Predicting Instability at Home and in Foster Care, Challenges and Opportunities

Publication Date

2-19-2026

Document Type

Article

Publication Title

Lecture Notes in Operations Research

Volume

Part F1488

DOI

10.1007/978-3-032-13116-4_2

First Page

15

Last Page

25

Abstract

The primary goal of foster care systems is to help foster children achieve the permanency stage through reunification with their family, adoption, or another suitable arrangement in the shortest possible time. Accomplishing this goal is challenging, both in terms of finding the best permanent option for foster children and providing them with a healthy, safe, and stable environment during their temporary foster care episode. Sometimes permanency planning decisions do not work as intended, leading to additional removals from home. Moreover, some children experience multiple placement settings during their out-of-home care. This paper provides an overview of the foster care ecosystem and its challenges, and how digital transformation can improve system performance to benefit its stakeholders and society. This paper also presents prediction models to examine factors associated with foster children’s high number of removals from home, as well high number of placement settings during their current foster care episode. The analysis indicates that the child’s age, race, ethnicity, clinical diagnoses, history of adoption, adoption age, circumstances associated with the child’s removal, caretaker family structure, and location based on state and rural-urban category have a relationship with the child’s placement instability in and out of care.

Funding Sponsor

U.S. Children's Bureau

Keywords

child welfare, Foster care, machine learning, placement instability

Department

Industrial and Systems Engineering

Link to Full Text

Share

COinS