- Email: [email protected]
The nature, meaning and measurement of health and illness: An economic viewpoint
.Soc.Sci. Med. Vol. 20, No. IO, pp. 1023~1027,1985 Printed
in Great
0277-9536/U $3.00 + 0.00 Pergamon Press Ltd
Britam
THE NATURE, MEANING AND MEASUREMENT OF HEALTH AND ILLNESS: AN ECONOMIC VIEWPOINT ALAN WILLIAMS Institute of Social and Economic Research, University of York, Heslington,
York YOl
5DD, England
Abstract-Health can be seen as a capital asset, subject to depreciation due both to the passage of time (ageing) and to ‘wear and tear’. It is valued for the flow of pain-free time and energy it offers us, and we can increase the quantity and quality of this ‘flow’ by appropriate care and maintenance. which may take the form of adopting a healthy life style or in seeking health care. The valuation of health (as a capital stock) can then be seen as the valuation of the time profile of the stream of pain-free time and energy we expect to get out of it. This in turn depends on the uses to which this time and energy can be put. Its use in paid work is relatively easily valued but unpaid work (especially work in the home) continues to be a source of difficulty, both in principle and in practice, in the search for appropriate money values. Leisure time is still more difficult, though for those in paid work whose remuneration is sensitive to the workileisure balance, foregone net earnings at that margin may be used in the marginal value of leisure time. But the appropriateness of any valuation depends on the use to which it is to be put, and it is important to distinguish whether it is to reflect ‘individual’ or ‘group’ values, and whether it is for ex post compensation associated with some particular event, or for use in ex ante decision making where it is not known who the particular gainers and losers will be.
SCENE-SETTING The hallmark of Economists’ thinking about health and illness is the attempt to characterise these phenomena by using existing economic concepts, reinterpreted with varying degrees of sophistication, in the hope that we can thereby gain some distinctive insight into the nature and meaning of health and illness, and use our measurement techniques (even though originally devised for other purposes) to measure (and particularly to value) health and illness. This inevitably prompts the question: is health not very different from other good things that people want to have (and is illness not very different from other bad things that people want to avoid)? At a trivial level the answer is obviously ‘yes’, otherwise we would not have a distinctive terminology and set of concepts to describe health (and illness). But in a more fundamental sense is having good health any different from having a nice house or a reliable car or fine clothing or a good education? Before proceeding further we must make an important distinction between good health as an objective (or as a source of welfare in itself) and health cure as one means of attaining that objective. In any hierarchy of ends and means, health is likely to come ‘high up’ and, for most people, higher up than the house, car or clothing. A good education (if by that we mean ‘treatment’ at School and University) is more like health care (i.e. ‘treatment’ at Consulting Rooms or Hospitals), but where the corresponding phenomenon to ‘health’ is ‘knowledge’ or ‘skill’. Thus at a high level, health is seen as an element in personal welfare, and in some extreme formulations (due to WHO rather than to economists) is virtually identical to welfare (who can ask for more in this world than ‘complete physical and mental wellbeing’?) When discussing the nature of illness (taken simply to mean a departure from good health, and including here injury) economists’ attention has been concen1023
trated on two characteristics which are important in interpreting its significance in an economic context. The first is its unpredictability and potentially catastrophic consequences, and the second is, whether or not it renders its sufferers unfit to judge their own best interests. Discussion of the first characteristic leads inexorably to consideration of insurance analogies, and the extent to which provision for illness can be made efficiently and equitably through insurance mechanisms (public or private). Discussion of the second characteristic is more worrying for economists, since most of our analytical structures rest on the notion of ‘consumer sovereignty’. If, in the health field, there is legitimate doubt whether consumers are (or ever will be) both sufficiently knowledgeable about the nature of illness, and unemotional enough to be able to process in a rational manner such information as they have, then their best strategy might be to place themselves in the hands of some knowledgeable and trusted person who will act on their behalf. This is what parents encourage their children to do, and what the medical profession encourages us to do, and has led health economists to see the link between illness, the seeking of health care, and consequent health status, not as the working of simple demand and supply models, but as the operation of an agency relationship in which ‘demand’ is ‘mediated’ by (imperfect) agents who are actually also suppliers of (some kinds of) care, and whose notion of health (or welfare) may not be precisely that of the ‘demander’. In order not to trespass on the territory staked out for others, I will not pursue these general observations further here, since they will surely be central to the discussions on health professions and professionalism. Instead I will go off in a different direction, and concern myself with the question: what is it that makes health valuable, and what value do we attach to it? I do not claim to be able to give definitive answers to these searching questions, but merely to
1024
ALAN WILLIAMS
indicate how economists have tackled them. Before doing so I am going to assume that we know what health is, and for anyone who feels that it is altogether too dismissive of an area of controversy which is both crucial and unsettled, I will make clear my own personal view, which is that health is best regarded as the product of quantity of life (measured by life-expectancy) and quality of life (measured by physical mobility, freedom for pain and distress, capacity for self-care, and ability to engage in normal social interactions). I know that some of these phenomena are socially conditioned, and not easy to measure, and that they constitute an important area of investigation in themselves, epitomised in the literature on health status indices. But, on the whole, economists have been concerned with a different set of issues, and it is to these that I now wish to turn.
