Conceptualizing the commercial determinants of dietary behaviors associated with obesity: A systematic review using principles from critical interpretative synthesis

Abstract Introduction Unhealthy diet is an important preventable risk factor for overweight and obesity. Identifying the key drivers of an unhealthy diet is an important public health aim. “Big Food” has been identified as an influential factor shaping dietary behavior and obesity, and their practices have broadly been labeled as the “commercial determinants of obesity,” but there is a lack of definitions and conceptualizations for these terms. This review aimed to synthesize literature on the commercial determinants of dietary behavior associated with obesity. It presents the development of an integrative definition and a conceptual framework involving potential influences on dietary behavior, and it examines the prevalence of certain narratives within papers that focus on children and adolescents. Methods Four electronic databases (Ovid MEDLINE, PubMed, Web of Science, and Scopus) were searched up to December 2020. Eighty‐one articles met the inclusion criteria: they were published in a peer‐reviewed academic journal, described a practice from the food/beverage industry in relation to dietary behavior or obesity. Data were integrated using critical interpretative synthesis. Results The commercial determinants of dietary behavior are conceptualized in terms of three corporate spheres of action—political and legal; production, processing and design; and marketing and preference shaping—which enable powerful food industry to successfully pursue their business, market, and political objectives. The most frequently reported sphere of action targeting children and adolescents was marketing and preference shaping. Conclusions In the included literature, the commercial determinants of dietary behavior associated with obesity have been conceptualized as being part of a complex system where corporate practices are enabled by power structures. The proposed framework can facilitate a structured identification and systematic study of the impact of specific aspects of food industry's strategies and increase opportunities for primary prevention by anticipating industry responses and by discouraging corporate practices that harm health.


| INTRODUCTION
The global increase in obesity is associated with the increased availability and consumption of energy-dense, nutrient-poor foods and beverages, many of which are "ultra-processed." [1][2][3] A key contributing factor is the continuing expansion and concentration of power of transnational food and beverage corporations ("Big Food"/food industry). 4,5 About 75% of the global food sales include processed foods, for which the largest producers hold over a third of the global market. [6][7][8] This has led to an accelerated "nutrition transition" from more traditional diets to highly processed foods. 9,10 Many authors have suggested that food industry practices have shaped the eating environment and determined food choices 11 through food availability, pricing, social, and cultural desirability. [12][13][14] Factors related to the food system that promote obesogenic dietary behaviors have broadly been labeled as the "commercial determinants of obesity." 15,16 The term "commercial determinants of health" (CDoH) was first used by West and Marteau, 17 who defined it as: "Factors that infiuence health which stem from the profit motive." Millar proposed the term "corporate determinants of health" 18 to describe how companies can act in ways that benefit society, but also how they can have negative influences on population health. In 2016, Kisckbusch et al. 19 further defined the term as "strategies and approaches used by the private sector to promote products and choices that are detrimental to health" and conceptualized health outcomes as being determined by the influence of corporate activities on the cultural and social environments. In 2018, Madureira-Lima and Galea presented 20 and applied 21 a framework to map corporate practices and its impact on health, and conceptualized power as the vehicle through which corporations exert their influence on preference shaping and on the political, knowledge, legal, and extra-legal environments.
Although there is a developing discussion on the commercial determinants as drivers of ill-health, there has not been a comprehensive review that conceptualizes and defines these factors and the ways in which they can directly and indirectly influence dietary behavior and obesity.
Two systematic reviews 22,23 and an overview 24 of the CDoH were recently published, showing that corporations use market (i.e., commodities themselves and production practices) and nonmarket practices (e.g., extensive supply chains, corporate political activities) to sell their products and secure a favorable regulatory environment, 22,24 as well as that the role of commercial actors as drivers of ill-health are frequently obscured, understated or simply absent in the existing frameworks of the determinants of health. 23 The current systematic review extends the work in previous reviews by addressing four specific, focused aims: (1)

| METHODS
The Cochrane Handbook for Systematic Reviews of Interventions 25 guided the methodology for this review. Risk of bias was assessed using Cullerton et al.'s 26 key guiding principles for population health researchers working with food industry. Principles from critical interpretative synthesis (CIS) were used to guide the data synthesis 27 and iteratively refine the research questions while searching and selecting from the literature. CIS allows integrating and interpreting a substantial body of data from different types of research evidence and across multi-disciplinary fields into a coherent conceptual framework ("synthesising argument"), grounded in the concepts identified in the included articles. The review protocol was registered with PROSPERO, registration number CRD42019137363.

