Maternal internal migration and child growth and nutritional health in Peru: an analysis of the demographic and health surveys from 1991 to 2017.

BACKGROUND: Peru has historically experienced high rural-to-urban migration. Despite large reductions in undernutrition, overweight is increasing. Elsewhere, internal migration has been associated with differences in children's growth and nutritional health. We investigated how child growth and nutritional status in Peru varied over time and in association with maternal internal migration.

METHODS: Using data from Demographic & Health Surveys from 1991 to 2017, we assessed trends in child growth (height-for-age [HAZ], weight-for-age [WAZ], weight-for-height [WHZ] z scores) and nutritional health (stunting, underweight, overweight) by maternal adult internal migration (urban [UNM] or rural non-migrant [RNM], or urban-urban [UUM], rural-urban [RUM], rural-rural [RRM], or urban-rural migrant [URM]). Using 2017 data, we ran regression analyses, adjusting for confounders, to investigate associations of maternal migration with child outcomes and the maternal and child double burden of malnutrition. We further stratified by timing of migration, child timing of birth and, for urban residents, type of area of residence. Results are given as adjusted predictive margins (mean z score or %) and associated regression p-values [p].

RESULTS: In 1991-2017, child growth improved, and undernutrition decreased, but large differences by maternal migration persisted. In 2017, within urban areas, being the child of a migrant woman was associated with lower WHZ (UUM = 0.6/RUM = 0.5 vs UNM = 0.7; p = 0.009 and p < 0.001 respectively) and overweight prevalence ((RUM 7% vs UNM = 11% [p = 0.002]). Results however varied both by child timing of birth (birth after migration meant greater overweight prevalence) and type of area of residence (better linear growth in children of migrants [vs non-migrants] in capital/large cities and towns but not small cities). In rural areas, compared to RNM, children of URM had higher HAZ (- 1.0 vs - 1.2; p < 0.001) and WAZ (- 0.3 vs - 0.4; p = 0.001) and lower stunting (14% vs 21%; [p < 0.001]). There were no differences by timing of birth in rural children, nor by time since migration across all children. The mother and child double burden of malnutrition was higher in rural than urban areas but no differences were found by maternal internal migration.

CONCLUSIONS: Migration creates a unique profile of child nutritional health that is not explained by maternal ethnic and early life factors, but which varies depending on the pathway of migration, the child timing of birth in relation to migration and, for urban dwellers, the size of the place of destination. Interventions to improve child nutritional health should take into consideration maternal health and migration history.