STOCKS
AND FLOWS
As any owner-occupier will testify, a house is both an asset and a liability. It is an asset in that it provides shelter, a pleasant environment in which to pursue leisure activities, sometimes (e.g. as I write this paper) a place to work, and, in another aspect entirely, a material investment to be converted into money at some later date. It is a liability in that it requires attention to the structure and fabric from time to time, the more so the older it gets and the more harshly it is treated by us or by the environment in which it stands. It also requires cleaning and heating and lighting if it is to provide the services we seek from it. Some of this maintenance and repair work we can do for ourselves (or some of us can!). We can reduce the need for it by being prudent in how we treat it (avoiding wild parties, or throwing things at other members of the household). But some of the work we have to get the professionals to do for us. We can also get rid of some of the financial risks (e.g. those associated with fire, storm damage, burglary) by insurance and this may make us marginally more relaxed (i.e. careless) about protecting ourselves from these hazards. We can also, of course, choose to save money and neglect the property, and do nothing till disaster strikes (e.g. the roof falls in), but that may prove to be very costly in the long run, and may well shorten the life of the property quite significantly. Mutatis mutandis, what has just been said about ourselves as owner-occupiers of our homes could equally be said about ourselves as owner-occupiers of our bodies. The terminology and balance of emphasis would of course be different, and I do not believe (as some people appear to do) that we can switch bodies as we can houses. A better analogy would be one in which we each inherit at birth a house in course of construction, and we are saddled with it the rest of our lives. But it is dangerous to push these picturesque comparisons too far, and the point I wish to make is simply that health can usefully be viewed a capital asset, which depreciates both with time (due to the aging process) and with use (due to life style and environmental stress) but which can be restored (to varying extents) by appropriate care and maintenance. What this ‘asset’ then provides us with is ‘healthy time’ (or, if it is not in very good shape,
‘unhealthy time’), and it is thisJEow which we need for all our activities (including sleeping) and which terminates at death. Some uses of time draw on our health capital more than others, however, (e.g. smoking is a very ‘health-capital-intensive’ activity, as are some hazardous occupations like labouring in the contruction industry). If these uses of time are seen as essential ‘inputs’ into everyday activities, and if healthy time is much more efficient than unhealthy time as such an input, then it is clear why we value health, and it also gives us a clue as to how we might set about valuing it. Let us consider first the simplest case, which is the activity of working for money. If it could be shown that better health is associated with higher earnings, then that clearly gives us a rather direct differential value for healthy over unhealthy time (and, indirectly, also indicates the maximum we would be prepared to pay for treatments which improve our health if there were no other benefits therefrom). Thus a great deal of the early work by economists was concerned with estimating the relationship between health and productivity, viewing the individual as a piece of human capital equipment, rather like a piece of physical capital equipment, to be valued by the value of the flow of services it produces. Even in this narrow context it was noted there is a difference between the value to the community of that individual’s contribution to output (which is the value so added to the final product) and its value to the individual (which is his or her net take home pay). It is also to be noted that this sort of calculation does not (or should not) purport to measure the entire value of health to the individual (or indeed to the community of which he or she is a part) since it relates to only one segment of (some) individuals’ activities. It has, nevertheless, played a dominant role in the value-of-health literature, probably because it deals with an intrinsically interesting and important phenomenon, and because there is fairly readily accessible data about it, waiting to be exploited, (which unfortunately is not the case with most of the other phenomena to be considered later in this paper). Work which is not done for money has always created great problems for economists. We know that housework ought to be in GDP, and it is only because of practical difficulties that it is not. Attempts to value it by what it would (or does) cost to get equivalent work done by hiring people to do it, often come up with large sums which people find it hard to believe, because they reckon that most people could not actually afford such sums, so it must be wrong. I think this is a mistaken objection, because the value of (unpaid) housework should be counted both as imputed income and as imputed expenditure by the household (just as the rental value of owner-occupied houses should be counted as both income and expenditure of a household) and if this were done most households would show up as much richer than they are if only money income is counted. A more telling objection is that ‘marketed’ housework is such a small (and probably unrepresentative) proportion of all housework that it is unlikely to provide a suitable basis for valuing the rest. If it is only the relatively rich and/or desperate who pay others to do a
The nature, meaning and measurement of health and illness significant amount of housework for them, and since there is also some social stigma attached to that kind of work on the supply side, it is possible that the market values are rather an unreliable guide as a basis for general valuation of such activities. On the other hand, the growth of homehelp services, meals-onwheels, etc. to assist ‘unhealthy’ people to cope with housework, may give us data on which we can draw with more confidence when estimating the value of health to the houseworker. To the extent that sick (unpaid) houseworkers draw on other (unpaid) houseworkers to help out, or simply leave tasks undone, this would then indicate that the costs of these ‘marketed’ services are to be regarded as an upper limit on the value attached to these unpaid (or unperformed) tasks. The third broad category of activities requiring time, where being healthy may be more valuable than being sick, is leisure activity (loosely termed ‘consumption’ in economics). Being unhealthy may actually make some consumption activities impossible, or at least highly dangerous, (e.g. marathon running, party going, smoking), but whether that is so or not, it will more pervasively diminish the pleasure to be found from them (e.g. eating and drinking, conversation, reading, etc.) (just think of toothache!). Since work is viewed in economics as instrumental, and (at the margin at least) as diminishing welfare, whereas consumption is the ultimate objective and the sole source of positive welfare, then the consumption benefits of health are seen as giving rise to its ‘intrinsic’ worth, and as closely related to the value of leisure time (which, at the margin, will also be systematically related to the value of working time especially for those able to adjust the balance between the two fairly closely in accordance with their preferences). So, by and large, economists have linked the ‘stock’ concept ‘health’ with its ‘flow’ concept ‘healthy time’, then sought to estimate the differential value of healthy over unhealthy time in its various uses (remunerated work, unremunerated work and leisure). The work uses are sometimes called the investment benefits of health (since ‘work’ is a process by which time is invested either in earning money with which to enhance consumption possibilities, or in producing such consumable goods and services directly), whilst the leisure uses are viewed as consumption benefits. Indirect and direct consumption might be a better terminology, reserving the term ‘investment’ in this context for activities designed to increase the health stock itself (e.g. sleeping, consulting a doctor, or jogging. assuming, as is undoubtedly true in my case, that none of these is intrinsically pleasurable). One final observation at this general level, before looking more closely at measurement problems, is that, like everyone else, economists have found it easier to focus on issues of life and death than on issues of quality of life. Just as medical practice has been greatly concerned with mortality, and hence with life-threatening situations to the relative detriment of non-life-threatening situations, so have economists tended to interpret the value of health as the value of life, and then equate life with life expectancy, ignoring quality of life. There are, of course, a few honourable exceptions, but modesty forbids.
WHAT
IS THE
1025 QUESTION?
There are many reasons for wanting to measure and value health, and generally it will be the case that the appropriate way of measuring and valuing health will depend on the purpose to which the measure is to be put, i.e. on what question we are asking. We might be interested in predicting individual or group behaviour, because we want to know, say, whether increasing prescription charges will adversely affect people’s health or merely reduce their consumption of useless or harmful drugs, or change volume of GP consultations (or even their nature, so that there is more talk and less Valium). We might be interested in estimating what represents proper compensation for someone whose health has been adversely affected by the negligence or wilful behaviour of someone else (a doctor, a drug company, an employer, a fellow employee, a motorist, a criminal). We might be interested in whether it would be advantageous to redeploy resources from ‘curative’ to ‘preventative’ activities. We might be interested in whether too much or too little is being spent on the NHS. We might be interested in why poor people seem on average to have worse health than rich people. All these ‘interests’ require us to form some kind of view about the value of health (though forming such a view will not be suficicient to solve any of these problems for us). If we are not to get into a hopeless muddle we need a set of analytical distinctions which will enable us to see what kind of question we are asking and in what context so that we can be pointed in the right direction for an answer. The first of these distinctions turns on whether the focus of interest is some past event or some future decision. The second turns on whether the future decision is an individual decision or a group decision. The third turns on whether the group sees itself as acting on behalf of one individual, several individuals, or on behalf of some broader notion of ‘public’ or ‘social’ interest. And the fourth turns on whether the future decision is to be taken in a context in which the value of an outcome can be separated from the probability of its occurrence, or whether the value of the outcome and the uncertainty surrounding it are to be jointly assessed. In the next section some illustrative material will be presented designed to bring out the importance of these distinctions.
WHAT
IS THE
ANSWER?
First of all let us deal with the case where we are looking back at some past event where the value of health to some identifiable person is to be estimated, say in connection with a claim for damages before a court. In this case we need to distinguish further between death and injury, for although in both cases it is the effects upon living that are of interest, it will be clear that in many cases the damages for premature death will be much less than the damages for permanent injury simply because in the former case the principal loss bearer is beyond compensation by the courts. This creates a superficially curious, but quite understandable, distortion in the relative valuations of death and injury. As regards what is actu-
1026
ALAN WILLIAMS
ally to be valued, the courts are happier with the pecuniary (or ‘objective) losses than with the nonpecuniary (or ‘subjective’) ones. Here loss of income (so far as it would have benefitted those still alive) is admissible, as are any additional expenses directly attributable to the event in question. These must not however be too ‘remote’, and apart from a surviving victim, only nearest and dearest are recognized as having an ‘interest’ (employers, for instance, are not). Historically the courts have not been happy about awarding more than nominal sums for intangible things like loss of expectation of a happy life, loss of companionship, bereavement. or for pain grief and suffering, though things are changing. The courts are also worried about compensation being so generous that people might actually be better off than before the event in question, and also about doing anything that might reduce people’s incentive to minimise the event’s impact (the courts want people to cooperate with rehabilitative regimes to seek the best possible employment in their new situation). There are also complications about exemplary damages and contributory negligence. which need not concern us here, but all in all there are plenty of reasons why court awards are unlikely to be more than a crude answer even to the questions they themselves seek to answer, and are not very relevant to other questions one might be trying to answer. The rest of this paper concentrates on forwardlooking problems, i.e. where the damage to health is prospective. If such damage is unpredictable at individual level, and also largely unavoidable by individual actions, but affects a large number of people, then it may be insurable. So we might consider whether individuals’ insurance behaviour gives us any clue to the value they place on their own health. Unfortunately the answer is no, because such insurance is almost exclusively concerned with the pecuniurv consequences of loss of health, either to the ind(vidual (i.e. the first party) or to dependents (third parties). So only a small segment of the value of health will be covered, and only by that segment of the population which has recourse to such markets. For our third case let us look at the situation where the loss of health is still unpredictable at individual level. but the individual could do something (costly) to reduce the risks (for instance, better fire proofing in our homes. buying a safer car or replacing tyres more often). Unlike the previous cases, we are now considering a situation where the individual is not ‘paying’ to compensate others (or himself) for damage to his health, but to prevent such damage, so this approach is likely to be more helpful if we are trying to find out how people value their own health from al/ the viewpoints described earlier. If we could find enough cases where that kind of behaviour is observable and measurable, we could come up with a good estimation of the value people place on their own lives gioen the information and resources at their disposal (two very important conditions of which more anon). Suppose now that collective action is required to reduce some particular risk to an individuals health (e.g. reducing pollution. or improving roads) we might then ask: should the public authority simply value health the way the aflected individuals do? The
first problem is that we do not know who the affected individuals are, since they are virtually drawn in a lottery (it could happen to you, but we console ourselves with the thought that it is more likely to be somebody else!). If we know how the probabilities are distributed across the population, and know what the characteristics are of the different groups, we could find out how individual valuations vary with these characteristics, and calculate some weighted average accordingly. This would be a purely individualistic valuation, but with the probabilities (weights) that are used in the aggregation process being established by the ‘experts’. But suppose now that the policymakers decide that such individual valuations are not ‘legitimate’ for this purpose, because, say, they will tend to give more weight to the health of the rich than to that of the poor (assuming that willingness and ability to spend more to protect or improve one’s health is positively associated with income, as it seems to be). We now need to be careful how we unpick this little puzzle. First of all, we could say that the loss of output (as opposed to the loss of own