The capabilities approach developed by Amartya Sen¹ and expanded further in conjunction with Nussbaum and others,² emerged as a response to theoretical difficulties in the conventional approach to welfare economics. One key insight is that welfare economics (whether applied to health or not) works by identifying optimal allocations of inputs and outputs whereas individuals are often concerned about other issues, such as rights, which form legitimate claims independent of the outcomes to which they might lead.³ In a sense, this was a surprising criticism of welfare economics because choice is so important to economic thinking. Nevertheless, the capabilities approach has been influential in a number of academic literatures as well as in policy making circles concerned with development and is now beginning to make substantial inroads into the fields of health and medicine.^{4,5,6,7,8,9,10} This paper contributes to the debate by examining the extent to which different aspects of the approach can provide insights into medical decision making. I will examine two areas of application: one on the micro level and the other on the macro level. These are not related in any direct or significant substantive sense, but they constitute a pair of choice problems that illustrate the diversity of such problems to which the capabilities approach might be applied as well as the way in which different aspects of the approach are highlighted in various applications.

AN OVERVIEW OF THE CAPABILITIES APPROACH

Capabilities, functionings, and freedom

Sen’s theory, which subsequently was developed in conjunction with the philosopher and political theorist Martha Nussbaum has, at its core, the idea that what people can do (capabilities), as opposed to what they actually do (functionings), should be the focus of wellbeing evaluations and government policy. Functionings can be overly materialistic, for instance, a person might live in opulence but be miserable. (This point is not particular to the capabilities approach, of course, although it is highlighted by Sen and others as one of the motives for emphasising what a person can do.) Furthermore, the choice context matters in our evaluation of functionings: a person not eating food because they are fasting is regarded differently from another person who cannot find anything to eat

This emphasis on freedom has been criticised by the philosopher Gerry Cohen as leading to an account of morality that is overly athletic. Cohen’s objection was that Sen’s characterisation of capabilities placed too much importance on the creation of opportunities and freedoms by human agents as opposed to the opportunities and freedoms themselves. This issue is still discussed from time to time by theorists although it has not been an obvious problem in the practical application to the measurement of economic development. Cohen’s¹¹ charge of athleticism was made early on during the development of the capabilities theory. Sen’s response was not to argue for athleticism but rather to suggest that Cohen’s was not an intended reading. Furthermore, when one examines the nature and force of Cohen’s points, it is clear his avowed emphasis is on concept clarification and that there is substantial overlap between Cohen’s version of “access to advantage” and Sen’s concept of capability. Indeed, whether Cohen’s careful analysis leads to a distinction of practical significance is moot as he allows that Sen is “supremely concerned” about the aspects he claims Sen’s account exclude. Certainly both agree on the broad issue, namely about the need to give accounts of freedom in the analysis of fairness and for that reason may be regarded as variations on a shared theme.

Nussbaum’s list of capabilities

Although Sen has refrained from providing a definitive list of capabilities, Martha Nussbaum has given such an account. Her list comprises 10 items, which she elaborates to varying degrees, as follows: normal life span; bodily health; bodily integrity; senses, imagination and thought; emotions; practical reason; affiliation; other species; play; and control over one’s environment. She claims these are all essential for a good life although many would say they are just one account, at a fairly abstract level, of capabilities likely to be regarded as important albeit to varying degrees.¹² Like most researchers in the field, I think this claim is too strong. I would argue, for instance, that the relative weight given to these capabilities would vary enormously with age, among people, and across cultures. That said, many such lists have been drawn up by psychologists, philosophers, and others and the commonalities are perhaps more striking than the differences (see, for instance, Sabine Alkire’s excellent survey¹³). Most development experts agree an implication is that the state should monitor opportunities in a range of life domains such as health and education as well as income. As has been argued persuasively elsewhere,¹⁴ the research problem is likely to dictate the account of substantive capabilities required: our use, which only requires a way of picking out a variety of concrete capabilities, is in no way tied to the universalistic claims which others have found difficult to justify. On the other hand, it is consistent with a more modest robustness claim, namely that many of the issues do emerge in recognisably similar forms in many different accounts of substantive values, a claim that Alkire’s survey¹³ could be used to support.

Differences between the capabilities approach and welfare economics

Some of the main differences between the capabilities approach and welfare economics are as follows:

• an emphasis on the options people have, as opposed to the activities they undertake

• the variety of material capabilities gives rise to a multivariate approach to wellbeing

• capabilities may make interpersonal comparisons easier than is the case under some ethical approaches—many economists are willing to take for granted that interpersonal comparisons of utility are impossible

• the capabilities approach is driven partly by the inadequacy of preferences (desires) as a measure of a person’s interests, in particular, difficulties arising from adaptive preferences.

