Implementation mechanisms used in national efforts to improve community services to keep individuals with mental illness out of local jails.

Background: Little is known about effective implementation processes by which counties can improve community services to keep people with mental illness and substance use disorders out of the local jails. This study examines hypothesized implementation mechanisms (relationship building, performance monitoring, interagency coordination, capacity building, and infrastructure programming) as predictors of outcomes (improved community services) and as mediators of the effects of a national implementation intervention (Stepping Up [SU]), on community services. Methods: A survey was conducted of mental health, substance use, jail, and probation administrators in 519 U.S. counties, of which 328 counties participated in a national jail reform effort (SU). Survey data were combined with descriptive data from the U.S. Census Bureau. Predictors included hypothesized implementation mechanisms (performance monitoring, interagency coordination teams, creating integrated systems of care, capacity building, relationship building, and quality programming). Covariates included county sociodemographic characteristics (e.g., size of county, size of jail, etc.) and general county service characteristics (e.g., primary care physicians per capita, Medicaid expansion). Implementation outcomes included number of evidence-based practices (EBPs) and evidence-based mental health treatments (MH-EBTs) for individuals with mental illness involved with justice systems. Multilevel regression analyses examined cross-sectional: (1) effects of Stepping Up on outcomes; (2) effects of implementation mechanisms on implementation outcomes; and (3) implementation mechanisms as mediators of the effects of Stepping up on implementation outcomes. Findings: SU was found to significantly predict the number of EBPs and MH-EBTs controlling for various demographic characteristics of the counties. When implementation mechanisms were added to these models, SU is no longer statistically significant. Instead, two implementation mechanisms (performance monitoring and interagency coordination) and Medicaid funding significantly predicted the availability of both EBP and/or MH-EBT. Other factors that predicted MH-EBTs include relationship building size of the county, rate of primary care physicians, rate of MH providers in the county, and jail population size. Mediation models found that SU significantly predicted these evidence-based outcomes through implementation mechanisms except interagency coordination. Conclusions: Little is known about the implementation mechanisms to decarcerate and build programming for MH services in a county. SU is an important attribute to facilitate reform both directly and indirectly through implementation mechanisms. Counties can benefit from use of relationship building activities to advance policy and service reform efforts, identifying performance metrics of their system, and having infrastructure available to improve the availability of EBPs. Overall, policy changes are possible, but an emphasis should be on strategies that increase the availability of EBPs and MH-EBTs.