The Journal of Political Philosophy: Volume 15, Number 1, 2007, pp. 46–66

What’s Wrong with Health Inequalities?* Daniel M. Hausman Philosophy, University of Wisconsin-Madison

I

T seems obvious that facts about health are of great importance to egalitarians concerned with social policy.1 Health is both an important component and cause of well-being. So if inequalities in well-being or resources are of interest from the perspective of justice – as seems evident – then health inequalities are of interest. What information about health inequalities should egalitarian policy makers seek? Until recently, egalitarians focused on health differences across different social classes, ethnic groupings, or geographical regions. Following the usage of staff at the World Health Organization, I shall call these “social group health differences.” Recently, as part of their effort to compare health and health systems, staff members at the World Health Organization sought to compare the extent of inequality in health across individuals in different countries and regions.2 They proposed to study the extent of inequality in health across individuals, just as economists study inequalities in income or wealth across individuals or families. One important factor that made this proposal attractive to the WHO staff is that relevant social groupings vary from society to society, which makes it difficult to compare health differences among an array of groups in one country to health differences among other groups in another. Measures of inequality across individuals, in contrast, permit international comparisons. In addition, members of the WHO staff argued that measuring individual health inequalities is in several ways superior to focusing exclusively on social group health differences. I have discussed these arguments elsewhere3 and will not comment on them here. This article instead argues that information about inequalities in health across individuals as opposed to information about social group health differences is *I am indebted to audiences at the University of Wisconsin, Stanford University and especially the Harvard School of Public Health for searching comments and questions. Special thanks are also due to Yukiko Asada and Dan Wikler for challenging me to address questions concerning health inequalities and particularly to Geert Demuijnck who offered extremely valuable criticisms of an earlier draft. Asada’s article “Is Health Inequality across Individuals of Moral Concern?” comments on an earlier unpublished version of this article. I am also indebted to Bob Goodin and to anonymous referees for their comments and criticisms. 1 Daniels, Kennedy and Kawachi 2000; Marchand, Wikler and Landesman 1998. 2 Murray et al. 1999, 2000; Gakidou et al. 2000. 3 Hausman, Asada and Hedemann 2002. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

47

generally of little use to egalitarians. The context I shall be concerned with is social policy, and I shall assume that the state’s responsibility for inequalities in outcomes is limited by a respect for individual choice and responsibility. Since health bears so heavily on opportunity, health inequalities are of particular importance in the context of social policy. Nevertheless, this paper argues that data concerning individual health inequalities do not provide egalitarian policy makers with the information they need. These data typically do not license conclusions about social justice. Health inequalities across individuals are often not prima facie unjust. Three remarks about this thesis may help avoid misunderstanding. First, this paper is concerned about the justice or injustice of inequalities in health across individuals, not about whether inequalities in health care or health-related resources are unjust. Second, this article is concerned with the justice or injustice of health inequalities, not with whether egalitarians should be concerned about health or its causes or consequences. An egalitarian might reasonably lavish attention on matters relating to health, because of its fateful consequences for life prospects or well-being, and an egalitarian might enthusiastically support specific health policies, such as vaccinations or prenatal care, because of their enormous impact. Third, this article is concerned with what inferences concerning injustices can be drawn from the data concerning health inequalities that the WHO hopes to provide, not with whether health inequalities involve injustices. There are many different kinds of inequalities in health, and many different ways in which inequalities in health can be conceptualized and measured. It is obvious that some health inequalities involve grave injustice and that some do not. For example, it is clearly unjust that African American infants die at three times the rate of American infants of European ancestry, but the fact that members of the military are, at least in peace time, healthier than stockbrokers or lawyers reveals no injustice. It is questionable whether there is any single fact which reports how unequal health is in some population.4 Since it is unclear what the phrase “the extent of inequality in health” refers to, it is doubtful whether there is anything to be said about the implications of the extent of health inequality. This article is concerned with the implications of particular information about health inequalities rather than the implications of health inequalities themselves. Given that some health inequalities involve injustices and that some do not, what use can egalitarians make of the sort of information the WHO seeks to provide concerning health inequalities? The most plausible case for the significance of data concerning individual health inequalities relies on an analogy between that data and data on individual 4 Temkin (1993) provides an extended argument for the multiple ambiguity of claims about inequalities.

48

DANIEL M. HAUSMAN

inequalities in wealth or income. Why not measure individual health inequalities as a proxy for overall inequalities in just the same way that investigators have measured individual wealth and income inequalities? Section V tackles this question, but the comparison between measures of individual health inequalities on the one hand and measures of individual inequalities in income and wealth will be relevant at several points in the argument. I. WHAT IS AT ISSUE? Several authors have argued that inequalities in health are themselves unjust regardless of whether they reflect overall inequalities in moral status or life prospects. Robert Veatch maintains that “justice requires that persons be given an opportunity to have equal health status insofar as possible.”5 Sudhir Anand maintains that inequality in health is itself a bad thing, indeed less tolerable than inequalities in income or wealth.6 Anthony Culyer argues that “An equitable health care policy should seek to reduce the inequality in health . . . at every stage of the life-cycle.”7 Because Martha Nussbaum distinguishes “bodily health” as one of the ten central human functional capabilities, one might also attribute a “specific egalitarian” view to her;8 but she does not call for equalizing the central human functional capabilities.9 Egalitarian theorists disagree about which inequalities are of moral importance, and the view defended by Veatch, Anand, Culyer and implicitly by the WHO staff is not uncontroversial. One benefit of asking whether inequalities in health are unjust is that doing so raises general questions concerning the relations between inequalities and injustices in a new and, I believe, enlightening context. Another benefit lies in the perspective it provides on the relevance of measurements of inequalities in income and wealth. In this article, I shall argue that information about inequalities in health across individuals permits inferences concerning justice in two circumstances. Either knowledge of the health inequalities all by itself justifies conclusions concerning overall inequalities of moral status or life prospects, or one already has enough information concerning the distribution of other goods that one can tell whether the health inequality aggravates or ameliorates the overall inequality.10 Information about isolated inequalities – that is, inequalities in individual goods such as health, education, nutrition or income – is of use to egalitarians only insofar as it permits inferences concerning one’s status as a person or one’s overall life prospects. Because of the fungibility of wealth (at least from the 5

Veatch 1981, p. 83. Anand 2002, pp. 485–6. 7 Culyer 2001, p. 281. 8 Tobin 1970. 9 Nussbaum 2000, p. 78. 10 Similar positions have been defended by Peter (2001), Woodward and Kawachi (2000) and, with some important caveats, Asada (2006). 6

WHAT’S WRONG WITH HEALTH INEQUALITIES?

