional detail in Customer Retail Experience below). However, some smaller producers and distributors were able to transition to selling food directly to the public. In addition to shifting supply, there has been a redirection in the workforce as grocery stores have had to hire more employees and restaurants in turn have laid off workers. The pandemic has also precipitated an increase in "gig" economy positions including food delivery for popular restaurant and grocery delivery apps like Instacart and Grubhub [60]. Despite being labeled as essential workers, many food retail and delivery jobs have limited benefits such as hazard pay or sick leave and may leave these workers vulnerable to food insecurity themselves [61].

#### **5. Business Models**

Many retail food businesses have been forced to change and adapt their business models to both serve the needs of their communities and ensure their survival. In addition to the rapid expansion of online shopping and delivery services observed in grocery stores and restaurants, many retail food businesses have transitioned their products or target markets. For example, some restaurants have transitioned to mission-driven work (e.g., providing food for hospital workers or laid-off restaurant workers) and are relying on donations from customers to keep their businesses afloat [42]. Others have pivoted by creating prepare-at-home food or by opening a farm stand or mobile grocery store which has allowed them both to sell-off excess stock and take some of the burden off of over-crowded grocery stores [62]. Many restaurants have enhanced outdoor seating thanks to parking lots and sidewalks that were turned into make-shift patios.

While the large majority of grocery stores remained open with adjustments, some non-traditional retailers (mobile markets, farmers' markets, community-supported agriculture) initially shut down over fears of not being able to safely serve their customers, especially those who served primarily senior populations [44]. This has resulted in increased sales for grocery stores while 74% of farmers' markets say that they have lost income [63]. Other farmers and small markets have found ways to stay open by converting to pre-packaged foods (i.e., bundle or box models) that eliminate having customers touch the food or spend extra time making selections. Many small stores (corner stores, bodegas, etc.) in urban settings were already outfitted with plexiglass partitions and had existing practices that limited the number of clients entering the store; therefore, these small stores were paradoxically better prepared for the pandemic in some ways compared to larger food stores [64,65]. There has also been a rise in direct sales by producers including food manufacturers' bulk shipping of canned goods, snacks, and other shelf-stable items directly to customers [66]. Notably, in May 2020, PepsiCo launched Snacks.com and PantryShop.com to sell its products directly to consumers. On a smaller scale, farmers and fishermen started community-supported farm or fishery programs [62].

#### **6. Customer Retail Experience**

COVID-19 has had an unprecedented impact on all aspects of the customer retail experience, most notably availability and prices. Customers have reported having to visit multiple retailers to find desired foods and beverages or say they are not able to find the types of foods and beverages their families prefer [67]. These reductions in food availability disproportionately impact communities that already had reduced access to retail food sources like rural communities and communities of color [68]. Supply chain issues limited the availability of many foods and beverages, including infant formula, but empty grocery store shelves were more a result of consumer behavior [69]. Many consumers started stocking up on food, either because they were afraid they would not be able to find the items or wanted to limit the number of times they had to leave their home [70]. These customer behaviors inadvertently deepened inequities in food access. People with limited income do not generally have enough disposable income to make bulk purchases and may be limited in how much they can buy if they rely on public transit. Limited stock may force them to make multiple trips or rely on more expensive small stores closer to home. Although flexibilities were available in some states due to the pandemic, WIC participants are limited to shopping in WIC-approved stores, are generally limited to specific sizes and brands, and cannot substitute when a WIC-approved brand is sold out, essentially making them unable to redeem much-needed benefits.

The shift to more grocery shopping and at-home preparation has led to some price increases (Table 3). In April 2020, the cereal and bakery index saw the largest monthly price increase ever recorded by the Bureau of Labor Statistics (3.1%) [71]. In May 2020, consumer prices for meats and eggs rose 10% which was the largest 12-month percentage increase since 2004; these increases were partially due to changes in meat supply due to COVID outbreak-related closures in meat processing facilities [71]. In addition to higher prices, consumers have seen fewer promotions and advertisements during the pandemic as stores are already dealing with increased demand and want to decrease traffic. Some food companies even pledged not to offer any promotions for at least the initial pandemic months [72]. However, marketing may have shifted to different venues; Kraft Heinz, Kellogg's, and McDonald's were forced to temporarily cease advertising on online learning platforms after advocacy groups raised concerns over ads for unhealthy foods being advertised to children [73].


