INTRODUCTION

Alcohol use disorders (AUDs) are common, chronic, and undertreated.1 , 2 This is particularly true for U.S. veterans, for whom the estimated lifetime prevalence of AUD is 32%,3 relative to an estimated lifetime prevalence of 29% in the U.S. general population.4 Although specialty addiction treatment improves outcomes for patients with AUD,5 , 6 most never receive it,7 and many do not want to attend specialty addiction programs.7 Experts have called for expansion of evidence-based treatment options offered to patients with AUD and increased management of AUD in non-specialty settings, especially primary care (PC).8,9, – 10

Three medications for AUD are approved by the U.S. Food and Drug Administration (FDA).11 , 12 They include disulfiram, which causes an adverse reaction if patients drink alcohol; naltrexone, an opiate antagonist that is available for both oral administration and via monthly injections; and acamprosate, a glutamate modulator that reduces symptoms associated with abstinence from alcohol.13 Other medications, such as topiramate, have strong meta-analytic support for treating AUD.14 A meta-analysis of the evidence for the effectiveness of AUD medications found the strongest evidence for acamprosate and oral naltrexone, with no clear difference in effectiveness between them.14

Medications for treating AUD can be prescribed and managed in PC.14 While the majority of the evidence for AUD medications originates from specialty treatment settings, and a “gold standard” for AUD medication management in PC has not been established, some studies have evaluated interventions that could be adapted to PC settings.14 For instance, the nine-arm COMBINE trial11 , 15 demonstrated that AUD medications with ongoing “medical management” could be as effective as a state-of-the-art behavioral intervention.11 Further, the emergence of new models of PC–mental health integration and/or patient-centered medical homes may enable PC providers to prescribe and integrate these medications into existing care systems.16 , 17 However, there are likely multiple barriers to PC providers’ prescribing these medications. Previous research in addiction treatment settings has identified barriers related to knowledge, skills, and beliefs.18 However, barriers to the use of AUD medications in PC have not been described previously, and are likely to include additional barriers unique to the PC setting, as care for AUD has historically been separated from PC.14

Conceptual models of implementation suggest a strong role for peer leaders in the diffusion of innovations.19 , 20 Similarly, social marketing theory21 , 22—focused on identifying and changing underlying emotions that influence behavior—employs business marketing techniques, such as “segmenting the market,” to identify groups that may lead change. To optimize implementation of AUD medications in PC, it may be important to identify barriers specific to “segments” of PC providers based on willingness to prescribe. Addressing barriers among those most willing to prescribe may result in a segment of peer leaders who can lead implementation across groups.

Therefore, using social marketing theory, we conducted this qualitative study of PC providers in the Veteran’s Health Administration (VA) to describe barriers to and facilitators of the use of AUD medications, and secondarily to describe barriers and facilitators specific to subgroups based on willingness to prescribe.

METHODS

Participants and Setting

This study took place at five distinct, freestanding PC clinics associated with a single large VA healthcare system in the northwestern United States. Three clinics were located within two large urban centers, which serve approximately 21,000 patients each (both ~88% male). Two clinics were VA-managed community-based outpatient clinics (CBOCs), which are more rural and serve approximately 4000 (~90% male) and 6500 patients (~92% male), respectively. PC providers with prescribing privileges were eligible.

Recruitment

We first identified a key point of contact at each clinic in order to introduce the study and determine the best way to opportunistically recruit and conduct interviews with participants, while minimizing disruption to clinic operations and patient care. Clinical contacts and participants were not blinded to the purpose of the study. Depending on clinic preference, recruitment strategies involved emailing invitations to all providers in the clinic to request and arrange a time for interviews, or coordinating with clinic leadership to be on site to recruit participants opportunistically. The latter was the most frequent method of recruitment. Potential participants were given a verbal overview of the study, provided with an information sheet, and asked to provide verbal consent. The study, including a waiver of written informed consent, was approved by the VA Puget Sound Institutional Review Board.

