Various chronic diseases, such as chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF) and end-stage liver disease (ESLD), have such unpredictable courses that it is difficult to anticipate when death will occur in patients with these illnesses. Fox and other investigators for the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) tested the ability of currently used hospice criteria to identify patients whose prognosis for survival is less than six months.

Eligibility for hospice requires that the patient's physician and the hospice medical director certify that the patient has a life expectancy of less than six months given the normal course of the illness. Recently, the Department of Health and Human Services Office of the Inspector General has, in some cases, been requiring reimbursement to Medicare if a hospice patient survives beyond six months. Determining a patient's length of survival, then, is critical; otherwise, appropriate referrals to a hospice will be curtailed because of the perceived threat of reimbursement. The National Hospice Organization (NHO) has drafted disease-specific guidelines for determining prognosis. Although these guidelines are widely used, they have not been validated and are to be used with clinical judgment. This study compares the accuracy of various methods of determining prognosis in noncancer illness.

Data from the SUPPORT study were used. Adult patients with various chronic, non-cancer illnesses, specifically COPD, CHF and ESLD, were included unless they died or were discharged within 48 hours of enrollment. Chart reviews were performed to collect information about disease history and whether patients were discharged to hospice or home care. Readmission data were also collected. Interviews with patients and their families elicited information about palliative care preferences. Three sets of inclusion criteria (broad, intermediate and narrow) were established to provide low, medium and high thresholds for determining hospice eligibility.

Of the 9,105 patients in SUPPORT, 2,954 were diagnosed with one of the illnesses in question. Of those, 2,607 survived hospitalization and were, therefore, potential hospice care candidates, but only 2 percent were actually discharged to a hospice service. One quarter of the potential hospice care candidates died within six months of discharge. When the broad inclusion criteria were used, 70 percent of the 923 patients identified as meeting the criteria lived longer than six months. The intermediate inclusion criteria identified 300 patients with a 65 percent survival rate at six months, and the narrow criteria identified 19 patients and a 53 percent survival rate at six months. Patients referred to hospice care had a median survival of 23 days (versus 842 days for patients not referred to a hospice). The authors conclude that using the most restrictive criteria would successfully exclude patients who lived longer than six months (false-positive rate: 5.1 percent) but would also exclude almost all patients (false-negative rate: 99 percent) who would be eligible for and most likely benefit from hospice care.

It remains difficult to predict a life expectancy of less than six months in gravely ill patients. Further study is needed to determine other ways to identify patients whose prognosis is poor and who would meet hospice eligibility criteria. In an accompanying editorial, Finucane discusses some of the difficulties of caring for a dying patient. It is not the technical aspect of symptom control but the task of predicting the future that is so difficult. Even when an outcome is poor and certain (e.g., in the case of an illness such as metastatic colon cancer or non-small cell lung cancer), studies have found that patients choose to have their life extended, even if it means experiencing more pain. Helping patients make the transition from fighting for a longer life to accepting the reality of death will require the physician to provide the best prognostication possible, support, communication and realism for the patient and the family.

editor's note: One of the most difficult times for a physician comes when the family of a gravely ill patient needs guidance about whether to continue treatment that may improve function or lengthen life, or whether to maximize a dying patient's comfort. Other studies have attempted to provide guidelines for helping patients and families determine whether comfort, function or prolonged life are most important to the patient. These studies guide the practicing physician in thinking about end-of-life issues in helpful ways. Unfortunately, it is unlikely that any guidelines, including the National Hospice Organization guidelines, will ever make these issues easy or straightforward.—g.b.h.