Reliability and validity of a new accelerometer-based device for detecting physical activities and energy expenditure

Abstract

BackgroundObjective assessments of sedentary behavior and physical activity (PA) by using accelerometer-based wearable devices are ever expanding, given their importance in the global context of health maintenance. This study aimed to determine the reliability and validity of a new accelerometer-based analyzer (Fibion) for detecting different PAs and estimating energy expenditure (EE) during a simulated free-living day.MethodsThe study consisted of two parts: a reliability (n = 18) and a validity (n = 19) test. Reliability was assessed by a 45 min protocol of repeated sitting, standing, and walking (i.e., 3 × 15 min, repeated twice), using both Fibion and ActiGraph. Validity was assessed by a 12 h continuous sequence tasks of different types (sitting, standing, walking, and cycling) and intensities (light [LPA], moderate [MPA], and vigorous [VPA]) of PA. Two Fibion devices were worn on the thigh (FT) and in the pocket (FP), respectively and were compared with criteria measures, such as direct observation (criterion 1) and oxygen consumption by a portable gas analyzer, K4b2 (criterion 2).ResultsFT (intra-class correlation coefficients (ICCs): 0.687–0.806) provided similar reliability as the Actigraph (ICCs: 0.661–0.806) for EE estimation. However, the measurement error (ME) of FT compared to the actual time records indicated an underestimation of duration by 5.1 ± 1.2%, 3.8 ± 0.3% and 14.9 ± 2.6% during sitting, walking, and standing, respectively. During the validity test, FT but not FP showed a moderate agreement but lager variance with the criteria (1 and 2) in assessing duration of sitting, long sitting, LPA, MPA, and VPA (p > 0.05, ICCs: 0.071–0.537), as well as for EE estimation of standing, LPA, MPA, and VPA (p > 0.05, ICCs: 0.673–0.894).ConclusionsFT provided similar reliability to that of the Actigraph. However, low correlations between subsequent measurements of both devices indicated large random MEs, which were somewhat diminished during the simulated 12 h real-life test. Furthermore, FT may accurately determine the types, intensities of PA and EE during prolonged periods with substantial changes in postures, indicating that the location of the accelerometer is essential. Further study with a large cohort is needed to confirm the usability of Fibion, especially for detecting the low-intensity PAs.

The Validity of MotionSense HRV in Estimating Sedentary Behavior and Physical Activity under Free-Living and Simulated Activity Settings

“…Additionally, the reliability of the ActiGraph accelerometer for measuring sedentary behavior and PA under free-living conditions was confirmed in a previous study (standard error of the measurement < 11.2%) [ 35 ]. Accordingly, GT9X was previously used as a criterion measure to investigate the validity of consumer-based activity monitors in estimating PA in free-living conditions [ 36 , 37 ]. For that reason, the current study chose GT9X as a criterion measure to examine the validity of the MotionSense HRV in estimating time spent in sedentary behavior and PA during free-living conditions.…”

Does free public transit increase physical activity and independent mobility in children? Study protocol for comparing children’s activity between two Finnish towns with and without free public transit

“…Fibion® is valid in detecting sitting and continuous uninterrupted sitting periods against direct observation, and light physical activity, MVPA and total energy expenditure against indirect calorimetry [21]. Moreover, Fibion® gives estimates of activity types including standing (mean difference 17,2 min/12 h day, limits of agreement (LoA) -12,9 to 47,3 min), walking (mean difference − 17,3 min/12 h day, LoA − 47,4 to 12,8 min) and cycling duration (mean difference − 6,8 min/12 h day, LoA − 18,7 to 5,0 min) [34]. Fibion® has been used to monitor children's daily activity previously [35].…”

Wearable Activity-Tracking Device Feature Preference Amongst South African Generation Y Students

“…Furthermore, 17 percent highlighted wearable activity tracker acceptance, adoption, and abandonment; 10 percent focused on self-monitoring; and 6 percent pertained to privacy. These studies, along with more recent research, focus mostly on wearable device reliability and/or validity, comparing different brands and specific models of devices, the accuracy of such devices, medical attributes and patient treatment by means of using wearable technology, promoting physical activity or physical activity intervention, device acceptance and factors influencing the adoption or abandonment of wearable activity trackers, and the privacy concerns posed by using these types of devices (Bassett, Freedson & John, 2019;Bunn, Wells, Manor, & Webster, 2019;Muller, 2019;Jones et al, 2018;Lamont, Daniel, Payne & Brauer, 2018;Muller, de Klerk & Bevan-dye, 2018;Yang, Schumann, Le & Cheng, 2018;Chu et al, 2017;Shinde et al, 2017;Steinert, Haesner & Steinhagen-Thiessen, 2017;Diaz et al, 2016;Kaewkannate & Kim, 2016;Lamb, Huang, Marturano & Bashir, 2016;Roe, Salmon & Twiggs, 2016;Wang et al, 2016;Cadmus-Bertram et al, 2015;Case, Burwick, Volpp & Patel, 2015;Kooiman et al, 2015;Fulk et al, 2014;Lee, Kim & Welk, 2014;Fausset et al, 2013;Noah, Spierer, Jialu & Bronner, 2013). However, the majority of these studies used smaller samples (between 0-100) and the research and data collection methods were mostly observational, experimental, interventional, or qualitative in nature, with the exception of a few that were quantitative and survey-based.…”