Farm-level production diversity and child and adolescent nutrition in rural sub-Saharan Africa: a multicountry, longitudinal study.

BACKGROUND

Child malnutrition remains widespread in sub-Saharan Africa, especially in rural areas where many households are involved in subsistence farming. Increasing farm-level production diversity (FPD) is often considered a useful strategy to improve child diets and nutrition, but the empirical evidence is mixed. Most studies have investigated associations between FPD and dietary diversity. We therefore aimed to analyse associations between FPD and child and adolescent nutritional status.

METHODS

In this multicountry, longitudinal study, we used representative panel data from four countries in sub-Saharan Africa (Ethiopia, Malawi, Tanzania, and Uganda) to test the hypothesis that higher FPD is positively associated with child and adolescent nutritional status. The data were from the Living Standards Measurement Study-Integrated Surveys on Agriculture collected between 2008 and 2019. We included data from all children and adolescents aged 0-18 years with available anthropometric data who were living in households involved in farming activities for home consumption, market sales, or both. FPD was measured in terms of the number of different crop and livestock species and food groups produced on each farm. Child and adolescent nutritional status was measured in terms of height-for-age Z scores (HAZ). We estimated panel data regression models with correlated random effects to control for confounding factors and time-invariant unobserved heterogeneity.

FINDINGS

The total sample size included 50 689 child and adolescent observations. In combined models, with data from all countries included, we found one additional species produced on the farm (crop and livestock combined) was associated with a mean 0·015 SD greater child or adolescent HAZ (p<0·0001). The role of FPD tended to decrease with better market access (in more remote locations mean 0·020 SD [p<0·0001] and in less remote locations mean 0·008 SD [p=0·091]). In individual-country models, the effects were smaller and statistically insignificant in three of the four countries. Livestock diversity had larger positive associations with HAZ than crop diversity (livestock diversity effect on HAZ mean 0·085 SD [p<0·0001] and crop diversity effect on HAZ mean 0·007 SD [p=0·080]). In Tanzania and Uganda, higher crop diversity was negatively associated with child and adolescent HAZ.

INTERPRETATION

Our findings suggest that further farm-level diversification is not a suitable general strategy to improve child and adolescent nutrition but might be useful in some situations. Livestock production seems to be conducive for improving child and adolescent nutrition on average. Context-specific approaches need to be developed.

FUNDING

None.