The Dissemination and Implementation of Contingency Management for Substance Use Disorders: A Systematic Review

Abstract

Contingency management is one of the most effective behavioral interventions for substance use. However, the implementation of contingency management has not been as widespread as might be expected given its efficacy. This review summarizes literature that examines the dissemination and implementation of contingency management for substance use in community (e.g., specialized substance use treatment) and clinical (e.g., primary care) settings. A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) statement. Databases including Google Scholar, World of Knowledge, PsycINFO and PubMed were searched. Search results yielded 100 articles and after the screening of titles and abstracts 44 were identified. Full-text articles were examined for eligibility and yielded 24 articles that were included in this review. Of the 24 articles included in the review, the majority (n=11) focused on implementing contingency management in methadone clinics and opioid treatment programs. Training methods, implementation strategies, fidelity assessments, and attitudes towards the implementation of contingency management are discussed in greater detail. These findings highlight the importance of organizational input and ongoing supervision and consultation and that there is a need for additional research that is guided by theoretical frameworks and use rigorous study designs.

INTRODUCTION

Approximately 20 million individuals in the U.S. were diagnosed with substance use disorders (SUD) in 2017 (Substance Abuse and Mental Health Services Administration, 2018). In the U.S., SUDs account for roughly $200 billion annually on the health care system and an additional $500 billion on the criminal justice system and social services combined (Florence, Zhou, Luo, & Xu, 2016; Lushniak, Samet, Pechacek, Norman, & Taylor, 2014). Individuals diagnosed with a SUD are more likely to experience adverse health events (e.g., cardiovascular disease, overdose), mental health problems, homelessness, and incarceration (Galea & Vlahov, 2002). Although SUDs have a negative impact on quality of life and are a financial burden on the economy, only 15% of adults with SUDs receive treatment and the vast majority of addiction treatment is not evidence-based (Epstein, Hourani, & Heller, 2004; Watkins, Burnam, Kung, & Paddock, 2001).

Contingency management (CM) is a behavioral treatment that provides reinforcers (e.g., vouchers, prizes) for a targeted behavior such as abstinence from substance use (Petry et al., 1998). Nancy Petry (2000), deployed one of the most well-known and widely used CM procedures, “the fishbowl method,” which involves delivering variable magnitudes of reinforcement (Petry, Martin, Cooney, & Kranzler, 2000; Petry & Martin, 2002). Using this method, clients who provide a negative drug test result draw a token from a ‘fishbowl’ for a chance of winning a prize equal to the value of the token, that generally ranges from “good job,” “small,” “large,” or “jumbo” (i.e., $0, $1, $20, $100). The ‘fishbowl method’ popularized by Dr. Petry, as well as other CM procedures such as the voucher-based reinforcement therapy developed by Higgins and colleagues (Higgins et al., 1991; Higgins et al., 1993), are supported by a large body of literature that has demonstrated the effectiveness of CM in reducing tobacco, alcohol, and other drug use. (Kaminer, Burleson, Burke, & Litt, 2014; Kirby, Marlowe, Festinger, Lamb, & Platt, 1998; McDonell et al., 2013; McDonell et al., 2017; Petry et al., 2000; Petry & Martin, 2002; Roll, Higgins, & Badger, 1996; Roll et al., 2006). While CM is one of the most effective treatments for SUDs, only 10–25% of clinicians utilize CM in clinical practice (McGovern, Fox, Xie, & Drake, 2004).

In recent years, there has been a drive to accelerate implementation and dissemination of evidence-based practices for substance use treatment. According to the National Institute of Health (2018), dissemination is defined as the “targeted distribution of information and intervention materials to a specific public health or clinical practice audience,” and implementation is defined as “strategies to adopt and integrate evidence-based interventions into clinical and community settings.” The widespread dissemination and implementation of CM into clinical and community-based settings has been limited, contributing to the attenuation of its impact and sustainability. The paucity of CM implementation is perpetuated by 1) limited funding for CM-associated urine tests and reinforcers, 2) limited provider understanding and knowledge of CM, 3) a lack of practical implementation tools, and 4) scarcity of documented implementation strategies in a real-world setting (Kirby, Benishek, Dugosh, & Kerwin, 2006).

Several reviews highlight the practical barriers to CM, hindering the widespread implementation of CM (Petry et al., 2001; Petry, 2010; Roll, Madden, Rawson, & Petry, 2009). These review studies have highlighted various aspects that apply to the implementation (i.e., provider beliefs, adoption) of CM, however, there has yet to be a review examining strategies that have been utilized to further CM dissemination and implementation efforts. To this end, the purpose of this systematic review is to examine current literature on dissemination and implementation strategies of CM for substance use in community and clinical (e.g., primary care) settings.

