Method: This study used treatment and matched comparison data to test foster care exit patterns of families with children in foster care due to parental substance use. Treatment group data was collected on a sample of 64 children with open dependency cases in an integrated FDC in a rural Midwestern town. From a pool of 18,657 children in the State’s child welfare database, propensity score matching on thirteen predictive variables, including county population density, was used to generate a comparison group. Stata version 15.0 was used for propensity score nearest neighbor one-to-two matching within a caliper, with 0.25 of a standard deviation as caliper size (Guo & Fraser, 2009). This resulted in a matched comparison group of 102 children. This study used survival analyses to examine time to, and likelihood of, exiting to reunification and permanency.
Results: Findings suggest that FDC participation significantly influenced foster care exits. By 22 months after entering foster care, 42% of the FDC group experienced reunification compared to 12% of the comparison group. At study end, 50% of the FDC group reunified compared to 16% of the comparison group. The Cox proportional hazards model indicates that FDC children were significantly more likely to reunify than comparison cases (H.R.=4.63, p<.001). The standardized coefficient reveals that FDC children experienced a 363% increased hazard of reunification versus comparison children. The permanency model revealed that FDC children experienced a 140% (p<.001) increased hazard of permanency.
Implications: Using the most rigorous quasi-experimental evaluation tools at the disposal of researchers in child welfare—a setting where randomized controlled trials are often unfeasible—this study adds to the growing body of literature suggesting that family drug courts are an effective service delivery strategy for meeting policy driven goals with this challenging population. As family drug courts continue to expand, the role of social work within the court structure, and implications for social work practice models need to be the subject of future research.