Evidence, ethics and inclusion: a broader base for NICE
- First Online:
- Cite this article as:
- Wilmot, S. Med Health Care and Philos (2011) 14: 111. doi:10.1007/s11019-010-9256-1
- 109 Views
The National Institute for Health and Clinical Excellence (hereafter NICE) was created in 1998 to give guidance on which treatments should be provided by the British National Health Service, and to whom. So it has a crucial role as an agent of distributive justice. In this paper I argue that it is failing to adequately explain and justify its decisions in the public arena, particularly in terms of distributive justice; and that this weakens its legitimacy, to the detriment of the National Health Service as a whole. I argue that this failure arises from the fact that NICE works within the frameworks of positivist science and liberal ethics, largely to the exclusion of other perspectives. This narrowness of view prevents NICE from properly connecting with the range of moral concerns represented in the population. I argue for NICE’s deliberations to become more inclusive, both in terms of epistemology, and also in terms of ethical perspectives. And I suggest a range of perspectives that could usefully be included. Finally I offer a framework of structures, philosophies and discussion process that will enable competing perspectives to be debated fairly and productively in this process.
KeywordsNational health serviceNational institute for health and clinical excellenceIndependent regulatory agencyLegitimacyMediaPositivist scienceLiberal ethicsCommunitarianismConstructionismDialecticPluralism
My general theme in this paper is the legitimacy of a publicly funded health care system, and the significance of distributive justice for that legitimacy. I intend to consider how distributive decision-making in health care might gain in legitimacy by giving wider representation to the diversity of perspectives in contemporary western society. I shall focus on the UK National Health Service (hereafter the NHS), and in particular on that part of the NHS most relevant to my concerns—the part that exists most explicitly to deliberate and advise on resource allocation. I refer to the National Institute for Health and Clinical Excellence (hereafter NICE). I shall start by identifying a problem that exists in the debating and justifying of NICE’s guidance in the public arena, and the implications of this problem for the legitimacy of the system. And I shall consider how a more inclusive process might lead us to a solution to this problem.
The problem for NICE
The task of giving guidance on which treatments should be provided out of taxes by the NHS and which should not, was given to NICE on its creation by the UK government in 1998. NICE is an example of an Independent Regulatory Agency, a species of public body of fairly recent origin, which has proliferated in Europe in recent decades, though it has a longer history in the United States (Gilardi 2008). As such, NICE has a fair degree of independence (I shall discuss that more later), and a responsibility to government and public to make fair and informed judgements. In making these judgements, NICE takes account of a range of evidence, and formulates its advice on the basis of the effectiveness of the treatments under consideration (according to the evidence), taking account also of considerations of fairness (Culyer 2006). NICE operates on a scientific model, using systematically gathered evidence, analysed and evaluated, before weighing that evidence against other considerations, and reaching a decision. NICE includes non-quantifiable evidence in its deliberations, and it makes provision for the use of qualitative research evidence. But it is clear from its own guidelines (NICE 2009) that the core and substance of its collective thinking relies on quantitative evidence, gathered by methods that are validated by positivist science.
However, value judgements also form an essential part of NICE’s processing of its evidence. In deciding on the effectiveness of a treatment NICE must have a view on what constitutes effectiveness, and in making that judgement it needs to be clear what value it will give to the benefits and disbenefits of the treatment. To illustrate this point briefly, let us suppose that NICE has good evidence that drug A relieves pain in sufferers of a particular chronic condition more effectively than drug B, but carries a greater risk of shortening life expectancy if used over a long period. NICE must use that evidence to decide what advice to give, and the choice between A and B clearly involves a value judgement between the different predicted outcomes. Do we prefer a shorter life with less pain or a longer life with more pain? As with its assessment of the evidence, NICE seeks to make such value judgements as rationally and systematically as possible. This can be seen particularly in its use of the principle of QALYs (Quality Adjusted Life Years) which involves the numerical weighting of health states in terms of public preference, and the calculation of the expected benefits of a treatment on the basis of the quantifiable gains in health state that (on the evidence) it will achieve (Hasman 2008). However, NICE also makes what it calls ‘social value judgements’, concerned with fairness and equality; and it has acknowledged the importance of these judgements in its guidelines (NICE 2008). Rawlins and Culyer (at different times leading members of NICE) both emphasise the effort made in NICE’s deliberations to properly and openly take account of distributive justice, and other ethical considerations (Rawlins 2005; Culyer 2005, 2006).
