Child safety measure as a proxy for child maltreatment: Preliminary evidence for the potential and validity of using ICD-9 coded hospital discharge data at the community level

Objective: The purposes of the study, carried out in the context of Strong Communities for Children in the Golden Strip (Strong Communities) in South Carolina, were to identify a set of ICD-9 codes (The International Classification of Diseases, Ninth Revision, which represents the standard nomenclature for classifying mortality and morbidity data), to establish base rates for ICD-9 coded child safety in a community measure and to undertake a preliminary analysis of the validity of ICD-9 coded injuries as a measure for child safety at the neighborhood level. Method: The study drew on data from two sources. Data on child injuries were extracted from the linked agency administrative data base maintained by the South Carolina Budget and Control Board. Data on neighborhood distress were derived from census data in the census block group level. An initial pool of ICD-9 codes for children's injuries was provided for the 2002-2003 and 2003-2004 reporting periods. An expert panel consisting of five physicians was convened to review the ICD-9 codes to assess their relationship to child safety, and to collectively decide on the codes to be used as the measure and proxy for child maltreatment. Results: The findings offer a glance to the utility and potential of using ICD-9 coded injuries as a measure of child safety. The results present a final list of 323 codes of the ICD-9 codes (230 codes of child physical safety, 57 codes of child sexual abuse, and 36 codes of child safety not otherwise classified). Moderate to strong correlations were found between the suggestive codes and the Neighborhood Distress Index. Further, these data served well to establish a baseline to monitor the prevalence of child safety and injuries. Conclusions: The finding offers preliminary support for the validity of ICD-9 coded child injuries as a measure of child safety. There is great potential in this approach for monitoring the prevalence of injuries of children as well as children safety in the community level.