Detection of Behavioral Health Cases from Sensitive Police Officer Narratives

Early detection of and intervention in behavioral health cases, including mental health, is crucial to prevent harm to one’s self and others. Police reports generated by officers on duty or in response to 911 calls remain an untapped resource for identifying such incidents. To expedite the detection process, we propose a workflow that involves collaboration between experts to manually annotate cases and correct model predictions. This approach can improve both initial annotation and model performance. Therefore, we advocate for the incorporation of manual annotations from experts, natural language processing (NLP), active learning, and advanced machine learning techniques to detect behavioral health cases within police reports. The experimentation suggests that a CNN-LSTM model achieves the best performance with an accuracy of 86.67% and an F1-score of 0.82 in detecting behavioral health issues.