Fostering MOUD use in justice populations: Assessing the comparative effectiveness of two favored implementation strategies to increase MOUD use

Abstract

Coaching is a favored strategy for the implementation of medications for opioid use disorder (MOUD), yet research has not adequately tested or assessed coaching dosages and mediums for overall effectiveness, nor have coaching doses been widely studied within criminal justice settings (CJS). Scaling up the use of MOUD, particularly in CJS, presents a challenge given the stigmatization of substance use disorder, funding for MOUD, availability and capacity of community-based treatment providers, leadership support, and the historical preference for behavioral therapy–based treatment practices. The University of Wisconsin’s Center for Health Enhancement and Systems Studies (CHESS) and George Mason University’s Center for Advancing Correctional Excellence! (ACE!) are conducting a randomized controlled trial to determine the optimal combination and dosages for two different coaching strategies to disseminate MOUD in justice-involved populations; those strategies are the NIATx model for process improvement and Extension for Community Healthcare Outcomes (ECHO) model. NIATx coaches provide technical assistance in MOUD implementation and organizational change to help justice and treatment organizations to implement and disseminate MOUD for justice clients. The ECHO platform focuses primarily on the clinical provider by connecting the provider with expert MOUD prescribers to promote high-quality MOUD practices. The trial will have four study arms that compare high-dose and low-dose coaching, with and without ECHO. This will be the first trial that assesses the comparative effectiveness of two types of coaching methods at varying dosages for justice-involved individuals. The trial will be conducted with 48 jails and community-based treatment provider sites that handle justice-involved persons with opioid use disorder (OUD).

1. Introduction

The U.S. criminal justice system (CJS) has the largest concentration of adults with behavioral health disorders. Approximately half of state and federal inmates meet criteria for a substance use disorder (Mumola & Karberg, 2004); however, less than 1 percent of the more than 5,000 U.S. prisons and jails that incarcerate more than two million inmates, allow access to FDA-approved medications to treat opioid use disorder (OUD; Vestal, 2018). While OUD is prevalent among individuals in the CJS in the United States, and two-thirds (63%) of sentenced jail inmates meet the criteria for drug dependence or abuse (Bronson, Stroop, Zimmer, & Berzofsky, 2017), few justice-involved individuals can access medication for opioid use disorders (MOUD) while in jail or prison. And, justice agencies rarely have systems in place to transition individuals with OUD to community-based, medication-based treatments. The risk of death within the first 2 weeks after release is over 12 times more than for individuals with OUD in the general population. The leading cause of death is fatal overdose (Binswanger, Stern, Deyo, et al., 2007), due to a tolerance loss during periods of opioid or other drug abstinence while in prison/jail.

While administering MOUD, jails encounter common barriers: the stigmatization of substance use disorder (Lam, Lee, Truong, et al., 2019; Chou, Korthuis, et al., 2016; Samuels, 2015; Grell, Ostile, Scott, Dennis, & Carnavale, 2020), funding for MOUD (Chou, Korthuis, et al., 2016; Ferguson, Johnston, Clarke, et al., 2019), institutional design making it difficult to administer MOUD (Grell, Ostile, Scott, Dennis, & Carnavale, 2020; Ferguson, Johnston, Clarke, et al., 2019), leadership support (Samuels, 2015; Ferguson, Johnston, Clarke, et al., 2019; Rogers, 2020), policy and administrative procedures (U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, 2019; Nunn, Zaller, Dickman, Trimbur, Nijhawan, Rich, 2009; Friedmann, Koskinson, Gordon, et al., 2012), availability and capacity of community-based treatment providers (Ferguson, Johnston, Clarke, et al., 2019; U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, 2019; Friedmann, Koskinson, Gordon, et al., 2012), and communication about MOUD’s effectiveness at achieving both public safety and personal recovery (Samuels, 2015; U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, 2019; Friedmann, Koskinson, Gordon, et al., 2012). MOUD is more than dispensing new medication, it requires major changes in internal work processes in correctional settings and partnerships with community agencies (Rogers, 2020). Jails typically do not have the medical resources to administer the medication and lack staff to screen for OUD, conduct case management, provide behavioral treatment, and engage in discharge planning. Similarly, community providers also lack resources and staff to work with jails and/or prisons on reentry-related issues.

Organizational coaching is an implementation strategy to address barriers and implement innovations (Fixsen, Blase, Naoom, & Wallace, 2009). For study purposes, we refer to coaching as providing expertise in MOUD implementation and systems change through well-structured technical assistance and mentoring to help sites adopt, implement, and increase the use of MOUD for justice-involved individuals. This study will use two different coaching strategies, as well as implement two different dosage amounts: NIATx,¹ a process improvement model and the Extension for Community Healthcare Outcomes (ECHO) model. Existing evidence supports high-dose NIATx coaching, and this type of coaching focuses on organizational change in non–justice settings (Gustafson, Sainfort, Eichler, Adams, et al., 2003). ECHO focuses on the MOUD provider’s knowledge and self-efficacy of MOUD care (The University of New Mexico, 2020) to increase confidence in the use of MOUD. The general literature finds coaching to be an effective technical assistance tool, but its effectiveness in justice settings is understudied. And studies do not delineate which type of coaching would yield improved reach and advance integration within the system. Both types are needed to increase the acceptance for the use of MOUD for justice-involved clients by agency staff. Also, the general organizational literature has not widely tested the dosing of coaching across implementation settings (Owens, Lyon, Brandt, Warner, et al., 2014).

This implementation effectiveness trial will be conducted with 48 jails and community-based substance use treatment providers around the nation. The trial will have four study arms that compare low-dose and high-dose NIATx coaching, with and without ECHO, to increase access to OUD medications. This study hypothesizes that sites assigned to the study arm that includes high-dose NIATx coaching and ECHO will be the most successful in implementing or expanding MOUD use.

2. Interventions

The study will have four arms that sites will randomly be assigned to: 1) high-dose NIATx coaching with ECHO, 2) low-dose NIATx coaching with ECHO, 3) high-dose NIATx coaching only, and 4) low-dose NIATx coaching only. High dosage refers to twelve monthly coaching calls, whereas low dosage will have four calls or one per quarter. The study will provide coaching for a 12-month period with a 12-month follow-up period to measure sustainability. Each site will have a change team that consists of up to eight staff members with whom the coach will work with. See Table 1 below.

TABLE 1.

Overview of NIATx and ECHO coaching.

ARM	NIATx Coach	ECHO
High-dose NIATx coaching and ECHO	• Four-hour, Virtual Kick-Off Meeting* split into two days with Study Team & Coaches • 12 monthly (one-hour) coaching calls with Change Leader/Team	• Prescribers participate in 12 monthly (one-hour) scheduled video conference calls
Low-dose NIATx coaching and ECHO	• Four-hour, Virtual Kick-Off Meeting* split into two days with Study Team & Coaches • Four (one-hour) coaching calls at months 1, 4, 8, and 12 with Change Leader/Team	• Prescribers participate in 12 monthly (one-hour) scheduled video conference calls
High-dose NIATx coaching only	• Four-hour, Virtual Kick-Off Meeting* split into two days with Study Team & Coaches • 12 monthly (one-hour) coaching calls with Change Leader/Team
Low-dose NIATx coaching only	• Four-hour, Virtual Kick-Off Meeting* split into two days with Study Team & Coaches • Four (one-hour) coaching calls at months 1, 4, 8, and 12 with Change Leader/Team

^*The Kick-Off Meetings were originally planned to be an in-person event but was restructured to be held virtually due to Covid-19.

NIATx coaching.

NIATx is an organizational change approach aimed at gauging performance, increasing efficiency, and informing process improvement (Gustafson, Sainfort, Eichler, Adams, et al., 2003). Each participating site will have a NIATx coach who will work with a site’s change team to review the current functional level, identify processes to change, help to implement changes, and monitor performance. Coaches will help the sites to assess their work processes and consider key issues and facilitate change team decisions regarding how to improve the use of MOUD.

ECHO (Extension for Community Healthcare Outcomes).

The ECHO model is focused on the MOUD prescribers and clinical staff in the jail and community-based treatment provider (CBTP) sites to improve their competence in using MOUD with justice-involved individuals. The emphasis is on case conferencing to address the issues that clinical staff must navigate, such as clients with OUD and other behavioral or medical needs. ECHO is provided monthly through videoconferences. Experienced MOUD prescribers address topics such as counseling strategies; patients’ complex needs; medication storage and diversion; and transitioning from buprenorphine, naltrexone, and methadone. Prescribers will participate in 12, one-hour monthly ECHO videoconference calls.

2.1. Outcome measures

The implementation effectiveness study design uses the RE-AIM framework, which will be used to measure the impact on MOUD utilization and uptake, as Table 2 shows.

TABLE 2.

