1) It is an experiment. The point is to test something that nobody is sure about. Whether the result is expected or unexpected there isn't really such a thing as "unsuccessful"; even duplication of work is considered useful.
2) If you think the cost-benefits are bad, my advice is don't sign up to be experimented on. Nobody has to be experimented on if they don't want to be.
1) Experiments, by their nature, must spend resources in order to produce tangible results. In this case the relevant results are saved lives, and the expenditure is often going to be lives. You framed your comment in these terms, I’m arguing the costs very well may not be justifiable by the benefits. You’d need to make a case for that.
2) What if the families of experimentees receive payment? Allowing that would be a short way down the slippery slope from allowing the experimentation, and would make the matter of consent more difficult to assess.
2) If you think the cost-benefits are bad, my advice is don't sign up to be experimented on. Nobody has to be experimented on if they don't want to be.