End-of-life care values and treatment preferences among Japanese breast cancer patients and families: A dyadic cross-sectional study.

BACKGROUND: Patient-family discordance in end-of-life values complicates advance care planning (ACP) but is rarely quantified in Asian oncology settings. We measured patient-family differences in values and treatment preferences in breast cancer and correlates of discordance. METHODS: In this single-center cross-sectional study at a Japanese breast-cancer unit (August 2024-May 2025), 470 patients completed a questionnaire on Life Priorities and Values (14 items) and End-of-Life Care Preferences (6 items); 218 were matched to 324 family respondents. Analyses used generalized linear mixed-effects models, Cohen's κ, multiple correspondence analysis (MCA), an actor-partner interdependence model (APIM) on MCA dimension scores, and a total discordance score. RESULTS: Patients prioritized autonomy ("Independence in ADL," OR = 0.20; "Not being a burden to family," OR = 0.31), families intervention ("Longevity," OR = 2.15; "Treatment until satisfaction," OR = 4.50; all p < .001). Concordance was uniformly poor (κ 0.03-0.32) and directional: 88.5% of disagreements on treatment pursuit were on the family-endorsement side. Prior ACP discussion was associated with lower discordance (median 6 vs 7, p < .001), chiefly for palliative items, and chemotherapy with higher discordance (p = .012). Exploratory APIM showed strong actor associations and two partner associations for family Preference Dim2: patients' greater value engagement and longevity orientation were each associated with higher family endorsement of active treatment (p = .022 and 0.002). CONCLUSIONS: Systematic, directional patient-family discordance was present across nearly all end-of-life domains. Prior ACP discussion was associated with lower discordance, particularly for palliative preferences, supporting structured dyadic communication in breast cancer care.