METHODS

Database Search Methodology

To guide the selection process, the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) Systematic Review Checklist was used (Moher, Liberati, Tetzlaff, & Altman, 2009). A systematic literature search on published studies examining the dissemination and implementation of CM was conducted using PubMed, Web of Science, Google Scholar, and PsycInfo. Search terms were related to 1) contingency management; 2) motivational incentives; 3) dissemination; 4) implementation; and 5) SUDs. A secondary search included a hand search of reference lists and citations of potential articles included in the review. Titles and abstracts were screened by two independent researchers and full papers were read in detail and excluded according to the inclusion and exclusion criteria.

Inclusion and Exclusion Criteria

Articles that were included needed to be specific to substance use populations. Additional inclusion criteria were 1) peer-reviewed, 2) published in English, 3) published between January 2000 and December 2018, 4) clinician delivered, and 5) assessed or described implementation strategies for CM or included details of dissemination efforts. Studies were excluded if 1) delivered by research staff, 2) there were no details on the dissemination or implementation of CM, 3) were systematic reviews, and/or 4) the study’s main purpose was to examine the effectiveness of CM.

Data Abstraction

Three independent researchers extracted data and information that included sample size, settings, study methodology, and study design that are all noted in study characteristics.

Synthesis of these articles revealed major themes that were consistent among all independent researchers. These constructs were organized and summarized into four distinct constructs related to dissemination and implementation: 1) training strategies; 2) implementation fidelity generally defined as a measure of adherence or competence to the standards and principles of delivery an intervention (Bond, Evans, Salyers, Williams, & Kim, 2000); 3) implementation strategies defined as a more than one method and technique to enhance adoption (Curran, Bauer, Mittman, Pyne, & Stetler, 2012); and 4) staff and administrator perceptions toward CM after implementation. Several articles were limited to only one of these constructs, however the majority presented information that were classified into more than one construct.

RESULTS

Search Results

Figure 1 displays the literature searches that yielded a total of 107 results and an additional seven articles were identified through a hand search of references. Duplicate articles (n=14) were removed and the titles and abstracts of 100 articles were initially screened for inclusion that resulted in 44 full-text articles to be reviewed for inclusion. Of the 44 full-text articles that were assessed for eligibility, 20 were excluded because there was no dissemination or implementation data, or descriptive information related to implementation efforts was not included, was a review article, or was not a peer-reviewed article (i.e., book chapter, technical report). Thus, a total of 24 articles were included in the review.

Figure 1.

Study Selection Flow for Systematic Reviews (PRISMA)

Study Characteristics

The study characteristics of the 24 articles included in this review are described in Table 1. Overall, articles concentrated primarily on the process of the adoption and implementation of CM or focused on the fidelity and supervision of CM during implementation. Four studies implemented CM in methadone treatment centers (Andrzejewski, Kirby, Morral, & Iguchi, 2001; Kropp, Lewis, & Winhusen, 2016; Petry, Alessi, & Ledgerwood, 2012b; Roman, Abraham, Rothrauff, & Knudsen, 2010) and seven were in opioid treatment programs (Becker et al., 2016; Hartzler, Jackson, Jones, Beadnell, & Calsyn, 2014; Hartzler, 2015a; Hartzler, Beadnell, & Donovan, 2017; Helseth, Janssen, Scott, Squires, & Becker, 2018). Four studies focused their implementation within the Veterans Administration system (DePhilippis, Petry, Bonn-Miller, Rosenbach, & McKay, 2018; Hagedorn et al., 2014; Petry, DePhilippis, Rash, Drapkin, & McKay, 2014; Rash, DePhilippis, McKay, Drapkin, & Petry, 2013; Ruan, Bullock, & Reger, 2017). Three articles focused their dissemination and implementation efforts statewide. The first two articles concentrated on substance use and mental health organizations in South Carolina (Henggeler et al., 2008; Henggeler, Chapman, Rowland, Sheidow, & Cunningham, 2013; Squires, Gumbley, & Storti, 2008) and the third focused on chemical dependency treatment services in New York (Kellogg et al., 2005). One study implemented CM for adolescents with SUD in the juvenile justice system (Henggeler et al., 2006) while another study was implemented in multisystemic therapy programs for adolescent cannabis abuse (Henggeler, Sheidow, Cunningham, Donohue, & Ford, 2008). The remaining studies evaluated the implementation of CM in the context of general substance use treatment programs for adults (Petry, Alessi, & Ledgerwood, 2012a; Petry, Alessi, Ledgerwood, & Sierra, 2010).

Table 1.

