Developmental Patterns in Foster Care Reentry: A Longitudinal Analysis

Abstract

Purpose: With the adoption of the Adoption and Safe Families Act (ASFA) in 1997, child welfare agencies began to shift emphasis from child protection efforts to family preservation initiatives (D'Andrade & Berrick, 2006). However, evidence suggests that in some cases, these reunification efforts are unsuccessful and children re-enter out-of-home care after being returned to live with their parents (Courtney, 1995; Courtney, Pilavin & Wright, 1997; Wells & Guo, 1999; Shaw, 2006). Theories of developmental victimization (Finkelhor, Cross, & Cantor, 2005; Finkelhor & Hashima, 2001) suggest that there is a reason to anticipate developmental differences in placement and reunification patterns among youth in out-of-home care. Existing literature emphasizes that age (Courtney, 1995; Courtney, Pilavin and Wright, 1997; Shaw, 2006) and maltreatment history (Jones, 1998; Wells & Guo, 1999; Terling, 1999) may predict reentry into care following reunification. Prior reports of child maltreatment have also been found to be higher among children who re-enter care following reunification (Frame, Berrick and Brodowski, 2000). This longitudinal analysis examines re-entry outcomes among a first time entry cohort of children in out-of-home care in a New England state. Specifically, this analysis matches child-level National Child Abuse and Neglect Data System (NCANDS) variables with longitudinal out-of-home care spell files developed with the Chapin Hall Center for Children at the University of Chicago. We hypothesized that for this first time entry cohort, child age at the beginning of placement spell, initial maltreatment allegation, number of prior substantiated reports, and number of prior unsubstantiated reports would be significant predictors of foster care reentry.

Method: This analysis was conducted using matched child files from two archived data sources for one New England state. The State's NCANDS data from 2001-2006 were restructured to identify the alleged maltreatment types and total number of substantiated and unsubstantiated allegations of maltreatment during this time period. Cases identified in these NCANDS submissions were then merged with a longitudinal out-of-home care spell file, which The Chapin Hall Center for Children created using the State's administrative data. The use of a longitudinal file avoids possible bias inherent in cross-sectional analyses of foster care data (Wulczyn, Hislop, & Goerge, 2000). This study utilized Cox regression modeling to identify child and case characteristics associated with foster care re-entry. The working file included both maltreatment and placement data for 1,163 children in out of home during the observation window.

Results: These findings suggest that child age is predictive of re-entry into out-of-home care, even when controlling for type of placement, number of previous allegations and substantiations of maltreatment, and type of alleged maltreatment. Children between the ages of 6-14 were about two times more likely to re-enter care than infants.

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. In addition, this analysis illustrates an innovative approach to using administrative child welfare data to examine outcomes and direct policy and programming.