Predictors of advance care planning in 11 high‐income nations

Abstract

Background

Elderly population is increasing in high‐income countries. For instance, by 2050, 21.4% of the United States population is expected to be 65+, thus making advance care planning (ACP) increasingly important. We aim to identify predictors of ACP completion in 11 high‐income countries and explore relationships between ACP and utilization factors.

Method

Using the 2021 International Health Policy (IHP) survey data, we assessed the relationship between sociodemographic factors, healthcare utilization, and ACP. The primary outcome variable was a composite of three ACP activities. A generalized linear mixed model (GLMM) was used to identify predictors of ACP completion.

Results

Analyses included 18,677 older adults who answered at least one ACP question. Only 5126 (27.4%) reported completion of three ACP activities. Germany (64.7%) showed the highest completion rates, while Sweden (5.0%) and France (5.0%) showed the lowest completion rates. Predictors of ACP completion identified in the GLMM were: increasing age (incidence rate ratio [IRR] range between 1.2 and 1.5), completion of high school education or more (IRR: 1.1, 95% CI: 1.1–1.1), higher income (IRR: 1.1, 95% CI: 1.1–1.2), presence of two or more health conditions (IRR: 1.1, 95% CI: 1.0–1.1), hospital stay in the past 2 years (IRR: 1.1, 95% CI: 1.1–1.1), and access to quality primary care (IRR: 1.0, 95% CI: 1.0–1.1). Male gender (IRR: 0.9, 95% CI: 0.8–0.9) had a negative association with ACP activity completion.

Conclusion

Several patient‐specific and health system utilization factors were identified as predictors of ACP activity completion, which clinicians and policymakers could use to enhance ACP completion.

Key points

• Among the older adults included in this survey, nearly two‐thirds (64.9%) reported completion of at least one ACP activity, but only slightly more than one in four (27.4%) reported the completion of all three ACP activities.

• Completion of all three ACP activities ranged from a high of 64.7% in those surveyed in Germany to a low of 5.0% in those surveyed in Sweden and France.

• Several patient‐specific and health system utilization predictors of completion of ACP activities were identified in a generalized linear mixed model.

Why does this paper matter?

ACP has been shown to help align treatment with patient preference, reduce end‐of‐life hospitalization and costs, and support family members during surrogate decision making and bereavement. By understanding the patient‐specific and health system utilization predictors of ACP and the nations with the highest and lowest rates of ACP, clinicians and health policymakers can develop informed and targeted programs to improve ACP practices.

INTRODUCTION

Elderly population is increasing in high‐income countries. For instance, by 2050, 21.4% of the United States population is expected to be 65 years of age or older, surpassing the percentage of individuals aged 18 and younger. ¹ Additionally, Medicare spending increased to $944.3 billion in 2022 ² and is projected to account for up to 5% of the U.S. gross domestic product by 2035. Given the demographic shift and rising costs, advance care planning (ACP) is increasingly pertinent as such plans can reduce hospitalization ³ and costs associated with end‐of‐life care, ⁴ , ⁵ and has been promoted by healthcare providers in developed countries after seeing low rates of individuals with such plans. ⁴ As described by the National Institute on Aging, ACP is a process of arranging plans for health care in advance of illness or injury in case the patient cannot communicate with providers at the time of hospitalization. ⁶ The process can involve speaking with providers or family members about care preferences or describing preferences in written documents. Such discussions or documents may include preferences on intubation, resuscitation, treatment methods, and hospitalization, as well as power of attorney and living wills. ACP has been found to benefit patients and families by allowing healthcare providers to give more patient‐centered care and reducing intensive treatment, which has been associated with improved overall quality of life. ⁷ It has been shown to make decision‐making more accessible for families and help them know they made choices a loved one would have wanted. ⁸ , ⁹ ACP has also been shown to reduce the emotional distress of family members following the death of a loved one, as well as provide other significant benefits. ¹⁰ ACP can be used by all adults to provide security for their healthcare treatment preferences.

