Objective Coeliac disease (CD) is a lifelong condition requiring strict adherence to a gluten-free (GF) diet and good availability of GF foods is critical to this. Patients with CD from lower socioeconomic groups are recognised to have higher treatment burden and higher food costs may impact this. Therefore, we aimed to assess the availability and cost of GF food in supermarkets and via the internet.
Design Supermarkets and internet shops delivering to homes in a single city (UK) were analysed between February and March 2014. Stores were identified with comprehensive internet searches. Ten commonly purchased items were analysed for cost and compared with standard non-GF alternatives. Direct measurement of the number of GF foods available was compared between stores which were categorised according to previously published work.
Setting Supermarkets covering the whole of Sheffield, UK.
Results None of the budget supermarkets surveyed stocked any GF foods. Quality and regular supermarkets stocked the greatest range, each stocking a median of 22 (IQR 39) items (p<0.0001). All GF foods were at least four times more expensive than non-GF alternatives (p<0.0001). GF products are prevalent online, but 5/10 of the surveyed products were significantly more expensive than equivalents in supermarkets.
Conclusions There is good availability of GF food in regular and quality supermarkets as well as online, but it remains significantly more expensive. Budget supermarkets which tend to be frequented by patients from lower socioeconomic classes stocked no GF foods. This poor availability and added cost is likely to impact on adherence in deprived groups.
- PUBLIC HEALTH
Statistics from Altmetric.com
Coeliac disease (CD) is an autoimmune condition affecting the small bowel, occurring in genetically predisposed individuals following the ingestion of gluten. CD is known to affect up to 1% of the population1 and has seen a fourfold increase in incidence over the past two decades.2 A gluten-free (GF) diet is currently the only treatment for CD and with incidence increasing, the demand for GF food has increased. Untreated CD can lead to potentially serious complications, including lymphoma, osteoporosis and anaemia,3 but the risk of developing these may be normalised with strict adherence to a GF diet. Gluten is ubiquitous in the Western diet and items such as bread and pasta are a staple part of many diets. GF alternatives are therefore an important component of dietary treatment for CD. Furthermore, demand for GF food is further enhanced by those without CD, including those with non-coeliac gluten sensitivity.4
Public awareness of CD and other gluten-related disorders has grown markedly over the past decade, with public awareness increasing from 44.2%5 to 74.4%6 between 2003 and 2014. This surge in awareness has helped fuel the recent growth in the GF food market which is now estimated to be worth £0.5 billion in the UK.7 A US report estimates as many as a fifth of the population are now buying GF food and media coverage of the trend is also increasing.7
It has previously been shown (2007, 2008 and 2011) that GF products are more expensive and availability is limited.8–10 This could preclude optimal adherence to a GF diet in certain groups and put them at an increased risk of developing complications. High costs impact those of a lower socioeconomic status, while limited availability may also prevent less mobile groups from accessing GF food, such as elderly and disabled patients. Finally, people from lower socioeconomic classes tend to shop in smaller local shops in a more chaotic and reactive manner.11 ,12 Studies have already shown that those of a lower socioeconomic status report a higher treatment burden such as reduced enjoyment of food and reduced social interaction.13 ,14 A higher perceived burden of disease is a predictor of poor adherence and higher rates of persistent villous atrophy.15 Furthermore, persistent villous atrophy has been shown to be more common in elderly patients with CD.
It is important to appreciate the role of prescription GF food when analysing the potential economic burden a GF diet could have on patients with CD,16 with 90% of patients with CD using prescribed foods in the UK.17 Currently, a 12-month prescription prepayment certificate costs £10418 (£2 per week) and adults are eligible to receive up to 18 ‘units’ of GF food each month. Based on UK consumer habits,19 half of these units are likely to be used on bread and pasta, with the remaining quota free to receive some breakfast provisions and snacks.
No previous studies have looked at the bearing that internet shopping may have on the market for GF food. Internet shopping has become more prevalent in recent years and we hypothesised that an increased range of GF foods would be available online and that costs may be lower which could reduce the economic burden of a GF diet.
