Probability and Risk: Is vaccine efficacy a statistical illusion?

The article explain why studies with a placebo branch are important. You have a lot of known unknown and unknown unknown. In particular, as discussed in the article, counting the number of cases/death is hard.

The main problem with the article is that all the vaccines are approved after a RCT were the vaccine is compared to a placebo. For example, look at the slide #6 in https://investors.biontech.de/static-files/53f0968a-279b-4f8... and more details in https://en.wikipedia.org/wiki/Pfizer%E2%80%93BioNTech_COVID-...

As was mentioned in the comment section of the article, it would be interesting to see if this effect really exists (by trying to shift the numbers backwards).

Furthermore, I think it is a striking reminder that even (seemingly) simple data like mortality can be subject to strong bias.