Advance directives are another essential part of ACP. They are legal documents, such as a living will and healthcare power of attorney, which will help providers understand a patient's wishes in the event they are unable to communicate and friends or family cannot provide the information. ⁶ This ensures that individuals have control over their care and improves their overall quality of end‐of‐life care. ⁷ , ⁸

Controversies and issues related to ACP are common in high‐income nations. While some countries, such as Sweden, have found benefits from implementing ACP, ⁹ other nations experience problems associated with implementation. In a Canadian study, researchers reported that communication with healthcare professionals and patients was insufficient, diminishing the positive effects of instituting ACP, including outlining comfort and end‐of‐life preferences for older adults. ¹⁰ In addition, the rate of ACP remains low in these countries, meaning many individuals do not have plans in place. For instance, results from a meta‐analysis of studies in the United States involving participants 18 years and older indicated that only 36.7% of individuals completed some advanced directive. ¹¹

Given the lack of ACP completion identified in the literature, the objective of this study was to identify predictors of the completion of ACP activities in older adults in 11 high‐income countries. Using data from the most recent 2021 Commonwealth Fund's (CMWF) International Health Policy (IHP) survey of older adults, we explored the relationship of ACP to patient‐specific and health system utilization factors, while identifying whether differences exist across 11 high‐income nations. This is an update of a 2018 study by Sable‐Smith et al. ¹² that analyzed data from the 2014 IHP survey of older adults. Briefly, their findings indicated that just over half of all respondents completed at least one ACP activity and less than one‐fifth completed all included activities. Also, higher education, multimorbidity, informal caregiving, and inpatient hospitalization, among others, were associated with completion of more ACP activities.

METHODS

Data

This analysis used the 2021 CMWF IHP survey of older adults. ¹³ The survey collected information on access to care, care coordination, use and experience with the health system, and socioeconomic conditions from adults 65 years of age and older living in Australia, Canada, France, Germany, the Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom, and adults 60 years of age and older in the United States. Using the probability sampling design, the survey aimed to provide national‐level estimates for each country. It was conducted via landline and mobile telephone or online between March 1 and June 14, 2021. More details about the study sample and methodology have been published elsewhere. ¹³

Variables

The primary outcome variable was a composite of three ACP questions asked in the “End of life care wishes” section of the survey. The three questions were: (1) “In the event you become very ill or injured and you cannot make decisions for yourself, have you had a discussion with family, a close friend, or with a healthcare professional about what healthcare treatment you want, or do not want?,” (2) “Do you have a written plan or document describing the healthcare treatment you want or do not want at the end of your life?” and, (3) “Do you have a written document that names someone to make treatment decisions for you if you cannot make decisions for yourself?” Counting the number of positive responses to these questions, a composite variable was created with values 0, 1, 2, and 3. The “Not sure/Decline to Answer; WEB ONLY: Blank” responses to all three questions were excluded from the count.

To define predictor variables, we primarily followed the methodology used by Sable‐Smith et al. ¹² We only departed from their methodology whenever we discovered a change in the related survey question. We defined age as a categorical variable, with the categories being 60–64, 65–69, 70–74, and 75+ years of age. The respondents' gender was categorized as male or female. The education variable was a binary variable defining those who completed a high school education or higher or otherwise. The income variable categorized respondents into two categories: those with income 50%–60% of their national median or below. Respondents' general health status was categorized as “Excellent/very good/good” and “Fair/poor.” Multimorbidity was captured as a dichotomous variable, with respondents with two or more chronic diseases grouped against those with one or no disease. The hospitalization and emergency department (ED) visits in the past 2 years were also dichotomized (having at least one visit vs. having no visits). Following Sable‐Smith et al., ¹² our quality of primary care index was developed using positive responses to the key functions of primary care. The questions in our index were related to access to care and care coordination (a total of seven questions). Unlike the CMWF IHP 2014 survey, the 2021 survey did not include questions on primary care provider communication, interpersonal relations, and cultural competence (7 questions) for all countries (most of them were asked only for Swedish residents). For similar reasons, we could not include informal caregiving as a predictor.

Analytical approach

Besides descriptive statistics, we used a generalized linear mixed model (GLMM) with a log link function to model the count data on our predictor variable, assuming having a Poisson distribution. The overdispersion (and, thus, if Poisson distribution was appropriate) was checked by using the method described by Gelman and Hill. ¹⁴ The respondent's residing country was treated as a random effect to account for within‐country variation. All the predictor variables were included a priori as their significance had already been reported by Sable‐Smith et al. ¹² Each covariate's incidence rate ratio (IRR) was calculated, and statistical significance was considered at p‐value <0.05. Survey weights provided with the dataset were used to derive the population‐level estimates. The descriptive analyses and visualization were performed using “pollster,” ¹⁵ “gtsummary,” ¹⁶ “fmsb,” ¹⁷ and “ggplot2” ¹⁸ packages, whereas GLMM modeling was performed using the “lme4” ¹⁹ package in R statistical software version 4.3.2. ²⁰

Ethics approval

Augusta University's Institutional Review Board (IRB) exempted this study from full review as it is based on secondary de‐identified data (application no. 2071045‐1).

RESULTS

Our analysis is based on 18,677 older adults who answered at least one ACP question in the survey. As can be seen in Table 1, 6560 (35.1%) respondents reported the completion of no ACP activities, 4624 (24.8%) reported the completion of one ACP activity, 2368 (12.7%) reported the completion of two ACP activities, while 5126 (27.4%) reported the completion of all three ACP activities. Respondents from Germany (64.7%), the United States (44.0%), and Switzerland (38.3%) reported the highest completion rates for all three ACP activities while older adults in Sweden (5.0%), France (5.0%), and Norway (5.6%) reported the lowest completion rates for all three ACP activities. Moreover, 62.6% of those surveyed in Norway and 62.5% of those surveyed in Sweden reported that they had not completed any of the three ACP activities. The percentage of individuals completing none versus all three ACP activities for each nation can be seen in the radar graph in Figure 1.

TABLE 1.

Bivariable analysis for composite variable for advance care planning activities (weighted).

Characteristics	Number of advance care planning activities				p‐value b
	0	1	2	3
	N = 6560 (35.1%) a	N = 4624 (24.8%) a	N = 2368 (12.7%) a	N = 5126 (27.4%) a
Age (years) (n = 18,677)					<0.001
55–64	161 (32.3%)	127 (25.5%)	59 (11.8%)	152 (30.5%)
65–69	2097 (39.3%)	1402 (26.3%)	569 (10.7%)	1269 (23.8%)
70–74	1688 (37.1%)	1150 (25.3%)	545 (12.0%)	1168 (25.7%)
75+	2613 (31.5%)	1944 (23.4%)	1196 (14.4%)	2536 (30.6%)
Gender (n = 18,677)					<0.001
Female	3239 (31.8%)	2636 (25.9%)	1297 (12.7%)	3022 (29.6%)
Male	3321 (39.1%)	1988 (23.4%)	1071 (12.6%)	2104 (24.8%)
Education (n = 18,013)					<0.001
High school or greater	2038 (30.8%)	1527 (23.1%)	889 (13.4%)	2169 (32.8%)
Less than high school	4327 (38.0%)	2934 (25.8%)	1397 (12.3%)	2731 (24.0%)
(Missing)	195	162	82	225
Income (n = 18,677)					<0.001
No, income is not 50%–60% of median household income	2464 (42.0%)	1538 (26.2%)	717 (12.2%)	1151 (19.6%)
Yes, income is 50%–60% of median household income	4097 (32.0%)	3085 (24.1%)	1651 (12.9%)	3975 (31.0%)
Multimorbidity (n = 18,622)					<0.001
Two or more health conditions	3208 (31.8%)	2580 (25.6%)	1359 (13.5%)	2937 (29.1%)
Less than two	3350 (39.0%)	2044 (23.8%)	1007 (11.7%)	2178 (25.4%)
(Missing)	3	0	2	11
General health status (n = 18,565)					0.051
Excellent/very good/good	4844 (34.9%)	3414 (24.6%)	1726 (12.4%)	3912 (28.2%)
Fair/poor	1664 (35.6%)	1187 (25.4%)	630 (13.5%)	1190 (25.5%)
(Missing)	52	23	13	24
Hospital stay in past 2 years (n = 18,580)					<0.001
No	5113 (36.3%)	3493 (24.8%)	1747 (12.4%)	3737 (26.5%)
Yes	1404 (31.3%)	1107 (24.7%)	610 (13.6%)	1369 (30.5%)
(Missing)	43	24	11	19
Emergency department visit in past 2 years (n = 18,386)					0.014
No	4623 (35.5%)	3250 (25.0%)	1573 (12.1%)	3576 (27.5%)
Yes	1791 (33.4%)	1304 (24.3%)	758 (14.1%)	1511 (28.2%)
(Missing)	147	69	38	38
Average score on quality of primary care index (n = 18,677)	4.7 (1.1)	4.7 (1.1)	4.8 (1.1)	5.0 (1.2)	<0.001
Nation (n = 18,677)					<0.001
Australia	136 (27.1%)	112 (22.4%)	127 (25.3%)	127 (25.3%)
Canada	860 (20.0%)	1026 (23.8%)	782 (18.2%)	1638 (38.1%)
France	891 (51.1%)	542 (31.1%)	224 (12.8%)	87 (5.0%)
Germany	147 (13.4%)	221 (20.2%)	18 (1.6%)	708 (64.7%)
The Netherlands	229 (36.7%)	206 (33.0%)	86 (13.8%)	103 (16.5%)
New Zealand	136 (27.2%)	155 (31.0%)	109 (21.8%)	100 (20.0%)
Norway	313 (62.6%)	98 (19.6%)	61 (12.3%)	28 (5.6%)
Sweden	1879 (62.5%)	788 (26.2%)	188 (6.3%)	151 (5.0%)
Switzerland	732 (28.3%)	639 (24.7%)	225 (8.7%)	993 (38.3%)
United Kingdom	850 (45.4%)	388 (20.7%)	298 (15.9%)	336 (17.9%)
United States	390 (20.0%)	449 (23.1%)	251 (12.9%)	855 (44.0%)

