Anomalies in the spatial distribution of health and child mortality: an illustration from Accra, Ghana.

Most analyses of differentials and trends in health and mortality find the anticipated relationship between measured outcomes and a range of socio-economic status variables including education, income status of employment and wealth. From a multi-year longitudinal study of health and mortality in Accra called the Women’s Health Study of Accra (WHSA), we produce a much finer grained analyses of health and mortality differentials within an urban area (Weeks, Hill et al. 2006; Douptcheva and Hill 2011). First, in some of the communities designated as ‘slums’ by UN Habitat, child mortality and adult morbidity are surprisingly good. The socio-demographic profile of the population, shaped heavily by migration, modifies the risk profile (Fink, Weeks et al. 2012). Secondly, the anticipated differentials in both self-reported and objectively measured dimensions of adult health are very narrow and often in the opposite direction anticipated (Darko, Adanu et al. 2012). Thirdly, there is evidence of pronounced neighbourhood effects where even for those poorer than the mean for the whole neighbourhood, place of residence can mitigate some of the otherwise adverse health outcomes associated with poverty (Weeks, Getis et al. 2010).These empirical facts present a challenge to conventional theory on the determinants of health outcomes.