Comparative analysis of 6-Lead and single lead consumer-grade electrocardiograms: diagnostic accuracy, numerical agreement, and inter-rater reliability

Abstract Background With the increasing availability of consumer-driven cardiac monitoring devices, there is a need to validate their diagnostic accuracy and numerical agreement against the clinical gold standard, the 12-lead electrocardiogram (ECG). Most commercially available consumer devices provide only a single lead ECG, potentially limiting their diagnostic capabilities. Purpose This study aimed to compare the diagnostic accuracy and numerical agreement of ECG readings from the 6-lead hand-held mobile device and the single lead smart watch against a standard 12-lead ECG system. The study also evaluated inter-rater reliability for each device to assess the consistency of diagnosis across clinical expertise. Methods A total of 194 patients participated in this study. The 6-lead and 12-lead ECGs were recorded simultaneously, while single lead ECGs were obtained with a slight delay. Two electrophysiology specialists interpreted all ECGs. Diagnostic accuracy for arrhythmia detection was assessed for 12-lead vs. 6-lead and 12-lead vs. single lead, using sensitivity analysis. Numerical agreement of PR interval, QRS duration, QT interval, and corrected QT interval (QTc) values was assessed using the Bland-Altman method. Additionally, inter-rater reliability on diagnosis for each device was assessed using Cohen’s kappa coefficient. Results Diagnostic Accuracy: The 6-lead mobile device demonstrated a sensitivity of 98.45% (Rater 1 and 2) in detecting arrhythmias compared to the 12-lead ECG, while the single lead smart watch achieved 93.81% (Rater 1) and 92.78% (Rater 2) (Table 1). Notably, the 6-lead ECG exhibited superior performance in detecting atrial fibrillation and other arrhythmias. Furthermore, the 6-lead ECG had 0.52% (Rater 1) and 0% (Rater 2) ungradable reading, whereas the single lead ECG had 2.06% (Rater 1) and 3.61% (Rater 2). Numerical Agreement: The proportion of beats that could not be labeled was 4.12% for the 6-lead ECG and 8.76% for the single lead ECG. The outlier rates for PR interval, QRS duration, QT interval, and QTc were comparable between the 6-lead and single lead ECG. However, PR, QT and QTc outlier rates were consistently lower in the 6-lead ECG across both raters. Inter-Rater Reliability: Cohen’s kappa values for diagnostic agreement between the two independent readers were 0.89 for 12-lead, 0.85 for 6-lead, and 0.78 for single lead, showing that 6-lead provided more consistent interpretations compared to the single lead. Conclusion The 6-lead ECG exhibited superior sensitivity for arrhythmia detection and numerical agreement compared to the single lead. The inter-rater reliability analysis indicated that interpretations from 6-lead ECGs were more concordant than those from the single lead smart watch. These findings suggest that consumer-grade ECG devices, particularly the 6-lead mobile device, could serve as reliable alternatives for preliminary cardiac assessments where full 12-lead ECG access is limited.