Automatic ECG and arrhythmia interpreter--an almost analog device.

Summary Automatic ECG diagnosis requires a computer with a vast amount of calculation and memory capacity. Moreover, arrhythmias, which occur unexpectedly, should also be included in the subjects for diagnosis in daily practice. Accordingly, data from a considerable number of the cardiac cycles must be treated with a substantial increase in computation and resultant data. In this regard the authors have developed a special purpose computer of hybrid type for automatic ECG diagnosis. With this system data reduction is carried out by analog treatment of ECG and the subsequent processes are dealt with in digital form. Thirty-two consecutive heart beats in six leads, (I, II, III, V 1 , V 3 and V 5 ) are fed into the computer. For data reduction the following procedure is employed. The differentiated ECG waveform is transformed into square waves with Schmitt triggers. Time values, such as QRS duration, RR intervals and PR intervals, are measured using the intervals from the beginning to the end of each square wave or between the beginnings of two square waves, and then converted in a digital form. The amplitude of the waves are obtained by use of peak holders which operate simultaneously with the leading edges of the squared waves obtained from original ECG waveforms. The results show complete agreement between computer diagnosis of the wave shape and physician's interpretation in 79.2% of 221 cases, partial disagreement in 9.0%, and disagreement in 11.8%. In left bundle branch block and incomplete right bundle branch block agreement was obtained in all cases. Disagreement was seen once in left ventricular hypertrophy, myocardial infarction, right ventricular hypertrophy, and complete right bundle branch block. For the diagnosis of arrhythmias, f and F waves detector and a sinus rhythm detector are used. Six items—RR interval, PP interval, PR interval, QRS duration, polarity of the P wave and number of P waves found in the TQ interval—are coded in a binary system and diagnosis is made with the use of AND Gates. The diagnosis of the type of arrythmia was correct in 91.5% of 318 cases. These results indicate that this system is useful not only for mass screening, but also for diagnosis in daily hospital practice.