Background

High and imbalanced dietary intakes of salt, saturated fat, and sugar are modifiable risk factors for overweight, obesity, and diet-related noncommunicable diseases (NCDs) [1,2,3]. The rising global trend in overweight and obesity is mirrored in Ireland, with 21% of children, 24% of adolescents and 57% of adults now living with overweight and obesity [4,5,6]. Diets in Ireland, like many other countries worldwide, exceed the World Health Organization (WHO) recommended intake for saturated fat of less than 10% of total energy (TE), conditional recommended intake for free sugar of less than 5%TE and salt intake of < 5 g/day [7].

The determining factors of a healthy diet are varied and complex [8]. Environmental, social, commercial, physical, psychological, and economic determinants influence what people eat [9,10,11,12]. Food environments are an important component of an individual’s choice and access to foods [13]. Evidence shows that the foods available and accessible to a person in the community and the consumer food environment can influence food choice [14]. Interventions such as front of pack nutrition labelling, food reformulation, and sugar-sweetened beverage tax have been pursued to improve the healthiness of the consumer food environment [15, 16].

More recently, the retail food environment has been highlighted as an amenable environment to intervention to improve the community and consumer food environment and enable consumers to choose foods with healthier nutrient profiles [17, 18]. Winkler et al. define the retail food environment as “the environment where all food and beverages are purchased by consumers, including food service operations such as restaurants” [19]. The retail food environment comprises of different types of food outlets such as restaurants, cafés, takeaways, street food vendors, pubs, convenience stores, grocery stores and supermarkets [19].

Evidence shows that the community and consumer food environments can influence food choice and contribute to subsequent diet-related health outcomes [14]. Evidence suggests that supermarkets can influence dietary behaviours [20], and they are gaining influence in the retail food environment globally, as the amount of food purchased from supermarkets increases [21, 22]. In Ireland, during the time period of this study supermarkets made up 89% of grocery market share and in 2010, the most recent data available, 78% of meals prepared were consumed at home [23, 24]. Supermarkets are recognised for their important role in providing access to healthy fresh produce; however, they have also been shown to be a significant source of high fat, sugar and salt foods (HFSS) [25]. In-store supermarket interventions to reduce the availability and prominence of unhealthy food whilst increasing availability and prominence of healthy food have been shown to be effective in supporting consumers to make healthier food choices and purchases [26]. In light of this, voluntary and mandatory measures have been introduced to limit the availability of HFSS foods in supermarkets. For example in Ireland, a voluntary measure was introduced to limit the availability of HFSS foods at supermarket check-outs [27]. More recently, regulation has been introduced in England, The Food (Promotion and Placement (England) Regulation 2021, to restrict the prominence and promotion of HFSS foods [28]. Early evaluation of this legislation has indicated it is a good starting point, however its impact on reducing consumer purchase of HFSS foods has yet to be fully evaluated [29].

Several researchers to date have considered the healthfulness of the supermarket food environment using the International Network for Food and Obesity / Non-communicable Diseases Research, Monitoring and Action Support (INFORMAS) Protocol: Food Retail - Food availability in supermarkets [30]. These studies have found unhealthy foods are more available and prominent in supermarkets than healthy foods. Vandevijvere et al. reported that for every meter (100 cm) of shelf space of unhealthy food in New Zealand supermarkets, there was an average of 42 cm of healthy food [31]. Schultz et al. reported in supermarkets in Victoria, Australia, 63.2-72.7% of shelf space was allocated to discretionary (unhealthy) foods, depending on the retailer [32]. Supermarkets in Flanders, Belgium were found to have 3.6 m of shelf space allocated to healthy foods, for every 10 m of shelf space measured [33]. These studies also found a higher proportion and prominence of shelf space allocated to unhealthy foods in areas of higher deprivation and the proportion of shelf space allocated to unhealthy food, as well as its prominence varied across retailers [31,32,33]. Despite the important role supermarkets play in consumer food choice, the healthfulness of the consumer food environment in Irish supermarkets has not been investigated. This study aimed to address this evidence gap by determining the proportion of available shelf space allocated to healthy and unhealthy foods, represented by a proxy indicator, and the prominence of this shelf space in supermarkets, by area-level deprivation in County Dublin, Ireland. A secondary aim was to determine the proportion of available shelf space allocated to healthy and unhealthy foods, represented by a proxy indicator, and the prominence of this shelf space in supermarkets overall and by retailer.

Methods

Design and sample

A cross-sectional study design was applied, where supermarkets in County Dublin, Ireland (including the city and its suburbs) were considered. County Dublin is in the Leinster region, located at the midpoint of Ireland’s east coast and where Ireland’s capital city Dublin is located. It covers an area of 922 square kilometres and is the most densely populated area of Ireland, with a total population of 1,458,154 and 3,659 people per square kilometre, with the national average being 73 people per square kilometre [34]. For the purposes of this study a supermarket was defined as per the INFORMAS protocol: “primarily self-service grocery stores selling food and other household items with a minimum of 2 or more checkout stations/aisles”. For pragmatic purposes a decision was made to include the five leading supermarket chains in Ireland who dominate the market and had, on average, 89% of grocery market share in 2021 [23].

The five large supermarket chains in Ireland were invited to participate in the study. Three consented to participate, and their stores accounted for n = 99 (54%) out of n = 185 of grocery stores across County Dublin, Ireland. The three grocery retailers who agreed to participate were a combination of retailers and discount supermarkets. All retailers agreed to participate in an anonymous manner, meaning no single retailer would be identifiable in any analysis or presentation of results. Ethical approval was provided by University College Dublin Office for Research Ethics, reference code LS-E-21-24-Omahony-Gibney. Consent was received from an appropriate representative of each retailer taking part.

To determine the density of supermarkets by area-level deprivation and population, the addresses of all stores, including postcodes, were collected from the retailer’s websites and validated for locational accuracy by a second researcher using Google Maps. Electoral Division (ED) boundaries were sourced from the Irish Central Statistics Office [35]. The deprivation index of these EDs was sourced from the Irish Pobal HP Deprivation Index 2016, and mapped onto a geographical map of County Dublin, Ireland, in the form of polygons using the ArcGIS Online Geographic Information System [36]. Using their validated geo-referenced address point locations, each supermarket was then allocated an area-level deprivation score as per the existing categorisation set out in the Irish Pobal HP Deprivation Index 2016 (Very Affluent, Affluent, Marginally Above Average, Marginally Below Average, Disadvantaged, and Very Disadvantaged) based on the ED in which it was located.

To select supermarkets for inclusion in the study, the following two grouped area-level deprivation categories were created: “Low Deprivation” (consisting of the Very Affluent, Affluent, Marginally Above Average groups) (n = 68) and “High Deprivation” (consisting of Marginally Below Average, Disadvantaged, and Very Disadvantaged groups) (n = 31). Two groups, high deprivation and low deprivation, were created to determine a more reasonable sample size.

The primary objective of this study was to compare the proportion of relative available shelf space allocated to healthy and unhealthy food and the prominence of this shelf space in supermarkets between the low and high deprivation groups. To do this with a high degree of precision, an equal sampling strategy was used, as the lowest estimation error is achieved by using equally sized samples within the two deprivation groups. In this study, n = 36 supermarkets were sampled, with n = 18 supermarkets being sampled from the “Low Deprivation” group and n = 18 from the “High Deprivation” group. Assuming (conservatively) that the standard deviation in relative availability is σ = 0.115, this sample size ensured (with 95% probability) the estimated difference in average relative availability would be within 0.075 units of the true difference. This comes from the fact that the margin of error in the 95% confidence interval is

$$\:1.96\times\:SE=1.96\times\:\frac{2\left(0.115\right)}{\sqrt{36}}=0.075$$

The assumed σ = 0.115 was determined from a conservative assumption that the relative availability value for a given supermarket could be anywhere between 0.3 and 0.7. Therefore, from a uniform distribution (which assumes maximum uncertainty), the standard deviation was defined as

$$\:\sigma\:=\frac{0.7-0.3}{\sqrt{12}}=0.115$$

The secondary research question related to overall availability and prominence of unhealthy foods in supermarkets. In order to ensure findings were representative, and not influenced by sampling for the primary objective, a weighting scheme representing the ratio of supermarket availability in areas of high deprivation (27% of all supermarkets in County Dublin were located in areas of high deprivation) and low deprivation (73% of all supermarkets in County Dublin were located in areas of low deprivation) was applied.

