Morrison RS, Meier DE. High Rates of Advance Care Planning in New York City’s Elderly Population. Arch Intern Med. 2004;164(22):2421-2426. doi:10.1001/archinte.164.22.2421
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
Previous studies have demonstrated low rates of advance care planning (ACP), particularly among nonwhite populations, raising questions about the generalizability of this decision-making process.
To explore factors that may influence patients’ willingness to engage in ACP.
Thirty-four randomly selected New York City senior centers.
A total of 700 African American (n = 239), Hispanic (n = 237), and white (n = 224) adults 60 years and older.
Participants were administered a 51-item survey that assessed attitudes, beliefs, and practices regarding ACP.
Main Outcome Measures
Attitudes and beliefs about physicians’ trustworthiness, fatalism, beliefs about surrogate decision making, and comfort discussing end-of-life medical care; factors associated with health care proxy completion; and health care proxy completion rates.
More than one third of the participants had completed a health care proxy. There were no significant differences in completion rates across the 3 ethnic groups. Respondents who had a primary care physician (odds ratio [OR], 2.0; 95% confidence interval [CI], 1.3-3.2), were more knowledgeable about advance directives (OR, 2.0; 95% CI, 1.4-2.9), or had seen a friend or family member use a mechanical ventilator (OR, 1.5; 95% CI, 1.02-2.1) were significantly more likely to have designated a health care proxy. Respondents who were only comfortable discussing ACP if the discussion was initiated by the physician (OR, 0.6; 95% CI, 0.0-0.8) were significantly less likely to have completed a health care proxy.
African American, Hispanic, and white community-dwelling, older adults had similarly high rates of advance directive completion. The primary predictors of advance directive completion involved modifiable factors such as established primary care physicians, personal experience with mechanical ventilation, knowledge about the process of ACP, and physicians’ willingness to effectively initiate such discussions. Some of the racial/ethnic differences in desire for collective family-based decision making that were observed in this study have implications for the evolution of ACP policy that respects and operationalizes these preferences.
Advance care planning (ACP) is the process of planning for future medical care under circumstances of impaired decisional capacity. Although sometimes equated with the completion of a living will or treatment directive or health care proxy or durable power of attorney, ACP is ideally a process that involves a structured discussion in which patients and physicians explore goals of care in the context of current and hypothetical illness states, discuss treatment options in the context of these goals of care, and finally articulate and document treatment and care preferences.1 Acceptance of ACP among the public, particularly among older adults and other adults living with serious and chronic illness, appears to be increasing.2- 5
Many studies have found that African American patients are less likely to complete advance directives (specifically living wills), are less likely to have discussed preferences regarding cardiopulmonary resuscitation with their physicians, and are more likely to feel that they would be treated differently or cared for less if they had a living will compared with white patients.2,4- 8 Although some have interpreted these results as indicating a lack of trust by minority patients in ACP, living wills, and the health care system in general, the reasons behind the observed ethnic differences in advance directive completion rates have not been well defined by empirical research.5,9,10
Advance care planning is based on several underlying premises. First, it requires a belief that planning for the future will effectively influence future events. Second, it requires a willingness to engage in discussions that could be perceived as uncomfortable and difficult (eg, imagining a situation in which one is seriously ill). Third, particularly with respect to the designation of a health care proxy, it requires one to assign decision making to 1 or at most 2 individuals and presumes that these individuals will speak for the incapacitated person. Finally, it requires trust in both physicians and the health care system that the wishes expressed by the patient will be honored and treatments will not be withdrawn prematurely.
In a previous study5 of patients followed up in a geriatrics group practice, we reported significant differences among white, African American, and Hispanic patients with respect to their comfort discussing advance directives, trust in physicians, beliefs about the burdens placed on family members by proxy designation, and knowledge of advance directives. This study was limited by the fact that it was performed in one medical setting in which all patients had access to a single primary care physician and in a setting in which physicians were strongly encouraged to engage in ACP discussions. The present study builds on this previous work by exploring factors that may influence a willingness to engage in ACP in a large sample of community-dwelling, older white, African American, and Hispanic adults.
