Provide High-Cost Medicare Beneficiaries with Highly Valued Care

To hear Eugene Allen tell it, he was nearly dead when he was wheeled into the Special Care Center in Atlantic City, New Jersey.¹ He had just been discharged from the hospital, where he’d spent three months on a mechanical ventilator in the intensive care unit, unconscious most of the time. As a rule, Allen avoided doctors unless it was a medical crisis or his wife nagged him to see someone. So, when his right leg swelled up, he continued his work as a chef and didn’t seek care—until the blood clot travelled into his pulmonary artery and collapsed his lung. His chances at survival didn’t look great; in addition to the collapsed lung, he battled a hospital-acquired infection and weighed 480 pounds at the time. But Allen is just the kind of patient for whom the Special Care Center is made. There, health care providers work to engage patients as partners in their care with a focus on lifestyle changes, such as diet and exercise. To help patients succeed, they receive unlimited, free access to the Special Care Center, which assigns health coaches to help keep patients on track, and often receive reduced-cost medications. Despite his initial misgivings, Irma, Allen’s health coach, was patient with him, helping him work through medication side effects and encouraging him to walk, up to 14 miles a week, which he credits with helping him lose 100 pounds.

This kind of care coordination and flexible benefit design could benefit millions of Medicare beneficiaries in the fee-for-service program. Medicare beneficiaries want and deserve highly valued care—care that is person-centered, convenient, efficient, and delivered in innovative ways to help each patient achieve their life goals. But Medicare’s administrators need to make explicit goals on spending and health performance as well as be held accountable for achieving them. Doing so will deliver highly valued care to the highest-cost Medicare fee-for-service beneficiaries and can save the federal government as much as $80 billion over 10 years.

This idea brief is one of a series of Third Way proposals that cuts waste in health care by removing obstacles to quality patient care. This approach directly improves the patient experience—when patients stay healthy, or get better quicker, they need less care. Our proposals come from innovative ideas pioneered by health care professionals and organizations, and show how to scale successful pilots from red and blue states. Together, they make cutting waste a policy agenda instead of a mere slogan.

What is Stopping Patients from Getting Highly Valued Care?

Medicare is in the process of transforming itself from a passive bill payer to an active program manager. Much of this effort focuses on beneficiaries like Eugene Allen whose high costs can be avoided through better coordination and accountability. But this transformation will not be complete without a clear description of the spending and health performance goals that Medicare expects to achieve for beneficiaries and taxpayers—as well as accountability for achieving those goals.

The Affordable Care Act launched a period of massive experimentation in how Medicare pays providers. And, in the Medicare Access and CHIP Reauthorization Act, a broad, bipartisan majority in Congress established new payment models as an alternative to the fee-for-service model for physicians. These changes come on the tail of the creation of private Medicare Advantage Special Needs Plans enacted as a part the Medicare Modernization Act of 2003, which fine-tune benefits and services for high-cost beneficiaries who have special needs.² Congress identified three categories of beneficiaries with special needs, resulting in high costs of care, that would benefit from a model of care built for them:

• Dual-eligible beneficiaries;
• Beneficiaries living in a nursing home or requiring nursing home-level care; and
• Beneficiaries with certain severe or disabling chronic conditions.

As a result of these and other directives from Congress, the Centers for Medicare & Medicaid Services (CMS), and the Center for Medicare & Medicaid Innovation, in particular, has created many great initiatives, but these efforts are missing a key piece—an overall strategy that will lead to the delivery of highly valued care.

Highly valued care is individualized, convenient, and efficient care delivered in innovative ways to help each patient achieve their specific goals.

Other than thresholds for moving away from fee-for-service payments into alternative payment models, Congress and CMS have not set overall patient outcome and cost goals—key metrics for advancing highly valued care. For example, the kind of personal care that Eugene Allen received from the Special Care Center can reduce repeat hospitalizations and increase levels of patient satisfaction. The lack of explicit goals in Medicare can be traced to the origins of the program, which was not conceived as an organized system of care when it was enacted in 1965. At that time, it was a single, public health plan that paid for health care with a fee for every service—plus supplemental coverage from private health plans, also known as Medigap.

