Introduction
Health equity is one of the main avowed objectives of public health policy across the world. Yet economic evaluations in public health (like those in health care more generally) continue to focus on efficiency objectives – in particular, improving population health – whilst ignoring explicit consideration of equity objectives in particular, reducing health inequalities. For the most part, this disparity is not due to limitations of methodology or data availability as much more could be done with methods and data that currently exist, nor yet (as far as we can tell) to any anti-egalitarian ideological bias on the part of economists.
The National Institute for Health and Clinical Excellence (NICE) has recently taken on responsibility for the evaluation of public health interventions. The NICE guidance for the appraisal of public health interventions differs slightly from the standard reference case for applying to drugs, devices, or medical procedures. The perspective on costs is broader, encompassing all costs falling on the public sector, thus recognizing the broader, inter-sectoral, nature of most public health interventions. In addition, whilst the Quality-adjusted Life-Year (QALY) is still the preferred measure of health gain, it is recommended that a ‘cost-consequence’ approach is used in order to take account of the complexity and multi-dimensional character of public health interventions. This allows explicit consideration of multiple, non-health related and/or outcomes that are difficult to quantify. It also means that issues such as equity and distribution, which are key to public health policy, can be used to inform the analysis (NICE, 2006a).
Some aspects of income distribution and equity have been explored in HM Treasury’s Green Book (http://greenbook.treasury.gov.uk/). It is proposed that ‘distributional weights’ might be applied to monetary costs and benefits to reflect diminishing marginal individual value of income – i.e. that a gain or loss of a thousand pounds matters more to a poor individual than a rich individual. However, in focusing on an individual’s income and converting that into individual value, there is no consideration of the social value of improving the health of different individuals – for example, a poor individual with a relatively low life expectancy compared to that of a rich person with a relatively high life expectancy.
A systematic review of 424 economic evaluation studies of health interventions published in 1987, 1992, 1995, 1996, and 1997 concluded that ‘Distributional effects seem to have been completely neglected in existing economic evaluations’ (Sassi et al., Reference Sassi, Archard and Le Grand2001). Our own systematic review of 154 economic evaluation studies of public health interventions published from 2000 to 2005, came to much the same conclusion (Drummond et al., Reference Drummond2006).
The neglect of equity concerns is ubiquitous across all economic evaluations in the health sector, in both clinical medicine and public health. However, in public health evaluations, concern for health inequality reduction takes on particular importance as a policy objective. In contrast to clinical medicine, policy-makers in the public health sphere are more focused on concerns about population health inequalities. Therefore, there is potentially greater demand for such approaches to be applied in practice within public health. Issues underlying the disparity between clinical medicine and public health include the individualized nature of clinical medicine and the principle of equality of access. In clinical medicine, tackling health inequalities between population groups would typically require discrimination between those groups (e.g. by postcode, income, age, sex, ethnicity) in the volume and quality of care provided to people who are currently suffering from debilitating and possibly life-threatening conditions. This form of discrimination violates the principle of equality of access and is generally considered to be both ethically and politically unacceptable. For example, the report by NICE on social value judgements explicitly rejects discrimination of this kind in relation to its guidance in clinical medicine (http://www.nice.org.uk/). Whereas in public health, tackling inequality requires discrimination, for example in the volume and quality of preventative programmes between groups of relatively healthy people not currently suffering from severe illness. This form of discrimination, or ‘targeting’, is frequently considered to be politically acceptable up to a point. For example, restricting access to free child-care and stop-smoking facilities to families living in disadvantaged areas is more acceptable than restricting access to the latest life-saving breast cancer drugs to those same families.
Despite the importance of health inequality in public health, in practice economic evaluations in public health still typically fail even to identify and measure health inequality impacts – i.e. to provide factual information about how an intervention might change existing patterns of health inequality between equity-relevant population subgroups – let alone to value them – i.e. explicitly to quantify how much a particular measured reduction in health inequality is worth, compared with other uses of scarce resources.
In this paper, we discuss four main approaches for incorporating equity considerations into economic evaluation in public health: (i) review of background information on equity, (ii) health inequality impact assessment, (iii) analysis of the opportunity cost of equity, and (iv) equity weighting of health outcomes. All four approaches can be used to address concerns for equitable outcomes (in particular, health inequality reduction); approaches (i) and (iii) can also be used to address concerns for equitable procedures (e.g. non-discrimination, respect for individual liberty). We conclude that the first three approaches can readily be implemented using standard methods of health technology assessment. The fourth approach, equity weighting, requires further value elicitation research, focusing on equity considerations and contexts relevant to public health, as opposed to health care more generally. In particular, research is warranted on equity considerations relating to socio-economic status, the degree of voluntariness or personal responsibility for health risk, the value of treating current ill-health versus preventing future health risk, and the aspects of health inequality that the general public is most concerned about.
Background: equity considerations in economic evaluation
We can distinguish at least three different types of equity consideration that decision-makers may have in mind when taking decisions about the allocation of scarce resources in public health. First, concern to reduce a particular type of health inequality – such as inequality in quality-adjusted life expectancy between different socio-economic groups (e.g. http://www.ash.org.uk/html/policy/inequalitiesfacts.pdf). Second, concern to give priority to improving the health of a particular group – such as children or the severely ill. Third, concern to respect a particular ethical rule or procedure – such as a prohibition on race discrimination or respect for individual liberty, e.g. in cases of mandatory public health measures such as water fluoridation (Shackley and Dixon, Reference Shackley and Dixon2000).
Equity considerations may sometimes conflict with a strict cost-effectiveness goal of maximizing the improvement of population health, giving rise to ‘equity–efficiency trade-offs’. For example, it may be possible to save more lives in total with a given smoking cessation budget by focusing on ‘easy-to-reach’ affluent smokers rather than ‘hard-to-reach’ deprived smokers. Equity considerations may also conflict with one another, giving rise to what we might call ‘equity–equity trade-offs’. For example, smoking restrictions in public places may involve trade-offs between individual freedom to smoke versus reduced socio-economic health inequality (as well as improved health). Another example, giving blanket priority to improving the health of all children rather than all adults, may sometimes conflict with the goal of reducing socio-economic inequality in health, since some children are considerably wealthier and healthier than some adults.
There is a substantial theoretical literature on health equity considerations, which seeks to clarify the nature of alternative equity considerations by teasing out the logical relations (and contradictions) between them and by developing arguments about which equity consideration or set of considerations should be viewed as most important in which circumstances. This health equity literature draws on the wider literature on equity and justice, with contributions from a number of different disciplines, including economics, epidemiology, philosophy, and others (Anand et al., Reference Anand, Peter and Sen2004). Some of this theoretical literature examines how different health equity considerations might be defined in a quantitative manner suitable for weighing up against one another, and against efficiency considerations, within an economic evaluation framework (Culyer and Wagstaff, Reference Culyer and Wagstaff1993; Williams and Cookson, Reference Williams, Cookson, Culyer and Newhouse2000).
Economic evaluations implicitly embody important value judgements about equity, in at least three ways. First, the selection of interventions to be evaluated may involve equity judgements. Some interventions with important health inequality impacts may not be selected for evaluation because they are too expensive or too controversial, or because the relevant health inequalities are too low on the current political agenda. Secondly, economic evaluations typically exclude certain costs and benefits from the analysis on equity grounds. In the health care context, for example, the equity judgement that medical treatment should not depend on the patient’s social role (e.g. the doctor with young children versus the childless tramp) precludes examination of effects on gross national output and the welfare of the patient’s dependents, both of which may ultimately influence population health. Thirdly, and perhaps most importantly, health economic evaluations typically embody the quasi-egalitarian principle that ‘a QALY is a QALY’, that is a QALY has equal weight, no matter to whom it accrues. The conventional definition of efficiency in economics, known as ‘Pareto efficiency’, is a situation in which no one can be made better off without some one else being made worse off. This conventional definition does not make a strong value judgement about equity. It does not arbitrate, for example, on whether it is worth funding an intervention that improves the health of one group by 1,000 QALYs while reducing by 900 QALYs the health of another group who would otherwise have benefited from this funding. Yet the principle of ‘a QALY is a QALY’ embodies the value judgement that the first intervention is unequivocally worth funding since it improves total population health by 100 QALYs. Under this framework, severity of illness is not a consideration. Under the equally weighted QALY formula, improving an individual’s health by 0.1 QALY, say, would be equally valued, whether the improvement was to change the patient’s health state from 0.2 to 0.3 or 0.9 to 1.0 QALY, other things equal. The standard health economic definition of ‘efficiency’ in terms of sum total population health therefore embodies a number of strong value judgements about equity (Nord, Reference Nord1999).
