Abstract
Context:
Advance Care Planning (ACP) is a process of understanding and sharing preferences regarding future medical care.
Objective:
To explore individual and national stability of end-of-life treatment preferences among a sample of older adults.
Methods:
National Health and Aging Trends Study (NHATS) is a nationally representative sample of older adults. In 2012, a random sample, and in 2018, the entire sample were queried on end-of-life treatment preferences defined as acceptance or rejection of life prolonging treatment (LPT) if they had a serious illness and were at the end of their life and in severe pain or had severe disability. Using a cohort design, we explored individual trends in preferences for LPT among those with responses in both waves (pain scenario: N=606, disability scenario: N=628) and, using a serial cross-sectional design, national trends in LPT among the entire sample (1,702 older adults in wave 2 and 4,342 in wave 8).
Results:
In the cohort study, individual preferences were stable over time (overall percent agreement = 86% for disability and 76% for pain scenarios), particularly for older adults who would reject LPT in wave 2 (overall agreement 92% for disability and 86% for pain). In the serial cross-sectional study, national trends in preferences for receipt of LPT were stable over time in the pain (27.4% vs. 27.0%, p=0.80) and disability (15.8% vs. 15.7%, p=0.99) scenarios.
Conclusions:
We found that national trends in preferences for end-of-life treatment did not substantially change over time and may be stable within individual older adults.
Keywords: advance care planning, older adults, end-of-life preferences
Introduction
The goal of ACP is to ensure that people receive care consistent with their values and preferences throughout illness and at the end of life.(1) ACP rests on the notion that patients’ preferences when engaging in ACP will reflect their preferences when treatment decisions must be made.
Much of the work exploring the stability of end-of-life preferences focused on older adults with advanced disease,(2, 3) at the end-of-life,(4) who are hospitalized,(5, 6) or are physicians.(7) Overall, studies of the stability of preferences have been limited by varied study populations, timing to preference assessments, and definitions of stability.(8) For example, in the same cohort of older adults with advanced disease, the stability of treatment preferences differed based on how preferences were assessed.(2, 3) Stability of preferences among the general population of older adults may differ from those with advanced disease because of their health status or possible decreased knowledge of ACP.(9)
While these data have focused on stability of individual’s preference, data from national samples on the stability of end-of-life preferences over time are lacking.(2, 3, 10) It may be that national sociopolitical events impact end-of-life preferences for the population or a subset of the population such as specific race/ethnicity groups. For example, the use of the phrase “death panels” that surrounded the Affordable Care Act in 200 (11) with lasting impact such that in 2014 40% of Americans still believe the Affordable Care Act created death panels (12) may influence national end-of-life treatment preferences. Or possibly the constellation of factors that have led to the decline in Americans’ trust in government, the medical system, and each other over the last 20 years.(13, 14) Thus, in a national sample of older adults, we sought to explore the stability of individual and national end-of-life treatment preferences from 2012 to 2018.
Methods
Data Source and Study Population
The National Health and Aging Trends Study (NHATS) selected a nationally representative sample of Medicare aged 65+ beneficiaries in 2011, replenishing the sample in 2015. In-person interviews are conducted annually. In the 2nd NHATS wave, 2012, a random 1/3 of the NHATS sample received a module inquiring about end-of-life plans and care (N=2000), and this module was repeated in the 8th wave, 2018, across the whole NHATS sample (N=5547). Within the module, two separate hypothetical scenarios are posed to older adults who did not require a proxy — whether older adults would want to “receive” or “stop/reject” life-prolonging treatment (LPT) if they were at the end of their life and had a serious illness and either they were in “constant, severe physical pain” with preserved cognition and function (pain scenario) or they could not “speak, walk or recognize others” but were not in pain (disability scenario).
We report the results of two separate, but related studies using two study populations and two separate study designs. First, to analyze individual LPT preferences, we limited the sample to older adults who responded to wave 2 and wave 8 (pain scenario: N=606, disability scenario: N=628) in a cohort study. Notably, of the wave 2 population, 25% died, and 31% were lost to follow-up before wave 8. Second, to analyze national trends in LPT preference, we expanded the study population to include older adults who responded to either wave 2 or wave 8 (1,702 older adults in wave 2 (85.1% response rate) and 4,342 older adults in wave 8 (78.3% response rate) in a serial cross-sectional study. This study was approved by the University of Michigan Institutional Review Board.
