Predicting out-of-home placements of children and adolescents with serious emotional disturbance (SED)
Abstract
The purpose of the present study was to examine a set of indicators and factors to predict future out-of-home placements for children and adolescents with serious emotional disturbances (SED). Using characteristics of children and families at intake, this study predicted future out-of-home placements after participation in the Children’s Partnership, a systems of care program funded by the Center for Mental Health Services (CMHS) that serves children and adolescents with SED and their families in Travis County, Texas. A series of hierarchical logistic regression analyses were conducted to evaluate both individual predictors and conceptual models. Contrary to expectation, descriptive indicators (diagnostic information and risk factors) and protective indicators (the BERS and the FAD) were not statistically significant predictors of future out-of-home placements. Only two pathological indicators, as a set, showed a significant contribution to predicting future out-of-home placements. The CAFAS, which is measuring functional impairment of children with SED, demonstrated a strong individual relationship with the dependent variable even after controlling all the other indicators in the model. In addition to examining a set of indicators to predict out-of-home placements for children with SED, this study also explored profile scores of each predictor at intake for children and adolescents with high risk of future out-of-home placements. Results of independent t-tests were quite consistent with the findings observed in the multivariate logistic regression analysis. The children who had outof-home placement at follow-up period showed much severer functional impairment at intake measured by the CAFAS, compared those did not have any out-of-home placement. Overall children in placement group enrolled into the Children’s Partnership with worse symptoms and lower levels of protective factors, compared to children without any out-of-home placement. The findings of the study help clinicians identify children with high risk of out-of-home placement from the beginning and it assists them utilize profile information for their service planning and the early intervention. With several limitations, the study also suggests combining both multivariate and univariate analysis technique is preferable to get a better understanding of each relationship observed in both methods.