RESUMO
BACKGROUND: Food-assisted maternal and child health and nutrition programs are a widely used approach to address undernutrition. Little is known about the effects of these programs' combined household and individual food rations on household and individual food consumption. Tubaramure in Burundi targeted women and children during the first 1000 d of life, and included: 1) food rations (corn-soy blend and micronutrient-fortified vegetable oil); 2) health services strengthening and promotion of their use; and 3) behavior change communication on nutrition, hygiene, and health practices. OBJECTIVES: The objectives were: 1) to assess Tubaramure's impact on household food consumption and food security, maternal dietary diversity, and infant and young child feeding practices; 2) to explore the role of the food rations; and 3) assess 6-8 mo impacts around 8 mo after the end of the program. METHODS: We used a 4-arm cluster-randomized controlled repeated cross-sectional design (11,906 observations). The treatment arms received the same food ration but differed in the ration timing and duration: 1) the first 1000 d; 2) from pregnancy through 17.9 mo of age; or 3) from birth through 23.9 mo of age. RESULTS: Tubaramure significantly (P < 0.05) improved the percentage of food secure households [from 4.5 to 7.3 percentage points (pp)], and increased household energy consumption (from 17% to 20%) and micronutrient consumption. The program had a positive effect on maternal dietary diversity (+0.4 food groups, P < 0.05) and increased the proportion of children aged 6-23.9 mo consuming ≥4 food groups (from 8.0 to 9.6 pp, P < 0.05). The effects on many outcomes were attributable to the food rations. Postprogram effects (P < 0.05) were found on household food security, maternal dietary diversity, and younger sibling's complementary feeding practices. CONCLUSIONS: Programs such as Tubaramure have the potential to improve food security and household and individual energy and micronutrient consumption in severely resource-constrained populations, as seen in rural Burundi. This trial was registered at clinicaltrials.gov as NCT01072279.