**Table 3.** Twelve-month percent change in the US Consumer Price Index for food at home Jan–May 2020.

Note: U S Bureau of Labor Statistics data for urban consumers [74].

#### **7. The Customer and Individual Dietary Intake**

The RFE changes described here have likely had profound and lasting effects on the shopping behavior and diets of customers. Customers have reduced in-person shopping frequency; only 20% of customers reported multiple shopping trips each week (down from 28% in 2019) [44]. Preliminary data collected during the pandemic indicates increases in cooking at home, following a diet, snacking, and eating plant-based foods [47]. Importantly, these effects appear to be differentially felt depending on the individual, interpersonal, and household level characteristics of each customer. Individuals or households who live in communities with greater access to a variety of food sources, and who have sufficient resources, steady employment, and credit are more able to adapt to the pandemic food environment. In addition to financial savings, the positive benefits of transitioning to fewer prepared foods may include consumption of more locally produced foods (especially local produce) and decreased reliance on ultraprocessed foods [75,76].

Changes in food intake are likely very different among the economically and nutritionally vulnerable members of our society. Individuals experiencing food insecurity are more likely to buy less expensive and less healthy foods and beverages, such as packaged and ultraprocessed items. A study of the early effects of COVID-19 indicated that 41% of food-insecure individuals reported buying fewer fresh items (i.e., milk, meat, fruits and vegetables) compared to 21% of food-secure individuals [77]. Food-insecure individuals are also more likely to need to access other components of the food system, including food pantries and free meal distribution sites, such as school meal distribution. Lower diet quality associated with food insecurity could potentially exacerbate already higher rates of diet-related diseases that resulted in part from historic structural barriers in food access [15,78]. This is especially problematic as diet-related diseases, such as hypertension, obesity, and diabetes, are associated with COVID-19 hospitalizations and higher mortality [79,80].

#### **8. Future Directions**

The COVID-19 pandemic has put unexpected strains on our nation's food system, upending many traditional food supply and access strategies employed by retailers and customers, and establishing new ones. It will not only be important to study the direct impact of changes to the RFE during the pandemic, but to look at the implications for building more resilient food systems following the pandemic. This work in particular should focus on our most under-resourced community members, including low-income communities and communities of color, whose access to healthy foods was already limited. Recommendations for further research at the consumer, retail, community, and policy levels are outlined.

Little is yet known about the effects of the pandemic on consumer behavior, including shopping frequency and the types of foods purchased. Fewer shopping trips due to pandemic exposure concerns may have led to less purchasing of fresh foods, such as fruits and vegetables [81]. It is likely that far greater quantities of ultraprocessed and long-term storage foods were purchased, but the longer-term impacts of these food system shifts on the diet of households and individuals are not known. Future research is needed to understand how changes in shopping patterns due to the

pandemic have affected what people buy and eat. The pandemic accelerated an existing trend towards online ordering of both groceries and prepared foods. Research on online food purchasing behaviors is still nascent and there is an urgent need to better understand who is buying online and from what types of sellers (i.e., retailers, manufactures). We also need to better understand the role of marketing and other behavioral economic factors, especially given the reports of increased surveillance and marketing of unhealthy foods to SNAP participants shopping online [32,33]. It is also critical that accessibility needs for underserved populations, including rural customers and SNAP and WIC recipients, be better understood and addressed.

