Developing a Tool for Predicting Changes in Parameters of Bioimpedance Analysis of Human Body Composition

The parameters of bioimpedance analysis of human body composition (BIA) are investigated to diagnose current medical conditions and assess the efficacy of therapeutic interventions. In some cases, it is important to predict changes in these parameters during ongoing treatment or rehabilitation. The purpose of this study was to develop a graphical tool for the visual assessment of the most probable movements of data points representing current BIA parameter pair values on plots of two-dimensional bioimpedance analysis of human body composition (2DBIA). The baseline data for the development of 2DBIA references were derived from records of 1,083,275 patients aged 5 to 80 years, who underwent body composition assessments at health centres between 2009 and 2015. The benchmarks for predicting the direction of changes in 2DBIA plots can be either the rectangular domains of the medical norm or the boundaries of selected reference centiles (e.g., the 50th percentile) for the population in question. To evaluate the adequacy of predicting BIA parameter values, the study utilised data from 130 adult patients at the Ivanovo Regional Clinical Hospital (Ivanovo, Ivanovo Region, Russian Federation) who were monitored for fat reduction and had at least three BIA measurements. The median interval between consecutive observations was 37 [33; 48] days, with an interquartile range of 33 to 48 days, and the median number of observations was 5 [4; 8]. A total of 798 observations were recorded. The percentage of points on the 2DBIA plot that fell outside the predicted region ranged from 0.6% in the Total Body Water/Basal Metabolic Rate pair to 11.7% in the Fat Mass Index/Fat-Free Mass Index pair. Points falling outside the prediction zone almost always indicated non-compliance with the prescribed treatment regimen by the patients.