Validity of heart rate, pedometry, and accelerometry for predicting the energy cost of children's activities

Heart rate telemetry is frequently used to estimate daily activity in children and to validate other methods. This study compared the accuracy of heart rate monitoring, pedometry, triaxial accelerometry, and uniaxial accelerometry for estimating oxygen consumption during typical children's activities. Thirty Welsh children (mean age 9.2 +/- 0.8 yr) walked (4 and 6 km/h) and ran (8 and 10 km/h) on a treadmill, played catch, played hopscotch, and sat and crayoned. Heart rate, body accelerations in three axes, pedometry counts, and oxygen uptake were measured continuously during each 4-min activity. Oxygen uptake was expressed as a ratio of body mass raised to the power of 0.75 [scaled oxygen uptake (s(V) over dot O-2)]. All measures correlated significantly (P < 0.001) with s(V) over dot O-2. A multiple-regression equation that included triaxial accelerometry counts and heart rate predicted s(V) over dot O-2 better than any measure alone (R-2 = 0.85, standard error of the estimate = 9.7 ml.kg(-0.75).min(-1)). The best of the single measures was triaxial accelerometry (R-2 = 0.83, standard error of the estimate = 10.3 ml.kg(-0.75).min(-1)). It is concluded that a triaxial accelerometer provides the best assessment of activity. Pedometry offers potential for large population studies.