Dying in America: Improving Quality and Honoring Individual Preferences Near the End of Life (2015)

Appendix D

Financing Care at the End of Life and the Implications of Potential Reforms

Haiden A. Huskamp, Ph.D.¹

David G. Stevenson, Ph.D.²

As with other health care services, the manner in which end-of-life care is financed in the United States has a substantial impact on the care that is delivered. In the following paper, we examine the implications of financing and payment methods for end-of-life care for utilization, quality, and expenditures of individuals with advanced illness. After presenting context about service utilization, expenditures, and insurance coverage at the end of life, we highlight key limitations of current financing approaches, including incentives for overutilization, fragmentation, and inattention to quality of care. We discuss possible reforms to end-of-life care financing and the potential trade-offs involved, paying particular attention to bundled payment approaches and targeted changes to eligibility and payment policies. We conclude with broad guidance about factors to consider in advancing public policy in this important area.

FINANCING CARE AT THE END OF LIFE AND THE IMPLICATIONS OF POTENTIAL REFORMS

Introduction

As with other health care services, the manner in which end-of-life care is financed in the United States has a substantial impact on the care that is

_______________

¹Department of Health Care Policy, Harvard Medical School, 180 Longwood Avenue, Boston, MA 02115, huskamp@hcp.med.harvard.edu, 617-432-0838.

²Department of Health Policy, Vanderbilt University School of Medicine, 2525 West End Avenue, Suite 1200, Nashville, TN 37203, David.Stevenson@vanderbilt.edu, 615-322-2658.

delivered. Not only does coverage in public and private insurance programs shape the services for which individuals are eligible, but also the interaction between coverage and approaches to payment exerts a strong influence on the intensity, setting, and quality of care that is provided. In this appendix, we examine the implications of financing and payment of end-of-life care in the United States and discuss potential reforms and their possible impacts. We focus primarily on the Medicare program, given its prominent role in paying for and shaping end-of-life care. We detail relevant research findings where possible, while also identifying gaps in the knowledge base. We conclude with broad guidance about factors to consider in advancing public policy in this important area.

UTILIZATION AND EXPENDITURES FOR INDIVIDUALS AT THE END OF LIFE

Approximately 2.5 million individuals die each year from a variety of causes, including sudden acute illness, accident, suicide, homicide, and long-term chronic conditions (Minino, 2013). Approximately three-quarters (74 percent) of deaths occur among persons aged 65 and older (Minino, 2013). In contrast to nonelderly decedents, death among the elderly is most likely to occur after a diagnosis of advanced chronic illness as opposed to a sudden, unexpected death. Of all deaths in 2011 among persons aged 65 and older, 26 percent were due to heart disease, 22 percent to cancer, 7 percent to chronic lower respiratory diseases, 6 percent to stroke, and 5 percent to Alzheimer’s disease (Minino, 2013). For the one-quarter of all deaths among the nonelderly, a much larger proportion is due to accidents, homicide, suicide, or acute episodes of illness. For example, 38 percent of all 2011 deaths among those aged 1-24, 26 percent among those aged 25-44, and 7 percent among those aged 45-64 were due to accidents (Minino, 2013).

Services Used at the End of Life

Individuals suffering from advanced illness may use a variety of life-prolonging and palliative services after their diagnosis. While the inpatient share of Medicare expenditures for decedents aged 65 and older in a given year has dropped dramatically over the past 30 years, inpatient care still accounted for half of all Medicare spending—50.2 percent—among elderly decedents in 2006 (see Table D-1) (Riley and Lubitz, 2010). Physician services accounted for 18.8 percent, skilled nursing facility (SNF) care for 10.4 percent, hospice for 9.7 percent, other outpatient services for 6.8 percent, and home health care for 4.1 percent (Riley and Lubitz, 2010). Although inpatient spending as a proportion of total spending at the end of life has

TABLE D-1 Percent of Medicare Payments by Service Type for Medicare Decedents Aged 65 and Older

SOURCE: Riley and Lubitz, 2010.

declined substantially along with an increased role for hospice care, it is important to note that many Medicare beneficiaries still have intensive service use at the end of life (Barnato et al., 2004; Riley and Lubitz, 2010). In fact, many Medicare beneficiaries access the hospice benefit only after spending time in the hospital and the intensive care unit (ICU), and then do so only for short periods of time (Teno et al., 2013).

Approximately one-quarter of Medicare spending is incurred by individuals in their last year of life, a proportion that has remained virtually unchanged since the late 1970s (Riley and Lubitz, 2010). Focusing only on Medicare expenditures provides a narrow picture of health care spending at the end of life, particularly for nursing home residents, approximately two-thirds of whom are dually eligible for Medicare and Medicaid and receive Medicaid-financed long-term services and supports. While some have pointed to the relatively large share of Medicare spending devoted to care in the final year of life to support an argument that growth in health care spending is driven largely by the increased use of high-cost aggressive treatment for individuals near death, Scitovsky (1984) argued that changes in technology and intensity of treatment had not disproportionately affected utilization and spending for patients at the end of life relative to other elderly Medicare patients who are ill, and thus that spending should not be a disproportionate target of cost containment efforts. Lubitz also concluded that Medicare beneficiaries in their final year of life did not account for a larger share of Medicare expenditures after the emergence of new medical technologies, demonstrating that the dying were utilizing care in ways similar to those of other patients (Lubitz and Riley, 1993). It is also important to note that calculations of spending in the last year of life can be made only by looking backward from the decedent’s date of death. These calculations do not necessarily reflect “real-time” decision making by patients and families about care in the final year of life, as 1-year survival is extremely difficult to predict.

Elderly individuals with advanced illness and their families face considerable financial risk from out-of-pocket health care expenditures in the final years of life (Kelley et al., 2013b). Studying elderly Medicare beneficiaries who died between 2002 and 2008, Kelley and colleagues found that average out-of-pocket spending in the 5 years before death was $38,688 (in 2008 dollars). Expenditures were highly skewed, however, with a 90th percentile of $89,106. For one-quarter of decedents, out-of-pocket expenditures exceeded the household’s baseline total assets. Reaffirming the importance of including long-term services and supports in calculations of spending at the end of life as mentioned above, nursing home costs accounted for half of the expenses for those in the top quartile of spending (Kelley et al., 2013b).

Variation in Utilization and Spending at the End of Life

Although spending on end-of-life care is uniformly high, the Dartmouth Atlas documented substantial geographic variation in use of end-of-life care services and spending by hospital referral region (HRR) over time, which researchers and policy makers viewed as evidence of wide regional differences in physician practice patterns (Goodman et al., 2011). For example, in 2007, the average number of days spent in an ICU for chronically ill Medicare beneficiaries in the last 6 months of life varied from 0.7 in Minot, North Dakota, to 10.7 in Miami, Florida (Goodman et al., 2011). In this same population, the percentage dying in a hospital varied from 12.0 percent in Minot, North Dakota, to 45.8 percent in Manhattan, New York, and the average number of days spent enrolled in hospice varied from a low of 6.1 in Elmira, New York, to a high of 39.5 in Odgen, Utah (Goodman et al., 2011).

A July 2013 report by the Institute of Medicine’s Committee on Geographic Variation in Health Care Spending sheds new light on the literature on variations in health care spending and utilization at the end of life (IOM, 2013). First, the report documents large variation in health care spending at all levels of geography studied, including HRRs; hospital service areas; core-based statistical areas (CBSAs); physician practices; and even the level of an individual physician after controlling for demographic characteristics, insurance plan factors, and market-level characteristics. Importantly, the variation in total Medicare spending was driven largely by the utilization of post-acute services, including SNF services, home health care, hospice, inpatient rehabilitation, and long-term acute care; if there were no variation in post-acute care expenditures, then variation in total Medicare spending would decrease by 73 percent (IOM, 2013). One potential implication of these findings is that the integrated payment and bundled payment demonstrations we describe below (e.g., where acute, post-acute, and other health care services are more integrated in their financing and delivery) could have

greater relevance than simple geographic adjustments to administered payments alone in introducing efficiencies into the health care system.

WHO PAYS FOR CARE AT THE END OF LIFE?

As of 2011, the majority of the U.S. population (54 percent) had private health insurance coverage either through an employer (49 percent) or an individual/nongroup policy (5 percent) (KFF, 2011). Sixteen percent were covered by the Medicaid program, 13 percent by Medicare, and 1 percent by other public programs, with approximately 16 percent being uninsured (KFF, 2011). Although the Patient Protection and Affordable Care Act (ACA) will likely decrease the number of uninsured individuals and increase the proportion who have Medicaid and private coverage beginning in January 2014, the Medicare program is—and will remain—the predominant payer for end-of-life care in the United States, primarily because of the older ages at which most Americans die. Moreover, Medicare’s role in shaping end-of-life care in the United States likely goes beyond the proportion of Americans who die as Medicare beneficiaries, given that older individuals are disproportionately likely to die from advanced illness as opposed to an accident or sudden acute event.

With some exceptions, such as the Medicare hospice benefit (described later in this appendix), insurance coverage for individuals with advanced and terminal illnesses reflects coverage that is available to enrollees more generally. The traditional Medicare program covers a broad range of preventive, acute, and post-acute care services for approximately 49 million beneficiaries (KFF, 2012d). Medicare Parts A and B cover hospital services; post-acute SNF care, home health, and rehabilitative services; physician services; durable medical equipment; and ambulance services. For those who elect to join a Part D plan, outpatient prescription drugs are also covered. Medicare’s biggest coverage gap relates to the lack of coverage for long-term services and supports. In addition, cost-sharing requirements can be substantial for patients with high levels of service use. Approximately 12 million of the 49 million Medicare beneficiaries are enrolled in Medicare Advantage plans, which receive a capitated payment to cover all Part A and B services in addition to any supplemental services that the plan chooses (e.g., dental, vision) (KFF, 2012c). Of particular relevance to care at the end of life, Part A has covered hospice since 1983 for individuals with terminal illnesses who have an expected prognosis of 6 months or less and who agree to forgo curative treatment for the terminal condition.

