Youden index and Associated Cut-points for Three Ordinal Diagnostic Groups

Area of Science:

• Biostatistics
• Medical Diagnostics
• Health Informatics

Background:

• The Youden index is vital for assessing diagnostic accuracy by optimizing sensitivity and specificity.
• Current Youden index applications are limited to binary diagnostic outcomes.
• Many diseases present with an intermediate stage requiring early recognition for effective intervention.

Purpose of the Study:

• To extend the Youden index for assessing diagnostic accuracy with three ordinal diagnostic groups.
• To develop parametric and nonparametric methods for estimating the optimal Youden index and cut-points.
• To evaluate the performance of proposed methods through simulations and a real-world example.

Main Methods:

• Extension of the Youden index to accommodate three ordinal diagnostic categories.
• Development of parametric and nonparametric statistical approaches for estimation.
• Conducting extensive simulation studies under various distributional assumptions.
• Application to a real-world diagnostic dataset for validation.

Main Results:

• Proposed methods effectively estimate the optimal Youden index for three-group classification.
• Parametric and nonparametric approaches demonstrate robust performance in simulations.
• Confidence intervals for optimal cut-points provide valuable uncertainty quantification.
• The extended Youden index proves useful in evaluating diagnostic test discriminating ability.

Conclusions:

• The extended Youden index offers a valuable tool for diagnostic accuracy assessment in scenarios with intermediate disease stages.
• The proposed parametric and nonparametric methods provide reliable estimation of optimal cut-points.
• This work enhances the utility of the Youden index in complex biomedical diagnostic practice.