Abstract
Urban–rural comparisons of childhood undernutrition suggest that urban populations are better off than rural populations. However, these comparisons could mask the large differentials that exist between socio-economic groups in urban areas. Data from the Demographic and Health Surveys for 11 countries from three regions were used to test the hypothesis that intra-urban differentials in child stunting are greater than intra-rural differentials, and that the prevalence of stunting among the urban and the rural poor is equally high. A socio-economic status (SES) index based on household assets, housing quality, and availability of services was created separately for rural and urban areas of each country, using principal components analysis. Odds ratios (OR) were computed to estimate the magnitude of differentials in stunting (height-for-age Z scores < −2) between urban and rural areas and between the lowest and highest SES quintiles within areas. The prevalence of stunting was lower in urban than in rural areas for all countries, but rural–urban odds ratios were relatively small (< 3.3). As hypothesized, the gap between low and high SES was markedly larger in urban (median OR, 4) than rural (median OR, 1.8) areas, and differentials were statistically significant (interaction between area and SES in logistic regression) in all but three countries. Within-urban ORs as high as 10 were found in Peru and the Dominican Republic, whereas within-rural ORs were smaller than 3.5, except in Brazil. In most countries, stunting in the poorest urban quintile was almost the same as that among poor rural dwellers. Thus, malnutrition in urban areas continues to be of concern, and effective targeting of nutrition programmes to the poorest segments of the urban population will be critical to their success and cost-effectiveness.