Intended for healthcare professionals

Education And Debate

The rationing debate: Defining a package of healthcare services the NHS is responsible for The case for

BMJ 1997; 314 doi: (Published 15 February 1997) Cite this as: BMJ 1997;314:498
  1. Bill Newa, senior research officer
  1. a King's Fund Policy Institute London W1M 0AN


    The use of tattoo removal as an example of NHS rationing is now so common that it is in danger of trivialising an important debate. Behind such questions as “Should the NHS be devoting resources to tattoo removal?” lies a more fundamental issue: what kinds of benefit should the NHS provide?

    Most readers will assume that defining a healthcare “package” is a means of rationing healthcare resources. In other words, faced with the task of managing the limited NHS budget one option is to exclude some services altogether. But my case rests on a different interpretation of a package and involves asking a preliminary question. Before deciding how to ration, we need to know what to ration: What is the range of services relevant to the role of the NHS? What “business” is the NHS in? Is it the NHS's job to provide fertility treatment, physiotherapy for sports injuries, long term nursing care, gender reassignment, adult dentistry, and cosmetic surgery? Or should these services be provided by local authorities, voluntary agencies, or the private sector? The question does not rely on clinical judgment. It is about the boundary of a public institution's responsibilities. And it is a question which has been muddled up with issues of rationing proper.

    Defining the boundaries

    The need to address this question derives from a growing sense of confusion and uncertainty about what it is reasonable to expect from the NHS. For example, where one lives can have a decisive effect on whether or not NHS treatment is available. The 1991 NHS reforms were an important catalyst in this process: purchasing authorities now concentrate on commissioning health care for their resident populations, rather than on management issues. Wishing to be seen to be making the best use of financial allocations, some took the view that certain services were not a priority and therefore not worth purchasing. For example, the availability of fertility treatment depends on the whim of purchasing health authorities—and increasingly general practice fundholders. In addition to this uncertainty over regional variations is the apparent removal of some primary care services from NHS provision altogether. Adult dentistry is subsidised by central government, but only for a fraction of its cost. In some areas it is hard to find dentists who offer even this minimal NHS cover. There has been no explicit national debate about why dentistry is apparently not an NHS responsibility.

    It is inequitable that one's place of residence should determine access to care. Levels of service provision will inevitably vary from one part of the country to another, in response to varying need or because some providers are more efficient than others. But this is different from removing availability altogether: infertility is not at zero levels in those areas where in vitro fertilisation treatment it is not available. Ad hoc developments such as these can serve only to promote uncertainty and a sense of unfairness quite out of proportion to the quantity of resources at stake. Furthermore, the vigilance of the news media has had a significant impact on the public's perception of health delivery. Activity, and inactivity, in the NHS is now scrutinised and reported daily. This is welcome, but awkward: old issues, once hidden, must now be tackled if the NHS is not to fall into disrepute.

    Nothing to do with saving money

    To be clear about what devising a healthcare package would seek to achieve, we must be clear about what it is not trying to do. Firstly, it is not (necessarily) about saving money. The case for a centrally defined package has been associated with easing the pressure on resources. However, this is not the purpose of the proposals outlined here—the desire is to promote equity, collective understanding, and reassurance. The package considered relevant to NHS business is just as likely to be more extensive than that available now. The point is that it should not vary from one area to another and that it should be derived as the result of an explicit, democratic process.

    Secondly, drawing up such a package is not an attempt to avoid additional rationing. In the well known Oregon initiative in the USA, all those services which might possibly be provided collectively are ranked and the line drawn where resources allow. Above the line everyone has access; below no one does (unless privately financed). The line moves up and down depending on the availability of resources. Rationing health care is therefore a centrally undertaken activity, specifying a package to which everyone has access.

    But the approach presented here is a preliminary to rationing. It is about deciding what should appear on Oregon's list in total, not about where the line should be drawn. Of course, there will be resource implications from this decision. For example, if long term nursing care was considered an NHS responsibility then resources might need to be reallocated from private households to the NHS through taxation. But specifying a package to promote reassurance and geographical fairness does not involve deciding how NHS resources should be distributed between individuals who make a claim on them. Defining a package does not imply a right to treatment.

    Cost effectiveness not relevant to establishing “the package”

    However, if the need, in principle, for a package is accepted then there must be a coherent and practical means of establishing it. For this to be successful the focus must shift from criteria which guide rationing decisions to criteria which help establish the range of relevant services which are to be rationed. Trying to do both at once results in doing very little at all, as international experience testifies.

    New Zealand and Holland have both tried to establish packages of healthcare services, and both have had little success. In New Zealand, “core” services were intended both to clarify what the population could expect and limit the financial burden on the state. Ultimately, though, the planners were forced to concede that everything that was currently provided would form the core—hardly the result they were looking for. In Holland, four criteria—necessity, effectiveness, efficiency, and individual responsibility—were used to define a package. The Dutch also found it difficult to specify precisely which services ought to be excluded. Why so little success?

    These strategies were trying to do too much at once. Issues of equity, reassurance, and clarity about the responsibility of the state were mixed up with a desire to contain costs and ration more systematically. The root of the difficulties lies in the inclusion of effectiveness and cost effectiveness as criteria for establishing a package.

    Whether a service is relevant to a healthcare system has nothing to do with effectiveness. For example, cosmetic surgery for enhancement where there is no severe psychological distress is rarely supported for collective healthcare provision. But no one suggests that the plastic surgeons who work miracles on Hollywood stars are not effective. So questions of effectiveness and cost effectiveness should be left aside at this stage of the debate.

    Their importance comes when deciding how to allocate resources between all the services that are relevant and between the people who can benefit from them. The result of such deliberations may also lead to a package, but it will be of an entirely different kind. In Oregon, cost effectiveness was the basic principle at work (although with many refinements and alongside other criteria). But the resulting package did not address whether fertility treatment, residential care, dentistry, and so on were relevant to Oregon's public healthcare provision.

