Network Meta-analysis

Discuss statistics related things

by Xenophon » Wed Oct 09, 2019 11:46 am

Is there any rule of how many studies per arm should I have in a network meta-analysis or the number of arms and the appropriate number of studies included in each arm?

thank you in advance.
Xenophon
 
Posts: 9
Joined: Fri Mar 29, 2019 3:22 pm

by Nancysara_123 » Sat Dec 28, 2019 12:40 pm

Generally, network meta-analysis defined as to combine the result of several studies (generally randomized trails) or comparing several interventions or treatments is called network analysis. Previously network meta-analysis usually was done in a Bayesian setting, but nowadays it has possible to perform in a frequentist setting. A minimum of three studies or treatments is essential for the performance of network meta-analysis.
Nancysara_123
 
Posts: 2
Joined: Sat Dec 28, 2019 12:25 pm

by Xenophon » Mon Dec 30, 2019 8:20 am

Thank you Nancysara_13 for your reply. Do you have a reference for your last sentence "A minimum of three studies or treatments is essential for the performance of network meta-analysis."?
Xenophon
 
Posts: 9
Joined: Fri Mar 29, 2019 3:22 pm

by Nancysara_123 » Tue Apr 28, 2020 11:07 am

Network meta-analysis (NMA), called multiple treatment meta-analysis or mixed treatment comparisons, was created as a pair meta-analysis extension to enable comparisons of more than two interventions in a single, coherent analysis of all related randomized controlled trials (RCTs). Its main advantages are that, in a single study, it provides accurate estimates of the relative effects of all treatments compared to each other using both direct and indirect data, thus correctly integrating the relative impact of trials with more than two arms. This results in greater precision of treatment effect estimates and the ability to rank all the interventions coherently. This results in higher accuracy of treatment effect estimates and the ability to list all the responses cohesively.
Nancysara_123
 
Posts: 2
Joined: Sat Dec 28, 2019 12:25 pm

by cjfoote » Sat May 30, 2020 12:47 pm

Yes, I'm aware of all what was said above. However, is this type of analysis supported by Jamovi?
cjfoote
 
Posts: 1
Joined: Sat May 30, 2020 12:40 pm


Return to Statistics