Computational application of semiotic principles in intensive care units

Abstract

Intensive care units (ICU's) are places where full-time vigilance is required for patients. Sometimes, the attention and the medical care of the staff are not adequately distributed to all beds. The eventual lack of attention can be due to fatigue, stress, bad judgment and inadequacy of the alerts arriving from multi-parametric monitors. Since the monitors do not analyze the clinical context of the patient and their alarms do not consider the relationships between various parameters, it would be necessary to provide a computational mean able to cross information from the clinical history of the patients and the monitored vital signs, to help providing good therapeutic. To address this issue, a computational interface semiotic-based was implemented to construct knowledge from the patient records. After, it uses the constructed knowledge to try predicting clinical events. The results show that the interface was capable to identify relationships between parameters and clinical events.