Potentially, all four aspects could be relevant to medical decision making and ethics although in this paper I concentrate on implications for problems related closely to the first two of these points.

APPLYING THE CAPABILITIES APPROACH TO INDIVIDUAL MEDICAL DECISION MAKING

Quality adjusted life years (QALY) maximisation has tended not to be used directly at the level where the patient and clinician interact. Even the staunchest proponents of health maximisation do not argue that a doctor should insist a patient take the QALY maximising treatment. So in reality we observe a two stage ethical process. Measures of health gain are used to prioritise (in theory and sometimes in practice) the options that can be made available to patients on essentially utilitarian grounds. Then these are put (actually by the clinician, although acting on behalf of the state in many countries) to patients who have a right to refuse all the options presented. I submit that considerations of patient autonomy are to the fore over primarily utilitarian concerns about patient benefit and if that is the case, then what we see in practice, say in countries such as the UK, is a moral hybrid. Of course an indirect utilitarian rationalisation could be given, but if one were to examine the evidence such as it is, it seems plausible that most people would agree that patient choice is driven—in most healthcare systems—substantially by concerns about autonomy. And if that is true, the fact supports the theoretical case for an account of this hybrid which Sen outlines in his arguments for moving beyond old-style welfare economics.³

The capabilities approach, by introducing an emphasis on the opportunities a person has, makes it possible to give an account of ethical systems that value autonomy as it might feature in discussions about informed consent to the right to die. For this reason, the approach provides a conceptual basis, which is missing from health maximisation yet is necessary to describe ethical issues that are important in many medical choice problems.

An example: the right to die

Let us consider the right to die as an illustration of the way in which the capabilities approach might be applied. To be fair to the traditional approach of welfare economics, which can be regarded as one attempt to operationalise utilitarianism, one should recall that the Benthamite approach to rights was that they were “nonsense on stilts”, a phrase sometimes used to suggest that Bentham was against rights, although incorrectly. In fact, Bentham was against the idea of natural rights but held that people had legal rights that should be established so as to maximise overall happiness.¹⁵ If the net impact on the wellbeing of patients, carers, and the medical profession was positive, then a right to die could be given a Benthamite justification. So we should recognise that, as with any right, a utilitarian justification is not impossible but is contingent on what would bring about the most happiness.

Now compare this with the capabilities approach, in particular, the insights that might derive from a list of capabilities such as Nussbaum’s. The first three capabilities—to do with life, bodily health, and bodily integrity—are all potentially relevant to medical decision making in a number of ways. However, in each case, they suggest rights to positive health states in life or to duties of others to avoid bringing about such health states. In no case do they give any clear or direct indication about what we should think in right to die cases, and it is not until we move on to capability six, concerning practical reason, that we find a clear and unambiguous link. This capability is concerned with a person’s ability to form a concept of good and to engage in critical reflection about planning of one’s own life. If a person has principles governing treatment in the face of severe, debilitating, or terminal illness then it would be reasonable to regard those principles as part of a life plan. Capabilities concerning other species (capability eight) are not obviously relevant, but this leaves five capabilities unaccounted for. Capability four concerns abilities “to imagine, think and reason … in a ‘truly’ human way … [and] to be able to avoid non-beneficial pain”. Clearly, some right to die cases do centre around pain avoidance, but those cases in which people are concerned about the prospect and quality of their lives are not all such cases. So although capability four provides a rather direct possible justification for right to die, it has the capacity to do so, unlike capability six, only in cases where pain avoidance cannot be achieved by other interventions.

Capabilities concerning emotional expression (capability five) although related to capability four, are probably not relevant to the right to die as they concern the articulation of attitudes rather than pragmatic deliberation and choice behaviour. This leaves capabilities concerning affiliation (capability seven), play (capability nine), and environmental control (capability ten). Nussbaum divided affiliation into two parts, the first relating to the ability to live with, and towards, others, and this too seems irrelevant. On the other hand, the second part, “having the social bases of self-respect and non-humiliation … [which involve] being treated as a dignified being whose worth is equal to that of others” could be relevant. In some cases, a right to die would provide the social basis patients wanting to exercise such a right feel they need to maintain their self-respect and avoid humiliation. Capability ten, the ability to control one’s political and material environment adds little to the autonomy based justification of a right to die which can be found in capability six, but one can say that, at least, capabilities six and ten appear to argue in the same supportive direction for such a right. The capabilities approach is not committed in general to the right to die because one could reasonably object to Nussbaum’s list. However, the fact that two items on her list can be directly related to such a right, and in the same supportive direction, does seem to reflect a way in which many advocates of the capability approach would use it to form a view about the right to die. The claim seems rather robust as many accounts of substantive value have something like autonomy in them.

