Physician-identified barriers to and facilitators of shared decision-making in the Emergency Department: an exploratory analysis

Abstract

Objectives

Shared decision-making (SDM) is receiving increasing attention in emergency medicine because of its potential to increase patient engagement and decrease unnecessary healthcare utilisation. This study sought to explore physician-identified barriers to and facilitators of SDM in the ED.

Methods

We conducted semistructured interviews with practising emergency physicians (EP) with the aim of understanding when and why EPs engage in SDM, and when and why they feel unable to engage in SDM. Interviews were transcribed verbatim and a three-member team coded all transcripts in an iterative fashion using a directed approach to qualitative content analysis. We identified emergent themes, and organised themes based on an integrative theoretical model that combined the theory of planned behaviour and social cognitive theory.

Results

Fifteen EPs practising in the New England region of the USA were interviewed. Physicians described the following barriers: time constraints, clinical uncertainty, fear of a bad outcome, certain patient characteristics, lack of follow-up and other emotional and logistical stressors. They noted that risk stratification methods, the perception that SDM decreased liability and their own improving clinical skills facilitated their use of SDM. They also noted that the culture of the institution could play a role in discouraging or promoting SDM, and that patients could encourage SDM by specifically asking about alternatives.

Conclusions

EPs face many barriers to using SDM. Some, such as lack of follow-up, are unique to the ED; others, such as the challenges of communicating uncertainty, may affect other providers. Many of the barriers to SDM are amenable to intervention, but may be of variable importance in different EDs. Further research should attempt to identify which barriers are most prevalent and most amenable to intervention, as well as capitalise on the facilitators noted.

INTRODUCTION

Shared decision-making (SDM) is formally defined as the ‘collaborative process that allows patients and their providers to make health care decisions together, taking into account the best scientific evidence available, as well as the patient’s values and preferences.’¹ SDM has been promoted in primary and outpatient care for decades, but has only recently made inroads in acute care settings.^2,3 SDM has been shown to increase patient knowledge and engagement, increase adherence, decrease resource utilisation and, in some scenarios, improve outcomes.^4–6 SDM has also been shown to enhance trust between patients and physicians, a challenge for emergency physicians (EP) due to the brevity of the clinical encounter.⁷

Although EPs report that they use SDM in the ED,⁸ a number of questions remain about SDM use in the ED. With increased attention focused on the delivery and promotion of patient-centred care, the use of SDM in the ED requires further characterisation and understanding.² SDM is a conversation between patients/families and clinicians, but is usually initiated by the clinician, in the setting of emergency care.⁹ That makes the understanding of the perspective of the physician-as-stakeholder paramount to efforts to promote SDM. Interventional studies intended to test the value of SDM, such as ‘Chest Pain Choice,’ have used stakeholder engagement methods in the context of decision aid development to understand the perspectives of those who will be using an intervention.⁶ However, an in-depth understanding of the overarching barriers and facilitators that ED physicians face in their day-to-day use of SDM is lacking. A broader approach to understanding EPs’ perspectives could yield insights that could be generalised to many SDM scenarios, facilitating the uptake and use of SDM for a range of clinical encounters. While studies of other physicians outside of emergency medicine suggest that time constraints, the clinical scenario and certain patient characteristics are the biggest barriers to SDM, to date no exploratory research exists investigating the barriers and facilitators to SDM that EPs deem relevant.¹⁰ For example, while SDM has been purported as particularly relevant in low-evidence, high-uncertainty clinical situations,^11,12 little is known about how uncertainty affects EPs’ use of SDM.

While we have previously published an analysis regarding EPs’ motivations,¹³ the objective of the current study was to specifically examine the EP-identified barriers to and facilitators of the use of SDM in the ED, with the idea that this analysis could guide future research examining these barriers and facilitators in a wider and more representative sample of EPs.

METHODS

Study design

We conducted semistructured interviews with practising EPs and used established qualitative methods for our analysis.^14,15 The study was granted exempt status by the local Institutional Review Board, but used written informed consent because we recorded the interviews. Participants were reimbursed $25 for their time. The study was designed to comply with published standards for reporting qualitative research.^14,15

Interview guide

We based the interview guide on an integrative theoretical model that combined the theory of planned behaviour and social cognitive theory. This framework (figure 1) organises the factors that influence an individual’s performance of a behaviour, such as initiating an SDM conversation.¹⁶ We also incorporated concepts from previous qualitative literature that included interviews with non-EPs, as there are no published qualitative data regarding EPs’ attitudes.^10,17–19 The interview guide was piloted and we made minor additions to question probes following the first several interviews, but no changes were made to the guide after the fifth interview (online supplementary appendix 1). Both interviewers took field notes during interviews and discussed these notes after interviews.

Figure 1.

