Impact of intervention surveillance bias on analyses of child welfare report outcomes

Abstract

Two studies using data from child maltreatment intervention outcome evaluations were conducted examining several aspects of surveillance bias, including directly measuring rates of surveillance reporting, comparing characteristics of surveillance versus nonsurveillance reports, examining differences across service types and doses, and testing how ignoring versus removing surveillance reports in the data affects survival analysis. The net effect of surveillance bias was small in the studies examined. Home-visiting services were not biased more than center-based services, and bias was not greater among intervention compared to prevention cases. Surveillance reports were equally as severe as nonsurveillance reports, failing to support the hypothesis that surveillance serves as early detection of less severe maltreatment. However, surveillance bias was far more substantial during time periods when participants were actively engaged in services. Therefore, the net impact of surveillance could vary with service engagement rates and the relative duration of service engagement and postservice follow-up times.