Provider perspectives on emergency department initiation of medication assisted treatment for alcohol use disorder

The Behavior Change Wheel (BCW) is a framework used in Implementation Science to guide complex implementation design and has been used successfully in the healthcare setting [27, 28]. The framework acknowledges that at the core of any successful intervention is behavior change on the part of the stakeholders. In order to accomplish the tasks of understanding behavior and identifying appropriate intervention options, the BCW uses the COM-B model of behavior. The COM-B model assumes that a specified target behavior (B) is the result of the interaction between one’s physical and psychological capability (C), physical and social opportunity (O), and automatic and reflective motivation (M). Once identified, each of these behavioral elements can be mapped to relevant intervention functions on the BCW in order to design successful interventions. For a visual of the BCW and its COM-B components, we encourage the reader to visit the authors’ website [29].

Our target behavior was defined as the initiation of treatment of AUD from the ED, including prescribing and administering Naltrexone, counseling, and referring patients to further resources. The BCW framework proposes three stages of development for a behavior change intervention: Stage 1 focuses on understanding behavior; Stage 2 identifies intervention options; and Stage 3 identifies behavior change techniques and modes of delivery [28]. The BCW framework was chosen a priori to assist in mapping qualitative findings to candidate interventions and thus optimize the chances of successful implementation of AUD treatment in the ED.

This is a qualitative study using an implementation science approach consisting of semi-structured interviews of Emergency Department staff to solicit their knowledge and perceptions of AUD treatment in the Emergency Department, and to inform the design of an ED-based AUD treatment program. Our team of clinicians and researchers met regularly throughout the process in order to review collected data, revise our interview guide, and integrate that evidence to shape the implementation plan. IRB approval was obtained from the Olive View-UCLA Medical Center IRB prior to commencement of any research.

The purpose of the provider interviews was to understand the barriers and facilitators present prior to the implementation in order to address each with targeted interventions. We developed an interview guide to explore staff perspectives on AUD, Naltrexone, and our proposed implementation. Questions focused on the challenges in caring for AUD patients, prior knowledge of NTX for AUD, concerns about the initiation of Naltrexone in the ED, and their preferences on the format for NTX-related training and education. Demographic information was collected including staff role, years of experience, and number of years at the institution. The interview guide developed for this study is provided as Additional File 1.

Olive View-UCLA Medical Center is a public, safety-net Emergency Department with an academic affiliation, emergency medicine residency, and an annual census of approximately 60,000 patients. The ED has dedicated pharmacists and social workers available 24 h a day, 7 days a week, and a substance use navigator during business hours. A substance use navigator (SUN), under the California Bridge model of care, performs the role of a care coordinator for patients experiencing SUD [30]. As part of this model, a Bridge Clinic exists within our institution that allows patients with SUD to receive care until a primary care clinic with experience in treating SUD can be assigned.

Purposive sampling was used to identify participants from each of the following categories: attending physicians, resident physicians, nurse practitioners (NPs), nurses (RNs), nursing supervisors, clinical social workers (CSWs), and pharmacists. Participants were approached face-to-face in the Emergency Department for in-person interviews, or via email for telephone interviews. Interviews continued in each of the categories until thematic saturation was reached.

In-person interviews were held in a private office or sub-waiting room. Telephone interviews were held at the subject’s convenience. All interviews were audio-recorded after the interviewer introduced themselves, made an attempt to establish rapport with the subject, explained the study and its goals, and obtained verbal consent. No field notes were made during interviews. Interviews were intended to range from 15 to 30 min.

The research team was comprised of emergency physicians and medical students. One emergency physician is an expert in SUD and is the co-director of the Bridge Clinic for opiate use disorder (MM), another is trained in research and implementation science (BRT), and a third has a prior background in harm reduction strategies (EF). During analysis, we reflected as a team on our positions and background and how they might impact the coding and interpretation of the data.

