Dear All,

Many thanks in advance for reading this - any help is gratefully received!

I'm trying to combine the incidence rates for a complication of surgery (a ratio that i'm using as r). I am using the Hedges method Random Effects Model (as i have hetrogeneity in my studies). I want to combine each studies incidence (a ratio of cases/exposures) to get an overall mean incidence with confidence intervals (CI) and credibility intervals(CR)*.

After computation my lower CI and CR are both negative.

I don't really understand the CR so don't know if it's reasonable for this to be negative. but with the CI I plan on just using 0 as my lower interval as you can't have less than 0% complication rate.

In summary I guess i want to know:
a) is it ok to put 0 as my lower CI?
b) What do I do with the credibility interval?

Fingers crossed,
RDF
(General novice)

*There are a few studies that compare treatment A vs. treatment B to prevent the complication but I was going to treat each arm as a separate study and then combine all A, all B etc. but again i'm not really sure if that's ok

example
Study x - 8 complications in 100 patients = 0.08
study y - 4 compplications in 200 patients = 0.02
study z - 9 complications in 300 patients = 0.03

Mean complication rate (after weighting) = ?? (CI ?-?, CR ?-?)