The fact that the capabilities approach does not rule out a right to die is a feature it shares with utilitarianism, and this is therefore not something that can be used to distinguish between the approaches. Utilitarianism focuses on wellbeing and mandates such a right if it promotes the total good, whereas on the other hand, the capabilities approach is not complete as a political theory, in the sense that it points to elements of an evaluative space without being linked to a particular method for valuing particular capabilities. (This is true even in Nussbaum’s version which, by virtue of her claims about the universality of her list is more objective than most.) It seems difficult, therefore, to say that the capabilities view would necessarily argue for a right to die. On the other hand, it gives structure to the concepts of capability and wellbeing which helps identify issues to which proponents of a right to die are likely to appeal. In this sense, the capabilities approach (Nussbaum’s version) is more explicit about wellbeing than its utilitarian counterpart: if these capabilities, though not their relative weights, are universal, then this explicitness is valuable. Hence the capabilities approach gives a more detailed and direct account of factors that might be relevant to a right to die than that which we might expect to obtain from the utilitarian derived approach to welfare economics.

Concept of “right” and role of rights in the capabilities approach

This analysis uses the concept of right, although without giving a specific account of rights and duties: to do this is neither my aim nor is there an obvious account on which one might draw. However, a couple of points are worth noting. Whether a right implies a co-relative duty has been much debated by philosophers though the issue remains open. In this particular case, our intuitions might be such that even if we accept that a person has a right to die, it is less obvious that the person’s doctor has a duty to hasten that person’s death in general (that is, to realise the plan on the patient’s behalf if he or she was incapable of realising it without assistance). For instance, a clinician might be obliged to respect a request to avoid or desist from treatment and yet not be obliged to actively intervene by administering a drug, say, so as to hasten death. This is still not incompatible, however, with the maintenance of a logical link between duty and right as it could be that co-relative duties fall on the state and not on any particular clinician. (This is a question for the capabilities approach, yet it is one that most other ethical views would have to address too.)

The second and somewhat related point is about the role of rights within the capabilities framework. The early usage of the term “capability rights” is currently waning though the underlying sense and motivation, that people should be entitled to choose to do a variety of things usually of value to people by virtue of their humanity, is not. It is said that the capabilities approach is not a theory of justice—which accurately reflects its technical development in recent years—but it is clear that in practice people are concerned about capabilities as opposed to just functionings because they believe they matter to issues of fairness or equity, as well to questions of efficiency. One approach to fairness that is consistent with the emergence of capabilities as a response to difficulties in utilitarian welfare economics, is the idea that what matters in decision making where there is more than one agent are legitimate claims and that rights constitute one kind of legitimate claim. This view, which has been argued far more extensively elsewhere,^5,7 uses the term “right” in a different sense from that which the term “right to die”, as used here, requires. In brief, the idea is that there are a small number of legitimate claim types (for example, consequences, rights, due process) which when aggregated serve to specify and prioritise allowable policies. This approach allows that a capability (such as that implied by a right to die) might be warranted because it brings about desirable states of affairs, it respects fundamental human rights, or it reflects due process (perhaps society has voted for a right to die) or some combination of these desiderata. Without any such theory of justice, the capabilities approach, as currently developed, leads one to concentrate on the value of capabilities as the justification for any rights to capabilities, which at first pass suggests a consequential justification of any rights to capabilities that one might posit (though a fuller account would clearly need to consider how best to deal with the intrinsic value of freedom¹⁶).

CAPABILITIES, RESOURCE ALLOCATION, AND FAIR INNINGS

Health maximisation

As an approach to economic welfare, the capabilities approach can be expected to have insights at the macro level too. The traditional approach to welfare economics can be viewed as an attempt to operationalise utilitarianism. (In general, philosophy tends to concentrate on the identification and grammar of values whereas economics has focused on the mathematical aspects of specification.) Health economics has developed a second interpretation (alternative but related to conventional welfare economics) of utilitarianism that does not depend on market assessments of value but uses assessments of the quality of life (QALY). It has been argued by health economists that treatments should be offered so as to maximise the total number of QALYs produced by the total population. This social choice rule, also referred to as health maximisation, is appealing until one examines some of its implications in more detail. In general, the view taken here (and consistent with views taken by an increasing amount of work in health economics^4,7,17–21) is that:

health maximisation can be justified either as a useful first order approximation to the social objective function for health or even just as a first attempt to make ethical principles for healthcare rationing explicit.

