Ties Boerma and I have just published a short paper in BMC Medicine (http://www.biomedcentral.com/1741-7015/13/50) which discusses the role of WHO and other global health agencies in the preparation and publication of global, regional and country-level health statistics. There has been considerable recent debate around two key questions:

• Our member states want to know why we do not simply publish their latest national statistics, and in some cases, can be quite unhappy that we are not doing so
• Some academic groups also publish global health statistics in research journals, and recently, the Institute for Health Metrics and Evaluation, funded by the Bill and Melinda Gates Foundation at the University of Washington in Seattle, has been publishing comprehensive health statistics which disagree substantially in some areas with those of the UN agencies.

Our just-published paper builds upon previous publications and commentaries on global health statistics, for example:

• Rudan I, Chan KY. Global health metrics needs collaboration and competition. Lancet. 2015 Jan 10;385(9963):92-94.

• Boermat T, Mathers C, Abouzahr C. WHO and global health monitoring: WHO and the way forward. PLoS Med 2010, 7(11): e1000373.

• , Chan M, Kazatchkine M, Lob-Levyt J et al. Meeting the demands for results and accountability: a call for action on health data from eight health agencies. PLoS Medicine 2010,7(1): e100023

• Stein C, Kuchenmueller T, Hendrickx S, Pruess-Ustun A, Wolfson L, Engels D, Schlundt J. The global burden of disease assessments – who is responsible? PLoS Med 2007, 1(3): e161.

The production and dissemination of health statistics for health action at the country, regional and global levels are core WHO activities mandated by the Member States in the Constitution. WHO figures carry great weight in national and international resource allocation, policy-making and programming, because of WHO’s reputation as being unbiased (impartial and fair), global (having a worldwide remit and responsibility) and technically competent (drawing on leading research and policy institutions and individuals). WHO works closely with countries, partners and global experts to produce health statistics of the greatest possible accuracy. The estimates have been of considerable value in generating an overview of the health situation and emerging trends and for reporting on country and global progress towards international goals and targets such as the MDGs.

However, data are often lacking, especially in the low and middle income countries where many health problems are typically the biggest. To fill the gaps, statistical modelling is frequently used to produce comparable health statistics across countries that can be combined to produce regional and global statistics. We work closely with a number of academic collaborating groups and with other agencies to prepare such statistics, and they undergo a process of consultation with Member States before release.

The country consultation process provides a platform for member states to understand how estimates are derived, and for WHO to identify additional data sources that can be used to improve the accuracy of estimates. It is a consultation process, not a clearance, meaning that WHO and country best estimates may differ because of differences in data used and methodology. WHO aims to produce comparable statistics (ie. Consistent methods, data and assumptions are used across groups of countries) and to adjust appropriately for known biases (for example, incomplete recording of numbers or causes of death). National statistics may use a variety of methods, and in some cases, do not take biases completely into account. For these reasons, WHO statistics may not always be consistent with official national statistics, although the differences are usually very minimal for countries with high quality data.

Statistical estimates for populations always have uncertainty and the fewer quality data points, the greater the uncertainty. Different researchers can easily come up with different estimates for the same country, region or globally and this has happened on many occasions. In recent years, the Institute for Health Metrics and Evaluation (IHME) at the University of Washington has started to publish estimates for many health indicators globally and for countries. Sometimes these estimates are very different from those published by WHO and UN agencies. Our paper gives several examples of this:

• Malaria deaths for ages 5+: 655,000 (IHME) compared with <100,000 (WHO).
• Child tuberculosis cases: < 200,000 (IHME) compared with 530,000 (WHO)
• Older child deaths (ages 5-14 years): 817,000 (IHME) versus 1.44 million (WHO).

WHO reviews methods and estimates developed by IHME and other organizations, and may make use of them in cases where scientific rigor can be evaluated. For instance, the publication of the IHME Global Burden of Disease (GBD) 2010 Study in 2012 led to consultations and exchange of data between IHME and WHO and other UN agencies and their expert groups. Some results from the IHME GBD 2010 study have also been used by WHO in preparing its global health statistics (www.who.int/gho).

The existence of divergent estimates for the same indicator has led to increased awareness of major data gaps, especially in low and middle income countries. There are now considerable efforts underway in WHO, World Bank and other global agencies to mobilize resources and policy focus on improving death registration systems in low and middle income countries, and in addressing other gaps in health information.

Globally, the existence of multiple estimates for the same indicator has led to some concerns among the global health agencies and donors. Countries seem less worried: to-date no country has challenged WHO draft estimates during the country consultation process on the basis of the existence of a competing estimate. In several areas there is convergence in terms of methods and results of the estimation modelling. Examples include child mortality, maternal mortality and etiology of pneumonia. In others, more work is needed to discuss data inputs, methods and discrepant results.

I do not regard this situation as necessarily competitive. WHO and other UN agencies will continue to prepare and report on global health indicators to fulfill their mandate from Member States, and to be accountable to those States through a transparent process, reproducible methods, and country involvement. Academic inputs are needed to improve data collection, compilation and sharing, analytical methods, and communication of global health indicators. For many years this has almost exclusively occurred in the context of WHO or UN expert groups, and now this work is also taking place in independent academic research institutions. The resulting debates on data interpretation, methods and results, can be healthy and productive and will hopefully lead to improvements in data and in methods and results.