The research hypotheses in the study are as follows; 1) There are racial and ethnic differences in AD completion after controlling for covariates identified by Critical Race Theory (CRT) and Andersen’s Behavioral Model of Health Services, and 2) The effect of race and ethnicity on AD completion is mediated by SES, after controlling for covariates.
Methods: To test the research hypotheses, the present study conducts a secondary data analysis of Health and Retirements Study (HRS) 2004-2014 wave. The sample includes 6,518 participants who have died at age of 65 or above. The sample for the HRS is selected using a stratified multi-stage area probability sampling.
The independent variable is measured by a question regarding the respondents’ race and ethnicity with the following response categories: non-Hispanic whites, non-Hispanic blacks, and Hispanics. The dependent variable is measured by the completion rate of two forms of AD, living wills or Durable Power of Attorney for Health Care (DPAHC). The variable of SES is assessed by the deceased participant’s educational attainment and income.
First, hierarchical logistic regression examines the racial and ethnic effect on AD completion, while controlling for 12 covariates: gender, age, marital status, religion, place of birth, educational attainment, income, cognitive function, limitations in physical functioning, geriatric syndromes, and the number of progressive chronic disease. Next, in the mediation analysis, Karlson, Holm, and Breen (KHB) method examines the mediating effect of SES, education and income, on the racial and ethnic disparities in AD completion.
Results: The results show that race and ethnicity was a significant predictor of AD completion (p≤0.001). As compared to non-Hispanic whites, Non-Hispanic blacks were least likely, followed by Hispanics, to have living wills or DPAHC.
Mediation analysis reveals that both education and income explained 14.4% of racial and ethnic differences in completion of living wills for non-Hispanic blacks and 17.6% for Hispanics. Similarly, education and income accounted for 17.6% of racial and ethnic disparities in completion of DPAHC for non-Hispanic blacks and 20.9% for Hispanics. In particular, education had a larger mediating effect on the outcome variables than income.
Discussion: The present study provides evidence for the causal mechanism generating racial and ethnic disparities in AD completion. Based on the findings, social workers should make targeted efforts for people of color with lower education levels and/or lower income to promote executing AD which is vital for receiving higher quality EOL care.