Both Medicaid and private insurance plans typically cover a set of services similar to those covered by Medicare, although there is some variation. All state Medicaid programs are required to cover a set of man-

dated services, including hospital services (both inpatient and outpatient), physician services, home health services, and (unlike Medicare) long-term services and supports. State Medicaid programs are permitted—but not required—to cover prescription drugs, hospice, and personal care services. While an optional benefit, all 50 states and the District of Columbia cover prescription drugs, and all but one—Oklahoma—include hospice care as a covered benefit for adults. In early 2013, the Louisiana Department of Health and Hospitals announced plans to discontinue Medicaid coverage of hospice services as part of a broad set of budget cuts intended to balance the state’s budget (Adelson, 2012). After opposition was raised, these plans were dropped (Adelson, 2013).

Private insurance coverage, including both covered benefits and cost-sharing requirements, varies greatly by plan, with some policies offering generous coverage and others offering more limited benefits. Private plans that will be offered through the ACA-created health insurance exchanges must cover services in 10 broad categories of “essential health benefits” (EHBs), using the state’s specified “benchmark” plan (a private plan marketed in the state) as the guide for the generosity of coverage of these 10 types of services. States may also choose to mandate that exchange plans marketed in the state cover additional services. Hospice does not fall under one of the broad categories of EHBs that must be covered by all plans offered on the exchanges, so states are not required to insist that exchange plans cover it. Although hospice is currently covered by the benchmark plans for all 50 states and the District of Columbia, only 11 of the 50 states and the District of Columbia require hospice as a covered benefit for all plans offered through the exchanges.

KEY LIMITATIONS OF CURRENT FINANCING APPROACHES

The payment approaches used most commonly by Medicare, Medicaid, and commercial payers have a number of limitations that can contribute to suboptimal care at the end of life. Given its prominence in financing end-of-life care in the United States, we focus primarily on the role of Medicare, including the Medicare hospice benefit.

The Traditional Medicare Program

As noted above, Medicare finances health care for around 70 percent of the individuals who die each year in the United States. The program does not finance all services used at the end of life (as noted, for example, it does not cover long-term supportive services), but it pays for the vast majority of acute medical care and hospice that beneficiaries receive and thus plays a substantial role in shaping how health care providers deliver care for those who die. Although it can be difficult to discern the extent to which financing

and payment alone lead to shortcomings in the provision of end-of-life care, researchers and other stakeholders often point to systemic incentives of the traditional Medicare program when discussing challenges such as burdensome, high-intensity treatments delivered at the end of life; fragmentation across payers and settings, which often leads to poor coordination of care; and benefit design features that inhibit delivering care that is in the best interest of patients.

Issues with the General Financing Approach

Excessive health care utilization at the end of life can be burdensome for patients and offers little clinical value. Previous studies of the Medicare population have shown high rates of hospitalization and use of intensive procedures at the end of life (Hogan et al., 2001; Kwok et al., 2011; Lubitz and Riley, 1993; Teno et al., 2013). In addition to the cultural and professional norms that shape physician behavior, a key determinant of older patients’ end-of-life care may relate to the fee-for-service (FFS) payment system that is the foundation of reimbursement for services used by the nearly three-quarters of Medicare beneficiaries enrolled in the traditional Medicare program (IOM, 1997). FFS payment provides incentives to deliver more—and often more aggressive—care, and can lead to fragmentation in financing and delivery. More generally, the predominance of FFS payment in U.S. health care financing is often identified as a key impediment to addressing problems of low-value, poor-quality care, as well as rapidly growing health care expenditures (Schroeder and Frist, 2013).

Beyond incentivizing the delivery of more services and procedures, the traditional Medicare program historically has done little to encourage coordination across settings or benefit categories. Medicare generally uses a “silo” approach to reimbursement, employing a separate payment approach for a given provider type for a specific type of service. This approach ignores interrelationships between providers and the care they deliver and can distort clinical decision making. For example, hospitals and post-acute care providers are paid prospectively established rates for inpatient, SNF, and home health care. Hospitals receive a fixed payment for an inpatient hospital stay based on the diagnosis-related group (DRG) methodology, regardless of length of stay or costs incurred. The DRG payment system encourages hospitals to discharge patients as early as possible; in fact, implementation of the DRG system was associated with a decline in hospital lengths of stay and more frequent rehospitalizations (Lave, 1989). In contrast, nursing homes receive a fixed per diem amount for SNF, based on residents’ resource utilization groups (RUGs) acuity score, an approach that incentivizes providers to capture fully residents’ acuity and therapy needs but not necessarily to limit their lengths of stay. In somewhat of a hybrid of these two approaches, home health agencies are paid a prospectively deter-

mined rate for 60-day episodes of care. Importantly, none of these payment approaches gives providers any incentive or mechanism to coordinate the services individuals receive, despite the interrelated trajectories of patients between these settings.

The Medicare program’s fragmented approach to payment interacts in negative ways with its largely disjointed approach to determining eligibility across service categories (e.g., eligibility and payment are defined separately for post-acute care services such as SNF care and home health care, despite overlap in populations served and services offered). This fragmented dynamic is confounded further when individuals are dually eligible for Medicare and Medicaid. The classic example of this disjuncture and its potential negative impact on beneficiaries is the perverse incentive that nursing homes have to hospitalize dually eligible residents who, in some instances, could be treated more successfully and efficiently in the nursing home. The financial incentive to shift residents onto the Medicare SNF benefit where possible is created by the disparity between Medicare SNF payments and the generally much lower Medicaid nursing home room and board payments. This challenge has received increasing attention in recent years as the financial and health costs of avoidable hospitalizations and rehospitalizations have become better understood, and interventions and strategies to address them have become more widespread (Lipsitz, 2013; Segal, 2011; Teno et al., 2013).

A related dynamic of particular importance to end-of-life care is the potential barrier the SNF benefit can present for nursing home residents’ enrollment in hospice (residents may not enroll in hospice while receiving Medicare-financed post-acute care, unless the two services are treating distinct conditions). In particular, nursing homes and beneficiaries face financial disincentives to enrollment in Medicare hospice instead of Medicare-financed SNF care when both are an option. For individuals being discharged from the hospital who are eligible for either the hospice or SNF benefit, the nursing home generally receives much lower reimbursement for hospice-enrolled residents (whose nursing home care is typically paid for by Medicaid) relative to residents who are utilizing the Medicare SNF benefit. Moreover, residents not Medicaid eligible are liable for paying room and board costs if they choose hospice instead of SNF care (dually eligible residents are not subject to cost sharing for nursing home care during a hospice/long-term care stay or SNF stay). Calculating precise numbers of residents enrolled in SNF care who could benefit from earlier admission to hospice is difficult; however, previous research has identified a sizable minority of individuals who transition from SNF care to hospice within 1 day of SNF discharge, possibly suggesting that financial factors influence the timing of referral (Aragon et al., 2012; Hoffmann and Tarzian, 2005; Miller et al., 2012; Zerzan et al., 2000). The clinical implications of these incentives for

residents and the nature of the transition from skilled-rehabilitative care to hospice care are unclear.

The Hospice Benefit as the Primary Financing Mechanism for Palliative Care

The primary mechanism for financing palliative services in the Medicare program (and, for that matter, in most state Medicaid programs and commercial insurance plans) is the hospice benefit. As discussed above, a beneficiary is eligible for the Medicare hospice benefit only if two physicians (one of whom can be the hospice physician) certify that the individual has a prognosis of 6 months or less should the illness run its natural course and if the beneficiary agrees to forego treatments intended to cure the illness or prolong life. These two requirements often limit timely enrollment in the hospice benefit, especially in the context of how the benefit is currently used. Defining hospice eligibility relative to the 6-month prognosis mark can be quite difficult, especially for individuals with noncancer diagnoses (Christakis and Lamont, 2000; Sachs et al., 2004). Moreover, limiting hospice to individuals who agree to forego curative therapies creates an artificial distinction between potentially life-prolonging and palliative therapies and could impede both enrollment and quality of care (Meier, 2013; Temel et al., 2010).

Together these eligibility requirements can serve to delay or prevent enrollment in the Medicare hospice benefit for some beneficiaries, effectively denying them access to palliative care services. Of the almost half (45.2 percent in 2011) of Medicare decedents who use the hospice benefit before their death, approximately one-quarter enroll 5 or fewer days before death (MedPAC, 2013), a period that most agree does not allow the individual or his/her caregivers to obtain the full benefits of hospice services (Bradley et al., 2004; Iwashyna and Christakis, 1998; Kelley et al., 2012; Taylor et al., 2007). Equally troubling is that many short-stay hospice users enroll in the benefit only after a hospitalization, and often after a hospitalization that includes an ICU stay. For example, one study estimates that 40 percent of individuals who used hospice for 3 or fewer days in 2009 had a hospitalization with an ICU stay prior to hospice admission (Teno et al., 2013). In other words, even though an increasing number of Medicare beneficiaries are using the hospice benefit, many do so only after exhausting high-intensity services.

