Development and Validation of a Wearable-Based Monitoring System for Predicting Psychiatric Crises in Inpatients: Study Protocol

This study introduces a protocol to evaluate the feasibility and clinical utility of the Mental Instability and Risk Early Detection Monitoring System (MIR-Med System), a novel, digital monitoring platform designed for real-time detection and prevention of self-harm, aggression, and psychiatric instability among psychiatric inpatients. The MIR-Med System integrates wearable devices and machine learning algorithms to collect and analyze digital phenotyping data, such as circadian rhythm patterns, heart rate variability, and sleep quality. By leveraging a three-layer risk assessment model—historical risk factors, circadian rhythm disruptions, and immediate physiological indicators—the system enables proactive, real-time risk evaluation. The study will be conducted in both secure and open psychiatric wards, involving initial clinical assessments by medical staff and continuous data collection through wearable devices (e.g., Fitbit). Primary objectives include assessing the system’s stability in data collection, predictive accuracy, and ability to generate timely alerts for preemptive interventions. Secondary objectives include evaluating the system’s scalability, usability, and impact on clinical workflows. The feasibility of the MIR-Med System will be determined by comparing its alerts to verified clinical outcomes, with performance metrics such as sensitivity, specificity, and F1 score. By addressing limitations in current risk assessment practices, this study aims to enhance patient safety, optimize clinical workflows, and reduce the burden of self-harm and violence in inpatient psychiatric settings. Findings will inform the development of scalable, real-time monitoring solutions applicable to diverse clinical environments.