Intended for healthcare professionals

Education And Debate

The rationing debate: Maximising the health of the whole community. The case for

BMJ 1997; 314 doi: (Published 01 March 1997) Cite this as: BMJ 1997;314:667
  1. A J Culyer, professora
  1. a Department of Economics and Related Studies University of York York YO1 5DD


    It seems a pity to compromise what seems uncompromising, but let us begin with some health warnings. Firstly, “principal” does not mean “only” and some of the other things the NHS does (and ought to do) turn out to be necessary anyway if it is to achieve this prime objective. Moreover, efficiency (which is what maximising is about) needs always to be tempered by consideration of equity in both process and outcome. Secondly, let's remind ourselves that most moral objectives (of which this is one) do not lose their force by virtue of being impossible to attain–one of the reasons for having moral rules about anything is that they provide bases for judging how well one is doing with respect to what one ought to be doing.

    Thirdly, let's remember that there are good reasons for our having taken health care out of the “ordinary” market place. These include a solidarity-type case that ensures no one is excluded from benefit on grounds of lack of portable, transparent, and comprehensive entitlement; protection from professional dominance in the determination of both general healthcare priorities and specific patient-doctor relations (in any system of health care it is primarily the doctor who determines the demand for care, not the patient); equity in funding arrangements, processes, and outcome (mainly health); and the provision of care that is more likely to confer benefit than harm.

    Fourthly, maximising such an objective involves not only a commitment to the ethicality of that which is being maximised but also embodies within it a host of other ethical issues; these often take the form of trade offs, whose exposure, discussion, and resolution by people with legitimate rights to be involved is important. Fifthly, maximising anything implies the need for particular sorts of knowledge: for information about health status, changes in it, its decomposition into relevant population subgroups, and believable attribution of such changes to causes (whether they lie in the delivery of health care or through other means).

    Finally, the desirability of measurement in general ought to be distinguished from the suitability and acceptability of any specific measure. One desideratum of any measure of health or health gain is that it should enable interpersonal comparisons of health gain (or loss) to be made; and this is one of the striking departures from the more general utilitarian objectives customarily set by economists in evaluating the advantages and disadvantages of various institutions and policy options. A common objection to health measurement is not so much an objection to outcome measurement per se as to either a particular measure of it (for example, that it misses something important out) or to a particular way of using a measure (for example, not weighting prospective health gain, or prospective health gainers, differentially according to morally relevant factors). One of the attractions of explicit measures of prospective outcome is that they clearly expose sins of commission and omission. Thus they enable the explicit discussion and implementation of equity based desiderata, rather than leaving them to the uninformed whim of individuals and committees with influence.

    NHS ought to be about maximising health

    There can be no doubt that a principal objective of the NHS is to maximise health. We have ministerial authority for that. The more interesting, non-factual, assertion is that it ought to do this. The ethical underpinnings for my view that it ought lie in the importance of good health for people to lead flourishing lives, which I take as an ultimate good. We can all think of individuals with terrible handicaps of ill health who seem to flourish but these are not persuasive counterexamples. Such people excite our admiration and are seen as exceptional.

    In general, I take it that flourishing is an ultimate good and that good health is in general a necessary condition for achieving this ultimate good. In short, health is needed in the twin senses that it is both necessary (just as my possessing a Rolls Royce is a necessary sign of my personal success in life) and serves an ethically commendable end. This gives an otherwise merely technical relation between means and ends its ethically persuasive quality and raises the need for health to high ethical significance (in a way that is not true for my need for a “roller”).

    To take the argument further, health care (including medical care) may be a necessary (though not sufficient) condition for realising better health. If so, it too is needed (that is, is necessary if improved health is to be attained) and it too derives its ethically compelling character from the ethicality of the flourishing that is the ultimate good. So, not only may it be reasonably assumed that individuals want health care; they also need it in an ethically persuasive sense of the word.

    If all that is accepted, maximising the health of populations becomes an ethical objective. So does being efficient so that the resources used in health care are used to maximise health outcomes. This is not the same as maximising the use of beneficial health care–or effectiveness. It differs from it principally in that delivering only that care which is most effective takes no account of the opportunity cost of such care (a highly effective but very costly treatment may rightly be given lower priority than a less effective but much cheaper one) when both cannot be delivered to all who might benefit. Distributive justice also acquires a high priority: in my view (which is not that equity is sufficiently served by maximising some equity weighted outcome measure) this is best tackled in terms of seeking to identify and move towards a more equal distribution of health across the population while at the same time ensuring that each procedure offered to patients is that believed (on the best evidence available) to be the most cost effective. This will not usually imply an equal distribution of resources, nor will it imply a curmudgeonly equality in which everyone gets nothing (equally). It actually implies, given current knowledge of the way medical technology is deployed, both a rise in the average health of people and a more equal distribution of health. There are twin problems for social decision makers here. One (for healthcare commissioners and providers) is the selective use of their resources to achieve objectives efficiently. Others (for higher level decision makers) involve trading off other ultimately good things which we might legitimately seek in pursuit of flourishing lives but which compete with health care in the battle for resources. There is no room for absolutism here, for there is more than one means to the great ethical end of flourishing. Nor can every desirable thing be done for everyone. Conflict, and the need to choose, is inevitable.

