Greystones

Researchers are conducting a multinational observational study to evaluate the feasibility of collecting speech and self-report data across six languages for AI-based relapse risk estimation in people with psychotic disorders. This study includes individuals at risk of relapse and healthy controls, aiming to explore new speech markers for predicting relapse while minimizing participant burden. It involves an international, multicenter design across diverse healthcare settings to ensure broad applicability. Participants will use a smartphone app to provide weekly speech recordings and self-report data over 12 months. These recordings are securely transferred and analyzed by an AI backend to calculate relapse risk scores, which are stored but not shared with treating clinicians. The study includes a Human-in-the-Loop quality control process with two independent reviewers assessing data accuracy and providing relapse risk estimates without influencing clinical care. Participants will attend three study visits at baseline, 6 months, and 12 months for additional assessments. During the study, researchers will evaluate user adherence, transcription quality, recording usability, overall system usability, and the performance of the monitoring system for relapse prediction. Assessments include clinical evaluations, functional and quality-of-life measures, and usability questionnaires. The total participation duration is 12 months, with ongoing monitoring and data collection to validate the AI system and discover novel speech markers associated with relapse risk.