Although palliative care can be introduced at any point in a person’s illness to manage symptoms and maximize quality of life, Medicare offers little explicit coverage of palliative care outside of hospice. The Medicare Prescription Drug Improvement and Modernization Act of 2003 (the MMA) authorized a one-time payment to a hospice for evaluation and

counseling services provided by a hospice physician for a beneficiary who has not elected the hospice benefit and has a prognosis of 6 months or less (fiscal year [FY] 2005 payment rate = $54.57). In addition, physicians may provide some palliative care consultation in the context of physician services financed by Medicare Part B. While these options provide some financing for physician discussions with patients about end-of-life care preferences and planning, the often time-intensive discussions are poorly reimbursed, in part because of lower value units assigned to the provision of evaluation and management services relative to the provision of procedures under the Resource Based Relative Value Scale (RBRVS) used to determine Medicare physician reimbursement rates for different types of services. Outside of the hospice benefit and these narrow provisions for physician consultation, there is no direct financing stream for palliative care services under the Medicare program (CAPC, 2009). Yet despite the lack of direct financing, many U.S. hospitals currently offer palliative care consultation services. Researchers report that the number of such programs in hospitals with 50 or more beds increased from 658 (24.5 percent) to 1,486 (58.5 percent)—a 125.8 percent increase—from 2000 to 2008 (CAPC, 2010).

Legislation has recently been introduced in both houses of Congress that would provide Medicare and Medicaid reimbursement for advance care planning. In March 2013, Congressman Earl Blumenauer introduced the Personalize Your Care Act (HR 5795), which would provide Medicare and Medicaid coverage for voluntary consultations with health care professionals about advance care planning every 5 years or after a change in health status. In August 2013, Senators Mark Warner and Johnny Isakson introduced the Care Planning Act (S 1439), which would provide Medicare and Medicaid coverage for voluntary discussions about treatment goals and options for patients with advanced illness that result in a documented care plan. The prospects for passage of these two bills are unclear.

Specific Concerns About the Current Hospice per Diem Payment Approach

Currently, for 97 percent of Medicare-financed hospice days, Medicare pays the hospice agencies an all-inclusive per diem payment (i.e., the FY 2014 routine home care rate of $156.26 per day, which is adjusted for differences in local wage rates) to provide all care related to the terminal condition; the other 3 percent of days are billed under one of three other allowable categories (continuous home care, general inpatient care, or inpatient respite) (MedPAC, 2013). In contrast to Medicare reimbursement for other providers, hospice payments are not adjusted for case mix or setting of care, and there is no provision for additional reimbursement for particularly high-cost cases (i.e., outlier payments) (Huskamp et al., 2010).

While this approach to payment may have been appropriate when hospice was used almost exclusively by cancer patients living at home (i.e., when the benefit was first implemented in the 1980s), it is less sufficient in the context of the much greater diversity in the diagnoses and settings of current hospice users, and of hospice providers as well.

The Medicare Payment Advisory Commission (MedPAC) and others have raised a number of concerns about the current payment approach for the Medicare hospice benefit. First, while hospice costs are generally higher for the first and last days of a hospice stay, the routine home care per diem payment is uniform throughout the stay, making longer stays more profitable (Huskamp et al., 2001, 2008, 2010). MedPAC has documented dramatic increases in mean length of stay over the past decade, driven largely by increases in duration of very long hospice stays (MedPAC, 2012). Importantly, these increases are not just a result of the greater portion of hospice users with noncancer diagnoses, as increased lengths of stay extend across diagnosis categories (MedPAC, 2012). In 2009, MedPAC recommended that the payment system be changed such that per diem payments are higher for days at the beginning and end of a hospice stay and lower for the middle days as length of stay increases (MedPAC, 2009). Second, the structure of the hospice benefit—with no adjustments for particularly high-cost stays—limits access for individuals with high-cost palliative care needs (Huskamp et al., 2001; Lorenz et al., 2004). A national survey of hospices found that 78 percent had at least one enrollment policy that could restrict access for individuals with high-cost palliative care needs (Carlson et al., 2011). Third, 18 percent of Medicare hospice stays in 2010 ended in live discharge from hospice, raising concerns about the quality of care received by these beneficiaries and questions about whether hospices are following the eligibility criteria for the benefit (MedPAC, 2013). Fourth, increasing numbers of hospice recipients are using the much more expensive general inpatient (GIP) care category of hospice services, raising questions about the appropriateness of such use (HHS OIG, 2013). Finally, hospice costs are lower on average for hospice users living in nursing homes than for those in the community, suggesting potential efficiencies in joint management of care by the hospice and nursing home (HHS OIG, 1997, 2013; Huskamp et al., 2010; MedPAC, 2013). In addition, hospice staff members provide more aide visits but fewer nurse visits to nursing home residents than to community-based residents, raising questions of duplicative payments because room and board fees paid by Medicaid or by patients themselves are intended to cover aide services needed by residents (Miller, 2004).

Researchers have argued that the current structure of the Medicare hospice benefit may be a particularly poor fit in the nursing home setting. Beyond the barrier that the SNF benefit can create for nursing home residents’ enrollment in hospice and the need to pay appropriately for services

that can overlap with nursing home care discussed above, several features of the nursing home population pose challenges vis-à-vis the hospice benefit: diagnoses of noncancer terminal conditions (for which, as noted, prognostication can be even more difficult than for cancer patients) are typically more common than in the community; levels of cognitive impairment are often high, and many residents do not have family members involved in their care to assist with the hospice election process; and physicians are often based off site, making it more difficult to discuss hospice with patients and family members (Huskamp et al., 2010).

Medicare Managed Care

Relative to the traditional Medicare (TM) program described above, providers serving the nearly 30 percent of beneficiaries enrolled in the Medicare Advantage (MA) program may be better positioned to promote the use of recommended services at the end of life while discouraging the use of unnecessary invasive procedures (Stevenson et al., 2013). MA plans generally are paid on a per-person—rather than per-service—basis, thereby rewarding plan efforts to manage chronic disease and to minimize unnecessary treatment intensity at the end of life. Importantly, hospice is one of the few benefits “carved out” of Medicare’s managed care program. When managed care enrollees enter hospice, FFS Medicare becomes the payer for both hospice care and care unrelated to the terminal condition; health plans remain liable only for any supplemental benefits they provide beyond those in TM, such as vision or dental care. This policy creates a strong financial incentive for plans to promote hospice enrollment among their more expensive terminally ill enrollees, while also diminishing—at least somewhat—incentives to develop integrated, high-quality palliative care networks for people with advanced illness. MedPAC voted in January 2014 to end this hospice carve-out policy, recommending that MA plans begin to cover hospice services for the first time. The likelihood that this recommendation, to be released in the March 2014 report, will be implemented remains unclear.

Previous studies using data from the 1990s confirmed higher rates of Medicare hospice enrollment in managed care versus TM while concluding that this elevated use did not appear inappropriate (McCarthy et al., 2003; Riley and Herboldsheimer, 2001; Virnig et al., 2001). Yet these data are now almost two decades old and preceded passage of the MMA, which has led to markedly increased enrollment of Medicare beneficiaries in managed care plans (Afendulis et al., 2012). Outside of comparing hospice enrollment between MA and TM enrollees, few studies have characterized the intensity or quality of end-of-life care in the MA program. One recent study analyzed end-of-life care for MA and TM decedents matched on

age, sex, race/ethnicity, and geography (Stevenson et al., 2013). Although the study could not assess the appropriateness of service use or the quality of end-of-life care delivered, its findings suggest that MA plans may do a better job of minimizing high-intensity procedures at the end of life. MA enrollees used hospice more frequently at the end of life than those being cared for in traditional Medicare, although this difference narrowed over the 2003-2009 study period. After accounting for differential enrollment in hospice, MA enrollees also used fewer inpatient services overall and had markedly lower emergency department use at the end of life compared with matched TM enrollees.

Other Integrated Financing Models

In addition to the MA program, other approaches to integrated financing and delivery have relevance for Medicare beneficiaries with advanced illness and offer potential advantages over Medicare’s traditional FFS program. For instance, the Program of All-Inclusive Care for the Elderly (PACE) is an integrated model of financing and delivery for dually eligible older people who have nursing home–level clinical needs while also having the potential to be cared for in the community with adequate supports. Providers receive capitated payments from Medicare and Medicaid and offer enrollees comprehensive, interdisciplinary care across the health care continuum. A number of evaluations have assessed the potential of PACE to keep enrollees out of nursing homes and hospitals and to generate savings (Mukamel et al., 2007); however, few have focused specifically on the potential of such models to improve care at the end of life. Although studies have noted some benefits to PACE enrollment with respect to end-of-life care (e.g., reduced hospitalizations, improved end-of-life care planning, patient-centered care) (Famakinwa, 2010; Mukamel et al., 2002), wide variation also has been observed across PACE sites. Other integrated models of financing—including MA Special Needs Plans, focused on specific, high-risk populations, and more recent state integrated care demonstrations—have similar potential to improve the coordination of care. However, little solid evidence has documented the fulfillment of this potential in these types of models to date (Grabowski, 2007, 2009).

