End-of-life care discussions with healthcare providers and dying experiences: A latent class analysis using the health and retirement study.

BACKGROUND: The impact of provider-patient end-of-life care conversations on the dying experience as a multi-dimensional concept among non-White population is understudied. The study examines whether such discussions are effective at improving end-of-life experiences among U.S. older adults with diverse backgrounds.

METHODS: The analytic sample featured 9,733 older adults who died between 2002 and 2019 in the Health and Retirement Study. Latent class analysis was used with sixteen end-of-life indicators, including service utilization of seven aggressive and supportive care, symptom management, and quality of care. Multinomial logistic regression was conducted to estimate the effects of provider-patient end-of-life discussions on the predicted membership.

RESULTS: Three types of end-of-life experiences were identified. People in "minimum service user with good death" (44.54 %) were least likely to use any type of medical care, either aggressive or comforting, and had best end-of-life symptom management and quality of care. Intensive care users (20.70 %) are characterized by very high use of aggressive treatments and low use of supportive care. "Extensive service user with uncomfortable death" (34.76 %) had high likelihoods of using both aggressive and comforting care and had the worst dying experience. Older adults who discussed their end-of-life wishes with providers were 49 % and 51 % more likely to be an intensive care user and extensive service user with uncomfortable death, respectively, rather than a minimum service user with good death.

CONCLUSION: Discussing end-of-life care wishes with providers is associated with worse end-of-life experiences. Efforts are needed to facilitate early initiation and effectiveness of the provider-patient end-of-life care conversation.