Organized Around Impact: Safer Children, Stronger Families, and A Home for Every Child - Project Title: Organized Around Impact: Safer Children, Stronger Families, and A Home for Every Child Applicant: DISTRICT OF COLUMBIA Child and Family Services Agency (CFSA) Address: 200 I Street, SE, Washington, D.C. 20003-3317 • Phone: 202-727-7197 • Email: stephanie.peguero@dc.gov • Website: https://cfsa.dc.gov/ Project Director: Stephanie Peguero, Administrator • Authorized Organization Representative: Dr. Brandynicole Brooks, Deputy Director for CFSA The DISTRICT OF COLUMBIA Child and Family Services Agency (the Agency or CFSA), the District’s public child welfare agency, requests $600,000 over a 36-month project period to serve as a national demonstration site for the responsible design, implementation, testing, validation, and monitoring of predictive analytics in child welfare. The Agency served 669 children in foster care and 2,037 children through in-home services in Fiscal Year (FY) 2025, screened in 3,799 hotline maltreatment reports during that same period, and manages a federally compliant Comprehensive Child Welfare Information System (CCWIS), Stronger Together Against Abuse and Neglect in DC (STAAND), built by Microsoft. CFSA will deploy five predictive artificial intelligence (AI) capabilities. These models, spanning service needs predictions, placement‑stability scoring, and maltreatment recurrence predictions, are organized around A Home for Every Child. Need and population. CFSA must make high-stakes safety, service, and placement decisions under limited time and information. Manual risk and safety tools and unstructured clinical judgment — particularly amid a 29% vacancy rate among the Office of Hotline & Investigations — may indirectly contribute to inconsistent decisions, late service connections, and preventable maltreatment recurrence of 9.7%. The population of focus is all children and families who come to the Agency attention across the District, particularly Wards 5, 7, and 8. Proposed services and approach. With Microsoft and Chapin Hall as committed partners, the Agency will design, implement, test, and continuously monitor three predictive analytics strategies, or Pillars, drawn from ACF's stated areas of interest: (1) a Family Service Need model that predicts the services families need, including identification of candidates for foster care under the Agency’s Title IV-E Prevention Program; (2) Placement Stability models that align children's strengths and needs with family characteristics and environmental factors to improve first-time-best placement match and foster home retention; and (3) Maltreatment Risk models that support consistent, accurate hotline intake decisions, in coordination with professional judgment, clinical overrides, and human decision-making authority. The project will build sustainable in-house analytics capacity within CFSA’s Performance Accountability and Quality Improvement Administration (PAQIA), implement a rigorous model and performance evaluation, and disseminate replication toolkits, policy briefs, and peer-reviewed findings nationally. Expected outcomes include a targeted 3‑percentage‑point (pp) improvement in CFSA’s A Home for Every Child's 1:1 ratio over three years, a reduction in placement disruptions and in the recurrence of maltreatment, faster connection of families to prevention services, and a documented, scalable, CCWIS-embedded model architecture.
