Access to Healthy Wheat and Maize Processed Foods in Mexico City: Comparisons across Socioeconomic Areas and Store Types

The contributions of processed foods to the overweight and obesity problem in Latin America are well known. Engagement with the private and public sectors on possible solutions requires deeper insights into where and how these products are sold and the related implications for diet quality. This article characterizes the diversity of wheat and maize processed foods (WMPFs) available to consumers in Mexico City. Data were gathered across nine product categories at different points of sale (supermarkets, small grocery stores, convenience stores) in high and low socioeconomic (SE) areas. We assessed WMPFs based on Nutri-Score profile, price, and health and nutrition claims. Roughly 17.4% of the WMPFs were considered healthy, of which 62.2% were pastas and breads. Availability of healthy WMPFs was scarce in most stores, particularly in convenience stores Compared to supermarkets in the low SE area, those in the high SE area exhibited greater variety in access to healthy WMPFs across all product categories. In the low SE area, healthy WMPFs were priced 16–69% lower than unhealthy WMPFs across product categories. The extensive variety of unhealthy WMPFs, the limited stock of healthy WMPFs in most retail outlets, and the confusing health and nutrition claims on packaging make it difficult for urban consumers to find and choose healthy WMPFs.


Introduction
Food processors and food retailers target specific products to specific subgroups of consumers in response to heterogeneity in economic, demographic, and behavioral factors. In providing food products tailored to meet consumer demands [1][2][3], segmentation strategies potentially deliver higher sales and profits for processors and retailers. In North America and Europe, research has supported these strategies by shedding light on consumer-related trends and issues related to, for example, perceptions on GMObased foods [4], attitudes towards nutrition [5], food safety and quality (including organic and locally produced) [6], and acceptance of new food and agricultural products [7][8][9]. Discussions have explored potential solutions to address the private sector's targeting of poorer areas with less-healthy and lower-cost foods, including the promotion of farmers' markets in lower-income areas [10,11]. However, limited research has examined how the food industry in developing and emerging economies targets different products to contributed to the rapid increase in obesity occurring in Mexico and elsewhere in Latin America [29]. In Mexico, overweight and obesity are present in 35.6% of school-aged children, 38.4% of adolescents, and 76.8% and 73% of adult women and men [30], and lowering these levels is a major public health priority. Among foods produced in Mexico, WMPFs represent roughly 30.7% of total food processing sector sales [31]. Maize tortillas; maize products (20.6%); and sweet, white, and wheat breads (11.5%) account for 32.1% of total energy intake in the Mexican diet [32]. Ultraprocessed foods based on wheat and maize ingredients (i.e., cookies, pastries, salty snacks, and ready-to-eat cereals) comprise a major component of the Mexican diet (16.2% of total energy intake) [33]. A research group at the National Institute of Public Health of Mexico recently reported that maize tortillas and tortilla products contributed 20.6% of total energy intake in the diet in the Mexican population, while fruits and vegetables, by comparison, contributed only 5.6% [32]. This study seeks insights into the following two questions: (1) What are the characteristics of WMPFs in terms of nutritional quality, health and nutrition claims, and price? (2) How do these vary by levels of socioeconomic attainment and points of sale? The findings here will allow for insights on the nutritional challenge and possible entry points for future engagement with public and private sectors on potential options for promoting increased access to healthy WMPFs for different types of urban consumers.

Study Design and Sampling
This cross-sectional study presents data on access to WMPFs collected in Mexico City and its surrounding area between March and June 2019. Principal component analysis was used to characterize "basic geographical areas" (BGAs)-administrative units in Mex- This study seeks insights into the following two questions: (1) What are the characteristics of WMPFs in terms of nutritional quality, health and nutrition claims, and price? (2) How do these vary by levels of socioeconomic attainment and points of sale? The findings here will allow for insights on the nutritional challenge and possible entry points for future engagement with public and private sectors on potential options for promoting increased access to healthy WMPFs for different types of urban consumers.

