Department Seminar Series | PsyMetRiC: From Clinical Prediction Model to Regulatory Certified Medical Device and Beyond

Clinical prediction models are used routinely in the general population to support primary prevention of adverse cardiometabolic outcomes. Yet, a wide body of evidence has shown that existing population-based tools are inadequate for people with severe mental illness (SMI), particularly younger populations with early SMI where elevated cardiometabolic risk is already detectable. Moreover, despite a huge growth of research activity in clinical prediction models in psychiatry over recent decades, there is a dearth of translation into clinically available tools. 

In this talk, Ben will present the latest work from the psychosis metabolic risk calculator (PsyMetRiC) project. First developed in 2021, PsyMetRiC was originally developed to predict up to six-year risk of developing metabolic syndrome in young people with psychosis spectrum disorders. The original PsyMetRiC models have subsequently been validated in  11 countries across four continents. Now, as part of an NIHR Advanced Fellowship, the PsyMetRiC project has been transformed with multi-disciplinary components spanning data science and the inclusion of over 25,000 early psychosis cases to refine the PsyMetRiC models; health economic modelling; qualitative components focused on risk communication and behaviour change; comprehensive stakeholder engagement and involvement including training lived experience co-researchers; software design emphasizing  comprehensive graphical and animated results visualisations; and collaboration with regulatory medical device experts. The resulting work, due for upcoming publication in LancetPsychiatry, will coincide with the release of the new PsyMetRiC web application, which is among the first regulatory certified prediction tools in the field of psychiatry, thus one of the first in the field to be available for routine clinical use. PsyMetRiC can support a shift in cardiometabolic care for young people with SMI from reactive management - associated with persistenly poor outcomes - to earlier proactive prevention supported by shared decision-making. This work also provides a translational template for moving prediction models from statistical equations to regulated medical devices that can be trusted by clinicians and patients. 

This seminar is hosted online only:

• https://zoom.us/j/93311812405?pwd=9kbjSbEcO2fa7n7gFLZVqrChvr467B.1

• Meeting ID: 933 1181 2405
• Passcode: 169396