As Littlejohns et al. (2009) point out, NICE’s decisions sometimes generate public controversy, conducted, inter alia, in the news media. NICE’s explanations of its decisions unsurprisingly emphasise the systematic and rational nature of their process and outcome (NICE 2009), and when its decisions are controversial, its press releases and other documents defend these on the same basis. And it is this feature that embodies NICE’s problem. Opposition to NICE’s decisions, typically focused in sectors of the British press, is often expressed in very different terms from NICE’s explanations. For example, the controversy over Herceptin, a drug with some positive effects against breast cancer, for which NICE recommended restricted access, is documented by Wilson et al. (2008). According to that study, adverse press comment generally focused on the experiences of specific individuals in the wake of NICE decisions. Typically a few individuals were identified, and were then followed through their experiences over a number of months or years. Their situation was often presented in emotive terms and implicitly or explicitly, their moral claim to receive Herceptin was expressed in the relevant articles and reports. This, of course, is not untypical of the way the popular press deals with health issues. For example Thomson et al. (2008), in a study of press treatment of another health issue also identifies a tendency for the press to focus emotively on individual protagonists involved in health issues, rather than on the issues themselves.
There are no surprises here. The mass media in the UK have long recognised that, though the public can be impressed by the selective use of statistics, issues and events are most readily apprehended by most people through accounts of individual experience. By contrast, on the NHS side of the debate, discussion and decision-making in health care take the opposite form, seeking to be as impersonal as possible. This is shaped by a number of cultural influences, in particular those of the civil service and the medical profession, wherein impartial rational analysis and the use of impersonal quantitative evidence are highly valued. It reflects the continuing influence of rationalist and positivist values among the relevant elites in the health care system, underpinned by the long-held self-perception of the medical profession as science-based, and reinforced by the more recently developed ideology of evidence-based medicine. One rationale for those values can be found in the philosophy of utilitarianism, which brings together judgements of evidence and judgements of value. Since the social and health reforms of the nineteenth century, utilitarianism has provided a rationale for basing quantitative ethics on quantitative evidence. The utilitarian emphasis on maximising well-being encouraged the early reformers to measure the state of the population systematically, and provided a value-base to evaluate that state. And NICE partakes of this heritage. As Dolan (2001) argues, utilitarian thinking remains a major element in the principles and application of QALYs in NICE’s evaluations. He suggests that the QALYs principle is, at the least, 40% utilitarian, and at most, 100% so. This suggests that utilitarianism can be seen as a major influence on NICE.
So the discourses of NICE and its media antagonists differ fundamentally. The result, in the case of the Herceptin debate and in other comparable cases, is that the protagonists and their arguments collide, but do not connect. In fact often there is little debate worthy of the name in the wider media. To be sure, opposing arguments engage more coherently in the specialist journals, where NICE has also been strongly criticised from a distributive justice point of view (e.g. Harris 2005). But for the general public there is little opportunity to observe that process, let alone participate in it. NICE’s difficulty in explaining its decisions to the public is identified as an issue by Hasman (2008). NICE’s clear ethical (and in particular justice) remit hardly features at all in public discussions of NICE, except through accusations of injustice. This is despite the clear acknowledgement of that remit by Rawlins (2005) and Culyer (2006) and in NICE’s working documents; and despite its efforts to connect with the public through representative structures (particularly its citizens council). Despite all this NICE is generally portrayed in the media as a body carrying out what are essentially complex technical tasks. Far from being seen as an ethics body, it has been accused of being ethically illiterate (Harris 2005). Its rationales for its decisions evidently do not engage with the ethical concerns of its critics.
The lack of a coherent debate that is comprehensible to the public potentially weakens the legitimacy of the decisions made. This legitimacy issue has been identified by several commentators including Hasman (2008) and Milewa (2005). NICE was created by government to deliberate on the distribution of certain health care resources, and it is the state’s agent in this. But NICE is acting for the benefit of the population as a whole—Culyer (2005) states that NICE’s client is ‘everyone’—and much of its legitimacy comes from its mandate, to maximise the benefit gained by the population, and to do this in a fair way. Taylor-Gooby (2008) argues that institutions such as the NHS can no longer expect automatic trust, and that in societies characterised by social uncertainty, trust must be actively sought. He goes on to emphasise that a public perception of service-users being well treated is a crucial basis for trust. So it is important that NICE’s deliberations are understood by the public, and are comprehensible within generally held perceptions of what is fair and what is desirable. The process I have described above is failing to facilitate any such common understanding. I shall now attempt to identify the roots of this problem.
Legitimacy and evidence
I would suggest that over- reliance on positivist science explains part of NICE’s problem. With regard to its choice of evidence, I would argue that rational decision-making cannot be achieved on the near-exclusive basis of quantitative evidence. The positivist epistemology of western science favours evidence of the sort that can be replicated, compared and manipulated in a rational, systematic way. But this is a partial view of rationality. The view that we can only rationally comprehend the world by quantifying, reducing, comparing and replicating is rooted in the empirical tradition of Locke and Hume. From Kant and Hegel onward a different argument has developed; that we cannot properly apprehend the world, far less comprehend it, simply by choosing those phenomena that can be quantified and replicated. Such a narrow view of knowledge, runs this argument, actually works against rationality because it encourages a simplistic and reductionist view of the universe, and excludes a great deal of information that would allow us to understand the world. From its roots in Kant and Hegel, and also from contemporary philosophers such as Rockmore (2005), who see our knowledge partly as a reflection of our own processes, psychological, social and political, comes a view that rational understanding of the world requires a grasp of evidence which is both holistic and specific, including the context of evidence, and including also ourselves as knowers and makers of knowledge. And from this come the theories that legitimise qualitative research and qualitative evidence. These theories are identified in a number of ways, with many overlaps and common threads. Guba and Lincoln (2004), for instance identify critical theory and constructivism as focal philosophies, both generating epistemologies that are transactional and subjectivist, supporting dialectical methodologies and qualitative methods. Palys (2003) identifies constructionist epistemology, phenomenalist research methodology, and qualitative methods. There are other versions, but I shall use Palys’ terminology henceforth.
Qualitative research seeks to gain a deep and complex view of the particular experience or situation under study. It provides evidence on the depth and complexity of human experience and perception in relation to particular phenomena. It can provide rich data for example about quality of life, a dimension which has considerable relevance to judging the effectiveness of many treatments. The mass media, with their human interest stories, understand the impact of this, and provide the illusion of richness of evidence, without discipline or methodology. By contrast, the procedures developed by qualitative researchers for the validation of their data, while in many respects different from those used with quantitative data, are equally demanding.
With regard to the potential implications of this for the legitimacy of decisions, Scharpf (1999) makes a useful distinction between input-oriented legitimacy, derived from a perception that there is a high level of public participation in their decision-making; and output-oriented legitimacy, derived from a perception that the decision-making body is using its own expertise to make effective and rational decisions. I would argue that the use of a wider epistemological base by NICE would particularly enhance output-oriented legitimacy, in that the perceived rationality and justice of its decisions would be enhanced by a perception that NICE was able and willing to process a wide range of different kinds of evidence, including qualitative evidence. I would add here that in my view the validation methods of qualitative research are sufficiently comprehensible to the lay person for the public to be reassured also that the evidence is properly weighed and processed.
Legitimacy and ethical analysis
I shall now move onto the second factor that in my view helps to explain NICE’s problem. I said at the outset that I am concerned with the evidence, and also with how value judgements are made about the evidence. And just as I have argued that NICE is too reliant on one kind of evidence, I shall also now argue that it is too reliant on one kind of ethics in making its value judgements. I have already identified the significance of utilitarianism as a formative tradition in health care, and as an important part of NICE’s philosophical heritage. Utilitarianism can also be seen as part of a group of ethical and political perspectives focusing on individual rationality and responsibility, and the systematic application of rational principles to moral and political questions; perspectives that can be identified with Locke, Kant, Bentham, Mill, and more recently with Brock, Singer, Dworkin and Rawls, among many others. A number of writers (Brown 2009; Callahan 2003) associate these philosophies with the broader political philosophy of liberalism, because they are characterised by individualism, moral egalitarianism, an emphasis on individual autonomy, and a belief in the universality of certain moral principles. Callahan (2003) and Turner (2004), argue that these perspectives have dominated ethical thinking in health care in the US, at the expense of alternative perspectives. And the influence of these ethical theories can be seen also in the UK health care system in general, and NICE in particular. In its documentation NICE explicitly subscribes to the four principles (Benificence, non-Maleficence, Justice and autonomy) which probably constitute the most widely applied version of liberal ethics in health care.(NICE 2008). These principles are more modest and flexible in their requirements than utilitarian or Kantian ethics, but are still essentially liberal, being universalist, and primarily concerned with the characteristics of the action rather than of the actor.
Liberal ethics fits comfortably with NICE’s preferred epistemology. It deals well with quantitative data, because, as Moon (1993) points out, liberal ethics and political theory work on a ‘thin’ concept of the person which focuses on universal characteristics. This corresponds closely to the strategy of the quantitative investigator of human beings who must likewise ‘thin’ down their specification of their subjects in order to allow quantitative measurement. Bryman (1992) argues that positivist research methodology as applied to humans is essentially individualist. My reading of this is that it is individualist in the same sense that liberal ethics is individualist; expressed by Bentham’s dictum, that ‘each counts for one and none for more than one’ (Bentham 1962). Where the units of data are humans, or human experiences, they must be comparable. So the egalitarianism and universalism of liberal ethics give value to generalisable evidence, and the need for quantification and comparison in utilitarian ethics leads also toward quantitative evidence.
However, liberal ethics is not so well suited to dealing with qualitative evidence, particularly where human action and experience are viewed as unique to one person or group, or one situation; and not systematically comparable with others. For instance, if I perceive my own experience of happiness as unique and not comparable to anyone else’s, I cannot subscribe to a utilitarian analysis of my position, which must compare my happiness to that of others to make a meaningful judgement. Specific affiliations to family, community, class, gender or culture, however important they may be to the individual’s own moral judgements, are irrelevant to the universal moral status of person which largely provides the basis of contemporary liberal ideas of distributive justice. And the concept of duties arising from these specific affiliations cannot easily be squared with the broader duties of deontological ethics. These features limit what we can do with qualitative evidence if we confine ourselves to this family of ethical theories.
Like its epistemology, NICE’s ethics limits its ability to communicate effectively with the community. Its reliance on quantitative evidence impoverishes its ability to justify its decisions, and its adherence to liberal ethics exacerbates this impoverishment. The ‘thin’ concept of the person which works well for NICE’s deliberations, does not carry an adequate ethical message to the general public. Its message comes over into the public domain as narrow and ‘managerialist’—a term Braun (2005) uses to describe the discourse characteristic of liberal ethics. It is this same poverty of discourse that disables NICE from properly performing its ethics role.
So NICE needs a wider ethical repertoire that will enable it to provide a more robust rationale for its decisions. One counterbalance to liberal ethics presents itself in communitarianism, which gives more weight to features specific to particular persons or groups of persons (communities, families, political units) and to the specific historical roots of these. However to maximise its credibility NICE would need to take account of a more representative range of positions on the ethico-political spectrum, starting on the left with liberation theory (relating to class, race, gender and other identities) and including, on the right, conservative political ideologies. These non-liberal traditions differ enormously from one another, but have in common that they tend to give weight to the particular features of the individual and/or the group, and are more likely to accommodate specific moral solutions to particular issues rather than to seek the application of the same criteria to every person and every situation. And they have in common a concern with specific historical, political or social circumstances which entail specific moral judgments and responses. This common ground between radical and conservative philosophies is convincingly delineated by Braun (2005) in her discussion of ethical controversy in Germany. She presents the liberal tradition somewhat negatively as impoverished and ‘managerialist’ whereas she describes the non-liberal ‘republican’ tradition (bearing many similarities to communitarianism), as more authentically focused on the depth and complexity of moral and political issues.
Bringing them together
How are the different kinds of evidence and the different moral frameworks to be brought together in a coherent structure and process so that inclusive deliberation can actually be achieved, and sustained? Can exponents of different fundamental views on the judgement of fact and value agree on acceptable ways of debating from their positions to a conclusion? A combination of qualitative evidence and non-liberal ethics in NICE’s deliberations would generate distinctive issues and arguments, in some cases highly challenging to those accustomed to positivist evidence and liberal ethics. It may for instance, involve giving serious consideration to using health care distribution as an instrument for achieving stability or change in social norms, social institutions and relationships. In qualitative evidence of health states, family roles and relationships might emerge as important in health and illness experiences, and relevant to the just distribution of care. Study of narratives of health behaviour (both lay and professional) may indicate that ideas of desert play an important part in distributive justice, despite the discomfort with which that concept is viewed by utilitarians and Kantians. Many other points of conflict are likely to emerge.
With regard to the debate on evidence, it is unlikely that positivists and constructionists will have fundamental disagreements about decision-making process and structure. Their disagreements are likely to focus on the content of their discussion, and their philosophical differences will not necessarily carry over into the politics of decision-making. With liberals and non-liberals in the ethics debate, on the other hand, the politics of decision-making are likely to be contentious in themselves. So in the next section I shall look at liberal and non-liberal views of the process of ethics deliberation, in the hope of establishing which circles need to be squared to have a properly inclusive process. I shall draw mainly on communitarian writers for the latter, but acknowledge that there are other important perspectives also on both left and right of the political spectrum.
Issues of deliberation
If we start with the liberal tradition, one of the best known models for deliberation on justice issues is that of the late John Rawls. He focused on modern societies with a wide diversity of beliefs, which nonetheless have to agree on the basics of a fair distribution of goods (Rawls 1993). He argued that where people hold rational systems of moral, political and religious belief (what he termed ‘Comprehensive Doctrines of the Good’), it is possible through ‘overlapping consensus’ to arrive at a model of distributive justice that would be acceptable within each Comprehensive Doctrine. He argued that a specific model of justice would be appropriate for all rational Comprehensive Doctrines in contemporary society, and as a separate (and earlier) discussion he set out what he thought such a rational model of justice would look like (Rawls 1971).
On this question there is reason to suppose that non-liberals will not have as much faith in overlapping consensus as a general principle. On the non-liberal side there are distinct communitarian counterarguments to Rawls on the question of rationality, which is basic to the idea of overlapping consensus. For instance Macintyre (1988) argues that definitions of rationality are contingent on the social and historical context of the purportedly rational agent—that rationality for an ancient Athenian is legitimately different from rationality for a contemporary European. Rationality, like justice, must be judged ‘internally’ within the society. A universal model of rationality, which is required by Rawls, has no place in this thinking. Likewise, Macintyre’s view of communal value systems, as rooted in the specific circumstances and practices of that community, provides no reason to expect overlap.
Certain other principles are important to Rawls’ argument. For one, citizens arrive at their Comprehensive Doctrines of the Good (hereafter CDGs) on the basis of a number of factors, but these CDGs should be seen as acquired and adhered to by choice. Also, the state needs to be neutral between the various CDGs subscribed to by its citizens. The only moral commitment justified for the state is to justice, and state activities should not involve endorsing or challenging anyone’s CDG beyond that commitment. These requirements have important implications for the processes and structures of deliberation that will enable the overlapping consensus on justice to be realised and applied. The fact that citizens choose their CDG means that citizens are responsible for the range of CDGs and conflicts between them. They are not constituted by their CDG. The moral neutrality of the state means that no CDG has the advantage of being the ‘state ideology’, and therefore all moral communities are, in that respect at least, on a level playing field.
To start with the second of these points first, the neutral state is rejected by a range of non-liberal viewpoints. Macintyre (1988) and Taylor (1992) both take up this point from a communitarian position. From non-liberal viewpoints a desirable state needs a number of moral features other than justice. These features vary with the particular position involved. And if the state does in fact embody a number of values other than justice, those whose CDGs put them at odds with the values of the state are likely to feel somewhat marginalised. This leads to a second complication relating to the first point above. Non-liberals do not necessarily separate the individual from their values as liberalism does. For instance communitarian theory sees values as part of the makeup of the individual—as helping to constitute the individual, rather than as an acquisition by choice (Mulhall and Swift 1996). So where conflicting values are juxtaposed, the adherents of those values potentially find their very selves at stake in the debate, rather than their opinions only. It follows from at least some non-liberal perspectives that people are exposed to the potential for profound personal change in the process of debate, on the principle that a clash and comparison of values will frequently lead to some mutual influencing.
Structure for deliberation
Staying focused in this section on the ethics debate, we need to ask whether NICE’s status as an Independent Regulatory Agency fits it to accommodate these contending perspectives and engage them in a productive and legitimate process of deliberation. As a starting-point, I would argue that it is appropriate for the state to take responsibility for enabling this kind of deliberation to take place, and for ensuring that the outcomes of those deliberations are put into effect. Neutral or not, there is no other body which can legitimately shoulder this responsibility. Irrespective of whether health care is provided from taxes paid to the state or from other payments made to other agencies, if there is to be a coherent framework of justice and legitimacy in the allocation of resources, only the state can carry that responsibility, because that responsibility by its nature must be to all citizens.
It is appropriate, then, that this function is undertaken by a state body. However, there is a distinction to be made between state and government, a distinction that is important to liberals and some non-liberals. The state is a system of institutions whose ongoing functions and relationships are defined by the (written or unwritten) constitution, whereas government is that part of the state apparatus responsible for overall direction of policy at any given time. NICE derives its mandate and legitimacy from the state. But its independence is also an important basis for legitimacy, and that must mean, among other things, independence from government. This involves protection from the short-term pressures of political expediency that determine most of the actions of modern governments in western democracies. Without that protection, NICE’s existence as an Independent Regulatory Agency is impossible. The notion of state bodies being relatively independent of government is consistent with the idea of the neutral state, and sits easily with liberal political theory, although radical and conservative views are generally more sceptical. In practice, particularly among those with political experience, there is likely to be a continuum of views rather than sharp divisions, and independence from government is likely to be seen by most radicals, liberals and conservatives as a matter of degree rather than an either-or matter, with some bodies achieving more real independence than others. In fact it seems that, compared with the general run of Independent Regulatory Agencies in western countries, NICE has a reasonable level of independence from government. For instance, senior positions in NICE are appointed to by an independent appointments commission (Pearson and Rawlins 2005), a feature which scores well on Gilardi’s criteria for independence of such agencies (Gilardi 2008).
For liberals, the nearer to independence NICE approaches, the more closely it conforms to their ideal of the agent of a truly neutral state. For many radicals all state bodies and some non-state bodies are contaminated by being part of a hegemonic power-structure, but, as most radicals do not see hegemonic systems as omnipotent or watertight, they may well still see it as worthwhile to participate in such bodies. For many conservatives, NICE’s link to the state will be an acceptable manifestation of the principle that the state does, and should, reflect dominant social forces and moral principles.
However, independence and legitimacy are not simply a matter of being protected from government pressure. It has been argued by a number of commentators, including Thatcher and Stone Sweet (2002), that Independent Regulatory Agencies can enhance their input-oriented legitimacy by being more open and accessible than government bodies, which are often relatively distant and inaccessible. The importance of input-oriented legitimacy based on accessibility to a wide range of opinion is confirmed by other evidence of public perceptions of decision-making; evidence accumulated over many years, for instance, that perceived fairness in the method of decision-making in formal bodies is particularly important in acceptance of a decision as fair by those whose interests were not necessarily served by that decision (see for instance Lind et al. 1993). To know one’s perspective was considered must help in acceptance of a decision one does not like. This principle is also supported by Daniels and Sabin’s model of accountability for reasonableness (Daniels and Sabin 1997) which requires the bases of decisions of resource allocation bodies in health care to be publicly known. Again, that information enables people to know whether their perspective was taken into account. This is crucially important for a body such as NICE, whose decisions are often controversial from a justice point of view. It needs a decision-making process that is perceived as fair and open, and the visible inclusion of as wide a range of views as possible in the making of the decision is an important part of a procedure perceived as fair. NICE has the opportunity to achieve a degree of input-oriented legitimacy in this way.
