Meta-analysis to compare 3 randomized control trials?

ACAD

New Member
#1
I'm proposing to evaluate an outcome in three completed randomized controlled trials. The primary goal is to analyze a new outcome in each data set separately (i.e., 3 separate papers). A secondary aim is to compare across the three trials to see which intervention type was most effective. Because the trials differ in various ways: different populations (but same age range), different intervention plans, different outcome tasks, different time lengths; I thought that all I could propose was an overall "visual" type comparison (i.e., Study 1 was significant, Study 2 wasn't...therefore the intervention type in Study 1 is most effective).

However, I'm now second guessing myself and the thought of doing a meta-analysis to compare the 3 studies' effects has come to mind. I only have 3 studies though, so is a meta-analysis even reasonable? I've read one study saying no, and another saying the "visual" comparison isn't a good choice so a meta-analysis is the best alternative. I also considered combining the data from all 3 studies and doing a MANOVA where a study x time interaction would the main value of interest.

I appreciate any insight or advice on this issue. Thanks.