Techniques for improving overall consistency of serial ECG analysis

The techniques that improve the overall repeatability of computer interpretation of electrocardiograms (ECGs) that have been recorded several minutes apart from patients in a clinically stable condition are described. Estimates of the normal amounts of variability present in many ECG parameters that are used in the identification of a variety of cardiac abnormalities have been adopted in conjunction with smoothing techniques to form the basis of the new methodology When applied to the Glasgow ECG analysis program, these new methods improve overall repeatability by about 31% when tested on a set of 263 pairs of ECGs. Randomly generated noise was added to the test set and an additional technique aimed at removing noise from the ECG tracings was used in conjunction with the smoothing methods. The observed improvement over the original repeatability was 63%.