Aligning a Household-Level Service Array Through a Jurisdiction-Wide Child Maltreatment Prevention Effort: Protocol...

Area of Science:

• Public Health
• Geospatial Analysis
• Machine Learning

Background:

• Child maltreatment leads to severe individual and societal negative outcomes.
• Children exposed to maltreatment face increased risks of psychological disorders, substance use, delinquency, suicidality, and educational problems.

Purpose of the Study:

• To prevent and ameliorate child maltreatment through optimized service allocation.
• To identify high-risk areas using geospatial smoothing and risk terrain modeling (RTM).
• To determine optimal service reallocation using counterfactual modeling for improved outcomes.

Main Methods:

• A 3-stage iterative process combining RTM and counterfactual modeling.
• RTM uses machine learning to map maltreatment risk based on historical data and built environment factors.
• Counterfactual modeling optimizes service allocation at the household level, considering constraints and fairness.

Main Results:

• A participating jurisdiction is currently being recruited for study implementation by September 2025.
• The methodology allows for flexible application across different jurisdictions.

Conclusions:

• This novel approach optimizes supportive service allocation for families at risk of child maltreatment.
• It integrates data-informed strategies with outcomes-based counterfactual modeling for enhanced child welfare.
• The flexible methodology can be adapted to various jurisdictional preferences and constraints.