Tools
Tools
Nikoloulopoulos, Aristidis K. 
ORCID: https://orcid.org/0000-0003-0853-0084
(2024)
Joint meta-analysis of two diagnostic tests accounting for within and between studies dependence.
Statistical Methods in Medical Research.
ISSN 0962-2802
Preview |
PDF (Nikoloulopoulos-2024-SMMR)
- Published Version
Available under License Creative Commons Attribution. Download (491kB) | Preview |
Abstract
There is an extensive literature on methods for meta-analysis of diagnostic test accuracy, but it mainly focuses on a single test. A multinomial generalised linear mixed model was recently proposed for the joint meta-analysis of studies comparing two tests on the same participants in a paired tests design with a gold standard. In this setting, we propose a novel model for joint meta-analysis of studies comparing two diagnostic tests which assumes independent multinomial distributions for the counts of each combination of test results in diseased and non-diseased patients, conditional on the latent vector of probabilities of each combination of test results in diseased and non-diseased patients. For the random effects distribution of the latent proportions, we employ a one-truncated D-vine copula that can provide tail dependence or asymmetry. The proposed model includes the multinomial generalised linear mixed model as a special case, accounts for the within-study dependence induced because the tests are applied to the same participants, allows for between-studies dependence, and can also operate on the original scale of the latent proportions. The latter enables the derivation of summary receiver operating characteristic curves. Our methodology is demonstrated with simulation studies and a meta-analysis of screening for Down’s syndrome with two tests: shortened humerus and shortened femur.
| Item Type: | Article |
|---|---|
| Additional Information: | Software: R functions to derive estimates and simulate from the multinomial one-truncated D-vine CMM for meta-analysis of two diagnostic tests accounting for within and between studies dependence are part of the R package CopulaREMADA.52 The data and code used in Section 5 are given as code examples in the package. |
| Uncontrolled Keywords: | mixed models,multinomial generalised linear mixed model,multiple diagnostic tests,multivariate meta-analysis,summary receiver operating characteristic curves,vine copula models,epidemiology,statistics and probability,health information management ,/dk/atira/pure/subjectarea/asjc/2700/2713 |
| Faculty \ School: | Faculty of Science > School of Mathematics (former - to 2024) |
| UEA Research Groups: | Faculty of Science > Research Groups > Statistics Faculty of Science > Research Groups > Data Science and AI |
| Related URLs: | |
| Depositing User: | LivePure Connector |
| Date Deposited: | 06 Aug 2024 10:30 |
| Last Modified: | 02 Oct 2024 09:30 |
| URI: | https://ueaeprints.uea.ac.uk/id/eprint/96162 |
| DOI: | 10.1177/09622802241269645 |
Downloads
Downloads per month over past year
Actions (login required)
 |
View Item |