Where There’s a Will: The Link Between Estate Planning and Disparities in Advance Care Planning by White and Black Older Adults

Abstract

Data from the Health and Retirement Study (n = 6,946) were used to test whether differences in estate planning accounted for disparities in advance care planning between White and Black older adults. White participants were more likely to have advance directives after controlling for demographic, health, and financial variables. When estate planning was also controlled, the odds of having an advance directive were equal for White and Black participants. In contrast, Whites remained more likely to discuss end-of-life preferences after controlling for demographic, health, financial, and estate planning variables. White participants were almost four times as likely to have wills or trusts. Wealth, income, and home ownership were predictive of estate planning. Financial disparities contributed to lower rates of estate planning which in turn explained in large part why Black older adults were less likely to have advance directives but did not account for race disparities in advance care discussion.

Keywords

advance care planning advance directives estate planning medical decision making race disparities

Approximately 40% of older adults in the United States require surrogate decision making at the end of life because they become unable to formulate or articulate instructions about their health care (Silveira, Kim, & Langa, 2010). When no information about a person’s preferences is available, the general legal presumption is that the individual would choose to receive all medical treatment necessary to sustain life (Institute of Medicine, 2015). However, if a person adequately communicates his or her preferences about life-sustaining treatment in advance while still competent, this presumption no longer applies and the patient’s expressed wishes should guide medical decisions. The ongoing process of reflecting on and informing loved ones and health care providers about how future medical decisions should be made in the event of incapacity is called advance care planning (ACP) (Pearlman, Cole, Patrick, Starks, & Cain, 1995).

ACP is associated with some important end-of-life quality indicators, including dying at home rather than in a hospital and receiving hospice care before death (Bischoff, Sudore, Miao, Boscardin, & Smith, 2013; Teno, Gruneir, Schwartz, Nanda, & Wetle, 2007). Numerous studies have observed that Black older adults are significantly less likely to engage in ACP compared to their White counterparts (Gerst & Burr, 2008; Koss & Baker, 2016; Kwak & Haley, 2005; Rao, Anderson, Lin, & Laux, 2014). This is particularly concerning given the relationship between ACP and end-of-life care. Black elders are less likely to receive hospice or other palliative care at the end of life and are less likely to have their pain effectively managed (Bullock, McGraw, Blank, & Bradley, 2005; Crawley et al., 2000).

Proposed explanations for relatively low rates of ACP among Black older adults include mistrust of doctors and the health care system, religiosity, reluctance to acknowledge terminal prognosis, greater desire for life-sustaining treatment, lower health literacy, and doubt about the efficacy of ACP (Kwak & Haley, 2005; Ladd, 2014; Sanders, Robinson, & Block, 2016). Despite numerous studies testing many of these theories, the underlying mechanisms for race disparities in ACP are still not well understood. One possible contributing factor that has yet to be examined is disproportionality in estate planning (i.e., the making of a will, trust, or other legal document to dispose of property after death). Having a will is highly predictive of having an advance directive (Carr, 2012; Kelly, Masters, & Deviney, 2013; Su, 2008). Although the association between ACP and estate planning has been suggested as a possible reason for race disparities (Carr, 2012), it has not been empirically assessed.

Using data from the Health and Retirement Study (HRS), the present study addresses this gap in the literature by testing whether ACP disparities between White and Black older adults can be explained by differences in estate planning. We also examined the relationships between estate planning and demographic, health, financial, and ACP variables. We hypothesized that accumulative financial disadvantages would account for lower rates of estate planning among older Blacks. Estate planning was predicted to explain race disparities in ACP, particularly advance directive completion.

Literature Review

ACP

ACP is an ongoing process of contemplating one’s wishes about medical care, discussing preferences with others, documenting instructions, and periodically revisiting and revising one’s choices as circumstances change (Sudore et al., 2008). The process often results in a legal document called an advance directive, the format of which varies but usually consists of two parts: a living will and a durable power of attorney for health care (IOM, 2015). The living will allows individuals to state what medical treatments they would or would not wish to receive under certain conditions such as a persistent vegetative state or irreversible and severe cognitive impairment. The durable power of attorney for health care is used to nominate one or more health care proxies to act on one’s behalf in the event of incapacity. Individuals may also communicate their wishes verbally instead of or in conjunction with completing advance directives.

ACP is associated with a reduction in aggressive life-sustaining interventions at the end of life (Brinkman-Stoppelenburg, Rietjens, & van der Heide, 2014; Teno et al., 2007; Tschirhart, Du, & Kelley, 2014, Wright et al., 2008). Patients who engage in ACP are more likely to be admitted to hospice and to receive hospice care for longer periods before death (Bischoff et al., 2013; Brinkman-Stoppelenburg et al., 2014; Greiner, Perera, & Ahluwalia, 2003; Wright et al., 2008). Advance directive completion has been found to lower the odds of dying in the hospital, although it does not seem to reduce hospital admissions in the final months of life (Bischoff et al., 2013; Kessler & McClellan, 2004; Silveira et al., 2010; Teno et al., 2007). Among patients at high risk of death, those who complete advance directives and/or discuss their end-of-life wishes with physicians are less likely to be admitted to intensive care units. (Khandelwal et al., 2015; Wright et al., 2008).

Psychological benefits of ACP for both patients and family decision makers have also been documented (Wright et al., 2008). At least one third of those who make treatment decisions on behalf of an incapacitated patient experience anxiety, stress, or other negative emotional responses as a result (Wendler & Rid, 2011). Knowing what the patient would have wanted has been found to reduce these negative effects (Abbott, Sago, Breen, Abernethy, & Tulsky, 2001; Tilden, Tolle, Nelson, & Fields, 2001; Wendler & Rid, 2011).

