Guidelines For Using Accelerometer Ambient Light Sensor Data To Assess Young Children’s Outdoor Time

PURPOSE: Develop analytic guidelines for processing accelerometer-derived ambient light sensor data to quantify children’s outdoor time, focusing on identifying optimal LUX cutoffs to distinguish indoor from outdoor time and considering device placement (hip vs. wrist). METHODS: Data were obtained from 1073 children ages 3-5 years who participated in two cluster randomized controlled trials in early care and education centers (ECE). Both trials collected ActiGraph GT3X+ acceleration and light sensor lux data, with children wearing the device on the hip in one trial and on the wrist in the other, alongside full-day ECE behavioral observations with timestamped indoor and outdoor periods. A SAS macro was used to determine lux values cutoffs to distinguish indoor/outdoor classifications to identify the range that optimized percent agreement (%agree), Kappa, sensitivity + specificity (SENSPC), and minimized false positive(FP) and false negative (FN) rates. Area under the curve (AUC) and error estimates for minutes of outside time were also computed. Analyses were stratified by placement (hip vs. wrist) to examine differences. RESULTS: In total, there were 556,543 minutes of observation and matched accelerometer LUX data, with 13% classified as outside time. Optimal cutoffs for hip worn devices was 37 LUX and 223 LUX for wrist worn. These cutoffs for hip/wrist wear resulted in good agreement: AUC (0.886; 0.946), %agree (92.9%; 96.7%); kappa (0.678, 0.832), SENSPC (1.677; 1.832), and low balanced FP/FN rates (1.64%; 3.56%). Observation and Lux derived estimates of outdoor time were very similar for both hip and wrist device wear, with mean lux derived outdoor time of 51.9 (±38.4) minutes (hip) compared to 51.6 (±37.4) observed minutes. For wrist placement, lux derived outdoor time was 50.5 (±43.7) minutes compared to 50.5 (±45.5) minutes based on observations. CONCLUSIONS: Findings show differences in lux cutoffs to determine indoor and outdoor time in young children based on wear placement. With growing evidence of the importance of outdoor time on youth health and wellbeing, along with the ubiquitous use of accelerometry, these results can guide future research on accurately classifying outdoor exposure with these devices. Supported by: 5R21HL170275-02