General Lack of Focus on Quality of Care in Current Financing and Regulatory Approaches

Whether in the TM or MA context, an important impediment to improving the financing and delivery of end-of-life care for beneficiaries is the lack of established quality measures. It has become increasingly common for both public and private payers to include in provider contracts financial

incentives for providers to meet specified performance standards in an effort to improve the quality of care delivered to a population. However, existing contracts that include performance standards rarely include standards relevant to the provision of high-quality end-of-life care. For example, the Medicare accountable care organization (ACO) Shared Savings Program (MSSP) and Pioneer ACO demonstrations (discussed later) include 33 performance metrics in contracts with ACOs, not one of which is related to the provision of high-quality end-of-life care (Architecture for Humanity, 2012). Similarly, Blue Cross Blue Shield of Massachusetts’ (BCBSMA’s) Alternative Quality Contract, an innovative model that gives large provider organizations a global payment that covers all care plus bonuses of up to 10 percent of the global payment for meeting specified performance standards, includes a total of 64 performance standards (32 related to ambulatory care and 32 to hospital care), none of which focuses on end-of-life care (BCBSMA, 2010).

One sees a similar lack of focus on palliative and end-of-life care in public reporting systems that are intended to assist patients in selecting high-quality providers while also encouraging providers to improve quality of care. For instance, although the National Quality Forum (NQF) recently endorsed a set of quality measures with relevance to palliative and end-of-life care (NQF, 2012), the Healthcare Effectiveness Data and Information Set (HEDIS) measures that are used to assess and monitor MA plans historically have not included such measures. Although nearly 28 percent of Americans die in a nursing home (Teno et al., 2013), the Nursing Home Compare website reports information on few clinical or other measures relevant for assessing the quality of end-of-life care (it does report the percentage of long- and short-stay residents who self-report moderate to severe pain), instead focusing on measures of functional outcomes (Huskamp et al., 2012). The lack of measures appropriate to end-of-life care on the Nursing Home Compare website may not be surprising given the deemphasis on such measures in nursing home inspection surveys, which focus more on measures related to restoration and maintenance of function for residents. Yet the almost exclusive focus on the latter types of measures has the potential to impede appropriate end-of-life care for residents, because some measures that may address natural symptoms experienced in the dying process (e.g., functional decline, weight loss, dehydration) could be interpreted as implying poor-quality nursing home care (Huskamp et al., 2012). Perhaps as a further indication of this focus, the most recent version of the nursing home resident assessment form (the Minimum Data Set, Version 3.0) drops any reference to advance care planning, something that was included in the previous iteration.

Although progress has been made in developing quality measures for end-of-life care, as evidenced most prominently by the recent NQF endorse-

ment (NQF, 2012), incorporating these measures into provider payment and oversight will be essential as health care is shaped increasingly by more integrated financing and delivery systems such as MA plans and newer innovations such as ACOs and patient-centered medical homes. Policy development in this area needs to ensure adequate provider networks for patients (e.g., including access to palliative care specialists), suitable quality measurement for oversight, and sufficiently flexible financial incentives to foster coordination of care and mitigate incentives for selection or for stinting on needed care.

As required by Section 3004 of the ACA, all Medicare-certified hospice agencies must report a set of hospice quality measures starting in FY 2014 or face payment reductions (a 2 percentage point decrease in the market basket update for that year). These measures include information about pain screening and assessment, dyspnea screening and assessment, the percentage of opioid users who are offered or prescribed a bowel regimen, and documented discussions about treatment preferences and patients’ beliefs and values. The ACA also stipulates that hospice quality measures ultimately will be publicly reported (the timetable has yet to be announced), as the Centers for Medicare & Medicaid Services (CMS) already does with quality measures for other types of Medicare providers.

POSSIBLE FINANCING REFORMS AND POTENTIAL IMPLICATIONS FOR END-OF-LIFE CARE

A number of payment and delivery reforms that are planned or currently under way would impact the utilization, cost, and quality of services for patients with advanced illnesses, perhaps addressing some of the problems described above, but also possibly introducing other concerns. Some of these reforms are being implemented as Medicare demonstration programs authorized under the ACA; others are being considered or adopted more broadly in the Medicare program, the commercial market, and/or state Medicaid programs.

Bundling of Payments to Providers

There is now broad national interest on the part of payers, including private insurers, Medicare, and Medicaid, in identifying alternatives to FFS payment (Kirwan and Iselin, 2009; Schroeder and Frist, 2013). Growing attention is focusing on bundled payment models as an alternative that could lead to lower expenditures and increased efficiency (Cutler and Ghosh, 2012). Instead of reimbursing each provider individually for every service delivered to a patient, bundled payment models entail payments for bundles of related services. The bundle of services can be defined relatively nar-

rowly (e.g., to include both physician and nonphysician services delivered during an acute inpatient stay) or more broadly (e.g., to include all acute inpatient care and post-acute care related to an index hospitalization), with the broadest bundle including all services provided to an individual over the course of 1 year (i.e., a global budget).

Depending on how they are structured, models that bundle payments across types of providers have the potential to remove some existing incentives related to setting of care that result from a “silo-based” payment system (e.g., a nursing home’s incentive to hospitalize residents to receive higher payments upon discharge). Bundled payments can also give providers greater flexibility to tailor service delivery, as well as incentives for improved coordination of care (Hackbarth et al., 2008). However, these arrangements also raise some concerns. Paying health care providers a fixed fee to cover a bundle of services provides strong incentives for the efficient delivery of services within the bundle. However, this arrangement also creates incentives for providers to increase the volume of bundles delivered as a way to increase revenue, and encourages them to direct care to services not included in the bundle where possible, limiting the potential for savings (Weeks et al., 2013). More important, relative to an FFS system, paying providers a fixed rate that covers a bundle of services can create incentives for them to stint on care or attempt to select patients who have lower-than-average expected costs, creating potential access problems for relatively sicker patients—something that is seen currently in the Medicare hospice benefit (Aldridge et al., 2012). Even with risk adjustment methods to adjust the bundled rate for observable characteristics related to a patient’s expected costs, variation in expected spending across individuals within a bundle that is not accounted for by risk adjustment methods will remain, creating incentives for selection. Also, bundling payment for services delivered by different types of providers conceivably could restrict the patient’s choice of provider (Sood et al., 2011). For example, if Medicare began paying hospitals a bundled rate for acute and related post-acute care, hospitals might limit their network of post-acute providers in order to negotiate favorable rates (in exchange for volume of patient referrals) with a subset. As discussed in more detail below, a more recent emphasis of bundled payment approaches is the incorporation of quality metrics to incentivize providers to balance quality and efficiency concerns in their provision of care.

In response to the strong incentives that would be created by a pure bundled payment approach for these services, Feder has called for caution in implementation, suggesting a hybrid payment approach that involves the sharing of both risk and savings on the part of providers (Feder, 2013). This type of mixed payment method, which has been proposed for use in health care reimbursement for decades (Newhouse, 1994), could help temper incentives for selection and stinting while still creating an incen-

tive for increased efficiency (in a sense, splitting the difference between the positive and negative incentives created). Others have called for an outlier payment system, as is used in the Medicare reimbursement system for acute inpatient care (Carter et al., 2012). In addition, most agree that performance measures with financial incentives for achieving high-quality care should be included in efforts to bundle payments (see the discussion below) (Schroeder and Frist, 2013; Sood et al., 2011).

Medicare demonstrations of two important models that involve the bundling of payments have recently been undertaken: (1) the Medicare Bundled Payments for Care Improvement Initiative (BPCII) and (2) the MSSP and the Pioneer ACO Program.

In January 2013, CMS announced the provider organizations that will participate in the BPCII, a demonstration that will test several models for bundling provider payments. These models include one that bundles all post-acute services delivered after hospital discharge, one that bundles an acute inpatient admission with post-acute care delivered within 180 days of discharge, and one that bundles an acute inpatient admission with any related readmissions within 30 days of discharge (CMS, 2013a).

The BPCII is not Medicare’s first effort at bundling payments. In the 1990s, Medicare conducted the Heart Bypass Center Demonstration, which paid seven hospitals an all-inclusive, per-discharge, bundled rate covering all hospital and physician services provided during the inpatient stay for coronary artery bypass graft (CABG) surgery. Over the 5-year demonstration period (1991-1996), Medicare saved approximately 10 percent on in-hospital spending for CABG surgery recipients, with no evidence of a worsening in health outcomes (Health Care Financing Administration, 1998). In January 2011, CMS implemented a change in the prospective payment system for end-stage renal disease (ESRD), which involved expanding the bundle of services for which dialysis providers are paid to include dialysis-related lab tests and injectable medications such as relatively high-cost erythropoiesis-stimulating agents (ESAs, which some argued were being overused), iron, and vitamin D analogs (Chambers et al., 2013). In addition to the expansion of the bundle, Congress required a Quality Incentive Program (QIP), which reduced payments to providers that failed to meet certain performance standards. Preliminary analyses of the impact of the changes suggest that ESA use dropped by approximately 15-20 percent immediately after implementation, while use of home-based therapies such as peritoneal dialysis increased (Collins, 2012; Fuller et al., 2013; Gilbertson et al., 2012). These data are preliminary and do not elucidate the impact on health outcomes; a longer-term, more detailed study is needed. Nevertheless, these early findings suggest that a change in the bundle of services could have important implications for both spending and quality of care.

While the experience of these previous demonstrations is informative, it does not necessarily generalize to the likely impacts of many of the bundling models being implemented or discussed today. Both the CABG and ESRD bundling demonstrations focused on relatively narrow bundles of services delivered by a single provider organization, while many of the current bundling models would involve bundling services across multiple types of providers.

