Sources of variation in under-5 mortality across sub-Saharan Africa: a spatial analysis

The ongoing decline in under-5 mortality ranks among the most significant public and population health successes of the past 30 years. Deaths of children under the age of 5 years have fallen from nearly 13 million per year in 1990 to less than 6 million per year in 2015, even as the world's under-5 population grew by nearly 100 million children. However, the amount of variability underlying this broad global progress is substantial. On a regional level, east Asia and the Pacific have surpassed the Millennium Development Goal target of a two-thirds reduction in under-5 mortality rate between 1990 and 2015, whereas sub-Saharan Africa has had only a 24% decline over the same period. Large differences in progress are also evident within sub-Saharan Africa, where mortality rates have declined by more than 70% from 1990 to 2015 in some countries and increased in others; in 2015, the mortality rate in some countries was more than three times that in others.

What explains this remarkable variation in progress against under-5 mortality? Answering this question requires understanding of where the main sources of variation in mortality lie. One view that is implicit in the way that mortality rates are tracked and targeted is that national policies and conditions drive first-order changes in under-5 mortality. This country-level focus is justified by research that emphasises the role of institutional factors in explaining variation in mortality—factors such as universal health coverage, women's education, and the effectiveness of national health systems. It is argued that these factors, which vary measurably at the country level, fundamentally shape the ability of individuals and communities to affect more proximate causes of child death such as malaria and diarrhoeal disease.

An alternate view has focused on exploring the importance of subnational variation in the distribution of disease. In the USA, studies on the geographical distribution of health care and mortality have been influential for targeting of resources and policy design. Similar studies in developing regions have shown the substantial variability in the distribution and changes of important health outcomes such HIV, malaria, and schistosomiasis—information that can then be used to improve the targeting of interventions. Nevertheless, the relative contribution of within-country and between-country differences in explaining under-5 mortality remains unknown. Improved understanding of the relative contribution of national and sub-national factors could provide insight into the drivers of mortality levels and declines in mortality, as well as improve the targeting of interventions to the areas where they are most needed.