The second workshop session, which featured a presentation by Gilles Pison, focused on changes in the patterns and trends of mortality and causes of death in sub-Saharan Africa. Work done by Masquelier, Reniers, and Pison prior to the workshop showed that, after decades of declining mortality rates in sub-Saharan Africa during the latter half of the twentieth century, such progress is no longer the norm. Several countries have experienced important setbacks which can be traced to such factors as political instability, the advent or resurgence of infectious diseases, and possibly also an increase in mortality rates from noncommunicable diseases. The lack of full-fledged vital registration systems in most sub-Saharan countries has made it difficult to perform detailed assessments of mortality rates. Adult mortality measurements in particular are difficult to perform in the absence of vital registration statistics, and for that reason most United Nations (UN) agencies estimate adult mortality by matching indices of child survival to model age patterns of mortality.
Pison provided an overview of the trends in mortality among children under age 5 (5q0) and for individuals between the ages of 15 and 60 (45q15) in sub-Saharan Africa which were based on data on the survival of children and siblings collected in Demographic and Health Surveys. They used generalized additive models to smooth mortality rates and made adjustments to correct for recall bias. Except for some conspicuous periods of stagnation in several countries in the 1990s, child mortality mostly declined over the past 30 to 40 years. Adult mortality patterns, by contrast, were much more diverse. Most Eastern and Southern African countries witnessed enormous surges in adult mortality (with 45q15 > 0.67 in some populations) that were clearly correlated with earlier increases in HIV incidence. In Eastern Africa, adult mortality levels have begun to decline, and, surprisingly, in some cases that happened even before the large-scale expansion of antiretroviral therapy programs in 2005. Even more surprising is the lack of sustained improvements in adult survival that have been observed in some African countries without severe HIV epidemics.
Most sub-Saharan African countries have registered considerable progress in child survival over the past 30 to 40 years but are still far removed from reaching the UN Millennium Development Goals. Overall the region has been characterized by an increasing convergence in mortality rates for children under age 5. Adult mortality rates and trends are characterized by a much greater heterogeneity and sometimes move in opposite directions from the child mortality trends in the same country. AIDS mortality is dominating many of these all-cause mortality patterns, but the first indications of a reversal in adult mortality rates in Eastern African countries are now visible.
During this session Pison discussed what might be learned about child mortality from the experience of Senegal, a country with low AIDS prevalence. He noted that Senegal experienced a rapid decrease in child mortality in the 1970s and 1980s followed
by a period of no change in the 1990s and then a return to progress in the 2000s. The earlier decrease was due largely to success with vaccination programs, he said, while the lack of improvement in the 1990s can be attributed to increases in malaria mortality and to stagnation in vaccination efforts. Renewed efforts to combat malaria (new treatments, rapid diagnostic tests, and impregnated bed nets) and progress in vaccine coverage were largely responsible for the resumption of the decline in mortality during the past decade. The presenters acknowledged that improvements in socioeconomic conditions might also have had an influence, but the results of that influence would be visible only over the long term and would not explain relatively rapid changes such as the mortality decline in the 2000s. A key takeaway point of the session was that vaccinations and efforts to combat malaria have “non-specific” effects, meaning that in addition to lowering mortality due to the specific disease being addressed, they also tend to lower mortality due to other causes.