Children involved in the child welfare system are at high risk for developing behavioral health problems, but only a fraction of these children receive specialty behavioral health care. System reforms that routinize behavioral health screening and assessment, and strengthen linkages between child welfare and behavioral health systems have potential to improve identification of behavioral health service needs, and facilitate service access. This study examines the effectiveness of Gateway CALL - a system demonstration project that implemented a behavioral health screening-assessment-referral model to impact child welfare-involved children’s behavioral health, safety, and permanency outcomes.
This quasi-experimental study included 489 children (birth-18 years) who entered child welfare custody and were placed in out-of-home care through 8 experimental and 8 comparison (service as usual) intake units between February 2015-July 2016. We observed children for up to 24 months by drawing on linked child welfare, Medicaid billing, and behavioral health assessment records [results from the Child Behavior Checklist (CBCL), which was administered to caregivers every 90 days while a child was in custody)]. We used propensity score matching techniques to refine the composition of our comparison group based on demographics and behavioral health risks. We conducted intent-to-treat regression analyses to examine the effect of the intervention on safety (number of screened in calls, substantiated allegations, and returns to custody), permanency (adoptions, reunifications, and placement moves), and well-being (whether behavioral health services received, and number of visits) outcomes. For those in the experimental group, we also used a one-group longitudinal design, and generalized linear mixed models to examine the relationship between number of behavioral health service visits and change in behavioral health assessment scores.
In terms of safety, children who received the intervention were more likely than comparison-group children to have screened-in reports of abuse, and physical abuse, but not substantiations. In terms of placement permanence, there were no observed effects. In terms of well-being, intervention-group children were no more likely to receive behavioral health care services than those who received services as usual, but children who received the intervention attended nearly three times the number of behavioral health provider visits (IRR=2.98, SE=1.04, p<.01). Children’s internalizing, externalizing, and total problem behaviors declined significantly over time; and the number of behavioral health service visits was significantly associated with internalizing (b =-.02, SE=.01, p =.019) and total behavior problems (b =-.02, SE=.01, p =.045) over time.
Given that children in the intervention were more likely to have subsequent abuse reports, but not substantiations, greater service engagement may have created a surveillance effect on children’s safety. The intervention did not appear to improve children’s likelihood of receiving behavioral health care, suggesting the need for closer alignment of child welfare and behavioral health services to address unmet needs. However, for children who did receive treatment, the intervention appeared to increase the number of behavioral health service visits. Moreover, the more services children received over time, the more their behavioral health symptoms improved suggesting a positive effect of community-based services when children can access them.