Facilitators and barriers to utilization of medications for opioid use disorder in primary care in South Carolina

Abstract

Objective:

Utilization of medications for opioid use disorder (MOUD) has not been widely adopted by primary care providers. This study sought to identify interprofessional barriers and facilitators for use of MOUD (specifically naltrexone and buprenorphine) among current and future primary care providers in a southeastern academic center in South Carolina.

Method:

Faculty, residents, and students within family medicine, internal medicine, and a physician assistant program participated in focus group interviews, and completed a brief survey. Survey data were analyzed quantitatively, and focus group transcripts were analyzed using a deductive qualitative content analysis, based upon the theory of planned behavior.

Results:

Seven groups (N = 46) completed focus group interviews and surveys. Survey results indicated that general attitudes towards MOUD were positive and did not differ significantly among groups. Subjective norms around prescribing and controllability (i.e., beliefs about whether prescribing was up to them) differed between specialties and between level of training groups. Focus group themes highlighted attitudes about MOUD (e.g., “opens the flood gates” to patients with addiction) and perceived facilitators and barriers of using MOUD in primary care settings. Participants felt that although MOUD in primary care would improve access and reduce stigma for patients, prescribing requires improved provider education and an integrated system of care.

Conclusions:

The results of this study provide an argument for tailoring education to specifically address the barriers primary care prescribers perceive. Results promote the utilization of active, hands-on learning approaches, to ultimately promote uptake of MOUD prescribing in the primary care setting in South Carolina.

Introduction

In the United States (US), the opioid crisis has created alarm with nearly 48,000 deaths involving opioid overdose in 2017.^1,2 The 2017 opioid-related overdose death rate (15.5 per 100,000) in South Carolina (SC) was higher than the national rate (14.6 per 100,000),^3,4 and SC residents fill opioid prescriptions at 1.5 times the national average.⁵ In 2017, only 13.7% of those with opioid use disorder received treatment within a substance treatment program in SC.⁶ The downstream consequences of untreated opioid use disorder are significant and include higher morbidity and a tenfold increase in mortality, as well as financial strain on the healthcare system.^2,7,8 Of the 14 states that did not expand Medicaid programs, 50% are located in the Southeastern US, making access to care for certain populations potentially challenging in this region.^9,10 Nearly 80% of patients with opioid use disorder do not receive treatment with the FDA-approved medications for opioid use disorder (MOUD): naltrexone, buprenorphine, or methadone.^11,12

The primary care setting provides a unique opportunity to detect and treat opioid use disorder. Many patients present to primary care as their first point of contact for treatment, and primary care providers (PCPs) are more available to the general population than specialists.^13–15 When substance treatment is combined with primary care, additional medical issues may be identified and treated, which has been associated with decreased costs and higher rates of remission from substance use.^16,17 Primary care has demonstrated an ability to engage patients with higher self-stigma in substance abuse treatment, and patients with previously untreated substance use disorders may prefer primary care to specialty addiction treatment.^18,19 When it comes to treatment of opioid use disorder, primary care has similar retention rates compared to those in specialty care.^20–22 There has been a call for increasing MOUD availability in primary care to increase treatment capacity.^14,23 However, research suggests that MOUD is under-utilized by PCPs.²⁴ For instance, only 2.2% physicians in the US are waivered to prescribe buprenorphine and of those, less than 37% were of a primary care specialty.²⁴ Studies have examined primary care perspectives regarding MOUD and have described varied attitudes and barriers to better understand adoption of practice, though few if any have been conducted specifically in the southeastern US.^25–30

Following the 2018 Opioid Abuse Prevention Study by the SC House of Representatives, there has been a call for academic medical centers to increase training on treatment for opioid use disorder and pain as one of many efforts to attenuate the SC opioid crisis.⁵ This study aims to provide modern perspectives about facilitators and barriers to MOUD in primary care settings from current and future PCPs in an academic medical center in SC. Findings may inform future training and education.

Methods

Participants

Participation was elicited via e-mail from academic medical center faculty providers with teaching responsibilities, residents, and students who were, or anticipated being, in primary care.³¹ All participants engaged in a focus group separated by discipline/training: Family Medicine (FM) faculty, Internal Medicine (IM) faculty, Physician Assistant (PA) faculty, FM residents, IM residents, IM/Psychiatry (IMP) residents, and PA students; with a 10-participant limit to each group (a convenience sample). Volunteers provided written informed consent and received $40 gift cards for their time.

Procedures

Participants engaged in a 60-minute focus group consisting of a survey (15-minutes) and semi-structured interview (45-minutes) on perspectives on prescribing MOUD.^32–36 The university’s institutional review board (Pro00067643) reviewed and approved all procedures as exempt from review. The focus groups occurred 9/2017–2/2018 in private academic space.

Survey and interview development.

