Turner BJ, Laine C, Lin Y, Lynch K. Barriers and Facilitators to Primary Care or Human Immunodeficiency Virus Clinics Providing Methadone or Buprenorphine for the Management of Opioid Dependence. Arch Intern Med. 2005;165(15):1769-1776. doi:10.1001/archinte.165.15.1769
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
Federal initiatives aim to increase office-based treatment of opioid dependence, but, to our knowledge, factors associated with willingness to deliver this care have not been defined. The objective of this study was to describe clinics’ willingness to provide methadone hydrochloride or buprenorphine hydrochloride for opioid dependence.
The design of the study was a survey conducted in New York State. Two hundred sixty-one directors of primary care and/or human immunodeficiency virus specialty clinics (response rate, 61.1%) that serve Medicaid enrollees were questioned. Outcomes were willingness to provide methadone and buprenorphine. Predictors included clinic characteristics, attitudes about drug users and their treatment, and reported barriers and facilitators to treatment.
Clinics were more willing to provide buprenorphine than methadone treatment (59.8% vs 32.6%; P < .001). Clinics offering human immunodeficiency virus specialty care (adjusted odds ratio [AOR], 2.16; 95% confidence interval [CI], 1.18-3.95) or a safe location to store narcotics (AOR, 2.99; 95% CI, 1.57-5.70) were more willing to prescribe buprenorphine and more willing to provide methadone. Willingness was positively associated with continuing medical education credits for training, but negatively associated with greater concern about medication abuse. Immediate telephone access to an addiction expert was associated with willingness to provide buprenorphine (AOR, 2.08; 95% CI, 1.15-3.76). Greater willingness to provide methadone was associated with a belief that methadone-treated patients should be seen along with other patients (AOR, 6.20; 95% CI, 1.78-21.64), methadone program affiliation (AOR, 4.76; 95% CI, 1.64-13.82), and having more patients with chronic pain in the clinic (AOR, 2.80; 95% CI, 1.44-5.44).
These clinics serving Medicaid enrollees were more receptive to buprenorphine than methadone treatment. Willingness to provide this care was greater in clinics offering human immunodeficiency virus services, treating more chronic pain, or affiliated with methadone programs. Accessible addiction experts and continuing medical education for training may facilitate adoption of this care.
Long-term treatment for opioid dependence has yielded dramatic health benefits for drug users.1,2 Society also benefits from drug users’ reduced criminal activity3,4 and lower costs of care.5,6 The United States has had a limited-capacity system of segregated specialized clinics dedicated to treating drug dependence, while in Australia and Europe, primary care physicians deliver office-based management of opioid dependence.7- 15 Because the US treatment capacity does not meet demand and the office-based model of treatment for opioid dependence has been successful abroad and in domestic pilot programs,16- 21 modified federal regulations permit management of opioid dependence by qualified, trained, office-based physicians.22,23
Methadone hydrochloride and buprenorphine hydrochloride are the primary pharmacological treatments for opioid dependence. The latter drug is a partial opiate agonist approved by the Food and Drug Administration in 2002. Generalists’ willingness to participate in office-based management of opioid dependence with either of these medications has been evaluated in only a few providers (ie, physicians and nurse practitioners or physician assistants),24 and only a few attitudinal issues have been addressed.25 To our knowledge, no studies have focused on a broad set of factors that prevent or promote primary care management of opioid dependence.
We surveyed a random sample of directors of New York State (NYS) primary care or human immunodeficiency virus (HIV) specialty services clinics that serve Medicaid enrollees. We compared clinic directors’ willingness to provide methadone with willingness to prescribe buprenorphine. We also examined associations of clinic characteristics and attitudes with willingness to offer these treatments.
New York State Department of Health officials provided contact information for all primary care and/or HIV clinics treating at least 20 Medicaid enrollees in 2000. Of the 710 state-identified clinics, we excluded 151 duplicate or closed clinics, 38 in a qualitative survey, and 94 reporting no primary care services. Of the 427 remaining eligible clinics, 261 (61.1%) completed surveys.
