New study: PRIMED

In collaboration with researchers at Linnéuniversitet, Royal Institute of Technology and University of North Carolina at Chapel Hill we are planning a new study called PRIMED: Predicting Response to CBT in Mental Disorders using multimodal data and machine learning.

Common mental disorders such as depression, social anxiety, panic disorder and OCD have a point prevalence of 15%. CBT is the treatment of choice, yet 30-60% of patients undergoing CBT will not respond to treatment. Identifying non-responders before treatment would allow alternative treatment choices.

Predictor studies have so far not been able to reach an acceptable predictive power to guide clinical decisions so adding fine-grained and multiple types data, including the genetic footprint, may be a way forward.

Hence, the aim of PRIMED is to better predict treatment response to Cognitive behavioral therapy (CBT) in depression, social anxiety disorder, panic disorder, and OCD by using clinical, register-based, and genetic data from 6000 individuals in Sweden.

In PRIMED, potential predictors include clinical information, register based data and genetic variation. Outcomes are short-term clinical data, long-term register based medical data e.g. prescriptions, diagnoses, suicide attempts and social data such as unemployment, sickness absence, and disability pension.

icon primed transparent

In addition to study predictors on these different levels, we also want to add artificial intelligence, or Machine Learning to identify individuals at risk of not responding to treatment. So-called ‘Learning Machines’ can learn from historical cases and then apply what it has learnt to predict outcomes in a single, new case. Rather than considering the effect of one variable on an outcome of interest at a group level, Learning Machines identify patterns of information that can be used to predict the outcome for an individual.

The Learning Machine will learn from a training dataset consisting of 4 000 individuals and then validate its ability to correctly identify patients at risk on 2 000 new patients.

This project pursues a novel research area that can only be developed by a multidisciplinary team involving experts in disparate fields: psychiatry, psychology, epidemiology, AI, and genetics. We have assembled a strong interdisciplinary team from four different universities in two countries and are very excited about this collaboration!