What is known about the topic?
Goals of care discussions (GOC) are critical to reflecting patients’ preferences in the provision of acute care, yet these discussions can be challenging to have in the emergency department (ED) setting.
What did this study ask?
What are emergency physicians’ perspectives on barriers and facilitators to GOC discussions?
What did this study find?
In this survey of emergency medicine attending and resident physicians, the majority reported feeling comfortable and adequately trained to conduct GOC discussions. However, they identified time constraints, environmental factors, and patient expectations as barriers. Fifty-four percent of respondents believed that it was primarily the responsibility of admitting services to conduct GOC discussions.
Why does this study matter to clinicians?
This study suggests that dedicated ED resources for palliative care, such as a palliative care ED pathway, and addressing structural factors, such as a way to dedicate time and private space to GOC discussions, would be promising avenues for improvement. Training did not appear to be a barrier.
Emergency medicine (EM) is a specialty that was born out of the need to provide timely access to acute resuscitative care. End-of-life care and symptom palliation were not part of this initial vision, but they are now widely understood as essential elements of providing acute care.Reference Lamba and Mosenthal 1 , Reference Lamba 2 One study showed that half of all older Americans receive care in an emergency department (ED) in the last month of life, highlighting the importance of the EM’s role in providing end-of-life care.Reference Smith, McCarthy and Weber 3 However, a survey of over 100 American EM residency directors found that only 59% of programs include palliative care training in their program, although no data are available on the quantity and quality of that training.Reference Kraus, Greenberg, Ray and Dy 4 This deficiency may extend to the Canadian setting. In response to these patient population needs, several initiatives have been implemented in the United States and Canada to integrate palliative care into the ED setting with variable success in scaling.Reference Grudzen, Richardson and Johnson 5 - Reference Quest, Herr, Lamba and Weissman 8
ED care happens at the nexus between community and hospital care, and goals-of-care discussions are critical in orienting proposed therapies within the priorities and wishes of the patient.Reference Sinuff, Dodek and You 9 , Reference Fowler and Hammer 10 Goals-of-care discussions transpire between a patient (or their substitute decision-maker) and the treating physician, and are defined as a process of communication about decision-making that outline an individual’s plan of care in an institutional setting.Reference You, Downar and Fowler 11 , Reference You, Fowler and Heyland 12 Patients and their caregivers identified five critical elements to goals-of-care discussions: 1) preferences for care in the event of a life-threatening illness; 2) values; 3) prognosis; 4) fears or concerns; and 5) questions about goals of care.Reference You, Dodek and Lamontagne 13 , Reference Heyland, Ilan and Jiang 14 Clear documentation of these discussions is also paramount so that patients’ wishes can be respected throughout their care journey.Reference You, Fowler and Heyland 12 Presently, few physicians discuss these elements with their patients, and there is limited concordance between an individual’s stated preferences and their prescribed treatments.Reference You, Dodek and Lamontagne 13 However, both patient satisfaction and alignment of preferences with goals of care increase with the number of elements discussed.Reference You, Dodek and Lamontagne 13
Establishing goals of care begin with advance care planning (ACP), a process by which individuals clarify their preferences for future medical care. These discussions are context-specific. ACP optimally occurs in the outpatient setting, largely with primarily the treating physicianReference Chiarchiaro, Arnold and White 15 but may also take place upon hospital discharge for future care.Reference You, Fowler and Heyland 12 In some instances, an advance directive is created – a written document that guides the patient or his or her substitute decision-maker by outlining the level of “in-the-moment” care that an individual would want,Reference Sudore and Fried 16 including preferences for life-sustaining treatments (sometimes referred to as the patient’s code-status, e.g., full code, do not resuscitate (DNR), or comfort care).Reference You, Downar and Fowler 11 , Reference You, Aleksova and Ducharme 17 ED visits are often triggered by an acute change in the patient’s health condition and require review of an individual’s specific goals of care for that crisis (using an advance directive as a guide if available), resulting in particular challenges and urgency.Reference Martin, Hayes, Gregorevic and Lim 18 - Reference Seow, Barbera and Pataky 20 As front-line providers, EM physicians and residents are called upon to conduct initial goals-of-care discussions with patients who present to the hospital. However, there is a limited understanding of the barriers and facilitators to conducting goals-of-care discussions in the ED setting.
