GE HEALTHCARE B20, Руководство пользователя

Страница: 199 / 320

ECG
11-17
Description of the arrhythmia algorithm
When an ECG signal is detected at the start of monitoring, the arrhythmia detection algorithm
begins acquiring and analyzing QRS complexes in the leads used for arrhythmia detection. This
phase is known as learning. Once learning is complete, the dominant QRS complex is stored as
a reference template. Reference template is used as a normal morphology of that patient and
it is compared with incoming beats to identify possible arrhythmias.
The algorithm uses continuous correlation, incremental template updating and contextual
analysis. Continuous correlation attempts to find the best match between each incoming
complex and the set of stored (learned) templates. Incremental template updating allows
information from each beat, that correlates over time, to be reflected in the associated
template. Contextual analysis uses information from neighboring QRS complexes along with
existing template measurements to make the best possible decision regarding the beat’s origin
(e.g., early, wide).
NOTE: A physician must analyze the arrhythmia information in conjunction with other
clinical findings.
Test results of arrhythmia algorithm testing
The algorithm testing has been performed per EC57 by using the following databases:
− AHA (The American Heart Association Database for Evaluation of Ventricular Arrhythmia
Detectors)
− MIT-BIH (The Massachusetts Institute of Technology- Beth Israel Hospital Arrhythmia
Database)
QRS numeric accuracy
•
QRS sensitivity: >97.5%
•
QRS positive predictivity: >97.5%
VEB numeric accuracy
•
VEB sensitivity: >90%
•
VEB positive predictivity: >90%
VF numeric accuracy
•
VF episode sensitivity: >95%
•
VF episode positive predictivity: >95%