49

perspective of social policy) – that is, the range of things for which wealth can be exchanged – wealth inequalities are typically not in this way “isolated.” To forestall misunderstanding, let me emphasize that the pressing practical case for a redistribution of health-related resources toward the poor and oppressed does not depend on whether information concerning health inequalities permits inferences concerning injustice. Even if one had no concern for justice at all and were only interested in increasing well-being or alleviating suffering, there would be an overwhelming case for redistribution. In some parts of the world, such as sub-Saharan Africa, nothing has a higher moral priority than addressing health needs. But even though improving life prospects and mitigating health inequalities often in fact go hand in hand, under other circumstances there will be tradeoffs. So it is important to ask which inequalities are of moral concern and why. In addition, as I shall explain in section 2, the thesis of this paper, that data concerning health inequalities across individuals typically do not tell egalitarian policy makers what they need to know, depends crucially on the content of the data. The WHO has developed a generic health measure, in terms of disability adjusted life years (DALYs), while most other generic health measures quantify health in terms of quality adjusted life years (QALYs). The differences between DALYs and QALYs are not germane to the questions with which this paper is concerned, and for the sake of brevity I shall refer to generic health measures as measures of QALYs.11 As I shall argue below, taken by itself, the fact that one person has enjoyed more QALYs than another typically tells an egalitarian policy maker very little. Only in extreme cases does it license any conclusions concerning justice, even prima facie. Other information about inequalities, in contrast, may justify conclusions about inequalities in life prospects. Consider, for example, information about infant mortality. If differences in infant mortality could be addressed by public health measures, nutrition or maternal care, then they are prima facie glaring injustices. If, as I have so far only asserted, moral concern with inequalities in health derives from moral concern of policy makers with inequalities in moral status or life prospects, then the WHO’s focus on inequalities in general health scores across individuals is misplaced. As I shall argue below, merely knowing the distribution of QALYs across individuals (as opposed to knowing the distribution of income or wealth) tells one little about inequalities in moral status or life prospects. On the other hand, when knowledge is available – as it often is – about differences in income, wealth, social status, opportunities and so forth across 11 In measuring DALYs, full health (a state of no disabilities) is assigned a measure of zero and death is assigned a measure of one. In most other generic health measures, the quality adjustment for full health is one and the number assigned to death is zero. Health is then measured in QALYs by multiplying the weights assigned to health states by the length of time individuals occupy health states. The common feature in all these systems of health measure is the effort to combine mortality and morbidity into a single measure.

50

DANIEL M. HAUSMAN

social groups, then information about differences in QALYs across such groups bears on whether data on other group differences exaggerate or understate the overall differences in life prospects. Information about social group health differences will thus often be relevant to conclusions about justice, not because group differences matter and individual differences don’t, but because information about differences in QALYs between well-studied social groups will often license conclusions about the fundamental inequalities that egalitarians care about. Because there are in many societies well-known differences in incomes, status and opportunities across geographic regions and across ethnic and religious boundaries, knowing in addition the extent of the health differences can tell one a great deal about the overall inequalities in a society. For example, the differences in quality adjusted life expectancy between Americans of African and European ancestry highlight how serious are the overall inequalities between European-Americans and African-Americans. Data concerning the variance across individuals of health outcomes or expectations may suggest questions and give rise to inquiries whose findings could be of great importance to egalitarians. But simply knowing the distribution of QALYs licenses very few moral conclusions. A state of affairs in which those who are otherwise worse off are healthier than those who are otherwise more fortunate is more just rather than less just than a state of affairs which is exactly the same except that health is equally distributed.12 Apart from identifying those who have extremely low QALY scores, information concerning inequalities in QALYs in a society does not permit even prima facie judgments concerning whether that society is just or unjust. II. INCOMPENSABLE HEALTH INEQUALITIES My central claim concerning the limited relevance of information about individual health inequalities carries an important qualification which requires some explanation. Actions that lessen health inequalities are of two kinds: remediation and compensation. Let us say that a health inequality is “remediable” if it is possible to eliminate the inequality in health itself and that a health inequality is “compensable” if it is possible through some means other 12 One might object that analogous remarks apply to inequalities in wealth or income. For the purposes of argument, suppose that well-being consisted of a mental state of happiness. Isn’t a society in which the rich are less happy than the poor more just rather than less just than a society that is exactly the same except that wealth is equally distributed? I maintain that the answer is “no” because I am assuming that the state’s responsibility for individual happiness is limited (even if happiness could be measured without an unacceptable incursion upon the privacy of citizens). Obviously, many of the poor are happy and many of the rich are unhappy, but if the social determinants of life prospects and moral status have been equalized, differences in happiness or other outcome measures are not a concern of liberal egalitarian social policy. If one held a more intrusive view of social policy as aiming directly to equalize well-being, there might be a case for concentrating on health inequalities rather than income or wealth inequalities. Unlike income and wealth, which are only weakly correlated with well-being, poor health is strongly linked to unhappiness. Part of the linkage is analytic: as explained in section 5, the measurement of health is arguably a measurement of well-being.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

