The objective of this study is to find out about and evaluate food purchases made by Red Cross users, who participate in various projects and received a prepaid card to buy in Bon Àrea stores at some point during a period of seven years (2016-2022). Only the nutritional aspect of the food purchased with this assistance is analysed, and not that of the total food consumed by the households.

Specifically, records relating to 18,449 cards have been used, as well as 450 stores throughout Catalonia, for a total of 4,742,146 foodstuffs purchased, aggregated in monthly periods. In order to be able to analyse the data, a prior task of validating, coding, and searching of nutritional information has been carried out, as well as generating food groups based on the grouping of items into different categories. Sociodemographic data linked to each card are also available, such as gender, age, nationality, employment status and level of education, although the information on the last two variables is limited.

In order to be able to evaluate both the quality of food and that of the purchases from different viewpoints, five complementary indexes have been used. Given that in all cases they have a certain comparability, they have been re-scaled between 0 and 100. Values closer to 100 imply healthier purchases. The indexes used are: a) Mediterranean Diet Quality Index (M-DQI); b) Índice de Alimentación Saludable en España (‘Index of Healthy Food in Spain’, IASE); c) Programme national nutrition santé (PNNS), d) Healthy Purchase Index (HPI), and e) Food Standards Agency Nutrient Profiling System Dietary Index (FSA-NPS-DI).

The M-DQI index computes the percentage of purchases that adhere to the Mediterranean diet pattern, assigning between 0 and 2 points to each of the seven food and nutrient groups in accordance with the current recommendations of the Mediterranean diet. These seven groups are: a) saturated fatty acids (SFA) (% energy); b) cholesterol (mg); c) meats (g); d) olive oil (ml); e) fish (g); f) fruits and vegetables (g), and g) cereals (g). The scores of the groups are added together to obtain a total score for the M-DQI index with a range between 0 and 14.

The IASE index is formulated based on nutritional recommendations for the Spanish population established by the Spanish Society for Community Nutrition (SENC) and is calculated based on portions of food purchased. To calculate it, foods are classified into 10 groups (each food is assigned a specific score) and distributed according to their frequency of consumption, divided into five categories. In this study, the quantities of food purchased have been converted into standard portions for adults. The final score of the IASE index is calculated by adding together the score obtained in each of the different categories, which allows a theoretical maximum of 100 points to be obtained.

The PNNS index takes into account the recommendations of the Programme national nutrition santé (‘National Programme of Nutrition and Health’) in France. As in the case of the previous index, it is also calculated based on the frequency of consumption of foods, which are classified into 13 groups. Some are healthy, and are assigned between +1 and 0 points; others, in contrast, are not recommended, and are assigned between 0 and -1 points. The final score ranges from -17 to +14.25 points.

HPI – another of the few indexes available to evaluate the household shop – is the result of the sum of scores based on the diversity and quality of purchases made based on a classification into 9 food groups and 23 subgroups. On the one hand, the diversity of purchases consists of a score that reflects the presence in the diet of five food groups: a) fruits; b) vegetables; c) starches; d) dairy products, and e) meat, fish, and eggs. On the other hand, quality is calculated based on seven sections: a) fruits and vegetables; b) sweets; c) cheese; d) sugary drinks; e) added fats and seasonings; f) refined cereals, and g) fish. The final HPI score is based on a maximum score of 15 points, reflecting the highest quality purchase.

Finally, the FSA-NPS-DI index evaluates the nutritional profile according to the recommendations based on the nutrient profiling system of the Food Standards Agency (FSA) of Great Britain. To do this, it characterizes the individual diets of users based on a final score that takes into account the consumption of certain food groups, nutrient intake, sociodemographic variables, and lifestyle. Thus, it assigns a positive score (from 0 to +10 points) to some specific characteristics of foods – such as the energy they provide (kcal) and the sugars (g), fats (g), and sodium (mg) they contain – and a negative score (from 0 to -5 points) to fruits and vegetables, nuts, fibre, and protein. The lower the score, the higher the quality of consumption.

In addition to using these five indexes, the ratio has also been calculated to find out, based on the NOVA classification of the Food and Agriculture Organization of the United Nations (FAO), the volume of processed and ultra-processed foods purchased, as well as the percentage of foods of vegetable/animal origin. To be able to calculate them, the weights of the items have been converted into portions, or the monetary spend or the content of macro- and micronutrients have been taken into account. The NOVA classification considers the intensity of the processes to which foods are subjected, while the percentage of plant origin allows calculation of their sustainability.

Given that all these indexes have a certain degree of comparability, they have been re-scaled between 0 and 100 to make them easier to understand. Values closer to 100 imply healthier purchases made by the users of the Red Cross cards.

Figure 1 shows the values obtained from the different indexes – already adapted to a scale from 0 to 100 – and compares them with the situation, evidenced by different studies that calculate the same indexes, across France and Spain as a whole. As can be observed, the values obtained by the Spanish population are always below the reference values. In the case of the M-DQI index, the Spanish population is 11 points below the monitoring values of the Mediterranean diet identified by Bondia-Pons et al. (2010). With regard to the IASA reference index, the values are significantly below those evidenced by Norte-Navarro and Ortiz-Moncada (2011). The rest of the indexes have been compared with the French study by Tharrey et al. (2019) and that of Julia et al. (2016).

The study by Norte-Navarro and Ortiz-Moncada (2011) shows significant but not alarming differences in the values of the IASE index by income level (73.0: high income; 71.4: low income). These differences are corroborated by education level (75.6 vs. 69.8). Therefore, there is an income gradient in terms of the purchase of food. For this reason, the study also analyses the variability in the IASE index according to sociodemographic variables. The estimation method used is a generalized linear model (GLM), because the variable is on a scale from 0 to 100. As shown in figure 2, in which Spaniards have been taken as the base category, statistically significant differences are observed according to nationality. Specifically, Americans (5.0%), Maghrebis (13.5%), other Africans (8.2%) and, above all, Asians (12.0%) show higher values in the index, which means that they make healthier food purchases. In contrast, neither sex nor age have statistical significance.

The results of the study indicate that the quality of food purchases made with unconditional assistance could be lower than that of the population as a whole, which would entail the need to provide sufficient tools so that people have sufficient autonomy to make adequate food choices. This snapshot of purchases made with unconditional assistance is a prior step to conducting a randomized controlled experiment in order to improve the food purchasing decisions of the beneficiaries of the Red Cross cards. This type of analysis is fundamental in order to correctly design support policies aimed at the most vulnerable groups.

Since these are purchases for amounts granted with a spending limit, indexes cannot be considered for the total household purchases or for the overall diet of the individuals involved.

BONDIA-PONS, I., et al. (2010): «Diet quality of a population sample from coastal north-east Spain evaluated by a Mediterranean adaptation of the diet quality index (DQI)», Public Health Nutrition, 13(1), 12-24. 

JULIA, C., et al. (2016): «Validation of the FSA nutrient profiling system dietary index in French adults - Findings from SUVIMAX study», European Journal of Nutrition, 55, 1901-1910. 

NORTE-NAVARRO, A. I., & R. ORTIZ-MONCADA (2011): «Calidad de la dieta española según el índice de alimentación saludable», Nutrición Hospitalaria, 26(2), 330-336. 

THARREY, M., et al., (2019): «Development of the Healthy Purchase Index (HPI): a scoring system to assess the nutritional quality of household food purchases», Public Health Nutrition, 22(5), 765-775.