1. THE PRIVATE VALUE OF HEALTH AND THE CONTRIBUTION OF HEALTH TO WELL-BEING

‘Health’ is a matter of how well the parts and processes in our bodies and minds are functioning. However, for allocational decisions, we do not need a measure of health but a measure of the value of health. This value depends on how health interacts with individual and social objectives and the environment. According to Hausman, there are deep problems with defining the private value of health states as the contribution of health to well-being (something related to what health economists and epidemiologists would call health-related quality of life). He also argues that the alternatives focusing on autonomy or capability are even more troublesome.

It is hard to say what well-being consists in. In line with his earlier work (e.g. Hausman and McPherson Reference Hausman and McPherson2009; Hausman Reference Hausman2010), Hausman claims that it consists in neither the satisfaction of preferences, nor subjective feelings of happiness. Hausman argues that the most attractive way of defining well-being is in terms of human flourishing, which depends ‘on how the things that make human lives good (such as friendships, happiness, health, or a sense of purpose) are integrated into the dynamic structure of that individual's life’ (141). Two questions then come to the fore. Is it possible to separate the effect of health on well-being from the effects of the other factors? And how to measure this separate effect?

With respect to the first question, Hausman rejects Broome's (Reference Broome, Murray, Salomon, Mathers and Lopez2002) argument that the effect of health cannot be separated from the effects of other factors. A simple formalization helps to understand the argument. Suppose well-being W is a function of health h and other factors f. Even if this function W(h_i, f_i) is the same for all individuals i, the value of a change in health from h _i0 to h _i1, given by ΔW_i = W(h _i0, f_i) − W(h _i1, f_i), will obviously depend on the value taken by f_i, unless the function W(.) is separable. To show that it is still possible to measure the value of health, Hausman proposes to take either the average direct contribution of health to well-being (in my notation $( {\frac{1}{n}} )\mathop \sum_i ( {W( {{h_{i0}},{f_i}} ) - W( {{h_{i1}},{f_i}} )} )$) or the direct contribution of health to well-being in some standard environment (in my notation $W( {{h_{i0}},\overline{f}} ) - W( {{h_{i1}},\overline{f}} )$).

With respect to the measurement question, both subjective experience and preferences may give useful evidence. Subjective experiences such as pleasures and pains are indicators of well-being, but they are often unreliable, especially in the case of cognitive, emotional and sensory problems. Preferences are ‘total subjective comparative evaluations of alternatives’ (75). They can serve as useful evidence, but only under strict conditions. If a person is a competent evaluator, then a third party is justified in drawing conclusions concerning the person's well-being from the person's preferences whenever that person (a) is free from significant deliberative flaws; (b) is well informed; and (c) her preferences are self-interested.

In the light of these conditions, Hausman formulates a devastating (and in my view convincing) criticism of the present practice of health economists. Their concept of preferences is primitive. The actual surveys (e.g. the EQ-5D or HUI(3)) that are used to assign values to health states do not encourage careful reflection, and they do not screen out answers that reveal rational flaws. They do not ensure that respondents have the information they need. Because the survey questions are difficult and in some cases not well defined, the quick answers given by respondents do not reveal settled valuations but only reflect gut feelings. Further evidence that the values inferred from preference elicitations are unreliable is given by the systematic differences in the values assigned to similar health states in different surveys. In eliciting preferences in order to assign values to health states, health economists are delegating the task of evaluating health states to the survey respondents, but these are not in a better position than the economists themselves to make accurate evaluations.

So far so good. My summary until now has brought me to Chapter 12 in the book. Well-being is human flourishing. The personal value of health is more than only its effect on well-being, but as a first approximation this is not too bad. We should be aware that the contribution of health to well-being is difficult to measure and that the techniques that are used in the present literature are deeply flawed, but we know in what direction we could go in order to improve them. In fact, Hausman offers a list of suggestions on how to design better studies (151–2). When the reader has arrived at this stage, it is as if Hausman has laid out an interesting research programme.

There are some technical questions with respect to that research programme. Hausman claims that we need a cardinal scalar measure of health to guide allocational decisions. This is not obvious (cardinality would not be needed if we adopted the leximin principle) and, in any case, when discussing preference measurement it is essential to distinguish ordinal preferences (the ordering of different states) from the specific cardinalization used to give a scalar value to health. It is well documented that the scaling of self-assessed health may depend on aspirations and expectations of respondents, and that these aspirations may differ between individuals, e.g. because some individuals will adapt more rapidly and more completely to changes in their health situation. Present techniques are not careful about the distinction, but an adequate research programme should take into account the scaling issue. Moreover, Hausman regularly states that preference relations have to be complete to be useful. This is not true. In some cases, incomplete preference relations are sufficient to give definite answers. As a matter of fact, Hausman himself points to the possibility that ‘health states should be assigned an interval of values rather than a single value’ (154).

There are two more important issues. First, even if we accept Hausman's argumentation that it is possible to value health as a separate component of well-being, this does not imply that this is also desirable. The question remains why one should not focus on well-being as such. Hausman argues that it is easier to measure the value of health than to measure well-being. In my view, his argument is not very convincing. If well-being is not separable, the value of health will be conditional on other dimensions of well-being and measuring this conditional value seems particularly difficult.

Second, how to tackle the issue that different individuals will give a different weight to health? Underlying Hausman's proposal to work with an average contribution of health or with the contribution of health in some standard environment is his desire to obtain a unique value for each health state which is the same for all decisions. Yet, as soon as we focus on the contribution of health to well-being, the value of this contribution will certainly be different for different individuals and it will change over time if the other factors f_i change. It is hard to see what would be the optimal ‘standard’ environment, or over which population and at which point of time one should take the average. Why not just accept that there is a distribution of personal values at any moment in time (to be estimated with a representative sample of the population) and then base allocational decisions on the characteristics of that distribution?

2. THE PUBLIC VALUE OF HEALTH

Hausman does not give a convincing answer to these questions, but for him this is not a real problem. Indeed, while he might be interpreted as sketching a research programme, it turns out that he does not find this research programme interesting. The cards come on the table in Chapter 13. Hausman argues in that chapter that we should not in the first place be interested in the personal value of health (‘the contribution to whatever the individual cares about or should care about’ (158)), but in its public value. Of course, it would be a coherent ethical position to state that the public value of health should be in one way or another an aggregate of personal values (possibly taking into account distributional issues), but Hausman rejects this view in favour of a liberal political view. ‘State action should not be governed by private interests or by private views of what makes life good. The central responsibility of government is to create an environment that secures the basic prerequisites for common activities and competences’ (159–60). He then distinguishes two main dimensions in the value of health states: first, the activity limitations they impose as these necessarily lead to a restriction of individual opportunities, and, second, the suffering they involve. Indeed, a good state also has duties of care and compassion and the reduction of suffering is therefore one of its legitimate roles.

Based on this view of what aspects of health states matter from the perspective of social policy, Hausman offers a very speculative account of how to assign public values to health states. Criticism is easy here. The classification of activity limitations (based on ADL and IADL) and the classification of distress (four levels from ‘okay’ to ‘severe suffering’) are not more refined than the levels in the traditional health economic approaches. Moreover, the normative assumptions made in the valuation exercise (e.g. additivity of the two dimensions) and the measurement methods are extremely doubtful: measuring the value of activity limitations in terms of the median wages of people with different sets of activity limitations is a more economistic approach than most health economists would endorse. Of course, Hausman himself keeps emphasizing that his proposal is at best a first input into a collective deliberation procedure.

Putting implementation issues aside, the proposal to interpret the public value of health from a liberal political perspective is in my view by far the most original and interesting part of the book. Once one takes this position, the two issues mentioned before no longer seem relevant. Measuring overall well-being is obviously no longer needed and it is evident that one has to look for an ‘objective’ determination of the value of health states. However, since I am a liberal welfarist (a position that Hausman qualifies as ‘attractive’ but ultimately wrong), I do believe that more can be said about this.