So NICE’s position as an Independent Regulatory Agency seems to provide a viable framework for involving different perspectives in a potentially productive dialogue; and an opportunity to enhance its legitimacy through this process, as well as, arguably, contributing in a modest way to the democratising of the UK political system. However, other elements would need to be in place to facilitate this, and it is these I shall now consider.
Philosophies for deliberation
In considering the philosophical framework for deliberation, I shall return initially to the debate on the evidence. The main protagonists here will be positivists who have confidence in quantitative scientific evidence, and constructionists who give credibility to qualitative evidence. Some protagonists will be relatively middle-of the road on the epistemological issues, others will be partisan. But on the whole they will inhabit different points on the same spectrum. The debate with regard to value judgements will be rather more complicated, with adherents of many different moral and political positions involved. Some of these may adhere to theories of ethics in the sense that I have been discussing them, but others will draw their moral position from other sources, including religious beliefs, community value systems, professional ethics and personal values. Some of the protagonists in the two debates will overlap, but there needs to be a clear distinction between the different kinds of debate that will need to go on and the arrangements that will be needed to contain them.
Those in charge of the discussion on evidence will need a robust philosophy and methodology of their own to sustain their demanding role. An adversarial winner-takes-all approach to the arguments and the process is unlikely to work, because both sides of the argument need to be kept engaged with the process. At the same time, reliance on finding consensus will not suffice, because consensus will sometimes not be found. Their philosophy and methodology will need to be predicated on getting the best result out of conflict, rather than necessarily achieving consensus. The concept of dialectic, as originally developed by Hegel, provides a philosophical framework for the bringing together of opposing perspectives in order to achieve a higher understanding. And the method of dialectical inquiry, developed by Churchman (1971) and elaborated by Malhotra (2000), and by Berniker and Mcnabb (2006), provides a Hegelian framework for the conduct of inquiry. Within this framework, where contradictory propositions on an issue can both be argued convincingly, this indicates incomplete understanding of the issue. A higher level of understanding can be achieved through reaching a ‘synthesis’ which resolves the contradiction. According to this principle, conflicting perspectives can be juxtaposed in order to expose and transcend their assumptions and enable the achievement of a more sophisticated analysis. We cannot suppose that this, any more than any other philosophical position, can provide a completely reliable formula for success in arriving at the best outcome on controversial issues. But it provides a positive set of goals with a coherent rationale, and as such can greatly fortify the position of those in charge of discussions.
An empowering principle for the leaders of the ethics discussion would probably best be defined as value pluralism. I have taken this term from Moon (1993), who states that this position “… rests on a denial of monistic theories according to which all values can be rationally ordered, but that very denial presupposes the truth or genuineness of our various and incommensurate goods and aims” (p. 53) This position allows the possibility of different moral positions to be comparable to one another, and to have different, comparable degrees of validity. There is no expectation that one position will finally be found to have exclusive validity, but there is a recognition that some positions may prove preferable to others. It is worth adding here that willingness to see the justice of one’s decision as provisional rather than final, and willingness to accept a review of that decision if it is coherently challenged, are implicit in both perspectives, and in accord with Daniels and Sabins model of accountability for reasonableness (Daniels and Sabin 1997) arguing in favour of the retention of NICE’s appeal mechanism.
Boundaries of deliberation 1: the evidence
Arguing, as I did earlier, for a more inclusive process inevitably raises the question of whether that inclusivity should be unlimited. For both the evidence debate and the ethics debate I shall argue that limits have to be placed, and that some exclusions are required. However, the limits will be very different in the two areas.
Taking first the debate on the evidence, the fundamental requirement is for an agreement at the outset as to the kinds of evidence that NICE would take into account. It is clear that among participants in general there is bound to be disagreement as to the worth of some evidence. Different kinds of evidence will be given very different weightings by people of different backgrounds. If the discussion is to be properly inclusive, participants who question the validity of a piece of evidence must nonetheless be committed to taking part in the discussion of its implications. If the evidence is admitted to the discussion, it must be discussed from all viewpoints. To enable this, it will be necessary for participants to commit beforehand to a set of criteria for admissible evidence, and to adhere to this whatever their own epistemology might prompt them to exclude.
