We can't assume that there are no biases, but we can certainly use it to quantify the difference. ETA: And, if someone now wants to argue that the vaccine does

We can't assume that there are no biases, but we can certainly use it to quantify the difference. ETA: And, if someone now wants to argue that the vaccine does not work, they will have to come up with a plausible 20x confounder. Which in itself would be quite an achievement.