The theoretical framework drawn from an integrative model: social cognitive theory and the theory of planned behaviour¹⁶ demonstrating how various factors affect a behaviour (SDM, shared decision-making).

Participants were given a standard definition of SDM.¹ They were then asked to discuss scenarios where they used SDM techniques. They were asked further questions about what helped them use SDM at that time (facilitators), and they were then asked to think about and discuss scenarios where they could have used SDM but chose not to, and asked to elaborate on the reasons why they did not use SDM (barriers). After they discussed scenarios, they were asked whether they could think of other factors that prevented them or their colleagues from using SDM more frequently.

Study setting and selection of participants

We chose a purposive sample of attending EPs based on gender, years in practice since residency, region (rural/suburban vs urban) of primary employment, academic versus community practice setting and location of training (inside vs outside of the state) based on the idea that these characteristics may affect attitudes towards SDM.²⁰ Physicians who met these criteria were identified via professional networks and enrolment continued until thematic saturation was achieved (the point when additional interviews fail to lead to new concepts).²⁰ Residents (EPs in training) were not included.

Data collection

Participants provided written informed consent and provided demographic and background information. Interviews were conducted in person at private locations. The interview team consisted of two female practising EPs (EMS and TRE) who trained and piloted interviews under a senior investigator with qualitative methods experiences (SLG). All interviews were audio recorded and transcribed. After verbatim transcription, member checking was performed by providing participants with a short summary of the major points they discussed and asking them for their agreement, disagreements or comments.¹⁵

Data analysis

Transcripts were entered into Dedoose qualitative data management and analysis software (Dedoose V7.0.18, Los Angeles, CA: SocioCultural Research Consultants). Coding was performed by three research team members (EMS, ERK, KEP). The codebook was developed using a directed approach to qualitative content analysis: we combined a priori codes drawn from previous literature and our theoretical framework with emergent codes that came directly from line-by-line coding of the transcripts.²¹ We coded transcripts of interviews in an iterative process; transcripts were recoded as the codebook was refined. Each transcript was coded at least twice by at least two coders, and agreement was calculated for final coding iterations. Disagreements were discussed until consensus was reached. The codebook is available in online supplementary appendix 2.

Our previous analysis of this data focused on fully exploring physicians’ motivators, however the interview guide was designed to facilitate a discussion of barriers and facilitators to SDM in the ED, and this analysis was planned a priori. Although lack of motivation could be interpreted as a barrier, emergent themes were distinct between ‘motivators’ and ‘barriers and facilitators’, and the research team felt that an in-depth analysis of barriers and facilitators was warranted.¹³

Research team and reflexivity

The interviewers knew many, but not all, of the participants prior to the interviews. Most of the participants did not know the goals of the research prior agreeing to be interviewed, but the goals were stated during the semistructured interview and it was made clear that the interviewers sought honest attitudes (online supplementary appendix 1: semistructured interview guide).

RESULTS

Characteristics of study subjects

Between June 2015 and November 2015, we interviewed 15 EPs. No distinct new codes emerged from the last three interviews, indicating that thematic saturation was achieved.¹⁵ Participant characteristics are described in table 1.

Table 1.

Participant demographics

Participant characteristics (n=15)	n
Age, mean (range)	46(31–65)
Female	6 (40%)
Race/ethnicity
White	10(67%)
Black	1 (7%)
Asian/Indian	2(13%)
Did not answer	2(13%)
100% of clinical effort at academic hospital	6 (40%)
100% of clinical effort at community hospital	4 (27%)
Combined academic/community practice	5 (33%)
Years since residency, mean (range)	13.1 (1–30)
Residency in emergency medicine	13(87%)
Total different practice sites where participants had worked within past year	14

Measures of validity

Member checking did not lead to any additional new codes, and coders were able to reach consensus on the generation and application of codes.

Main results: themes identified

Although our interview guide was designed to touch on all aspects of the theoretical model (figure 1), a number of themes emerged as more relevant to EPs. Figure 2 depicts the relative importance of the domains, based on the frequency that EPs brought up particular themes and the importance that they gave them in discussion. The domains of environmental constraints, attitudes towards patients, norms and skills emerged as more prominent in our analysis. The domains and themes as well as the representative quotes are listed in table 2 and further explained in the text.

Figure 2.

The theoretical framework: the themes cited as most important to emergency physicians (EP) are in bold. SDM, shared decision-making.

Table 2.

