Stata is statistical software that combines programming and un-programming, which is easy to operate, of high efficiency and good expansibility. In performing meta-analysis, Stata software also presents powerful function. The mvmeta package of Stata software is based on a multiple regression model to conduct network meta-analysis, and it also processes "multiple outcomes-multivariate" data. Currently, the disadvantages of mvmeta package include relatively cumbersome process, poor interest-risk sorting, and lack of drawing function in the process of conducting network meta-analysis. In this article, we introduce how to implement network meta-analysis using this package based on cases.
ObjectiveTo evaluate the association between tumor necrosis factor alpha (TNF-α) gene -308G/A polymorphism and the risk of endometriosis (EM). MethodsDatabases including PubMed, EMbase, CBM, CNKI, VIP, and WanFang Data were searched to collect case-control studies about the association between TNF-α gene -308G/A polymorphism and the risk of EM from inception to October 2014. Two reviewers independently screened literature, extracted data, and assessed the risk of bias of included studies. Then Meta-analysis was performed by using Stata 12.0 software. ResultsA total of seven case-control studies were included, which involved 687 patients and 877 controls. The results of meta-analysis showed that, AA genotype carriers presented 2.06-fold (OR=2.06, 95%CI 1.10 to 3.83, P=0.02) and 1.94-fold (OR=1.94, 95%CI 1.18 to 3.18, P<0.01) higher risk of EM than GG genotype carriers and AG+GG genotype carriers, respectively; but sensitivity analysis showed that the robustness of these results was unstable. No statistical significance was found in the allele model, co-dominant model, and dominant model (A vs. G: OR=1.37, 95%CI 0.87 to 2.17, P=0.18; AG vs. GG: OR=1.09, 95%CI 0.77 to 1.53, P=0.63; AA+AG vs. GG: OR=1.29, 95%CI 0.95 to 1.77, P=0.11). While excluding studies that controls were not in HWE, no significant association was observed in these five genetic models; and no significant association was found in the results of subgroup analysis by ethnicity. ConclusionThe AA genotype of TNF-α gene -308G/A polymorphism might contribute to the risk of EM, but the association between other genotypes and EM susceptibility is unclear. In addition, due to the limited quantity and quality of included studies, more high quality studies are needed to verify the above conclusion.
The key for performing meta-analysis using WinBUGS software is to construct a model of Bayesian statistics. The hand-written code model and Doodle model are two major methods for constructing it. The approach of hand-written code is flexible and convenient, but the language programming is fallibility. The Doodle is complicated, but it is benefit to understand the structure of hand-written code model and prevent error. This article briefly describes how to construct the Doodle model for binary and continuous data of head to head meta-analysis, indirect comparison and network meta-analysis, and ordinal variables meta-analysis.
NetMetaXL is a macro command to conduct network meta-analysis in the frame of Microsoft Excel on basis of Bayesian theory. This macro command, which was officially launched in 2014, integrates data extraction and entry, analysis results output and graph plotting as a whole. Currently, this version contains enough optional models, and all operations are through menu and easy to conduct; however, it is appropriate only for the network meta-analysis based on dichotomous variables, which still has fairly a lot to be enhanced and improved. This article gives a brief introduction based on examples to implement network meta-analysis using NetMetaXL.
ITC (Indirect Treatment Comparison) software and indirect procedure of Stata software are especially used for indirect comparison nowadays, both of which possess the characteristics of friendly concise interface and support for menu operation. ITC software needs the application of other software to yield effect estimation and its confidence interval of direct comparison firstly; while Stata-indirect procedure can complete direct comparison internally and also operate using commands, which simplifies complicated process of indirect comparison. However, both of them only perform "single-pathway" of data transferring and pooling, which is a common deficiency. From the results, their results are of high-degree similarity.
ObjectiveTo evaluate the relationship between tumor necrosis factor-α (TNF-α) gene promoter-308 G/A polymorphism and ankylosing spondylitis (AS) in Chinese population by meta-analysis. MethodsThe casecontrol studies about the correlation between TNF-α gene polymorphism and AS in Chinese population were retrieved from PubMed, EMbase, CNKI, CBM, WanFang Data and VIP database by two researchers. The retrieval time was from their establishment to December, 2015. After the paper screening, data extraction, and assessment of bias risk, the metaanalysis was conducted by Stata 12.0 software. ResultsA total of 11 case-control studies involving 1 154 AS patients and 1 458 controls were included. The results of meta-analysis showed that, for Chinese population, there was no significant association between TNF-α-308 G/A polymorphism and AS susceptibility (A vs. G: OR=0.96, 95% CI 0.63 to 1.47, P=0.86; AA vs. AG: OR=0.97, 95% CI 0.51 to 1.84, P=0.93; AA vs. GG: OR=0.92, 95% CI 0.32 to 2.61, P=0.87; AA+AG vs. GG; OR=1.04, 95% CI 0.60 to 1.80, P=0.89; AA vs. AG+GG: OR=1.03, 95% CI 0.58 to 1.82, P=0.92). ConclusionTo date, it has not found the relationship between TNF-α gene promoter-308 G/A polymorphism and AS in Chinese population. For the quantity and quality limitation of the included studies, the conclusion has to be verified by more large-scale highquality studies.
The mada package is a type of package that is especially used for implementing meta-analysis of diagnostic accuracy tests. This package is developed on basis of classical statistical theories and it can be used to calculate all relevant effect size of diagnostic accuracy tests; however, it does not provide pooled values of sensitivity and specificity. This article uses an example to introduce the whole functions of mada package in implementing meta-analysis of diagnostic accuracy tests, including data preparation, calculation implementation, result summary, and plots drawing.
ObjectiveTo systematically review the association between myopia and open-angle glaucoma (OAG). MethodsForeign and domestic articles published from January 2000 to May 2013 were searched in PubMed, EMbase, WanFang Data, VIP and CNKI for observational studies on the association between myopia and OAG. According to the inclusion and exclusion criteria, the studies were screened, the data were extracted, and the method quality of included studies was assessed. Then meta-analysis was performed using Stata 12.0 software. ResultsEleven cross-sectional studies including 45 996 participants were finally included. The results of meta-analysis showed that, myopia increased the risk of OAG. Besides, the results of subgroup analysis by the severity of myopia showed that, low-degree myopia (more than-3.00 D) increased 1.52 times the risk (OR=1.52, 95%CI 1.23 to 1.88) while middle/high-degree myopia (no more than-3.00 D) increased 2.41 times (OR=2.41, 95%CI 1.91 to 3.03). ConclusionIndividuals with myopia have increased risk of OAG.
Restricted cubic spline function is an ideal model in trend approximation, which is widely used in doseresponse meta-analysis. The spline function, based on parameter technique, is a smoothly joined piecewise polynomial of each knot, with a cubic polynomial in each sub-interval of the slope which fits well in the non-linear trend by changing the number and (or) the sites of the knots. We have introduced the methodology of linear and non-linear slope model in dose-response meta-analysis in the previous article, and in this one, we will give a more detailed discussion on restricted cubic spline function mainly in the following aspects: model building, parameters pooling and knots selecting.
Dose-response meta-analysis, an important tool in investigating the relationship between a certain exposure and risk of disease, has been increasingly applied. Traditionally, the dose-response meta-analysis was only modelled as linearity. However, since the proposal of more powerful function models, which contains both linear, quadratic, cubic or more higher order term within the regression model, the non-linearity model of dose-response relationship is also available. The packages suit for R are available now. In this article, we introduced how to conduct a dose-response meta-analysis using dosresmeta and mvmeta packages in R.