Summary of CM Dissemination and Implementation Articles

Article/Author	Sample	Setting	D&I Component	Findings
Andrzejewski et al., 2001	N = 10 drug treatment counselors	Methadone treatment clinic	Training: Two-hour long group training sessions that include program manual, role-playing exercise with practice negotiating and writing a ‘good’ behavioral objective. Supervision for compliance implementing CM protocol.	Providing graphical feedback and performance-contingent incentives improved adherence to CM protocol.
Becker et al., 2016	N = 60 treatment providers M = 41.7 years ± 12.4 68% were female 66% were white	Opioid addiction treatment programs	Training Method: Addiction Technology Transfer Center’s Science to Service Laboratory (SSL)	Over the 12-month period, SSL-trained providers were more likely to adopt CM and had faster rate of CM adoption compared to TAU providers.
DePhilippis et al., 2018	N = 94 SUD treatment programs.	Department of Veterans Affairs	Training Method: CM Coaching Calls and Implementation Forms	Roughly 75% participated in at least 5 Coaching Calls for at least 12 months. High fidelity relating abstinence to prize draws and low fidelity with the availability of same-day test results. Patients attended more than 50% of CM sessions.
Hagedorn et al., 2014	N = 29 staff members and leadership	Outpatient SUD treatment clinics from Veterans Health Administration	Implementation/effectiveness Hybrid ‘type 1’ trial	60% of potential participants scheduled an intake appointment. Staff and leadership suggestions focused on education and persuasion strategies in sites where staff are skeptical of CM, and the need to assist leadership in developing strategies for financing and staffing.
Hartzler, 2015a	N = 19 clinicians M = 59.3 years ±12.7 89% were female 79% were white	CTN-affiliated private, non-profit opioid treatment program	Instrument: Helpful Responses Questionnaire for CM (HRQ-CM)	The HRQ-CM instrument demonstrated strong scoring reliability, internal consistency, concurrent and predictive validity, test-retest reliability and sensitivity to training effects.
Hartzler, 2015b	N = 5 managerial staff	CTN-affiliated private, non-profit opiate treatment program	Managerial staff opinions and experiences after implementation	Collaborative intervention design allows community partners to participate in shaping the design and how CM will be delivered.
Hartzler et al., 2014	N = 19 counseling staff M = 59.3 years ±12.7 89% were female 79% were white	CTN-affiliated private, non-profit opioid treatment program	Implementation/effectiveness Hybrid ‘type 3’ trial - Training: Four weekly ½ day sessions (16-hours) onsite. Included trainer demonstrations (fishbowl method) and dyadic trainee role-play. Materials included handouts and recording of training sessions. Prior to each session trainers meet with managerial staff for 30 minutes.	The training increased knowledge on CM and improved skills to deliver CM and CM adoption readiness. Adoption of CM increased the attendance of clients.
Hartzler et al., 2017	N = 19 direct care clinicians M = 59.3 years ±12.7 89% were female 79% were white	CTN-affiliated private, non-profit opioid treatment program	Post-training implementation domains	Clinicians’ CM skillfulness robust predictor of client outcomes.
Helseth et al., 2018	N = 60 treatment providers M = 41.7 years ±12.4 68% were female 66% were white	Opioid addiction treatment programs	Training conditions: Addiction Technology Transfer Center’s Science to Service Laboratory (SSL) Consolidated Framework for Implementation Research (CFIR)	SSL-trained providers adopted CM quicker and applied CM to their clients more frequently than TAU-trained providers.
Henggeler et al., 2006	N = 16 therapists N = 161 families	Juvenile justice system	Drug court/multisystemic therapy/CM	CM adherence was highest for the condition that incorporated CM protocol into the intervention
Henggeler et al., 2008	N = 30 therapists M = 33 years ±8 83% were female 57% were white	Multisystemic therapy programs	Training conditions: Workshop Only vs. Intensive Quality Assurance included ongoing quality assurance	Intensive Quality Assurance increased the use of CM techniques. However, it did not increase the use of CM monitoring techniques among IQA-trained therapists compared to workshop-only therapists.
Henggeler et al., 2013	N = 161 therapists M = 37 years 82% were female 51% were white	Department of Alcohol and Other Drug Abuse Services and the Department of Mental Health organizations	Training conditions: Workshop and resources (WS+), WS+ and computer assisted training (WS+/CAT), or WS+/CAT and supervisory support (WS+/CAT/SS).	All training conditions increased CM use, knowledge, and implementation adherence among providers.
Henggeler, Sheidow et al., 2008	N = 430 practitioners 81% were female 56% were white	Department of Alcohol and Other Drug Abuse Services and the Department of Mental Health organizations	Training conditions: Workshop Only (WSO) vs. Intensive Quality Assurance (IQA) included ongoing quality assurance	IQA-trained practitioners were likely to implement CM cognitive-behavioral techniques compared to WSO-trained practitioners. However, no difference in the use of CM-monitoring techniques between IQA-trained and WSO-trained practitioners.
Kellogg et al., 2005	-	Chemical Dependency Treatment Services of the New York City Health and Hospitals Corporation	Tool: Promoting Awareness of Motivational Incentives (PAMI). Local clinic presentations that focus cover the theory, practice, and research findings on contingency management interventions in substance-abuse settings, as well as patient experiences from other CM projects within the CTN.	Using CM increased patient motivation for recovery, supported therapy and goal attainment, improved staff morale and rapport between staff and clients.
Kropp, Lewis, Winhusen, 2016	N= 544 clients 55% were female 94% were white	Methadone treatment clinic	Tool: Promoting Awareness of Motivational Incentives (PAMI)	CM intervention developed and implemented by the staff of a methadone program, without research support. Analysis of group attendance data suggests that CM did not increase attendance of clinician-led groups but did increase in-clinic MA group attendance, an increase which was sustained at least 3 months after initial implementation period.
Neale, Tompkins, Strang, 2016	N = 10 clinic staff; 20 patients	Injectable opioid treatment (IOT) clinic	View and experiences of delivering and the effectiveness of CM.	The evaluation suggests contextual factors to improve CM implementation such as stakeholder input, consistent eligibility criteria and rules, stable treatment settings, and trust.
Petry et al., 2010	N = 35 therapists N = 181 cocaine-using patients M = 46.2 years ± 12.4 48.6% were male 71.4% were white	Substance use treatment programs (methadone and outpatient psychosocial programs)	Instrument: CM Competence Scale (CMCS) to help assure appropriate CM implementation.	CMCS is a reliable and valid measure of administering CM. Adherence to items on the general subscale of the CMCS was significantly associated with patients’ treatment outcomes.
Petry, Alessi, Ledgerwood, 2012a	N = 15 clinicians M = 45.6 years ± 11.5 40 % were male 53.3% were white	Outpatient substance use treatment programs	Training: Two ½ day training that include didactics, demonstration of urine sample monitoring procedures, role plays. ≥ 16 correct on CM knowledge test, participate in least three mock CM sessions and achieve average score of 4 on CMCS (training phase), and deliver CM to pilot clients and achieve 4 on CMHS in three sessions and 10 on CM adherence checklists (supervision phase).	Therapists improved knowledge about CM throughout training and training was rated positively. High levels of fidelity in competence in the delivery of CM and adherence to CM protocol. During supervision phase competence and increased over time.
Petry, Alessi, Ledgerwood, 2012b	N = 23 clinicians M = 45.7 years ± 12.4 60.9% were male 84% were white	Methadone treatment clinics	Training: Two ½ day training that include didactics, demonstration of urine sample monitoring procedures, role plays. ≥ 16 correct on CM knowledge test, participate in least three mock CM sessions and achieve average score of 4 on CMCS (training phase), and deliver CM to pilot clients and achieve 4 on CMHS in three sessions and 10 on CM adherence checklists (supervision phase).	Majority (96%) of therapists passed knowledge tests. Adherence to CM protocol with tracking samples and number of prize draws was high. Client outcomes improved during the CM implementation phase
Petry et al., 2014	N = 187 substance abuse providers	Veterans Health Administration	Workshop: 1 and ½ day training to: (1) ensure understanding of the basic principles underlying CM treatments, (2) provide evidence of the efficacy of CM, (3) assist participants in adapting a CM protocol consistent with behavioral principles in their setting, and (4) practice CM delivery.	77 clinics from 75 of the 108 CM funded stations implemented CM between August 2011 and May 2014.
Rash et al., 2013	N = 159 clinical leaders 50% were male 84% were white	Veterans Health Administration	Workshop: 1 and ½ day training consisted of three general areas: 1) background and efficacy of CM, 2) designing a CM program, and 3) implementation issues.	Training workshops significantly increased knowledge on CM, which helped correct clinicians’ common perceptions of CM.
Roman et al., 2010	N = 241 program administrators	Methadone treatment clinics	Adoption and implementation of interventions (MI/CM and buprenorphine) were assessed at 48-month follow-up.	Relationship scores between Wave 1 and Wave 2 did not increase significantly. There was no significant change in the rate of CM adoption, from 35.6% in Wave 1 and Wave 2 to 34.0% in Wave 3.
Ruan, Bullock, Reger, 2017	-	Veterans Health Administration	Local challenges, solutions, and lessons learned during implementation	This paper presents commentary on efforts to implement and disseminate CM at a VA facility. No results included.
Squires, Gumbley, Storti, 2008	N = 241 program administrators	54 community-based organizations	Training Method: Addiction Technology Transfer Center’s Science to Service Laboratory (SSL)	52% of agencies completed SSL components and 96% of those agencies adopted and implemented CM. Both agencies that completed and dropped out of the SLL rated it favorably. Dropout agencies were more likely than completer agencies to report turnover in staff positions that were critical to training.

