Where Americans Die — Is There Really “No Place Like Home”?

Where Americans die has shifted over the past 15 years, away from hospitals and nursing facilities and toward home. By 2017, 31% of Americans died at home, making it the most common site of death for the first time in decades.¹ Obituaries often describe a “good” death as one that occurred “peacefully at home, surrounded by family.” Yet reality frequently diverges from this scenario. Unprepared family caregivers are routinely tasked with managing distressing symptoms, including pain, agitation, and dyspnea; administering medications; and providing intimate personal care, including bathing and toileting assistance, to bedbound patients. In other care settings, these tasks are performed by trained clinicians. As death nears and the realities of dying at home sink in, patient and family preferences for home death can change rapidly.²

The perception that a good death is one that occurs at home is ingrained in our cultural and social history. Patient surveys have revealed a general preference for home death, though preferences may vary with patient characteristics, including race and ethnic group.³ Many physicians assert that hospital death isn’t in keeping with high-quality end-of-life care and recommend home death as the best alternative. But palliative care clinicians and others have begun to question this assumption.⁴

Alternative locations for death, such as freestanding inpatient hospice facilities and skilled nursing facilities (SNFs), can provide around-the-clock nursing and personal care that is particularly critical for certain physically and emotionally burdensome end-of-life needs that are difficult to manage with intermittent home-hospice visits (e.g., incontinence or agitated delirium). In these inpatient settings, medications can also be rapidly titrated or changed to control symptoms, often more effectively than at home. Hospital-based palliative care has grown substantially over the past 20 years, although there are disparities in access. Increasingly, hospitalized patients can elect comfort-focused end-of-life care, sometimes in dedicated palliative care or hospice units, and with access to palliative interventions (e.g., transfusions and high-flow oxygen) that are rarely available in nonhospital settings.

So why does the notion persist that — in the language of The Wizard of Oz and other palliative care researchers — “there’s no place like home” to die?⁵ We believe that, in addition to the effects of culturally shaped preferences, financial incentives and publicly reported quality measures influence key stakeholders (including insurers, hospital and hospice agency administrators, clinicians, patients, and families) to view home as a “one-size-fits-all” destination for death, while failing to adequately consider and support individualized and unpredictable needs, regardless of setting.

Home deaths tend to reduce health care costs by averting expensive hospital stays and transferring costs to families in the form of unpaid caregiver labor and out-of-pocket spending. Payers are aware of these dynamics, and payment policies often discourage hospital death. For instance, insurers typically pay hospitals a lump sum based on the diagnosis-related group, a category that reflects the medical complexity of the patient’s diagnosis (which may be low for patients near the end of life), regardless of length of stay. Hospitals therefore have an incentive to discharge patients as quickly as possible. Home is often the quickest and easiest discharge destination. To counterbalance the incentive to discharge patients too early, the Centers for Medicare and Medicaid Services (CMS) penalizes hospitals for high readmission rates. Patients who are enrolled in home hospice or who are near death may be expected to die before being readmitted, however, which could attenuate the readmission-penalty disincentive.

Hospitals also are subject to quality-measurement programs that create indirect incentives for home death. CMS’s Hospital Value-Based Purchasing Program penalizes hospitals with high 30-day mortality, measured from the admission date. This penalty is a strong incentive to avoid admitting patients for what may be a “terminal admission,” because they could increase the hospital’s death rate. Mortality also figures prominently in ranking systems used by both CMS and U.S. News and World Report, and hospitals generally strive to optimize their rankings. CMS excludes from ranking-related calculations patients who were enrolled in hospice at any time in the 12 months prior to the admission, including the first day of the admission, but U.S. News and World Report does not, which may create confusion. Staff members who are unaware of exclusion rules may believe it’s safer to avoid admitting patients who are near death — regardless of their hospice status — to preserve the hospital’s reputation.

Hospice agencies, too, are affected by financial incentives favoring home death. Under the Medicare hospice benefit, so-called routine home care is reimbursed at $150 to $200 per day and accounts for 98% of hospice days, according to reports from the Medicare Payment Advisory Commission (MedPAC). Because hospice agencies are required to take full responsibility for costs related to the diagnosis that qualified a patient for hospice enrollment, hospital admissions can reduce their profits. Hospices can avoid these costs if patients disenroll from the program; however, CMS monitors disenrollments and may investigate hospice agencies with high “live-discharge” rates, which gives hospices an incentive to keep patients enrolled and at home. Medicare’s general inpatient hospice provision pays hospices roughly $1,000 per day to cover inpatient hospice care when symptoms cannot be controlled in an outpatient setting. This provision accounts for only 1% of hospice days, in part because eligibility criteria are restrictive. Hospice agencies, meanwhile, tend to lose money under the general inpatient care provision but earn profits for providing routine home care. With inpatient care, hospices pay the costs of around-the-clock nurses and aides. With home care, they can maintain lower nurse-to-patient staffing ratios and provide relatively few hours of home health aide services — relying instead on unpaid family caregiving, family-financed out-of-pocket spending, or Medicaid payments for low-income patients.

What changes are needed to provide high-quality end-of-life care in whatever setting is most in keeping with patients’ individual goals and needs? First, policymakers could increase support for patients and families who choose home death, particularly underserved populations including low-income households and communities of color. One such step would be to expand financial support for family caregivers. Compensation covering around-the-clock care is unrealistic, but the Veterans Health Administration (VA) and some state Medicaid and paid-family-leave programs provide models. Another step would be to change Medicare’s hospice regulations to classify care delivered by home health aides as a “core” service (similar to nursing, social work, and chaplain services), with a minimum number of hours required to be available daily. Another intervention would be to expand access to continuous home care, Medicare’s level of hospice care that provides extended in-home nursing services for patients with acute, uncontrolled symptoms, which currently accounts for only 0.2% of hospice days.

Second, policymakers could decrease barriers to providing high-quality end-of-life care in inpatient settings. One strategy could involve increasing access to general inpatient hospice care in nonacute care settings, including freestanding hospice facilities and hospice units within SNFs or hospitals, by loosening eligibility criteria and increasing the payments. The VA, which embeds hospice units within its SNFs, provides a model of a successful inpatient-hospice program with less-restrictive eligibility criteria than Medicare’s general inpatient care provision. We also suggest aligning all quality-measurement systems, most notably U.S. News and World Report’s ratings, with CMS’s system that excludes hospice patients from mortality reporting. This exclusion would need to be well-publicized to avoid confusion.

Third, research is necessary to support goal- and need-concordant end-of-life care, with the aim of developing shared decision-making tools that not only elicit patient preferences, but also assess patient and family needs to match patients to appropriate support and care settings. This research should actively engage groups that have historically been underrepresented to account for differences in preferences and to avoid persistent disparities in access to end-of-life care. Finally, a cultural shift that reduces stigma surrounding nonhome death would be needed to ensure the success of these policy changes.

These proposals aren’t cost neutral; they would shift some burden and expenses away from individual families and toward the government and other payers, with the goal of providing equitable access to high-quality care during the dying process. For some patients and their families, there’s no place like home at the end of life; for others, a hospital, nursing facility, or inpatient hospice facility — where staff can manage symptoms and provide personal care, thereby enabling families to be families and maximize quality time together — may be the best place for a good death.