End-of-life discussions, preference concordance, and good death: insights from a nationwide mortality follow-back survey.

PURPOSE: End-of-life (EOL) discussions are integral to EOL care. However, whether EOL discussions improve the concordance between care preferences and care received, and their association with a good death, remains uncertain. This study aimed to examine the association among EOL discussions, concordance with patient preferences for EOL care, and family-reported good deaths in patients with cancer.

METHODS

This observational study analyzed 42,379 responses from national mortality follow-back survey targeting bereaved families of patients. Measurements included EOL discussions regarding resuscitation preferences (i.e., code status) and care settings, advance directives (ADs), concordance between patient preferences and received care, and family-reported good-death scores.

RESULTS

EOL discussions between patients and physicians increased the concordance for resuscitation preferences and care settings. Concordance for do-not-attempt resuscitation (DNAR) order increased (odds ratio [OR], 1.73; 95% confidence interval [CI], 1.46-2.06), as did full code (OR, 2.75; 95% CI, 2.03-3.74). Regarding care settings, preferred home care showed the highest concordance (OR, 4.60; 95% CI, 4.31-4.90). ADs were associated with concordance for full code preferences (OR, 1.63; 95% CI, 1.17-2.28) but not for DNAR. Preference concordance was associated with higher family-reported "good death" scores, with the largest difference observed for home care (71.4 vs. 60.2; difference, 11.2; 95% CI, 10.9-11.5).

CONCLUSION

EOL discussions enhanced preference-concordant care for resuscitation preferences and care settings, which were associated with better family-reported good deaths. These findings underscore the importance of integrating EOL discussions into routine oncology care to ensure high-quality EOL care that is consistent with patient preferences.