^a Row percentages are calculated for categorical variables by excluding the missing values; mean (standard deviation) values are presented for continuous variables.

^b Chi‐squared test with Rao and Scott's second‐order correction; Wilcoxon rank‐sum test for complex survey samples.

FIGURE 1.

This radar graph shows the number of respondents from each of the 11 high‐income nations who completed no advanced care planning activities, and those who completed all 3 advanced care planning activities (discussed desired treatments, created a written plan, appointed a healthcare decision maker), as well as the combined totals for all countries.

Table 1 shows bivariable associations between the ACP completion activities and the included predictor variables. The number of ACP activities completed varied significantly across age, gender, education, income, multimorbidity, and respondents' general health status. Similarly, healthcare system utilization‐related covariates like hospital stay, and ED visits in the past 2 years showed a significant association to ACP. The average score on the quality of primary care index also varied significantly among those who reported differing numbers of completion of activities. Variations across nations were also observed.

The GLMM model results are presented in Table 2, and several predictors of completion of ACP activities were identified. Among patient‐specific factors, the rate of activity completion increased with age. Respondents 65–69, 70–74, and 75+ years old had higher rates of completion, 20.0% (CI: 1.1–1.3; p‐value <0.001), 30.0% (CI: 1.2–1.4; p‐value <0.001), and 50.0% (CI: 1.4–1.6; p‐value <0.001), respectively. The completion rate was also higher for those with a high school education or greater (IRR: 1.1, 95% CI: 1.1–1.1; p‐value <0.001) than those without at least a high school education. Similarly, respondents with income at least 50%–60% of the national median had a 12.0% (CI: 1.1–1.2; p‐value <0.001) higher completion rate than the low‐income respondents. Lower completion rates were reported among males than females (IRR: 0.9, 95% CI: 0.8–0.9; p‐value <0.001). Individuals with two or more chronic diseases had a 10.0% (CI: 1.0–1.1; p‐value <0.001) higher completion rate than those with only one or no morbidity. An individual's general health status did not significantly affect the ACP.

TABLE 2.

Generalized linear mixed model for advance care planning.

Characteristics	IRR (95% CI)	p‐value
Age‐group (years)
55–64	–
65–69	1.2 (1.1–1.3)	<0.001
70–74	1.3 (1.2–1.4)	<0.001
75+	1.5 (1.4–1.6)	<0.001
Gender
Female	–
Male	0.9 (0.8–0.9)	<0.001
Education
Less than high school	–
High school or higher	1.1 (1.1–1.1)	<0.001
Income
No, income is not 50%–60% of median household income	–
Yes, income is 50%–60% of median household income	1.1 (1.1–1.2)	<0.001
Multimorbidity
Less than two morbidities	–
Two or more morbidities	1.1 (1.0–1.1)	<0.001
General health status
Fair/poor	–
Excellent/very good/good	1.0 (0.9–1.1)	0.41
Hospital stay in past 2 years
No	–
Yes	1.1 (1.1–1.1)	<0.001
Emergency department visit in past 2 years
No	–
Yes	1.0 (0.9–1.0)	0.36
Average score on quality of primary care index	1.0 (1.0–1.1)	<0.001

Abbreviations: CI, confidence interval; IRR, incidence rate ratio.

