インテリジェントアラームのための治療看護手技認識手法

概要

論文の詳細を見る
Abstract In intensive care units, most patient monitors generate many nuisance alarms. There are many artifacts in vital signs measured by a patient monitor and current patient monitors are not able to recognize the difference real emergencies and those artifacts. Currently alarm information is not generated based upon the clinical staff behavior, but of only the patient's condition. To achieve a sophisticated alarm generating system, the patient monitor has to recognize the clinical activities of clinical staff. This paper describes a method of recognizing clinical care by combining simultaneous vital signs'data and video image data analysis. Clinical events are recognized by time transition analysis of vital signs and cross-correlation of video images on a computer. The computer is connected to the patient monitor and the video recorder in order to collect vital signs and video images simultaneously. The video camera is a small dome type camera positioned on the ceiling above the patient to observe his/her surroundings. Especially developed algorithms allow the computer to recognize clinical events performed by the clinical staff such as aspiration, drawing blood, IV dosage, and manual blood pressure measurement. Using the developed method, a patient monitor can generate alarms based upon clinical care.
ライフサポート学会の論文
2000-03-25