«Edited by ANNE MASON Research Fellow, Centre for Health Economics University of York and ADRIAN TOWSE Director, Ofﬁce of Health Economics Radcliffe ...»
whether the authors assume that it should have a precise relationship to a QALY. There appears to be a considerable degree of ‘content’ overlap between this and the EQ-5D, but without a clear indication, for example, of which dimensions of the EQ-5D will be covered by the ﬁrst proposed dimension of ‘physical health’ in this new instrument, or how the ‘mental health problems’ dimension relates to the EQ-5D dimension of ‘anxiety/ depression’. It is curious, but probably relatively unimportant, that Dolan and Tsuchiya suggest four levels for each of the dimensions, when the EQ-5D instrument currently uses three and the EuroQol group is busy attempting to remodel the instrument with ﬁve levels (Kind and Macran, 2002). More fundamentally, it is unclear to me whether the four dimensions of ‘mental health problems’, ‘constraints on behaviour’, ‘relationship to others’ and ‘feelings of vulnerability’ are conceptually sufﬁciently separable to make a subsequent valuation process meaningful. Other readers would no doubt identify additional concerns of detail.
However, this is essentially an ‘ideas’ paper, in which much of the detail is not yet ﬁnalised, and my main focus therefore is not on the detail but on the bigger, and ultimately more important, picture. Of course, I can see a real value for the Home Ofﬁce in what the authors propose: it would provide a useful measure of effectiveness against which to compare different policies to reduce the extent, or effect, of crime. But in the bigger schema, how would a cost per QALY from a health-sector investment be compared with a cost per SALY from a Home Ofﬁce investment? Would not Alan, who was rarely daunted by the magnitude of a task he saw as important, have wanted this instrument alongside his much loved QALY to be capable of addressing this comparison?
Essentially the EQ-5D provides a means to record and value the adjustment to length of life due to health-related quality of life. It has long been criticised by some economists because it may not adequately represent individual preferences for health outcomes (for example, Mehrez and Gafni (1999)) and by others in that it does not directly address all the factors, such as information, that would appear in an individual’s utility function (for example Protiere et al., 2004). Rather it provides a partial measure focusing on the impact of changes in health outcomes. The focus on health outcomes is most often justiﬁed in terms of its use in helping to estimate the effectiveness of health interventions funded predominantly from health service budgets.
Dolan and Tsuchiya now offer us the potential for a new quality adjustment reﬂecting security or the non-pecuniary effects of crime. Rather than a SALY – a security-adjusted life year – it is perhaps more useful to think of it as a crime-related QALY. Again the argument for this focus is that for the Home Ofﬁce this measure potentially offers a method to show how interventions within its control can change quality of life. Essentially it provides another sectoral and partial QALY.
It does not involve a great ﬂight of imagination to see the possibilities for
DISCUSSION OF ‘THE MEASUREMENT AND VALUATION OF PUBLIC SAFETY’ 149other partial or sectoral QALY measures. It is easy to imagine an environmentrelated QALY or an education/opportunities-related QALY, which might provide broadly equivalent measures for other areas of public spending. But before we get to that point – indeed arguably before Dolan and Tsuchiya go further with crime-related SALY – we need to consider how these various QALYs should relate to each other.
One possible model is to actively encourage the development of a series of non-overlapping sectoral descriptors of the way in which, and extent to which, quality of life is diminished by sectoral factors – health, crime, environment, education and opportunities, etc. Each would need to be scaled (directly or indirectly) in a similar way. Tentatively, I would suggest that might be from the absence of any diminution of quality of life with respect to that ‘sector’ (value 1) to ‘natural death’ (value 0). If they are to be used together it would be desirable that the scales did not overlap, and so the SALY here would not include any descriptors or dimensions of health that appeared in, or were already covered by, the health-related QALY. (In principle it would be possible to allow for some common dimensions, if described identically.) An additional partial QALY could be developed for any set of ‘sectoral factors’ not hitherto covered, so clearly an initial task might be to consider what the full range of sectoral QALYs might look like.