We aimed to assess a range of supermarkets in a single UK city (Sheffield, UK; see box 1) and any association of internet shopping with the cost and availability of GF food. Furthermore, we aimed to consider the association of socioeconomic status with the availability of GF food by analysing levels of deprivation in supermarket locations and by classifying supermarkets into groups associated with different food prices.
Demographics of Sheffield
Population of approximately 550 000
An ethnically diverse city with 19% of its population from black or minority groups
The third largest metropolitan authority in England
The 56th most deprived local authority in England (out of 326)20
Supermarkets were analysed across the whole of Sheffield, including a range of store categories, sizes and locations between February and March 2014. Supermarkets and corner shops were identified using internet searches of individual Sheffield postcodes and supermarket names as well as by searching the websites of individual supermarket companies. Stores were divided into five categories: budget supermarkets, regular supermarkets and quality supermarkets and corner shops and health food shops, as has previously been categorised by Singh and Whelan (see box 2).8 Size categorisations were based on the total food aisle length of each store: less than 100 m was small, 100–250 m was medium and more than 250 m was large.
Categories of shop
Budget supermarkets—limited variety and range of low-cost food with many ‘own brand’ labels
Regular supermarkets—wide variety and range of both budget and quality foods at competitive cost
Quality supermarkets—stock a wide variety and range of selected quality foods, many specialist products and generally at a higher cost
Corner shops—small and locally situated shops stocking mostly essential and convenience foods
Health food shops—stock specialist foods, including many organic products
Store surveys were carried out by medical students from the University of Sheffield. The survey process involved measuring the total food aisle length for each store to calculate store size. Each shop was then systematically searched to identify any GF sections. If a GF section was identified, its length was measured in order to calculate the proportion of the store which was GF. Subsequently the total number of GF items was recorded. In addition, 10 items were selected to represent a range of commonly purchased GF foods and the price and availability of each of these were recorded. The items were chosen as they are staple items where no naturally occurring GF options are available. The items chosen were sliced bread, bread rolls, spaghetti, pasta, flour, flaked cereal, savoury biscuits, sweet biscuits, oats and gravy. If any of these were found their prices were recorded (all prices were recorded as pence/100 g), along with the cost of their gluten-containing counterparts. Wherever multiple versions of a product existed, the least expensive item (per 100 g) was recorded. If any item could not be found, a member of staff was asked to confirm if the store stocked them.
Analysis of deprivation
Deprivation scores (see box 3) for each store were recorded to determine whether location had any effect on the cost of GF food and its availability. An average index score for each store was derived from the postcode of the store itself in conjunction with all of its bordering postcodes.21
UK deprivation index
The UK deprivation index is a multifactorial scale of poverty which ranks all postcodes in England (1=most deprived, 32 482 =least deprived).
The deprivation index of each store was calculated using data from the open data communities organisation.21
The deprivation index takes into account seven different domains to assess deprivation:
Education, skills and training
Health and disability
Barriers to housing and services
In order to find online retailers of GF food, a Google search (UK based) was carried out for relevant keywords. Information from coeliac organisations and other GF-related web forums was also used. Stores were included if they sold GF food and delivered goods to the study area (Sheffield, UK). Stores visited as part of the survey of supermarkets were included in this analysis. Several online retailers were specialist GF/health food outlets and as a result did not sell any gluten-containing alternatives to provide a cost comparison. In this instance the prices and availability of the 10 preselected items were still recorded in order to compare costs between supermarkets and online retailers. The total number of GF items available for sale on each website was also recorded. Costs recorded were exclusive of any delivery costs associated with purchase.
No patients were involved in the study and there were no relevant ethical considerations.
The average cost and number of available GF foods was compared for each of the five store categories and online. The impact of deprivation was compared using average deprivation indices for each of the store categories. Correlation between number of GF foods available and the deprivation index for each shop was analysed.
All statistical analyses were performed using SPSS V.22 (IBM). Continuous parametric data were compared across groups using ANOVA with Bonferroni correction or the independent Student’s t test. For continuous non-parametric data the Kruskal–Wallis and Mann–Whitney U tests were used. Correlations between the deprivation index and number of GF foods available was performed using a Spearman's rank coefficient. A p value of <0.05 was considered significant and all values are two-sided.
