Evaluating a decision making system for cardiovascular dysautonomias diagnosis

Area of Science:

• Cardiovascular Health
• Neurology
• Medical Informatics

Background:

• The autonomic nervous system (ANS) regulates vital bodily functions and its malfunction can cause cardiovascular dysautonomias.
• Accurate diagnosis of cardiovascular dysautonomias is crucial for patient management.

Purpose of the Study:

• To develop and evaluate a decision tree-based prediction system for cardiovascular dysautonomias.
• To implement and clinically validate a prototype system using real-world patient data.

Main Methods:

• A dataset of 263 records from the Avicenne hospital's ANS unit was utilized.
• The C4.5 decision tree algorithm was employed for system development.
• A Java Enterprise Edition platform was used for prototype implementation and clinical validation.

Main Results:

• The decision tree system achieved average error rates of 1.46% (training), 2.24% (test), and 0.89% (validation).
• The prototype system was successfully deployed and clinically validated within the ANS unit.

Conclusions:

• The decision tree-based prediction system demonstrates encouraging performance for diagnosing cardiovascular dysautonomias.
• Further replicated studies are recommended to confirm these findings and enhance clinical utility.