Computational methods to simultaneously compare the predictive values of two diagnostic tests with missing data: EM-SEM algorithms and multiple imputation

TL;DR

This paper develops and compares EM-SEM algorithms and multiple imputation methods to evaluate and compare the predictive values of two diagnostic tests with missing data, demonstrated through simulations and Alzheimer's disease diagnosis.

Contribution

It introduces computational methods for comparing predictive values of diagnostic tests with missing data, including implementation in R and application to Alzheimer's diagnosis.

Findings

EM-SEM and multiple imputation methods perform well in simulations

Rules for applying the hypothesis tests are provided

Methods are demonstrated on Alzheimer's disease data

Abstract

Predictive values are measures of the clinical accuracy of a binary diagnostic test, and depend on the sensitivity and the specificity of the test and on the disease prevalence among the population being studied. This article studies hypothesis tests to simultaneously compare the predictive values of two binary diagnostic tests in the presence of missing data. The hypothesis tests were solved applying two computational methods: the EM and SEM algorithms and multiple imputation. Simulation experiments were carried out to study the sizes and the power of the hypothesis tests, giving some general rules of application. Two R programmes were written to apply each method, and they are available as supplementary material for the manuscript. The results were applied to the diagnosis of Alzheimer's disease.