Given these benefits, considerable effort is made in the United States to encourage ACP, particularly by older adults. All hospitals, long-term care facilities, hospices, home health agencies, and health maintenance organizations that receive Medicare or Medicaid funding are mandated by the 1990 Patient Self Determination Act to provide information about advance directives at the time of admission or enrollment. In 2015, the Centers for Medicare and Medicaid Services approved new reimbursement codes permitting physicians and other qualified health care providers to bill Medicare for time spent with patients discussing medical preferences and completing advance directives (Sabatino, 2015).

Approximately half of U.S. adults age 65 and older report having completed a living will or advance directive, and rates of advance care discussion are somewhat higher (Pew Research Center, 2009; Rao et al., 2014). Factors that have been found to be positively associated with ACP include being White, older age, widowhood, female gender, education, income, net worth, hospitalization, and nursing home admission (IOM, 2015; Pew Research Center, 2009; Rao et al., 2014; U.S. Government Accountability Office, 2015).

Black older adults are much less likely than Whites to complete advance directives or engage in advance care discussion (Gerst & Burr, 2008; Koss & Baker, 2016; Kwak & Haley, 2005; Rao et al., 2014). Sanders, Robinson, and Block (2016) recently conducted an extensive literature review and concluded that current data are not able to provide definitive explanations as to why Blacks are less likely to engage in ACP. One study found that, although no single factor accounted for lower rates of advance directive completion among Black participants, the combination of greater preference for life-sustaining treatment, higher levels of discomfort discussing death, greater mistrust in the health care system, and spiritual beliefs did explain race disparities (Johnson, Kuchibhatla, & Tulsky, 2008). Another study found the differences in the odds of engaging in advance care discussion for White and Black older adults were no longer statistically significant once belief that God determines the timing of death was controlled (Carr, 2011). Controlling for belief also substantially reduced the race gap in advance directive completion. In contrast, one of the few studies on the subject using a nationally representative data set found that the relatively low rate of advance directive completion among Black older adults was only modestly explained by religiosity, personal health values, and demographic variables (Huang, Neuhaus, & Chiong, 2016).

Relationship Between ACP and Estate Planning

ACP, particularly written documentation of medical wishes in the form of an advance directive, is frequently carried out with the assistance of an attorney or financial planner as part of estate planning (Kelly et al., 2013; Pollack & Williams, 2010). Many of the same variables have been found to be predictive of both estate and health care planning, including race, age, education, and socioeconomic status (Carr, 2012; Goetting & Martin, 2001; Kelly et al., 2013; Su, 2008). Various types of financial resources may influence estate planning in different ways. For example, greater overall wealth may motivate tax planning while home ownership, regardless of property value, may cause some to engage in estate planning to ensure the smooth transfer of the home. Those with substantial retirement incomes are likely to derive a portion of their incomes from investments and may therefore be more inclined to seek professional financial advice. Although it is generally acknowledged that Black older adults are less likely to engage in estate planning, participants in prior studies have tended to be predominantly White, and there is little known about factors influencing estate planning among Blacks or other racial or ethnic groups.

The potential relationship between race disparities in estate planning and ACP can be understood through accumulative disadvantage theory, which suggests that differences in opportunities and obstacles throughout the life course may lead to increasing inequalities (O’Rand, 1996). These differences—resulting from both structural forces (e.g., discriminatory policies and practices) and individual life trajectories (e.g., events that determine the range of available options)—manifest in financial, social, technological, and informational resource disparities that grow over time. These resource inequities contribute to persistent health disparities by influencing health behaviors and outcomes (Phelan, Link, & Tehranifar, 2010).

Accumulative disadvantages lead to Black older adults entering later life with substantially fewer financial resources than Whites. Differences in educational and employment opportunities across the life span result in poverty rates that are twice as high among older Blacks (17.6%) than Whites (7.4%) (Administration on Aging, 2014). The median income for households headed by older Whites in 2013 was more than US$11,000 higher than for households headed by Black older adults. Housing discrimination and segregation contribute to racial inequalities in home equity (Krivo & Kaufman, 2004). This study is the first to examine how these accrued economic disadvantages may account for ACP disparities between White and Black older adults through the intermediary mechanism of estate planning.

Research Design

Sample and Data

Data from the HRS, a longitudinal survey of older adults sponsored by the National Institute on Aging (Grant NIA U01AG009740) and conducted by the University of Michigan, were used for this investigation. Summary financial data were obtained from the RAND HRS data file (Version O), a subset of the HRS that is processed and aggregated across multiple waves. The sampling design of the HRS was accounted for by adjusting for subjects’ clustering, stratification, and most recent sampling weights. When these adjustments are made, the HRS data are representative of the noninstitutionalized U.S. population over the age of 50 (Sonnega et al., 2014).

Analyses were conducted on a subset of 2012 HRS data made up of 6,946 non-Hispanic Blacks and non-Hispanic Whites age 65 and older who resided in the 50 United States and had been assigned a respondent-level sampling weight. One person was randomly selected from each household with more than one respondent.

Measures

Advance directive completion

Advance directive completion was measured with two questions: (1) “Have you provided written instructions about the care or medical treatment that you want to receive if you cannot make those decisions yourself? This is sometimes called a ‘Living Will.’” and (2) “Have you made any legal arrangements for a specific person or persons to make decisions about your care or medical treatment if you cannot make those decisions yourself? This is sometimes called a ‘Durable Power of Attorney for Health Care.’” Participants who responded yes to either or both of these questions were coded 1. Those who responded negatively to both questions were coded 0.

Advance care discussion

Engagement in advance care discussions was measured with a single item yes/no question: “Have you ever discussed with anyone the care or medical treatment you would want to receive if you were to become seriously ill in the future?”

Estate planning

Estate planning was measured with two questions: (1) “Do you currently have a will that is written and witnessed?” and (2) “Have you [and your] [husband/wife/partner] put any of your assets into a trust?” A positive response to either question was coded 1. Participants responding in the negative to both questions were coded 0.

Household wealth

Data on net household assets were obtained from the RAND file. RAND calculated net household wealth by summing the values of all assets (including the home, other real property, bank accounts, investments, and vehicles) and subtracting all debt (Chien et al., 2014). These values may be less than 0 if total debt exceeds total wealth. We averaged the total household net assets across the 2006, 2008, 2010, and 2012 HRS waves to capture a more complete picture of wealth over time, including periods before and after the 2008 economic crisis. Negative average values were changed to 0. A nominal $250 was added before taking the natural log.

Household income

Data on household income were taken from the RAND file. RAND derived total annual household income by summing all types of income received during the previous calendar year by respondents and their spouses, including employment, Social Security, pensions, government benefits, and investment income (Chien et al., 2014). Total annual household income was averaged across 2006, 2008, 2010, and 2012 waves. A nominal $250 was added before taking the natural log.

Home ownership

Home ownership was assessed based on a single question asked in 2012: “Do you [and your] [husband/wife/partner] own your home, rent it, or what?” Those who reported owning their home or being in the process of buying were coded as home owners. Participants who reported renting their homes, living with relatives or friends, or having other living arrangements were coded as non-home owners (reference group).

Race

Race, the primary predictor variable, was coded as either non-Hispanic White or non-Hispanic Black (reference group).

Covariates

Covariates included gender, age in years, education (less than high school, high school/General Educational Development Test (GED), some college, college or more), marital status (married/partnered, widowed, separated/divorced, never married), self-rated health status (reverse coded so that 1 = poor and 5 = excellent), and whether the respondent had spent at least one night in a hospital or nursing home during the prior two years.

Analyses

Weighted descriptive statistics were calculated for the total sample and stratified by race. A set of three logistic regression models were estimated to examine to what extent accumulated financial disparities may account for race disparities in estate planning. The first model regressed estate planning on race, age, gender, education, marital status, self-rated health, prior hospitalization, and nursing home stay. The three financial predictor variables of household income, household net assets, and home ownership were added to the second model. Although we hypothesized that estate planning would explain race disparities in ACP, we tested the reverse possibility that ACP may contribute to race disparities in estate planning by including advance directive completion and advance care discussion as predictor variables in the third model.

To test the hypothesis that estate planning would account for race disparities in ACP, we first regressed advance directive completion on race controlling for age, gender, education, marital status, self-rated health, prior hospitalization, and nursing home stay. Because advance directive completion and advance care discussion are correlated, we also included advance care discussion as a covariate. In the second model, we added the three financial variables. The estate planning variable was included in the third model. A similar set of three logistic regressions were estimated to test the hypothesis that estate planning would explain race disparities in advance care discussion, with advance directive completion being included as a covariate.

STATA Version 14 was used to prepare data for analysis and generate descriptive statistics. Multivariable logistic regression models were estimated in Mplus Version 7 with robust maximum likelihood estimation. Missing data were handled with full information maximum likelihood estimation (Enders, 2010).

Ethical Considerations

This study was exempted from institutional review board review by the University of Kansas.

Results

Descriptive Statistics

Weighted descriptive statistics are presented in Table 1. The sample included 1,198 non-Hispanic Black and 5,748 non-Hispanic White older adults. Approximately two thirds of the White participants had completed advance directives (63%) or discussed their medical care preferences (65%). In constrast, fewer than half of the Black participants had advance directives (36%) or engaged in advance care discussions (40%). There were also substantial differences in rates of estate planning, with 73% of Whites and 27% of Blacks having a will and/or trust (Figure 1).

Table 1.

Sample Characteristics (Weighted).

Variable	Total (n = 6,946)	Black (n = 1,198)	White (n = 5,748)	p Value
Female	58	62	57	.02
Age	75.0	74.2	75.1	<.001
Education
Less than high school	16	37	13	<.001
High school/GED	38	32	38
Some college	23	19	23
College or above	24	12	25
Marital status
Married	47	30	49	<.001
Divorced	15	26	14	<.001
Widowed	32	34	32	.32
Never married	6	10	5	<.001
Self-rated health
Excellent	8	4	8	<.001
Very good	32	20	33
Good	33	38	32
Fair	20	27	19
Poor	8	11	8
Hospitalized in past 2 years	31	33	31	.20
Nursing home in past 2 years	6	5	6	.35
Household net assets^a	215,887	59,600	276,000	<.001
Household income^b	34,869	21,592	37,805	<.001
Home ownership	79	63	80	<.001
Will and/or trust	69	27	73	<.001
Advance directive	59	36	63	<.001
Advance care discussion	62	40	65	<.001

Note. Means are shown for age; medians are shown for household net assets and household income; percentages are shown for all other variables. Probability values are based on χ² tests for categorical variables, point biserial correlations (unweighted) for continuous variables, and Mann–Whitney U tests (unweighted) for ordinal variables.

^aNet value of household assets averaged across 2006, 2008, 2010, and 2012. ^bHousehold income averaged across 2006, 2008, 2010, and 2012.

Figure 1.

Weighted percentages of Black and White participants 65 and older who reported in 2012 having a will and/or trust, advance directive, and advance care discussion (n = 6,946).

