Background

Opioid Use Disorder (OUD) is a problematic pattern of opioid use that results in impairment or distress of clinical significance [1]. OUD is a cause of significant and increasing harm, with 14,700 opioid-related deaths in Canada between January 2016 and September 2019 [2]. During this time period, 1 in 4 patients seen in EDs for opioid-related poisonings were admitted, resulting in 19,490 hospitalizations [2]. Since opioid-users frequently present to the ED in withdrawal, following an overdose, or with complications of injection drug use, it is an important opportunity to initiate treatment for OUD.

A recent Canadian clinical guideline recommends the use of the opioid agonist medication buprenorphine-naloxone (trade name Suboxone), from here on referred to as buprenorphine, as first-line treatment for OUD [3]. Buprenorphine offers advantages over methadone, including a ceiling effect which limits the abuse potential and risk of diversion, reduced stigma, and milder withdrawal symptoms [4, 5]. Buprenorphine is increasingly accessible in Ontario after being included on the Ontario Drug Benefit formulary [6]. Buprenorphine induction in the ED has been shown to be a feasible and effective approach to begin medication for addiction treatment [7,8,9]. A randomized clinical trial found that ED-initiated buprenorphine significantly increased engagement in addiction treatment and decreased use of inpatient addiction services compared to referral or brief interventions in the ED [7]. Retention on buprenorphine treatment has been shown to reduce ED utilization [10].

Widespread adoption of new practices takes time and has particularly lagged in the treatment of substance use disorders [11]. Previous studies have examined attitudes of primary care providers towards prescribing buprenorphine [12,13,14,15] as well as those of emergency physicians [16,17,18]. Yet there is limited literature on attitudes, knowledge and behaviours surrounding OUD and initiating buprenorphine in Canadian EDs [16]. We sought to explore Canadian emergency physician attitudes and experiences in order to inform communications and quality improvement strategies during the opioid crisis and help support patients and providers alike as buprenorphine prescribing expands across Canada.

To achieve this objective, we used grounded theory, a systematic qualitative methodology which is used to generate theories about social processes, experiences, and interactions [19, 20]. A constructivist approach to grounded theory recognizes that there can be multiple perspectives of reality which are shaped by outcomes, experiences, and interpretation within a changing context [21]. Constructivism also draws on researchers’ expertise and experiences to inform data collection and analysis. Constructivist grounded theory is increasingly used to understand health care processes, particularly in nursing research [22].

Methods

The objective of this study was to assess the attitudes, motivation, experiences, and practices in the treatment of OUD in an ED in Toronto, Ontario. Our qualitative study design was informed by constructivist grounded theory to allow us to generate theory rooted in the data [20], while remaining reflexive during data collection. The research team consisted of two medical trainees with health policy backgrounds (DW, PH) and two academic emergency and addictions physicians (JH and HS). This approach thus allowed us to acknowledge our assumptions throughout, and that data analysis is a co-constructed process [21, 23].

Sampling

We purposefully sampled ED physicians from one urban academic teaching hospital in Toronto, Ontario. This department was chosen as a department ‘in transition’ which had some exposure to OUD management: all physicians had received educational rounds on OUD management, an introduction to an opioid withdrawal order set, and the hospital had established a rapid-access addictions clinic for follow-up. We randomly invited half of the physicians from that group of eighty physicians to participate. We performed a simple randomization process using a random number generator to minimize selection bias. We limited invitations to half of the department’s physicians to ensure that all interviews could be completed within the study period.

Data collection

We drafted and field tested a flexible semi-structured interview guide with topic experts, and modified interview questions to ensure developing categories were explored.

The guide consisted of open-ended questions about clinicians’ experiences with patients with opioid use disorder. We prompted if needed on their conceptions of the role of the ED in addressing opioid addiction, preventing future overdoses, and counselling around opioid use, treatment options, and harm reduction. The final portion elicited clinicians’ experiences and perceived barriers to prescribing buprenorphine, their training, interest, and resources required to facilitate prescribing buprenorphine. The interview guide is available for reference [see Additional file 1].