For healthcare utilization variables, individuals who reported hospitalization in the past 2 years had 10.0% (CI: 1.1–1.1; p‐value <0.001) higher rates than those who did not report any hospitalization. Those with at least one ED visit in the past 2 years did not have statistically significant associations with ACP. With every one‐unit increase in the average score on the quality of primary care index, the ACP completion rate increased by 5.0% (CI: 1.0–1.1; p‐value <0.001). Using the method described by Gelman and Hill, ¹⁴ we estimated a dispersion ratio of 0.96, justifying the use of Poisson distribution for modeling.

DISCUSSION

We examined the ACP practices of older adults from 11 high‐income countries using the 2021 CMWF IHP survey. Our results were similar to the 2018 study by Sable‐Smith et al., ¹² with ACP increasing with age, and higher rates of ACP observed among females, more educated individuals, those with higher income, those with multiple morbidities, and individuals who were hospitalized within the last 2 years. Moreover, the effect of the average score on the quality of primary care index showed a statistically significant positive association. Unlike the 2018 study, however, we did not find statistically significant relationships between ACP and ED visits or health status.

Age and multimorbidity are closely related, ¹⁹ , ²⁰ with older adults being nearer to the end of their expected lifespan and multimorbidity having increased mortality. ²¹ The higher mortality in this population may lead to ACP being more cogent among both patients and providers. Higher educational attainment is associated with health literacy, ²² and low health literacy is associated with health behavior factors, including lower preventive care utilization such as disease screening. ²³ More educated, and therefore, more health literate individuals may be more likely to request ACP, leading to increased utilization. Similarly, there are well‐established associations between higher income and better health, including morbidity, mortality, insurance, access to care, and others. ²⁴ While lower‐income populations are more likely to have higher rates of morbidity and mortality, limited access to care may prevent ACP utilization. Like Sable‐Smith, we find a positive statistically significant relationship between primary care quality and ACP. However, as they note, we cannot conclude if ACP activities happened in primary care settings or higher ACP activities are common in health systems having quality primary care networks.

While respondents in the current study who were hospitalized in the past 2 years had higher rates of ACP, previous work has found that several barriers exist in hospitals for successful completion of ACP, including lacking time and understanding of ACP, limited communication, and others. ²⁵ It may be, then, that the association is a function of the population examined. Older adults ²⁶ and those with multimorbidity ²⁷ are more likely to be hospitalized and as noted above, may have traits which make ACP more likely to occur. Thus, ACP is more likely to occur for this population, and is observed more frequently in hospitals due to higher hospitalization rates.

ED visits were not associated with higher rates of ACP. It is possible that the same barriers that apply to hospitals also apply to ED visits. Regarding the ED specifically, lack of training among ED staff, prioritization of certain patients, and long wait times have been noted by some providers as barriers to palliative care in the ED. ²⁸ This may also be the case regarding ACP. Furthermore, studies have shown that those who have ACP in place are less likely to visit the ED. ²⁸ , ²⁹ ED visits may not be significantly related to ACP because providers may be unable or unwilling to provide ACP, and those with plans already in place are visiting EDs less frequently. Additionally, general health status was not significantly related to increased ACP. Sable‐Smith et al.'s ¹² findings from the 2014 dataset were significant. Even though insignificant (perhaps due to variation in sample size), our result was in a similar direction (IRR: 0.99). The contrasting results for multimorbidity and general health status may be due to the intent of the survey question for the latter. As noted above, multimorbidity increases healthcare encounters, providing opportunities to complete ACP activities. The general health status captured the respondents' perceptions of their health at the time of the survey. It is possible that despite having multimorbidity, the respondents might have reported excellent or good health in the survey, probably because they were accustomed to an altered quality of life.

Between countries, we found the highest completion rates for all three ACP activities in Germany, the United States, and Canada, and the lowest rates in Sweden, Norway, and France. Our findings align with previous work (e.g., Osborn et al., 2017). ³⁰ These differences may be due to country‐level characteristics. For example, Germany introduced a policy in 2018 that allows providers from certain facilities to be reimbursed by insurance for providing ACP to patients. ³¹ While the policy does not cover all individuals and providers, ³² allowing even some providers to bill for ACP may be, at least in part, contributing to Germany's higher rates. Rates are likely higher in Canada and the United States for similar system‐level reasons. Canada has a national initiative which raises ACP awareness among the public and providers. ³³ In the United States, the Patient Self Determination Act of 1990 ensures that Medicare and Medicaid patients in a variety of settings (e.g., hospitals, nursing facilities, hospice care) understand their rights, are regularly questioned about ACP, and have their wishes honored. ³⁴ In 2016, another act allowed physicians to be reimbursed by Medicare for providing ACP. ³⁵ These acts, along with major legal cases and campaigns, have dramatically increased ACP. ³⁶

