End-of-Life Care Discussions with Doctors: Evidence from the United States and China

TitleEnd-of-Life Care Discussions with Doctors: Evidence from the United States and China
Publication TypeThesis
Year of Publication2023
AuthorsLou, Y
Academic DepartmentSocial Work
UniversityColumbia University
CityNew York City
Keywordsdiscrimination in medical care, health planning, Hospice Care, Medicare, minority older people-care, social service, Terminal Care

Chapter 1 presents the backgrounds and research questions of the dissertation project. Although everyone hopes to die with dignity, a “good death” is often privileged. Inequality in end-of-life care in the United States is striking; older minorities are 3 times more likely to receive intrusive care and experience hospital readmission and less likely to use hospice care and die at home. With the hope to address disparities in end-of-life outcomes, as part of the Affordable Care Act, Medicare expanded its benefit and started to reimburse end-of-life care planning services from 2016. This dissertation examines the roles of health insurance policy and its advocated intervention in ensuring person-centered dying experiences in the United States and the public readiness for establishing a similar end-of-life care policy in China.

The second chapter examines whether physician–patient advance care planning (ACP) discussions, the “intervention” embedded in the policy, is effective at prompting end-of-life care planning behaviors and improving end-of-life experiences. I also studied the heterogenous effects by race and ethnicity. I used exit data about 7,282 deceased Medicare-eligible respondents from the Health and Retirement Study (HRS) from 2012 to 2020, a nationally representative biennial survey of U.S. older adults. I conducted latent class analysis to identify different end-of-life experiences and used multinomial logistic regression models to study the relationship between having ACP discussions with doctors and different end-of-life experiences in full and stratified samples by race and ethnicity. Then I used path analyses to test whether completing legal documentation initiated by ACP services explains the effectiveness of the intervention. The results show three types of end-of-life experiences among White, Black, and Hispanic older adults; significant effects of end-of-life care discussions with doctors on uncomfortable dying experiences and extensive use of invasive treatments in the older White population; and the partial mediating role of completion of legal health care documents on the relationship between ACP discussions and end-of-life outcomes. This study identified the limited effectiveness of physician–patient ACP discussions among minority older adults and emphasizes the importance of quality, content (e.g., to cover legal documentation during the conversation), and cultural appropriateness of ACP consultations led by doctors.

The third chapter examines the heterogenous intent-to-treat effect of the Medicare policy on older adults’ motivation to plan for care by exploiting policy intervention as a natural experiment. Taking advantage of the longitudinal nature of the HRS, I used core data with 47,982 observations nested in 15,647 individuals across 9 years (2012–2020). I used two quasi-experimental models—interrupted time series analysis and difference-in-difference—to study the care-planning outcomes before and after the policy implementation between older adults eligible or not eligible for Medicare. I studied the heterogenous effects in stratified samples by race and ethnicity, immigration background, and socioeconomic status proxied by education and wealth. I used coarsened exact matching to address the potential bias of the comparison group not being the ideal counterfactual control unit. I found that in the total population, the Medicare policy was associated with a significant increase in the proportion of older adults who completed a living will and designated a durable power of attorney for health care, but not having an ACP conversation. Furthermore, although there were immediate positive policy effects, the increases in ACP outcomes driven by the ACA Medicare benefit expansion were not sustained. Finally, I found no evidence that the ACA Medicare benefit expansion was effective in improving the rates of ACP among Black, Hispanic, Asian, Native American, or non-U.S.-born older adults. I also identified heterogeneous policy effects by socioeconomic status. I discuss promising recommendations for health care policy and practice to improve the participation rate of end-of-life care planning.

Chapter 4 explores the feasibility of encouraging physicians to discuss end-of-life care wishes with older patients in China. Knowledge of a medical diagnosis is critical for end-of-life decision making. However, a patient’s right to know their diagnosis is neither guaranteed nor protected by law in China. Doctors typically inform family members of the patient’s illness diagnosis and prognosis and let them decide whether to disclose this information to the patient. This study examined middle-aged and older Chinese adults’ preferences for disclosing their and significant others’ diagnoses and whether sociodemographic, economic, and cultural factors were associated with these preferences. The research team surveyed 571 adults aged 50 or older in Shanghai from 2021 to 2022. The outcome measures are preferences for diagnosis disclosures, including whether and to whom diagnoses should be disclosed. I characterized preference types using latent class analysis and estimated multinomial logistic regression to identify the covariates associated with each preference type. Three latent classes were identified. The transparent group (34%) believed patients, whether self or a significant other, should be informed of their diagnoses. The contradictory group (50%) held conflicting attitudes; they preferred to know their diagnosis but preferred that significant others not be informed of their diagnoses. The avoidant group (16%) preferred not to know their diagnosis or to disclose significant others’ diagnoses to them. Familism attitudes were positively associated with holding contradictory views. Experiences with hospitalization and medical decision-making for family members were associated with holding transparent views. I discuss the importance of illness disclosure for honoring patients’ autonomous decision-making.

Citation Key13329