“Don’t Talk to Them About Goals of Care”: Understanding Disparities in Advance Care Planning

Abstract

Background

Structurally marginalized groups experience disproportionately low rates of advance care planning (ACP). To improve equitable patient-centered end-of-life care, we examine barriers and facilitators to ACP among clinicians as they are central participants in these discussions.

Method

In this national study, we conducted semi-structured interviews with purposively selected clinicians from 6 diverse health systems between August 2018 and June 2019. Thematic analysis yielded themes characterizing clinicians’ perceptions of barriers and facilitators to ACP among patients, and patient-centered ways of overcoming them.

Results

Among 74 participants, 49 (66.2%) were physicians, 16.2% were nurses, and 13.5% were social workers. Most worked in primary care (35.1%), geriatrics (21.1%), and palliative care (19.3%) settings. Clinicians most frequently expressed difficulty discussing ACP with certain racial and ethnic groups (African American, Hispanic, Asian, and Native American) (31.1%), non-native English speakers (24.3%), and those with certain religious beliefs (Catholic, Orthodox Jewish, and Muslim) (13.5%). Clinicians were more likely to attribute barriers to ACP completion to patients (62.2%), than to clinicians (35.1%) or health systems (37.8%). Three themes characterized clinicians’ difficulty approaching ACP (preconceived views of patients’ preferences, narrow definitions of successful ACP, and lack of institutional resources), while the final theme illustrated facilitators to ACP (acknowledging bias and rejecting stereotypes, mission-driven focus on ACP, and acceptance of all preferences).

Conclusions

Most clinicians avoided ACP with certain racial and ethnic groups, those with limited English fluency, and persons with certain religious beliefs. Our findings provide evidence to support development of clinician-level and institutional-level interventions and to reduce disparities in ACP.

Advance care planning (ACP) promotes patient autonomy by allowing patients to express their preferences for medical care in the event of serious illness. ACP is an essential component of patient-centered end-of-life care and has been associated with patient and family satisfaction and hospice utilization (1–5). Yet, few patients report discussing ACP, and only 1 in 3 adults documents their wishes in advance directives (6–9). Black, Indigenous, and people of color (BIPOC), persons with low income, and those with limited health literacy experience even lower rates of ACP completion (10–15). We will use the term structurally marginalized or minoritized to highlight the role of institutions in creating and maintaining power imbalances and to combat the notion that disparities are inherent, fixed, or natural. In a national sample, Black and Hispanic patients were 70% less likely than non-Hispanic White patients to have completed an advance directive (16). Ensuring high-quality end-of-life care for BIPOC and other structurally marginalized groups is a moral imperative, as well as an urgent national priority as the United States is projected to become a “majority–minority” nation (ie, <50% of the population will identify as White) in the coming decades (17).

Understanding the reasons for lower ACP among structurally marginalized groups is critical for improving equitable access to care (18). Despite the impact of clinicians and institutions in availing ACP to patients, to date, most studies have focused on patient factors, such as religiosity and distrust, to explain differential rates of ACP, and these have been limited to a single center or region (10,11,17,19–23). However, the role of clinicians has not been well investigated, despite evidence of clinician factors contributing to other health care disparities (24–27). Moreover, studies examining how clinicians decide whether to engage structurally marginalized patients in ACP discussions have largely been limited to palliative care providers (14,17). Gaps remain in understanding which groups of patients may be excluded from ACP discussions, why clinicians avoid engaging some patients in ACP conversations, and factors assisting clinicians in overcoming these barriers.

Using a national sample of clinicians from multiple care settings and specialties, we sought to identify patient populations that are less often approached for ACP alongside clinician-level barriers and facilitators. Improving access to ACP for structurally marginalized populations is a necessary step to ensuring patient autonomy and equity in end-of-life care.

Method

This qualitative study employed case study methodology to examine ACP practices across a range of health systems using semi-structured interviews with clinicians (28,29).

Participant Selection

Two investigators (K.L. and J.W.) conferred with an advisory panel consisting of leaders in palliative care and health services research to purposively sample health systems based on national reputation for ACP, size, region, type (eg, academic, public), and religious affiliation. At each site, clinicians were identified in collaboration with site champions (Chief Medical Officer or a senior palliative care physician), and purposively sampled based on role (eg, physician, nurse), specialty, and experience with ACP. Interviews continued until thematic saturation was confirmed through deliberation (30).

Data Collection

A social scientist with qualitative methods and palliative care expertise (K.L.) guided development of a semi-structured interview guide based on an appraisal of the literature (31). Open-ended questions probed clinicians’ approaches to ACP among all patients, characterization of populations with whom clinicians had difficulty discussing ACP or avoided discussing ACP, if any, and barriers and facilitators to ACP among structurally marginalized patients (Supplementary Figure 1).

Clinicians were recruited by e-mail. Between August 2018 and June 2019, investigators (D.C.A, P.K.G., A.G., S.P., A.J.R., J.T., J.S.W., or K.L.) with formal training in qualitative interviewing conducted interviews at the participant’s workplace or by phone and took field notes. All interviews were audio recorded and transcribed. Participants also completed a demographic survey (Supplementary Figure 2). Participants provided verbal consent to being recorded and taking the survey. This study was approved by the Partners HealthCare Institutional Review Board.

Analysis

Transcripts were uploaded to NVivo 11 (QSR International, Melbourne, Australia). D.C.A. and K.L. created a preliminary codebook based on the questioning structure of the interview guide (32). D.C.A. and N.D. iteratively revised the codebook through joint deliberation after independently coding 14 transcripts (18.9% of total transcripts) and allowing new codes to emerge inductively (32). K.L. served as an arbiter. The codebook was finalized after intercoder agreement of over 90% was reached. All interviews were coded using the final codebook through a thematic analysis approach (30). Codes were then iteratively organized into themes to capture the full range of narratives and counternarratives. All transcripts were reviewed again to categorize identified barriers to ACP as patient-, clinician-, or health system-level barriers. Findings were shared with site champions (32). This study follows the Consolidated Criteria for Reporting Qualitative Health Research (Supplementary Figure 3) (33).

Results

The final sample included 74 clinicians from 6 health systems across 4 states (Table 1). Among these, 66.2% were physicians, 16.2% were nurses, 13.5% were social workers, and 4.1% were chaplains. Their mean age was 49.0 years (SD 10.4), 54.5% were female, and they had an average of 20.0 years (SD 11.7) of professional health care experience. They most commonly worked in primary care (35.1%), geriatrics (21.1%), and palliative care (19.3%) settings (Table 1).

Table 1.

Characteristics of Included Health Systems and Clinicians

Health system characteristics
Health systems, no.	6
Region, no.
West	3
South	2
Northeast	1
Type, no.
Academic	4
Public	1
Other	1
Religious affiliation, no.
Yes	2
No	4
Bed count, mean (SD)	1650.8 (1480.0)
Total physician count, mean (SD)	1526.8 (819.2)
Clinician characteristics
Clinicians, no.	74
Age, mean (SD)a	49.0 (10.4)
Gender, N (%)a
Female	36 (45.5)
Male	30 (54.5)
Years in practice, mean (SD)a	20.0 (11.7)
Location of clinical training, N (%)a
United States	64 (98.5)
Other	1 (1.5)
Type, N (%)
Physician	49 (66.2)
Nurse	12 (16.2)
Social worker	10 (13.5)
Chaplain	3 (4.1)
Specialty, N (%)a
Primary care	20 (35.1)
Geriatrics	12 (21.1)
Palliative care	11 (19.3)
Other	18 (31.6)
Hispanic ethnicity, N (%)a	3 (4.5)
Race, N (%)a
White	61 (92.4)
Asian	5 (7.6)
Black	1 (1.5)
Languages spoken, N (%)a
English only	50 (73.5)
English and another language	18 (26.5)
Spanish	15 (22.1)
Other	7 (10.3)
Interview type, N (%)
In-person	49 (77.0)
Phone	25 (23.0)
Interview time, mean (SD)	41.0 (10.3)

Note: ^aAge, gender, years in practice, location of clinical training, specialty, Hispanic ethnicity, race, and languages spoken were missing for 13, 8, 7, 9, 17, 8, 8, and 6 clinicians, respectively.

