This is not the calculation being made. Using your numbers, experimenting on little girls needs to be at least 1.001 times as effective as experimenting on chimpanzees or mice to be worthwhile (because then you save an extra thousand lives for your thousand girls sacrificed.) It's not a flat "little girls versus millions of malaria deaths."

Point taken.

This is, quite frankly, not clear to me, and I'd want to call in an actual medical researcher to clarify.

Well, yes. I doubt that JAD has particular expertise in malarial research, I don't and neither do you. To know whether a malarial research programme would benefit scientifically from a supply of humans to experiment on with no more restraint than we use with chimpanzees, one would have to ask someone with that expertise. But I think the hypothesis prima facie plausible enough to conduct the hypothetical argument, in a way which merely saying "suppose you could save millions of lives by torturing some children" is not.

After all, all medical interventions intended for humans must at some point be tested on humans, or we don't really know what they do in humans. At present, human testing is generally the last phase undertaken. That's partly because humans are more expensive than test-tubes or mice. (I'm not sure how they compare with chimpanzees, given the prices that poor people in some parts of the world sell their children for.) But it is also partly because of the ethical problems of involving humans earlier.