A total of 700 people 60 years and older who were self-identified as African American, Hispanic, or white were recruited from Manhattan (New York City) senior centers. Senior centers were selected from a list obtained from the New York City Department of Aging. Of the 90 potential senior centers, 13 centers refused to participate on initial contact and an additional 10 centers were eliminated because most of their clients were not included in the 3 ethnic groups being studied. From the remaining 67 centers, we randomly selected 34 (51%) for inclusion. The directors of these 34 centers were again contacted and a specific recruitment date was established. The directors also agreed to publicize the study through verbal announcements to the centers’ clients. Two research assistants, trained by one of us (R.S.M.), visited each of the 34 centers on the specified recruitment date. People attending the senior centers on the specified recruitment date were approached for informed consent and administered a 51-item questionnaire. The research assistants were bilingual in English and Spanish. Completion of the questionnaire typically required 25 to 30 minutes.
We designed a questionnaire to examine the relationship between a series of hypothesized barriers to completion of health care proxies and participants’ actual appointment of a surrogate decision maker. The content of the questionnaire was based on previous work that has been described in detail elsewhere.5,9 In brief, the final instrument consisted of 51 closed-ended questions. Participants were initially asked whether they were familiar with the term health care proxy. Participants who responded affirmatively were then asked 6 additional questions based on the New York State Proxy Law. Following this section, a comprehensive explanation of health care proxies was provided to all participants. Additional components of the questionnaire included patient demographics and characteristics, experience with the health care system (including exposure to life-prolonging technology such as intensive care units and mechanical ventilation), access to health care, knowledge of health care proxies, beliefs about trustworthiness of physicians, fatalism, burdens on proxy decision makers, relevance of formally appointed decision makers, comfort discussing end-of-life medical care, whether the respondent had ever formally appointed a surrogate decision maker (health care proxy), and if the respondent had not formally appointed a surrogate decision maker, whether he/she intended to do so.
The questionnaire was written in simple language at a sixth grade level, forward-translated into Spanish, and then back-translated into English to confirm the accuracy of initial translation. Questionnaires were administered in the participant’s preferred language.
Analysis of variance and χ2 procedures were used to examine differences in the independent variables (Table 1) across the 3 ethnic groups. Multiple logistic regression analyses were performed to assess differences regarding attitudes and beliefs about the trustworthiness of physicians, fatalism, beliefs about surrogate decision making, and comfort discussing end-of-life medical care across the 3 ethnic groups. Covariates entered into the regression models included age, sex, education, self-perceived health status, Medicaid insurance, hospitalization in the past year, prior history of mechanical ventilation, prior history of intensive care, prior history of seeing a friend or family member receive mechanical ventilation, and prior history of seeing friend or family member in an intensive care unit.
We used multiple logistic regression to examine the relationship between ethnicity and completion of a health care proxy. To develop this model, given the large number of possible covariates, we first performed a series of univariate analyses, including each variable of interest (Table 1 and Table 2), with completion of a health care proxy as the dependent variable. Although the questionnaire included a question about income level, 47% of participants declined to answer this question. As a surrogate for income status, enrollment in the state’s Medicaid insurance program was used. All variables with P≤.15 in the single-variable analysis were examined jointly in the next step of model building. Variables that were no longer of borderline significance (P>.10) when the other variables were entered were eliminated from the model. After controlling for variables retained in the model, variables not previously significant in the single-variable analysis were added to the model and those of borderline significance (P≤.10) were retained. In addition, variables that had been found in prior studies to be associated with advance directive completion (eg, ethnicity, health status, income) were retained in the model. The study protocol was approved by the Institutional Review Board of the Mount Sinai School of Medicine.
Seven hundred eighty eligible people were approached for consent to participate, of whom 700 (90%) agreed to be interviewed. Participant characteristics are given in Table 1. Two hundred thirty-nine participants were white (response rate of 91%), 237 were African American (response rate of 91%), and 224 were Hispanic (response rate of 86%). Of Hispanic participants, 80% were born in Puerto Rico or the Dominican Republic and had lived in the United States for a median of 38.5 years (range, 1-82 years). White participants were significantly older, more educated, and less likely to have Medicaid. African American and white participants rated their health significantly better than Hispanic participants. Overall, 35% of the sample had completed a health care proxy, and there were no significant differences in proxy completion across the 3 ethnic groups.
Results from the multivariate analyses regarding attitudes and preferences regarding ACP are given in Table 2. More than 80% of respondents in all 3 racial/ethnic categories reported that they felt comfortable talking to their physician about ACP and that they would like to have a conversation about treatments at the end of life. However, compared with white participants, Hispanic respondents were more likely to want to have these discussions initiated by their physician (odds ratio [OR], 1.9; 95% confidence interval [CI], 1.2-3.1) or family member (OR, 1.6; 95% CI, 1.0-2.4) rather than raising the topic themselves. Hispanic participants were significantly more likely to believe that it was the family’s role to make decisions about life-sustaining treatments (OR, 2.1; 95% CI, 1.2-3.7) even in the presence of a designated health care proxy (OR, 2.2; 95% CI, 1.3-3.5) and to value collective rather than individual family decision making (OR, 2.5; 95% CI, 1.5-4.2) compared with white participants. Finally, Hispanic participants were significantly more likely to be worried that they would not be treated as well as other people (OR, 3.1; 95% CI, 1.9-4.9) and that physicians might prematurely discontinue life-sustaining treatments (OR, 2.4; 95% CI, 1.5-3.8). Responses by older African American participants were not significantly different than those of white respondents.
Table 3 lists factors associated with assignment of a health care proxy. Controlling for other variables, respondents who had a primary care physician (OR, 2.0; 95% CI, 1.3-3.2), were more knowledgeable about advance directives (OR, 2.0; 95% CI, 1.4-2.9), or had seen a friend or family member use a mechanical ventilator (OR, 1.5; 95% CI, 1.02-2.1) were significantly more likely to have designated a health care proxy. Respondents who were only comfortable discussing ACP if the discussion was initiated by the physician (OR, 0.6; 95% CI, 0.4-0.8) were significantly less likely to have completed a health care proxy. Race/ethnicity was not significantly associated with completion of an advance directive.
This study of 700 community-dwelling, older adults from 3 major racial/ethnic groups found significant differences in attitudes toward ACP for Caribbean Hispanic participants compared with African American and white participants. Despite these differences in attitudes, this study found no significant differences in health care proxy completion rates across the 3 ethnic groups studied. Factors that predicted health care proxy completion rate were related to access to a known primary care physician, access to a physician who would initiate an ACP discussion, knowledge of advance directives, and prior experience with mechanical ventilation. These findings enhance our understanding of how race/ethnicity relates to ACP and have important implications for end-of-life decision making.
A substantial portion (35%) of the sample indicated completion of an advance directive, a number that did not vary significantly by ethnicity. Prior studies2- 4 conducted in recent years have shown a growing increase in the prevalence of advance directives and have identified a lack of knowledge about advance directives as the predominant barrier to their completion. The results from prior studies2- 4 and the study reported herein suggest that awareness and employment of the ACP process is gradually increasing, at least in New York City. Thus, prior studies11,12 that point to limitations and low utilization of ACP may have been conducted too early in the introduction of this method to accurately assess its longer-term impact. The present study involved a large sample of older adults from randomly selected senior centers in New York City. Most prior work has focused on individual academic centers, where the investigators focus on ACP and are likely to have influenced the awareness and utilization of this method. The fact that this study recruited people from outside an academic center amplifies the importance and generalizability of its findings.
Advance care planning, as currently practiced and codified in the United States, is based on several underlying premises.13,14 It requires a belief that planning for the future will effectively influence future events; a willingness to engage in ACP discussions and to discuss emotionally laden decisions; a willingness to assign 1 person as the primary decision maker; and trust on the part of patients that their wishes will be carried out in accordance with their preferences (that desired treatments will not be withheld and undesired treatments will not be applied).13,14 Our results suggest that although white and African American people appear to hold to these premises, a substantial number of Hispanic people may not.
