In patient monitoring, the set of conclusions that is possible to make is larger when combining information from several variables than when making conclusions on separately applied single variables. During patient monitoring, a high amount of physiological variables is already produced electronically but the information they bear is not fully utilized. The present patient monitoring equipment typically just represent measured variables without combining information that can derived from them. In literature, several prototype systems have been developed to generate high level assessments of the patient's state (e.g., hypovolemia) on a multivariate basis. The development of these systems has not yet resulted in commercial products, probably partly because these systems easily grow complex. Modularization of a system makes the development easier and enables gradual development of the system. In our architecture for patient monitoring, source data is first preprocessed in order to provide representative and informative numerical features for further processing. Preprocessed variables are transformed into a standard scale to separate the physiological interpretation of single variables and the combining of them. The applied standard scale of the physiologically interpreted variables makes the development of patient's state assessment easier. On top of the architecture, the assessment of the patient's state is performed by assessing first the subsystems of the patient's physiology.
|Journal||Acta Anaesthesiologica Scandinavica|
|Issue number||Suppl. 100|
|Publication status||Published - 1993|
|MoE publication type||A1 Journal article-refereed|
|Event||22nd Congress of the Scandinavian Society of Anaesthesiologists - Kuopio, Finland|
Duration: 28 Jun 1993 → 2 Jul 1993