I'm comparing 3 treatment arms with each others. Usually, I use RevMan for comparing 2 groups. But I can't use it for network meta-analysis. What is the best alternative?
Dear
I suggest you "SBML Software".
The starting point is an appreciation that computational modeling of biological systems is no longer a fringe activity—it's a requirement for us to make sense of our vast and ever-expanding quantities of data. This reality is acknowledged and reinforced by a vast increase this decade in the number of journals, books and articles having computational and systems biology emphases.
At its most basic, computational modeling is no different from modeling as it's practiced by all scientists, whether in biology or elsewhere. The extra but crucial step is casting the model into a formal, computable form that can be analyzed rigorously using simulation and other mathematical methods.
Different representations of models are useful for different purposes. Graphical diagrams of biological processes are useful for visual presentation to humans, but at the level of software, a different format is needed for quantifying a model to the point where it can be simulated and analyzed. That's where the Systems Biology Markup Language (SBML) comes in.
Simply put, SBML is a machine-readable format for representing models. It's oriented towards describing systems where biological entities are involved in, and modified by, processes that occur over time. An example of this is a network of biochemical reactions. SBML's framework is suitable for representing models commonly found in research on a number of topics, including cell signaling pathways, metabolic pathways, biochemical reactions, gene regulation, and many others.
http://sbml.org/SBML_Software_Guide/SBML_Software_Summary
I would recommend the netmeta R-package available. It is written by Gerta Rücker and Guido Schwarzer with contributions of Ulrike Krahn. All three have published on methods for network meta-analysis and you can conduct fixed and random effects network meta-analysis with the package.
The drawback is, that you need to learn some R (never a bad thing though). The worst part will probably be importing your data in the first place (I recommend the .csv format which is simple to import if your use Rstudio). The examples in the netmeta package are very helpful.
http://cran.r-project.org/web/packages/netmeta/index.html
http://www.rstudio.com/
Hi. The Cochrane Comparing Multiple Interventions Methods Group provides an overview of different types of packages available to conduct network meta-analysis, links to more detailed descriptions, as well as supporting references. It can be found at the link below.
http://cmimg.cochrane.org/network-meta-analysis-toolkit
I tried them all.
STATA: mvmeta: unfortunately it is very difficul to get dataset as it needs y and v with same reference (see White 2009, White 2011)
R:
gemtc: the best and the easiest if you have event and sample size
mvmeta: similar to STATA command
netmeta: frequentistic approach, but unclear how it uses TE ans seTE
Good news - The free software MetaXL version 4.0 will be on www.epigear.com sometime next week. This will be the easiest and most robust implementation of network meta-analyses yet and should by and large make this methodology accessible freely to all mainstream researchers. the WMD effect can easily be used with this software and if you need help setting this up please send me a message and will forward an example.
There are few free software for network meta-analysis and are easy to use:
1. MetaInsight: easiest to use. It is essentially a web based app. R based using gemtc package and possibly JAGS. It is basically point and click so you do not need to code or be fluent in R. Gives you all the plots and data you need for publication. The creators of this app are very helpful and will answer any questions quickly by email.
2. Stata: fairly easy as long as your are comfortable with Stata. Stata 16 has a point and click feature for meta-analysis. Hopefully, they will do the same for Network meta-analysis in the future. Otherwise, the commands it and some do files are available free on line. The software Stata is however not free.
3. BUGSnet: this is a R based, using gemtc package and uses JAGS. It is very good and comprehensive. The authors also offer a web-based R studio platform for academic researchers free of charge. However, the packages are easy to load into R and work with. You may not get any help or support with this package.
4. ADDIS; online tool. I have not used it. Looks cumbersome. It is R/gemtc based.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software