Sixty-seven stores were identified. Of these, 27 (40.3%) stocked no GF items, with the remaining 40 (59.7%) stocking at least one. The median number of GF items available for sale overall was eight (IQR=0–32.0) items per store, but this varied greatly across store category and size (table 1).
Effect of store type and size
None of the budget or corner shops surveyed stocked any GF items (n=14) and the single health food shop visited stocked just one. Regular and quality supermarkets had a far greater range (p<0.0001), the median number of items sold were 22 (IQR=17.5–109.5) and 22 (IQR=0–39.0), respectively. When analysing the effect of store size, large supermarkets stocked the most GF items in total (p<0.0001). Of the 10 food items surveyed, 31 (46.3%) stores stocked none of the 10 items, with the remaining 36 (53.7%) stocking at least one. The median proportion of surveyed items available across all stores was 40.0% (IQR=0–80.0) (4 of the 10 surveyed items, table 1). Large, regular supermarkets offered the best availability of the 10 surveyed items, stocking 97.0% of the items. Out of the five categories, quality and regular supermarkets stocked the most, with medians of 70.0% (IQR=45.0–80.0) and 70.0% (IQR=0–90.0), respectively (p=0.880); this is significantly higher than that of the other store categories (p<0.0001).
Cost of GF food
Overall, all of the 10 GF items surveyed cost significantly more than their gluten-containing counterparts (p<0.0001, table 2), with GF items costing on average 4.1 times more. Bread, sweet biscuits and oats were the most expensive, costing on average 5.5 times more. Difference in costs between store categories was minimal. The main comparison between regular and quality supermarkets showed only one item to differ significantly in cost, with bread rolls being more expensive in quality supermarkets (p=0.041). There were no significant differences between any other GF items.
The association of online availability and cost of GF food
In total 21 online retailers were identified. All online stores surveyed sold at least one GF item. The median number of GF items available for sale from each online store was 123 (IQR=64.5–231.0), compared with eight for stores (p<0.0001). The proportion of the 10 surveyed items available for sale was also significantly higher, with a median of 80.0% (IQR=55.0–100.0) sold (p<0.001). Despite a larger range, the average cost of GF items online was generally higher than in stores (table 3). Costs of over half the items did not differ significantly between supermarkets and online retailers; however, spaghetti, pasta, sweet biscuits and savoury biscuits all cost significantly more online (p<0.05, table 3).
The effect of deprivation on GF food availability
The median deprivation index score for all of the surveyed stores was 12 712 (IQR=8417–20 003); the median deprivation score in England is 16 241. The median deprivation score for the stores which stocked all (100.0%) of the surveyed items (n=8) was 18 527 (IQR=12 364–20 076) and for those that stocked no items (n=31) was 15 760 (IQR=9379–22 142) (p=0.878). The deprivation scores did not vary significantly by store category (p=0.126) or by store size (p=0.056). No correlations were found between GF food prices and deprivation score.
We have demonstrated that the availability of GF food in budget supermarkets and corner shops remains poor and GF food remains significantly more expensive. This is the first study to assess online availability of GF food and we have shown that, although GF food is more available, prices for half of our surveyed items were significantly more expensive through online retailers. This could prevent patients with CD in lower socioeconomic groups from fully adhering to a GF diet, which is vital for both short-term and long-term overall health.
It is clear from our results that despite increased awareness of CD the availability of GF food remains poor in some areas. Availability of GF food in budget supermarkets and corner shops was exceptionally poor. This is particularly important when considering the socioeconomic status of patients who are more likely to use budget supermarkets and local convenience stores.22 Regular supermarkets and high-end stores offered the best range and were in reach of patients from deprived areas. However, GF items in these stores cost as much as 5.7 times more than their regular counterparts. There was almost no difference in costs between the regular and quality supermarkets, with 9/10 surveyed items costing the same on average. It was interesting to note the average proportion of the surveyed GF products found across all stores was 40.0%, a figure consistent with findings from 4 years ago (41%).8 Interestingly, in our study there was no difference in the deprivation index between any of the stores surveyed. This may be down to the overall levels of deprivation seen in Sheffield (see box 3) and further studies in more diverse populations should be considered. However, in our population this does suggest that patients from lower socioeconomic groups are within range of supermarkets that do stock significant amounts of GF food.