Study Design and Sampling
This cross-sectional study presents data on access to WMPFs collected in Mexico City and its surrounding area between March and June 2019. Principal component analysis was used to characterize "basic geographical areas" (BGAs)-administrative units in Mexico City for which data were available. The analysis included 14 variables on socioeconomic (SE) characteristics: housing materials and access to public utilities (e.g., floor type electricity, running water), material goods (e.g., ownership of televisions and refrigerators), and population characteristics (e.g., years of education, access to health care services, and employment). BGA-level data and information on frequency and location of stores were obtained from the National Statistic Directory of Economic Units and the National Population and Household Census 2010. Those BGAs which did not contain at least two supermarkets and two independently owned grocery stores were excluded. BGAs were categorized into low and high SE BGAs using their first and last deciles of the first principal component of SE characteristics. These criteria resulted in eight high SE areas and eight low SE areas in Mexico City and its surroundings; i.e., the high SE was in the Polanco area, the low SE was in the Chimalhuacán area. Additionally, field reconnaissance of each selected area was conducted to assess diversity in the type of food retailer and ensure the minimum number of food retail outlets, as well as their distribution, for the final selection of the two areas for the study.
Four types of food retail outlets were considered for each area: chain supermarkets, independently owned grocery stores, convenience stores, and corner stores. Trained personnel visited each store in a selected SE area and included adjacent BGAs if a specific type of food retailer was not accessible to the data collection team. Twenty-seven food retail outlets were confirmed within each SE area. We were able to increase the final sample to 2 convenience stores (which were not originally included), 2 neighborhood corner stores, and 20 independently owned grocery stores in both areas and 3 supermarkets in the high SE area and 2 supermarkets in the low SE area. Most low SE areas did not contain a chain supermarket, thus requiring adjustments in the field to ensure adequate supermarket coverage.

Data Collection and Management
We distinguished nine categories of WMPFs for inclusion in the study: ready-to-eat foods, breakfast cereals, salty snacks, cookies, flour, breads, bars and pastries, pastas, and maize tortillas and tortilla products. Within each category, we considered only packaged and labeled products, regardless of origin, which listed wheat or maize among the first three ingredients. We excluded nonpackaged products from traditional stores such as bakeries and tortillerias. Data were collected by taking photographs of information available on each product in the selected product categories at the store. A standardized methodology [34] was applied for the assessment of each product where a set of photographs captured the following: front of package, content, serving size, number of servings, bar code, nutritional information, ingredients, health and nutrition claims, and price.
Each product was classified according to its Nutri-Score profile [35]. The score considers calories, nutrients, and ingredients. Based on the score, each food is classified into one of five Nutri-Score profiles (from the healthiest to the least healthy): A (score ≤ −1), B (score 0-2), C (score 3-10), D (score [11][12][13][14][15][16][17][18], and E (score ≥ 19). Nutrients and ingredients that negatively impact health include calories, sugars, saturated fats, and sodium, while those with a positive impact include fruits, vegetables, fiber, and protein. Negative nutrients and ingredients can take 10 points each while positive ones can add 5 points each [35] As a result, the healthiest products should not only contain nutrients and ingredients with a potentially positive impact on health but also be free of nutrients and ingredients with a potentially negative impact on health. The higher extent to which a product contains more nutrients and ingredients with a potentially negative impact on health and/or lacks nutrients and ingredients with a potentially positive impact on health, the more likely it will be classified as "unhealthy" (i.e., group C, D or E).
Prices obtained from each food category were converted from Mexican pesos to US dollars (USD) by using the exchange rate during the collection period (USD 1 to MXN 19.15) [36]. Prices of each WMPF were scaled by 100 g of product and by package.

Ethical Considerations
The study protocol and procedures were approved by the Ethics Committee of the International Maize and Wheat Improvement Center (CIMMYT). Consent to conduct in-store data collection was obtained by email or in person from store representatives.

