Advancing Health Equity Through The CMS Innovation Center: First Year Progress And What’s To Come

In October 2021, the Centers for Medicare and Medicaid Services (CMS) Innovation Center released its new 2030 vision: “A health system that achieves equitable outcomes through high quality, affordable, and person-centered care.” To achieve this vision, the CMS Innovation Center outlined five strategic objectives, one of which is to “Advance Health Equity.” CMS defines health equity as “the attainment of the highest level of health for all people, where everyone has a fair and just opportunity to attain their optimal health regardless of race, ethnicity, disability, sexual orientation, gender identity, socioeconomic status, geography, preferred language, or other factors that affect access to care and health outcomes.”

To meet this objective, the CMS Innovation Center launched a new health equity initiative in 2022, proposing to: 1) develop new models and revise existing models to promote and incentivize equitable care; 2) increase participation of safety-net providers; 3) increase collection and analysis of equity data; and 4) monitor and evaluate models for health equity impact. The CMS Innovation Center has made meaningful progress in each of these areas, which we believe will help mitigate health inequities within CMS’s beneficiary populations. This article provides a one-year look back at what we have accomplished and describes additional areas of focus moving forward.

1) Develop New Models And Revise Existing Models To Promote And Incentivize Equitable Care

In the past year, three revised models and one new model have been announced, all of which have a prominent focus on addressing health equity. One revised model is the Medicare Advantage (MA) Value-Based Insurance Design Model, which builds upon the statutory authority permitting CMS to waive the uniformity requirements with respect to supplemental benefits provided to chronically ill enrollees. CMS is using this authority to permit MA plans to target and provide supplemental benefits to beneficiaries with chronic illnesses and also allow plans to target beneficiaries with health-related social needs. Participating plans can provide targeted supplemental benefits such as food, transportation, and housing assistance; reduced copayments; and rewards and incentives in connection with participation in activities that focus on promoting improved health, preventing injuries and illness, and promoting efficient use of health care resources.

Beginning in July 2023, we will collect detailed information from participating plans about the supplemental benefits provided, so that we can assess their use and effectiveness in improving care and outcomes. In addition, model participants will be required to develop and submit health equity plans to ensure they understand the needs of their beneficiary populations, identify any disparities within these populations, and develop and implement interventions to remedy these disparities; and form networks of Medicare hospice providers who have a history of serving underserved communities, have strong relationships with their local communities, actively collaborate with organizations that may help meet the health-related social needs of patients, and demonstrate cultural competency.

The second revised model is the ACO Realizing Equity, Access, and Community Health (ACO REACH) Model, which is a revamped and renamed version of the former Global and Professional Direct Contracting Model. This model, which launched in January of this year, also requires development and submission of health equity plans, as well as sociodemographic data collection and screening of beneficiaries for health-related social needs. In addition, the ACO REACH Model is testing a novel “health equity benchmark adjustment,” to remove the disincentive for providers serving disproportionate numbers of underserved beneficiaries to participate in the model. Underserved beneficiaries are defined in ACO REACH as those dually eligible for Medicare and Medicaid (dually eligible individuals) and living in areas characterized by a high area deprivation index (ADI). Such populations are more medically and socially complex and may require additional staff and more intensive services for the provision of high-quality care.

We also revised the Maryland Primary Care Program, a component of the third model, the Maryland Total Cost of Care Model. In January 2022, the Maryland Primary Care Program began to offer a “Health Equity Advancement Resource and Transformation (HEART)” payment to provide additional support to primary care participants and federally qualified health centers (FQHCs) furnishing enhanced care management services to socioeconomically disadvantaged and complex patients with high clinical risk in counties with high ADI relative to other counties in the state.

In addition to these three revised models, we announced a new model, Enhancing Oncology Model (EOM), last year; it is expected to begin in July 2023. The EOM focuses on several cancers that disproportionately affect underserved populations, including multiple myeloma and colorectal cancer, both of which disproportionately affect African Americans. The CMS Innovation Center plans to embed health equity in the EOM by requiring model participants to collect sociodemographic data, screen beneficiaries for health-related social needs, and develop health equity plans as part of use of data for continuous quality improvement efforts. In addition, model participants serving dually eligible individuals will receive an enhanced per beneficiary per month “Monthly Enhanced Oncology Services (MEOS)” payment to support care management and coordination, and address potential disincentives for caring for the underserved. This enhanced MEOS payment for dually eligible individuals will not count against a model participant’s cost-of-care calculations, which are used to determine whether model participants are meeting cost targets.