Prior work suggests that barriers to providing palliative care, of which the clarification of goals of care is a critical component, include the ED environment, a lack of time, family dynamics in the acute setting, and difficulty accessing an individual’s complete medical history.Reference Lamba, Nagurka, Zielinski and Scott 21 - Reference Lakin, Isaacs and Sullivan 23 In a U.S. study, EM physicians reported that patient complexity, acuity, and communication gaps pose challenges in patients and families making goals-of-care decisions in the ED.Reference Fassier, Valour, Colin and Danet 24 The creation of targeted interventions to facilitate goals of care in the ED is hindered by a lack of understanding of the nature and importance of these barriers, particularly in the Canadian setting. This study sought to describe the perceived barriers and facilitators to conducting goals-of-care discussions in the ED from the perspective of EM physicians and residents.
Setting and population
Certified EM physicians employed in six Toronto, Ontario-area hospitals as well as residents at the University of Toronto’s Royal College of Physicians of Canada (FRCPC) EM residency program were invited to participate by email. Four academic and two community high-volume urban EDs were included in the study. None of these EDs had access to a direct 24-hour palliative consult service. (Institutional ethics approval was obtained.)
An interdisciplinary team of physicians with training in EM, Palliative Care, Internal Medicine, and dual EM and Palliative Care Certification developed a 21-question survey to assess perceived barriers and facilitators to conducting goals-of-care discussions with patients arriving in the ED (Appendix 1). The survey consists of 3 multiple choice, 17 Likert-scale, and 1 open-ended question, created through a review of the EM and palliative care literature, including a previously published survey of U.S. EM physicians and hospital administrators.Reference Lamba, Nagurka, Zielinski and Scott 21 Questions covered four domains of goals-of-care discussions identified from the literature, as follows:
3) ED environment
4) Personal beliefs of the patients and caregivers
To qualitatively assess for common themes reported by EM physicians that were not captured by our survey, a single general open-ended question was included in the survey (What do you find challenging about having these [goals of care] discussions?). The survey was initially pilot tested with 16 EM physicians at a single site. Electronic and paper-based surveys were distributed at the pilot site’s monthly departmental meetings and through email. Participant feedback was solicited on survey length, and to ensure question clarity and relevance. The survey was then further revised based on these preliminary findings.
An email invitation was sent to 224 academic EM physicians and 49 FRCPC-EM residents with a link to the survey (SurveyMonkey, Palo Alto, CA). Consent was implied by way of survey completion. Reminder emails were sent to all eligible participants at 10, 21, and 50 days.Reference Burns, Duffett and Kho 25
Descriptive statistics are presented for all multiple choice and Likert-scale questions and were performed using Microsoft Excel (Redmond, WA). Responses to the open question were analysed using the content analysis method as described by O’Cathain and Thomas.Reference O’Cathain and Thomas 26 Based on a preliminary content analysis of comments, a coding scheme with eight common themes was devised.Reference Sutradhar, Barbera and Seow 19 The comments were coded by two independent reviewers using the initial coding scheme, and conflicts were resolved through a case-by-case discussion of every code and by consensus. Two themes were added in the course of this process, and one was removed. The frequency of each theme was calculated as a percentage of respondents who mentioned it. A single-respondent comment could be coded into multiple themes.
Of the 273 eligible EM physicians, 108 EM staff (108/224, 48%) with similar representation from all six sites and 23 EM residents (23/49, 47%) participated. Five respondents did not identify as staff or resident, and their responses were included in the overall analysis and in the calculation of the response rate, except where results were reported by practice status.
Study participant characteristics are presented in Table 1. EM staff had practiced medicine for an average of 9.8 years (median 5, range 1-46), whereas EM residents were typically in their second to third year of postgraduate training (median, 2; range year, 1-5). Responses to questions other than frequency of discussion were very similar between EM staff and residents, and thus results were collapsed and combined as “EM providers” in Tables 2 and 3.
* Proportion of staff response rate from each site were similar. Data not displayed.
† Five respondents did not specify resident or staff affiliation. Their data were included in “all respondents.”
