Digital bathroom scales with open source software provide valid dynamic ground reaction force data for assessment and biofeedback

Abstract

BACKGROUND: Measuring dynamic vertical ground reaction force allows for assessment of important clinical and physical capacity factors such as weight bearing asymmetry, force distribution, and rate of force development. However, current technologies for accurately assessing ground reaction force are typically expensive.

RESEARCH QUESTION: The aim of this study was to examine the validity and reliability of obtaining static and dynamic ground reaction force data from low-cost modified digital bathroom scales.

METHODS: Four modified bathroom scales, two units each of two different brands, were examined. Repeated mechanical loading trials were performed with known loads ranging from 0.01 to 65 kg, with acquired data compared against the known loading to calculate accuracy, hysteresis, and non-linearity. Dynamic trials consisting of 5 times sit-to-stand and weight-shifting were performed by 32 adults. Absolute and relative agreement intraclass correlation coefficient, and Pearson's and Spearman's correlations were performed to determine validity and reliability for the mechanical tests. Bland-Altmann plots were created for each comparison. Mean absolute error (MAE) and unbiased cross-correlation were performed on the dynamic data, comparing the calibrated data to the known values from a Bertec force platform.

RESULTS: Hysteresis and non-linearity were excellent (<0.2 % full-scale), and mechanical test results showed excellent reliability and validity. Cross-correlation results for the dynamic data were excellent, however MAE for the more rapid sit-to-stand task was higher than the slower weight-shifting test. This may have been due to the low default sampling rate for the lowest noise setting of the HX711 amplifier (10 Hz).

SIGNIFICANCE: In summary, our results suggest that digital bathroom scales can be easily and inexpensively modified to obtain accurate vertical ground reaction force data, with sensitivity to detect changes of as little as 0.01 kg.