Comment on Bozio et al, the CDC study claiming vaccine efficacy better than prior infection
This is a repost of a comment I made at Steve Kirsch’s Substack relative to the CDC study claiming vaccine efficacy > immunity from prior infection. This is getting into the statistical weeds, but this is about a fairly consequential statistic. Please see link to Steve’s article for context. The comment is:
Agreed. Setting aside issues with statistical adjustments etc, the study essentially uses the rates of prior infection and vaccination in patients hospitalized with COVID-like illness (CLI) as a surrogate for community rate of prior infection and vaccination respectively. Fair enough. But in their study they find 1020 prior-infection and 6328 prior-vaccination subjects CLI subjects for the time period Jan-Sept, 2021. That corresponds to a vaccinated-to-prior-infected ratio of 6.2 in the community. In other words, the claim is that "on average", over that time period, 6.2 times more people were vaccinated than infected in the community. That is hard to believe. That ratio is almost certainly less than 1 at the start of the trial period and probably between 1-3 by the end of the trial period (estimate of < 2.3 on July 26 here: https://www.theguardian.com/us-news/2021/jul/26/us-covid-cases-undercounted-study). The premise of the study is not credible.
I make the case that the mis-science by the CDC and other federal medical authorities is not a bug but a feature: