Integrated Services to Better Meet the Needs of Dually Eligible Beneficiaries: State of the Field

Tyler Shives, Center for Health Care Strategies
Jose Figueroa, MD, MPH

Today, across the United States, more than 12 million people are dually eligible for both Medicare and Medicaid. Each of these programs is governed by its own delivery, financing, and administrative policies and dually eligible individuals are largely left to navigate a fragmented uncoordinated system of care on their own. Integrated care models are designed to help dually eligible individuals maneuver this confusing landscape.

There are currently three different types of fully integrated care models: the Program of All Inclusive Care for the Elderly (PACE); Fully Integrated Dual Eligible Special Needs Plans (FIDE-SNPs); and Medicare-Medicaid Plans (MMPs) created under the federal Financial Alignment Initiative. While there is a growing evidence base for the effectiveness of these programs, relatively few dually eligible beneficiaries are enrolled in one of these programs.

The Better Care Playbook recently spoke with Jose Figueroa, MD, MPH, assistant professor of health policy and management at Harvard School of Public Health, assistant professor of medicine at Harvard Medical School, and associate physician at Brigham and Women’s Hospital, to explore who is enrolled in integrated care programs and what is known about their impacts.

Q: Why is fully integrated care critical for people who are dually eligible for Medicare and Medicaid?

A: Integrated care models mean one entity or organization is responsible for coordinating all Medicare and Medicaid services and benefits for dually eligible beneficiaries, and importantly, is responsible for the total costs of care across both programs. About 90 percent of dually eligible individuals are enrolled in non-integrated plans. They are left to navigate the benefits and services provided by both programs separately. When care is not integrated, it's challenging for these individuals to coordinate care across both programs. This bifurcated system is also confusing, inefficient, ineffective, and arguably, unsafe for dual eligible populations.

One other major concern is that there are perverse incentives to cost shift between the programs. For example, the Medicare side is not responsible for what happens on the Medicaid side, so it is possible that Medicare-only plans might shift some costs to the Medicaid program or be less willing to invest in interventions that lead to savings in Medicaid, such as reduced nursing home stays and increased home- and community-based services.

Q: What do we know about the impacts of integrated care programs?

A: Two important questions are: Do integrated care models work? And if so, how do they work?

First, it’s important to note that there is a lack of empirical evidence that isn’t based on observational studies. For example, there are no randomized control trials about whether enrollment into these programs work, or do not work, relative to a non-integrated plan. Most of the evaluations are limited to a few years of data, so long-term evaluations of integrated care plans are also missing. But, if you were to step back and look at the data to date, the evidence is promising, just not fully definitive. Here’s what we know in three key areas:

  • Improving utilization. The data supports that integrated care models are associated with reduced utilization of long-term nursing home stays. Integrated care models seem to be keeping people at home longer, for example through the PACE program, instead of institutional care settings. For reductions in acute care utilization, the data shows a more mixed picture. However, people who are dually eligible with complex needs often need more health care services. It is entirely possible that enrollment into an integrated care model may better help individuals connect with necessary services, and in some cases, lead to more utilization. We should be clear that utilization of health care services, even if it's higher, does not necessarily imply worse quality of care.
  • Improving quality of care. Most evaluations are based on utilization, but few studies have examined the impact of integrated care models on care coordination or quality of care. The data that does exist, however, leans in a positive direction for integrated care models although it is not a “slam dunk” that integrated care models substantially improve quality of care. There are a few studies that suggest enrollment in integrated care models may be beneficial in reducing mortality, but those studies have really important limitations. These include concerns for selection issues in terms of who's enrolling into integrated care models versus who is not. We don’t fully understand the impact on mortality or patient-reported outcomes in the absence of really good, randomized control trials, but generally, the data is at least pointing in the right direction.
  • Reducing unnecessary spending. People who are dually eligible are among the costliest individuals to take care of in our country. They incur about a third of the budget across both the Medicare and Medicaid programs. Much of the interest in expanding integrated care models is influenced by the idea that these programs may achieve savings or reduce costs of care for dual eligibles. If, for example, we fix fragmented care, we might reduce avoidable or unnecessary care, and that may mean we achieve savings. But unfortunately, currently, the data on whether integrated care plans achieve savings are not very clear. Most of the studies do not have comprehensive Medicare and Medicaid spending data to definitively answer this question. In some cases, some studies suggest that providing integrated care to some individuals may actually be costing more because of the upfront costs of care coordination. But stepping back, we have to remind ourselves that dual eligibles are individuals with complex health and social needs and they require extensive care. We should therefore not necessarily be prioritizing whether these programs save a lot of money. Instead, we should be ensuring that these programs are improving coordination, quality, and outcomes of care for some of the sickest people in our country.

Q: Based on your recent research, which groups of dually eligible individuals are the most likely and the least likely to be covered by integrated care programs?

A: There are important subgroups that are often left out of integrated plans. The first group that is less likely to be a part of these programs is younger dual eligibles. This group tends to have a higher percentage of individuals with a serious mental illness or major disability. One factor that limits enrollment is that PACE programs only enroll individuals who are 55 and older. While FIDE-SNPs and MMPs do not have an age restriction, they do have physician network restrictions, and a lot of studies about managed care plans suggest that any kind of provider or service restrictions may lead certain people who need specific services to disproportionately not enroll.

Notably, Black and Hispanic dually eligible beneficiaries had higher odds of enrolling in integrated care programs than white beneficiaries, which is an important finding that may help address health disparities.

Finally, geography limits access, since only some states have access to integrated care models like the PACE program, MMPs, and FIDE-SNPs.

Q: What high-level takeaways from this research are most relevant for policymakers across the country?

A: One of the most important questions has to be: how do we make integrated plans available as an option for dually eligible individuals everywhere? It's a pretty complicated issue, because both federal and state level coordination is needed to make these plans accessible. Medicaid rules and regulations are predominantly set at the state level. There's a lot of interest in figuring out how to increase support to states that are interested in expanding access to integrated care programs.

A second issue is that there are a lot of threats to integrated care currently in the form of, for example, non-conventional, non-integrated MA plans, also referred to as “lookalike” plans. I recently co-authored a study published in Health Affairs showing one in three dually eligible individuals who enrolled into look-alike plans were previously enrolled in more integrated models in other words, they’re taking them away. This growth in non-integrated models is part of the reason why so few dually eligible individuals are enrolled in integrated care programs. In an effort to curb enrollment into these lookalike plans, CMS recently introduced a policy to eliminate enrollment into certain lookalike plans, but our data suggest that these efforts may not be enough. Fortunately, there are ongoing discussions and proposed regulation about being much more aggressive at eliminating the growth of non-integrated plans.

The third thing that's a major issue is the lack of data availability that can help us ask better questions and conduct better studies about if these programs work and who they work for. We've been talking about dually eligible individuals as if it's one homogeneous group, but they are not. They’re a very heterogeneous population with complex multimorbidity and not every program is going to be a good fit for individuals within this group. Identifying which program is effective for whom depends on the context and on the person's needs, so we need better data to help guide these decisions. Most of the data out there does not look at subgroups of dually eligible individuals.

One example of a promising potential change is that CMS is currently considering taking action to improve quality reporting at the plan level, rather than just the contract level.  I think with more time, the data, and in turn, the studies will get better.