Counties with Fewer Than 1,000 Child Maltreatment Reports: Enigmas and Opportunities

Background: Child maltreatment reporting and substantiation are associated with contextual conditions related to location, including rates of poverty, unemployment, and housing challenges, as well as access to resources such as mental health care (Freisthler, et al., 2006; Maguire-Jack, 2014). Researchers studying child maltreatment often rely on data from the National Child Abuse and Neglect Data System (NCANDS). For questions related to geospatial context, NCANDS includes state and county indicators, but county identifiers are shared publicly only for counties with more than 1,000 report units (i.e., a child on a report). At the child level, it is important to mask data that could jeopardize confidentiality, but many questions could be addressed with aggregate county-level data, posing no threat to confidentiality. This study addresses the question: How do counties identified in NCANDS differ from counties without county identifiers? That is, what are researchers and service providers missing by not having access to NCANDS county identifiers for predominantly rural counties?

Methods: The study used demographic, health, and services data from all US counties. County data were obtained from County Health Rankings and Roadmaps, which consolidated 2020 county-level data from a range of primary sources (e.g., US Census, National Center for Health Statistics). A 2020 indicator of whether county was identified in NCANDS was computed from the NCANDS 2022 child file. Aggregated county-level child maltreatment data were obtained from the National Data Archive on Child Abuse and Neglect for 2015 through a special request. To compare counties having and not having a county-level identifier, t-tests were conducted for a range of demographic, health, and services characteristics. Mean rates of child maltreatment reporting and substantiation were also compared.

Results: Of 3,163 US counties, 773 (24%) had an NCANDS county identifier in 2020. About 80% of children lived in counties with a county identifier, including about 74% of white children, 85% of Black children, and 90% of Latinx children. Unsurprisingly, most counties without identifiers were predominantly rural (>50% rural per Census Bureau). Among other differences, compared to counties with identifiers, counties without identifiers had statistically higher mean rates of child poverty, food insecurity, housing problems, teen births, child mortality, and infant mortality, and statistically lower rates of mental health care providers. Paradoxically, child maltreatment report and substantiation rates were statistically higher in rural counties, but counties without county identifiers had statistically lower report and substantiation rates.

Conclusions: Most children live in counties publicly identified in NCANDS, but these results show that identified counties differ substantially from predominantly rural counties that are not identified. On average, counties without identifiers have higher rates of the contextual challenges that place children at greater risk of child maltreatment and other poor outcomes. Community responses to child maltreatment also seem to differ in perplexing ways in counties without identifiers. Researchers and practitioners have much to learn about how rural contexts influence child maltreatment and responses to it. Access to aggregate data with county identifiers for all counties would facilitate investigation of the enigmatic nuances of child maltreatment in rural areas.