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.

Chan B, Gean E, Arkhipova-Jenkins I, et al. Retention Strategies for Medications for Addiction Treatment in Adults With Opioid Use Disorder: A Rapid Evidence Review [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2020 Jul.

Cover of Retention Strategies for Medications for Addiction Treatment in Adults With Opioid Use Disorder: A Rapid Evidence Review

Retention Strategies for Medications for Addiction Treatment in Adults With Opioid Use Disorder: A Rapid Evidence Review [Internet].

Show details


Our search retrieved 1,580 unique titles and abstracts from which we reviewed 258 full–text articles for eligibility and included 2 SRs and 39 unique primary studies (Figure 2). No included studies were identified in the gap search (Figure 2). MAT retention duration ranged from 3 to 24 months. See Appendix C for quality ratings of all included studies.

This figure depicts the literature search results organized by intervention category. The initial systematic review (SR) search yielded 168 results; with 152 results from Medline, 12 from the Cochrane Database, and 4 from other sources. The initial primary study search yielded 757 results; with 750 from Medline, 6 from other sources, and 1 from a topic expert. The initial literature search for financial support yielded 709 records from Medline. After the initial search, 54 duplicates were removed, leaving 1,580 studies that were reviewed at the title and abstract level. From there, 1,322 studies were excluded for not meeting the PICOTS criteria. There were 258 full text-articles assessed for relevance. Of those, 217 full-text articles were excluded for the following reasons: 8 excluded because the population not of interest; 99 excluded because intervention not of interest; 6 excluded because comparator not of interest; 12 excluded because outcome not of interest; 1 excluded because duration of MAT was less than 3 months; 71 excluded because study design not of interest; 1 excluded because country not of interest; 1 excluded because study date not of interest; 3 excluded due to a lack of quality assessment; 2 excluded because they did not meet SR criteria; 7 excluded because they were in an included SR; 6 excluded for other reasons. In total, there were 41 included reports, with 2 SRs and 39 primary studies. In the care settings intervention, there was 1 included SR and 11 primary studies, 1 of which is also included in contingency management. In contingency management, there was 1 SR that was also included in psychosocial support and 8 primary studies, one of which was also included in care settings. For health IT, there were seven primary studies. In extended-release medications, there were five primary studies. In psychosocial, there were one SR, also included in contingency management, and nine primary studies. There were no included studies for financial support.

Figure 2

Literature flow diagram.

Care Settings, Services, and Logistical Support

We defined care settings and services as interventions that provide MAT in alternative settings or integrated models, compared with TAU conditions that offered MAT alone through specialty treatment programs. We defined logistical support as interventions that changed the process of MAT initiation and maintenance as compared with TAU, or interventions that provided MAT in conjunction with logistical supports, such as housing with health and social services assistance. We used an existing framework58 to further categorize care setting interventions into the following:

  • Interventions that initiate MAT for soon-to-be-released incarcerated populations (1 SR of 21 studies,1 and 2 additional RCTs2, 3)
  • Integration of MAT with psychiatric or primary care services (PC) (three RCTs46)
  • Integration of MAT in emergency department (ED)/hospital settings (two RCTs7, 8)

We identified one good quality SR1 and 11 primary studies (one good quality;9 nine fair;38, 1012 and one poor2) that investigated care setting/logistical support interventions.27, 912 Most were downgraded for unblinded outcome assessment or high rates of crossover.

Key Question 1. Effectiveness and Comparative Effectiveness of MAT Retention Strategies

Table 2 provides a summary of findings across care settings/logistical support, and Appendix B Table 1 and Appendix C Table 1 provide details for included studies.

Table 2. Summary of findings for care settings / services / logistical support.

Table 2

Summary of findings for care settings / services / logistical support.

