Daily fall risk assessment for older adults using a balance sensor with machine learning

Area of Science:

• Gerontology
• Biomedical Engineering
• Artificial Intelligence in Healthcare

Background:

• Fall-risk assessment is crucial for preventing falls in older adults.
• Conventional balance tests are often time-consuming and impractical for frequent clinical use.
• There is a need for rapid, objective, and easily implementable fall-risk assessment tools.

Purpose of the Study:

• To evaluate a 30-second quiet-standing assessment using a balance sensor and machine learning.
• To classify fall history and assess fall risk in older adults.
• To determine the feasibility of this method for frequent, real-world use.

Main Methods:

• 394 older adults participated across laboratory, hospital, and home-based cohorts.
• A balance sensor captured plantar pressure data during quiet standing.
• Machine learning models analyzed center-of-pressure (CoP) and center-of-gravity (CoG) features for fall-risk classification.

Main Results:

• The random forest model achieved an AUC of 0.71 (eyes-open) in the laboratory cohort.
• External validation in the hospital cohort yielded an AUC of 0.68 and 72.0% accuracy.
• The home-based study demonstrated feasibility for repeated, short-duration balance measurements.

Conclusions:

• A brief sensor-based standing assessment with machine learning offers a practical approach for fall-risk assessment in older adults.
• This method shows potential as an improvement over conventional functional assessments.
• Further prospective, multi-center studies are required to confirm predictive validity and clinical utility.