a b s t r a c t

Background: There is strong evidence for emergency department (ED)-initiated treatment of Opioid use disorder . However, implementation is variable, and ED management of OUD may differ by clinical presentation. Our aim was to use mixed methods to explore variation in ED-based OUD care by patient clinical presentation and understand barriers and facilitators to ED implementation of OUD treatment across scenarios.

Methods: We analyzed Treatment outcomes in OUD-related visits within three urban, academic EDs from 12/ 2018 to 7/2020 following the implementation of interventions to increase ED-initiated OUD treatment. We assessed differences in treatment with medications for OUD (MOUDs) by clinical presentation (overdose, with- drawal, others). These data were integrated with results from 5 FoCUS groups conducted with 28 ED physicians and nurses January to April 2020 to provide a richer understanding of clinician perspectives on caring for ED patients with OUD.

Results: Of the 1339 total opioid-related visits, there were 265 (20%) visits for overdose, 123 (9%) for withdrawal, and 951 (71%) for other OUD-related conditions. 23% of patients received MOUDs during their visit or at dis- charge. Treatment with MOUDs was least common in overdose presentations (6%) and most common in with- drawal presentations (69%, p < 0.001). Buprenorphine was prescribed at discharge in 15% of visits, including 42% of withdrawal visits, 14% of other OUD-related visits, and 5% of overdose visits (p < 0.001). In focus groups, clinicians highlighted variation in ED presentations among patients with OUD. Clinicians also highlighted key as- pects necessary for successful treatment initiation including perceived patient receptivity, provider confidence, and patient clinical readiness.

Conclusions: ED-based treatment of OUD differed by clinical presentation. Clinician focus groups identified sev- eral areas where targeted guidance or novel approaches may improve current practices. These results highlight the need for tailored clinical guidance and can inform health system and policy interventions seeking to increase ED-initiated treatment for OUD.

(C) 2023

1. Introduction

* Corresponding author at: 122 Blockley Hall, 421 Guardian Drive, Philadelphia, PA 19104, United States of America.

E-mail addresses: [email protected] (M.K. Delgado), [email protected] (J. Perrone), [email protected] (R. McFadden), [email protected] (R.A. Xiong), Nicole.O’[email protected] (N. O’Donnell), [email protected] (C. Wood), [email protected] (G. Solomon), [email protected] (M. Lowenstein).

Opioid Use Disorder and opioid-related morbidity and mor- tality continue to rise rapidly in the United States, with annual drug overdose deaths exceeding 100,000 in recent years [1]. Medications for OUD (MOUDs), particularly opioid agonist therapy with buprenor- phine or methadone, reduce overdose and all-cause mortality, illicit drug use, and transmission of infectious diseases as well as improving Treatment retention and quality of life [2-4]. Use of MOUDs is also asso- ciated with reduced opioid-related acute care visits compared with

https://doi.org/10.1016/j.ajem.2023.01.009

0735-6757/(C) 2023

other treatment strategies [5,6]. However, there is a large evidence-to- practice gap, with a minority of patients receiving any treatment for OUD and even fewer receiving treatment with MOUDs [7].

Emergency departments (EDs) are increasingly recognized as a crit- ical touchpoint for MOUD initiation. OUD-related ED visits are increas- ing [8], and the ED is a place where out-of-treatment patients seek substance use assessments as well as receive care following overdose or for complications of drug use. randomized control trials have shown ED-initiated buprenorphine more than doubles treatment en- gagement at 30 days compared with referral alone and is cost-effective [9,10]. Recent guidelines have advocated for ED-initiated buprenor- phine to become standard of care, and some EDs have developed path- ways for MOUD initiation and harm reduction strategies like Naloxone distribution [11-13].

While MOUDs are the standard of care, treatment initiation is not a one-size-fits all approach, and guidelines and protocols for clinical man- agement may require adaptation depending on the clinical scenario. In the ED setting, buprenorphine is the most commonly used MOUD be- cause it can be both administered on-site in the ED and prescribed at discharge. Until recently, this required a DATA 2000 waiver (“X-waiver”), but recent Policy changes now allow buprenorphine to be prescribed by anyone with a DEA registration. However, standard bu- prenorphine initiation typically is done when patients are experiencing mild to moderate Opioid withdrawal, limiting the ability to administer buprenorphine during an ED stay for patients who are not in with- drawal. There is a great deal of heterogeneity in OUD-related ED visits, including presentations for overdose, Acute withdrawal, or associated Medical complications. In prior work, we have seen a high degree of provider-level variability in buprenorphine prescribing among clini- cians in the ED [14] and in other settings [15]. However, less is known about how care differs based on clinical presentation or strategies for tailoring treatment implementation across clinical scenarios. A better understanding of the presentation-specific barriers and facilitators to buprenorphine initiation may help develop clinical Care pathways strat- egies that are flexible enough to meet the needs of heterogenous patients.

visit characteristics“>In prior work, our team reported on outcomes following implemen- tation of multi-component strategy in three urban, academic EDs that included X-waiver training and education, automated consultation to peer recovery specialists, and culture change approaches. These inter- ventions resulted in significant increases in buprenorphine administra- tion and prescribing, but missed opportunities remained. For this study, we sought to understand how patient clinical factors were associated with differences in MOUD treatment with the goal of identifying targets for subsequent care pathways or interventions. Our aim was to use mixed methods to explore variability in ED care based on clinical pre- sentation for OUD and barriers and facilitators to ED-initiated OUD treatment from clinician perspectives.