But with chronic disease, health care isn’t always a matter of going to the doctor to get a quick fix. Medicare beneficiaries with the most expensive health care needs are the most at risk in a health system that fails to appreciate and deal with the obstacles they face with taking care of themselves. The most costly 10% of beneficiaries enrolled in fee-for-service Medicare account for 57% of that program’s spending.³ These high-cost beneficiaries have diverse needs but tend to share a few common characteristics, notably having multiple chronic conditions, being hospitalized (often repeatedly), and being dually eligible for Medicare and Medicaid.⁴ These beneficiaries are often lost, trying to navigate an unorganized system, attempting to coordinate their own care among multiple providers who may provide conflicting recommendations, experiencing sub-standard outcomes and costing taxpayers more every time they experience the inevitable outcomes of all this confusion—emergency room (ER) visits, repeat hospitalizations, and more.

If the fee-for-service world is challenging for these high-cost beneficiaries, the Medicare Advantage world of capitated payments has its own challenges, despite starting with a strong foundation. Medicare Advantage health plans receive a monthly capitated rate that is intended to cover all the costs of care for patients. That gives health plans an incentive to coordinate care and help beneficiaries avoid unnecessary hospitalizations. But the capitated rate must be high enough to cover the costs of beneficiaries who have higher than expected costs due to their health condition. Otherwise, plans will have an incentive to encourage high-cost beneficiaries to leave the plan by limiting their care or not enroll them in the first place.

Over the last decade, CMS has improved the process of adjusting the capitated rate based on expected costs for each beneficiary, a process called risk adjustment.⁵ These changes have significantly boosted payments for high-cost beneficiaries, but technical questions remain about whether the current risk adjustment process is as accurate as it could be.⁶ Beyond the technical questions, CMS has not incorporated risk adjustment as part of an overall strategy to improve health outcomes for high-cost patients that makes these technical issues highly relevant to public policy goals. Instead, CMS has sometimes eliminated factors in risk adjustment, such as stage 1-3 chronic kidney disease and dementia codes,⁷ despite the fact that a National Institutes of Health-supported study found health care costs for people with dementia were equal to or greater than costs for heart disease and cancer, which are accounted for in the current model.⁸

One type of Medicare Advantage plans, Special Needs Plans (SNPs), may feel the impact of this lack of refinement in risk adjustment more acutely than other Medicare Advantage plans, given their enrollment population. SNPs are permitted to limit enrollment to, and tailor their benefit packages for, beneficiaries with certain special needs who tend to have higher health care costs. While SNP enrollment grew in 2014, the plans cover just 4% of the total Medicare population⁹ and likely remain underutilized relative to beneficiaries who qualify to enroll. In addition to risk adjustment concerns, an often hidden enrollment process and the absence of risk adjustment for quality measures that drive the plans’ star ratings likely contribute to this underutilization. These problems also affect Medicare Advantage plans because they use the same risk adjustment system as the SNPs.

For example, while SNP-qualifying beneficiaries have more enrollment opportunities than non-SNP beneficiaries, Medicare’s Plan Finder does not automatically display these plan options; users must proactively choose to view SNPs as they refine their plan results online instead of being directed to those options based on their situation. With regard to star ratings, the Medicare Payment Advisory Commission (MedPAC) and others have noted a relationship between plans with a low star rating and more high-cost beneficiaries, specifically dual eligible beneficiaries.¹⁰ Dual eligible enrollment in SNP Medicare Advantage plans is four to 12 times more common than regular Medicare Advantage plans.¹¹ There is also an association between low star ratings and the proportion of plan enrollees under age 65. Relative to non-SNP Medicare Advantage plans, SNPs for duals and for those with chronic conditions serve many more of these younger beneficiaries who qualify for Medicare on the basis of disability or end stage renal disease.¹²

Medicare has all the individual parts in place to deliver highly valued care to these beneficiaries. What is missing is a management plan that brings all the pieces together.

Where Are Innovations Happening?

Large health plans, medical group practices, large employers, and even third parties throughout the country provide effective examples for how Medicare can deliver highly valued care that improves beneficiary outcomes and reduces costs.

