EDITORIALS
Health statistics: time to get serious
Kenji Shibuya; Suzanne Scheele; Ties Boerma1
Measurement and Health Information Systems, Evidence and Information for Policy, World Health Organization, 20 Avenue Appia, 1211 Geneva, Switzerland
Health statistics, including both empirical data and estimates related to health such as mortality, morbidity, risk factors, health service coverage, and health systems, are the basis for every aspect of health planning. The demand for better health statistics is rapidly growing. More money is being spent on global health than ever, and donors are keen to know their return on investment. In the financial sector, credibility and accountability are everything. Why not in the health sector? Tracking progress towards the Millennium Development Goals and performance-based funding, are promoting greater demand. We cannot afford to continue "stumbling around in the dark" any longer (1).
The supply side is not sitting still either. The Health Metrics Network, launched in May 2005, is a new international partnership that aims to improve health information at all levels primarily by strengthening health information systems in countries. The Ellison Institute, affiliated with Harvard University and expected to be launched in early 2006, will focus mainly on "improving world health through accountability" (2). There is now a great opportunity to improve both the quality and quantity of health statistics. The next key step for WHO is to better define its role among the rapidly proliferating health statistics constituency.
First and foremost, WHO's comparative advantage must be reconsidered. The Organization's work in health statistics needs to build upon its constitutional and legitimate links with Member States and its convening power to reach consensus and facilitate harmonization at the country and regional levels with lead partners in the health field.
The flip side of this comparative advantage is the intense political pressure to which WHO figures are subject and the often poor quality of country-level data provided by its Member States upon which estimates are based. Recently, the appropriateness of WHO as a global health monitoring body has been questioned due to its multiple roles in advocacy, technical assistance to countries and monitoring and evaluation, against a backdrop of political links with its Member States (3).
WHO's role in health statistics has always been the subject of debate. In the 1980s the emphasis was primarily on reporting of country data and there was very little attempt to adjust for bias and to fill gaps in data. During the 1990s, substantial progress was made in the analysis of the global and regional burden of disease and in the development of composite health measures, relying heavily on modelling. Since 2003, WHO has put more emphasis on strengthening country data and information systems, and developing partnerships with other organizations and institutions. Better empirical data support better modelling efforts and vice versa. We need both to complement each other.
International agencies are fully aware of the need to harmonize health statistics. For example, WHO, UNICEF, the World Bank and the UN Population Division are working on child mortality estimates. WHO is developing one-stop-shop access to publications and websites for core health statistics, such as the World Health Statistics publication (4).
To ensure accuracy and transparency of health statistics, WHO is improving its approach to producing estimates, particularly at country level. A country consultation process has been in place since 2001, following the publication of the World Health Report 2000, which resulted in criticisms concerning the credibility of WHO estimates. The consultation process is now supported by a four-step framework when clearing official WHO estimates: 1) a publicly accessible database of all data sources; 2) independent review by a group of experts such as the Child Health Epidemiology Reference Group (CHERG) and the UNAIDS Reference Group on Estimates, Modelling and Projections; 3) well-documented, preferably peer-reviewed and published, methods of estimation; and 4) internal WHO clearance by the Evidence and Information for Policy (EIP) cluster.
Countries are the major producers and users of health information. WHO will need to step up its efforts to provide assistance to countries. This involves the strengthening of the availability, quality and uses of health information, and building country capacity for modelling and generating estimates. While the country consultation and clearance procedures place constraints on the timeliness of health statistics, the engagement of countries and subsequent strengthening of their capacity to produce reliable statistics should eventually improve the efficiency of production and quality of estimates.
WHO statistics are produced by disease-specific programmes as well as by the EIP cluster. The experts of individual programmes vary and their work is often fragmented. Strengthening internal capacity across programmes is a necessary step to maintain WHO's leading role as the global health monitoring body. However, given an increasing number of international players in the field of health statistics, WHO should simultaneously strengthen its role as a facilitator and coordinator of leading experts in health statistics. One example is the recently established WHO High-Level Panel on Health Statistics (5). If estimation of health statistics can be done jointly with academics and other organizations and institutions, WHO should be a fully-fledged partner in the process of generating health statistics.
Finally, the value of information is judged by those who use it, not those who produce it. Now that the field of health statistics has become more visible, the success or failure of this work in the international health arena will depend greatly on how WHO asserts itself with its member countries and its research partners. It is truly time to get serious. 
References
Web version only, available from www.who.int/bulletin
1 Correspondence should be sent to Ties Boerma (email: boermat@who.int).