Generating the numbers

Generating global estimates of the burden of diabetes and impaired glucose tolerance (IGT) relies heavily on the availability of data from studies and surveys conducted in communities all over the world. These raw data provide a basis for modelling estimates at a national and global scale.

Sources review

The data used in this edition of the IDF Diabetes Atlas came from a variety of sources, including peer-reviewed literature, national and regional health surveys, personal communication provided from investigators in the IDF network and official reports by multinational organisations, such as the United Nations or the United States Centers for Disease Control and Prevention. A total of 565 data sources were reviewed for the estimates of diabetes in adults and IGT, of which 170 sources from 110 countries were included. An additional 88 sources were selected for the estimates of diabetes in the young.

After the selection process, the information gathered from the data sources is transformed to a standard format using statistical software that fills in any gender and age gaps. The estimates for diabetes in adults take into account the proportion of people that live in urban areas and also correct for the number of expected undiagnosed cases. The estimates of prevalence (%) and incidence are then used to determine the number of people with diabetes using the latest population data available from the UN Population Division for each country and territory. The same method is used to generate estimates for 2011 and 2030. Projections take into account changes in population structure and urbanisation, but do not explicitly include changes in the prevalence of diabetes risk factors.

To estimate the mortality due to diabetes, two sets of data are used. The first are World Health Organization estimates of the total number of deaths in each country. The second are published regional estimates of the relative risk a person with diabetes has of dying compared to those without diabetes. That risk is then applied to total mortality estimates taking into account the prevalence estimates for diabetes.

Finally, mortality and prevalence data are used to estimate the healthcare expenditures due to diabetes. Total healthcare expenditures for each country are taken from publicly available data. These numbers are then used together with information on the ratio of expenditures for people with diabetes compared to those without, adjusting for the number of people with diabetes in a country and adjusting for increased mortality-related costs.

A number of countries do not have information on diabetes available. When there are no data available for a given country, information is taken from other countries matched on ethnicity, income level, and geography.

The lack of reliable data is a particular problem for low- and middle-income countries, but even some high-income countries do not have recent data. In addition, reliable mortality estimates and reporting on health expenditures are also vulnerable to the availability of good data.

Barriers to finding usable data are not only tied to source quality, but to the way in which data are presented in the data sources. For example, data may be presented separately by age group and by gender, but ideally the data would be presented by age group for each gender.

All the estimates presented in the IDF Diabetes Atlas are sensitive to the quality of the data behind them. Efforts were made to contact investigators and validate numbers whenever needed. However, data and surveillance for diseases like diabetes must be representative, and regularly improved to ensure the estimates derived are reliable and sound.