Visual Analytics for Pattern Discovery in Home Care. Clinical Relevance for Quality Improvement

Area of Science:

• Gerontology
• Health Informatics
• Data Visualization

Background:

• Large datasets in home health pose interpretation challenges.
• Visual analysis can simplify complex information and reduce cognitive load.
• Understanding patient characteristics and outcomes is crucial for effective care.

Purpose of the Study:

• To apply visual analysis techniques to home health data.
• To identify clinically significant associations between patient characteristics and health outcomes.
• To examine older adult home health patients during their service period.

Main Methods:

• Utilized a dataset of 988 de-identified patients from 15 home health agencies.
• Employed SPSS Visualization Designer v1.0 for visual analysis using heat maps and histograms.
• Statistically evaluated significant patterns using correlation analysis.

Main Results:

• 150 visualizations yielded 69 potentially meaningful patterns, with 21 statistically significant associations.
• 14 associations between episode length, Charlson comorbidity index, and urinary issues remained significant.
• Visual analysis highlighted adverse associations between longer episodes, higher comorbidity, and impaired urinary function outcomes.

Conclusions:

• Demonstrated visual analysis for discovering novel patterns in older adult home health data.
• Identified high-needs subgroups within the patient population.
• Visualizations can guide clinicians in identifying improvement areas and optimal intervention timing.