A Learning Agent for Stress Multi-Level Diagnostics, Personalised Stress Profiles and Interventions in the Work Context
Oliver Jung, Mathias Schmoigl-Tonis, Christoph Schranz, Beatrix Kunas, Anton-Rupert Laireiter, Julian Mehlis, Michael Beer (2025): A Learning Agent for Stress Multi-Level Diagnostics, Personalised Stress Profiles and Interventions in the Work Context In: Studies in Health Technology and Informatics, Volume 324: dHealth 2025
Work-related stress affects 39% of Austrians, contributing to mental health issues like depression and burnout, driven by factors such as workload and lack of control.
The Relax project aims to develop a holistic stress management framework using continuous stress assessment and personalized interventions based on physiological, behavioral, emotional, and cognitive indicators.
The study combines wearable sensors (e.g., Polar Verity Sense), psychological methods, and technical strategies, with a longitudinal design to assess the app’s usability and effectiveness.
Usability was hindered by technical issues, but stress data visualization was well-received. The Micro-Model aligned well with stress estimates, while the Macro-Model faced data limitations. Methodological constraints suggested further refinement.
The study highlights the need for individualized, multimodal stress management, with future research focusing on system improvements, technical refinement, and validation through an anonymized dataset.
