U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Chou R, Korthuis PT, Weimer M, et al. Medication-Assisted Treatment Models of Care for Opioid Use Disorder in Primary Care Settings [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2016 Dec. (Technical Briefs, No. 28.)

Cover of Medication-Assisted Treatment Models of Care for Opioid Use Disorder in Primary Care Settings

Medication-Assisted Treatment Models of Care for Opioid Use Disorder in Primary Care Settings [Internet].

Show details


The Technical Brief integrates discussions with Key Informants with searches of the published literature and grey literature to inform the Guiding Questions.

Discussions with Key Informants

We identified and interviewed 11 Key Informants (8 nonfederal and 3 federal) to represent broad and balanced perspectives relevant to MAT, with a focus on people with expertise or experience related to implementation in primary care settings, including rural or other underserved settings. The Key Informants represented the following stakeholder areas: researchers, clinicians (including primary care providers and experts in management of addiction), health policy, implementation, professional societies, patient groups, and federal representatives. Potential Key Informants were asked to disclose conflicts of interest prior to participation. The Agency for Healthcare Research and Quality (AHRQ) Task Order Officers reviewed conflicts of interests; we extended invitations to potential Key Informants who did not have conflicts of interest that precluded participation.

We organized and facilitated small group telephone discussions with the Key Informants (2 to 4 per call) to gain input on the Guiding Questions; group calls maximized efficiency and the relatively small number of Key Informants on each call allowed all representatives the chance to provide input. Members of our research team and the AHRQ Task Order Officers also attended the calls. On the calls, we interviewed Key Informants using a semi-structured approach. Key Informants were asked to respond to predetermined questions targeted to different Key Informant perspectives, share more general insights, and interact with each other (Appendix A). The questions were used as a guide, but we asked additional or supplemental questions based on interviewee responses. We asked which MAT models of care are in use in primary care and other related settings, including models of care which are not described in the published literature, and asked Key Informants to describe the different components of the models and which components were particularly effective or promising, the current challenges or barriers to implementation, patient preferences, and future directions, including promising new and innovative models and strategies for implementation. We also asked about specific issues to be aware of when reviewing the literature, such as outcomes to be prioritized, meaningful length of followup, study design issues, and how MAT models of care vary in terms of intensity, goals, and components of care. Because we were particularly interested the feasibility and applicability of models of care implemented in one setting or population compared with others and about identifying models of care that may be particularly suitable for specific settings, including rural and other underserved settings, we focused the questions and discussions in that area. The calls were recorded, and the key points were summarized and shared with the group for clarification and additional input. We reviewed all of the Key Informant input regarding successful and promising MAT models of care and developed a framework for categorizing the different types of components in MAT models of care, to help organize and provide a structure for future research and discussions in this area. We then integrated feedback from the Key Informants with the expertise of our project team and evidence identified from the published and unpublished literature.

Grey Literature Search

To identify grey literature, the AHRQ Evidence-based Practice Center (EPC) Scientific Resource Center sent email notification to relevant stakeholders about the opportunity to submit Scientific Information Packets via the Effective Health Care Web site.

In addition, we conducted searches of the grey literature. Specifically, we searched ClinicalTrials.gov and Health Services Research Projects in Progress (HSRProj) for ongoing research, as well as Google Scholar, NIH Reporter, and Web sites of government agencies with MAT initiatives. The grey literature searches were used to primarily inform Guiding Question 3, but if information relevant to the other Guiding Questions was identified, it is also discussed in the report.

Published Literature Search

We searched, reviewed, and summarized the available literature on MAT for OUD in primary care settings to address Guiding Question 3. An experienced research librarian created search strategies for the following databases: Ovid Medline, PsycINFO, the Cochrane Library, SocINDEX, and CINAHL. The search strategies are available in Appendix B. Since OUD with opioid agonists could not be treated in the primary care/nonaddiction treatment settings after the passage of the Harrison Narcotic Act in 1914 until the year 2000, with the passage of the Drug Addiction Treatment Act (DATA) 2000, and due to the focus of the report in primary care settings and the large volume of abstracts, we restricted the start date for the searches to the year 1995 and later (to mid-June 2016). The search was also used to identify contextual evidence to supplement the Key Informant input obtained for Guiding Questions 1, 2, and 4. We also reviewed the reference lists of identified publications and solicited additional references from Key Informants to supplement electronic searches. Searches will be updated while the report is undergoing peer and public review in order to capture any recently-added publications. If any new studies are identified from the update searches or arise as suggestions from the peer or public review, they will be added to the report prior to finalization.

We applied predefined screening criteria to identify the most relevant and authoritative evidence on MAT models of care in primary care settings. For Guiding Question 3, we focused on the following sources of evidence: (1) high-quality Cochrane systematic reviews of MAT; (2) randomized trials and cohort studies on the effectiveness of MAT models of care in primary care settings; (3) randomized trials evaluating the effectiveness of newer pharmacological therapies for MAT that could impact implementation or future models of care; and (4) randomized trials on the effectiveness of more intensive versus less intensive psychological interventions with MAT in primary care settings. To provide context for the other Guiding Questions, we also identified published and unpublished studies describing MAT models of care in primary care settings, including the setting for the model of care (e.g., urban vs. rural), patient characteristics (e.g., age, presence of comorbid conditions, OUD related to prescription opioids for chronic pain versus nonprescribed opioid use), and intervention characteristics (e.g., components of MAT models of care, including degree of coordination and intensity of psychosocial interventions). We also identified studies that provided contextual information on implementation strategies and barriers in primary care settings, including rural and other underserved settings. We excluded trials that focused on the dose or duration of pharmacological therapy, as the focus of this report was on MAT models of care, not on details regarding how pharmacological therapy should be provided.

All titles and abstracts identified through searches were independently reviewed for eligibility against our inclusion/exclusion criteria organized by PICOTS (population, intervention, comparator, outcome, timing, study design) (Table 1) by a trained member of the research team. Studies marked for possible inclusion by any reviewer underwent a full-text review. For abstracts without adequate information to determine inclusion or exclusion, we retrieved the full text and then made the determination. All results were tracked in an EndNote® database (Thomson Reuters, New York, NY). Each full-text article was independently reviewed by two trained members of the research team for inclusion or exclusion on the basis of the eligibility criteria. If the reviewers disagreed, conflicts were resolved by discussion and consensus or by consulting another member of the review team. Results of the full-text review were also tracked in the EndNote® database, including the reason for exclusion for excluded full-text publications when they did not meet the eligibility criteria.

Table 1. Inclusion and exclusion criteria for Guiding Question 3 on the efficacy and safety of MAT for OUD.

Table 1

Inclusion and exclusion criteria for Guiding Question 3 on the efficacy and safety of MAT for OUD.

For Guiding Question 3, we summarized information from systematic reviews and primary studies that met inclusion criteria in summary tables. For systematic reviews, we summarized information on year of publication, the purpose of the review, search dates and databases searched, the number of studies included, populations and settings in the trials, MAT intervention characteristics, the type of studies included, how quality was rated for included studies, methods of synthesis, the total number of patients included, main findings (including harms), and limitations (including whether the studies were primarily performed in an OTP or addiction specialty settings, whether the studies were conducted outside the United States, and other limitations). For randomized controlled trials, we summarized information on year of publication, comparisons evaluated, duration of followup, sample size, population characteristics, MAT model of care components, setting (including provider type and staffing if that information was provided), outcomes evaluated, and main findings.

For Guiding Question 1, we summarized data sources for the various MAT models of care, including published sources (with citations), unpublished sources (with URL information), and Key Informant input.


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.0M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...