The majority of the sample was female for both race groups. The mean age for Whites (75.1) was slightly higher than for Blacks (74.2). The White participants tended to have more education and were more likely to be married than the Black participants, whereas a larger percentage of Blacks were divorced or never married. An approximately equal proportion were widowed. White participants were on average healthier, but there was no significant difference in rates of prior hospitalization or nursing home stay.

Home ownership was less common among Blacks (63%) than Whites (80%). The median annual household income of Black participants was over $16,000 less than the median household income of White participants ($21,591 vs. $37,805, respectively). Median household net worth was more than four times lower for Black participants ($59,600) than White participants ($276,000).

Multivariable Logistic Regressions

Estate planning

The results from the logistic regresson models predicting estate planning are presented in Table 2. Model 1 showed the odds of having a will or trust were almost six times as high for Whites compared to Blacks (odds ratio [OR] = 5.88, p < .001) after controlling for age, gender, education, marital status, self-rated health, hospitalization, and nursing home stay. Female gender, age, education, better self-rated health, and nursing home stay were each positively associated with higher odds of engaging in estate planning. Compared to married older adults, those who were divorced or never married were less likely to have wills or trusts.

Table 2.

Logistic Regression Models Predicting the Odds of Estate Planning.

	Model 1	Model 2	Model 3
Variable	OR [95% CI]	OR [95% CI]	OR [95% CI]
White (ref = Black)	5.88* [5.07, 6.80]	4.19* [3.57, 4.91]	3.74* [3.17, 4.41]
Female	1.27** [1.13, 1.42]	1.35* [1.20, 1.53]	1.22** [1.08, 1.37]
Age (years)	1.08* [1.07, 1.08]	1.08* [1.07, 1.09]	1.07* [1.06, 1.08]
Education	1.69* [1.56, 1.83]	1.34* [1.24, 1.45]	1.21* [1.11, 1.32]
Marital status (ref = married)
Divorced	0.51* [0.42, 0.61]	1.06 [0.85, 1.31]	0.95 [0.76, 1.18]
Widowed	0.91 [0.79, 1.04]	1.35** [1.16, 1.56]	1.13 [0.95, 1.33]
Never married	0.41* [0.31, 0.55]	0.71 [0.53, 0.95]	0.68 [0.48, 0.96]
Self-rated health	1.20* [1.14, 1.27]	1.05 [0.99, 1.12]	1.07 [1.00, 1.14]
Nursing home stay in past 2 years	1.32*** [1.05, 1.67]	1.39*** [1.05, 1.82]	1.03 [0.77, 1.37]
Hospitalized in past 2 years	1.17 [1.02, 1.33]	1.24*** [1.08, 1.44]	1.10 [0.96, 1.27]
Household net assets^a		1.48* [1.42, 1.55]	1.39* [1.32, 1.47]
Household income^b		1.18*** [1.06, 1.33]	1.21*** [1.06, 1.37]
Home ownership		1.34** [1.13, 1.60]	1.84* [1.51, 2.24]
Advance directive			7.02* [6.13, 8.03]
Advance care discussion			1.12 [0.97, 1.30]
R ² (McKelvey and Zavoina)	.27	.39	.51

Note. n = 6,946. OR = odds ratio; CI = confidence interval.

^aNet household assets averaged across 2006, 2008, 2010, and 2012, log-transformed. ^bTotal household income averaged across 2006, 2008, 2010, and 2012, log-transformed.

*p < .05. **p < .01. ***p < .001.

Adding the three financial variables of household net assets, household income, and home ownership to Model 2 narrowed the gap between White and Black participants, but Whites were still more than four times as likely to engage in estate planning (OR = 4.19, p < .001). Those who were female, older, more educated, and had spent at least one night in a nursing home in the past two years were still more likely to have a will or trust. The relationships between estate planning and self-rated health as well as being divorced or never married became nonsignificant while widowhood now displayed a positive association. Persons who were hospitalized in the past two years were also more likely to have a will or trust.

When advance directive completion and advance care discussion were added to Model 3, the race disparity was somewhat reduced but the likelihood of having a will or trust remained significantly higher for Whites (OR = 3.74, p < .001). The odds of having a will or trust if one had an advance directive were seven times the odds for those without an advance directive. The relationship between estate planning and advance care discussion was not statistically significant. Widowhood, hospitalization, and nursing home stay were no longer significantly related to having a will or trust. The effects of the other covariates remained significant.

Advance directive completion

The results from the logistic regressions predicting advance directive completion are presented in Table 3. Controlling for age, sex, education, marital status, self-rated health, prior hospitalization, nursing home stay, and advance care discussion, Model 1 showed the odds of having an advance directive were two times as high for Whites (OR = 1.96, p < .001). Older age, higher education, widowhood, prior hospitalization, nursing home stay, and advance care discussion were each significantly associated with a higher likelihood of having an advance directive.

Table 3.

Logistic Regression Models Predicting the Odds of Advance Directive Completion.

