Advance Care Planning Documentation Completeness and End-of-Life Care: Trends and Associations Using HRS 2010-2022 Data

Abstract

Background

Advance care planning documentation allows individuals to communicate preferences for medical care and designate surrogate decision-makers. It is unknown if comprehensive documentation confers additional benefits.

Objectives

This study examined additive associations between comprehensive advance care planning (ACP) documentation and end-of-life care (EOL) outcomes among older adults in the United States.

Setting/Subjects

Data from the Health and Retirement Study exit interviews (2010-2022) indicated ACP documentation status, based on proxy-reported completion of a living will (LW) and/or durable power of attorney (DPOA) for 5622 decedents representing 23.2 million individuals.

Measurement

Documentation was operationalized as a binary variable (any document, no document) and an ordinal variable (no documentation, one document, two documents). EOL outcomes included binary indicators of intensive care unit use in the last two years of life, use of life-sustaining treatments, hospice utilization prior to death, and location of death (out-of-hospital, hospital).

Results

About 42.7% decedents had two documents and 28.9% had none, documentation increased substantially around 2014. Compared with no documentation, having any documentation was associated with lower likelihood of life-sustaining treatment (aRR = 0.85, 95% CI: 0.74-0.98) and higher likelihood of hospice utilization (aRR = 1.43, 95% CI: 1.28-1.60) and out-of-hospital death (aRR = 1.11, 95% CI: 1.06-1.18), but not ICU use. Having two documents showed similar patterns. Associations were stronger among decedents with expected death and attenuated among those with unexpected death.

Conclusions

ACP is associated with less aggressive EOL care and greater hospice use, although the incremental benefits of having both documents are modest.

Keywords

advance care planning living will durable power of attorney end-of-life care

Key Message

Advance care planning documentation is associated with less aggressive end-of-life care and greater hospice use among older adults. However, having multiple documents offers only modest additional benefit, suggesting that communication and care processes, rather than documentation alone, are central to achieving goal-concordant end-of-life care.

Introduction

Advance care planning (ACP) is central to align end-of-life care with patient preferences. ACP allows individuals to articulate their future medical treatment and designate surrogate decision-makers, thereby promoting care aligned with patients’ values and goals.^1-4 Two key components of ACP include living wills, which usually document healthcare preferences such as life-sustaining treatments, and durable power of attorney (DPOA) for healthcare, which appoints a surrogate decision-maker to act on the patient’s behalf.⁵ Together, these comprehensive ACP documentations are intended to support shared decision-making and reduce uncertainty during critical clinical situations when patients may lack decisional capacity.^5-7

Despite longstanding policy and clinical emphasis on ACP, evidence regarding its effectiveness in improving end-of-life outcomes remains mixed.^8-10 Earlier studies suggested that ACP is associated with reduced use of aggressive interventions (eg, intensive care unit [ICU] admission, mechanical ventilation, cardiopulmonary resuscitation) and increased hospice utilization.^11-16 However, more recent research highlights substantial heterogeneity in these associations, with some studies finding limited or context-dependent effects, particularly after adjusting for patient preferences, illness trajectories, and healthcare system factors.^17-19 For example, recent analyses using nationally representative datasets have shown that ACP documentation is associated with greater hospice use and a higher likelihood of dying in their preferred setting (eg, outside of a hospital), but not with reductions in potentially burdensome care such as gastrostomy tube insertion, intubation/mechanical ventilation.¹⁰ One factor contributing to this heterogeneity is that most prior studies have operationalized ACP documentation as a binary construct (any vs none), potentially obscuring important differences between partial and comprehensive ACP documentation.^20-22 Emerging evidence emphasizes that ACP is an ongoing process rather than a one-time event, and that effective ACP requires integration of documentation with communication and clinical decision-making processes.^23-27 In this context, having both a living will and a DPOA may represent a more complete and actionable form of ACP, potentially leading to more goal-concordant end-of-life care.¹⁵

Additionally, whether these additive associations differ by the anticipated trajectory of death is further unresolved question.²⁸ When death is anticipated, there may be a greater opportunity for ACP to influence care decisions through communication and planning. In contrast, in cases of unexpected death (eg, sudden death from a heart attack), the opportunity for ACP to shape care may be more limited, and the presence of a surrogate decision-maker may play a more critical role than documented preferences alone. However, to our knowledge, few prior studies explicitly examined whether the associations between ACP and end-of-life care differ by the expectedness of death.

