As an epidemiologist and physician, I’ve long been concerned about how health interventions for marginalized communities are evaluated. A new study explores that concern and offers some solutions for making cost-effectiveness analyses work for everyone.
Many of my patients live in San Francisco’s Bayview neighborhood, which is disproportionately Black, cut off from the rest of the city by freeways, and has a toxic waste problem from a nearby shipyard. The neighborhood often fails to secure new investments required to maintain health clinics or develop mixed-income housing or grocery stores due, in part, to cost-effectiveness analyses that show investment in the community won’t drive returns.
For many people who struggle with the difficult decision between paying for medications or covering essentials like food, rent, and transportation, the consequences of these analyses and the investment choices they inform can have a profound impact on their health and well-being.
In a peer-reviewed article just published in the journal Health Affairs, Atheendar Venkataramani, Dean Schillinger, and I make the case that standard approaches to cost-effectiveness analysis in health economics may inadvertently perpetuate health disparities. This issue deserves the attention of not just academics and researchers but also policymakers, payers, and providers who care about improving health equity.
Cost-effectiveness analyses are a cornerstone of decision-making. They help determine which interventions provide the most benefit per dollar spent. But the current methods have a critical flaw: they tend to undervalue interventions that would primarily benefit disadvantaged groups.
This problem stems from three main factors:
First, marginalized populations often face higher risks of dying from multiple possible causes. This phenomenon, known as competing risks, can make interventions targeting a specific disease appear less cost-effective for these groups even if the intervention is equally or more effective for single-risk groups. If a person doesn’t die from one disease, they can still die from another and, according to standard cost-effectiveness analyses, the benefits from preventing the first disease are diminished.
Second, because disadvantaged individuals are often covered by Medicaid (which pays less to physicians and hospitals than Medicare or commercial insurance) or don’t have insurance at all, they tend to have lower baseline health care expenditures. This can make interventions designed to prevent costly complications appear to save less money among marginalized populations than they would among privileged populations, who often have more expensive insurance plans.
Third, lost work productivity is a common element of cost-effectiveness calculations. But because of wage disparities, interventions that prevent lost work among marginalized groups may appear to have lower economic benefits for society because they have less economic impact than among more privileged groups (put bluntly, in economic calculations, the health and lives of marginalized populations are literally less valued).
To illustrate why these assumptions are problematic, here’s a real-world example: evaluating a program that helps people afford nutritious food and participate in physical activity to reduce their risk of diabetes. This intervention might be especially beneficial for Black Americans, who face a higher risk of diabetes. A standard cost-effectiveness analysis following the current national guidelines, however, would likely conclude this intervention is less cost-effective for this population for several reasons. Black Americans face higher competing risks of dying from other causes like asthma before developing diabetes and its many complications. They’re also more likely to be uninsured or covered by Medicaid, so preventing complications appears to save less money. And their lower average wages mean preventing lost work productivity appears less economically beneficial.
Deeming an intervention that could significantly benefit a disadvantaged population as too costly (or insufficiently cost-effective) to implement not only fails to address existing health disparities but may actually contribute to their persistence.
To address this issue, my colleagues and I propose several approaches:
First, cost-effectiveness analyses shouldn’t assume that current health disparities will exist indefinitely. This makes it possible to consider scenarios where inequities are reduced over time. Second, calculation methods that place higher value on health gains for disadvantaged groups should be explored, recognizing that there is potentially greater value to spending a dollar to help those who are facing the most adversity over and above providing the same dollar to those who already have extensive resources. Cost-effectiveness analysts must also broaden their lens. This means accounting for how interventions might affect multiple health risks simultaneously, rather than focusing on a single disease in isolation. For example, environmental waste in Bayview is potentially contributing to multiple diseases that range from various cancers to respiratory disease, not just one condition. Opening the lens also entails incorporating broader societal benefits beyond health care costs and productivity, such as improving education outcomes or community well-being, into cost-effectiveness analyses.
Practitioners of cost-effectiveness analyses need to have open discussions about the ethical implications of current approaches to this task. Baking existing disparities into an analysis risks perpetuating a system that has historically undervalued the health of marginalized communities. This goes against principles of health equity and social justice that many in the medical and public health spheres hold dear.
Health researchers and policymakers have a responsibility to ensure that their analytical tools promote equity rather than hinder it. Evolving and fine-tuning methods can help ensure that cost-effectiveness analyses become a force for reducing health disparities, not reinforcing them.
This won’t be an easy task. It requires rethinking long-standing economic assumptions, challenging highly entrenched and powerful academics who have made their careers publishing studies using existing methods, and asking honest questions about the ethical implications of studies and methods that often don’t include the people who are most affected by them. But it’s a necessary step if we’re serious about creating a more equitable health care system.
In the end, this isn’t just about improving analytical methods. It’s about recognizing the full value of every life and every community. Doing so can lead to better decisions that serve all members of society, especially those who have been historically underserved. It’s time for research approaches to catch up with aspirations for health equity.
Sanjay Basu, M.D., Ph.D., is a primary care provider, epidemiologist, and head of clinical at Waymark, a public benefit company dedicated to improving access and quality of care for people receiving Medicaid benefits. He provides primary care, substance use, HIV, and hepatitis C treatment to patients at San Francisco’s HealthRight360 Integrated Care Center for marginally housed adults.