Good question--especially as no trials have ever shown that even screening the general population for real diabetes has an effect of mortality. The first RCT to attempt this was just recently released:
http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(12)61422-6/abstract
