Abstract
Background: Overdoses and deaths from synthetic opioids grew sharply in the past decade. Most people with opioid use disorder (OUD) do not receive recommended evidence-based treatment: nationally, 72% to 87% of people who need OUD treatment do not receive medication for opioid use disorder (MOUD). Little is known about practice teams’ experiences with home, office, and telehealth induction for MOUD, particularly in primary care.
Methods: We conducted semistructured interviews with primary care clinicians and staff from February through September 2023 to understand experiences providing MOUD via home, office, and telehealth induction. Interviews were part of a PCORI-funded trial, Home versus Office versus telehealth for Medication Enhanced Recovery (HOMER). We used template and editing coding styles to categorize text according to deductive codes derived from research questions and inductive codes derived from multiple readings of transcripts. We used immersion-crystallization to iteratively review coded text and identify interview themes.
Results: Thirty-eight clinicians and staff from 21 US primary care practices participated in interviews. Home induction is increasingly common and preferred by patients and practice teams, social determinants of health affect induction and maintenance in treatment, clinicians and staff use honest communication to build trusting relationships with patients, practices identified patients as MOUD candidates through word-of-mouth and referrals, and an evolving OUD landscape are causing practices to adapt their care.
Conclusion: Primary care practices are committed to offering MOUD. Findings offer insights about the challenges facing primary care practices in their efforts to deliver MOUD to address a rapidly evolving opioid epidemic.
- Buprenorphine
- Opioid Use Disorder
- Opioids
- Primary Health Care
- Qualitative Research
- Social Determinants of Health
- Telemedicine
Introduction
Opioid overdoses and dependency are a US public health epidemic, with sharp growth in overdoses and deaths from synthetic opioids in the past decade.1 In 2021, an estimated 2.5 million people had experienced opioid use disorder (OUD) in the previous year.2 Rates of opioid-related drug overdose deaths have increased every year from 2013 to 2022, despite no significant change in overall drug overdose deaths between 2021 and 2022.3 Most people with OUD do not receive evidence-based treatment for it: nationally, 72% to 87% of people who need treatment for OUD do not receive medication for opioid use disorder (MOUD).4,5 Barriers to accessing MOUD include a shortage of MOUD prescribers, cost, insurance coverage, and transportation.6 In 2023, the federal government eliminated the requirement that clinicians complete training and obtain explicit Drug Enforcement Agency (DEA) authorization through a waiver to prescribe MOUD,7 but this has been insufficient to expand access to care. US opioid treatment programs have limited openings and are serving 80% or more of their capacity.8 Primary care can help increase MOUD access.9 However, only 6 to 10% of primary care clinicians prescribe MOUD (ie, buprenorphine, naltrexone).10,11 Barriers to prescribing MOUD include lack of clinician knowledge and experience, lack of confidence in ability to manage opioid addiction, lack of institutional support, lack of mental health and psychosocial support for patients, time constraints, and resistance from colleagues.12–15 Induction itself acts as a barrier to offering MOUD due to difficulties of timing induction appropriately to minimize withdrawal symptoms, titrating medication, and the risk of precipitated withdrawal.14,16–18 Stigma and mistrust toward people with OUD and MOUD treatment persist, negatively influencing clinicians’ attitudes toward prescribing.19,20 Induction of MOUD, the brief, intense period when patients receive initial buprenorphine doses,21 has traditionally occurred in clinical settings and observed by a medical professional.22 The Substance Abuse and Mental Health Services Administration (SAMHSA) recognizes that induction can occur in clinical or home settings22 and home induction can be performed safely with proper education and support.22,23 Challenges to office induction include logistics and timing of the patient’s visit during the appropriate stage of withdrawal and office hours, and the extensive time, space and other resources required of the practice.13,17,24 Unobserved, asynchronous home induction allows the patient to begin MOUD at the time of their choosing in their home. Home induction has occurred since at least 200325 and become common practice, with studies showing that between 42% and 65% of prescribers perform home induction.17,26 However, some clinicians still cite concerns regarding home induction, such as the potential for medication diversion, precipitated withdrawal, and practicing counter to guidelines recommending observed induction.27 Telehealth induction involves observed, synchronous phone or video communication between the patient and clinician during induction. Home and telehealth induction became more frequent in 2020 after relaxed regulatory restrictions due to COVID-19.25 Soon after, SAMHSA recognized that induction is feasible and safe in both observed, office and unobserved home settings.28 While home and telehealth induction present potentially more convenient options than office induction, the uptake and maintenance of MOUD treatment for different types of patients across induction locations is unknown.29 Primary care clinicians are responsible for the largest share and fastest growth in buprenorphine treatment from 2010 to 2018.30 Further, community-level protective factors such as rates of primary care visits may contribute to lower opioid mortality.31 These facts underscore the need to understand effective treatment for OUD in primary care settings.