Conceptually, this study is guided by contemporary models of ACP that view ACP as a multistep process rather than a discrete documentation event. The Advance Care Planning Process Framework posits that formal documents, such as living wills and durable powers of attorney for healthcare, serve as mechanisms through which patient preferences are communicated, surrogate decision-makers are prepared, and shared decision-making can occur when decisional capacity is lost.^4,27 In this framework, documentation alone does not directly determine end-of-life care; rather, its influence is mediated through communication, surrogate engagement, and the accessibility of patient preferences during clinical decision-making. Consequently, more comprehensive ACP documentation may facilitate goal-concordant care by strengthening these underlying processes. We further posit that these mechanisms may operate differently depending on whether death is anticipated, as expected illness trajectories provide greater opportunity for communication and care planning than sudden or unexpected deaths.

To address these gaps, we used nationally representative data from the Health and Retirement Study (HRS) to examine additive associations between comprehensive ACP documentation (living wills and DPOA) and end-of-life care outcomes among U.S. older adults. We evaluated whether having both documents is associated with differences in intensive care use, life-sustaining treatment, hospice utilization, and location of death, compared with having one or no documents. We further examined whether these associations vary by the expectedness of death and assessed temporal trends in ACP documentation. By distinguishing between partial and comprehensive ACP documentation, this study aims to provide new insights into the mechanisms by which ACP documentation may influence end-of-life care and to inform policies and interventions that promote goal-concordant end-of-life care.

Methods

Data Source and Study Population

Guided by the Advance Care Planning Process Framework, ACP documentation was conceptualized as a structural component of advance care planning that may facilitate communication, surrogate preparedness, and goal-concordant decision-making at the end of life. To embody this framework, we conducted a retrospective cross-sectional study using HRS data, a nationally representative longitudinal survey of U.S. adults aged 50 years and older.^29,30 We analyzed exit interview data collected from 2010 to 2022, generally covering decedents who died one to three years before the interview. The interviews were administered to proxy respondents following a participant’s death and provide detailed information on end-of-life experiences, healthcare utilization, and advance care planning.³¹ The analytic sample included decedents aged 50 years or older with complete information on ACP documentation (living will and/or DPOA) and having positive sampling weights (n = 6986). After excluding respondents with missing data on sociodemographic and death-related variables, the main analytic sample consisted of 5622 decedents, representing approximately 23.2 million individuals nationwide. The University of Florida Institutional Review Board determined that this study involved non-human subjects research using publicly available, de-identified data and was therefore exempt from review. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.

Measures

The primary exposure was the comprehensive ACP documentation completion status, based on proxy-reported information on whether the decedent had a living will and/or a DPOA for healthcare. We operationalized it using two approaches. First, a binary indicator distinguished individuals with any documentation (living will and/or DPOA) from those with none. Second, an ordinal variable categorized individuals into three mutually exclusive groups to ensure adequate cell sizes for multivariable regression analyses: no documentation, one document (either a living will or DPOA), and two documents (both a living will and DPOA). This categorization allowed examination of the additive associations of having both components of ACP documentation.

We examined four end-of-life care outcomes reported in the exit interview: ICU use in the last two years of life, use of life-sustaining treatments, hospice utilization prior to death, and location of death categorized as out-of-hospital vs hospital death. All outcomes were treated as binary variables (yes, no).

Demographic characteristics included age at death (50-64, 65-84, ≥85 years), sex (male, female), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, non-Hispanic Other), educational attainment (GED or lower, high school graduate, some college, college graduate or above), marital status (married or living with a partner, divorced or separated, widowed, never married), income categorized into quartiles (Q1–Q4), with Q1 representing the lowest income group, cause of death (cancer, cardiovascular disease, or other causes), and whether the death was expected at about the time it occurred (yes/no). These variables were selected based on prior literature demonstrating their associations with advance care planning and end-of-life decision-making.^32,33

Statistical Analysis

We first compared decedents’ characteristics by ACP documentation status using chi-square tests. We then estimated the annual prevalence of ACP documentation completion by year of death from 2008 to 2021 and assessed linear trends over time. These analyses were additionally stratified by expectedness of death to examine potential differences in temporal patterns. To further characterize trends, we conducted joinpoint regression analyses to estimate annual percent change (APC) in ACP documentation completion and to evaluate the statistical significance of changes in trend slopes across distinct time periods.³⁴

Next, we calculated the prevalence of each end-of-life outcome across ACP documentation groups. To examine associations between ACP documentation and end-of-life care outcomes, we used modified Poisson regression models with robust variance estimation to estimate adjusted risk ratios (aRRs) and 95% confidence intervals (CIs) adjusting for aforementioned covariates. This approach was selected to provide valid and interpretable estimates of relative risk for common binary outcomes.³⁵ Models were specified using both binary (any vs none) and ordinal (none, one document, two documents) exposure definitions, with no ACP documentation as the reference category. We further conducted stratified analyses by expected vs unexpected death to assess whether associations varied by the anticipated nature of death.

