Interesting that so many are saying it's always problematic. Does this mean we should never meta-analyse epidemiological (i.e. cohort, case-ctrl etc) data? If not, why?
-
-
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Bear in mind I2 is not one size fits all. A MA with only small studies with huge CIs would have a much smaller I2 value compared to one with only big studies with the same effect estimates and small CIs. 1/n
-
More, an MA with huge studies with tiny CIs can have a massive I2, even if the studies have virtually identical effect estimates. That's just how I2 functions. Of course, most MAs will have a mix, but I2 will be higher if your big studies disagree. 2/n
- Show replies
New conversation -
-
-
This Tweet is unavailable.
-
No worries
End of conversation
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.