[eluser]jedd[/eluser]
I'm not sure how expensive those calls are, so it might be easier / comparably fast to just have a function in your method, that you call at the end of your insert() function, that does a lookup for the data you just inserted. Because it's a PK, and consequently both columns are indexed (and you might get some benefit from the fact the disk was just there, at the file system level) it should be very fast.
A new thought, however, pops into my head.
If your PK is a compound key, and you just used the two components of that key to generate your INSERT query .. why don't you just use the two components that you presumably have to hand in your variables?
Usually you want to look up the auto-incremented ID column (ie. a non-predictable PK) for a row you've just inserted - but this obviously doesn't apply in your situation.