In August, 2016, a 93-year-old woman—the grandmother of one of this Blog post’s authors—died of congestive heart failure, five weeks after she underwent surgery to receive a pacemaker. There were alternative care options, but they were not offered to her and her family in a timely manner, at least in part because of Medicare’s long-standing payment rules that value procedures over discussion of goals and alleviation of symptoms. Medicare paid for the surgery and pacemaker with no questions asked, even though the procedure was, in retrospect, unproductive, wasteful, and even harmful from the family’s perspective.

The better alternative would have included palliative care, which helps patients and family members to call a “timeout,” step back, and decide what course of treatment is best for addressing symptoms and deteriorating quality of life in persons with advanced, life-limiting illness. However, there was no palliative care team available to disrupt the default, “do everything” course of care during this family member’s last hospitalization. Nor was such care offered at any of the four previous hospital visits she experienced in 2016, or through her primary and specialist care providers.

Medicare will cover some elements of palliative care, but payments are generally less than the cost of delivering the service because Medicare’s approach to reimbursement values procedures over time spent with patients and families. Hospitals have begun investing in palliative care teams because of the acute disease burden their patients experience near the end of life, but the availability of such care varies widely. Increasing the system-wide use of palliative care will require a new payment approach for Medicare, the insurer for over eight in 10 decedents annually in the United States.

As part of the new Health Affairs Blog series highlighting serious illness and end-of-life care, this post describes a community-based, interdisciplinary palliative care model that spans care settings and is being evaluated in a Center for Medicare and Medicaid Innovation (CMMI) demonstration project. It provides some early lessons on how alternative payment models for palliative care might be developed to increase its provision in Medicare.

The model is being demonstrated by Four Seasons Compassion for Life, a nonprofit hospice and palliative care provider operating with a network of community partners in the Western regions of North and South Carolina. Supported by $9.5 million in funding from the Center for Medicare & Medicaid Innovation (CMMI), the project began in September 2014 and, by the end of its evaluation period in August 2017, approximately 5,200 Medicare patients will have enrolled in the model. Evaluation of the project will include patient outcomes and the total cost of delivering care through the model. From these results, we will propose a new payment model for palliative care that could then be tested in other locations.

Current Medicare Payment Design Discourages Palliative Care

Expanding access to palliative care should be a key policy priority given the growing body of evidence that suggests that palliative care both improves patient outcomes and limits costs. From a policy perspective, new interventions or treatments are typically evaluated by determining whether their marginal benefits are worth the increased costs.

A landmark randomized controlled trial in 2010 found that palliative care improved quality of life, increased survival, and reduced costs for patients with stage 4 lung cancer. If someone developed a drug that did this, that person would be a billionaire. This combination of outcomes sounds too good to be true but may be possible for palliative care, in part because of how expensive the default care patterns are for elderly persons near the end of life.

Palliative care’s cost-saving potential lies in the ability of upstream care coordination to reduce low-value care just before death. A key question is whether the results of this randomized controlled trial can be replicated in other settings. Our CMMI demonstration project will help determine the answer.

Even if our model proves successful, Medicare’s current benefit and payment design architecture would remain an impediment to the promotion and spread of the model’s interdisciplinary approach to palliative care. Medicare Part B pays for some of the provider visits and treatments associated with palliative care, but only at a substantial loss to the provider due to the Part B bias that reimburses procedures more amply than time spent with patients.

Medicare will need a new payment model that adequately compensates the coordinated approach to delivering this time-intensive, procedure-light care.

The Four Seasons Hospice Model of Palliative Care

Patients with advanced, life-limiting illness can enter the Four Seasons model via a hospital consult, a stay in a long-term care center such as a skilled nursing or assisted living facility, a referral to an outpatient clinic by an attending physician, or a home visit. A palliative care team then delivers needed follow-up care across all of these sites to address symptoms and help patients and families make health care decisions. Such care supplements the regular, disease-modifying care provided by their primary physicians.