Care Setting: MAT for Soon-To-Be-Released Incarcerated Populations

A good quality SR1 of 21 studies (6 RCTs and 15 observational) and 2 additional RCTs2, 3 examined interventions that initiate MAT in soon-to-be-released incarcerated OUD patients. The SR found that initiating MAT in this population was associated with high levels of postrelease treatment entry and retention compared with TAU controls who did not initiate MAT prior to release (retention rates 50% [range 27–75%] vs. 5% [range 0–9%]).1 The review assessed quality using Cochrane guidelines, with the majority of studies of fair quality (nine), followed by good and inadequate (six each). Two additional RCTs conducted in the United States (one poor2 and one fair quality3) similarly reported improved retention with initiation of MAT in soon–to–be–released incarcerated populations. One was a small 6-month study (n=15) that randomized inmates to be offered the first monthly dose of XR injectable naltrexone prior to being released from prison compared with TAU and found that the intervention group had higher retention by multiple measures (mean number of injections received 2.8 (standard deviation (SD)=1.9) vs. 1.3 (SD=1.9; 22% (2/9)) received all 6 monthly injections vs. 0% (0/6); and 46% vs. 22% treatment appointments attended, p-values not reported (NR)).2 The second study was a 2x2 factorial design (n=213) that assessed the effect of prerelease buprenorphine treatment compared with an office–based buprenorphine or traditional outpatient treatment program and found that the prerelease buprenorphine group had higher mean number of days retained in treatment at 12 months (65.9 days (standard error (SE) 12.2) vs. 21.8 (SE 7.6), p=0.005).3 Together, these studies suggest that initiating MAT prior to release from incarceration improves retention.

Care Setting: Integration of MAT With Psychiatric and Primary Care Services

Three fair–quality RCTs provide conflicting evidence on the effectiveness of integrating psychiatric or PC interventions on MAT retention.46 All were small studies (n range = 94–316): two of three involved integrating psychiatric care with substance use treatment (one methadone and one buprenorphine) and were conducted in the United States, and one was a study from France that integrated methadone treatment with PC. The largest study is a U.S. study offering psychiatric services on site at a traditional methadone treatment program compared with separate nonintegrated psychiatric and substance use care. While initiation and total days of psychiatric care were improved for onsite integrated services, there was no statistically significant difference in MAT retention at 12 months (n=316, 41% vs. 41%, p=0.96).4

The second study integrating psychiatric care (n=94) was a fair–quality three-arm U.S. trial that offered buprenorphine treatment in psychiatric clinics, a manually matrixed psychosocial model with cognitive behavioral therapy (CBT) in private clinics, and a specialized opioid treatment program. This study reported a significant association between treatment site and retention at 20 weeks (p=0.05), with 33.3 percent retained in the psychiatric setting, 51.5 percent retained in the psychosocial model, and 21.4 percent retained in the specialized outpatient treatment (retention).6 The one study in which treatment was integrated into PC was a noninferiority trial conducted across 10 sites in France (n=195) that compared methadone treatment integrated with PC with specialized outpatient methadone treatment programs and found no statistically significant differences at 12 months (88% retention in PC vs. 69% in urgent care, p=0.13).5

Care Setting: MAT in ED/Hospital Settings

We identified two RCTs of fair quality that examined retention after introducing MAT in ED or hospital settings.7, 8 One study enrolled patients at a U.S. safety-net hospital (n=139) to initiation of buprenorphine with linkage to outpatient treatment within 7 days of discharge compared with TAU (medically supervised withdrawal and community referral to treatment). Patients in the hospital-initiated group reported a 2.4 times higher rate of buprenorphine or methadone use over 6 months (incidence rate ratio (IRR), 2.44, 95% confidence interval [CI], 1.99 to 3.36). At 6–month followup, 12 (16.7%) patients randomized to linkage compared with two (3.0%) to TAU had continued on MAT (p=0.007).7 The second study was a followup study conducted in the United States of a 3–arm RCT (n=290) of adults with OUD randomized to either 1) ED–initiated buprenorphine with linkage to PC within 72 hours, 2) referral to treatment (TAU), or 3) brief intervention in ED, and assessed retention at 6 and 12 months, defined by self–reported engagement in formal addiction treatment, and found no differences in retention at 6 (53% vs. 60% vs. 51%, p=0.546) or 12 months (49% vs. 49% vs. 63%, p=0.136).8

