Generating an average of 3 x variables with a rule that at least 2 need to not be missing

Hi Folks, simple question here that I can't get my head around.

I have 3 x blood pressure values per participant. I can take an average like below, which ignores missing values (good):

Code:

egen SBPavg=rmean(BPSys1 BPSys2 BPSys3)

However, if a requirement is that I need AT LEAST 2 x blood pressure values to generate a valid average, what is a robust way to code for this (i.e. just 1 x blood pressure value is not sufficient, but taking an average of 2 or 3 values is valid and what I want)?

Thanks,
Patrick

Share on FacebookPlus on Google+