Qualitative interviews with 27 NYS primary care clinic directors and officials at the Center for Substance Abuse Treatment guided the development of the survey. It was pilot tested with 6 Philadelphia primary care clinic directors.
The survey asked about clinic characteristics, including affiliations, types of providers and patients, receipt of enhanced payments for HIV care, experience with narcotics, management of drug users, and substance abuse treatment expertise. Questions also addressed knowledge and attitudes about buprenorphine, methadone, opioid-dependent persons, and federal efforts toward office-based treatment for drug users. As facilitators, we asked about continuing medical education (CME) credits and payment for training, payment for this care, and access to addiction experts. Regarding barriers, we asked about knowledge about these drugs, personnel, financing, medicolegal risks, paperwork, patient problems, staff resistance, and community resistance.
Mathematica Policy Research, Inc, Princeton, NJ, conducted the survey in 2003. To introduce the interview, Mathematica Policy Research, Inc, sent letters to clinic directors signed by one of us (B.J.T) and a senior official of the NYS Office of Alcohol and Substance Abuse Services. The letter briefly described office-based methadone maintenance and buprenorphine, citing several recent publications,26,27 and introduced our survey. Respondents were offered $60. If the director managed multiple clinics, this individual completed separate interviews for each site or named an alternate to respond for each clinic.
Our primary end points were defined as responding “extremely” or “somewhat likely” to the following questions: “If federal officials offer financial and other incentives for methadone treatment, what is the likelihood that your clinic will implement it?” and “Given what you know about buprenorphine, what is the likelihood that at least some physicians in your practice will be trained to use it?”
Independent variables included academic affiliation, New York City location, affiliation with a methadone program, and on-site narcotics storage. Human immunodeficiency virus specialty care was defined as receiving enhanced Medicaid payment rates for HIV care or having HIV expertise. Enhanced payment for HIV care started in 1989 when the NYS Medicaid program created Designated AIDS Centers in institutions meeting service availability, accessibility, expertise, and quality-of-care criteria.28 Several years later, the HIV Primary Care Medicine Program was created for other hospital and ambulatory care sites that meet quality-of-care standards set by the NYS AIDS Institute (information available at: http://www.idsociety.org/Template.cfm?Section=Home&CONTENTID=9698&TEMPLATE=/ContentManagement/ContentDisplay.cfm; accessed February 28, 2005). These providers could be generalists with HIV expertise and infectious disease experts. Because several numerical variables had noncontinuous associations with the outcomes, these data were analyzed by quartile. Because the lowest quartile was consistently the outlier, we analyzed the following variables as number of clinic physicians (lowest quartile, ≤2; and highest, >2), patient visits per week (low, ≤127; and high, >128), clinic patients needing chronic pain management (low, ≤5%; and high, >5%), and clinic patients actively abusing narcotics (low, <1%; and high, ≥1%). The latter 2 variables were also divided at the highest quartile (>30% for patients with chronic pain and >5% for narcotic abusers). Although significant bivariate associations with the outcomes appeared, these did not persist after adjustment; therefore, we retained the lowest quartile cut points.
We defined favorable attitudes toward methadone as the sum (0-4) of “strongly agree” or “agree” responses to 4 items: well tolerated, effective for narcotic dependence, few adverse effects, and prevents criminal activity. Similarly, we summed (0-2) strongly agree or agree responses to 2 questions about negative aspects of methadone: keeps patients narcotic dependent and continued too long. We similarly categorized responses to items about positive opinions of patients receiving methadone (ie, want to stop taking drugs, want to combine primary care and methadone maintenance, and no different than other patients) and negative items (ie, noncompliant, severe mental health problems, and threatening to other patients and staff). We also summed 4 actions that could be taken with patients who abuse prescribed narcotics: confront and tell the patient not to do it, tell the patient to enter a drug treatment program, ask the patient to leave the practice, and change to nonaddictive pain medications. Questions about concerns regarding methadone maintenance were grouped as extremely or somewhat vs slightly or not at all.
