Predicting balance and quantifying Romberg quotient reference ranges in older adults

Balance control can be quantified with measurements from force plates or wearable accelerometers. The purposes of our study were (1) to assess the explanatory power of accelerometer measures to predict center-ofpressure (CoP) sway-area rate across four standing balance tests, (2) to provide expanded Romberg quotient (RQ) reference ranges for older adults, and (3) to perform an agreement analysis between RQs derived from accelerometer measures with those from CoP sway-area rate. Twenty-six healthy older adults completed four 30s s tests standing on rigid and foam surfaces with eyes opened and closed on two occasions. Tests were performed on a force plate (1 variable) while wearing accelerometers (5 variables). The outcomes were assessed with linear mixed-effects models, RQs, and agreement analyses. CoP sway-area rate could be significantly predicted by different combinations of accelerometer variables for each balance condition (R2>0.45). Despite some accelerometer RQs not being significantly different (p>0.12) from corresponding CoP sway-area rate RQs, there was low agreement between these RQs as determined by intraclass correlation coefficients, Bland-Altman plots, and mountain plots. Clinicians with access to wearable accelerometers but not force plates can use our models to estimate CoP sway-area rate for each standardized test of standing balance. The expanded RQ reference ranges provide information about potential sensory impairments that can influence balance control in older adults.