The movement of food retail towards alternative (and in some instances, more community-minded) models combined with the closure of many restaurants underscores a need for research looking at business sustainability, especially in lower-income and minority communities. Food retail already operates on very slim margins and the US has some of the cheapest food in the world in relation to income with low costs coming at the expense of small farmers, food workers, and our environment. Research to understand what it will take to change this exploitive relationship is needed. This is particularly relevant as business interest expands outside profit to include public health, supporting local producers and economies, and simultaneously being able to advance equity through affordable pricing and paying a living wage. Outlining how private companies can meet these goals will help retail food actors move past the purely profit-driven model which has contributed to current health disparities. More diversity within business models and retail actors (e.g., minority and immigrant-owned businesses) who live and support the communities they serve may be one step towards this goal. Research is needed to understand how the effectiveness of existing policies and community programs (e.g., business incubators, healthy food financing, public procurement) can support business diversity at the local level. We also need more community-engaged research to understand the impact of food policy councils and other forms of community representation, particularly among Indigenous peoples and communities of color.

At the federal level, we need to advance our understanding of the impacts on families and retailers, among others, of the quick and large-scale expansion of federal nutrition assistance. While there has been an increase in SNAP enrollment and some SNAP benefits, current benefit levels are inadequate and legislative attention is needed to define, calculate, and provide adequate SNAP benefits, especially since increasing benefits has been shown to help stabilize the economy [32]. Understanding how flexibilities made to programs like WIC and SNAP in certain states affected program effectiveness can help with creating future resilience and understanding which adaptations should remain during "normal times" [13,82]. We also need to understand the impact of the USDA Farmers to Families Food Box program on both food security and the broader RFE. It is still unclear how this large influx of free food affected small retailers and distributors in lower-income communities.

In thinking about how we can build resilience in our food system and be better prepared for future crises, one possible food system adaptation may be referred to as "smaller is better." This would include enhancing and supporting local production and shorter supply chains [83]. Although more research is required, smaller enterprises may respond to market disruptions more effectively as they have more ability to shift to new markets and products as they gain insights from an engaged customer base (e.g., the Scale paradox) [84]. We see that when food supply chains are developed to only serve one type of food business (e.g., restaurants), it has ripple effects that head back to the source and negatively affect farmers and producers. Efforts to improve local food control and sovereignty have the potential to enhance food system resilience, but the format and impact of these efforts are just beginning to be explored [85].

#### **9. Conclusions**

The RFE and Customer Interaction Model provided a useful framework for outlining adaptations and research needs related to the US RFE during the COVID-19 pandemic. Using this model helped highlight vulnerabilities in our food system and future research needs. However, we emphasize that

#### *IJERPH* **2020**, *17*, 7397

many of the challenges that COVID-19 has brought to the forefront are not new, but instead are the result of a deepening of previously existing inequities, notably in communities of color. Using the model, we were able to identify potential strategies that could help build a more equitable RFE, which may not only benefit our country during normal times but could help build resilience against future pandemics or similar crises.

**Author Contributions:** Conceptualization, J.G., L.A.L., S.F. and B.A.-S.; investigation, J.G., L.A.L., S.F., B.A.-S. and K.H.; resources, B.B., M.W. E.R.; writing—original draft preparation, L.A.L., K.H., S.F., B.A.-S. and J.G.; writing—review and editing, all authors. All authors have read and agreed to the published version of the manuscript.

**Funding:** All authors are members of the Healthy Food Retail Working Group leadership team, jointly supported by Healthy Eating Research (HER), a national program of the Robert Wood Johnson Foundation, and the Nutrition and Obesity Policy Research and Evaluation Network (NOPREN). NOPREN is supported by Cooperative Agreement No. 5U48DP00498-05 from the Centers for Disease Control and Prevention (CDC), Prevention Research Centers (PRCs) Program. All authors receive a stipend from HER for their leadership role with the working group. Support for MRW's effort was also provided by the National Heart, Lung, and Blood Institute (NHLBI), grant number K99HL144824 (Principal Investigator: MRW). Publication fees were supported by Healthy Eating Research, a national program of the Robert Wood Johnson Foundation. The findings in this study are solely the responsibility of the authors and do not necessarily represent the official views of CDC, HER, or NOPREN.

**Acknowledgments:** We would like to thank Alyssa Moran for reviewing this paper and providing feedback.

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

#### **References**


© 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*
