Unhealthy food availability, prominence and promotion in a representative sample of supermarkets in Flanders (Belgium): a detailed assessment

Introduction The supermarket food environment is a key setting for potential public health interventions. This study assessed food availability, prominence and promotion in a representative sample of supermarkets in Flanders (Belgium). Methods A sample of 55 supermarkets across five chains and 16 Flemish municipalities was selected in 2022, about 64% in the most deprived socioeconomic areas. Healthiness indicators related to food availability (ratio of cumulative linear shelf length for healthy versus unhealthy foods), prominence (proportion of unhealthy foods at check-outs and end-of-aisle endcaps), and promotion (food marketing on food packages) were measured. Results Overall, the average ratio of healthy/unhealthy foods in supermarkets in Flanders was 0.36, meaning that for every 10m of shelf length of unhealthy foods there was 3.6m of healthy foods. There was a large variation in ratio’s across supermarket chains. Of all foods available, 97.5% were ultra-processed at the check outs, while 72.2% and 58.5% were ultra-processed at the front and back end-of-aisle end-caps, respectively. Confectionery and sweet biscuits were the food categories with on average the highest number of marketing messages on pack per 10m of shelf length. Conclusion Supermarket in-store food environments in Flanders were found generally unhealthy, with those located in low income areas having unhealthier in-store food environments than supermarkets located in medium and high income areas. Despite commitments of all large supermarket chains in Flanders to promote and create healthier in-store food environments, our findings indicate that currently consumers are incentivized to buy unhealthy rather than healthy food products. Supplementary Information The online version contains supplementary material available at 10.1186/s13690-023-01175-3.


Introduction
Overweight and obesity rates are increasing across Europe [1].According to the World Health Organisation (WHO), 59% of adults are currently living with overweight or obesity across the European region [2,3].In Belgium the prevalence of overweight was 49% and the prevalence of obesity was 16% in 2018, a significant increase since the first health survey in 1997 (when the prevalence of overweight and obesity were 41% and 11%, respectively).In Flanders the prevalence was slightly lower, about 48% and 15%, respectively in 2018 [4].
One of the main drivers of overweight and obesity are unhealthy obesogenic food environments [5].These environments have been described as "the collective physical, economic, policy and socio-cultural surroundings, opportunities and conditions that influence people's food and beverage choices and nutritional status" [6].Internationally, there is growing interest in improving the healthiness of food environments in order to improve population diets.To date, food environment policies have mainly focussed on food reformulation, fiscal policies (e.g.taxation of sugar sweetened beverages), healthrelated front-of-pack labelling and improving in-school food environments [7].However, an additional important setting for potential public health interventions are the supermarket instore food environments, defined as the environments that consumers encounter when buying foods, including the cost, quality, and availability of foods [8].In Europe, as well as in Belgium, supermarkets are the most important retailers when it comes to grocery shopping [9].The five leading Belgian retailers (Colruyt, Carrefour, Delhaize, Aldi and Lidl) have a market share of 25% for the sales of packaged foods and 9% for the sales of non-alcoholic beverages for their own-brand products according to Euromonitor data 2018 [10].Finally, the fact that 59% of the market share is in the hands of those five chains shows that a limited number of supermarkets could influence the purchasing behaviour of a significant proportion of the population through (in-store) health promotion interventions [11].
Previous research already evaluated the nutritionrelated commitments of the leading Belgian supermarket chains, the healthiness of their own-brand product portfolios and the food products promoted in their circulars [12,13].Results from these studies showed that supermarkets made several commitments regarding the reformulation of their own-brand product portfolio, front-of pack nutrition labelling and limiting marketing toward children.Commitments to improve the healthiness of the instore food environments however, such as dedicating a minimum amount of shelf space to healthy products or limiting the placement of unhealthy products in high prominence areas (e.g.front end-of-aisle endcaps and cash registers), were however often lacking from existing pledges [12].It has also been shown that median product portfolios of the leading Belgian supermarkets (own-brand products) consisted for 49% of ultra-processed food products, for 71% of foods not-permitted to be marketed to children and for 41% of products with a Nutri-Score D or E [12].In addition, foods promoted across circulars consisted for 52% of ultraprocessed products on average, with substantial variations across supermarket chains [13].
Nonetheless, to date, no research in Belgium has assessed the in-store food environments of the biggest supermarket chains.Yet, because of their central role within the Belgian grocery landscape, their in-store food environments present a major opportunity to influence the food choices of a noteworthy proportion of the population.
As such this study aims to, for the first time, assess instore food availability, prominence and promotion in a representative sample of supermarkets in Flanders, Belgium.

