Center for Child Health, Behavior, and Development, Seattle Children's Research Institute, Seattle, WA, USA
Registration Is Open! Early Bird Pricing Expires June 30th
ActiGraph Digital Data Summit 2021November 4 - 5 | Learn more
Indoor versus outdoor time in preschoolers at child care
- Published on January 15, 2013
Being outdoors may have health benefits including being more physically active. Understanding the relationship between outdoor time and health is hampered by the difficulty of measuring outdoor time.
To examine the accuracy and validity of light-sensor and GPS methods for quantifying outdoor time among those aged 3-5 years at child care.
A total of 45 children (mean age 4.5 years, 64% boys) from five child care centers wore portable accelerometers with built-in light sensors and a separate GPS device around their waists during child care, providing 80,648 episodes (15 seconds each) for analysis. Direct observation (gold standard) of children being outdoors versus indoors was conducted for 2 days at each center. GPS signal-to-noise ratios, processed through the Personal Activity and Location Measurement System were used to define indoor versus outdoor locations. Receiver operating characteristic (ROC) analyses were used to determine thresholds for defining being indoors versus outdoors. Data were collected in Fall 2011, analyzed in 2012.
Mean observed outdoor time was 63 [±44; range: 18-152] minutes/day. Mean light-sensor levels were significantly higher outdoors. The area under the ROC curve for location based on light sensor for all weather conditions was 0.82 (range: 0.70 on partly cloudy days to 0.97 on sunny days); for GPS, it was 0.89. The light sensor had a sensitivity of 74% and specificity of 86%. GPS had a sensitivity of 82% and specificity of 88%.
A light sensor and a GPS device both distinguish indoor from outdoor time for preschoolers with moderate to high levels of accuracy. These devices can increase the feasibility and lower the cost of measuring outdoor time in studies of preschool children.