Logistical Support

As a whole, studies of logistical interventions enrolled patients with higher addiction severity and did not show improved MAT retention compared with a standard treatment setting. We identified one good- and three fair–quality RCTs (total n=709) that tested changes in MAT prescribing procedures, such as lessened MAT participation requirements (“low-threshold”) or expedited initiation onto MAT treatment (“Script in a Day”), or the provision of housing with social/medical supports, compared with TAU.912 A U.S. trial of a low-threshold, patient–centered methadone intervention (optional counseling, modified clinic rules, no discharge for administrative violations) compared with TAU found no differences in retention at 12 months (n=300, 48.6% vs. 46.3%, p=NR).9 Similarly, another trial that involved low-threshold initiation intervention that only required once monthly counseling found no statistically significant differences in retention at 90 (n=212, 35% vs. 31%, p=not significant (NS)) and 180 days (37% vs. 29%, p=NS).11 An innovative intervention offering immediate access to methadone from a U.K. syringe exchange followed by transfer to office–based methadone compared with TAU similarly found no differences in retention at 3 months (n=100, 51% vs. 47%, p=NR). Finally, a trial of a Housing First intervention in which participants in Canada experiencing homelessness were assigned to one of three housing with health/social services groups compared with TAU (referral to housing assistance programs and outpatient specialty treatment), assessed retention using medication possession ratio (MPR) over 2 years and found no differences between groups (n=97, MPR 0.52 vs. 0.57, p=0.60).12 (See Appendix B Table 1.)

Key Question 2. Harms of MAT Retention Strategies

Most studies of MAT retention strategies did not evaluate possible harms. Only 4 of the 10 trials of care settings/services/logistical support reported serious harms or adverse events—of these, two did not specify in which arm the events occurred. A trial of a patient-centered methadone intervention reported 67 non–study–related hospitalizations and two non–study–related deaths, one from methadone overdose, out of 149 intervention participants. This compares with 59 hospitalizations, 4 non–study–related deaths, and 2 overdoses out of 151 TAU participants.9 In the study of prerelease MAT with XR-naltrexone, six patients in the prerelease MAT arm reported adverse events (two serious, not specified) compared with two (none serious) in the control arm.2

Key Question 3. Participant Characteristics Associated With MAT Retention

The majority of studies did not examine particular population characteristics to explain variation in responses to care settings/services/logistical support. As noted above, patients with OUD in criminal justice populations may benefit from prerelease initiation of MAT. The study of the prerelease MAT with office-based buprenorphine in a PC setting found no gender differences in retention outcomes.3

Contingency Management

Table 3 provides a summary of findings across CM interventions, and Appendix B Table 2 and Appendix C Table 2 provide details for included studies.

Table 3. Summary of findings for contingency management.

Table 3

Summary of findings for contingency management.

Contingency management interventions involve providing a reward contingent upon the achievement of specified criteria. While CM is sometimes used in conjunction with other psychosocial interventions, we analyzed it separately, similar to the approach used in a prior SR on psychosocial interventions16.

One good–quality SR,16 and seven RCTs (one with a followup study)11, 1315, 1719 of CM interventions assessed retention outcomes among nonpregnant adults who received MAT for OUD.59 Contingency management improved retention on antagonist MAT, but not for opioid agonist MAT (Table 3). Appendix B Table 2 provides study details. Sample sizes per study for the seven identified RCTs ranged from 35 to 252 participants and we rated one study as good quality,18 and seven fair quality.11, 1315, 17, 19, 59 Most were downgraded due to lack of evidence of allocation concealment, lack of similarity of groups at baseline, and lack of blinding.

Key Question 1. Effectiveness and Comparative Effectiveness of MAT Retention Strategies

A 2011 SR of 35 studies (n=4319) that included heterogeneous CM interventions that rewarded opioid abstinence in individuals receiving agonist/partial agonist MAT (methadone, buprenorphine, or levo-alpha acetyl methadol [LAAM]) found no statistically significant difference in retention compared with agonist/partial agonist therapy (methadone, buprenorphine, or LAAM) alone (14 trials, n=1616, risk ratio (RR) 1.02[0.96,1.08]).16 The evidence included in the SR was assessed with Grading of Recommendations Assessment, Development and Evaluation (GRADE) as being high strength of evidence.