	Model 1	Model 2	Model 3
Variable	OR [95% CI]	OR [95% CI]	OR [95% CI]
White (ref = Black)	1.96* [1.70, 2.25]	1.53* [1.34, 1.75]	0.96 [0.82, 1.12]
Female	1.07 [0.93, 1.24]	1.09 [0.94, 1.26]	1.02 [0.88, 1.18]
Age (years)	1.07* [1.06, 1.08]	1.06* [1.05, 1.07]	1.04* [1.03, 1.05]
Education	1.35* [1.27, 1.43]	1.22* [1.15, 1.30]	1.15* [1.08, 1.22]
Marital status (ref = married)
Divorced	0.95 [0.79, 1.13]	1.14 [0.93, 1.39]	1.18 [0.95, 1.46]
Widowed	1.32** [1.14, 1.52]	1.43* [1.21, 1.69]	1.40** [1.17, 1.67]
Never married	0.89 [0.62, 1.27]	1.01 [0.69, 1.46]	1.14 [0.76, 1.72]
Self-rated health	1.06 [1.01, 1.12]	1.00 [0.95, 1.06]	0.98 [0.92, 1.04]
Nursing home stay in past 2 years	2.23* [1.69, 2.94]	2.24* [1.69, 2.96]	2.29* [1.72, 3.06]
Hospitalized in past 2 years	1.23** [1.10, 2.94]	1.27** [1.13, 1.42]	1.23** [1.11, 1.37]
Advance care discussion	6.03* [5.41, 6.72]	6.02* [5.40, 6.70]	5.79* [5.17, 6.50]
Household net assets^a		1.29* [1.23, 1.35]	1.15* [1.10, 1.21]
Household income^b		0.94 [0.85, 1.04]	0.90 [0.80, 1.01]
Home ownership		0.63* [0.52, .75]	0.48* [0.39, .59]
Will and/or trust			7.04* [6.16, 8.04]
R ² (McKelvey and Zavoina)	.32	.35	.45

Note. n = 6,946. OR = odds ratio; CI = confidence interval.

^aNet household assets averaged across 2006, 2008, 2010, and 2012, log-transformed. ^bTotal household income averaged across 2006, 2008, 2010, and 2012, log-transformed.

*p < .05. **p < .01. ***p < .001.

The three financial variables of net assets, income, and home ownership were added as predictor variables in Model 2. The odds of having an advance directive for White participants were 1.5 times the odds for Black participants (OR = 1.53, p < .001). There was little change in the effects of the covariates. Household net assets were associated with a higher likelihood of having an advance directive (OR = 1.29, p < .001), whereas there was a negative relationship between home ownership and advance directive completion (OR = 0.63, p < .001). The effect of income was not significant.

When having a will or trust was added to Model 3, the effect of race became nonsignificant and the odds of having an advance directive were approximately the same for Whites and Blacks (OR = .96). Having executed a will and/or trust increased the odds of having an advance directive sevenfold (OR = 7.04, p < .001). Age, education, widowhood, nursing home stay, prior hospitalization, and net assets remained positively associated with advance directive completion and home ownership continued to exhibit a negative effect.

Advance care discussion

Results from the logistic regressions predicting advance care discussion are presented in Table 4. Model 1 estimated that Whites were almost twice as likely to have discussed medical care preferences after controlling for age, gender, education, marital status, self-rated health, hospitalization, nursing home stay, and advance directive completion (OR = 1.83, p < .001). Female gender, higher education, hospitalization, and advance directive completion were each significantly associated with higher odds of advance care discussion. Age displayed a negative relationship but the effect was small.

Table 4.

Logistic Regression Models Predicting the Odds of Advance Care Discussion.

	Model 1	Model 2	Model 3
Variable	OR [95% CI]	OR [95% CI]	OR [95% CI]
White (ref = Black)	1.83*** [1.56, 2.14]	1.78*** [1.53, 2.08]	1.75*** [1.50, 2.03]
Female	1.66*** [1.44, 1.90]	1.71*** [1.49, 1.97]	1.71*** [1.49, 1.96]
Age (years)	0.98*** [0.97, 0.99]	0.98*** [0.97, 0.99]	0.98*** [0.97, 0.99]
Education	1.34*** [1.27, 1.41]	1.29*** [1.21, 1.37]	1.28*** [1.20, 1.37]
Marital status (ref = married)
Divorced	1.06 [0.89, 1.26]	1.15 [0.95, 1.38]	1.15 [0.96, 1.38]
Widowed	1.11 [0.95, 1.30]	1.19 [1.00, 1.41]	1.19 [1.01, 1.42]
Never married	.85 [0.61, 1.18]	0.94 [0.68, 1.30]	0.95 [0.68, 1.31]
Self-rated health	.98 [0.93, 1.04]	0.97 [0.91, 1.03]	0.97 [0.91, 1.03]
Nursing home stay in past 2 years	.90 [0.68, 1.19]	0.89 [0.67, 1.18]	0.89 [0.67, 1.18]
Hospitalized in past 2 years	1.19* [1.03, 1.38]	1.19* [1.03, 1.38]	1.19* [1.03, 1.37]
Advance directive	6.04*** [5.41, 6.73]	6.01*** [5.39, 6.69]	5.81*** [5.17, 6.53]
Household net assets^a		0.99 [0.95, 1.02]	0.98 [0.95, 1.02]
Household income^b		1.22** [1.10, 1.35]	1.21** [1.09, 1.35]
Home ownership		0.89 [0.77, 1.04]	0.89 [0.76, 1.03]
Will and/or trust			1.10 [0.96, 1.27]
R ² (McKelvey and Zavoina)	.25	.26	.26

Note. n = 6,946. OR = odds ratio; CI = confidence interval.

^aNet household assets averaged across 2006, 2008, 2010, and 2012, log-transformed. ^bTotal household income averaged across 2006, 2008, 2010, and 2012, log-transformed.

*p < .05. **p < .01. ***p < .001.

Adding the three financial variables to Model 2 did not substantially change the race disparity in the odds of engaging in advance care discussion (OR = 1.78, p < .001). The ORs for the covariates remained stable as well. Income was positively associated with advance care discussion (OR = 1.22, p < .001), while the effects of the two other financial variables were not statistically significant. When having a will or trust was added to Model 3, the likelihood of advance care discussion was virtually unchanged (OR = 1.75, p < .001), as were the effects of the other covariates. The relationship between estate planning and advance care discussion was not statistically significant.