All analyses accounted for the HRS complex survey design, incorporating last wave interview sampling weights, strata, and primary sampling units to generate nationally representative estimates. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.) and Stata version 19 (StataCorp), and statistical significance was defined as a two-sided P-value less than 0.05.

Results

Overall, among the 5622 decedents, 28.9% of decedents had no ACP documentation, 28.3% had one document, and 42.7% had two documents (Table 1). ACP documentation differed significantly across sociodemographic and clinical characteristics. Decedents with both documents were more likely to be older, with nearly half aged 85 years or older, compared with less than one-fifth among those without any documentation. They were also more likely to be female (53.5%), non-Hispanic White (90.2%), and have college or above educational attainment (23.8%). In addition, individuals with two documents were more likely to have experienced an expected death compared with those without documentation (66.1% vs 43.6%).

Table 1.

Respondent-Reported Characteristics of Decedents by Dementia Status Before Death (n = 5622, Weighted n = 23,194,490)

Characteristics	Advance care planning documentation (weighted %)								P
	None		Living will only		DPOA only		Two documents (living will and DPOA)
	No.	%	No.	%	No.	%	No.	%
Unweighted, n	1730		1604				2288
Weighted, n (%)	6 712 744	28.9	6 574 328	28.3			9 907 418	42.7
Age at death (Year)									<.001
50-64	368	26.0	194	14.8			134	7.9
65-84	996	55.4	823	49.4			1090	46.4
85+	366	18.7	587	35.8			1064	45.7
Sex									<.001
Male	916	55.7	762	48.5			1052	46.5
Female	814	44.3	842	51.5			1236	53.5
Race/ethnicity									<.001
NH white	945	65.7	1101	77.7			1973	90.2
NH black	507	20.2	323	12.1			178	4.7
Hispanic	237	11.1	142	6.8			99	3.3
NH other	41	3.0	38	3.5			38	1.8
Education									<.001
GED or lower	748	38.8	596	33.6			576	22.5
High school graduate	504	31.8	495	30.8			700	29.3
Some colleges	305	18.8	318	21.6			533	24.4
College and above	173	10.6	195	14.0			479	23.8
Marital status									<.001
Married or living with a partner	955	54.6	744	48.4			1033	44.4
Divorced/separated	247	16.3	185	13.1			217	10.6
Widowed	423	21.0	616	34.1			971	40.1
Never married	105	8.1	59	4.4			67	4.9
Urbanicity									.520
Urban	789	43.5	756	44.7			1122	47.1
Suburban	430	23.8	374	23.4			514	23.1
Ex-urban	511	32.7	474	31.9			652	29.8
Income (quartile)									<.001
Q4	345	21.5	380	24.7			738	34.0
Q3	93	5.3	92	6.8			153	6.5
Q2	736	44.0	744	45.2			1002	43.1
Q1 (Poorest)	556	29.2	388	23.3			395	16.5
Death cause									0153
Cancer	418	24.7	452	28.5			610	27.5
CVD	629	36.5	514	31.8			784	34.0
Other	683	38.8	638	39.7			894	38.5
Death expected									<.001
Yes	771	43.6	1010	63.5			1510	66.1
No	959	56.4	594	36.5			778	33.9
Year of death
2008-2015	1154	62.7	951	55.0			1404	56.3
2016-2021	576	37.3	653	45.0			884	43.7

Abbreviations: NH, non-Hispanic; CVD, cardiovascular disease.

In unadjusted analyses, ACP documentation was consistently associated with less aggressive end-of-life care and greater use of hospice services (Table 2). After adjustment for sociodemographic and clinical covariates, having any ACP documentation was associated with significantly lower likelihood of life-sustaining treatment (aRR = 0.85, 95% CI: 0.74-0.98) and higher likelihood of hospice utilization (aRR = 1.43, 95% CI: 1.28-1.60) and out-of-hospital death (aRR = 1.11, 95% CI: 1.06-1.18) (Table 2). There was no statistically significant association between having any ACP documentation and ICU use. When the exposure was modeled ordinally, an additive pattern was observed, although differences between one and two documents were attenuated in adjusted models. Compared with having no ACP documentation, having two documents was associated with a lower likelihood of life-sustaining treatment (aRR = 0.84, 95% CI: 0.71-0.98), as well as higher likelihood of hospice utilization (aRR = 1.48, 95% CI: 1.32-1.66) and out-of-hospital death (aRR = 1.12, 95% CI: 1.05-1.19). Although point estimates generally suggested less aggressive care patterns among those with both documents compared with one document, these differences were modest and not consistently statistically significant after adjustment.

Table 2.

