JAMA October 07, 2026

Advance Care Planning for Black and White Older Adults With Serious Illness

Kimberly S. Johnson et al.

Bottom Line

In a cluster randomized trial of older adults with serious illness, a facilitated advance care planning intervention was not significantly more effective than a self-guided approach within Black or White subgroups, though an interaction effect suggested the facilitated approach was relatively more effective for Black participants compared to White participants.

Key Findings

1. Among Black participants, advance care planning (ACP) completion within 1 year was 71% (152/213) in facilitated clinics compared to 62% (131/213) in self-guided clinics, which was not a statistically significant difference (adjusted RR, 1.15 [95% CI, 0.83-1.59]).

2. Among White participants, ACP completion was 68% (133/196) in facilitated clinics versus 78% (131/167) in self-guided clinics, also showing no significant difference (adjusted RR, 0.88 [95% CI, 0.65-1.18]).

3. In the overall pooled sample of 789 participants, 547 (69%) completed ACP, with 130/789 completing formal ACP and 523/668 completing informal ACP.

4. A significant interaction was observed between race and the intervention group (adjusted RR, 1.32 [95% CI, 1.03-1.69]), indicating that the facilitated approach was more effective than the self-guided approach among Black relative to White participants.

5. Racial concordance between the participant and the ACP facilitator was not significantly associated with ACP completion in either racial group.

Study Design

Design

Cluster RCT

Open-Label

Sample

791

Patients

Duration

12 mo

Median

Setting

Multicenter, US South

Population Community-dwelling Black and White adults 65 years or older with serious illness.

Intervention Respecting Choices First Steps (a facilitated advance care planning approach delivered by lay facilitators) with randomization to a racially concordant or discordant facilitator.

Comparator Five Wishes (a self-guided advance care planning approach delivered by lay facilitators) with randomization to a racially concordant or discordant facilitator.

Outcome Advance care planning (ACP) completion within 12 months, defined as new formal documentation (advance directive, clinician note, or POLST) and/or new informal ACP (participant-reported discussion with family).

Study Limitations

• The inclusion of informal ACP—defined as participant-reported discussions with family—relies on self-reporting, which may introduce recall or social desirability bias.

• The study was conducted across 10 clinics exclusively in the Southern United States, which may limit the generalizability of the findings to older adults in other geographic regions or healthcare systems.

• Although a significant interaction effect was noted, the trial may have been underpowered to detect smaller, yet clinically meaningful differences in ACP completion between the intervention arms within each specific racial subgroup.

Clinical Significance

The EQUAL ACP trial demonstrates that structured advance care planning interventions can achieve high completion rates (nearly 70%) among both Black and White older adults with serious illness. Although a facilitated approach did not outright beat a self-guided approach within either racial subgroup independently, the significant racial interaction indicates that Black patients may derive greater relative benefit from facilitated ACP than White patients. This highlights the clinical importance of tailoring ACP delivery strategies to demographic groups to maximize engagement and help bridge known racial disparities in end-of-life care planning.

Historical Context

Historically, Black adults in the United States have had significantly lower rates of advance care planning (ACP) participation compared to White adults. This disparity has contributed to well-documented differences in end-of-life care, including lower utilization of hospice services and higher rates of aggressive, often non-beneficial, interventions at the end of life. Previous ACP interventions have largely been applied universally without examining their differential effectiveness across racial lines. The EQUAL ACP trial (NCT03516994) was specifically designed to evaluate whether a culturally sensitive, facilitated ACP approach could better overcome these barriers compared to a standard self-guided approach, and to rigorously test the impact of racial concordance between facilitators and patients.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

What are the foundational components of advance care planning (ACP), and what historical and sociocultural factors might contribute to differences in how Black and White older adults engage with self-guided versus facilitated ACP interventions?

Key Response

ACP involves identifying a surrogate decision-maker and articulating values for end-of-life care. Students must understand that medical mistrust stemming from historical injustices, differing spiritual beliefs regarding end-of-life, and systemic healthcare inequities heavily influence ACP engagement. These factors explain why simply providing self-guided materials may be insufficient, making culturally sensitive, relationship-based facilitated approaches critical for certain populations.

Resident

Given the study's finding that facilitated advance care planning showed a relatively greater impact on Black patients compared to self-guided methods, how should this influence your approach to goals-of-care conversations when admitting a Black older adult with a serious illness?

Key Response

Residents should recognize that merely providing a standardized form or pamphlet is often inadequate for meaningful ACP. Taking the time for a facilitated, relationship-based discussion to explore patient values, address potential mistrust, and clarify medical information is an essential clinical skill, especially for populations that experience higher baseline systemic barriers to healthcare engagement.