All 24 articles ranged in methodology. Three were qualitative studies (Hartzler, 2015b; Kellogg et al., 2005; Neale, Tompkins, & Strang, 2016), five were mixed-methods studies (Becker et al., 2016; DePhilippis et al., 2018; Hagedorn et al., 2014; Hartzler et al., 2014; Roman et al., 2010), two were commentaries related to implementation processes (Petry et al., 2014; Ruan et al., 2017), and the remaining and overwhelming majority (n=14) of studies were quantitative studies (Andrzejewski et al., 2001; Hartzler et al., 2015a; Hartzler et al., 2017; Helseth et al., 2018; Henggeler et al., 2006; Henggeler et al., 2008; Henggeler et al., 2008; Henggeler et al., 2013; Kropp et al., 2016; Petry et al., 2012a; Petry et al., 2012b; Petry et al., 2010; Rash et al., 2013; Squires et al., 2008). Of the 24 articles, only two used implementation/effectiveness hybrid designs. The first study used a hybrid ‘type 1’ trial guided by the Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) framework and Promoting Action on Research Implementation in Health Services (PARIHS) framework (Hagedorn et al., 2014). The RE-AIM and PARIHS are both implementation frameworks used to guide the uptake of an intervention and increase sustainability. The second study used a hybrid ‘type 3’ trial guided by Proctor and colleagues’ (2011) conceptual framework (Hartzler et al., 2014).

