The Role of Risk in Child Welfare Decision-Making: A Longitudinal Examination of Transfers to Ongoing Services

Background and Purpose: Analyses using provincial and national incidence study data in Canada have identified several factors associated with the provision of continued child welfare services after an investigation, including children’s age, race/ethnicity, and functioning concerns; caregiver’s age and risk factors; and socioeconomic circumstances. Importantly, determinations that children are at risk of future maltreatment have been found to influence investigative decision-making over above all other family and case characteristics. This study extends this foundational work by using longitudinal administrative data to assess the decision to transfer a family to ongoing services within twelve months of an initial investigation.

Methods: We developed an entry cohort from agency-level administrative data that were extracted, harmonized, and stored by the Ontario Child Abuse and Neglect Data System (OCANDS). The dataset included all families investigated during 2012 (January to December) across eight child welfare agencies, and these cases were followed for twelve months to determine which families were transferred to ongoing child welfare services. Variables included characteristics of the child, family, and investigation, and information generated from actuarial safety and risk assessments. We conducted descriptive comparative analyses, comparing families that were transferred within twelve months of an investigation to those that were not, and used these results to generate a generalized linear model that assessed predictors of the decision to transfer.

Results: Of the 27,135 families investigated in Ontario agencies in 2012, 23% were transferred to ongoing child welfare services. Rates of transfer were highest among investigations in which immediate safety was a concern; future risk of maltreatment was identified; child behavior and parent-child conflict, caregiver capacity, or exposure to intimate partner violence (IPV) were the primary reason for the investigation; and investigation concerns were verified by the investigating worker. Family size, age of the youngest child, and caregiver and household risk factors were also associated with higher transfer rates. In the multivariable model, assessment of future risk was the most significant and substantial predictor, followed by investigation reason, and initial determinations about immediate safety. Families considered high risk (RR: 5.61; 95% CI: 4.93, 6.38, p<0.001) or very high risk (RR: 6.74; 95% CI: 5.83, 7.79, p<0.001) were far more likely to be transferred, as were families where concerns about child behavior, child and caregiver conflict, and caregiver capacity (e.g., substance abuse, mental health concerns) were identified.

Conclusions and Implications: Decisions about more intensive involvement with families were largely based on safety and risk assessments, which may be measuring the very same constructs that bring a family to the attention of child welfare in Ontario. It is unclear whether child welfare services are the appropriate mechanism for effectively addressing circumstances that may harm children, particularly when the structure and delivery of services have been designed to intervene when children have already been harmed. Considerations for practice and policy will be discussed.