CONTEXT

Late hospice enrollment exacts a substantial toll from patients, families, hospices, and society. The relationship between the propensity for late hospice enrollment and aggressive health services received at the end of life (EOL) has been underinvestigated.

OBJECTIVES

To identify determinants of hospice enrollment within the last three days of life.

METHODS

Retrospective population-based cohort study using administrative data for 31,529 Taiwanese cancer decedents who used hospice care in their last year of life.

RESULTS

Rates of hospice enrollment within the last three days of life (16.80%-18.73%) remained constant over 2001-2006. After adjustment for patient demographics and disease characteristics, physician specialty, availability of health care resources at the hospital and regional levels, and historical trends, late hospice enrollment was more likely if Taiwanese cancer patients received chemotherapy, had multiple emergency room visits or hospital admissions, and used the intensive care unit in their last month of life (adjusted odds ratio [95% confidence interval] (AOR [95% CI]): 1.61 [1.44-1.80], 1.40 [1.29-1.52], 1.78 [1.51-2.09], and 1.45 [1.19-1.76], respectively). Late hospice enrollment was less likely for patients with hospital stays>14 days or who received cardiopulmonary resuscitation in their last month of life (AOR [95% CI]: 0.51 [0.45-0.58] and 0.41 [0.25-0.65], respectively).

CONCLUSION

Aggressive EOL care played a more significant role than patient, physician, or hospital characteristics in determining the propensity of Taiwanese cancer patients to be enrolled in hospice care within their last three days of life. Clinical and health policies should aim to avoid aggressive care when it will not benefit patients but may preclude timely hospice enrollment.