A web app now predicts diabetes and obesity risk for young people with psychosis

University of Birmingham

People with severe mental illness die, on average, 15 years younger than the general population. The leading causes are not psychiatric. They are cardiovascular disease, diabetes, and obesity: conditions that are largely preventable if caught early enough. But the prediction tools that GPs use to assess cardiovascular risk were designed for middle-aged adults, not twenty-somethings with psychosis. They do not work accurately in this population.

The result is a detection gap. The very patients who face the highest cardiometabolic risk are the least likely to be identified and offered early intervention.

PsyMetRiC: built for the patients who need it most

A new web-based tool called PsyMetRiC aims to close that gap. Developed by Benjamin Perry, Associate Clinical Professor of Psychiatry at the University of Birmingham, and published in The Lancet Psychiatry in March 2026, PsyMetRiC predicts three specific outcomes for young people with psychosis: clinically significant weight gain within one year, metabolic syndrome within six years, and type 2 diabetes within ten years.

The algorithms were developed and tested using routine anonymized health data from over 25,000 young people with psychosis in the UK, followed for more than 20 years. The tool requires only simple, routinely recorded clinical information to make predictions, nothing exotic or hard to obtain.

From metabolic syndrome to outcomes patients understand

An earlier version of PsyMetRiC, published in 2021, predicted metabolic syndrome, a cluster of changes including high blood pressure, weight gain, high cholesterol, and elevated blood sugars. Clinician and patient feedback indicated a problem: many clinicians were not familiar with what metabolic syndrome means, and patients found it even less intuitive.

The updated PsyMetRiC2 predicts outcomes that both groups understand: weight gain, diabetes, and metabolic syndrome. The shift was driven by focus groups with clinicians, carers, and young people with lived experience of psychosis, organized through The McPin Foundation in collaboration with The Centre for Mental Health and Equally Well.

This patient-centered design extends to how results are communicated. The app presents risk information in multiple graphical and numeric formats, including an adaptation of the "heart age" score approach that has shown promise in motivating behavioral change like reducing smoking or improving diet.

Testing for fairness across populations

Perry's team took deliberate steps to test whether PsyMetRiC works equitably for people from different backgrounds. Drawing on the work of the STANDING Together collaboration at Birmingham, which highlighted how health datasets can inherit societal biases, the researchers evaluated the tool's performance across demographic subgroups to ensure it does not systematically disadvantage underserved populations.

The tool has been validated in academic studies across Spain, Switzerland, Finland, the Netherlands, Canada, Hong Kong, and Australia, with funding recently received to test it in the United States.

MHRA certified for real clinical use

PsyMetRiC is certified by the UK's Medicines and Healthcare products Regulatory Agency (MHRA) as a Class 1 Medical Device. It is among the first prediction tools in psychiatry to achieve this status, meaning it can be used in actual clinical practice rather than remaining a research prototype.

Health professionals can sign up for a free account after confirming their status and agreeing to use terms. The tool is designed for both primary and secondary care settings.

A prompt for conversation, not a prescription

PsyMetRiC is explicitly designed to inform clinical decisions, not dictate them. It generates risk estimates intended to prompt conversations between clinicians and patients about preventive measures, lifestyle changes, and monitoring. The downloadable risk communication guide, co-produced with patients, aims to make these conversations more meaningful and accessible.

Perry described the goal as expanding the conversation between doctors and patients about physical health risks, encouraging shared decision-making about interventions before conditions become entrenched.

What PsyMetRiC cannot do

The tool predicts population-level risk, not individual outcomes. A high-risk score means a patient belongs to a group with elevated probability of developing cardiometabolic conditions, but individual trajectories will vary.

The algorithms are based on UK data. While international validation studies have been encouraging, healthcare systems, dietary patterns, and population genetics differ across countries, and performance may vary in settings that differ substantially from the training data.

PsyMetRiC currently addresses cardiometabolic risk only. Other physical health risks in psychosis, including respiratory disease and some cancers, are not covered.

The tool does not itself deliver interventions. Identifying high-risk patients is only useful if healthcare systems follow through with appropriate preventive care, and the resource constraints that contribute to the current detection gap will not be solved by a prediction tool alone.

Impact evaluation studies, health economic analyses, and qualitative assessments of clinical uptake are planned. The tool will receive iterative updates shaped by new research and user feedback.