ABSTRACT
Physical activity has been measured with self-report instruments such as questionnaires and interviews, mainly in large-scale research. In children and adolescents, these instruments must be used with care, because of their inability to accurately recall intensity, frequency and duration of the activities (Sirard and Pate, 2001). However, gold-standard methods such as doubly labeled water and indirect calorimetry are expensive and unfeasible in large-scale epidemiological studies. Therefore, advancements in technology have increased in both number and type of objective physical activity measurement devices including accelerometers. Accelerometers are lightweight electronic devices, able to measure and store accelerations in one to three axis. Nevertheless, the accelerometer output value, typically called “counts”, still remains without a biological meaning. Thereby, different cut-points of physical activity intensity and predictive energy expenditure equations have been published. We examined the validity of two METs regression models of ActiGraph, the most widely used accelerometer (De Vries et al., 2009).