income) occasioned by anyone’s loss of health should still be valued in the conventional way, since that loss affects us all, and will be mediated by the market. The same would apply to any resources we as a group contribute to caring for the person in question or for anyone else consequentially in need of care. The problem is then concentrated on the subjective value to the individual of the other affected items (essentially own health used in consumption) and if we want some sociul valuations of this we have to start from a different stand point. One possibility could simply be to postulate a value to be applied to everybody, i.e. have a ‘standard tariff for a year of lost life expectancy, for each kind of injury, for each duration, and so on. These ‘postulated’ values would be expressions of the social value attached to avoiding those outcomes (over and above the ‘economic’ costs already mentioned, which we assume are valued separately). This is essentially what happens to the ‘warm blooded’ or ‘subjective’ costs (pain grief and suffering) in investment appraisal for the road programme. Alternatively. a technique more sensitive to individual variations in the relative value attached to different kinds of illhealth, would be to postulate that a year of healthy life expectancy is of equal value to everybody (again, aside from the associated economic costs falling on others) but individuals might then be left to indicate how they value various states of imperfect health relative to the state of being healthy. Such relative scales (often implicitly) lie behind crude indicators such as ‘quality-adjusted-life-years’ as a summary measure of health, to which a money value may or may not be attached. In all of the foregoing cases it was left to individuals, or policymakers, to arrive at the valuation of health states by some acceptable method, and these valuations then enter a calculation in which the probabilities of people being in certain health states rather than others are estimated separately (and possibly by different people). For instance, the road engineers might estimate the changes in accident rates, and the value of avoiding accidents of different sorts would be calculated separately in the manner
The nature,
meaning
and measurement
described above. The problem with this approach is that the actual situation is that all road users experience a small reduction in the risk of accidents, and if they are risk averse they will be willing to pay more for this than the actuarially determined ‘stake’ in a ‘fair gamble’ implied above. So it has been argued that if you want such solutions to reflect the improved sense of wellbeing which the reduction in a known risk to health generates, then it would be better to discover more directly what people are, or would be, willing to pay for such reductions in risk. In the rather complex experiments that have been conducted along these lines, the valuations which have emerged are certainly much higher than the numbers currently in use in policymaking (and, in the transport field, have led to substantial upward adjustments to these figures). The complications are not yet quite over, however, because we still have the problem that willingness to pay is conditional upon the knowledge and resources that a person has. In an experimental setting these could be controlled for, by telling people what the probabilities are, and giving them all the same ‘budget’ to work with. But it has been argued that people do not really understand probabilities, so in a valuation process which jointly values the change in probability and the change in health, it will be hard to disentangle the latter if the former is not known with certainty. And as regards standardising incomes hypothetically, this worries some people because we can then no longer be sure that their responses will be ‘realistic’ (i.e. what would happen if they actually had that income). This is a more general problem, however, since it applies equally, if not more, to the problems people may have valuing health states they have never experienced. One final complication to be noted is that if people actually have ‘erroneous’ views about the probabilities, which make them more or less anxious than they should be about the likelihood of loss of health, should we not be estimating the value of the changes in their perceptions of risks rather than the actud reductions? CONCLUSIONS
At this point the bemused reader may be forgiven for wondering why economists go on struggling with such an apparently intractable. if not insoluble, problem as the valuation of health. The answer, paradoxically, is that we have to struggle with it simply because it is ‘solved’, everyday by everybody. Each day, as individuals, we make decisions which imply a valuation of our own health and of the health of others. Each day, some professional person or public official makes a decision which involves assigning resources in such a way that someone’s health will
of health
and illness
1027
consequently be improved and/or someone else’s made worse. If these decisions are to be made consistently, and if those making them are to be effectively accountable for such decisions, then we need to know what valuations they are using and how they arrived at them. So it is not so much that such valuations cannot be made, but that we have difficulty eliciting them, and when we do elicit them, there will remain the problem of which actual values should be used for which specific purposes, but, fortunately, that was not part of my remit. POSTSCRIPT
Since this has been a sketchy survey of a great deal of ground, designed as a non-technical exposition of rather complex material, it will doubtless cause affront to some of my colleagues who will be aware that I have simplified some of the issues in a rather brutal manner, and added insult to injury by offering my own sweeping generalisations and judgements of worth without anyone else’s views getting a look in. It is all true, and I could only redress the balance by citing an enormous bibliography of more learned works than this, so as to cover my tracks. I have resisted that temptation, because I believe that it would be inappropriate in an interdisciplinary context such as this in which my task is to try to convey to sociologists what it is that economists have done and are doing in this field. One area I reluctantly excluded, however, is that of health indicators and health status measurement. I did so because in most of that work the economists who have made contributions have on the whole worked in a multi-disciplinary context, and rather outside the mainstream of economic thinking, whereas the value-of-life literature is much more central to the corpus of welfare economics, and, in my judgement at least, conveys much more accurately and dramatically the distinctive flavour of economics as a discipline to be applied to valuation questions. For those wishing to dig more deeply into this territory may I recommend, first, Gavin Mooney’s book The Vuluution qfHuman Life (Macmillan, New York, 1977) which is a good general introduction, and comprehensive yet not very technical. Secondly, a stimulating and challenging set of essays by Victor R. Fuchs entitled Who ShuN Lice? (Basic Books, New York, 1974). For the hardier spirits, Michael JonesLee’s The Value of L(fe (Martin Robertson. London, 1976) and a recent set of essays reflecting current controversies, edited by Jones-Lee, and called The Vulue of L@ and SGfety (North-Holland, Amsterdam, 1982) are good exemplars of the economists at work on a problem, in the latter case with a (small) leavening of other disciplines.
