Why Did Certain Language Communities Steer Clear Of Medicare-Medicaid Integration Demonstrations?

EDITOR’S NOTE

This article is the latest in the Health Affairs Forefront major series, Medicare and Medicaid Integration. The series features analysis, proposals, and commentary that will inform policies on the state and federal levels to advance integrated care for those dually eligible for Medicare and Medicaid.

The series is produced with the support of Arnold Ventures. Included articles are reviewed and edited by Health Affairs Forefront staff; the opinions expressed are those of the authors.

The series will run through August 30, 2022; submissions are accepted on a rolling basis.

Ensuring equity in outreach and in access to program benefits should be a central consideration when designing programs serving people who are dually eligible for Medicare and Medicaid benefits (dual eligibles). It also is a daunting task with this diverse, high-needs population. Experience with the Medicare–Medicaid Financial Alignment Initiative (FAI), a highly integrated demonstration serving dual eligibles, has exposed some of the challenges in reaching language and cultural communities. Learnings from the FAI also provide a roadmap of questions to ask so that, moving forward, program models can better reach out to—and better serve—diverse communities.

The FAI, starting in 2013, enrolled more than 424,000 dual eligibles in 10 states into capitated managed care models in which a single managed care plan provides all Medicare and Medicaid benefits using a blended funding stream from both programs. It has been a rich resource for learnings about how to design programs and implement beneficiary protections for dual eligibles. Eight states participating in the FAI passively enrolled dual eligibles into the demonstration with an opportunity to opt out of participation.

One issue that stood out at the start of enrollment was the decision of significant numbers of individuals in certain language and ethnic communities to opt out of the demonstration. Evaluations of the initial phase of the demonstration commissioned by the Centers for Medicare and Medicaid Services (CMS) observed notably high opt-out rates in ethnic communities in New York (with particular note of the opt-out of Russian communities) and most notably California where the state carefully tracked opt-out rates by language community. Advocates observed similar patterns in other demonstration states.

The data from California are stark. In Los Angeles County, more than 90 percent of Russian speakers opted out of the FAI, as did 80 percent of Armenian, Farsi, and Korean speakers. Across the state, Mandarin and Cantonese Chinese speakers opted out at significantly higher rates than English speakers. In Orange County, it was Vietnamese speakers who topped the opt-out list. In contrast, Spanish speakers opted out at some of the lowest levels, even lower than English speakers. As an organization partnering with these communities and working to advance equity, we sought to figure out what was going on.

High Opt-Out Rates Raise Questions

With nearly 44 percent of the dual-eligible population identifying as members of racial/ethnic minorities, these opt-out figures invite deeper investigation. What do these opt-outs tell us about the ability or the perceived ability of integrated plans to adequately serve racial and ethnic communities, many of which have large numbers of limited-English proficient (LEP) individuals, in a way that is culturally competent? What can be learned about outreach to these communities and to the providers who serve them? Are there specific aspects of plan design and operation that can be tailored to address the unique circumstances of providers serving LEP communities and ensure robust culturally competent networks?

The answers to these questions can inform design of integrated models moving forward, particularly Dual-Eligible Special Needs Plans (D-SNPs), a type of Medicare Advantage plan created to address the benefit coordination needs of dual eligibles. D-SNPs are growing in dominance as a vehicle for integrating care, with more than one in four dual eligibles nationwide currently enrolled. Furthermore, the FAI demonstration, which proved complex to administer, is winding down, with most participating states transitioning their FAI plan enrollees into D-SNPs. Both CMS and the states are applying learnings from the FAI to the design and oversight of D-SNPs. Lessons from the opt-outs in the FAI are particularly important in light of the authority CMS offers states to use default enrollment in connection with D-SNPs.

The evaluations of the demonstration provide some insights. Evaluators found that the primary reason for the large opt-outs was that trusted providers serving these communities, especially primary care doctors, were not part of the networks for FAI plans. When told that they could not continue with their provider, many dual eligibles opted out, choosing to remain with their current coverage. While ability to continue with trusted providers was a primary reason for opt-outs across ethnicities, the levels seen with certain language communities were well above those seen with English speakers.

Both CMS evaluators and on-the-ground advocates reported that, in some cases, the actions of providers to discourage enrollment could be quite aggressive. Advocates reported that one provider group serving a particular language community created a YouTube presentation explaining exactly how to opt out. There were also reports of providers buying time on in-language radio programs and placing ads in local non-English language newspapers urging dual eligibles to opt out.

The evaluators reported that respondents they interviewed believed that the reluctance of providers in these communities to participate in the demonstration networks was grounded primarily in their antipathy to managed care. That explanation, however, does not seem to be the entire answer. These providers may not be fans of managed care, but that position is hardly unique. It is easy enough to find managed care critics throughout the medical community. Furthermore, because these providers are serving disproportionately low-income communities, they are not typically concierge practices and must necessarily be engaged with managed care with patients enrolled in employer plans, Medicaid plans, and other insurance.

