Revman help

Sm29

New Member
#1
Hi,

My stat knowledge is not great at the best of times so I'm hoping someone can help me.

I'm doing a systematic review where I'm comparing the effects of group based exercise to individual based exercise in Parkinson's disease. As not specific RCT's have been published on the topic. I have articles that compare individual based exercise to usual care and group based exercise and usual care. Is it possible on Revman to compare these two interventions?

Also, when entering data do I put in baseline or follow up data? Or do I do mean difference?

Thanks
 

hlsmith

Not a robit
#2
So you want to compare two treatment which have never been directly constrasted but both have been compared to a similiar third treatment? If so this would be call a network meta analysis. Not sure if revman covers this yet, it is a new topic.
 

Sm29

New Member
#3
Yes, that's what I thought.

If I was to just compare group based exercise with usual care, what data would I need to enter into the table? baseline or follow up data? or would I have to calculate mean difference (SD)?

Thanks
 

Sm29

New Member
#4
Yes, that's what I thought.

If I was to just compare group based exercise with usual care, what data would I need to enter into the table? baseline or follow up data? or would I have to calculate mean difference (SD)?

Thanks
Can anyone help me with this?

I'm really struggling and have a deadline coming up soon
 

hlsmith

Not a robit
#5
So you no longer want to do a network meta analysis. You now want to do a typical meta analysis?


Side note, I know SAS has a macro for network MA, and I would be surprised if R didn't have something.
 

Sm29

New Member
#6
So you no longer want to do a network meta analysis. You now want to do a typical meta analysis?


Side note, I know SAS has a macro for network MA, and I would be surprised if R didn't have something.

Yes, just a typical meta analysis.

Just confused on the values i input into Revman. Is it the baseline data, follow up data or do I have to calculate mean difference and standard deviation as most of my included studies do no present their data in that way.
 

hlsmith

Not a robit
#7
Do all of the studies present data as format and units. Probably need to standardize. How do you want results presented.
 

hlsmith

Not a robit
#9
So you currently have a list of all of the studies you are going to use along with mean differences.


Were study groups in studies matched or not? Do you have standard deviations and sample sizes? Do you have standardized mean differences?


You should find a copy of Borenstein et al. book, "Introduction to Meta-Analysis" to help guide you. Side note, I downloaded RevMan5 last weekend, so I should be able to help with software questions once you get that far.
 
#10
I currently have a list of all my studies however only 3 provide mean difference and standard deviation between pre and post interventions. The rest present a mean and SD.

The studies compare usual care against individual exercise.

I still haven't got the hang of Revman, in terms of what details to input - At first I thought mean difference but the studies obviously just provide the mean and not the raw data needed to calculate SD of the mean difference as well.


Just very confused at the minute