Could you use more than one software for a meta-analysis?

Question

Im currently working on a meta-analysis of interventional randomized controlled trials.

In the protocol, I mentinoned that I will be using Review Manager 5 for both the analysis and assesement of publication bias.

However, after the analysis, I noted that there is asymmetry of the funnel plot, indicating potential publication bias.

Now I would like to confirm the publication bias and correct the effect with trim-fill method.

In Review Manager, there is no way to do this, so I have to do another software like JASP.

My question is: Is using another software a violation of the protocol of the study?

If not, how should I report this? Should I just mention that I will be using JASP instead of Review Manager in the method?

Answer 1

There is no objection to using a different piece of software that I can think of. Obviously you would mention why you did it in your report.

Having said that you do realise that the trim and fill method has been deprecated? Schwarzer and colleagues in a paper entitled Empirical evaluation suggests Copas selection model preferable to trim-and-fill method for selection bias in meta-analysis point out an alternative. You might also be interested in my answer to an earlier question Publication bias: Egger's Test significant depending on subset. Why doesTrim and Fill not add studies?