Longitudinal analysis of the intrahousehold distribution of foods in rural Nepal: Effectiveness of a community-level development intervention

Inadequate child dietary quality is a problem of public health significance in rural Nepal. This study explores whether and how dietary patterns within households changed over a four-year time period with the introduction of a randomized community development intervention in rural Nepal. Individual-level dietary data within households is rarely observed over extended periods of time, which limits our understanding of within-household food distribution dynamics, especially in the context of impact evaluations. Six rural communities of Nepal with predominantly agricultural livelihoods were selected to participate in the phased implementation of a long-term community-level development intervention. Households (N=414 at baseline) and children (N=951 at baseline) in each community were surveyed at baseline; and the 116-item follow-up surveys were implemented at 6 months, 12 months, 18 months, 24 months, and 48 months. Detailed data on food consumption were collected at the household-level and for individual children older than 6 months of age using a 24 hour recall for 17 foods and food groups; parents responded for children. Child-level dietary diversity and consumption of animal sourced foods were the outcomes of interest. Fixed-effects analysis of the resulting panel data indicates that there are disparities in the responsiveness of child dietary quality with respect to household dietary quality, as measured by elasticities. Results indicate that there are no differences in the responsiveness of child dietary quality to household dietary quality between girl and boy children, but there are measurable disparities in dietary quality responsiveness across age groups of children and across regions of Nepal. As the length of time of exposure to the community development intervention increased, so did the responsiveness of child dietary quality to household dietary quality, as measured by elasticities. This pattern holds during both times of household stress and times of household prosperity, as indicated by the household-level dietary diversity differenced from the mean across all six time periods. The long-term, community-level development of rural women’s groups may have increased women’s status in the study sites and resulted in the improved diets for children, but measurement of women’s status over time is necessary to test that hypothesis. These results stress the importance of measuring and addressing intrahousehold dynamics – in particular across age cohorts – during community development projects, and caution against assuming the presence of sex bias in the distribution of foods within households.