[jamez1]

They've made a terribly flawed assumption which drives that 90% number. VE is modeled using a beta binomial model with a fixed prior of an infection rate of 1.5% per year. Obviously this number is going to vary a lot depending on the sample. [1]

By putting a fixed point vs the appropriate wide ranging distribution as the prior, the model scores the results far higher than otherwise.

This is a non-result and it was wildly inappropriate to publish this as proof of any finding.

[1] https://pfe-pfizercom-d8-prod.s3.amazonaws.com/2020-11/C4591... Page 15

[laGrenouille]

I don’t think that’s what the report is saying: The 1.5% is used to determine how long the trial will take to reach the required number of cases. Not sure what prior they were using, but my guess is that it’s conservative.

[awillen]

I have no idea what this means, but it sure sounds like a legitimate concern. Can anyone who actually understands this confirm?

[sanxiyn]

As far as I can tell attack rate assumption is there only to calculate desired sample size, and statistical analysis of vaccine efficacy is independent of attack rate assumption.

It seems to me exact prior is unspecified in protocol. The usual choice is uniform prior.