Training Strategies

A fundamental step to successful implementation is the dissemination of information which is accomplished through provider training prior to CM implementation. Seventeen of the 24 (71%) studies specifically described the training methods used (Andrzejewski et al., 2001; Becker, Stein, Curry, & Hersh, 2012; Hartzler et al., 2014; Hartzler, 2015b; Hartzler et al., 2017; Helseth et al., 2018; Henggeler et al., 2006; Henggeler et al., 2008; Henggeler et al., 2008; Henggeler et al., 2013; Kellogg et al., 2005; Kropp et al., 2016; Petry et al., 2012a; Petry et al., 2012b; Petry et al., 2010; Rash et al., 2013; Squires et al., 2008). All of the studies utilized in-person didactic individual or group training methods, led by individuals with expertise in the delivery of CM. Overall, trainings covered content on the basic principles and foundations of CM, including the presentation of previous and current research, CM procedures (i.e., prize draws), fidelity, and common implementation issues. Of these studies, four prominent training methods were identified.

Low Intensive Training:

The first training method was developed and structured to be a less intensive training strategy that consisted of four half day sessions (one session per week) delivered in-person and onsite (Hartzler et al., 2014; Hartzler, 2015b; Hartzler et al., 2017). Sessions included live trainer demonstrations of the CM procedures, role play activities, performance-based feedback, and small group discussions. All materials and audio-recordings of training sessions were provided as a continuous resource for administrators and providers. The second training method was implemented and specific to the Veterans Administration and were one-and-a-half-day training workshops (Rash et al., 2013). Similar to the first training, these workshops utilized role play, small group discussions, live demonstrations and performance feedback. Because of the large number of substance use programs within the Veterans Administration, CM procedure design (i.e., selecting target behavior, choosing reinforcers, behavioral parameters) was emphasized in addition to the standard training topics. This training method was rated favorably and valued as an important aspect to CM implementation among providers. Further, CM knowledge and provider preparedness significantly increased.

High Fidelity Training:

As opposed to the previously described training methods that were based on participation and attendance, the third training method developed by Petry (2012a; 2012b; 2010) offered a more intensive performance-driven training, marked by three stages. The first stage involved participation in training sessions that included a majority of the methods noted already and a passing score (≥ 16 correct) on a 20-item CM knowledge test. In the second stage, providers participated in at least three mock CM sessions to demonstrate competency in three main skills: 1) informing new clients of CM; 2) increasing reinforcement for negative tests; and 3) withholding reinforcement for positive tests. To assess competency, the 12-item CM Competence Scale (CMCS) was used and providers had to achieve an acceptable rating on CMCS items. After this stage, providers progressed to the final supervision phase which included the audio recording of provider CM sessions with 2–3 of their clients for up to 12 weeks. Each audio recording was rated, and providers were given weekly feedback on performance. Similar to the competence rating during mock sessions, providers had to achieve acceptable scores on audio recordings and adherence checklists prior to utilizing CM with their entire caseload. This training method was rated favorably among community-based SUD treatment and methadone treatment providers, while also demonstrating the importance of fidelity monitoring.

