Comparison of three Bayesian methods to estimate posttest probability in patients undergoing exercise stress testing

Area of Science:

• Cardiology
• Medical Diagnostics
• Biostatistics

Background:

• Bayesian methods are crucial for estimating posttest probability in medical diagnostics.
• Refinements in Bayesian approaches aim to enhance diagnostic accuracy.
• Exercise stress testing is a common diagnostic tool for coronary artery disease (CAD).

Purpose of the Study:

• To compare the diagnostic ability of three Bayesian methods for estimating posttest probability of CAD after exercise stress testing.
• To evaluate the impact of incorporating more variables into Bayesian models.
• To assess the performance of CADENZA, a 15-variable Bayesian method, against simpler models.

Main Methods:

• Three Bayesian methods (A=5 variables, B=6 variables, C=CADENZA=15 variables) were used to estimate posttest probability of CAD.
• Data from 436 patients undergoing stress testing and coronary arteriography were analyzed.
• Sensitivity and specificity were compared across different posttest probability thresholds.

Main Results:

• CADENZA yielded significantly higher mean posttest probabilities for CAD compared to methods A and B.
• All methods, particularly CADENZA, showed overestimation of CAD at posttest probabilities >= 60%.
• CADENZA demonstrated significantly higher sensitivity but lower specificity than methods A and B.

Conclusions:

• CADENZA is a more effective screening tool at lower probability thresholds due to its higher sensitivity.
• Simpler Bayesian methods (A and B) retain value in confirming higher CAD probabilities suggested by CADENZA.
• Bayesian method refinements can improve diagnostic performance, but careful interpretation is needed across different probability ranges.