Clinicians most frequently reported avoiding ACP discussions with patients identified as BIPOC (specifically Black, Hispanic, Asian, and Native American patients) (31.1%), non-English or non-native English speakers (24.3%), those with certain religious beliefs (Catholic, Orthodox Jewish, and Muslim) (13.5%), low health literacy (12.2%), and low income (8.1%). Clinicians were more likely to attribute barriers to ACP completion to patients (62.2%) than to clinicians (35.1%) or to health systems (37.8%) (Table 2), although some attributed barriers to more than one group.

Table 2.

Clinician-Identified Barriers to Advance Care Planning Among Structurally Marginalized Patients

Barrier	Number of Clinicians, N (%)
Patient-level barriers	46 (62.2)
Reluctance to discuss ACP due to religious, cultural, or personal beliefs	21 (28.4)
Distrust of clinicians or health systems	19 (25.7)
Deference to family for decision making	10 (13.5)
Low health literacy	9 (12.2)
Lack of capacity to make medical decisions	5 (6.8)
No available surrogates	4 (5.4)
Clinician-level barriers	26 (35.1)
Implicit bias	12 (16.2)
Discomfort discussing ACP with structurally marginalized patients	9 (12.2)
Limited time to discuss ACP due to medical complexity	5 (6.8)
Health system-level barriers	28 (37.8)
Insufficient language interpretation services	23 (31.1)
Limited health system access due to underinsurance	4 (5.4)
Insufficient workforce diversity	1 (1.4)

Note: ACP = advance care planning.

Four themes emerged characterizing clinicians’ approaches to ACP among structurally marginalized patients. Three of the themes describe barriers to completing ACP (preconceived views of patients’ preferences, narrow definitions of successful ACP, and insufficient institutional resources), while the fourth theme identified key facilitators from clinicians who regularly overcame barriers described via a tailored and patient-centered approach (Table 2).

Barriers to ACP With Structurally Marginalized Populations

Preconceived views of patient preferences

Clinicians’ preconceived views about discussing ACP with structurally marginalized patients resulted in their avoiding ACP altogether or undertaking cursory explorations of these patients’ care goals. One said (female [F], physician), “[In] non-White cultures… there is this idea that if you talk about death, it’s going to make it happen.” Another participant (F) remarked, “In… Hispanic culture… you can’t tell grandma that she has cancer because she’ll give up and die.” A physician (male [M]) shared, “Patients who grew up more grounded in religion… Catholicism for example, might say… ‘I don’t want to talk about this because God will take me when God is ready’.” Another (F, physician) remarked, “It is uncomfortable… to be accused, ‘You think you can undo God’s will’. If you get that often enough, you subconsciously are going to avoid that conversation again.” A female physician said, “We see people from many countries. We don’t know their practices. We might offend someone by asking.”

Others discussed their beliefs about patients’ distrust of the health care system. One physician (F) shared, “I love my [African American] patients, but the minute that I bring up [dying], I feel this… hit the brakes, and I don’t know if it’s my own insecurities, but… it brings up… not trusting me.” Another (F, physician) shared, “Providers are worried about ACP with non-White patients because of… trust. And ironically they do it less because of that.”

A participant (F) even suggested that some clinicians may overtly discriminate against Latinx patients, “[Clinicians] seem more interested in helping your typical White person. [Hispanics] tended to be assumed to not want to complete directives. I could tell [patients] were being blown off.” A physician (F) similarly shared, “What you are taught about African American families is that you don’t talk to them about their goals of care.” Another (F, physician) stated, “If I have an African American woman who obviously has a strong foothold in the church, I will bring [ACP] up, but… in another normal patient I would say, ‘Are you sure you want CPR (cardiopulmonary resuscitation)? Let’s talk about what that would actually look like for you and why I might recommend against that’. I might not be as deep into that conversation [with the African American woman]. I say, ‘ Yes, ma’am’. And… and we move on.”

Narrow Definitions of Successful ACP

Many clinicians defined successful ACP by completion of forms, including specifying a surrogate decision maker and documenting medical directives. A physician (M) said, “Part of my success with ACP conversations is… that I can do them really fast. I don’t think it has to be [a conversation]. I can do an MDPOA (medical durable power of attorney) using [my state’s] form in 30 seconds.” A participant (F) explained, “Our objective is to check the box, have a document that we can scan into the medical record.”

Yet, these narrow definitions were often discordant with patients’ goals for ACP. Clinicians reported that many structurally marginalized patients preferred to cede decision making to family, receive fully restorative care, and decline to complete legal documents, which often led clinicians to conclude that ACP efforts were futile. For example, one physician (M) said, “I have [Spanish-speaking] patients who… don’t want to burden their family. The decision to be DNR (do-not-resuscitate)… is counter to what their family’s expectation is for them. So they will choose to be full CPR. That is bewildering to me because it doesn’t seem consistent with their values.” A clinician (F) discussed documentation of care wishes, “[Hispanic patients] think when they sign that piece of paper… they no longer have a choice and it’s kind of like a sign of giving up.” Another (F, physician) described the conflict between family and clinician goals, “[The patient’s wife] was very frustrated because… the [doctors] keep bringing the same things up, and she is very clear of her goals. And they are different from what… the [doctors] want her husband’s goals to be. It’s about cultural differences… I know spirituality was a large driver of what her preferences were.”

Insufficient Institutional Resources

Many clinicians cited insufficient institutional resources as a barrier to engaging structurally marginalized populations in ACP. They reported lack of in-person interpreters with training in discussing ACP, and limited availability of ACP documents in multiple languages or designed for users with limited health literacy. One clinician (F) said, “We have… an iPad virtual translator. Sometimes the volume isn’t loud enough. It’s hard to read the body language. We used to have in-person translators, and that was… better.” Another physician (M) remarked, “There’s a state website on ACP. But that’s not available in Spanish, and half of our patients are Spanish speaking.” A participant (F) shared, “In Texas, we have a lot of Spanish-speaking folks, but you [cannot] print the DNR in Spanish. Any legal document has to be in English.” Regarding an ACP website, a physician (F) shared, “We haven’t found it terribly useful… it requires a pretty high health literacy.” Other clinicians discussed limited availability of social workers and chaplains. A participant (M) said, “It would be great to consult [a chaplain]… but pastoral care doesn’t come over to the outpatient world. They just don’t have the funding.”

Facilitators to ACP With Structurally Marginalized Populations

Clinical strategies to improve equitable access to ACP

Acknowledging bias and rejecting stereotypes.

—Few clinicians described a patient-centered and inclusive approach to ACP. These clinicians adopted a more nuanced approach, assessing individual preferences for ACP rather than assuming patients’ values conformed to cultural stereotypes. One said (F, physician), “It’s a bigger pitfall to make an assumption about… you’re Orthodox Jewish, so you have this belief. You have to leave the door open for everybody.” Another (M, chaplain) explained, “A Muslim from Indonesia… might think about medical decisions differently than Muslims from Kuwait. Realizing there is… these broad strokes of religious traditions… but it ultimately comes down to this family and this person because there is such a diversity of how they might apply it.”

Mission-driven focus on ACP and overcoming discomfort.

—Clinicians who overcame difficulty discussing ACP with structurally marginalized patients also described their duty to conduct high-quality ACP with all patients rather than being limited by personal discomfort. One physician (M) said, “Certain people come from backgrounds… that cause them to be distrustful and when you’re already like, ‘I don’t want to talk with them, it will make them upset’. And then you have someone tell you, ‘I’m gonna scream at you if you talk about this’. [I say,] Probably the biggest barrier is meeting initial resistance when you’re already resisting [ACP] yourself.”

Another physician (F) likened ACP to other medical procedures requiring practice to gain proficiency, “I don’t really believe that fear of the conversation or not having the skill set to do it is… [an] answer. I am not trained in hip injections, but I don’t really have an excuse not to do them. If I don’t prioritize it, I don’t learn to do it.”

Acceptance of all preferences.

—Unlike clinicians who defined ACP based upon completion of forms, clinicians who routinely completed ACP with all patients believed that successful ACP need only elicit patients’ values. A physician (F) noted, “Do you speak for you? Alright, I’m with you. On the other hand, if son is the one who speaks for you, I have to respect that as well.” A participant (M) said, “Some people… show love and care through… medical interventions. And culturally… their values are going to be different than mine and realizing it can make sense in those cases.” Another (F) described that successful ACP may result in respecting patients’ wishes to avoid ACP altogether, “Advance care planning tends to preference a world view that anticipates a future. That is very… autonomy and control focused and doesn’t resonate with all people.” A summary of contrasting approaches to ACP among clinicians reporting or overcoming difficulty discussing ACP with structurally marginalized patients is shown in Table 3.

Table 3.

Contrasting Approaches Among Clinicians Reporting or Overcoming Difficulty Discussing Advance Care Planning With Structurally Marginalized Patients

Stage of ACP	Clinicians Reporting Difficulty…	Clinicians Overcoming Difficulty…
Initiation	Rely on preconceived notions regarding patients’ willingness to participate in ACP	Reject stereotypes and assess individual preferences for engaging in ACP
Process	Feel anxious discussing ACP with structurally marginalized patients	Believe developing strategies for ACP with all patients is a professional duty and embrace anxiety in difficult conversations
Outcome	Seek to achieve specific documentation-based ACP outcomes	Value a range of of ACP outcomes, including conversations, documentation of preferences, completion of DNR forms, etc.

Note: ACP = advance care planning.

Discussion

In this national study, we found that many clinicians systematically avoided comprehensive ACP discussions with certain racial and ethnic groups (African American, Hispanic, Asian, and Native American), non-English or non-native English speakers, and those with certain religious beliefs (Catholic, Orthodox Jewish, and Muslim). Barriers to conducting ACP with these patients included clinicians’ preconceived notions that patients were reluctant to participate in ACP, narrow beliefs about successful ACP, and insufficient institutional support to optimally engage structurally marginalized patients in ACP. Regardless of patient preferences, these clinician- and institution-level barriers may ultimately lead to clinicians disproportionately withholding ACP from minoritized patients (Table 4).

Table 4.

Themes and Notable Quotes

Themes and Selected Relevant Quotes	Participant Descriptorsa
Preconceived views of patients’ preferences
[We have a] large African American lower SES population. There’s a lot of mistrust in the medical profession [in this city]. So, when you start probing, “What are your goals for your health care?” “Well, I’m just going to leave it in the Lord’s hands.” Then I’m like, okay, I’m backing off.	M, physician
Within a lot of the rural areas, you don’t have discussions of a personal nature, [and] they’re going to take care of it on their own. “We’ll discuss in the family. We’re not going to have this discussion with you folks.”	M, nonphysician clinician
I think that within specific cultures… there are different considerations. We have a large Latino population, most of whom are Catholic… and taking somebody off life support is not what the Pope endorses. And then there are specific kind of sub-populations where it feels like… do you think I’m going to die when you bring it up.	F, physician
I would never introduce withdrawal of life support… to a highly Orthodox family of either Nation of Islam or Jewish or several other religions.	M, physician
Within the Hispanic community, there is a reliance that whatever will happen is God’s intent and that’s just how it needs to be.	M, physician
There are African American families where there is an attitude… and I can see where some physicians might be leery about approaching the prognosis.	M, physician
I see specifically within the Black community… who may be a little more religious. When providers hear them start to speak about “God will take care of me”… I see sometimes [providers] get a little uncomfortable or they kind of then don’t broach the subject about that.	M, nonphysician clinician
Narrow definition of successful ACP
I think people that are of different ethnicities… they don’t mind telling you who their person is that makes medical decisions for them but if you ask them to fill out a form then they get a little uncomfortable with that.	F, nonphysician clinician
And I’m asking the questions like, “Are there things that you can do now that you wouldn’t want to live without doing?”—“I just think it’s God’s will, doc.” I was left with, well I tried it. Now I’ve got documentation in the chart that really doesn’t help anybody… and we haven’t really achieved anything.	M, physician
Insufficient institutional resources
What do we know about patients who are having these very confidential discussions with an interpreter they have never met and they can’t see? What’s [the interpreter’s] training in having this level of dialogue in a sensitive way?	F, physician
They tell us not to use family members as interpreters, so it gets a little sticky, especially if I’m in the nursing home, because I don’t have as much access to those services. So I basically called on my cell phone, put it on speaker… but there are several degrees of separation there.	F, physician
Patient-centered approach to ACP
Rejection of stereotypes
If this was a Navajo who was living in Phoenix and not on the reservation and he’s 2 generations removed, he’s probably going to act a lot like me and may or may not care about the elders within the family making decisions for him. You never go with an assumption that every person in a culture has to be exactly following the rules of that culture.	M, nonphysician clinician
Age is a culture. My 85-year-old Russian woman is more like my 85-year-old Chinese woman than they might be to their own grandchildren. Because they have taken a shift in their independence, functionality, autonomy, economic stability. They can relate to one another sometimes more than they can their own grandchildren, who might not speak their language, know anything about the culture that grandma is living in.	F, physician
There was a Greek family… whose sister said, “Until the last second, until the last breath” to every question. Was it because they were Greek? I don’t think so, I know other Greek families.	M, physician
Acceptance of all preferences
I’ve had patients… from other cultures, where they said, “Don’t tell daddy he has cancer, because he will give up.” And I think there are these gray areas that we’re just not comfortable with… in Western society we value self-determination and patient autonomy… but that’s not necessarily true in other cultures.	F, physician
Every family has its own culture and what we do is ask, “Who do you want to be involved” because it may be just the patriarch or the patient may never want to know the diagnosis. You acknowledge some cultures don’t talk about death… and that’s okay, and then you kind of have to explore what can we talk about and how can we support you. You have to respect whatever boundaries they have established, and also acknowledge this is the law… and how can we come together?	F, nonphysician clinician
If the patient wants to sign the advance directive, “Everything should be decided by my son,” that’s still valuable information. The worst thing… we can do is alienate the patient by pushing our agenda and then we don’t get to interact with them at all.	F, nonphysician clinician
To me success means the patient is well-informed and they make a decision that is consistent with their values. Whether or not they choose what I may think is best.	M, physician
I’m fine if someone dies in the ICU. I think it’s more their peace with us. Is there a conversation and acceptance of the family?	F, nonphysician clinician

Notes: ACP = advance care planning; F = female; ICU = intensive care unit; M = male; SES = socioeconomic status.

^aParticipant descriptors include gender and role.