Four of the seven additional trials (of fair to good quality) that we identified also used agonist/partial agonist MAT (methadone or buprenorphine) and confirmed findings of the SR.11, 1719 The remainder, three fair–quality studies, reported improved retention in MAT.1315 These three studies were conducted at the same site, but with separate populations. They used antagonist therapy (naltrexone), had similar criteria for accessing the reward (participants were required to take naltrexone that was provided free of charge and in close proximity to access to the reward), and provided a similar type of reward (access to a therapeutic workplace where participants earned vouchers to exchange for preferred goods and services). The studies differed in the type of naltrexone formulation administered (two using monthly XR-naltrexone and one using daily naltrexone). Regardless of formulation, retention outcomes were greater in contingency participants compared with participants who were permitted access to the workplace without contingency: injectable, XR-naltrexone (two studies, n=35, 66% vs. 35% received all scheduled injections, p=0.02613; n=38, 74% vs. 26% received all scheduled injections, p=0.004);14 and daily naltrexone (n=67, 54% vs. 16% retention at end of study, p<0.01).15 In an additional followup study, the group differences in retention were no longer observed 6 months after the intervention ended (n=67, p=0.66).59

Key Question 2. Harms of MAT Retention Strategies

The SR16 sought data on adverse events but did not report any, and only one of the eight RCTs reported harms. The 1 RCT that reported harms used daily naltrexone and reported adverse events in 8 out of 67 participants, 6 of whom were in the contingency group.15 The adverse events included one lethal opioid overdose in a contingency group participant that occurred a month after study conclusion. Other adverse events included sexual dysfunction, abdominal problems, headache, sleep problems, opioid withdrawal, nausea, chills, rapid heart rate, and shakiness.

Key Question 3. Participant Characteristics Associated With MAT Retention

The identified studies did not assess differential effectiveness of CM on retention by defined participant characteristics (e.g., gender/sex, socioeconomic status, etc.).

Health Information Technology

We adopted the Office of the National Coordinator for Health IT Playbook definition and categorization of health IT that defines the scope of IT to include electronic medical record interventions such as prescription drug monitoring, phone apps, e-prescribing, telehealth/teleconsult, and computerized decision aids.56 Figure 3 presents an overarching framework, adapted from the Health IT Playbook, for the ways in which health IT interventions are proposed to increase retention in MAT.

The figure displays a spectrum of health information technology (IT) interventions proposed to increase MAT retention. The spectrum spans from healthcare-based IT, with fixed times, to on-demand patient-based IT. Outside of the spectrum is ePrescribing, which helps prevent against drug misuse and diversion while improving provider workflow. The most healthcare-based end of the spectrum begins with electronic health record (EHR) system integration which makes it quick and easy to check data prior to prescribing. Next comes telehealth, which can help expand access to addiction services by providing care remotely. Some additional components of telehealth that may explain why telehealth is effective are that it involves connecting people to people, whether by connecting providers to providers, groups support, patient to providers, or peers to peers. After this comes computer-based education/support tools in the EHR, which can help support appropriate opioid prescribing and improve outcomes. On the patient-based side of the spectrum are phone apps, which can provide convenient and practical tools to help an individual. Phone apps act by connecting people, but also through a certain degree of automation of care delivery. Phone apps enable automation of education, resource locators, reminders, and wearables, monitoring, and prediction - which can be further separated into physiological and symptom monitoring or situational monitoring through global positioning system (GPS).

Figure 3

Spectrum of IT interventions proposed to increase MAT retention. Apps=applications; CBT=cognitive behavioral therapy; EHR=electronic health record; GPS=global positioning system This figure adapted from the Office of the National Coordinator for Health (more...)

Seven unique completed studies (three fair–quality RCTs, one poor–quality RCT, and three fair–quality cohort studies),2026, 60 met the inclusion criteria for health IT interventions, including an active control involving in-person MAT as a comparator, and MAT retention as an outcome. Table 4 provides a summary of findings across health IT interventions and Appendix B Table 3 provides details for included studies.

Table 4. Summary of findings for health IT.

Table 4

Summary of findings for health IT.

