Association of a novel restrictive housing diversion program with rates of mental health and self-injury in prison.

Restrictive housing for control purposes (RHCP, a form of solitary confinement) is used in prisons in response to disruptive behaviors, including violence. North Carolina prisons introduced the rehabilitative diversion unit (RDU) in 2016 as an alternative to and step-down from RHCP. We compared rates of psychiatric treatment and self-injury among men enrolled in the RDU and men eligible for the RDU but placed in RHCP. We used Poisson regression to calculate rate ratios (RRs) and 95% confidence intervals (CIs) for these outcomes, using inverse probability of treatment weights to adjust for confounders. The cohort included 1,225 men in the RDU group and 3,059 in the RHCP group. In RHCP, the adjusted psychiatric treatment rate was 2.6 times (95% CI: 1.8, 3.8) and the adjusted self-injury incident rate was 1.2 times (95% CI: 0.6, 2.8) that in RDU. Nearly all self-injury incidents in RDU occurred during "non-participating time" (i.e., in a restrictive housing setting and not actively participating in RDU). After excluding non-participating time from the RDU group's person time, the adjusted RR for self-injury incidents was 23.5 (95% CI: 8.6, 64.2). These results further knowledge of potential benefits of diversion from restrictive housing. Continued development, implementation, and evaluation is needed.