2) Increase Participation Of Safety-Net Providers

Increasing safety-net health care provider participation in value-based care is a key component of our strategy to expand the diversity of beneficiaries served through CMS Innovation Center models. To this end, we have engaged multiple safety-net provider stakeholders through one-on-one interviews and roundtable listening sessions to identify strategies for addressing barriers to safety-net provider participation in CMS Innovation Center models. We also looked inward and conducted qualitative analyses of our model application requirements to identify specific barriers to safety-net provider participation. Furthermore, we are working across CMS and with the Health Resources and Services Administration to share our findings and operationalize strategies for including more providers that serve Medicaid beneficiaries and FQHCs in our models.

These engagement and analytic efforts have been critical in increasing our understanding about model elements that may encourage or hamper participation by safety-net providers, such as level and timing of funding support, requirements for downside risk, quality measurement systems, data/technology capabilities, and provision of technical assistance. Importantly, we are promoting multipayer alignment as much as possible to harmonize model elements across Medicare, Medicaid, and commercial payers. We anticipate that multipayer alignment will facilitate greater safety-net provider participation in models as well. These activities may already be having an impact—we are encouraged by the more than 700 FQHCs participating in ACO REACH. To encourage even greater participation in future models, we have begun to conduct real-time application analyses to track inclusion of providers who serve dually eligible individuals.

3) Increase Collection And Analysis Of Equity Data

Collecting and analyzing data to monitor and evaluate CMS Innovation Center models is critical, and evaluation findings inform future model design. CMS relies upon data imputation to estimate beneficiaries’ demographic backgrounds and uses data proxies, such as dual eligibility for Medicare and Medicaid, to measure “underservedness.” At the population level, these can be useful short-term approaches until we have sufficient self-reported “gold standard” data that can support more robust analyses, including individual beneficiary-level analyses.

The ACO REACH Model requires accountable care organizations (ACOs) to collect sociodemographic data, and health-related social needs data when feasible, in accordance with the voluntary US Core Data Interoperability standards promulgated by the Department of Health and Human Services (HHS) Office of the National Coordinator for Health Information Technology. Over time, such data will help enable us to better stratify and evaluate model outcomes for beneficiary populations defined by factors such as race, ethnicity, gender, geography, disability, sexual orientation, and gender identity.

Collecting and reporting such data can be challenging for some ACOs and their providers who do not all have the same electronic health record systems, leading to difficulties capturing this data within existing provider workflows and extracting and compiling it for the entire ACO. As such, we are pilot-testing mechanisms to facilitate data collection by model participants that minimize administrative burden. These mechanisms represent both high-tech and low-tech solutions, ranging from data extraction from electronic medical records to manual data entry and submission of Excel files. Other federal efforts to support adoption of updated health information technology systems, interoperable data exchange of standardized data elements, and additional staff training will help ACO providers (and other model participants) as well.

It bears noting that in two older models, CMS is relying upon administrative data from other federal departments to obtain both demographic information and data on health-related social needs. For example, the Integrated Care for Kids Model merges Medicaid claims data with data from the Department of Housing and Urban Development’s Office of Policy Development and Research and the Department of Agriculture’s Women, Infants, and Children program and Supplemental Nutritional Assistance Program. In the Maternal Opioid Misuse Model, Medicaid claims data are merged with vital records such as birth certificates, which have complete and accurate race and ethnicity data, and socioeconomic information, such as parents’ education levels. In addition, we are able to collect sociodemographic data from state partners as well. Under the Maryland Total Cost of Care Model, the CMS Innovation Center entered into a Memorandum of Understanding with the state of Maryland that commits the state to expanding and standardizing the collection and use of sociodemographic data to improve the ability of primary care participants and Maryland hospitals to address health disparities.

4) Monitor And Evaluate Models For Health Equity Impact

Evaluating models to understand and measure the impact of interventions on underserved populations follows a two-pronged approach. For older models, we have conducted and will be publishing findings from a retrospective, cross-model review focused on underserved populations. In addition, where possible, we are expanding analytic work for current models to capture health equity impact, especially for models in which external studies or early signals suggest that further investigation may be needed.