1. Methods

We used a mixed methods approach to describe variation in treat- ment for OUD based on ED clinical presentation and barriers and facili- tators to management across clinical scenarios. This is a secondary analysis of previously collected data describing the impact of a multi- component strategy for increasing implementation of ED-initiated treatment of OUD. Details of that intervention have been published pre- viously [14]. For the current study, we used a sequential explanatory mixed methods approach, in which the qualtitative data provide in- depth understanding of the quantitative results [49]. First, we analyzed outcomes among a cohort of patients with OUD presenting to three urban, academic EDs following the implementation of our interventions describing variability in OUD treatment by clinical presentation. Quanti- tative data were then integrated with results from focus groups con- ducted among of a sample of ED physicians and nurses from the same

institution to provide a richer understanding of clinician perspectives on caring for patients with OUD across varying clinical presentations.

• 1. Study setting

The study was conducted in a large, urban, academic health system in Philadelphia, Pennsylvania, which has one of the highest overdose Death rates of any large U.S. city [16]. We analyzED patient visits from EDs in three academic hospitals, including a tertiary referral center, a level 1 trauma center, and another hospital with a psychiatric crisis cen- ter. Together, these 3 EDs receive approximately 120,000 visits annually and >2000 visits for patients with OUD each year. Early in the study period, the study EDs implemented OUD treatment guidelines and a program to incentivize physicians to obtain their DATA 2000 waiver (“X-waiver”), resulting in >90% of physicians being credentialed to pre- scribe buprenorphine [17]. Study EDs also incorporated a team of peer recovery specialists to support treatment and referral for patients pre- senting for OUD-related concerns, and patients are referred for OUD fol- low-up to primary care and specialty substance use treatment providers within and outside the health system for longitudinal buprenorphine treatment [14]. The multi-component strategy resulted in increases of 20% in total buprenorphine use, including 13% increase buprenorphine administration in the ED and 14% increase in buprenorphine prescrip- tions at discharge. However, there was still substantial variability across prescribers and patients.

• 1. Selection of patient cohort

For our quantitative analysis, we included visits with an OUD-related ICD-10 code from adult patients (>=18 years old) seen and discharged from the study EDs between December 2018 and July 2020. OUD-re- lated encounters were identified using ICD-10 codes for opioid use disorder and overdose (Appendix) [14]. All health system data were obtained from the electronic health record via Clarity, a reporting database for Epic (Hyperspace 2017; Epic Systems Corporation, Verona, WI).

• 1. Patient and visit characteristics

We characterizED patient encounters in terms of presentation type: overdose, withdrawal, and other based on ICD-10 codes (Appendix). We chose to stratify by presentation type because these clinical scenar- ios often require different management and present unique challenges that must be addressed when developing ED-based OUD interventions. Within each group, we assessed several measures of ED-based OUD treatment. First, we analyzed the rate of MOUD treatment per opioid- related ED encounter, a composite metric that included both buprenor- phine or methadone administration in the ED and/or a buprenorphine prescription at discharge. We also assessed proportions of patients re- ceiving MOUD administered in ED (methadone and/or buprenorphine), a buprenorphine prescription at discharge, and a naloxone prescription at discharge for overdose prevention. In addition,e we measured ED length-of-stay , 30-day ED revisits and 30-day hospital readmis- sions and extracted patient demographic and clinical characteristics including age, sex, race/ethnicity, and insurance status as well as calcu- lating Charlson Comorbidity Index (CCI) based on previously coded diagnoses in patient records.

• 1. Quantitative data analysis

We used descriptive statistics to characterize the sample and com- pared differences in variables between the 3 presentation subgroups using chi-squared tests for categorical variables, Kruskal-Wallis for ordi- nal variables, and analysis of variance for quantitative normal variables. Analyses were conducted using Stata (Version 15.1; StataCorp, College Station, TX) and R statistical software [18].

• 1. Focus group design

Between January and April 2020, we conducted five focus groups with ED physicians and nurses from two of the study EDs. The interview guide focused on participant experiences caring for patients with OUD in the ED, including: 1) typical clinical presentation of patients with OUD, 2) impact of clinical presentation on identification and treatment, and 3) barriers and facilitators to treatment based on clinical scenario. Participants were also asked for feedback on several strategies to help identify and treat patients with OUD; these results are reported else- where [19].

We recruited participants by email. The first two focus groups were conducted in person and final three were conducted via videoconfer- encing platform due to pandemic restrictions. After obtaining informed consent from each participant, focus groups were moderated by trained clinician members of the research team (RM, MKD, ML) who did not work directly with participants. We collected demographic and other participant characteristics through a short survey. Sessions lasted approximately an hour, and participants received a $50 incentive.

• 1. Qualitative data analysis

Focus groups were audio recorded and transcribed by a professional transcription service. Each transcript was independently coded by two trained research assistants (CW, GS) using Dedoose qualitative analysis software [20], and any disagreements were resolved through discussion with a third team member (ML). We analyzed the transcripts using thematic content analysis using a combined inductive and deductive approach that incorporated prespecified and emergent themes [21]. Results from the thematic analysis of the Patient presentations are reported here.

Integration of quantitative and qualitative data was done during analysis and manuscript preparation, using a constant comparative method, by key team members. The study was approved by the Institu- tional Review Board of the University of Pennsylvania.