This development also opens up the opportunity to consider what characteristics of the process, by which services are delivered, contribute to individual and collective well-being, and to include these in a component of the superQALY. For example reduced waiting times for appointments with the health system (even if they have no impact on outcomes) values, like faster police response times, appear to have a real value to the public (Cave et al., 1993).
Together they would enable us to describe, measure and value what I would call a ‘super-QALY’* – a super-QALY being equivalent to a year of life undiminished by limitations in any major aspect of life (or at least on any contributing QALY instrument). A super-QALY would thus have to combine adjustments for limitations on any scale, so that there would need to be a cross-sectoral relative valuation exercise to establish how values for one factor compare to another. In principle, a cross-sectoral exercise could seek to set values for each of all the combined states that the system describes.
Of course in practice, the partial QALYs which contribute to the superQALY are not exclusively produced by or attributable to any one sector of the economy or public sector spending. It would in principle be possible to generate any type of QALY in any context. Thus for example, crime reduction might well generate improvements that would be measured on the health-related, as well as the crime-related, QALY scales. Similarly health * I am aware that others have used the term super-QALY in different ways, particularly to describe an equity-weighted QALY, as for example by Anand (1999). I would contest that my use of the term is more appropriate and consistent with both the correct and common usage of the adjective, but I confess bias in this respect.
150 THE IDEAS AND INFLUENCE OF ALAN WILLIAMSinterventions for children might well generate health-related and education/ opportunities-related QALYs. Full-blown evaluations of an intervention could consider all dimensions; more limited evaluations could focus on the most important and relevant sectoral scale.
Obviously such a broader vision needs to be thoroughly debated. There would clearly be those advocates of cost–beneﬁt analysis, who would argue that it is all unnecessary and that it just emphasises the need to use monetary valuations of changes, without complex, intermediate multidimensional scaling. There would be others who might argue that we should proceed on a piecemeal basis sector by sector, as implicitly accepted in this paper, because that way useful progress can be made more quickly and the pursuit of the ‘holy grail’ will only stand in the way of more modest but valuable progress. There will be those existing users of health-related QALYs who will appreciate that such an approach has important implications for our existing measures, and may resist rethinking aspects of them. For example, the concept of a natural death, introduced in the Dolan and Tsuchiya paper, brings into question what has been implicitly assumed, for example in the EQ-5D valuation exercises, but potentially offers a system that can much more readily reﬂect what many would believe, that valuable quality of life can be gained by allowing death to take place in a more natural way or in a more natural environment – the belated recognition of which would receive a cheer from the hospice and homecare movements.
This brief discussion is not the place to attempt to anticipate that vigorous debate, but the Dolan and Tsuchiya paper provides an excellent incentive, and an urgent reason, to begin such a debate. I would not be so bold as to claim I know what position Alan Williams would have taken, but I’m conﬁdent he would have enjoyed, encouraged and taken an active part in that debate, and he would not have been daunted by the magnitude of the potential research agenda it might create.
REFERENCES Anand P (1999) QALYS and the integration of claims in health-care rationing, Health Care Analysis 7: 239–53.
Cave M, Burningham D, Buxton M et al. (1993) The Valuation of Changes in Quality in
the Public Services. Report prepared for HM Treasury by Brunel University. London:
Kind P and Macran S (2002) Levelling the Playing Field: increasing the number of response categories in EQ-5D. Proceedings of the 19th Plenary Meeting of the EuroQol Group, University of York, September 2002. http://gs1.q4matics.com/euroqol_zoek_ test/DocumentBekijken.rbm?DocID=110117&KlantID=125 (accessed 9 March 2007).
Mehrez A and Gafni A (1989) Quality adjusted life years (QALYs), utility theory and healthy years equivalent. Medical Decision Making 9: 142–9.
