Primary care provides medication for opioid use disorder: findings from the HOMER study

Abstract

Context

Changes in regulations related to medication for opioid use disorder (MOUD) have expanded access to MOUD in primary care. However, there has been concern that primary care practices are unwilling or unable to treat patients with OUD.

Objective

To describe the practices and patients enrolled in the Patient-Centered Outcomes Research Institute (PCORI)–funded HOMER (Comparing Home, Office, and Telehealth Induction for Medication Enhanced Recovery) research study who delivered MOUD as part of routine primary care practice.

Results

A total of 79 practices from 25 states expressed interest in participation. Sixty-two practices signed up for HOMER. Practices were typical of US primary care, accepting a variety of payers, including commercial insurance, Medicaid, Medicare, and uninsured patients, and caring for patients of across a spectrum of adult ages, races and ethnicity, education, and income. The majority had health insurance (82%). Most patients reported using prescription opioids (59%), while 41% reported other opioid use. Greater than 40% of participating patients reported no prior medication treatment for OUD.

Conclusion

The finding that nearly half of patients had no prior treatment supports the importance of primary care as a crucial component of MOUD. Practices in HOMER were similar to practices across the country. Patients enrolled were typical of family practice patients. Policies that support primary care MOUD may improve access to patients.

Introduction

The opioid epidemic has continued despite aggressive efforts to limit prescribing opioids, decrease the flow of illicit opioids, and increase access to treatment for opioid use disorder (OUD). While deaths from opioid overdose have fallen, there are still over 80 000 deaths per year from drug overdose.^1-3 Changes in regulations related to medication for OUD (MOUD) have expanded access to buprenorphine. Emergency room programs may offer short-term treatment for patients presenting in withdrawal and referral to ambulatory MOUD programs.⁴ Formal addiction treatment centers are not equipped to manage the huge influx of patients seeking treatment.⁵ Soon after expanding access to buprenorphine there was optimism that patients could receive initial and long-term treatment in primary care.^6-8 However, many patients may not know that their primary care team can provide MOUD.⁹ Other reported barriers to primary care providing MOUD include lack of resources, lack of interest, and lack of time in an already busy clinic schedule.^10-12 Early-career family physicians may not feel adequately trained to provide MOUD.¹³ As primary care consolidates into larger health systems, there may be less interest from hospital system administrators to offer MOUD.^14,15 Taking on new patients with OUD may be less attractive to primary care practices.¹⁶

More recently, there has been a concern that primary care practices are becoming less willing or able to treat patients with OUD.¹⁷ State and federal opioid policy has focused on temporary pop-up clinics and mobile treatment vans^18,19 and telehealth.^20-23 While well intended, these efforts syphon resources away from primary care and disrupt existing continuity relationships.²⁴ There appears to be an ongoing misconception that primary care may not be interested in offering MOUD.^25-27 Several recent publications report on low uptake of MOUD among trained primary care professionals.^28,29 Because some primary care physicians are now refusing to prescribe opioids to patients with chronic pain, there may be a belief they are also unwilling to provide MOUD.³⁰

However, many primary care practices can and do offer medication for OUD.^8,31-33 More information about primary care practices that provide MOUD and the patients they reach is needed to reinforce clinical best practices, help patients understand their treatment setting options, and inform policies that support primary care delivery of treatment of OUD. As part of a national Patient-Centered Outcomes Research Institute (PCORI-)–funded comparative effectiveness study comparing patient treatment outcomes by home, office, and telehealth buprenorphine induction, we engaged dozens of primary care practices around the nation.³⁴ This article describes the practices and the patients that engaged in the HOMER (Comparing Home, Office, and Telehealth Induction for Medication Enhanced Recovery) study. The purpose of this study is to provide evidence to inform policies that support primary care as an important component of a comprehensive strategy to address OUD.

Data and methods

The HOMER study was designed as a patient-randomized comparative effectiveness trial to compare outcomes among patients starting MOUD using home- vs office-based vs telehealth induction. The initial protocol has been described elsewhere.³⁴ Due to feedback from participating practices, the trial was modified to include patients for whom there was a strong preference for induction approach, resulting in a randomized study group and a nonrandomized study group.³⁵ HOMER was conducted by the State Networks of Colorado Ambulatory Practices and Partners (SNOCAP) at the University of Colorado in partnership with the American Academy of Family Physicians National Research Network housed at DARTNet Institute. Both networks sent invitations to network member clinicians and practices to elicit interest. Over 250 practices expressed interest in participating. Participation required that the practice had conducted at least 1 buprenorphine induction prior to starting in the study and the ability to use the 3 induction approaches (home, office, or telehealth).

