¹ Otsuka, Michael and Voorhoeve, Alex, ‘Why It Matters that Some are Worse Off than Others: An Argument against the Priority View’, Philosophy and Public Affairs 37 (2009), pp. 171–99CrossRefGoogle Scholar. All otherwise unattributed page references are to this article.

² The canonical statement of the view is Parfit, Derek, ‘Equality or Priority?’, Bioethics, ed. Harris, John (Oxford, 2001), pp. 347–86Google Scholar.

³ Otsuka and Voorhoeve adopt the health state descriptions from Nord, Erik, Pinto, Jose Luis, Richardson, Jeff, Menzel, Paul, and Ubel, Peter, ‘Incorporating Societal Concerns for Fairness in Numerical Valuations of Health Programmes’, Health Economics 8 (1999), pp. 25–393.0.CO;2-H>CrossRefGoogle ScholarPubMed.

⁴ See Nord et al., ‘Incorporating Societal Concerns for Fairness in Numerical Valuations of Health Programmes’. For a recent review of the empirical literature, see Shah, Koonal K., ‘Severity of Illness and Priority Setting in Healthcare: A Review of the Literature’, Health Policy 93 (2009), pp. 77–84CrossRefGoogle ScholarPubMed.

⁵ Otsuka and Voorhoeve assume that in these examples all people have the same life expectancy, their overall utility is affected only by their health states, those who are in the same health state are at the same level of utility, and they remain in the health state they end up in for the rest of their lives. These strong assumptions raise a worry about the artificiality of considering the situation of the young adult in isolation. Shouldn't egalitarians take into account the fact that in real life the person in the Intrapersonal Case is not isolated and hence she would be worse off than others if she chose to take the drug against the slight impairment and ended up with the very severe impairment?

⁶ Of course, in practice the best we can hope for are approximations, since there is likely to be considerable variation between individual valuations. The differences might be systematic depending on whether the respondents have experience of the health states that they are asked to evaluate. One well-known way that experience influences individual valuation is due to adaptation. Patients who have adapted to living with a chronic condition tend to judge the health state associated with it less bad than those who do not have the condition. There is also some evidence that people evaluate health states by using their own health as a reference point. For those who are currently in poor health or have past experience of illness, more severe conditions seem relatively less bad compared to their reference point. Given their own health as a baseline, however, less severe conditions seem relatively worse. That is, respondents who treat their own experience of poor health as a reference point use a narrower range of values to evaluate health states. One implication of this phenomenon is that not only the levels of health, but also changes in health are evaluated differently depending on a person's health experience. I set these complications aside in this article. For further discussion and empirical studies, see Dolan, Paul, ‘The Effect of Experience of Illness on Health State Valuations’, Journal of Clinical Epidemiology 49 (1996), pp. 551–64CrossRefGoogle ScholarPubMed and Menzel, Paul, Dolan, Paul, Richardson, Jeff, and Olsen, Jan Abel, ‘The Role of Adaptation to Disability and Disease in Health State Valuation: A Preliminary Normative Analysis’, Social Science & Medicine 55 (2002), pp. 2149–58CrossRefGoogle ScholarPubMed.

⁷ More precisely, the assumption is that people have a constant risk attitude towards health states. (For simplicity, I discuss constant risk-neutrality, which is also the standard assumption.) The need for such an assumption is widely recognized. See, for instance, Broome, John, ‘Qalys’, Ethics Out of Economics (Cambridge, 1999), pp. 196–213CrossRefGoogle Scholar, and Loomes, Graham and McKenzie, Lynda, ‘The Use of QALYs in Health Care Decision Making’, Social Science & Medicine 28 (1989), pp. 299–308CrossRefGoogle ScholarPubMed.

