Can we achieve Millennium Development Goal 4? New analysis of country trends and forecasts of under-5 mortality to 2015

The researchers found that global under-5 mortality has fallen from 110 per 1000 in 1980 to 72 per 1000 in 2005. Child deaths worldwide have decreased from 13.5 million in 1980 to an estimated 9.7 million in 2005. However, despite this progress, global under-5 mortality is expected to decline by only 27% from 1990 to 2015, substantially less than the MDG4 target.

 

Looking at child mortality compared to neonatal and post-neonatal mortality, the researchers also found that from 1970 to 2015, the estimated annual number of childhood deaths dropped by 1.7 million (44%), the number of post-neonatal deaths fell by 3.5 million (47%), and the number of neonatal deaths decreased by 1.3 million (26%). By 2015, it is expected that neonatal deaths will make up 37% of all under-5 deaths, up from 31% in 1970.

 

Regional analysis indicates that several regions in Latin America, North Africa, the Middle East, Europe, and Southeast Asia have had consistent annual rates of decline in child mortality in excess of 4% over 35 years. By 2005, child deaths in the Americas, Europe, and the Middle East accounted for only 7.2% of all child deaths. Asia and South Asia experienced significant declines in the number of deaths from 1990 to 2005, while the subregions of West, East, and Central Africa had a rise in death rate by more than 25% between 1970 and 2005, which can be attributed to minimal reductions in the total fertility rate and rising total population. Global progress on MDG4 is dominated by these slow reductions in sub-Saharan Africa, where it is estimated that 56% of all under-5 deaths will occur by 2015. 

 

According to the researcher’s estimates, 49 countries (17 high-income, 28 middle-income, and four low-income) have a greater than 75% chance of achieving MDG4, and 94 countries (13 high-income, 37 middle-income, and 44 low-income) have a less than 20% chance of meeting the target. 

Researchers merged available child mortality databases maintained by UNICEF and WHO as of April 2007. Duplicate measurements from these two sources were removed. Additional data, including registration system data, surveys, and censuses were included from other sources. The resulting database for analysis included 8,243 empirical observations for 172 countries. The researchers applied Loess regression to estimate past trends and forecast to 2015 for all 172 countries. They also developed uncertainty estimates based on different model specifications and estimated levels and trends in neonatal, post-neonatal, and childhood mortality. 

Annual or periodic assessments of global child mortality are produced by UNICEF, WHO, the United Nations Population Division, and the World Bank. These assessments are based on a mixture of data sources, including complete vital registration systems for some countries, partial registration systems, complete birth histories from the Demographic and Health Surveys or similar surveys, survey or census questions, and sample registration systems. However, these organizations have not always agreed on past estimates, predictions for a recent period, or forecasts. Furthermore, these organizations often fail to use all data available, do not use transparent and reproducible methods, do not distinguish predictions from measurements, and provide no indication of uncertainty around point estimates. Sound measurement is needed so that we can learn from national success and identify countries where extra efforts are needed to accelerate the rate of decline in under-5 mortality. The researchers, therefore, aimed to develop new reproducible methods and reanalyze existing data to elucidate detailed time trends.

This study highlights the need to further improve the quality and timeliness of child-mortality measurements by using existing datasets more comprehensively and applying standard analytical strategies. Ensuring that all available evidence is used for global monitoring and that better data collection efforts are conducted should be a key priority. Additionally, considerably more research is needed to investigate effectively the possibility of regional policy effects that impact mortality rates. For example, common patterns in Latin America and the Middle East suggest that regional convergence has taken place of various policies and practices, perhaps mediated through common language, regional institutions, or informal policy networks. Likewise, when diversity of rates of decline happens within regions with similar culture, history, and institutions, opportunities could exist to identify effective policy adoption and implementation. Finally, further investment in measurement of cause-specific mortality in sub-Saharan Africa is urgently needed if policy questions are to be answered.