51

than mitigating the health inequality to eliminate the overall inequality to which the health inequality contributes. The contrast between compensation and remediation is not perfectly sharp. For example, is providing someone who has lost a leg in an accident with a prosthetic limb a matter of compensation or remediation? But the contrast is clear enough for my purposes. Health inequalities are often only partially rather than fully remediable or compensable. Some health states such as sensory deficits are arguably (fully) compensable, whether or not remediable. Others such as a diabetic coma are remediable but incompensable. Some health inequalities may be both compensable and remediable, while others may be neither compensable nor remediable. The notion of “compensating” for a health deficiency requires further clarification. First, I shall regard a health state as compensable if and only if it is “positively” compensable – that is, if and only if the overall inequality resulting solely from the health inequality can be eliminated by providing those in poor health with other advantages. Providing these advantages may slightly diminish the prospects of those who are healthy, but equalizing via “leveling down” does not count as compensation. The fact that an inequality between those with cognitive deficiencies and those without any deficiencies might be eliminated by limiting the education of those who are cognitively unimpaired would not imply that these cognitive deficiencies were compensable.13 Second, in speaking of compensation (and remediation), what I have in mind is practical. A health state that was compensable only by means of the provision of a private yacht staffed with cooks, entertainers and care givers would count as incompensable. Compensation in such a case would be so costly that it would amount to leveling down. Third, since the context is social policy, the standard of compensability is social, not individual. What is at issue is whether a health state would be compensable for a representative or typical individual in a population, not whether individuals, with all their idiosyncrasies, can be compensated. Fourth, what constitutes compensation depends on what one takes the egalitarian “currency” to be. Compensating for an inequality in moral or political status caused by health inequalities is not the same as compensating for an inequality in well-being. From the perspective of an individual, whether a health deficiency is compensable will depend on how it bears on what matters to that person, while from a social perspective, it is not obvious whether or how to take account of the range of objectives individuals may have. Finally, the possibility of compensation depends on some sort of benchmark. In an affluent and egalitarian society a health deficit such as persistent back pain might not be compensable – 13 In adopting this asymmetric view of compensation, I am assuming that egalitarians will not favor leveling down policies that lessen health inequalities by means of worsening the circumstances of those who are relatively advantaged. If one cared only about equality, leveling down would not be objectionable, but egalitarians do not care only about equality. Similar comments apply to remediation. Lobotomizing those without cognitive deficiencies does not count as an admissible way to remedy the inequality between those with and without cognitive handicaps.

52

DANIEL M. HAUSMAN

that is, no matter what other accommodations are made (short of leveling down), people with persistent back pain will be worse off than most people are. In an impoverished society or one with great overall inequalities, on the other hand, it may be relatively easy to compensate someone with persistent back pain and bring them to a level of overall advantage that compares with the pre-redistribution position of the median. When health inequalities are incompensable, then all by themselves they imply overall inequalities in status or life prospects. There is no way to compensate someone who dies at age one for his poor health. The only way to address the inequality is to address the cause of death. Similarly, someone with severe cognitive limitations cannot have the same set of effective rights and liberties or the same moral status. If the cognitive limitations are remediable, then there is a serious injustice here. Egalitarians should regard health inequalities that are both remediable and incompensable as prima facie injustices. It may help to make the general point more abstractly. Let X be a measure of what egalitarians want to equalize. I have spoken vaguely of “status” or “life prospects,” and later in section 4, I will clarify to some extent what I mean. But there is no need for precision at this stage of the argument. Depending on one’s favorite version of egalitarianism, X could be some measure of moral status, welfare, resources, capabilities, functionings, access to advantage or opportunity for welfare. Whatever X is, it will depend on specific categories of goods. To illustrate the conceptual point, let us oversimplify and suppose that X is itself a scalar, whose value depends on six arguments: health, education, friendship, liberty, social status and wealth. In other words X = g(h,e,f,l,s,w). The relationship between X and these inputs is intricate. The contribution that any specific input makes to X varies depending on the values of the other inputs. For any two individuals, George and Ann, an egalitarian is interested in how much of X George has compared to how much of X Ann has. Since it is usually the case that better health increases X, an egalitarian will prefer a state of affairs in which George’s health is just as good as Ann’s health over a state of affairs in which George’s health is worse than Ann’s, if both have equal quantities of the other inputs into X. If the other inputs vary and shortfalls with respect to some inputs can be compensated for by advantages with respect to other inputs, then an egalitarian can reach no conclusions concerning overall inequality or justice by comparing how much of one input George has to how much of that input Ann has.14 14 Exactly the same thing might be said about wealth. But there is a difference. Wealth, unlike health, will typically buy education, (effective) liberty and social status. Consequently it is unlikely that abundant health and friendship will compensate for poverty, poor education, low social status and less effective liberty. In most cases inequalities in wealth will indicate inequalities in X. Moreover, there are fewer social interventions that enable the state to provide more of the other arguments of the X function to compensate those who are poor.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

53

If, on the other hand, one has evidence that a difference with respect to one input is not compensable (like the difference between a healthy person and someone with very serious cognitive limitations), then that difference implies an inequality in X. In that case one can infer from the health inequality that there is an overall inequality, and if the health inequality is remediable, it is a prima facie injustice, not because egalitarians care about health inequalities themselves, but because in such a case the health inequality determines the inequality with respect to what it is that egalitarians do care about.15 If health inequalities were typically incompensable, then information about health inequalities would imply the existence of overall inequalities, and remediable health inequalities would be prima facie injustices. But, as I shall argue below in section 4, health inequalities are often compensable, and differences in QALYs, unless very large, justify no inferences concerning whether health inequalities are compensable and whether they imply overall inequalities. III. CAN HEALTH DISTRIBUTIONS BE JUST OR UNJUST? One might question the relevance of information about health differences to questions of justice on the grounds that health is not the sort of thing whose distribution could be just or unjust. Health depends on nature, luck and individual choice; the possibilities for changing the distribution of health by social intervention are limited. The principle behind the suggestion that irremediable and incompensable health inequalities are not unjust is that states of affairs that do not derive from social arrangements or the actions of people and are not subject to control or remediation by human action are neither just nor unjust. I find this principle plausible.16 When someone who is overweight complains, “It isn’t fair. I only have to look at a piece of cake and I put on pounds, while others eat and eat and stay slim,” we should not take this use of the word, “fair” seriously. The distribution of metabolisms among people is not the sort of thing that is fair or unfair. If an incompensable health inequality such as that between a healthy person who lives to 85 and a child who dies in infancy cannot be mitigated – that is, remedied or compensated – by any human actions, then the inequality is not an injustice. 15 Some of the sentiments favoring “specific egalitarianism” that Tobin (1970) discusses arguably rest on a recognition of the incompensability of certain inequalities. Similarly, Nussbaum’s insistence that everyone should be guaranteed at least some minimum level of each of the central human functional capacities might be defended on the ground that below some minimum level a deficiency with respect to any one of these capacities is not compensable. 16 Not everyone agrees. Consider Larry Temkin’s remarks which are quoted below. When an injury is the result of an intentional wrong, then obviously an injustice has been done. The resulting pattern of advantages and disadvantages is, however, no more or less unjust than if the pattern resulted from a blameless act of nature, and if the injury is incompensable and irremediable, then there is nothing unjust about the continued existence of its consequences. Accordingly the text focuses on whether health states are compensable and remediable, not on what caused them.