| Inclusion criteria
Articles had to fulfill the following criteria: published in a peerreviewed journal (including commentary pieces), books or book sections; written in English or Spanish; refer to humans; propose a definition, or describe a mechanism (e.g., influencing policymakers to maintain a business-friendly regulatory environment), framework, or practices (e.g., lobbying) from the food and/or beverage industry that relates to their commercial or corporate activities in relation to dietary behavior and/or its link with health or obesity; refer to food and beverage industry exclusively.

| Article screening
Database searches were managed using EndNote X9 and screened using Rayyan QCRI. One author (Y.C.U.) retrieved and screened titles for eligibility. Y.C.U. and a second reviewer (P.H.) screened titles and abstracts selected after the first title screening. Full text screening was done independently by the two reviewers. Reasons for exclusion were documented and discussed until agreement was reached.

| Data extraction
Data extraction was done by Y.C.U. and reviewed by Y.C.U., Z.T., R.J., and F.D.V. Data were extracted for author and year, article name, publication type, country/region, income level, field of study, population age group, mention of adolescents (10-19 years), health outcome, definition or mechanism, conflict of interest reported, and details for potential conflict of interest. To capture the whole range of definitions, key terms, and mechanisms that have been used in the academic literature, there was no restriction on age groups. Once the final list of included articles was selected, a subgroup of articles was created that focused on adolescence (10-19 years). For the purposes of this review, if the article referred to "youth" or "young people" without specifying age, they were considered to be ≤19 years old and were included as part of the "adolescent" group to identify the dominant narratives and most frequently mentioned practices targeting this age group.

| Quality assessment and risk of bias
Risk of bias and other aspects of study quality were not assessed since CIS recommends prioritizing relevant articles rather than selecting study types. 28 Conflicts of interest (CoI) reported by the authors were documented, discussed, and explored using Cullerton et al.'s guiding principles to help identify, prevent, and manage actual or perceived CoI. Even if authors reported not to have any CoIs or competing interests, these guiding principles still enabled critical appraisal of potential CoIs. For example, reporting not having competing interests but being employed by the food industry or by an industry funded organization could influence study findings. [29][30][31][32] If any potential CoIs were identified, findings from the study were critically appraised to identify if, for example, the study funding source or collaboration with food industry could have influenced the study results.

| Data synthesis
Guided by CIS 27 principles, data synthesis covered the following steps. Fragments of text that made reference to corporate activities in relation to dietary behavior and obesity were extracted and coded by the first author. A second researcher (Z.T.) independently doublecoded thirty percent of the included articles, and any discrepancies were discussed until agreement was reached. Codes that explained similar ideas were iteratively grouped into themes and subthemes.
Based on the themes and subthemes, authors developed higher-level conceptual themes (data-driven themes). The relationship between the data-driven themes, themes, and subthemes enabled the creation of the conceptual framework showing how the commercial determinants of diet and obesity operate.

| Search results
A PRISMA flow diagram ( Figure 1) documents the search, screening, and selection process of the 81 included articles.

