No. It is that the med under test could perfectly treat the small number diagnosed who actually have the illness it addresses, and have nil or negative effect

No. It is that the med under test could perfectly treat the small number diagnosed who actually have the illness it addresses, and have nil or negative effect on everybody else given the same diagnosis who have an actually different illness. According to the RCT, it just failed. Validity of RCT results is limited absolutely by validity of diagnosis. Bad diagnostics, bad results. That doesn't stop people from running invalid RCTs, so it should stop people from citing invalid RCTs. But it doesn't.