Though, in that analogy, we've been used to the test for water damage being to turn on the hose briefly to soak an item, perhaps with a small bucket of water and a stirrup-pump, then stop it, take the bucket and pump away and spend time looking at the item, seeing that it doesn't seem to do much and discussing this in a meeting across town, then getting an item that's smouldering and turning the hose on again (perhaps a bigger hose and a few buckets of water) to drench it, before ceasing again while meetings are had about what happened, then finding some item that's on fire and again starting with the squirting.... Etc, until eventually you're allowed to put out fires around a whole house (so long as the fires are set upon items muchike your test ones) and/or dampen things next to the current flames to stop them catching, this time having called out a fire-tender and/or hooked up to the water mains.
What has been rone here is to quickly agree what problems there are (e.g. don't squirt water at oil-fires or electrical items), based upon prior fire-fighting experience, and what signifies good or bad quenchings, then start with a small amount of water-flow, see that it gets a sample sofa (not on fire) wet, but no obvious issues that prior sofa-drenching has proved to be troublesome, re-aiming on the smouldering easy-chair to see that it stops smouldering, then direct the stream onto the rest of the furniture that's in danger, thus saving more of the house from fire, knowing that there still might be irecoverable water-damage for some things, but nothing that should cause neighbours the worry that a whole burning house would and a lot more rectifiable than if most of your precious things are now a mix of simple/complex oxides in the air or on the ground, and most of the rest melted at least cracked.
If the hosepiping was shown to be actually harmful (you were trying to use gasoline, not water) or the household items themselves unexpectedly problematic to pour water on (those chesspieces you had cast from pure lithium might not be helping!) then you would have rushed to stop the water (but not send the fire-tender away) and work out what else you can do (powder? foam?).
And if the fire, unbeknownst to you, has ready reached the neighbouring roofspave then you can quickly drench that and/or get some of the firefighters to make preventative fire-breaks exactly how they've been trained.
In reality, the vaccine programmes were not rushed, merely compressed out of expediency. Factories were set up to start producing vaccines very early on, while the initial in-human studies were still being looked at, at the risk that the designed-for vaccine would be scrapped (as they could easily be) and the massive production investment would be wasted, or need to be reallocated. The same levels of scrutiny (more or less, and possibly more!) were cast upon the trial results and some candidate treatments/protectives were indeed abandoned (as they would have been in the more long and drawn-out versions of investigations). Even some (most?) of the successful ones had to have pauses during or between certain phases of study, just to check that unrelated deaths (in at least one case, I think it was a car-accident!) were in fact truly unrelated.
Which is not to say that the fast-track (or, rather, concertinaed) study plan should be used for everything from now on. Novel painkillers don't need this as much. HIV treatments probably can be developed without, given the current arsenal of preventatives and mitigating treatments. (In both examples, a very few hard cases may be temptingly used to urge the 'rushing' of ways to help individuals who aren't responding to these existing ones. That's for the experts to argue over, though.) Yet it should show what can be done, without making things go bad through insufficient regulation or other errors.