Validity Of Wearable Physical Activity Monitors During Activities Of Daily Living.

Abstract

There are a great number of new wrist worn physical activity monitors that have been developed in the last 5 years, but there is little or no information available to substantiate the validity of these wearable activity monitors in sleep period time (SPT) or energy expenditure (EE) in daily living. PURPOSE: The purpose of this study is to evaluate the validity of wearable activity monitors in SPT and EE under free-living environment. METHODS: Sixteen (27.1+5.4 years) healthy males (n=12) and females (n=4) participated in this study. Total SPT and EE were measured by eight monitors; Nike+Fuel Band SE (NFB), Garmin VivoFit (VF), Misfit Shine (MF), Fitbit Flex (FF), Jawbone UP (JU), Basis B1 (BB1), Polar Loop (PL), and Sense Wear Armband Mini (SWA). The monitors were worn for at least 23 hours to be included in final data analysis and no PA restriction was applied. The SWA and a sleep log were used as a criterion measure for SPT and EE, respectively. RESULTS: Total 24 hours of EE (Kcal) (means±SD) were 3123.1±730, 2355.71±528, 2393.86±612, 2240.33±273, 2075.47±87, 2901.5±658, 2926.58 ±576, 2695.25±600 for VF, JU, PL, BB1, FB, NFB, and MF, respectively. Mean absolute percent errors (MAPE) were calculated (means±SD) 24.4%±8.1, 23.3%±10.7, 15.6% ±17.4, 31.8% ±87.1, 9.5%±9.6, 5.7%±7.6, and 11.4%±11.2 for the VF, JU, PL, BB1, FB, NFB, and MF, respectively. SPT in minutes (mean±SD) were 464.0±66.4, 401.0±72.9, 460.7 ±61.4, 457.4±73.8, 504.3±85.5, 439.0±120.8, 443.8±49.2, and 457.8±112.8 for the log, SWA, VF, JU, PL, BB1, FB, and NFB, respectively. MAPE were calculated for SPT (mean±SD) 13.57% ±9.7, 0.71% ±7.6, 1.41 %±11.1, 8.69%±28.6, 5.38%±81.9, 4.38%±25.9, and 1.34%±69.8 for the SWA, VF, JU, PL, BB1, FB, and MF, respectively. ANOVA and post-hoc analyses with LSD indicated no significant differences were found with the FB, NFB, and MF in EE estimates. Additional post-hoc analyses with LSD for SPT revealed no significant difference (P>.05) in all monitors except SWA. CONCLUSION: The present study indicates that the FF, MS, and NFB are the most accurate wearable activity monitors when estimating EE and all monitors provide reasonable estimates of sleep period time, except SWA. Additional research is needed to validate these monitors utilizing Doubly-Labeled Water with a large sample.