Accelerometer Cutpoint Non-Equivalence in Preschool Children

Abstract

There are no universally accepted accelerometer cutpoints to quantify moderate-to-vigorous physical activity (MVPA). Estimates of MVPA from one set of cutpoints cannot be directly compared to MVPA estimates using different cutpoints, even when the same outcome units are reported (MVPA min·d-1). PURPOSE: To illustrate the utility of an equating system that translates MVPA estimates from one set of cutpoints into another set of cutpoints. Such an equating system would have a major impact on systematic reviews and would allow for reported data to be maximally utilized to inform public health policy. METHODS: The data from a large preschool project (N = 419, 3-5yr-olds, CHAMPS) was used to conduct the analyses. Conversions were made among the four most common ActiGraph cutpoints used to derive MVPA in preschool children: Pate (PT), Sirard (SR), Puyau (PY) and Freedson Equation (FR). PT cutpoints served as the criterion MVPA estimate, on which SR, PY, and FR estimates were equated, based on methodological quality of PT's initial calibration study - indirect calorimetry, free-living activities, preschool-age sample. A 10-fold cross-validation procedure was used to develop prediction equations using MVPA from PT as the dependent variable, with MVPA estimated from SR, PY, and FR serving as the predictor variable. Absolute percent error (APE) were calculated and Bland Altman (BA) plots constructed to evaluate agreement across the conversions. RESULTS: The mean levels of MVPA for the total sample ranged from 39.5 (SR) to 269.0 (FR) min·d-1. Across the prediction models, the mean proportion of variance explained (R2) was 0.78 (range 0.71-0.87). The median APE was 15.8% (range 12%-20.2%). The BA plot mean difference across conversions was: PY to PT 0.26 minutes (limits of agreement [LOA] -29.0 to 29.5); FR to PT -0.47 minutes, LOA -38.2 to 37.3; and SR to PT 0.22 minutes (LOA -32.3 to 32.7). CONCLUSION: The prediction equations developed in this study provide a way to make direct comparisons between studies employing different cutpoints in preschool aged children. Using the prediction equations developed here gives public health researchers and policy makers a more concise picture of physical activity levels of preschool aged children. *Supported by a grant from the National Institutes of Health (NIH), R01 HD043125.