The MSSP and the Pioneer ACO Program, both implemented in 2012, define a broad service bundle—all services financed by Medicare Parts A and B—in setting spending targets for participating organizations. Under the MSSP, implemented in 2012, participating ACOs are put at risk in one of two ways: (1) use of a “one-sided” approach to risk, whereby the ACO shares in any savings achieved relative to the benchmark spending level (calculated using Medicare Part A and B spending data from the previous 3 years, inflated to the performance year for beneficiaries assigned to the ACO) but is not subject to risk if expenditures exceed the benchmark; or (2) use of a “two-sided” approach to risk, whereby the ACO shares in any savings achieved while also sharing responsibility for spending that exceeds the benchmark (Cao et al., 2013). For the one-sided model, ACOs share in up to 50 percent of any savings, depending on their performance with respect to the 33 performance measures related to quality that are used by the program. ACOs that accept the two-sided model can share in up to 60 percent of savings, again depending on the achievement of performance standards. To date, only a small number of MSSP ACOs have opted for two-sided risk sharing, although current regulations state that all will be required to accept downside risk in the second contract period.

The Pioneer ACO model is an alternative to the MSSP that CMS is testing in 32 “advanced” organizations with experience operating under ACO-type risk-sharing arrangements (CMS, 2012). With the Pioneer model, ACOs can achieve higher levels of rewards and are subject to higher levels of risk than is the case under the MSSP during years 1 and 2 of the demonstration. Starting in year 3, ACOs are eligible to receive a prospective, per-beneficiary per-month payment instead of some or all FFS payments received under the current system (CMS, 2012).

In some respects, ACOs have a strong incentive to adopt care management practices that optimize palliative care and hospice use for individuals with advanced illness, provided these individuals remain assigned to the ACO (i.e., by using sufficient services with an ACO-contracted primary care physician). More specifically, previous studies have shown that early integration of palliative care for individuals with advanced illness has the potential to reduce health care costs overall (Morrison et al., 2008). At the same time, this incentive is tempered by the ability of ACOs to refer individuals outside of the ACO, for example, to a hospice agency or a primary

care provider who does not contract with the organization. One key indicator of how ACOs respond to these incentives will be determined by the adequacy of the provider networks available for patients within the ACO with advanced illnesses (e.g., including access to palliative care specialists). A related indicator is the timing of palliative care and hospice utilization for individuals at the end of life. For instance, relative to trends currently seen in the Medicare program (described earlier in this appendix), will ACOs achieve hospice lengths of stay that allow individuals to realize the strengths of the benefit more fully, and will individuals access these services before they enter the hospital and ICU?

Early results from the first year of the MSSP show that nearly half (54 of 114) of the MSSP ACOs that began operation in 2012 had lower spending than projected, and 29 produced savings relative to the target that were large enough to allow them to share savings with Medicare (CMS, 2014c). Early results from the first year of the Pioneer ACO model suggest that Medicare spending per beneficiary for individuals enrolled in these organizations grew at a slower rate overall than spending for beneficiaries enrolled in the FFS program in their area, although results differed across Pioneer ACOs (L&M Policy Research, 2014). Relative to FFS beneficiaries in the same area, 8 Pioneer ACOs had significantly lower Medicare spending growth per beneficiary, 1 had significantly higher Medicare spending growth per beneficiary, and 23 had no statistically significant difference (L&M Policy Research, 2014). Pioneer ACOs performed better than FFS Medicare on 15 quality measures for which published data on FFS beneficiaries were available, and 25 of the 32 had lower risk-adjusted readmission rates relative to the benchmark rate for FFS beneficiaries (CMS, 2013e; Toussaint et al., 2013). However, none of the 33 quality performance standards relates specifically to end-of-life care.

In the commercial market, some payers have also begun experimenting with bundled payment models similar in many ways to the Medicare ACO demonstration programs. In 2009, BCBSMA adopted its Alternative Quality Contract (AQC), which gives provider organizations a risk-adjusted prospective payment for all primary and specialty care provided to a fixed population (the global budget) for a 5-year period. As noted above, AQC organizations are eligible for bonuses of up to 10 percent of their budget based on their performance on 64 outpatient and hospital measures, again none of which is particularly relevant or meaningful for end-of-life care.

AQC implementation was associated with a modest slowing of total spending growth, particularly among organizations that had previously been paid under FFS by BCBSMA, over the first 2 years of the contract (Song et al., 2011, 2012). The savings were driven primarily by a shift of outpatient care to providers with lower fees. However, effects appeared to differ based on prior risk contracting experience with BCBSMA; lower use

explained about half of savings for enrollees of providers without prior BCBSMA risk contracting experience. AQC implementation was also associated with some improvements in the contract’s performance standards, which were larger in year 2 than in year 1. The AQC evaluation was unable to assess the impact on care at the end of life because of relatively small numbers of decedents each year in this nonelderly commercial population.

If a key goal of bundled payment models (regardless of how broad the bundle) is to maintain or improve quality while increasing provider efficiency—something that was not a primary focus of many previous models intended to increase efficiency—then the success of these efforts will depend on the ability to measure and monitor quality of care for patients with advanced illness. As noted above, none of the new models being implemented or debated includes performance measures appropriate to measuring quality of care for patients with advanced illness.

As policy makers seek to incorporate quality measures into bundled payment and other coordinated care efforts, the role of such measures can be viewed as twofold. First, by integrating quality metrics into the financial incentives for providers, policy makers can help ensure that providers are delivering care that is aligned with expected standards. It should be noted, however, that while the literature on pay-for-performance (P4P) strategies suggests that P4P often does result in improved quality as measured by the metrics used in the P4P systems, the improvements are often relatively small in magnitude and may be somewhat narrowly focused on the clinical areas that are targeted through the measures (Colla et al., 2012; Mullen et al., 2010; Werner et al., 2013; Wilensky, 2011). Even if good measures of end-of-life care quality were to be incorporated in these arrangements, the extent to which overall quality of end-of-life care would improve in response to the incentives is unclear. A perhaps even more important role for quality measures in the context of financing reforms will be as part of a broad effort of oversight and monitoring of organizations responsible for providing end-of-life care. If policy makers detect important quality deficiencies that result from reforms to the financing of care, modifications can be made to such financing arrangements or to the oversight and compliance requirements for providers.

Another new initiative that is related conceptually to the idea of better integrating care individuals receive through the Medicare program involves integrating Medicare and Medicaid financing for individuals who are dually eligible for both programs. Also created by the ACA, the State Integrated Care and Financial Alignment Demonstrations for Dual Eligible Beneficiaries allow states to use one of two models to coordinate services for dually eligible individuals, something that has been challenging historically. Although few states have begun to implement their programs, 26 states have submitted applications to CMS for approval (CMS, 2013c). These

programs have the potential to improve the coordination of medical and supportive services for dual eligibles at the end of life, but several caveats should be kept in mind when considering their possible future impact. First, although many states have expressed interest in developing these demonstrations, few programs are under way, and not all states will ultimately move forward with the initiative. Second, previous research has shown that achieving savings in the context of these programs is difficult, in part because states and provider organizations have relatively little experience in coordinating acute and supportive services for a frail population. Finally, and more specific to the context of end of life care, it appears that the state demonstration proposals either carve out Medicare-financed hospice care or explicitly exclude hospice enrollees from the demonstration (i.e., if individuals elect hospice, they are no longer enrolled in the demonstration) (CMS, 2013b; Grabowski, 2007; KFF, 2012a,b). Although palliative care is listed as an included benefit in some state proposals, most fail to mention palliative care services explicitly. One notable exception is South Carolina’s proposal and memorandum of understanding with CMS, which details a new palliative care benefit for enrollees who may not meet hospice eligibility criteria (CMS, 2013d). The benefit is designed to provide care earlier in the continuum of illness or disease process, and the care can be provided in conjunction with potentially life-prolonging therapies.

In the context of discussing the potential value of coordinating long-term services and supports with other acute and post-acute care services, it is important to note that current provisions of the ACA do little to address the financing and delivery of long-term services and supports for non-Medicaid-eligible individuals, let alone how these services relate to the broader health care system. The Community Living Assistance Service and Supports (CLASS) Act (Title VIII of the ACA) could have bolstered the financial protection of individuals from long-term care costs, but it was repealed as part of the “fiscal cliff” deal in January 2013 (i.e., The American Taxpayer Relief Act of 2012). Although the Commission on Long-Term Care, which was created through the same legislation, was unable to reach consensus on any alternative approaches to financing of long-term services and supports, it did highlight the need to identify approaches that could better integrate these services and supports with other acute and postacute services, including through bundled payment and interdisciplinary workforce development initiatives (U.S. Senate Commission on Long-Term Care, 2013).

Concurrent Care Models

The ACA calls on the Secretary of Health and Human Services to create a Medicare Hospice Concurrent Care demonstration program under which

beneficiaries will no longer be required to forego curative therapies if they meet other eligibility criteria for the hospice benefit. Budget neutrality is required during the 3-year demonstration period, meaning that total Medicare expenditures under the demonstration must not exceed what Medicare spending would have been in the absence of the demonstration.

In March 2014, CMS released a request for applications for this demonstration program, called the Medicare Care Choices Model (CMS, 2014b), with applications due no later than June 2014. CMS plans to select at least 30 Medicare-certified hospice programs, including hospices that serve rural areas and those that serve urban areas, for participation in the demonstration. Beneficiaries who meet Medicare hospice eligibility criteria and have advanced cancers, chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), or HIV/AIDS are eligible for enrollment in the new model (CMS, 2014b). Participating hospices will provide services available under the Medicare hospice benefit for routine home care and inpatient respite levels of care (CMS, 2014a). CMS will pay participating hospices $400 per beneficiary per month for these services and for related care coordination activities (CMS, 2014a). Providers that deliver curative services to beneficiaries enrolled in the demonstration will be allowed to bill Medicare for the reasonable and necessary services they deliver (CMS, 2014a).

For a subset of its commercial clients, Aetna has used a concurrent care model for almost a decade. In 2004, Aetna expanded its hospice and palliative care benefits in two key ways: (1) by allowing members to receive curative therapies while enrolled in hospice (i.e., concurrent care) and (2) by requiring a prognosis of 12 or fewer months for hospice eligibility (as opposed to the 6-month prognosis requirement for the Medicare hospice benefit, which would still apply under the Medicare Concurrent Care demonstration) (Krakauer et al., 2009; Spettell et al., 2009). At the same time, Aetna implemented for all commercial and MA members a comprehensive case management program in which services are provided by a nurse care manager with extensive training in palliative care, using predictive modeling to identify potential enrollees for the program.

In a retrospective cohort study that matched current enrollees with historical controls, Spettell and colleagues (2009) compared expenditures, hospice use, and inpatient use for three groups of members who died in 2005, 2006, or 2007: (1) commercial enrollees who received the specialized case management and the traditional hospice benefit; (2) commercial enrollees who received both the case management and the expanded hospice benefit; and (3) MA enrollees who received the Medicare hospice benefit and case management. They found that hospice enrollment and mean number of hospice days for hospice users was substantially higher for all groups relative to the historical controls (who died in 2004). In contrast, the rate of inpatient stays was lower for the intervention groups relative to

the controls: 17 percent of those receiving case management plus enhanced benefits had an inpatient stay versus 40 percent of their matched controls, and 23 percent of those receiving case management and traditional hospice benefits were hospitalized versus 43 percent of their controls. Commercial members with both case management and the expanded benefit had longer mean hospice stays than those who received case management and the traditional hospice benefit (37 versus 29 days, respectively), and both groups had longer mean stays than the historical controls.

Krakauer and colleagues (2009) estimate that the increase in hospice use and decrease in acute care service use resulted in a 22 percent decrease in spending compared with historical controls for the commercial member case management/traditional hospice benefit group (the authors provide no estimates for the other groups, nor do they discuss detailed methods). Given that hospice use was increasing during this period, the use of historical controls likely overstates the impact of the interventions. While the results of the Aetna experiment are informative, it is not possible to estimate accurately the potential impact of the Medicare concurrent care demonstration on expenditures using data from a program that offered both concurrent care and expanded hospice eligibility for a commercial under-65 population.

Outside of the Aetna program, no published studies shed light on the expected costs of concurrent care. Studies of cost savings associated with use of the Medicare hospice benefit document savings associated with stays of fewer than 30 days and stays lasting from 53 to approximately 105 days (Kelley et al., 2013a; Taylor et al., 2007). On the basis of these results, one might expect overall savings to the extent that concurrent care resulted in stays lasting fewer than 105 days. There are no data on the relationship between use of the Medicare hospice benefit and Medicare spending for longer stays because of smaller sample sizes in the upper tail of the distribution of stay duration. As a result, the extent to which demonstration sites might be able to meet the budget neutrality requirement could depend on who enrolls in the program and the duration of stays that result from the implementation of concurrent care.

While the Medicare Concurrent Care demonstration program was not implemented initially (with applications accepted starting only in spring 2014), a concurrent care requirement for children in Section 2302 of the ACA was implemented immediately after the act’s passage. Effective March 2010, Medicaid and State Children’s Health Insurance Programs (SCHIPs) may no longer require children up to age 21 to agree to forego curative therapies to be eligible for the hospice benefit. As noted above, although hospice is an optional benefit for Medicaid and SCHIP programs, the Early Periodic Screening, Diagnosis, and Treatment (EPSDT) provision requires that Medicaid and SCHIPs operating as Medicaid expansions cover hospice

for children up to age 21 for whom a physician certifies a prognosis of 6 months or less, and the concurrent care requirement would apply to all of these children. Lindley and colleagues (2013) report that 31 of 50 states had implemented concurrent care for children by 2012, but there have been no published evaluations of the impact of the policy change on utilization of hospice services, spending, or quality of care.

If implemented broadly, concurrent care models should reduce barriers to accessing hospice. Existing concurrent care models would not, however, address any barriers to high-quality palliative care created by the 6-month prognosis requirement.

ACA-AUTHORIZED CHANGES TO MEDICARE HOSPICE REIMBURSEMENT

The ACA calls on the Secretary of Health and Human Services to implement revisions to the payment methodology for hospice services no earlier than October 1, 2013. The legislation requires that such changes be budget neutral in the fiscal year of their implementation. Medicare hospice payment changes could help reduce barriers to access and make payments more efficient, depending on the specific changes implemented by the secretary. For example, an outlier payment system for hospice care, whereby hospices would receive somewhat higher payments for particularly high-cost stays, could help increase access for patients with high-cost palliative care needs. Similarly, payments could be adjusted for case mix and/or setting to ensure that they reflect the true cost of services delivered to hospice recipients. Of course, any reform of payment methodology could produce both intended and unintended consequences. Also, any given change implemented in isolation could produce very different outcomes than a package of individual changes combined to meet the budget neutrality provision of the law. Absent the more substantial reforms detailed above, it will be important to ensure that hospice payments are as fair and efficient as possible so as to facilitate both access to the benefit and its long-term sustainability.

CONCLUSION

Current approaches to financing services used by patients with advanced illness have a number of limitations that often lead to limited access to hospice and palliative care services and poor quality of care at the end of life. The ACA authorized a number of payment reforms, and payers are adopting or considering other changes as well. These reforms could address some—but likely not all—of the current limitations in financing (for example, in most cases they would not add the role of long-term services

and supports for individuals at the end of life or at other points in their health trajectories). Policy makers should anticipate both the intended and unintended consequences of these reforms when structuring their design and implementation. To this end, we identify the following key elements as essential considerations for policy makers as they formulate and implement relevant reforms:

• As Medicare, Medicaid, and commercial payers move forward with efforts to bundle payment for groups of services, these models should incorporate performance metrics that are appropriate for patients with end-of-life care needs to ensure that the models do not result in lower quality of care for individuals with advanced illness. These measures would serve as the foundation for performance incentives in this area and should also be used in oversight and monitoring efforts to ensure high-quality care.
• In creating bundled payment systems, payers should consider mixed payment methods that involve providers sharing both risk and savings with the payer (as opposed to paying a fixed rate per bundle), especially while end-of-life care quality measures are in an early stage of development and use.
• Given the special concerns inherent in the financing of care for nursing home residents at the end of life, payment models that bundle acute, post-acute, and end-of-life care should be explored, again using mixed payment methods. Some package of hospice and palliative care services should be made available to nursing home residents while they are on the Medicare SNF post-acute care benefit.
• In the context of integrated care programs of all types, including MA programs, PACE, and ACO programs with risk-based payment such as the Medicare ACO demonstrations and the AQC, hospice and palliative care services should be included in the package of services for which these organizations are paid and held accountable.
  • – Alternatives to carve-outs of hospice and palliative care services for individuals with advanced illness that ensure access to high-quality end-of-life and palliative care should be explored. Although prospects for implementation are unclear, MedPAC voted in January 2014 to recommend ending the hospice carve-out within the MA program.
  • – Policy makers should ensure that these programs have adequate provider networks for patients (e.g., including access to palliative care specialists) and that they provide high-quality care to patients.

• Nursing homes should be held accountable for the quality of end-of-life care provided for all residents, including both those who do and do not use hospice. To support such expectations, nursing home survey processes, public reporting efforts such as Nursing Home Compare, and P4P efforts should incorporate performance measures appropriate for patients with advanced illness. In particular, the Nursing Home Compare tool could implement such improvements in the near term to ensure that its focus is not exclusively on restoration and maintenance of functioning.
• Changes in Medicare hospice benefit payment authorized by the ACA and implemented by the Secretary of Health and Human Services should attempt to match expected costs and payments for different types of hospice stays while ensuring access to high-quality end-of-life care for all beneficiaries with advanced illness, including those with high-cost palliative care needs. The impact of such changes on both expenditures and quality of care should be monitored on an ongoing basis.

REFERENCES

Adelson, J. 2012. Louisiana cuts health care, Medicaid and hospice programs to rebalance budget. The Times-Picayune, December 14, 2012. http://www.nola.com/politics/index.ssf/2012/12/louisiana_cuts_health_care_med.html (accessed December 18, 2014).

Adelson, J. 2013. Gov. Bobby Jindal backs down on Medicaid hospice elimination. The Times-Picayune, January 23, 2013. http://www.nola.com/politics/index.ssf/2013/01/gov_bobby_jindal_backs_down_on.html (accessed December 30, 2014).

Afendulis, C. C., M. B. Landrum, and M. E. Chernew. 2012. The impact of the Affordable Care Act on Medicare Advantage Plan availability and enrollment. Health Services Research 47(6):2339-2352.

Aldridge Carlson, M. D., C. L. Barry, E. J. Cherlin, R. McCorkle, and E. H. Bradley. 2012. Hospices’ enrollment policies may contribute to underuse of hospice care in the United States. Health Affairs (Millwood) 31:2690-2698.

Aragon, K., K. Covinsky, Y. Miao, W. J. Boscardin, L. Flint, and A. K. Smith. 2012. Use of the Medicare posthospitalization skilled nursing benefit in the last 6 months of life. Archives of Internal Medicine 172:1573-1579.

Architecture for Humanity. 2012. Design like you give a damn [2]: Building change from the ground up. New York: Abrams Books.

Barnato, A. E., M. B. McClellan, C. R. Kagay, and A. M. Garber. 2004. Trends in inpatient treatment intensity among Medicare beneficiaries at the end of life. Health Services Research 39:363-375.