1. Results
   1. Patient characteristics

There were a total of 1339 OUD-related visits over the study period (Table 1). This included 265 visits for drug overdose (20%), 123 for opi- oid withdrawal (9%), and 951 for other conditions (71%). Other

Table 1

Patient characteristics.

Patient Characteristics Overall

n = 1339

Age, mean (SD) 40.86 (14.2)

Male Gender, n (%) 907 (68%)

Race, n (%)

White 703 (53%)

Black/African American 488 (36%)

Other/Unknown 68 (5%)

Hispanic ethnicity, n (%) 80 (6%)

Insurance Status, n (%)

Medicaid	863 (65%)
Commercial	207 (16%)
Medicare	159 (12%)
Other/Uninsured	110 (8%)
Charlson Comorbidity Index, mean (SD)	0.85 (1.9)
ED visits in prior 12 months, mean (SD)	3.2 (8.9)
Hospital admissions in prior 12 months, mean (SD)	0.45 (1.6)
Presentation Overdose	265 (20%)
Withdrawal	123 (9%)
Other	951 (71%)

ED = Emergency department.

conditions were a heterogeneous group, including patients seeking sub- stance use treatment, those with medical complications of their OUD such as infections, or a variety of routine clinical scenarios unrelated to OUD. The majority of patients were male, middle-aged, and publicly insured. 53% of patients identified as White and 36% identified as Black.

• 1. Visit and treatment characteristics by clinical presentation

Overall, 23% of patients over the study period received MOUDs dur- ing or after their ED visit (Table 2). The most common MOUD was bu- prenorphine. 21% received buprenorphine during and/or after their ED visit, with 15% being administered buprenorphine in the ED and 15% re- ceiving a prescription at discharge. MOUD receipt differed significantly by presentation type; treatment with MOUDs was least common with overdose presentations (6%), followed by other presentations (21%), and most common in presentations for withdrawal (69%; p < 0.001). 63% of patients with withdrawal presentations received either bupre- norphine or methadone in the ED. Buprenorphine was prescribed at dis- charge in 15% of visits overall, including for 42% of withdrawal visits, 14% of other OUD-related visits, and 5% of overdose visits (p < 0.001). Participants prescribed naloxone at discharge in 31% of visits, including 45% of overdose visits, 37% of withdrawal visits, and 27% of other visits (p < 0.001).

ED length of stay averaged 5.3 h and did not significantly differ by presentation type. 35% of patients had an ED revisit and 6.6% had a hos- pital readmission within 30 days of the of the index visit, with these rates differing across presentations (p < 0.001). Revisits and readmis- sions were most common among those in the “other” category.

• 1. Qualitative analysis

We supplemented quantitative data with qualitative data from focus groups with ED physicians and nurses. Participants included 28 clini- cians in two of the study EDs, including attending physicians (n = 9), resident physicians (n = 10), and nurses (n = 9) (Table 3). Focus group participants were primarily male (57%) and white (82%) 72% of participants had ordered (physicians) or administered (nurses) bupre- norphine in the ED. Of the physicians, all attending physicians had ob- tained an X-waiver, and 24% reported prescribing buprenorphine at ED discharge to one or more patients. Across clinical presentations, we explored the themes of clinician confidence in identification and treat- ment of OUD, ED clinical management, and clinician-perceived patient receptivity to treatment. We also highlight themes that emerged from clinician focus groups that were not captured in quantitative data, par- ticularly challenges with patients presenting with acute or chronic pain.

• • 1. Overdose presentations

Participants expressed comfort identifying patients with opioid overdose, mentioning both the clinical acuity and large volume of pa- tients presenting with overdose as facilitators. Participants reported high levels of comfort and motivation to discuss OUD with patients fol- lowing overdose, both because the diagnosis felt obvious and the grav- ity created a reachable moment. One noted:

“It’s easy when they’ve just had an overdose and they’ve nearly died.“

(Attending physician)

Despite perceived receptivity, participants noted that overdose sur- vivors were rarely clinically ready to start MOUDs, particularly bupre- norphine. One reported experiencing this barrier in initiating treatment for a patient following an overdose:

“I wanted to help, [but] unfortunately he was still high… you have to wait for them to start withdrawing.” (Attending physician)

Table 2 Differences in treatment by clinical presentation.
	Overall		OVERDOSE		WITHDRAWAL		OTHER
	n = 1339		n = 265		n = 123		n = 951	p-value
Total MOUD, n (%)	302 (23%)		16 (6%)		85 (69%)		201 (21%)	<0.001
Any MOUD administered in the ED, n (%)	215 (16%)		4 (1.5%)		78 (63%)		133 (14%)	<0.001
Total buprenorphine, n (%)	282 (21%)		15 (6%)		81 (66%)		186 (20%)	<0.001
Buprenorphine administered in the ED, n (%)	195 (15%)		3 (1%)		74 (60%)		118 (12%)	<0.001
Buprenorphine prescribed at discharge, n (%)	202 (15%)		14 (5%)		51 (42%)		127 (14%)	<0.001
Methadone administered in the ED, n (%)	20 (1.5%)		1 (0.4%)		4 (3%)		15 (1.6%)	0.087
Naloxone administered in the ED, n (%)	63 (4.7%)		27 (10%)		0 (0%)		36 (4%)	<0.001
Naloxone prescribed at discharge, n (%)	417 (31%)		120 (45%)		45 (37%)		252 (27%)	<0.001
ED LOS in hours, mean (SD)	5.3 (4.8)		5.3 (3.8)		5.3 (5.1)		5.3 (5.0)	0.972
ED Revisit within 30 days, n (%)	467 (35%)		63 (24%)		44 (36%)		360 38%)	<0.001
Hospital readmission within 30 days, n (%)	87 (6.5%)		8 (3%)		1 (0.8%)		78 (8.2%)	<0.001

MOUD = Medication for opioid use disorder; ED = Emergency department; LOS = length of stay.

Another echoed patient clinical status as a barrier to MOUD initiation:

“Do we start talking about a long-term – a good plan? The times I have, they’re genuinely not interested or very defensive because they’re not in withdrawal. They’re not really seeking help for that.” (Attending physician)

Finally, one participant shared:

“I go months without seeing one of these patients that’s perfect for buprenorphine.” (Resident physician)

Participants also highlighted heterogeneity among patients present- ing with overdose and perceieved that patients who experienced

Table 3

Focus group participant characteristics.

Participant Characteristics n, %

multiple overdoses may not be as receptive to treatment as those who overdosed for the first time.

“I find that a lot of my older patients in their 50’s who have had near or multiple overdoses are just ‘I’ve been through this song and dance be- fore. I don’t want to talk to you about it.‘ And it’s hard to get through to them, and it’s hard for social work to get through to them. So, I find that it’s the first-time overdose, the 20-year-old that I find to make a big impact.” (Attending physician)

Overall, participants expressed confidence in identifying and treating opioid overdose, although initiation of MOUDs was limited by participant assessment of patients’ clinical readiness. Despite feeling like overdose was a potentially reachable moment, participants re- ported barriers in engaging with patients in treatment, especially after repeated overdose.

• • 1. Withdrawal presentations

Participants also expressed high levels of comfort in recognizing opi- oid withdrawal. Well-described clinical signs and symptoms as well as overt patient disclosure facilitated diagnosis. However, non-specific presentations such as undifferentiated abdominal pain were mentioned as less readily identifiable. In addition, participants described mixed success in MOUD initiation for this population. Participants felt rela-

Professional Group
Attending Physician	9 (32%)
Resident Physician	10 (35%)
Nurse	9 (32%)

Age		tively confident in offering MOUDs for patients in withdrawal because
<30	9 (32%)	they provide symptomatic relief and patients were receptive to treat-
30-39	12 (43%)	ment. In reference to engaging patients in long term treatment, one

usly

40-49	6 (21%) participant noted:
>50	1 (4%)
Gender
Female	12 (43%) “I feel like we make a difference…[in] the ones that come in obvio
Male Race	16 (57%) saying I’m here because I’m in withdrawal” (Attending physician)

White	23 (82%)
Black/AA	1 (4%)	However, participants still cited barriers. For patients who presented
Asian/Pacific Islander	3 (14%)	for medical workup, there were challenges in forming a therapeutic al-
Hispanic/Latino Ethnicity Years since graduation from medical or nursing school 1-4	1 (4%) 12 (43%)	liance and having meaningful conversations regarding treatment. This was exacerbated by inadequate symptom control or non-specific
5-9	7 (25%)	presentations that delayed identification.
10-14	3 (11%)
15+	6 (21%)
Percent time spent in clinical care		“Usually by the time you get to the end of doing the medical workup,

<20 1 (4%)

20-50 1 (4%)

51-75 3 (11%)

75+ 23 (82%)

Has X-waiver (physicians)

Yes 7 (37%)

No 9 (47%)

Completed training, no DEA? 3 (16%)

* Resident physicians do not all have a DEA number.

they’re so disgruntled by the fact that you’ve allowed to be withdrawal that long that they’re a much harder population to help.” (Attending physician)

Participants also spoke extensively about patients who were inter- ested in treatment but whose withdrawal was not severe enough to safely initiate buprenorphine. This led to long waits and eventually

resulted in patients leaving due to the extensive time commitment and physical discomfort of increasing withdrawal.

“It’s rare that I – at least personally – that I have a patient who is with- drawing enough to start on bup[renorphine]. Most of them, unfortu- nately, I feel come in and are still not somebody you would initiate the bup[renorphine]” (Attending physician)

Overall, participants felt most confident in starting buprenorphine in patients presenting with withdrawal and recognized that these visits as an opportunity to initiate treatment. However, the physical discomfort of withdrawal and initial lack of adequate Withdrawal symptoms were still substantial barriers to care.

• • 1. Patients with other presentations

Among those patients with OUD-related visits not related to over- dose or withdrawal, presentations and participant experiences were variable. Some patients came specifically seeking substance use treat- ment, either for inpatient services or MOUD initiation. Others presented with medical complications of their opioid use such as infection, and some presented for a broad range of other Medical issues Participants highlighted different management challenges for each group.

When patients who came seeking substance use treatment, includ- ing short- or long-term inpatient treatment, participants described increased motivation and ease in supporting them. One shared:

“…some are very explicit. Like I have opioid use disorder and I need help, and I’m super collaborative and I get excited about using my X-waiver and I do it in house and discharge and connect them.” (Attending physician)

However, many participants indicated that efforts to link to inpa- tient treatment were limited by a lack of resources and patient social needs. One shared an illustrative example of a typical patient challenge:

“…and then like you can come back in the morning, use our phone if you need to call this number, or you can go to somewhere else, but we don’t have detox available here. It’s the middle of the night. They say well, I don’t have a phone. How am I going to get back?” (Nurse)

While interested in connecting patients to resources, participants felt frustrated by limitations and the differences between available re- sources and patient expectations. Notably, participants perceived that many patients were primarily seeking inpatient detox or long-term res- idential treatment and less interested in outpatient MOUDs.

Participants also described many patients presenting with complica- tions of OUD such as skin and soft tissue infections. They noted that this group of patients would frequently be admitted, and while life-threat- ening complications were a motivator to engage with medical care and substance use treatment, participants reported challenges in man- agement of acute pain in patients with OUD. Clinician participants expressed significant ambivalence around acute pain treatment, partic- ularly those with repeat presentations. One shared:

“…people that constantly come in and out of the ED, how do we help them by not just giving them tons of drugs and then sending them on their way” (Nurse)

There was also uncertainty about how to provide care for patients with OUD and acutely painful conditions and a recognition that this could be a source of anxiety and mistrust among patients seeking care. One participant reported:

“They’re here for another medical reason and [OUD is] part of their med- ical history, but not why they’re here, which can be challenging because

historically, I don’t think the medical system has done a great job of car- ing for these patients…And so, they’re here for their ankle fracture and they’re worried about how they’ll be perceived and treated and how that part of their care will be managed.” (Attending physician)

Participants noted that for many patients with OUD, mistrust in the medical system and clinician uncertainty on how to adequately manage OUD concurrently with other medical problems led to challenges in dis- cussions of long-term care as well as patient-directed discharges.

• • 1. Patients with chronic pain

Another challenging area reported by focus group participants was patients where an OUD diagnosis was less clear, such as those with chronic pain or treated with long-term opioid therapy. Given the evolv- ing nature of chronic pain treatment and the sometimes unclear line be- tween chronic opioid use and OUD, participants often felt unprepared to broach the subject. One reported:

“I have been finding there’s a broader population there, or a little grayer and fuzzier about whether we bring it up, whether we intervene, whether we do something. We see just coming in for some sort of chronic pain.” (Attending physician)

Several participants felt that some chronic pain conditions were not legitimate, or perceived patients on chronic opioids to be drug seeking. One shared:

“I think you have a lot of very significant people who use a previous ex- cuse of whatever the source for the chronic pain is as a crutch to stay on the medicine, and that’s a difficult patient population.” (Attending physician)

Others recognized the impact of participant stigma on management decisions for this population and advocated for a more proactive approach. One reported:

“A lot of what I see is more of the labeling as drug seeking and not prop- erly trying to get them into different pain management programs that can better manage their pain other than opioids or the like.” (Nurse)

Conversations aimed at diagnosing OUD versus Opioid dependence in patients on long-term opioid therapy were viewed as time-consum- ing and challenging, and participants often avoided these conversations amidst other competing priorities. Given the sometimes uncertain diag- nosis of OUD, there was also discomfort in broaching OUD treatment in this patient population. One participant shared,

“It’s a lot more difficult if they’re there for back pain and you perceive that the real issue may be opioid dependence to mount the enthusiasm and the time that it takes to really sit down and have that conversation with somebody. Because otherwise, those are pretty quick encounters.” (Attending physician)

Participants who were able to engage in these conversations cited the importance of being open and honest with patients and would ask them directly about their own interest in reducing opioid use.

“I’m pretty honest with patients like hey, I looked through your chart. I saw this on your chart. What’s the status of that?” (Nurse)

Participants advocated for clearer guidelines and resources for treat- ment of chronic pain in patients on long term opioid therapy who pres- ent to the ED. There was a recognition that some patients might benefit

from OUD treatment but lacked clarity on how to determine this and which patients would be appropriate.

“What are the resources that we have that we can use for the cancer pa- tient on chronic opioids versus the patient who might be a candidate and may be a willing candidate for treatment for OUD, and maybe hav- ing two separate pathways or different pathways to be able to go down for those.” (Resident physician)

1. Discussion

In this mixed methods study focused on ED OUD care, we found sig- nificant variability in ED-initiated treatment based on the patient clini- cal presentation. We saw missed opportunities for MOUD initiation across all presentations, particularly those presenting after non-fatal overdose. Focus group discussions contextualized findings, and we found variability in clinician confidence, patient readiness, and clini- cian-perceived patient receptivity to treatment across clinical scenarios. Our mixed methods approach to highlights specific implementation gaps and demonstrates that ED OUD treatment is not a one-size-fits- all approach. Prior studies have described numerous barriers to ED-ini- tiated buprenorphine including Patient engagement, clinician comfort with medication, perceived patient motivation for treatment, and link- age to next stages of care [22,23]. Our results suggest that incorporating tailored guidance based on clinical scenarios, such as treatment of opi- oid withdrawal, management of acute pain, or initiating buprenorphine in patients who are not in withdrawal in the ED – might further close the gaps in treatment.