Prevalence and Relative Risks of End-of-Life Care Outcomes Among Decedents Using Survey-Weighted Modified Poisson Regression

Characteristics	Prevalence			P	Regression
Characteristics	%	95% CI		P	aRR^a	95% CI
Intensive care unit use (n = 4482, WEIGHTED n = 18 354 145)
ACP documentation (binary)				.004
None	51.7	48.6	54.8		1.00
Any (living will and/or DPOA)	46.4	44.4	48.3		1.03	0.96	1.10
ACP documentation (ordinal)				.008
None	51.7	48.6	54.8		1.00
Living will only	48.3	45.0	51.7		1.04	0.96	1.13
DPOA only
Two documents (living will and DPOA)	45.0	42.4	47.7		1.02	0.93	1.11
Life-support use (n = 4470, weighted n = 18 328 959)
ACP documentation (binary)				<.001
None	38.9	35.6	42.3		1.00
Any (living will and/or DPOA)	25.8	23.8	27.8		0.85	0.74	0.98
ACP documentation (ordinal)				<.001
None	38.9	35.6	42.3		1.00
Living will only	28.1	24.4	31.8		0.87	0.75	1.02
DPOA only
Two documents (living will and DPOA)	24.3	22.3	26.3		0.84	0.71	0.98
Hospice use before death (n = 3679, weighted n = 16 005 182)
ACP documentation (binary)				<.001
None	29.3	26.2	32.4		1.00
Any (living will and/or DPOA)	55.9	53.2	58.7		1.43	1.28	1.60
ACP documentation (ordinal)				<.001
None	29.3	26.2	32.4		1.00
Living will only	51.6	47.7	55.6		1.37	1.20	1.55
DPOA only
Two documents (living will and DPOA)	58.8	55.4	62.2		1.48	1.32	1.66
Out-of-hospital death (n = 5619, weighted n = 23 182 189)
ACP documentation (binary)				<.001
None	59.1	56.1	62.1		1.00
Any (living will and/or DPOA)	72.3	70.4	74.3		1.11	1.06	1.18
ACP documentation (ordinal)				<.001
None	59.1	56.1	62.1		1.00
Living will only	71.0	68.5	73.6		1.11	1.05	1.17
DPOA only
Two documents (living will and DPOA)	73.2	70.6	75.8		1.12	1.05	1.19

Abbreviation: ACP, advance care planning; aRR, adjusted risk ratio.

^aModels were adjusted for age at death, sex, race/ethnicity, educational attainment, marital status, urbanicity, income level, cause of death, expectedness of death, and year of death.

In stratified analyses (Table 3), associations between ACP documentation and end-of-life care varied by expectedness of death. Among decedents whose death was expected, ACP documentation were associated with lower likelihood of life-sustaining treatment and higher likelihood of hospice utilization and out-of-hospital death. Specifically, having both documents was associated with a 19% (aRR = 0.81, 95% CI: 0.66-0.99) lower likelihood of life-sustaining treatment, a 35% (aRR = 1.35, 95% CI: 1.21-1.51) higher likelihood of hospice use, and a 20% (aRR = 1.20, 95% CI: 1.10-1.30) higher likelihood of out-of-hospital death compared with having no ACP documentation. In contrast, among decedents with unexpected death, associations were generally attenuated, particularly for ICU use and out-of-hospital death. However, ACP documentation remained strongly associated with hospice utilization even in this group, with those having both documents demonstrating substantially higher likelihood of hospice use compared with those without documentation (aRR = 1.88, 95% CI: 1.29-2.74). Across both strata, there were no significant associations between ACP documentation and ICU use after adjustment.

Table 3.

Relative Risks of End-of-Life Care Outcomes Estimated via Modified Poisson Regression, Stratified by Death Expectedness

Characteristics	Expected Death			Unexpected Death
Characteristics	aRR^a	95% CI		aRR^a	95% CI
Intensive care unit use
ACP documentation (binary)
None	1.00			1.00
Any (living will and/or DPOA)	0.96	0.85	1.08	1.08	0.98	1.18
ACP documentation (ordinal)
None	1.00			1.00
Living will only	0.97	0.84	1.11	1.10	0.98	1.24
DPOA only
Two documents (living will and DPOA)	0.95	0.83	1.09	1.05	0.93	1.19
Life-support use
ACP documentation (binary)
None	1.00			1.00
Any (living will and/or DPOA)	0.81	0.68	0.96	0.86	0.74	1.00
ACP documentation (ordinal)
None	1.00			1.00
Living will only	0.81	0.67	0.99	0.90	0.75	1.07
DPOA only
Two documents (living will and DPOA)	0.81	0.66	0.99	0.82	0.67	1.00
Hospice use before death
ACP documentation (binary)
None	1.00			1.00
Any (living will and/or DPOA)	1.30	1.16	1.46	1.80	1.29	2.49
ACP documentation (ordinal)
None	1.00			1.00
Living will only	1.25	1.09	1.42	1.72	1.22	2.41
DPOA only
Two document (living will and DPOA)	1.35	1.21	1.51	1.88	1.29	2.74
Out-of-hospital death
ACP documentation (binary)
None	1.00			1.00
Any (living will and/or DPOA)	1.19	1.10	1.29	1.08	0.98	1.19
ACP documentation (ordinal)
None	1.00			1.00
Living will only	1.17	1.08	1.28	1.09	0.99	1.21
DPOA only
Two documents (living will and DPOA)	1.20	1.10	1.30	1.07	0.96	1.20

