Predictors of live hospice discharge: data from the National Home and Hospice Care Survey (NHHCS).

In the context of the current regulatory environment, patients may be discharged from hospice if their condition stabilizes or improves over time and the certifying physician is unable to provide a conscientious recertification of the six-month prognosis. Little is known about the characteristics or outcomes of patients who are determined by physicians to no longer be eligible for hospice care. This retrospective study was designed to characterize the rates and predictors of live hospice discharge using data from the 1996 and 1998. National Home and Hospice Care Survey (NHHCS). We compare records of live hospice discharge with hospice discharges due to death from the discharge patient files of the 1996 and the 1998 NHHCS. Of the 807,733 patients in the combined 1996 and 1998 NHHCS discharge patient file who met study inclusion criteria, 761,858 (94 percent) were deceased and 45,875 (6 percent) were discharged alive. Those who were discharged alive were more likely to be female, have received hospice care for more than 60 days, and to have had a noncancer diagnosis, particularly advanced cardiopulmonary or neurologic disease. Mean age and total number of assistive medical devices used did not differ significantly between patients who were discharged alive and those who died in hospice care. Factors most associated with live hospice discharge, using bivariate analyses, were length of service greater than 60 days (OR, 6.60; 95 percent CI, 6.47-6.73), cardiopulmonary diagnosis (OR, 3.24; 95 percent CI, 3.19-3.30), and neurologic diagnosis (OR, 2.73; 95 percent CI, 2.67-2.79). Multivariate logistic regression identified length of service greater than 60 days, cardiopulmonary diagnosis, neurologic diagnosis, female gender, worse functional status, and living in an institutional setting as being independently associated with live hospice discharge. We found that patients who were discharged alive from hospice care were more likely to have longer lengths of service, noncancer diagnoses, and better functional status than those who died while receiving hospice care. Age was not associated with discharge disposition. Given the demonstrated differences between these patients and those who died while receiving hospice care, these data provide further impetus for careful study of the appropriateness of current hospice eligibility criteria, the determinants of hospice discharge and, most important, the outcomes of patients who are discharged alive from hospice and the impact of hospice discharge on patients and their families.