The Special Care Clinic, part of AtlantiCare, a health system in Atlantic City, NJ, was established in 2007 to help two self-insured health plans—a casino workers’ union and AtlantiCare Medicare Center—reduce their health care costs.¹³ The clinic is exclusively available to workers who have (or who are likely to have) very high medical costs. These patients were incentivized to join the clinic with unlimited, free access—no co-payments, no insurance bills—and with a clinic designed around providing very sick patients what they need, such as guaranteed same-day appointments for the acutely ill and health coaches to help them meet their goals. Reduced-cost medications are also often available. Patients typically see their doctor at the clinic 6-8 times each year, but have 35-40 visits with their health coach.¹⁴ Health coaches operate in a similar manner to community health workers—they come from the same communities as their patients, speak their native language, and work to connect with patients—understanding the challenges they face, setting goals, and developing a plan together to meet those goals.

The Special Care Clinic is different from other clinics on the back end, too. It receives a flat, monthly fee for every patient, rather than a fee for every service performed.¹⁵ Because this is the only reimbursement the clinic receives, it has no billing department, eliminating a huge administrative expense. Physicians must focus on service to retain patients and, thus, their monthly fees. And, clinic staff meet each morning to review the charts of the patients scheduled to be seen that day, discuss those in the hospital, and identify those who must be contacted regarding lab results or for other reasons.

This model of care has both reduced costs and improved quality. During the clinic’s first 12 months:

• ER visits and hospital admissions dropped by more than 40%;
• Surgeries dropped by more than 25%;
• Only two of 503 patients with high blood pressure were in poor control;
• Patients with high cholesterol experienced, on average, a 50-point drop in their cholesterol levels; and
• 63% of smokers with heart and/or lung disease had quit smoking.¹⁶

Comparing these patients with a similar group demonstrated a 25% reduction in the cost of their care.

Another innovative example is CareMore—an integrated health care company which offers Medicare Advantage plans, including special needs plans and a disease management program, and operates health centers that deliver care to members.¹⁷ CareMore’s Medicare Advantage health plans receive a risk-adjusted, capitated payment for each enrollee, and its health care centers only serve its plan’s members.¹⁸ In this way, CareMore has moved away from fee-for-service payments, which allows it to build specialized programs for high-risk and high-cost patients. For example, in order to ensure patients make their doctors’ appointments, CareMore provides free transportation to and from its clinics. Other nonmedical services to improve patient compliance include health care professionals who assess patients’ homes to mitigate fall hazards or ensure they have scales to track their weight, pill boxes equipped with an alarm to remind patients to take their medications, and other wireless monitoring devices.

CareMore employs different clinical models for its frail and chronically ill population and for its non-frail population.¹⁹ The frail and chronically ill population represents about 20% of CareMore members, and they account for 60% of its medical costs.²⁰ The chief difference between the two models is the use of an “extensivist” who manages care for frail and chronically ill patients. As conceived by CareMore, an extensivist is a physician who connects a patient’s hospital care with their outpatient follow-up and continuing care. This ensures that the patient understands the care plans and instructions outlined by the multitude of providers they may see and works to reconcile those plans when they inevitably conflict. CareMore offers additional programs designed for specific chronic conditions, such as back pain, chronic kidney disease, diabetes, heart failure, and more.²¹ All of this care coordination is supported by what CareMore calls a “longitudinal patient record,” which collects information from 8-10 different sources—electronic health record information is just one source of many—to build a complete picture of each member’s health and health care.²²

CareMore’s model seems to be working. The company reports medical costs for its Medicare members are 18% below the industry average, which it achieves with a hospitalization rate 24% below average, hospital stays that are 38% shorter than average, and an amputation rate for diabetics 60% below average.²³

SCAN Health Plan, operating in California and Arizona, serves high-cost Medicare beneficiaries through its Medicare Advantage SNPs.²⁴ SCAN takes a holistic approach to evaluating beneficiaries’ needs and improving their health, offering a range of home- and community-based services to help beneficiaries live in their communities. Part of their approach involves sending case managers, who may be nurses or social workers, to beneficiaries’ homes to help coordinate the range of health and social services a member may need to safely remain in their home. This may involve helping beneficiaries access meal and food services, apply for Medicaid or subsidized housing, or make doctors’ appointments and arrange for transportation.²⁵ SCAN’s efforts result in fewer repeat trips back to the hospital, higher member satisfaction ratings, and a 26% reduction in readmission to nursing facilities.²⁶

Integra ServiceConnect is one example of a non-provider, non-health plan service that connects individuals to the health care and social services they need to live healthier, more productive lives. Integra partners with health plans, health systems, accountable care organizations, and even government programs (such as state Medicaid programs) to improve health outcomes.²⁷ By deploying teams of community health workers, who are from the communities in which the targeted individuals live, Integra is able to find, engage, and connect with people with high health care costs, helping them become invested in their health and health goals, and connecting them to a variety of services.

