A Validation Study: Fitbit Charge 2 Heart Rate Measurement at Rest and During Cognitive-Emotional Stressors

Abstract

Previous studies emphasize concurrent validity of wearable devices during physical activity. We investigate device agreement between Fitbit Charge 2 (FBC2) Heart Rate (HR) and electrocardiography (ECG) under sedentary conditions. In a sample of 32 post-secondary students, agreement was observed during rest, negative affect images, deceptive pattern solving task, and a listening task. ECG HR and FBC2 HR was sampled continuously and transformed into 1120 paired 1-min epochs throughout a 35-minute protocol. Bland-Altman were adopted to evaluate agreement. FBC2 HR underestimated ECG HR as indicated by a mean error bias = -1.1 bpm (95% limits of agreement, -5.62 to 3.42), mean absolute percentage error 1.34 (SD = 1.85) %. Bland-Altman limits of agreement plot indicated minimal systematic error in the measurement range of 54 bpm to 117 bpm. When compared to gold-standard ECG HR the FBC2 HR demonstrated good agreement in healthy young adults.