Can you do a meta-analysis of meta-analyses?

[RedMorning]

For instance, pretend there are multiple meta-analyses for different drugs in a certain class of drugs, (ie. one analysis on Lasix, one on a beta-blocker, another on calcium channel blockers, and one on ace inhibitors, etc) to determine a rate of side effects for each category. Can I then do a meta-analyis of those articles to give an overall rate of that side effect for the entire class of drug (ie hypertensive medications)?

Disregarding the above example, as that wouldn't really say anything useful. Has this ever been done before? I did a lit search and didn't come up with much. What methods are available...or is this just stupid and that's why its never been done?

Thanks for any help,
RM

---

Members don't see this ad.