Abbreviation: ACP, advance care planning; aRR, adjusted risk ratio.

^aModels were adjusted for age at death, sex, race/ethnicity, educational attainment, marital status, urbanicity, income level, cause of death, and year of death.

Temporal trends indicated a substantial increase in any ACP documentation, either Living will or DPOA, and two documents over time (P < .05), with overall higher prevalence in the expected death group than unexpected death group (Figure 1). The prevalence of any ACP documentation remained relatively stable between 2008 and 2016 but increased significantly thereafter, reaching more than 86% by 2021. Joinpoint analysis identified a significant inflection point around 2014, after which the prevalence increased at an annual rate of approximately 2.7% (Figure 2). Similar patterns were observed for durable power of attorney and completion of both documents, with evidence of a reversal from declining or stable trends prior to approximately 2012 to significant increases thereafter. These findings suggest a growing uptake of advance care planning, particularly comprehensive documentation involving both treatment preferences and surrogate designation.

Figure 1.

Trends in advance care planning documentation (living wills and durable power of attorney [DPOA]) before death among decedents, 2008-2021.

Figure 2.

Jointpoint analysis of any advance care planning documentation (living wills and/or durable power of attorney [DPOA]), living will, DPOA, and two documents (living will and DPOA) before death among decedents, 2008-2021.

Discussion

In this nationally representative study of U.S. older adult decedents, we observed a substantial increase in ACP documentation over time, particularly after a joinpoint inflection around 2014. Having any ACP documentation was associated with more favorable end-of-life outcomes compared with having no documentation, though the incremental benefit of having both a living will and a DPOA relative to a single document was modest after covariate adjustment. In stratified analyses, associations with life-sustaining treatment and out-of-hospital death were more pronounced among individuals with expected death and attenuated among those with unexpected death; however, ACP documentation remained strongly associated with hospice utilization regardless of death expectedness. Together, these findings suggest that while ACP documentation is associated with less aggressive end-of-life care patterns, much of the observed benefit may be driven by broader processes underlying ACP rather than the accumulation of documents alone.

Temporal Trend in ACP Documentation

These findings are consistent with contemporary conceptual models that distinguish ACP documentation from the broader ACP process. Under this framework, living wills and durable powers of attorney function as tools that support communication, surrogate activation, and preference implementation rather than as independent determinants of care. The relatively modest incremental benefit observed for dual documentation suggests that the pathways linking ACP to end-of-life outcomes may depend more heavily on the quality of communication and preparedness generated by the ACP process than on the number of documents completed.

The observed increase in ACP documentation prevalence over time, particularly the acceleration identified by joinpoint analysis around 2014, likely reflects a confluence of policy, clinical, and cultural shifts that have elevated the prominence of advance care planning in routine care. The introduction of Medicare reimbursement for ACP discussions in 2016, alongside growing emphasis on patient-centered care and serious illness communication, may have incentivized clinicians to engage patients in these conversations and formalize documentation.^23,25 In addition, broader public awareness of end-of-life planning, driven by initiatives such as The Conversation Project and increased attention to goal-concordant care may have contributed to rising uptake. Despite these encouraging trends, prior work suggests that disparities in ACP completion persist across racial, socioeconomic, and educational groups,²² which is consistent with the sociodemographic gradients observed in our sample. Taken together, these findings suggest that while ACP is becoming more common, equitable access to and engagement in ACP remain important areas for intervention.

Modest Incremental Benefit of Having Both Documents

Although having both a living will and a durable power of attorney represents a more comprehensive form of ACP, our findings indicate that the incremental benefit of dual documentation (beyond having a single document) is modest after adjustment for sociodemographic and clinical factors. This pattern aligns with emerging conceptualizations of ACP as a dynamic, communication-driven process rather than a static documentation event.^27,36,37 The presence of any documentation may serve as a proxy for underlying patient engagement, preferences for less aggressive care, or prior discussions with clinicians and family members, which may be the primary drivers of observed differences in end-of-life outcomes. In this context, the marginal gains associated with additional documentation may reflect diminishing returns when preferences are already known or when surrogate decision-makers are sufficiently prepared. These findings are consistent with prior studies demonstrating that ACP’s effectiveness depends not only on the existence of documents but also on their integration into clinical workflows and real-time decision-making.^10,17,38 Accordingly, interventions focused solely on increasing documentation rates may have a limited impact unless paired with efforts to enhance communication quality, surrogate preparedness, and accessibility of ACP information at the point of care.

