Sedation Evaluation Based on Support Vector Machines Using IMU Data

An escalated level of sedation has been associated with adverse outcomes, including mortality, delirium, and delayed extubation. Hence, maintaining a sedation level corresponding to a Richmond agitation–sedation scale (RASS) value of 0 is paramount in clinical settings. This article presents an innovative approach to sedation assessment, aiming to administer a minimal sedative dose while accounting for the patient’s 3-D movements. Data acquisition was carried out utilizing an inertial measurement unit (IMU), with real-time evaluation of patient RASS values achieved through support vector machine (SVM) classification. Five SVM classifiers were developed to distinguish RASS levels 0–4. Test outcomes showcased an average accuracy rate of 91.30%. Particularly noteworthy was the attainment of a 94.40% accuracy rate when the RASS value was 0.