Discussion

ACP is associated with important quality of death indicators including less intensive medical care at the end of life, increased hospice enrollment and longer periods of hospice care, reduced chances of dying in a hospital, and psychological benefits for patients and family members. There remain disparities in ACP and quality of care at the end of life across diverse race and marginalized populations. This is the first study to empirically test the possibility that accumulative financial disadvantages may contribute to differences in estate planning by White and Black older adults, which may in turn account for race disparities in ACP.

Before including financial or ACP covariates in the model, White participants were almost six times as likely to have engaged in estate planning as Black participants. Home ownership, household net assets, and household income were each associated with a higher likelihood of having a will or trust. However, despite controlling for financial as well as demographic and health variables, older Whites were still more than four times as likely as older Blacks to have engaged in estate planning. Although controlling for advance directive completion and advance care discussion somewhat reduced these differences, White older adults remained 3.7 times more likely to have a will or trust. These results indicate that financial inequities only partially explain race disparities in estate planning. Furthermore, although completion of an advance directive was highly predictive of having a will or trust, differences in advance directive completion only accounted for a small portion of race disparities in estate planning.

Consistent with other studies, findings from this investigation showed that Whites were twice as likely as Blacks to have completed advance directives even after controlling for demographics, health status, health care utilization, and advance care discussion. This gap narrowed after including income, assets, and home ownership as covariates in the model, but White older adults were still 1.5 times as likely to have advance directives. When having a will or trust was added to the final model, the odds of having an advance directive became equal for White and Black older adults. Prior research has shown that the lower likelihood of advance directive completion by individuals with little or no assets is largely attributable to lower rates of estate planning, an activity that frequently triggers health planning (Carr, 2012). Our results suggest that the lower odds of advance directive completion by Black older adults are likewise largely attributable to their lower rates of estate planning.

Also consistent with prior studies, White participants were almost twice as likely as Black participants to have discussed their medical care preferences after adjusting for demographics, health status, health care utilization, and advance directive completion. The disparity was not substantially reduced by the addition of financial predictors. Having a will or trust was not significantly related to advance care discussion. In contrast to the findings regarding advance directive completion, estate planning did not explain persistent disparities in advance care discussion between older Whites and Blacks.

Household wealth appears to be positively associated with advance directive completion both directly and indirectly through estate planning. Those with higher household incomes were more likely to have completed wills or trusts and were more likely to have engaged in advance care discussion, but there was no direct relationship between income and advance directive completion. Home ownership displayed somewhat puzzling effects, being positively associated with estate planning but negatively associated with advance directive completion. One possible explanation for these mixed results is that some older adults may not own their own homes because they have downsized or moved into more supportive housing. Declines in physical or cognitive abilities may trigger such moves and also motivate older adults to complete advance directives. Home ownership may also be less common in urban areas or in different regions of the country, and these geographic differences may also play a role in ACP.

There are a few study limitations that should be acknowledged. First, because data are self-reported, some participants may have been mistaken about whether or not they had estate and/or health planning documents due to confusion about the nature of and differences among these and other legal forms. Second, there are no data regarding when advance directives were completed or when advance care discussions took place. This lack of chronological information poses less of an issue for more constant predictors such as race or education but is a significant limitation in the interpretation of the effects of event-based variables such as hospitalization or estate planning. We can observe associations among the variables but cannot reach conclusions about causation or mediation. Third, the HRS core interview does not ask about life-sustaining treatment preferences, so it was not possible to control for the potential effects of these preferences on ACP.

The study likewise has a number of strengths. A major limitation of most research on race differences in ACP is that it is often based on small, nonrepresentative samples (Kwak & Haley, 2005; Ladd, 2014). Furthermore, the relationships among ACP, financial resources, estate planning, and race are not well understood and have not been empirically examined. This is the first study to focus on the intersection of race, estate planning, and ACP using a large, nationally representative data set. Our consideration of advance care discussion as well as advance directive completion contributes to the limited research on informal ACP.

Conclusion

Historically, society has been plagued with race differences and inequities that continue to place many Black people at a disadvantage, especially older adults who grew up during a time when “every penny counted” and saving for long-term care or retirement was not an option. Lower generational inheritance, higher unemployment, and differing rates of homeownership and marriage (Pew Research Center, 2013) have left Black older adults financially disadvantaged. This, of course, is no fault of the individual, but rather a systemic issue that extends across multiple generations. Older Blacks were not afforded the same opportunities to accumulate wealth over their lifetimes. As a result, White older adults are in a better position to access and benefit from estate planning which may, in turn, trigger advance directive completion and its protective effects of higher quality end-of-life care.

This study illustrates one of many pathways by which economic inequities may shape health behavior and manifest as disparate medical care at the end of life. Financial disparities contribute to lower rates of estate planning. Estate planning in turn explains in large part why Black older adults are less likely to have advance directives. Advance directives are associated with better quality end-of-life care. One strategy to address at a more systemic level race disparities in advance directive completion and, in so doing, improve end-of-life care for older Blacks would be to reduce financial inequality in older age.

Distinct mechanisms appear to be responsible for race disparities in advance directive completion and advance care discussion. Differing rates of estate planning explain disparities in advance directive completion by White and Black older adults. In contrast, disparities in advance care discussion are not directly accounted for by estate planning. The factors influencing advance care discussion are less frequently studied and even less well understood than advance directive completion. Persistent race disparities in end-of-life discussion are somewhat surprising considering the strong oral traditions among many Black communities (Banks-Wallace, 2002). More research is needed to identify the specific underlying causes of lower rates of advance care discussion among Black older adults.