Qualitative data were anonymized and transcribed verbatim using a professional online transcription service. Transcripts were edited within a week of interview to ensure accuracy, then analyzed in Dedoose, a collaborative, cloud-based qualitative data analysis platform [31, 32]. We used both inductive and deductive approaches to analysis. We used open coding and constant comparative analysis to identify emergent patterns in the data and also applied the BCW COM-B model of behavior to identify barriers and facilitators to the target behavior. Combining these approaches to coding and analysis allowed us to link themes to the BCW and identify the corresponding intervention functions. The qualitative researchers (TP, EF, CD, BRT) met regularly during the coding process to discuss the coding, refine and consolidate identified themes, and resolve coding discrepancies. Member checking and deviant case analysis were used to enhance trustworthiness. All results are reported according to the COREQ standards for reporting of qualitative research and is included as additional file 2 [33].

A total of 25 interviews were conducted with various members of the emergency department, including attending physicians (n = 8), registered nurses (n = 5), resident physicians (n = 3), pharmacists (n = 3), social and clinical social workers (n = 3), nurse practitioners (n = 2), and a nurse supervisor (n = 1). Years of experience ranged from new RN graduates that had worked less than a year in the medical field, to attending physicians with up to 33 years of clinical practice. Due to COVID-19 distancing measures, on-site interviews were halted after the initial 9 interviews, and the remainder were completed virtually. No repeat interviews were conducted.

Forty-two emergent themes were identified. Themes were consolidated as specific barriers or facilitators to implementation, and further categorized as relating to capability, opportunity, or motivation under the BCW COM-B model. The complete findings, relating the facilitators and barriers, their COM-B behavioral source, and their corresponding intervention function(s) can be found in Table 1 (Capability barriers), Table 2 (Opportunity barriers and facilitators), and Table 3 (Motivation barriers and facilitators).

Capability barriers describe the physical and psychological capacity of staff to engage in the treatment of AUD. The most frequent barrier identified was a lack of familiarity with treatment options for AUD:

“I just don't think that a lot of us feel very comfortable with it [Naltrexone] because we don't really have a lot of training in it. It’s not something covered in medical school and it's not something that we currently do in the emergency department.”

ED staff also felt they lacked capacity to identify patients with AUD and to discern the patient’s interest in AUD treatment or stage of change. In order to address these gaps in knowledge, these barriers were mapped to educational, training, and enabling interventions to build providers’ capacity. Details of the educational, training, and enabling interventions can be found in Table 1.

Opportunity barriers and facilitators describe factors external to providers that impede, prompt, or modify their treatment of AUD. Ambiguity in responsibilities between medical providers and clinical social workers created confusion about who was screening for AUD and assessing patients’ readiness to change:

“It’s mostly by consult. Essentially, the doctor or the nurse contacts the social worker if the patient is interested in resources. [...] So they know we're stopping by. Because a lot of times they [the providers] consult us without asking the patient and we [clinical social workers] show up and they're like, ‘Why are you here?’”

Many physicians and clinical social workers assumed the other was responsible for initiating a discussion on AUD with patients. The BCW suggests an environmental restructure to address this barrier, which can be fulfilled by clearly defining the roles of all staff with regards to AUD treatment, and increasing dialogue between providers and clinical social workers. Some raised concern about the role of EDs in providing care that would require long-term management:

“It’s easy to start things but if we prescribe naltrexone, you say, ‘Here's your prescription,’ and they're coming back in a week for refills or two weeks because there is poor substance use disorder assistance in the community, then we're becoming a continuity clinic, which– no. It's not how the ED is supposed to be.”

Although many expressed support for MAT and were enthusiastic about the possibility of offering more for AUD patients from the ED, others were hesitant about the ED becoming a source of long-term management for SUD. To address this barrier, the BCW suggests modeling and enablement interventions. Selecting champions to model the behavior of NTX prescription to demonstrate the potential of NTX is one potential remedy. Better processes that enable transition to the primary care setting by coordinating AUD treatment and sharing educational materials with primary care colleagues are additional options.

Lack of resources and options for follow-up was an additional concern. The specific lack of capacity at SUD treatment centers, as well as insufficient social support for patients such as reliable housing, food were mentioned. Access to a phone was considered crucial:

"I'm wasting my time on doing something, if no one else help me to do something, because I can't get a hold of the patient. Giving resources, it works probably 10% of the time, if that, because maybe 10% of patients have a phone"

Enablement in this case might entail bringing phone access programs into medical centers.

Interviews with pharmacists revealed multiple logistical barriers. The hospital’s mandatory pre-approval form (referred to as prior authorization in interviews) for ordering IM NTX was problematic. Providers agreed that having to complete a pre-approval form would be a major disincentive to ordering IM NTX. Additionally, pharmacy staff considered the cost of the intramuscular form of Naltrexone prohibitive.

“With the intramuscular injection, it’s ten times more expensive per month, so that’s one issue right there. And I’m not sure in terms of insurance how well that would work for some patients.”

Modifying costs and streamlining processes are potential solutions. We sought out pharmaceutical company cost reduction programs and recommended removing the pre-approval form in addition to focusing the intervention design on the oral version of naltrexone, which was significantly more affordable and did not require authorization.

Facilitators are themes or elements that are already in place at our ED that may contribute to care for AUD patients. Identified facilitators were mostly components of opportunity and were factors external to the individual provider. They include access to regularly scheduled continuing education for physicians and nurses, the availability of ED clinical social workers, pharmacists, and substance use navigators, and the proximity of the hospital to SUD treatment centers.

Motivational barriers and facilitators describe the thought processes and emotional responses that providers associate with the treatment of AUD. We found that providers’ limitations in their perceived or real capabilities and opportunities (such as those described above) triggered emotional reactions of powerlessness when treating patients with chronic AUD:

“I have my streamlined protocol but I guess I'm aware that my streamlined protocol works for me for short-term management of severe withdrawal symptoms, but doesn't address the underlying problem. I guess I find it a challenge that I'm good at managing short-term, but I feel basically helpless in helping them get more definitive management of their problem.”

This emotion was reinforced by an experience of futility and a stigma against patients with AUD, assuming that such patients either did not want to quit or were unable:

“You know, it’s a loop. We see it all the time. They get out of the ER, they go to 7-Eleven, they buy the alcohol at 7-Eleven, they go outside, they drink it, they get drunk, 7-Eleven calls 911, they come back to the ER, and then from the ER, they go back to 7-Eleven.”

Furthermore, providers expressed fear that if the ED started offering more AUD services, that primary care clinics would start referring more patients with AUD to the ED, creating an influx of “difficult patients”:

“Anytime you have a service like this, the clinics [...] outside the hospital in different cities, basically, just refuse to deal with it and send everybody to the hospital.”

The BCW mapped these barriers to intervention functions of modeling, enablement, and persuasion. We will augment trainings to educate ED providers about the capability of our primary care providers to treat AUD in the clinics, draw analogies to how AUD treatment is similar to other chronic diseases, like diabetes, and celebrate examples of providers and patients who made progress through using the new treatment and referral pathways.

We found that providers had developed mental heuristics to summarize many of these emotions, communicate them to each other, and avoid frustration/complexity.

“There’s that coarse and vulgar term used that goes ‘metabolize to freedom,’ where you try to get them to the point where, clinically, they're relatively sober in their clinical state and they’re safely able to discharge. Oftentimes I feel like the piece that's missing is, once you reassess them, they can eat, they could walk, they’re safe to be discharged, I feel like oftentimes we miss that opportunity to ask them, "Hey, now that I’m finally cognitively meeting you for the first time, are you interested in quitting?"

We propose modeling using alternative catch-phrases such as “metabolize to screening” to emphasize the need to reevaluate intoxicated patients once they have sobered, including screening for AUD and assessing patient interest in treatment. A significant facilitator related to motivation is the pre-existing culture that views MAT positively, likely due to an existing successful program for opiate use disorder.