APHSA Annual Report 2018

The prototype is a high performing model using machine learning methods that when compared to current call screening practice, shows significant potential to improve decision-making by screening in more high-risk referrals to direct preventive resources earlier and prevent future maltreatment as well as screen out more low-risk referrals to prevent undesirable intrusion into family’s lives. The screening score will predict the likelihood of placement/court action/re-referral within one year and two years of the original call. The higher the score, the greater the likelihood. The development of a prototype model was led by the Auckland University of Technology (AUT) with research partners at the University of Southern California and Carnegie Mellon University. AUT used two years of Colorado child welfare data to help build a prototype that identifies children that are at risk of being removed from their home. The prototype was found to show good levels of accuracy in identifying children who are at the highest and lowest risk of placement and has created the opportunity for leadership to refine screening protocols and direct efforts to children and families most in need of services. Douglas County has made clear that this supportive decision-making tool is just that—a tool. It will not replace human review and assessment by child welfare professionals. Instead, it will speed up the current manual process of reviewing data from various systems/ cases to assess a given referral. The next stage in the project is to implement the predictive model in the Douglas County IT system using current individual data. The final stage is an evaluation—test the effectiveness of the model by using a Random Control Trial. STATE OF VIRGINIA In 2016, the Virginia Department of Social Services (VDSS) convened a workgroup of approximately fifty stakeholders from

across Virginia to evaluate the child care needs of specific target populations (children in underserved areas, infants and toddlers, children with special needs and children in need of care during nontraditional hours).The workgroup identified potential strategies to increase the supply and quality of child care, shared ideas and formed collaborations to implement selected strategies. The workgroup formulated thirty-two recommendations addressing these areas. As a result of the needs assessment, VDSS initiated Child Care Provider STEPS (Shared Training, Education and Professional Services), a pilot project to establish shared services networks for family day homes in the Western, Piedmont and Eastern Regions of the state. The networks will focus on improving the capacity and sustainability of family, day and home child care providers to increase the availability of quality child care for infants and toddlers, children with disabilities and children in need of care during non-traditional hours. VDSS has contracted with a state higher education institution to operate a network “hub” in each region. Specific services offered through the “hubs” will be determined by the participating family day homes and may address such areas as business operations (billing, accounting, etc.); special use permits; connections to discount pricing for materials/services; marketing; assistance with child care licensing, Child Care Subsidy Program, Virginia Quality and Child and Adult Care Food Program participation; and improving competencies through professional development. Connections to existing services and support organizations (such as Virginia’s Infant and Toddler Specialist Network) are a critical part of the strategy. Through participation in STEPS, providers become stronger, more efficient, more financially sound and better positioned to deliver higher-quality child care services.

8