Integra’s community health workers, called community coordinators, help individuals establish contact with a primary care physician, support them in advocating for themselves and ensuring all their questions are answered at health care appointments, and teach them about appropriate ER use, among many other responsibilities. On the social services side, community coordinators may connect individuals with local food banks or housing services to meet their needs. One of Integra’s partnerships with a Medicare Advantage dual-eligible special needs plan yielded an 80% reduction in ER visits.²⁸ Spending for members of the plan with a behavioral health primary diagnosis dropped by 12% through a combination of decreased medical claims and increased pharmacy claims as patients presumably learned better self-management skills.²⁹ The cost of care for the engaged cohort dropped by approximately 60%.³⁰

How Can We Bring Solutions to Scale?

In order to provide highly valued care to Medicare beneficiaries with the highest costs, Medicare’s administrators need to make explicit goals on spending and health performance as well as be held accountable for achieving them. To codify this process, policymakers need to take the following five steps:

First, Congress should direct Medicare to develop and implement a management plan every two years for getting high-cost beneficiaries highly valued care. The plan should include both spending and health outcome goals. Medicare has outlined laudable and aggressive goals for moving Medicare payments from a volume to a value basis but, thus far, has not defined what those goals mean for patients and how achieving those goals will impact spending. Through a management plan, Medicare would outline its role as an active program manager and specify how it will deliver highly valued care for the costliest 10% of fee-for-service beneficiaries. Ideally, this management plan will integrate the many outstanding initiatives Congress has directed CMS to implement, ranging from special needs plans to accountable care organizations, in a way that prioritizes outcomes over processes. It should also include an overall strategy to handle the diversity of needs among high-cost beneficiaries targeted by geographic regions.³¹ CMS can leverage ongoing efforts to create and share new knowledge about how to make care highly valued.³²

Second, Congress should tie the spending and health goals Medicare establishes in its management plan to performance criteria for Medicare administrators. To further incentivize Medicare administrators to implement the management plan and achieve spending and quality goals, Congress should tie these goals to the existing performance rating and bonus system for political appointees and members of the Senior Executive Service (SES).³³ In this way, performance bonuses will be available to appointees and SES members whose work directly contributes to achieving Congress’ goals.

Third, Congress should provide Medicare with the benefit design flexibility needed to meet the spending and health goals established under the management plan. Some Medicare beneficiaries need more help than others to achieve their best health. For example, the costliest 10% of Medicare fee-for-service beneficiaries may benefit from the intensive management described earlier, but this level of care coordination is not necessary or useful for beneficiaries who do not have these care needs. In order to meet the needs of this expensive population, Congress should provide Medicare with the authority to employ creative solutions (such as enhanced benefits or cost-sharing elimination) to improve health outcomes and reduce spending on the costliest beneficiaries who often face social challenges that exacerbate their health problems.

For example, food insecurity, difficulty securing transportation to appointments, or even lack of storage for temperature-sensitive medications can contribute to poor health outcomes. But, Medicare doesn’t pay for meals, transportation, or refrigerators, and care coordination and navigation is nonexistent for fee-for-service beneficiaries who are not aligned with an accountable care organization or patient-centered medical home. As outlined in legislation introduced during the 114th Congress by Representative Cathy McMorris Rodgers (R-WA), benefit flexibility may include provision of benefits not currently covered under Medicare, such as in-home care, transportation, or meal services, and lower out-of-pocket costs to ensure beneficiary compliance with treatment plans or pharmaceuticals.³⁴ By offering flexibility in the benefit design for a highly targeted population, Congress can give Medicare the tools needed to achieve Congress’ goals.

