The Effectiveness of Service Integration: Studying the Crossover Youth Practice Model

Background/Purpose: Prior studies indicate that crossover youth who had a history of involving in both the child welfare and juvenile justice systems, have higher rates of reoffending than delinquent youths without a history of child welfare involvement. This is particularly concerning, since the number of prior juvenile arrests is positively associated with future arrest, a finding that holds true for future arrests during both adolescence and adulthood. Little was known, however, about which interventions could effectively direct crossover youth from continuous involvement in the justice system. The Crossover Youth Practice Model (CYPM) is the only existing practice model for serving crossover youth. The CYPM provides a conceptual map and organizational ideology regarding how staff in the juvenile justice system can partner with families, service providers, and other stakeholders in the delivery of services. The CYPM has been implemented in 73 counties in 25 states. However, the evaluations have studied only a limited number of jurisdictions. This study evaluated CYPM implemented in South Florida.

Methods: Data and Samples: We use quasi-experimental design with group assignment at the county level. We selected one county in South Florida as the experimental county, as CYPM has been implemented there since 2010, and another county in South Florida as the control county, as CYPM is not used there. The two counties are similar in terms of their percentage of children and youth in the population, and child welfare statistics. Our sample includes 86 crossover youth from the experimental county, and 87 crossover youth from the control county. Both groups had juvenile justice referrals around 2013.

Measures: We collected administrative data from both the child welfare and juvenile justice systems, including their demographics, delinquency history, experiences with juvenile justice processing, experiences with dental and medical services referred by child welfare agencies.

Results: To identify confounding variables to include in the multivariate model, we used chi-square statistics to test differences on categorical variables, and used t-test to test differences on numerical variables. In the multivariate analysis, we control for the significant differences. Using cox regression to model time to re-offending, the results show that three variables were statistically significant in most models: as compared with their White peers, Hispanic crossover youth is associated with increased risk of re-offending within a year; as compared with the crossover youth from the control county, the crossover youth from the experimental county is associated with lower risk of re-offending within a year; as compared with the crossover youth with no prior juvenile justice referrals, the ones with three or more prior referrals is associated with increased risk of having new referrals within a year. However, as we add two service experiences variables, neither group membership nor prior referrals were statistically significant.

Conclusions and Implications: Our results show that indicate that CYPM can have an effect on reducing the risk of juvenile recidivism. The difference was mediated by receiving dental and medical services referred by the child welfare agencies. We compare our findings with previous literature, and provide interpretations.