| Descriptive information
The largest number of identified studies focused on high income countries (n = 37; 46%). The field of study was mainly focused on public health (n = 51; 63%) and health policy (n = 17; 21%), followed by nutrition (n = 3; 4%), law (n = 2; 2%), business (n = 2; 2%), anthropology (n = 1; 1%), and sociology (n = 1; 1%). Only 23 articles (28%) referred to adolescents (10-19 years) with the majority (n = 58; 72%) not specifying an age group. About half of the studies (n = 38; 47%) focused on obesity, while 32 articles (40%) focused on diet related noncommunicable diseases. Potential CoIs were found in four articles (5%), 33-36 but three of those took measures to explicitly manage these, 33,35,36 for example, limiting the involvement of the funder in any aspects of the project 33,36 ; explicitly reporting the nature of funding received from the food industry 33,35,36 ; and including findings that were unfavorable to the funder. 33,35,36 The conclusion drawn from these studies was that there were no CoIs since, even when being employed by the food industry or by an CHAVEZ-UGALDE ET AL. industry funded organization, the funding source or collaboration with the food industry should not have had an influence on study findings and results did not seem to enhance industry's reputation or influence over the evidence base of diet and obesity. In contrast, in one study 34 the author was employed, and the study was funded, by the food industry, but these were not reported as this being a potential CoIs, and study findings and results seemed to only enhance industry's reputation. A detailed table with descriptive information on the 81 included articles and details on CoI can be found in Tables S2 and S3. Extracted quotes of definitions and mechanisms found in the 81 articles included can be found in Table S4.

| A conceptual framework for the commercial determinants of dietary behavior and obesity
Three data-driven themes were developed and fell under the concept of "food industry's spheres of action": 1. political and legal; 2. production, processing and design; and 3. marketing and preference shaping.
The framework was developed by expanding on these three datadriven themes and resulted in 12 themes ("corporate strategies"), 26 subthemes ("corporate practices"), and 85 mechanisms. A table with details on the themes, subthemes, and mechanisms can be found in Table S5. A visual representation of the framework (i.e., the relationship between themes and subthemes) is presented in Figure 2

Strategy 1.1: Framing evidence and debate
Through this strategy, the food industry aims to frame the evidence and debate of diet and obesity as an issue of individual and societal choices and responsibilities. [37][38][39][40] This framing has the potential to shift the focus away from dietary behavior (e.g., emphasizing physical activity over diet and calorie intake on obesity) and limit the perception that policymakers have on the food industry's responsibility for the products they produce, promote, and sale, particularly ultra-processed foods which have been linked with excess calorie intake and weight gain 41 . This strategy is enacted by shaping narrative and debate of health and disease and through the creation of evidence.

Practice 1.1 (a): Shaping narrative and debate of health and disease.
This was a commonly reported practice which allowed food-related corporate interests to shift focus away from health and reframe regulatory efforts (e.g., soft drink taxes) as an issue of consumer rights and to highlight these efforts as a restriction to people's freedom of choice. 15,24,[42][43][44][45][46][47][48][49][50] (…) food industry selectively produces and disseminates information that would be beneficial to its activities, to infiuence public policies and public opinion in ways favourable to its companies 48

Practice 1.2 (a): Participating in foreign direct investments (FDIs).
Corporate economic and political power allows the global food industry to have unrestricted capital flows in emerging markets, grow through mergers, and joint ventures. This allows them to have an increased control over different levels of the food system (production, processing, distribution, retail). 20,55,[57][58][59][60] Transnational food companies powerfully shape the supply, demand, and consumption of food and beverage products. (...) Transnational food companies are moving quickly into markets in developing countries, using strategies such as foreign direct investment to increase production and sales. 58

Practice 1.2 (c): "Revolving doors" between regulatory agencies & food
and agriculture industries. There are national agencies whose remit includes setting the governance rules of food production, trade, and investment. Officials from these regulatory agencies are sometimes recruited from food industry and agribusinesses (or vice versa), and in some cases then move on to become lobbyists in favor of the food industry interests. 5 This becomes a "revolving door" between public and private sectors that gives key access to decision makers and valuable knowledge and relationships that allows them to shape governance systems. 5,21,39,[63][64][65] There is a long history of USDA leaders and leaders of other agencies being recruited from food and agriculture industries and then returning to businesses like lobbying firms when their government service ends. 5
Within this strategy, four corporate practices were identified.