Opinions about ways to promote office-based treatment for opioid dependence were examined as follows: very or somewhat high, neutral, and very or somewhat low. Because we thought that only strongly held opinions would motivate clinics to participate, we analyzed responses to facilitators to office-based treatment as very important vs all other groups. We defined opinions about barriers as favorable vs neutral-negative opinions.
The unit for all analyses is clinic. We compared the 261 respondents with the 157 nonrespondents using NYS Department of Health data, including location, HIV specialty status, and freestanding vs hospital-based location. Respondents for 6 of the 261 sites did not answer more than one third of the questions and were excluded. Of the remaining 255 clinics, the final analytic sample includes 248 clinics responding about willingness to provide methadone and 249 responding about willingness to prescribe buprenorphine. We used generalized estimating equations to account for the clustering of clinics in 178 health care systems. All reported P values account for clustering, and significance is set at <.05. Our initial logistic regression model considered variables associated with willingness at P < .20 in bivariate analyses. From these initial models, we selected variables associated with willingness at P < .20 plus region (simplified into New York City vs rest of the state) in the final multivariate models. The item about the opinion of buprenorphine was highly correlated with other key independent variables and excluded from the final model. For our final multivariate models, we imputed missing values for survey items using multiple imputation.29 Models with imputed values did not differ substantively from those excluding missing values, so we report the former. To evaluate goodness of fit, we examined the Hosmer-Lemeshow statistic from logistic regression models unadjusted for clustering. It was not significant, suggesting that lack of fit was not an important concern. Analyses were performed using SAS statistical software, version 8.0 (SAS Institute Inc, Cary, NC).
Respondents did not differ from nonrespondents in hospital affiliation, geographic location, or HIV specialty status (P> .10). Participating clinics come from a range of geographic locations, affiliations, clinic size, and experience with narcotics and drug abuse (Table 1). Overall, one third of the clinics reported willingness to provide methadone treatment, while 59.8% reported willingness to prescribe buprenorphine (P< .001). Predictors of both outcomes included affiliation with a methadone program, storage of narcotics on site, a secure place to store narcotics, and HIV specialty care. Clinics with more clinic patients who abused narcotics were more willing to provide methadone, while clinics with a high volume of patient visits were more willing to prescribe buprenorphine.
Clinic directors with more positive opinions of persons taking methadone (P = .02) and less concern about abuse or overdose of methadone (P = .03) were more likely to be willing to provide methadone (data not shown). Although half of the respondents replied that methadone keeps patients dependent on narcotics and treatment is continued too long, these negative attitudes were not associated with willingness to provide methadone. Only 25.4% of clinic directors endorsed all 4 negative statements about persons receiving methadone (ie, noncompliant, severe mental health problems, threatening to other patients, and threatening to staff), and these responses were not associated with willingness to provide methadone. Most clinic directors judged that giving a month’s supply of methadone was reasonable for some (50.8%) or most (30.2%) stabilized patients, but 87.1% were concerned about street diversion of methadone. Of the clinics, 12.5% would see patients receiving methadone at a separate time or location from other patients.
Of the respondents, 48.2% had heard about buprenorphine, and they tended to be more willing to prescribe it (P = .05). Greater willingness to prescribe buprenorphine was related to an opinion that buprenorphine is a good idea (P < .001), more positive opinions about persons receiving methadone (P = .01), and less concern about abuse (P = .004). More than two thirds of respondents affirmed the importance of sharing management with an addiction expert or immediate telephone access to such an expert. Of the respondents, 61.4% endorsed the importance of CME credits for buprenorphine training.