These criteria will be crucial, as they will have a limiting effect on who participates in these proceedings. Whatever criteria are used, some people will be excluded because their preferred epistemology falls outside of the limits set by those criteria. That can only be acceptable if the criteria give the best prospect of legitimacy within the wider society. In that respect we might ask what kind of evidence we should reasonably expect a person to trust, when that evidence relates to their health, but they lack the skill to interpret that evidence themselves, and must therefore accept the results of another person’s interpretation, that person being unknown to them. It is one thing to entrust one’s health to ones own interpretation of a piece of evidence, or to trust that interpretation to a known and trusted clinician. But what kinds of evidence would we regard as appropriate for complete strangers to use in making such decisions? We would of course include evidence that we are competent to interpret ourselves, so that we can double-check. But for most of us, there won’t be much of that, compared with the range of evidence that may need to be brought to bear on our health. Much of the time we must rely on experts whom we cannot check up on and evidence that we cannot ourselves process.
What kind of criteria would allow us to trust those expert judgements? The NHS’s stock in trade is western medicine and related disciplines, modern, largely post-enlightenment systems of knowledge. These provide the subject-matter for NICE’s judgements, and the criteria for admissible evidence would need to reflect this. I would suggest that such evidence would need two characteristics. First it would need a clear epistemological base providing a rationale for validation of data, and that base would need to be locatable within the canon of western philosophy as it has developed since the enlightenment. Second, its rationales and procedures for validation would need to allow for systematic scrutiny and testing of the observations and interpretations of one investigator by others knowledgeable in that field and methodology. Those observations and interpretations would need to be open, verifiable and falsifiable within their particular conceptual system, each of which should be justifiable rationally. So we would exclude those forms of evidence and exploration that are ‘closed’ and cannot be replicated, validated or falsified. Divine or spiritual guidance through revelation would thus be excluded. There are grey areas, inevitably. For instance aesthetic experience is seen by some as verifiable, by others as a purely personal ‘inner’ experience that cannot be validated or falsified. However, in general it is possible to draw a boundary around those kinds of enquiry with open methods of validation and a secure base in the canon of western philosophy.
A system which confines itself to these kinds of evidence has a reasonable claim on the commitment of participants, and the trust of the public. Because these procedures represent a rational response to the problem of evaluating evidence pertinent to a modern health care system, they provide a common ground for everyone who is able to exercise reason. However, It is clear that my criteria are not universal, or culture-free. Clearly those methods of verification I have mentioned are substantially (though not exclusively) products of western culture. But by virtue of that they provide reliable and stable common ground in a modern health care system, a ground that people can share, whose other beliefs may take them in many different directions. For instance I personally may believe that the word of God trumps research evidence. I may know that those with whom I co-operate give no credence to the word of God, as I give no credence to the theory of Atlantis. But we have the prospect of being able to communicate, and co-operate, if we both accept that certain formal procedures adequately verify certain types of evidence.
Boundaries of deliberation 2: ethical analysis
With regard to the ethics debate, again we need to have an agreement on what moral positions should be included. The criteria for inclusion would need to be much more accommodating than the criteria for admissible evidence discussed in the previous section. The rationale for this is as follows. The criteria for admissible evidence are specific to the health care system, which is rooted in a particular set of epistemological traditions and involves a high degree of dependence on expert judgement. The main concern is to find criteria that provide a rational and reasonable basis to trust the experts. By this means we hope, using Scharpf’s terminonlogy (Scharpf 1999) to enhance output-oriented legitimacy. By contrast, moral judgements draw on the faculties and resources of the average person, and there is no single credible body of knowledge, expert or otherwise, which allows the validation of such judgements. Every person with the relevant faculties has a claim to make their own judgements, and evaluate the judgements of others. Ethics is, in that sense, much more democratic than epistemology and output-oriented legitimacy is not a realistic goal. So the main basis of legitimacy for moral deliberations must be the breadth of representation of beliefs held in the community, and by implication of the people holding those beliefs; in Scharpf’s terms, input-oriented legitimacy which rests on public accessibility. Beliefs based on faith and on cultural traditions would need to be admitted into the conversation, as they are likely to represent the thinking (or feelings) of a significant range of the population. However, I would argue that certain exclusions can be justified. Because justice is the goal of the deliberations, belief-systems which devalue or disregard the concept of distributive justice could properly be excluded. Perspectives on justice will differ widely among those who value it, and much of the process of deliberation will involve attempting to connect those perspectives. But those who give no weight to any version of justice cannot meaningfully participate.
A question arises as to whether it is possible in principle to include apparently non-rational beliefs in the discussion. In the Rawlsian framework, if we are seriously seeking a consensus on justice (i.e. a just distribution of resources) only adherents of rational CDGs could achieve that. At a practical level the application of this strict criterion would probably exclude too may viewpoints for the goal of legitimacy to be achieved, in a political culture where inclusiveness and equality of treatment are important means to legitimacy. And at the level of principle a compromise is needed between liberal and non-liberal viewpoints. For instance, though not all communitarians adhere as strongly as Macintyre to the view that rationality is contingent on history and society, it is nonetheless a powerful challenge to the Rawlsian model of overlapping consensus depending on shared rationality. So a compromise on this particular issue between the ‘strong’ communitarian position (exemplified by Macintyre 1988) and the liberal position would certainly require a more accommodating approach to the inclusion into the deliberation process of CDGs that look non-rational from the ‘outside’. On this matter I think I part company with Daniels and Sabin, whose narrower criteria of acceptable bases for decision reflect, I believe, their Rawlsian heritage, which is less accommodating of non-liberal perspectives (Daniels and Sabin 1997).
