Psychosis Metabolic Risk Calculator may Predict Cardiometabolic Risk in Young Adults with Psychosis: Study

This retrospective, multicohort clinical prediction study utilized large-scale primary care datasets, including the Clinical Practice Research Datalink and QResearch, along with secondary care data from the South London and Maudsley NHS Foundation Trust. Participants were aged 16–35 years at first diagnosis of a psychosis-spectrum disorder between January 1, 2005, and December 31, 2015, with follow-up extending to December 31, 2020, in primary care cohorts. Secondary care participants were enrolled between January 1, 2012, and December 31, 2024.

The models developed by PsyMetRiC were revised and updated to increase the accuracy of prediction with the addition of more relevant variables, namely family history of cardiometabolic disease, antidepressant use, systolic blood pressure, and HbA1c levels. Three key outcomes were also modeled, namely, the development of metabolic syndrome in 1 to 6 years using logistic regression, type 2 diabetes in 10 years using Weibull regression, and significant weight gain in 1 year using logistic regression. Full models with biochemical predictors and partial models without biochemical predictors were developed to increase the applicability of the models in different settings.

Key findings:

• The participants were 25,850 in number, comprising 13,614 males (52.7%) and 12,236 females (47.3%).

• Their mean age was 26.7 ± 5.4 years. In terms of ethnicity, 16,445 participants (63.6%) were White European ethnicity, while 9,405 participants (36.3%) were of Black African/Caribbean, South Asian, mixed, East Asian, or other ethnicity.

• The results obtained using the PsyMetRiC2 model were excellent in terms of discriminative ability.

• In terms of the model for metabolic syndrome, external validation gave a C-statistic of 0.81 (95% CI, 0.77–0.84) for the full model and 0.79 (95% CI, 0.76–0.83) for the partial model.

• In terms of type 2 diabetes prediction, internal validation gave a C-statistic of 0.86 (95% CI, 0.76–0.95) for the full model.

• External validation gave a C-statistic of 0.81 (95% CI, 0.71–0.88) for the full model.

• In terms of clinically significant weight gain, internal validation gave a C-statistic of 0.78 (95% CI, 0.73–0.82) for the full model and 0.77 (95% CI, 0.72–0.80) for the partial model.

• Calibration plots were acceptable across all models, and decision curve analyses indicated clinical usefulness at all plausible risk thresholds.

The PsyMetRiC models facilitate the accurate and clinically useful prediction of cardiometabolic risk in young people with psychosis and promote the development of preventive, personalized, and integrated care.

Reference:

Perry, B. I., Osimo, E. F., Si, S., Hitchins, K. V. B., Lewis, C., Laws, B., Griffin, S. J., Khandaker, G. M., Murray, G. K., Shiers, D., Chew-Graham, C. A., Jones, P. B., Denniston, A. K., Bardus, M., Jowett, S., Walsh, A. E. L., Arshad, S., Formanek, T., Pillinger, T., McCutcheon, R. A., … PsyMetRiC Network (2026). Cardiometabolic prediction models for young people with psychosis spectrum disorders in the UK (PsyMetRiC 2.0): a retrospective, multicohort clinical prediction model study. The lancet. Psychiatry, 13(4), 291–303. https://doi.org/10.1016/S2215-0366(25)00398-0