“But for the blind spot”: Accuracy and diagnostic performance of smart watch cardiac features in pediatric patients

Abstract

BackgroundThe Apple Watch™ (AW) offers heart rate (HR) tracking by photoplethysmography (PPG) and single-lead electrocardiographic (ECG) recordings. The accuracy of AW-HR and diagnostic performance of AW-ECGs among children during both sinus rhythm and arrhythmias have not been explored. ObjectiveThe purposes of this study were to assess the accuracy of AW-HR measurements compared to gold standard modalities in children during sinus rhythm and arrhythmias and to identify non–sinus rhythms using AW-ECGs. MethodsSubjects ≤18 years wore an AW during (1) telemetry admission, (2) electrophysiological study (EPS), or (3) exercise stress test (EST). AW-HRs were compared to gold standard modality values. Recorded AW-ECGs were reviewed by 3 blinded pediatric electrophysiologists. ResultsEighty subjects (median age 13 years; interquartile range 1.0–16.0 years; 50% female) wore AW (telemetry 41% [n = 33]; EPS 34% [n = 27]; EST 25% [n = 20]). A total of 1090 AW-HR measurements were compared to time-synchronized gold standard modality HR values. Intraclass correlation coefficient (ICC) was high 0.99 (0.98–0.99) for AW-HR during sinus rhythm compared to gold standard modalities. ICC was poor comparing AW-HR to gold standard modality HR in tachyarrhythmias (ICC 0.24–0.27) due to systematic undercounting of AW-HR values. A total of 126 AW-ECGs were reviewed. Identification of non–sinus rhythm by AW-ECG showed sensitivity of 89%–96% and specificity of 78%–87%. ConclusionsWe found high levels of agreement for AW-HR values with gold standard modalities during sinus rhythm and poor agreement during tachyarrhythmias, likely due to hemodynamic effects of tachyarrhythmias on PPG-based measurements. AW-ECGs had good sensitivity and moderate specificity in identification of non–sinus rhythm in children.