[innomatics]

It is great news.

I didn't see the case numbers but I've inferred to arrive at 90% that there were 9 cases in vaccinated group vs 85 in controls. (95% C.I. 80-99%)

Very interested to see forthcoming results from competitors that don't face the -80 degrees issue.

[StavrosK]

Out of 45,000 people, only 86ish people got infected in the trial arm? Isn't that too few? Even for the control arm, that's still only 774 people infected.

[BurningFrog]

No, 94 people total got infected. At least 85 were in the control group.

Those are the numbers that matter, statistically. If the total number of people in the study were 2 thousand or 2 billion doesn't enter into the math.

[OJFord]

That's not true, population size absolutely 'enters into the math'. 9 vs. 85 does not give the same result independently of whether N=94, 200, 2k, or 2B.

It affects confidence, and potentially the result of the test entirely. I won't be more specific because stats was never a strong suit and my memory's hazy, but I remember enough to know it matters!

[innomatics]

I'm thinking that 85/21,750 or 0.4% of the controls developed the disease (vs 0.04% of those who received vaccine).

It does seem low for the US where you have 3% of the population had disease already in 2020. Perhaps more data came from Germany? Or it could be they measure not just PCR +ve results, but antibody levels and/or severe symptoms?