Impact of healthy food and diet information on household food security: A randomized control trial in Kampala, Uganda

Abstract

Healthy diets are unaffordable for almost 3.1 billion people worldwide, and in 2018, already 43% of the African population were urban dwellers. Our food systems are changing rapidly, coupled with rising dietary aspirations. What are healthy diets and how can healthy diets be reached and increased in an urban context was the question of project NOURICITY. Looking into different information channels to deliver information on healthy diets in a low-income area in urban Kampala, Uganda, two packages of interventions were developed based on information on the selected focus group discussions with parish representatives. The first package consisted of a flier including graphic information on healthy diets and three food groups (treatment group 1). The second included the same flier plus interactive voice responses (IVRs) as a phone call to deliver the flier information in a different format (treatment group 2). For the study, we targeted 450 randomly selected households, which were randomly assigned into three groups (control, treatment group 1, and treatment group 2). Respondents from Kanyanya (a low-income parish of Kampala city) were randomly selected. They were visited two times in March as well as in December of the year 2021, while the intervention was rolled out in the period from September to November 2021. Healthy diets are measured using the Household Dietary Diversity Score and the food variety score is based on a 7-day food consumption recall, while dietary quality is measured for a subsample of women and the minimum dietary diversity for all of the selected women. The results indicated statistically significant changes per group at household food consumption in March 2021 compared to December 2021. However, the results of the difference-in-difference method between the control and the treatment groups did not display any significant difference at the household level. However, increased dairy and meat consumption in the treatment groups was observed. Over 90% of the households in treatment group 2 listened to any type of IVR message. Although no significant treatment effects were found, almost all households receiving intervention package 2 mentioned that the IVR calls were easy to follow, while almost 80% indicated sharing the information with their neighbors and 92% enjoyed receiving the message. However, the intervention has potential but needs to be improved upon.