Effectiveness of an integrated agriculture, nutrition-specific, and nutrition-sensitive program on child growth in Western Kenya: a cluster-randomized controlled trial.
Wegmüller R., Musau K., Vergari L., Custer E., Anyango H., Donkor WES., Kiprotich M., Siegal K., Petry N., Wirth JP., Lewycka S., Woodruff BA., Rohner F.
BACKGROUND: Stunting rates remain unacceptably high in many regions, including sub-Saharan Africa. Agricultural programs have led to increased yields and household incomes but showed limited success in improving nutritional status. OBJECTIVES: We assessed whether linear growth could be improved through a potentially scalable, integrated program adding nutrition-specific and nutrition-sensitive components to an existing agricultural program. METHODS: In this cluster-randomized controlled trial in rural Western Kenya, we randomized children aged 6-35 months from farming families to an agricultural intervention without (control group) or with a bundle of interventions (intervention group), including distribution of micronutrient powders (MNP), poultry to increase egg consumption, seeds of greens and onions, and soap and chlorine solution, as well as provision of monthly behavior change trainings. The primary outcome was the change in height-for-age z-score (HAZ) over 2 years of follow-up. We assessed safety through active morbidity and passive adverse event monitoring. We conducted an intention-to-treat analysis, followed by per-protocol and prespecified subgroup analyses. RESULTS: From March to April 2018, we enrolled 1927 children from 126 clusters (control, 942 children in 63 clusters; intervention, 985 children in 63 clusters). Data on HAZ were available for 1672 (86.6%) children after 2 years. Adherence was >80% for use of MNP, chlorine, and greens and receipt of soap, and ∼40% for egg and red onion consumption. The intention-to-treat analysis indicated a greater change in HAZ over 2 years in the intervention group (adjusted effect size, 0.11; 95% CI: 0.02-0.19). We found a slightly stronger effect in the per-protocol analysis (adjusted effect size, 0.15; 95% CI: 0.06-0.24). Dietary diversity and consumption of iron-rich foods were improved in the intervention group, and reported instances of fever, lower respiratory tract infections, and diarrheal episodes were lower in the intervention group. CONCLUSIONS: This study found a modest improvement in linear growth, indicating the need for multiple, integrated interventions to achieve benefits. The trial was registered with clinicaltrials.gov as NCT03448484.