Research That Matters (January 17 - 20, 2008)

Purpose: Children's experiences in the child welfare system reflect the unique characteristics of their chronological age and corresponding developmental stage in life (Wulczyn, Barth, Yuan, Harden, & Landsverk, 2005). Theories of developmental victimization suggest that there is a reason to anticipate developmental differences in both trajectories of maltreatment and child welfare system intervention (Finkelhor, Cross, & Cantor, 2005; Finkelhor & Hashima, 2001). Existing research has provided a limited examination of developmental differences in trajectories of foster care placement, with an emphasis on outcomes for infants and toddlers (Frame, 2002; Jaudes & Ekwo, 1997). It is likely that foster care placement trajectories and associated outcomes may vary for adolescents, given their characteristic task of identity development (Erikson, 1950) and other developmental features. This longitudinal analysis of New Hampshire Adoption and Foster Care Reporting System (AFCARS) data files examined characteristics of youth entering foster care between 2001 and 2003. Specifically, it examined the impact of child and case characteristics on exits from all types of foster care and on exits to reunification with family, with an emphasis on comparing exit outcomes for adolescents and other children. We hypothesized that for all youth in care, child age, prior removal history, placement setting, and reason for removal would be significant predictors of foster care exit. For adolescents in care, we hypothesized that characteristics related to the developmental period of adolescence would be significantly related to the likelihood of exit from care. These characteristics include group or institutional placement settings, removal for child behavior problems, and previous removals from the home. Method: This analysis was conducted using New Hampshire Division for Children, Youth, and Families AFCARS data submissions in 2001, 2002 and 2003. The use of a longitudinal file reflecting one entry cohort avoids possible bias inherent in cross-sectional analyses of foster care data (Wulczyn, Hislop, & Goerge, 2000). This study utilized Cox regression modeling to identify characteristics associated with foster care exit generally, and more specifically with family reunification during this period. The working file included 1,625 children removed during the observation window. Results: These findings generally support this study's hypotheses regarding the factors that predict length of stay in foster care. The analysis finds that children likely to exit all types of foster care are more likely to be above the age of 14, experiencing their first removal, placed in non-relative care, and removed for child behavior problems. We find that children experiencing their first removal and those removed for physical abuse or child behavior problems are most likely to be reunified with families. Among adolescents, those removed for behavioral problems were more likely to exit care and to be reunified with parents. Implications: These findings support the importance of considering developmental age, as well as other child and case characteristics, in child welfare service provision and in assessing foster care outcomes. This analysis finds support for consideration of adolescence as a developmental period with unique child welfare implications.