in Great
0277-9536/U $3.00 + 0.00 Pergamon Press Ltd
Britam
THE NATURE, MEANING AND MEASUREMENT OF HEALTH AND ILLNESS: AN ECONOMIC VIEWPOINT ALAN WILLIAMS Institute of Social and Economic Research, University of York, Heslington,
York YOl
5DD, England
Abstract-Health can be seen as a capital asset, subject to depreciation due both to the passage of time (ageing) and to ‘wear and tear’. It is valued for the flow of pain-free time and energy it offers us, and we can increase the quantity and quality of this ‘flow’ by appropriate care and maintenance. which may take the form of adopting a healthy life style or in seeking health care. The valuation of health (as a capital stock) can then be seen as the valuation of the time profile of the stream of pain-free time and energy we expect to get out of it. This in turn depends on the uses to which this time and energy can be put. Its use in paid work is relatively easily valued but unpaid work (especially work in the home) continues to be a source of difficulty, both in principle and in practice, in the search for appropriate money values. Leisure time is still more difficult, though for those in paid work whose remuneration is sensitive to the workileisure balance, foregone net earnings at that margin may be used in the marginal value of leisure time. But the appropriateness of any valuation depends on the use to which it is to be put, and it is important to distinguish whether it is to reflect ‘individual’ or ‘group’ values, and whether it is for ex post compensation associated with some particular event, or for use in ex ante decision making where it is not known who the particular gainers and losers will be.
SCENE-SETTING The hallmark of Economists’ thinking about health and illness is the attempt to characterise these phenomena by using existing economic concepts, reinterpreted with varying degrees of sophistication, in the hope that we can thereby gain some distinctive insight into the nature and meaning of health and illness, and use our measurement techniques (even though originally devised for other purposes) to measure (and particularly to value) health and illness. This inevitably prompts the question: is health not very different from other good things that people want to have (and is illness not very different from other bad things that people want to avoid)? At a trivial level the answer is obviously ‘yes’, otherwise we would not have a distinctive terminology and set of concepts to describe health (and illness). But in a more fundamental sense is having good health any different from having a nice house or a reliable car or fine clothing or a good education? Before proceeding further we must make an important distinction between good health as an objective (or as a source of welfare in itself) and health cure as one means of attaining that objective. In any hierarchy of ends and means, health is likely to come ‘high up’ and, for most people, higher up than the house, car or clothing. A good education (if by that we mean ‘treatment’ at School and University) is more like health care (i.e. ‘treatment’ at Consulting Rooms or Hospitals), but where the corresponding phenomenon to ‘health’ is ‘knowledge’ or ‘skill’. Thus at a high level, health is seen as an element in personal welfare, and in some extreme formulations (due to WHO rather than to economists) is virtually identical to welfare (who can ask for more in this world than ‘complete physical and mental wellbeing’?) When discussing the nature of illness (taken simply to mean a departure from good health, and including here injury) economists’ attention has been concen1023
trated on two characteristics which are important in interpreting its significance in an economic context. The first is its unpredictability and potentially catastrophic consequences, and the second is, whether or not it renders its sufferers unfit to judge their own best interests. Discussion of the first characteristic leads inexorably to consideration of insurance analogies, and the extent to which provision for illness can be made efficiently and equitably through insurance mechanisms (public or private). Discussion of the second characteristic is more worrying for economists, since most of our analytical structures rest on the notion of ‘consumer sovereignty’. If, in the health field, there is legitimate doubt whether consumers are (or ever will be) both sufficiently knowledgeable about the nature of illness, and unemotional enough to be able to process in a rational manner such information as they have, then their best strategy might be to place themselves in the hands of some knowledgeable and trusted person who will act on their behalf. This is what parents encourage their children to do, and what the medical profession encourages us to do, and has led health economists to see the link between illness, the seeking of health care, and consequent health status, not as the working of simple demand and supply models, but as the operation of an agency relationship in which ‘demand’ is ‘mediated’ by (imperfect) agents who are actually also suppliers of (some kinds of) care, and whose notion of health (or welfare) may not be precisely that of the ‘demander’. In order not to trespass on the territory staked out for others, I will not pursue these general observations further here, since they will surely be central to the discussions on health professions and professionalism. Instead I will go off in a different direction, and concern myself with the question: what is it that makes health valuable, and what value do we attach to it? I do not claim to be able to give definitive answers to these searching questions, but merely to
1024
ALAN WILLIAMS
indicate how economists have tackled them. Before doing so I am going to assume that we know what health is, and for anyone who feels that it is altogether too dismissive of an area of controversy which is both crucial and unsettled, I will make clear my own personal view, which is that health is best regarded as the product of quantity of life (measured by life-expectancy) and quality of life (measured by physical mobility, freedom for pain and distress, capacity for self-care, and ability to engage in normal social interactions). I know that some of these phenomena are socially conditioned, and not easy to measure, and that they constitute an important area of investigation in themselves, epitomised in the literature on health status indices. But, on the whole, economists have been concerned with a different set of issues, and it is to these that I now wish to turn.
