95%CI <0 and >1 with meta-analysis?

[TessaJ]

Hello,

I'm using the Major - proportions analysis to conduct a meta-analysis on single-arm event rates.
However, Jamovi calculates 95%CI that contain negative proportions and/or proportions exceeding >1.0
Am I doing something wrong in data entry or data analysis?

For example, with the following 3 studies:
Study - No. patients - No. events
A - 94 - 92
B - 66 - 60
C - 14 - 13

Jamovi calculates the following proportions with 95%CI (using the DerSimonian Laird model, measuring raw proportions)
A - 0.98 (0.95-1.01)
B - 0.85 (0.76-0.93)
C - 0.93 (0.79-1.06)

I'm a bit of a newby when it comes to meta-analyses, but this can't be correct right?
Hopefully you know what I need to do, to correct this!

Thank you very much in advance!

[jonathon]

hi tessa,

try reaching out to kyle hamilton ... you should be able to find his email somewhere. he doesn't spend much time in these forums.

cheers

jonathon