Fourth, Congress should direct Medicare to improve payments to health plans for high-cost beneficiaries as part of the management plan. These changes should include revisions to the risk adjustment system in order to more accurately account for the predictable costs of high-cost beneficiaries in Medicare Advantage. The current risk adjustment model combines beneficiary demographic characteristics and medical diagnoses into categories (called hierarchical condition categories, or HCC) to predict a beneficiary’s health care costs. However, costs vary substantially even among beneficiaries with the same HCC classification.³⁵ This fact poses three problems for risk adjustment.

First, according to MedPAC, the model over-predicts actual costs for very low-cost beneficiaries and under-predicts actual costs for very high-cost beneficiaries.³⁶ A recent analysis by Avalere documents a similar problem for Medicare beneficiaries with multiple chronic conditions.³⁷

A second problem stems from the fact that the current risk adjustment model uses demographic data from the same year for which costs are predicted but draws health condition data from the previous year.³⁸ This timing in the collection and use of risk adjustment data accentuates situations where costs grow over time, such as end-of-life care. Research sponsored by CMS in 2002 (which needs to be updated) showed that end-of-life care costs expand well beyond the costs expected for any given disease.³⁹

A third problem for risk adjustment that MedPAC has identified is the use of a single adjustment factor for both full- and partial-benefit dually eligible beneficiaries.⁴⁰ While partial-dual eligibles share many characteristics with full-dual eligibles, they tend to be healthier, resulting in an average risk score nearly 15% lower than that of full duals and average annual spending nearly $3,000 less than for full duals.⁴¹ MedPAC has considered a number of changes that may improve the model’s accuracy, which present CMS with a solid starting point for reform.⁴²

Fifth, Congress should require the Medicare management plan to encourage better anticipation of and care for high-cost beneficiaries in health plans. Medicare should reward, through risk sharing or similar agreements, health plans that can improve the identification of beneficiaries who are likely to experience high health care costs. Private health plans are already investing in predictive modeling that identifies which beneficiaries might need early interventions to improve their health outcomes and avoid higher costs.⁴³ For example, patients with a neurological disease, like dementia, have a 9% greater chance of being high-cost.⁴⁴ Such private investment could also have a public benefit if it were used to improve risk adjustment models and identify high-cost patients in original, fee-for-service Medicare. Congress should provide for temporary, three-year funding for risk-sharing through reinsurance or a similar mechanism, which would be more than the offset from the projected savings under the other provisions of this proposal. At the end of a three-year risk sharing period, the new method for identifying beneficiaries at risk for having higher costs would become part of the risk adjustment process.

Potential Savings

Research conducted by Avalere concluded that, for example, using a capitated payment rate 5% less than fee-for-service rates, managing care for high-cost beneficiaries would decrease federal spending by more than $80 billion dollars over the 2015-2024 federal budget window.⁴⁵ Using a capitated payment rate of 1% less than fee-for-service rates yields savings to the federal government of almost $17 billion for this same period, according to Avalere estimates. An estimate of the savings from the specific provisions in this proposal would render a more precise estimate of savings.

Questions and Responses

How would you identify the costliest 10% of Medicare fee-for-service beneficiaries?

Analysis of the persistency of health care costs over time demonstrates the importance of prospectively identifying the costliest 10% of Medicare beneficiaries in a given year. Only about 45% of beneficiaries with spending in the top 10% in one year retain this status in the following year.⁴⁶ Identifying high-cost beneficiaries early is critical because, as their spending begins to rise, they will often need help to utilize health care services more efficiently. Claims data alone will not accurately predict high-cost beneficiaries, and some non-medical characteristics also increase a beneficiary’s probability of being high cost.⁴⁷ CMS should take a sophisticated, big data approach to identifying the costliest 10% of beneficiaries, combining clinical, claims, and non-medical data to develop prediction tools that identify beneficiaries who are likely to be high cost. One framework for this kind of analysis separates the high-cost population into three subgroups—those with advanced illness, those with persistently high spending patterns, and those with episodic high spending—and focusing interventions on the first two groups.⁴⁸

How does this proposal relate to existing organizations that focus on high-cost beneficiaries, such as Accountable Care Organizations (ACOs) and Special Needs Plans?