Variation by Expectedness of Death

The stronger associations between ACP documentation and life-sustaining treatment, hospice utilization, and out-of-hospital death observed among decedents with expected death are consistent with the premise that anticipated dying trajectories create greater opportunity for iterative discussions, prognostic awareness, and alignment of care with patient preferences.^39,40 In cases of unexpected death, such as acute events or rapid clinical deterioration, the window for ACP to influence care may be limited, and decisions may be driven more by emergent clinical circumstances than by previously documented preferences. A notable exception to the general attenuation pattern was hospice utilization, for which ACP documentation showed its strongest association in those with unexpected death. One interpretation is that among decedents whose deaths were expected, hospice utilization is relatively common regardless of documentation status, leaving less room for documentation to differentiate care patterns. In contrast, when death is unexpected, documented preferences or a designated surrogate may function as a more decisive signal that redirects care toward hospice even in acute clinical situations where extended deliberation is not feasible. Our findings suggest that the mechanisms through which ACP documentation influences hospice referral differently from other end-of-life care decisions; surrogate activation and the documentation signaling may be more important than preference specification alone when death is not expected.⁴¹ Further investigation of these pathway-specific effects is warranted.

Policy Implications

Our findings suggest that future ACP initiatives should move beyond document completion as the primary marker of success. While recent policy efforts such as Medicare reimbursement for advance care planning have likely contributed to increased documentation,^23,39,42,43 our results suggest that expanding documentation alone may be insufficient to meaningfully improve end-of-life care. Policies should therefore move beyond document-completion metrics and promote systems that support ongoing communication about patient goals, values, and treatment preferences across care teams and care transitions. Ensuring that ACP information is routinely reviewed, updated, and accessible across settings may be particularly important for translating preferences into care decisions during serious illness.

Our findings also highlight persistent sociodemographic differences in ACP documentation. Targeted outreach and education efforts focused on historically underserved populations, including racial and ethnic minority groups and individuals with lower educational attainment, may help reduce disparities in ACP engagement. In addition, clinician-facing interventions, including structured communication training, decision aids, and technology-supported ACP programs, have demonstrated success in improving ACP engagement and documentation and may represent scalable approaches for health systems seeking to improve goal-concordant care. Collectively, these strategies may enhance patient-centered decision-making, improve care quality, and promote more efficient use of healthcare resources near the end of life.

Although documentation was associated with less aggressive end-of-life care and greater hospice utilization, the modest incremental benefit of having multiple documents indicates that communication and decision-making processes may be more important than documentation quantity alone. Health systems should prioritize interventions that improve communication among patients, families, surrogate decision-makers, and clinicians, while ensuring that ACP information is readily accessible across care settings. Future research should identify which components of the ACP process—including surrogate preparedness, communication quality, and timing of discussions—most effectively promote goal-concordant care.

Limitations

These findings should be interpreted in light of the following limitations. First, the use of proxy-reported exit interview data may introduce recall bias or misclassification, particularly for ACP documentation and end-of-life care experiences. Although proxy reports are widely used in end-of-life research and are often the only feasible source of information after death, their accuracy may vary depending on the respondent’s knowledge and involvement in care. Second, our measures of ACP documentation capture the presence of living wills and durable power of attorney but do not reflect the content, quality, timing, or accessibility of these documents, nor whether they were available or adhered to at the point of care. As a result, we were unable to assess whether care was truly goal concordant. Third, outcomes were measured as broad indicators of healthcare utilization and may not fully capture the complexity of end-of-life decision-making or the appropriateness of care in specific clinical contexts. Fourth, stratification by expectedness of death was based on proxy assessment and may be subject to misclassification, particularly in cases with uncertain illness trajectories. Finally, the observational study design precludes causal inference. Individuals with ACP documentation may differ systematically from those without in ways not fully captured by measured covariates, such as health literacy, preferences for care, or engagement with the healthcare system, potentially leading to residual confounding. Despite these limitations, this study leverages a large, nationally representative dataset and robust analytic methods to provide new insights into the role of ACP documentation components in shaping end-of-life care.