Four approaches to incorporating equity considerations
We discuss four broad methodological approaches that analysts may take to explicitly address equity considerations within economic evaluation:
Review of background information on equity
Health inequality impact assessment
Opportunity cost analysis of equity
Equity weighting of health outcomes
The first approach is the most modest: it does not aim to generate new quantitative evidence, but merely to clarify the equity considerations at stake and to review background information that decision-makers may find helpful. For example, in relation to concerns about a particular health inequality, decision-makers may be interested to know about the existing patterns and causes of the health inequality in question, about stakeholder views on the relative importance of reducing this particular health inequality (as compared with other beneficial uses of scarce resources), and about the effects on health inequality of related interventions in other settings. Collating such background information together in one place is relatively cheap and easy to do; it requires no new methodologies or datasets; and would be a useful first step towards incorporating equity concerns into economic evaluation.
The second approach, health inequality impact assessment, aims to generate quantitative evidence about the impact the intervention is likely to have on a particular health inequality (or set of inequalities). This may take the form of data on how the effectiveness and/or cost-effectiveness of the intervention are likely to vary between different equity-relevant sub-groups of the population (such as socio-economic status, geographical location, ethnicity, age, gender, and so on). Tugwell et al. (2006) have proposed a method for this, using epidemiological studies.
Such data could in principle be generated using standard evaluation methods. The key methodological difficulty lies in the paucity of primary effectiveness data on many equity-relevant characteristics. Clinical trials and other primary studies are typically designed and powered to detect average effects across the whole sample, not sub-group effects. Furthermore, although data on some equity-relevant patient characteristics are routinely collected and reported in primary studies, data on many other equity-relevant characteristics are not – in particular, socio-economic status and ethnicity. A second methodological challenge lies in the selection of ‘equity-relevant’ characteristics to examine, from the many possible characteristics that may be of interest to decision-makers. There is a cost attached to collecting, analysing and disseminating additional sub-group data, and so hard choices have to be made about which characteristics are likely to be most important in which contexts.
A more ambitious form of health inequality impact assessment would involve simulation modelling of the magnitude of any reduction in health inequality likely to arise were the intervention to be implemented. This simulation modelling would combine data on existing patterns of health inequality with data on sub-group (cost-) effectiveness of the intervention. Simulation work of this kind would raise important methodological issues about the choice of which health inequality index to use, since different indices can generate quite different results (e.g. Wagstaff et al., Reference Wagstaff, Paci and van Doorslaer1991; Low and Low, Reference Low and Low2006).
The third approach, opportunity cost analysis, would aim to estimate the opportunity cost of a particular equity consideration in terms of population health sacrifice – for example in terms of QALYs forgone by pursuing the equitable option compared with the QALY maximizing option. Every departure from a strict QALY maximizing cost-effectiveness approach on grounds of equity has an opportunity cost in terms of sum total QALYs forgone. The size of that opportunity cost is a test of how important that equity consideration is deemed to be. This approach can be implemented using standard methods of cost-effectiveness analysis, simply by comparing the cost-effectiveness and total QALY impact of alternative programmes that satisfy a particular equity principle to a greater or lesser extent. For example, it can address equity considerations such as a fair innings objective of equalizing lifetime health by comparing the health maximizing option with an option resulting in a more equal distribution of lifetime health. The opportunity cost is calculated as the difference in benefit from choosing the more equitable option. The opportunity cost approach can also be implemented in a more full blown and systematic manner using methods of linear programming (Epstein et al., Reference Epstein2006).
One important methodological advantage of this approach is that it is flexible, and can be used to address any kind of equity consideration – not just concerns about reduction of health inequality. In particular, it can examine considerations of procedural justice – such as anti-discrimination principles – by formulating them as equity constraints. For example, during the 1990s the UK Standing Medical Advisory Committee advised local health authorities against adopting a racially selective policy on screening for sickle cell anaemia – which is more prevalent in certain ethnic minority groups – even though this may have been the most cost-effective strategy (Sassi et al., 2001). Imposing the equity constraint of non-discrimination was thus shown to impose a health sacrifice in terms of cost-effectiveness.
A disadvantage of the opportunity cost approach, of course, is that this approach only looks at the cost of the equity consideration, not the benefit. It only measures the equity–efficiency trade-off that is implied by a particular decision (a factual matter), it does not value the trade-off that policy-makers ought to make (a moral matter). That is, it does not help the decision-maker decide how large a sum total sacrifice (in terms of health and/or non-health benefits) is worth making in order to pursue a particular equity consideration. This runs the risk of lack of transparency and inconsistency across decisions, since decision-makers are then free to make implicit judgments about how much QALY sacrifice is worth making in pursuit of a particular equity principle – and to vary those judgments from one decision to another without giving any explicit justification.
