Improving Safety in a Maximum Security Female Juvenile Justice Corrections Facility

Purpose: Serious juvenile delinquency is a significant and costly problem in our society. However, custodial environments often exacerbate current problems and promote recidivism. Girls’ delinquency, in particular, may call for trauma-informed approaches within organizations that serve the most serious offenders. This study explores whether implementation of a trauma-informed intervention that aims to change the therapeutic stand of the organization, the Sanctuary Model®, corresponded with improved indicators of physical and psychological safety of staff and youth at the only female maximum-security corrections facility in PA. Results suggest that statistically significant improvements in safety for both youth and staff were realized.

 Methods: This study utilizes administrative and OJJDP developed Performance-based Standards data routinely collected at the facility. The period of focus for the study is from 2008, when the model was first introduced, through 2012, when the model had been fully implemented. The measures examined focused on staff and youth measures of safety and perceived safety with three interests: 1) did measures of safety indicate improvement at the study program, 2) how did change at the study program compare with change in other juvenile corrections facilities during the same time period, and 3) how did the study program rank against the comparison group at both time points. The measures, which are proportions (or percentages), were compared using Fisher’s Exact Test in R. 

 Results: In 2008, the program had a rate of 6.6 incidents of youth misconduct resulting in injury, confinement, and/or restraint per 100 person-days of youth confinement; by 2012 this rate had dropped to 1.0 (p < .001). In the field comparison, the average rate of incidents of youth misconduct increased during the same period from 1.2 in 2008 to 2.3 in 2012 (p < .001). A similar statistically significant comparison pattern is seen in the number of physical restraints, incidents of isolation, room confinement or segregation. For three measures - confirmed cases of institutional abuse or neglect over the last six months per 100 person days of youth confinement, the number of injuries to youth per 100 person days of youth confinement, and the number of injuries to staff per 100 staff days of employment, none showed statistically significant differences between 2008 and 2012 for the NCSTU Girls Program. For the entire field, all these measures showed minor changes for the worse. Assaults by youth on youth decreased for the program, from .3 per 100 person days of youth confinement in 2008 to 0 in 2012 (p < .05). The field had a small increase. Assaults on staff decreased at the program from 1.4 in 2008 to 0.4 in 2012 (p< .01) compared to the field which saw a rise from 0.07 to 0.2.

 Implications: Findings suggest that the facility was a safer place for both residents and staff after implementation of the model. Its safety indicators also compare favorably to those of the juvenile justice correctional field in general.