Previous studies regarding advance directives have focused predominantly on examining attitudes and beliefs of African American compared with white people. Relatively few studies5,9,15- 17 have focused on attitudes and beliefs of Hispanic people, and most of those have enrolled Mexican American rather than Caribbean Hispanic people. Our results suggest that Caribbean Hispanic people, like Mexican American Hispanic people, appear to prefer family-centered and collective decision making rather than individual and autonomous decision making.15- 17 Furthermore, our results also suggest that Hispanic people were less likely to believe that their actions could change the future and were more worried that life-sustaining treatments would be withdrawn prematurely compared with white and African American people. These data suggest that the current and almost sole focus on advance directives and ACP as a means of guiding medical decision making for patients without capacity may not serve the best interests of most Hispanic patients and that alternative models that include collective family decision making should be explored.
Our study also confirms and extends previous work relating to attitudes and beliefs of African American patients regarding ACP. Several studies5,8,18 have demonstrated lower prevalence rates of advance directives in African American patients and reported that African American patients are less likely to specify that life-sustaining treatments be withheld or withdrawn in their living wills compared with white patients. Other studies,4- 6 however, that have specifically explored willingness to engage in ACP, including designation of a health care proxy, however, have not reported differences in advance directive completion rates attributable to race when other factors (eg, knowledge of advance directives) have been controlled for. The present study enrolled more than 200 community-dwelling, African American adults and found that almost one third had completed a health care proxy and that health care proxy prevalence rates were no different than those observed in elderly white and Hispanic respondents. Furthermore, this study did not observe differences in trust, comfort with ACP, the surrogate decision making process, or fatalism compared with white adults.
This study also amplifies our understanding of barriers to ACP within the 3 racial/ethnic groups studied. Similar to our prior study, which enrolled patients in an ambulatory care geriatrics/internal medicine group practice, we found that race/ethnicity per se did not predict completion of an advance directive. Rather, this study identified specific and relevant attitudes and barriers to advance directives that can guide physicians in providing culturally competent health care. Specifically, our data suggest that the most important factors that predict completion of an advance directive include having a physician who is willing to initiate an ACP discussion, access to a known primary care physician, knowledge about advance directives, and experience with mechanical ventilation or having a friend or family member who has used mechanical ventilation. Interestingly, these factors appeared to be stronger predictors of a willingness to engage in ACP than the attitudinal differences described herein.
There are several limitations to this study. First, our study was undertaken in New York City, and the results may not be generalizable to other regions. Second, we recruited participants from senior centers, which may have resulted in selection bias, and thus our results may not be generalizable to the general population. Third, we relied on self-reports as our measure of health care proxy completion and did not confirm responses by examining the actual document. Fourth, although we attempted to eliminate selection bias by obtaining oral consent and interviewing people in their preferred language, it is possible that some participants may have felt uncomfortable or refused to participate because the interviewer was not of the same race/ethnicity. Finally, since ethnicity is not a unidimensional concept, caution is mandated in the interpretation of our results.
In summary, this study of African American, Hispanic, and white community-dwelling, older adults selected from a random sample of senior centers in New York City had a high rate of advance directive completion. No significant differences in completion rates were observed across the 3 ethnic groups. The primary predictors of advance directive completion involved modifiable factors, such as an established primary care physician, personal experience with mechanical ventilation, a basic fund of knowledge about the process of ACP, and physicians’ willingness to effectively initiate such discussions. Some racial/ethnic differences in desire for collective family-based decision making that were observed in this study have implications for the evolution of ACP policy that respects and operationalizes these preferences.
Correspondence: R. Sean Morrison, MD, Department of Geriatrics, Campus Box 1070, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029 (firstname.lastname@example.org).
Accepted for Publication: January 27, 2004.
Financial Disclosure: None.
Funding/Support: Dr Morrison is a Paul Beeson Faculty Scholar in Aging Research, an Open Society Institute Project on Death in America Faculty Scholar, and the recipient of a Mentored Clinical Scientist Development Award (K08AG00833-01) from the National Institute on Aging (Bethesda, Md) at the time this study was conducted. Dr Meier is an Open Society Institute Project on Death in America Faculty Scholar and the recipient of an Academic Career Leadership Award (K07AG00903) from the National Institute on Aging. This project was supported by the Nathan Cummings Foundation (New York).