Early Alert of Elderly Cognitive Impairment using Temporal Streaming Clustering.

Area of Science:

• Gerontology
• Neurology
• Data Science

Background:

• Over 44 million people globally have dementia, with numbers projected to triple.
• Cognitive impairment (CI) is significantly underdiagnosed, impacting the aging population's quality of life.
• Early CI detection and progression understanding are vital for effective treatment.

Purpose of the Study:

• To address the lack of unsupervised methods for characterizing temporal health changes in CI.
• To explore the potential of streaming clustering for early CI detection.
• To identify distinct health change patterns in older adults using clinical visit data.

Main Methods:

• Utilized a streaming clustering model on clinical visit data from the Mayo Clinic Study of Aging.
• Analyzed temporal health information to identify patterns associated with cognitive changes.
• Evaluated the model's efficacy in generating early alerts for potential CI incidents.

Main Results:

• The streaming clustering model successfully identified distinct health change patterns in older adults.
• Temporal characteristics integrated into the model showed promise in predicting CI.
• The approach demonstrated potential for early CI detection through unsupervised analysis of health data.

Conclusions:

• Streaming clustering offers a promising unsupervised approach for analyzing temporal health data.
• This method has the potential to enhance early identification and prediction of cognitive impairment.
• Incorporating temporal dynamics in health data analysis can improve CI diagnosis and management strategies.