⁸ The full scale consists of the following conditions: (1) full health; (2) slight impairment; (3) moderate impairment; (4) considerable impairment; (5) severe impairment; (6) very severe impairment; (7) complete disability; (8) dead. (For details, see the Appendix to Otsuka and Voorhoeve and Nord et al., ‘Incorporating Societal Concerns for Fairness in Numerical Valuations of Health Programmes’.) The reason differences between adjacent steps are considered to represent equal intervals is that even though initially the scale was developed using direct scaling methods, the results were later corroborated by other preference elicitation methods, including the standard gamble. When direct scaling methods are used, respondents are asked directly to evaluate health states at a high level of precision (e.g. on interval or ratio scales). These methods assume that respondents can make these comparisons, but the empirical evidence that they can do so consistently and in a manner that correlates with results obtained by indirect methods (including the standard gamble and the person trade-off) is notoriously inconclusive (see, e.g., Froberg, Debra G. and Kane, Robert L., ‘Methodology for Measuring Health-State Preferences – II: Scaling Methods’, Journal of Clinical Epidemiology 42 (1989), pp. 459–71)CrossRefGoogle ScholarPubMed. Hence the fact that the health state utilities on the scale used in these studies are consistent with results obtained by the standard gamble is significant – especially given that many health economists regard the standard gamble as the benchmark preference elicitation method because it is based directly on the axioms of expected utility theory. See also Nord, Erik, ‘The Trade-Off between Severity of Illness and Treatment Effect in Cost-Value Analysis of Health Care’, Health Policy 24 (1993), pp. 227–38CrossRefGoogle ScholarPubMed and Nord et al., ‘Incorporating Societal Concerns for Fairness in Numerical Valuations of Health Programmes’, pp. 29–30 for a more detailed explanation of the way the scale was derived. (I thank Alex Voorhoeve for clarification of the way they interpreted these studies, and Erik Nord for discussion on the development of the scale.)

⁹ Nord, Erik, Cost-Value Analysis in Health Care: Making Sense out of QALYs (Cambridge, 1999), pp. 115–16CrossRefGoogle Scholar. But it must be noted that this is a minority view. For instance, the widely used burden of disease measure, originally developed by the World Bank and the World Health Organization, uses a person trade-off protocol to assign ‘disability weights’ to different health conditions. These weights are not intended to reflect distributive considerations; they represent the burden that different health conditions impose on a person relative to full health. For details, see Murray, Christopher J. L., ‘Rethinking DALYs’, The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020, ed. Murray, C. J. L. and Lopez, A. D. (Harvard School of Public Health on behalf of the WHO and the World Bank, 1996), pp. 1–98Google Scholar.

¹⁰ In terms of the ranks on the scale, described in note 8, the first group can be moved from step 6 to step 2, and the second group can be moved from step 6 to step 4.

¹¹ Nord, Erik, Undrum Enge, Anja and Gundersen, Veronica, ‘QALYs: Is the Value of Treatment Proportional to the Size of the Health Gain?’, Health Economics 19 (2010), pp. 596–607CrossRefGoogle ScholarPubMed, at 606. The experiment that is summarized here was only one of four to study the hypothesis that distributive considerations enter into personal valuations of health gains as well.

¹² There might be, however, a question about how far the correspondence between people's judgments and prioritarianism goes. Most of the relevant studies focus on ‘concern for severity’ – that is, the concern for giving priority to those whose health condition is worse. But perhaps the concern for severity is narrower than the prioritarian concern for the worse off, hence the empirical results cannot ultimately be used to support the prioritarian view. A recent study, however, found that similar results are obtained when the better and worse off patient groups are identified in non-health terms as being ‘advantaged’ and ‘disadvantaged’. In fact, the study found that there is a stronger preference for providing treatment to the worse off group when it is described in more general terms. These results may suggest that the concern for severity can indeed be interpreted as a general prioritarian concern for the worse off. (Green, Colin, ‘Investigating Public Preferences on “Severity of Health” as a Relevant Condition for Setting Healthcare Priorities’, Social Science & Medicine 68 (2009), pp. 2247–55.)CrossRefGoogle ScholarPubMed

¹³ Both quotes are from p. 179, with capitals added to case names.

¹⁴ For a take of empirical researchers on this issue, see Menzel, Paul, Gold, Marthe R., Nord, Erik, Pinto-Prades, Jose-Louis, Richardson, Jeff, and Ubel, Peter, ‘Toward a Broader View of Values in Cost-Effectiveness Analysis of Health’, Hastings Center Report 29 (1999), pp. 7–15CrossRefGoogle Scholar. See also Walker, Rebecca L. and Siegel, Andrew W., ‘Morality and the Limits of Societal Values in Health Care Allocation’, Health Economics 11 (2002), pp. 265–73CrossRefGoogle ScholarPubMed, for a more sceptical view.