Alternative Response in Child Protective Services: Evaluating Implementation in Nebraska

Abstract

Background

Through a Title IV‐E waiver demonstration project, the Nebraska Division of Children and Family Services (DCFS) planned to improve the service outcomes with Alternative Response (AR), as opposed to Traditional Response (TR). AR allowed for Nebraska’s child welfare system to engage with families in a non‐investigative and more collaborative way, based on the severity of initial allegations. This family‐centered response aims to improve the outcomes for participating children and families.

DCFS contracted with the Center on Children, Families, and the Law (CCFL) at the University of Nebraska-Lincoln and UCLA to conduct this program evaluation. This evaluation of the AR program contributes to an understanding of how the demonstration leads to enhanced outcomes for the children and families in the system. As part of the full evaluation, this study mainly focused on comparing incident recurrence, family safety outcome, and out-of-home placement between the AR and the TR tracks.

Methods

The AR program was evaluated through a randomized controlled trial. After the initial assessment, over 4000 AR-eligible cases were randomly assigned to either AR or TR tracks. We followed families assigned to either AR or TR tracks using a unique master case number across a variety of Nebraska DCFS administrative data sources from 2014 to 2019.

Family safety was measured by the safety assessment tool within the Structured Decision Making (SDM) model and was compared between groups using ordered logistic regression. Subsequently, we used survival (time to event) analysis to compare the incident recurrence between the AR and TR tracks. The time of failure was calculated as the period between the first and the second report (if ever) within the same family, and the “event” in this case refers to the recurrence of the second report within the same family. Some of the covariates included were substantiation decision, family risk level, and allegation type.

Results

Overall, 97% of AR-eligible families were assessed as “safe” based on the SDM assessment tool, compared to conditionally safe or unsafe. AR families are almost twice as likely to be evaluated as “safe” compared to AR-eligible TR families (OR = 1.96, p < 0.001). When controlling for risk level, a significant increase in probability (Hazard Ratio = 1.2, p < 0.001) of repeated reports was observed for TR families compared to AR families. Additionally, although the overall relationship between out-of-home placement and track assignment was not significant at the family level, it was significant at the individual level (Χ²(DF=1,N=927) = 4.93, p = 0.027).

Conclusion

The results suggest that families in the AR program are more likely to be deemed safe using the SDM tool. The analysis also indicated a higher chance of re-entry to the system among TR families and a significant difference in out-of-home placements between individuals in the two programs. Such findings urge continued effort to implement and evaluate AR programs and call for further research on effectiveness of differential response in child welfare.