Early gait segments may be sufficient: fall risk assessment does not require steady-state walking

 What is your competent? doctors EXACT FALL PREVENTION PROTOCOL? Doesn't have one I bet; A FIREABLE OFFENSE! We have forcefully clean out a lot of incompetent dead wood in stroke.

Early gait segments may be sufficient: fall risk assessment does not require steady-state walking

• Jianlei Fang,
• Peng Wu,
• Yibin Li,
• Rui Song,
• Damiano Zanotto,
• Sunil K. Agrawal &
• Huanghe Zhang 

We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.

Abstract

Falls remain a significant health concern for older adults, highlighting the need for efficient and accurate fall risk screening. Although wearable inertial measurement units provide accessible gait analysis, it remains unclear whether fall-history classification requires gait parameters computed over fully stabilized walking sequences or whether discriminative information may already be present in the early portion of the walking sequence before parameters converge. This study analyzes foot-mounted IMU data from two independent cohorts: the publicly available GSTRIDE dataset and a private dataset collected by our team. After preprocessing, the analytical samples included 147 GSTRIDE participants (71 fallers and 76 non-fallers) and 95 participants from our dataset (16 fallers and 79 non-fallers) recruited from senior living facilities. Faller status was defined using retrospective fall-history labels. Across cumulative and sliding window feature extraction strategies, variability-based gait parameters required a large number of strides to achieve stable reliability, particularly among fallers. Nevertheless, strong discriminative potential was consistently observed using gait segments obtained prior to full parameter stabilization. Window-based statistical analyses further showed that significant early-window variability differences were present in the GSTRIDE dataset but not in our dataset, despite comparable classification trends. These findings indicate that full parameter stabilization is not a prerequisite for effective fall-history classification. Instead, gait segments from the early portion of walking sequences can provide useful discriminative information, offering a practical alternative to conventional approaches that rely on prolonged steady-state walking recordings.