The first author (DW) conducted semi-structured interviews in person and over the phone based on participant preference between July and September 2018. Data collection and analysis continued until we reached theoretical thematic saturation after 19 participant interviews. Interviews ranged from fifteen to forty-five minutes in length. We recorded and transcribed the interviews verbatim, using pseudonyms during transcription to ensure anonymity. We obtained informed verbal consent from all participants and maintained confidentiality. The University Health Network Research Ethics Board approved the protocol as part of a quality improvement project to improve the care of patients with OUD.

Data analysis

Constructivist grounded theory guided a step-wise approach to simultaneous data collection and analysis [21]. The first four transcripts were reviewed independently by DW, JH and HS to develop familiarity with the data and consensus on a broad coding framework. Two reviewers (DW, PH) independently organized and coded the transcripts through multiple readings to identify meaningful patterns, using an iterative constant comparative and interpretative approach. DW and PH reviewed the themes to ensure representativeness of the data and to reach a consensus on discordant views. A third reviewer (JH) reviewed and triangulated the thematic analysis with the two main reviewers. Data analysis was organized using Dedoose software (version 8.0.35, SocioCultural Research Consultants LLC, Los Angeles, CA).

Results

Nineteen physicians participated in the study. Participants were predominantly male (n = 12; 63%) Almost half of participants (n = 9) were in their first five years of practice, five were mid-career having practiced for 6–20 years, and five had been in practice for over 20 years. Almost half (n = 9; 47%) were family physicians with additional training in emergency medicine (CCFP-EM), six were Royal College trained in emergency medicine (FRCP), and four were family physicians (CCFP).

Predominant concepts and themes are summarized here, with supportive quotes outlined in Tables 1, 2, and 3. Results were grouped under broad categories of barriers and facilitators to providing evidence-based treatment of opioid use disorder in the ED. Overall, we found wide variation in the current approaches to treating OUD in the ED. Barriers to treating OUD and prescribing buprenorphine in the ED included physician reports of limited knowledge and experience, inadequate nursing support, safety concerns, patient factors, and logistical challenges. Facilitators to managing opioid withdrawal included the presence of physician champions and positive experiences with opioid agonist therapy, as well as system-level factors such as the availability of order sets and timely access to follow-up for patients.

Table 1 Participant quotes supporting variability in OUD management of OUD
Full size table
Table 2 Participant quotes supporting facilitators to treating withdrawal and initiating buprenorphine
Full size table
Table 3 Participant quotes supporting barriers to treating withdrawal and initiating buprenorphine
Full size table

Variable management of opioid withdrawal

Participants varied in their approach to managing opioid withdrawal and opioid use disorder. Several participants indicated that opioid withdrawal management was frustrating and required nuanced decision-making (Table 1.1i, ii). Many physicians recognized that their approach to medical management and counselling practices likely differed from their colleagues (1.1iii). While some physicians counseled patients extensively on treatment options, harm reduction practices and available resources, others stated that they counselled patients minimally.

Some physicians offered and initiated buprenorphine (1.1iv) while others used non-agonist treatments to target symptoms (e.g. clonidine, non-steroidal anti-inflammatories, anti-diarrheals, benzodiazepines) (1.1v, vi). Participants recognized that opioid medications might be inappropriately prescribed in some cases to patients in opioid withdrawal (1.1 vii). One participant did not feel that there were effective medications for opioid withdrawal (1.1viii). Nevertheless, most physicians interviewed were familiar with buprenorphine and pointed out that its use in the department was increasing, although still in its early stages (1.1ix).

Facilitators to evidence-based treatment of opioid withdrawal in the ED

Departmental champions and support

Physicians frequently stated that their colleagues helped them change their approach to managing opioid withdrawal in the ED including prescribing buprenorphine. These physician champions educated their colleagues about opioid withdrawal and the benefits of buprenorphine (Table 2.1i). Physicians frequently referred to mentorship and informal discussions about their patient cases with more experienced colleagues as important factors in increasing their comfort with prescribing buprenorphine (2.1ii, iii, iv). Some also felt that hearing success stories of patients who had decreased their opioid use after presenting to the ED in withdrawal served as an important motivator to learn more (2.1v).

Physician Empowerment & Patient Satisfaction

A few physicians who had initiated buprenorphine in the ED gave accounts of significant patient satisfaction due to resolution of symptoms (Table 2.2i, ii),. These accounts contrasted sjarply with reported experiences using other non-agonist medications which did not resolve symptoms effectively and raised safety concerns (2.2iii). Most physicians with experiences prescribing buprenorphine felt it empowered them to provide a meaningful intervention to a patient population for whom they had no effective options previously (2.2iv).