Primary Outcome
Preference for LPT was determined if the participants responded that they would want to “receive” LPT or they would “stop/reject” LPT for either pain and disability scenarios.
Covariates
Race/ethnicity was defined as self-reported Non-Hispanic White (White), Non-Hispanic Black (Black), Hispanic, Other, and missing and was included due to race differences in ACP.(15) Participant demographics (age [measured using linked Medicare Beneficiary Summary File as a continuous measure], sex, marital status [married/living with a partner vs. others], and education), religiosity, depression, and anxiety. We also included variables to capture change in function consisting of physical functioning (limitations in activities of daily living [ADLs]: self-care and mobility), cognitive functioning (probable/possible dementia vs. no dementia), pain limits activities (for pain scenario only), and exposure to illness (self-reported number of hospital stays in last 12 months) were considered as other covariates for the study.
Statistical Analysis
We conducted descriptive analyses to characterize the study population and compared participants who responded to the LPT pain and disability scenarios in wave 2 and wave 8. Then, in the cohort study, for respondents who answered both wave 2 and wave 8 survey questions, we calculated individual trends in LPT preferences for the pain and disability scenarios. Respondents in both waves were more likely to be White, to have dementia, and reject LPT in the pain scenario. Thus, data were not missing at random, and standard analytic strategies to account for missingness (i.e. multiple imputation) were not valid. Therefore, we did not conduct further analysis to explore predictors of changes in individual LPT preferences.
In the serial cross-sectional study using the entire wave 2/wave 8 sample, we then explored national trends in end-of-life treatment preferences for each scenario. In addition, we examined predictors of receipt of LPT preferences using multiple logistic regression (time only, race and time and interaction, and all covariates) models, applying round-specific analytic weights and clustering within respondents to account for respondents appearing in both waves. We included a race-time interaction to assess whether sociopolitical trends may influence LPT preferences by race.
Results
There were 1702 round 2 and 4342 round 8 older adults who responded to at least one LPT scenario (either disability or pain scenario). For the disability scenario, 1645 round 2 and 4163 round 8 older adults responded (Table 1). For the pain scenario, 1613 round 2 and 4060 round 8 older adults responded (supplemental Table 1). Compared with respondents in wave 2, those in wave 8 were more likely to be older (76.6 vs. 75.6, p<.01), less likely to be White (78.6% vs. 81.3%, p<0.05), complete post-high school education (58.9% vs. 49.7%, p<.01, pain scenario), less likely to have anxiety (8.1% vs. 11%, p<0.05), depression (8.8% vs. 11.6%, p<0.01) and dementia (13.5% vs. 15.9%, p=0.02 disability scenario) in both pain and disability scenarios (Table 1 and supplemental Table 1). There were differences in wave 2 respondents who and did not respond in wave 8. We found that respondents in both waves were younger, more likely to be white, married, have higher education, have less dementia, depressive symptoms, anxiety symptoms and experience less disability and fewer hospital stays than respondents in wave 2 only (supplemental Table 2).
Table 1.
Demographic characteristics for older adult respondents of life prolonging treatment disability scenario, cross-sectional study
| Wave 2 (N=1645) | Wave 8 (N=4163) | χ2p-value | |||
|---|---|---|---|---|---|
|
| |||||
| N | Weighted % | N | Weighted % | ||
| Sex | 1.00 | ||||
| Female | 928 | 54.8 | 2375 | 54.8 | |
| Male | 717 | 45.2 | 1788 | 45.2 | |
