The previous three posts, while I think addressing important ideas in themselves, were intended to set up another point. Most people agree that the standards for testing experimental medicine on people should be very strict, that informed consent should always be explicitly obtained, and that no one ought to be forced to be a experiment subject against their will. Most of us believe this even when we consider that our medical capacities would progress far more quickly, and we could save countless lives, if we were much more devious in our experimental guidelines.
Of course, you could say that the negative consequences of living in a world with such duplicitous testing methods would end up outweighing any potential benefits. This, however, is uncertain, and it depends on how much more lax our standards of practice would become. Regardless, there's an easy fix to this theoretical complication. Imagine a world in which the duplicitous testing methods were unknown to most people, perhaps only done on the orphaned, or on isolated populations. Still, I think, most people would find this extremely perverse. The ends, as they say, cannot justify such foul means.
Which is all just to point out that those who point to the successes of non-human animal experimentation, in its defense, are missing the point. It's easy to acknowledge that there have been positive outcomes from unjustified forms of medical experimentation. We have learned a lot, and benefited a lot, from various forms of injustice (depending on who you count as "we"). This does not get you on the path to justification, as most people would agree in the human case. There are certain practices we agree we should not engage in, despite the potential for positive outcomes.
All of which doesn't prove my point, which is that most non-human medical experimentation is wrong. But it does rebut the most common defense of such practices. The work a defender of such practices has to do is show that these positive outcomes for some reason serve as justification in the non-human case, when they wouldn't in the case of humans. (It's also important to remember that much non-human animal experimentation is not even justified on utilitarian grounds. That is, the potential human benefits of many experiments often do not even come close to balancing out the costs in terms of non-human welfare and life. Many utilitarians oppose many forms of animal testing.)
A further complication for the defender of animal testing is this. The closer an animal is in terms of similarities to us (e.g., primates), the more useful it's going to be as an experimental model. But primates and animals that have a lot of similarities to humans are often thought to have a closer moral status to humans than more distantly related animals(say, birds.) But often, the farther an animal gets in terms of similarities to humans, the less helpful it is as a model for humans.
One final note: all of this discussion assumed that legitimate humans benefits would be gained for any animal testing. This assumption is often dubious, if not wholly illegitimate. I've heard of no one who plausibly defends harmful animal testing for frivolous need (make-up, hygiene products), and yet the vast majority of people support such practices.