You seem to be saying, in essence: allowing gays to donate doubles the number of instances of transfusing HIV while only increasing the number of lives saved by 7-10% (or 20%, or 2% - it doesn't matter). But that's the wrong way to look at it. Even if 10% of all transfusions from gays resulted in an additional HIV infection (thanks to testing etc., I'd guess it'd be well below 1%), that still means that, for any given transfusion candidate, you have a 90% possibility of saving their life without transfusing HIV.
The problem is that the percentages are higher than that. The biggest problem is that testing won't always pick up HIV and AIDS, because these diseases can lie "dormant" for quite a while, but yet they are still in the blood (undetectable) and will cause infections somewhere down the line.
Actually, excluding 7-10% from the pool may not mean any lives lost. It will just decrease the available blood supply, which can be compensated through many other means.
Excluding certain categories is a simple and effective administrative solutions - and doesn't really make any sense to confuse it with anything else.
You seem to be saying, in essence: allowing gays to donate doubles the number of instances of transfusing HIV while only increasing the number of lives saved by 7-10% (or 20%, or 2% - it doesn't matter). But that's the wrong way to look at it. Even if 10% of all transfusions from gays resulted in an additional HIV infection (thanks to testing etc., I'd guess it'd be well below 1%), that still means that, for any given transfusion candidate, you have a 90% possibility of saving their life without transfusing HIV.