The validation and accuracy of wearable heart rate trackers in children with heart disease: a prospective cohort study (Preprint)

BACKGROUND Wearables are increasingly used in pediatric cardiology for heart rate (HR) monitoring due to advantages over traditional heart rate monitoring, such as prolonged monitoring time, increased patient comfort and ease of use. However, their validation in this population is limited. OBJECTIVE This study investigates the HR accuracy and validity of two wearables, the CardioWatch bracelet and Hexoskin shirt, in children attending the pediatric cardiology outpatient clinic. In addition, factors that influence HR accuracy, the Hexoskin shirt's arrhythmia detection efficacy, and patient satisfaction are investigated. METHODS Children indicated for a 24h-Holter ECG were equipped with a 24h-Holter ECG (gold standard), together with both wearables. HR accuracy was defined as percentage of HRs within 10% of Holter values and agreement was assessed using Bland-Altman analysis. Subgroup analyses were conducted based on body mass index (BMI), age and time of wearing, among other factors. A blinded pediatric cardiologist analysed Hexoskin shirt data for rhythm classification. Patient satisfaction was measured using a 5-point Likert-scale questionnaire. RESULTS Thirty-one participants (mean age 13.2±3.6 years; 45% female) and thirty-six (mean age 13,3±3,9) participants were included for the CardioWatch and Hexoskin measurements respectively. Mean accuracy was 84.8% (±8.7%) for the CardioWatch and 87.4% (±11.0%) for the Hexoskin shirt. Hexoskin shirt accuracy was notably higher in the first 12 hours (94.9±7.4%) compared to the latter 12 (80.0±16.7%, P<.001). Higher accuracy was observed at lower HRs (low vs. high HR: CardioWatch: 90.9±9.3% vs. 79.0±10.6%, P<.001; Hexoskin shirt: 90.6±14.0% vs. 84.5±11.8%, P<.001). Both wearables demonstrated good agreement in their HR measurement with Holter readings (CardioWatch bias: –1.4 beats per minute [BPM]; 95% Limits of Agreement [LoA]: –18.8 to 16.0. Hexoskin shirt bias: –1.1 BPM; 95% LoA: -19.6 to 17.4). Correct classification of the Hexoskin’s shirt rhythm recordings was achieved in 86% (31/36) of cases. Patient satisfaction scores (median[range]) were significantly higher for both the CardioWatch (3.8[3.5–4.3], P<.001) and Hexoskin shirt (3.7 [3.0–4.0], p<0.001) compared to the Holter ECG (2.6 [2.1-3.2]). CONCLUSIONS The Corsano CardioWatch and Hexoskin shirt demonstrate good accuracy in pediatric HR monitoring and provide higher patient comfort than conventional monitoring. Both wearables show good agreement in relation to the gold standard device. However more research is needed to explore the reasons for inaccuracy during higher heart rates. The Hexoskin shirt also shows potential in arrhythmia detection. While further development is warranted, these wearables show promise in enhancing diagnostics, therapeutic monitoring and patient safety in pediatric cardiology.