54

DANIEL M. HAUSMAN

However, the principle that inequalities can be unjust only if they are subject to mitigation by human action does not imply that health inequalities cannot be unjust, because health inequalities are often subject to mitigation by human actions. Social factors and human actions strongly influence people’s health. In addition to the growing efficacy of health care, consider factors such as nourishment, education, clean water, living space, occupational hazards, public health measures and wars, which account for a great deal of the variance in health. Recent work suggests that the extent of inequalities in a society coupled with a person’s position in the socio-economic hierarchy also have major influences on health.17 When remediation is not possible, overall inequalities can often be mitigated by means of compensation, and there is a prima facie egalitarian case for compensation. For example, although it may be impossible for a victim of a car accident to walk again, wheelchair ramps and elevators will permit her to participate more fully in a wider range of occupations and activities. Even if nothing can be done to extend the life of someone suffering from a disease that brings about an early death, more resources may help make this short life more rewarding. Information about health differences is relevant to egalitarians only when it conveys information about X. Since information concerning incompensable health differences across individuals may determine a difference in X, more or less regardless of the values of the other factors, such information is important to egalitarians. In circumstances in which people have very short lives or lives in extremely poor health, QALY differences will be huge, and it will be safe to conclude that the health differences are incompensable. Other goods cannot compensate for a quality adjusted life expectancy of five or six years. So very large health inequalities imply inequalities in X, and those health inequalities, if remediable, are prima facie injustices. Less extreme inequalities in QALYs across individuals, on the other hand, say almost nothing about whether health differences are compensable or not. Coupled with knowledge concerning other inequalities across groups, information about the differences in average QALY scores across social groups will, however, convey information to policy makers about the extent of overall inequalities. Information about compensable health inequalities among individuals (or among groups about which nothing else is known), like moderate differences in QALY scores, is by itself of no use to an egalitarian. IV. SPECIFIC EGALITARIANISM Those who study inequalities in QALYs across individuals might respond in four ways to this argument that the information they provide is not of use to egalitarian policy makers. This section discusses the first three of these responses, 17

Wilkinson 1996; Marmot 2004.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

55

while the next section discusses the fourth response. First, one might argue that because inequalities in health are sometimes remediable, while at the same time generally incompensable and independent of other inequalities, they are best addressed separately. However, the premises in this argument for the relevance of information about inequalities in QALYs are indefensible. Though there are some health states that are remediable and incompensable, there is ample and growing evidence that there are intricate causal connections between inequalities in health and other important inequalities, and it is not true that health inequalities typically can or should be addressed separately from other inequalities. Ill-health often has social causes and its consequences can frequently be mitigated by social actions. Moreover, policies that mitigate inequalities in health often do so by providing non-health resources or goods. Constructing wheelchair ramps and elevators helps to alleviate inequalities in opportunities and well-being, not inequalities in health. What makes plausible the view that health inequalities are generally incompensable is the understandable tendency to focus on very serious and life-threatening illness. But there is no serious empirical case to be made for the thesis that health differences are generally incompensable. Moderate differences in QALYs are often compensable, and accordingly they do not imply differences in life prospects or status and so license no conclusions concerning overall inequality. Though there is no controversy about this point among health economists – which is unsurprising, given the tendency economists have to regard everything as having a price – perhaps some further comments are in order for non-economists. Health professionals have devised a variety of systems for classifying health states. Consider the Health Utilities Index (Mark 3), which is one of the two most detailed of these health classification systems.18 This index identifies eight “dimensions” of health: vision, hearing, speech, ambulation, dexterity, emotion, cognition and pain; and along each dimension it specifies five or six levels of functioning, for a total of 972,000 distinguishable health states. The fact that the self-reported quality of life of those who are deaf or blind or unable to walk is not much lower than those without these disabilities suggests that deficiencies along the dimensions of sight, hearing or ambulation are compensable.19 Though I do not have data on self-reported quality of life among those who are unable to speak, it is hard to believe that speech deficits are less compensable than the loss of sight or hearing. Severe limitations on dexterity might be incompensable, 18 The other is the “SF-36,” which also has eight dimensions. The SF-36 would be more complicated to discuss, because it does not have a simple list of levels along its dimensions. 19 If egalitarianism aimed to equalize well-being or quality of life, it might be that no compensation was needed. But this is not a plausible version of egalitarianism and in any event conflicts with the liberal view of the limited concerns of egalitarian social policy. The inability to see or to hear or to walk also shapes the opportunities that are available to people. Whether they leave uncompensated individuals with life prospects that are (on some plausible metric) “equivalent” is debatable. I suspect that compensation is needed to equalize life prospects.

56

DANIEL M. HAUSMAN

perhaps at the fifth of the six levels of severity.20 That leaves the three dimensions of emotion, cognition and pain. Along each of those dimensions, the boundary between what is compensable and what is not probably lies between the third and four levels.21 Deficiencies along different dimensions that are compensable by themselves are not necessarily compensable in combination. Even so, it should be obvious from this discussion that diminished health states are not typically incompensable. The individual health inequalities that the WHO would have us attend to are differences in the scores of health states multiplied by the amount of time individuals occupy those health states. Unless the QALY difference is very large, knowing people’s actual or expected QALYs implies nothing about whether their health states are compensable. So a moderate QALY difference between individuals does not permit an egalitarian to reach any conclusions about distributive justice. Second, those who argue that health inequalities are themselves prima facie injustices might point out that information about health inequalities may raise important questions and contribute to research whose results are relevant to egalitarians. I have already granted this claim: all sorts of research might yield information that turns out to be relevant to egalitarians. What is at issue is the relevance of the findings to egalitarian social policy, not the significance of their consequences for further inquiry. Third, and much more importantly, those interested in health inequalities across individuals might challenge the assertion that egalitarians should focus on overall inequalities, such as inequalities in moral status or life prospects. Shouldn’t egalitarians be concerned about inequalities in significant goods, regardless of whether inequality in the possession of that good correlates with other inequalities or is linked to inequalities in status or life prospects? There are two versions of this specific egalitarian objection. The first, which I will call “health exceptionalism,” holds that health inequalities, unlike other specific inequalities, are themselves of moral concern. Health exceptionalists are not making the implausible claims that health inequalities are always unjust, that only health inequalities are unjust, or that health inequalities are always more important than other inequalities. The view is instead that inequalities in health are prima facie unjust merely in virtue of being inequalities with respect to health. In that case, information about health inequalities would obviously be of use to egalitarians. On such a view, a society in which (for example) resources were