Central among the barriers identified by clinicians was a distancing of BIPOC, people facing language or health literacy challenges, and those of certain religions by classifying them as more difficult and less interested in ACP due to distrust of the health care system and unique cultural or religious beliefs. However, political events of the last decade would suggest that this belief is misguided, given widespread aversion to discussing dying nationally such that a proposal to reimburse ACP in the Affordable Care Act was equated to establishing “death panels.” (34) Moreover, evidence from clinical trial settings demonstrates that Black patients are as likely as White patients to complete advance directives when given equal opportunities to do so, challenging the belief that systematic differences exist in patients’ willingness to participate in ACP (35,36).

Many clinicians described a narrow perspective of successful ACP that prioritized outcomes such as completion of ACP documents and refusal of intensive therapies near the end of life, choices they believed were more likely to conflict with preferences among structurally marginalized patients. Contrary to this approach, a Delphi panel of ACP experts recommended that clinicians simply support patients in creating advance care plans, rather than being prescriptive about the form or content of those care plans (37). Furthermore, clinicians should be wary of using ACP document completion as a metric for successful ACP in any patients, as stated preferences in ACP documents are inconsistently associated with actual care received by patients. Less than 10% of physician orders for life-sustaining therapy (POLSTs) are reviewed by clinicians in emergency settings, and one-third of patients receive care that is discordant with wishes described in their POLSTs (38,39). In addition, mounting evidence suggests that many patients, including seriously ill physicians who are arguably among the most informed consumers of health care, prefer to receive intensive therapies near the end of life, possibly because use of hospital resources enables more effective management of symptoms than is possible in home- and community-based care settings (40–42).

Consistent with existing evidence, clinicians were more likely to identify barriers to ACP attributable to patients than health systems or clinicians themselves (43,44). In contrast, seriously ill hospitalized patients and their caregivers reported an equal number of patient-, clinician-, and health system-level barriers in a prior study (7). This discrepancy suggests clinicians may not be aware of or may be hesitant to discuss additional mechanisms by which they and health systems contribute to underutilization of ACP. Alternatively, it is possible that the same barrier may be classified differently by clinicians and patients. For example, although many clinicians reported low health literacy as a patient-level barrier to ACP, patients may perceive this barrier as originating from clinicians or health systems whose communication is not appropriately tailored to patients with low health literacy. Such distinctions are important, as they influence the feasibility and target of future interventions (18).

The primary strength of this study is inclusion of a large sample of clinicians purposively selected to ensure diversity of geography, role, and specialty. There are also several limitations. First, as discussed previously, the themes identified here are limited by clinician insight and willingness to share information with the study team. However, the transtheoretical model of behavioral change suggests that mechanisms uncovered through interviews may be most modifiable, as clinicians have acknowledged these heuristics or behaviors as potentially contributing to disparities in ACP (45). Second, the impact of identified barriers and facilitators on processes of care and patient-reported outcomes is not known. Third, the sample comprised an overwhelming majority of White clinicians, potentially omitting unique viewpoints from BIPOC clinicians. Given the lack of diversity in the clinician workforce (46), future studies should purposively sample BIPOC clinicians to ensure adequate representation.

In summary, we found widespread clinician- and institution-level barriers to discussing ACP with BIPOC, patients with limited health literacy, and non-native English speakers that may lead to disparities in ACP completion and end-of-life care. These findings provide preliminary data for future studies focused on developing clinician-targeted interventions and implementing institutional structural changes to promote patient-centered care near the end of life. These interventions can draw upon the distinct ACP approaches of clinicians who routinely engaged all patients in ACP, offering insights into an inclusive, tailored approach.

Funding

This work was supported by the National Institutes of Health/National Institute of Nursing Research (grant number R017034). The funder had no role in the design, conduct, or reporting of the study.

Conflict of Interest

None declared.

Author Contributions

J.S.W., K.L., P.K.G., J.T., A.G., D.C.A., A.J.R., S.D.H., and S.P. designed and carried out the study. J.S.W. obtained funding. N.D., D.C.A., and K.L. conducted data analysis. D.C.A., N.D., and K.L. contributed to writing the first draft. All authors contributed to and approved the final draft.