Predicting fall risk in elderly ındividuals: a comparative analysis of machine learning models using patient...

Area of Science:

• Gerontology
• Biomedical Engineering
• Machine Learning in Healthcare

Background:

• Falls are a significant risk for elderly individuals, leading to injury and decreased quality of life.
• Accurate fall risk assessment is crucial for implementing preventive strategies.
• Traditional methods may not always capture the complexity of fall risk factors.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting fall risk in the elderly.
• To compare the predictive accuracy of models using different combinations of patient data.
• To determine if advanced balance assessment tools are necessary for accurate fall risk prediction.

Main Methods:

• A cohort of 120 elderly individuals was studied.
• Data collected included fall status, physical characteristics, medical history, audiometry, functional balance tests, and sensory organization test (computerized dynamic posturography).
• Machine learning models were trained and validated using these datasets.

Main Results:

• A machine learning model integrating co-morbidities, physical characteristics, and functional balance tests achieved 100% accuracy in predicting fall risk.
• Models utilizing only co-morbidities and physical characteristics, functional balance tests, or the sensory organization test showed accuracies of 87.5%, 83.34%, and 91.66%, respectively.
• The comprehensive model significantly outperformed models using limited datasets.

Conclusions:

• Advanced balance assessment systems are not essential for effective fall risk prediction.
• Simple functional balance tests, combined with patient characteristics and co-morbidities, provide a highly accurate method for fall risk assessment using machine learning.
• This approach offers a cost-effective and efficient alternative for clinical practice.