For modified or new models, the CMS Innovation Center is designing robust evaluations that are intended to provide a thorough understanding of the model’s health equity impact. These evaluations have multiple dimensions, from seeking to understand beneficiary-level health impact and experience of care to better understanding provider impact, including how the model helped (or constrained) providers’ capacity to care for underserved populations. Other analyses will focus on the impact of additional funding, including the uses of such funding and whether the level, structure, and timing of funding facilitated improvements in care or better outcomes, as possible. We anticipate that the newly required health equity plans will provide information about participants’ intentions and identify where additional technical assistance, data, and other supports from CMS may be needed to help participants succeed.

Importantly, in addition to comprehensive evaluations, we are developing “leading indicators” and other internal, interim analyses that can provide early signals about model performance. For example, as demonstrated in exhibit 1, we have developed “heat maps” for the Primary Care First (PCF) Model to assess if and to what extent PCF participants are equitably located in both advantaged and underserved communities.

Exhibit 1: Primary Care First Model participation in Los Angeles, 2021

Source: CMS Innovation Center data.

Notes: Participants seem to be concentrated in the more affluent areas, with fewer beneficiaries in low socioeconomic, primarily Hispanic census tracts.

What To Expect In 2023

We will continue to focus our health equity work in each of the four areas described above, while working to refine and improve upon current efforts. For example, we have prepared detailed guidance for optimizing screening and referrals for addressing beneficiary health-related social needs (HRSNs), with expectations that new models will include HRSN screening and referral components when feasible. In addition, we have developed recommendations for measuring quality of care for underserved populations, in alignment with cross-agency efforts. These activities are ongoing and will evolve over time. To increase safety-net provider participation, we are disseminating outreach strategies to model teams, developing an expanded safety-net provider definition to encompass Medicaid beneficiaries and younger populations, and describing Medicaid policy levers that both model design teams and model applicants may reference.

We are expanding upon these foundational efforts in a number of ways. First, we intend to sharpen our focus on the needs of other underserved populations that are not well represented in our models. For example, exploratory work is underway to identify priorities and opportunities to embed a focus on people with disabilities in new models or initiatives.

Second, we are exploring options to build upon and support earlier investments in rural health, to address the unique access and health care needs of rural and geographically isolated communities. We continue to engage with stakeholders to inform solutions, particularly as rural providers report financial pressures affecting access to care.

Third, we will continue to explore mechanisms for social risk adjustment of payment. Some models are currently testing use of the ADI and dual-eligibility status to better account for social risk. Other indices, such as measures of frailty, eligibility for the Part D Low Income Subsidy, or life expectancy, may merit additional scrutiny.

Fourth, in response to the Executive Order on Lowering Prescription Drug Costs for Americans, the secretary of HHS selected three new models for testing by the CMS Innovation Center to help address the high cost of drugs, promote affordability and accessibility to life-changing drug therapies, and improve quality of care. These models could benefit some underserved populations. As one example, the Medicare High Value Drug List Model may help lower-income beneficiaries because it would provide access to a standard set of high-value generic drugs with a maximum copayment of $2 for a month’s supply. Additionally, through the Cell and Gene Therapy Access Model, CMS would help coordinate and administer multistate outcomes-based agreements with manufacturers to help Medicaid beneficiaries gain access to certain novel cell and gene therapies.

Fifth, and finally, to identify and remedy model designs that may lead to inequities in our models, we have developed an implicit bias tool that is based on findings from a previous CMS Innovation Center implicit bias initiative. This tool is being pilot tested on new models in the pipeline, and once finalized, we anticipate embedding use of this tool into the model review and approval processes.

Our equity “to-do” list is ambitious but necessary if we are to ensure that our long-term transformational work equitably benefits all beneficiaries we serve. We will hold ourselves accountable, with new metrics specific to the “Advance Health Equity” strategic objective’s primary areas of focus.

President Joe Biden has stated, “We need to make equity and justice part of what we do every day—today, tomorrow, and every day.” The CMS Innovation Center has accepted this charge and will continue to lead efforts that drive meaningful progress in advancing health equity today and in the years to come.