A gauge repeatability and reproducibility study applied to the measurements from the ENSANUT‐2012 anthropometric training

Abstract

We performed a Gauge Repeatability and Reproducibility study to data from the anthropometry training prior to the Mexican National Health and Nutrition Survey of 2012. A crossed random effects model was specified to estimate variance components of weight (kg) and height (cm) adult measurements. A study of bias and linearity was also performed taking a gold standard anthropometrist as reference. The measurement system accounted for 0.1% of the total weight variation and 0.4% of the total height variation. Compared to the gold standard, a small overall bias of − 0.024kg (IC 95%: −0.048, −0.001) was observed for weight and did not significantly changed with the size of the part measured (IC95%: −0.001, 0.001). No bias was found for height (IC95%:− 0.037−0.048) and it was not linearly associated with the size of the part measured (IC95%:−0.004, 0.005). Our results allow assessing the quality of the gauge system used in ENSANUT‐2012 for anthropometric measurements. Overall the measurement system is adequate in terms of its precision and no relevant biases were found.