Intro
Many countries, including all high-income ones, maintain vital registration systems that provide data on the number of deaths by cause, sex, and age. Some countries also report years of life lost because of premature mortality (YLL) due to each cause, a number that depends on the age of death and on the choice of an algorithm for how YLL should depend on the age of death. The tracking of stillbirths, however, is often incomplete and variable.
As of the early 1990s, no estimates of YLL were available for many developing countries or for regional groupings of such countries. The World Bank (1993), as part of the preparation for its World Development Report 1993: Investing in Health, initiated an effort to provide estimates of deaths by age and cause, and hence YLL, for around 100 conditions for eight regional groupings, including all low- and middle-income countries. By adding years of healthy life lost as a result of disability (YLD) to YLL, the World Bank was able to generate estimates of the global burden of disease measured both in deaths by cause and in disability-adjusted life-years (DALYs) (Murray, Lopez, and Jamison 1994; World Bank 1993, appendix B). Murray and Lopez (1997) provide updated and extended results and a complete description of methods. Global burden of disease estimates have subsequently been used to help guide resource allocation in the health sector and to inform debates about national and international disease control priorities (see chapter 1 in this volume); however, the global burden of disease literature currently provides little insight into the importance of deaths near the time of birth.
The purpose of this chapter is to explore the sensitivity of results within the Global Burden of Disease (GBD) framework to alternative approaches to encompassing the large number of deaths that occur near the time of birth, namely almost 4 million neonatal deaths and 3.3 million stillbirths. The sensitivity analyses in this chapter thus complement those of chapter 5, which explore the effect of variations in discount rates, age weights, and disability weights. Chapter 3 in this volume describes the GBD framework and provides estimates of deaths and DALYs by cause for 2001 using the World Bank regional grouping of countries. (Map 1, inside the front cover of this volume, shows the World Bank regional groupings used throughout this book.) This chapter uses the same framework and numbers to the extent possible, but with the following exceptions:
-
We divide the newborn through age 4 category into neonatal (newborn through 27 days), postneonatal (28 days to less than 1 year), and child (1 through 4 years).
-
We aggregate the 136 causes noted in chapter 3 into 35 causes.
-
We allocate the substantial number of neonatal deaths attributed to pneumonia or sepsis to the chapter 3 category of respiratory infections.
-
We explore the sensitivity of the results in chapter 3 to adding stillbirths as a new age category.
-
We explore the sensitivity of the results to alternative ways of assigning YLL to deaths under the age of five.
The first section of this chapter deals with mortality: all-cause and cause specific. It uses the results presented in chapter 3, but adds to them estimates of the level of stillbirths and of the level and causes of neonatal mortality. The second section deals with estimation of the burden of disease in DALYs. The inclusion of stillbirths in the analysis highlights the more general issue of how to deal appropriately with deaths at different ages in constructing a measure of YLL.
As emphasized throughout this volume, data on causes of death and disability are fragmentary and are often inconsistent for many regions of the world. This is particularly true for the neonatal period and for stillbirths. One clear implication is the desirability of more and better data. Another implication is that any effort to construct an overall picture of population health must aggregate data of variable, often low, quality and completeness. In some instances this is done essentially as a political process, with various disease advocacy groups advancing their claims to policy makers and in the press. Alternatively, summary measures can be constructed systematically in a way that eliminates internal inconsistencies, describes methods carefully, and imposes the discipline of demographically derived totals into which cause-specific estimates must fit. This is the nature of our work on the global burden of disease.