Table 2 shows barriers and potential facilitators to office-based methadone maintenance. More than half of the clinic directors were concerned about physicians’ lack of knowledge about methadone, lack of social workers or counselors on staff, inadequate access to psychiatrists or psychologists, complex problems of opioid-dependent persons, inadequate financial reimbursement, medicolegal risks, and burdensome paperwork. Less than one third were concerned about community resistance. More than two thirds of clinic directors endorsed the following potential facilitators as important: training about methadone, immediate telephone access to an addiction expert, extra support personnel, and ability to refer difficult patients to traditional methadone programs. Slightly less than half of the clinic directors regarded an on-site psychiatrist or greater payment as important.
In multivariate analyses, clinics that would see methadone-maintained patients with other patients or that were affiliated with a methadone program had roughly 5 times greater adjusted odds of being willing to provide methadone (Table 3). Clinic directors were more than 2 times more likely to be willing to provide methadone if their clinic was characterized by HIV specialty care, more than 5% of clinic patients needed chronic pain management, and their clinic was a secure site to store narcotics. Directors who viewed CME for training as important were 2 times more likely to be willing to provide methadone maintenance. Directors who were more concerned about overdose with methadone had more than two-thirds lower adjusted odds of being willing to offer methadone.
Human immunodeficiency virus specialty care was also associated with 2-fold greater adjusted odds of willingness to train to prescribe buprenorphine (Table 4). Other predictors with 2- to 3-fold higher adjusted odds of willingness to prescribe buprenorphine were a secure storage site for narcotics and CME credits for training or immediate telephone access to an addiction expert regarded as important. Greater concern about abuse of buprenorphine was strongly negatively associated with willingness to prescribe it.
Although the feasibility and effectiveness of office-based methadone maintenance has been shown,17,20,30- 32 integration of this treatment into physicians’ offices is in its infancy in the United States. Successful approaches to expand this alternative for opioid-dependent persons could substantially increase treatment capacity33,34 and potentially improve health outcomes.35 European providers’ attitudes toward treating drug users have been examined,12,36- 40 but similar information about US providers’ attitudes are needed to guide policy makers’ efforts to increase acceptance in this country. In our statewide sample of roughly 250 medical directors of primary care and HIV specialty clinics, only one third were willing to provide methadone to stabilized drug users, while 59.8% expressed willingness to provide buprenorphine.
Buprenorphine was regarded as having lower potential for abuse41 and was less stigmatizing than methadone. But many barriers to buprenorphine treatment exist. Physicians must complete training and obtain a waiver from the federal Controlled Substances Act. After completing this training, there has been a cap on the number of patients who can be seen at a site. In addition, many clinic directors also expressed concerns about adequacy of knowledge and resources and complexity of the patient population.
Clinics offering HIV specialty care composed 41.4% of our sample and were twice as likely to report willingness to provide methadone or buprenorphine as clinics without this care. These HIV specialty clinics receive an enhanced payment for HIV services and may be staffed by general internal medicine or other primary care physicians who are also experts in HIV and/or infectious diseases. These HIV providers are likely to be accustomed to treating drug users because intravenous drug use is the HIV transmission risk factor for more than 40% of all AIDS cases in NYS.42 Human immunodeficiency virus specialists have also been long aware that methadone maintenance prevents the spread of HIV.43 Clinics with a safe location to store narcotics were significantly more likely to be willing to offer methadone or buprenorphine, but only one quarter of our sample had such a facility. In some countries, pharmacies provide methadone.44 Similar arrangements with US pharmacies might increase receptiveness of providers to offering methadone treatment.
Clinics affiliated with methadone maintenance programs had roughly 5-fold greater adjusted odds of being willing to provide methadone, but only 10.0% of our sample had such an affiliation. Conversely, most of our study clinics estimated that more than 5% of their patients needed chronic pain management, and these clinics were nearly 3 times more willing to provide methadone. Chronic pain management is a challenge for primary care providers.45 Methadone seems to be viewed as useful for patients with chronic pain,46,47 increasing receptiveness to training to use this medication.