DISCUSSION OF ‘THE MEASUREMENT AND VALUATION OF PUBLIC SAFETY’ 151Protiere C, Donaldson C, Luchini S et al. (2004) The impact of information on nonhealth attributes on willingness to pay for multiple health care programmes, Social Science and Medicine 58: 1257–69.
Williams A (1997) The measurement and valuation of health: a chronicle. In:
Culyer AJ and Maynard AK (eds). Being Reasonable about the Economics of Health.
Cheltenham: Edward Elgar Publishing Limited, pp. 136–75.
Sugden, Robert van Hout, Ben collaborations with Williams 8–13, 58 on Alan Williams 131–2 differences of opinion 9–11 on QALYs 127–9 education and early papers 8–9 on valuation methods 129–31 interpretations of Williams’ works 13–17 Viagra 131 further critiques and analysis 19–25 Visual Analogue Scales (VAS) 80, 116, 121, summary measures of population health 130 (SMPH) 112 super-QALYs 81, 149–50 Wagstaff, A and van Doorslaer, E 104 supply mechanisms Ware, JE and Sherbourne, CD 115 actual practices 32 warm-glow effects 139 ideological perspectives 30, 36–9 Watts, Vincent 133 tackling inefﬁciencies 49–54 Weinstein, MC and Stason, WB 115 SWB see subjective well-being (SWB) measures welfare economics and CBA, historical Szende, A and Williams, A 89 perspectives 9–10 welfarist approaches see decision-maker tariff policies see payment by results (PbR) approaches; Paretian welfare economics technological advances 34, 50 well-being frameworks (Sen) 77 Thatcher, Margaret Wennberg, J 38 introduction of public choice theories Wennberg, J and Gittelson, A 53 10 Whitby–Scarborough railway closures 8 introduction of ‘value for money’ Why are Some People Healthy and Others not?
micromanagement 27–8 (Evans et al.) 108 Thurstone, LL 113 Wildavsky, Aaron 6–7, 10 time trade-off (TTO) measures 113, 118–19, Williams, Alan (AW) 121, 130–1 on age discriminators 90–8, 106 for crime and loss valuations 140–1 on cost–beneﬁt analysis (CBA) 11–17, 20–1 Tobin, J 59 on equity and fairness in health care 88–98 Torrance, GW et al. 115, 119 on establishing QALYs 114–17, 128–9 Towse, Adrian, on resource allocation on expenditure drivers 50–1, 54–5 approaches 75–82 on Patient Reported Outcome Measures 39 Tsuchiya, A 90 on public–private health care challenges Tsuchiya, A and Dolan, P 27–44 on equity and fairness in health care on resource allocation approaches 11–17, 85–99 20–1, 58–60 critiques and further analysis 103–8 on TTO measures 131 Tsuchiya, A and Williams, A 98 personal recollections 1–3, 131–2 TTOs see time trade-off (TTO) techniques willing to forego (WTF) measures 92–4, 98–9 willingness to accept (WTA) measures 138–9 United States willingness to pay (WTP) measures 138–9 1980s healthcare systems 28 concepts and comparisons 10–12, 75–6, GDP investment–appropriate care 37 79–80 practice guideline effectiveness studies 53 cf. egalitarian systems 14–17 use of HMOs 54 crime risk studies 139 user charges 35–6 Winkielman, P et al. 137 ‘utility’ concepts 129 Wiseman, Jack 57 utility measures 61, 67–8, 71, 115, 117–20 Wright, C et al. 53 WTA see willingness to accept (WTA) measures valuation mechanisms WTF see willing to forego (WTF) measures market vs. paternalistic forces 7–8, 10–11, WTP see willingness to pay (WTM) measures 13–17, 22–3 see also health indicators Yip, WC and Hsiao, WC 35 The Value of Life (Harris) 90 Yi, Y et al. 35 ‘value for money’ policies 27–8