Identification and Case-Ascertainment of Self-Harm in Electronic Health Records (EHR): Development of Machine Learning-Led Tool Using CRIS Data

Self-harm is the strongest risk factor for suicide and is associated with many other adverse outcomes. It is an important outcome for secondary mental health care (specialised services provided by medical professionals, such as psychiatrists or clinical psychologists).

This project aims to improve the identification of self-harm cases within Electronic Health Records (EHR) using advanced technology. EHR contains valuable data about healthcare visits, but most of it is stored in a complex, unstructured format. By employing artificial intelligence and natural language processing, we're creating a tool to automatically identify and organise self-harm information from these records. This will contribute to identifying self-harm cases more consistently, evaluating interventions to reduce self-harm, reducing health inequalities, and potentially lowering the economic burden of self-harm on the NHS and the wider economy. The work will benefit individuals with mental health conditions and improve healthcare delivery nationally.