Diagnostic accuracy of cardiologists compared with probability calculations using Bayes' rule

Area of Science:

• Cardiology
• Medical Diagnostics
• Probability Theory

Background:

• Probability analysis offers insights into diagnostic tests for coronary artery disease (CAD).
• Recent advancements may enable clinical application for individual patient risk assessment.
• Validation of existing probability calculation methods is crucial for clinical utility.

Purpose of the Study:

• To independently validate two probability calculation methods for CAD.
• To compare the diagnostic accuracy of these methods against cardiologists' assessments.
• To assess the efficacy of probability analysis in diagnosing CAD.

Main Methods:

• Ninety-one cardiologists evaluated clinical summaries of 800 randomly selected patients.
• Cardiologists estimated CAD probability based on history, physical exam, and treadmill tests.
• Two probability methods (tables and computer program) using Bayes' rule were applied.
• Coronary angiography served as the gold standard for diagnostic accuracy assessment.

Main Results:

• Average diagnostic accuracy: Cardiologists (80.2%), Tables (78.0%), Computer program (83.1%).
• Computer-based probability analysis showed statistically significant higher accuracy than cardiologists (p < 0.01).
• Table-based analysis was also significantly more accurate than cardiologists (p < 0.05).

Conclusions:

• Detailed probability analysis can achieve diagnostic accuracy comparable to cardiologists.
• Computerized probability calculations demonstrate superior accuracy in CAD diagnosis.
• Further studies on the clinical efficacy of probability analysis in patient care are warranted.