Validity of Wearable Activity Monitors for Estimation of Resting Energy Expenditure in Adults

Abstract

Accurate assessment of resting energy expenditure (REE) is a key to estimating total daily energy expenditure since REE contributes to 60-70% of daily energy expenditure. Commercially available wearable activity monitors have the function to estimate REE by utilizing their own proprietary algorithms. However, no evidence exists about the relative accuracy of the monitors in estimating REE in relation to measured REE. PURPOSE: The purpose of this study was to evaluate the validity of the Fitbit Flex (FF) and SenseWear Mini Armband (SWA) in REE estimates in adults. METHODS: Sixty healthy (26.4±5.7 yrs) males (n=30) and females (n=30) volunteered to participate in the study. The REE measurement was performed in the morning (i.e., 6:00-9:00am) after a 10-hour fast, following previously published guidelines. Estimates of REE from the FF and SWA monitors were obtained from the corresponding software and website. These REE estimates were compared to REE measured from open-circuit indirect calorimetry (IC) and estimated using the Institute of Medicine (IOM) and World Health Organization (WHO) prediction equations. RESULTS: Analyses of covariance (ANCOVA) showed no significant effects of gender for any of the comparisons with REE from IC; therefore, males and females were combined for all analyses. REE from FF, SWA, IOM, and WHO were (kcals/day) (means±SD): 1554.3±249.3, 1587.1±247.7, 1528.0±213.0, and 1559.0±232.0, respectively. Mean absolute percentage errors were: 10.85±8.8%, 9.53±8.2%, 9.31±8.4%, and 10.8±8.7% for the FF, SWA, IOM, and WHO, respectively. Pearson correlation coefficients for the FF, SWA, IOM, and WHO in relation to IC were 0.635, 0.640, 0.657, and 0.683, respectively. No significant differences (p-values < 0.05) were observed between the measured REE, FF, SWA, IOM, and WHO in REE estimates. CONCLUSION: The estimates of REE from the FF, SWA, IOM, and WHO equation were similar to measured REE. The relatively high accuracy of the FF and SWA in estimating REE suggests that they have great potential to be utilized in intervention and surveillance studies aimed at precisely estimating total daily energy expenditure.