Domains, subdomains and representative quotes

DOMAIN
Subdomain, explanation of barriers and facilitators (facilitators in italics)	Representative quotes (facilitators in italics)
Environmental factors
Emotional
The effects of certainty and uncertainty Uncertainty is a barrier to using SDM, although sometimes SDM can be used to mitigate uncertainty. Tools, such as risk stratification methods, that help with certainty or the communication of uncertainty can facilitate SDM.	‘In our culture, we have an extraordinarily low risk tolerance because we don’t cause most people don’t understand risk.’ ‘I’m never 100%, you know, ‘here’s my best guess of your chances,’ which if I’m talking to them, it’s usually a few percentage points, and I talk to them a bit about what they feel comfortable with.’ ‘Some patients, in my experience, have been offended because… I’m even offering the possibility that they might tolerate some risk.’ ‘I’ve used probably some sort of guide, whether it’s some sort of score or just clinical decision making, I say ‘you know, I think your risk is pretty low.’ “Well, with the usual tools that we all have available to us… the risk assessment tools, the current one that I like to use more than the others is the HEART score, those current tools are good for medical legal documentation…I don’t really need the tools, but… lean at least say ‘this is what the predictor says.’”
Fear of a bad outcome The stress caused by worrying about a bad outcome may keep physicians from using SDM. The use of SDM may mitigate the likelihood of litigation if there is a bad outcome, since the patient understood that there was a chance of this outcome.	‘I think you can potentially can expose yourself to more risk with shared decision making, and I think that’s probably one of the highest disadvantages and why a lot of people might not… I think that’s why they don’t do it.’ ‘[our job is] getting everything right 100% of the time and move as fast as possible…’(therefore being wrong is unacceptable, leading to more testing) ‘I still think about some of them (patients I do SDM with), I worry a little more about some of them than the patients I CT.’ ‘It is mentally easier (to admit the patient instead of using SDM) because I’m the type of person that would go home and think about that…a catastrophic event…a little bit.’ ‘(I didn’t use SDM earlier in my career because) I didn’t like trust myself, my clinical instincts, and even the patient’s instincts on things…and you’re afraid of the medical-legal implications.’ “[Interviewer: What is it that the doctors don’t feel comfortable with?] ‘Missing something, getting sued, I think bad outcomes and all the trimmings.’“ I think that there’s also some…mitigation of the risk, because at least you’re making the patients aware of that…you’re able to document that shifting responsibility to some degree, and that you’ve had a conversation and you’re documenting patient preference. So, whether or not that holds up in any way…I’m not sure, but it feels like it does when you ‘re doing it.’ ‘Yea, actually, there is a benefit to this system. In general, I think it provides…decreased risk, so it’s a risk mitigation technique.’
The challenge of confidence and trust building with ED patients Because of wait times, patient and physician stress and the demands of other patients, it can be hard to build confidence and trust with patients, which is a barrier to effective SDM. EPs noted that SDM done well is likely to improve communication with patients, as well as increase patient satisfaction and engagement.	‘I think that people want to know that the doctor that they talked to had found something or was confident in this is what’s going on, and so I think that if I don’t do a good job, of that or come in too shared decision making-oriented, where ‘maybe it’s this, maybe it’s that,’… I don’t want to sound too wishy-washy.’ “I don’t feel like that (SDM) instills confidence into the patient, like ‘what do you want to do?’“ ‘I think if we have these conversations it gives the patient some understanding about what we’re doing, why we’re doing it, maybe some say into what gets done to them or for them, and sorta brings them into more of an active role rather than a passive role in their care. ‘ I think patients are sometimes overwhelmed with the prospect of having to make those decisions, but also usually fairly appreciative that you take the time to explain things to them and that they get a say’. ‘ I gotta assume there’s increased satisfaction in the provider when they involve you this way. It’s probably going to decrease the liability, it’s gonna increase my satisfaction with the job, because …if I have a patient that’s happy it always feels good to walk away from that, no matter what happened.’
Logistical
Time Time constraints as a barrier were noted by nearly every participant.	‘Time is perhaps the major constraint’ ‘… you can’t go in there knowing you only have 5 min cause it’s never gonna take just 5 min.’
Scenario Often the scenario is not appropriate for SDM.	‘I mean sometimes… it’s too acute of a scenario.’
Follow-up care Concern regarding reliable follow-up was a frequently noted barrier. Reliable follow-up was noted as a facilitator.	‘If she had already had established an oncologist and PCP [primary care provider] (I) probably would have discussed with her ‘what do you want to do, do you want to stay in the hospital? All we’re going to do is pain control and nausea medicine, it’s up to you.’ I usually tell them I’ll do whatever you want, at that point, whatever you’re comfortable with, and they’re usually pretty happy with either…’ [Interviewer: With this patient you didn’t even approach that conversation?] ‘No.’ “Then I kind of talk with the person factoring in ‘do they have follow up? Are they reliable?’“ ‘(Can I) easily contact their primary care doctors and talk about things?’
Frequent interruptions The frequent interruptions expected by physicians were seen as a barrier to SDM.	‘Depending on the number of interruptions it can be challenging.’ ‘You don’t want to rush, and then you get phone calls and then you have to go do other things, and in the back of your mind you’re thinking about the other 8 patients you have to go do stuff for, but you’re also trying not to rush with this thing. Cause if it’s rushed, you can’t really do it right.’
Challenges due to multiple clinicians Contributions from multiple providers (both due to residents, sign-outs and consults) make SDM harder.	‘I also think that it’s harder to do if you are then going to transition care to someone else.’ ‘I have a personal relationship with the patient (that I don’t have when I work with a resident) and that personal relationship plus your own gestalt or impression makes a big difference.’ ‘I’d say it’s harder with the residents, I do it less.’ ‘…I’m often seeing the patient second, or sometimes third if there’s a student involved also, or if it’s sorta like another team signed out to another team, and they’ve seen so many caregivers and the plans have already been essentially determined.’