Education and training
Most staff physicians (64/108) conducted goals-of-care discussions several times per month or more, 74.1% (80/108) of the time versus 35% (8/23) of residents. Over 90% of all respondents (92.4%, 121/131) felt that goals-of-care discussions were within their scope of practice, and most noted that they are adequately trained to conduct them (73.1%, 95/130). However, only 24% (33/136) of all respondents felt that EM providers are the best service to carry out goals-of-care discussions. Over half of the respondents (54%, 74/136) felt that the admitting physician is the most appropriate healthcare provider to conduct goals-of-care discussions, and 55% (72/130) of respondents felt that the availability of a 24-hour palliative care service would facilitate the ability to conduct goals-of-care discussions in the ED; however, only 6% (8/136) of respondents felt that palliative care should be specifically consulted to have goals-of-care discussions (these data are not displayed in a table).
Of all respondents, 34% (46/135) indicated that they never or almost never find it difficult to initiate these conversations in the ED. Only 42% (55/131) of all providers believed that patients are ready and willing to have goals-of-care discussions in the ED. The majority of participants (76.1%, 102/134) reported being unable to communicate with a patient’s primary care provider prior to initiating a goals-of-care discussion.
Most EM staff and residents (76%, 98/129) felt that the ED environment is not conducive to having goals-of-care discussions. Only 5% (6/130) reported having sufficient time in a busy ED to communicate with patients regarding their goals of care.
Forty-five percent (60/134) of respondents disagreed that advance directives were readily available for patients arriving from a long-term care facility, and 8% of respondents (115/131) disagreed that goals-of-care discussions are unnecessary for long-term care patients because of the availability of advance directives.
Personal beliefs of patients and caregivers
Fifty-five percent (74/134) of all respondents were comfortable having goals-of-care discussions with their patients. However, 70% (90/130) of all respondents felt that they never or almost never have effective prognostic tools for patients with non-cancerous disease to facilitate discussions.
One hundred two participants of 130 responded to the open question – What do you find challenging about having goals-of-care discussions? – which contributed to 192 thematic statements outlined in Table 4. The most commonly reported challenges to having goals-of-care discussions in the ED were lack of time (46.1% of respondents) and patient personal and cultural beliefs about their medical care and condition (22.5% of respondents). Several participants noted that patient expectations in the ED, including the perception of bad outcomes when goals-of-care conversations are undertaken, pose a challenge in the acute care environment. One physician commented, “Once I start talking about goals of care, families and patients just assume they’re going to die today in the ED – that their prognosis is set in stone.”
* n = 102 responses into 192-themed statements (total % exceeds 100%)
Another physician commented, “We do not have an ongoing relationship with our patients and so it’s a hard conversation to initiate.” Multiple respondents noted other perceived challenges to engaging in goals-of-care discussions in the ED, including lack of goals-of-care consideration and discussions prior to the ED visit (20.6% of respondents), level of patient acuity (17.6% of respondents), working environment not being conducive to goals-of-care discussions (16.7% of respondents), difficulty contacting substitute decision-makers (16.7% of respondents), and challenges communicating with the patient due to an emotional state and level of consciousness or language barrier (10.8% of respondents). Examples of reported challenges with the working environment included a high noise level, a lack of quiet spaces, and frequent interruptions. One responder commented that “the ED does not lend itself to quiet, thoughtful discussions.”
In this survey-based study of 136 EM staff and resident physicians, we identified several important themes relating to the perceived challenges and barriers to having goals-of-care discussions in the ED. Respondents reported frequently conducting goals-of-care discussions and also identified several challenges and limiting factors related to the ED practice setting. Consistent with prior research, our findings point to ED patient characteristics, including their acuity, expectations, lack of relationship with ED physicians, and unavailable family or a substitute decision-maker as major barriers to goals-of-care discussions in the ED.Reference Lakin, Isaacs and Sullivan 23 , Reference Fassier, Valour, Colin and Danet 24 , Reference Platts-Mills, Richmond and LeFebvre 27 Many of these obstacles could be helped by outpatient healthcare providers ensuring ACP, but this is not sufficient and is outside of the ED providers’ scope of intervention.Reference Rolnick, Asch and Halpern 28 Some respondents highlighted the difficult role that they play in conducting goals-of-care conversations with patients without prior advance directives with whom they have little or no relationship.Reference Smith, Fisher and Schonberg 29 , Reference Stone, Mohanty and Grudzen 30 Efforts to improve communication between EM physicians and primary care providers, as well as ensuring that advance directives accompany patients to the ED as a means to enhance continuity, appear as promising areas for future intervention.