Primary and secondary outcomes.

Type	Name	Aim & Time Frame	Brief Description
Primary	Reach	% of OUD patients initiated and engaged with MOUD use during the 12 months, by site	Adoption of MOUD by clients during intervention period
Secondary	Effectiveness	% of patients that are not arrested in 12 months postrelease	Re-arrest rates
Secondary	Adoption	% of Clinicians using MOUD during the 24 months	Adoption of MOUD by clinicians for their clients
Secondary	Implementation	NIATX/ECHO Fidelity Measures	Adherence to NIATx and ECHO protocols
Secondary	Maintenance	% of OUD patients initiated onto MOUD and engaged with MOUD use during the months 13–24, by site	Adoption of MOUD by clients during sustainability period

2.2. Study sites

The study will be carried out by county jail clusters that include the jail and any community-based treatment provider (CBTP) sites with which the jails collaborate. The MOUD services can be provided in the jail, community, or both. The randomization will occur by clusters, and the study will collect data by the jail and CBTP site, with the primary outcome variable being the % of patients with OUD receiving MOUD by site. The anticipated breakdown of site participants is 8 jail-only clusters and 20 jail-CBTP clusters (n=20 jail sites & n=20 CBTP sites), for a total of forty-eight (48) study sites. There will be two cohorts, with sites currently enrolled from Hawaii, Maine, Virginia, and Wisconsin.

2.2.1. Recruitment.

The primary target for recruitment efforts are jails seeking technical assistance to implement or expand MOUD practices within their site, and/or increase MOUD use postincarceration. When a jail joins the study, the study team encourages the site to invite the CBTP site(s) with which they work to join the study. Although beneficial, it is not mandatory for the jail to include a CBTP.

The research team is recruiting jails through a variety of methods, including networking and distributing promotional materials through national networks such as the Justice Community Opioid Innovation Network (JCOIN) and the Bureau of Justice Assistance (BJA), accessing state databases of jails and cold-calling, and outreach to national organizations such as the American Jail Association.

2.2.2. Change teams.

Each participating site will identify an executive sponsor, change leader, and change team. The executive sponsor will represent their respective jail or CBTP site, such as a director, CEO, sheriff, or warden. The executive sponsor will be responsible for identifying a change leader or site liaison (someone in a management role) who will coordinate study components. The executive sponsor and change leader will identify up to seven staff members to be a part of a change team. Members of the change team can hold a variety of positions, including criminal justice staff (jail or probation), health provider representatives, medical providers/prescribers (i.e., nurse, physician), counselors, and other stakeholders, to ensure that the team reaches various pertinent audiences. The change teams, with guidance and technical assistance from their assigned NIATx coach, will work on quality process improvement projects to implement or improve MOUD practices and policies within their sites.

3. Study procedures

3.1. Study phases.

The study has three phases. Phase 1 is the implementation phase, during which the executive sponsor, change leader, and change team members from each study site will participate in two study onboarding calls with the research team. Each site’s change team will engage in prestudy workgroup processes, which will include a walk-through exercise and a prioritization of issues regarding MOUD use in their jurisdiction. Phase 1 will also include a kick-off meeting, to be done virtually, covering NIATx methods, MOUD promising practices, and receive assistance from their NIATx coach in developing their project charter. During Phase 2, the coaches will meet with their teams per the assigned arm (see Table 1). The assigned coach will collect implementation data to gauge the progress of the team. If a site is assigned to two of the four study arms that include ECHO, the clinicians on the change teams will concurrently attend monthly ECHO sessions. Phase 3 will be 12 months postimplementation, and this phase will involve monitoring the use of MOUD and the goals that the group had selected. This timeframe is considered our follow-up period. During all three phases, the executive sponsor, change leader, and change team members will be asked to complete surveys.

NIATx coaches.

Six trained NIATx coaches will provide coaching and technical assistance. The coaches attended ten-hours of virtual training and workgroup sessions to ensure standardization of study protocol and coaching practices.

4. Study design and power

This study employs a randomized block design, where a 2x2 factorial design will be implemented within each block of four homogeneous sites. The blocks will be defined based on if the jail has a CBTP (as part of the jail-CBTP combination), population of the county where the jail is located, whether the jail is currently providing MOUD to justice-involved populations, and type of MOUD provided. The blocking procedure will reduce variability within intervention conditions and yield more precise estimates of treatment effects. Each jail and CBTP within a block will be treated as a distinct site as each operates by their own unique policies and procedures and has their own process improvement change team. However, they will be placed in the same intervention arm as they will work collaboratively during the trial to increase MOUD engagement and retention postincarceration. For example, if a jail has one CBTP in the trial, this will be two total sites.

