The Demographics: Mortality and Population research team is developing and advancing methods for quantifying fundamental indicators of population health. Through country-specific estimates of premature mortality by age and sex, we can begin to address the causes of premature death. The ideal method to capture death data is a vital registration system where all deaths in a population are documented. However, less than half (approximately 44%) of the world’s population has their births and deaths recorded in a vital registration system.
In the absence of such systems, mortality estimates must be derived from other sources of data, such as censuses and surveys. IHME compiles and synthesizes data across each of these types of sources and systematically generates estimates of mortality levels for 187 countries since 1970. Providing comparable estimates of mortality rates is necessary to help inform global and national health policy and research aimed at preventing premature deaths in both children and adults.
Key Activities
- Improve demographic estimation methods
We are developing new methods to improve data collection, analysis, and estimation, particularly in the areas of survey design, forecasting, synthesis of data from multiple sources, and single-source data analysis (including death distribution, birth histories, and sibling/parental survival). For example, we have fielded household surveys as part of the Population Health Metrics Research Consortium (PHMRC) in an effort to refine sibling survival and parental survival methods, and have developed and tested alternative methods for mortality forecasting.
Currently, IHME relies upon population estimates generated by the United Nations Population Division. To ensure the highest internal consistency among results, IHME must produce its own population estimates. The research team will draw upon lessons learned from developing the methods used to derive mortality estimates and incorporate these advancements into improved methods for population estimation. The team will need to undertake a full systematic analysis of fertility data to produce these estimates.
- Produce subnational mortality estimates
Subnational estimates of mortality by age and sex are an important input into national policy debates. In Brazil, we generated infant mortality estimates at the microregion and municipality levels from 2000 to 2010. We faced two main methodological challenges: vital registration is incomplete in many places, and the small size of many microregions and municipalities may lead to unstable estimates of mortality rates. The research team addressed the gaps in these data by comparing under-5 mortality from surveys and censuses with data from birth and death registries and using a small area model to make mortality rate estimates more reliable, even with small sample sizes.
Given the health disparities that exist within large populations, the Demographics: Mortality and Population research team is currently producing subnational mortality estimates for several countries, including Zambia, India, China, and the United States. We expect that significant differences will be revealed from one locality to the next within each of these countries, highlighting inequalities resulting from a range of factors, such as socioeconomic and geographic indicators.
- Create a visualization tool that allows users to assess and compare mortality estimates and data sources
The research team will design a data visualization tool, conceptually similar to iCoD, that allows the user to examine for each country the data that have gone into estimates by age and sex. To the degree possible, data sources will be labeled with full citation information and, to the degree possible, original data sources will be accessible to the user. This tool will both serve internal analytic purposes and help researchers to explain results to external audiences.
- Continue to provide timely updates to adult and child mortality estimates by country and five-year forecasts
As in past years, IHME will continue its routine updates to adult and child mortality estimates by country, using the highest number possible of newly accessible datasets. Going forward, IHME intends to publish updates in appropriate policy reports, using established methods.
- Complete a subnational analysis of mortality for India and Zambia
In both India and Zambia, the research team will finalize subnational estimates, aiming to calculate mortality at the district level in both countries. The Zambia results will be presented to the Ministry of Health as part of the ongoing collaboration in that country, and the India estimates will be computed in collaboration with colleagues at the Public Health Foundation of India.
- Evaluate options for developing forecasting models for mortality
The research team will conduct a preliminary assessment of possible approaches to explore better forecasting models. The ultimate desire is to forecast mortality and causes of death 30 years into the future. The scope and possible parameters of this task will be determined in FY13, with the actual methods development implemented thereafter.
Impact
We continue to refine new demographic methods to provide systematic and comparable mortality estimates for the global health community. In the future, we plan to use these mortality estimates along with other data sources to produce a series of population estimates.