Online GF retailers offered the greatest choice, but were more expensive than supermarkets. This would again favour those on a higher income who can both afford the higher product costs and are more likely to have internet access. While the cost of products is the primary issue for patients on a low income, the effects of a limited choice should not be underestimated. Previous studies have shown that the perceived treatment burden of CD is substantial, with both cost and lack of variety contributing to this.13 ,14 The study by Shah et al14 found that this was directly linked to poor dietary adherence in groups with a low household income, with high income groups still able to maintain a strict diet despite a perception of high treatment burden. This may be due to differences in education with a failure to understand the consequences of poor adherence. More intensive dietary advice, improved variety and availability of GF food may go some way to reducing these differences.23 Prevalence of depression and anxiety are also known to be elevated in CD and cognitive, emotional and sociocultural factors have all been linked to adherence,23 ,24 meaning that it may also be necessary for primary care monitoring of these factors if issues persist.
Despite the provision of prescription GF foods there is still a significant risk that people are left with out-of-pocket expenses by their diagnosis. A study by Whitaker et al13 looked into the perceived burden on patients with CD and all patients included in the study had been prescribed a GF diet and given dietary advice by a dietician. Despite this, approximately half (46%) said they estimated their weekly food costs had risen by £10, with a fifth stating this to be a problem. The Low Pay Commission estimates there are 1.4 million minimum wage jobs in the UK, where an employee could expect to earn £213 per week,25 meaning that a £10 rise in weekly food costs could amount to 5% of their income. This proportion could rise significantly in the case of an unemployed patient with CD having to claim Job Seeker's Allowance of between £57.35 and £72.40 per week.
There are some limitations to our work. First, this study was only performed in a single UK city and may not be truly representative of the country as a whole. Nonetheless, in covering the whole city we were able to take in a wide range of boroughs with differing levels of deprivation. Second, we were not able to take account of naturally occurring GF foods that may form part of a GF diet. These products are undoubtedly of great importance when assessing the ‘healthiness’ of a GF diet. However, prior studies have shown that the majority of patients with CD do purchase significant quantities of GF alternative products and patients consider certain products such as bread to be critical to consuming a ‘normal diet’.26 Finally, we have not taken into account the impact of GF prescriptions that are available on the National Health Service in the UK. Nevertheless, despite prescriptions being available many patients do supplement this with additional products particularly as prescription foods cover only a small range of products. Future work may need to look at the real cost of CD for patients taking into account prescription charges. This may be best achieved by prospectively comparing actual shopping bills and shopping habits between different patient groups. Finally, consideration of methods to improve skills in preparing GF meals, from naturally occurring GF foods, also needs evaluating.
Despite an increase in prevalence and general public awareness, the availability of gluten-free (GF) food remains limited and costs high.
The emergence of online GF sellers has opened a new area of availability, but has done nothing to lower costs.
Shops more commonly frequented by lower socioeconomic groups, such as corner shops and budget supermarkets, provided no GF products.
Policies for treatment provision in different socioeconomic groups must be evaluated and methods to reduce disparity must be sought to improve treatment outcomes.
Current research questions
Does increased availability of gluten-free (GF) food improve adherence to a GF diet?
How does cost of GF food impact on shopping habits for patients with coeliac disease (CD)?
Does better availability of GF food have an impact on the nutritional status of patients with CD?
Twitter Follow Mitchell Burden at @mitchellburden
Contributors DSS and PDM conceived the study. RJB, WLW and DRC-D initiated the study design and carried out the bulk of data collection. MB also contributed to data collection. MB drafted the final manuscript and conducted statistical analysis in conjunction with PDM. All authors approved the final manuscript, with DSS giving final approval.
Competing interests None declared.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.