Practice 1.3 (b) Constituency building. Through constituency
building, food industry seeks to get involved in the community, establish relationships with key stakeholders and highlight media and public opinion that support industry's position. 45,65,68 This practice includes philanthropic activities, promoting public-private partnerships and public relationships to recruit supporters and detract opposition. 50,65,68 Constituency building… attempts to infiuence public opinion and public policies and programmes. 48

Practice 1.3 (c):
Funding key stakeholders/opinion leaders. Giving financial incentives to key stakeholders and opinion leaders (e.g., election campaigns, health and nutrition organizations, opposition groups) creates a supportive environment for food industry activities and helps to maintain a business-friendly regulatory environment. 36  The production, processing, and design corporate sphere of action aims to optimize cost viability. To achieve this, corporations employ two corporate strategies (themes) and five practices (subthemes).

Strategy 2.1: Reducing processing/manufacturing costs
The food industry can reduce production costs using optimization practices (e.g., mass production and economies of scale), 64 reformulating and manufacturing products with low-cost ingredients that enhance palatability (e.g., fat, sugar, salt, caffeine). 57,59,67,85,86 Practice 2.1 (a): Optimizing food manufacture and processing.
Technological advancement and the usage of economies of scale has enabled massive manufacturing and processing of energy-dense/low nutritional value foods. 80,87,88 These foods are highly palatable, attractive to the consumer due to their convenience for purchase, and consumption and cheaper to produce. 81 Changing product recipes may be good brand protection but has little population dietary impact (…). They favor a technical approach to nutrition to justify the products they produce and sell. 57

Practice 2.1 (c): Increasing product appeal with low-cost ingredients.
Products high in fat, sugar and/or salt have a high sensory appeal, increase shelf-life, and by being cheap, generate large profit margins, especially with high-volume sales. These ingredients are commonly used in high proportions to manufacture energy-dense and ultraprocessed foods. 35,89 "Obesogenic" food companies maximize their profits by maintaining or increasing sales and prioritizing both types of addictive mechanism (e.g. "value deals" and addictive properties of sugar, salt, fat and caffeine on foods). 86

Strategy 2.2: Increasing market share
Increasing food industry's market share is both an outcome and a driver for commercial profit-making strategies and practices.
Increased corporate growth due to sales and profit-margins increases corporate power which allows continued market penetration in emerging markets and enables them to take advantage of cheaper production costs while continuously optimizing their production and processing costs. 24

Strategy 2.3: Agribusiness food/ingredient supply
The food value chain begins with the production input (i.e., materials for crop production and seeds), followed by farmers, growers, and agribusinesses that provide raw agricultural commodities. Therefore, the nutritional quality of the food environment is strongly influenced by the ingredients that the food and beverage industry use to manufacture their products. 69 Additionally, which and how much of these ingredients are produced is determined by regulations and targets set for agricultural production, economic performance, and competitiveness for agribusinesses. 5 -479 prioritize the production of commodities with higher productivity and that will generate higher profits. 79 Such is the case of soybeans and corn. 94 With increased productivity, the price of these commodities reduces. 79 Although this can increase farmers' profits momentarily, eventually, this surplus will result in a reduction in prices. This has two effects: first, food industry can have continuous access to cheap ingredients, making food manufacturers prioritize these ingredients over others; and second, the need to find new applications for corn, soy, and their by-products. Today, most ultra-processed food contains some form of corn or soy. 94 …the low cost of high calorie foods with little nutritional value is due, in part, to federal subsidies for production of corn and soybeans. 79

| Sphere of action 3: Marketing and preference shaping
The marketing and preference shaping corporate sphere of action aims to increase brand loyalty and enhance consumers' desire for their product. To achieve this, five corporate strategies (themes) and nine practices (subthemes) were identified.

