Association of Prospective Falls in Older People With Ubiquitous Step-Based Fall Risk Parameters Calculated From...

Area of Science:

• Gerontology
• Biomedical Engineering
• Data Science

Background:

• Traditional fall risk assessments often rely on dedicated, body-fixed inertial sensors, limiting user adoption.
• Ambulatory gait monitoring is increasingly recognized as a valuable complementary assessment tool.
• Current sensor-based gait analysis methods face challenges with sensor placement variability.

Purpose of the Study:

• To propose novel step-based fall risk parameters derived independently of sensor location.
• To evaluate the association of these sensor-placement-agnostic parameters with prospective falls in older adults.
• To leverage a ubiquitous step detection method for gait analysis.

Main Methods:

• Reanalysis of ambulatory inertial data from 301 community-dwelling older adults (StandingTall study).
• Utilized the Smartstep method for sensor-placement-agnostic step detection and gait parameter calculation (variability, complexity, intensity, quantity).
• Employed stepwise backward elimination for parameter reduction and negative binomial regression to assess fall association (AUC metric).

Main Results:

• A predictive model achieved an Area Under the Curve (AUC) of 0.69, comparable to fixed-sensor models.
• Increased fall risk correlated with higher gait variability (stride time coefficient of variance), intensity (cadence), and quantity (steps).
• Lower gait complexity (sample entropy, fractal exponent) was also associated with higher fall risk.

Conclusions:

• The proposed method enables accurate fall risk assessment irrespective of sensor placement.
• This sensor-agnostic approach holds significant potential for ambulatory gait and fall risk monitoring.
• Findings support the use of consumer-grade devices for effective fall risk assessment.