The two factors, NIATx coaching (high dose vs. low dose) and ECHO (present vs. absent) will be fully crossed within each block. The number of jails and CBTPs per arm are expected to be balanced, with an anticipated seven jails and five CBTPs per arm. This strategy will allow for the study team to examine the effects of high- versus low-dose coaching and ECHO on MOUD use rates in large and small settings. There will be a 12-month implementation period followed by 12 months to track sustainability.

The study conducted a power analysis based on two previous studies that implemented NIATx coaching and ECHO to similar populations. Specifically, Molfenter et al. (under review; 2013) found an effect size of Cohen’s d= .51 for high- vs. low-dose coaching, and Komaromy M, Duhigg D, Metcalf A, et al. (2016) found large effects of ECHO, including a 10-fold increase in buprenorphine-waivered physicians. Coaching effects of d= .46 to .56 can be detected at a power range of .87 to .96 with the 48 sites in this study with a Type I error rate of 0.05, with corresponding power being greater for ECHO due to larger effect sizes.

4.1. Data sources

4.1.1. Staff surveys.

The study will collect data from each change team. The study will ask the executive sponsor, change leader, and change team at each jail and CBTP site to complete three main surveys based on their role(s) within the study: organizational survey, staff survey, and prescriber survey (See Table 3). The study will administer the surveys at baseline, 12-month, and 24-month timepoints. The surveys are adapted from the Chestnut Survey of Jails (Scott & Dennis, 2019), a working relationship scale from the National Criminal Justice Treatment Practices Survey (Taxman et al., 2007), organizational readiness (Gustafson et al., 2003), staff attitudes (Knudsen et al., 2005), Program sustainability assessment (Luke et al., 2014), stress (Gomez, et al., 2014), NIATx Fidelity (Gustafson et al., 2011), and the Physician Worklife Survey (Williams et al., 1999). The Chestnut survey instruments document the practices of the jail and health facilities regarding opioid use screening, management, and technical assistance needs. The study will send site staff participants links to complete the survey(s) via REDCap, a secure web platform for managing online databases and surveys. The instruments will be used to measure adoption of MOUD (percent of clients using the medication), implementation based on the organizational readiness for MOUD, and sustainability efforts.

TABLE 3.

Organizational and staff surveys.

Survey	Time Frame	Staff	Survey Measures	Number of Surveys
Organizational Survey	Baseline, 12-months, 24-months	Executive Sponsor	Working Relationship Scale (Taxman et al., 2007), Opioid Use Screening (Scott & Dennis, 2019), Opioid Withdrawal Management (Scott & Dennis, 2019), MAT for Pregnant Women (Scott & Dennis, 2019), Jail-Based Services (Scott & Dennis, 2019), NIATx Fidelity Scale (Gustafson et al., 2011), Staff Attitudes (Knudsen et al., 2005), MAT for Opioid Use Disorders (Scott & Dennis, 2019), Other Jail Substance Use Treatment Services (Scott & Dennis, 2019), Technical Assistance Needs Related to MAT (Scott & Dennis, 2019)	48 unique staff participants at baseline, and 4 new people per round at postintervention (m12) and post-sustainability (m24), assuming 10% turnover at each round. This will result in a total of 56 unique participants and 144 total surveys.
Staff Survey	Baseline, 12-months, 24-months	Change Leader & up to 7 Change Team Members	Organizational Readiness for Implementing Change (ORIC) (Gustafson et al., 2003), Staff Attitudes (Knudsen et al., 2005), Program Sustainability Assessment Tool (Luke et al., 2014), Stress (Gomez et al., 2014), NIATx Fidelity Scale (Gustafson et al., 2011)	Each survey cycle will include up to 384 people. The 384 staff will include up to 8 participants from the 48 sites including the Change Leader and up to 7 team members. Since many of the same staff will likely complete surveys in successive cycles, 536 unique staff are expected for survey completion across all three cycles. The 536 unique people projected assumes 20% or 76 new people per round occurs postintervention and post sustainability round. There will be a total of 1,152 surveys.
Prescriber Survey	Baseline, 12-months, 24-months	Up to two Prescribers from each Change Team	MAT Usage, Physician Satisfaction, Organizational Attitudes (Williams et al., 1999)	This will result in 96 unique prescriber participants at baseline, and 9 new people per round at postintervention (m12) and post-sustainability (m24), assuming 10% turnover at each round. This will result in a total of 114 unique prescribers and 288 surveys.

