JMIR Publications

ABSTRACT

Background:

Parkinson’s disease (PD) is the second most common neurodegenerative disease and affects approximately 1% of the world’s population. Increasing evidence suggests that aerobic physical exercise can be beneficial in mitigating both motor and non-motor symptoms of the disease. In a recent pilot study of the role of exercise on PD we sought to confirm exercise intensity by monitoring heart rate (HR). For this purpose, we asked participants to wear a chest strap heart rate monitor (Polar Electro Oy, Kempele) and the Fitbit Charge 4 (Fitbit Inc) wrist-worn heart rate monitor as a potential proxy due to its convenience. Polar H10 has been shown to provide highly accurate R-R interval measurements, therefore, we treated it as the gold standard in this study. It has been shown that Fitbit Charge 4 has comparable accuracy to Polar H10 in healthy subjects. It has yet to be determined if the Fitbit is as accurate as the Polar H10 in subjects with PD during rest and exercise.

Objective:

To compare the Fitbit Charge 4 to the Polar H10 for monitoring HR in PD subjects at rest and during an intensive exercise program.

Methods:

A total of 596 exercise sessions, from 11 subjects (6 males) were collected simultaneously with both devices. Subjects with early-stage PD (Hoehn & Yahr <=2), were enrolled in a 6-month exercise program designed for PD patients. Subjects participated in 3 one-hour exercise sessions per week. They wore both Fitbit and Polar H10 during each session. Sessions included rest, warm-up, intense exercise, and cool-down periods. We calculated the bias in the HR of the Fitbit Charge 4 at rest (5min) and during intense exercise (20min) by comparing the mean HR during each of the periods to the respective means measured by the Polar (HRFitbit – HRPolar). We also measured the sensitivity and specificity of the Fitbit Charge 4 for detecting average HRs that exceed the threshold for intensive exercise, defined as 70% of an individual’s theoretical maximum HR. Different types of correlations between the two devices were investigated.

Results:

The mean bias was 1.68 beats-per-minute (bpm) at rest and 6.29 bpm during high intensity exercise, with an overestimation by Fitbit Charge 4 in both conditions. The mean bias of the Fitbit across both rest and intensive exercise periods was 3.98 bpm. The sensitivity of the device in identifying high intensity exercise sessions was 97.14%. The correlation between the two devices was non-linear, suggesting a tendency of Fitbit to saturate at high values of HR.

Conclusions:

The performance of Fitbit Charge 4 is comparable to Polar H10 for assessing exercise intensity in a cohort of PD subjects (mean bias: 3.98 bpm). The device could be considered a reasonable surrogate for more cumbersome chest-worn devices in future studies of clinical cohorts.