The role of health economics.

One group is likely to benefit especially from the proposed changes in the National Health Service: economists. As more emphasis is placed on market transactions so the demand for economic analysis will grow. The age of "econocrats" is about to begin.' Hence the importance of looking critically at the discipline and its claim to use rational techniques ofanalysis to resolve the complex problems ofdecision making in the NHS. In this, a study produced by a group of sociologists in York2ironically also the home of health economics in Britainprovides a useful starting point. Ashmore, Mulkay, and Pinch examine not only the role of health economists in the NHS but also some of the techniques used. In terms of the social sciences-where economists see themselves as representing a "hard" scientific discipline as against "soft" messy disciplines such as sociology or social policy-this is like the sheep turning on the sheep dog. But in the event the herbivores manage to score well off the carnivores, even ifthey do not quite drive them from the field. In the process they underline the importance of everyone concerned in NHS policy making gaining an understanding of just how economists work in order to be able to test their claims. How credible, in short, is their assertion that they represent scientific rationality in a world dominated by the prejudices of doctors and politicians, where decisions reflect the power of the participants rather than the force of reason? In practice, the ability of economics to resolve disputes by analysis is limited. This is well illustrated by the York sociologists in their chapter dissecting the attempt, led by their colleague Professor Alan Williams, to develop a technical tool for determining priorities among different claims on resources. This is the concept of the quality adjusted life year (QALY). The QALY approach purports to answer the question of where the NHS should put its money by evaluating on a common denominator different procedures or interventions. This is done by establishing the valuation of the consequential state of the patienttaking into account both disability and distress-on a scale from 0 to 1. The worse the quality of life the lower the rating. This valuation is then applied to the predicted life expectancy after the procedure or intervention to yield a QALY score. If, for example, patients are likely to live for 10 years in perfect health after operation A they notch up 10 QALYs; if, however, they are likely to live for 20 years after procedure B but at a quality of life rated at only 0-2 they notch up 4 QALYs. But the calculations do not stop there-the underlying assumption is that an efficient health care activity is one in which the cost for each QALY is as low as possible.3 Hence the final stage in the process is to reckon the cost of producing each QALY. If operation A is expensive-say, £20 000-then each QALY gained costs £2000. If procedure B is cheaper-say, £1000-then each QALY gained costs £250. So in this entirely imaginary example the NHS decision maker would, if following the efficiency maximising principle, give priority for resources to procedure B. There are problems in making such calculations. The first is how to establish the valuation of different states of health or conditions after medical intervention. Constructing such indices is now an academic industry,4 and various products are on offer. What they have in common is that the ratings are elicited with survey instruments, which rather than tapping the spontaneous judgments of respondents force them, as the York sociologists point out, to fit their views to the "abstract and artificial" categories of whoever designed the questionnaire. This method questionably assumes "that the numbers assigned by respondents have some real meaning for them in ordinary situations and that these numbers express quantifications that are already implicit in individual respondents' scales of preference." Further, the whole exercise presumes that such valuations are stable over time and that they would not change in a different context. Another fundamental problem with QALYs is that information about outcome is often inadequate. We simply may not know enough about life expectancy (let alone about the quality of life) after many procedures or interventions, particularly new ones. The point is illustrated by the report on a joint project between the University of York and the North Western Regional Health Authority, which used the QALY technique "to provide a new criterion for use in determining resource allocation. "I This examined treatment in end stage renal failure, replacing an arm joint, the use of ceftazidime in the treatment of cystic fibrosis, and the surgical treatment of scoliosis. In all four cases adequate data about expected life expectancy or quality of life were lacking. In all four cases, therefore, some bold assumptions had to be made. Nevertheless the report concluded, with surprising precision, that the cost for each QALY was £194 for an operation to correct scoliosis in neuromuscular illness, £592 for a shoulder joint replacement lasting for 10 years, and £1412 for a kidney transplant lasting for 10 years. The problems of using QALYs in decision making do not stop with the credibility of the processes for generating such figures. Can interventions that prevent death be measured on