Using algorithms to prevent suicide

Her motivation to take up research is rooted in the conviction that there is great untapped potential in using data and machine learning in psychiatry. Dimitra Rentouka is a data warehouse architect and business intelligence strategist. She divides her time between Region Uppsala and Uppsala University.

At the Region she works in Beslutsstöd, a decision-support unit that provides quality-assured regional data for various kinds of follow-up and development. Since the spring she has been a doctoral student at the Uppsala Centre for Preventive Psychiatry. Dimitra sees a great advantage in having a foot in both camps.

“My two roles complement each other. In my research I always think about how the results can be used in everyday clinical practice, and in the regional setting about how research can help us develop our services further. Collaboration is a key component for success, not least when it comes to implementation, she says.

Developing risk assessment tools

Dimitra's research concerns the use of algorithms for suicide prevention. In her first study she will validate data from the risk assessment tool Oxford Mental Illness and Suicide Tool (OxMIS), developed by Seena Fazel, Professor of Psychiatry at the University of Oxford.

OxMIS is a web-based tool for predicting suicide risk in people with schizophrenia spectrum disorders or bipolar disorder. Dimitra will validate the data in a newly developed version of OxMIS and then use the same dataset to develop new algorithms.

Swedish datasets – a goldmine

The validation will draw on electronic health record data and register data collected from four Swedish regions between 1997 and 2025. Access to datasets of this kind requires ethical approval.

“The most interesting thing about the data is how extensive and high in quality it is – a real goldmine, in fact. It offers great scope for building different algorithms, examining which factors matter most for risk assessment and analysing different outcomes,” she says.

The first studies focus specifically on people with schizophrenia spectrum disorders and bipolar disorder respectively. The idea is for the doctoral project to take a holistic view of suicide prevention, from the design and development of algorithms to validation, implementation and acceptance.

“For that to become reality, further research is needed into the barriers and facilitators for implementation. This concerns, for example, how and when the information should be presented, who receives it and who acts on what. In the final part of the doctoral project I will focus on processes and organisation. That completes the circle,” says Dimitra Rentouka.

Prevention is the key word

The Uppsala Centre for Preventive Psychiatry is a newly established interdisciplinary centre. It is a setting that can generate a range of synergies, Dimitra believes.

“Prevention really is the key word for me. With suicide, there is no treating it after the fact. The only way to act is to be able to proactively identify the people who are at high risk. Because my project is part of the Centre's work, I will have the opportunity to draw on more perspectives and on other people's knowledge and experience, and to enrich my work on suicide prevention. Hopefully I will also be able to contribute my own experience to the work of others – that is something I very much look forward to.”

Maria Bergenheim