Advancing Predictive Modelling and Transdiagnostic Intervention Research in Persistent Somatic Symptoms (SOMA.STATS.2)

• Applicant: Professorin Dr. Antonia Zapf
• Subject Area: Public Health, Healthcare Research, Social and Occupational Medicine ; Epidemiology and Medical Biometry/Statistics
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 445297796

Background: The first funding phase of the SOMACROSS Research Unit 5211 aimed to identify biomedical, psychosocial, and sociodemographic factors contributing to persistent somatic symptoms (PSS). Previous studies had described numerous risk factors and mechanisms, but these were mostly disease-specific and lacked a transdiagnostic perspective. The Z-Project (SOMA.STATS) addressed this gap by distinguishing transdiagnostic from diagnosis-specific predictors, applying advanced statistical modeling approaches such as structural equation modelling (SEM) based on high-quality datasets. Results of the first funding phase: In SOMA.STATS, a central database was established, integrating data from 1,343 participants across nine of ten medical conditions and 62 healthy controls. Cross-project analyses identified both transdiagnostic and diagnosis-specific predictors of PSS. A risk score was developed to predict PSS severity and enable early identification of high-risk patients. Mediation analyses in P2 (SOMA.GUT) showed that modifying dysfunctional expectations and illness-related anxiety can reduce gastrointestinal symptom severity – providing a starting point for interventions. Objectives: Building on these findings, SOMACROSS.2 shifts the focus from observational studies to the development of psychological interventions. In SOMA.STATS.2, risk scores and prediction models will be validated in new cohorts, and interventions targeting modifiable mechanisms such as dysfunctional expectations, symptom-related distress, depression, and avoidance behavior will be tested. The aim is to evaluate empirically grounded, mechanism based interventions for PSS across different medical conditions. Work programme: SOMA.STATS.2 will integrate data from intervention and cohort studies into a harmonised dataset. Prediction models, SEM, and risk scores from the first funding phase will be validated with new data using longitudinal models and compared with machine learning approaches. Intervention effects will be analysed across conditions using mixed regression models and single-case experimental designs (SCED). In addition, SOMA.STATS.2 will provide methodological support, project-specific analyses, and training for early career researchers. Expected impact: In the second funding phase, SOMA.STATS.2 will ensure methodological consistency and integrity of research data. Risk scores will be validated to enable the early identification of vulnerable patients, while longitudinal mechanisms of symptom persistence will be investigated. By evaluating intervention modules, both transdiagnostic and condition-specific mechanisms will be identified. In doing so, SOMA.STATS.2 will advance the overarching goal of SOMACROSS.2: the development of empirically grounded, transdiagnostic models and interventions for PSS.

• DFG Programme: Research Units
• Subproject of: FOR 5211: Persistent SOMAtic Symptoms ACROSS Diseases: From Risk Factors to Modification (SOMACROSS.2)
• Co-Investigator: Professor Dr. Bernd Löwe