Unlike the opportunity cost approach, the fourth and final approach, equity weighting analysis, does aim explicitly to value health inequality reduction (or other equity concerns) and to guide the decision-maker about how much sum total QALY sacrifice is worth making in order to reduce health inequality. There is a small but growing body of methodological work in this area, which develops methods for setting quantitative ‘equity weights’ on health gains accruing to different people in different circumstances (Nord, Reference Nord1995, Reference Nord1999; Dolan et al., Reference Dolan2005). These methods remain at a developmental stage and have not yet been applied in practice.
The essential idea is to set weights on health gains (e.g. QALY gains) accruing to people with different equity-relevant characteristics, based on values elicited from a relevant stakeholder group (e.g. the general public, policy-makers). Therefore, unlike the opportunity cost approach, the measure of benefit is adjusted for equity considerations. The value elicitation techniques applicable in this area are similar to those used in the health measurement and valuation literature: the basic idea is to confront respondents with ‘hard choices’ that require explicit quantitative trade-offs between competing objectives. One major difference, however, is that value elicitation questions generally ask respondents to adopt a ‘citizen’ or ‘policy-maker’ perspective, involving trade-offs between the health of groups of people, rather than an individual or consumer perspective involving trade-offs between ones’ own health and/or wealth (Dolan et al., Reference Dolan2003).
The three main value elicitation techniques used to date are ‘Willingness to Pay’ (WTP), ‘Person-Trade-Off’ (PTO), and ‘Discrete Choice Experiment’ (DCE). DCE involves a series of ‘hard choices’ between interventions benefiting groups that vary in terms of multiple specified attributes (e.g. social class, ethnicity, smoking status, health gain, initial health, etc.). Regression analysis is then used to derive the ‘weight’ on each attribute implied by the pattern of responses. PTO involves a ‘hard choice’ between two interventions benefiting different groups of the same size that vary by a single attribute, usually the health condition that they are suffering from (Nord, Reference Nord1995). The number of individuals in one group is then varied until the respondent reaches a point of ‘indifference’ between the two interventions (i.e. values them roughly equally). The relative number of individuals left in each group is then an indication of the weight the respondent places on this attribute. Finally, the WTP approach involves eliciting individual willingness to pay amounts for interventions that benefit particular groups, which may or may not include the individual respondent (Donaldson, Reference Donaldson1999). These groups may vary across one attribute only (as in PTO) or across multiple attributes (as in DCE); although in the latter case it is harder to tease out the equity weight associated with each attribute.
There have been a number of studies aimed at eliciting equity weights from members of the general public (e.g. Nord, Reference Nord1999; Dolan et al., Reference Dolan2005; NICE, 2006b, report on health inequalities http://www.nice.org.uk/page.aspx?o=410781). Despite considerable noise and variation in the results, two consistent findings seem to emerge. First, that the public value health gains to children more highly than health gains to adults, and, second, that the public value health gains to individuals currently suffering from severe illness more highly than health gains to individuals currently in relatively good health. In terms of reduction of health inequalities, early findings indicate that people think that social class inequalities are more inequitable than those by smoking status, with inequalities between the sexes somewhere in between (Williams et al., Reference Williams, Tsuchiya, Dolan, Smith, Ginnelly and Sculpher2004). Findings with respect to other health inequalities and other personal characteristics, such as degree of responsibility for own ill-health, tend to be more equivocal. Although quantitative equity weights can be calculated from these data, at present the data base is not comprehensive enough to do this credibly for current policy purposes (Williams and Cookson, Reference Williams and Cookson2006).
Practical ways forward
When is explicit evaluation of equity likely to be most worthwhile?
Careful choices will have to be made about when to conduct explicit economic evaluations of equity concerns, given their likely cost and methodological difficulty. In some cases, there may be no need for economic evaluation explicitly to address equity considerations. For example, imagine that a straightforward yes or no decision is required about a well-specified public health intervention that targets disadvantaged individuals or communities. If this intervention is demonstrated to be cost-effective, then it is likely also to reduce health inequalities between advantaged and disadvantaged groups. Under these circumstances, the decision-maker may not need to know by how much health inequality is reduced, or how far the general public would value a health inequality reduction of this kind. It may simply be obvious that the intervention is worthwhile, without having to undertake costly exercises in reviewing background information, assessing health inequality impacts, and analysing opportunity costs and equity weighting health gains.