Established guidelines for developing surveys and semi-structured interviews from the Theory of Planned Behavior (TPB) framework were utilized.²² TPB is an evidence-based behavior change theory used to understand medical provider perspectives and design strategies to increase evidence-based practices.^37,38 TPB posits that enhancing provider attitudes towards a practice, subjective norms (social pressure to perform/not perform a behavior), and perceived behavioral control (confidence about performing a behavior), will increase intention to perform an actual behavior. This framework is useful for understanding factors affecting use of MOUD in primary care.

Quantitative data

Surveys.

Participants completed an 18-item Likert-scale survey (designed using TPB guidelines for quantitative direct measures) evaluating perspectives about MOUD utilization in primary care (interview guides and survey available upon request).³⁷

Analysis.

Multiple imputation was used to address item-level missing data (0.6% values missing; missing completely at random via Little’s MCAR test).^39–41 Analyses and parameter estimation were conducted using pooled values of the three datasets. Averages for responses to close-ended questions about TPB variables were calculated. Analysis of residuals revealed that normality could not be assumed; non-parametric Kruskal-Wallis tests were used to determine group differences. A general linear factorial model with gamma distribution and identity link function examined the role of the TPB factors and basic demographics in predicting intent to prescribe MOUD (averaged across buprenorphine and naltrexone). SPSS v24 (IBM, Armonk, NY) was used for all analyses.

Qualitative data

Semi-structured interview.

TPB qualitative elicitation study guidelines were used to design a semi-structured focus group interview guide educing perspectives on implementation of MOUD.^{32–34,36,37} Two female facilitators who were part of a HRSA-funded project to improve education around treating opioid use disorders in primary care conducted each group, taking field notes. A licensed clinical Ph.D. psychologist (L.C.) or a board-certified D.O. physician (K.B.) with experience in semi-structured interviews and qualitative methodology served as primary facilitator; neither had a formal relationship established with participants prior to the study; participants were introduced to the team/purpose of the project prior to the group. An IMP resident (S.O.) or a FM research coordinator (C.B.) were co-facilitators; nobody else was present.

Groups were audio-recorded and transcribed verbatim with identifying information removed using Nuance Dragon Professional Individual English version 15 (Nuance, Burlington, MA) software. Project staff (C.B.) reviewed and verified transcriptions.

Analysis.

A phased qualitative content analysis was performed on focus group transcripts utilizing deductive analyses.^36,42,43 Team members individually reviewed transcripts/field notes, identified subthemes within general TPB categories based on emergent themes until the team determined saturation occurred, and then met to develop consensus. A codebook was created including definitions of emergent and sub-themes and responses exemplifying each. The finalized code structure (Table 4) was applied to the data in Microsoft Excel, with the meaningful unit of analysis being each participant’s whole conversation turn.⁴⁴ A negotiated consensus approach to coding was utilized; three coders reflecting the different levels of training of participants (K.B. [IM faculty], S.O. [IMP resident], B.P. [MD student]) reviewed all transcripts independently and assigned codes, reviewed by a final coder (L.C.) who identified discrepancies and assigned a summary code. The coding team reviewed the discrepancy resolution and in negotiated consensus, resolved final disagreements resulting in a single agreed-upon application of the final codebook.^44,45

Table 4.

Qualitative themes and illustrative quotes, with frequency of mention in overall conversation turns (N = 731 conversation turns) during focus groups.

