Rethinking Depression Relapse: Modelling the Dynamic Interplay of Functioning, Depressive Symptoms, and Psychological Mechanisms

This research is supported by the SNSF under its project funding scheme. The project will start in 2027.

Depression often follows a chronic, relapsing course, where each episode increases the risk of recurrence—leading to higher hospitalization rates, healthcare costs, reduced work performance, and prolonged individual suffering. Even after successful depression treatment, relapse is common and remains a major public health challenge. Current relapse prevention approaches focus primarily on symptom reduction, overlooking daily functioning. Yet, impairments in functioning often persist during remission and may interact with psychological mechanisms to shape long-term recovery or impairment. Understanding these processes is key to improving relapse prediction and guiding personalized interventions. The aim of this project is to increase our understanding of the intrapersonal and interpersonal dynamics of depression deterioration and relapse, with a novel focus on the dynamic role of functioning. Two levels of granularity will be used to model how functioning interacts with symptoms and depression-related psychological mechanisms over time, and to use these models for predicting relapse in patients who have undergone acute-phase depression treatment. To address this aim, a prospective, multi-center observational study will be conducted in a sample (N=200) of individuals who have terminated acute phase treatment for depression in one of the participating clinical study sites. The study uses an intensive longitudinal design in which ecological momentary assessments (EMA) will capture daily fluctuations in functioning, symptom severity, and depression-related mechanisms during the first five months post-treatment. Three EMA blocks will be scheduled each lasting two weeks with five daily prompts resulting in a total of 42,000 observations. Follow-up assessments will assess relapse over a total observation period of 12 months, enabling analysis of short-term dynamics and long-term relapse risk. Speech diaries will allow participants to report daily positive and negative experiences. Primary analyses will rely on Continuous-Time Dynamic Models (CTDM), estimated within a Bayesian framework. This approach integrates intensive longitudinal data across domains, capturing autoregressive and cross-lagged relationships while accounting for within- and between-person variability. To reduce the participant burden of this intensive study design a participatory approach is embedded in the study, informing design, feasibility, and results' interpretation through structured input and co-design. This study addresses critical gaps in relapse prevention research by (1) incorporating daily functioning as a central component of relapse prediction, (2) leveraging multimodal daily-life data through dynamic modeling to capture real-world complexity, and (3) embedding lived experience perspectives. These innovations are expected to yield a more nuanced understanding of relapse mechanisms, enable identification of individualized dynamic risk profiles, and advance methodological approaches for integrating functioning into longitudinal models which may inform the development of personalized relapse prevention strategies with meaningful implications for patients' recovery, rehabilitation, and well-being.

Team

PIs: Prof. Dr. Birgit Watzke, Department of Psychology, University of Zurich

Project Partners: Prof. Dr. Charles Driver; Department of Psychology, University of Zurich, PD Dr Anja Frei; Population Research Center, University of Zurich, Prof. Dr. Viktor von Wyl; Institute for Implementation Science in Health Care, University of Zurich

Contact: Dr. Catalina Nuñez;  Department of Psychology, University of Zurich