probably some hint of the problem in defeating, as used in the title, you could probably gauge some statistically indicated benefit by comparing various groups, but can’t know exactly of whatever single example that the benefit was from use of the drug, add that there maybe adverse effects from the drug, which may be difficult to separate from immune system damage, by the immune system, or more generally somewhat enhanced by complex chemical insult. So there’s a dimension of chemical insult potentially.
so where there’s possible effects from chemical insult you’d expect medicine to proceed with caution
back to the use of the word defeating, does a person speculatively apply that to an individual example that used the drug and recovered, then apply it as a more general cure likely to help most people with infection (or with rapidly declining health from the infection)
so, i’d guess the problem is the medical people can’t measure and know in advance what the chemical insult from using the drug might be, in any particular individual, they can’t really give reliable advice even of a statistical nature that’d represent the possible risks