Given the existence of an educational and income gradient in the prevalence of obesity, people with the lowest income are those who generate the highest healthcare costs related to obesity. At the same time, people living in poverty are more likely to suffer from malnutrition. For both reasons, interventions in these population segments are essential. Previous research has implemented several strategies to encourage healthy food choices among low-income families. These strategies include educational programmes, interventions fostering healthy food, and financial incentives. Some studies used small samples given the difficulty of approaching these less advantaged households. Additionally, some of these approaches tackled the question using lab experiments rather than field ones. Moreover, given that habits are easily malleable for children, a large proportion of this previous research focused on households with children or specifically on children.

Given the limitations of previous studies, the current research aimed to measure the effect of educational workshops, informational SMS messages, and cash aid on the food purchases of families receiving cash food assistance in Catalonia. The final database contained 1,429 cash-aid beneficiaries from Red Cross Catalonia who had previously consented to participate. Of all the beneficiaries listed by the Red Cross, only 23 individuals did not consent. Of the total, 1,406 used their regular vouchers at any supermarket over 2023, 1,213 over the pre-treatment period and 1,356 over the post-treatment period. In total, 1,163 beneficiaries used their vouchers during both periods. Once accepted, participants were given data protection information, an informed consent document, and an initial questionnaire that was administered by Red Cross staff. The main objective of the questionnaire was to assess the socio-economic and educational status as well as the eating habits of the participants.

Four arms were designed, including a control arm and three exclusive interventions: (i) provide workshops on nutritional guidelines; (ii) send eight SMS messages that promote healthier food purchases; and (iii) provide cash aid consisting of an additional purchase credit beyond the existing one. Stratification was performed at the branch level to avoid contagious effects between arms. Out of the 1,163 final participating beneficiaries, the distribution into the different arms was as follows: 345 were not treated, 180 received workshops, 322 received SMS messages, and 316 received additional cash aid.

Three complementary indexes were used to evaluate the healthiness of the food purchased from different viewpoints. To allow comparability, they were re-scaled to 0-100. Values closer to 100 imply healthier purchases. The indexes used were the Índice de Alimentación Saludable en España (Healthy Eating Index for the Spanish Population, IASE), the Healthy Trolley Index (HETI), and the Healthy Purchase Index (HPI). The IASE measures how well people’s diets follow Spain’s dietary guidelines. The HETI looks at how much money is spent on different food groups compared to what is recommended, creating a score to show how closely people follow these guidelines. Finally, the HPI evaluates the nutritional quality of household food purchases by combining two scores: one that looks at the variety of foods bought and another that assesses how healthy those foods are. Alternative indexes were also considered. However, the research team ultimately chose the IASE because it strictly refers to the Spanish context using nutritional information. The HETI and the HPI were included because they are complementary approaches that assess the healthiness of purchases and analyse the ratio of healthy to unhealthy foods in monthly purchases. The former was developed for Australia and the latter for France.

The research team estimated the average treatment effects (ATEs) separately by intervention arm. The results indicated that all three interventions had a positive impact on the IASE (that is, on the nutritional content of the purchased food), with the greatest impact coming from the workshops and the smallest from additional cash aid. However, for the HETI index, the channels that showed a significant positive impact were the ones that provided information (workshops and SMS messages). The effect was stronger from the workshop than the messaging compared to the control arm. In contrast, the additional cash aid had no impact on either index that proxies the quality of purchases (HETI and HPI). 

The workshops had a higher impact but were less cost-efficient than the SMS messages, which were free to implement. On average, a training session costs 50€ (depending on the travel distance) because they were provided by trained nutritionists already working for the Red Cross.

Once the ATE was determined, estimating the conditional ATEs based on several characteristics is strongly policy relevant. Some groups of individuals based on a specific covariate might show a different impact from the intervention. This approach made it possible to examine treatment effects without stratifying continuous covariates such as age and average voucher amount over the pre-post treatment periods. Those experiencing small changes in the voucher amount showed a higher impact from the intervention than those with significant changes in their aid.

This study contributed to previous literature on healthy food choices by low-income families, running a randomised controlled trial in low-income families experiencing extreme vulnerability. Three kinds of interventions (nutritional workshops, SMS messages, and additional cash aid) were provided to beneficiaries of vouchers provided by Red Cross Catalonia for food purchases during 2023. The results pointed to the treatment impacting the IASE and the HETI, but not the HPI. Disaggregating by intervention arm, the results indicated that the treatment’s positive impact on the IASE (adherence to nutritional guidelines) came through all interventions. However, the effect was not always significant when using the HETI, and, when using the HPI—which accounts not only for healthiness but also variety—the intervention effect disappears. The results may be useful for health policymakers in Catalonia and other contexts to address low-income families with extreme vulnerability who are already receiving cash-aid vouchers for food.

The present study has some limitations. First, only the short-term impact of the interventions was measured. Second, individuals received the intervention from the same institution that provides their regular aid. Fear of losing their present aid may have conditioned their food choices during the trial.

LIST, J. A., & SAMEK, A. S. (2015): «The behavioralist as nutritionist: Leveraging behavioral economics to improve child food choice and consumption». Journal of Health Economics, 39, 135-146.

LOEWENSTEIN, G., PRICE, J., & VOLPP, K. (2016): «Habit formation in children: Evidence from incentives for healthy eating». Journal of Health Economics, 45, 47-54.

MORA, T., & LOPEZ‐VALCARCEL, B. G. (2018): «Breakfast choice: An experiment combining a nutritional training workshop targeting adolescents and the promotion of unhealthy products». Health Economics, 27(2), 306-319.

SAPIO, S., & VECCHIO, R. (2024): «The effect of nudges on healthy food choices of individuals with a low socio-economic position: A systematic literature review». Trends in Food Science & Technology, 104475.