I'm conducting a meta-analysis, and some of the studies that I have included only report the number of occurrences of the event of interest without an accompanying standard deviation. As such, I am using the formula $\text{SD} = \sqrt{(p(1-p)/n)}$, where $p$ represents the sample proportion and $n$ represents the sample size, to estimate a standard deviation for these studies. However, two of the studies report 0 events of interest. This yields $p=0$ and $\text{SD}=0$, which is obviously not very helpful in constructing a confidence interval.
Is there another estimate of standard deviation I can use in these cases or any other way I can construct a confidence interval with these values?