Goals of deliberation
The two debates are rather different, in that the protagonists in the debate on the evidence are likely to occupy different positions on the same epistemological spectrum, whereas the protagonists in the ethics debate will draw their perspectives from many very different sources. So the outcomes aimed at will need to reflect this. For the debate on the evidence, the ideal outcome would be a deeper shared understanding of a particular health outcome, through the dialectical juxtaposing of different kinds of evidence. This can be achieved in one research project, through triangulation of different kinds of data, and given appropriate leadership, I suggest that it could be achieved also through deliberation on compared evidence. However, realistically it will not always be possible to do this, and in that case compromise will be needed on the relative weight given to different kinds of evidence. This is clearly the less creative option, but it can be done rationally and fairly, again given appropriate leadership.
We can see views on the ethics debate along a spectrum. At one end we have the view that deliberation should work toward a rational consensus. Rawls’ (1993) view belongs there, as he envisages an authentic moral consensus, arrived at rationally. At the other end of the spectrum is Trotter’s model for ethical deliberation in health care (Trotter 2006) which envisages a ‘Modus Vivendi’ as the appropriate goal for deliberation between different positions. The Modus Vivendi is a compromise arrived at for the practical purpose of enabling the system to operate, and allowing different moral communities to work together in the same system. It would not be a moral agreement, based on the discovering or creation of shared moral ground and commitment. It would be more like a truce, adhered to for the prudential reason of maintaining a balance of power between different moral communities.
I have suggested it is unlikely that the range of CDGs that would need to be included in the process would all overlap in the way Rawls’ model requires. If I am right, the aim of moral consensus may be unrealistic. On the other hand Trotter’s Modus Vivendi, making no attempt at moral common ground, carries the danger of inadvertently allowing disagreements that are initially fluid and provisional to become solidified. Pursuit of genuine moral consensus by contrast recognises the contingent and provisional nature of at least some moral disagreements. So a position between the extremes seems to be worth finding. In terms of the possibilities of agreement, Kuczewski’s argument that there is probably enough overlapping ground between different positions to allow agreement to be reached on a number of issues (Kuczewski 2001), is relevant here. Kuczewski argues that agreement would be partial and contingent, and accepts the existence of areas where there is no common ground, or not enough to allow significant consensus. If he is right that some agreement is possible in some areas for the reasons stated, then the goal must be to stretch that area as far as it can go. Philosophical differences tend to be in the first instance differences of general principles, as indeed are religious differences, and differences of general moral belief. And often they are disagreements about the relative importance of different principles rather than whether those principles should be recognised at all. It has been suggested that focusing on these disagreements of principle is less useful in ethical deliberation than focusing on the application of general principles to specific situations. Ipperciel (2003), and De Marco and Ford (2006) both recommend a situationist approach to ethical deliberation, facilitating a process where adherence to general principles is modified by a particularist, situation-based approach to their application. De Marco and Ford argue for a ‘balancing’ approach where the application of different principles is not seen as absolute and immutable, but as contingent on the situation being considered. Ipperciel commends reasonableness in ethical deliberation—the balanced and judicious taking into account of the specifics of a situation—as against rationality, which is the following through of the logic of general principles. These suggestions offer something helpful. Disagreements about application tend to be more contingent and provisional than disagreements about principle and are thereby likely to be more amenable to negotiation. Also, psychologically there is less investment in specific applications than there is in the principles being applied, and generally it is likely that participants will be more willing and able to give ground and compromise on specific applications, where they would not on the principles involved.
None of the above provides a comprehensive formula for successful deliberation. The principles and practices I have identified provide partial responses to specific problems and in aggregate, I suggest, will help significantly toward effective deliberation. But the process remains highly challenging and complex and there will no doubt be many areas of difficulty. However, given the prospect of a richer, more inclusive process, and the possibility of outcomes that are acceptable to most or all involved, I suggest that the effort is justified.
I have identified a serious limitation which NICE suffers at present in performing one of its basic functions, in that it cannot effectively present the justifications of its decisions in the public arena, and cannot therefore legitimise its function. I have also attempted to explain the root of this problem. And I have suggested, in very rough terms, a structure and process which would work to resolve that problem. My suggestions would require NICE to take on a bigger agenda and a more explicit philosophy, and probably to become a somewhat larger, more complicated and more expensive organisation than it is at present. But its basic powers and status would not need to change greatly. If it can raise its profile as an agent of distributive justice, and effectively explain itself to the public in that role, it will not only secure its own legitimacy and enhance that of the NHS. It will also achieve a valuable goal of public education. That in itself might justify the extra resources.