The most common diagnoses are cancer, heart failure, chronic obstructive pulmonary disease, and dementia. When patients reach a six-month life expectancy and are eligible for hospice, they are given the option to enroll. The palliative care team includes physicians, nurses, social workers, and chaplains. The amount, type, and site of care delivered in the model depends on the needs of patients and their families.

Palliative Care’s Relation to Hospice

Hospice is a subset of palliative care that is provided only in the last six months of life for patients who have chosen to cease curative treatments — eligibility rules that have been stable for over three decades in the Medicare program. The hospice benefit was implemented in 1983, and has proved a critical policy reform that has helped to mainstream a multidisciplinary, team-based approach to caring for the dying. The use of hospice has risen greatly from less than one in 10 Medicare decedents receiving such care in 1990 to almost five in 10 currently. All hospice care is palliative, but not all palliative care is hospice.

Provision of palliative care before hospice election has grown in larger hospital-based settings where illness is more acute and the costs of palliative care can be subsidized by other clinical areas, but there are gaps in the availability of palliative care in most other settings. Delivering palliative care further upstream in the disease progression before patients require inpatient care and before they reach a hospice election decision point may unlock the potential for realizing even greater improvements in patient outcomes and reductions in cost.

A palliative care payment model that reduces overall Medicare costs will likely increase Part B spending, while reducing Part A spending

Although we will not have preliminary results until the end of 2017 and final results until a year later, a key objective of our CMMI project is to estimate the magnitude of increased Part B payments needed to make upstream palliative care break even across sites of care, and then to determine if patients receiving palliative care in the model cost Medicare less compared to controls. Increased upstream palliative care spending can achieve overall cost savings only by reducing downstream hospitalizations and emergency department visit costs, which are financed by Medicare Part A. This general arithmetic is the essence of any palliative care payment model and essentially any care coordination approach that circumvents the high costs of acute care.

From the perspective of the Medicare program as a whole, reducing overall costs while maintaining or improving patient outcomes and quality of care would be a clear success. However, the differences in financing mechanisms between Medicare Part A and Part B mean that such a change, although cost-saving on the whole, will raise issues related to the redistribution of cost burden between taxpayers and beneficiaries.

Any consequential “increase in Part B to reduce Part A” will bolster the Part A Hospital Insurance Trust Fund, extending the solvency of the account funded primarily by payroll taxes. This development should be a win for the working population that funds the account and hopes to benefit from it whenever they reach Medicare eligibility. But Part A trust fund solvency is a diffuse benefit that is not widely understood by non-experts, and the threat of insolvency in more than a decade, while concerning to some politicians, is proof for others that no immediate intervention is needed.

Part B, on the other hand, is financed through general revenues and premiums paid by current Medicare beneficiaries. Therefore, any significant increase in Part B spending for palliative care could result in more immediate and salient demands on the federal budget and increases in beneficiary premiums, which must cover 25 percent of Part B costs, and they are updated annually. Calls from concerned Medicare beneficiaries are widely understood by members of Congress, and will surely begin the day that premium increases are announced.

This calculus will have critical implications for navigating political and popular support for palliative care.

Could palliative care be financed through an Alternative Payment Model?

Alternative Payment Models are designed to transition from fee-for-service to value-based payments, such as episodic bundling of care, in which a single “bundled” payment is made for all care delivered by multiple providers over a specified timeframe or “episode.” Bundled payment designs have been proposed for naturally discrete bundles of services such as elective joint replacement and maternity care. Hospice is arguably the original bundle, providing a per diem payment covering all Medicare-financed care for terminal illness. Similar to hospice, we will propose a bundled payment model for palliative care that is based on the utilization and cost results of our project. It is an intuitive design since it would provide a single payment to flexibly cover the costs of the multiple providers and care settings incorporated into the model.

However, the range of covered services and procedures under the hospice benefit is limited since patients elect to forgo curative treatments, and the median length of hospice use is around 17 days, with one in four beneficiaries using it for the last five days or less of life. By contrast, our CMMI project’s model seeks to engage patients up to one year or more before death. A “last year” perspective is a much longer episode timeframe than most bundled approaches tested so far and would have to cover the full complement of the Medicare benefit package. Constructing a single daily, weekly, or monthly, let alone one-year payment amount for palliative care would be complicated by large fluctuations in cost as patients’ care needs change and they transition through care settings.

