You are obliged to conduct a natural experiment. We do this thing (vaccination) we expect will work, on some fraction of a test population, and then we wait. Some other tricks could work, in theory but they're often unethical and unlike this experiment they aren't a direct model of the real world. "Does it actually work?" is an incontrovertible test.
But it should take years to do this, because obviously hardly anybody is going to get some horrible deadly infectious disease right?
Unless, as happened in most of the world, you have exponential pandemic case growth as measures to prevent spread are half-arsed or just not implemented at all.
Then it turns out - good news - that's a great environment for proving your vaccine works quickly. Shame about all the dead people though.
I recommend https://www.deepl.com/translator for english translation, as I don't have the time right now unfortunately.
I don't have any background in this, so if anybody can chime in with more knowledge than me it would be appreciated.
But to me the point she is making sounds reasonable.
They had a placebo arm and a vaccine arm. Look at figure 2 on page 30. You can see that new COVID cases in the vaccine group virtually stop on day 12 while the placebo arm continues to grow.