I think that's partly a survivor (the disease surviving, not the patient) bias effect.

Things that could be solved by random discoveries are no longer considered the highest-grade nasties. There were a lot more intractably fatal conditions in 1870 than there are today.

So the likelihood of there being answers that could have been randomly discovered by medics with 1870 or 1920 levels of knowledge is tiny. At the same time, the sum of human knowledge has expanded so rapidly since then, it's not impossible for stuff to get missed.