Gait variability predicts real-life falls in high-functioning stroke survivors

Area of Science:

• Neurology
• Biomechanics
• Rehabilitation Science

Background:

• Over 60% of adults with stroke experience falls annually.
• High-functioning individuals with mild motor impairments face the greatest fall risk.
• Current predictors of falls are insufficient for this population.

Purpose of the Study:

• To identify sensitive predictors of falls in high-functioning adults with stroke.
• To compare the predictive power of gait variability versus traditional balance and mobility tests for falls.
• To determine the relative contribution of gait variability in predicting real-life falls.

Main Methods:

• Twenty-four independent walkers with stroke performed overground walking, Timed-up and go, and Berg balance scale.
• Gait speed, stride length variability, and stride time variability were quantified.
• History of falls in the past year was recorded.

Main Results:

• Stride length variability and Berg balance scale score were associated with previous falls.
• Multivariate analyses revealed stride length variability as a significant predictor of past falls (OR=2.73, p=0.03).
• A stride length variability cut-off of 3.98% demonstrated 75% sensitivity and 91.7% specificity for predicting falls (AUC=0.83).

Conclusions:

• Stride length variability during overground walking strongly predicts past falls in high-functioning adults with stroke.
• Gait variability is a more sensitive fall risk predictor than traditional balance and mobility tests in this cohort.
• Findings emphasize gait variability's importance for accurate fall risk assessment post-stroke.