Abstract
INTRODUCTION: Mobile monitoring devices offer an opportunity to characterize physical health recovery in children who survive critical illness. METHODS: To validate the BioIntelliSense BioButton® as a pediatric activity monitor, we studied healthy children (2-17 years-old) who wore the BioButton® device and an ActiGraph wGT3X-BT accelerometer, and a study team member documented activity in 1 min intervals (gold standard) during 45 min of scripted activities. In two-thirds of the cohort (derivation cohort), we identified BioButton activity count thresholds to differentiate activity levels based on highest Youden indices. Thresholds were applied to the remainder of the cohort (validation cohort) to determine sensitivity and specificity [95% confidence interval (CI)]. We also evaluated BioButton activity designations compared with accelerometer designations and calculated agreement between BioButton-measured body position and the activity log. RESULTS: Forty-five participants provided a median 43 (IQR 41, 44) analyzable minutes. Sensitivity and specificity of derived BioButton thresholds were 0.78 (95% CI: 0.69, 0.88) and 0.95 (95% CI: 0.90, 0.97) to identify moderate or vigorous activity (MVPA) and 0.91 (95% CI: 0.87, 0.95) and 0.98 (95% CI: 0.98, 0.98) to identify sedentary behavior. Sensitivity and specificity compared with the accelerometer were 0.52 (95% CI: 0.45-0.60) and 0.88 (95% CI: (95% CI: 0.84, 0.93) to identify MVPA and 0.92 (95% CI: 0.89-0.96) and 0.70 (95% CI: 0.67, 0.73) to identify sedentary behavior. The BioButton accurately identified position during 1,125 of 1,432 (78.6%) minutes. DISCUSSION: The BioButton device accurately identified physical activity and body position in children and may be a useful tool to quantify physical activity as an outcome in future trials.