[mansr]

Both groups have (approximately) the same size, in the tens of thousands. The first evaluation is done when 90 or so cases have been confirmed, not knowing which group they belong to. The blinding of these subjects is then unsealed, in this case revealing that 90 of the 95 confirmed infections occurred in the control group. This distribution means the vaccine has been highly effective at preventing infection.

[StavrosK]

Sorry, I meant just 90 people infected (obviously the vaccine arm has much fewer), but after running the numbers that's actually about right for the prevalence rate we see in the US.

[mansr]

This isn't the final evaluation. There will be another at ~150 cases. The numbers may seem small, but this is enough for the purpose of deciding for or against approval. Anything over 80% effectiveness would likely be approved, so it doesn't matter whether the true value (insofar one exists) is 90%, 95%, or something in between. It's useful regardless.

[StavrosK]

Oh, I didn't know that, thanks. I'm sure it's enough for statistical significance given the difference, I was just wondering why there were so few cases in 30k people, but it looks like that's just the normal case rate.