BCBSMA (Blue Cross Blue Shield of Massachusetts). 2010. The Alternative Quality Contract. http://www.bluecrossma.com/visitor/pdf/alternative-quality-contract.pdf (accessed July 28, 2014).

Bradley, E. H., H. Prigerson, M. D. Carlson, E. Cherlin, R. Johnson-Hurzeler, and S. V. Kasl. 2004. Depression among surviving caregivers: Does length of hospice enrollment matter? American Journal of Psychiatry 161:2257-2262.

Cao, X. Y., X. L. Jiang, X. L. Li, M. C. Hu Lo, and R. Li. 2013. Family functioning and its predictors among disaster bereaved individuals in China: Eighteen months after the Wenchuan earthquake. PLoS ONE 8(4):e60738.

CAPC (Center to Advance Palliative Care). 2009. Making the case for hospital-based palliative care. 2009. http://www.capc.org/building-a-hospital-based-palliative-care-program/case (accessed September 19, 2013).

CAPC. 2010. Analysis of U.S. hospital palliative care programs: 2010 snapshot. http://www.capc.org/news-and-events/releases/analysis-of-us-hospital-palliative-care-programs-2010-snapshot.pdf (accessed July 28, 2014).

Carlson, M. D., C. Barry, M. Schlesinger, R. McCorkle, R. S. Morrison, E. Cherlin, J. Herrin, J. Thompson, M. L. Twaddle, and E. H. Bradley. 2011. Quality of palliative care at US hospices: Results of a national survey. Medical Care 49:803-809.

Carter, C., A. B. Garrett, and D. Wissoker. 2012. Reforming Medicare payments to skilled nursing facilities to cut incentives for unneeded care and avoiding high-cost patients. Health Affairs (Millwood) 31:1303-1313.

Chambers, J. D., D. E. Weiner, S. K. Bliss, and P. J. Neumann. 2013. What can we learn from the U.S. expanded end-stage renal disease bundle? Health Policy 110:164-171.

Christakis, N. A., and E. B. Lamont. 2000. Extent and determinants of error in doctors’ prognoses in terminally ill patients: Prospective cohort study. British Medical Journal 320:469-472.

CMS (Centers for Medicare & Medicaid Services). 2012. Pioneer Accountable Care Organization Model: General fact sheet. http://innovation.cms.gov/Files/fact-sheet/Pioneer-ACO-General-Fact-Sheet.pdf (accessed July 28, 2014).

CMS. 2013a. Bundled Payments for Care Improvement (BPCI) initiative: General information. http://innovation.cms.gov/initiatives/bundled-payments (accessed September 9, 2013).

CMS. 2013b. Financial alignment initiative. http://www.cms.gov/Medicare-Medicaid-Coordination/Medicare-and-Medicaid-Coordination/Medicare-Medicaid-Coordination-Office/FinancialModelstoSupportStatesEffortsinCareCoordination.html (accessed September 19, 2013).

CMS. 2013c. Medicare-Medicaid enrollee state profiles. http://www.cms.gov/Medicare-Medicaid-Coordination/Medicare-and-Medicaid-Coordination/Medicare-Medicaid-Coordination-Office/StateProfiles.html (accessed November 5, 2013).

CMS. 2013d. Memorandum of understanding between the Centers for Medicare & Medicaid Services (CMS) and the state of South Carolina. https://www.cms.gov/Medicare-Medicaid-Coordination/Medicare-and-Medicaid-Coordination/Medicare-Medicaid-CoordinationOffice/Downloads/SCMOU.pdf (accessed July 25, 2014).

CMS. 2013e. Pioneer Accountable Care Organizations succeed in improving care, lowering costs. http://www.cms.gov/Newsroom/MediaReleaseDatabase/Press-Releases/2013-Press-Releases-Items/2013-07-16.html (accessed February 10, 2014).

CMS. 2014a. Fact sheet: Medicare Care Choices Model. http://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2014-Fact-sheets-items/2014-03-18.html (accessed March 25, 2014).

CMS. 2014b. Medicare Care Choices Model. http://innovation.cms.gov/initiatives/Medicare-Care-Choices (accessed March 25, 2014).

CMS. 2014c. Medicare’s delivery system reform initiatives achieve significant savings and quality improvements—off to a strong start. Press release. http://www.cms.gov/Newsroom/MediaReleaseDatabase/Press-Releases/2014-Press-releases-items/2014-01-30.html (accessed February 10, 2014).

Colla, C. H., D. E. Wennberg, E. Meara, J. S. Skinner, D. Gottlieb, V. A. Lewis, C. M. Snyder, and E. S. Fisher. 2012. Spending differences associated with the Medicare Physician Group Practice Demonstration. Journal of the American Medical Association 308:1015-1023.

Collins, A. J. 2012. ESRD payment policy changes: The new “bundled” dialysis prospective payment system (PPS) in the United States. National Kidney Foundation Spring Clinical Meeting, Washington, DC.

Cutler, D. M., and K. Ghosh. 2012. The potential for cost savings through bundled episode payments. New England Journal of Medicine 366:1075-1077.

Famakinwa, A. B. 2010. End-of-life care in a PACE program: Respecting the patient’s wishes while supporting the caregiver. Journal of the American Medical Directors Association 11:528-530.

Feder, J. 2013. Bundle with care—rethinking Medicare incentives for post-acute care services. New England Journal of Medicine 369:400-401.

Fuller, D. S., R. L. Pisoni, B. A. Bieber, B. W. Gillespie, and B. M. Robinson. 2013. The DOPPS Practice Monitor for US dialysis care: Trends through December 2011. American Journal of Kidney Diseases 61:342-346.

Gilbertson, D., A. Collins, and R. Foley. 2012. Transition in service utilization: Vascular access, injectables, hemoglobin levels, and transfusions. Ann Arbor, MI: United States Renal Data System.

Goodman, D. C., A. R. Esty, E. S. Fisher, and C.-H. Chang. 2011. Trends in variation in end-of-life care for Medicare beneficiaries with severe chronic illness. Hanover, NH: The Dartmouth Institute for Health Policy and Clinical Practice.

Grabowski, D. C. 2007. Medicare and Medicaid: Conflicting incentives for long-term care. Milbank Quarterly 85:579-610.

Grabowski, D. C. 2009. Special needs plans and the coordination of benefits and services for dual eligibles. Health Affairs (Millwood) 28:136-146.

Hackbarth, G., R. Reischauer, and A. Mutti. 2008. Collective accountability for medical care—toward bundled Medicare payments. New England Journal of Medicine 359:3-5.

Health Care Financing Administration. 1998. Medicare Participating Heart Bypass Center Demonstration. https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Reports/downloads/oregon2_1998_3.pdf (accessed July 25, 2014).

HHS OIG (U.S. Department of Health and Human Services Office of Inspector General). 1997. Hospice patients in nursing homes. Washington, DC: HHS OIG.

HHS OIG. 2013. Medicare hospice: Use of general inpatient care. Washington, DC: HHS OIG.

Hoffmann, D. E., and A. J. Tarzian. 2005. Dying in America—an examination of policies that deter adequate end-of-life care in nursing homes. The Journal of Law, Medicine & Ethics 33:294-309.

Hogan, C., J. Lunney, J. Gabel, and J. Lynn. 2001. Medicare beneficiaries’ costs of care in the last year of life. Health Affairs (Millwood) 20:188-195.

Huskamp, H. A., M. B. Buntin, V. Wang, and J. P. Newhouse. 2001. Providing care at the end of life: Do Medicare rules impede good care? Health Affairs (Millwood) 20:204-211.

Huskamp, H. A., J. P. Newhouse, J. C. Norcini, and N. L. Keating. 2008. Variation in patients’ hospice costs. Inquiry 45:232-244.

Huskamp, H. A., D. G. Stevenson, M. E. Chernew, and J. P. Newhouse. 2010. A new Medicare end-of-life benefit for nursing home residents. Health Affairs (Millwood) 29:130-135.

Huskamp, H. A., C. Kaufmann, and D. G. Stevenson. 2012. The intersection of long-term care and end-of-life care. Medical Care Research and Review 69:3-44.

IOM (Institute of Medicine). 1997. Approaching death: Improving care at the end of life. Washington, DC: National Academy Press.

IOM. 2013. Variation in health care spending: Target decision making, not geography. Washington, DC: The National Academies Press.

Iwashyna, T. J., and N. A. Christakis. 1998. Attitude and self-reported practice regarding hospice referral in a national sample of internists. Journal of Palliative Medicine 1:241-248.

Kelley, A. S., S. L. Ettner, R. S. Morrison, Q. Du, and C. A. Sarkisian. 2012. Disability and decline in physical function associated with hospital use at end of life. Journal of General Internal Medicine 27:794-800.

Kelley, A. S., P. Deb, Q. Du, M. D. Aldridge Carlson, and R. S. Morrison. 2013a. Hospice enrollment saves money for Medicare and improves care quality across a number of different lengths-of-stay. Health Affairs (Millwood) 32:552-561.

Kelley, A. S., K. McGarry, S. Fahle, S. M. Marshall, Q. Du, and J. S. Skinner. 2013b. Out-of-pocket spending in the last five years of life. Journal of General Internal Medicine 28:304-309.

KFF (The Henry J. Kaiser Family Foundation). 2011. Health insurance coverage of the total population. http://kff.org/other/state-indicator/total-population (accessed July 25, 2014).

