There is no printed copy available to order.
The literature has firmly established the link between urban/rural place of residence and health indicators, and the association of poor health outcomes and health services in rural areas. However, few studies have looked at gradients of urban areas or urbanicity, which have been described as the impact of living in those areas at a given time. In this report, we used several measures of urbanicity to study four health indicators: using modern contraceptives (mCPR); having four or more antenatal care (ANC4) visits for the most recent birth among women; completing three doses of an immunization that protects children against diphtheria, pertussis, and tetanus (DPT3); and providing the minimum acceptable diet (MAD) for children. The urbanicity measures include the urban/rural dichotomous measure, SMOD, which uses satellite data and population density to identify rural, peri-urban (suburbs), and urban centers; nightlights, which measures the level of luminosity; and a variable constructed from DHS data which splits urban areas into urban poor and urban non-poor clusters. Data from 30 countries with a recent DHS were used for the analysis. Of these, an in-depth analysis was also performed on six DHS surveys: Bangladesh 2014, Democratic Republic of the Congo (DRC) 2013-14, India 2015-16, Kenya 2014, Nigeria 2016, and Senegal 2016. The in-depth analysis examined the urbanicity variables and used unadjusted and adjusted regression to examine associations between the urbanicity variables and the health indicators. The analysis of the 30 surveys used adjusted regressions to examine the association between SMOD and the urban poor cluster variable with the health outcomes. The findings show few significant associations, particularly those with DPT3 and MAD. The urban poor cluster variables exhibited the most significant associations, particularly with mCPR. However, results were country specific, with some countries exhibiting large significant differences that favored the urban non-poor or urban centers. For example, in Haiti and Burundi, there were more than 80% lower odds of MAD for children who live in urban poor clusters compared to the urban non-poor. Some limitations of the analysis include small sample sizes for certain categories of the urbanicity variables and for the DPT3 and MAD outcomes. In addition, the analysis included health services indicators, which may exhibit few differences within an urban environment compared to health outcomes indicators.