Statistical Analysis
We described the diversity of access to WMPFs by SE area and the type of store with the percentage of distinct WMPFs that were found in both SE areas, in low SE areas exclusively, or in high SE areas exclusively. The same analysis was used to describe the availability of WMPFs by Nutri-Score and food product categories by SE area including all products in the sample.
We applied a specified a multiple linear regression model for analyzing price differences between SE areas. The log of price of WMPFs per 100 g was specified as the dependent variable. Explanatory variables included three categorical variables and all their two and three way interactions, these variables were: SE area (low/high), Nutri-Score profile (coded as healthy/unhealthy) and food group. We obtained geometric means for each combination of SE area, Nutri-Score profile, and food group category, as well as geometric mean ratios between SE areas, by exponentiating the corresponding linear combination of model coefficients. We derived 95% confidence intervals by back transforming the confidence limits of the linear combinations. Additionally, we fitted a separate model for each type of retail store with the log of price per package as a dependent variable and the same specification of predictors as previously described and obtained the aforementioned differences by SE area and their confidence intervals. We reported the use of health and nutrition claims as percentages for each combination of food group category, SE area, and Nutri-Score profile, with their corresponding logit-transformed 95% confidence intervals [37], and calculated their differences in percentage points by SE area. Categories with less than 5 observations were excluded from analyses. Robust standard errors from all models were adjusted for data clustering within retail stores [38]. All analyses were performed in Stata (version 14).

Food Supply and Stock of Healthy WMPFs across Socioeconomic (SE) Areas and Food Retail Outlets
We recorded a total of 7567 WMPFs available in both SE areas, including unique and repeated products. Of these, 2462 were unique WMPFs. Among unique WMPFs, only 28.8% of them were classified as healthy (Table 1). Supermarkets offered the greatest variety of healthy products while convenience stores showed the least (614 vs. 29). However, the number of healthy WMPFs offered in supermarkets and convenience stores varied greatly between SE areas: 79.3% of unique WMPFs in supermarkets were available exclusively in the high SE area whereas 8.6% were available exclusively in the low SE area. In convenience stores, 44.8% of healthy WMPFs were in the high SE area exclusively, 3.5% were in the low SE area exclusively, and 51.7% were in both areas. Cookies, salty snacks, pastas, and bars and pastries were the WMPF categories with the greatest variety (636, 387, 290, and 258 products, respectively). Breakfast cereals, tortillas and tortilla products, flour, and breads were the WMPF categories with the least variety of products (203, 157, 155, and 154 products, respectively). Independently owned grocery stores provided the most homogeneous set of WMPFs in both low and high SE areas (from 86.1% to 100% similarity within WMPF categories). Across WMPF categories, cereal bars and pastries (38.7%), salty snacks (36.7%), and breakfast cereals (29.1%) were more frequently found in both SE areas.

Availability of WMPFs across Socioeconomic (SE) Areas by Nutri-Score Profile
Looking across the sample, only 11.7% and 5.6% of available WMPFs were included in the A and B Nutri-Score profiles, respectively. We found considerably greater availability for healthy WMPFs (i.e., Nutri-Score profiles A and B) in the high SE area (Table 2). Indeed, healthy items in the cookies and bars and pastries categories were only available to consumers in the high SE area. Consumers in the low SE area had basically no nearby access to healthy (i.e., included in groups A and B) breads, bars, and pastries. Unhealthy WMPFs (i.e., Nutri-Score profiles C, D, and E) were widely available in both SE areas.  1 Nutri-Score profile is a scale that divides food products into 5 classes (A to E) to distinguish healthier foods (A score) from the less healthy (E score). 2 Based on the Nutri-Score profile, includes all those products which had a score ≤ −1. 3 Based on the Nutri-Score profile, includes all those products which had a score between 0 and 2. 4 Based on the Nutri-Score profile, includes all those products which had a score between 3 and 10. 5 Based on the Nutri-Score profile, includes all those products which had a score between 11 and 18. 6 Based on the Nutri-Score profile, includes all those products which had a score ≥ 19.

Price Variation of WMPFs across Socioeconomic (SE) Areas and Retail Outlets
The geometric mean (GM) prices per 100 g and per package of healthy and unhealthy WMPFs food profiles were significantly higher in the high SE area compared to the low SE area (Table 3). For example, GM price per 100 g of healthy cookies was roughly 3-fold in the high SE area compared to the low SE area (GM ratio = 3.01; 95%CI: 2.15, 4.21), for unhealthy cookies GM price per 100 g was about 50% higher in the high SE area compared to the low SE area (GM ratio = 1.53; 95%CI: 1.18, 1.99). Within the low SE area, GM price per 100 g for healthy WMPFs was 16-69% lower compared to GM price per 100 g for unhealthy WMPFs (e.g., USD 0.20 95%CI: 0.018, 0.23; USD 0.64 95%CI: 0.59, 0.71, for total WMPFs, respectively). Supermarkets charged the highest prices for healthy and unhealthy WMPFs per package across all retail outlets and WMPF categories in both SE areas (Table 4). GM prices in supermarkets for healthy and unhealthy WMPFs were 165% (GM ratio: 2.56; 95%CI: 1.41, 4.97) and 86% (GM ratio: 1.86; 95%CI: 1.42, 2.44) more expensive in the high SE area compared to the low SE area. GM price differences for all WMPFs in other food retailers were smaller and most of them not significant between SE areas.

Health and Nutrition Claims Displayed in WMPFs across Socioeconomic (SE) Areas
Health and nutrition claims were found on 23.2% of all WMPFs. They were more frequently found on healthy products in the high SE area compared with the low SE area for cookies (+32.9 p.p., 95%CI: 4.6, 61.3) and maize tortillas and tortilla products (+32.7 p.p., 95%CI: 9.9, 55.6) ( Table 5). A considerable proportion of unhealthy WMPFs employed health and nutrition claims, particularly in the high SE area (20.1%, 95%CI: 14.8, 26.5). Unhealthy WMPF categories where health and nutrition claims were used in both SE areas included breakfast cereals (>45%) and flour (>25%). In the high SE area, 40.4% (95%CI: 30.9, 50.5) of unhealthy maize tortillas and tortilla products displayed health and nutrition claims.    1 Classification based on the combination of Nutri-Score profiles A and B. 2 Classification based on the combination of Nutri-Score profiles C, D, and E. 3 Geometric mean price ratio between high and low SE, reference is low SE. Estimates do not exactly match the ratios from shown geometric means because of rounding.

Discussion
In this study, we collected data on the diversity of WMPFs available to consumers in Mexico City paying attention to variation in access by socioeconomic areas, retail outlets, nutrient profile, and health and nutrition claims employed. The study was motivated by the need for data on what WMPFs were accessible to consumers, as an input for future discussions with the private sector on their role in marketing available healthy and nutritious WMPFs for urban consumers. Our findings highlighted various issues in the retail landscape for WMPFs that have important implications on the dietary options available to consumers, especially those in lower-income neighborhoods.

Most WMPFs Are Not Healthy
Cereals, including wheat and maize, can contribute essential amino acids, minerals, and vitamins and beneficial phytochemicals and dietary fiber components to the human diet. Although wheat and maize have traditionally been processed for human consumption, current WMPFs normally exceeded the daily recommended intake of sodium, saturated fat, trans fats, and added sugars established by WHO to prevent overweight, obesity, and chronic diseases [33,39], and their consumption is associated with lower-quality nutrient profiles [40,41] and often leads to unhealthy diets and overall poorer health conditions [42]. We found a majority and considerably greater variety of unhealthy WMPFs in both SE areas and across all food retail outlets. WHO's current global plan is a 30% reduction in salt/sodium population intake along with policy measures that engage food retail outlets to improve the availability of healthier products with reduced content of salt/sodium, saturated fats, and free sugars [43]. Our findings showed the urgent need for more progress towards these goals, as only 11.7% and 5.6% of available WMPFs were from the A and B categories of the Nutri-Score in the whole sample. The current need for food systems transformation for improved nutrition and health is alarming [44].

