Recently, the advancement in data analytics including various machine learning methods lead to designing effective clinical decision support systems (CDSS) in the clinical sector. However, a CDSS including noisy clinical data can provide unhelpful (if not misleading) decision support to clinical staff as well as clinically dangerous decisions to already suffering patients. Therefore, in order to design a better and effective CDSS, such noisy data caused by patient movements and clinical protocols should be identified and filtered from the inputs in in the training dataset for CDSS development. For this purpose, we propose MediSense a system designed to identify and classify different patient motions on the bed and filter out physiological signal data points collected when patient motion occurs using sensor-based motion classification results. Essentially, MediSense can be considered as ``glasses" for the third eye in accurate-sensitive clinical domain is an intelligent embedded wireless sensing system for supporting a CDSS and consists of a motion classifier, a wireless network and localization techniques. To evaluate our system, we deploy MediSense in intensive care units (ICUs) at the Ajou University Hospital Trauma Center, a major hospital facility located in Suwon, South Korea, and evaluate its each system component's performance from real patient traces collected at these ICUs through a 4-month pilot study. Our results show that MediSense successfully classifies patient motions on the bed with >90% accuracy, shows 100% reliability in determining the locations of beds within the ICU, and each bed-attached sensor achieves a lifetime of more than 33 days, which satisfies the application-level requirements suggested by our clinical partners.
Furthermore, a simple case-study with arrhythmia patient data shows that MediSense can help improve the clinical diagnosis accuracy.
