New tools aim to improve risk assessment in psychiatry
John Wallert’s research aims to develop tools that can be used in psychiatry to predict the risk of serious incidents, so-called prediction models.

Text: Annika Lund, first published in Medicinsk Vetenskap nr 2 2026
“With better decision support, we could work less reactively and in a more tailored way. We are developing a tool to assess risks when discharging people who have been subject to compulsory care. Suicide is one such risk, acute readmission another,” he says.
The research is based on register data from more than 70,000 people who have received compulsory psychiatric care. This makes it possible to compare risk factors in relation to different outcomes.
The aim is to create a simple tool where the clinician in charge can input variables such as age, diagnosis, clinical history and living situation, and obtain risk assessments to support decisions on continued care and treatment.
“We know that people’s ability to assess risk is limited and subjective. A tool like this would provide precise and consistent support for clinicians’ decision-making. A person at high risk of a particular outcome, such as a suicide attempt, should be managed differently from someone at lower risk. There is room for improvement here,” he says.
His team is also developing a tool that estimates the likelihood of a patient benefiting from CBT for depression and anxiety. Here too, the researchers are drawing on extensive registry data, as well as genetic analyses carried out with AI support.
“The aim is to be able to assess the likely outcome even before treatment begins. If CBT is not deemed sufficient, for example, medication might need to be added from the start.”
About John Wallert
Occupation: Psychologist and Associate Professor at the Department of Clinical Neuroscience.
Research focus: How to use statistical and AI-based models to understand and predict adverse outcomes in psychiatry – and thereby improve care and treatment.