Consumers in Richer Areas Have Access to a Greater Diversity of WMPFs
The high SE area had much more diversity in terms of WMPFs available as compared to the low SE area. Previous studies had found that in-store healthy foods were less likely to be found in low-income neighborhoods in cities, and our findings were consistent with these [16][17][18][19]. However, in our study, the high SE area had the most diversity of not only in-store healthy WMPFs but also unhealthy ones. As a result, both areas offered limited options for healthy WMPFs, restricted to few WMPF categories. A recent panel survey analysis of purchase and intake data in urban areas in Mexico showed that higher SE groups from urban areas had greater purchases and intakes of less-healthy foods than lower SE households [45]. Our findings are consistent and further suggest that consumers in the high SE area have access to and probably purchase a greater variety of healthy products, but also unhealthy ones, than lower SE households.
Mexico's 8% sales tax on unhealthy foods resulted in a 5.8% decline in the purchase of taxed foods among households of medium SE status and a 10.2% decline among households of lower SE status [46]; thus, impact on high SE has not been measured or acknowledged as a risk, especially because complex urban surroundings such as those of Mexico City imply people moving around different areas to access their foods, and food supply in high SE areas may be seen as a model to be followed or symbol of status, setting the example of procurement for vendors stocking WMPFs in other SE areas.

Specialization in Retail Stores across Socioeconomic Areas
We found specialization in retail stores, with convenience stores and neighborhood corner stores specializing in WMPFs with Nutri-Score profiles D and E (unhealthy) WMPFs. Supermarkets had the greater stock of products with profiles A and B (healthy groups). In the low SE area, neighborhood corner stores and supermarkets could provide, in theory, access to healthier WMPFs; however, supermarkets are scarce in low SE areas. Similarly, in high SE areas, neighborhood corner stores and supermarkets could provide greater access to healthier WMPFs, but neighborhood corner stores, as well as independently owned grocery stores, tended to be scarce. Yet, increasing in-store healthy products in small food retail outlets is feasible and can be achieved with support and understanding of healthy product criteria [47] or incentives [48].
Interventions to improve healthy food offerings in small stores should consider the diverse environments, stocking practices, and supply mechanisms with innovative distribution practices of small stores, particularly nontraditional food retailers [49]. For example, retail outlets in the low SE area could stock cookies, bars and pastries, and breads classified in the healthy categories (A and B groups) and provide consumers access to them.

Higher Prices Prevail for Unhealthy WMPFs
The consumption of healthy food is mostly consistently related to income, contrary to the intake of unhealthy ones [50]. However, we found, in general, that prices per 100 g and per package of unhealthy WMPF products were higher than prices for healthy products in most WMPF categories, but prices per package differed by SE area. By package, healthy products tended to be more expensive, particularly in the high SE area. In the low SE area, for some WMPF categories, unhealthy products were more expensive than healthy ones. It is possible that expenditure elasticity differed by SE area, and higher elasticity is not always expected for healthy products when healthy foods are considered a necessity and unhealthy foods a luxury [51]. In addition, higher prices for the same WMPF category in the higher SE area likely reflect higher operational costs, while larger stores such as supermarkets may also have the marketing savvy to better segment the market as compared to smaller stores.
Although prices varied greatly across WMPF categories (as expected), higher prices for some foods also may predict lower consumption of such groups, as has been previously demonstrated in several countries [52]. While higher prices may be positive for some foods, such as the ones with added sugars as a result of the sugar tax policy, for example [53][54][55], they may have a negative impact for foods with positive ingredients and nutrients for health and nutrition outcomes [52].