Audit tool

To determine the relative availability and prominence of healthy and unhealthy foods in supermarkets, the INFORMAS Protocol: Food Retail - Food availability in supermarkets FULL V1.1 was applied in the Irish supermarket context [30]. The protocol describes a methodology based on two previously validated measures, the first is to determine availability and the second is to determine prominence.

Determining the proportion of relative shelf space allocated to healthy and unhealthy food using a proxy indicator

The study used a proxy indicator where healthy foods were represented by two food categories: fresh fruit and vegetables and frozen fruit and vegetables, and unhealthy foods were represented by four food categories: confectionery (which for the purposes of this study was split into chocolate confectionery and non-chocolate confectionery in order to align with food categories included in the Irish national consumption surveys), sugar-sweetened carbonated beverages (including energy drinks), sweet biscuits, and savoury snacks. Additional details on the food categories included are outlined in Supplementary File 1, Table 1. This indicator was developed and validated in New Zealand to overcome the impracticalities of measuring the availability of foods across the entire supermarket [37]. In this study Vandevijvere et al. concluded that, in supermarkets, cumulative linear shelf space is a more sensitive measure of food availability than measuring the variety of food for sale [37]. An investigation into the validity of five shelf length ratios previously used in the scientific literature, against a gold standard indicatorFootnote 1, found the indicator which was closest to the gold standard indicator included two food categories to represent healthy foods: fresh fruit and vegetables and frozen fruit and vegetables, and four food categories to represent unhealthy foods: confectionery, sugar-sweetened carbonated beverages (including energy drinks), sweet biscuits, and savoury snacks [37]. Shelf length (cm) and shelf height (cm), or depth (cm) (depending on shelf type e.g. height was measured for a standard shelf, and depth was measured for an angled fruit and vegetable bin), and type (aisle, special display, endcap etc.) were collected for food categories included in the validated proxy indicator.

Determining the prominence of shelf space allocated to healthy and unhealthy food using a proxy indicator

In store location prominence (in terms of visibility of the area to the customer) was determined using the validated GroPromo tool, which is previously described in the scientific literature [38]. As per the GroPromo tool, shelf locations were classified into high prominence (endcap A (end of aisle facing store checkouts), checkout side, and checkout end), medium prominence (endcap B (end of aisle facing rear of supermarket), aisle, island and, entrance), and low prominence (supermarket edge), this is shown in Fig. 1 and described in Supplementary File 1, Table 2.

Fig. 1
figure 1

Prominence of shelf types and locations, adapted from the GroPromo tool and as per INFORMAS Food Availability in Supermarkets Protocol [30, 38]

Full size image

In store data collection

A pilot study was completed in May 2021, to test the data collection tools and measurement approach. Following this pilot, a decision was made to measure length (cm) and height (cm) for standard shelves, and length (cm) and depth (cm) for basket and basket-type displays where fruit and vegetables are often displayed. This approach ensured the shelf area visible to the customer was best represented. In October 2021, the refined measurement approach was tested in four supermarkets by two data collection team leads, providing an opportunity to make decisions on how to measure certain types of special displays such as floor stacks of sweets and biscuits. Data was then collected across all selected supermarkets (n = 36) by a team of six data collectors made up of two team leads and four data collection assistants. Shelf length (cm) and height (cm) or depth (cm) were measured using a retractable five-metre measuring tape. To avoid seasonal influence on food availability in December due to Christmas promotions and in January due to New Year healthy eating promotions, data was collected in two-time periods; October-November 2021 and February 2022. The data collection tools, outlining further detail of variables collected are provided in Supplementary File 1, Table 3.

Data entry and analysis

Data was entered into Microsoft Excel by three researchers. All data entered was reviewed by a second researcher and where entered data differed from the data collection form it was corrected. The cumulative shelf space (m2) was calculated by multiplying shelf length (cm) by shelf height (cm) or depth (cm) e.g. length (cm) × height (cm) or length (cm) × depth (cm), summing all shelf space for each food category and dividing by 100. The relative availability of the shelf space allocated to unhealthy food was determined by calculating the proportion of total shelf space measured, allocated to unhealthy food. To determine the prominence of unhealthy foods, the calculation for relative availability was repeated and stratified for low, medium, and high prominence. The relative availability and prominence of unhealthy food were further stratified by area-level deprivation and retailer.

