Personalised medicine with AI; strategic in improving mental health

Personalised medicine is transforming mental health with more accurate, effective and human responses. Four groundbreaking projects - FarmaPRED, PERMEPSY, PERMANENS and deep brain stimulation (DBS) - which were presented today at the Mental health research: personalised medicine for the future conference at Parc Sanitari Sant Joan de Déu demonstrate that bespoke medicine in mental health is already saving lives. Featuring algorithms which predict the response to antipsychotics, training that tailors psychological treatment to each patient with a psychotic disorder, systems which anticipate the risk of self-harm and individualised neuromodulation techniques, all of them show that integrating advanced technology with clinical models and active user participation are a paradigm shift in mental health.

This was the centrepiece of the fourth event in Mental Health Week 2025 at Sant Joan de Déu Health Park which runs from 6 to 10 October.

Collaboration, progress in data technology

Three of the projects are based on a collaborative database platform which helps to generate joint knowledge by aggregating information on a large scale to enhance treatments and deliver bespoke care to the public.

In data collection, experts urge factoring in not just biomarkers but also the whole human spectrum (such as gender, environment and life experiences) and for this information to be included in artificial intelligence tools. The active involvement of both patients and professionals needs to be ensured.

Transforming the approach

The projects presented show how technology, clinical science and patient input are coming together to redefine care. A prominent example is FarmaPRED, which uses machine learning algorithms to combine genetic, clinical and environmental data to predict the response to antipsychotics in people experiencing a first psychotic episode. This tool makes it possible to anticipate side effects and optimise prescriptions, thereby addressing a critical issue: variability in response to treatment which affects up to 30% of patients. It can also prevent unnecessary exposure to ineffective drugs and enhance confidence in the treatment. The idea is to "create a tool which can predict clinical outcomes and forecast whether a specific treatment will be effective or may have side effects,” said Dr Sergi Mas, a researcher at Hospital Clínic Barcelona.

In lockstep, the European PERMEPSY project led by Parc Sanitari Sant Joan de Déu is developing a digital platform to personalise psychological therapy through Metacognitive Training (MCT), which adapts to the cognitive and emotional characteristics of each patient. This tool draws on data from numerous clinical trials and machine learning techniques to predict therapeutic response. From a clinical standpoint, it provides a validated alternative with positive effects on delusions, self-esteem and cognitive insight while its group- and manual-based application makes it easy to roll out in a variety of settings. In terms of patient experience, it stands out for subjective satisfaction and improved adherence owing to personalisation. Dr Susana Ochoa, coordinator of the Research Unit at Parc Sanitari SJD, pointed out that “the project taps artificial intelligence techniques to identify the profiles that benefit most from this approach and also includes a digital platform to tailor the treatment to each person." Featuring the participation of institutions from five countries, PERMEPSY is an innovative step towards more effective, empathetic and patient-centred mental health care. This intervention has been in place at Parc Sanitari Sant Joan de Déu for several years in most mental health services and has benefited over 300 people.

Also in prevention, PERMANENT harnesses predictive models to identify the risk of self-harm and suicide in emergency departments, working with medical records to provide alerts, safety plans and psychoeducational support. Its collaborative design, which includes patients, families and professionals, improves decision-making in highly complex emergencies and furnishes clinical tools that lessen subjectivity and step up healthcare safety. It also promotes empathy training and safeguards the emotional health of professionals. “The project's ambition is for this technology to help reduce clinical uncertainty, enhance empathy and prevent lives being lost to suicide," commented Dr Philippe Mortier from Hospital del Mar Research Institute. Usability testing will begin in November with 10 mental health professionals from various backgrounds and around 20 patients.

Finally, deep brain stimulation (DBS) is a therapeutic option for patients with severe, treatment-resistant obsessive-compulsive disorder (OCD). This approach includes advanced neuroimaging to individualise stimulation parameters, and even though it is an invasive technology, its evolution points towards more precise and safer personalisation. With a 60-70% response rate, DBS is effective in both the short and long term. “It is essential to have multidisciplinary teams working in specialist centres to ensure the best outcomes," argued Dr Pino Alonso, head of Psychiatry at Hospital Universitari de Bellvitge. It can offer real hope for patients as long as clear communication and ongoing emotional support are ensured.

Parc Sanitari SJD is a partner centre in the PharmaPred and PERMENS projects and leads the PERMEPSY project.

AI in mental health needs to be ethical, equitable and well-resourced

The panel participants emphasised the need to be cautious when using algorithms while acknowledging their potential. The consensus is that technology applied to mental health has to be designed and implemented ethically, equitably and to help people recover "following the community mental health model and keeping the person at the forefront," said Ángel Urbina, an industrial engineer, data scientist and mental health activist at the Salut Mental Catalunya Federation.

Finally, the experts agreed that these innovations need to be implemented in real-world environments with training, resources and equity, and their accessibility must be ensured.