Background: Prints of electrocardiograms (ECGs) are often sent to core laboratories, where they are scanned, converted to a digital format, and read on-screen. These ECGs may differ from the original ECG because of variability introduced by the printer, scanner, or digitization software.
Methods: Digital ECGs were recorded in 50 volunteers simultaneously using electrocardiographs from 2 different manufacturers. QT and RR intervals were measured on-screen on the digitized ECGs. To study the contribution of individual steps in the digitization process, differences in RR interval between 2 prints each of 50 digital ECGs, 2 scanned files of 50 prints, 2 digitized files from 50 scanned files, and 2 readings of 50 digitized ECGs (intrareader variability) were analyzed.
Results: Repeatability coefficient for RR interval measurement was 18.5 milliseconds for machine 1 and 21 milliseconds for machine 2. Contributions of the printer were 6.5 milliseconds for machine 1 and 9.0 milliseconds for machine 2, digitization process was 5.5 milliseconds, and reader variability was 8.0 milliseconds. Variability of the scanner was negligible.
Conclusions: The printer and digitization process account for significant differences in interval measurements in digitized ECGs.