Key Question 1. Effectiveness and Comparative Effectiveness of MAT Retention Strategies

Four RCTs (three fair quality2426 and one poor23) and three fair–quality retrospective cohort studies2022 assessed IT interventions for MAT retention. No RCT found IT intervention to be less effective than in-person approaches.

The largest RCT (five times the size of the others combined) of fair quality (n=1426) is an industry–sponsored study where patients new to buprenorphine treatment were randomized to receive in-person buprenorphine MAT (TAU) or buprenorphine MAT plus a Here-To-Help (HTH) IT intervention consisting of calls from care coaches and access to online educational materials, treatment calendars, and self-reported information from previous participants.26, 60 At 12 months, ITT analysis indicated no significant difference between groups (55.0% HTH vs. 56.1% TAU, p-value=NR). Planned post-hoc analysis revealed that people who completed a greater number of intervention calls with coaches had a greater probability of retention in MAT (adjusted logistic regression Exp (β)=1.01, p<0.001) and 64 percent of participants in the HTH program who completed at least three HTH sessions remained in buprenorphine MAT compared with 56 percent in TAU (p<0.025).26

Health IT is relatively new to healthcare and not part of the provision of MAT in the United States, so it is not known, if offered, the degree to which people would use it. While OUD treatment differs widely from the United States, a large retrospective Canadian study (n=3733) suggests that up to half of patients might choose health IT as part of their treatment if offered, without worsened retention.20

Key Question 2. Harms of MAT Retention Strategies

As a whole, harms were not well reported across health IT studies, with the majority not mentioning whether they assessed potential harms. One RCT reported that 12 adverse events occurred in a study of 82 participants (15% overall; 17% in intervention group and 12% in TAU), but no details were provided about the nature of the events.23

A small feasibility study that did not report on retention and therefore was not included in our review provides helpful detail on some challenges and harms that may be unique to IT.61 These issues included difficulties with computers or internet problems, lack of study staff training in technical support with patient connectivity, difficulties with compatibility with organizational security systems, patient privacy, and loss of clinical care time due to IT issues.

Key Question 3. Participant Characteristics Associated With MAT Retention

There is inconclusive information to determine which patients are most likely to benefit from IT and which types of IT might be most appropriate for specific patients (e.g., patient-provider communication, treatment, and/or counseling, educational support, reminders, monitoring). The large retrospective cohort study from Ontario, Canada, found a significant association between sex, clinic region (northern vs. southern), age, and peak methadone dose but not for clinic rurality.20 Patients who chose treatment delivery via predominantly telemedicine (greater than 75% of appointments) came from both urban (77%) and rural (23%) populations. Findings from a secondary analysis of an RCT that replaced the second half-hour of every in-person counseling session with an IT Therapeutic Educational System found that including a Therapeutic Educational System as part of care was better for patients who were employed, highly anxious, ambivalent about opioid abstinence, and had crack cocaine use in the past 30 days.62

Commercially Available Mobile Apps

Six apps were identified via the Google Play and Apple store (Opioid Addiction Recovery Support, Pear Re-SETO, Thrivee, A-CHESS, COR-12, FlexDek: MATList). There were no published studies about these apps with regard to OUD treatment. All apps have some educational component. Five of six apps contain some feature that allow patients to connect with peers, either through interaction in forums, virtual group therapy, or direct contact, and allow users to track their progress through treatment. At least two apps (A-CHESS and FlexDek MAT) provide patients with resources to identify local Narcotics Anonymous meetings, either through direct Global Positioning System (GPS) locating or by acting as a repository for schedules. We only identified evidence for the A-CHESS app, which was initially developed for alcohol use disorder, but an ongoing RCT is evaluating A-CHESS for OUD.

Extended-Release Medication–Based Treatment for OUD

Extended-release is a long-acting form of MAT delivered as either injectable or implant–based formulations. The only three XR formulations currently approved by the U.S. FDA for the treatment of OUD are a 1-month buprenorphine injection, a 6-month buprenorphine implant, and a 1-month naltrexone injection.

We identified a total of five fair-to-good–quality RCTs2731 comparing XR formulations (naltrexone injection, buprenorphine injection, and buprenorphine implant) head-to-head against daily MAT formulations (naltrexone, buprenorphine/naloxone, and buprenorphine). Most were downgraded for unblinded treatment or outcome assessment.

Table 5a provides a summary of treatment retention findings for XR and daily formulations for the same drug, while Table 5b summarizes retention findings for XR and daily formulations for drugs across different agonist/antagonist categories and formulations (see Appendix B Table 4 and Appendix C Table 4 for details).

Table 5a. Summary for extended-release versus daily MAT formulations within the same agonist/antagonist drug categories.

Table 5a

Summary for extended-release versus daily MAT formulations within the same agonist/antagonist drug categories.

Table 5b. Summary for extended-release versus daily MAT formulations across different agonist/antagonist categories.

Table 5b

Summary for extended-release versus daily MAT formulations across different agonist/antagonist categories.

Key Question 1. Effectiveness and Comparative Effectiveness of MAT Retention Strategies

One fair–quality study (n=60)27 compared XR-naltrexone monthly injection with daily naltrexone. Retention was defined as documented clinical contact at 6 months and favored the XR-naltrexone group (57.1% vs. 28.1%, hazard ratio (HR)=2.18, 95% CI=1.07, 4.43).

Two studies compared XR and daily buprenorphine formulations, though neither assessed retention as a primary outcome.28, 29 One was a good–quality multisite trial (n=177) that randomized participants who were stable on a low dose of daily buprenorphine prior to enrollment to a buprenorphine 6-month implant versus daily sublingual (SL)-buprenorphine and found no difference in rates of treatment retention at 6 months (93.1% vs. 94.3%, p-value NR).29 The other was a larger fair–quality trial (n=428) of a weekly, followed by monthly, buprenorphine injection compared with daily SL-buprenorphine/naloxone that enrolled treatment–seeking participants and found no difference in retention at 24 weeks (56.8% vs. 58.1%, p-value NR).28

Two studies comparing XR-naltrexone injection versus daily SL-buprenorphine/naloxone had inconsistent results.30, 31 One was a good–quality study conducted in Norway (n=159) that recruited patients from outpatient and inpatient settings.31 Treatment retention was defined as mean days until dropout from the study medication and did not differ between the two groups (mean (SD) days 69.3 (25.9) XR-naltrexone vs. 63.7 (29.9) days SL-buprenorphine/naloxone, p–value NR) The other was a fair–quality U.S. study that recruited 570 patients from community outpatient treatment programs.30 Retention was defined as percentage of patients receiving MAT at 3 months and was lower with XR-naltrexone compared with SL-buprenorphine/naloxone using ITT analysis (33.9% vs. 40.0%, p-value NR).

Our synthesis identified study design issues that may have impacted the retention results, particularly for XR-naltrexone formulations. The study that found that XR-naltrexone injection improved retention compared with SL-buprenorphine/naloxone randomized only those participants who successfully completed medically supervised opioid withdrawal. In contrast, the other study that found lower retention with XR-naltrexone randomized patients prior to their completion of supervised withdrawal and had high rates of treatment induction failure (72% vs. 94%, p<.0001) in the XR–naltrexone compared with the SL-buprenorphine/naloxone group; this could explain the lower retention rates reported. Restrictive study inclusion criteria that exclude participants with alcohol dependence or polysubstance use28, 31 are also likely to affect generalizability of the results.

Key Question 2. Harms of MAT Retention Strategies

Studies of XR formulations reported a variety of adverse events, ranging from serious fatal and nonfatal adverse events to adverse events presumed secondary to treatment medication. All studies reported nonserious adverse events at XR-naltrexone and XR-buprenorphine injection sites.

Sullivan et al. 2019 (n=60) reported one severe adverse drug-related event when one of 28 participants developed hives after receiving XR-naltrexone injection and was removed from the study.27 Lee et al. 2018 (n=570) reported a total of 28 overdose events among 23 participants.30 While these overdose events were not categorized as drug related, 18 (64%) of the 28 events were among participants randomized to XR-naltrexone, and included eight participants who had failed treatment induction and never received an injection. Five overdose events were fatal and included two participants in the XR-naltrexone arm and three in the SL-buprenorphine/naloxone arm.

