Aim: This prospective study explored whether self-report measures of parenting self-efficacy, mood, anxiety, stress, emotional regulation or substance use were useful in predicting subsequent notifications of harm to Child Protection Services (CPS). Methods: CPS-involved parents living in Christchurch, New Zealand were recruited to the study, all of whom retained the custody of their youngest child at the time of assessment. Comprehensive data regarding the aforementioned constructs, along with demographic data, were gathered during a baseline assessment. Notifi cations of harm were prospectively retrieved from the national child protection database. Methods of analysis included logistic regression and Receiver Operating Characteristics (ROC) analyses. Results: No individual psychological constructs, nor cumulative risk scores, were found to be associated with notifications of harm among the study sample (N = 26). The Area Under the Curve (AUC) values ranged from 0.49 to 0.65, indicating that these scores generally performed no better than chance. Two background factors predicted CPS notifications, namely: having more children and the parent reporting a personal history of neglect. Conclusion: Self-report measures of mood, emotional regulation, parenting self-efficacy and substance use are unlikely to be useful for PCA when conducted in the context of child protection casework or court assessments.