Methods
This is a cross sectional study.Ethics approval for this study was obtained from the Human Participants Ethics Committee of the University of Ghent (reference number ONZ-2022-0138).Informed consent and cooperation for the study was gained from the head offices of the five major Belgian retailers: Delhaize, Colruyt, Carrefour, Lidl and Aldi.Store managers were called by telephone beforehand to discuss the time of the visit.The study methods were explained to them before data collection.

Sampling of municipalities
This study forms part of a larger study measuring food environments across diverse settings in Flanders.Based on feasibility, a sample of 16 municipalities out of a total of 300 was selected in Flanders.The following criteria were taken into account when making a representative selection of municipalities: LOGOs (local health care regions [14]), province, population size of the municipality (derived from Statistiek Vlaanderen [15]), demographic data of the municipality (derived from STATBEL [16]), prevalence of overweight and obesity among children (derived from Vlaams Agentschap Zorg en Gezondheid [17]), total number of supermarkets per 1000 inhabitants in the municipality (derived from the Locatus database [18]) and the type of municipality (industrial, agricultural, coastal, urban, rural, etc) (derived from the BELFIUS municipality classification [19]).The final selection of municipalities and some of their socioeconomic indicators are shown in Fig. 1 and Table 1.The level of urbanization of the municipality was derived from Departement Omgeving Vlaanderen [20] and the tertile of median disposable household income was derived from STATBEL [21].

Sampling of stores
A sample of supermarkets was selected across the five biggest supermarket chains in terms of market share (Colruyt, Carrefour, Delhaize, Lidl, Aldi).One supermarket of each chain (if available) was randomly selected in each selected municipality.and across different urbanization levels and tertiles of median household income for the statistical sectors where the supermarkets were located.In total 55 supermarkets were selected for inclusion and their characteristics are presented in Table 2 and Table 3.The size of the supermarkets (in cm²) was derived from the Locatus database [18].

Measures
Based on recommendations from the 'International Network for Food and Obesity/NCDs Research Monitoring and Action Support' (INFORMAS) [8], the following indicators on the supermarket in-store food environment were measured in this study: • The ratio of cumulative linear shelf length for healthy versus unhealthy foods • The cumulative linear shelf length of healthy and unhealthy foods, corrected for supermarket size • Proportion of foods displayed at the checkouts that are unhealthy • Proportion of foods displayed at the front end-ofaisle endcaps that are unhealthy In addition, the extent of marketing on food packages instore, as well as in different store areas, was assessed.

Data collection
Data collection was carried out between July and October 2022.For the data collection, five field workers were recruited, one per province in Flanders.
The fieldworkers received a Tablet with the app Kob-oCollect.All of the data were entered via KoboCollect   datasheets, as well as the pictures taken (as explained below).The data were uploaded on a daily basis to the KoboCollect server, after which they were downloaded as excel files and pictures and stored on a local server.The total assessment took between 2 and 3 hours per store, depending on the size of the store and the food products available.A training of two full days was given to the fieldworkers with a practical field work session before the actual start of the field work.

Shelf length
For measuring shelf length, six food categories were measured as previous research has shown that these categories are a valid proxy for the overall availability of (un)healthy products in a supermarket [22].Linear shelf length of 1) fresh fruit and vegetables, 2) frozen fruits and vegetables, 3) soft drinks and energy drinks, 4) crisps and snacks, 5) sweet biscuits, and 6) confectionery was measured by two researchers in centimetres using a MAKA MK201 60M Laster Distance Meter, either along the shelf or along the floor in front of the shelf.The definition of the food categories is given below in Table 4.
The number of shelves (of equal measured length) on which the food was displayed was recorded and multiplied by the linear shelf length to obtain the cumulative shelf length for each food category.If shelf length for a particular food category was different across different shelves, the shelf length was measured and recorded for each shelf separately and then summed to produce a total shelf length.For shelving units that did not have a physical shelf (e.g., units with hanging confectionery), rows of hanging products were counted as a single shelf.
Displays that contained multiple rows of different products (e.g., deli meats or dividers between frozen food) were also counted as multiple 'shelves' .Measurement of islands/freestanding bins was performed by measuring the exposed sides from which customers could pick products, in line with previous studies [23,24].For round, freestanding bins, the circumference was measured.