Protocols and order sets

Most ED physicians indicated that having an order set for buprenorphine induction was helpful in the management of opioid withdrawal. The order set prompted physicians to consider prescribing buprenorphine, provided inclusion and exclusion criteria as well as dosing recommendations. Physicians felt that the order set instilled greater confidence and was well-integrated into their workflow (Table 2.3i, ii). The order set is available for reference [see Additional file 2].

Timely access to follow-up care

Physicians identified timely follow-up as an important factor in their willingness to initiate buprenorphine in the ED (Table 2.4i). Physicians felt that the follow-up should be available the next day, though some were comfortable if patients could access it within the next few days (2.4ii). Limited follow-up during the weekend reduced their comfort with prescribing buprenorphine in some cases (2.4iii).

Barriers to evidence-based treatment of opioid withdrawal

Lack of experience and confidence among the care team

Most physicians still felt inexperienced with managing opioid withdrawal, with some concerned about missing subtle presentations (Table 3.1i) and the risk of precipitating withdrawal when initiating buprenorphine therapy (Table 3.1ii). Others noted uncertainty about the long-term safety of buprenorphine as a barrier to prescribing (Table 3.1iii). Many physicians hesitated to use the opioid withdrawal order set as the nursing staff lacked familiarity with buprenorphine and the Clinical Opiate Withdrawal Scale (COWS) (Table 3.1iv).

Some physicians who were interested in prescribing buprenorphine experienced challenges when they encountered more complex cases (3.1.v). Factors such as medical or psychiatric comorbidities (3.1vi) or a patient’s lack of interest in treatment (3.1vii) presented additional challenges. Some physicians encountered difficulties with buprenorphine induction for patients who used methadone or had recently used other opioids (3.1vii, viii).

Negative patient interactions

In some cases, physicians had experienced negative interactions with patients which affected their motivation to counsel and treat OUD in the ED. Physicians spoke of cases of aggressive behaviours which made for stressful encounters (3.2i). Tensions between patient and physician arose in cases where the patient wanted to leave after an overdose but the physician felt it was unsafe (3.2ii) or when the patient requested an opioid medication which the physician did not believe was indicated `3.2iii). These adversarial interactions limited the establishment of rapport and presented a barrier to further counselling and discussion of buprenorphine therapy. In some cases, patients declined due to previous poor experiences with the medication (3.2iv).

Logistical constraints

Physicians commonly cited time constraints as a barrier to treating opioid withdrawal in the ED. Some physicians identified waiting for the patient to reach an appropriate level of withdrawal and subsequent monitoring upon initiating the medication as barriers (Table 3.3i). Concerns around time requirements for proper counselling were common (3.3ii). Some physicians disagreed that time constraints were a barrier to the management of withdrawal and rather focused on the potential to reduce future visits by intervening (3.3iii).

The physical environment of the ED was cited as an important barrier to comprehensive management of opioid withdrawal, including overcrowding (3.3iv), “hallway medicine” and a lack of privacy (3.3v). Physicians felt that both time and physical constraints made extensive counselling difficult to take on in the ED.

Initiating a chronic medication in an acute care setting

Some physicians did not feel that prescribing a long-term medication in an ED setting was appropriate (Table 3.4i) with comparisons to hypertension and hyperlipidemia management (3.4ii). The need for medication titration, uncertainty of immediate follow-up and long-term management were reported (3.4iii). Some perspectives changed when they considered the potential of this medication to effectively address opioid withdrawal in the acute setting (3.4iv).

Furthermore, many physicians felt uncertain about the effectiveness of their intervention in managing opioid withdrawal. Those that referred patients to the outpatient follow-up clinic identified that they had no feedback on whether the patient had attended (3.4v). Most felt that it would be helpful to know if patients were attending (3.4vi).

Discussion

This is a qualitative study describing barriers and facilitators to addressing OUD and prescribing buprenorphine in an urban Canadian ED setting. The Canadian guidelines recommend buprenorphine as first-line for management of OUD [3]. Our findings suggest highly variable approaches to managing opioid withdrawal and the need for a targeted knowledge translation strategy to address common barriers and shift ED culture to one that celebrates treatment of acute presentations to affect long-term outcomes. In Canada, there are no regulatory barriers or exemption programs needed in order to prescribe buprenorphine [24].