Theme #	Theme name Subtheme	n	%	Illustrative quote(s)
	Attitudes towards specific MOUD interventions in primary care
1	Perspectives supportive of prescribing buprenorphine in primary care	43	5.9%	• “I think one of the benefits of Buprenorphine is that it is a partial agonist, so it does alleviate some of the, at least the cravings, withdrawals craving somewhat, and it also, at least depending upon the drug product, has potential for, you know, deterrent of abuse.” (PA faculty)
2	Perceive challenges of prescribing buprenorphine in primary care	43	5.9%	• “So there’s some, I think a lot of misperception in patients and providers. And there’s also a stigma I think a little bit perceived with regards to Buprenorphine. Why do I need to go to an 8-hours class for Buprenorphine, but I don’t for Naltrexone? It must be something about Buprenorphine that’s scary.” (PA faculty) • “I don’t know if this might be inaccurate, but if it’s just Buprenorphine, potentially it could be diverted and sold and kind of used as an opioid type of medication.” (FM resident) • “I guess theoretically those [patients] are also more motivated, saying I’ve already gone through a detox process.” (FM resident)
3	Perspectives supportive of prescribing naltrexone in primary care	29	4.0%	• “I guess theoretically those [patients] are also more motivated, saying I’ve already gone through a detox process.” (FM resident)
4	Perceive challenges of prescribing naltrexone in primary care	20	2.7%	• “Is the Naltrexone addictive in itself? Are there any like cons of taking it long-term?” (asked one PA student) • “Sometimes verifying that they’ve actually been through a treatment program can be difficult, if we’re not sharing records so that may limit my use of Naltrexone… specifically.” (IM faculty)
	Attitudes supportive of MOUD in primary care
5	MOUD treats a disease and is within PCP scope of practice	44	6.0%	• One PA faculty member described the irony of asking a patient to stop opioids when addicted, by drawing a comparison to managing diabetes: “When are going to get off your insulin? Come on, step it up you’ve got to get off that insulin. I mean, I think that is the key, it’s a misperception of this is a disease, versus just a choice.”
6	MOUD leads to improved access to treatment for addiction	39	5.3%	• “There’s, you know, a lot more primary care physicians than there are like pain specialists and things like that… they are also aware of things to watch out for and things like that as well.” (IM/Psych resident) • “I mean primary care, I think, is more accessible, from my standpoint. I can get a new patient in within a day or two versus psychiatry, which I know you guys are majorly backed-up and have limited resources and that kind of thing, so it’s not great but it’s better I think primary care relatively speaking.” (PA faculty)
7	MOUD in PCP office reduces stigma of receiving treatment	16	2.2%	• “And because it is such like a stigma with people who use or overuse, they might be nervous or embarrassed to get help so they have like a family care provider that they’re comfortable with, maybe they’ll be more like open to go talk to them about it because they know them and they won’t feel as embarrassed.” (PA student)
8	Benefit of prescribing MOUD in the context of established relationships and trust with PCPs	19	2.6%	• “I think also like whenever you have a trusting relationship between a physician and a patient, probably withdrawing from opioids and going through this treatment is a really big decision for them, and they may be more likely to make that decision if someone that they trust is talking to them about it, encouraging them.“ (FM resident)
9	MOUD is a life-saving intervention	21	2.9%	• “Yeah, and seeing people who do well, like I think one of the things that changed my perspective was when I worked in this … clinic and you have patients that are stable on [buprenorphine], it changes their lives, it’s like oh wow this is amazing you know.” (IM/Psych resident)
10	MOUD reduces healthcare system burden	9	1.2%	• “I think the hospital would be in agreement, because it might increase our patient population instead of those people who we cut off of pain medicines and then leave the clinic. We would have continuity of care of those patients and I’d assume the emergency departments that we work with would appreciate people not coming in when they are cut off from their medications for all their pain complaints trying to get the next fix. So I think that all sides, all the other providers and the hospital, I think would be in support of it.” (FM resident)
	Expressed challenges towards MOUD use in primary care
11	MOUD “opens the flood gates” to “addicts”	21	2.9%	• “I think the other disadvantage I see is that we don’t want to simply treat opioid use disorder, and so that is one hesitation I think within our practice, because we don’t want to be known for maybe a Center for Opioid use disorder and we want to still be focusing on other comorbidities.” (IM faculty) • “…they also think that if they get their waiver for [buprenorphine] it’s going to open up the floodgates for people calling them wanting treatment… the learning for just routine medical problems the primary care could become compromised because you’re turning into [sic]an addiction type clinic.” (IM/Psych resident)
12	MOUD is “replacing one drug for another”	17	2.3%	• “I think it’s a similar issue with Buprenorphine, how are you going to get off of that? I feel like it’s a step in the right direction, but I’ve also had patients come in and say well when are you planning on getting off of your [buprenorphine]? I have no plan to get off of [buprenorphine]. So you’re just going to go 60 more years on this?” (PA faculty)
13	“Whose job is it?” to prescribe MOUD	42	5.7%	• “And some of the sub-specialists that have people with chronic pain. They would have a built in gain from it, I don’t think that they necessarily are nasty, I think they just have a blind spot. I think they believe they are doing good work and this is what I do and this is what, but it doesn’t work, it’s just kicking the can down the road, so there’s a problem with it.” (FM faculty) •“I’d be willing to take a heavier load, but we’ve got to have an understanding of how there’s an exchange here going, it’s not just more dumped on primary care…” (FM faculty)
14	“Pills without skills is a fool’s errand” (there is a need for comprehensive treatment including therapy in addition to MOUD)	23	3.1%	• “…Pills without skills is a fool’s errand. The medicines are fine, they’ll help, but if you don’t have a context for treating, you’re wasting your time.” (FM faculty) • “I think in primary care there needs to be, to build a more multidisciplinary approach at baseline because even it’s not opioid use disorder, psychosocial issues, which are chronic, affect my ability to do your hypertension, your diabetes. If you PHQ-9 is 23 I can’t fix your diabetes and then you, it is time constrained and you can’t just throw pills at it.” (IM resident)
15	MOUD requires all providers in practice are on board	36	4.9%	• “And I think because we work in a group practice, we would, even in our own private panels, I feel that, we would probably all need to be comfortable doing it, because we’re going to be, you know, out sick, on vacation, different things are going to come up, so coverage would be an issue and we would all need to be comfortable doing that yeah.” (IM faculty)
16	Time limitations in PCP practice make prescribing MOUD hard	55	7.5%	• “I think the biggest disadvantage for me is that we treat a lot of patients with complex medical problems and the time that it takes to appropriately cover this and do the counseling that’s required by it. We just don’t have that in our visits you know and my priority is to take care of their medical problems. Even though this is compounding it, it’s you know, it’s a time issue…Time is a huge factor.” (IM faculty)
