Substantiation and recidivism.

This article reports rates of recidivism among initially substantiated and initially unsubstantiated child maltreatment events to determine if substantiation status is associated with higher risk of recidivism. This is an important question given recent concerns that unsubstantiated cases may have as high or almost as high a risk of recidivism as do substantiated cases. The data are analyzed at both the victim level and the case level, divided by type of maltreatment, and followed for 4.5 years. The data used are administrative and combine a series of state databases with census data. Analyses are performed at the bivariate and multivariate (Cox proportional hazards model) levels. The main finding is that unsubstantiated cases are at high risk for recidivism, in many cases as high a risk as substantiated cases. Implications for practice, policy, and research are presented with a focus on the importance of providing preventative services to unsubstantiated cases.