4.1.2. Site data.

Each month the study will ask jails and CBTP sites to provide data tracking reports. The data that the study will capture for the jails and CBTPs will include MOUD treatment data; the number of individuals screened or referred for OUD; the number of individuals receiving MOUD; the number of individuals who received naltrexone, buprenorphine, and methadone; and the number of injections or slots. This will be the source for measuring data. The study will capture arrest records from jail data to assess effectiveness as determined by re-arrest rates.

4.1.3. Qualitative Interviews.

The study will gather qualitative data from twelve randomly selected Jail-CBTP or Jail-only clusters, three clusters within each study arm, at two points: mid-intervention (6 months) and postintervention (12 months). Study staff will complete qualitative phone interviews individually with up to five change team members from each Jail-CBTP or Jail-only cluster (n=12), including the executive sponsor(s), change leader(s), and prescribers from the jail and CBTP, with a total of up to 60 interviews completed at baseline. Interviews will be semi-structured and gather information that is not included within the organizational surveys that will be completed throughout the study. This includes information about site structure and process, culture surrounding justice-involved persons, motivation for serving OUD populations, and approaches to organizational change.

5. Data analysis

The study will conduct initial exploratory analyses for each site examining MOUD baseline practices and changes over time. A series of summary statistics will characterize the 48 participating sites, including patient age, race, ethnicity, nature of SUD, criminal justice status, organizational readiness for change, and MOUD practices. The study will use Chi-square, t, and F tests to test for statistically significant baseline differences. Descriptive statistics and figures will display the distributions of analytic variables collected.

Although site assignment will be random, the randomization will occur within blocks, and thus the blocks should be accounted for in the data analysis. In addition, baseline characteristics of the jails and CBTPs will be collected and considered to control for additional sources of variation across the study sites and to make the evaluation of the interventions valid and precise. We will examine experimental findings for the four study conditions using general linear models.

The study will repeatedly measure rates and frequency of MOUD use (buprenorphine, methadone, and injectable naltrexone), and we will correlate these values over time. The study will use growth curve models with time-dependent errors (e.g., auto-correlated or Toeplitz-structured) to examine the patterns of change during the intervention and sustainability periods as well as to compare the effects of the interventions over time.

6. Response to COVID-19

COVID-19 led to some challenges, particularly with recruitment of study sites. The study was originally projected to begin in April 2020, but COVID-19 slowed the recruitment of jails and CBTPs. Sites that had previously shown interest withdrew from the study after experiencing a lack of staff resources and time, a shift in their agency’s primary focus, and a decrease in jail populations and arrests. In July 2020, the research team received feedback from potential sites confirming that they were now functioning better within their agency’s COVID-19 protocols. The research team reinitiated recruitment efforts. Although progress is slow, 21 jail and CBTPs have joined the study.

The research team has adjusted the protocol to include a shift in the study timeline to be adaptable and has changed the format of the study rollout. These changes are:

• The study converted the NIATx coaches’ training, typically a two-day in-person training, into a virtual training format that included five, two-hour sessions that spanned over a two-month period.

• The study will be rolled out in two cohorts. The first cohort of sites will begin implementation in January 2021 with a second cohort starting sometime in March or April 2021.

• The study will conduct in-person kick-off meetings and on-site coach visits virtually for the foreseeable future.

Highlights.

• Approximately half of state and federal inmates meet criteria for a substance use disorder, however fewer than one-percent of the more than 5,000 U.S. prisons and jails that house more than two-million inmates, allow access to FDA-Approved medications to treat opioid use disorder (OUD).

• While administering MOUD, jail settings encounter a common set of barriers: the stigmatization of substance use disorder, funding for MOUD, institutional design, leadership support, policy and administrative procedures, and availability and capacity of community-based treatment providers.

• Organizational coaching focused on goal setting has been shown to be an effective implementation strategy to addressing barriers and implementing innovations, but effectiveness of coaching in justice settings are understudied.

• This is the first trial that assesses the comparative effectiveness of two coaching methods, the NIATx method, aimed at organizational processes, and the ECHO method, aimed at clinical providers, within justice settings for the purpose of increasing the use of MOUD.

• The first cohort of study sites is projected to roll-out in January 2021 with a second cohort beginning in April 2021.