We conducted interviews with primary care clinicians and staff to understand their experiences providing MOUD via home (asynchronous, unobserved), office (synchronous, observed), and telehealth (synchronous video or phone contact, observed) induction. Interviews were conducted as part of a larger trial, Home versus Office versus telehealth for Medication Enhanced Recovery (HOMER).29,32 HOMER study aims were to 1) compare short-term stabilization and long-term maintenance in MOUD treatment across patients receiving office induction (synchronous, observed), home induction (asynchronous, unobserved), and telehealth induction (synchronous phone or video contact, observed) and 2) identify patient and practice characteristics associated with maintenance in long-term treatment.29 This information is particularly important because induction acts as a barrier to primary care clinicians offering MOUD,14,16–18 and practice member interviews offer context to understand the challenges and strategies necessary for MOUD implementation and maintenance at the organizational level. Guided by the RE-AIM framework,33 interviews explored MOUD implementation and maintenance across induction locations through the following research questions:
What are primary care clinician and staff experiences with and opinions about home, office, and telehealth modes of MOUD induction?
What are patient perceptions of home, office, and telehealth modes of MOUD induction?
What factors affect primary care practices’ implementation and ongoing maintenance of home, office, and telehealth modes of MOUD induction?
Methods
Study Design
This is a qualitative exploratory study from a larger pragmatic comparative effectiveness trial. HOMER was led by the State Networks of Colorado Ambulatory Practices & Partners (SNOCAP) in partnership with the American Academy of Family Physicians (AAFP) National Research Network (NRN). This study was funded by the Patient-Centered Outcomes Research Institute (PCORI). It was reviewed and approved by the Colorado Multiple Institutional Review Board under expedited procedures. Qualitative methods and results are reported in alignment with COnsolidated criteria for REporting Qualitative research (COREQ).34
Data Collection
We conducted semistructured key informant interviews with individuals from participating practices from February through September 2023. Participants were purposefully sampled35 to select interviewees likely to have rich insights about implementation of induction approaches across varying organizational environments. We aimed for representation from multiple roles of the care team from each sampled practice, including physician or advanced practice clinician, clinical staff, and nonclinical staff. We also selected practices representing varied organizational settings, sizes, and geographic location (rural/nonrural, state, region of US).
Clinicians and staff were invited to interview through an initial e-mail from the corresponding study PI (LZ) connecting practice contacts to the interview team. Then, a member of the interview team coordinated scheduling and asked the contact to nominate any additional team members who could provide insight about offering MOUD. Interviewees participated voluntarily, provided verbal informed consent, and received a $50 gift card following study participation. We continued interview recruitment until reaching saturation, when few or no new concepts arose on topics of interest across multiple interviewees.36
The interview guide included questions about treatment protocols and workflows for providing MOUD, challenges and successes in offering MOUD, experiences with and perceptions of home, office, and telehealth induction; patient feedback about home, office, and telehealth induction; and the feasibility of continuing to provide each induction mode. Interviews took place by Zoom meeting or phone call and lasted between 20 and 50 minutes. Interviews were audio recorded and professionally transcribed.
Each interview was conducted by a team of 2 interview team members. No other persons were present outside of researchers and study participants. The interview team consisted of 4 members of the study team (TH, DF, KC, VJ), 3 female and 1 male. Three interview team members had at least master’s degree-level formal training and multiple years of experience conducting qualitative data collection and analysis and 1 team member was a physician researcher with qualitative training. Interview team members had general knowledge of MOUD induction approaches through several years of research experience in this area. Interviewers did not have prior working relationships with interview participants before this study, though many practices had established relationships with the 2 study PIs and recruitment team through previous collaborations. Interviewers completed summary forms37 following each interview to capture topics of discussion, subjective reflections, potential interview guide revisions, and considerations for future interviews. We provided background information to participants about the HOMER study in the e-mail invitation, and again verbally before beginning the interview. This included a general statement about the researchers’ interest in learning the most patient-centered and effective ways to offer MOUD induction.
