The Health Economics of Palliative Care

We can afford to die well, but it will require coordinated care with someone in charge. . .

An aging population that demands more treatment[1] and new technology[2] to stay alive will continue to pressure the current health-care system. As much of palliative care concerns federally funded Medicare and Medicaid, all of us have a share in this "medical commons." Currently, about 25% of Medicare dollars are spent on patients in their last 60 days of life,[3,4] and money spent for palliative care cannot be spent for prevention, screening, or other treatment.

At the same time, the quality of cancer care that we purchase is suboptimal.[5,6] Examples include studies showing that high-volume or specialty centers/providers are associated with 5% to 10% better breast cancer survival, twofold better survival for testicular cancer patients, less mortality for prostatectomy, and up to fourfold better survival for complicated cancer surgery. The Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) showed that half of all dying patients had unnecessary pain in their final days of life in academic hospitals.[7] Poor pain relief has been found in academic practice[8] and in minority populations[9]—indeed, anywhere searched.

We have identified some important questions about economic outcomes and palliative care, which are listed in Table 1. The first question—what outcomes justify treatment—although probably the most important (and clearly the most contentious), is beyond the scope of this manuscript. The American Society of Clinical Oncology (ASCO) could not describe a minimum benefit that would justify treatment, but did state that some benefit must be demonstrable, and that cost and toxicity should be considered along with benefit.[10] We will address the other issues with data in this manuscript, based on a review of Medline from 1970 to the present and selected searches within bibliographies.

Hospice vs Nonhospice Care

Hospice has been recommended by ASCO as the most applicable model for good end-of-life care in the United States,[11] but there are few solid data that show hospice improves care or reduces costs (Table 2).[12-19] The only randomized clinical trial of hospice vs conventional care showed no important differences by any measured benchmark (pain, ability to perform activities of daily living).[15] Patients still used many hospital days (48 for control group, 51 for hospice group), but more of the hospice patients were hospitalized on the hospice unit. There was no difference in diagnostic procedures or in total costs of about $15,000 per patient.

Hospice care may be cost-saving in the last month of life.[16,17] For those who enrolled during that interval—typically, over half of Medicare patients—Medicare saved $1.26 to $1.65 for each $1 spent. Those who elected hospice tended to use more resources in the months from diagnosis until about 3 months before death, so the total disease management costs were equal.

Hospice may actually not be saving total disease management costs, but just shifting them to costs not captured by our current accounting systems. In our own study of Medicare hospice use in Virginia, matching hospice use with Medicare files, total disease management costs in the last year of life were not changed by hospice care (unpublished data, T.J. Smith et al). In the same study of hospice use in Virginia, multivariate logistic regression showed that patients who were very old, white, well-to-do, and without comorbidities more often elected hospice than those who were not (unpublished data, T.J. Smith et al). This group is also able to absorb more home care costs, out-of-pocket drug costs, and so forth. The data are consistent with an affluent group of patients using all the resources needed for treatment, then electing hospice. There are few published data on whether the poor or minorities use hospice, how much those patients will cost the system, or total costs.[12]

Studies from databases of hospice use have shown similar results, with one study finding a 39% reduction in end-of-life care costs—by prevention of hospitalization—if patients were in hospice more than 2 weeks.[18] Hospice patients were more likely to receive home nursing care and spend less time in the hospital than conventional care patients.[19] Conventional care was the least expensive approach when overall disease management costs were calculated, but hospital-based hospice ($2,270) and home care hospice ($2,657) were less expensive than conventional care ($6,100) in the last month of life. Not surprisingly, when full-cost accounting is done, home care provided by relatives is not much different ($4,563 for each 3-month period) than care in a nursing home or similar setting. Costs were lowest when the patient and caregiver lived in the same household.[20,21]

Use of Advance Directives

No cost savings have been associated with the use of either typical advance directives or do not resuscitate (DNR) orders done in the hospital,[13,14,22] as shown in Table 3.[22-24] In one randomized study of 204 patients with life-threatening diseases, the execution of an advance directive had no significant positive or negative effect on a patient’s well-being, health status, medical treatments, or medical treatment charges.[23]

For patients who had advanced directives prior to intervention in the SUPPORT trial,[22] there was a 23% reduction in cost associated with advance directives—$21,284 with one, compared to $26,127 without one—suggesting that advance directives may save money if executed early in the disease course. Nevertheless, Emmanuel and Emmanuel[13,14] estimated that if all Medicare recipients had advance directives and elected hospice, savings would only amount to 3%.

Changes in Process or Structure of Care

Role of the Nurse

Coordinated care with a nurse in charge of resource use has been shown to preserve palliative care quality and lower cost. A randomized clinical trial of a nurse coordinator for terminally ill patients in England maintained outcomes: Most patients still had some unrelieved symptoms, but patient and family satisfaction was helped slightly.[25] Total per patient costs were reduced from £8,814 to £4,414 (-41%), primarily as a result of decreased hospital days.[26]

The US Medicare Hospice Benefit requires similar nurse coordination, team management, easy access to low per diem hospital beds for respite or temporary care, and expanded drug coverage,[12,27] but at present only 15% of the eligible population uses it. (For the most recent statistics, see Basic Statistics About Hospice on the National Association of Hospice Care's website, http://www.nahc.org/Consumer/hpcstats.html.)

In a Canadian study,[28] home nursing care was associated with more patients dying at home. It is difficult to predict whether US physicians (or, for that matter, patients) would accept mandatory nurse coordination—for instance, the nurse could direct the doctor to make a home visit—and be willing to lose patient control, reduce supportive care use, and lose income.

Increased Access to Care

The Regional Palliative Care Program (RPCP) in Alberta, Canada, increased access of terminally ill patients to palliative care with an additional four full-time physicians and nurse consult teams, 56 hospice beds at three different sites, and greater availability of home nursing care. Fees for family physician home and hospice visits were increased, and the palliative care team offered extensive patient education.

On analysis, the RPCP significantly increased access to palliative care, increased family practice physician participation in palliative care, and reduced costs of care (Table 4).[29] Approximately 85% of patients discharged from the cancer center chose to continue care under the direction of their own family physician, with the RPCP available as needed. The authors estimated that the program saved the province a total of $2,500,000 as a result of decreased use of acute care facilities.[29]

Our own group developed the Rural Cancer Outreach Program (RCOP) between rural hospitals and the Medical College of Virginia’s Massey Cancer Center to bring state-of-the-art cancer care to medically underserved rural patients. The key to the program is the coordination of all aspects of care by nurses at the rural centers and at Massey Cancer Center. State-of-the-art care, clinical trials, and palliative care have improved at the rural centers.[30] Pre- and post-RCOP financial data were collected on 1,745 cancer patients. The cost for each rural patient admitted to the Medical College of Virginia fell from $12,268 to $7,370 (-40%), compared to only a 2% decrease for all other cancer patients—consistent with other coordinated programs.[31]

Other Coordinated Care Approaches

The City of Hope changed the culture of pain management with enhanced institutional education programs, a highly visible and respected consultative team, and a pain resource center for nurses and families. This was associated with a decrease in admissions for pain control with marked cost savings (Table 5).[32] The study was not randomized and could not account for other significant changes such as the growth of managed care with restricted admission policies. However, a reasonable conclusion is that this program led to better pain management and probably saved money.

Coordinated ethics and intensive care unit planning for patients who are likely to die also matches resource use to care and save money. An ethicist in the surgical intensive care unit (SICU) taught the staff about issues of patient choices in dying and the ethics of futile care. For those who died, there was a decrease in length of SICU stay from 28 to 16 days and a decrease in SICU days from 2,028 to 1,003 days, far greater than was observed in other parts of the hospital; savings were estimated at $1.8 million.[33]

Similarly, Dowdy and colleagues conducted mandatory proactive ethics consultations for all patients who had been mechanically ventilated beyond 4 days. They found a reduced use of the intensive care unit—due to either the discontinuation of futile care or the transfer of patients to lesser-intensity units—and a decrease in costs.[34]

Coordinated care that uses a less expensive setting reduces costs. Despite higher costs for drug equipment and nursing, home narcotic infusions produced lower total costs due to fewer hospital costs.[35]

We recently started an inpatient palliative care program based on the following principles: Only expert specialized caregivers are used, care is standardized with algorithms, and a high-volume unit is employed. Preliminary analysis supports the premise that the quality of care can be maintained and that costs can be substantially reduced compared to conventional care (unpublished data, T.J. Smith et al).