Whose values should count – those of patients or the general public – when adopting the quality-adjusted life year (QALY) framework for healthcare decision making is a long-standing debate. Specific disciplines, such as economics, are not wedded to a particular side of the debate, and arguments for and against the use of patient values have been discussed at length in the literature. In 2012, Sinclair proposed an approach, grounded within patient preference theory, which sought to avoid a perceived unfair discrimination against people with disabilities when using values from the general public. Key assumptions about general public values that beget this line of thinking were that ‘disabled states always tally with lower quality of life’, and the use of standardised instruments means that ‘you are forced into a fixed view of disability as a lower value state’ (Sinclair, 2012). Drawing on recent contributions to the health economics literature, we contend that such assumptions are not inherent to the incorporation of general public values for the estimation of QALYs. In practice, whether health states of people with disabilities are of ‘lower value’ is, to some extent, a reflection of the health state descriptions that members of the public are asked to value.
- allocation of healthcare resources
- health economics
- quality/value of life/personhood
Statistics from Altmetric.com
In the economic evaluation of healthcare treatments and interventions, ‘benefits’ are often expressed in terms of quality-adjusted life years (QALYs). QALYs value the benefits of healthcare in terms of time spent in a series of quality-weighted health states, incorporating the impacts on survival and quality of life in a single measure.1 The quality-adjustment component of the QALY is based on preferences for health states, and a long-standing debate has focused on whether these preferences should be those of patients (ie, individual-level values derived from those who are experiencing the health states pertinent to the studyi) or the general public (average values, often referred to as community-derived or population-based).
In a 2012 Journal of Medical Ethics paper, Sinclair proposed an approach, grounded within patient preference theory, which sought to avoid a perceived unfair discrimination against people with disabilities when using values from the general public.2 The accusation of discrimination is based on the notion that members of the general public will, on average, assign a lower value to a health state that comprises a disability compared with an individual who experiences that same health state. Indeed, many studies have reported that health state values elicited directly from patients exceed community-derived values for the same health states.3 4 The following hypothetical example – similar to the one used by Sinclair (p158) – describes how this discrimination could manifest itself in a resource allocation decision context. Consider two groups of individuals that are identical with the exception that those in Group A live with a disability and those in Group B do not. Now consider that all individuals contract a life-threatening illness, and a curative treatment (the only treatment available) will return individuals to their pre-illness health state. After receiving treatment, average life expectancy in the two groups is the same, but the quality-adjusted life expectancy will differ if those in Group B were assigned a higher value by the general public for their pre-illness health state. Assuming treatment costs are the same for all individuals, this means that providing treatment for Group A is less cost-effective compared with Group B. In the 1980s, Harris referred to this eventuality as a problem of ‘double jeopardy’ because individuals with pre-existing conditions suffer a second misfortune due to their reduced capacity to benefit from life-saving treatment.5 6
The objective of this Response is not to critique the ‘new method of researching public opinion’ offered by Sinclair (p158) or to add support to one side of the patient values versus public values debate (see Versteegh and Brouwer (2016) for a comprehensive discussion on this topic4). Instead, we wish to challenge the key assumptions that prompted Sinclair’s pursuit of a patient-focused approach – assumptions that also align with Harris’ double jeopardy argument – in particular, that the incorporation of general public values guarantees that ‘disabled states always tally with lower quality of life’ (Sinclair, p159) and that the use of standardised instruments means that ‘you are forced into a fixed view of disability as a lower value state’ (Sinclair, p159).ii
Incorporation of public preferences
Prior to examining the existence of inherent disability discrimination when using public preferences, it is pertinent to revisit how general public values are incorporated in the estimation of QALYs. The most common approach is through the use of generic preference-based health-related quality of life (HRQoL) instruments. These standardised instruments (eg, Assessment of Quality of Life 8-dimension (AQoL-8D), EQ-5D-3L, EQ-5D-5L, Health Utilities Index Mark 3 (HUI-3), and SF-6D7) comprise questions and response options that enable a respondent to describe their current health state. The valuation of the described health state comes from an ‘off the shelf’ set of scores obtained from a nationally representative valuation study. In national valuation studies, members of the general public are asked to value a sample of the finite number of health states defined by an instrument. For example, an instrument comprising five questions, each with five levels of response, defines 3125 (55) unique health states, and a member of the public participating in the valuation study could be asked to value a subset of these states. Through a combination of statistical modelling and participants valuing different subsets of health states, the end product of a national valuation study is a complete set of general public values for all possible response permutations.
Inherent disability discrimination?
