What to do with multiple conditions from 1 study in the analysis (2 intervention groups, 1 sham group). Sham group is then counted twice in the analysis...
I have a very elaborate analysis (different subgroups), so within the analysis there are maybe 5 studies that contribute >1 condition. It's impossible to describe all the differences when 1 condition is left out or included..
This is not an uncommon situation. The attached PDF will hopefully answer your question. If you are using the Comprehensive Meta-Analysis software, it can handle studies with multiple comparisons though it does make simplistic assumptions about the correlation between comparisons.
I agree this is common. You can find solutions and examples in the Cochrane handbook (Chapter 16.5.4 "How to include multiple groups from one study").
Thanks a lot for your quick responses! Great that there are options available to handle these studies. If I inderstand it correctly; it is important to determine which intervention groups are relevant for the analysis, and if they are, you can create a combined effect (ignoring the differences among the treatments)!
The two groups model is not enough to find solutions in the real world. Put them aside, follow your heart to find a solution to solve the real world problem. Then, you can find the answer.
Such as the US use only two groups to do a clinical trial. China uses three groups to do the clinical trial: The new way to treat a diseases vs the traditional Chinese medicine and western medicine total three groups to compare which group is superior.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis