Methods. Data were collected from 88 counties in Ohio over 12 years (2004 to 2015) for a total sample size of 1056 space time units (n = 88 * 12). Data on child abuse and neglect were obtained by the National Child Abuse and Neglect Data System (NCANDS). Rates of naloxone/narcan administration (as a proxy for overdose) were obtained Ohio Emergency Medical Services Incident Reporting System and geocoded to point locations. We included a covariate for county type (rural, urban, suburban, and Appalachian) to assess whether or not changes in maltreatment due to overdoses disproportionately affected rural and Appalachian areas. We controlled for county-level sociodemographic information, amount of child welfare spending, and the presence of a tax levy designed to provide additional funding for child welfare services. Data were analyzed using Bayesian space-time models to adjust for the temporal spatial structure of the data.
Results. Current year rates of naloxone administration was not related to current year rates of child substantiated maltreatment. However, having higher rates of naloxone administration in last year was related to more substantiations (per referrals). We also found county type moderated the relationship between current year naloxone administration and child maltreatment. Higher rates of naloxone administration in rural and Appalachian counties were related to lower substantiated rates of child maltreatment.
Implications. Our study does not find that these overdoses are related to child maltreatment rates in the current year. We do find a year lag between rates of naloxone administration and child maltreatment. This lag presents an ideal opportunity for secondary prevention efforts designed to support families where drug misuse may be occurring. For example, this greater understanding of the spatial patterning of drug markets and rates of maltreatment might serve as an early detection system for child welfare practitioners, similar to those systems developed for infectious disease control. To the extent that drug distribution systems are related to or are a precursor of child maltreatment, knowing the locations of these activities may provide an opportunity to implement prevention activities in populations prone to drug abuse or dependence in an attempt to reduce future maltreatment when drug markets begin to develop.