Conclusions

Comprehensive ACP documentation was associated with less aggressive end-of-life care and greater hospice utilization among U.S. older adults, although the additional benefit of having both a living will and durable power of attorney was modest. These findings suggest that ACP effectiveness may depend less on the accumulation of documents and more on the communication, surrogate preparation, and decision-making processes that documentation facilitates. Efforts to improve end-of-life care should therefore focus on strengthening the quality and implementation of ACP conversations, as well as on promoting documentation completion.

Footnotes

ORCID iD

Molly Jacobs

Ethical Considerations

The HRS is a publicly available, de-identified dataset, the Institutional Review Board (IRB) determined that this study was non-human subject research.

Author Contributions

ZX: Conceptualization; methodology; software; formal analysis; original draft; writing – review and editing.

MJ: Conceptualization; original draft; supervising; writing – review and editing; corresponding author.

JL: Conceptualization; writing – review and editing.

BP: Conceptualization; writing – review and editing.

YH: Conceptualization; methodology; supervising; writing – review and editing; senior author.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

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

Data Availability Statement

The data supporting the findings of this study are openly available on the Health and Retirement Study website. These data were derived from the following resources available in the public domain: .

References

McMahan

Tellez

Sudore

. Deconstructing the complexities of advance care planning outcomes: what do we know and where do we go? A scoping review. J Am Geriatr Soc. 2021;69(1):234-244.

Murray

Butow

. Advance care planning in motor neuron disease: a systematic review. Palliat Support Care. 2016;14(4):411-432.

Fried

Bullock

Iannone

O'leary

. Understanding advance care planning as a process of health behavior change. J Am Geriatr Soc. 2009;57(9):1547-1555.

Sudore

Lum

You

, et al. Defining advance care planning for adults: a consensus definition from a multidisciplinary Delphi panel. J Pain Symptom Manag. 2017;53(5):821-832. e1.

Sedini

Biotto

Crespi Bel’skij

Moroni Grandini

Cesari

. Advance care planning and advance directives: an overview of the main critical issues. Aging Clin Exp Res. 2022;34(2):325-330.

Rosca

Karzig-Roduner

Kasper

Rogger

Drewniak

Krones

. Shared decision making and advance care planning: a systematic literature review and novel decision-making model. BMC Med Ethics. 2023;24(1):64.

Hamilton

. Advance care planning in general practice: promoting patient autonomy and shared decision making. Br J Gen Pract. 2017;67(656):104-105.

Kirkland

Tran

Aziz

Banerjee

Kuehl

. Advanced clinical practitioner-led palliative care heart failure model within the national health service: a prospective study of patient experience, healthcare costs, and end-of-life outcomes. Eur J Cardiovasc Nurs. 2026;zvag058. https://doi.org/10.1093/eurjcn/zvag058

Kunonga

Thomson

Beyer

Dawson

Brittain

Craig

. A mixed methods systematic review of the effectiveness, acceptability, and feasibility of advanced care plans for older adults without life-limiting conditions. Journal of Long-Term Care. 2024;383-402.

10.

Wolff

Scerpella

Giovannetti

, et al. Advance care planning, end-of-life preferences, and burdensome care: a pragmatic cluster randomized clinical trial. JAMA Intern Med. 2025;185(2):162-170.

11.

Ashana

Chen

Agiro

, et al. Advance care planning claims and health care utilization among seriously ill patients near the end of life. JAMA Netw Open. 2019;2(11):e1914471.

12.

Crooks

Rizk

Simpson-Greene

, et al. Evaluating outcomes of advance care planning interventions for adults living with advanced illness and people close to them: a systematic meta-review. Palliat Med. 2025;39(8):02692163251344428.

13.

Song

Amatya

Voutier

Khan

. Advance care planning in patients with primary malignant brain tumors: a systematic review. Front Oncol. 2016;6:223.

14.

Walczak

Butow

Clayton

. A systematic review of evidence for end-of-life communication interventions: who do they target, how are they structured and do they work? Patient Educ Counsel. 2016;99(1):3-16.

15.

Silveira

Kim

Langa

. Advance directives and outcomes of surrogate decision making before death. N Engl J Med. 2010;362(13):1211-1218.

16.

Wright

Zhang

Ray

, et al. Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA. 2008;300(14):1665-1673.

17.

Jimenez

Tan

Virk

Low

Car

AHY

. Overview of systematic reviews of advance care planning: summary of evidence and global lessons. J Pain Symptom Manag. 2018;56(3):436-459. e25.

18.

Pimsen

Kao

Hsu

Shu

. The effect of advance care planning intervention on hospitalization among nursing home residents: a systematic review and meta-analysis. J Am Med Dir Assoc. 2022;23(9):1448-1460. e1.

19.

Wang

Zhou

Peng

, et al. Effects of advance care planning on end-of-life decisions among community-dwelling elderly people and their relatives: a systematic review and meta-analysis. Ann Palliat Med. 2023;12(3):57183-57583.