Computer-Assisted Training:

Among the two studies that examined the use of technology to assist implementation, Henggeler and colleagues (2013) randomized organizations to three CM training methods and Kropp and colleagues (2016) utilized the Motivational Incentives package developed by the National Institute on Drug Abuse (NIDA) and the Substance Abuse and Mental Health Services Administration (SAMHSA) Blending Initiative. Henggeler and colleagues (2013) was the only study to incorporate a computer assisted training (CAT) that included training modules, troubleshooting tips, suggestions for engagement, and additional scripts and video examples component as an ongoing support for providers. The first method was a one-day CM workshop with resources focused on practicing the delivery of CM (i.e., CM protocol, worksheets, scripts, drug test cups, $100 in vouchers). The second method was the CM workshop plus CAT. The third method was the CM workshop and CAT plus supervisory support, designed to train supervisors and provide performance feedback of audio-recordings sessions. Although all training methods improved CM knowledge, providers who received CAT showed significantly greater increases in CM knowledge compared to providers who received the workshop only training methods.

One study leveraged the Promoting Awareness of Motivational Incentives (PAMI), an introductory training product of the NIDA/SAMHSA Motivational Incentives package. The NIDA/SAMHSA Blending Initiative consists of two other products in addition to PAMI: 1) Motivational Incentives: Positive Reinforcers to Enhance Successful Treatment Outcomes (interactive online course), and 2) Motivational Incentives Implementation Software (implementation support on CD-ROM) (Stitzer, Petry, & Peirce, 2010). Providers in this study received a one-hour training from a PAMI trained staff member (Kropp et al., 2016). However, no descriptive details were noted about the training method used. Although not specifically mentioned, Kellogg and colleagues (2005) were the first to utilize the Motivational Incentives package, however no formal evaluation was conducted on this tool or training components.

Implementation Fidelity

While the majority of implementation studies evaluated effectiveness of CM training for the purposes of adoption and implementation, seven studies (29%) additionally ventured into both fidelity to intervention (Andrzejewski et al., 2001; DePhilippis et al., 2018) and psychometric properties of CM delivery scales (Hartzler, 2015a; Petry et al., 2010). Fidelity to CM procedures have been directly tied to the successful implementation of CM and provider satisfaction. Specific monitoring and adherence instruments have been designed or adapted for CM. As an example, Henggeler and colleagues (2008; 2008; 2013) used the CM Therapist Adherence Measure (CM-TAM) to monitor the use of CM by providers and Petry and colleagues (2012b) developed the CM adherence checklist to rate adherence to CM procedures. The Helpful Responses Questionnaire for CM (HRM-CM) was developed to assess communicative aspects needed to delivery CM (Hartzler, 2015a). Petry and colleagues (2010) developed the CMCS as a valid and reliable tool to assess the delivery of CM. The CMCS is a 12-item scale, rated from 1 (poor) to 7 (excellent) Likert scale. To date, the CMCS is the only instrument available to evaluate the level of competence among providers for the delivery of CM. One study found that provider competence in the delivery of CM, assessed using the CMCS, was associated with improved client outcomes (Petry, 2010). Monitoring fidelity using the CM-TAM and CMCS requires the resources to audio-record and rate provider sessions, which can be a limitation. Two articles offer an alternative to the CM-TAM and CMCS for organizations with limited resources through use of the Implementation Form (DePhilippis et al., 2018). The Implementation Form can be used to monitor CM after initial implementation and collects information on setting, structure, outcomes, and fidelity. DePhilippis and colleagues (2018) found high fidelity with regard to providers’ ability to relate abstinence to prize draws and low fidelity with regard to same-day results specifically for organizations who outsourced laboratory testing.

One study assessed whether graphical feedback on performance would increase provider compliance with implementing a CM protocol for substance use (Andrzejewski et al., 2001). This study demonstrated a 71% increase in performance ratings among providers who received graphical feedback from supervisors and an 81% increase among providers who received an additional reward for positive performance ratings.

Implementation Approaches

Nine of the 24 studies (37%) articulated approaches to implementation that used sequential steps to creating climates both suitable for CM implementation and the subsequent adoption of CM (Becker et al., 2016; DePhilippis et al., 2018; Hartzler et al., 2014; Hartzler, 2015b; Hartzler et al., 2017; Helseth et al., 2018; Petry et al., 2014; Ruan et al., 2017; Squires et al., 2008). The studies described the process of planning and changing organizational structure in addition to the process and effectiveness of training providers on the delivery CM.

High Intensive Strategy:

The three studies utilized the New England Addiction Technology Transfer Center (ATTC) Science to Service Laboratory (SSL) strategy to increase implementation of CM (Becker et al., 2016; Helseth et al., 2018; Squires et al., 2008). SSL consists of didactic training in CM procedures conducted by a CM expert, followed by nine months of additional support involving four major elements. First, is the assignment of a technology transfer specialist with experience in supervision and management who serves as an external technical coach to support implementation (e.g., monthly conference calls with providers and innovation champions). Second, programs identified an internal innovation champion (e.g., treatment director) to support adoption and to work closely with the technology transfer specialist. Third, in addition to CM training (prior to SSL components) all providers, including innovation champions, received training on the change process. Innovation champions received an additional four-day training focused on organizational change, adoption of evidence-based practices, and strategies for integration. Last, innovation champions and technology transfer specialists led a half day training for providers focused on evidence-based practices, the theory of change, and an overview of the adoption process. The SSL strategy demonstrated a greater likelihood of CM adoption and was rated favorably among participating programs (Squires et al., 2008). Specifically, providers began delivering CM at least four weeks earlier and the adoption rates were approximately two times higher than providers who received training as usual (CM workshop) (Becker et al., 2016; Helseth et al., 2018).

