However, the reported energy expenditure from these devices should be interpreted with caution, given their potential bias and error. Energy expenditure estimations of single speeds were overestimated by the Fitbit One and underestimated by the Garmin Vivofit.Įnergy expenditure reported by the devices distinguished between walking and running, with a general increase as exercise intensity increased. Bland-Altman plots showed proportional bias for the Fitbit One and Garmin Vivofit. However, only the Garmin Vivofit correlated significantly (p< 0.05 r = 0.346) with energy expenditure estimations from criterion across running speeds (2.22-3.33 ms - 1). Fitbit One, Garmin Vivofit and Jawbone UP correlated significantly (p < 0.05 r = 0.729 0.711 0.591) with criterion across all walking speeds (0.70-1.80 ms - 1). Mean percentage difference assessed magnitude of difference between estimated and criterion energy expenditure for each speed.Įnergy expenditure estimates from the Fitbit One and Garmin Vivofit correlated significantly (p< 0.01 r= 0.702 0.854) with criterion across all gait speeds (0.70-3.33 ms -1). Paired t-tests, correlation coefficients and Bland-Altman plots assessed agreement and proportional bias. Estimation of energy expenditure from each tracker was compared to measurement from indirect calorimetry (criterion). Determining validity of such trackers will inform the interpretation of the data they generate.įourteen adults walked at 0.70, 1.25, 1.80 ms -1 and ran at 2.22, 2.78, 3.33 ms -1 on a treadmill wearing a Fitbit One, Garmin Vivofit and Jawbone UP.

To determine the validity of energy expenditure estimation made by the Fitbit One, Garmin Vivofit and Jawbone UP activity trackers during treadmill walking and running.

Garmin vivofit