20 The fifth level is “Limitation in the use of hands or fingers, requires the help of another person for most tasks (not independent even with the use of special tools).” 21 The third level of the five levels of emotion is “somewhat unhappy,” and the fourth is “very unhappy.” The third level of the six levels of cognition is “Somewhat forgetful, but able to think clearly and solve day to day problems,” while the fourth is “Somewhat forgetful, and have a little difficulty when trying to think or solve day to day problems.” The third level of the five levels of pain is “Moderate pain that prevents a few activities,” and the fourth is “Moderate to severe pain that prevents some activities.”

WHAT’S WRONG WITH HEALTH INEQUALITIES?

57

equalized by providing unequal non-health resources to compensate for health inequalities would be less just than a society in which health and non-health resources were separately equalized. Should egalitarian policy makers be concerned with specific inequalities in health, as the health exceptionalist urges, rather than extending their concern to other significant specific inequalities or, alternatively, should they limit their concern to overall inequalities in prospects and status? I see no way to justify singling out health inequalities as prima facie unjust, as the health exceptionalist does. After all, health itself consists of very different dimensions. If the fact that one individual is in pain while another has limited mobility raises no question of justice when their health is equally good, why should the fact that one individual is in poor health while another is poor raise questions of justice when the two individuals have equivalent life prospects?22 A more defensible form of specific egalitarianism rejects health exceptionalism and argues for the prima facie injustice of health inequalities on the grounds that all significant specific inequalities are prima facie unjust. In other words, the specific egalitarian denies that X tells egalitarians everything they need to know. The specific egalitarian denies that alternative ways of equalizing X across people are equally just. If one way involves fewer inequalities with respect to specific significant goods, then egalitarians should prefer it. Why? To offer a conclusive critique of specific egalitarianism would require an examination of the foundations of egalitarianism, which would be out of place in a paper devoted to a specific issue such as this one. But a good deal can be said even at a relatively superficial level. First of all, it is important to distinguish questions concerning the (distributive) justice of patterns of holdings from other questions concerning justice. Consider a parent who arbitrarily gives one child a larger allowance than another or a police officer who avoids giving speeding tickets to drivers of green Subarus. If one is thinking about justice in general, all sorts of inequalities may matter, even if the goods at stake are of small importance and the resulting pattern is egalitarian. What is at issue in this article is the justice or injustice of the pattern of health outcomes or expectations, not of the actions or process that led to that pattern.23 Why should an unequal distribution of any specific good be of moral concern apart from its role with respect to overall inequalities in life prospects or moral 22 This ad hominem argument does not apply to Anand, who maintains explicitly that we should pay attention to inequalities in specific dimensions of health (2002, p. 486). 23 Peter’s argument for what she calls an indirect approach to health distribution (2001) is in this regard ambiguous. Consider the following remarks from an article she co-authored with Evans: “it identifies as unjust those class, gender, race, regional, and other inequalities in health that originate in the basic structure of society . . .” (Peter and Evans 2001, p. 29). In denying that health inequalities are themselves unjust, Peter seems to be arguing both that the object of an egalitarian’s distributional concern should not be health and that it is a mistake to be concerned about distributions rather than the procedures and institutions that give rise to them.

58

DANIEL M. HAUSMAN

status? One possibility is that the inequality with respect to a specific good is important as evidence of other injustice. For example, it could be that the best explanation for an inequality in access to running water is that people have been treated unfairly, and so the unequal distribution is evidence of unjust treatment. But if there is a moral concern with patterns of holdings of specific goods and not just processes, the inequality needs to have some non-evidential importance. What could it be? The answer presumably lies in the rationale for egalitarianism. Egalitarians sometimes sound as if they believe that undeserved and unchosen inequalities with respect to any good are unfair. Larry Temkin writes, “Egalitarians generally believe that it is bad for some to be worse off than others through no fault or choice of their own. This is because, typically, if one person is worse off than another through no fault or choice of her own, the situation seems comparatively unfair . . .”24 One could read this as condemning a state of affairs in which one person is worse off than another with respect to any specific good – whether the good be children or transportation. But it is more plausible to suppose that Temkin is concerned with overall advantage – that is, with X, which may be equal even though holdings of particular goods are unequal. People live differently and encounter different environments, and egalitarians do not maintain that there is anything prima facie unjust if some lives possess more of some goods while other lives possess more of other goods, unless the inequalities in some particular goods are incompensable. To maintain otherwise is to risk making egalitarianism into a monstrous fantasy of utter uniformity. In an egalitarian utopia, all would be equal or equivalent with respect to some overall measure, but people’s lives would not all be the same. And if this were a liberal egalitarian utopia, outcomes would not be equal either. There seems to be no way to defend specific egalitarianism without finding diversity itself prima facie unjust. Egalitarians are concerned with equalizing X, and equalizing X does not imply separately equalizing the factors that jointly determine X. Egalitarians disagree about how to characterize X. According to Temkin, a distribution of specific goods will be relevantly unequal if and only if there are inequalities in well-being that are undeserved and unchosen. Richard Arneson, in contrast, holds that egalitarians should be concerned about inequalities in opportunities for welfare.25 Ronald Dworkin seeks to lessen inequalities in resources.26 G.A. Cohen takes the currency of egalitarian concern to be “access to advantage,”27 while for Amartya Sen and Martha Nussbaum it is functionings and capabilities.28 Elizabeth Anderson argues that more fundamental commitments to the democratic equality of persons ground distributive

24

Temkin 2003, p. 767. Arneson 1989. 26 Dworkin 1981. 27 Cohen 1989. 28 Sen 1992; Nussbaum 2000. 25

WHAT’S WRONG WITH HEALTH INEQUALITIES?

