Precision psychiatry powered by real-world data
Mental disorders constitute one of the greatest public health challenges in Europe. Despite the associated healthcare costs, progress in treatment has been limited, particularly for individuals with severe and complex conditions. Many patients continue to experience complex care pathways, poor outcomes and reduced quality of life.
Unlocking the value of real-world data
The aim of the EU-funded REALMENT(opens in new window) project was to address this gap by harnessing the untapped potential of real-world data to support more effective and personalised treatment strategies in psychiatry. By integrating electronic health records, health registries and genomic data from biobanks across Europe, the project has created a unique data ecosystem. “By combining and harmonising data across countries, we capture a much broader range of biological and environmental variation. This allows us to train predictive models with the accuracy needed for precision psychiatry,” explains the project coordinator Ole Andreassen. This large-scale integration moves beyond traditional clinical trials, enabling researchers to analyse treatment outcomes across diverse populations and real-life conditions.
Artificial intelligence for precision psychiatry
The consortium applied artificial intelligence and machine learning to uncover patterns within vast datasets. By analysing population-scale data from cohorts spanning Nordic, Baltic and other European populations, researchers have developed predictive models capable of identifying treatment response, side effects and disease trajectories. “These models are trained on real-world data and validated against clinical trial datasets, providing proof of principle that our approach can support more personalised treatment decisions,” emphasises Andreassen.
Genomic insight and treatment response
A key innovation of the project is the integration of genomic data into predictive modelling. Through large-scale genome-wide association studies, researchers have identified genetic variants linked to treatment response and resistance. For example, the discovery of rare variants in pharmacogenes, such as CYP1A2(opens in new window) which affect clozapine metabolism, implies potential pathways that may govern how patients respond to antipsychotic medications. The project has also demonstrated the utility of polygenic risk scores(opens in new window) in predicting treatment trajectories. “Our goal is precision psychiatry: using your unique biology to get the right treatment, the first time,” highlights Andreassen. Project findings reveal shared genetic links between mental and physical conditions, supporting a more integrated view of patient health. This has important implications for managing multimorbidities and tailoring interventions to the broader biological profile of each patient.
From data to clinical application
To translate these advances into practice, REALMENT has developed a clinical management platform that integrates predictive algorithms with patient data. This platform supports clinicians in evaluating treatment options, anticipating adverse effects and monitoring patient outcomes over time. Importantly, the project has also contributed to methodological advances, including recommendations for defining treatment outcomes using real-world data(opens in new window) in psychiatric disorders across Europe. Moreover, one of the partners implemented pharmacogenetic reporting within a national biobank, where patient feedback has helped refine communication and usability. Looking ahead, efforts are underway to further develop and validate predictive models, expand the platform’s capabilities and support regulatory pathways for clinical adoption. “Ultimately, our goal is to move away from trial-and-error prescribing and provide clinicians with tools that enable the right treatment for the right patient at the right time,” concludes Andreassen.