As part of the practice enrollment process, interested practices were asked to complete a Practice Characteristics Survey, which collected information such as practice type, size, and patient population, and a modified ITMATTTRs Implementation Checklist⁸ to describe the MOUD elements implemented across practices. The HOMER study launched in the summer of 2020, well into the COVID-19 pandemic lockdown era. The onset of COVID-19 resulted in several practices canceling their planned participation due to patient volume, staffing, and financial instability.

Patient enrollment occurred between March 2021 and January 2024. Demographic and other characteristics were collected at the time of enrollment and in a baseline participant survey. Descriptive statistics were generated for practice and patient characteristics. The study protocol was reviewed and approved by the Colorado Multi-Institutional Review Board (COMIRB) and other local IRBs as deemed necessary by participating practices and institutions.

Results

After initial contact, 79 primary care practices expressed ongoing interest and completed a practice enrollment survey, 62 practices signed up for the study, and 45 practices ultimately had at least 1 patient enroll in the study. The 79 practices were located in 25 states. Those that went on to participate were located in 21 states, and those with enrolled study participants were located in 18 states. One practice had not provided MOUD initially, but had just been trained, was prepared to offer MOUD, and was accepted into the study (see Figure 1).

Figure 1.

Primary care practices providing MOUD interested in HOMER. Blue dots represent primary care practices that reported offering MOUD, expressed interest in participating in HOMER, and provided baseline practice descriptions. Abbreviations: HOMER, Comparing Home, Office, and Telehealth Induction for Medication Enhanced Recovery; MOUD, medication for opioid use disorder.

Table 1 provides a description of the practices that were interested in or signed up to participate in HOMER. Overall, practices included independent private practices (25%), academic practices (20%), and federally qualitied and other community health centers (40%). A total of 928 clinicians provided care in the interested and participating practices.

Table 1.

Characteristics of practices offering medication for opioid use disorder.

	No. (%) (total = 79a)
Practice type (practices could choose all that apply)
Federally qualified health center	26 (34.7%)
Hospital-based or health system–owned clinic	25 (33.3%)
Private practice/clinician-owned solo or group practice	19 (25.3%)
Community health center	3 (4.0%)
Academic medical center	15 (20.0%)
Nonfederal government clinic	2 (2.7%)
Free and charitable clinic	1 (1.3%)
Rural health clinic	7 (9.3%)
Other clinic type	5 (6.7%)
University affiliate	26 (32.9%)
Residency program	35 (44.3%)
Practice staff	Median number of staff (range)
Clinicians (MD, DO, NP, PA)	5 (1, 69)
Registered nurses	1 (1, 24)
Licensed practice nurses	1 (1, 15)
Medical assistants	3 (1, 25)
Certified nursing assistant	1 (1, 12)
Front desk staff	3 (1, 30)
Practice managers/administrators	1 (1, 18)
Behavioral health providers	1 (1, 40)
Medical records staff	1 (1, 9)
Total number of employees	18 (1209)
Practice patient characteristics (estimated by practice)	Median % (range)
Asian	1% (0%-20%)
American Indian or Alaska Native	1% (0%-20%)
Black or African American	7% (0%-74%)
Native Hawaiian or Pacific Islander	0% (0%-20%)
White	73% (0%-99%)
Multiracial	2% (0%-49%)
Hispanic or Latino	18% (0%-80%)
Male	45% (6%-65%)
Female	54% (30%-94%)
Nonbinary or transgender	1% (0%-30%)
Payer mix	No. (%)
Medicaid
0%	5 (6.4%)
1% to 50%	52 (66.7%)
>50%	21 (26.9%)
Medicare
0%	2 (2.6%)
1% to 50%	74 (94.9%)
>50%	2 (2.6%)
Private insurance
0%	3 (3.8%)
1% to 50%	61 (77.2%)
>50%	14 (18.0%)
No insurance/self-pay
0%	7 (9.2%)
1% to 50%	67 (88.2%)
>50%	2 (2.6%)
Practice community size	No. (%)
Metropolitan	53 (67%)
Micropolitan	7 (9%)
Small town	13 (16%)
Rural	6 (8%)
Telehealth visits prior to COVID-19	12 (15.4%)
Telehealth visits since COVID-19	77 (97.5%)

^aSeventy-nine practices expressed interest and reported they were providing MOUD, 62 practices joined HOMER, and 45 practices enrolled a patient into the randomized trial. Not all practices provided data for every characteristic. One practice had not conducted an induction prior to sign up but had just been trained and was planning buprenorphine treatment soon.

Abbreviations: HOMER, Comparing Home, Office, and Telehealth Induction for Medication Enhanced Recovery; MOUD, medication for opioid use disorder.

