Abstract
Social Determinant of Health (SDoH) Indices, such as the Social Vulnerability Index (SVI) or the Area Deprivation Index (ADI), are increasingly used in child maltreatment research either to explore the degree which SDoH is associated with maltreatment or as a control variable to more accurately understand other associations of interest. To date, however, there have been no published explorations of which SDoH measures are most predictive of child maltreatment outcomes. This paper compares a range of publicly available SDoH measures in terms of their ability to explain variability in rates in National Child Abuse and Neglect Data System (NCANDS) data. We examine correlations between SDoH measures and a range of child welfare outcomes (report rate, substantiation rate, foster care rate) as well as types of maltreatment reports (e.g. neglect, physical, sexual). We find some SDoH measures and median household income have strong correlations with NCANDS data (e.g. the ADI) while others (e.g. the BRIC) are much weaker. Associations tend to be consistent across both outcomes and maltreatment types. These associations were stronger when within-state clustering was applied. Our findings highlight the importance of accounting for state-level clustering and evaluating the predictive performance of SDoH indices when using NCANDS data.