Estimating model of sedentary behavior with tri-axial accelerometer in elementary school children

Abstract

Several recent studies reported that a lack of moderate to vigorous physical activity (MVPA) in combination with a high degree of sedentary behavior (SB) is associated with health problems including overweight and obesity in children, as well as psychosocial stress. Therefore, it is important that methods are developed to objectively evaluate both MVPA and SB. The aim of this study was to redevelop the existing equation for estimating SB to improve its accuracy. Healthy boys (n = 42) and girls (n = 26) attending primary school were invited to participate in this study. Participants were asked to perform 2 SB tasks, which were desk work and Nintendo DS, 2 light intensity activities such as sweeping up and clearing away, and higher intensity activities such as sweeping up and throwing a ball, with a few minutes of recovery time between tasks. The tasks and activities were performed in order of PA intensity (lower to higher). All participants wore a triaxial accelerometer on their waist. In addition, they wore a facemask connected to a Douglas bag to gather respiratory gas samples while performing each activity. First, we proposed the two linear regression equations (TL), including an equation for SB, and another equation for light or higher intensity activities with a fixed intercept of 0.9. Moreover, we redeveloped a quadratic polynomial (QP) equation that takes into account all activities. Both models were demonstrated to improve the accuracy of estimations of PA (about 0.2 to 0.3 METs), including SB (about 1.0 METs) other than sweeping up and wiping floor, compared to the existing model.