| Race | 0.04 | ||||
| White | 1148 | 81.3 | 2969 | 79.1 | |
| Black | 351 | 8.1 | 812 | 7.6 | |
| Other | 42 | 3.0 | 105 | 4.0 | |
| Hispanic | 92 | 6.7 | 229 | 7.3 | |
| Missing | 12 | 0.9 | 48 | 2.1 | |
| Marital Status-Married | 812 | 56.5 | 1983 | 55.2 | 0.45 |
| Education | <0.01 | ||||
| Less than high school graduate | 391 | 19.8 | 759 | 14.2 | |
| High school graduate | 479 | 29.1 | 1066 | 24.4 | |
| At least some post high school | 762 | 50.1 | 2291 | 59.5 | |
| Missing | 13 | 1.0 | 47 | 1.9 | |
| Dementia (Probable/Possible) | 336 | 16.2 | 762 | 13.6 | 0.02 |
| Anxiety | 193 | 10.9 | 368 | 8.2 | <0.01 |
| Depression | 205 | 11.4 | 440 | 8.6 | <0.01 |
| Religious | 975 | 57.1 | 2426 | 54.8 | 0.16 |
| Preferences in Disability | 0.99 | ||||
| Stop all treatments | 1331 | 84.3 | 3443 | 84.3 | |
| Receive all treatments | 314 | 15.7 | 720 | 15.7 | |
| Weighted Mean (SD) | Weighted Mean (SD) | t-test p-value | |||
| Age | 1645 | 75.59 (6.98) | 4163 | 76.64 (5.80) | <0.01 |
| ADL Score | 1645 | 0.41 (1.16) | 4163 | 0.47 (1.07) | 0.09 |
| Hospital Stays Count | 1645 | 0.34 (0.92) | 4163 | 0.32 (0.96) | 0.51 |
Individual Trends — Cohort Study
Among older adults who responded in wave 2 and wave 8, in wave 2, 27% of older adults would receive LPT for the pain scenario and 12% for the disability scenario. In wave 8, 23% would receive LPT for the pain scenario and 13% for the disability scenario. Amongst older adults who would receive LPT at wave 2, 50% (95%CI: 42%–59%) would receive LPT at follow-up for pain scenario, and 59% (95%CI: 46%–71%) would receive LPT for disability scenario (Table 2). While amongst older adults who would reject LPT at wave 2, 87% (95%CI: 84%–90%) would reject LPT at follow-up for pain scenario, and 93% (95%CI: 91%–95%) would reject LPT for disability scenario. Overall percent providing the same response in both waves was 76% (95%CI: 72%–79%) for pain scenario and 86% (95%CI: 83%–89%) for disability scenario.
Table 2:
LPT Preferences among Older Adults who Responded in Round 2 and Round 8, cohort study
| Pain Scenario (N=606) | |||
|---|---|---|---|
|
| |||
| Receive LPT in round 8 | Reject LPT in round 8 | Total | |
| weighted Row % (95%CI) | weighted Row % [95%CI] | weighted Col % [95%CI] | |
| Receive LPT in round 2 | 50.4 (95%CI: 41.6–59.1) | 49.6 (95%CI: 40.9–58.4) | 27 (95%CI: 23.3–31) |
| Reject LPT in round 2 | 12.6 (95%CI: 9.6–16.4) | 87.4 (95%CI: 83.6–90.4) | 73 (95%CI: 69–76.7) |
| Total | 22.8 (95%CI: 19.2–26.9) | 77.2 (95%CI:73.1–80.8) | |
|
| |||
| Disability Scenario (N=628) | |||
|
| |||
| Receive LPT in round 8 | Reject LPT in round 8 | Total | |
| weighted Row % (95%CI) | weighted Row % (95%CI) | weighted Col % (95%CI) | |
| Receive LPT in round 2 | 59.1 (95%CI: 46–71) | 40.9 (95%CI: 29–54) | 12.4 (95%CI: 10.1–15.1) |
| Reject LPT in round 2 | 6.8 (95%CI: 4.9–9.3) | 93.2 (95%CI: 90.7–95.1) | 87.6 (95%CI: 84.9–89.9) |
| Total | 13.3 (95%CI: 11.1–15.8) | 86.7 (95%CI: 84.2–88.9) | |
National Trends — Serial Cross-sectional study
There was no significant national trend in preference to receive LPT for pain or disability scenario when comparing the cross-section of responses in wave 2 to the cross-section of responses in wave 8 (pain: 27.4% vs. 27%, p=0.80; disability: 15.8% vs. 15.7%, p=0.99) (Table 3 and Supplemental Table 3). When examining the trend by race/ethnicity, the changes between waves within race/ethnicity groups were not significant. For example, 41.3% of Black people in wave 2 would receive LPT in the setting of pain compared with 42.4% of NH black in wave 8 (p=0.77). After accounting for sociodemographics, physical and cognitive functioning, pain limitations, and exposure to illness, there was no interaction between race/ethnicity and time for the pain and disability scenarios (Table 4 and Supplemental Table 4).
Table 3.
