Intelligent visualization and exploration of time-oriented data of multiple patients

Area of Science:

• Medical Informatics
• Health Data Science
• Clinical Data Visualization

Background:

• Clinicians and researchers need intuitive tools for processing large volumes of time-oriented, multi-source patient data.
• Aggregated patient views are crucial for clinical trial analysis and quality assessment.

Purpose of the Study:

• To design and develop the VISualizatIon of Time-Oriented RecordS (VISITORS) system.
• To combine intelligent temporal analysis with information visualization for multi-patient data processing.

Main Methods:

• Developed an interactive, ontology-based exploration module for visualizing and exploring raw and abstracted patient data.
• Utilized a knowledge-based temporal-abstraction method to derive meaningful interpretations from time-oriented data.
• Implemented an ontology-based temporal-aggregation specification language for patient population selection.

Main Results:

• Evaluated on over 1000 oncology patients, users answered clinical questions with high accuracy (98.7%) in a mean of 2.5 minutes.
• Demonstrated feasibility across absolute and relative time lines without significant differences in response time or accuracy.
• System Usability Scale (SUS) showed the VISITORS system to be usable (mean score 69.3), though clinicians rated usability lower than informaticians.

Conclusions:

• The VISITORS system provides a feasible, functional, and usable solution for intelligent visualization and exploration of longitudinal multi-patient data.
• The system supports efficient retrieval, visualization, and analysis of complex temporal health information.
• Further usability refinement may be needed to fully meet clinician needs.