The charge of discrimination asserts that individuals living with a disability will assign a higher valuation to their health state compared with the average valuation obtained from a representative sample of the general public. Before addressing this point directly, it is important to understand what a comparison of these health state valuations means in practice, appreciating that such data are not a like-for-like comparison. For illustration, taking the EQ-5D-3L as the preference-based instrument and the time trade-off (TTO) method as the preference elicitation technique,7 a comparison of patient versus public values would comprise consideration of two sets of health state values: (1) the TTO values elicited from the patients and (2) the general public valuations of the health states described by the patients when completing the EQ-5D-3L. Divergence between patient values and public values is not, in itself, problematic.4 There are multiple reasons why these two sets of values may differ, ranging from fundamental differences in the valuations of individuals who experience a health state compared with individuals asked to consider a hypothetical health state, to methodology-related differences such as the administration of the preference elicitation task or the statistical model used to estimate general public values. However, even after taking these reasons into account, one should expect the two sets of health state values to be different. This is because the patient is valuing their own health state, whereas it is the patient’s description of their health state that is valued by the public, where the health state descriptions are restricted to the standardised dimensions, items and response options contained within the respective preference-based HRQoL instrument.
The question of whether individuals living with a disability assign a higher value to their health state compared with the values of the general public is now placed within its correct context. It is not possible to attribute a difference between patient values and public values to an inherent disability discrimination within the QALY framework. Public values may be lower than patient values because the descriptive system of a preference-based HRQoL instrument constrains the patient to describe their health in an inaccurate way (inaccurate from the patient’s perspective), resulting in the description of the health state to appear worse than the lived experience. Similarly, public values may be higher than patient values if the descriptive system does not allow the patient to indicate a pertinent impairment, resulting in the description of the health state to appear better than the lived experience. Such eventualities are not criticisms of the QALY framework, they are consequences of the preference-based HRQoL instrument used to estimate QALYs. Accordingly, any concern about disability discrimination is an empirical question about the extent to which a standardised descriptive system enables respondents to adequately describe their state.
Recent contributions to the health economics and quality of life literature have highlighted the relative importance of descriptive systems in understanding the validity of preference-based HRQoL instruments.8–10 With specific regard to disability research, Whitehurst and colleagues have shown that variation in the framing of mobility-related items is a key driver of differences in index scores derived from preference-based instruments in the context of spinal cord injury.10 The study identified decrements in health state valuations that were imposed on respondents by virtue of the wording of response options to the Mobility dimension of the EQ-5D-5L and the Ambulation dimension of the HUI-3, where all response options are framed around the respondent’s ability to walk. Conversely, individuals were not prevented from reporting health state values at the upper end of the scoring scale when completing the AQoL-8D or SF-6D instruments because response options to the respective mobility-related items are not framed around the respondent’s ability to walk. Whether or not to frame mobility-related items in terms of respondents’ ability to walk is a normative question that requires further attention. However, what is evident from this research is that you are only ‘forced into a fixed view of disability as a lower value state’ (Sinclair, p159) if the descriptive system of the preference-based HRQoL instrument restricts the respondent’s ability to provide an accurate description of their health state.
There are numerous theoretical, methodological and ethical challenges regarding the use of QALYs in healthcare resource allocation decision making. We believe that denouncing the concept of QALYs because of inherent disability discrimination is misdirected — the discrimination lies, if anywhere, in the descriptive systems of preference-based HRQoL instruments.
We would like to thank Dr. Hareth Al-Janabi (University of Birmingham, UK) for his helpful comments and suggestions on an earlier draft of this paper.
↵ i The term ‘patient’ is used throughout the paper, referring to the individual who is experiencing the health state in question. This is in keeping with the language used by Sinclair and the general health economics literature concerning the debate around patient versus public valuations of health states.
↵ ii Sinclair made this statement in reference to standardised instruments conceptually linked to Amartya Sen’s capability approach, rather than preference-based HRQoL instruments that are widely used to estimate QALYs. It is our assertion that this statement applies to standardised instruments per se (capability instruments and preference-based HRQoL instruments) because both approaches comprise a fixed list of attributes that, in the opinion of Sinclair, ‘(do not) support the nuanced approach we are looking for’ (p159). It is also important to note the context of Sinclair’s statement. It is difficult to argue against the view that ‘you are forced into a fixed view of disability as a lower value state’ when considering the valuation of health states (any health states) relative to ‘full health’ because of how preference elicitation tasks are framed. For example, a TTO exercise requires respondents (patients or the general public) to consider how much time they would be willing to trade in order to achieve full health, suggesting that the health state being valued is less desirable than full health. However, this is not the context of the issue raised by Sinclair. Sinclair’s proposition is that patient values should be the focus of research rather than the values of the general public, which implies that it is the use of general public values that imposes a lower valuation (lower, presumably, relative to a patient’s own valuation).
Contributors Both authors were involved in the conception and framing of the Response and have read and approved the final manuscript for resubmission. DGTW wrote the first draft and subsequent revisions. LE was involved in the review and critical revision of the content prior to submission.
Funding This paper was written without financial support from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests DGTW is a member of the EuroQol Group.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.