59

concerns,29 while Michael Walzer and David Miller favor “complex equality” where inequalities with respect to specific domains (other than fundamental rights and political status) are unimportant, provided that they are not correlated with one another.30 For the purposes of this paper, there is no need to choose among these versions of egalitarianism, because the thesis that I defend is common to them all. Apart from incompensable shortfalls and a few special cases, such as citizenship or political liberties – which are intimately connected to moral status and equality of respect – egalitarians are not concerned with the distribution of specific social goods. V. HEALTH INEQUALITIES AS INDICATORS OF OVERALL INEQUALITIES The fourth way in which one might defend the moral relevance of data concerning individual health inequalities is more practical and also more plausible. Even if health inequalities matter only to the extent to which they reflect overall inequalities in moral status or life prospects, the same thing is true of income and wealth. Yet data concerning income and wealth are thought to be of enormous importance to egalitarians as indicators of overall inequalities. In just the same way, one might argue that health is so important and so tightly correlated with other inequalities that information about health inequalities can serve as a proxy for information on overall inequalities. As Sudhir Anand puts the point, “Social inequalities in health may be seen as a sensitive barometer of the fairness of the underlying social order.”31 The fact that health depends heavily on individual choices, effort and luck does not automatically disqualify health inequalities as proxies for overall inequalities in moral status or life prospects, because accepted proxies such as income and wealth also strongly depend on individual choices, effort and luck. It is enough that health, like income and wealth, is strongly correlated with overall inequalities. Just as those who are more affluent tend to be better nourished, to attend better schools, to have better educated parents, and so forth, so do those who are healthier. In much the same way that one studies the distribution of income and wealth as indicators of these other inequalities, why shouldn’t one study the distribution of health? The idea might seem promising. Indeed, three crucial disanalogies between health on the one hand and income and wealth on the other support the proposal. Anand points out the first two.32 First, health – unlike income – is of intrinsic rather than merely instrumental value, and so its distribution seems more ethically significant than the distribution of income. Second, there are reasons to 29

Anderson 1999. Walzer 1983; Miller and Walzer 1995. Peter and Evans 2001, p. 29. 32 Anand 2002, pp. 485–6. 30 31

60

DANIEL M. HAUSMAN

welcome income inequalities that do not carry over unproblematically to health inequalities. In particular, income inequalities may serve as incentives to socially beneficial effort, and may consequently be deserved. Third, unlike income or wealth, health is essentially an individual trait. Though health is difficult to measure, each individual has his or her own state of health, while income and wealth are often held by groups and especially families. Individuals may be rich in virtue of their family connections, even though they have no income or wealth of their own. Whether Lear was rich or poor after he divided his kingdom among his daughters cannot be determined by examining his treasury. There are thus serious problems in determining how income and wealth are distributed among individuals, which have no parallel in the case of health. Since health is so important, so clearly attached to individuals, and so clearly correlated with other factors relevant to overall inequalities, there seems to be a strong case for studying health inequalities across individuals in just the same way that economists study inequalities across individuals of income or wealth. I shall dispute this case. Five considerations count against measuring individual health inequalities as a proxy for overall inequalities. (1) Health is difficult to measure. (2) Measuring health involves evaluating health, which disqualifies measures of health inequalities from serving as proxies for measures of inequalities in what one values. (3) Health is a worse proxy for overall inequalities than wealth or income. (4) Because health is not easily redistributed, information about health inequalities has a lesser strategic value to policy makers. And (5) if one already has proxies, does one need another one? A. TECHNICAL ISSUES Before turning to the argument against measuring individual health inequalities as a proxy for overall inequalities, I should say a few words about the technical complications of measuring health inequalities, most of which echo problems that arise with respect to measuring inequalities in wealth or income. Notice first that nobody seeks to equalize health (or income) hour-by-hour or day-byday. For example, lots of people in Wisconsin have colds in mid-February, 2006. These health differences are not of concern to those interested in health inequalities. This is not to deny that there could be relevant health inequalities even with respect to the lowly common cold. Those who are malnourished or who have poor housing get colds more often and have them longer. But it could turn out that those who generally suffer worst from colds are underrepresented among the actual suffers in Wisconsin in mid February 2006. Those who are in worse health at that particular time might be in better health overall. Many of the differences between people’s health at any particular moment are irrelevant to any ethical concerns one might have concerning health inequalities.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

61

Other differences in health – like differences in income and wealth – reflect the relationship between health and age or life-stage. Rachel is healthier not just in mid-February, but throughout 2006 than is her father, who is nearly 80, because she is probably healthier now than she will be when she is nearly 80, and because he is less healthy now than he was when he was in his 50s. She probably has a larger income, too. Inequalities such as these arise from relatively uniform intrapersonal inequalities between different ages. Intrapersonal health inequalities, like intrapersonal inequalities in income and wealth, might be morally significant, but it is tricky to say what their significance is, and they raise difficulties for a naive use of health or wealth or income inequalities as indicators of overall inequalities. Equality of health (like equality in income or wealth) across life stages may not be desirable, and in addition (unlike equality in income or wealth) it is not possible. Indeed, equality of health across age groups may be meaningless, if measures of health are relative to activities and objectives which change as one ages. Since interpersonal health inequalities may reflect inequalities across life stages, it seems that the relevant health inequalities must be either inequalities over a whole life or inequalities among people occupying the same life stage. However, both ways of construing health inequalities across individuals are problematic. If one examines inequalities across people at different stages of life, one needs lots of data and a non-arbitrary way to divide life into stages. One then either has to take the egalitarian goal to be equalizing health at every life stage,33 which lacks a clear justification and may be unreasonably demanding, or one has to explain to what extent inequalities in one life stage can compensate for inequalities at other life stages. The difficulties with focusing on whole-life inequalities are even greater. The fact that one is bracketing within-life inequalities means that overall measures might be misleading – to what extent can inequalities at one stage of life fully make up for inequalities at another? In addition, and more seriously, over 70 or 80 years, the effects of chance can easily overwhelm other factors that contribute to health inequalities. As Gakidou, Murray and Frenk show,34 realized health outcomes differ little between a hypothetical population where everyone’s health expectancy is 56.5 years and a second hypothetical population in which health expectancies vary between 47 and 82 years. Whole-life outcomes are bound to be quite unequal no matter how equally society distributes those social factors that contribute to health. Equalizing whole-life health outcomes is not a sensible moral objective. Either one has to disaggregate and focus on inequalities 33 Culyer 2001, p. 281. Once again there are parallel difficulties with respect to health and income. Indeed inter-temporal comparisons of wealth and income are more difficult because health unlike income or wealth has a natural maximum. These additional problems in comparing wealth and income do not arise for whole-life or life-stage comparisons of income or wealth within a single cohort, but egalitarian social policy must also address inequalities across cohorts. 34 Gakidou, Murray and Frenk 2000, p. 45.

