Individuals who commit acts of violence in prisons are often placed in highly controlled environments called restrictive housing (i.e., solitary confinement), which can have severe physical and mental health consequences and does not reduce violence. As such, North Carolina prisons have introduced the rehabilitative diversion unit (RDU) to reduce the use of restrictive housing and reduce violence in prison. We evaluated the effect of the RDU on prison infractions. We compared rates of infractions by type (including violent infractions) among men enrolled in the RDU and men who were eligible for the RDU but placed in restrictive housing for control purposes (RHCP). We also evaluated sustained program impacts by comparing the hazard of first infraction among these same two groups of men after program completion, when they had returned to the general prison population. Finally, we compared the hazard of first promotion to a less restrictive custody level (medium custody) when these men had returned to the general prison population. The primary analytic cohort was made up of 3128 men contributing 897,822 person-days. Adjusted rates of violent infractions were lower in the RDU than in RHCP (adjusted rate ratio: 0.6; 95% CI: 0.4, 1.1). All other categories of infractions, including drug-related infractions, occurred at higher rates during RDU, as compared to RHCP. In analyses of sustained program impacts, for most categories of infractions, there were no differences in the hazard of first infraction post-RDU and post-RHCP. However, the hazard of violent infraction post-RDU was higher (adjusted hazard ratio: 2.1; 95% CI: 1.1, 4.0) than post-RHCP. The hazard of promotion to a less restrictive custody level was higher post-RDU (adjusted hazard ratio: 17.4; 95% CI: 7.2, 42.2) than post-RHCP. We found the RDU program may be effective in reducing violence for men enrolled in the program, but that these benefits were not sustained. Continued programming may be a useful tool to transition men from the programmatically intensive environment of the RDU to the general prison population. Additionally, we recommend the expansion of evidence-based treatment for substance use disorder.