These results add to the literature in several keys ways. First, we found that while OUD treatment varied by clinical presentation, there were missed opportunities across all scenarios. Prior work has demon- strated that fewer than 5% of commercially insured patients had a claim for an MOUD within 90 days of an ED-visit for a nonfatal overdose [24]. Although rates in this study were higher than those reported in na- tional data, there were still substantial treatment gaps, especially given the 5% mortality rate in the year following a non-fatal overdose [25,26]. Participants noted varying degrees of confidence in a range of important aspects of OUD care including patient identification, communication, and management based on clinical presentation. Some of these differ- ences may be explained by patient factors, such as interest in treatment or social context. However, prior studies have shown substantial vari- ability among clinicians, suggesting that participants approach these same clinical scenarios differently [14]. Proposed strategies to reduce variability include implementing recovery coaches or navigators to facilitate patient engagement, tailored electronic health record Clinical decision support, more comprehensive screening and identification, or other design approaches to influence ED clinician behavior [14,19,27,28]. Our findings suggest that strategies for ED OUD treatment may be most effective if they can incorporate the variability in clinical presentations and differing management needs among patients with OUD.

In addition, our data suggest that clinicians may not always recog- nize opportunities to initiate treatment. Participants described chal- lenges finding the “perfect patient” even when they felt confident and motivated to treat patients with OUD. In particular, overdose visits were perceived as reachable moments, yet we saw low rates of treat- ment initiation and heard that patients were not clinically ready for bu- prenorphine inductions. Similarly, participants expressed frustration with lack of resources for patients seeking intpatient treatment services but may not have recognized this population as potential candidates for ED buprenorphine. One potential strategy to overcome this gap in care is wider implementation of home or off-site induction from the ED.

[29] This practice is safe, effective, and commonly used in outpatient treatment settings [30,31]. Education, best-practice guidelines, and

institutional protocols may be needed to support off-site inductions when patients are not clinically ready in the ED.

Additionally, ED clinicians can consider novel strategies such as low- dose or ultra-low dose inductions [32,33], sometimes called microinductions, in which low doses of buprenorphine are initiated for patients not yet in withdrawal [34,35]. There are also high-dose in- ductions, in which higher starting doses of buprenorphine are used than for traditional induction strategies, that have been shown to be safe and well-tolerated in an ED context [36]. While low-dose ap- proaches have not been well-studied in ED or outpatient populations, future studies should consider testing this approach in ED patients pre- senting for care who are not clinically ready for buprenorphine induc- tion using Traditional methods. Alternative induction strategies are increasingly relevant given reports that fentanyl, the dominant opioid in the U.S. drug supply, is creating challenges with buprenorphine in- ductions following standard approaches [37,38]. Finally, although much of the focus has been on ED-initiated buprenorphine, there is in- creasing interest in ED-initiation of methadone. While outpatient meth- adone can only be provided in licensed opioid treatment programs (OTPs) [39], short-term methadone administration can be done in ED and hospital settings under the “72-hour rule” and then patients can be handed off to OTPs [39]. Although this is rarely done in practice, ex- ploratory studies have found ED initiation of methadone to be desired by patients [40], and the practice was feasible from a hospital-based bridge clinic [41]. Further work developing, implementing, and testing ED-initiated methadone treatment and direct OTP referral may be a promising avenue to create more patient-centered models of care and overcome some of the challenges associated with buprenorphine.

Another key finding was the challenge of managing acute and

chronic pain conditions in the context of potential OUD. Participants felt ill-equipped to discuss the intersection of OUD and chronic pain and poorly prepared to manage acute pain in patients with OUD. Prior studies have cited unmanaged pain as a driver of patient-directed dis- charges in patients with OUD [42], and despite the existence of guide- lines [43], clinicians may feel ill-equipped to manage acute pain in patients with OUD, especially if pain management involves opioid ago- nist treatment. Further, while chronic pain is a common driver of ED visits [44], it is not always prioritized by ED clinicians [45]. Our results suggest that implementation and dissemination efforts for ED-initiated of OUD treatment may need to include explicit guidance around discussing and managing pain in patients with OUD. This may need to include the use of short-acting opioids for treatment of acute pain and withdrawal given some of the limitations of MOUDs for rapid, effective pain and withdrawal management [46,47].

Our study has several limitations. First, we use ICD-10 to identify and categorize OUD diagnoses, which may not accurately capture all pa- tients with OUD [48]. Second, study EDs are highly engaged and well- resourced for OUD treatment compared to many settings and results are based on local context and drug supply at the time the study was conducted, so findings may not generalize to all locations or settings. Fi- nally, although participants discuss perceptions of patient perspectives, our study lacks the important perspective of patients themselves.

1. Conclusions

In conclusion, we found high rates of variability in ED-initiated OUD care based on clinical presentation and highlight nuances in ED-initi- ated OUD treatment. Using mixed methods and implementation science approaches, our results highlight the need for tailored clinical guidance and can inform health system and policy interventions seeking to in- crease ED-initiated treatment for OUD.

Author contributions

Study concept and design: SF, MKD, ML. Acquisition of the data: ML, MKD.

Analysis and interpretation of the data: SF, MKD, RAX, CW, GS, ML. Drafting of the manuscript: SF, ML. Critical revision of the manuscript for important intellectual content:

MKD, JP, RM, NO.

Statistical expertise: RAX. Acquisition of funding: ML, MKD.

Financial support

This work was supported by the Penn Injury Science Center (CDC 19R49CE003083). Dr. Lowenstein was also supported by the National Institute on Drug Abuse (grant K23DA055087).