    Reliance on measures of cost and effectiveness has meant that most commentators believe strategies for defining a package are doomed to failure, or at least likely to disappoint those who promote them.1 The reasons given are now reasonably well accepted: health interventions are extremely variable in their effects on individuals. Just about every treatment therefore does some good for someone, even if it is “ineffective” in general. Making blanket exclusions on this basis will inevitably be a blunt instrument and will antagonise doctors, who feel their clinical freedom is curtailed. As a consequence, the Oregon experiment has proved to be extremely controversial and has generally not been considered relevant to the NHS.

    A qualitative approach

    But the following approach for establishing relevance to the NHS does not rely on effectiveness. Instead it proposes a qualitative approach which avoids the difficulties of variable individual response to treatment. The approach has been described in more detail elsewhere,2 but the central proposition is this: those characteristics which define healthcare's special nature, and which in general terms make it unsuitable for economic exchange, should determine whether or not individual services are relevant to the NHS.

    Taken together, three characteristics set health care apart: fundamental importance, information imbalance, and uncertainty. Health is clearly of fundamental importance; there is little certainty about how our health will develop in the future; and typically we know little about the nature of our needs for health care or of the likely effects of the treatments available to us. In short, we are not good consumers. We must trust doctors who, under free market conditions with insurance markets, have an incentive to provide as much care as possible.

    It is the combination of characteristics which makes health care special. Fundamental importance is not sufficient on its own—food, after all, is important but no one suggests that we should have a National Food Service because people are perfectly able to choose what and how much food they need. Neither are information imbalance and uncertainty sufficient on their own. Car repair services display these characteristics, but no one is suggesting a National Car Repair Service—because cars are not of fundamental importance.

    However, not all services which are related to health care display these characteristics in combination. Residential care for the elderly is fundamentally important but does not suffer from significant information imbalances; cosmetic surgery for enhancement is not generally considered to be of fundamental importance. Services of this kind should not be an NHS responsibility. On the other hand, curative dentistry, fertility treatment, and intensive nursing care do seem to satisfy the criteria; consequently, they should form part of the NHS's range of available services. All health authorities and GP fundholders should be required to provide some level of these services. Note that defining this package does not imply any judgment about effectiveness or whether any individual service represents good value for money. It simply clarifies what the NHS should be in the business of doing.

    Meeting the objections

    There are a several practical objections to such a proposal. The first argues that such an approach necessitates drawing up a long list of individual services which would need to be continually updated. Every existing and new treatment for cancer, for example, would need specification. In fact, such an enterprise would be unnecessary. Specification is required only at a general level: treatment for cancer, AIDS, infertility, or whatever, not individual drugs or surgical interventions. Indeed, it may be sufficient simply to concentrate on exclusions rather than long lists of inclusions. Furthermore, whether or not a particular treatment is effective is irrelevant—the NHS may decide not to purchase a drug to treat AIDS until it proves its effectiveness, but it would be clear that drugs of this kind are relevant to NHS business.

    The second objection argues that any attempt to centralise decision making will fail to accommodate individual cases and exceptional circumstances. Clearly, defining the range of relevant services will restrict clinicians' ability to use their skills to certain ends. But this will not be as inflexible as critics suggest.

    A typical example cites an individual with extreme psychological distress caused by a tattoo mistakenly purchased when young. In circumstances where cosmetic surgery for enhancement is outside NHS responsibilities, surely such central codification would deny legitimate treatment? Not if treatment is correctly focused on the nature of the condition. In the case of psychological distress the correct course of action is referral to a psychiatrist (psychiatry, let us assume, is within NHS responsibilities). The specialist might well provide appropriate treatment herself or may recommend removal of the tattoo as the preferred means of treating the distress. In this way tattoo removal could still be undertaken quite legitimately on the NHS. But at the same time some restriction has been placed on clinical freedom: cosmetic procedures for conditions that do not involve psychological distress are outside the scope of the NHS—no exceptions. If a real improvement in the clarity of the NHS's role and equity in the availability of its services is to be achieved then such restrictions are inevitable.

    The final objection asserts that introducing legal specifications of what the NHS should provide will simply allow the nominal provision of services. So, for example, if infertility were to be included it would probably be sufficient for a health authority to provide one round of treatment per year to satisfy its legal obligations. Thus health authorities could pay lip service to the policy but continue to decide the range of responsibilities for themselves. This may indeed turn out to be how purchasers act. But it is equally reasonable to suppose that health authorities simply want clear guidance on the range of their responsibilities and have no wish to indulge in gamesmanship with policy directives. In any event, individual purchasers would be clear about their role—they would still decide on the extent of provision depending on local circumstances, but would now provide services secure in the knowledge that they were in step with all other purchasers in the NHS.

    One final point. Criteria such as those suggested here for guiding the specification of a healthcare package are just that: guidance. They cannot replace debate and political compromise. So it is not possible simply to read off a list of relevant healthcare services—people will inevitably disagree over the degree to which fertility treatment, for example, satisfies information imbalance and fundamental importance. But once a decision has been made, openly and with reference to coherent criteria, what we stand to gain from clarity and equity will surely outweigh the awkward processes involved. The alternative—allowing the NHS increasingly to ignore the principles on which it was founded—risks losing mass popular respect for a successful and valued public institution.


    The BMJ, BMA, King's Fund, College of Health, and RAG (Rationing Agenda Group) will be holding a conference entitled “Rationing in the NHS: Time to get real” on 10 and 11 July at Kensington Town Hall. For further information ring Jane Lewis 0171 383 6605 (Email: 106005.2356{at}


    1. 1.
    2. 2.