Children's physical activity assessed with wrist- and hip-worn accelerometers.
Medicine and Science in Sports and Exercise 2014 ; 46: 2308-16.
Rowlands AV, Rennie K, Kozarski R, Stanley RM, Eston RG, Parfitt GC, Olds TS
DOI : 10.1249/MSS.0000000000000365
PubMed ID : 24781890
PMCID : 0
Abstract
Recently, triaxial raw acceleration accelerometers have become available from GENEActiv and ActiGraph; both are designed for wrist and hip wear. It is important to determine whether the output from these monitors is comparable with the wealth of data already collected from the hip-worn, epoch-based, uniaxial ActiGraph.
This study aimed to assess the concurrent validity of measures of total activity and time spent at different activity intensities from the GENEActiv relative to the ActiGraph GT3X+.
Fifty-eight children age 10-12 yr wore two accelerometers at the hip (ActiGraph GT3X+ and GENEActiv) and one at the wrist (GENEActiv) for 7 d. Wear time was matched for all monitors before analysis.
Mean daily accelerometer output, time spent sedentary, and time in moderate-to-vigorous physical activity (MVPA) from the hip- or wrist-worn GENEActiv were strongly correlated with the corresponding output from the hip-worn ActiGraph (r > 0.83, P < 0.001). However, less time was estimated to be sedentary and more time was estimated to be MVPA using the hip- or wrist-worn GENEActiv (Phillips cut points) than that when using the Evenson vertical axis cut points with the hip-worn ActiGraph. Output from the vertical axis ActiGraph cut points could be predicted with 95% limits of agreement, equating to 23%-28% and 33%-35% of the mean value, by the hip- and wrist-worn GENEActiv, respectively.
The assessment of children's activity level, time spent sedentary, and time in MVPA estimated from the hip- or wrist-worn GENEActiv seems to be comparable with that of the uniaxial ActiGraph. On the basis of the strong linear correlations, ActiGraph output can be predicted from the hip- or wrist-worn GENEActiv for comparative purposes at the group level. However, because of relatively wide limits of agreement, individual-level comparisons are not recommended.