Missouri Child Welfare Predictive Analytics Demonstration Project - The Missouri Department of Social Services Children’s Division (DSS/CD) is responsible for protecting children from abuse and neglect and promoting their safety, permanency, and well-being through collaboration with families, communities, and partner agencies. To better support this mission, address limitations of legacy systems, and alignment to federal Executive Orders and priorities, Missouri is implementing new modular technology. Through enhanced data sharing, real-time access to information, streamlined workflows, and improved decision support, Missouri aims to strengthen service coordination, reduce technical barriers, and enable data-driven decision-making advancing efficiencies, and the overall improvement of the program. Missouri has invested in a state-of-the-art analytics platform for use across the Department of Social Services. Leveraging this asset as the foundational analytics platform for the data modernization work, a set of use cases were evaluated based on the current child welfare modernization goals and includes the following use cases: • Placement Stability Management • Smart Case Summary Generator • Intelligent Service and Placement Mapping • Supervisory Oversight of Open Investigations • Intake & Hotline Screening • Risk-Based Staffing & Quality Assurance Targeting • Following an assessment of business value, implementation complexity, and expected outcomes, Placement Stability Management and Smart Case Summary Generator emerged as the highest-priority opportunities and are being targeted as the primary use cases that will be developed and implemented through the grant funds. Approval of this grant application will help Missouri advance its goal of placement stability management by enabling the State to leverage integrated data and decision-support capabilities to identify children at risk of placement disruptions earlier, allowing caseworkers to proactively intervene and improve permanency outcomes. This directly supports proactive intervention and improvement of child stability and permanency outcomes. The proposed solution will leverage existing data extracts, narrative analysis, and historical placement patterns to detect early indicators of placement stress and elevated risk of placement disruption before a crisis occurs. The solution analyzes case notes using natural language processing, incident reports, visitation logs, historical data, and behavioral documentation as part of the model definition. The predictive model will provide recommended action steps and calculates a placement disruption score that can be used as a data driven decision point for case management functions. It will also support a Smart Case Summary Generator by providing a centralized, interoperable data environment that can automatically synthesize key case information, reducing administrative burden, improving caseworker efficiency, and ensuring timely access to actionable information needed for informed decision-making. Combined these AI predictive models and tools will reduce crisis driven workload, enable more efficient resource allocation, reduce workforce burden, and improve consistency across caseworkers. Funding of this use cases are expected to deliver the greatest near-term return on investment by helping reduce placement disruptions through earlier identification of at-risk placements while also significantly reducing caseworker administrative burden through automated case summarization, enabling staff to spend more time supporting children and families and less time on documentation and information gathering. Based on preliminary calculations, funding of this project will be highly impactful. It is estimated to indicate a potential disruption reduction by 30% - 45% and time savings of 20% - 35% per caseworker.
