Olly’s ISP seminar

Today, Olly held her individual study plan (ISP) seminar for her doctoral studies. The overall purpose of the studies is to enhance the prediction of health and socioeconomic outcomes in patients with common psychiatric disorders, such as depression and anxiety.

We know that 25-50% of patients with depression and anxiety disorders don’t respond to CBT. These patients may benefit from alternative or tailored treatment formats. We want to explore ways to help them, but in order to do that we need to predict who they are. Existing predictive models of treatment outcome are usually based on clinical data with a small set of predictors which results in low predictive power and limited clinical utility.

Olly’s PhD project will use a big sample from Psykiatri Sydväst’s Internet Psychiatry and predict treatment outcomes for multiple disorders. It will use clinical, genetic and register data to predict both treatment response and long-term socioeconomic disadvantage such as labour market marginalization.

The work consists of four studies.

  • Study 1: Clinical, socioeconomic and genetic predictors of treatment outcomes in internet-delivered cognitive behavioral therapy for depression and anxiety disorders
  • Study 2: Development of a machine learning model to predict response to ICBT treatment at the individual subject level
  • Study 3: Development of a machine learning model aimed at predicting labour market marginalization
  • Study 4: Validation of the model from study 2 in clinical practice

Olly’s main supervisor is Christian, and her co-supervisor is John.

We look forward to this!

Welcoming Olly Kravchenko

We are happy to announce that Olly Kravchenko just joined the Rück research lab as our new PhD student!

Photo: Micke Sandström

Olly has a background in project management, product testing, and user experience research in tech start-ups. She also has a MSc in German and English Language & Literature (Kyiv National Linguistic University) and a MSc in Public Health Sciences (Stockholm University). She recently finished an internship at the Department of Medical Epidemiology and Biostatistics at Karolinska Institutet, working on a genetically informed research project using data from the Swedish Twin Registry.

Here at the Rück Research Lab, Olly will mainly be working with the PRiSMED project, with John Wallert as main supervisor and Christian Rück as co-supervisor.

About the project
PRiSMED: Predicting health and socioeconomic outcomes in patients with common psychiatric disorders

Depression, social anxiety disorder, panic disorder, and obsessive-compulsive disorder are common mental disorders with a combined point prevalence of 15%. Almost 50% of diagnoses leading to sick-leave in Sweden are psychiatric and their share is increasing. Cognitive behavior therapy (CBT) is first-line treatment for these conditions. Yet, 30-60% of patients will not respond to CBT. Several predictors of CBT outcome have been proposed. Results are mixed and lack predictive acuity to guide clinical decisions. Prediction studies using larger samples and multimodal data (clinical, register, genetic) are urgently needed. Coupled with advanced modelling, we could augment precision psychiatry allowing for tailored intervention and cost-effective resource use.

The project main purpose is improved outcome prediction for common mental disorders in routine clinical care. We will build prediction models for both clinical CBT outcomes (e.g., remission) and long-term outcomes (e.g., poverty) using multimodal data on 5,000 genotyped and register-linked patients treated with CBT – the hitherto largest data collection of its kind.

The project has three aims. Aim 1 uses traditional statistical methods to identify group-level predictors for these outcomes. Aim 2 applies Machine Learning (ML) to identify which individual will experience these outcomes. Aim 3 will plan and initiate a trial comparing a developed ML algorithm versus clinicians at predicting remission in prospective CBT patients.