Three-Dimensional Visual Patient Based on Electronic Medical Diagnostic Records

Area of Science:

• Medical Informatics
• Radiology
• Data Visualization

Background:

• Medical diagnostic records (MDRs) are complex and voluminous.
• Current systems like PACS/RIS can be inefficient for comprehensive patient data review.

Purpose of the Study:

• To introduce a novel 3-D anatomical graphic pattern, Visual Patient (VP), for indexing, representing, and rendering MDRs.
• To enhance physician access to patient's current and historical medical status.

Main Methods:

• Developed a VP data model (Visual Index Object - VIO) and a graphic model (Visual Anatomic Object - VAO).
• Implemented methods to parse and extract key information from MDRs to populate the VIO.
• Designed a VP system (VPS) to map VIO to VAO, creating a patient-specific VP instance.

Main Results:

• A prototype VPS was implemented and evaluated in a simulated hospital PACS/RIS environment.
• Over 70% of participating radiologists expressed willingness to use the VPS.
• VPS demonstrated a 2.24x increase in review efficiency compared to PACS/RIS, with a minor 4% accuracy difference.

Conclusions:

• The developed VPS effectively displays patient organ status using a 3-D anatomical graphic pattern.
• Radiologists favor the VPS for its improved efficiency in reviewing MDRs, with acceptable accuracy.
• VP offers a new, efficient method for accessing and interacting with extensive patient records in the big data era.