Timing and Initiators of End-of-Life Discussions and Their Association With Family Members’ Understanding and Fulfilment of Patients’ Hopes and Wishes

Objective This study aimed to identify who initiated end-of-life discussions (EOLD) and to examine how the timing and initiator of these discussions are associated with bereaved family members’ understanding and fulfilling of patients’ medical and care-related hopes and wishes. Method This retrospective cross-sectional study analyzed anonymous self-administered questionnaires completed by bereaved family members of patients with cancer in Japanese palliative care units. The survey included items on the occurrence, depth, timing, and initiator of EOLDs, as well as family-reported understanding and fulfilling of the patient's medical and care-related hopes and wishes. Two-way analysis of variance was conducted to examine the main effects and interaction between timing and initiator. Results Of 995 distributed questionnaires, 523 valid responses (52.6%) were obtained. Among 507 respondents included in the primary analysis, 26.6% reported fully understanding the patient's hopes and wishes. EOLDs were most frequently initiated during cancer treatment, with 41.3% initiated by healthcare providers. No significant interaction was observed for family-reported understanding. In contrast, a significant interaction between timing and initiator of EOLD was observed for fulfilling patients’ hopes and wishes (partial η 2 = 0.03). In particular, healthcare provider-initiated EOLDs were associated with higher fulfilling scores when conducted early. Discussion The timing and initiator of EOLDs were associated with bereaved family members’ reports of fulfilling of patients’ hopes and wishes, although the effect sizes were modest. These findings suggest that earlier EOLDs—particularly when initiated by healthcare providers—may be associated with better outcomes in terms of fulfilling patients’ hopes and wishes. Both the timing and initiator of EOLD should be considered to support effective EOLD.