Practices reported a median of 3569 active patients—54% female, 73% White, 7% Black or African American, and 18% Hispanic/Latino/a/x. Practices had a median of 5 clinicians (range, 1–69), with a median of 18 total employees (ranging from 1–209). Practices reported accepting a wide variety of payers including commercial insurance (96%), Medicaid (94%), Medicare (97%), and uninsured patients (91%). Only 15% of practices reported any telehealth clinical visits prior to COVID-19, while 98% reported using telehealth since COVID-19. Fifty-three practices were metropolitan, 7 micropolitan, 13 small town, and 6 rural. The practices that participated in HOMER were similar to practices described in several nationally representative reports on primary care and family medicine.^36,37

At the start of the HOMER study, prescribing buprenorphine for OUD required a Drug Enforcement Administration (DEA) waiver. Practices reported a median of 4 clinicians with a DEA waiver, with a range of 1 to 38. This represents approximately 43% of the clinicians in the practices. Ninety-seven percent of practices reported doing buprenorphine induction for OUD prior to joining the HOMER study, with 65% reporting more than 20 patients receiving MOUD in the past year, 16% reporting 11–20 patients, and 17% reporting 1–10 patients.

A wide variety of patients enrolled in the HOMER study. Table 2 shows patient characteristics. Almost half were high school graduates, and 9% were college graduates. The majority had health insurance (82%). There were a variety of marital statuses, with a plurality of patients reporting they were single–never married (39%), followed by married/partnered (33%). A majority of patients reported using prescription opioids (59%), while 41% reported other opioid use (fentanyl, heroin, etc). Seventy-three percent reported experiencing a significant traumatic life event, just over 5% had military service, and 35% reported that religious practice and spirituality meant a great deal to them. Nearly 40% of participating patients reported no prior buprenorphine treatment for OUD.

Table 2.

HOMER patient characteristics—self reported.

Patient characteristics	No. (%) (n = 294)a
Age, mean (SD), y	39.5 (11.8)
Sex assigned at birth
Female	148 (50.7%)
Male	144 (49.3%)
Gender
Man	145 (49.7%)
Woman	145 (49.7%)
Prefer not to answer	2 (0.7%)
Race/ethnicity
American Indian or Alaska Native	22 (7.5%)
Asian	2 (0.7%)
Black or African American	24 (8.2%)
Pacific Islander	2 (0.7%)
White	233 (79.3%)
Other	16 (5.4%)
Hispanic, Latino/a, or Spanish origin	53 (18.4%)
Education
Some high school or less	50 (17.1%)
High school graduate	138 (47.3%)
Some college or technical school	78 (26.7%)
College graduate	26 (8.9%)
Annual household income
Less than $15 000	112 (44.3%)
$15 000 to less than $35 000	66 (26.1%)
$35 000 or more	75 (29.6%)
Health insurance
Yes	210 (81.7%)
No	29 (11.3%)
Don’t know/not sure	17 (6.6%)
Marital status
Never married	99 (38.5%)
Married/partnership	84 (32.7%)
Divorced/separated/widowed	74 (28.8%)
Living situation (select all that apply)
Living alone	34 (13.3%)
Living with significant other/partner/spouse	105 (41.0%)
Living with children (including foster)	66 (25.8%)
Living with roommate(s) or housemate(s)	38 (14.8%)
Living with extended family	58 (22.7%)
Other living situation	30 (11.7%)
Household health conditions
Partner/children with chronic health issue	41 (16.0%)
Live with person who uses/abuses opioids	39 (15.2%)
Type of opioids currently using
Prescription (prescription opioids)	147 (58.8%)
Other opioids (heroin, fentanyl, etc)	103 (41.2%)
Experienced traumatic event in life	187 (73.0%)
Military service	14 (5.5%)
Meaning of spiritual/religious practice
None at all	41 (16.0%)
Not much	51 (19.8%)
Some	76 (29.6%)
A great deal	89 (34.6%)
Previous buprenorphine OUD treatment	177 (60.6%)

^aNot all patients answered every question; therefore, characteristic totals may not be the same as the overall total.

Abbreviations: HOMER, Comparing Home, Office, and Telehealth Induction for Medication Enhanced Recovery; OUD, opioid use disorder.