By contrast, there would be a more pressing need for explicit approaches to addressing equity considerations if the public health intervention were found not to be cost-effective. Then the difficult questions arise. By how much does this intervention reduce health inequality? How large is the sacrifice to sum total population health if this intervention were to be funded? How highly does the general public value this health inequality reduction in comparison to the sacrifice to population health?
These difficult questions can also arise when, as is often the case, multiple options are on the decision-making table (e.g. different modes of programme design, delivery, and targeting). Some of the programme options may be more cost-effective than others and/or reduce health inequality by more than others. In such cases, explicit equity analysis may again prove fruitful to choose among the multiple options.
Explicit evaluation of equity considerations may also be worthwhile in relation to screening and health promotion programmes, and other preventive public health interventions whose effectiveness depends crucially on eliciting some positive action or behaviour change on the part of the intended beneficiaries. Although such programmes may be designed primarily to increase sum total population health, they may also have the unintended and unfortunate consequence of increasing health inequalities if take-up is lower among more disadvantaged ‘hard-to-reach’ populations. There is considerable evidence that this is a widespread phenomenon – for example, in relation to breast and cervical cancer screening (Katz and Hofer, Reference Katz and Hofer1994, Alexander et al., Reference Alexander1999), immunization for diphtheria and pertussis (Reading et al., Reference Reading1994), the ‘Sure Start’ programme for children and families in deprived areas (National Evaluation of Sure Start, 2005), educational outcomes from the US children’s TV programme ‘Sesame Street’ (Berk et al., Reference Berk1985), anti-smoking and smoking cessation interventions (Secker-Walker et al., Reference Secker-Walker2002, Rice et al., Reference Rice1994), a Canadian school-based bicycle helmet promotion (Parkin et al., Reference Parkin1993), child abuse prevention programmes in New Zealand and Ireland (Briggs and Hawkins, Reference Briggs and Hawkins1994, MacIntyre and Carr, Reference MacIntyre and Carr1999), and an inner city visual screening programme for preschool children in Glasgow (Williamson et al., Reference Williamson1995).
Explicit evaluation may help to re-design such initiatives, so that the benefits are spread more evenly between different social groups. For example, a UK policy introduced in 1990 of explicit financial incentives for GPs to undertake screening for cervical cancer resulted in greater uptake among women in relatively advantaged communities, hence increasing health inequalities (Sassi et al., Reference Sassi, Archard and Le Grand2001). Based on their own retrospective equity analysis, Sassi and colleagues suggest that a more equitable policy would be to offer stronger financial incentives in more deprived areas, possibly financing this by increasing the screening interval to five years rather than three years.
Which approaches are likely to be most worthwhile?
Since the four approaches have so rarely been used in economic evaluation studies, let alone used to inform an actual policy decision, it is premature to make any attempt to formulate advice on which approach is likely to be most helpful in which context. Instead, what is required is a period of development and piloting of all four approaches, so that decision-makers and other stakeholders can provide feedback as to which approaches are most helpful in practice. For the most part, the technical tools required to undertake explicit equity analysis are more or less extensions of existing statistical and modelling methods of economic evaluation. For example, health economists are well accustomed to performing sub-group analyses of cost-effectiveness on characteristics such as stage of disease.
Tools exist for mathematical programming analysis of opportunity cost. For the first three approaches, it may be possible and worthwhile to go directly to the ‘field testing’ stage and to pilot some of these approaches in the context of real policy decisions by performing equity analysis within relatively short time frames.
There may be grounds for additional methodological work, however, before piloting the equity weighting approach in real-world cases. Most economists would agree that the first three approaches all have merit, and that a comprehensive economic evaluation of equity considerations would contain all three elements. However, there is some disagreement among the profession about the merits of equity weighting analysis. The main problem lies with (i) identifying what the key elements of equity status are and (ii) the quality and comprehensiveness of the evidence base for determining the magnitude of equity weights, and (iii) philosophical disagreements about the role of economic evidence to inform political decision-making. In addition, Tsuchiya and colleagues argue that there are so many potential dimensions of equity that use of equity weights could substantially complicate the analysis (Tsuchiya et al., Reference Dolan and Tsuchiya2007). This is because ideally equity weights would not simply be applied in a blanket fashion at the aggregate level of whole disease areas or programmes – for example, weighting all health gains from all anti-cancer programmes twice as highly as health gains from programmes in other disease areas. Rather, weights would be applied to the individuals targeted by those programmes – for example, according to their socio-economic status, age, severity of illness or other equity-relevant characteristics. This means that the overall effect of applying equity weights to the benefits of a particular programme requires a process of computation, which depends on the observed characteristics of the individuals within the programme, and that the result may not be immediately obvious before that computation is performed. The process of computing weights for each individual in the programme may become complex, especially as some individual characteristics – for example, age – will or may change over time.
In relation to the philosophical disagreements, the case for equity weighting is that transparency and consistency in public policy-making would be better served if public policy-makers were forced to be explicit about all the value judgements that go into their decision-making, including quantitative value judgements about equity–efficiency and equity–equity trade-offs (Williams and Cookson, Reference Williams and Cookson2006). Economists have long made the case for explicitness about value judgements concerning efficiency – in particular, value judgements about the trade-offs between different dimensions of quality of life and length of life as embodied in the QALY approach. There is therefore a case for extending this same principle of explicitness to value judgements concerning equity.
Recent interest in social value judgements is illustrated by development of The Citizens Council of the National Institute for Health and Clinical Excellence (NICE). The Council, made up of 30 ordinary members of the public, reflecting the age, gender, socioeconomic status, and ethnicity of people in England and Wales, has been asked by the NICE Board to consider and report on issues such as clinical need, and whether there are circumstances in which age should be taken into account when making decisions. In 2004 NICE asked the Council to consider the issues related to the funding of orphan drugs, ‘to advise on whether or not the NHS should be prepared to pay premium prices for drugs to treat patients with very rare diseases’ (http://www.nice.org). To our knowledge little work has been undertaken to explore approaches to obtain decision-makers’ value judgements on equity, although one example was found (Dolan and Tsuchiya, Reference Dolan and Tsuchiya2007). It remains unclear as to whose equity values count. In addition, it remains unclear whether the cost-effectiveness threshold value should or has been varied according to the equity status of the affected individuals.
A number of economists prefer that analysts focus on efficiency objectives rather than equity objectives. Sassi et al. (2001) distinguish ‘positive’ versus ‘normative’ approaches to addressing equity considerations. The first three approaches discussed here are ‘positive’; whereas the equity weighting approach is ‘normative’ – it seeks to value equity–efficiency trade-offs (e.g. to stipulate how much total health should be sacrificed to achieve a given health inequality reduction), not just to measure them. Sassi and colleagues are pessimistic about the prospects for ‘normative’ approaches, and argue that decision-makers should be presented with factual information about health inequality impacts and then left to make their own value judgements about how much weight to place on them. In a similar vein, Culyer (2006) distinguishes ‘deliberative’ versus ‘algorithmic’ approaches to making decisions. The former aims to specify a suitable process of deliberation and consultation that will help the decision-maker to judge what weight to place on particular considerations, whereas the latter aims to specify a suitable algorithm that will explicitly determine what weight to use. Culyer advocates an ‘algorithmic’ approach to valuing health-related quality of life but a ‘deliberative’ approach to valuing equity considerations. He argues that equity considerations (unlike quality of life considerations) are too thoroughly context-sensitive to be adequately addressed using a pre-defined algorithm.
Even if one accepts the case against ‘algorithmic’ approaches to equity considerations, however, there may still be an important role for equity weighting research. Data emerging from value elicitation studies aimed at generating equity weights need not necessarily be employed in an algorithmic fashion. That is, equity weight data may merely be used to help inform decision-makers about what equity considerations are generally valued most highly by the general public, while leaving plenty of room for deliberation and judgement on the part of the decision-maker about how much weight to place on these data in particular circumstances.
Conclusions
Given the importance of equity as a policy objective in public health interventions, the impact on equity, as well as efficiency, of these interventions needs to be explored in economic evaluations. In this paper, four approaches for conducting equity analysis have been outlined. Three of the approaches could be added, with little additional cost. The fourth, explicit equity weighting of outcomes, requires further methodological research.
Acknowledgements
This paper arises from work undertaken as part of the Public Health Research Consortium (PHRC). The Consortium is funded by the Department of Health Policy Research Programme. The views expressed in the publication are those of the authors and not necessarily those of the Department of Health. Information about the wider programme of the PHRC is available from www.york.ac.uk/phrc
The authors acknowledge the contributions of colleagues at the PHRC, as well as Dr Mark Petticrew for sending useful examples of relevant studies, Dr Aki Tsuchiya for commenting on the paper, and Adam Oliver for helpful editorial comments.
While doing this work, Richard Cookson was funded by MRC Health Services Research Special Training Fellowship G106/1145.