CRediT authorship contribution statement Sophia Faude: Formal analysis, Writing – original draft, Conceptual-

ization. M. Kit Delgado: Conceptualization, Formal analysis, Funding ac-

quisition, Writing – review & editing. Jeanmarie Perrone: Writing – review & editing. Rachel McFadden: Investigation, Writing – review & editing. Ruiying Aria Xiong: Formal analysis, Writing – review & editing. Nicole O’Donnell: Writing – review & editing. Christian Wood: Data curation, Writing – review & editing. Gabrielle Solomon: Data curation, Formal analysis, Writing – review & editing. Margaret Lowenstein: Conceptualization, Data curation, Funding acquisition, Supervision, Writing – review & editing.

Declaration of Competing Interest

SF, MKD, JP, RM, RAX, NO, CW, GS and ML report no conflicts of interest.

MKD, JP, RM, and ML report grant support from the Centers for Disease Control for this research. MKD received grant support from PCORI and NIH outside the submitted work.

JP received grant support from NIH, City of Philadelphia, SAMHSA, and Independence Blue Cross outside the submitted work. ML received grant support from NIH, the City of Philadelphia, and the Pew Charitable Trusts outside the submitted work.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2023.01.009.

References

1. Provisional Drug Overdose Death Counts. Accessed November 2022, at https:// www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm; 2022.
2. Mattick RP, Breen C, Kimber J, Davoli M. Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database Syst Rev. 2014:Cd002207.
3. Sordo L, Barrio G, Bravo MJ, et al. Mortality risk during and after opioid substitution treatment: systematic review and meta-analysis of cohort studies. BMJ (Clinical research ed). 2017;357:j1550.
4. National Academies of Sciences E, Medicine. Medications for opioid use disorder save lives. Washington, DC: The National Academies Press; 2019.
5. Wakeman SE, Larochelle MR, Ameli O, et al. Comparative effectiveness of different treatment pathways for opioid use disorder. JAMA Netw Open. 2020;3:e1920622.
6. Biondi BE, Zheng X, Frank CA, Petrakis I, Springer SA. A literature review examining primary outcomes of medication treatment studies for opioid use disorder: what outcome should be used to measure opioid treatment success? Am J Addict. 2020; 29:249-67.
7. Krawczyk N, Rivera BD, Jent V, Keyes KM, Jones CM, Cerda M. Has the treatment gap for opioid use disorder narrowed in the U.S.?: A yearly assessment from 2010 to 2019. Int J Drug Policy. 2022;110:103786.
8. Vivolo-Kantor AMSP, Gladden RM, et al. Vital signs: trends in emergency depart- ment visits for suspected opioid overdoses — United States, July 2016-September 2017. MMWR Morb Mortal Wkly Rep. 2018;67:279-85.
9. D’Onofrio G, Chawarski MC, O’Connor PG, et al. Emergency department-initiated buprenorphine for opioid dependence with continuation in primary care: outcomes during and after intervention. J Gen Intern Med. 2017;32:660-6.
10. Busch SH, Fiellin DA, Chawarski MC, et al. Cost-effectiveness of emergency depart- ment-initiated treatment for opioid dependence. Addiction. 2017;112:2002-10.
11. Hawk K, Hoppe J, Ketcham E, et al. Consensus recommendations on the treatment of opioid use disorder in the emergency department. Ann Emerg Med. 2021; 78 (3):434-442.
12. Martin A, Mitchell A, Wakeman S, White B, Raja A. Emergency department treat- ment of opioid addiction: an opportunity to lead. Acad Emerg Med Acad Emerg Med
13. D’Onofrio G, McCormack RP, Hawk K. Emergency departments – a 24/7/365 option for combating the opioid crisis. N Engl J Med. 2018;379:2487-90.
14. Lowenstein M, Perrone J, Xiong RA, et al. Sustained implementation of a multicom- ponent strategy to increase emergency department-initiated interventions for opi- oid use disorder. Ann Emerg Med. 2022;79(3):237-248.
15. Duncan A, Anderman J, Deseran T, Reynolds I, Stein BD. Monthly patient volumes of buprenorphine-waivered clinicians in the US. JAMA Netw Open. 2020;3:e2014045.
16. Unintentional Drug Overdose Fatalities in Philadelphia, 2020. Accessed October, 2021, at. https://www.phila.gov/media/20210603100151/CHARTv6e5.pdf; 2021.
17. Foster SD, Lee K, Edwards C, et al. Providing incentive for emergency physician X- waiver training: an evaluation of program success and postintervention buprenor- phine prescribing. Ann Emerg Med. 2020;76(2):206-14.
18. Team RC. R: A language and environment for statistical computing. Vienna, Austria: Computing RFfS; 2018.
19. Lowenstein M, McFadden R, Abdel-Rahman D, et al. Redesign of opioid use disorder screening and treatment in the ED. NEJM Catalyst. 2022;3.
20. Dedoose Version 7023 web application. Socio Cult Res Consult. 2016;LLC.
21. Pope CM. Nicholas, Qualitative Research in Health Care. Hoboken, NJ: BMJ Books; 2000.
22. Lowenstein M, Kilaru A, Perrone J, et al. Barriers and facilitators for emergency department initiation of buprenorphine: a physician survey. Am J Emerg Med. 2019;37(9):1787-90.
