Signal quality of resting electrocardiograms

Artifact may cause errors of technical origin when ECGs are interpreted by automatic methods. Baseline shift and high-frequency noise content of minimal and typical length ECG records from pediatric and adult populations were measured to allow prediction of both the likelihood of interpretation errors of technical origin and the number of reacquistions needed to obtain an artifact-free record. Ages of the 708 subjects in this study ranged from 2 weeks to 27 years. When a baseline shift of 0.25 mv (exceeded in 7% of the R-R intervals in the database) or a noise content greater than 15 muv RMS (exceeded in 6% of the R-R intervals in the database) within six seconds of three simultaneous leads was declared an unacceptable artifact, then 68% of the records from 0-4 year olds and 31% of the records from adults (greater than 19 years), were rejected on the basis of technical quality. These failure rates mean that, on the average, 3.1 tries would be needed to obtain an artifact-free record from 0-4 year olds; 1.4 tries would be needed for adults. If acquisition is done interactively, the measurement time for a 6-second, 3-lead group would be increased by 13 seconds for 0-4 years olds and by three seconds for adults in order to assure adequate signal quality for computer-assisted analysis.