Impact of a Homestead Food Production Program on Dietary Diversity: Seasonal and Annual Results from the Cluster-Randomized FAARM Trial in Sylhet, Bangladesh

Abstract

Background: Rice-predominant diets are common in Bangladesh, leading to widespread nutritional deficiencies. The Food and Agricultural Approaches to Reducing Malnutrition (FAARM) cluster-randomized controlled trial in rural Sylhet, Bangladesh evaluated a homestead food production intervention implemented 2015-2018 through Helen Keller International, aiming to improve child growth.

Objectives: We estimated the intervention's impact on women's and children's dietary diversity, a secondary trial objective.

Methods: We collected dietary diversity throughout the trial (March 2015 to June 2020) at multiple times each year using standard, United Nations-endorsed, self-reported measures for women (10-food group scale) and children (7-food group scale). We included 28,282 observations of 2701 women (out of 2705 enrolled) and 17,445 observations of their 3257 children (aged 6-37 mo) in 96 settlements, 48 of which received the intervention. We estimated the intervention's impact on dietary diversity by year of intervention, overall periods following the start of the intervention, and seasonally, using multilevel regression with the control group as the counterfactual, controlling for seasonality, baseline dietary diversity, and clustering by settlement and repeated measures.

Results: At baseline, approximately one-third of women and children consumed a minimally diverse diet. Over the entire intervention and postintervention period, women's and children's odds of consuming a minimally diverse diet nearly doubled (odds ratio [OR] 1.8, P < 0.001, for both). This benefit was barely present in the first year, increased in the second, and peaked in the last intervention year (OR 2.4 for women, OR 2.5 for children, both P < 0.001) before settling at around double the odds in postintervention years (P < 0.001). Dietary improvement was observed throughout the year for both women and children with incremental increases in nearly all food groups.

Conclusions: The nutrition-sensitive agriculture intervention successfully increased dietary diversity in women and children, and these impacts persisted after the project closed, including during the early COVID 19 lockdown period. This trial was registered at clinicaltrials.gov as NCT02505711.

Keywords: gardening; infant and young child feeding; minimally diverse diet; nutrition-sensitive agriculture; seasonality.