Key Question 3. Participant Characteristics Associated With MAT Retention

There is limited information to determine which patients are most likely to benefit from XR formulations. None of the studies we reviewed noted differences in participant characteristics that predicted retention for XR formulations.

Psychosocial Support

We used a prior review16 to define psychosocial support interventions, which include psychiatric care, psychotherapy, counseling, and social work services that provide psychological support ranging from structured psychotherapies such as CBT and supportive expressive therapy to behavioral interventions. We analyzed CM interventions separately above.

One good–quality SR16 and nine additional RCTs3240 (sample size range=49–653) examining psychosocial supports for MAT met inclusion criteria; an additional two studies included IT interventions and are reviewed in that section.24, 26 We rated two studies as good quality39, 40; four as fair3538, and three as poor.3234

Most were downgraded due to unblinded outcome assessment. Table 6 provides a summary of findings and Appendix B Table 5 provides study details. Quality ratings for individual studies are reported in Appendix C Table 5.

Table 6. Summary of findings for psychosocial interventions.

Table 6

Summary of findings for psychosocial interventions.

Key Question 1. Effectiveness and Comparative Effectiveness of MAT Retention Strategies

An SR with meta-analysis found that adding structured psychosocial interventions to MAT did not improve retention compared with standard MAT (27 studies, n=3124, RR 1.03, 95% CI 0.98 to 1.07).16 The SR assessed the strength of the body of evidence with GRADE and rated it as high quality. Of the nine additional RCTs, eight reported no statistically significant effect of psychosocial interventions.3340 However, many of the studies reviewed included some form of counseling in the control groups, which may explain the lack of effect. One poor–quality study reported significantly higher buprenorphine MAT retention in the intervention group, which included a Community Reinforcement Approach (CRA) component that makes use of an individual’s social networks to encourage adherence to the program compared with CM alone (n=170, 80% CRA vs. 64% CM-alone; odds ratio (OR) 2.3, 95%CI 1.15-4.60).32

Key Question 2. Harms of MAT Retention Strategies

The one study in which there was a statistically significant benefit of psychosocial intervention did not report any significant harms.32 Another study (n=300) reported significant potential harm from psychosocial intervention.34 In this study, when controlling for number of days in treatment, more exposure to counseling was significantly associated with increased frequency of heroin use, cocaine use, and criminal activity (all p<0.01). The authors suspected that this association was due to confounding by indication as clinicians tended to insist on participation in additional services for patients who were not progressing. This same study also reported higher self-reported burden on participants and found the intervention group reported increased burden compared with the control (p<0.05). Looking across studies, we found differences in time commitment between intervention and control conditions in six studies that may have affected intervention fidelity, attrition rates, and retention outcomes due to increased treatment burden.32, 34, 3740

Another study (n=542) of pharmacist-delivered motivational interviewing versus normal practice for methadone patients found that self-reported physical health was statistically poorer for the intervention group; they theorized these results were due to increased awareness of health due to increased conversations between participant and pharmacist in the treatment group.33 Other studies reported adverse events and non–overdose–related deaths that the study authors judged to be unrelated to the interventions studied.35, 38

Key Question 3. Participant Characteristics Associated With MAT Retention

Few included studies provided information about which patients, if any, are most likely to benefit from psychosocial interventions. A study (n=125) of XR-naltrexone coupled with a Behavioral Naltrexone Therapy (BNT) intervention compared with XR-naltrexone coupled with Compliance Enhancement reported a significant interaction (p=0.03) between condition and severity of OUD.36 For low-severity patients (less than six bags of heroin per day), retention was highest in the BNT group (60% at 6 months). BNT adapted elements of CRA in encouraging positive social reinforcement.

Financial Support

We defined financial support as individual and system–level interventions to lower financial barriers to MAT, ranging from financial subsidies to assist enrollment in MAT programs to expanding MAT coverage by Medicaid and private health plans. We did not identify any SRs or RCTs of financial interventions to improve retention in MAT; expanded searches of observational studies also did not yield relevant studies.


Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...