However, beyond that there is room for debate, even within economics. Here, I discuss an aspect of the comparison between the capabilities approach and health maximisation. The argument is that the fair innings defence of health maximisation is flawed, not as a justification of age related rationing, but rather as defence of health maximisation. In contrast, the capabilities approach is capable of providing a more natural and flexible justification (that is, non-ageist basis) of age based rationing.

The fair innings argument

One of the implications calling health maximisation into question concerns the charge of “ageism”.²² Health maximisation is not directly age related in that QALYs are maximised whoever produces them, old or young. However, it is accepted that age enters indirectly, and significantly, in that young people will tend, ceteris paribus, to produce more QALYs. If ageism involves the inappropriate use of (old) age as a criterion for exclusion, then an important ethical question is whether the indirect age relatedness of health maximisation is ageist.^23–26 Consider the following situation: suppose we can treat one of two people, both of whom if treated would fully recover but otherwise die. If these patients were equal in all material respects except age, then QALY maximisation would advocate treating the younger person because that person has the greater amount of life left and will generate more QALYs. A case for this is the so called fair innings argument which holds that older people, because of the length of life they have enjoyed compared with younger people, should yield priority to those who are younger. The argument has been much debated but for present purposes it will be useful to mention three particular difficulties, which might be thought to apply to an authoritative version of fair innings, put forward by John Harris²⁷ (see in addition the work by the philosopher, Michael Rivlin,²⁸ who also uses this account).

First, the fair innings argument only supports some of the age discriminations that QALY maximisation makes. Specifically, it applies to comparisons where the age differences between those in need are substantial and where, therefore, one party might be said to have had a fair innings and another not. However, what are we to do when making comparisons between say a 35 year old and a 50 year old? QALY maximisation would clearly prioritise the younger patient on the grounds of her or his greater life expectancy. This could be called a form of ageism but one could hardly say that it was one justified by the fact that a 50 year old had had a fair innings. (At least proponents of the fair innings argument do not make such a claim as they put the fair innings mark at 70 years.) The fair innings argument sets up a partial (which might also be fuzzy) ranking whereas the QALY approach is inherently complete so the fair innings argument cannot justify all the age related rationing that health maximisation allows. Second, a fair innings approach does not support, or at best sits awkwardly with, the sum-maximisation foundations of utilitarianism, welfare economics, and health maximisation. The fair innings argument is one based on equity between individuals whereas the justification for QALY maximisation is that it does most good—so the former does not provide a justification that goes to the heart of the latter. Third, when combined with facts about differences between life expectancies of men and women as Tsuchiya and Williams acknowledge,²⁹ the fair innings argument appears to argue for prioritising the treatment of men over women, and to many this consequence is unappealing.

Capabilities approach versus fair innings

If we take a capabilities perspective, however, all three difficulties may be avoided. First, if capabilities are to play a central role in defining the space in which actions are evaluated, it is crucial to ask what capabilities we wish to promote. The capabilities that people actually wish to develop, in general, seem to vary across the population with respect to a number of factors. Abilities to succeed in sport, for example, depend on health status but physical endowments which we accept, normatively, will vary. Age is another factor—and indeed recent evidence suggests that people accept a decline in mobility with age while remaining unaccepting of pain.³⁰ These variations in the desired levels of functionings fit naturally with the view that the capabilities we expect or hope for will vary with age (amongst other variables). Proponents of the capabilities approach have emphasised the fact that this approach to ethics combines objective and subjective elements. Broadly speaking, there is a tendency to be objective with respect to the dimensions of capabilities and functionings while allowing the relative weights and trade-offs between these dimensions to be subjective. So if society agreed that the acceptable levels of functioning were dependent on age and other personal factors, this could have an impact on the way in which treatments are allocated. Note that with this approach, the capabilities view discriminates with respect to age, but for reasons that are to do with the social notion of desirable health states. Such discrimination is not self-evidently discriminatory and further arguments would have to be produced to show that such discrimination is inappropriate. Furthermore, if people did not accept that such ideals should vary with age, the capabilities approach would not be able to accommodate an age indifferent approach. We would ask about the capabilities regardless of age. (However, it seems so embedded in the concept of ageing that physical and mental capabilities do decline, that the age indifferent approach seems somewhat implausible.)