17	Patient factors affect willingness to prescribe MOUD	58	7.9%	• “I think that goes back to the point of who’s doing it. No one’s doing it. I mean we’ve reach the point in society where people who have addiction to alcohol, we can say ok we can deal with that, but when you’re addicted to heroin and you’re addicted to other opioids that’s still an ugly thing and those people are manipulative I’m around a whole bunch of them who are now in recovery, but still they’ve had rough lives, and they’re tough to deal with, I agree, you know.” (FM faculty)
18	Cost, insurance, and financial considerations for MOUD for patients	34	4.7%	• “Well especially if you take the, you know, you go through the trouble you go through the emotional trauma of identifying this patient telling them that they’re misusing that they are addicted and by the way we want you to be on this medicine oh and by the way it’s not covered.” (IM faculty)
	Subjective norms surrounding prescribing MOUD in primary care (Important figures)
19	Provider and supervisor support	53	7.3%	“I think with PAs too, our dependence on supervising physicians, somewhat has a huge influence over, whether we are given, on whether that becomes part of our scope of practice.” (PA faculty) • “To kind of jump off that, until recently we haven’t had supervising physicians in our internal medicine and in our, you know, psychiatry clinics, except for a few that were Bup providers, so it becomes a supervision issue. I think we had up until recently, one Suboxone provider in UIM clinic, which is our primary care clinic, so you know, from a medical legal standpoint we have to have supervision.” (IM/Psych resident)
20	Institution and community support	22	3.0%	• “I agree with that too, I said that’s one of my things was because I would need the support of the staff and the institution and all these other people before I would do it.” (FM faculty)
21	Practice/Patient support	28	3.8%	• “May be the folks in the waiting room wouldn’t want to be surrounded by drug addicts.” (FM faculty) • “I think there’s lack of continuity between providers in the clinic that if you’re going to have building of a relationship in the setting of treatment of opioid use disorder, you need to have kind of a constant provider or, you know one or two people you built that relationship with, because otherwise I don’t think that it would be as beneficial for the patient, they wouldn’t be receiving the care that they needed to do so effectively, you know, and I think there needs to be some way from a scheduling standpoint to allow so at least initial longer visits. As previously mentioned, because I think it would be hard to address that, you know and starting that and initial precautions in a 20-minute visit.” (FM resident)
22	Other important players	29	4.0%	• “I think the lawyers and the business people in charge of our health care system would have trouble with it, because you look at the bottom line, they’re not bad people but they look at the bottom line, is this squeezing something out that makes money or not. But I’d bet X at some point would have a problem with it depending on what is it affecting the bottom line or are we making our two to three percent? Honest to God, they’re not bad people, but it’s complex and we’ve got this massive system and every decision you make, it’s going to hurt some, it’s going to help some, it’s going to hurt some, it’s going to help. And so, I mean you have to have a basic fundamental level commitment that we’re going to do this because it is important to do it and if not it becomes a fool’s errand, you’re just sort of spinning in the wind [?] but you’re not going to make much headway on it.” (FM faculty)
	Perceived behavioral control around prescribing MOUD in primary care: What would enable prescribing?
23	MOUD requires an integrated system of care to support primary care	62	8.5%	• “I am going to take a step back. We got in this mess by pushing pharmaceuticals and now we are pushing pharmaceuticals to solve a problem that is so much more complex than that. The medicine may be part of the solution but we need a system of care. In a way, I think it has a use, but without a context of care where you have people to back you up in a system of care for the patient or patients I think it’s unethical. Actually just prescribing pills and say good luck I’ll see you next week or next month, I think we’re just going down another road that’s another set of problems, so it could be part of the solution but there’s got to be an organized system of care involved. And they could involve primary care, but none of these questions have to do with that system of care, so I say: “Well, where is the system of care? How does that look?” (FM faculty)
24	MOUD requires protocol/guidelines and dedicated visits	78	10.7%	• “And it is very hard with patients, to say “at this visit we are only discussing this” you know they do come in with multiple things and so there is only so much that you can do that gets absorbed, and you know, in one visit, and so I think it is that’s a challenge too, that they know they’re coming to their doctor, they have a lot of things they want to address.” (IM faculty)
25	MOUD use requires more education and training	146	20.0%	• “I agree that would be the most beneficial not to do it through our primary care clinic but to do it through an addiction clinic. Where they do this every day and they’re the ones who can tell us kind of the nuances of prescribing and how we can integrate that into our own practice, rather than adding it to a practice without having that.” (IM resident) • “… I think it would make the physicians more comfortable if there was somebody that was very knowledgeable about these medications that could help us, you know, become more comfortable and kind of seeing it in practice. Make us more comfortable with prescribing them. So some mentorship specifically from a physician who’s had a wealth of experience.” (IM resident) • “…the vast majority of patients we’re treating have chronic pain and that’s why they are on opioid anyway, just having the screening tools for misuse… I don’t have the training for that and I would need to get up to speed on that. And… this conversation [with patients] introduces an emotionally charged element that can halt your entire day and be very disruptive to practice as well. So just getting into that conversation there can be barriers there too. Starting the conversation too can be very difficult.” (IM faculty) • “One the other questions and I think keeps coming up for me too is what is the scope of the problem in our practice. And I know what some of the numbers are nationally and statewide, but it might be interesting just to get better handle on what we’re talking about… Yeah, I mean how many people are we talking about? How big of a problem is this?” (IM faculty) • “Yeah, I think didactics from prescribers that have done it or that are currently doing it and have talked about what it’s like, what their practice is like. I feel like getting firsthand experience is probably valuable…” (FM resident) … “…this is primary care we’re talking about, but I mean if everyone in the office is prescribing it, then it may be important … to have kind of like a staff training, just to, because these are also difficult conversations to have with patients and a lot of time they don’t want to talk about this in the first place, like the best approach to be able to deal with it, to talk to them about it, encourage them to do this, probably would be good to have kind of like a staff training sort of thing.” (PA student)
26	Financial, institutional, and legal/political/state supports need to be in place to prescribe MOUD in primary care	65	8.9%	• “I think that, you know, having gone to talk with lobby, with state legislators, and they have, there’s clearly this concern about opioid overuse and dependence, and what the legislature seems to think that is they have PAs prescribing opioids, that this is just going to magnify the problem in their state. What they actually don’t seem to realize, and I’ve had legislators actually tell me that.” (PA faculty)

Abbreviations: MOUD = Medication for Opioid Use Disorder, IM = Internal Medicine, FM = Family Medicine, PA = Physician Assistant, IMP = Internal Medicine/Psychiatry.

Validity and trustworthiness.

Efforts were made throughout to obtain valid, credible, and trustworthy data and to be transparent in reporting methodology.^44,46–48 Guidelines for reporting of qualitative data were followed, using the consolidated criteria for reporting qualitative research checklist.⁴⁹ For data triangulation, data was obtained from multiple perspectives –students, residents, and faculty providers – and from different types of PCPs – IM, FM, and PAs.^31,50 Established guidelines were used for implementation of the focus groups, multiple coding, and qualitative analyses.^35,37,42 Data journals and field notes/memos were used to document steps in the analysis process, coding disagreement resolution, and consensus over time.⁴⁹ Quotes and illustrative summaries are presented to allow the reader to fully experience the richness of the data.⁴⁷

Results

Description of sample

Seven focus groups were conducted (N = 46; no refusals or drop outs): FM faculty (n = 9), IM faculty (n = 5), PA faculty (n = 6), FM residents (n = 9), IM residents (n = 4), IM/Psychiatry (IMP) residents (n = 4), and PA students (n = 9). Participants were 61% female, aged 35 (11.9), 77.8% Caucasian with 7.2 (9.6) years of experience in practice (Table 1). One provider (2.1%) had a buprenorphine waiver.

Table 1.

Demographic information for focus groups (N = 7).

	Age (M, SD)	Year in practice (M, SD)	Gender	Ethnicity
Faculty
IM (N = 5)	39.8, 11.19	10.1, 10.61	80 % F 20 % M	100 % Caucasian
FM (N = 9)	48.22, 11.21	19.78, 11.63	44.4 % F 55.6 % M	22.2 % Asian, 11.1 % African American, 66.7 and Caucasian
Residents
IM (N = 4)	27.5, 1.0	1.88, 0.85	75 % F 25 % M	33.3% Asian, 33.3 % African American, 33.3 % Caucasian
FM (N = 9)	28.89, 2.47	2, 0.75	55.6 % F 44.4 % M	11.1 % Asian, 11.1 % Latino,77.8 % Caucasian
IMP (N = 4)	28.5, 1.29	3.13, 1.32	25 % F 75 % M	25 % Latino, 75 % Caucasian
PA faculty (N = 6)	47.4, 10.97	11.4, 4.45	50 % F 50 % M	100 % Caucasian
PA students (N = 9)	24.56, 1.42		88.9 % F 11.1 % M	11.1 % Asian, 11.1 % African American,77.8 % Caucasian
Total (N = 46)	35, 11.9	7.2, 9.62	60.9 % F 39.1 % M	11.1 % Asian, 4.4% Latino, 6.7 % African American, 77.8 % Caucasian

Abbreviations: IM = Internal Medicine, FM = Family Medicine, PA = Physician Assistant, IMP = Internal Medicine/Psychiatry.