Effort SDM often required more work on the part of the physician than the alternatives.	‘I can spend two seconds and look at that person and know that now we’re gonna admit him versus… ‘he really wants to go home.’ Now I have to think about that for a bit more. Now I have to sit down, and if there are three codes going on, it kind of gets to, ‘well, sometimes we just admit folks.’ [Interviewer: You’re saying the easiest and safest thing is the more conservative thing, to not use SDM?] ‘It becomes the default mode, I guess.’
Challenges due to physical space Hallway beds and lack of privacy were seen as barriers.	[regarding barriers] ‘Having a real conversation in the hallway, it’s not private…can’t sit down…’
Norms
Culture of the institution The culture of the institution could be both a barrier and a facilitator, depending on the institution. Guidelines, reinforcing the norms and culture of the institution, were seen as helpful and ‘protective’by some, and a burden by others.	‘What is more powerful is the culture of the institution, right? Where I trained before, at a county hospital, we didn’t admit anybody for chest pain… you’d talk to them about the risk and… that was what the institution, and… the population, expected. Those were the resources, that was what was given. I think… especially at a more conservative place, where you can just admit everyone, it’s a lot harder to have the impetus to engage and kind of put yourself at risk.’ ‘Having guidelines that are, that there’s at least some buy-in from other providers and a practice approach where you can feel protected both medico-legally and, you know, reputation wise. So consistency in the practice so that the community perception of an institution is appropriate and not so individually affected.’ ‘I think like we’re inundated with the guidelines…They feel more like rules than guidelines most of the time, so …that would not work for me… not looking for more guidelines.’ ‘I am pro-guidelines. I feet like people stilt need their autonomy, but you also still need to make sure that there’s a standard of care and that there’s expectations met when patients come to the Emergency Department.’
Attitudes towards patients
Stubborn or aggressive patients If the patient is perceived as aggressive or difficult, or has their own agenda, the provider may avoid SDM. An upset patient/family may benefit more from SDM.	‘If (the patients’) expectations exceed what I think is appropriate then I may…not…approach it with them for (SDM).’ ‘If they have started to sort of dig their heels in about something, prior to that conversation, then I have a sense that I know what their preference is already, I won’t always take part in shared decision making at that point.’ ‘Although sometimes if they are somewhat more hostile I will get them more involved just so they can be… (less hostile).’
Capacity If the patient lacks capacity or competency or is not reliable regarding following instructions (returning, medications, follow-up), this is a barrier.	‘Are they reliable?’ ‘Their ability to understand that they really, really need to come back if something different happens, is really important for me to involve them in the decision making process.’
Desire to be involved Many physicians noted that many patients do not want to be involved (barrier). When patients explicitly asked about alternatives, this is a facilitator.	‘When they say, ‘what would you like to do doc’… it doesn’t work, it’s doesn’t work because … the patient’s not either able to understand or willing to truly participate.’ “There are people who are gonna be overwhelmed. Who you can tell are getting overwhelmed when you start to talk about things and who say, ‘You know what? You just tell me what to do.’“ ‘Sometimes patients just want to be told what to do.’ “Others have clearly expressed to me that they don’t want to have any part in that decision, ‘(You’re) the goddamn doctor, why don’t you make a decision?’“ [Interviewer: Is there anything that can … push you toward shared decision-making?] ‘They ask questions.’
Education/intelligence Some patients are perceived as difficult to engage due to education level/intelligence.	‘Even when I do it in a plain language, there are just some people that lack education and might not be able to understand what I have to say.’ ‘They have their fixed opinion and you’re not going to change it. We had a guy in a car accident, wasn’t wearing his seatbelt, said ‘yea man, those seat belts kill people, you know!’ I’m not going to touch that, no SDM with that guy.’ [Interviewer: Why did you choose not to use SDM with that patient?] ‘It was health literacy.’
Cultural/language barriers Language and cultural barriers lead EPs to avoid SDM as they feel it will be too hard or they will not be understood.	‘If there’re huge language barriers, unfortunately even with a translator, sometimes those nuances are lost.’ ‘They think it’s because of who they are, that they don’t have insurance, that’s why we’re not admitting them… I think (with) that specific population I have a very hard time doing (shared) decisionmaking with.’
Patient does not tolerate or understand risk or uncertainty	‘The main thing that stands in my way is the patient’s inability to understand risk.’ ‘There are some patients who don’t want to know what’s going on or what’s the chances of a bad outcome.’
Skills
Clinical skills Lack of confidence in clinical skills makes providers hesitant to allow options. Improving clinical confidence (due to experience) was noted as facilitator, as was having communication skills.	‘You have to have clinical confidence that this is probably not anything big and be willing to take that risk, again, with the decision or the input of the patient.’ ‘After many years one gets to see the many different ways of doing it and so I’m more comfortable with those.’ ‘…until you have your clinical gestalt grounded (in experience), how can you have a discussion?’ ‘I think my use of shared decision making has changed based on my level of training. I feel like I use it a lot more now that I am an attending.’ ‘I think I do that a lot more now than I did when I started (because of my clinical experience).’
SDM skills No providers had formal training in SDM. Those who reported using SDM felt that each positive interaction encouraged them to use SDM more.	‘I’ve not had any formal training in it… I’m very comfortable in it but I don’t know if it matches with the techniques that others use.’ ‘I’ve done it more than most and therefore am comfortable, not that I’m doing it right or anything.’ ‘I think part of it’s just confidence with their own ability, decision making, and also communication ability’. ‘I think you have to be pretty good at it. I think you have to be pretty good at talking to patients and their family, and so sometimes when I hear more junior people trying to do something in this line I think they’re using the wrong words, big words, or wrong concepts, or… I think it’s something you practice and learn.’

