Child Welfare Outcomes and Risk Assessment Scores

Background and Purpose:

The Differential Response System (DRS) is a set of policies establishing multiple pathways for families reported for child maltreatment; serious reports are still investigated, and families with allegations involving low to moderate risk reports with no safety factors are offered services to meet their needs. Typically, DRS has two tracks, a traditional Investigation/Intake and a Family Assessment Response (FAR) for families with a low-moderate safety risk.

The statewide DRS system that is the focus of this study utilizes the Structured Decision Making (SDM) Risk Assessment of Abuse/Neglect tool to identify factors related to future risk of harm for FAR families.  The risk assessment results inform recommendations and decisions relative to the intensity of the services provided and the nature of DCF's continued involvement.

The aims of this study are to 1) examine the relationship between risk level score and subsequent and substantiated subsequent reports of child abuse and/or neglect for FAR families who participate in a voluntary, family driven program after completing the FAR protocol, and 2) model specific predictors of subsequent and substantiated subsequent reports of child abuse/neglect for the program families.

Methods:  

As part of a comprehensive evaluation of a statewide DRS system, we examined the CPS agency’s electronic administrative admissions and services records for over 4,600 families from 2012-2015. Two sets of Cox regressions (also known as Cox proportional hazards model; Cox, 1972) for survival analysis were carried out to determine risk factors associated with subsequent reports and substantiated subsequent reports. Factors included in the first set of Cox regressions were: race/ethnicity of the care givers, employment status, families with younger children (< = 5 years), CPS regions, and overall SDM risk scores.  To allow us to estimate the likelihood of subsequent reports associated with each risk assessment item, the second set of Cox regressions include the same factors (excluding overall SDM risk scores) and each individual risk assessment items from the SDM Risk tool.

Results:

Of the program families (n=4,656), 33.9% have a subsequent report and 7.7% have a substantiated subsequent report within 12 months of starting CSF.  The majority of subsequent and substantiated subsequent reports occurred within the first six months.  Risk factors include unemployment, having children under the age of five, increased risk assessment score, and race (white).  Individual items from the SDM Risk Assessment of Abuse/Neglect tool that were risk factors for subsequent and substantiated subsequent reports include; prior CPS history, homelessness, a delinquency history, and a caregiver with a current or past alcohol problem. 

Conclusions and Implications:  

Risk assessment scores provide some insight into predicting the outcomes related to subsequent reports of child abuse and neglect.  Examining the relationship of individual items from the SDM Risk Assessment of Abuse/Neglect tool provides valuable information to identify and secure the appropriate supportive services to better address the needs of the family and to improve family functioning and well‐being.