Validity of an ultrasound device to measure bone mineral density.

Abstract

This study aims to examine the validity and reliability of the UltraScan650™, a portable ultrasound device, used to measure BMD at the 1/3rd radius position. Fifty-two female first responders and healthcare providers were assessed using DXA (forearm, femur, lumbar, and total body) and the UltraScan650™. Fat and lean mass were also assessed using the DXA. Pearson correlations, Bland-Altman plots, t-tests, and linear regressions were used to assess validity. Intra-class correlation (ICC) coefficients were used to assess reliability. Inter-rater reliability and repeatability were good (ICC = 0.896 [0.818; 0.942], p < 0.001) and excellent (ICC = 0.917 [0.785; 0.989], p < 0.001), respectively. BMD as measured by the UltraScan650™ was weakly correlated to the DXA (r = 0.382 [0.121; 0.593], p = 0.0052). Bland-Altman plots revealed that the UltraScan650™ underestimated BMD (-0.0569 g/cm2), this was confirmed with a significant paired t-test (p < 0.001). A linear regression was performed (0.4744 × UltraScan650™ + 0.4170) to provide more information as to the issue of agreement. Bland-Altman plots revealed a negligible bias, supported by a paired t-test (p = 0.9978). Pearson's correlation revealed a significant relationship (r = -0.771 [-0.862; -0.631], p < 0.0001) between adjusted UltraScan650™-DXA and the average of the two scans (i.e., adjusted UltraScan650™ and DXA), suggesting a proportional constant error and proportional constant variability in measurements of BMD from the UltraScan650™. The UltraScan650™ is not a valid alternative to DXA for diagnostic purposes; however, the UltraScan650™ could be used as a screening tool in the clinical and research setting given the linear transformation is employed.