EP, emergency physician; SDM, shared decision-making.

Domain 1: environmental constraints: emotional and logistical challenges

Our analysis suggested that the environment of the ED presented the most significant barriers, from the perspective of EPs. From our discussions, we included in ‘the environment’ the physical space and logistic issues of the ED, as well as the emotional factors that EPs attributed to their particular work setting. We felt that the environment of the ED set the emotional tone, and therefore the emotional challenges were inextricably linked to the environment of emergency care.

Emotional challenges

Theme: the effects of certainty and uncertainty

Many comments were made about uncertainty, and many noted that communication of uncertainty was a challenge, and that tools that facilitated the communication of uncertainty, by quantifying risk, were facilitators. Most physicians reported that they attempted to use methods to risk-stratify patients prior to engaging in SDM, and noted that SDM was harder to do if the had no good way to quantify their uncertainty. They reported that they often did not know what a patient’s true risk of a bad outcome was, and many suspected that they themselves or their peers were overestimating patients’ risks of bad outcomes, and this was a barrier to SDM, as it prevented them from involving patients in decisions. Many also noted that both a systemic and cultural intolerance of uncertainty or their patients’ intolerance of uncertainty was a barrier to SDM.

Facilitator: (numbers) make you more comfortable having the conversation… with some degree of knowledge associated it with as opposed to a gestalt for what we think it is.

Barriers: [at that hospital there was] less shared decision-making, as the tolerance for ambiguity was less.

When we assess risk for adverse outcomes, we as emergency physicians usually overestimate the potential harms that can occur. That’s our training.

Theme: fear of a ‘Bad Outcome’

EPs reported that there was some amount of emotional stress in involving patients in decision-making due to the possibility that a less ‘conservative’ (or less resource-intensive) decision could be made, potentially leading to a missed diagnosis or bad outcome. For example, many physicians reported using SDM to discuss whether or not to use a CT scan in young patients who had abdominal pain but were considered low risk for appendicitis. However, they reported that when they did not get the scan, they worried about the patient after their shift, whereas when they did order the scan, they did not go home and ruminate on how much radiation exposure they caused by exposing a low-risk patient to a CT scan. Conversely, several did note that they hoped that SDM was protective in the setting of a bad outcome. They hoped that that by involving the patient, they shared the responsibility for the choice, and therefore might not be as liable, should a less conservative decision be made and a subsequent bad outcome occur.

Facilitator: I don’t know if this pans out in actual data, but it feels like you’re protected from bad outcomes better if the patient feels like they had a say in that choice.

Barriers: Medico-legally I don’t think this (SDM) helps us a lot.

It is so much easier just to admit people (rather than having an SDM conversation). It’s sad, and I don’t want to do it for everyone, but god it just takes that liability off.

Theme: the challenge of confidence and trust building with ED patients

Physicians reported both logistical and emotional challenges to building trusting relationships and having open conversations with patients in the ED. They noted that frustrated patients want ‘answers,’ not discussions, particularly after long wait times. Several reported that they felt that offering a discussion of options sometimes left patients feeling less confident in the physician, and one noted receiving a patient complaint alleging exactly that.

Barriers: I don’t want—by providing the option to choose—I don’t want that to be interpreted as being indecisive on my end.

Despite these challenges, the majority of the physicians did report that patients often left an SDM encounter more satisfied with their care. However, while physicians recognised this, it did not seem to be a consistent facilitator: less than half of physicians reported that patient satisfaction was a top priority for them, among the competing demands of a busy shift.

I think for the most part people feel better with more information and feeling as if they had a say in what happens.

Regarding patient satisfaction as a priority: ‘Getting the diagnosis right probably has to be first, getting teaching done efficiently, getting through the waiting room, and then the patient satisfaction is going to be—not at the bottom—but at the middle of the pack of things I’m interested in getting done because… I can’t let it get in the way of getting diagnoses right. They don’t have to like me; they have to just get better.’

Logistical challenges

Previous research has demonstrated that ‘time constraints’ and ‘inappropriate clinical scenario’ were two of the most frequent barriers noted by physicians of other specialties.¹⁰ Although these barriers were cited by our participants as frequent issues, our analysis revealed other logistical constraints, many of which are specific to the ED and new to the literature. Most notably, ‘reliable follow-up’ (having rapid access to a primary care provider) was noted to be crucial by two-thirds of the participants.

Theme: time constraints

All but one physician noted that time is a barrier, with several noting that being pressed for time could lead to a biased or incomplete conversation, which was no better than no SDM at all.

It takes a little more time, sometimes it’s easier just to shotgun (send more tests).

Theme: clinical scenario

Several physicians noted that when there was no clinical equipoise or preference-sensitive decision to be made, SDM was not appropriate.

[Interviewer: Tell about times you don’t use SDM?] ‘STEMIs [ST segment elevation myocardial infarction], I’m not asking a lot of questions, I’m going forward.’

Theme: follow-up care

As many of the decisions that physicians discussed with patients involved sending the patient home with less testing, reliable and accessible follow-up care was seen as critical prior to SDM. This was a hard stop for many physicians—a lack of access to follow-up primary or specialty care or concern that the patient would not follow through with accessing this care definitively meant that SDM would not be offered.

Facilitator: At—hospital, with next day follow-up, (we) could engage the patient in more SDM as I knew they would not get lost in the system.

Barriers: Often it depends on primary care follow-up… you have to determine, does the person have a primary care physician, if they do, is it somebody they actually have a relationship with and could see in the office?

Theme: frequent interruptions

Several physicians noted that having an important conversation regarding a clinical decision was challenging with frequent interruptions.

And interruptions…that’s a little bit crazy to try to talk and have a conversation.

Theme: challenges due to multiple clinicians

Physicians noted that it was easier to have discussions if only they, individually, and the patient and family were involved, but that decisions were much more complex when other clinicians were involved. This was noted to include residents, who often saw patients first and set expectations, as well as consulting physicians and oncoming EPs, to whom the patients were signed out. The reasons for this as a challenge were different for different clinicians. When it came to working with residents, EPs noted mistrust of the resident’s interpretation of the patient’s history, and therefore concern that the patient was not appropriate for SDM. For consultants, they noted that specialists often had specific plans that did not allow for SDM. Regarding ‘sign-outs,’ the transfer of care between EPs, they noted that a certain standard of care was expected and while they could deviate if they were the only practitioner, they could not deviate if they were then transferring care at the end of their shift.

Theme: effort

Several providers pointed out that SDM is almost never ‘the path of least resistance’; it almost always requires more work than the alternative. This means that when providers are stressed by high volumes, high acuity or time constraints, SDM is even less likely to occur.

On a slower shift, (I) might have thought more about a test that I ordered, but you know, if there’s a lot going on and a disposition needs to be made and the patient is either outright or essentially asking me for, again, their 53rd CAT [computerised axial tomography] scan, that is, you know, two clicks of a button for me…

Theme: challenges due to physical space

Several physicians also noted that caring for patients in hallway beds, as opposed to actual ED rooms, made having important conversations more difficult, if not impossible.

This is not a hallway thing.

Domain 2: attitudes towards patients

Every physician noted that there were some patients whom they would avoid engaging in SDM, and a few mentioned patient characteristics that would encourage them to initiate SDM. The most common influential patient characteristics noted are listed in table 4. Patient behaviours that would encourage SDM included asking about alternatives and appearing engaged. Physicians noted they avoided SDM in patients who lacked decision-making capacity, risk tolerance, capability of understanding risk, English language fluency, or follow-up, and also with patients who seemed to have an agenda or were too anxious to meaningfully participate.

