Allegheny Family Screening Tool
Researchers led by Rhema Vaithianathan modelled, designed and supported implementation of this world-first child welfare predictive analytics tool. The Allegheny Family Screening Tool (AFST) uses rich administrative data to generate a screening score for incoming calls alleging child maltreatment and neglect. The score is an additional piece of information that helps call screeners as they decide whether to open an investigation. Allegheny County introduced this decision support tool with the aim of improving accuracy and consistency of call screening decisions. An independent impact evaluation of the tool was completed by researchers at Stanford University in March 2019. The evaluators’ findings included that use of the tool improved the accurate identification of children in need of services and was associated with a modest reduction in racial disparities in case openings.
- Vaithianathan R, Putnam-Hornstein E, Chouldechova A, Benavides-Prado D, Berger R. Hospital Injury Encounters of Children Identified by a Predictive Risk Model for Screening Child Maltreatment Referrals: Evidence From the Allegheny Family Screening Tool. JAMA Pediatr. Published online August 03, 2020. doi:10.1001/jamapediatrics.2020.2770
- Key AFST documents including Methodology and Methodology V2, Ethics Report, Frequently Asked Questions and Impact Evaluation (March 2019) along with press coverage are available on the Allegheny Analytics site.
- Vaithianathan, R., Kulick, E, Putnam-Hornstein, E & Benavides Prado, D. (2019). Allegheny Family Screening Tool: Methodology, Version 2. Centre for Social Data Analytics. Auckland: Centre for Social Data Analytics.
- Chouldechova, A., Putnam-Hornstein, E., Benavides-Prado, D., Fialko, O., & Vaithianathan, R. (2018). A case study of algorithm-assisted decision making in child maltreatment hotline screening decisions. Proceedings of Machine Learning Research, 1-15.
- Vaithianathan, R., Jiang, N., Maloney, T., Nand, P., & Putnam-Hornstein, E. (2017). Developing predictive risk models to support child maltreatment hotline screening decisions: Allegheny County methodology and implementation. Auckland: Centre for Social Data Analytics.
- Cuccaro-Alamin, S., Foust, R., Vaithianathan, R., & Putnam-Hornstein, E. (2017). Risk assessment and decision making in child protective services: Predictive risk modeling in context. Children and Youth Services Review, 79, 291-298.
- Child Protection and the Promise of Predictive Analytics, July 7, 2017
- CSDA takes Danish experts behind the scenes of US projects, November 28, 2017
- “What is predictive risk modeling and how can it be applied to child welfare?”, Casey Family Programs Child Safety Convening, May 2018
- Bias Detectives: the researchers striving to make algorithms fair" - CSDA’s work in Nature news, July 2018
- Can an algorithm keep kids safe? Innovation Hub, June 2018
- Analysis of screening tool wins best paper award at international fairness conference, February 15, 2018
- Can an Algorithm Tell When Kids are in Danger?" - our research in the NYT, January 12, 2018
- Allegheny Family Screening Tool evaluation: Improved decision accuracy, reduced disparities, May 3, 2019
Location: Allegheny County – PA, United States
Partner/s: Children’s Data Network (University of Southern California), Carnegie Mellon University, Allegheny County Department of Human Services