Health and Nutrition Claims Frequent in Healthy and Unhealthy WMPFs
The use of health and nutrition claims can inform consumers through short and clear messages about the energy and nutrient content of a product and its potential health benefits [56]. However, this strategy has been regulated in some countries due to its public health implications [57]. First, consumers perceive products displaying health and nutrition claims as healthier (health halo effect) without considering nutritional quality, and this influences their preferences and purchases [58]. Second, the regulation of these claims is limited and generally independent of nutritional quality [59]. Our results show that unhealthy WMPFs also displayed health and nutrition claims on the front-of-pack label; this suggested that consumers who make their food choices based on the information available on the label were vulnerable to misperception about the nutritional quality of these products, mainly for less-healthy groups such as breakfast cereals. Generally, WMPFs such as cookies, breads, and pastries are fortified in products that use wheat flour (mandatory regulation in Mexico). Some studies have shown the use of these practices in less-healthy foods to promote the use of health and nutrition claims [60]. Improving front-of-pack labeling regulations that include specifications for the use of health and nutrition claims contributes to helping consumers make healthy food choices, particularly in low SE areas.

From Traditional Healthy Foods to Unhealthy WMPFs
Some WMPF categories included in our study comprised traditional food ingredients and foods such as flour, tortilla, bread, and pastas; others comprised other nonbasic food commodities such as snacks and ready-to-eat foods, which have become part of the modern diet in Mexico. The WMPFs from the first categories tended to be lower in price and of better nutritional value in both SE areas. Maintaining healthy properties and ingredients of WMPFs, including dietary fiber, prebiotics [61], resistant starch content, low glycemic index, and micronutrient content, as well as reducing ingredients with a negative impact on health, is essential for WMPFs to contribute to a healthy diet, particularly among the poorer segments of the population, given that they derive a larger share of their energy from staple foods [62]. Moreover, high consumption levels of low-cost maize flour and cornmeal, especially in areas where micronutrient deficiencies exist, make this food a vehicle for fortification [63].

Conclusions
Healthy WMPFs have an important role to play in healthy diets [64], given their relatively low cost and concentrations of energy and essential nutrients. Transformation to healthy diets will require a greater than 50% reduction in unhealthy foods [61]. However, as highlighted in this study, the current urban retail environment is not conducive to rapid change towards greater consumption of healthy WMPFs. The extensive variety of unhealthy WMPFs available, the limited stock of healthy WMPFs in most types of retail outlets, and the confusing health and nutrition claims on packaging make it difficult to find and choose healthy WMPFs.
The findings raise important questions about the private sector's formulation and marketing of WMPFs and the implications for promoting healthier diets. In many developing regions, income growth will continue to push demand towards more processed foods and more foods purchased at supermarkets. The Food Systems Summit in 2021 stated that "the food choices people are faced with and the choices they make are profoundly determined by the food system of which they are part" [65], and adding other domains of the food environment to the marketing strategies used can be a win-win for the food processors, vendors, and food choices we have in Mexico City.
This study represents a much-needed initial step to initiate dialogue between researchers, the private sector, and government agencies on how to improve access to healthier WMPFs. It highlights issues of differential access by consumers from higher and lower socioeconomic strata to healthier processed foods.
Strengths in our study include the careful selection of the high and low socioeconomic areas (i.e., first and last deciles of the socioeconomic areas within Mexico City and surroundings) and the variety of food retail outlets visited. To the best of our knowledge, no other study has included all available food retail outlets that comply with our inclusion criteria in a given area. Future work in this area should look to engage the consumers, private sector, government agencies, and civil society on options for increasing access to and purchases of healthy WMPFs. Funding: This work was undertaken as part of the CGIAR Research Programs on Wheat and Agriculture for Nutrition and Health (A4NH). We thank PIM for funding the study and all donors who support the Wheat and A4NH programs through their contributions to the CGIAR Fund.
Institutional Review Board Statement: The study was approved by the Internal Research Ethics Committee IREC.2019.001 and was waived from further revisions.

Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the research team continuing with further analyses of this study.