Facilitators: If they’re somebody who’s clearly paying attention to what’s going on with their own life then I assume that they’re going to be more involved.

Barriers: The main thing that stands in my way (of using SDM) is the patient’s inability to understand risk.

There are families who…very aggressively advocate for their loved ones and are not as receptive to um… I don’t know…grey areas, medical uncertainty, diagnostic uncertainty, and who just who have very strong ideas about what is right and what is wrong.

Early on you can kind of tell that they’re just not going to be the ones that are deciding…they don’t want that kind of responsibility.

Much of the time I don’t think the patient’s capable of understanding these highly complex issues.

I don’t know that they trust a white female who doesn’t speak Spanish.

Domain 3: skills

Two related skill sets influenced the EPs’ use of SDM: clinical skills and SDM/communication skills. No physicians reported having any formal training in SDM, though none expressed a desire for further training, and none admitted that this lack of skills might be a barrier for themselves.

Theme: clinical skills

Many physicians noted that increasing confidence in their clinical skills, or clinical acumen, encouraged their use of SDM (and conversely, when they were less confident, they were hesitant to use SDM).

Earlier on in my career I didn’t trust myself, you know, my clinical instincts…So I think getting confident in the process, right? And that’s just I think, you just do it (SDM) more and more and you believe that it’s better.

Theme: SDM/communication skills

Regarding specific SDM skills, all providers noted they had no training in SDM, and though most said they were comfortable having SDM conversations, several noted that they had no idea if they were ‘doing it right’ or similarly to how others engaged.

I think I’m fairly ok in doing it, I think that there’s definitely people that would benefit from formal education in it, but I feel like those are the same people that you identify with having difficulty communicating to patients in general.

Domain 4: norms

Theme: institutional norms and culture

The culture of the institution could be both a barrier and a facilitator, depending on the institution itself. In places where tolerance for uncertainty was high and resource utilisation scrutinised, SDM was felt to be more of the norm and was used more often. When there was no pressure to judiciously use resources, SDM was less the norm. While the culture of the institution could be a barrier or facilitator, EPs generally did not express the idea that their colleagues’ behaviour affected their own—their reasons for using or not using SDM rarely related to others’ use. Additionally, physicians were very mixed about guidelines— some felt institutional guidelines would ‘support’ them in using SDM, and others felt that guidelines limited their independence.

Barrier: I feel like it’s hard to go against the institutional norms that are there. You could, but it makes for a persistent headache.

Facilitators: More powerful is if the institution says, ‘no…we’ll back you,’ and here are…the guidelines that we have, and we will say, ‘If this person is low risk as you determine, and you do this, we’re ok with that and that’s part of our culture,’ (that) would be much more beneficial.

DISCUSSION

This is the first study to examine EP’s self-reported barriers to and facilitators of the use of SDM via exploratory methods. Physicians recognised numerous barriers and facilitators regarding their use of SDM in the ED, and many of these are newly reported and specific to emergency medicine. While logistic concerns such as time, space and lack of follow-up care are clearly important barriers, the emotional barriers such as the stress of uncertainty and the fear of a bad outcome are likely to be more difficult to address—and appear to be significant obstacles for many of the EPs interviewed. While our previous analysis focused on what motivates EPs to use SDM, the results we present here explore a different side of the issue: despite a theoretical acceptance of the idea of SDM, and several clear motivators, SDM continues to be a challenge both logistically and emotionally.¹³ The logistical challenges suggest concrete, although difficult, solutions: increased privacy and time for encounters and greater access to follow-up care. The emotional barriers—that physicians are more comfortable ordering more tests and admitting patients rather than risking a bad outcome or missed diagnosis—do not lend themselves to simple answers. The combination of these obstacles may be seen as insurmountable for many physicians, and further research is needed both to understand which obstacles are most salient in a particular setting and how to go about mitigating them.

Our findings regarding the emotional stress of uncertainty warrant a closer look. Although many believe that situations of increased clinical uncertainty are exactly the place for SDM,^11,12 research suggests that the communication of uncertainty may decrease decision satisfaction.²² Portnoy et al demonstrated that physicians do have varying degrees of comfort with uncertainty, and that their perception of their patients’ ‘ambiguity aversion’—their tolerance of uncertainty—predicted their likelihood of using SDM: physicians were less likely to offer choices to patients whom they perceived to be intolerant of ambiguity.²¹ Our participants echoed this but phrased it differently, they noted that the patient needed to be able to tolerate risk in order to participate in SDM, and a failure to understand or accept risk led the physician to avoid SDM. It has been previously noted that for patients to participate in SDM, they need to ‘appreciate that there is uncertainty in medicine.’²³ Our results suggest that EPs would agree with this and may withhold SDM for patients they believe will have difficulty with uncertainty. What is unknown is how much uncertainty or risk ED patients are willing to accept, although it is likely this varies considerably by patient and by clinical scenario.²⁴ Although research exists to bolster accurate risk communication, it is unclear whether we can affect our ‘ambiguity aversion’ as a culture.²⁵ It is possible that EPs may be overestimating their patients’ ambiguity aversion, and avoiding SDM in situations where ambiguity might be acceptable to patients. Conversely, patients may have higher ambiguity aversion in the ED, and EPs may withhold SDM because they recognise this.