Collaborative Design Strategy:

Three studies drawn from a single parent study described an approach that utilizes a collaborative intervention design with four implementation steps (Hartzler et al., 2014; Hartzler, 2015b; Hartzler et al., 2017). Step 1 involved engaging programs and administrators (leadership staff) in order to identify a target population (e.g., adults) and target behavior (e.g., alcohol, attendance), gift cards in $5 increments to local stores, and develop a reinforcement system prior to initial implementation. After these key aspects were defined, expert feedback was provided on how to establish a reinforcement schedule (e.g., warmup phase, reinforcement escalation/reset). Once CM procedures were defined, administrators then met to anticipate and resolve potential fiscal or logical barriers of implementation. Step 2 involved four 30-minute consultative planning meetings identify onsite implementation leaders who were responsible for general oversight (e.g., purchasing reinforcers, proper documentation, case reviews). Step 3 required the identification of a start date for CM. Step 4 involved the initial implementation of CM for 90 days.

Consultation Driven Strategy:

Three studies highlighted the large-scale implementation process of CM within the Veterans Administration system beginning in 2011 (DePhilippis et al., 2018; Petry et al., 2014; Ruan et al., 2017). This approach involved training workshops followed by a pre-implementation phase consisting of planning conference calls. Calls were used to discuss clinical structure, client characteristics, CM protocol design, expert feedback, and modifications to implementation procedures. Upon initial implementation of CM, programs participated in coaching calls until the first CM client completed treatment. Roughly 75% of all programs (n=94) participated in at least five coaching calls within 12-months of implementation (DePhilippis et al., 2018).

Staff and Administrator Perceptions of CM after Implementation

Only seven of the 24 studies (29%) included in this review examined provider and administrators (e.g., directors) attitudes toward CM after initial implementation using qualitative methods (Hagedorn et al., 2014; Hartzler et al., 2014; Kellogg et al., 2005; Kropp et al., 2016; Neale et al., 2016; Ruan et al., 2017) while only one assess barriers after implementation using quantitative methods (Henggeler et al., 2008). The qualitative methodologies used in each of these studies resulted in less comparable, but often more descriptive attitudes and perceptions across studies. Review of these studies revealed that staff and administrators had generally more positive attitudes and perceptions of CM and its implementation in various settings, although none of these studies assessed differences in attitudes pre- and post CM implementation. For instance, Kellogg and colleagues (2005) noted that provider opinions toward CM were viewed more positively after client’s attendance had increased and clients showed appreciation for a more positive environment compared to other SUD programs.

Neale and colleagues (2016) identified a number of barriers in their evaluation of CM delivery at a supervised injectable opioid treatment clinic. They found that implementation of eligibility criteria, rules and processes, stable treatments settings, trusting staff-patient relationships, and stakeholder input into CM design were critical to successful CM implementation. In their descriptive assessment of client, staff, and leadership perspectives following implementation of CM at SUD clinics at two Veterans Administration medical centers, Hagedorn and colleagues (2014) found that all three groups reported favorable experiences of CM, but staff and leadership also reported barriers to implementation. These barriers included a need for secure funding and more flexible staff time. Additional barriers that providers endorsed were the low priority of substance use compared to other clinical issues and lack of time (Henggeler et al., 2008).

In contrast to the previous studies that found funding and staff time to be a barrier, several studies indicated providers and administrators were not overly concerned or burdened by the cost (i.e., cost reinforcers or staff time) of implementing CM (Hartzler et al., 2014; Hartzler, 2015b; Kropp et al., 2016). Through conversations with managerial leadership, Hartzler (2015b) found leadership perceived CM to have a relative advantage through therapeutic benefit and cost-effectiveness, to be compatible to with existing infrastructure, to have simplicity in design, and to have observable positive impacts for staff and patients. In addition, the study found that the 90-day trial enhanced the implementation of CM.