Prominence
To measure the prominence of different foods, two different methods were used.
In-store areas (eight areas: 1) check-out end, 1) checkout side, 3) endcaps front, 4) endcaps back, 5) islands, 6) aisles, 7) entrance, and 8) the edges) were categorized into high/medium/low prominence based on the validated Gro Promo tool [25].The Gro Promo tool measures the locations of the products and assigns them a weighting according to their location.High prominence areas included check-outs (end and side) and endcaps front (endcaps facing the check-outs), medium prominence areas included endcaps back, islands, aisles, and entrance, and the edges of the supermarket were considered low prominence areas [25].Check-outs included self-check-outs.
Photographs were taken of all food products (and thus not just the selection of food categories taken for the shelf length measures) offered through the checkouts and the end-of aisle endcaps.These were then uploaded in KoboCollect.

Marketing on-pack
Marketing was identified for the products photographed in the high prominence locations (check-outs and endof-aisle endcaps) and for the products within the six food categories for which shelf length was measured.One photo per product of the front of the packaging was taken, taking into account that all extras such as gifts, games, etc. were included in this photo.The brand of the food products was registered.
According to the INFORMAS methodology [8], the following categories were taken into account as marketing: 1. Cartoon character/character owned by the brand (e.g.M&Ms)

Food classification
Foods and non-alcoholic beverages were coded into one of the 17 categories of the WHO Europe nutrient profile model that distinguishes between food products permitted and not permitted to be marketed to children.The model covers all foods and non-alcoholic drinks marketed to or for children aged36 months or older [26].In addition, all food products were coded according to the extent and purpose of food processing using the NOVA classification [27].Ultra-processed foods (UPF) are products made mostly or entirely from substances extracted from foods or derived from food constituents with little, if any, intact food.They often contain flavours, colours and other additives that mimic or intensify the sensory qualities of foods or culinary preparations made from foods [27].

Data analysis
Data were analyzed using SAS statistical software version 9.4 (SAS Institute Inc., 2013).For locations within supermarkets where the number of shelves was not filled in, it was assumed to be 1 shelf.When shelf length was not filled in for some locations and food categories within supermarkets, this was assumed to be zero.When the location at shelf length was not filled in for some food groups, this was included for the total shelf length in the supermarket, but not for the location-specific shelf Descriptive statistics (mean, standard deviation and 95% confidence interval) were used to assess the healthiness of in-store food environments using aforementioned indicators.Analysis of variance tests were conducted to assess differences in indicators between different levels of socioeconomic deprivation (using the tertile of median household income of the areas where supermarkets were located).

