Validation Of A Wristband In Tracking Free-living Physical Activities Of Kindergarteners

PURPOSE: Heart rate (HR) was one of the most widely used indexes in determining the physical activity (PA) intensity, which could be measured by electrocardiogram (ECG) chest band as a criterion or by wristband as a field measure that is more practicable in daily life because of high convenience and low cost. While popular, the validity of the wristband have not been established for kindergarteners. The purpose was to validate a wristband designed specifically for kindergarteners (K-wristband). METHODS: A total of 46 kindergarteners (35 males & 11 females) aged 4 to 6 yr. (M ± SD: Age = 5.5 ± 0.5 yr , Height = 109.6 ± 5.0 cm , Weight = 19.0 ± 2.1 kg & BMI =15.8 ± 1.3) was recruited to wear Polar ECG chest straps and K-wristband on their right wrists for 30 min during free-living PAs in kindergarten, including sitting, drawing and aerobic dance, etc. Correlation and mean difference between ECG and K-wristband HR were examined. PA intensity levels (0-sedentary, 1-light, 2-moderate, 3-vigorous) were set based on the cutoff scores of 0.57, 0.64, 0.76 of the maximal hearts, which was established by the equation of “208 - .7(Age in yr.)”. Linear regression and receiver operating characteristic (ROC) were used to adjust K-wristband to find the optimal way to classify intensity, and Youden’s index was used to find the optimal cutoff scores in ROC method. RESULTS: The correlation between raw K-wristband data and ECG was high (r = 0.89), but the difference was also large (t = -68.98, ES = 0.25, p < .05). The intensity agreement (M ± SD of 46 participants = 0.80 ± 0.13, 0.79 in total) and Kappa (0.44 ± 0.20, 0.60) between K-wristband and ECG was checked on each participant as well as in total, i.e., merging all participants into one dataset. Linear regression was fitted with adjusted = 0.86*K-wristband+14.27, and ROC method gave the adjusted cutoff scores at 0.58, 0.61, 0.73 of HRmax. As a result, the agreement (M ± SD of 46 participants = 0.83 ± 0.11, 0.83 in total) and Kappa (0.46 ± 0.18, 0.64) of linear regression were higher than ROC (0.76 ± 0.19, 0.75; 0.40 ± 0.21, 0.53) with the ECG. CONCLUSIONS: With some statistical adjudication, heart rate information from K-wristband can be used to track and evaluate kindergarteners’ PA participation and new wristband may need doing similar statistical adjudication and validation before being used in practice.