The sample consisted of 805 children presenting at Children's Hospital of Oklahoma (CHO), a large urban children's hospital, affiliated with the University of Oklahoma Health Sciences Center and serving children from across Oklahoma. All 805 children in this study were identified as victims of abuse and neglect. The case records of these children were analyzed for the effects of several independent variables (family histories of domestic violence, alcohol and substance abuse, prior Child Protective Services involvement, child age, type of case, and different referral versus reporting sources) in correctly classifying cases confirmed as failure to protect (FTP, n = 201) on the part of a non-perpetrating parent vs. cases in which FTP was ruled out (n = 604). Logistic regression analysis yielded a model that correctly classified 37.88% of FTP cases and 92.96% of non-FTP cases. The variables of "family history of violence", "sexual abuse case" and "difference between disclosure and referral source" were found to contribute significantly to the model. Analysis on demographic and case outcome variables was also conducted. Limitations of the study are answered, as well as implications for further research. Findings yielded a model that can assist Child Protective Services workers and other professionals in identifying at-risk cases for FTP.