Quantitative results

General attitudes towards MOUD utilization in primary care were positive, supportive of its use, and did not differ significantly among the different focus groups for buprenorphine or naltrexone. Intention to prescribe was significantly higher for trainees. Subjective Norms and Perceived Behavior Control for Self-Efficacy and Controllability were more neutral overall and varied significantly among the groups for Buprenorphine only (Tables 2 and 3).

Table 2.

Average responses (mean, standard deviation), on a scale of I (strongly disagree) to 7 (strongly agree), regarding Buprenorphine and Naltrexone by 7 different focus group (Internal medicine and Family Medicine faculty and residents, Internal Medicine/Psychiatry Residents, Physician Assistant – faculty and Residents), based on pooled data after imputation.

	Attitudes		Subjective norms		Perceived behavior control		Intention
	Bupe	Nal	Bupe	Nal	Bupe	Nal	Bupe	Nal
Faculty
IM (N = 5)	5.03, 0.84	5.03, 0.78	2.50, 1.12	2.90, 1.04	3.65, 1.26	4.50, 1.57	2.27, 1.26	3.27, 1.52
FM (N = 9)	4.83, 1.44	5.38, 1.59	2.39, 1.06	2.83, 1.86	4.22, 1.06	4.44, 1.89	2.59, 1.78	3.78, 2.23
Residents
IM (N = 4)	5.36, 0.68	5.88, 0.98	3.13, 1.56	3.50, 1.84	3.5, 0.61	4.38, 0.60	3.92, 1.10	4.00, 1.88
FM (N = 9)	5.64, 0.92	5.75, 1.07	3.05, 0.59	3.16, 1.10	2.69, 0.80	3.31, 0.69	4.59, 192	4.59, 1.93
IMP (N = 4)	6.25, 0.96	6.21, 0.42	3.50, 1.17	3.63, 1.05	4.50, 1.65	5.19, 1.25	5.92, 0.88	6.00, 0.82
PA faculty (N = 6)	5.92, 1.25	5.97, 1.29	3.47, 1.16	3.45, 1.18	3.46, 0.75	3.83, 1.30	4.44, 1.61	4.39, 1.54
PA students (N = 9)	6.19, 1.23	6.22, 1.05	4.86, 0.88	4.89, 0.88	4.28, 0.69	4.39, 0.63	6.22, 1.03	6.33, 0.78
Total (N = 46)	5.59, 1.18	5.77, 1.15	3.32, 1.29	3.51, 1.44	3.73, 1.09	4.18, 1.27	4.30, 2.00	4.67, 1.87
χ2(6), p <	9.02, .18	5.24, .52	19.5, .01 *	12.1, .07	13.4, .03 *	9.74, .14	22.9, .01 *	15.1, .02 *

Abbreviations: Nal = Naltrexone, Bupe = Buprenorphine, IM = Internal Medicine, FM = Family Medicine, PA = Physician Assistant, IMP = Internal Medicine/Psychiatry.

^*Significant difference (p<.05) between groups detected by Kruskal Wallis Test.

Table 3.

Average responses (mean, standard deviation), on a scale of I (strongly disagree) to 7 (strongly agree), regarding Buprenorphine and Naltrexone by Faculty (Internal Medicine and Family Medicine) versus Residents (IM, FM, IM/Psych.), based on pooled data after imputation.

	Attitudes		Subjective norms		Perceived behavior control self-efficacy		Perceived behavior control controllability		Intention
	Bupe	Nal	Bupe	Nal	Bupe	Nal	Bupe	Nal	Bupe	Nal
Faculty (N = 14)	4.90, 1.23	5.26, 1.33	2.43, 1.04	2.86, 1.57	3.14, 1.41	3.93, 1.95	4.89, 1.81	4.89, 2.10	2.48, 1.57	3.59, 1.96
Residents (N = 17)	5.72, 0.88	5.89, 0.91	3.18, 0.96	3.32, 1.2	3.47, 1.45	4.06, 1.45	3.15, 1.77	3.94, 1.52	4.74, 1.66	4.78, 1.81
TOTAL (N = 31)	5.35, 1.11	5.61, 1.15	2.84, 1.05	3.13, 1.34	3.32, 1.42	4.0, 1.66	3.94, 1.97	4.37, 1.84	3.72, 1.96	4.25, 1.94
χ2(1), p <	3.69, .06	1.67, .20	4.13, .04 *	1.93, .16	0.39, .54	0.03, .87	6.11, .01 *	2.80, .09	2.86, .09	9.8, .002 *

Abbreviations: Nal = Naltrexone, Bupe = Buprenorphine, IM = Internal Medicine, FM = Family Medicine, PA = Physician Assistant, IMP = Internal Medicine/Psychiatry.

^*Significant difference between groups detected by Kruskal Wallis Test.

The model showed that more positive attitudes towards MOUD (p=.005) and being male (p = .03) were significantly predictive of participants’ intent to prescribe MOUD in the future.

