What do we know about health in Africa, and how do we know it? In this Leadership Forum, Eduard R. Bos, Lead Population Specialist, discusses the challenges to improving the health of the population in Sub-Saharan Africa. More attention is now focused on improving the health of the population in Sub-Saharan Africa than at any previous time. In the years since the 1991 publication of the first Disease and Mortality in Sub-Saharan Africa (DMSSA), numerous reports have been issued that address the health status of African populations and propose strategies to combat poor health with improved access to more effective health services. Prime among these was the World Bank's 2005 report Improving Health, Nutrition, and Population Outcomes in Sub-Saharan Africa - The Role of the World Bank, which gave rise to the publication of the second edition of the DMSSA (2006). In the just released second edition of the DMSSA, produced jointly by the World Bank and the South African Medical Research Council, a team of seven editors1 coordinated the inputs of more than 70 authors to update the epidemiological evidence underlying the strategies of national governments and development partners to prioritize interventions and monitor results. The volume documents that, despite the multitudes of reports and expanded investments in health over the past decade, life expectancy for the continent as a whole has decreased by almost five years since the 1991 publication. The main objective of the second DMSSA remained the same as that of the first edition: to provide policy makers with a comprehensive resource of information on the levels and trends in mortality and main causes of ill health. Without knowledge of disease incidence and prevalence, setting policies for prioritizing interventions risks misallocation of resources to combat causes of ill health that contribute little to the overall health status of a population. Good epidemiological information does not ensure good policy or effective implementation, but without reliable epidemiological data, efforts to design cost-effective strategies are made in a vacuum. Successful disease control interventions, such as a multi-country effort in southern Africa to eliminate measles mortality or onchocerciasis control programs, share an emphasis on disease surveillance and monitoring of interventions. But how does one obtain good epidemiological data? In the statistically developed high-income countries, aggregated reports of notifiable infectious diseases combined with surveillance and registration of major non-communicable diseases are the main instruments for estimates of the incidence and prevalence of diseases and conditions. Civil registration data of deaths by age and by cause completed by physicians are the basis for age-specific and disease-specific mortality estimates, as well as for derived mortality indicators such as the infant mortality rate or life expectancy. In high-income countries, estimates of the population at risk are available from population registers, or from projected census counts. In Sub-Saharan Africa, nationally representative estimates of the burden of disease and mortality from such sources do not exist in any country. Just two (Mauritius and the Seychelles) provide vital registration data that are deemed by the United Nations Statistics Department to be at least 90 percent complete. In the absence of such routine data collection systems, information on disease and mortality in Sub-Saharan African countries is based on the following sources, often used in combination to obtain a more complete picture: Household surveys, such as the Demographic and Health Surveys (DHS), the World Bank’s Living Standard Measurement Surveys (LSMS), and UNICEF’s Multiple Indicator Cluster Survey (MICS), have been used in many African countries to measure infant and under-five mortality by asking respondents about the survival of their children. Such interviews can yield estimates for as many as 15 years before the survey, and can therefore be used to compare across surveys conducted at different times. But due to limitations in the sample size, single-year mortality indicators cannot be precisely estimated, so that even the most recent survey reports averages of 3 to 5 years before the survey was conducted. Nevertheless, the proliferation of household surveys conducted since the first edition of DMSSA has vastly improved the data for infant and child mortality; the new volume includes an analysis of 153 data sets covering 42 countries in Sub-Saharan Africa. Household surveys are less practical for the measurement of adult mortality, or cause-specific mortality rates, including maternal mortality indicators. DHS have included modules to collect adult mortality rates by asking respondents to report on the survival of siblings, but, due to the relatively small number of deaths, the mortality rates computed from these reports are usually quite imprecise. Surveys such as the DHS have provided many other pieces of empirical evidence referred to in DMSSA. This includes the prevalence of childhood diseases, such as diarrheal disease and acute respiratory disease, as well as information on risk factors for several correlates of health outcomes, such as malnutrition indicators, immunization coverage, or maternal health, such as the use of antenatal and delivery care. Some household surveys have provided key information on risk factors associated with non-communicable diseases, including tobacco use, obesity, and alcohol consumption. More recently, household surveys have added HIV testing of blood samples from respondents, allowing cross tabulation of HIV prevalence with demographic characteristics and behaviors. Demographic surveillance sites (DSS) are used to monitor vital events (births and deaths) of a registered population in a geographically defined area where complete national vital registration is unavailable. Seventeen DSS from across Sub-Saharan African countries are linked by the INDEPTH