Explaining the variation in child and adult health and nutrition outcomes in Bangladesh's cities: Evidence from empirical decompositions

Bangladesh, a lower-middle income country with the eighth largest population in the world, has experienced rapid urbanization and the growth of large slums, accompanied by steeper improvements in rural health and nutrition... [ view full abstract ]

Bangladesh, a lower-middle income country with the eighth largest population in the world, has experienced rapid urbanization and the growth of large slums, accompanied by steeper improvements in rural health and nutrition outcomes. Attention is thus shifting to the health and nutrition status of urban residents. Until recently, urban health research was mostly qualitative or based on quantitative surveys of small purposive samples, mainly of the capital Dhaka. In this study, we use data from a recent multi-topic urban survey of over 12,000 households representative of all cities in the country and slum and non-slum neighborhoods within these cities. In addition to the advances in sample coverage, size, and representativeness, the survey atypically captures an array of health and nutrition outcomes for children and adults (including non-communicable diseases such as hypertension, diabetes, and mental health for the latter).

Apart from adult obesity, hypertension, and diabetes where the opposite pattern is found, the health and nutrition status of slum residents is poorer than for non-slum residents. Individual, household, and community characteristics that capture most elements in the WHO’s SDH framework are more favorable for non-slum residents. We estimate regressions of the relationship between the various health and nutrition outcomes and these characteristics, separately for slum and non-slum residents. Results suggest that the effects of most characteristics on outcomes are insignificant for either slum or non-slum residents. We decompose differences in mean outcomes between slum and non-slum residents into the portion due to differences in the effects of characteristics versus the portion due to differences in the mean levels of characteristics between slum and non-slum residents. Results suggest the dominant role of differences in mean characteristics. Ongoing analysis, to be completed shortly, explores which particular mean characteristics contribute most towards explaining differences in mean outcomes between slum and non-slum residents.