Μeta-analysis in Stata using gllammThere are several user written programs for performing meta-analysis in Stata. These include metan (univariate meta-analysis), metareg (meta-regression), mvmeta (multivariate meta-analysis), midas and metandi for diagnostic tests and glst (dose-response models). However, there are several cases for which these programs do not suffice. For instance, there is no software for performing univariate meta-analysis with correlated estimates, for performing a multilevel or hierarchical meta-analysis, or for meta-analysis of longitudinal and repeated measurements data. In this work, we are going to show with practical applications, that many seemingly unrelated models, including but not limited to the ones mentioned above, can be fit using gllamm. The software is very versatile and can handle a wide variety of models with applications in a wide range of disciplines. The method presented here, takes advantage of these modelling capabilities and makes use of appropriate transformations (based on the Cholesky decomposition of the inverse of the covariance matrix) known as generalised least squares (GLS), in order to handle correlated data. Most of the models described above can be thought as special instances of a general linear mixed model formulation, but to the author’s knowledge a general exposition in order to incorporate all the available models for meta-analysis as special cases, and the instructions to fit them in a single statistical package has not been presented so far.A Multivariate Method for Meta-Analysis and Comparison of Diagnostic TestsSeveral methods for meta-analysis of diagnostic tests have been proposed. However, when several diagnostic tests are evaluated, the situation is problematic since the within-studies correlation needs to be taken into account. We present an extension of the bivariate random effects meta-analysis for the log-transformed sensitivity and specificity that can be applied for two or more diagnostic tests. The advantage of this method is that a closed-form expression is derived for the calculation of the within-studies covariances. The method allows the direct calculation of sensitivity and specificity, as well as, the diagnostic odds ratio (DOR), the area under curve (AUC) and the parameters of the Summary Receiver Operator’s Characteristic (SROC) curve, along with the means for a formal comparison of these quantities for different tests. There is no need for individual patient data or the simultaneous evaluation of both diagnostic tests in all studies. The method is simple and fast, it can be extended for several diagnostic tests and can be fitted in nearly all statistical packages. The method was evaluated in simulations and applied in a meta-analysis for the comparison of Anti–Cyclic Citrullinated Peptide (anti-CCP antibody) and Rheumatoid Factor (RF) for discriminating patients with Rheumatoid Arthritis (RA), with encouraging results.International Journal Publications
Software
|
IntDaMuS > Work Packages >