EXAMINING VALIDITY OF CONSUMER-AVAILABLE ACTIVITY MONITORS IN MEASURING ENERGY EXPENDITURE, HEART RATE, AND STEPS

Abstract

The first study of this dissertation evaluated the accuracy of the Fitbit Surge (FBS), Garmin Vvofit (GVF), and SenseWear armband (SWA) in measuring energy expenditure (EE), heart rate (HR), and steps during treadmill and cycling activities performed at two intensities in healthy, physically active individuals. In the second study, the monitors were evaluated in measuring EE and HR during a gym-based routine that included aerobic and resistance training activities performed by healthy, physically active participants.

In the first study, the activity monitors underestimated EE compared to the Oxycon Mobile (OM) metabolic analyzer across all bouts and their accuracy declined with vigorous intensity, signified by higher measurement error. The HR analyses revealed that the FBS and GVF yielded lower average HR (HRavg), although the estimates were comparable to the Polar HR monitor (PM). The GVF had better HR accuracy over the FBS, however, the difference in accuracy was minimal during the moderate intensity treadmill bout. The same trend was observed for session maximal HR (HRmax). The step count analysis showed that all monitors accurately estimated steps during the vigorous intensity treadmill bout. During the moderate intensity bout, only the SWA had an equivalent step count with the video observation.

The results of the second study demonstrated that no monitor was equivalent to the OM in assessing EE. The FBS and GVF overestimated EE for all segments of the gym-based session. The SWA overestimated EE for the treadmill running bout but underestimated EE for the stationary cycling and resistance training bouts, which resulted in reasonable whole-session EE estimates (450.9 142.1 kcal) compared to the OM (470.6 106.0 kcal), further supported by low measurement error. Equivalency testing for HRavg data revealed that only the FBS did not agree with the PM during the stationary cycling bout. The GVF had superior accuracy in measuring HR indicated by lower measurement error across all segments of the session.

In conclusion, the activity monitors were the least accurate in measuring EE during common aerobic and resistance training activities. The monitors showed promising accuracy for measuring steps during treadmill walking and running. Lastly, the wrist-worn monitors demonstrated good potential in measuring HRavg and HRmax, although the GVF appears to have a lower measurement error than the FBS.