IFRA: A Machine Learning-Based Instrumented Fall Risk Assessment Scale Derived from an Instrumented Timed Up and Go...

Area of Science:

• Biomedical Engineering
• Rehabilitation Science
• Machine Learning in Healthcare

Background:

• Falls are a significant health issue for stroke survivors, requiring improved risk assessment.
• Traditional fall risk scales may miss crucial mobility measures.
• The Instrumented Fall Risk Assessment (IFRA) scale is proposed to address these limitations.

Purpose of the Study:

• To develop and validate the novel Instrumented Fall Risk Assessment (IFRA) scale.
• To utilize machine learning on Instrumented Timed Up and Go (ITUG) test data for fall risk stratification.
• To compare IFRA's performance against traditional clinical fall risk assessment tools.

Main Methods:

• A two-step machine learning approach was used to develop the IFRA scale.
• Predictive mobility features were identified from ITUG data (accelerations, angular velocity).
• IFRA's performance was evaluated against standard TUG and Mini-BESTest in 142 participants.

Main Results:

• Machine learning identified key predictors: vertical/medio-lateral acceleration and angular velocity.
• IFRA showed a significant association with fall status (p = 0.004).
• IFRA outperformed comparative scales by identifying more actual fallers as high-risk.

Conclusions:

• The IFRA scale shows potential as an automated tool for fall risk stratification in post-stroke patients.
• IFRA demonstrates promising capability in identifying individuals at high risk of falling.
• Further validation in larger cohorts is necessary before clinical implementation.