Strategy 3.1: Promotion to increase brand awareness and visibility
To increase sales of their products, food industry needs to increase brand awareness and visibility in targeted populations. 68,81 This can be achieved through various channels and strategies including integrated marketing and advertising 73,88 ; corporate social responsibility (CSR), 68 sponsorship, and branding 18,73,95 ; and by creating publicpartnerships with key stakeholders, opinion leaders, and influential people that will promote their brand enabling wide visibility and reach. 15

Strategy 3.2: Influencing consumers' perceptions of products
Food industry aims to influence consumers behavioral motivations. 86 Resulting patterns of consumption are influenced by consumers' perception of products, beyond the product itself, and can be heavily influenced by commercial companies, involving developing a brand image that is linked to emotional triggers and convenience. 68,85,86 Practice 3.2 (a): Packaging products with "added value" claims.
Focusing on a particular nutrient and labeling it as an "added value" health claim, generating brand differentiation to distinguish one line of products from another, and adding toys and appealing characters can elicit consumers desire for these products.
Health claims allow for a description of the relationship between a food product and its role in disease prevention. Food labelling is a significant marketing tool because of its impact on consumer confidence in food quality and the role it plays in the general discourse of diet and health 52

Strategy 3.3: Creating brand loyalty
Creating brand loyalty is key for the food industry to ensure consumers consistently purchase their products, 46,85,93 and they are particularly interested in forging long-lasting relations with children and adolescents to ensure brand loyalty. 85

Strategy 3.4: Product placement and distribution
Food industry has benefited from neoliberal policies and have managed to attain global presence by making use of extended supply chains and distribution channels, and by making ultra-processed foods increasingly available, accessible, and convenient for consumption worldwide. This was a commonly reported strategy. 16,24,54,59,74,[82][83][84]91,101 Practice 3.4 (a): Ubiquitous presence of ultra-processed foods. Ultraprocessed foods can be found almost everywhere at any time in urban and in high-and middle-income countries. 74 This combined with industry's marketing practices contributes to excessive consumption of ultra-processed foods. 33,62,74,80,89 The combination of high levels of promotion, widespread availability and low prices of these products (...) overwhelmingly drive the behaviours in the direction of positive energy balance. 80

Strategy 3.5: Pricing
Corporations also shape our environments by establishing the pricing strategies for their products, for example, cheap selling price at point of purchase, 24,84 discounts, bundle deals, price promotions, and coupons and reward programs. 16,62,[73][74][75]81 Consumption patterns are strongly dependent on price. 102 consumer's purchasing behavior by delivering more product for the same amount of money, in turn influencing consumer's purchasing preference. 54,79,86,93,97,103 Food and beverage marketers' and fast-food restaurants' ongoing marketing and sales promotion efforts, such as value pricing, psychological pricing, quantity discounts, and combo deals, which undermine portion control and healthy food choices. 79 The second most reported sphere of action was production, processing, and design, in particular, increasing product appeal through low-cost/addictive ingredients.

| Development of an integrative definition for the commercial determinants of dietary behavior associated with obesity
The inductive process of merging overlapping concepts across the  Table 1):

| DISCUSSION
An integrative conceptualization of the commercial determinants of dietary behavior associated with obesity has been developed using principles from CIS. 27  There was overlap between some themes and subthemes in the proposed framework showing that some strategies and practices could also be categorized as part of other spheres of action. For example, CSR has been theorized to be part of corporate political activities to advance corporations' interests in terms of regulation 65 ; however, CSR can also be seen as part of a marketing strategy to raise brand awareness in targeted populations. 68 Equally, product reformulation is part of the production, processing, and design sphere T A B L E 1 Aims and levels of influence of the three spheres of action  The frequency of mention of a themes or subthemes was not assessed as an indicative of a sphere of action or strategy's importance or the size of its impact in the overall food system. It does not necessarily reflect the strength of the evidence but may reflect the attractiveness of the topic for researchers and funding, the different specialized fields, or difficulties in accessing data.
Data analysis and thus the definition and framework proposed incorporated expert-opinion pieces and arguments supported by research data. A limitation of this approach is that the veracity of statements in commentary pieces could not be tested. However, expert-opinion pieces can provide valuable insight into issues of broad concern in global health, particularly those concerning policy issues. 109 Nonetheless, the sections in the framework that are dominated by opinion-based arguments should be tested empirically to make this framework fully evidence-based.