The latest publication from Rücklab is a new article first-authored by PhD student Olly Kravchenko, with Ralf Kuja-Halkola, Christian Rück, and John Wallert among the co-authors. The study was published in the Journal of Medical Internet Research.
The study investigated whether machine learning can help predict, before treatment begins, which patients are likely to experience clinically meaningful improvement from internet-delivered cognitive behavioural therapy (ICBT) for depression, panic disorder, and social anxiety disorder. Using data from 1,790 patients, the researchers combined clinical and sociodemographic information with data from national registers and genetics to develop and validate several predictive models.
Models incorporating register data performed better than a simpler model based on screening data alone, while genetic polygenic scores did not provide additional predictive value. The findings are a promising step towards tools that could help identify patients who may need more tailored support from the outset. The next step is prospective validation in new patients.
Congratulations to Olly and all co-authors!
Read the full article: https://www.jmir.org/2026/1/e100162