Results
Data were collected for 55 supermarkets in Flanders, of which 31 (56.4%) were in urbanized areas and 33 (60%) in areas with relatively low median household incomes (household incomes within the lowest three deciles) (Table 5).
The ratios of total cumulative linear shelf length for healthy (fresh and frozen fruit and vegetables) versus unhealthy (soft drinks, confectionery, salty and sweet snacks) foods, as well as the total cumulative linear shelf length for healthy and unhealthy foods, corrected for size of the supermarket, are shown by supermarket chain in Table 6 below.
Overall, the average ratio of healthy/unhealthy foods in supermarkets in Flanders is 0.36 meaning that for every 10m of shelf length of unhealthy foods there is 3.6m of healthy foods (Table 6).There is large and significant variation in ratio's across supermarket chains.The discounter Aldi is on average having the highest ratio of healthy versus unhealthy food linear shelf length while Carrefour is having the lowest average ratio.
Corrected for size of the supermarket, Carrefour and Colruyt offer the highest total cumulative linear shelf length for fresh and frozen fruits and vegetables, while this is the lowest for the discounter Aldi, while similarly for unhealthy foods, Carrefour is offering the highest total cumulative linear shelf length and Aldi the lowest (Table 6).
It can also be observed that supermarkets located in more deprived areas have a lower ratio than those located in less deprived areas; for example the shelf length ratio healthy/unhealthy foods is 0.35 on average for supermarkets in areas with the lowest median household incomes while it is 0.43 for supermarkets in areas with the highest median household income (Table 6).
The ratios for total cumulative linear shelf length for healthy (fresh and frozen fruit and vegetables) versus unhealthy (soft drinks, confectionery, salty and sweet snacks) foods, as well as the total cumulative linear shelf length for healthy and unhealthy foods are shown by location in the supermarket and supermarket chain in Table 7 below.
From Table 7 it can be observed that the shelf length ratio healthy/unhealthy foods is lowest in the high prominence areas such as check-outs and front end-of aisle endcaps while it is highest in the low prominence areas in the supermarket (e.g.along the edge).The differences are substantial.There is also substantial variation across  *For 3 supermarkets, the number of check-outs was not entered as it was forgotten by the fieldworkers supermarket chains with Colruyt having the highest ratio in the high prominence areas, probably due to the fact that they have a commitment not to have junk food at the check-outs, while Delhaize has the highest ratio in the low prominence areas (Table 7).When we look more closely at the foods available in the high prominence locations in the supermarkets (front end-of-aisle endcaps and check-outs), we can see that on average 81% of products available are foods and nonalcoholic beverages at the check-outs, while 44% and 52% are foods and beverages for the end-caps front and back, respectively.Of all foods available, 97.5% are ultraprocessed at the checkouts, while 72.2% and 58.5% are ultraprocessed at the end-caps front and back, respectively (Table 8).At the check-outs, the most commonly available foods are chocolate and sugar confectionery (71.3%), while at the endcaps front and back the most commonly available foods are chocolate and confectionary (9.4%) and sugary drinks (7.2%) respectively (Table 8).
When we look at instances of marketing at those high prominence locations (excluding food packages), we observed 26 promotional characters and 97 premium offers for end-caps front, 36 promotional characters and 202 premium offers across the 55 end-caps back and 17 promotional characters and 104 premium offers across all check-outs across the 55 supermarkets investigated (data not shown).
The number of on-pack marketing ads found across the 55 supermarkets is shown below in Table 9.As it concerns marketing on packaging (responsibility of the companies themselves), no breakdown by supermarket is given.Confectionery and sweet biscuits were the food categories with on average the highest number of on pack ads per 10m of shelf length (Table 9).For frozen fruit and vegetables no on-pack marketing was found whilst for fresh fruits and vegetables 54 promotional characters and 1 premium offer were identified (data not shown).For all food groups a large share of the marketing could be attributed to a few brands (Table 9).