KFF. 2012a. Explaining the state integrated care and financial alignment demonstrations for dual eligible beneficiaries. Washington, DC: KFF.

KFF. 2012b. State demonstrations to integrate care and align financing for dual eligible beneficiaries: A review of the 26 proposals submitted to CMS. Washington, DC: KFF.

KFF. 2012c. Total Medicare Advantage (MA) enrollment. http://kff.org/other/state-indicator/ma-total-enrollment (accessed July 25, 2014).

KFF. 2012d. Total number of Medicare beneficiaries. http://kff.org/medicare/state-indicator/total-medicare-beneficiaries (accessed July 25, 2014).

Kirwan, L., and S. Iselin. 2009. Recommendations of the Special Commission on the Health Care Payment System. http://www.mass.gov/chia/docs/pc/final-report/final-report.pdf (accessed July 25, 2014).

Krakauer, R., C. M. Spettell, L. Reisman, and M. J. Wade. 2009. Opportunities to improve the quality of care for advanced illness. Health Affairs (Millwood) 28:1357-1359.

Kwok, A. C., M. E. Semel, S. R. Lipsitz, A. M Bader, A. E. Barnato, A. A. Gawande, and A. K. Jha. 2011. The intensity and variation of surgical care at the end of life: A retrospective cohort study. Lancet 378:1408-1413.

L&M Policy Research. 2014. Effect of pioneer ACOs on Medicare spending in the first year. Washington, DC: CMS.

Lave, J. R. 1989. The effect of the Medicare prospective payment system. Annual Review of Public Health 10:141-161.

Lindley, L. C., S. Edwards, and D. J. Bruce. 2013. Factors influencing the implementation of health care reform: An examination of the concurrent care for children provision. American Journal of Hospice and Palliative Care [epub ahead of print].

Lipsitz, L. A. 2013. The 3-night hospital stay and Medicare coverage for skilled nursing care. Journal of the American Medical Association 310(14):1441-1442.

Lorenz, K. A., S. M. Asch, K. E. Rosenfeld, H. Liu, and S. L. Ettner. 2004. Hospice admission practices: Where does hospice fit in the continuum of care? Journal of the American Geriatrics Society 52:725-730.

Lubitz, J. D., and G. F. Riley. 1993. Trends in Medicare payments in the last year of life. New England Journal of Medicine 328:1092-1096.

McCarthy, E. P., R. B. Burns, Q. Ngo-Metzger, R. B. Davis, and R. S. Phillips. 2003. Hospice use among Medicare managed care and fee-for-service patients dying with cancer. Journal of the American Medical Association 289:2238-2245.

MedPAC (Medicare Payment Advisory Commission). 2009. Medicare payment policy. Washington, DC: MedPAC.

MedPAC. 2012. Report to the Congress: Medicare payment policy. Washington, DC: MedPAC.

MedPAC. 2013. Medicare payment policy. Washington, DC: MedPAC.

Meier, D. E. 2013. Palliative care and the health care crisis in the United States: A candid conversation with Dr. Diane Meier. Syracuse, NY: Syracuse University Maxwell School of Citizenship and Public Affairs.

Miller, S. C. 2004. Hospice care in nursing homes: Is site of care associated with visit volume? Journal of the American Geriatrics Society 52:1331-1336.

Miller, S. C., J. C. Lima, and S. L. Mitchell. 2012. Influence of hospice on nursing home residents with advanced dementia who received Medicare-skilled nursing facility care near the end of life. Journal of the American Geriatrics Society 60:2035-2041.

Minino, A. M. 2013. Death in the United States, 2011. Washington, DC: Centers for Disease Control and Prevention.

Morrison, R. S., J. D. Penrod, J. B. Cassel, M. Caust-Ellenbogen, A. Litke, L. Spragens, and D. E. Meier. 2008. Cost savings associated with US hospital palliative care consultation programs. Archives of Internal Medicine 168:1783-1790.

Mukamel, D. B., A. Bajorska, and H. Temkin-Greener. 2002. Health care services utilization at the end of life in a managed care program integrating acute and long-term care. Medical Care 40:1136-1148.

Mukamel, D. B., D. R. Peterson, H. Temkin-Greener, R. Delavan, D. Gross, S. J. Kunitz, and T. F. Williams. 2007. Program characteristics and enrollees’ outcomes in the Program of All-Inclusive Care for the Elderly (PACE). Milbank Quarterly 85:499-531.

Mullen, K. J., R. G. Frank, and M. B. Rosenthal. 2010. Can you get what you pay for? Pay-for-performance and the quality of healthcare providers. The RAND Journal of Economics 41:64-91.

Newhouse, J. P. 1994. Patients at risk: Health reform and risk adjustment. Health Affairs (Millwood) 13:132-146.

NQF (National Quality Forum). 2012. NQF endorses palliative and end-of-life care measures. http://www.qualityforum.org/News_And_Resources/Press_Releases/2012/NQF_Endorses_Palliative_and_End-of-Life_Care_Measures.aspx (accessed July 25, 2014).

Riley, G. F., and C. Herboldsheimer. 2001. Including hospice in Medicare capitation payments: Would it save money? Health Care Financing Review 23:137-147.

Riley, G. F., and J. D. Lubitz. 2010. Long-term trends in Medicare payments in the last year of life. Health Services Research 45:565-576.

Sachs, G. A., J. W. Shega, and D. Cox-Hayley. 2004. Barriers to excellent end-of-life care for patients with dementia. Journal of General Internal Medicine 19:1057-1063.

Schroeder, S. A., and Frist, W. 2013. Phasing out fee-for-service payment. New England Journal of Medicine 368:2029-2032.

Scitovsky, A. A. 1984. “The high cost of dying”: What do the data show? Milbank Memorial Fund Quarterly: Health and Society 62:591-608.

Segal, M. 2011. Dual eligible beneficiaries and potentially avoidable hospitalizations. https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Insight-Briefs/downloads/PAHInsightBrief.pdf (accessed July 25, 2014).

Song, Z., D. G. Safran, B. E. Landon, Y. He, R. P. Ellis, R. E. Mechanic, M. P. Day, and M. E. Chernew. 2011. Health care spending and quality in year 1 of the alternative quality contract. New England Journal of Medicine 365:909-918.

Song, Z., D. G. Safran, B. E. Landon, M. B. Landrum, Y. He, R. E. Mechanic, M. P. Day, and M. E. Chernew. 2012. The “Alternative Quality Contract,” based on a global budget, lowered medical spending and improved quality. Health Affairs (Millwood) 31:1885-1894.

Sood, N., P. J. Huckfeldt, J. J. Escarce, D. C. Grabowski, and J. P. Newhouse. 2011. Medicare’s bundled payment pilot for acute and postacute care: Analysis and recommendations on where to begin. Health Affairs (Millwood) 30:1708-1717.

Spettell, C. M., W. S. Rawlins, R. Krakauer, J. Fernandes, M. E. Breton, W. Gowdy, S. Brodeur, M. MacCoy, and T. A. Brennan. 2009. A comprehensive case management program to improve palliative care. Journal of Palliative Medicine 12:827-832.

Stevenson, D. G., J. Z. Ayanian, A. M. Zaslavsky, J. P. Newhouse, and B. E. Landon. 2013. Service use at the end of life in Medicare advantage versus traditional Medicare. Medical Care 51(10):931-937.

Taylor, D. H., Jr., J. Ostermann, C. H. Van Houtven, J. A. Tulsky, and K. Steinhauser. 2007. What length of hospice use maximizes reduction in medical expenditures near death in the US Medicare program? Social Science & Medicine 65:1466-1478.

Temel, J. S., J. A. Greer, A. Muzikansky, E. R. Gallagher, S. Admane, V. A. Jackson, C. M. Dahlin, C. D. Blinderman, J. Jacobsen, W. F. Pirl, J. A. Billings, and T. J. Lynch. 2010. Early palliative care for patients with metastatic non-small-cell lung cancer. New England Journal of Medicine 363:733-742.

Teno, J. M., P. L. Gozalo, J. P. Bynum, N. E. Leland, S. C. Miller, N. E. Morden, T. Scupp, D. C. Goodman, and V. Mor. 2013. Change in end-of-life care for Medicare beneficiaries: Site of death, place of care, and health care transitions in 2000, 2005, and 2009. Journal of the American Medical Association 309:470-477.

Toussaint, J., A. Milstein, and S. Shortell. 2013. How the Pioneer ACO Model needs to change: Lessons from its best-performing ACO. Journal of the American Medical Association 310:1341-1342.

U.S. Senate Commission on Long-Term Care. 2013. Commission on Long-Term Care: Report to the Congress. Washington, DC: U.S. Senate Commission on Long-Term Care.

Virnig, B. A., E. S. Fisher, A. M. McBean, and S. Kind. 2001. Hospice use in Medicare managed care and fee-for-service systems. American Journal of Managed Care 7:777-786.

Weeks, W. B., S. S. Rauh, E. B. Wadsworth, and J. N. Weinstein. 2013. The unintended consequences of bundled payments. Annals of Internal Medicine 158:62-64.

Werner, R. M., R. T. Konetzka, and D. Polsky. 2013. The effect of pay-for-performance in nursing homes: Evidence from state Medicaid programs. Health Services Research 48:1393-1414.

Wilensky, G. R. 2011. Lessons from the Physician Group Practice Demonstration—a sobering reflection. New England Journal of Medicine 365:1659-1661.

Zerzan, J., S. Stearns, and L. Hanson. 2000. Access to palliative care and hospice in nursing homes. Journal of the American Medical Association 284:2489-2494.

This page intentionally left blank.