DISCUSSION

This review describes and highlights the findings of 24 articles relevant to dissemination and implementation strategies for CM and its integration within various substance use treatment programs. The most frequently examined and evaluated implementation constructs were training methods, length of training, implementation strategies, and fidelity assessments. Much of this work was conducted and influenced by Dr. Petry. Training components were fairly consistent across studies and focused on the foundations of CM, current and present research on the effectiveness of CM, CM procedures, and barriers to implementation. Taken together these findings further underscore the importance of organizational input and ongoing supervision and consultation to maximize uptake and sustained implementation of CM. Although not included in this review, several guides have been published that provide suggestions to the practical application of CM in community-based settings (Petry, 2000). This review may also inform researchers and administrators seeking strategies to implement CM.

Overall, the articles identified in this review serve as examples that exemplify the process of implementation. Common elements across the studies that increased adoption included the collaborative involvement to develop CM procedures, identifying an implementation leader, on-going supervision, and fidelity monitoring. To date there has been a considerable amount of research addressing staff and administrator attitudes towards CM (Aletraris, Shelton, Roman, 2015). Interestingly, studies included in this review seem to suggest positive views after implementation and administrators also suggested incorporating periodic booster sessions for providers to reaffirm CM procedures. However, additional barriers seem to remain after implementation related to the cost of CM for an organization and the burden on staff to incorporate another element to an existing heavy load. Along those lines, is the importance of monitoring fidelity that has been linked to increased acceptability and adoption of CM. Fidelity has also been linked to improved client outcomes, in addition to performance-based feedback. However, fidelity monitoring may not be cost-effective for smaller SUD programs seeking to implement CM. For programs lacking the capacity to audio-record and rate CM sessions, there are alternative checklists and forms that can accurately assess fidelity.

Given the emphasis of theoretical frameworks in implementation science, this review raises important questions related study designs. For instance, the studies included in this systematic review have demonstrated the limited use of implementation frameworks to guide the implementation process for CM. Such frameworks are utilized to address common barriers to implementation and bridge the research to practice gap (Nilsen, 2015). Furthermore, only two studies (1%) used hybrid implementation models, which were also the only studies that utilized implementation frameworks. Effectiveness-implementation hybrid designs present the opportunity through the use one of three designs to assess the effectiveness of CM, the delivery of CM, and the implementation in a real-world setting (Curran et al, 2012). Related to this, is the need for more controlled studies to improve the method of implementation and sustainment.

Future research and real-world implementation efforts should focus on the replication of the implementation strategies highlighted in this review, guided by implementation frameworks, with larger sample sizes and more rigorous implementation designs. Previous research has suggested that alternative study designs such as factorial designs, stepped-wedge designs, and interrupted time series may be best suited for dissemination and implementation studies (Mazzucca et al., 2018).

Although several studies have demonstrated that CM is a cost-effective evidence-based practice (Murphy et al., 2015; Olmstead, Sindelar, & Petry, 2007) and that there is an economic benefit to implementing CM, what remains one the biggest barriers to the implementation of CM is cost (Petry, 2010; Roll et al., 2009). As important as the implementation of evidence-based practices for substance use is, a concept that is equally important is the deimplementation of current practices that have no utility and do not have a significant impact on substance use behaviors or address mechanisms of change. Substance use programs with limited resources may undertake this process first to free up the necessary resources and alleviate provider burden to implement CM. Further, additional resources need to be developed for community-based substance use programs who may be interested in implementing CM. One possible suggestion that builds on the checklist for designing CM programs outlined by Petry (2000), is the development of a structured guide for administrators that outlines how to obtain a funding source, assess the cost/benefit of implementation, and build the infrastructure to financially support and sustain CM. Roll and colleagues (2009) also suggested that advocacy efforts may be needed to diminish the common misconception that CM is expensive.

One consistent area of focus between this review and previous CM reviews, is the attention given to the use of technology to improve CM through the development of apps and web-based tools (McPherson et al., 2018). Although CAT was the only technology-based tool used to provide ongoing support for providers, the Motivational Incentives package is another tool developed to aid dissemination and implementation efforts. The integration of such technologies offers a potential way to increase CM implementation. However, additional research is needed to further develop, refine, and implement technologies for clinicians to support training, supervision, and monitoring of CM in the real-world settings.

This systematic review has several strengths, but some limitations should be noted. Although, the 24 articles identified were relevant to dissemination and implementation of CM, there may be additional studies that trained clinicians to delivery and implement CM that were not included in this review. Due to the limited data and heterogeneity of articles, a meta-analytic review of studies was not feasible. Overall, this review revealed the wide scope of studies that address various constructs relevant to implementation in multiple settings. Moreover, the studies included in this review were centralized to one geographical location with the exception of one article conducted in Great Britain, which may limit generalizability.

Conclusions

There are a limited number of studies focused on the dissemination and implementation of CM. Although feasibility and effectiveness trials are essential to advancing the delivery of CM and improving reach, additional research addressing the real-world implementation of CM using rigorous study designs (i.e., hybrid models) is needed to increase uptake. Unlike previous reviews, the articles included in this review have the potential to inform to serve as examples for implementation.