Discussion
Supermarket in-store food environments in Flanders are generally unhealthy and do not nudge consumers toward healthier food choices.The measured indicators on availability, prominence and promotion all show a predominance of unhealthy foods, with supermarkets that are located in low-income areas scoring the worst.
Whilst previous research showed that Belgian supermarkets made some nutrition-related commitments to create healthier food environments [12], these commitments mainly relate to reformulation, such as reducing nutrients of concern such as sodium, saturated fat, trans fat, added sugar and energy content; implementing the Nutri-Score on their own-brand food products and committing to the Belgian Pledge to reduce marketing towards children.Nonetheless, commitments regarding product accessibility, such as to dedicate a minimum amount of floor space to healthy products or limit the placement of unhealthy products at high-traffic areas, were mostly lacking [12].Colruyt however has a commitment not to place unhealthy foods at the check-outs [28] and, while their definition might differ from that used in the study, we indeed observed a very low percentage of ultraprocessed foods at the check-outs for this chain compared to the other chains.
The differences between supermarket chains are stark.The discounters Aldi and Lidl had, together with Colruyt, the best ratio of healthy/unhealthy cumulative shelf length.This is partially in line with a study conducted in Australia, which reported that Aldi had the least amount of space devoted to unhealthy food, in particular at endof-aisle and checkout displays, in regards to independent and more high-end supermarkets [29].However, these findings are in contrast with a study from the UK that found that people who bought groceries in discounters bought significantly lower percentages of energy from fruit and vegetables and higher percentages of energy from unhealthy foods [30].The latter study however did not assess the in-store food environment of supermarkets in the UK and therefore cannot be directly compared with our findings.
That our study found that supermarkets in low income areas have unhealthier in-store food environments might be due to the supply and demand dynamics.Families with a lower socioeconomic position, according to income  level or education, may have less nutritional knowledge and material resources (e.g.food budgets and facilities to prepare food) [31,32] resulting in more unhealthy diets [33].However, a study which assessed the perceptions of participants with low socioeconomic position towards supermarket nudging (i.e.making healthier choices easier or more intuitive in a retail environment by, for example, placing healthy food items at eye level), found generally positive results regarding the perception of participants towards nudges (34).Unhealthy, ultra-processed foods in Belgium have been found to be less expensive than healthy, unprocessed foods [35,36].It is therefore likely that supermarkets located in low income areas change their product promotions and prominence based on their clients preferences.Marketing aimed at children for unhealthy food categories, such as confectionary and sweet biscuits, is still very common despite the long established scientific findings about its detrimental effects on children's nutrition knowledge, preference, purchasing behaviour and dietrelated health [37].More than half of the food products in high prominence locations (front end-of-aisle endcaps and checkouts) were ultraprocessed, effectively nudging consumers to buy more of these products [38].
Even though this is a logical action from the supermarkets' commercial perspective, from a public health point of view this will lead to more health issues of an already more unhealthy and vulnerable part of the population [39].These results underscore the importance of continued and repeated measurements of supermarkets instore food environments to evaluate the implementation of their commitments.
There are a range of evidence-based actions that food retailers can take to improve the healthiness of retail food environments, categorized broadly into four categories [40]: (1) 'Corporate strategy' , including actions related to overarching company strategies and goals related to nutrition and health; actively supporting relevant public health-related government interventions; and avoiding lobbying against public health regulations to address unhealthy diets, (2) 'Product development and labelling' , including actions related to: reformulation of existing own-brand products; introduction of new healthier own brand products; implementation of easy-to-understand food labelling on own-brand products; (3) 'Product availability and placement' , including actions related to product availability, allocation of shelf space, placement of products in prominent areas, (4) 'Promotional activities' , including actions related to pricing strategies, promotions in catalogues/circulars, in-store / online signage, images or branding that appeal to children, loyalty rewards [40].In most cases, however, mandatory government regulation is likely to be needed to remove commercial barriers (e.g.such as restrictions on price promotions for unhealthy foods) [41].
Strengths of this study include the representative sample of supermarkets included and the diversity of indicators measured covering food availability, prominence and promotion.When using the indicators in future monitoring, it is important to consider the frequency and period of measurements (seasonality, frequency of changes of products within stores).It needs to be acknowledged that the types of foods in some store areas (endcaps) may change more quickly than in others (e.g., check-outs).In addition, the shelf length ratio does not include all foods in-store (e.g., non-food products and products that are not considered either healthy or unhealthy are excluded).Even for the healthy and unhealthy food categories, some indicator food categories were included rather than measuring all foods in order to improve feasibility as the measures of shelf length are the most burdensome to include in the study.Some food products are not placed on a physical shelf (e.g., hanging confectionary or fruit in freestanding bins) and methods were slightly adapted to be able to measure shelf length for those products.

Conclusion
Supermarket in-store food environments were found to be generally unhealthy, with large differences between supermarket chains and the area they are located.Supermarkets located in low income areas were found to have unhealthier in-store food environments than supermarkets located in medium and high income areas.The cumulative shelf length of healthy to unhealthy food products was under 0.5 for all supermarkets and high prominence areas such as checkouts and front end-ofaisle endcaps predominantly contained unhealthy and ultra-processed foods.Marketing aimed at children was very common for unhealthy food categories such as confectionary and sweet biscuits.Despite commitments of all large supermarket chains in Flanders to promote and create healthier in-store food environments, our findings indicate that currently consumers are incentivized to buy unhealthy rather than healthy food products.

Fig. 1
Fig. 1 Selection of municipalities and supermarkets in Flanders, 2022

Table 1
Representativeness of the sample of municipalities according to population size, % of elderly, % of young people, yearly net average household income and the total number of supermarkets/1000 inhabitants,

Table 2
Total number of supermarkets in Flanders and in the selected municipalities, 2022.Total count and count by chain

Table 3
The distribution of selected supermarkets across levels of urbanization, income terciles of the statistical sectors where the supermarkets are located and the supermarket chains, 2022

Table 4
Food categories included in the shelf length measurements in-store

Table 5
Overview of supermarkets included in the study in Flanders, 2022

Table 6
Ratio of cumulative linear shelf length for healthy and unhealthy foods for different supermarket chains in Flanders, 2022

Table 7
Ratio of cumulative linear shelf length for healthy and unhealthy foods for different supermarket chains in Flanders by location in the supermarket, 2022

Table 9
Overview of on-pack marketing found across supermarkets in Flanders,