Data Analysis
We qualitatively analyzed interviews using a combination of editing and template coding styles.38,39 We first used template coding to apply a priori codes based on the overarching study framework in a deductive approach that allowed us to organize data for the next step of inductive coding.38 A priori codes were based on the study’s focus on the combined influences of induction mode, patient characteristics, and practice characteristics, and included codes such as induction-practice fit and patient induction preferences/experiences. We then began the editing coding38 stage to identify inductive codes emerging from the text through multiple reviews of transcripts. After segmenting the text using template coding, all members of the interview team reviewed a predetermined selection of 2 transcripts, then met to discuss possible emergent codes to apply to transcripts. Over a series of 3 meetings, we discussed a random selection of 3 additional transcripts to refine the coding structure. Emergent codes included social determinants of health and changing drug trends (prescription and street). See Table 1 for high-level codes and definitions of the final coding structure. The interview team then met every other week over 4 months to refine the coding structure, discuss emergent themes, and develop a summary of themes, while reviewing coded text between meetings. Throughout this process, we used immersion-crystallization to iteratively review full transcripts and coded text to identify patterns within and across codes and interview themes, moving between immersion in the data and reflection on analysis.40 Three members of the study team (TH, DF, VJ) coded interviews. These 3 team members coded each of the first 5 transcripts analyzed, with 2 coders assigned to each transcript, meeting periodically to compare coding and resolve conflicts. On reviewing double coding of the final set of shared transcripts, the majority of applied codes were the same across both coders for 2 transcripts. At this point the team had developed a relatively stable coding structure and assigned the remaining 33 transcripts to the 3 coders for individual coding. We annotated sections of transcripts with memos while reviewing to capture emergent themes and ideas for discussion. We then reviewed all quotations assigned to each code using ATLAS.ti quotation reports to identify common themes within and across interviews. Qualitative analysis of interview transcripts was performed using ATLAS.ti (Version 23, ATLAS.ti Scientific Software Development, GmbH). No formal member checking or transcript returns were conducted. The analysis team presented preliminary themes to the larger study team for interpretation and questions throughout the analysis process. Practices were assigned unique identification numbers for attribution of quotes; participant numbers with an “a” or “b” are used to distinguish quotations from different members of the same practice.
High-Level Codes and Definitions Applied to Transcripts of Interviews with Primary Care Clinicians and Staff
Results
Participant Characteristics
We invited representatives from 31 practices to participate in interviews and completed interviews with members of 21 practices from 13 states for a response rate of 68%. Out of 10 nonparticipating practices, 5 did not respond to interview invitations, 3 agreed to participate but did not respond to scheduling attempts, and 2 declined to participate (reasons: team is short-staffed, key informant has moved on to new role). A total of 38 practice members participated in interviews. Most participants were physicians or clinical staff. There was representation from multiple practice organization types and regions of the US across 17 states, including both rural and nonrural areas. See Table 2 for characteristics of participating and nonparticipating practices and individual clinicians and staff who took part in interviews.
Characteristics of Practices and Individual Participants of Key Informant Interviews, Home vs. Office vs telehealth for Medication Enhanced Recovery (HOMER) Study, 2023 (n = 38 Clinicians and Staff; n = 21 Practices)
Interviews with clinicians and staff elucidated themes related to the feasibility and acceptability of office, home and telehealth induction; approaches to communicating with patients who are candidates for MOUD, and adaptations to care in response to environmental changes related to changes in drug use patterns.
Fit of Home, Office, and Telehealth Induction
Home induction among study practices has become more common and widely accepted. When many programs first began offering MOUD, most inductions occurred in an office setting. Now, the default induction setting in many cases is home. Both prescribers and patients became more comfortable and experienced with buprenorphine. Patients’ increasing comfort is partially the result of illicit buprenorphine: “There’s so much Bup on the street that if they wanted to induct themselves, they could. There’s plenty of Bup on the street.” –Physician Assistant (Participant 3). Clinicians have become more confident with time and positive experiences that home induction works, buprenorphine is safe, and do not see it as particularly risky. COVID-19 contributed to this shift of necessity when office visits were no longer feasible: “Post-COVID, some of the stuff that HOMER was trying to answer has already been answered because we were forced into that situation.” –Physician (Participant 20). Practice members indicated that patients favor home over in-office inductions for reasons of comfort and logistics. Patients prefer to be home when experiencing withdrawal symptoms.
Office inductions are difficult for clinics for several reasons. Many do not have a pharmacy nearby, requiring either that practices keep buprenorphine onsite—which is undesirable and difficult due to regulatory requirements—or that patients travel long distances to the pharmacy and back to clinic for induction. Further, many practices lack the space and time necessary for office induction. Patients come to the office before starting to experience withdrawal symptoms to avoid arriving while actively withdrawing. This makes office inductions time intensive and leads to patients feeling uncomfortable for a long period of time while waiting to be ‘sick enough’ to be induced in the office. Clinicians noted that withdrawal from fentanyl can be especially uncomfortable for patients, adding complication to office induction.
Several participants reported limited to no experience with telehealth inductions. Some highlighted similarities between telehealth and home inductions due to the inability to measure vital signs. Others had incorporated telehealth into their practice, though it was more common for this to be part of maintenance treatment (such as alternating office and telehealth follow-up visits) than induction. Reliable phone, computer or internet access affects patients’ ability to do telehealth induction or follow-up visits. Telehealth visits are not practical for many people: “A lot of them do not have computers, so the only thing they have got is their cell phone, and so we end up—there are people we just cannot do Zoom with. And particularly in the induction phase.” –Practice Manager (Participant 2B).
Regardless of induction mode, other social determinants of health including transportation and housing hinder initiation and sustainment of MOUD treatment. Lack of transportation affects many patients’ ability to get follow-up treatment, get to the clinic to start induction or even to a pharmacy to pick up their prescription. Housing came up frequently as a factor in patients’ success with treatment regardless of induction mode – an unstable or unsupportive housing situation negatively affects their ability to sustain MOUD treatment. Affordability and insurance also affect patients’ ability to participate in treatment in general.
Clinicians and Staff Communicate to Build Trust with Patients
A common approach and mindset for MOUD among practice members and clinicians involved honest communication, trust, and clear consequences in the relationship with patients, without the threat of patients being “fired.” Peer support from staff members with lived experience with MOUD treatment helped contribute to this positive relationship in some practices. Clinicians and staff described reassuring patients that it is acceptable to return to treatment after relapse and that they would not be judged or punished for this by the care team. Multiple interviewees described this as “meet[ing] them where they are at.” In return, participants noted the importance of patients being honest with the care team about substance use and treatment adherence. Prescribers and care team members with this mindset view MOUD the same as treatment for other chronic diseases such as diabetes or heart disease. However, many also acknowledged that stigma toward the idea of providing care to people struggling with substance use disorders remains common among health care providers.
Primary Care Practices Are Adapting MOUD Care in Response to Drug Use Patterns
Prescribers reported that they need to prescribe varying doses of medication at induction depending on patient characteristics including previous experience with buprenorphine and current exposure to buprenorphine or fentanyl. Practice members also observed that they are performing fewer inductions in recent years, attributing this to increased inductions in other medical settings, the wide availability of illicit buprenorphine, and increased fentanyl use. Practice members describe patients receiving MOUD induction in emergency rooms, at treatment facilities, and during hospitalizations, then being referred to primary care clinics after discharge for maintenance and dose adjustments.
In addition, clinicians reported that a high proportion of patients seeking MOUD treatment have fentanyl in their system. This made induction dosing needs higher and more variable, making it difficult for many clinicians to tailor induction dosing appropriately. Fentanyl exposure has also made withdrawal and induction more difficult for patients. Injectable, extended-release buprenorphine is becoming increasingly popular among MOUD prescribers and patients due to its ease of use for patients, who do not have to remember to take oral medication on a daily basis, and benefits for prescribers, who do not have to worry as much about medication diversion or adherence as with other routes of administration. Patients that might fall into these categories are those who haven’t been able to consistently stay on oral routes of buprenorphine or for whom the prescriber has concerns about adherence or relapsing. Alternatively, some prescribers may direct patients with these characteristics to other treatment facilities in the community such as inpatient, detox or methadone prescribers. These factors reflect a field with rapidly changing conditions and treatment approaches to which practice teams must regularly adapt. See Table 3 for illustrative quotations for each theme.
Themes and Illustrative Quotations from Interviews with Primary Care Clinicians and Staff, Home vs. Office vs. Telehealth induction for Medication Enhanced Recovery (HOMER) Study, 2023
Discussion
This national study of practice teams in urban and rural locales and differing ownership models showed that primary care practices are committed to treating patients with opioid use disorder with buprenorphine. Findings of the present study indicate that primary care teams want to be patient-centered by providing buprenorphine treatment most effectively, comfortably and conveniently for the patient, in a nonjudgmental manner that is tailored to their social needs. Toward that end, to reach more people who could benefit from MOUD, primary care clinicians and staff must adapt to a constantly changing landscape and tailor treatment to patients’ specific needs— for instance, by offering home induction where appropriate, varying induction timing and dosage depending on fentanyl or other relevant exposures, and focusing on maintenance treatment when appropriate. Given these changes, primary care practices need ongoing support and treatment approaches that keep up with community needs. Active community surveillance to rapidly adapt to opioid use trends might be an important step for primary care and public health entities.
In addition to responding to environmental changes in opioid use, it is also important to tailor treatment approaches and induction location in response to specific social determinants affecting patients’ health. Lack of access to transportation and financial challenges make office induction and maintenance in treatment nearly impossible for some patients. Our findings align with previous research showing that inflexible job and work schedules, lack of transportation and long travel distances to clinics as barriers to accessing traditional, office-based MOUD.41 Primary care practices cannot address all patients’ possible social needs by themselves. They need to use strategies such as screening and referrals for social needs, data monitoring to identify health inequities, and partnering with community groups and leaders, medical-legal partnerships, social support navigators, and income security health promoters.42–45 More broadly, a systematic review found that the most commonly cited reason for not delivering substance abuse treatment was an inconducive institutional environment.46 Specifically, they cited needing trained staff or resources, staff time, acceptance of addiction interventions by staff or leadership, and reimbursement to account for added staff time and training.46 The issues of staff, resources, and time emerged most notably in our study in relation to the difficulty of office inductions compared with home or telehealth induction. These topics may not have emerged more strongly in our study because participating practices had already implemented and begun prescribing MOUD as part of study eligibility requirements, and may have already moved past these initial challenges of delivering substance use treatment. Still, the challenges of office induction underscore the need for primary care organizations to dedicate time, staff, and resources to support MOUD access.
Complementing evidence demonstrating the importance of building trusting relationship with patients,41 this study offers specific suggestions about how to build relationships through open communication, clear consequences, and empathy. There remains a need to address stigma toward MOUD and people with OUD among some clinicians. For instance, some clinicians demonstrate exaggerated fears of diversion and view people with OUD as inherently untrustworthy.20 Stigmatizing attitudes among primary care clinicians are associated with lower likelihood of delivering MOUD and lower support for policies to promote MOUD access.10 In contrast, emergency physicians view OUD as a disease and MOUD as a long-term medication, which facilitates buprenorphine prescription.6 This is consistent with our findings that practice teams that deliver MOUD viewed OUD as a chronic disease. This study’s findings about teams’ open communication and trusting approach to treating patients with OUD could be incorporated into trainings as a strategy to fight stigma or help practice teams develop a similar approach. More broadly, primary care’s traditional principles of medication-first treatment, behavioral interventions and referrals, and actively listening to patient narratives in making care decisions related to chronic disease can be applied to MOUD treatment.47 Interviewees’ approach to relationship-building with patients could be incorporated into one such chronic disease-modeled approach and leveraged to introduce resistant PCPs to concepts that may make MOUD more appealing or familiar. See Table 4 for a summary of key findings and implications.
Summary of Interview Findings and Implications
Limitations
This study interviewed clinicians and staff but does not include direct input from patients, instead relying on practice members’ reports and perceptions of patient opinions. While interviews included representation from western, midwestern, and southern regions of the US, they lacked input from practices in the northeast or pacific northwest. In addition, physicians and clinical staff were overrepresented compared with administrative staff and other health professionals, though this was an intentional sampling strategy to garner input from respondents most likely to have relevant input and description of various induction modes. This is a qualitative study that was not designed to be generalizable to all primary care practices, but rather to richly describe a specific phenomenon, various modes of MOUD induction, among practices with experience delivering MOUD.
Conclusion
These findings offer insights about the challenges facing primary care practices in their efforts to deliver MOUD to address a rapidly evolving opioid epidemic. Regulations and evidence have not yet caught up with the reality under which primary care clinicians now operate: buprenorphine is safe, and patients and practices generally prefer home induction. However, more evidence is needed to understand the overall effectiveness of home, office or telehealth induction for different patients. Future findings from the HOMER study will offer additional context about uptake and maintenance of MOUD for specific patient characteristics and practice settings.
Notes
This article was externally peer reviewed.
Funding: This study was supported by a grant from the Patient Centered Outcomes Research Institute (PCORI), IHS-2019C1-16167.
Conflict of interest: The authors report no conflicts of interest.
To see this article online, please go to: http://jabfm.org/content/38/3/539.full.
- Received for publication August 23, 2024.
- Revision received October 30, 2024.
- Revision received November 6, 2024.
- Accepted for publication January 13, 2025.