62

DANIEL M. HAUSMAN

at different life stages, or one needs to shift one’s focus from health outcomes to health expectancies.35 Defining health expectancies is a complicated business that I shall not discuss here. In a co-authored previous paper, I argued that the notion of individual inequalities in health expectancies is (i) misleading, on the grounds that health expectancies are in fact derived from claims about group outcomes and (ii) incoherent, on the grounds that an expectation cannot be defined for an individual qua individual, but only qua member of some group.36 (Neither of these are objections to comparing health expectations for different social groups). For these reasons, health inequalities across individuals should be understood as inequalities in life-stage health outcomes. Supposing that one has defined and measured people’s life-stage health outcomes, one can plot their distribution, just as one plots distributions of other properties, and the measure of health inequality would then be some measure of the dispersion in that distribution. B. MEASURING VERSUS VALUING HEALTH The difficulties of measuring individual health inequalities discussed in the previous subsection are similar to and apparently no more serious than the difficulties of measuring individual inequalities in wealth and income. But there are many other complications involved in measuring health in terms of QALYs that have no analogies with respect to measuring income or wealth. As mentioned above, generic measures of health take a person’s health to depend on the amount of time the person spends in one or another “health state.” Health states are in turn defined by their location along some set of dimensions which capture different aspects of health or its consequences, such as the dimensions of the Health Utilities Index. If P’s health dominates Q’s – that is, if it is as close to full health along every dimension as Q’s and closer along at least one dimension – then P is healthier than Q. It is relatively easy to compare health states along individual dimensions, and some would argue that these comparisons can be independent of people’s environment, objectives and values. Comparisons of the health of individuals who have deficiencies along different dimensions are more problematic and apparently more dependent on features of the environment and on the objectives of individuals. Suppose, for example, one wants to compare the health of someone with a learning disability to the health

35 It is possible to make inequalities in incomes and wealth, unlike inequalities in health, very small, and those who measure inequalities in income and wealth are not forced to disaggregate or to focus on expectations rather than outcomes. On the other hand, those who study income and wealth inequalities often do disaggregate and study income and wealth inequalities at different life stages. 36 Hausman, Asada and Hedemann 2002. The same applies to inequalities in expectations of income or wealth, but the problem is less serious for those who measure income and wealth inequalities, since they typically measure outcomes rather than expectations.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

63

of someone confined to a wheelchair. What is the cross-dimensional metric that permits one to compare the “distance” between the health state of the person confined to a wheelchair and full health to the distance between full health and the health state of someone with a specific learning disability? It seems inevitable that what one determines is not who is healthier, but whose health is better along some evaluative dimension.37 In common with most of the literature on health measurement, I take what are called measurements of health states to be evaluations of health states, which depend on people’s objectives, features of the environment in which they live, and the consequences of their health states. Construing health measurement in this way undermines the view that information about inequalities in health states could serve as proxies for information concerning X. Not only are evaluations of health states dubious, rough and controversial, but if in supposedly measuring health, one is in fact evaluating health states, then inequalities in health states will (except in the case of dominance) simply be inequalities in some other value in terms of which health states are ranked. This difficulty can be clarified by considering a widely accepted view defended by John Broome and Dan Brock among others.38 According to this view, health states should be evaluated in terms of their impact on well-being. A change in someone’s health is an improvement in health if and only if it is an improvement in well-being. QALYs are, in their view, measures of health-related well-being. If, as Broome and Brock argue, inequalities in health are inequalities in well-being, then they are obviously much more difficult to measure than inequalities in income and wealth. If inequalities in health are inequalities in well-being, measuring health is measuring well-being, rather than measuring a proxy for well-being. The technical issues in this subsection and the last highlight how difficult it is to measure health in a way that would permit health inequalities to serve as proxies for overall inequalities. Not only does one need to come up with some reasonable way to evaluate health states – which does not in turn require that one measure directly what it is that health inequalities are supposed to indicate – but one needs to figure out how to quantify the overall health someone experiences over a life stage (and possibly then to develop an aggregate wholelife measure). If the measure of health is, as many researchers think, its impact on well-being, then there is no reason to focus specifically on health inequalities. It is not easy to measure income or wealth either, but it is much easier than measuring health, and one can measure them without first measuring well-being.

37 For a general discussion of the problems of valuing health states, see Hausman (2006). These problems have no analogues in measuring income and wealth, which are not multidimensional. 38 Broome 2002; Brock 2002.

64

DANIEL M. HAUSMAN

C. ARE HEALTH INEQUALITIES A GOOD PROXY FOR THE INEQUALITIES MATTER?

THAT

Let us set aside these measurement problems and ask whether it would be worthwhile to measure health inequalities as a proxy for overall inequalities in status or life prospects. Consider two further differences between health on the one hand and income or wealth on the other. First, health is not transferable and cannot literally be distributed. Health inequalities respond more slowly to policy than do income inequalities, and health cannot be redistributed directly.39 Although health disparities can sometimes be remedied, they often can be addressed only via compensation. So information about health inequalities cannot be linked to policy as easily as information about income or wealth inequalities. Second, even though health, like income or wealth, is strongly correlated with other factors that are important to life prospects, there is a wider range of variation in health states than of income or wealth among those who are otherwise privileged (or underprivileged). Since wealth can usually buy a good education and access to high status occupations and communities, and high status occupations provide high salaries, there are few individuals who are uneducated and have a low status yet have a high income. But a great many of the relatively disadvantaged enjoy good health. Virtually none of the privileged are lacking in wealth or income, but many are unhealthy and all eventually die. (Death has, with good reason, been called “the great equalizer”). Inequalities in income and wealth are consequently much better proxies for overall inequalities than are health inequalities, despite the problems in linking income or wealth to individuals. Since none of those who are destitute will be privileged and few of those who are rich will be oppressed, income or wealth is a more reliable indicator than health. Furthermore, if economists already have indicators of overall inequality, which are relatively straightforward to measure, why do they need another and worse one, which is in addition very troublesome to measure?

