Trajectories of Risk Leading to out-of-Home Placements and Identifying Factors Related to Permanency

Introduction: Interactions with child welfare systems for maltreatment allegations can range from low risk to immediate removal a child. Continued referrals for maltreatment provide a profile of how observed risk develops for a child as more information is gathered through repeated contact. Children removed from homes due to maltreatment are then subject to experiencing further issues opposed to permanency. We introduce a risk model for predicting out-of-home placements within two years and correlate these scores with permanency outcomes. We examine the likelihood of permanency during adolescence by assessed risk prior to placement, as well as the trajectory of risk leading to placement.

Methods: Child Protective Services records from California’s Child Welfare Services Case Management System were used to identify all foster youth between ages 13 and 18 with referrals between 2010 and 2015 in Los Angeles county (N=14,884). Data included out-of-home placement information following these adolescents through 2017, allowing for two years of follow-up after the end of the referral observation period. Variables used as predictors for risk of out-of-home placement were obtained from the focal referral and all historical data related to individuals within the focal referral. Permanency outcomes were coded by identifying exit reasons for each placement episode. A predictive risk model was constructed using logistic regression, predicting the likelihood of an out-of-home placement within 2 years of the referral date. These risk scores were then used to identify the most recent risk score prior to an out-of-home placement and trajectories of risk over the observation period (e.g., increasing, level, decreasing). Multivariable regression models were then used to determine the relationship of permanency to modeled risk scores and other referral history characteristics. All analyses were conducted using the Stata statistical package version 17 and R version 4.0.2.

Results: The average risk score decile (1-10) for youth prior to entering an out-of-home placement was 8.0 (sd=2.36). Almost half of adolescents with out-of-home placements had 2 or more episodes (47.7%), where 7.5% experienced declining risk, 12.6% had no change in risk, and 27.6% had increased predicted risk. Youth were found to be less likely to experience permanency as risk decile score increased (OR=0.93, CI=(0.92,0.95)). When there was equal risk predicted across referrals (e.g., no change in risk over time), there was an increase in the odds of permanency compared with youth with a single referral (OR=1.17, CI=(1.01,1.36)) and no significant change for youth with declining or rising risk scores over time. Older aged youth (OR=0.84, CI=(0.81,0.86)) and more referrals (OR=0.92, CI=(0.88,0.95) both indicated a decrease in odds of permanency.

Conclusions/Implications. The dynamics of risk leading to an out-of-home placement and the effect on achieving permanency show that recent predicted risk for placement is a contributing factor. This is important for caseworkers to understand that while there is an immediate need, ongoing issues will need to be addressed for permanency to be successfully achieved. Supports such as additional case services should be examined to improve odds of permanency.