Accelerometry-based gait analysis predicts falls among patients with a recent fracture who are ambulatory: a 1-year prospective study.

The aim of this prospective study was to assess whether accelerometry-based gait analysis is associated with the incidence of falls among individuals who are ambulatory and recently had a fracture, and to compare the ability to discriminate body function and structure and activity measures. A total of 100 patients who fulfilled our inclusion criteria were enrolled. At hospital discharge, the following outcomes were assessed: timed-up-and-go test, five-times-sit-to-stand test, one-leg standing test, 5-m gait test, and gait analysis using a wireless three-axis accelerometer. Root mean square (RMS) and autocorrelation (AC) values were analyzed on the basis of the acceleration waveform. Follow-up assessment of falls and subsequent fractures was performed 12 months after the fracture. Eighty-five percent of the participants completed the follow-up at 12 months. During the 12 months of follow-up, 61 falls were reported by 34 patients (31 women, three men), for a fall rate of 40.0% (34/85). The fall group had significantly poorer body function and structure and activity measures compared with the no-fall group. In accelerometry-based gait analysis, RMS and AC in the vertical axis and RMS in the anteroposterior axis were significantly lower in the fall group compared with the no-fall group. Multivariate logistic regression analysis showed that only AC in the vertical axis was a significant predictor of falls. This finding suggests that trunk acceleration AC in the vertical axis showed good discriminative ability for predicting the incidence of falls among patients who are ambulatory and recently had a fracture.