It would be fairly easy to make it work normally, from a coding PoV, to just use a double or a float and then cast it to an int.
Note that this demonstration is fairly simple.
Instead of the current method, something like this:
int attack = 10;
int hp = 100;
attack = attack + 1;
attack = attack / 2;
hp = hp - attack;
It could be possible to do something more like this
int hp = 100;
double attack = 10;
attack = attack + 1;
attack = attack / 2;
hp = hp - (int)attack;
After those two examples, it's the same end hp. You then double both, and you have 10 and 11, respectively. Obviously the code doesn't look exactly like that; this just illustrates the point. Also, this may not be valid in Flash or whatever; it's a workaround in Java/C++, as those are the languages I know.