Improved Confidence Intervals for the Youden Index.

Area of Science:

• Biostatistics
• Medical Diagnostics
• Clinical Trials

Background:

• The Youden Index is a key metric for evaluating diagnostic test accuracy using receiver operating characteristic (ROC) curves.
• Existing confidence intervals for the Youden Index, including bootstrap and delta methods, have limitations in performance.
• Accurate confidence intervals are crucial for interpreting diagnostic test performance in clinical settings.

Purpose of the Study:

• To propose two novel confidence intervals for the Youden Index using the square-and-add limits based on the Wilson score method.
• To compare the performance of the proposed confidence intervals against existing methods through simulation studies.
• To illustrate the application of the new intervals using a real-world clinical trial example.

Main Methods:

• Development of two new confidence intervals utilizing the Wilson score method with square-and-add limits.
• Extensive simulation studies were conducted to compare the performance of the proposed intervals with existing ones.
• Application of the new confidence intervals to data from a prostate cancer clinical trial.

Main Results:

• The proposed confidence intervals demonstrated satisfactory performance in simulation studies.
• The new interval based on the empirical proportion estimate generally outperformed the interval based on the adjusted proportion estimate.
• Comparison revealed advantages of the proposed Wilson score-based intervals over traditional methods.

Conclusions:

• The proposed Wilson score-based confidence intervals offer a valuable improvement for estimating Youden Index accuracy.
• The empirical proportion estimate-based interval shows particular promise for enhanced diagnostic test evaluation.
• These new intervals provide a more reliable tool for analyzing diagnostic accuracy in clinical research, as demonstrated in the prostate cancer example.