Most physicians reported limited or no experience prescribing buprenorphine and expressed feelings of dissatisfaction and frustration with their current interventions for OUD in the ED. Similar perspectives have been reported in a recent survey of ED clinicians in a tertiary center in the United States [18]. A key difference though was that a minority of the interviewees in that study were in favor of ED-initiated buprenorphine whereas most participants in our study agreed that it should be considered [18]. In our study, physicians reported more favorable attitudes with increased clinical education and exposure. In both studies, respondents agreed that ED-initiated buprenorphine was feasible if appropriate system supports were implemented.

Our findings align with previous research showing that emergency physicians perceive that system level facilitators including order sets and departmental protocols increase their uptake of procedures in the ED [16, 25]. Our recent Canada-wide survey of ED physicians also highlighted the potential value of order sets in shifting culture and practice [26]. Despite the availability of these resources, participants suggested that additional guidance around challenging cases would be helpful as part of the knowledge translation strategy, including complexities around polysubstance withdrawal, comorbidities, and relative contraindications to buprenorphine induction.

There is also a role for mentorship, continuing education, and additional clinical exposure. Our study found that a significant facilitator to adoption of buprenorphine in the ED was the presence of physician champions who led departmental change, consistent with literature which shows that role modeling builds physician comfort with new practices [27, 28]. Having outpatient clinics available addresses concerns of initiating a chronic medication in an acute setting, titrating the dose, and providing ongoing care. Rapid access clinics demonstrate positive clinical outcomes and retention in care, forming a crucial component of ED-initiated treatment for OUD [9, 29]. ED leadership should work closely with outpatient clinics to ensure rapid follow-up and a smooth transition of care from the ED to the outpatient clinic.

Our study identified persistent perceived barriers to buprenorphine use in an academic ED which had already initiated these supports including order sets, training, and follow up for patients. Additional barriers reported included lack of patient interest, discomfort with counselling around buprenorphine, lack of nursing support, time constraints and safety concerns as perceived barriers for ED physicians [16,17,18].. These findings highlight the need for earlier, more robust education for ED physicians and nurses on the use of buprenorphine.

Various frameworks suggest that guidelines are not adopted due to either knowledge, attitude or behaviours [30,31,32,33,34]. We found all three types of barriers in our findings; physicians reported lack of knowledge, uncertainty about the impacts of using buprenorphine for opioid withdrawal as well as environmental and patient factors as barriers to uptake. Treatment of OUD presents unique challenges in the ED given frequently associated social and medical complexities.

Additionally, the treatment of OUD is often stigmatized by physicians which can impact clinical outcomes [35, 36] and hinder knowledge translation [37]. Future studies should further explore the relationship between stigma and attitudes and perceived barriers to treating OUD.

Departments should develop a knowledge translation and implementation strategy that directly addresses perceived barriers in order to optimize guideline adoption [38]. Future research may also focus on understanding perceived barriers of health care administrators and allied health care staff in supporting buprenorphine prescribing in the ED setting [32]. These findings could inform our knowledge translation strategies, as we suspect from this study that early involvement of allied health professionals, including nursing and social work, in buprenorphine counselling would vastly improve uptake.

Limitations

This qualitative study had some important limitations. We interviewed physicians from one organization at one point in time which limits the generalizability of our findings. Generalizability is further limited by constructivist grounded theory methodology which considers the unique context and experiences of participants in shaping their reality which may vary considerably among emergency physicians. Although we randomly invited physicians, only half of physicians in the department were invited to participate to ensure that all interviews could be completed during the study period. Participants self-selected to participate in interviews which introduced some selection bias. Inviting all emergency physicians in the department could have improved sample variation. Finally, although participants were assured of confidentiality, we cannot exclude the possibility of socially desirable responses.

Conclusion

This study describes barriers and facilitators to addressing OUD and prescribing buprenorphine in a Canadian ED setting. Our findings suggest highly variable practices and the need for a targeted knowledge translation strategy to address common perceived barriers. Additional research is needed to elucidate perceived barriers of health care administrators and allied health care staff in supporting buprenorphine prescribing in the ED setting. Our findings suggest a role for physician champions, involvement of allied health professionals in counseling, and expanded education.