Discussion

Despite challenges reported in the literature, HOMER practices and patients provide a snapshot of primary care practices that provide MOUD. The HOMER study demonstrates that many primary care and family medicine practices do provide MOUD as part of their daily clinical practice. While not a random sample of family medicine, the practices that participated in HOMER were recruited from several well-established, practice-based research networks, and were similar to practices described in the Family Medicine Factbook.³⁷ Practices in this study included a wide variety of types from solo practice to larger health system, independent private practice, community health centers, and hospital-based ambulatory practices in urban, rural, and suburban communities across 25 states. There was a broad spectrum of ownership, organization, and structure. Payer mix among HOMER practices was variable, similar to that described in the Factbook. Prescribing clinicians included physicians, advanced practice nurses, and physician assistants. The patients enrolled in this study were similar to primary care patients reported by Willis et al,³⁸ with a diverse demographic profile (age, gender, race/ethnicity), medical histories, comorbid conditions, and a variety of social determinants of health.

Practices that participated in the HOMER study may represent a group of practices with higher commitment to serving patients with OUD. While approximately 43% of the clinicians in the participating practices had a DEA waiver, the Family Medicine Factbook³⁶ reports that, among mid- to late-career family physicians, only approximately 7% prescribe buprenorphine for OUD. While prescribing for OUD is higher for early-career physicians (13%) and resident physician intentions to prescribe MOUD (25%), HOMER practices had a higher rate of buprenorphine prescribing physicians. Some practices included just 1 or 2 buprenorphine prescribers, so offering MOUD to their patients did not require the entire practice or many buprenorphine prescribers.

An important finding from HOMER was that 40% of those patients enrolled in HOMER reported no prior buprenorphine treatment for OUD. Buprenorphine treatment is considered standard first-line therapy for patients with OUD.³⁹ First-time access to buprenorphine treatment demonstrates the importance of primary care in identifying and treating patients who need MOUD. In addition, the majority of patients were using prescription opioids, supporting the importance of primary care as a site for MOUD. Policies that support primary care provision of buprenorphine treatment offer a crucial option for patients with OUD, particularly in rural and underserved urban communities. Limiting MOUD to specialists, addiction clinics, or other limited-scope clinical practices may create more barriers for patients with OUD.⁴⁰ Solutions such as mobile vans that offer MOUD in rural communities on an itinerant basis extract resources from the local primary care, where patients are more likely to receive comprehensive and continuity care. Including primary care in policy and funding decisions offers the best opportunity to address gaps in care and improve access to MOUD.

Several limitations deserve mention. First, while we sent an invitation to hundreds of practices, and while several hundred initially expressed interest, only 79 provided practice-level data, and just 62 formally signed up for participation in HOMER. Thus, our sample of participating practices was a convenience sample. However, primary care practice–based research networks, because they are more representative of clinicians and practices, have been found to be a crucial element of meaningful translational and effectiveness research.⁴¹ Our findings describe a variety of typical family practices from all over the United States that are providing MOUD. Second, HOMER was a research project that initially required willingness to randomize patients to a specific induction treatment arm. We found that many patients and clinicians were uncomfortable with a random assignment to treatment, opting for a shared-decision-making approach to starting MOUD.⁴² This resulted in a lower number of interested practices, and a significant number of practices not enrolling any patients in HOMER. While still offering MOUD to their patients, they were not willing to enroll them in a randomized trial. These findings ultimately prompted a change in protocol to include a comprehensive cohort study design.³⁵ This addition required significant analytic change to the comparative effectiveness trial but does not impact the findings reported in this study. Finally, results from the comparative effectiveness study comparing methods for initiating MOUD are beyond the scope of this study, which sought to answer the question of whether primary care practices, and family medicine specifically, are providing MOUD in their practice.

Implications for policy and research

• Because many patients with OUD find their first medication treatment in primary care, policies that support primary care MOUD may improve access to treatment.

• Primary care clinicians prefer a patient-centered approach with shared decision making, so policies should support a variety of mechanisms for buprenorphine initiation and long-term treatment.

• Understanding the impact of behavioral health, whether it is integrated into primary care or community based, may require policies that support behavioral health and OUD, and further research on the combination of medication and behavioral health on OUD outcomes.

• Additional research is needed to describe and optimize collaboration and communication between various MOUD providers—emergency room, primary care, mobile clinics, and community treatment organizations.

Conclusion

Many primary care and family practices offer MOUD to their patients in their communities and serve as the first access to buprenorphine treatment for a substantial number of people with OUD. We hope this description will encourage other practices to consider offering MOUD for their patients. Our findings provide evidence to support primary care as an important component of a comprehensive strategy to address OUDs. Policies that support MOUD in primary care may improve access to patients with OUD.

Disclaimer

The views expressed are solely those of the authors and do not necessarily represent official views of the authors' affiliated institutions or the Patient-Centered Outcomes Research Institute (PCORI).

Supplementary material

Supplementary material is available at Health Affairs Scholar online.

Funding

Research reported in this publication was funded through a Patient-Centered Outcomes Research Institute (PCORI) award (IHS-2019C1-16167).