Meta analytic approaches for incorporated diagnostic tests

Depending on the developments in science and technology, the importance of meta-analysis studies, which enable the synthesis of research results on similar subjects at different places and times and obtaining universal results through statistical analysis, is gradually increasing. Although there have been significant advances in the methodology used in meta-analysis of the accuracy of diagnostic tests, there are still shortcomings and controversies when using combined diagnostic tests. The aim of this study is to examine how diagnostic accuracy measures change for cases where two combined diagnostic tests are used, depending on factors such as the number of studies included in the meta analysis, the number of subjects in the primary studies and dependency/independence between tests. In this context, evaluations on the subject were made through scenarios created using artificial data sets obtained through simulation study and an application example was given with real data. As a result of the simulation study, it was found that the number of samples in the primary studies, the number of studies included in the meta analysis, and the relationship between the tests had an impact on the results of the meta analysis of the combined test performances. In general, it has been observed that errors tend to increase when the relationship between tests increases, when the number of subjects in the primary studies is low and the number of studies included in the meta-analysis is low, and when the relationship between the combined tests increases, prevalence is more effective on test performance measures. Keywords:meta-analysis, combined diagnostic tests, diagnostic accuracy,Boole algebra, conditional dependence.