Validation of remote anthropometric measurements in a rural randomized pediatric clinical trial in primary care settings

Abstract

Rural children are more at risk for childhood obesity but may have difficulty participating in pediatric weight management clinical trials if in-person visits are required. Remote assessment of height and weight observed via videoconferencing may provide a solution by improving the accuracy of self-reported data. This study aims to validate a low-cost, scalable video-assisted protocol for remote height and weight measurements in children and caregivers. Families were provided with low-cost digital scales and tape measures and a standardized protocol for remote measurements. Thirty-three caregiver and child (6-11 years old) dyads completed remote (at home) height and weight measurements while being observed by research staff via videoconferencing, as well as in-person measurements with research staff. We compared the overall and absolute mean differences in child and caregiver weight, height, body mass index (BMI), and child BMI adjusted Z-score (BMIaz) between remote and in-person measurements using paired samples t-tests and one sample t-tests, respectively. Bland-Altman plots were used to estimate the limits of agreement (LOA) and assess systematic bias. Simple regression models were used to examine associations between measurement discrepancies and sociodemographic factors and number of days between measurements. Overall mean differences in child and caregiver weight, height, BMI, and child BMIaz were not significantly different between remote and in-person measurements. LOAs were - 2.1 and 1.7 kg for child weight, - 5.2 and 4.0 cm for child height, - 1.5 and 1.7 kg/m² for child BMI, - 0.4 and 0.5 SD for child BMIaz, - 3.0 and 2.8 kg for caregiver weight, - 2.9 and 3.9 cm for caregiver height, and - 2.1 and 1.6 kg/m² for caregiver BMI. Absolute mean differences were significantly different between the two approaches for all measurements. Child and caregiver age were each significantly associated with differences between remote and in-person caregiver height measurements; there were no significant associations with other measurement discrepancies. Remotely observed weight and height measurements using non-research grade equipment may be a feasible and valid approach for pediatric clinical trials in rural communities. However, researchers should carefully evaluate their measurement precision requirements and intervention effect size to determine whether remote height and weight measurements suit their studies.Trial registration: ClinicalTrials.gov NCT04142034 (29/10/2019).