Using Distributional Cost-Effectiveness Analysis to Balance Equity and Efficiency

For decades conventional cost-effectiveness analyses (CEAs) have evaluated the trade-off between health benefits and healthcare costs without acknowledging who experienced the benefits or incurred the costs. The First Panel on CEA stated in 1996 that notions of equity should be highlighted in decisions about health policy along with conventional CEA rather than embedded in the CEA calculations themselves.¹ The 2^nd Panel on CEA in 2016 endorsed the use of methods that directly quantify the equity-efficiency trade-off.² Distributional CEA (DCEA) has emerged as the formulation that attaches more weight to interventions that reduce health disparities and less weight to interventions that exacerbate health disparities.³

In 2014 the ACC/AHA, which recognized that many interventions identified as high value using conventional CEA also improve equity, released a joint policy statement that endorsed conventional CEA methods to weigh the trade-offs between costs and cardiovascular health outcomes.⁴ In 2021, the ACC added health equity to its broad mission of improving population health.⁵ To balance efficiency and distributional equity when there are trade-offs between these objectives, decision-makers who adopt a population-based perspective (e.g., societal, health care system, hospital) can use DCEA.

DCEA has three steps: 1) define disparities among sub-groups, 2) quantify the subgroup-specific impact of competing strategies on health and cost outcomes and identify whether an equity-efficiency trade-off exists, 3) use an equity weight and social welfare function to weigh the trade-off and make a strategy recommendation.³ Researchers from the University of York have provided tutorials, examples, and training materials to help analysts carry out DCEA in practice.⁶ In this Viewpoint, we focus on the intuition and conceptual nuances of DCEA to highlight its potential role in cardiovascular research and policy decision-making.

Declining marginal utility, a concept from economic theory, predicts that health and cost changes have greater impacts when they are experienced by those who are worse off (such as those with lower income) compared to those who are better off (Figure 1). When the equity weight is 0 in a DCEA, every additional unit of health benefit, such as a quality-adjusted life year (QALY), is treated equally as an outcome, regardless of income level or baseline healthcare status. The same is true for every unit of cost. As a result, conventional CEA uses a linear function (Panel A) to describe the relationship between social welfare and health gains, which is often summarized as by the statement that, “a QALY is a QALY is a QALY.” However, when the equity weight is set to >0 in a DCEA, more importance is given to health gains and cost savings for groups that are worse-off than those that are better off, describing nonlinear relation, and the degree of nonlinearity depends on the value of the equity weight itself (Panels B and C).

Figure 1. Social Welfare as a Function of Cumulative Health (or Cost) Outcomes and Equity Weight Value.

Social welfare functions are shown for conventional cost-effectiveness analysis (CEA) and for distributional cost-effectiveness analysis (DCEA). Conventional CEA uses an equity weight of 0 and a linear function (Panel A) to show that each marginal change, such as one unit of health gain or one unit of cost savings, is equally valuable in a population, regardless income level or baseline healthcare status. When DCEA uses an equity weight value greater than 0, such as X (Panel B) or higher values such as Y (Panel C), more importance in the social welfare function is attached to outcomes experienced by groups that are worse-off at baseline (lower values on the horizontal axis) than those that are better off (higher values on the horizontal axis). The degree of non-linearity of the function is driven by the numerical value of the equity weight

What equity weight value should be used in DCEA? To answer this question, we have used the example of the cost-effectiveness threshold value. Most readers will be familiar with the range of $50,000 to $150,000 per QALY used by the ACC/AHA as conventional cost-effectiveness thresholds in conventional CEAs.⁴ For the choice of equity weight values (represented by X and Y in Panels B and C, respectively), some studies have empirically derived these values from discrete-choice surveys designed for specific settings. In one such survey a weight of 11.0 emerged for social deprivation quintiles in the UK,⁷ but most settings lack empirically estimated values for equity weights for specific health disparities that, for example, are disease-based or race/ethnicity-specific. In these situations, DCEA can quantify a threshold equity weight that can be compared to some standard or range. In the US, equity weights are often in the range of 0.5-3.0.⁸ Such a range can be used to interpret threshold equity weights calculated from DCEAs. Calculating the actual threshold equity weight value is beyond the scope of this paper, but in principle it is solved by setting 2 strategies, with one better on the equity dimension and the other better on efficiency, equal to each other in terms of social welfare in the DCEA. In practice, a low threshold value close to 0 implies that relatively little weight is needed to favor the equity-improving but cost-ineffective intervention, whereas a higher threshold value of, say, >3.0 in the US allows the more cost-effective intervention to seem acceptable when there is an equity-efficiency trade-off. As a drawback, DCEA cannot exclude subjectivity and bias in the choice of an equity weight value, but it provides quantitative bounds that can act as guardrails against extreme preferences in the same way that cost-effectiveness thresholds do in conventional CEA.

We previously published a DCEA evaluating gene therapy for sickle cell disease.⁹ Using a cost of $2.45 million, we showed that gene therapy was conventionally cost-ineffective compared to standard-of-care (incremental cost-effectiveness ratio of $176,000/QALY), but the threshold equity weight was 0.9, which was at the lower end of conventional values (0.5-3.0) used in the US.⁸ We concluded that gene therapy could be viewed as favorable per DCEA standards, signaling to decision-makers who value health equity that it could be reasonable to adopt this conventionally cost-ineffective intervention.

Investigators, clinicians, and policy decision-makers in cardiovascular medicine recognize that many health disparities exist across sex, geography, race and ethnic subgroups.¹⁰ Because addressing these disparities will require economic resources, interventions that improve equity will not always save the system money. Conventional CEA evaluates whether health gains are worth the costs required to achieve them. DCEA quantitatively incorporates distributional equity as an additional objective. DCEA does not, however, remove the need for qualitative deliberation when decision-makers confront equity-efficiency trade-offs. Decision-makers will need to choose an equity weight value (or interpret threshold equity weights) to use DCEA, which becomes a normative choice. Unfortunately, capturing non-quantifiable aspects of health equity such as autonomy, non-maleficence, and dignity is not possible in DCEA. We recommend using DCEA to weigh the equity-efficiency trade-off across quantifiable dimensions and recognize the importance of transparent, rigorous deliberations while allocating limited resources among the interventions that improve cardiovascular health.