Paul Kind attended the Joint United Nations Economic Commission for Europe and World Health Organization Meeting on Measuring Health Status co-hosted by Statistics Canada, held in Ottawa, Ontario, Canada, 2000 October 23-25 as the ISPOR representative.

Why would a meeting in Ottawa of 41 'experts' from several countries1 together with representatives of OECD, World Bank, WHO and the UN Statistics Division, be of interest to ISPOR or its membership? Up for discussion was the subject of the measurement of population health. Not of immediate interest, it might be thought, for those facing the difficult task of designing or running clinical trials, or those of us with a fascination (fixation one might say) for the measurement of outcomes in health care.

At the heart of the enterprise was the worthy objective of standardising the way that population health is measured in national surveys. A recurring message was the need for a summary measure for use in informing national policy and for cross-national comparisons. Government agencies were invited to support the principle of standardised health status measurement in WHO member countries and there appeared to be little or no dissent from that general principle. The value of international league tables is highly questionable at the best of times, providing little beyond newspaper headlines and politician's sound-bytes. This is especially so where questionable methodologies are used to construct such tables - last year's WHO report being a case in point.

Mercifully however, US and UK spokesmen indicated that there was little short-term prospect of significantly modifying national health surveys to take account of any WHO requests in this area. So what then, is the problem? There are two aspects to this that ought to concern us - as citizens and as scientists - since the political and scientific agendas have been cross-contaminated. Developing new health status measures, whether these are targeted or generic in nature, involves a significant investment research. Amongst those present at the Ottawa meeting were lead investigators responsible for the development of SF-36, QWB, HUI and EQ-5D. All could bear witness to the lengthy and difficult path that researchers have followed in producing instruments that are recognised as having demonstrable scientific pedigree. This process involves submitting material for proper public scrutiny and comment - not presenting the scientific community with a fait a compli. In the absence of any peer-reviewed publication it is difficult to judge the quality of the science that led to the choice of Pain / Affect / Cognition / Mobility / Self care / Usual activities as the proposed core domains of the WHO health status measure. Selecting the descriptive basis of any health status classification marks the introduction of the valuation process. All omissions count equally and have a zero weight. So why no mention of social functioning in the new measure? And who decides that this should be so? One might be disposed to be lenient about the absence of traditional forms of evidence were the political realities less troublesome. A WHO Executive Committee scheduled for January 2001 was to consider the following recommendations:

• Subsets of core health domains selected from ICIDH-2 should be the basis of the development of a common survey instrument for measuring health status in periodic surveys in general populations;
• Health state valuations...should be measured in population - representative samples in each Member State; and
• WHO should provide detailed technical advice to Member States to assist them to develop data sources...on non-fatal health outcomes.

The existence of a single standard method for measuring health status could have enormous consequences outside the relatively arcane world of health systems league tables. Regulatory bodies impressed by the "made in WHO land" label may feel understandably, if mistakenly, comforted by the status of outcomes data based on its use. Gone are the problems of translation, cultural adaptation, valuation and re-calibration. Clinical trialists engaged in multicentre international studies will inevitably be tempted to try out the new technology and once a critical mass of such work is in the literature it will become difficult to avoid a steamroller effect. Valuations for health states defined by a WHO descriptive system are to be based on population surveys and some evidence of this work is expected to be revealed during the coming months. But here again there is general agreement that so single methodology dominates all others. A WHOne-size fits all approach would be an unacceptable solution for many researchers and decision-makers.

A final note of caution. There are few, if any absolutes in the outcome measurement business, so that the prudent decision-maker will always seek out the relevant source material before leaping to a considered view.

Don't take this observer's view alone - check out the website that houses most of the documents http://www.unece.org/stats/documents/2000.10.health.htm. It may not be too late for us to make up our own minds before others do it for us.

1	Australia, Canada, Denmark, France, Italy, Netherlands, New Zealand, Spain, UK and USA

  Issues Index | 2001 Issues Index