Many EPs noted that reliable follow-up care was a prerequisite for SDM, and that patients without access to follow-up care received more tests and were more likely to be admitted. As avoidable ED care is more expensive than reliable primary care, this is yet an additional reason for the government, policymakers and insurers to improve access to primary care. Although we cannot draw conclusions from this small exploratory investigation, our participants clearly report that patients with reliable follow-up care are more likely to be offered SDM, which has the potential to improve their care and decrease costs. This is a potentially testable hypothesis, with profound indications— at least in the USA where many patients do not have reliable follow-up care.

Regarding physicians’ skills, two different skill sets emerged as relevant in relationship to SDM: clinical skills and communication skills. A number of physicians noted that they used SDM more frequently as their clinical skills improved over the course of their careers. This is likely due to improved clinical acumen and to their increased ability to tolerate uncertainty. Second, many noted that they had no training in SDM, but they recognised the importance of communication skills. Although a few physicians noted that they may not be ‘doing it right,’ no physician stated that they needed training in SDM, implying that the physicians themselves did not see this lack of training as a barrier. However, a lack of recognition of the need for formal training does not mean it is not warranted—in contrast, this lack of recognition can be seen as a barrier in itself, as these physicians are unlikely to seek training in this skill or recognise their own ability to improve at this skill.

Lastly, as was noted by Kanzaria et al, the perception exists that many patients do not want SDM or cannot understand the medical issues well enough to meaningfully participate.⁸ It is possible that the patient’s limited understanding represents a failing not on the part of the patient, but on the part of the physician’s ability to facilitate this conversation. Theoretically, improved SDM skills could expand the number of patients with the ‘capacity’ to meaningfully participate in SDM.

Our qualitative data suggest several next steps. For those wishing to promote SDM, it will be important to know which of the barriers discussed are most widespread and salient, so further research, using larger and more generalisable samples, may be helpful. The breadth of the barriers we identified suggests that multidimensional interventions may be necessary to foster greater use of SDM. Patients can be encouraged to ask about alternatives; institutions can promote a pro-SDM culture by putting SDM in clinical guidelines, making physicians feel supported; insurers and government can increase access to primary care and put in place reimbursement for important discussions.

LIMITATIONS

Although our study is the first to qualitatively explore the barriers and facilitators to using SDM in the ED, there are limitations to our analysis. First, while the physicians worked in a number of different clinical settings, all were practising in the USA at the time of the interview, and physicians from other regions may have responded differently. Interviewers were EPs themselves, and social desirability bias may have influenced responses. Interviews took place in 2015, and increased attention to SDM may have shifted attitudes since then. Individual systems may impose different barriers that did not emerge from this analysis. For example, those without integrated electronic health records may see their lack of access to patients’ histories as a barrier. Also, we examined only physician-identified barriers. Physicians may not recognise the same barriers that patients, nurses or other involved parties feel are important, and this warrants further exploration. Lastly, this analysis was exploratory and should be viewed as hypothesis generating.

CONCLUSIONS

In summary, EPs noted a number of barriers and facilitators regarding SDM in the ED. Although further evaluation of these barriers is necessary, it appears that improving access to follow-up care, promoting increased dialogue around uncertainty, improving physicians’ communication skills and finding ways to decrease logistic barriers could all potentially increase the frequency with which EPs feel they can use SDM in their practices.

Key messages.

What is already known on this subject

• Outside the ED, physicians most commonly cite time restraints, situational appropriateness and patient characteristics as barriers to their use of shared decision-making (SDM). In previous literature, we reported an exploratory analysis of ED physicians’ reported motivations for using SDM. However, an exploratory evaluation of barriers that ED physicians perceive has not been published.

What this study adds

• ED physicians enumerated a number of ED-specific emotional and logistical challenges to SDM, such as difficulty building trust, the challenges of uncertainty, lack of follow-up care and the physical space of the ED. They also noted a number of barriers that generalise to other settings, such as concern that patients do not understand the conversations and their own lack of skills and experience. This study adds to our understanding of how to help clinicians overcome the barriers that prevent the use of SDM in this complex setting.