Statistical analysis

The completed dataset was transferred to International Business Machines (IBM) Statistical Package for Social Sciences (SPSS) V27 for descriptive and statistical analyses. Supermarket availability and in-store outcome measures were compared between retailer and area-level deprivation using t-tests and one-way ANOVA. Statistical significance was defined as P < 0.05.

Results

In 2021, the five leading grocery retailers accounting for, on average, 89% of market share had n = 185 supermarkets spread across County Dublin, as shown in Fig. 2 [23].

Fig. 2
figure 2

Distribution of supermarkets (n = 185) across County Dublin, Ireland by area-level deprivation

Full size image

When considering all supermarkets within the County Dublin area, analysis by area-level deprivation showed there was no statistically significant difference (p = 0.4) in the density of supermarkets (n = 185) per 1,000 population in areas of high (0.14 per 1,000 population) and low (0.16 per 1,000 population) deprivation. This was mirrored in the density of supermarket stores (n = 99) for the three grocery retailers who agreed to participate in the study, as outlined in Table 1. The two areas in County Dublin that were “Very Disadvantaged” had no supermarkets present.

Table 1 Density of supermarkets in County Dublin for all (n = 5) and participating (n = 3) grocery retailers by area-level deprivation
Full size table

The characteristics of the supermarkets included in the study differed by retailer. Included retailers had between 6 and 11 checkouts. All grocery stores included had a floor space of more than 650 m2. Across these supermarkets, a total of 517,503.4 m2 shelf space was measured. The shelf space allocated to the seven food categories included in the proxy indicator were similar in the areas of high and low deprivation, as outlined in Table 2. Of all seven food categories fresh fruit and vegetables had the most relative shelf space allocated to it with 28.3%, followed by chocolate confectionery with 26.9%, savoury snacks with 18.8%, biscuits with 10.6%, sugar-sweetened carbonated beverages with 6.3%, non-chocolate confectionery with 6.1% and frozen fruit and vegetables with 2.9%. Almost as much relative shelf space was allocated to chocolate confectionery (26.9%) as fresh fruit and vegetables (28.3%).

Table 2 Mean proportion of total relative shelf space (m2) allocated to each food category overall and by area-level deprivation
Full size table

Our study observed for every metre of shelf space measured 32.0% (SD 10.6) was allocated to healthy food and 68.0% (SD 10.6) was allocated to unhealthy food (Table 3). There was no statistically significant difference between the proportion of shelf space allocated to unhealthy foods in areas of high (67.0% (SD 12.9)) and low (68.4% (SD 9.9)) deprivation (p = 0.73), hence, there was also no difference for healthy foods (as healthy and unhealthy foods are the complement of each other, summing to 100%). There was also no statistically significant difference (p = 0.66) in the availability of unhealthy foods in smaller groupings of area-level deprivation (disadvantaged, marginally below average, marginally above average affluent), shown in Supplementary Results File, Table 1. When examining individual retailers, there was a statistically significant difference between the proportion of shelf space allocated to healthy and unhealthy foods in Retailer A (p = 0.03) by area-level deprivation, with a higher proportion of unhealthy food being offered in areas of higher deprivation. Across the three retailers, there was a statistically significant difference (p < 0.001) between the availability of healthy and unhealthy foods, with Retailers A, B, and C, allocating 78.7% (SD 5.9), 66.0% (SD 6.1), and 54.3% (SD 4.0) of relative shelf space to unhealthy foods, respectively. The relative shelf space allocated to unhealthy foods by area-level deprivation and retailer is summarised in Table 3.