    Efficiency and equity aren't always in opposition

    Conflicts can, however, be overdone. One that is commonly overdone is the alleged clash between efficiency and equity. If we define efficiency in a health service as being the maximisation of probable health outcomes, and there is also an acceptable quantification of these outcomes across the variety of activity we call “health services,” then there exists, as a matter of logic, such a maximum for every possible distribution of resources to individuals. All these possible distributions are efficient. But all are most certainly not fair or equitable. Choosing between these possible distributions, all of which are efficient, cannot involve any conflict between efficiency and equity–unless you make the additional ethical judgment that the marginal unit of outcome is always of equal value to whomsoever it accrues. I see no compelling moral argument for such a judgment.

    Talking theoretically, though difficult, can sound glib. In practice one is in a sea of uncertainty, even in a world as conceptually simple as that just described. There is a deficit of usable relevant information on health itself, its distribution across population groups, on health gains (actual or projected), on the links between the activities of the NHS and their final impact on people's health, on the reasons for the huge variations that can be measured between practitioners or the variations in outcomes that individual practitioners achieve. As a practical example, the enormous clinically inexplicable variations in general practitioners' referrals within and across health authority areas are a source of both deep inequity and substantial inefficiency which only health authorities can address.

    For many in the research and development commissioning communities, these lacunae provide the (ethical) momentum for changes that have recently been set in train in the research and development programme, for the intelligent use of evidence based medicine, for outcome measurement, and for the partial separation of the activity of healthcare commissioning from healthcare delivery. There is an act of faith involved here, which is that more evidence relating to the components of the links in the flourishing healthcare cascade is a good thing. This involves a belief that more (relevant) information is better than less and a commitment to the principle that the best should not be allowed to become the enemy of the good.

    Information not a substitute for judgment

    Undoubtedly, the mere provision of information is insufficient–at the very least it will need interpretation in particular contexts by patients and professionals who understand enough of its limitations not to fall into the trap of supposing that information can ever be a substitute for judgment (including clinical judgment). Moreover, there is abundant evidence that the mere provision of even very good information is not itself sufficient to get the professionals to act on it. Further, issues of value pervade the entire decision structure. At one level it is impossible to define “health” without value judgments (whose should they be?); at another, it is usually impossible to determine the appropriate course of medical actions for a particular patient without making patient specific value judgments (whose, again, should these be?). There are values to be selected at all points in between.

    As I wrote at the beginning, improved health is not the only business of the NHS. In relations with patients a common task in both primary and secondary care is to provide information–and no more: information that a person does not have the disease he or she feared, about whom outside the NHS to contact for help with a problem, about healthy lifestyles, and so on. Plainly, such information serves an ethical end. Moreover, it may also serve the end of health maximisation–health education, for example. The institutional side of the NHS also provides hotel services, which ought also to be provided efficiently but which may not raise questions of distributive equity of the same compelling sort as does active medical care itself and might be left to private purchasing power and insurance arrangements without damaging the objectives of the NHS.

    Similarly, equity in the distribution of health (or of health gain, or of healthcare resources) does not exhaust what ought to be proper equity concerns in the NHS. Procedures and processes too must be fair. It is not fair to keep similarly placed people waiting avoidably different times; nor for professionals to be rude or inconsiderate; nor to treat professionals within the system as though they were employees in a command economy or to set them professional targets without also supplying the means by which they might meet them; nor to exclude those for whom the NHS exists from decisions about the values that are to be incorporated in the layers of this many tiered cake.

    Work on measures is needed

    Setting an objective of the sort postulated here is not the usual way that economists have approached issues of efficiency and equity. They have more usually had a particular and rather sophisticated branch of utilitarianism to set the conceptual rules for resource allocation which goes under the name “Paretian welfare economics.” This is the view that decisions ought to maximise subjectively perceived welfare, that the only identifiable improvements are those where no one loses such welfare and at least one gains some, and that in situations where some gain and others lose one can only sit on one's hands. Some of us have rejected this framework for health and health care not because we want to reject the respect for individual values which is enshrined within its ethical frame but because it fails to deliver practical guidelines with practical consequences and, where it does, does so with severe limitations. A particular weakness of the traditional Paretian approach is that it affords no leverage on choices that have to be made which involve some people losing while others gain–which is, sadly, the usual situation. The usual evaluative framework is also silent for choices that are based on considerations of equity.

    This is not true of the object set here (maximising health) provided that a suitable measure of the thing to be maximised is available. Twenty five years ago no such measure was available. That is no longer true. A battery of claimants exists, each of which has its advantages and disadvantages and some of which may be more appropriate to some types of choice than others. We need appropriate measures for all the outputs of the NHS that are of prime concern and indicators of the varied dimensions that equity takes. We also need a community of users of this information who can interpret and use it towards the NHS's objective and who can feed problems back to the consumer and the professional, managerial, and research communities so that improvements and refinements can be made and lacunae filled. All this entails comprehensive partnerships and dialogue across a spectrum of communities and interest groups. It also requires education, training, and research.

    The practical problem at all levels of the NHS is to be able to apply consistent and acceptable principles to answer questions like: Which services shall be available? To whom shall they be available? On what conditions shall they be available? These questions are all rationing questions, and the principles need to be practically useful and defensible by those who use them. If you don't find mine acceptable (at least they meet the requirements of consistency and applicability and are derived from a set of explicit ethical considerations), then what are your alternatives–and how would you expect ministers, the NHS Executive, NHS managers, and NHS professionals to implement them?