23. Hawk KF, D’Onofrio G, Chawarski MC, et al. Barriers and facilitators to clinician read- iness to provide emergency department-initiated buprenorphine. JAMA Netw Open. 2020;3:e204561.
24. Kilaru AS, Xiong A, Lowenstein M, et al. Incidence of treatment for opioid use disor- der following nonfatal overdose in commercially insured patients. JAMA Netw Open. 2020;3:e205852.
25. Schoenfeld EM, Westafer LM, Soares 3rd WE. Missed opportunities to save lives- treatments for opioid use disorder after overdose. JAMA Netw Open. 2020;3: e206369.
26. Weiner SG, Baker O, Bernson D, Schuur JD. One-year mortality of patients after emergency department treatment for nonfatal opioid overdose. Ann Emerg Med. 2020;75:13-7.
27. Beaudoin FL, Jacka BP, Li Y, et al. Effect of a peer-led behavioral intervention for emergency department patients at high risk of fatal opioid overdose: a randomized clinical trial. JAMA Netw Open. 2022;5:e2225582.
28. Butler K, Chavez T, Wakeman S, et al. Nudging emergency department-initiated addiction treatment. J Addict Med. 2022;16 (e234-e9).
29. Regan S, Howard S, Powell E, et al. Emergency department-initiated buprenorphine and referral to follow-up addiction care: a program description. J Addict Med. 2022; 16:216-22.
30. Lee JD, Grossman E, DiRocco D, Gourevitch MN. Home buprenorphine/naloxone induction in primary care. J Gen Intern Med. 2009;24:226-32.
31. Martin SA, Chiodo LM, Bosse JD, Wilson A. The next stage of buprenorphine Care for Opioid use Disorder. Ann Intern Med. 2018;169:628-35.
32. De Aquino JP, Parida S, Sofuoglu M. The pharmacology of buprenorphine microinduction for opioid use disorder. Clin Drug Investig. 2021;41:425-36.
33. Ahmed S, Bhivandkar S, Lonergan BB, Suzuki J. Microinduction of buprenorphine/ naloxone: a review of the literature. Am J Addict. 2021;30:305-15.
34. Spadaro A, Long B, Koyfman A, Perrone J. Buprenorphine precipitated opioid with- drawal: prevention and management in the ED setting. Am J Emerg Med. 2022; 58:22-6.
35. Moe J, O’Sullivan F, Hohl CM, et al. Short communication: systematic review on effectiveness of micro-induction approaches to buprenorphine initiation. Addict Behav. 2021;114:106740.
36. Herring AA, Vosooghi AA, Luftig J, et al. High-dose buprenorphine induction in the emergency Department for Treatment of opioid use disorder. JAMA Netw Open. 2021;4:e2117128.
37. Silverstein SM, Daniulaityte R, Martins SS, Miller SC, Carlson RG. “everything is not right anymore”: buprenorphine experiences in an era of illicit fentanyl. Int J Drug Policy. 2019;74:76-83.
38. Varshneya NB, Thakrar AP, Hobelmann JG, Dunn KE, Huhn AS. Evidence of buprenor- phine-precipitated withdrawal in persons who use fentanyl. J Addict Med. 2022;16 (e265-e8).
39. 21 CFR 1306.07. Administering or dispensing of narcotic drugs. Accessed October 2022, at https://www.ecfr.gov/current/title-21/chapter-II/part-1306/subject-group- ECFR1eb5bb3a23fddd0/section-1306.07.
40. Schoenfeld EM, Westafer LM, Beck SA, et al. “just give them a choice”: Patients’ per- spectives on starting medications for opioid use disorder in the ED. Acad Emerg Med. 2022;29:928-43.
41. Taylor JL, Laks J, Christine PJ, et al. Bridge clinic implementation of “72-hour rule” methadone for opioid withdrawal management: impact on opioid treatment pro- gram linkage and retention in care. Drug Alcohol Depend. 2022;236:109497.
42. Simon R, Snow R, Wakeman S. Understanding why patients with substance use dis- orders leave the hospital against medical advice: a qualitative study. Subst Abus. 2020;41:519-25.
43. The ASAM. NATIONAL PRACTICE GUIDELINE for the treatment of opioid use disor- der. Accessed October 3, 2022, at https://sitefinitystorage.blob.core.windows.net/ sitefinity-production-blobs/docs/default-source/guidelines/npg-jam-supplement. pdf?sfvrsn=a00a52c2_2; 2020.
44. Milbrett P, Halm M. Characteristics and predictors of frequent utilization of emer- gency services. J Emerg Nurs. 2009;35:191-8. (quiz 273).
45. Wilsey BL, Fishman SM, Ogden C, Tsodikov A, Bertakis KD. Chronic pain manage- ment in the emergency department: a survey of attitudes and beliefs. Pain Medicine (Malden, Mass). 2008;9:1073-80.
46. Thakrar AP. Short-acting opioids for hospitalized patients with opioid use disorder. JAMA Intern Med. 2022;182:247-8.
47. Kleinman RA, Wakeman SE. Treating opioid withdrawal in the hospital: a role for short-acting opioids. Ann Intern Med. 2022;175:283-4.
48. Coupet Jr E, D’Onofrio G, Chawarski M, et al. Emergency department patients with untreated opioid use disorder: a comparison of those seeking versus not seeking referral to substance use treatment. Drug Alcohol Depend. 2021;219:108428.
49. Palinkas LA, Aarons GA, Horwitz S, Chamberlain P, Hurlburt M, Landsverk J. Mixed method designs in implementation research. Administration and policy in mental. health. 2011;38:44-53.