STOCKS
AND FLOWS
As any owner-occupier will testify, a house is both an asset and a liability. It is an asset in that it provides shelter, a pleasant environment in which to pursue leisure activities, sometimes (e.g. as I write this paper) a place to work, and, in another aspect entirely, a material investment to be converted into money at some later date. It is a liability in that it requires attention to the structure and fabric from time to time, the more so the older it gets and the more harshly it is treated by us or by the environment in which it stands. It also requires cleaning and heating and lighting if it is to provide the services we seek from it. Some of this maintenance and repair work we can do for ourselves (or some of us can!). We can reduce the need for it by being prudent in how we treat it (avoiding wild parties, or throwing things at other members of the household). But some of the work we have to get the professionals to do for us. We can also get rid of some of the financial risks (e.g. those associated with fire, storm damage, burglary) by insurance and this may make us marginally more relaxed (i.e. careless) about protecting ourselves from these hazards. We can also, of course, choose to save money and neglect the property, and do nothing till disaster strikes (e.g. the roof falls in), but that may prove to be very costly in the long run, and may well shorten the life of the property quite significantly. Mutatis mutandis, what has just been said about ourselves as owner-occupiers of our homes could equally be said about ourselves as owner-occupiers of our bodies. The terminology and balance of emphasis would of course be different, and I do not believe (as some people appear to do) that we can switch bodies as we can houses. A better analogy would be one in which we each inherit at birth a house in course of construction, and we are saddled with it the rest of our lives. But it is dangerous to push these picturesque comparisons too far, and the point I wish to make is simply that health can usefully be viewed a capital asset, which depreciates both with time (due to the aging process) and with use (due to life style and environmental stress) but which can be restored (to varying extents) by appropriate care and maintenance. What this ‘asset’ then provides us with is ‘healthy time’ (or, if it is not in very good shape,
‘unhealthy time’), and it is thisJEow which we need for all our activities (including sleeping) and which terminates at death. Some uses of time draw on our health capital more than others, however, (e.g. smoking is a very ‘health-capital-intensive’ activity, as are some hazardous occupations like labouring in the contruction industry). If these uses of time are seen as essential ‘inputs’ into everyday activities, and if healthy time is much more efficient than unhealthy time as such an input, then it is clear why we value health, and it also gives us a clue as to how we might set about valuing it. Let us consider first the simplest case, which is the activity of working for money. If it could be shown that better health is associated with higher earnings, then that clearly gives us a rather direct differential value for healthy over unhealthy time (and, indirectly, also indicates the maximum we would be prepared to pay for treatments which improve our health if there were no other benefits therefrom). Thus a great deal of the early work by economists was concerned with estimating the relationship between health and productivity, viewing the individual as a piece of human capital equipment, rather like a piece of physical capital equipment, to be valued by the value of the flow of services it produces. Even in this narrow context it was noted there is a difference between the value to the community of that individual’s contribution to output (which is the value so added to the final product) and its value to the individual (which is his or her net take home pay). It is also to be noted that this sort of calculation does not (or should not) purport to measure the entire value of health to the individual (or indeed to the community of which he or she is a part) since it relates to only one segment of (some) individuals’ activities. It has, nevertheless, played a dominant role in the value-of-health literature, probably because it deals with an intrinsically interesting and important phenomenon, and because there is fairly readily accessible data about it, waiting to be exploited, (which unfortunately is not the case with most of the other phenomena to be considered later in this paper). Work which is not done for money has always created great problems for economists. We know that housework ought to be in GDP, and it is only because of practical difficulties that it is not. Attempts to value it by what it would (or does) cost to get equivalent work done by hiring people to do it, often come up with large sums which people find it hard to believe, because they reckon that most people could not actually afford such sums, so it must be wrong. I think this is a mistaken objection, because the value of (unpaid) housework should be counted both as imputed income and as imputed expenditure by the household (just as the rental value of owner-occupied houses should be counted as both income and expenditure of a household) and if this were done most households would show up as much richer than they are if only money income is counted. A more telling objection is that ‘marketed’ housework is such a small (and probably unrepresentative) proportion of all housework that it is unlikely to provide a suitable basis for valuing the rest. If it is only the relatively rich and/or desperate who pay others to do a
The nature, meaning and measurement of health and illness significant amount of housework for them, and since there is also some social stigma attached to that kind of work on the supply side, it is possible that the market values are rather an unreliable guide as a basis for general valuation of such activities. On the other hand, the growth of homehelp services, meals-onwheels, etc. to assist ‘unhealthy’ people to cope with housework, may give us data on which we can draw with more confidence when estimating the value of health to the houseworker. To the extent that sick (unpaid) houseworkers draw on other (unpaid) houseworkers to help out, or simply leave tasks undone, this would then indicate that the costs of these ‘marketed’ services are to be regarded as an upper limit on the value attached to these unpaid (or unperformed) tasks. The third broad category of activities requiring time, where being healthy may be more valuable than being sick, is leisure activity (loosely termed ‘consumption’ in economics). Being unhealthy may actually make some consumption activities impossible, or at least highly dangerous, (e.g. marathon running, party going, smoking), but whether that is so or not, it will more pervasively diminish the pleasure to be found from them (e.g. eating and drinking, conversation, reading, etc.) (just think of toothache!). Since work is viewed in economics as instrumental, and (at the margin at least) as diminishing welfare, whereas consumption is the ultimate objective and the sole source of positive welfare, then the consumption benefits of health are seen as giving rise to its ‘intrinsic’ worth, and as closely related to the value of leisure time (which, at the margin, will also be systematically related to the value of working time especially for those able to adjust the balance between the two fairly closely in accordance with their preferences). So, by and large, economists have linked the ‘stock’ concept ‘health’ with its ‘flow’ concept ‘healthy time’, then sought to estimate the differential value of healthy over unhealthy time in its various uses (remunerated work, unremunerated work and leisure). The work uses are sometimes called the investment benefits of health (since ‘work’ is a process by which time is invested either in earning money with which to enhance consumption possibilities, or in producing such consumable goods and services directly), whilst the leisure uses are viewed as consumption benefits. Indirect and direct consumption might be a better terminology, reserving the term ‘investment’ in this context for activities designed to increase the health stock itself (e.g. sleeping, consulting a doctor, or jogging. assuming, as is undoubtedly true in my case, that none of these is intrinsically pleasurable). One final observation at this general level, before looking more closely at measurement problems, is that, like everyone else, economists have found it easier to focus on issues of life and death than on issues of quality of life. Just as medical practice has been greatly concerned with mortality, and hence with life-threatening situations to the relative detriment of non-life-threatening situations, so have economists tended to interpret the value of health as the value of life, and then equate life with life expectancy, ignoring quality of life. There are, of course, a few honourable exceptions, but modesty forbids.