Qualitative results

Twenty-six distinct sub-themes emerged to describe possible factors influencing participants’ intentions around utilizing MOUD (Table 4). Perspectives shared by participants were either expressed (own perspectives) or experienced (observed in others).

Attitudes towards buprenorphine and naltrexone.

Perspectives supportive of prescribing buprenorphine (Theme 1) included the belief that it typically lends itself to good adherence and may prevent relapse better than naltrexone due to alleviating withdrawal symptoms. Attitudes supportive of using naltrexone (Theme 3) were that it has a lower frequency of administration (monthly), it does not require an 8-hour training, and patients may be more motivated after showing they can obtain abstinence. Perceived challenges to utilizing buprenorphine (Theme 2) and naltrexone (Theme 4) included concern about patients not attending follow-up appointments and about diversion or misuse of medication. The challenge of obtaining a buprenorphine waiver was noted, as well as the perception that buprenorphine may be “riskier” given the need for a waiver. Specific to naltrexone, concern was expressed about verifying abstinence prior to initiation and uncertainty as to whether it was as efficacious as buprenorphine or methadone.

Attitudes supportive of utilizing MOUD in primary care.

Participants shared the belief that MOUD treats a disease (addiction) and is within the scope of practice of PCPs (Theme 5), comparably to prescribing medications for other chronic medical conditions (e.g., diabetes). Participants cited that utilizing MOUD in primary care would improve access to care (Theme 6; highlighting specialty-care access challenges, e.g., psychiatry) and could reduce the stigma of seeking treatment with an addiction specialist (Theme 7). Participants felt the ongoing relationship/trust between a patient and PCP may help with engagement in MOUD (Theme 8) and that MOUD availability can reduce health care burden with patients obtaining outpatient treatment versus emergency room/hospital care (Theme 10). Participants believed that MOUD is a life-saving treatment to address the opioid crisis (Theme 9) and that MOUD positively affected patients by reducing overdose risk and improving functioning.

Challenges towards MOUD use in primary care.

Participants shared concerns from themselves and colleagues/supervisors that patients with addiction could “flood” primary care offices offering MOUD (Theme 11) and that the focus on opioid treatment may take “front stage.” Participants voiced preference for diverse caseloads providing MOUD only to already-established patients.

One of the most often-raised concerns for utilizing MOUD in primary care was limited time available in this setting (Theme 16). There was concern about the potentially emotionally-charged nature of conversations about pain and addiction that may disrupt patient flow and satisfaction (provider and patient). Participants shared concern that PCPs “inherit” patients who may have developed addiction as a result of treatment from other physicians, and that utilizing MOUD should not fall solely to PCPs (Theme 13). Participants felt a clinic structure that would facilitate frequent follow-ups and more time for MOUD visits would be beneficial. Participants also voiced concerns about patient factors of those active in addiction (Theme 17), including the possibility of disinterest in treatment and poor follow-up. Participants brought up the cost of the medications for patients and insurance authorizations as potential barriers to MOUD utilization (Theme 18).

Some participants were apprehensive that MOUD is “replacing one drug for another” without a treatment endpoint (Theme 12), noting that other treatment modalities, including psychotherapy, must be in place for MOUD to work successfully (e.g., “pills without skills is a fool’s errand;” Theme 14). They expressed a need for all providers within a practice to share a similar philosophy around utilizing MOUD (Theme 15) for both leave coverage and continuity of care.

Subjective norms.

Participants were asked to identify figures that would support or disapprove of utilizing MOUD in primary care. Participants noted the importance of peers in their practice all being “on board” with MOUD (Theme 19). Residents and PAs cited the necessity of support/waiver of their supervising physician to utilize MOUD, which also reduced their sense of control (see perceived behavioral control below). Sub-specialist support, colleague support, and appropriate supervision were also noted as important facilitators to MOUD utilization. Providers were sensitive to perceptions of other patients, and others working in the practice, in considering MOUD (Theme 21), citing concern that patients without use disorders may not feel comfortable around those active in addiction.

Perceived behavioral control.

Participants were asked what may be needed to improve their sense of control over and confidence in utilizing MOUD in primary care. The most frequently discussed topic among participants was a need for more education (Theme 25), including training on comorbid pain and how to manage conversations about addiction within a busy practice. Most wanted hands-on-training and teaching in real clinical settings by experienced practitioners versus presentations/didactics. Some requested more information about the scope of the opioid use disorder problem within their own practices, preferring a locally-informed approach to MOUD implementation. Other information (Theme 26) that would improve MOUD uptake included billing/productivity, availability of state-level supports, and the legal aspects unique to treating patients with addiction (e.g. confidentiality, DEA audits).

Participants voiced a need for an “integrated system of care” to support MOUD in primary care (Theme 23), stating apprehension that current practice structure would be ineffective. To incorporate MOUD effectively and confidently, suggestions included improving access to appropriately-trained support staff, social work and sub-specialists, ability to refer out to specialty settings if needed (requesting specific resources/arrangements for referrals) and dedicated visits/clinic guidelines/protocols for MOUD (Theme 24).

