Save the Date!
ActiGraph Digital Data Summit 2021November 4 - 5 | Sign Up for Event Updates Now
Predicting Children's Free-living Energy Expenditure
- Added on June 16, 2011
Introduction The assessment of energy cost in paediatric populations has typically occurred in laboratory settings, which usually obtain higher validity coefficients. However, laboratory based protocols may be limited in the assessment of free-living behaviours as creating environments that reflect such behaviours is difficult . Few studies have provided children the option of engaging in activities of their own volition. The aim of the study was to predict children’s energy expenditure using multiple sites and measures when engaged in a range of free-living behaviours.
Methods Twenty eight children (13 boys, 15 girls) aged 10-11 years from one North-West school participated in the project. Children were fitted with a heart rate monitor, four GT1M Actigraphs (2 hip mounted, 2 wrist mounted), and the Metamax 3b system. Resting energy expenditure and heart rate was measured during 15 minutes of supine rest. Following this, children engaged in self-paced walking, self-paced running (5 min each), drawing, free-choice games and playground games (10 min each) in a randomised order, with 5 min of seated rest between each activity. During the activities, each child was observed using a modified SOFIT protocol. All activities took place on the school playground, with the exception of the drawing activity. Energy expenditure measured via indirect calorimetry was used as the criterion measure. Prediction of energy cost using a combination of physical activity measures was completed using multiple linear regressions with a stepwise backward method.
Results Energy cost was higher, but not significantly, during free-choice games (508.3 J·kg-1·min-1) compared to playground games (465.5 J·kg-1·min-1) and self-paced walking (361.6 J·kg-1·min-1). Wrist-mounted accelerometer counts were significantly higher during free choice games and playground games than during self-paced walking (p<0.001). The single best measure of predicting energy cost for free-living behaviours was wrist-mounted accelerometry (R2 ≥ 0.17). For multisite activity assessment, the variance explained in energy cost during free-living behaviours was greater than single-site activity measures (R2 ≥ 0.43).
Discussion and Conclusion Multiple regression equations explained a greater amount of variance in the energy cost of children’s free-living behaviours than any single measure alone. Notably, wrist counts (both dominant and non-dominant hand) were included in single and multiple regression equations, indicating that the energy cost of children’s free-living behaviours may be under-estimated when hip-mounted accelerometry is used. Developing energy expenditure equations during free-living activities in field-based settings may improve the prediction of physical activity in children.
References  Nilsson A, Brage S, Riddoch C, Anderssen SA, Sardinha LB, Wedderkopp N, et al. Comparison of equations for predicting energy expenditure from accelerometer counts in children. Scand J Med Sci Sports 2008, 18:643-650.
ICAMPAM- Glasgow 2011