How to get p-value for a riskratio in meta-analysis

[rollanders]

I want to make meta-analysis on the impact of Covid on an outcome of a disease, expressed as a RiskRatio for complicated/noncomplicated disease during the pandemic compared with the prepandemic period. I use

meta esize ComplicatedCovid NoncomplicatedCovid ComplicatedPre NoncomplicatedPre , esize(lnrratio)
meta summarize, transform(exp) random subgroup(source)

This does not give me any p-value for the pooled effect size.

If I use only one group I get Test of theta = 0. As the expected RR=1 I want to test if theta=1

This works well
metan ComplicatedCovid NoncomplicatedCovid ComplicatedPre NoncomplicatedPre, rr random by(source)
This gives me Tests of subgroup effect size = 1:

I also want to analyse the impact on the number of observed cases during the pandemic compared to pre-pandemic period. In other words this is a ratio of two counts. Some would call it Standardized Incidence Ratio even though we have no infrmation on the catchment population.
I have used the smrby plugin to get SIR and CI and from there I can get the SE and finally

meta set SIR SE
meta summarize

But also here I get the test of Theta=0, instead of 1.

In both cases, the confidence interval for some subgroups includes 1 but I get significant p-values

How can I get the correct p-value for these kind of analyses.

[Tiago Pereira]

The incidence ratio (IR) estimates should be on the logarithm scale. The standard error should be the standard error of log(IR).