WHAT
IS THE
1025 QUESTION?
There are many reasons for wanting to measure and value health, and generally it will be the case that the appropriate way of measuring and valuing health will depend on the purpose to which the measure is to be put, i.e. on what question we are asking. We might be interested in predicting individual or group behaviour, because we want to know, say, whether increasing prescription charges will adversely affect people’s health or merely reduce their consumption of useless or harmful drugs, or change volume of GP consultations (or even their nature, so that there is more talk and less Valium). We might be interested in estimating what represents proper compensation for someone whose health has been adversely affected by the negligence or wilful behaviour of someone else (a doctor, a drug company, an employer, a fellow employee, a motorist, a criminal). We might be interested in whether it would be advantageous to redeploy resources from ‘curative’ to ‘preventative’ activities. We might be interested in whether too much or too little is being spent on the NHS. We might be interested in why poor people seem on average to have worse health than rich people. All these ‘interests’ require us to form some kind of view about the value of health (though forming such a view will not be suficicient to solve any of these problems for us). If we are not to get into a hopeless muddle we need a set of analytical distinctions which will enable us to see what kind of question we are asking and in what context so that we can be pointed in the right direction for an answer. The first of these distinctions turns on whether the focus of interest is some past event or some future decision. The second turns on whether the future decision is an individual decision or a group decision. The third turns on whether the group sees itself as acting on behalf of one individual, several individuals, or on behalf of some broader notion of ‘public’ or ‘social’ interest. And the fourth turns on whether the future decision is to be taken in a context in which the value of an outcome can be separated from the probability of its occurrence, or whether the value of the outcome and the uncertainty surrounding it are to be jointly assessed. In the next section some illustrative material will be presented designed to bring out the importance of these distinctions.
WHAT
IS THE
ANSWER?
First of all let us deal with the case where we are looking back at some past event where the value of health to some identifiable person is to be estimated, say in connection with a claim for damages before a court. In this case we need to distinguish further between death and injury, for although in both cases it is the effects upon living that are of interest, it will be clear that in many cases the damages for premature death will be much less than the damages for permanent injury simply because in the former case the principal loss bearer is beyond compensation by the courts. This creates a superficially curious, but quite understandable, distortion in the relative valuations of death and injury. As regards what is actu-
1026
ALAN WILLIAMS
ally to be valued, the courts are happier with the pecuniary (or ‘objective) losses than with the nonpecuniary (or ‘subjective’) ones. Here loss of income (so far as it would have benefitted those still alive) is admissible, as are any additional expenses directly attributable to the event in question. These must not however be too ‘remote’, and apart from a surviving victim, only nearest and dearest are recognized as having an ‘interest’ (employers, for instance, are not). Historically the courts have not been happy about awarding more than nominal sums for intangible things like loss of expectation of a happy life, loss of companionship, bereavement. or for pain grief and suffering, though things are changing. The courts are also worried about compensation being so generous that people might actually be better off than before the event in question, and also about doing anything that might reduce people’s incentive to minimise the event’s impact (the courts want people to cooperate with rehabilitative regimes to seek the best possible employment in their new situation). There are also complications about exemplary damages and contributory negligence. which need not concern us here, but all in all there are plenty of reasons why court awards are unlikely to be more than a crude answer even to the questions they themselves seek to answer, and are not very relevant to other questions one might be trying to answer. The rest of this paper concentrates on forwardlooking problems, i.e. where the damage to health is prospective. If such damage is unpredictable at individual level, and also largely unavoidable by individual actions, but affects a large number of people, then it may be insurable. So we might consider whether individuals’ insurance behaviour gives us any clue to the value they place on their own health. Unfortunately the answer is no, because such insurance is almost exclusively concerned with the pecuniurv consequences of loss of health, either to the ind(vidual (i.e. the first party) or to dependents (third parties). So only a small segment of the value of health will be covered, and only by that segment of the population which has recourse to such markets. For our third case let us look at the situation where the loss of health is still unpredictable at individual level. but the individual could do something (costly) to reduce the risks (for instance, better fire proofing in our homes. buying a safer car or replacing tyres more often). Unlike the previous cases, we are now considering a situation where the individual is not ‘paying’ to compensate others (or himself) for damage to his health, but to prevent such damage, so this approach is likely to be more helpful if we are trying to find out how people value their own health from al/ the viewpoints described earlier. If we could find enough cases where that kind of behaviour is observable and measurable, we could come up with a good estimation of the value people place on their own lives gioen the information and resources at their disposal (two very important conditions of which more anon). Suppose now that collective action is required to reduce some particular risk to an individuals health (e.g. reducing pollution. or improving roads) we might then ask: should the public authority simply value health the way the aflected individuals do? The
first problem is that we do not know who the affected individuals are, since they are virtually drawn in a lottery (it could happen to you, but we console ourselves with the thought that it is more likely to be somebody else!). If we know how the probabilities are distributed across the population, and know what the characteristics are of the different groups, we could find out how individual valuations vary with these characteristics, and calculate some weighted average accordingly. This would be a purely individualistic valuation, but with the probabilities (weights) that are used in the aggregation process being established by the ‘experts’. But suppose now that the policymakers decide that such individual valuations are not ‘legitimate’ for this purpose, because, say, they will tend to give more weight to the health of the rich than to that of the poor (assuming that willingness and ability to spend more to protect or improve one’s health is positively associated with income, as it seems to be). We now need to be careful how we unpick this little puzzle. First of all, we could say that the loss of output (as opposed to the loss of own