Early Identification and Prevention of Child Maltreatment: Cross-Agency Processes and Outcomes

In the U.S., 1 in 8 youth will experience child maltreatment by age 18. Young children, aged 0-8, are particularly vulnerable. Local social service agencies and health care providers routinely make decisions regarding a child’s risk for maltreatment. Yet, providers have limited information to guide their decisions and rarely receive feedback regarding the children’s long-term outcomes. However, research suggests that children reported for alleged maltreatment tend to have poor health outcomes and are often re-reported to social services. The proposed study aims to better predict which children who are the subject of alleged maltreatment experience poor outcomes in the health and social services systems. We hypothesize that by incorporating information for the focal child and their family members’ health and social service use patterns, and social risk factors (e.g., housing insecurity), we will be able to improve our understanding of a typology of children at-risk for poor outcomes. The lack of a feedback mechanism regarding how children fare following an alleged report for maltreatment prevents case workers from reflecting on prior decisions. Moreover, because caseworkers can only access information collected within their agency, they are missing critical information detected by the health care system. Information from health care providers could inform needed family preservation services that would improve child safety and keep families intact. This study aims to use predictive analytics to better predict who is at-risk for maltreatment and subsequent adverse outcomes. We will begin with children who are reported to child protective services for alleged maltreatment and incorporate information on prior (a) health care use of the focal child, his/her siblings and parents; (b) social services involvement of the focal child, his/her siblings, and parents; and (c) the parents’ exposure to other social risk factors including criminal justice system involvement and housing insecurity. These analyses will aid in the development of predictive algorithms to identify children at risk of poor outcomes and improve service delivery.