Comprehensive assessment of the accuracy of heart rate determination with the personal wearable device

Abstract

BACKGROUND: Increased resting heart rate (HR) is an independent risk factor for overall mortality, sudden cardiac death, and death from cardiovascular diseases (CVD) and a worsening factor for CVD patients’ prognosis. Moreover, HR is easy to measure and monitor independently using various devices. The accuracy of personal wearable devices used in assessing HR should be investigated. AIM: To evaluate the accuracy of HR determination with the personal wearable device (PWD). MATERIALS AND METHODS: An open, observational, single-center study was performed. Participants underwent physical examination, electrocardiography, HR recording using PWD, and electrocardiography Holter monitoring (HMECG). For each participant, the study duration was 14 days. The reliability of HR measured in everyday life was assessed using the PWD in comparison with HMECG as the “gold” standard. For statistical analysis, Microsoft Excel 2016 and STATISTICA 10.0 software were used. To assess the reliability of HR measured in everyday life using PWD in relation to HMECG, the hypothesis of the equality of average HR values measured by both methods was tested, a correlation analysis of time series of HR values was conducted, and Bland–Altman plots were used to visualize differences in HR values. RESULTS: The study involved 45 healthy individuals (22 men, 49.24±6.47 years). HR data were obtained over a 24-hour period, including the subjects' routine working day. The results demonstrated good convergence between data on HR using PWD and HMECG. The relative error in determining the HR of PWD did not exceed 3.2%. The smallest relative error was recorded at night and early morning hours (0.3÷1.1%). During the daytime it was slightly higher — from 2.0 to 3.2%. A significantly high positive correlation was obtained between HR recorded using both devices, both at night (MCorrP from 0.75 to 0.85; p 0.001) and during daytime (MCorrP from 0.77 to 0.85; p 0.001). CONCLUSION: The accuracy comparison of 24-hour HR measurements by PWD showed that they were mostly within the acceptable error range (less than 3.2%). Future studies should include HR PWD assessment in patients with various diseases.