20.

Zhu

Enguidanos

. Advance directives completion and hospital out‐of‐pocket expenditures. J Hosp Med. 2022;17(6):437-444.

21.

Wang

Huang

Zhang

Wang

Sun

Tao

. Association between advance directives and invasive life support use at end of life: a retrospective cohort study. J Pain Symptom Manag. 2026;72(1):75-80. doi:10.1016/j.jpainsymman.2026.02.017.

22.

Yadav

Gabler

Cooney

, et al. Approximately one in three US adults completes any type of advance directive for end-of-life care. Health Aff. 2017;36(7):1244-1251.

23.

Luth

Manful

Weissman

, et al. Practice billing for Medicare advance care planning across the USA. J Gen Intern Med. 2022;37(15):3869-3876.

24.

Gupta

Jin

Reich

, et al. Association of billed advance care planning with end‐of‐life care intensity for 2017 Medicare decedents. J Am Geriatr Soc. 2020;68(9):1947-1953.

25.

Mehta

Kelley

. Advance care planning codes—getting paid for quality care. JAMA Intern Med. 2019;179(6):830-831.

26.

Dingfield

Kayser

. Integrating advance care planning into practice. Chest. 2017;151(6):1387-1393.

27.

Hickman

Lum

Walling

Savoy

Sudore

. The care planning umbrella: the evolution of advance care planning. J Am Geriatr Soc. 2023;71(7):2350-2356.

28.

Wang

Huang

Wang

Liao

. Advance care planning for frail elderly: are we missing a golden opportunity? A mixed-method systematic review and meta-analysis. BMJ Open. 2023;13(5):e068130.

29.

Juster

Suzman

. An overview of the health and retirement study. J Hum Resour. 1995;30:S7-S56.

30.

Fisher

Ryan

. Overview of the health and retirement study and introduction to the special issue. Work, Aging and Retirement. 2018;4(1):1-9.

31.

Health and Retirement Study . RAND HRS Longitudinal File 2022 public use dataset. Produced and distributed by the University of Michigan with funding from the National Institute on Aging (grant number NIA U01AG009740). Ann Arbor, MI; 2022.

32.

Palmer

Jacobson

Enguidanos

. Advance care planning for Medicare beneficiaries increased substantially, but prevalence remained low: study examines Medicare outpatient advance care planning claims and prevalence. Health Aff. 2021;40(4):613-621.

33.

Virdun

Luckett

Davidson

Phillips

. Dying in the hospital setting: a systematic review of quantitative studies identifying the elements of end-of-life care that patients and their families rank as being most important. Palliat Med. 2015;29(9):774-796.

34.

Lin

Yang

Zhang

. Statistics and pitfalls of trend analysis in cancer research: a review focused on statistical packages. J Cancer. 2020;11(10):2957-2961.

35.

Zou

. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702-706.

36.

Sudore

Boscardin

Barnes

. A patient-facing advance care planning (ACP) website called PREPARE increases ACP documentation and engagement in a randomized trial of diverse older primary care patients at a VA medical center (TH307B). J Pain Symptom Manag. 2017;53(2):317-318.

37.

Hickman

Lum

Unroe

. Family caregiver perspectives on advance care planning discussions for residents with dementia led by trained nursing home staff: insights from the APPROACHES project. J Am Med Dir Assoc. 2026;27:106154.

38.

Jimenez

Tan

Virk

Low

Car

AHY

. State of advance care planning research: a descriptive overview of systematic reviews. Palliat Support Care. 2019;17(2):234-244.

39.

Ladin

Bronzi

Gazarian

, et al. Understanding the use of Medicare procedure codes for advance care planning: a national qualitative study: Study examines the use of Medicare procedure codes for advance care planning. Health Aff. 2022;41(1):112-119.

40.

Vellani

Green

Kulasegaram

Sussman

Wickson-Griffiths

Kaasalainen

. Interdisciplinary staff perceptions of advance care planning in long-term care homes: a qualitative study. BMC Palliat Care. 2022;21(1):127.

41.

Prater

Wickizer

Bower

Bose-Brill

. The impact of advance care planning on end-of-life care: do the type and timing make a difference for patients with advanced cancer referred to hospice? Am J Hosp Palliat Med. 2019;36(12):1089-1095.

42.

Weissman

Gazarian

Reich

, et al. Recent trends in the use of Medicare advance care planning codes. J Palliat Med. 2020;23(12):1568-1570.

43.

Belanger

Loomer

Teno

Mitchell

Adhikari

Gozalo

. Early utilization patterns of the new Medicare procedure codes for advance care planning. JAMA Intern Med. 2019;179(6):829-830.