2

I am looking to pool data for adverse events for a meta analysis and I wanted to see if my plan made sense

I want to pool my AE together based on the Clavien Dindo categorizations, as this will make the most sense clinically. I am running into the following issue:

Many of the trials report AEs like this: enter image description here

So if I were to pool all these together it one group, is it okay for me to do it like this:

Fentanyl: 30/160 F-Paracetamol: 10/155

I am just used to on forest plots seeing the N=the sample size, and the way I am thinking about doing it I am multiplying my sample size by the number of events.

I would love to hear people's thoughts.

Vicki Archer
  • 71
  • 1
  • If all the primary studies examine the same side effects then you could investigate multi-level meta-analysis using them as multiple outcomes. You could approach authors for the total of people who experience any AE but otherwise I think you are stuck with analysing each AE separately which might make more sense as the signal might just be concentrated in one AE although I agree that des not look like the way it is going. – mdewey Nov 05 '21 at 11:30

0 Answers0