Abstract: Objective: To explore the research and development of a multimodal decision-support system for physical restraint and evaluate its clinical application effects, so as to standardize clinical restraint assessment and promote the standardized implementation of physical restraint. Methods: A multimodal decision-support system for physical restraint was constructed by integrating multi-source heterogeneous data and intelligent algorithms, and homogeneous training on the system was conducted. Neurological critical care patients were divided into a control group (62 cases) and an observation group (61 cases) according to the order of admission. The control group received routine physical restraint management, while the observation group implemented restraint management in accordance with the multimodal hierarchical physical restraint scheme by applying the multimodal decision-support system. The physical restraint rate, incidence of delirium and unplanned extubation rate were compared between the two groups. The scores of nurses' knowledge and behaviors related to physical restraint before and after training, as well as nurses' satisfaction scores with the restraint management system were also analyzed. Results: The unplanned extubation rate was 0 in the observation group and 1.6% in the control group, with no statistically significant difference between the two groups (P>0.05). The physical restraint rate (14.8%) and incidence of delirium (3.3%) in the observation group were significantly lower than those in the control group (27.4% and 14.5%), with statistically significant differences (P<0.05). After training, nurses' total score of knowledge related to physical restraint increased from 10.83±0.47 to 15.17±0.53, and the total score of physical restraint behaviors increased from 49.79±2.85 to 68.65±3.00, with statistically significant differences before and after training (P<0.05). The average satisfaction score of 31 nurses with the system was 4.62±0.36, indicating a high level of satisfaction. Conclusion: The implementation of body restraint management through the auxiliary multimodal decision-making information platform can reduce the body restraint rate of patients with neurological disorders, as well as the incidence of delirium, and is conducive to the standard execution of body restraint by nurses, ensuring patients' safety and comfort.

Key words: neurological critical care; multimodal; decision-support system; hierarchical restraint; unplanned extubation; restraint rate; delirium