Graduation Year

2026

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

School of Aging Studies

Major Professor

Debra Dobbs, Ph.D.

Committee Member

William E. Haley, Ph.D.

Committee Member

Hyo Jung Lee, Ph.D.

Committee Member

Lindsay J. Peterson, Ph.D.

Keywords

advance care planning, end of life care, ethnicity, gender, race, mixed methods

Abstract

Persistent gaps remain in end-of-life care quality (EOLCQ) in the United States. Structural factors within the health care system, as well as discriminatory practices, affect the quality of care received by at-risk populations. This dissertation uses a mixed-methods research design to examine perceived EOLCQ across racial/ethnic and gender groups. Study 1 and Study 2 were guided by the intersectionality and Social Determinants of Health (SDOH) frameworks, and examined ten EOLCQ outcomes: pain, dyspnea, sadness or anxiety, having personal care needs met, having decisions made about care without enough input from decedent or family, whether the proxy or family were kept informed about decedents condition, care decisions or treatment the decedent would not have wanted, treatment with respect, religious belief discussion, and an overall EOLCQ rating. Study 1 and Study 2 used proxy-reported data on perceived EOLCQ during the last month of life (LML) from the 2013–2021 National Health and Aging Trends Study (NHATS). Study 1 analyzed hierarchical logistic and linear regression models to examine the impact of intersectional race/ethnicity and gender on EOLCQ outcomes. Study 2 analyzed the impact of advance care planning (ACP) on perceived EOLCQ and additionally explored if EOLCQ differed by life-limiting illness. Research on end-of-life care disparities often overlooks how intersectionality may impact EOLCQ outcomes. Study 3 used qualitative data collected from semi-structured interviews with seven Black family caregivers caring for a person with dementia. The purpose of the study was to examine caregivers’ perceptions of the quality of care provided by healthcare professionals, as well as their perspectives on decision-making, including ACP.

For Study 1, in fully adjusted models, significant findings were observed for pain, dyspnea, sadness or anxiety, having personal care needs met, and whether proxies or family were kept informed about the decedents’ condition. White women had higher odds of proxies reporting pain compared to White men. Black women had lower odds of proxy-reported sadness or anxiety and higher odds of proxies reporting they or family members were kept informed about the decedent’s condition. Hispanic women had greater odds of proxies reporting that their personal care needs were being met and of reporting being kept informed. Hispanic men had higher odds of proxy-reported dyspnea and lower odds of proxies reporting that personal care needs were met compared to White men. In fully adjusted models, Black men did not significantly differ from White men on perceived EOLCQ outcomes. While proxies for Hispanic and Black women reported some positive outcomes, these findings may reflect cultural factors that influenced approaches to rating EOLCQ items, including greater family involvement in care and decision-making.

In Study 2, accounting for ACP did not eliminate intersectional disparities, and ACP was not associated with EOLCQ after adjusting for SDOH, health status, and end-of-life care covariates. EOLCQ outcomes that remained significant in Study 2 included proxy-reported pain, sadness or anxiety, having personal care needs met, whether proxy or family were kept informed, treatment with respect, and overall EOLCQ ratings. Differences by race/ethnicity and gender seen between Study 1 and Study 2 included Hispanic women having higher odds of proxy-reported pain compared to White men, while Hispanic men were no longer associated with proxy-reported dyspnea. Hispanic women were also no longer associated with proxy-reported personal care needs being met or that patients or family were kept informed. Treatment with respect emerged as a significant outcome in the second study. Black men had lower odds of proxy-reported treatment with respect in the LML. Finally, Black men had lower overall proxy-reported EOLCQ compared to White men.

Life-limiting illness was also found to be associated with EOLCQ outcomes. Decedents with heart or lung disease had higher odds of proxy-reported dyspnea, and decedents with lung disease also had higher odds of proxy-reported sadness or anxiety. Decedents with heart disease had higher odds of proxies reporting that they or family were kept informed about the decedent’s condition and lower odds of reporting that care decisions were made without enough input. Lung disease was associated with proxy-reported lower overall EOLCQ ratings. Finally, a dementia diagnosis was associated with higher odds of proxy report of decedent receiving unwanted care or treatment. Black and Hispanic men had several worse proxy-reported EOLCQ outcomes compared to White men across Study 1 and Study 2. Overall, proxy-reported perceived EOLCQ varied on several outcome measures among intersecting race/ethnicity and gender groups, even after controlling for SDOH, health, and end-of-life covariates; however, fewer significant associations were found than were expected, suggesting that the gap between racial and ethnic minority groups and White men may be narrowing.

For Study 3, seven themes emerged from the data: caregiving dynamics, quality of life, navigating the healthcare system, care coordination, communication with healthcare professionals, decision-making through the disease progression, and end-of-life care discussions. Study 3 demonstrated the persistent and ongoing challenges faced by Black caregivers. Caregivers took on a high number of care responsibilities and described having to navigate a fragmented healthcare system often without adequate assistance. Caregivers reported relying on faith communities during difficult times and including family in decision-making when possible. Taken together, the findings from these three studies suggest that culturally competent, holistic approaches to end-of-life care may help address observed disparities.

Download

Share

COinS