Quantitative patient graph analysis for transient ischemic attack risk factor distribution based on electronic...

Area of Science:

• Medical Informatics
• Network Science
• Epidemiology

Background:

• Transient ischemic attack (TIA) affects millions globally, with individual risk factors known but lacking systemic quantitative analysis.
• Electronic medical records (EMR) offer vast data potential for understanding complex disease associations.
• A comprehensive, data-driven approach to TIA risk factor identification using EMR is needed.

Purpose of the Study:

• To quantitatively analyze TIA health factors using a data-driven graph approach.
• To identify and rank TIA risk factors by leveraging hospital EMR data.
• To integrate patient data with established medical knowledge for clinical utility.

Main Methods:

• Construction of a TIA patient health factor graph using hospital EMR data (737 TIA patients, 737 controls, 740 health factors).
• Calculation and ranking of connection delta ratios (CDRs) for all health factors.
• Visualization of the patient graph integrated with the TIA knowledge graph in the Unified Medical Language System.

Main Results:

• A quantitative distribution of TIA health factors was generated via graph CDR analysis.
• Fifty-six known TIA risk factors were confirmed within the distribution.
• Rhinosinusitis emerged as a potential novel risk factor requiring further validation.

Conclusions:

• Graph CDR analysis provides an effective method for quantifying TIA risk factor distribution.
• The identified risk factor distribution can inform the development of machine learning models for TIA prediction.
• Integrated graph visualization enhances clinical access to patient data and standardized medical knowledge.