Abstract
BACKGROUND: Bioelectrical impedance analysis (BIA) devices are widely used for body composition analyses, but their accuracy for body impedance measurements can vary significantly between BIA device models. Calibration of BIA device is essential to ensure the reliability of measurement data, particularly for clinical and research purposes. This study aims to evaluate the electric impedance measurement abilities of multi-frequency BIA devices using a home-built body impedance imitator (BII), designed to simulate human body impedance, and to assess the measurement uncertainty based on calibration with internationally accredited reference standards. METHODS: Two different models of multi-frequency BIA device with similar specifications and measurement schemes were evaluated using the BII as a reference standard. First, the impedance values of the BII were calibrated using an LCR meter, traceable to national impedance standards. Next, the impedance measurements were performed using both models based on the standard BII over a wide frequency range (1 kHz to 1 MHz) to analyze the accuracy, linearity, contact resistance effects, etc. of the BIA devices. RESULTS: Despite having similar specifications, two models of BIA device exhibited notable differences in their measurement performance. Model A showed better accuracy and consistency, with measurement uncertainty of approximately 1.4% at a 95% confidence level (k = 2). In contrast, Model B showed higher deviations, particularly in trunk impedance measurements. The study found that the BIA device calibration procedure could effectively quantify the measurement uncertainty and ensure the reliability of measurement data provided by the BIA device. CONCLUSION: This study demonstrates that BIA devices can be calibrated using a standard procedure with a stable reference such as the BII, which traces back to internationally accredited impedance standards. The results show that the calibration significantly improves the reliability of BIA device measurements, providing a level of confidence for the measurement data.