39 One might also question whether attempts to distribute health are ethically appropriate. Consider inequalities such as the number of friends one has or whether one has married successfully. These are genuine inequalities of considerable importance, and they are not entirely beyond the reach of social policy. Some such concern might, for example, justify laws against polygamy. Yet inequalities in friends and lovers, unlike inequalities in income, raise few questions of justice. Why? The answer cannot be that marriage, intimacy and friendship depend on luck and individual choice and effort, since income also depends on these. In some cases, the inequalities may be deserved, but desert is not the central issue. Even when not deserved, inequalities in friendships flow from intimate features of individuals with respect to which collective social interference is unreasonably intrusive. In the same way, one might argue that a concern with individual health inequalities is inappropriate or unreasonably intrusive. Health is more personal than income. Although it depends on social relations, health, unlike income and wealth, is not a social fact. I am not sure how much weight to place on this consideration. In any event the objection seems to be directed more to making the equalization of health across individuals a goal of social policy than to measuring the extent of individual health inequalities as a potentially useful way of measuring overall inequalities in resources or well-being.

WHAT’S WRONG WITH HEALTH INEQUALITIES?

65

VI. CONCLUSIONS To draw conclusions about social justice from information about health inequalities, one needs to know what health inequalities imply for overall inequalities such as inequalities in life prospects or moral status. When inequalities in QALYs are very large, their implications for overall inequality are unambiguous. Otherwise, one needs to know more than how health is distributed. In the case of inequalities across groups, one often does know more. As it happens, information about health differences between groups that are known to be otherwise unequal usually reveals that the overall inequalities are even worse than they appear when health is left out of the comparison. But it is possible that health differences between groups that are otherwise unequal might mitigate the other inequalities, as is the case of comparisons between the lower ranks of the military and law students. In either case, information concerning social group health differences would be relevant to egalitarians. No influential version of egalitarianism supports the view that health itself should be equalized. Apart from the special case of incompensable health differences, the implications of information concerning inequalities in health depend on whether they mitigate or, as is more typical, aggravate other inequalities. Nothing in this article suggests that health inequalities are unimportant! But data concerning inequalities in QALYs across individuals do not tell egalitarians what they need to know.

REFERENCES Anand, Sudhir. 2002. The concern for equity in health. Journal of Epidemiology and Community Health, 56, 485–87. Anderson, Elizabeth. 1999. What is the point of equality? Ethics, 109, 287–337. Arneson, Richard. 1989. Equality and equal opportunity for welfare. Philosophical Studies, 56, 77–93. Asada, Yukiko. 2006. Is health inequality across individuals of moral concern? Health Care Analysis, 14, 25–36. Brock, Dan. 2002. The separability of health and well-being. In Murray et al. 2002, pp. 115–20. Broome, John. 2002. Measuring the burden of disease by aggregating well-being. In Murray et al. 2002, pp. 91–113. Cohen, G.A. 1989. On the currency of egalitarian justice. Ethics, 99, 906–44. Culyer, Anthony. 2001. Equity – some theory and its policy implications. Journal of Medical Ethics, 27, 275–83. Daniels, Norman, Bruce Kennedy and Ichiro Kawachi. 2000. Is Inequality Bad For Our Health? Edited by J. Cohen and J. Rogers, New Democracy Forum Series. Boston: Beacon Press. Dworkin, Ronald. 1981. What is equality? Part 2: equality of resources. Philosophy and Public Affairs, 10, 283–345. Gakidou, Emmanuela E., Christopher J. L. Murray and Julio Frenk. 2000. Defining and measuring health inequality: an approach based on the distribution of health expectancy. Bulletin of the World Health Organization, 78 (#1), 42–54. Hausman, Daniel. 2006. Valuing health. Philosophy and Public Affairs, 34, 246–74.

66

DANIEL M. HAUSMAN

Hausman, Daniel, Yukiko Asada and Thomas Hedemann. 2002. Health inequalities and why they matter. Health Care Analysis, 10, 177–91. Marchand, Sarah, Daniel Wikler and Bruce Landesman. 1998. Class, health, and justice. The Milbank Quarterly, 76, 449–467. Marmot, Michael. 2004. The Status Syndrome. London: Bloomsbury Publishing. Miller, David and Michael Walzer, eds. 1995. Pluralism, Justice and Equality. Oxford: Oxford University Press. Murray, Christopher J. L., Emmanuela E. Gakidou and Julio Frenk. 1999. Health inequalities and social group differences: what should we measure? Bulletin of the World Health Organization, 77 (#7), 537–543. Murray, Christopher J. L., Emmanuela E. Gakidou and Julio Frenk. 2000. Response to P. Braveman et al. Bulletin of the World Health Organization, 78 (#2), 234–235. Murray, Christopher J. L., Joshua Salomon, Colin Mathers and Alan Lopez, eds. 2002. Summary Measures of Population Health. Geneva: World Health Organization. Nussbaum, Martha. 2000. Women and Human Development. Cambridge: Cambridge University Press. Peter, Fabienne. 2001. Health equity and social justice. Journal of Applied Philosophy, 18, 159–70. Peter, Fabienne and Timothy Evans. 2001. Ethical Dimensions of Health Equity. Pp. 25–33 in T. Evans, M. Whitehead, F. Diderichsen, A. Bhuiya and M. Wirth (eds), Challenging Inequalities in Health: From Ethics to Action. New York: Oxford University Press. Sen, Amartya. 1992. Inequality Reexamined. Cambridge, MA: Harvard University Press. Temkin, Larry. 1993. Inequality. Oxford: Oxford University Press. ___. 2003. Egalitarianism defended. Ethics 113, 764–82. Tobin, James. 1970. On limiting the domain of inequality. Journal of Law and Economics, 13, 263–78. Veatch, Robert. 1981. A Theory of Medical Ethics. New York: Basic Books. Walzer, M. 1983. Spheres of Justice. New York: Basic Books. Wilkinson, R. 1996. Unhealthy Societies: The Afflictions of Inequality. London: Routledge. Woodward, Alastair and Ichiro Kawachi. 2000. Why Reduce Health Inequalities? Journal of Epidemiology and Community Health, 54, 923–29.