The core of this proposal is a management plan that includes spending and health outcome goals for the costliest 10% of Medicare beneficiaries. ACOs, SNPs, and similar organizations are tools to target and support high-cost beneficiaries, and they need better management practices to support their approach. The management plan should focus on the results CMS will achieve in delivering highly valued care for Medicare fee-for-service beneficiaries, not on the tools to be utilized in achieving those results. The most effective tools will likely change over time, but the focus on delivering highly valued care should remain constant.

How would high-cost beneficiaries be enrolled in programs that aim to rationalize their care?

Legislation introduced during the 114th Congress by Representative Cathy McMorris Rodgers (R-WA) offers one example of how beneficiaries may be enrolled with organizations that seek to coordinate and rationalize their care.⁴⁹ H.R. 3244, the Providing Innovative Care for Complex Cases Demonstration Act, utilizes proactive outreach to identify high-cost beneficiaries, informing them of the program and the process for enrollment and disenrollment, as well as passive enrollment. When a beneficiary is located in a service area with more than one care coordination option, the legislation directs enrollment in the organization, which includes providers from whom the beneficiary received services in the previous year. Beneficiaries retain the right to withdraw from the program and return to fee-for-service Medicare.

Our proposal would modify the process outlined in H.R. 3244 slightly to further avoid disruption in care for this vulnerable population by auto-enrolling beneficiaries only if they are already an established patient with that organizations’ providers. If not, CMS would extend to that beneficiary’s providers an opportunity to join or become a care coordination organization.

Should Medicare develop a management plan that covers all beneficiaries?

Yes, but the place to start is with the highest-cost beneficiaries. Shifting Medicare from its historic role as a bill-payer to its new role as an active program manager is a massive change. Medicare administrators have their hands full implementing the payment reforms under its own policy goals and the Medicare Access and CHIP Reauthorization Act, but it is fundamentally important to show beneficiaries what they are going to get from all these changes. Administrators can start this process with the highest-cost beneficiaries who will face the biggest impact from moving Medicare from a volume-driven payment system to value-driven payments. Beneficiaries have a right to know how restraints on the volume of care delivered will improve their care. Moreover, finding a way to provide affordable, customized care to a sizable group of high-cost beneficiaries with diverse needs can create the systems for dealing with everyone in Medicare with diverse needs who are not necessary high-cost beneficiaries.

How would your proposal change the existing performance rating and bonus system for political appointees and members of the Senior Executive Service?

Under our proposal, Congress would establish meaningful and measurable health outcomes and spending goals that SES members working in Medicare would need to achieve in order to receive a performance award. Congress authorized performance-based pay systems for members of the SES in 2003,⁵⁰ and, in 2012, the Office of Personnel Management developed a new performance evaluation framework.⁵¹ The 7,900 members of the SES make up less than 1% of all federal civilian employees and serve in key positions just below presidential appointees.⁵² They are eligible for individual performance awards of 5-20% of their basic rate of pay, though total award payments are generally limited to 10% of aggregate basic pay within an agency.⁵³ A recent report found nearly half of SES members received the highest possible rating over four fiscal years, calling into question whether the performance rating system in place is meaningful and accurately rewards outstanding performance.⁵⁴

How would you address benefit equity issues raised by this proposal?

Our proposal advocates for flexibility in benefit design to better serve the costliest 10% of Medicare fee-for-service beneficiaries. As described above, this may mean providing these beneficiaries with enhanced benefits to help address issues such as food insecurity, difficulty securing transportation, lack of storage for temperature-sensitive medication, or other seemingly non-health related issues. Though our proposal focuses on the fee-for-service population, the recently-proposed Medicare Advantage Value-Based Insurance Design (MA-VBID) model provides one example for addressing benefit uniformity. Medicare Advantage plans are generally required to offer the same benefits and cost sharing for all enrollees.⁵⁵ Medicare Advantage plans in seven states will be eligible to participate in the MA-VBID model, which will include a limited waiver of MA and Part D uniformity requirements for enrollees with seven specific clinical conditions. In this way, the MA-VBID model will demonstrate whether supplemental benefits or reduced cost sharing leads to higher-quality, lower-cost care.

CMS regulations authorize states to vary cost-sharing for drugs, emergency room, inpatient, and outpatient visits based on clinical information, and states, such as Michigan, have begun to implement VBID in their Medicaid programs.⁵⁶