Objective Assessment of Media Screens Effect on Physical Activity in Preschool Children

Media screen time is linked to physical and psychosocial health outcomes in preschool children (≤5 years), but evidence largely relies on parent reports or a single device type. This study proposes the use of a novel sensor, Screen Time Tracker (STT), to perform an objective assessment of the effect of screen time on physical activity in preschool children under free-living conditions. Egocentric images (10-s intervals) and triaxial accelerometer data (128 Hz) were collected from 25 children aged 3-5 years using the STT device. Accelerometer data were processed to estimate activity Counts Per Minute (CPM). A deep learning model was used to detect multiple types of active screens in images and compute Screen Counts per Minute (SCM; range, 0-6). On average, children exhibited lower CPM during screen interaction (1181 for SCM > 0) than during screen-free minutes (1628 for SCM = 0). Among screen interaction minutes, CPM decreased progressively with increasing SCM (from 1629 at SCM = 1 to 132 at SCM = 6). A multilevel negative binomial model with a log link, accounting for minutes nested within days and days nested within children, showed that higher SCM above a child's own daily mean was significantly associated with lower activity counts in a typical child on a typical day (b = -0.228, p < .001, exp[b] = 0.796). Thus, for each one-count increase in SCM above a child's daily mean, expected CPM decreased by approximately 20.4%. Furthermore, higher sustained screen interaction minutes were associated with lower physical activity. Children who continuously engaged with screens for 2 minutes showed mean CPMs of 1421, while continuous engagement for 15 minutes or more had a mean CPM of 693. Multi-level analysis showed that longer SSIM bouts were associated with lower mean CPM (b = -46.677, p = .005). This proof-of-concept study objectively shows that increased screen engagement is associated with lower physical activity intensity in preschool children. These findings provide useful evidence for epidemiologists, pediatricians, and behavioral scientists studying the health effects of media screen use in children.