Discriminating Between Fallers and Non-Fallers Using Kinematic Data from the Heel2Toe™ Wearable Sensor

Most falls occur while walking making gait quality a logical therapeutic target. Many tempo-ro-spatial variables have been implicated in increasing fall risk, but these are dependent upon kinematic parameters of the joints involved in the gait cycle. The widespread availability of wearable sensors has made the acquisition of kinematic data feasible and those related to the an-kle are most relevant as they relate most closely to causes of falls, trips, slips, and missteps. The purpose of this study is to estimate the extent to which measures of ankle angular velocity (AV) during walking are associated with falls. This is a comparative study of ankle AV metrics between people who have or have not experienced a fall in the past year. Data came from experimental use of the Heel2Toe™ sensor in a variety of settings including demonstrations and clinical research studies. The sample comprised 387 participants of whom 68 (17.6%) self-reported falling in the past year. Logistic regression with a natural cubic spline with 3 degrees of freedom identified AV of the angle at heel-strike to be the best discriminator between fallers and non fallers and the re-gression parameters were used to propose an algorithm to estimate fall risk. Applying the algo-rithm to the existing data yielded a range of probabilities from 0.0480 to 0.7245 depending on age of the person assessed. Further testing of this algorithm in different samples is warranted.