Table 3 Proportion (mean (SD)) of relative shelf space allocated to unhealthy foods overall, by area-level deprivation and by retailer
Full size table

Examining the prominence of foods, our study found that shelf space in areas of high prominence (endcap A, checkout end, and checkout side) was more likely to be allocated to unhealthy foods than to healthy foods. Of the shelf space measured in this study, 98.5% (SD 6.1) of the high-prominence shelf space was filled with unhealthy foods, with 1.5% (SD 6.1) allocated to healthy food. Fresh fruit and vegetables filled most of the low prominence shelf space measured, 70.7% (SD 31.2). There was no statistically significant difference between the prominence of shelf space allocated to unhealthy food by area-level deprivation. There was a statistically significant difference between the allocation of unhealthy food across low, medium and high prominence shelf space overall, in areas of high deprivation and in areas of low deprivation. The proportions of low, medium, and high prominence shelf space allocated to unhealthy foods is summarised in Table 4.

Table 4 In store prominence of relative shelf space allocated to unhealthy foods by area-level deprivation
Full size table

When prominence of shelf space allocated to healthy and unhealthy food was disaggregated by retailer and area-level deprivation, Retailer A had significantly more unhealthy foods (p = 0.015) in areas of low prominence, in supermarkets located in areas of high deprivation, additional detail of this analysis is outlined in Supplementary Results File, Table 2.

Discussion

The aim of this study was to determine the proportion of relative shelf space allocated to healthy and unhealthy foods and its prominence in supermarkets in Co. Dubin, Ireland, by area-level deprivation. A secondary aim was to determine variation in the proportion of shelf space allocated to unhealthy foods and its prominence, overall and by retailer. The study found more than double the proportion of relative shelf space was allocated to unhealthy food (68.0% (SD 10.6)) than healthy food (32.0% (SD 10.6)). Whilst there was no statistically significant difference in the proportion of shelf space allocated to unhealthy food by area-level deprivation overall, there was a statistically significant difference within one retailer. The study also found shelf space allocated to unhealthy food is more likely to be in areas of high prominence such as checkout sides, checkout end and end of aisle. The study addresses a gap in the evidence on the healthfulness of the supermarket consumer food environment in County Dublin Ireland, by determining the proportion of shelf space allocated to unhealthy food and its instore prominence. The findings of this study are important as they demonstrate a need to improve the availability and prominence of healthy foods in supermarkets in Ireland to support consumers in making healthier food choices.

Overweight and obesity follows a social gradient in the Irish context and so we were interested to examine the availability and prominence of healthy and unhealthy foods in supermarkets by area-level deprivation [5, 39]. Although this study did observe a statistically significant difference within one retailer, overall, it did not find a statistically significant difference in the proportion of shelf space allocated to healthy and unhealthy foods in areas of high and low deprivation. This finding differs from similar studies in other countries using the INFORMAS Food Availability in Supermarkets protocol including, Victoria in Australia, Buenos Aires in Argentina, New Zealand and Flanders in Belgium which observed significantly more unhealthy foods in areas of high deprivation [31,32,33, 40]. Each of these studies grouped grocery stores by deprivation or socio-economic position using different measures and number of groupings e.g. using median household income and comparing between two groups [40], using median household income and comparing between three groups [33], ranking by area-level relative disadvantage with five groups and comparing the group with the highest relative disadvantage score to the other groups [32] and using area-level deprivation and comparing across three groups [31]. Reasons for the differences in our findings can be explained. Maguire et al., noted studies investigating the relationship between the food environment and deprivation may need to include multiple measures of deprivation rather than a single measure, and may need to use the data differently depending on the context [41]. As County Dublin, Ireland is a small area geographically, where areas of high and low deprivation are closely located, multiple measures of deprivation (such as household level measures or individual level measures of socio-economic status (SES) of people living in the supermarket catchment area) may be needed to determine the relationship between the foodscape and the deprivation of the service population. Our sample size calculations were based on examining differences between two groups e.g. high deprivation and low deprivation. This could have missed differences between smaller groupings of area-level deprivation scores e.g. disadvantaged vs. affluent which our study was not powered to measure. Furthermore, although this study included three out of the five leading grocery retailers in Ireland, a repeat benchmark of this protocol should be completed in all five of the leading grocery retailers and if possible multiple measures of deprivation with smaller groupings (area-level, household and individual) included.

The overall study findings indicate the consumer food environment in supermarkets in County Dublin, Ireland, does not make the healthy choice the easy choice. Unhealthy foods are widely available and more prominent in supermarket stores, resulting in consumers being nudged towards unhealthier food options [42]. This finding is in agreement with studies completed in Victoria in Australia, New Zealand, Buenos Aires in Argentina, England, and Flanders in Belgium, which found supermarkets had a higher proportion of unhealthy foods available and unhealthy foods were more likely to be located in areas of high prominence [31,32,33, 40, 43]. Globally, supermarkets have growing power and influence on consumer food choice, and this trend is mirrored in Ireland [21]. Given the important role of supermarkets in influencing consumer food choice, there is a growing body of evidence describing in-store supermarket interventions to make them health-enabling [17].

Evidence shows that supermarket healthy food interventions which target a number of components of the marketing mix, such as reducing price and increasing product availability, can increase the purchase of healthy food, which is the precursor to consumption [18, 44]. In a systematic review examining the influences of food store product placement on diet-related health outcomes, Shaw et al. concluded that reducing the availability and prominence of unhealthy foods, and increasing the availability and prominence of healthy foods, encourages consumers to choose healthier food options [26]. Adam et al. reached similar conclusions, reporting in-store interventions combining merchandising techniques to improve the purchase of healthy foods (including interventions that combine price, information, and easy access to and availability of healthy food) are likely to be most effective in changing purchase behaviour and food choice [44].

To date in Ireland, to the authors knowledge there is no study investigating the effectiveness of in-store healthy food interventions. With over two thirds of shelf space allocated to unhealthy foods across Irish supermarkets and almost all high visibility shelf space allocated to unhealthy foods, the current layout of supermarkets in Ireland needs to change. Further research is required to determine in-store healthy food intervention components and design that are successful in the Irish context. It is also important to note that retailers are not public health authorities but commercial enterprises, and so health-enabling interventions must also be considerate of business outcomes [45]. Vargas et al. describe a co-creation approach to developing and implementing supermarket healthy food interventions, which takes into account the sometimes conflicting agendas of public health bodies and commercial food retailers [18]. Similarly, a pilot study implemented by Voegel et al. found that increasing the availability and prominence of fruit and vegetables and removing the availability of unhealthy food, predominantly confectionery from checkouts, resulted in a reduction in confectionery sales and an increase in fruit and vegetable sales whilst not affecting overall household spend [46].

Despite the significant evidence showing the potential impact of healthy in-store food interventions for improving healthy food purchase, few countries have enacted mandatory legislation requiring the consumer food environments in supermarkets to change. England is an exception, where in October 2022, they introduced legislation to restrict the placement of HFSS foods in locations of high prominence in qualifying retailers. This legislation restricts the placement of HFSS foods in the instore environment in areas of high prominence such as the supermarket entrance, checkout end, end of aisle, and on online grocery stores homepages [28]. This approach provides an example of how legislation can be used to require improvements to the healthfulness of the supermarket food environment. To date within Ireland, there are no mandatory rules to restrict the availability and prominence of unhealthy foods in supermarkets. A voluntary code of practice on the placement of healthy foods at grocery outlet checkouts was introduced in 2020 [27, 47]. The outcomes of the legislation in England could inform a future mandatory measure to improve the healthfulness of the Irish supermarket consumer food environment.

Although not an aim of this study, we observed a non-significant trend towards more supermarkets being located in less deprived areas in County Dublin. Notably, the ED’s with the highest level of deprivation in the study area (very disadvantaged) had no supermarkets located there. This indicates the possibility of ‘food deserts’ in areas of high deprivation in County Dublin Ireland. Food deserts refer to geographical locations with low access to fresh fruit and vegetables whilst having high access to HFSS foods [48]. This is of concern as previous research in Ireland has found individuals who live closer to a supermarket or who live in an area with a high density of large grocery stores have a significantly better diet in terms of cardiovascular disease risk [49]. As well as this, recent evidence from the USA found food swamps (areas that are proliferated with food outlets selling HFSS foods and with limited availability of outlets selling fresh produce) better predict obesity rates than food deserts [50]. In Ireland, it is not known where people living in areas with limited availability of supermarkets access their healthy foods e.g. do they travel further to access a supermarket or do they access foods from local convenience stores, and what is the healthfulness of these outlets? Our study considered supermarkets only, and not convenience stores, which may be an important source of foods in areas where supermarkets are less available. Evidence of Ireland’s spatial food environment, and food outlet access by area-level deprivation is required to better understand the relationship between food outlet type and food related health outcomes such as overweight, obesity, and NCDs in the Irish context. The Priority Places for Food Index, compiled by the Consumer Data Research Centre in the UK is a good example of how food outlet availability can be mapped alongside dimensions of food insecurity such as fuel poverty and social safety net schemes [51]. This approach provides readily available and easily accessible information on communities most in need of food and nutrition policy support [51].

The findings of this study and the available scientific evidence, demonstrate there is a need to increase the availability and prominence of healthy food and reduce the availability and prominence of unhealthy foods in supermarkets in Ireland. The legislative approach introduced in England and the example of successful voluntary co-design approaches between supermarkets and research groups in Australia and the UK provide an important evidence-base to inform future interventions to address the healthfulness of the Irish supermarket consumer food environment [17, 43, 45].

Strengths and limitations of this study

A key strength of our study is, to the best of our knowledge, it provides the first assessment of the relative availability and prominence of healthy and unhealthy food in the consumer food environment of supermarkets in urban Ireland. The study followed a validated protocol that has been applied in other countries which facilitates cross-country comparison of findings. Our study also has a number of limitations which should be acknowledged. The use of a proxy indicator to represent the availability of healthy and unhealthy food means that the study findings are relative and not the absolute availability of healthy and unhealthy food in supermarkets and so the percentage availability and prominence should be interpreted within this context. A second limitation is in relation to the number of retailers who participated in the study which could be considered a design weakness. As only three of the five leading Irish supermarket retailers participated, efforts to encourage all the five leading grocery retailers to participate in a future study should continue. Our study was cross-sectional, reflective of one point in time, and did not show the influence of seasonal variation of foods available in supermarkets at different times of the year. Future research could investigate the seasonal influence of food offerings in supermarkets in Ireland. In relation to national representativeness, our study was located in County Dublin which is mostly an urban setting and so does not represent rural areas, which should be considered in future research and in particular if this affects the findings in relation to unhealthy food availability and area-level deprivation. Our study had two limitations relating to how deprivation scores were determined. Firstly, our study allocated a deprivation score to the supermarkets; however supermarkets may serve a wider community than those in the direct vicinity of the supermarket location and so the supermarket deprivation score may not necessarily reflect the deprivation profile of the surrounding service population for each supermarket. Secondly, in order to create a more reasonable sample size for the resources available to this study we grouped six categories of deprivation into two broad groups (high and low deprivation), which could have made our study less sensitive to differences in availability and prominence of unhealthy food at finer levels of area-level deprivation. Our study focused on in-store grocery retail environments only and did not evaluate the availability and prominence of healthy and unhealthy foods online. Instore grocery purchase remains dominant in the Irish context, with over 80% of households purchasing groceries in store [52]. However, since Covid–19 the online grocery retail offer has increased, with four of the five leading grocery retailers now providing an online grocery purchasing and delivery service. This has resulted in an increase in online grocery retail uptake and coverage. Future studies investigating the availability and prominence of unhealthy foods should also include an evaluation of online grocery retail.

Conclusion

On average, supermarkets in urban Ireland allocate 68.0% (SD 10.6) of relative shelf space to unhealthy food and unhealthy food is more likely to be in areas of high prominence than healthy food. The availability of unhealthy food differed significantly by retailer, indicating there is an opportunity for retailers with higher availability of unhealthy food to change their food offerings. Overall, the study did not find a statistically significant difference in the availability of supermarkets, and in store availability of healthy and unhealthy foods, between areas of high and low deprivation. The study findings indicate that a change is needed to the supermarket in-store food environment to support consumers to make healthier food choices in line with Food Based Dietary Guideline recommendations. This could be done through a voluntary co-design approach or by the introduction of mandatory legislation such as that introduced in England. Further research is needed on the Irish community and consumer food environment both overall and by area-level deprivation, to better understand if the Irish food environment influences the socio-economic disparities in overweight, obesity and diet-related NCD prevalence. Studies are also required to determine effective interventions to improve the healthfulness of the Irish supermarket food environment.