income) occasioned by anyone’s loss of health should still be valued in the conventional way, since that loss affects us all, and will be mediated by the market. The same would apply to any resources we as a group contribute to caring for the person in question or for anyone else consequentially in need of care. The problem is then concentrated on the subjective value to the individual of the other affected items (essentially own health used in consumption) and if we want some sociul valuations of this we have to start from a different stand point. One possibility could simply be to postulate a value to be applied to everybody, i.e. have a ‘standard tariff for a year of lost life expectancy, for each kind of injury, for each duration, and so on. These ‘postulated’ values would be expressions of the social value attached to avoiding those outcomes (over and above the ‘economic’ costs already mentioned, which we assume are valued separately). This is essentially what happens to the ‘warm blooded’ or ‘subjective’ costs (pain grief and suffering) in investment appraisal for the road programme. Alternatively. a technique more sensitive to individual variations in the relative value attached to different kinds of illhealth, would be to postulate that a year of healthy life expectancy is of equal value to everybody (again, aside from the associated economic costs falling on others) but individuals might then be left to indicate how they value various states of imperfect health relative to the state of being healthy. Such relative scales (often implicitly) lie behind crude indicators such as ‘quality-adjusted-life-years’ as a summary measure of health, to which a money value may or may not be attached. In all of the foregoing cases it was left to individuals, or policymakers, to arrive at the valuation of health states by some acceptable method, and these valuations then enter a calculation in which the probabilities of people being in certain health states rather than others are estimated separately (and possibly by different people). For instance, the road engineers might estimate the changes in accident rates, and the value of avoiding accidents of different sorts would be calculated separately in the manner
The nature,
meaning
and measurement
described above. The problem with this approach is that the actual situation is that all road users experience a small reduction in the risk of accidents, and if they are risk averse they will be willing to pay more for this than the actuarially determined ‘stake’ in a ‘fair gamble’ implied above. So it has been argued that if you want such solutions to reflect the improved sense of wellbeing which the reduction in a known risk to health generates, then it would be better to discover more directly what people are, or would be, willing to pay for such reductions in risk. In the rather complex experiments that have been conducted along these lines, the valuations which have emerged are certainly much higher than the numbers currently in use in policymaking (and, in the transport field, have led to substantial upward adjustments to these figures). The complications are not yet quite over, however, because we still have the problem that willingness to pay is conditional upon the knowledge and resources that a person has. In an experimental setting these could be controlled for, by telling people what the probabilities are, and giving them all the same ‘budget’ to work with. But it has been argued that people do not really understand probabilities, so in a valuation process which jointly values the change in probability and the change in health, it will be hard to disentangle the latter if the former is not known with certainty. And as regards standardising incomes hypothetically, this worries some people because we can then no longer be sure that their responses will be ‘realistic’ (i.e. what would happen if they actually had that income). This is a more general problem, however, since it applies equally, if not more, to the problems people may have valuing health states they have never experienced. One final complication to be noted is that if people actually have ‘erroneous’ views about the probabilities, which make them more or less anxious than they should be about the likelihood of loss of health, should we not be estimating the value of the changes in their perceptions of risks rather than the actud reductions? CONCLUSIONS
At this point the bemused reader may be forgiven for wondering why economists go on struggling with such an apparently intractable. if not insoluble, problem as the valuation of health. The answer, paradoxically, is that we have to struggle with it simply because it is ‘solved’, everyday by everybody. Each day, as individuals, we make decisions which imply a valuation of our own health and of the health of others. Each day, some professional person or public official makes a decision which involves assigning resources in such a way that someone’s health will
of health
and illness
1027
consequently be improved and/or someone else’s made worse. If these decisions are to be made consistently, and if those making them are to be effectively accountable for such decisions, then we need to know what valuations they are using and how they arrived at them. So it is not so much that such valuations cannot be made, but that we have difficulty eliciting them, and when we do elicit them, there will remain the problem of which actual values should be used for which specific purposes, but, fortunately, that was not part of my remit. POSTSCRIPT
Since this has been a sketchy survey of a great deal of ground, designed as a non-technical exposition of rather complex material, it will doubtless cause affront to some of my colleagues who will be aware that I have simplified some of the issues in a rather brutal manner, and added insult to injury by offering my own sweeping generalisations and judgements of worth without anyone else’s views getting a look in. It is all true, and I could only redress the balance by citing an enormous bibliography of more learned works than this, so as to cover my tracks. I have resisted that temptation, because I believe that it would be inappropriate in an interdisciplinary context such as this in which my task is to try to convey to sociologists what it is that economists have done and are doing in this field. One area I reluctantly excluded, however, is that of health indicators and health status measurement. I did so because in most of that work the economists who have made contributions have on the whole worked in a multi-disciplinary context, and rather outside the mainstream of economic thinking, whereas the value-of-life literature is much more central to the corpus of welfare economics, and, in my judgement at least, conveys much more accurately and dramatically the distinctive flavour of economics as a discipline to be applied to valuation questions. For those wishing to dig more deeply into this territory may I recommend, first, Gavin Mooney’s book The Vuluution qfHuman Life (Macmillan, New York, 1977) which is a good general introduction, and comprehensive yet not very technical. Secondly, a stimulating and challenging set of essays by Victor R. Fuchs entitled Who ShuN Lice? (Basic Books, New York, 1974). For the hardier spirits, Michael JonesLee’s The Value of L(fe (Martin Robertson. London, 1976) and a recent set of essays reflecting current controversies, edited by Jones-Lee, and called The Vulue of L@ and SGfety (North-Holland, Amsterdam, 1982) are good exemplars of the economists at work on a problem, in the latter case with a (small) leavening of other disciplines.