Fellow

The study identified a significant interaction effect suggesting facilitated ACP is relatively more beneficial for Black patients compared to White patients, despite null findings within the isolated subgroups. In the context of a palliative care consult service, how do you interpret this differential treatment response, and how might it influence resource allocation for ACP navigators?

Key Response

Fellows must grapple with heterogeneous treatment effects. The interaction suggests that White patients might reach a 'ceiling effect' with self-guided tools due to higher baseline health literacy or systemic trust, whereas Black patients specifically benefit from the relational aspect of facilitated ACP. This justifies directing scarce navigator and social work resources toward populations facing structural barriers, maximizing the impact of the intervention.

Attending

As a clinic director implementing an ACP program, how does the interaction effect demonstrated in this trial justify a targeted, equity-focused approach to patient outreach rather than a universal 'one-size-fits-all' model?

Key Response

Attendings must make system-level workflow decisions. Universal interventions often inadvertently widen disparities if advantaged groups utilize them more effectively. This study supports 'targeted universalism' or equity-focused resource allocation, demonstrating that intensive, facilitated models yield a relatively higher return on investment for minoritized populations who are less served by default, low-resource self-guided materials.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

In a cluster randomized trial where main subgroup effects are non-significant but the interaction term (Intervention x Race) is significant, what are the methodological challenges regarding statistical power and intracluster correlation coefficient (ICC) estimation, and how should future trials be powered to definitively test these subgroup interactions?

Key Response

Interaction tests are notoriously underpowered in clinical trials. A significant interaction despite null subgroup main effects suggests the study was likely only powered for the overall cohort main effect. PhD researchers must account for the ICC within clinics or providers when calculating sample sizes for interaction effects, which often requires exponentially larger sample sizes to avoid Type II errors in subgroup analyses.

Journal Editor

When reviewing a manuscript claiming a significant interaction effect by race in a cluster-randomized trial, what critical confounders related to clinic-level characteristics (e.g., clinic resources, geographic location, patient socioeconomic status) must be scrutinized to ensure the interaction is driven by patient demographics rather than the specific clusters to which they belong?

Key Response

Editors must look out for ecological fallacies or confounding by cluster. If Black participants were disproportionately clustered in under-resourced clinics where self-guided tools failed universally due to lack of staff follow-up, the interaction might reflect clinic-level structural disparities rather than racial differences in response to facilitation. Thoroughly controlling for cluster-level covariates is essential for the validity of the interaction claim.

Guideline Committee

Current guidelines broadly recommend ACP for all older adults with serious illness but rarely specify the delivery modality based on patient demographics. Does the evidence of a differential racial response to facilitated versus self-guided ACP warrant an update to geriatric and palliative care guidelines to explicitly recommend equity-tailored ACP pathways?

Key Response

Guideline committees must weigh whether a secondary interaction effect in one trial is sufficient to change broad practice guidelines. While current guidelines emphasize universal ACP, this trial suggests a need for equity-tailored pathways. The committee might consider issuing a moderate recommendation to prioritize high-touch, facilitated ACP modalities for populations with known historical disparities, moving away from a uniform standard of care that relies on self-guided advance directives.

Clinical Landscape

Noteworthy Related Trials

1995

SUPPORT Trial

n = 9,105 · JAMA

Tested

Nurse-mediated communication to improve end-of-life care

Population

Hospitalized patients with life-threatening illnesses

Comparator

Usual care

Endpoint

Incidence of do-not-resuscitate (DNR) orders and physician knowledge of patient preferences

Key result: The intervention failed to improve end-of-life care or patient-physician communication, showing that simple prognostic feedback is insufficient.

2010

Early Palliative Care in NSCLC Trial

n = 151 · NEJM

Tested

Early integration of palliative care alongside standard oncology care

Population

Patients with newly diagnosed metastatic non-small-cell lung cancer

Comparator

Standard oncology care alone

Endpoint

Change in quality of life at 12 weeks

Key result: Early palliative care improved quality of life, led to less aggressive end-of-life care, and prolonged median survival.

2018

PREPARE for Your Care Trial

n = 414 · JAMA Intern Med

Tested

Interactive PREPARE website plus easy-to-read advance directive

Population

Diverse older adults with serious chronic illness

Comparator

Easy-to-read advance directive alone

Endpoint

New advance care planning documentation

Key result: The PREPARE intervention significantly increased advance care planning documentation and patient engagement compared to the directive alone.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis