An one-factor copula mixed model for joint meta-analysis of multiple diagnostic tests

Aristidis K. Nikoloulopoulos

DOI: https://doi.org/10.48550/arXiv.2006.09278

2021-05-10

Abstract:As the meta-analysis of more than one diagnostic tests can impact clinical decision making and patient health, there is an increasing body of research in models and methods for meta-analysis of studies comparing multiple diagnostic tests. The application of the existing models to compare the accuracy of three or more tests suffers from the curse of multi-dimensionality, i.e., either the number of model parameters increase rapidly or high dimensional integration is required. To overcome these issues in joint meta-analysis of studies comparing $T >2$ diagnostic tests in a multiple tests design with a gold standard, we propose a model that assumes the true positives and true negatives for each test are conditionally independent and binomially distributed given the $2T$-variate latent vector of sensitivities and specificities. For the random effects distribution, we employ an one-factor copula that provides tail dependence or tail asymmetry. Maximum likelihood estimation of the model is straightforward as the derivation of the likelihood requires bi-dimensional instead of $2T$-dimensional integration. Our methodology is demonstrated with an extensive simulation study and an application example that determines which is the best test for the diagnosis of rheumatoid arthritis.

Methodology,Applications

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the multidimensionality problem encountered when conducting a joint meta - analysis of more than two diagnostic tests in a multi - test design. Specifically, existing models face the problems of a rapid increase in the number of parameters or the need for high - dimensional integration when comparing three or more diagnostic tests. These problems increase the complexity of the models, are computationally costly, and may affect the stability and accuracy of the models.

To overcome these challenges, the paper proposes a new one - factor Copula mixture model. This model assumes that the true positive numbers and true negative numbers of each test are conditionally independent and binomially distributed given the sensitivity and specificity of the 2T - dimensional latent vector. For the random - effect distribution, the model uses a one - factor Copula, which can provide tail - dependence or tail - asymmetry. Through this method, the maximum - likelihood estimation of the model becomes simple because the derivation of the likelihood function only requires two - dimensional integration rather than 2T - dimensional integration.

This new method not only reduces the number of model parameters but also simplifies the calculation process, thereby improving the efficiency and accuracy of the model. In addition, this model can be applied to actual data, such as for determining which test is most suitable for the diagnosis of rheumatoid arthritis.