Abstract
BACKGROUND: End-of-life discussions for patients with advanced cancer are internationally recommended to ensure consistency of end-of-life care with patients' values. This study examined the elements of end-of-life discussions associated with end-of-life care. MATERIALS AND METHODS: We performed a prospective observational study among consecutive patients with pretreated non-small cell lung cancer after the failure of first-line chemotherapy. We asked oncologists whether they had ever discussed "prognosis," "do not attempt resuscitation," "hospice," and "preferred place of death" with a patient at baseline. The quality of life (QOL) and depressive symptoms of patients were assessed using validated questionnaires at baseline and 3 months later. The end-of-life care that patients received was investigated using medical records. Oncologists' compassion and caregivers' preferences for hospice care were also assessed using questionnaires. Multiple regression analyses were conducted to examine the association between elements of end-of-life discussions and patient-reported outcomes as well as actual end-of-life care. RESULTS: We obtained 200 valid responses at baseline, 147 valid responses 3 months later, and 145 data points for medical care at the end-of-life stage. No element of the end-of-life discussion between the patient and their oncologist was significantly associated with patients' reported outcomes or actual end-of-life care. In addition, oncologists' compassion was significantly associated with improvement in both comprehensive QOL and depressive symptoms, and caregivers' preferences for hospice care and high educational level were significantly associated with hospice death. CONCLUSION: Oncologist-patient alliances and caregivers' involvement in end-of-life discussions may be influential in achieving optimal end-of-life care.