Conventional versus computer‐aided interpretation of long‐term ECG recordings–aspects of precision and economy

Abstract

In a sample of 60 12-h long-term ECG recordings we compared a computer-aided analysis method to a conventional scanning technique for diagnostic precision and time consumption. As regards diagnostic precision no significant difference between the methods could be found. Both were considered efficient in detecting episodes of arrhythmia. Using the computer-aided method a physician spent an average of 16 min on analyzing a recording and writing the diagnostic report. The conventional method required 69 min from an ECG technician and 9 min from a physician for each recording. From these figures we conclude that about 750 12-h recordings/year are required to outweigh the extra annual cost for the computer, if it is used only for long-term ECG analysis. If the computer can also be used for other purposes the system for long-term ECG analysis is profitable with a smaller annual number of recordings.

Adaptive QRS Detection Based on Maximum A Posteriori Estimation

“…Since the continuous-time signal is band limited, it is uniquely determined as (27) with R'(f) given by (23). For (28) should be evaluated for continuous k. The estimator could be implemented by using a number, P, of filters in parallel, ho(k), ho(k + l/P),-.. , ho(k + (P -1)/P).…”

“…RESULTS AND DISCUSSION Clinical experience from about 600 12-h recordings, and a study of the performance of our monitoring system as a whole indicate that the QRS detector works well in the clinical routine [23]. The performance of the detector has preliminarily been evaluated for a material consisting of several ECG's from different patients [21], [22].…”