A test for comparing conditional ROC curves with multidimensional covariates

TL;DR

This paper introduces a non-parametric test for comparing conditional ROC curves with multidimensional covariates, accounting for covariate effects in diagnostic accuracy comparisons.

Contribution

It proposes a novel projection-based non-parametric method for testing equality of dependent ROC curves conditioned on multidimensional covariates.

Findings

Method performs well in simulations

Effective in real patient data analysis

Handles multidimensional covariate effects

Abstract

The comparison of Receiver Operating Characteristic (ROC) curves is frequently used in the literature to compare the discriminatory capability of different classification procedures based on diagnostic variables. The performance of these variables can be sometimes influenced by the presence of other covariates, and thus they should be taken into account when making the comparison. A new non-parametric test is proposed here for testing the equality of two or more dependent ROC curves conditioned to the value of a multidimensional covariate. Projections are used for transforming the problem into a one-dimensional approach easier to handle. Simulations are carried out to study the practical performance of the new methodology. A real data set of patients with Pleural Effusion is analysed to illustrate this procedure.