Facilitating Opioid Care Connections: System Level Strategies to Improve Use of Medication-Assisted Treatment (MAT) and Movement Through the Opioid Care Cascade for Defendants in a New Opioid Court System

Aims. Guided by Exploration Preparation Implementation Sustainment (EPIS) model and Social Cognitive Theory (SCT), Specific Aims are: 1) To refine OCM-RISE using mixed-method formative work with 5 NYS "initial adopter" counties, who established opioid courts before 2019, that (a) identifies gaps in service provision by documenting OCM/opioid cascade outcomes; (b) identifies successes/challenges in operationalizing guidelines; and (c) characterizes the working relationships between county opioid court and treatment systems. 2). In a cluster-randomized, stepped-wedge design in 10 "new adopter" counties, which includes some counties who established opioid courts in 2019, compared to baseline treatment as usual (drug courts), test, (a) the implementation impact of OCM RISE in improving implementation outcomes along the opioid cascade (screening/identification, referral, treatment enrollment, MOUD initiation); and (b) the clinical and cost effectiveness of OCM RISE in improving public health (treatment retention/court graduation) and public safety (recidivism) outcomes, exploring moderators: defendant gender, age, charge; county urbanicity and county OD rates. 3) To characterize and compare advancement through the stages of implementation of the OCM in the 10 counties, elucidating the inner- and outer-level EPIS- and SCT-derived factors that influence delivery of implementation strategies to inform OCM scale up; and (b) to explore the relationship between implementation stage completion and all opioid cascade, public health and public safety outcomes.

Setting.The research team will leverage the real-world roll-out of the OCM by the NYS UCS to provide implementation strategies that will optimize the ability of diverse counties to effectively adopt this innovative model. Between 2017 and 2019, several counties have scaled up these courts, and the research team will use their experiences and outcomes in intervention refinement. The team will then evaluate OCM RISE in 10 "new adopter" counties, some if which have established opioid courts in 2019, that all have an existing adult drug (non-opioid) court. Counties are varied in their experience of opioid "burden" as well as the stage of the development of their court.

Design, Sample Size and Randomization. In the proposed cluster randomized stepped-wedge design, 10 new adopter counties will be randomized - stratified by population density/urbanicity (as a proxy for staffing and resource availability), (a metric developed by NYS Department of Health (DOH) that encompasses non-fatal ER visits and hospital discharges involving opioid use, OD deaths, which will be used as a proxy for treatment need), and county court activity (newly established versus not)- to one of 5 waves of OCM RISE at 2-month intervals. The 2-month intervals for randomization reflect the real world need for counties to meet the UCS mandate to establish opioid courts as well as our experience with managing multiple start-ups across NYS. The research team will examine the differences between matched defendants with opioid use (OU) or OUD at baseline (county drug court outcomes in the 3 months prior to randomization) with those after OCM RISE on opioid cascade outcomes and on recidivism. By randomizing sites across time, the control group is taken to be counties that have not yet been randomized by a particular time, i.e., treatment as usual (TAU), allowing us to control for secular changes over time. During the pre-randomization period the opioid cascade outcomes will be collected on defendants with OU or OUD who enter into the existing county drug courts. The team expects that introduction of the OCM will identify a broader case-mix of defendants with OU than those currently identified for drug courts.

Hence, the research team will take into account defendant level differences (e.g. heroin vs prescription drug use, age, OUD severity) when testing the effect of OCM by using propensity score matching. There are two limitations to cluster randomized deigns: potential of uneven distribution of potentially confounding variables within a cluster (i.e. county) as well as interrelationship between variables. Our proposed stratification plan and use of propensity matching will address issues related to confounding. The team calculated our ability to detect an effect of the OCM (i.e. power) using a conservative intraclass correlation coefficient (ICC) of 0.05 denoting a moderate positive correlation between outcomes within a county to address issues of interrelationship between variables at the county level

Participants. Participants will be court and treatment staff who participate in 1) staff surveys, 2) in-depth interviews, 3) focus groups, 4) the interagency change team (IACT) and/or 5) preparation activities (e.g. leadership meeting, orientation meeting). Approximately n=10 court and n=10 treatment staff will be recruited in each county for a total of 200 participants. Main study outcome data are to be drawn from court records of defendants, who are not considered human subjects and will not be enrolled in the study.