01 Prediction of Falls in Community-Dwelling Older People using an Accelerometer: The Results of a 1-Year Prospective Study

Abstract Background Fall preventive exercise for community-dwelling older people in Japan has a certain effect. The purpose of this study was to examine the predictors of fall risk in older people engaging in preventive exercise. Methods In this prospective study, we recruited 162 community-dwelling older people. We had them complete a self-report questionnaire containing items on weekly exercise frequency, exercise duration, exercise efficacy, amount of pain, number of diagnosis chronic diseases, health anxiety level, vision impairment, difficulty using the stairs, insomnia, fear of falling, history of falls, and forgetfulness, as well as the Falls Efficacy Scale. Walking speed and trunk acceleration during walking were also measured in all participants. From trunk acceleration, we calculated five gait parameters: fluctuation, weight shift, lateral balance, anteroposterior balance, and rhythm. Participants were classified as fallers or non-fallers according to the incidence of falls over a 1-year period from baseline assessment. Results Of all participants, 19 (11.7%) were classified as fallers and 143 (88.3%) as non-fallers. In comparing the fallers and non-fallers, we found significant differences in their exercise duration, exercise efficacy, amount of pain, history of falls, walking speed, and the gait parameters of weight shift and anteroposterior balance. A logistic regression analysis revealed that walking speed (odds ratio: 0.049, 95% confidence interval [CI]: 0.005-0.265, p=0.008) and weight shift (odds ratio: 0.021, 95%CI: 0.000-0.877, p=0.043) were predictors of falls. Conclusion The gait parameter of weight shift is the amplitude ratio of the auto-correlation function from vertical acceleration. This gait parameter is an original metrics and might show the temporal structure around the terminal stance. Measurement of gait parameters, in addition to physical performance, using an accelerometer seems beneficial for fall risk identification among community-dwelling older people engaged in fall preventive exercise.