The 130th Annual Meeting of APHA

Abstract #49565

Lisa Sanbonmatsu, PhD, NBER, National Bureau of Economic Research, 1050 Massachusetts Avenue, Cambridge, MA 02138, 617-613-1201, lsanbonm@nber.org

PURPOSE: The purpose of this study is to estimate trends in fatal injuries associated with child abuse and neglect. Changes over time in awareness of child maltreatment and in investigation and reporting practices present challenges for identifying these trends.

METHODS: I use Multiple Cause of Death Data for 1990 through 1998 to develop proxies for neglect and abuse fatalities involving children under age 5. Since maltreatment deaths are thought to be under ascertained and may not be consistently identified over time, I rely on injury codes and the type of incident rather than on the underlying cause of death. The proxy for abuse deaths was based on injuries associated with non-weapon homicides (but not necessarily requiring autopsy). The neglect deaths were identified based on externally caused deaths excluding abuse proxy deaths and categories of deaths not likely to involve a supervisory component. I also excluded categories of deaths that might in the past have been mistaken for Sudden Infant Death Syndrome and could thus bias trends.

RESULTS: Proxies for battered child deaths/homicides (n=6,793) and child neglect deaths (n=18,901) were constructed. Both show declines over time. Differences between trends using the proxies and trends using underlying cause codes are consistent with the hypothesized biases.

CONCLUSIONS: The proxies developed provide an alternative for estimating trends in fatal abuse and neglect that is less susceptible to changes in awareness and in procedures. However, this method needs to be validated using individual case data.

Learning Objectives: