A quick question: if i have two clinical trials, one comparing A and B and another comparing A and C, can i combine them in a meta-analysis (using individual patient data - IDP) to compare A and B (i am not interested in treatment C), i.e. use the data for B from the first study and data for A from both studies? Or is this treatment comparison then confounded with differences between the two studies? I believe i've seen statisticians lumping data together in this way - are they simply making some 'additional assumptions' ie that the two studies are sufficiently similar, or does the model account for it? Or do i need another statistical approach altogether such as 'network meta-analysis'? Apologies if this is a stupid question, i tried to find an answer using google first but one can always find good and bad analyses thus it's not such a useful guide.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|