Yampolskaya, S., Armstrong, M. I., et al. (2011). "Contextual and individual-level predictors of abused children's reentry into out-of-home care: A multilevel mixture survival analysis." Child Abuse and Neglect, 35(9), 670-679.
This study examines the effect of child and child welfare organization characteristics on the re-entry rate of children to out-of-home-care. The authors describe re-entry as an important indicator given that it may suggest that an additional allegation of serious maltreatment has occurred, that permanency has not yet been achieved, and the well-being of the child could be in jeopardy.
Using secondary data analysis of three administrative data sets, the study followed all children in the state of Florida who were discharged from out-of-home-care in the 2004-2005 fiscal year (n=17,695). Although others have examined the link between child characteristics and re-entry into out-of-home care, this study also examined contextual factors that have not been studied before.
Both individual and organizational characteristics were found to be associated with re-entry. Older children were more likely to re-enter care while children placed because of neglect and those with physical health problems were less likely to re-enter. Contextual factors associated with faster re-entry included lower average expenditure per child and contracting out of case management services. The authors recommend considering the use of programs to support older youth, given the higher risk of re-entry for this population. The authors also recommend assessing the appropriateness of the length of stay and type of placement for individual children. These factors may explain the effect of lower average expenditure on re-entry rates as some children may be requiring higher-level services.
The authors used three existing data sets in order to answer the research questions. Multilevel cox regression analysis was applied to the data. This technique allowed for the study of two levels of data (individual and contextual) to be measured simultaneously. This is the appropriate type of analysis given the research questions. Limitations include those related to the use of secondary data analysis. For example, the original purpose of the administrative data was not for research, the quality of data could be questioned and the researchers were limited to variables that were included in the original data set. The generalizability of the findings is also limited given the uniqueness of the child welfare system studied.