A study to match duration criteria and computer ECG measurements

Abstract

The differences between computer and 'conventional' manual ECG (electrocardiography) duration measurements are assessed. The ultimate goal is to adapt duration criteria designed for the visual interpretation of ECGs to analysis by computerized systems. Independently of any diagnostic indication, a set of 100 ECG signals has been used to measure several duration parameters of potential diagnostic significance, such as QRS, P, and Q duration and R peak time. The values obtained by computer (program CARD5) have been compared with the results given by two physicians on conventional gain/speed recordings. Human discrepancies have been resolved through a joint analysis of the difficult cases. The consensus results have been compared with computer values. Almost all differences appear to be statistically significant; however, for the P duration criteria only a correction taking into account the systematic difference is suggested.<>