Although estate planning accounted for the disparity in advance directive completion between White and Black participants, we still do not fully understand why Black older adults are less likely to engage in estate planning. Home ownership, wealth, and income are likely contributing factors, as demonstrated in this study. However, substantial disparities in estate planning remained even after controlling for these financial variables. Other possible factors worth exploring include access to legal services, prior experience with attorneys or the legal system, end-of-life planning by parents or other family members who may serve as role models, reliance on informal or non-testamentary methods of transferring assets such as pre-death gifting, and degree of willingness to contemplate or discuss death by older adults and/or family members.

This study has practical implications for efforts to encourage older Blacks to complete advance directives. To date, many of the interventions to promote ACP have targeted patients in health care settings and focused exclusively on medical decision making. Meanwhile, those with financial means—particularly those who also happen to be White—are encouraged and assisted by attorneys to complete advance directives as part of their estate planning. Making comprehensive financial and health planning more easily available may be an effective means for advocates of ACP to reach Black and other underrepresented groups. These types of outreach efforts would ideally involve health care providers and attorneys working in partnership to provide information and assistance in both medical and nonmedical settings.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded by the Borchard Foundation Center on Law and Aging.

References

Abbott

K. H.

Sago

J. G.

Breen

C. M.

Abernethy

A. P.

Tulsky

J. A.

(2001). Families looking back: One year after discussion of withdrawal or withholding of life-sustaining support. Critical Care Medicine, 29, 197–201. doi:10.1097/00003246-200101000-00040

Administration on Aging. (2014). A profile of older Americans: 2014. Retrieved from http://www.aoa.gov/Aging_Statistics/Profile/index.aspx

Banks-Wallace

(2002). Talk that talk: Storytelling and analysis rooted in African American oral tradition. Qualitative Health Research, 12, 410–426. doi:10.1177/104973202129119892

Bischoff

K. E.

Sudore

Miao

Boscardin

W. J.

Smith

A. K.

(2013). Advance care planning and the quality of end-of-life care in older adults. Journal of the American Geriatrics Society, 61, 209–214. doi:10.1111/jgs.12105

Brinkman-Stoppelenburg

Rietjens

J. A. C.

van der Heide

(2014). The effects of advance care planning on end-of-life care: A systemic review. Palliative Medicine, 28, 1000–1025. doi:10.1177/0269216314526272

Bullock

McGraw

S. A.

Blank

Bradley

E. H.

(2005). What matters to older African Americans facing end-of-life decisions? A focus group study. Journal of Social Work in End-of-Life & Palliative Care, 1, 3–19. doi:10.1300/j457v01n03_02

Carr

(2011). Racial differences in end-of-life planning: Why don’t Blacks and Latinos prepare for the inevitable? OMEGA-Journal of Death and Dying, 63, 1–20. doi:10.2190/om.63.1.a

Carr

(2012). The social stratification of older adults’ preparations for end-of-life health care. Journal of Health and Social Behavior, 53, 297–312. doi:10.1177/0022146512455427

Chien

Campbell

Hayden

Hurd

Main

Mallett

… St.Clair

(2014). RAND HRS data documentation: Version N. RAND Center for the Study of Aging. Retrieved from http://www.rand.org/labor/aging/dataprod/hrs-data.html

10.

Crawley

Payne

Bolden

Payne

Washington

Williams

(2000). Palliative and end-of-life care in the African American community. JAMA, 284, 2518–2521. Retrieved from http://doi.org/10.1001/jama.284.19.2518

11.

Enders

C. K.

(2010). Applied missing data analysis. New York, NY: Guildford Press.

12.

Gerst

Burr

J. A.

(2008). Planning for end-of-life care: Black-White differences in the completion of advance directives. Research on Aging, 30, 428–449. doi:10.1177/0164027508316618

13.

Goetting

Martin

(2001). Characteristics of older adults with written wills. Journal of Family and Economic Issues, 22, 243–264. doi:10.1023/A:1016699822855

14.

Greiner

K. A.

Perera

Ahluwalia

J. S.

(2003). Hospice usage by minorities in the last year of life: Results from the National Mortality Followback Survey. Journal of the American Geriatrics Society, 51, 970–978. doi:10.1046/j.1365-2389.2003.51310.x

15.

Huang

I. A.

Neuhaus

J. M.

Chiong

(2016). Racial and ethnic differences in advance directive possession: Role of demographic factors, religious affiliation, and personal health values in a national survey of older adults. Journal of Palliative Medicine, 19, 149–156. doi:10.1089/jpm.2015.0326

16.

Institute of Medicine. (2015). Dying in America: Improving quality and honoring individual preferences near the end of life. Washington, DC: Author.

17.

Johnson

K. S.

Kuchibhatla

Tulsky

J. A.

(2008). What explains racial differences in the use of advance directives and attitudes toward hospice care? Journal of the American Geriatrics Society, 56, 1953–1958. doi:10.1111/j.1532-5415.2008.01919.x

18.

Kelly

Masters

Deviney

(2013). End-of-life planning activities: An integrated process. Death Studies, 37, 529–551. doi:10.1080/07481187.2011.653081

19.

Kessler

D. P.

McClellan

M. B.

(2004). Advance directives and medical treatment at the end of life. Journal of Health Economics, 23, 111–127. doi:10.1016/j.jhealeco.2003.08.006

20.

Khandelwal

Kross

E. K.

Engelberg

R. A.

Coe

N. B.

Long

A. C.

Curtis

J. R.

(2015). Estimating the effect of palliative care interventions and advance care planning on ICU utilization: A systematic review. Critical Care Medicine, 43, 1102–1111. doi:10.1097/CCM.0000000000000852

21.

Koss

C. S.

Baker

T. A.

(2016). Race differences in advance directive completion: The narrowing gap between White and African American older adults. Journal of Aging and Health, 1–19. doi:10.1177/0898264316635568

22.

Krivo

L. J.

Kaufman

R. L.

(2004). Housing and wealth inequality: Racial-ethnic differences in home equity in the United States. Demography, 41, 585–605. doi:10.1353/dem.2004.0023

23.

Kwak

Haley

W. E.

(2005). Current research findings on end-of-life decision making among racially or ethnically diverse groups. The Gerontologist, 45, 634–641. doi:10.1093/geront/45.5.634

24.

Ladd

C. S.

(2014). Systematic review of research literature on African Americans’ end-of-life healthcare preferences. Journal of African American Studies, 18, 373–397. doi:10.1007/s12111-013-9276-z

25.

O’Rand

A. M.

(1996). The precious and the precocious: Understanding cumulative disadvantage and cumulative advantage over the life course. The Gerontologist, 36, 230–238.

26.

Patient Self-Determination Act of 1990. 1990. 554206, 4751 of the Omnibus Reconciliation Act of 1990. Pubs. No. 101–508.

27.

Pearlman

R. A.

Cole

W. G.

Patrick

D. L.

Starks

H. E.

Cain

K. C.

(1995). Advance care planning: Eliciting patient preferences for life-sustaining treatment. Patient Education and Counseling, 26, 353–361. doi:10.1016/0738-3991(95)00739-m

28.

Pew Research Center. (2013). Black incomes are up, But wealth isn’t. Washington, DC. Retrieved from http://www.pewresearch.org/fact-tank/2013/08/30/black-incomes-are-up-but-wealth-isnt/

29.

Pew Research Center. (2009). End-of-life decisions: How Americans cope. Washington, DC. Retrieved from http://www.pewsocialtrends.org/2009/08/20/end-of-life-decisions-how-americans-cope/

30.

Phelan

J. C.

Link

B. G.

Tehranifar

(2010). Social conditions as fundamental causes of health inequalities: Theory, evidence, and policy implications. Journal of Health and Social Behavior, 51, S28–S40. doi:10.1177/0022146510383498

31.

Pollack

K. M.

Williams

M. A.

(2010). The public’s perspectives on advance directives: Implications for state legislative and regulatory policy. Health Policy, 96, 57–63. doi:10.1016/j.healthpol.2010.01.004

32.

Rao

J. K.

Anderson

L. A.

Lin

F.-C.

Laux

J. P.

(2014). Completion of advance directives among U.S. consumers. American Journal of Preventive Medicine, 46, 65–70. doi:10.1016/j.amepre.2013.09.008

33.

Sabatino

C. P.

(2015). Summary of center for Medicare and Medicaid services (CMS) rule for covering cost of advance care planning services. Bifocal, 37, 19–21. Retrieved from http://www.americanbar.org/publications/bifocal/vol_37.html

34.

Sanders

J. J.

Robinson

M. T.

Block

S. D.

(2016). Factors impacting advance care planning among African Americans: Results of a systematic integrated review. Journal of Palliative Medicine, 19, 202–227. doi:10.1089/jpm.2015.0325

35.

Silveira

M. J.

Kim

S. Y. H.

Langa

K. M.

(2010). Advance directives and outcomes of surrogate decision making before death. The New England Journal of Medicine, 362, 1211–1218. doi:10.1056/NEJMsa0907901

36.

Sonnega

Faul

Ofstedal

Langa

Phillips

Weir

(2014). Cohort profile: The Health and Retirement Study (HRS). International Journal of Epidemiology, 43, 576–585. doi: 10.1093/ije/dyu067

37.

Y.-T.

(2008). Looking beyond retirement: Patterns and predictors of formal end-of-life planning among retirement age individuals. Journal of Family and Economic Issues, 29, 654–673. doi:10.1007/s10834-008-9130-y

38.

Sudore

R. L.

Schickedanz

A. D.

Landefeld

C. S.

Williams

B. A.

Lindquist

Pantilat

S. Z.

Schillinger

(2008). Engagement in multiple steps of the advance care planning process: A descriptive study of diverse older adults. Journal of the American Geriatrics Society, 56, 1006–1013. doi:10.1111/j.1532-5415.2008.01701.x

39.

Teno

J. M.

Gruneir

Schwartz

Nanda

Wetle

(2007). Association between advance directives and quality of end-of-life care: A national study. Journal of the American Geriatrics Society, 55, 189–194. doi:10.1111/j.1532-5415.2007.01045.x

40.

Tilden

V. P.

Tolle

S. W.

Nelson

C. A.

Fields

(2001). Family decision-making to withdraw life-sustaining treatments from hospitalized patients. Nursing Research, 50, 105–115. doi:10.1097/00006199-200103000-00006

41.

Tschirhart

E. C.

Kelley

A. S.

(2014). Factors influencing the use of intensive procedures at the end of life. Journal of the American Geriatrics Society, 62(11), 2088–2094. doi: 10.1111/jgs.13104

42.

U.S. Government Accountability Office. (2015). Advance directives: Information on federal oversight, provider implementation, and prevalence (No. GAO-15-416) (pp. 1–28). United States Government Accountability Office. Retrieved from http://gao.gov/products/GAO-15-416

43.

Wendler

Rid

(2011). Systematic review: The effect on surrogates of making treatment decisions for others. Annals of Internal Medicine, 154, 336–346. doi:10.7326/0003-4819-154-5-201103010-00008

44.

Wright

A. A.

Zhang

Ray

Mack

J. W.

Trice

Balboni

… Prigerson

H. G

. (2008). Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA, 300, 1665–1673. doi:10.1001/jama.300.14.1665