Turning now to the second issue, namely that of sum maximisation, we should note that in developing a capabilities defence of age related health care, no appeal was made to the sum maximisation aspect of health maximisation. The significance of capabilities depends on the weights assigned to them, something that may be done by experts, public consultation, or some combination of the two. We may wish to maximise these weighted capabilities, but by varying the weights we can obtain very different ethical prescriptions. Although from a technical viewpoint, weighted sum maximisation is not much more complicated than unweighted sum maximisation, ethically, the approaches could be quite different. Put in another way, weighted maximisation may reflect quite different ethical bases to the utilitarian approach which sum maximisation formalises.

Thirdly and finally, we turn to the issue of equity and discrimination in favour of men. There are two possibilities here, both consistent with previous remarks about variations in the social value of capabilities. One could hold that differences between men and women, which gave rise to different profiles of capabilities over the life course, were not such that they merited intervention. If, for example, lower life expectancy reflected the fact that more men choose to work than more women (many jobs reduce life expectancy), then the element of free choice would negate any moral pressure to reduce inequalities in life expectancy. Alternatively and in contrast, it might be argued (more plausibly to boot) that choice in matters of work force participation is constrained and that if wages and promotion opportunities were more equal between the sexes, women would choose higher levels of workforce participation. Even if this is true, it provides no reason to provide more healthcare resources for men. Tsuchiya and Williams argue, correctly in my view, that one has to look at the inequalities in the round, not just in the health arena. But neither the fair innings argument nor the principle of health maximisation that the former was designed to support, provides a reason for looking at inequalities in the round. On the other hand, the capabilities approach, particularly Nussbaum’s explicit and comprehensive version does just this.

One last point

Tsuchiya³¹ suggests, for example, that the fair innings argument is particularly supportive of the decision to treat a 30 year old before a 60 year old on the grounds that the 60 year old has had the opportunities which the 30 year old has not yet had but might have if treated. This, she claims is an equity argument rightly of course—but it is also, quite explicitly, an opportunities or freedom argument—that is, precisely the kind of issue that the capabilities was designed to articulate. Indeed, it serves to underline the way in which the capabilities approach has the potential to give a more general and comprehensive moral account of how QALY information could be used in practice.

HEALTH FUNCTIONING AND CAPABILITIES IN OTHER AREAS OF LIFE: FINAL CONSIDERATIONS

As noted, the capabilities approach emphasises the multivariate nature of human wellbeing and there are many lists of human values that share this perspective. There is a direct sense in which the QALY relate to capabilities: QALYs incorporate life expectancy—and longevity appears on a number of lists of human values and indicators of human progress. However, from a multivariate perspective it seems reasonable to suppose that core features of health care, such as mobility and pain, also have substantial implications for both capabilities and functionings in a range of areas. Indeed, it can be argued that health status is, like income, an indicator of capability. (However, if we accept this, we should also accept that, like income, it is a rather crude and imperfect measure. Just as some can do more with a fixed budget, some are able to overcome physical incapacity better than others.)

The point is that we cannot assume that health status is independent of wellbeing in other areas of life, though the identification of a separate health component, which is then maximised, will only necessarily contribute to the maximisation of overall wellbeing if independence can be assumed. The practical significance of this point for human wellbeing depends on the empirical relation between health states and non-health aspects of life. If the interrelation is empirically material, the QALY measure, which looks at health status, would be too narrow a measure on which to base priorities. For example, two people might have the same expected QALYs for the rest of their lives but quite different opportunities, because of their individual health status, to seek and to find paid employment. When it comes to deciding which of two to treat, we might hold that their capacity to find work should not affect their right to be offered treatment. On the other hand, the capabilities approach is consistent with the view that when it comes to the allocation of funds at a macro level, we would want to give more weight to those healthcare interventions that have significant benefits both in healthcare and non-healthcare terms.

Perhaps the moral is that when we compare QALY maximisation and the capabilities approach, we are not comparing like with like. QALY maximisation is a theory whereas the capabilities approach is broader and better thought of as a framework from which theories can be derived. As a generalisation of the utilitarian approach to welfare economics the capabilities must be able to accommodate things that utilitarian welfare economics cannot. Where practice is informed by theory, theory’s sharp corners are often smoothed out by use. The explicit and direct treatment of freedom that the capabilities approach offers and utilitarianism does not is often respected by practices that would have described themselves as utilitarian. What both approaches do share is a recognition that wellbeing is morally significant.