First, I think it is dangerous to create a separate sphere for health and to argue that the Ministry of Health should only care about health issues. There are indeed good pragmatic reasons that explain why democratic states have a separate Ministry for Health. And for many policy decisions it is certainly legitimate to take only health issues into account – if the effects in other domains are relatively less important or impossible to predict. But this is not always the case. A Ministry of Health deciding about the level of out-of-pocket payments certainly should take into account the income level of the patients, and it could be very misleading to completely neglect the labour market effects of different ways of organizing the health care for those who are approaching their retirement. On the other hand, other Ministries (of Labour, of Education, of the Environment) often cannot neglect the health effects of their policies. More generally, the separation of public policies in different spheres is only meaningful if the size of the government budget to be allocated to the different spheres is independent of the choices that have to be made within these separate budgets. This is the exception rather than the rule. A striking example of the problem is precisely to be found in the sphere of cost-effectiveness analysis. The most important policy issue nowadays is the optimal size of the government budget going to health care. This ‘optimal size’ obviously depends on the quality of the new treatments that have to be evaluated by cost-effectiveness analysis. An approach that takes the total health-care budget as given remains silent about this crucial choice. Moreover, restricting the ethical discussion about priority setting to choices within a given budget leads to a misleading framing of the issues. The real choice is not between health care for the elderly in the last months of their life and health care for the young – the real choice is between health care for the elderly in the last months of their life and more private consumption (a second car and a more expensive holiday trip) for the healthy. It is only when formulated in this way that values like ‘compassion’ and ‘solidarity’ get their full strength – or become really challenging. And it is only in this setting that we experience the real bounds to what we can achieve in health terms.

Second, we cannot neglect the issue of inter-individual variation either. Hausman claims that no aggregation is needed in his ‘public value’ approach. Yet, this is a bit optimistic, as people will have different opinions about how to value the different cells in the (activity limitations, suffering) matrix. How to reach consensus about this, i.e. how to aggregate opinions? Hausman is well aware of the dangers of manipulation that are inherent in any collective deliberation, but he remains silent about the conditions that such a deliberation should satisfy in order to lead to legitimate results. Moreover, there are difficult questions concerning the neutrality of the interventions of the liberal state in a world where citizens have widely different life plans. Hausman rightly admits that the liberal state cannot be fully neutral in the range of activities that it aims at offering, since some activities are more significant and significant to more people than others, and he then states that ‘a nation's prioritization of the use of resources depends on what people in that nation value doing, on what promotes citizenship, and on the importance that activities have in most human lives’ (162). But how to measure relative ‘importance’ and what is the exact meaning of ‘most’? There can be no doubt that a minimal level of health is a necessary prerequisite for any decent life, but the discussion in rich Western societies is not only about that minimal level – it is also about trading off some more months of life versus more private consumption.

On the basis of his own account, Hausman comes to a moderately positive stance on cost-effectiveness analysis. Given the importance of ethical constraints, the inevitable imprecision in health measurements, and the uncertainties surrounding the measurement of costs and (health) benefits, cost-effectiveness information provides little credible guidance. However, according to him even a little guidance is still worth having. In my view, this position is too easy. It seems important to try to improve the quality of cost-effectiveness analysis and to overcome its limitations. In a world where the real political debate is about the size of government intervention in the health-care sector, a technique that takes the budget as given (or the threshold for the incremental cost-effectiveness ratio as outside the formal analysis) is of limited value. It may even become dangerous in a world in which ‘realistic’ decision-makers not only claim that the health-care budget is limited (which is obviously true) but also that it cannot be expanded beyond what is spent now (which is obviously wrong).

Hausman has written a marvellous book, which is a must-read for everybody who is interested in the crucial challenges raised by allocational decisions in the health-care sector. His criticism of existing practice is fully convincing. His emphasis on public values in a liberal conception of the government is fascinating, but raises difficult questions. One can only hope that his book will contribute to a deeper discussion between philosophers and health economists on the values underlying cost-effectiveness analysis. This is urgent, as the actual techniques are close to becoming canonized – and that would be a big mistake.