Emerging, Hot Topic: The Prevention and Intervention Growth Strategy Project: Leveraging Big Data to Grow Strategic Investments in Child Maltreatment Prevention in Texas

Introduction: During its 85^th Legislative Session, the Texas Legislature required the Prevention and Early Intervention Division of the Texas Department of Family and Protective Services (PEI) to identify strategies and goals for increasing the number of families receiving prevention and intervention services each year. To meet this requirement, PEI partnered with UT Population Health to develop a data-driven methodology for identifying geo-spatial risk factors related to child maltreatment in Texas at the zip code level. The resulting project was released in January 2019, and has already been leveraged by decision-makers within and outside PEI to determine areas for future investment and growth of Texas’ child maltreatment prevention programs.

Methods: The Growth Strategy Project leverages cross-sectional data from several sources, including Census data and data from several Texas agencies. The project develops community profiles of risk associated with child maltreatment for different age groups; infants, 1 to 4 year olds, 5 to 9 year olds, 10 to 14 year olds, and 15 to 17 year olds at the Zip Code Tabulation Area (ZCTA) level. Specifically, an aspatial clustering method (K-Means) was utilized to determine which zip codes exhibited similar age-specific patterns of maltreatment exposure relative to other zip codes, used in conjunction with hotspot analysis.

Results: Several clusters of community-level risk factors were associated with child maltreatment risk at the zip code-level in Texas. These risk factors included: Families in Poverty, Health and Disability, Low-Income, Child Safety & Health, Low-Education, Infant Mortality Rates, Assaults Needing Medical Attention, and School Enrollment. While several of these factors have been linked to child maltreatment through conceptual and theoretical models of risk, the actual indicators that load onto these factors often differed from the indicators included in past needs assessments. Through the use of a big data approach to identify community level indicators, PEI and UT Population Health have shed light on new forms of risk and need that can drive investments in communities that may have previously been ignored.

Conclusions: Through the use of state and academic partnership and the use of big data, PEI and UT Population Health have developed a tool that can be used in communities across Texas to identify risks and needs for investments in child maltreatment prevention programs targeting distinct age groups. Through successful communication of methodologically rigorous analysis, the Growth Strategy Project bridges the gap between research and practice to guide strategic investment of resources in the prevention of child maltreatment.