Meta-analysis for less than 4 studies using metandi?

#1
Bearing in mind that metandi command requires a minimum of four studies, which is appropriate command in case of less then 4 studies?
There is a recommendation that separate pooled estimates of sensitivity and specificity should be done using random-effects logistic regression. In that case can I do it by using xtmelogit?
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Not a regular STATA user, I wonder if with < 4, you should just try to run your own random effects models? How many studies do you have?
 

hlsmith

Less is more. Stay pure. Stay poor.
#4
I have not seen the number of studies restriction on programs before. I wonder what the rationale is for it. Understanding that may help you. Is there documentation available to the STATA programs?