Ambient Monitoring of Gait and Machine Learning Models for Dynamic and Short-Term Falls Risk Assessment in People...

Area of Science:

• Gerontology
• Biomedical Engineering
• Artificial Intelligence

Background:

• Falls are a major cause of injury and death in older adults with dementia living in long-term care facilities.
• Accurate, up-to-date fall risk assessment is crucial for implementing timely preventive interventions.

Purpose of the Study:

• To develop and validate machine learning models for predicting the short-term risk of falls in older adults with dementia.
• To identify key predictors of falls, including clinical assessments, gait analysis, and medication data.

Main Methods:

• Longitudinal data from 54 participants with dementia were used to train machine learning models.
• Data included baseline clinical assessments, daily medication records, and continuous gait monitoring via computer vision.
• Model performance was evaluated using leave-one-subject-out cross-validation, sensitivity, specificity, and AUROC.

Main Results:

• The best model achieved an AUROC of 76.2, with 72.8% sensitivity and 73.2% specificity in predicting falls within 4 weeks.
• Ambient gait analysis was a critical feature, as models excluding it had significantly lower performance (AUROC 56.2).
• Differential contributions of clinical assessments, gait analysis, and medication intake to fall prediction were identified.

Conclusions:

• Machine learning models, particularly those incorporating ambient gait analysis, show promise for accurately predicting fall risk in older adults with dementia.
• This technology has the potential to enable targeted interventions, reducing fall-related morbidity and mortality in long-term care.
• External validation is the next step toward clinical implementation.