Another challenge is determining what kind of entity could hold responsibility for such a bundle. A single provider entity would need sufficient capital reserves to operate with the risk of the costs of care for patients exceeding the single bundled payment amount, as well as the capacity to distribute the single payment across potentially many providers delivering the full suite of care benefits covered by Medicare. Fully integrated health systems such as accountable care organizations could take this on, but smaller provider entities like community hospices or physician groups lack the requisite financial resources, infrastructure, and experience.

An alternative payment model designed with a single coordinating entity but with sub-bundles of benefits held by provider groups best positioned to deliver care covered by the sub-bundle would provide market-specific flexibility. Bundles could be organized as follows, with disease-specific payment rates: primary care, secondary/specialist care, hospital, pharmacy/Part D, and post-acute including hospice (Table 1). This would enable smaller community-based hospice providers, such as Four Seasons, to hold portions of an end-of-life bundle and for Medicare to realize the potential for improved outcomes with lower costs associated with a coordinated approach to end-of-life care.

The End of Life and Serious Illness Health Affairs Blog series kicks off as two trends converge: interest in moving from fee-for-service payment toward purchasing based on value in CMS programs, and a cultural insistence that care at the end of life be more patient- and family-centered, a major theme of the 2014 National Academy of Medicine report Dying in America. These two trends collide in the Medicare program. The end of life presents a key opportunity for developing alternative payment models that increase the value of care. Given the high cost of the default end-of-life care system, lower costs with better outcomes for Medicare, patients, and families could all be attainable.

Table 1. Policy landscape for financing palliative care services at end of life in Medicare

Medicare Benefit
Part A
Hospital Insurance
Part B
Physician Services
(Medical Insurance)

Part C
Commercial Medicare Advantage

Part D
Prescription Drug Coverage

FinancingTrust Fund payroll tax and other sourcesPremiums with deductibles and general revenue (income tax)Commercial premiums with deductiblesGeneral revenue (income tax) & premiums with state contributions
ServicesHospital, skilled nursing, long-term care, hospiceDoctor visits, lab services, durable medical equipment, therapyPrivate A + B + (D) + additional benefits
• 30 percent population
• Hospice carved out
Prescription drugs
Cost triggersReduce unnecessary utilizationIncrease care coordination and goals of careUnknown; unavailable claims for researchSymptom management outlay vs. curative
Current movementHospice “two-tiered” payments with service intensity last seven daysAdvanced Care Planning CPT codes

Transitioned Care Management codes

Chronic Care Management PBPM

PILOT: Medicare Care Choices Model (test $400 PBPM concurrent care for hospice-eligible beneficiaries)
Aetna Compassionate Care program for under 65 commercial

Numerous proprietary coordinated/ palliative care management programs underway
Review of access, medication reconciliation, polypharmacy, and discontinuation issues
Potential bundles as APMHospital-based palliative care services

Post-acute care (90-180 days) prior to hospice palliative care services
Primary care (CCM, CPC+, PCMH medical homes) additive for palliative services in PBPM

Specialty care (CCM, medical home) additive in PBPM
Proprietary build on HCC risk score methodologyPharmacy/drug benefit during episode transitions (90-180 days) prior to hospice
Implication of ACO-MSSPProvider groups managing Total Cost of Care (Parts A, B, D) with increasing risk models and flexibility to deliver care across settings where financial control can be leveragedExcluded from MSSP; MA program innovation increasing but not publically sharedClustered resourcing as part of Total Cost of Care

Abbreviations: ACO=accountable care organization; APM=alternative payment model; CCM=chronic care management; CPC+=comprehensive primary care plus; HCC=Hierarchical Condition Category; MSSP=Medicare Shared Savings Program PCMH=patient-centered medical home; PBPM=per-beneficiary per-month