National End-of-Life Treatment Preference Trends for the Disability Scenario by Race/ethnicity, cross-sectional study
| Wave 2 (N=1645) | Wave 8 (N=4163) | p value* | |||
|---|---|---|---|---|---|
|
| |||||
| Race/ethnicity | N | Receive LPT (row%, 95%CI) | N | Receive LPT (row%, 95%CI) | |
| Overall | 1645 | 15.8 (13.8 – 17.9) | 4163 | 15.7 (14.3 – 17.3) | 0.99 |
| White | 1148 | 11.9 (9.8 – 14.3) | 2969 | 12.1 (10.5 – 13.9) | 0.86 |
| Black | 351 | 36.1 (31.5 – 40.9) | 812 | 34.9 (30.7 – 39.4) | 0.73 |
| Other | 42 | 34.7 (18.9 – 54.8) | 105 | 24.7 (15.3 – 37.2) | 0.30 |
| Hispanic | 92 | 31.6 (23.1 – 41.6) | 229 | 31.6 (24.8 – 39.2) | 0.99 |
| Missing | 12 | 1.7 (0.2 – 12.7) | 48 | 11.8 (4.5 – 27.5) | 0.08 |
this is p value between waves within each race/ethnicity category
Table 4.
Factors associated with preference to receive life prolonging treatment for the disability scenario over time, cross-sectional study
| OR (95%CI) | p-value | OR (95%CI) | p-value | OR (95%CI) | p-value | |
|---|---|---|---|---|---|---|
| Race (base: White) | ||||||
| Black | 4.18 (3.09–5.67) | 0.00 | 3.75 (2.73–5.16) | <0.01 | ||
| Other | 3.94 (1.87–8.33) | 0.00 | 3.50 (1.63–7.53) | <0.01 | ||
| Hispanic | 3.43 (2.10–5.59) | 0.00 | 2.66 (1.62–4.37) | <0.01 | ||
| Missing | 0.13 (0.02–1.02) | 0.05 | 0.09 (0.01–0.98) | 0.05 | ||
| Time (base: wave 2) | ||||||
| Wave 8 | 1 (0.84–1.19) | 0.99 | 1.02 (0.81–1.28) | 0.86 | 1.10 (0.87–1.39) | 0.45 |
| Race#Time | ||||||
| Black, wave 8 | 0.93 (0.65–1.34) | 0.71 | 0.88 (0.61–1.28) | 0.51 | ||
| Other, wave 8 | 0.60 (0.23–1.56) | 0.30 | 0.62 (0.23–1.66) | 0.35 | ||
| Hispanic, wave 8 | 0.98 (0.54–1.76) | 0.94 | 0.95 (0.53–1.72) | 0.88 | ||
| Missing, wave 8 | 7.65 (0.74–78.88) | 0.09 | 7.81 (0.74–82.34) | 0.09 | ||
| Age | 0.97 (0.96–0.99) | 0.00 | ||||
| Education (base: Less than high school graduate) | ||||||
| High school graduate | 0.75 (0.57–0.98) | 0.03 | ||||
| At least some post high school | 0.79 (0.61–1.00) | 0.05 | ||||
| Missing | 1.13 (0.26–4.84) | 0.87 | ||||
| Marital Status - Married (base: No) | ||||||
| Yes | 1.02 (0.83–1.26) | 0.84 | ||||
| Sex (base: female) | ||||||
| Male | 1.33 (1.09–1.63) | 0.01 | ||||
| Dementia (base: No) | ||||||
| Probable/Possible | 1.59 (1.25–2.02) | <0.01 | ||||
| Anxiety (base: No) | ||||||
| Yes | 0.95 (0.68–1.34) | 0.79 | ||||
| Depression (base: No) | ||||||
| Yes | 1.26 (0.92–1.72) | 0.16 | ||||
| Religiosity (base: No) | ||||||
| Yes | 1.41 (1.16–1.72) | <0.01 | ||||
| ADL Limitations | 1.09 (1.02–1.17) | 0.01 | ||||
| Hospital Stay Count | 1.00 (0.94–1.07) | 0.91 |
Predictors of LPT Preference
Compared with White older adults, Black older adults were more likely to accept LPT (pain: aOR 2.04 [95%CI: 1.53–2.72], p<.01; disability: aOR 3.75 [95%CI: 2.73–5.16], p<.01). Respondents who were younger (aOR 0.97 [95%CI: 0.96–0.98] per year, p<.01) or had greater religiosity (pain: aOR 1.46 [95%CI: 1.24–1.71], p<.01; disability: aOR 1.41 [95%CI: 1.16–1.72], p<.01) were more likely to accept LPT in both scenarios. Those with more education were less likely to accept LPT in both scenarios. Respondents who were male (aOR 1.33 [95%CI: 1.09–1.63], p=0.01), Hispanic (aOR 2.66 [95%CI: 1.62–4.37], p<.01), with probable/possible dementia (aOR 1.59 [95%CI: 1.25–2.02], p<.01) or with more ADL limitations (aOR 1.09 [95%CI 1.02–1.17], p=0.01) were more likely to accept LPT when considering end-of-life disability scenario.
