By Karen Davenport
Health care costs continue to rise, with expenditures accounting for nearly 20 percent of the gross domestic product (GDP) in 2020. Innovations in provider payments and benefit designs grounded in the known “value” of different health services may help payers control escalating costs while improving care quality and efficiency. But these strategies might run afoul of other goals, such as health equity, by failing to reflect the needs, values, and preferences of certain patients. This tension is evident as the Department of Health and Human Services’ Office of Civil Rights (OCR) considers whether value assessment methodologies discriminate against protected groups, such as people with disabilities and older adults.
Focus on value
“Value” is a popular buzzword in health insurance and health policy. Insurance companies offer value-based insurance designs (V-BID), developing cost-sharing structures that encourage enrollees to eschew low-value care and seek high-value services. Medicare, Medicaid, and private payers incentivize clinicians and health care facilities to participate in value-based payment models, which typically reward health care professionals and institutions for achieving quality goals and, sometimes, cost savings. And expensive new treatments and services face scrutiny over whether they offer better value to patients and payers compared to the treatments and services available today. To support coverage decisions, benefit designs, and payment methodologies built around the concept of value, professional organizations, academics, and others have advanced a range of value assessment frameworks designed to determine and quantify value in health care. Yet as this work advances, other voices have cautioned that value assessment may perpetuate inequities and discrimination within the health care system.
Recently, OCR indicated in its Notice of Proposed Rulemaking that certain value assessment methodologies may violate Section 1557 of the Affordable Care Act (ACA), which applies civil rights protections to health programs and activities administered by federal agencies or receiving federal financial assistance. Section 1557 prohibits discrimination on the basis of race, color, national origin, age, disability, or sex (including, under the proposed rule, pregnancy, sexual orientation, gender identity, sex stereotypes, and sex characteristics). While OCR did not propose regulatory language specific to value assessment, the preamble requests input on the civil rights implications of value assessment for a range of health insurance-related activities, including utilization management, formulary design, price negotiations, and alternative payment models.
What is value assessment?
Value assessment—sometimes called health technology assessment, and closely related to cost-effectiveness analysis and comparative effectiveness reviews—applies clinical, economic, and other evidence to determine the relative costs and benefits of treatments and services. These analyses inform clinical decisions as well as coverage and payment policies; payers can devise economic incentives that encourage health care professionals and patients to choose effective, high-value services and treatments over lower-value care. With careful assessments in-hand, for example, an insurance plan could decide which high-value services would not carry a cost-sharing requirement, and which low-value services would require increased enrollee cost-sharing. Key elements of value assessment include the inputs of clinical studies and economic analyses, and the data, analytic assumptions, and criteria that serve as the foundation for these studies.
Could analysis and quantification of value be discriminatory?
Value assessment analyses may help curb health care spending and improve health outcomes, but methodologies relying on data, assumptions, and criteria that reflect embedded and implicit biases related to race, ethnicity, disability, sex, and age undermine these worthy goals. Just as clinical algorithms can reflect underlying inequities in the data informing the algorithm, incomplete or flawed analytic foundations can result in biased value assessment results. Trip wires may include the exclusion or underrepresentation of people of color, people with disabilities, or people with chronic conditions within clinical data; the types of metrics and decision criteria used in the value assessment analysis—such as which evidence is included, how is it weighted, and what assumptions lie beneath the analysis; and whether the preferences and perspectives of people who need care are built into the analytic framework.
A metric often used in value assessment, for example, is the Quality-Adjusted Life Year (QALY), which encompasses both the likely additional years of life a service or treatment may confer and the likelihood that the service or treatment will restore the patient’s health and function. QALY scores enable comparison of various treatments and services for a range of conditions, with one QALY equal to one additional year of life in perfect health. These comparisons provide a foundation for coverage decisions and payment incentives tied to value. The data that lies underneath the QALY calculation, however, can be compromised in multiple ways. First, information on likely outcomes—whether patients experience restored health and function—may be drawn from clinical studies that did not enroll or fully represent people with disabilities, people with chronic illnesses, people of color, or people experiencing health disparities, including those that are driven by social determinants of health. Second, the QALY rubric draws on a quality-of-life score that is informed by surveys that assess the general public’s attitudes toward life with disabilities or chronic illness in comparison to perfect health. These results, and subsequently the metrics that shape what is viewed as a “valued” service or treatment, can therefore be influenced by publicly held prejudices and stereotypes related to disability, illness, and longevity.
Advocacy groups perceive discrimination in value assessment methodologies
Many disability rights organizations outlined their concerns with value assessment methodologies in their comments on the proposed rule. Several groups cited the role of the QALY metric in value assessment, with some arguing that this metric is grounded in implicit bias about life with disabilities, chronic illness, or the effects of aging, thus devaluing treatments and services provided to people with disabilities. Other stakeholders noted that the QALY metric’s emphasis on time —that is, the additional years of life a treatment or service may permit—discriminates against older adults since this measure is in part a function of the patient’s age. One comment from a coalition of consumer organizations noted that value assessments “are powered to show results for a patient population that is largely white, middle-aged, non-disabled, and male” while raising concerns that the analytic studies informing value assessment typically rely on population-level averages instead of data for underserved communities.
Some commenters suggested alternative methodologies, such as multi-criteria decision analysis (MCDA), arguing that this approach better captures the complexity of coverage decisions. MCDA applies a range of criteria to potential alternatives, according to stakeholder and decision-maker preferences. Other commenters endorsed the inclusion of patient perspectives and preferences in the conceptual framework and methods that guide this analysis.
Potential implications for payers
If OCR determines that certain value assessment methodologies violate the non-discrimination provisions of the ACA, what would this mean for health insurance plans, public programs, and their enrollees? Comments from major payers, including Kaiser Permanente, UnitedHealth Group, CVS Health, and Cigna, did not share insights about how value assessment supports key insurance functions, or how changes to the underlying methodology of these analyses would affect their business. Nevertheless, payers that use value assessments to design utilization management programs restricting access to low-valued services, develop formularies with step-therapy requirements or higher cost-sharing for low-valued medications, or establish provider payment models rewarding the delivery of high-value care may see these coverage and payment strategies upended by future OCR rulemaking. Similarly, payment demonstrations sponsored by Center for Medicare & Medicaid Innovation (CMMI) at the Centers for Medicare & Medicaid Services (CMS), such as the Medicare Advantage Value-Based Insurance Design initiative, may also be affected by OCR’s decisions. OCR’s approach to discriminatory clinical algorithms in the proposed rule—that is, requiring covered entities to determine whether the algorithm is inherently biased and adjusting their use of the algorithm to ensure their decisions are not discriminatory—may foreshadow an approach OCR could also use on value assessment. Alternatively, OCR could encourage or require plans to use value assessment frameworks that meaningfully integrate patient and caregiver perspectives into the overall analysis.
Takeaway
One of the challenges of policymaking is reconciling conflicting policy goals. In this case, the promise value-based payments and coverage decisions hold for controlling health care spending and improving outcomes may be compromised by bias in the analytic underpinnings of value assessment. These approaches could therefore have a different—and potentially damaging—impact on population groups that are protected by the nondiscrimination provisions of the ACA. Should OCR find that value assessment methodologies discriminate on the basis of race, color, national origin, age, disability, or sex, it will need to find a delicate balance between these important objectives.