Data Collection

In-person semi-structured interviews were conducted between August 2014 and July 2015 by an interviewer trained in qualitative methodology (CEA or JPY). One interviewer (CEA) is also a PC provider. The interview guide (Appendix 1) was developed based on domains of behavior change articulated in a social marketing framework,22 previous barriers identified in addiction treatment settings,18 and hypothesized barriers in PC settings.14 The guide assessed knowledge about/experience with AUD medications, barriers to prescribing, perception of skills and resources needed to prescribe, optimism regarding the usefulness of AUD medications, possible facilitators, and willingness to prescribe. The interview included three sections of open-ended questions, with suggested follow-up questions. The first two sections addressed the participants’ general experience with/approach to patients with AUD and AUD medications, and are the focus of the present study. The third addressed issues related to pharmacogenetics, and findings have been reported previously.23 Interviews lasted approximately 20 min and were digitally recorded, transcribed, and reviewed for accuracy. Interviews were conducted until saturation of themes occurred (e.g., investigator consensus that no new information was identified).

Data Analysis

Data were analyzed iteratively using thematic analysis24,25, – 26 to identify both expected (a priori) and emergent themes. Midway through data collection, interim analysis was conducted to review early emergent themes, triangulate data with existing theory and research, modify procedures in response, and check for saturation.27,28, – 29 Preliminary analyses were conducted using the Rapid Assessment Process,27 in which data from each interview transcript were reduced to a templated summary corresponding to interview questions and distributed to the full multidisciplinary team for review. This process confirmed a priori themes, identified emergent themes, and served as the initial codebook for full data analysis. Consistent with social marketing theory, this process also allowed “market segmentation,” which included categorizing participants into one of two groups (more willing and less willing) based on consensus among investigators after independent data review.

After interviews were complete, all data were analyzed using ATLAS.ti.30 Two master’s level investigators trained in public health (JY and JR) and social work (JY) independently coded all data using the codebook established during interim analysis; both added new codes if evidence accumulated in support of new themes.24,25, – 26 Discrepancies in coding were resolved via consensus among investigators. Coded content was reviewed iteratively by the multidisciplinary investigator team, and coded data were grouped into distinct themes for all participants and across subgroups based on willingness to prescribe. All investigators reviewed coded data to finalize themes, check conclusions against the data, and identify prototypical content within each theme for presentation.31 , 32

RESULTS

Twenty-four PC providers were recruited. This included all providers practicing at both CBOCs (n = 11), and all we attempted to recruit at two of the three urban clinics associated with large medical centers (n = 13). Recruitment at the third urban clinic was challenging due to changes in leadership during the study, and ultimately only one provider from this clinic participated. Recruited providers included 19 medical doctors (MDs), one doctor of osteopathy (DO), and four nurse practitioners (NPs); 53% were female. Participating providers had been practicing in PC for a mean of 15 years (SD 11.9) and spent an average of 25.7 hours (SD 9.0) with patients per week.

Barriers to and Facilitators of Prescribing AUD Medications

Three barriers and three facilitators to prescribing AUD medications were identified in the sample of all participants and are described below.

  • Barrier 1: Limited knowledge of and experience with prescribing AUD medications

Participants reported limited knowledge of and experience with prescribing AUD medications. Some had heard of some of the medications and/or recognized the name of one or more, and some had treated patients who had received prescriptions from another provider. Participants were more familiar with some medications than others: Antabuse® (disulfiram) was the medication most commonly recognized/mentioned by providers, acamprosate the least. However, few had received any training or education on AUD medications, most reported lacking the practical knowledge needed to prescribe or manage them, and few reported having ever prescribed them. One provider (PCP1) said, “I don’t have the education or experience … I definitely wouldn’t do that [prescribe].” Participants attributed discomfort and/or reluctance to prescribe to their limited knowledge and lack of prescribing experience. One (PCP2) said, “I don’t think I’d feel comfortable enough even if the patient was started on it, to renew it …. if I’ve never prescribed it, it’s not in my scope of practice, then I don’t know what to tell them to look out for.”

  • Barrier 2. Concern that medications should not be offered without counseling and beliefs that specialty addiction treatment is the only option for treatment

Providers expressed concern that prescribing AUD medications without providing concurrent behavioral therapy represented substandard AUD treatment. For instance, one participant (PCP3) said, “I think if I’m prescribing naltrexone … that it’s a far inferior option than having somebody engaged with ATC [the addiction treatment center].” Another (PCP4) said, “Do I think you can start somebody on naltrexone and pat them on the back and then send them on their way? No, I don’t. … if this is where it stops [at AUD medications], we’re not going to be successful.”

Providers believed in specialty addiction treatment as the only option for treatment, and based on this perspective, some expressed reluctance to prescribe medications. One (PCP6) said, “I’m not convinced that they’re [medications] all that helpful, to be honest. …they’re more of an adjunct, they’re not the treatment … the main treatment is the program, right?” Some felt that the inclusion of a 12-step model of treatment was an essential component of successful AUD care. One (PCP7) said, “Maybe a medication alone could do it for some people, but it’s part of broader behavioral … reinforcement … all the 12 steps.”.

  • Barrier 3. Expressions of alcohol-related stigma

Multiple expressions of alcohol-related stigma, or prejudicial societal beliefs and attitudes that discredit individuals based on their alcohol use,33 were identified. Consistent with definitions of stigma in the literature,34,35, – 36 these included: perceptions of character flaws (e.g., untrustworthiness), social distancing (e.g., someone else should treat this condition), perceptions of control of and culpability for AUD (e.g., beliefs that patients are choosing their condition and can quit if they are willing to do the work), and labeling language (e.g., “those people” and “alcoholics”). Prototypical examples are presented in Table 1.

Table 1 Expressions of Alcohol-Related Stigma (Barrier #3) Articulated by Primary Care Providers*
Full size table

Facilitators

Facilitator 1: Training and Education

Participants articulated that they would feel more comfortable prescribing AUD medications with training and education. Content suggestions included training on medication characteristics (options, side effects, safety profiles, contraindications), effectiveness, prescribing (how to start and stop medications, dosing and adjustment, duration, lab monitoring, choosing the right medication for the right patient), and education about specialty care services and treatment. One participant (PCP8) summarized the need for information: “Give me the tools, give me the rules, give me the pros and cons. Let’s make sure everybody’s educated … then why wouldn’t you do it?”

Regarding the method of training delivery, participants suggested education via flyers, handouts, and PowerPoint presentations; facilitated training such as in-services or mini-residencies; and integration of AUD medication and prescribing information into existing electronic clinical decision support applications. They also identified potential leaders who should deliver training, including addiction and mental health providers or pharmacists.

Facilitator 2: Support for Prescribing

Many participants suggested having ongoing external support for prescribing AUD medications (e.g., from mental health or addiction specialists, pharmacists). One provider (PCP2) said, “I feel it does have to be spoon fed to us. I’m totally willing to do it, but it would need to be someone telling me what to prescribe.” Possibilities for this type of support ranged from self-initiated consults with specialists outside PC (e.g., teleconferences or non-visit consults, just to “run it by” someone), to requesting ongoing consultation with mental health providers or pharmacists (e.g., “I’d have to have my hand held a little bit … probably working with pharmacy” PCP18), to integrating staff with addictions expertise within PC (e.g., a “larger sort of addiction team [to provide] close and regular follow up, support, compliance” PCP5).

Facilitator 3: Provision of Behavioral Follow-Up Completed by Mental Health or Specialty Care Staff

Providers also suggested having mental health or specialty addiction providers onsite and available to provide behavioral follow-up for patients. One (PCP6) said, “I think you’d want somebody at the clinic itself here … managing the substance dependence.” Some expressed that having these resources would be essential to their comfort/ability to prescribe. One provider (PCP10) said, “I don’t want them to just be getting medications from me and then not having the backup of a therapist.” And another (PCP2) described wanting “someone doing the talk therapy … and then me just checking in just like I check in on their heart disease.”

Themes Identified within Provider “Segments” Based on Willingness to Prescribe

Social marketing advocates identification of population “segments” that differ in key attitudes or behaviors, and tailors interventions to those segments.37 We identified two segments—providers more and less willing to prescribe—that differed in their beliefs regarding the role of PC in AUD treatment, the effectiveness and role of medications in treating AUD, and the potential for prescribing AUD medications in PC. A summary of beliefs reflected in each group is provided below; prototypical examples within each group are presented in Appendix 2.

Those who were less willing to prescribe expressed beliefs that: medications cannot effectively treat AUD; treating AUD is not the role of PC; specialty addiction treatment providers are and should remain the experts; and substantial changes would be necessary for them to prescribe AUD medications. Additionally, these providers indicated that there was no time to treat AUD in PC. Specifically, they noted that: management of AUD is time-intensive and requires a great deal of follow-up; effective medication prescription would require providers to also address psychosocial issues with their patients, which is not practical; and PC is already handling too many agendas for the short time they have with their patients.

Those who were more willing to prescribe expressed beliefs that medications are a “tool that can be useful” or a “first step” in treating AUD. This was accompanied by the perspective that treating AUD involves a continuum of approaches and that medications can be one part of that approach, and may even help open the patient to behavioral or psychosocial treatment. These providers believed that PC represents an optimal setting in which to prescribe medications, because it reduces barriers to specialty addiction treatment, such as stigma or geographical or other logistical barriers, and may enable capitalizing on prime moments for behavior change given that PC is a “first line” of treatment. These providers also believed that prescribing AUD medications was within the scope of PC and would be similar to prescribing for other conditions, such as smoking or depression. Finally, they believed that the role of the PC provider was to choose anything that might help the patient (e.g., even if a combination of medication and therapy might be best, they would be willing to try medication alone if that is what the patient wanted) and that to not prescribe medication might deny patients the opportunity for change.

Discussion

This qualitative study identified barriers to and facilitators of prescribing AUD medications in PC, and applied social marketing theory to identify two groups of providers who were distinguished based on their willingness to prescribe and beliefs regarding AUD pharmacotherapy in PC. The findings suggest that lack of knowledge and experience, beliefs regarding the superiority of behavioral treatments for AUD provided in specialty settings, and alcohol-related stigma may constitute barriers to the provision of AUD medications in PC, whereas training, support, and integrated care models may facilitate provision.

While previous studies have described barriers experienced by specialty addiction care providers18 , 38,39,40, – 41 and providers in HIV PC clinics,42, this is the first to our knowledgeto have assessed barriers in general PC. The barriers that were identified overlap with those previously reported in other settings—specifically, lack of knowledge/training, optimism, and confidence in the use of medications, and a belief that medication treatment is not consistent with traditional treatment philosophy.18 , 38,39,40,41, – 42 PC providers reflected similar feelings and perceptions; they lacked knowledge, confidence, and optimism regarding both prescribing and managing AUD medications, and articulated beliefs that patients were best served by expert providers in specialty addiction treatment. In addition to barriers that may be common across settings, findings from the present study suggest hurdles that are unique to PC. Specifically, providers believed that medication prescribing in PC would only  have a chance of being effective if additional staff were available to provide behavioral support.

While specialty addiction treatment is effective for patients who engage, most patients with AUD (~85%)7 do not engage.43,44, – 45 Therefore, offering only specialty addiction treatment may contradict contemporary standards of evidence-based46 and patient-centered care.46 , 47 Provision of AUD treatment in PC is one way to increase patient-centered care for AUD.10 With the emergence of studies suggesting that effective AUD treatment could be provided in PC, there is potential to expand on treatment options. Specifically, in the COMBINE trial, patients were offered “medical management,” which was designed to be feasible in PC and included up to nine manual-guided counseling visits over 16 weeks following medication prescription.11 Though the first visit was longer—45 min—follow-up visits were approximately 20 min each, and the content of the counseling was focused on advice to reduce drinking, adverse effects, the importance of adherence, and encouragement for 12-step attendance if patients were interested.11 , 14 , 15 In conjunction with AUD medications, this model was as effective as specialty addiction treatment.11 , 15 In another study of PC patients with AUD, an alcohol care management intervention focused on treatment with naltrexone and delivered by behavioral health providers (nurses and psychologists) with support from a nurse practitioner resulted in high rates of patient engagement with medications and fewer heavy drinking days relative to usual care (referral to specialty treatment).48 Findings from these studies in recruited populations of patients with AUD suggest the possibility that—if given the time, support, and/or training to do so— PC providers and/or teams can provide the needed adjunct to medications, similar to prescribing medications for depression. Such approaches would be compatible with patient-centered medical homes,49,50, – 51 integrated primary and mental health care,52 and collaborative care models for other behavioral health conditions.53

Expressions of alcohol-related stigma were common in this study. Because stigma is deeply embedded in social and cultural norms,34 and AUD is one of the most highly stigmatized mental health disorders,54 these findings are unsurprising. Stigma influences how medical care is provided.35 , 36 , 55 For instance, in a study of 728 mental health providers randomly assigned to one of two vignettes that differed only in the terminology used to refer to a patient, providers randomly assigned to the term “substance abuser” were more likely than those assigned the term “substance use disorder” to agree with the notion that the patient was personally culpable for his condition and to suggest punitive measures be taken (as opposed to treatment).36 Stigma is also a barrier to treatment receipt for patients with AUD due to perceptions of control and culpability. Among persons with AUD who had considered seeking treatment in a national sample, 44% felt they should be strong enough to handle their disorder alone, 18% were too embarrassed to discuss it, and 10% “hated answering personal questions.”7 Because stigmatized attitudes toward patients with AUD influence the provision or quality of AUD care and/or patient response or openness to care, interventions focused on changing social norms and/or beliefs56 among patients and providers may be needed.

Despite identification of several barriers, findings from this study offer multiple reasons for optimism regarding provision of AUD pharmacotherapy in PC. First, participating providers described models of care for facilitating the provision of AUD pharmacotherapy in PC that were consistent with efficacious models.11 , 48 Moreover, providers offered several actionable suggestions, including training on the use of AUD medications in PC and support for prescribing from pharmacists or specialty addiction treatment providers. They also suggested specific modes of delivery for training and support, such as webinars and electronic consults with pharmacists. Moreover, this study identified a subgroup of providers who indicated greater willingness to prescribe and who shared beliefs that may facilitate provision of AUD pharmacotherapy. According to social marketing theory, if encouraged via interventions, this group of providers may serve as leaders of changes in beliefs and attitudes across providers.21 , 22 , 37 , 57,58,59,60,61, – 62

This study has several noteworthy limitations. While interviews were conducted with providers from five different PC clinics to the point of saturation, findings within segments may require further exploration. Additionally, findings may not represent the experiences of PC providers elsewhere. Previous research has suggested variability in the provision of AUD pharmacotherapy across VA facilities.63 It is possible that levels of comfort among providers at the study clinics differed from those at other sites. In addition, because participants were not blinded to the study purpose, responses may have been influenced by social desirability bias. Further, providers were asked only about prescription of medications. They may have responded differently if asked whether they would be willing to prescribe AUD medications if systems for assistance with care management and/or linkage with other services were in place.

Despite these limitations, this is the first study to have identified barriers and facilitators to the provision of AUD pharmacotherapy in PC. Because most patients with AUD do not receive evidence-based treatment for their AUD,7 and because PC may be the optimal care setting in which to treat AUD,8,9, – 10 identifying opportunities to address barriers and capitalize on facilitators serves as an important foundation for efforts to increase access to quality care for AUD.64 This study identifies several such opportunities. Training and augmenting PC with team members who can help willing PC providers manage AUD, and interventions aimed at modifying beliefs held by PC providers may be the optimal targets for moving treatment of AUD into PC.