NeuroResource Facilitation for Improved Re-Entry Outcomes for Offenders With Brain Injury: A Multi-Site Randomized Controlled Trial

The prevalence of brain injury (BI) is significantly greater in justice-involved populations, and a substantial proportion of prisoners have likely experienced a BI during their lifetime. A history of BI has a statistically significant association with increased use of correctional, medical and psychological services including crisis intervention; an increased frequency of institutional misconducts; and higher recidivism rates. Brain injury in justice-involved populations is frequently undiagnosed, and therefore untreated, which contributes to the cycle of recidivism. Identifying BI opens up new resources to ex- offenders that can more effectively help them to become productive in their lives in the community. In partnership with the Icahn School of Medicine at Mount Sinai, the Brain Injury Association of Pennsylvania, and the Pennsylvania Department of Corrections (PADOC), this project will implement a randomized clinical trial to determine if a case management intervention called Neurorehabilitation Facilitation (NRF) is effective in reducing recidivism by ten percent in offenders with acquired BI. Offenders in two Pennsylvania prisons (State Correctional Institution (SCI) sites in PA, SCI Phoenix, SCI Chester, SCI Frackville, and SCI Mahanoy) will be screened for BI and cognitive impairment, and those who screen positive will be randomized to receive the intervention or to receive standard of care (SoC). Two thirds of those randomized will receive the intervention and the remaining one third will serve as controls. All participants will be followed for up to three years following their release to examine the immediate and long-term reduction in recidivism.

The primary hypothesis is that NRF, as compared to SoC, will result in a 10% reduction in average 1-year recidivism, estimated by the 2013

Department of Corrections (DOC) report as 35-38%. Understanding that additional factors will contribute to the effect of NRF, the required sample was based on the primary analysis, a logistic regression of the binary response variable recidivism, adjusted for additional independent variables (described below) in the model with estimated R-squared=0.10. An effect size of 10% reduction corresponds to an odds ratio in this model of 0.643. With two thirds (67%) of cases randomized to NRF, and one third (33%) randomized to SoC, a total sample of N=688 provides 80% power at alpha set at 0.05. Anticipating that there will be a loss of approximately 10% of randomized cases to attrition, N=764 cases will be randomized, with n=509 cases randomized to NRF and 255 randomized to SoC.