Child abuse in Georgia: a method to evaluate risk factors and reporting bias.

From July 1975 through December 1979, the Georgia Department of Protective Services Central Registry recorded population-based data on confirmed, non-confirmable, and ruled-out child abuse reports. We propose that reporting biases are reflected in the differential characteristics of confirmed and ruled-out reports of child abuse. Characteristics, households, or groups equally or more prevalent in the latter category cannot necessarily be considered associated with increased risk of child abuse, even if they are represented in the confirmed abuse population more than in the general public. Important examples of factors or categories for which an association was suggested in other studies but which are not supported by this analysis are: urban residence, teenage motherhood, infancy, and mothers and other female perpetrators. These are conditions or categories associated with greater surveillance; therefore, risk assessment is not possible using the data in this registry. Certain households do appear to be at increased risk for child abuse. These include large families, families without a biological mother or biological father, and families ever needing Aid to Families with Dependent Children (AFDC). This increased risk, however, is slight. Improved surveillance requires identification of reporting biases. A comparison of confirmed and ruled-out reports is an inexpensive and system-specific step toward achieving this aim.