Discussion
In the cohort study of a national sample of older adults, we found that the preferences of individual older adults appeared stable, particularly among older adults who would reject LPT. However, we consider this finding hypothesis-generating given the substantial mortality and loss to follow-up. In the cross-sectional study, we found that national trends for LPT preferences were stable between 2012 and 2018. Given the overall stability of national LPT preferences, our findings support the importance of engaging in ACP before older adults face the end of their life to allow as many Americans as possible to relay their preferences. Our work also supports reviewing preferences when personal circumstances change, particularly for those who express a preference to accept LPT given these preferences were less stable than the preference to reject LPT over time.
Recently, some have questioned the value of ACP and have pointed to the lack of stable preferences as one contributor.(16) However, our results of overall stable individual LPT preferences, particularly those with preferences to forego LPT, are consistent with findings from a systematic review.(8) Few studies of LPT preference trends have been conducted among community-dwelling older adults.(8) Our results add to a North Carolina study conducted over 30 years ago which found overall stability of preferences over 2 years.(17) We add a larger sample size, contemporary sample, and longer duration of assessment of preference stability and the exploration of individual and national LPT trends.
We found that Black and Hispanic older adults and those who identify as other race/ethnicity were more likely to accept LPT than White older adults. The drivers of these differences are unknown and not explored in this study given the limited variables of interest. Further research could consider the role of racism, health-related social needs and childhood experiences on preferences for LPT. There were no differences in change in preferences over time, suggesting that national events did not have a differential impact on the stability of end-of-life preferences by race/ethnicity.
We found that older adults with probable/possible dementia and more ADL limitations were more likely to accept LPT than their counterparts. This finding contradicts previous research and expectations (18). Older adults with disability or dementia were more likely to be lost to follow-up some of which may be due to living in a long-term care facility. It could also be that older adults with disability or dementia who were more willing to accept LPT were more likely to be alive and therefore participate in the study. Further research is needed with preference ascertainment in mid-life would inform this finding.
Our study had limitations. In the cross-sectional study, several factors were modestly associated with LPT preferences. We also found evidence for differential loss to follow-up in the longitudinal study based on these factors, although the magnitude of these differences was not large. Consequently, it is unlikely that differential loss to follow-up will markedly influence our overall findings as large biases are only likely to emerge when large number of subjects are lost to follow-up, subjects lost to follow-up differ markedly from those that are not lost and the factors that influence lost to follow-up also influence the outcome markedly. However, as race, education and dementia were more strongly associated with LPT preferences in the cross-sectional study, our cohort study and the stability of individual LPT preferences should be interpreted cautiously in these subpopulations. The cross-sectional study accounts for these differences and allows some insights into the association between respondent characteristics and LPT preferences. This work is also limited by using scenario-based assessment of treatment preference without a report of underlying values or goals. While consistent with some common approaches to ACP, such assessments can be challenging to apply to other clinical situations or circumstances involving uncertainty or tradeoffs. Thus, initial and recurrent ACP discussions involving a potential proxy decision-maker that elicits personal values, life priorities, and treatment goals can better guide care plans over time and changing circumstances. Our results do not generalize to older adults who require a proxy to complete the survey.
In conclusion, we found that national trends in preferences for end-of-life treatment did not substantially change over time. Further research to explore the stability of individual end-of-life treatment preferences is warranted.
Supplementary Material
Table 1 describes the NHATS respondents to the pain scenario, cross sectional study. Table 2 presents the results of the National End-of-Life Treatment Preference Trends by Race/ethnicity for Pain Scenario, cross sectional study. Table 3 describes the factors associated with preference to receive life prolonging treatment for the pain scenario, cross-sectional study.
Key Message:
Our findings support the importance of engaging in ACP before older adults face the end of their life and reviewing preferences when individual circumstances change, particularly among older adults who express a preference for acceptance of LPT.
Acknowledgments:
The study was conceptualized and designed by (LES, CCL, JFB), critical revisions to the analysis plan by (ASK), analysis conducted by (CCL), analysis supervised by (JFB), funding obtained by (LES, JFB), manuscript drafted by (LES), critical revisions to the manuscript (CCL, ASK, JFB).
This work is funded by NIH/NIA R01AG059733. They had no role in the study.
Funding:
This work was supported by the National Institutes of Health NIH/NIA R01AG059733 to Drs. Skolarus and Burke
Footnotes
Paper presentation: None
All those who made significant contributions to the manuscript are included as authors. The authors have no conflicts.
References
- 1.Sudore RL, Lum HD, You JJ, et al. Defining Advance Care Planning for Adults: A Consensus Definition From a Multidisciplinary Delphi Panel. Journal of Pain and Symptom Management 2017;53:821–832.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fried TR, O’Leary J, Van Ness P, Fraenkel L. Inconsistency over time in the preferences of older persons with advanced illness for life-sustaining treatment. J Am Geriatr Soc 2007;55:1007–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Fried TR, Van Ness PH, Byers AL, et al. Changes in preferences for life-sustaining treatment among older persons with advanced illness. J Gen Intern Med 2007;22:495–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fried TR, O’Leary JR. Using the experiences of bereaved caregivers to inform patient- and caregiver-centered advance care planning. J Gen Intern Med 2008;23:1602–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kim YS, Escobar GJ, Halpern SD, et al. The Natural History of Changes in Preferences for Life-Sustaining Treatments and Implications for Inpatient Mortality in Younger and Older Hospitalized Adults. J Am Geriatr Soc 2016;64:981–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ditto PH, Jacobson JA, Smucker WD, Danks JH, Fagerlin A. Context changes choices: a prospective study of the effects of hospitalization on life-sustaining treatment preferences. Med Decis Making 2006;26:313–22. [DOI] [PubMed] [Google Scholar]
- 7.Straton JB, Wang NY, Meoni LA, et al. Physical functioning, depression, and preferences for treatment at the end of life: the Johns Hopkins Precursors Study. Journal of the American Geriatrics Society 2004;52:577–582. [DOI] [PubMed] [Google Scholar]
- 8.Auriemma CL, Nguyen CA, Bronheim R, et al. Stability of end-of-life preferences: a systematic review of the evidence. JAMA internal medicine 2014;174:1085–1092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fried TR, Redding CA, Robbins ML, et al. Stages of Change for the Component Behaviors of Advance Care Planning. Journal of the American Geriatrics Society 2010;58:2329–2336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Auriemma CL, Nguyen CA, Bronheim R, et al. Stability of end-of-life preferences: a systematic review of the evidence. JAMA Intern Med 2014;174:1085–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Frankford DM. The remarkable staying power of “death panels”. Journal of health politics, policy and law 2015;40:1087–1101. [DOI] [PubMed] [Google Scholar]
- 12.Kaiser Health Policy Tracking Poll: December 2014. Available from: https://www.kff.org/health-reform/poll-finding/kaiser-health-policy-tracking-poll-december-2014/. Accessed May 13, 2022. [Google Scholar]
- 13.Rainie L, Perrin A. Key findings about Americans’ declining trust in government and each other. Pew Research Center; 2019. [Google Scholar]
- 14.Hostetter M, Klein S. Understanding and ameliorating medical mistrust among Black Americans. The Common Wealth Fund Retrieved March 2021;10:2021. [Google Scholar]
- 15.Harrison KL, Adrion ER, Ritchie CS, Sudore RL, Smith AK. Low Completion and Disparities in Advance Care Planning Activities Among Older Medicare Beneficiaries. JAMA Internal Medicine 2016;176:1872–1875. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Morrison RS, Meier DE, Arnold RM. What’s Wrong With Advance Care Planning? JAMA 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Stability of Choices about Life-Sustaining Treatments. Annals of Internal Medicine 1994;120:567–573. [DOI] [PubMed] [Google Scholar]
- 18.Gjerdingen DK, Neff JA, Wang M, Chaloner K. Older persons’ opinions about life-sustaining procedures in the face of dementia. Archives of Family Medicine 1999;8:421. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.