2018/12/10 by Aristidis K. Nikoloulopoulos, Nikoloulopoulos, Aristidis K.
Computer Science · Mathematics · #Applications (stat.AP) #Data Analysis with R #FOS: Computer and information sciences #Statistical Methods and Bayesian Inference #Statistical Methods and Inference
paper · pdf · doi:10.48550/arxiv.1812.03685
openalex publication_date 2018/12/10 · openalex created_date 2022/08/01 · openalex updated_date 2026/07/28
A recent paper proposed an extended trivariate generalized linear mixed model\n(TGLMM) for synthesis of diagnostic test accuracy studies in the presence of\nnon-evaluable index test results. Inspired by the aforementioned model we\npropose an extended trivariate vine copula mixed model that includes the TGLMM\nas special case, but can also operate on the original scale of sensitivity,\nspecificity, and disease prevalence. The performance of the proposed vine\ncopula mixed model is examined by extensive simulation studies in comparison\nwith the TGLMM. Simulation studies showed that the TGLMM overestimates the\nmeta-analytic estimates of sensitivity, specificity, and prevalence when the\nunivariate random effects are misspecified. The vine copula mixed model gives\nnearly unbiased estimates of test accuracy indices and disease prevalence. Our\ngeneral methodology is illustrated by meta-analysing coronary CT angiography\nstudies.\n