Discussion

With calls to improve access to MOUD within primary care and for increased opioid use disorder education for academic medical centers in SC, it is vital to address facilitators and barriers to implementation from front-line providers.⁵ While other studies have examined barriers and facilitators related to provider uptake of MOUD, this study is one of the first exclusively in the southern US to include trainees and to focus on providers who primarily did not possess a waiver and were not already providing MOUD.^{24,25,27–29,51,52} This study demonstrates an overall positive view of providing MOUD in primary care, and that higher positive attitudes significantly predicted intent to prescribe. While attitudes did not vary among the groups, trainees seem to express higher intention to prescribe but generally lower perceived behavioral control for controllability.

Several factors were identified to improve provider confidence, self-efficacy, and intention to provide MOUD in primary care in SC. More education was one of the most frequent requests - excluding the required 8–24 hour buprenorphine waiver training (which was seen as a barrier to providing MOUD). There was a call for general education about MOUD basics coupled with hands-on experience and access to an experienced practitioner/mentor to increase confidence to initiate MOUD. These findings are consistent with other literature suggesting that limited education and lack of awareness of MOUD are barriers to utilization for physicians.⁵³ Additionally, it was noted that front-line primary care providers, unlike addiction specialists, often encounter patients who are treatment-seeking for pain or other (usually multiple) medical conditions, and not for opioid use disorder. Having a conversation with a patient that may be demonstrating symptoms of opioid use disorder, especially during a brief office visit, can be emotionally challenging, disrupt patient flow, and end with a dissatisfied patient. Perceptions of patients needing MOUD being “difficult” has been previously suggested as a barrier to providing treatment.⁵³ Educational and experiential opportunities should couple validating these challenges with (1) mentorship on navigating difficult conversations successfully, (2) discussion/demonstration of rewarding, life-changing aspects of MOUD (i.e., enhancing positive attitudes, decreasing stigma), and (3) acknowledgement that providing MOUD can be less complex than expected.

Programs such as Project Extension for Community Healthcare Outcomes (ECHO), which utilizes de-identified case-based learning coupled with brief didactics and “mini-fellowship” opportunities, and the Massachusetts Consultation Service for Treatment of Addiction and Pain (MCSTAP), which is a phone-hotline consultation service to assist medical providers with challenges involved in treating patients with chronic pain and/or substance use disorders, have been put into place to provide pragmatic, real-world, on-the-job support, and could be more broadly utilized in academic settings.^54–56 In training, providing examples of varied successful models of MOUD-delivery in primary care would be valuable, acknowledging that offices might see a relatively low volume of patients with opioid use disorder. For example, identifying one provider “champion” may be a better approach than training all providers. It is also important to note that approval/waiver of the supervising provider is necessary for a resident (and at this time PAs and Nurse Practitioners) to provide MOUD, which affects their perceived controllability to provide MOUD.

There was a call for an “integrated system of care” to improve the uptake of MOUD. A system ensuring all providers/staff are cohesive in mission could be encouraged by developing in-office protocols and providing training for staff. This integrated system of care also requires that options for counseling and patient support are available. Funding for this system is a concern that needs to be addressed, and in the context of limited Medicaid expansion, will continue to likely be a barrier to access to MOUD in the southeastern US.^5,10Systems-level challenges in implementing MOUD were identified by participants in this study and have also been identified as important targets for improvement in the literature. For instance, one study examining whether Project ECHO helped to increase uptake of MOUD in rural PCPs identified that low participation was ultimately influenced for some by poor practice-level support from leadership around use of MOUD.⁵⁷ Institutional and state-systems must prioritize MOUD efforts, and provide both financial and informational support for efforts to trickle down to individual PCPs.

Limitations for this study include the focus on experiences from one academic center in SC, though given the association between lack of Medicaid expansion and barriers to MOUD access in the literature, regional differences may be important to consider.¹⁰ Only one participant had a buprenorphine waiver, indicating a relatively low level of experience, which was appropriate for the study objectives of defining barriers to initiating use, but limits the ability to discuss barriers once MOUD prescribing is implemented. Another limitation was difficulty interpreting clinical significance of Likert-scale differences between groups (i.e. no pre-defined number on the Likert-scale correlates to absolute intention to action).

The perspectives of current and future PCPs provide a valuable snapshot of the opportunity and challenges of providing MOUD in primary care. These perspectives can inform the design and implementation of training for current and future PCPs so that increased access to care can be achieved. Future directions may include replication after MOUD implementation to re-evaluate which barriers were realized. Also, evaluating the role of gender, and predictors of transition from attitudes, subjective norms, and perceived self-efficacy to behavior could help provide direction for practice-based implementation strategies. Providing solid education and fostering attitudes supportive of MOUD in primary care may increase MOUD uptake, ultimately with the potential to address the treatment gap for those with opioid use disorder during this critical period in the evolving opioid crisis.