Validity Of A Sleep Accelerometer For Assessing Physical Activity In Toddlers

Abstract

Research grade accelerometers that are valid for sleep assessment are rarely assessed for validity for physical activity assessment in pediatric populations, due to differences in recommended monitor placements and/or settings.1-2 The accelerometer used in this study is widely used in sleep research, but has not been validated for measuring physical activity in toddler populations.4-6 It is currently unknown if this accelerometer could be used as a valid method of measuring physical activity in toddlers. PURPOSE: To determine the validity of this accelerometer for measuring physical activity in toddlers. METHODS: An accelerometer was used to assess physical activity in eleven toddlers across different settings. Toddlers were recruited from a childcare center and from a registry of families with toddlers. Accelerometers were placed on the ankle of each toddler, and at least 30-minutes of both indoor and outdoor free-play activity were video-recorded for each toddler in either the childcare center or home setting. Activity level (1 - 5; 1 = sedentary, 5 = fast) was coded using direct observation, following the Observational System for Recording Physical Activity in Children - Preschool (OSRAC-P).7 Accelerometer count data were sampled in 15-s epochs and aggregated to 60-s for analysis. Association with OSRAC-P was assessed using Spearman correlation (p < 0.05). RESULTS: Nine participants (5 males, 12 - 36 months) with 30 or more minutes of video data were included in analysis. Data were collected over 894 minutes (99 ± 44.2 minutes per toddler). A moderate correlation was found between OSRAC-P and the accelerometer (R = 0.65, p < 0.001). CONCLUSION: There is a moderate correlation between OSRAC-P activity levels and accelerometer counts. The correlation in this study was higher than previous results reported in toddlers between OSRAC-P activity levels and counts from a hip-worn accelerometer (R = 0.52).8 Another study using similar ankle placement in toddlers, reported a higher correlation between Child Activity Rating Scale (CARS) and accelerometer counts (R = 0.75).9 Preliminary data suggest that this device may be a valid method of physical activity assessment in a toddler population. Supported by NIH/NINR