Theme: Implementation

Project Description

Clinical prediction models are increasingly used in mental health to identify people at elevated risk of adverse outcomes. To improve care, these tools must link to effective interventions. How to do this remains a key unresolved question across mental health services. This project addresses it using violence prevention in psychosis, which has significant implications for patients, victims, health services and society, and is a vital need for an important minority of people with psychosis. Forensic services that specialise in violence risk management are usually involved only after serious violence has occurred. Opportunities for preventing a first incident lie predominantly in general mental health services, which lack specific resources and expertise for managing violence risk. One innovation addressing this is OxMIV, a scalable risk assessment tool validated in Early Intervention in Psychosis services in England. It offered net benefit over clinical assessment alone, with a sensitivity of 71% compared with 40% for unstructured clinical judgement. It is practically feasible and acceptable to patients, carers and clinicians. However, no standardised, evidence-based intervention pathway links risk assessment to action. 

This project will develop a decision support tool linked to OxMIV, guiding clinicians to target interventions to those most likely to benefit. First, a realist review of interventions addressing modifiable risk factors for violence in psychosis will identify how these interventions work, for whom and in what contexts, and barriers to their effectiveness in practice. Second, the decision support tool will be codeveloped with patients, carers and clinicians through workshops and qualitative methods. Third, a mixed-methods feasibility study will test the risk assessment tool and intervention pathway together as a complex intervention in psychosis services, examining its integration into routine care and whether a definitive trial is feasible. Beyond violence, the project will produce a transferable approach to linking prediction and assessment to intervention.