We believe that there is more to learn from these dramatic statistics and propose that plans, CMS, and states take additional steps to explore their significance and consider ways to address the issues they expose. These steps include:

Collecting Comprehensive Data

Although there was a general perception in several states that some language groups were opting out of the demonstrations at higher rates, California was the only state that closely tracked race and language enrollment activity and made it publicly available at the county level. To ensure that integrated programs serve all dual eligibles equitably—not just in enrollment but in every use and outcome measure—it is important to collect comprehensive granular data and publish the data so that it can be analyzed to determine if all dual eligibles are equally benefitting from programs and services. Without such data, aggregate measures can easily mask deficiencies and successes in serving specific populations.

Listening To A Diverse Set Of Beneficiaries

Gathering answers behind the numbers must begin with asking the affected beneficiaries themselves. In the FAI, with a few exceptions, most of the CMS-sponsored evaluations as well as outside studies limited consumer interviews to English and Spanish, primarily because of resource constraints. This strategy meant that, in the FAI, consumer voices from the communities with the most extreme opt-out experience were not heard. Focus groups and consumer surveys need to capture the experiences of all affected populations and should be conducted in languages beyond English and Spanish. Similarly, enrollee advisory committees, now a requirement for all D-SNPs, should cast a wide net. Equitable solutions must start with equitable listening.

Digging Deeper With Providers And Plans

Perhaps the area most deserving of further exploration is why providers serving certain language communities were so underrepresented in integrated plan networks. One evaluation reported that the plans responded to the finding by saying that they would reach out more to providers serving LEP communities. While outreach is certainly helpful, we need a nuanced analysis of the unique needs and perspective of medical providers serving language and cultural communities, as well as of plan perspectives on challenges in working with these providers. That analysis must start with outreach to the providers themselves, asking why they did not join FAI networks and what barriers they perceive to participation in integrated programs moving forward. After the first wave of passive enrollments in California, the state took significant steps in this direction, including concerted outreach to providers whose patients accounted for most of the opt-outs and consulting with organizations such as the Network of Ethnic Physician Organizations. In-depth discussions with managed care plans about their experiences in recruiting and working with providers serving language communities also must be a foundational step. An inquiry should address multiple questions:

• Do the size and organization of medical practices serving LEP communities play a role? Advocates report that many providers who are embedded in LEP communities are smaller practices, either solo or small groups. Is it harder for them to participate in a program that requires levels of technical and staffing resources and data-sharing capabilities that may be beyond their current capacity? If so, what can plans do to assist providers in these areas? We note possible parallels with the Business Acumen Initiative of the Administration for Community Living, a program that provided technical assistance to community-based organizations (CBOs) around contracting issues with managed care plans.
• On the plan side, do plans concentrate their network recruitment on large provider groups that have more sophisticated support systems and that can be managed more centrally? What can plans do to prioritize recruitment of medical practices that serve language communities with culturally competent care? Are there flexibilities that plans could adopt to better accommodate the needs of providers serving language communities? Are there ways to shape network adequacy rules to better incorporate equity goals?
• Is antipathy to managed care really the core issue? Provider groups serving language and cultural communities are not monolithic, and some are not small. For example, one large provider group operating in the heart of an ethnic enclave in Los Angeles with 50 primary care physicians and affiliates in 250 specialties specifically identifies itself as serving one language and cultural community. The group participates in many managed care plans, including five Medicare Advantage plans. Interestingly, the group is part of the network for a D-SNP operated by a plan sponsor participating in the FAI demonstration but does not participate in that sponsor’s FAI plan network. The group also participates in the network of at least one D-SNP “look-alike,” a plan that has an overwhelmingly dual-eligible enrollment but is not a D-SNP. Why did this group that already participates in plans serving a dual-eligible population not participate in the FAI demonstration? Were there specific aspects of the design of the FAI models that they found objectionable or difficult? Did plan sponsors recruit them for FAI participation?

Engaging And Collaborating With CBOs

In the FAI, plans and states, to varying degrees, engaged with CBOs serving language and ethnic communities, including outreach during rollout of the demonstration programs. Going forward, integrated programs can be stronger and more equitable if states and plans consistently incorporate CBO expertise in every phase—developing program design, rollout, and ongoing implementation.

CBOs understand the specific attitudes and concerns of their community and can help formulate responses to address them. They can assist in tailoring outreach that is culturally competent, not merely translated. They have relationships with the local provider community, and some are themselves providers, particularly of non-medical services that address social determinants of health. Importantly, they are trusted and well situated to help members of their community make informed decisions and navigate new and complex programs.

Low-to-the-ground CBOs can be highly effective partners, as their contributions in the COVID-19 response have amply demonstrated. But they need grant and contract opportunities to support their work. They also first need to be identified, which requires outreach by states and plans. Many effective CBOs, particularly for smaller language groups, are themselves small and often underfunded and may not be closely following initiatives at the state and federal level.

The Need For Equity

We have more questions than answers, but one thing is sure: To make a dent in current inequities in health outcomes requires starting with equitable delivery of health care services. We saw, for example, that network limitations in the California demonstration had real consequences for those LEP speakers who did enroll. In one evaluation, fully half of LEP enrollees reported that they could “never” get a medical interpreter. Integrated models must, from the outset, have a robust network that includes culturally competent medical providers attuned to the unique needs of the language and cultural communities they serve.

Ensuring that happens requires inquiry, outreach, and careful design. Mining the learnings from the FAI is a good place to begin.