EM staff and resident physicians often felt that the admitting services, such as Internal Medicine, should be conducting these discussions with their patients. This may reflect that without good prognostic tools, it is difficult to engage in goals-of-care discussions and that specialized knowledge of a disease process and in-hospital follow-up may be a good alternative to support the discussion. What respondents consider “prognostic tools” is also subjective, because laboratory diagnostic tools, such as serum lactates, are known to have acute implications on mortality. However, tools, or structured classifications or assessments meant to estimate chronic prognosis secondary to disease, are lacking. Nonetheless, the initiation of a consult to an admitting service may initiate a patient on a trajectory that is contrary to his or her goals of care and cannot be seen as an alternative to goals-of-care discussions in the ED.
Respondents also identified limited time availability for ED patients as a challenge, particularly noting in comments that this discussion cannot be rushed while ED physicians are “under enormous time constraints.” It was also mentioned by several respondent comments that admitting services often care for patients over many days during an admission and therefore are afforded more time to build rapport with patients. Although this was not formally assessed in our study, lack of time may have contributed to the belief that other services should be carrying out these discussions. Taken together, these findings support the need to discuss a focused framework for goals-of-care discussions in the ED that complement the ongoing discussions of the admitting service, without replacing them. Environmental factors should also be considered, such as ensuring a quiet space and blocking off even a few minutes separately with the family to address questions and concerns. These may not require additional background training, which physicians said was currently adequate.
Our results highlight the potential need to add targeted resources directed at facilitating goals-of-care discussions in the ED. These resources may include additional tools to improve access to medical records,Reference Lakin, Isaacs and Sullivan 23 quiet spaces to conduct goals-of-care conversations, and potentially standardized goals-of-care pathways that communicate to patients and families that goals-of-care discussions are about providing appropriate, aligned care. The creation of an alternate “emergency palliative care pathway” for patients who are identified to be near end-of-life, could signal to set aside some time to engage in goals-of-care discussions with a primary focus on prioritizing relief of burdensome symptoms and eliciting preferences.Reference Grudzen, Stone and Morrison 7 , Reference Quinn, Detsky and Smith 31
Patient education in the outpatient setting continues to play a key part in ACP and should include expectations of what could occur during an ED visit for an acute exacerbation, highlighting that an ED goals-of-care discussion is part of a process, not a single point in time. A discussion by the ED physician may only be one step of a longer conversation with the patient and his or her caregivers once admitting services become involved and after prior conversations with the primary care physician.
This was a self-reported survey of EM staff and resident physicians, which is subject to recall biases. Participation was voluntary and is therefore subject to response bias. Although our study was performed at multiple sites, it still was conducted in a single urban centre, which may limit generalizable to other practice settings where resources and the availability of subspecialty teams may be limited. We decided to analyse responses of residents and staff EM physicians together given the similarities in their responses, but it is possible that significant differences in perceptions between these groups may have been missed due to our small sample size. However, because these physicians practice together in the same environment, and many residents are trained by these supervising staff, we felt that their responses are likely representative of each other. Residents training through the Canadian College of Family Physicians (CCFP) EM program were not included in this study due to logistical reasons, which may have highlighted differences in their training model. However, staff physicians from both the CCFP and FRCPC streams were included in the survey. Finally, this survey represents the perceptions of emergency physicians. Clinical services that admit patients from the ED, including Palliative Care, as well as patients and their caregivers, may have differing perspectives as to who should conduct these discussions and the unique barriers they face in clarifying goals of care in the ED.Reference You, Downar and Fowler 11 , Reference You, Aleksova and Ducharme 17
Multiple barriers and possible enablers to conducting goals-of-care discussions in the ED were identified as reported by EM physicians. System-level interventions to encourage ACP improve communication with outpatient health providers, and the creation of an “emergency palliative care pathway” may facilitate goals-of-care discussions in the ED.Reference Quinn, Detsky and Smith 31 This study also highlights the nature of goals-of-care discussions as a process rather than a single point in time.
Authors NA and KLQ contributed equally to the manuscript.