Our survey offers insights into strategies to facilitate participation in office-based treatment of opioid dependence. Continuing medical education credits for training were viewed as important by clinics that were more willing to train to offer buprenorphine or methadone. Virtually all respondents endorsed the importance of being able to refer difficult patients back to traditional drug treatment programs. Immediate telephone access to an addiction expert was associated with a 2-fold increase in willingness to offer buprenorphine. In Scotland, addiction and mental health consultant support for primary care providers managing opioid dependence has been successful.9 Two thirds of our respondents reported that poor access to mental health specialists was a problem in adopting this model of care. Increasing expert consultant accessibility may promote office-based treatment of opioid dependence.
Nearly half of our respondents had at least 1 negative opinion about patients treated for opioid dependence, and 81.4% expressed negative opinions about methadone. Despite this stigma,48 neither negative opinions about methadone-maintained persons nor about methadone were significantly associated with willingness to provide treatment. Only 44.8% of respondents would demand greater reimbursement for the care of opioid dependence than for other types of care, but 56.9% reported a high level of concern about the adequacy of reimbursement.
Several limitations deserve mention. First, we surveyed clinics in a single state. However, of the 50 states, NYS has one of the largest numbers of opioid-dependent persons, of whom at least 44 000 are receiving methadone maintenance.49 Second, our survey focuses on primary care clinics that serve Medicaid enrollees. The generalizability of our findings to settings that serve more affluent patients is uncertain. However, Medicaid enrollees are among the most vulnerable of all insured patients and this publicly funded population is of interest to policy makers. Third, our measures of willingness to provide office-based treatment for opioid dependence come from clinic directors’ self-reports. When actually confronted with the decision to participate, more clinic directors may balk. Furthermore, directors’ attitudes may differ from those of clinic physicians. Fourth, it is uncertain whether policy makers can implement the identified facilitators to office-based treatment of opioid dependence in resource-constrained health care settings. Last, our response rate was 61.1%.50 Despite these limitations, ours are among the first available data on this issue from a large sample of real-life primary care clinics.
The greater receptiveness of clinics to buprenorphine-based therapy suggests that efforts should focus on integrating this treatment. To date, a few physicians have been trained to prescribe buprenorphine.51 A brief national poll conducted in 2003 with 419 primarily addiction experts trained to prescribe buprenorphine found that only two thirds had actually prescribed it and 19% had had trouble identifying a local source for this medication.52 Given the tenuous acceptance of this care in our sample, such frustrating obstacles must be overcome to avoid discouraging those who do train in the use of buprenorphine.
Federal officials and addiction experts have expressed hope that involving mainstream physicians will improve the care of opioid-dependent persons.35,41,53,54 In NYS, clinics delivering HIV specialty care may be particularly receptive to initiating this care. Our data support offering incentives to providers to train to prescribe buprenorphine and ensuring that adequate support services are available for mainstream physicians who manage opioid-dependent persons.
Correspondence: Barbara J. Turner, MD, MSEd, University of Pennsylvania School of Medicine, 1123 Blockley Hall, 423 Guardian Dr, Philadelphia, PA 19104-6021 (firstname.lastname@example.org).
Accepted for Publication: March 27, 2005.
Financial Disclosure: None.
Funding/Support: This study was supported by the Substance Abuse Policy Research Program, The Robert Wood Johnson Foundation, Princeton, NJ.
Role of the Sponsor: The funding body had no role in data extraction and analyses, in the writing of the manuscript, or in the decision to submit the manuscript for publication.
Additional Information: Dr Turner had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Acknowledgment: We thank Robert Lubran and Arlene Stanton, PhD, from the Substance Abuse and Mental Health Services Administration, Rockville, Md, for their assistance in developing the survey instrument; Peter J. Gallagher from the Department of Health, Albany, NY, and Alan Kott from the Office of Alcoholism and Substance Abuse Services, Albany, for their assistance in identifying the study sample and conducting the survey; and staff at Mathematica Policy Research, Inc, for performing the survey.