P1127Usability of single lead ECG from smartphones: the USELESS pilot?

Abstract

Background: Different wireless, single-lead real-time ECG devices are commercially available for electrocardiogram (ECG) screening, monitoring and possible on-demand diagnosis. Some devices have demonstrated acceptable detection of intervals and rhythms, and are increasingly available. Objective: We wanted to evaluate the accuracy and usability of Smartphone ECG (spECG) in both healthy controls and clinic patients. Our hypothesis was that the device would prove to be of poor clinical value. Methods: We used a commercially available spECG device, connected wirelessly to a tablet, to record 30-second lead I ECG waveforms, which were interpreted by two observers. We examined the inter-and intraobserver variability of five standard intervals (PQ, QRS, QT, QTc), frequency and rhythm (sinus, atrial fibrillation, SVES, VES etc.). In clinic patients, we compared variability between standard calculated 12-lead ECG (scECG) and spECG. We have currently studied intervariability in four groups: Young healthy individuals at rest (n=20) and immediately after vigorous exercise (n=20), healthy individuals over 50 years (n=20) and clinic patients (n=20). For evaluation of the spECG we studied clinic patients (n=27), and acquired scECG for comparison shortly after. The ECGs were anonymized, and the data handled in Stata/IC. Results: In 1068 measured intervals (PQ, QT, QRS, frequency), 9,2% were unreadable, whereas 23,2% of the spECGs had one or more unreadable interval. The PQ measurements accounted for 65,3% of the disturbance. We calculated intraobserver variability to be: Observer 1: 3,0%, Observer 2: 3,1%. We calculated interobserver variability for each of the four groups to be: Untrained: 2,5%, Trained: 2,6%, Elderly: 2,4% Clinic Patients: 3,2%. Ultimately this represents a standard error for all measured intervals of less than 10 ms or 3 beats/min. Inter-and intraobserver variability was comparable. These were also comparable to spECG and scECG-variation (3,1%), see Figure 1. After vigorous exercise the readability decreased slightly in PQ intervals: from 95,0% to 87,5%. All other intervals remained fully interpretable. By defining true positive as abnormality seen in both modalities we found the specificity to be 62,5% and the sensitivity to be 90,9%. Bland Altman plotting of the five intervals indicated that spECG and scECG were comparable in lead I. In the groups over 50 years old, 22,7% needed assistance in handling the device. Conclusion: The examined wireless, single-lead real-time ECG device is not useless. It accurately measures intervals and frequencies as well as detects abnormal rhythms, although specificity is suboptimal and inconclusive recordings occur. Vigorous activity has a minor influence on the readability of the recorded intervals, primarily the PQ interval. Elderly might have challenges recording a spECG correctly without assistance.