The Allegheny Family Screening Tool (AFST) is a predictive risk modeling tool used in Allegheny County, Pennsylvania, to assess the risk of future child maltreatment. Implemented in August 2016 by the Allegheny County Department of Human Services (DHS), the AFST analyzes numerous data points to generate a “Family Screening Score,” predicting the likelihood of future child welfare involvement and potential out-of-home placement.
How the AFST Works
The AFST integrates and analyzes hundreds of data elements from the DHS Data Warehouse for each individual involved in an allegation of child maltreatment. This data synthesis creates a visualization that helps predict the long-term likelihood of a child needing to be removed from their home. A higher score indicates a greater risk. Scores that meet the “mandatory screen in” threshold require investigation. In other cases, the score informs professional judgment but doesn’t replace it. Importantly, the Family Screening Score is used solely for call screening decisions and isn’t shared beyond this process.
Development and Implementation of the AFST
The AFST emerged from a multi-year process initiated in 2014. Following a competitive Request for Proposals (RFP), DHS partnered with a team from Auckland University of Technology (AUT). This team, led by experts in social data analytics, explored applying predictive analytics to improve decision-making in child welfare. The collaboration focused on developing a tool to aid in determining whether to investigate allegations of maltreatment. The implementation process included community engagement, validation studies, and independent ethical reviews.
Evaluating the AFST’s Impact
Independent evaluations of the AFST have assessed its implementation process and impact. Hornby Zeller Associates, Inc. conducted a process evaluation, examining changes in call screener experiences, policy implications, and perceptions of the tool. Separately, Jeremy Goldhaber-Fiebert of Stanford University evaluated the AFST’s impact on decision accuracy, disparity rates, referral rates, and workload. The DHS continues to refine the model based on these findings and ongoing feedback.
Addressing Concerns and Misconceptions
The AFST has been the subject of scrutiny and debate. DHS has responded to concerns raised in various reports and publications, clarifying inaccuracies and addressing criticisms regarding racial bias and the ethical implications of using such a tool. They emphasize the tool’s role in enhancing child safety and improving decision-making, while acknowledging the need for ongoing evaluation and refinement.
Conclusion
The Allegheny Family Screening Tool represents an innovative approach to assessing risk in child welfare cases. While it has generated controversy, the AFST aims to provide valuable information to call screeners, potentially leading to better outcomes for children in foster care. Continuous evaluation and transparency are crucial to ensuring its responsible and effective use. The tool highlights the growing role of data analysis in informing critical decisions within the child welfare system.
