Project Summary Recent large-scale trials have shown no significant benefit of pharmacological interventions in delirium patients, and non-pharmacological approaches remain the cornerstone of delirium prevention. Among those strategies, minimizing patient immobility and circadian desynchrony are particularly difficult to implement, as their assessment is dependent on sporadic human observations. The overall objective of this application is to develop ADAPT, the Autonomous Delirium Monitoring and Adaptive Prevention system using novel pervasive sensing and deep learning techniques. It will autonomously quantify patients’ mobility and circadian desynchrony in terms of nightly disruptions, light intensity, and sound pressure level. This will allow for integration of these risk factors into a dynamic model for predicting delirium trajectories. It will also enable adaptive action prompts aimed at increasing patients’ mobility, reducing nightly disruptions, optimizing ambient light, and reducing noise, based on precise real-time quantification. The rationale is that successful application of the proposed technology would augment clinical-decision making in the fast-paced ICU environment and would promote more targeted interventions. The overall objective will be achieved by pursuing three specific aims. (1) Developing and validating an interpretable deep learning algorithm for precise and dynamic prediction of the delirium trajectory, to determine if it is more accurate in predicting delirium trajectory transitions compared to existing tools, while providing interpretable information to the physician. (2) Developing a pervasive sensing system for autonomous monitoring of mobility and circadian desynchrony, to determine if it can provide accurate assessments compared to human expert and circadian biomarkers, and if it can enrich delirium trajectory prediction when combined with clinical data. (3) Developing and evaluating prompts for adaptive delirium prevention using real-time monitoring system, to determine if the system has acceptable satisfaction and perceived benefit among ICU physicians. The approach is innovative, because it represents the first attempt to (1) dynamically predict precise delirium trajectory, (2) autonomously monitor mobility and circadian desynchrony risk factors in the ICU, and (3) implement adaptive preventions in real time. The proposed research is significant since it will address several key problems and critical barriers in critical care, including (1) lack of precise and real-time delirium trajectory prediction models, (2) uncaptured aspects of mobility and circadian desynchrony, and (3) the need for novel approaches for non-pharmacological prevention. Ultimately, the results are expected to improve patient outcomes and decrease hospitalization costs, as well as lifelong complications.