network; typical DSS populations include at least 60,000 individuals per site. This approach uses continuous cycles of enumeration to identify births, deaths, and migration by age and sex to estimate fertility and mortality rates, and has resulted in the first set of empirical life tables for African populations. In addition, so called “verbal autopsies” (reports of family members on disease symptoms and other signs of recent deaths in a household) have been collected in a number of the sites to obtain information on the causes of death, since physician provided death certificates are not available. The INDEPTH DSS have provided a wealth of comprehensive information, adding to knowledge of age patterns of mortality and changes in causes of death. Their limitation is that the sites are still few in number and are not selected randomly, and therefore cannot be used for generalizing findings or constructing national estimates. Surveillance of diseases and conditions. In contrast to household survey data collection, which is generally carried out several years apart, surveillance refers to the systematic collection and analysis of disease data on a continuous basis. Given the weaknesses of African health systems, routine surveillance data for detecting the spread of infectious diseases that could develop into large-scale outbreaks (such as measles or cholera) are often incomplete. Sentinel surveillance sites, which consist of geographically dispersed health facilities designated to collect disease information, such as antenatal clinics equipped to test for HIV among pregnant women, have provided more useful information: most of what is known about the spread of the HIV/AIDS epidemic across Africa in the 1990s is based on data obtained from sentinel surveillance sites. Surveillance can also refer to tracking the incidence and mortality of non-communicable diseases and their risk factors, as is being done with cancer registries in Kampala, Uganda, the Zimbabwe Cancer Registry in Harare, and the Malawi Cancer Registry in Blantyre. Such registries currently cover just 8 percent of the population in Africa. Injury mortality registries exist in some South African cities, as well as in Kampala. Other health facility-based surveillance for non-communicable diseases and conditions, sometimes called “passive surveillance” is much less useful, as relatively few people have access to hospitals. Nevertheless, much of what is known about prevalence of diabetes and cardiovascular diseases in Sub-Saharan Africa is based on such information. Small areas and selected subpopulations have frequently been used to assess incidence of specific causes, such as developmental disabilities in school children, or through community-based surveillance for infectious diseases such as diarrheal diseases. While the locations and subpopulations are rarely representative of a country’s entire population, such studies have provided a wealth of data on disease incidence and mortality patterns. Modeling and triangulation of conflicting data is frequently necessary to extrapolate from small area studies, incomplete surveillance data, or because of conflicting data from different sources. For example, maternal mortality indicators are estimated on the basis of fertility rates and the proportion of births with skilled delivery attendance; and tuberculosis incidence estimates are based on estimates of the completeness of notification rates. HIV prevalence rates are adjusted for urban bias, as antenatal clinic sentinel surveillance sites are more commonly located in urban areas where HIV prevalence tends to be higher. Estimates of child mortality collected with different procedures in household surveys frequently show inconsistencies over time, requiring adjustments to the data. The DMSSA volume discusses how these different data collection and analysis strategies are being used to improve knowledge of the epidemiology of Sub-Saharan African populations. It reflects the growth in technical capacity for data collection and analysis of health statistics since its first edition in 1991. Many new sources of health and demographic information have become available as a result of unprecedented international interest in health conditions in Sub-Saharan Africa. Notwithstanding these advances in health statistics, a theme that emerges from all the chapters in the volume is that not enough is known about trends in diseases and conditions in order to monitor and evaluate the effectiveness of programs intended to produce better health outcomes. The continued improvement of disease surveillance and other regularly published health information remains an important priority for African health systems. Top 1 Dean Jamison, Professor, Institute for Global Health, University of California, San Francisco; Richard Feachem, Executive Director, Global Fund to Fight AIDS, Tuberculosis, and Malaria; Malegapuru Makgoba, Vice-Chancellor and Principal, University of Kwa-Zulu Natal, South Africa; Eduard Bos, Lead Population Specialist, The World Bank; Florence Baingana, Senior Health Specialist, The World Bank; Karen Hofman, Director, Division of Advanced Studies and Policy Analysis, Fogarty International Center, National Institute of Health, Washington, DC; Khama Rogo, Lead Specialist, The World Bank INDEPTH. 2002. Population and Health in Developing Countries. Volume I: Population, Health, and Survival at INDEPTH sites. Ottawa: International Development Research Center. Jamison, D. et al. (eds.) 2006. Disease Control Priorities in Developing Countries. Second Edition. Disease Control Priorities Project. Washington, DC , and New York : The World Bank and Oxford University Press Levine, R., Kinder, M., et al. 2004. Millions Saved. Proven Successes in Global Health. Washington, DC : Center for Global Development. Lopez, A. et al. (eds.). 2006. Global Burden of Disease and Risk Factors. Disease Control Priorities Project. Washington, DC , and New York : The World Bank and Oxford University Press.
|
RSS