The provisions similar to Germany, the United States, and Canada are absent or less structured in the countries with lower ACP. In Sweden, there are no laws regarding living wills. ³⁷ Without a legal structure in place, ACP practices have been inconsistently implemented and poorly integrated, and ACP providers are left without substantial support or training. ³⁸ While France does have legislation regarding living wills, they are not legally binding. ³⁷ A study of French general practice residents found residents were interested in providing information and helping with drafts of living wills, but they also indicated that only certain patients should be informed of ACP, that not every part of a living will or advance directive need be followed, that they would not share advance directives with a hospital at time of admission, and they would not consider advance directives for all patients. ³⁹ In Norway, ACP practices are encouraged but not legally enforced, and healthcare providers hold considerable authority over medical decisions, particularly when patients are acting in a reduced capacity. ⁴⁰ ACP receives limited practice, ⁴¹ despite the fact that most Norwegians want to be involved in ACP and have the final say in their care decisions. ⁴⁰ As noted previously, there is some evidence to suggest that ACP may reduce end‐of‐life care costs. ⁵ Findings from the limited research comparing costs between countries suggests additional factors may be influencing cost. One study found that, in the last 3 months of life among patients diagnosed with cancer, the highest spending occurred in Canada, Norway, and the United States, ⁴² despite the differences observed in ACP practices in the current study. Regardless, the aims of ACP extend beyond cost reduction and, as compiled from a qualitative analysis of the literature, include respecting patients' autonomy, improving care quality and relationships, helping patients prepare of end‐of‐life, and limiting unwanted treatment. ⁴³

The above few examples indicate that effective legislation could contribute significantly to improving and increasing ACP practices. However, simply having legislation in place and recommending providers follow guidelines may not be enough to ensure patients receive quality ACP. Making documents such as living wills and advance directives legally binding and enforcing laws regarding provision of ACP could further enhance ACP. Additionally, informing individuals about the importance of, and their rights surrounding ACP through outreach programs could also increase frequency of ACP. These are important considerations for healthcare providers and policymakers when addressing ACP in their respective countries.

While the current study focused on older adults since, as discussed above, they are more likely to have conditions making ACP imminently relevant, it should be noted that ACP can benefit adults of any age. It can ensure healthcare wishes are met in the event of an accident, sudden‐onset health condition, or other such situations. Additionally, younger adults with terminal or serious illnesses or conditions should engage in ACP to establish care preferences and nominate desired healthcare proxies or powers of attorney. ⁶

Strengths and limitations

The study's strength is using the IHP dataset, a nationally representative dataset for the 11 high‐income countries mentioned above. Lack of ethnicity data and information on the completion sequence of ACP activities mentioned by Sable‐Smith et al. ¹² also applies to our study. Moreover, it is possible that age groups excluded from the survey and analysis could behave differently while accomplishing the ACP activities. We also caution readers to read our results with consideration that the ACPs are based on self‐reported data, which is susceptible to a reporting bias. We could not, however, determine the directionality of the bias. Nevertheless, we provide the most updated analysis using the latest data from these countries for older adults who might need ACP in the near term.

Conclusion

Among the older adults included in this survey, nearly two‐thirds (64.9%) reported completion of at least one ACP activity, but slightly more than one in every four (27.4%) reported the completion of all three ACP activities. Completion of all three ACP activities ranged from a high of 64.7% in those surveyed in Germany to a low of 5.0% in those surveyed in Sweden and France. Several patient‐specific and health system utilization factors were identified as predictors of completion of ACP activities, which could be used by clinicians and health policymakers to enhance the completion of ACP.

AUTHOR CONTRIBUTIONS

NJM served as PI for the study. NJM and PNA developed and designed the study. PNA conducted data analysis and interpretation. NJM, PNA, ZTH, and KM wrote and revised the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

SPONSOR'S ROLE

No funding was received for this study. Data from the IHP survey were provided by The Commonwealth Fund. The Commonwealth Fund did not contribute to study design, methods, analysis, or preparation of the paper.

FINANCIAL DISCLOSURE

No funding was received for this work.

Ambade PN, Hoffman ZT, Mehra K, MacKinnon NJ. Predictors of advance care planning in 11 high‐income nations. J Am Geriatr Soc. 2024;72(12):3855‐3864. doi: 10.1111/jgs.19226