Updated Allegheny Family Screening Tool precedes fewer home removals

Stanford evaluators found that after Allegheny's LASSO screening model and policy changes, the 180-day home removal risk after a screen-in fell 2.7 points. Acceptance for services fell 8 points.

  • 2 a person

    Children in screened-in episodes were 2.7 percentage points less likely to be removed from home within 180 days.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported and intended benefits.
Evidence

foundstudiedreportedclaimed

Only the evaluators support the benefit.

Compared with

General protective service episodes before any tool, January 2015 to July 2016. Measured.

The system’s part
Not isolated. The evaluators compared periods before and after the tool, with child protective service episodes as a second comparison. They state that the tool's effects cannot be fully separated from policy changes and COVID-19. The report notes that the model was re-estimated with new weights in July 2019.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Challenged, 12 June 2023. An audit by ACLU and HRDAG staff found that the tool, measured as the county uses it, had lower accuracy, more false positives and wider racial disparities than the county reported.

Download data Cite Suggest a correction

Benefit type
Reduced Risk
System
Allegheny Family Screening Tool, LASSO version (predictive risk model for child welfare hotline screening)
Made by
Not named in the evaluation (built for the Allegheny County Department of Human Services)
Used by
Allegheny County Department of Human Services, Office of Children, Youth and Families
Where
Allegheny County, Pennsylvania, United States
Stage
In routine use
Date
Kind
Aggregate result
Broad benefit
Judging the Qualities of Objects, Services, or People, Making Decisions and Solving Problems
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Peer-reviewed study, by authors with no tie to the system. Context or contrary evidence.

    Gerchick M, Jegede T, Shah T, Gutierrez A, Beiers S, Shemtov N, Xu K, Samant A, Horowitz A. The Devil is in the Details: Interrogating Values Embedded in the Allegheny Family Screening Tool. Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency (FAccT '23), pp. 1292-1310, June 2023. doi:10.1145/3593013.3594081 (full text on aclu.org).

    But when we measure the AFST in a manner more closely aligned with how it is deployed — using a maximum score policy or a policy of assigning a “protocol” to each family — we sometimes see a lower AUC, a higher false positive rate, and greater racial disparities in performance results

  2. Evaluation report, by an evaluator that the makers or users commissioned.

    Goldhaber-Fiebert JD, Prince L. Impact Evaluation of a Predictive Risk Modeling Tool for Allegheny County (Phase 2). Stanford University, 26 March 2023. Published by the Allegheny County Department of Human Services.

    The 180-day risk of home removal after an index GPS screen-in episodes decreased by 2.7 percentage points with the implementation of LASSO and surrounding policy and practice changes.

Details

AI system

Workflow

The highest-scoring referrals triggered a High-Risk Protocol that defaults to a screen-in unless a supervisor overrides it. The model was re-estimated in July 2019 with new data elements and weights.

Follow-up

Conditions for replication

Referrals are processed quickly enough for the score to use current data. Supervisors review the protocol defaults.

Entry details

Sector

Child welfare.

Purpose

Developers working for Allegheny County's Department of Human Services built the LASSO version of the Allegheny Family Screening Tool to help call screeners decide which child welfare referrals to investigate.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). Updated Allegheny Family Screening Tool precedes fewer home removals (AIBD-0728). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0728.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions