1/ Important question: Once infected, what is the effect of vaccination status on death/severity?
Here is relevant statistics from Israel data.
Disturbing observations on the early period after 1st dose.
Here is relevant statistics from Israel data.
Disturbing observations on the early period after 1st dose.
2/ Table shows rate of death & severe disease, since the vaccination campaign began.
Notice that risks are higher in vaccinated than unvaccinated. That’s misleading. At any time, vaccinated were older than unvaccinated (and older, on average, in sequential vaccination groups.)
Notice that risks are higher in vaccinated than unvaccinated. That’s misleading. At any time, vaccinated were older than unvaccinated (and older, on average, in sequential vaccination groups.)
3/ The phenomenon is called confounding. Illustrated in a causal diagram. All comparisons are not valid. They don’t estimate effects.
4/ CAVEAT: data taken from this table. Looks official but no source. Number of cases matches OWID and DASHBOARD. But number of deaths does not. Table show ~400 fewer deaths. Unclear why.
5/ Nonetheless, we observe something peculiar: comparing 2nd dose (within 7 days) vs. 1st dose period, the probabilities do NOT increase. Even decrease.
6/
Why?
2nd dose within 7 days is not supposed to offer much benefit as compared with 1st dose. So, we would expect to see the usual confounding. Higher CFR/severity.
We don’t!
Why?
2nd dose within 7 days is not supposed to offer much benefit as compared with 1st dose. So, we would expect to see the usual confounding. Higher CFR/severity.
We don’t!
7/
Graphs illustrate the findings: observed trend vs. hypothetical trend. The latter shows the kind of monotonicity we expect to see in the data. (Again, recall: these trends don’t show effects, but rather trends in confounded associations.)
Graphs illustrate the findings: observed trend vs. hypothetical trend. The latter shows the kind of monotonicity we expect to see in the data. (Again, recall: these trends don’t show effects, but rather trends in confounded associations.)
8/ There is more.
Table: data for age 60+ (previous reports).
Similarly, last week of 1st dose is not expected to show much benefit as compared with the first 2 weeks. So, we expect to see the usual confounding: higher CFR/severity percentage. Why are we observing the opposite?
Table: data for age 60+ (previous reports).
Similarly, last week of 1st dose is not expected to show much benefit as compared with the first 2 weeks. So, we expect to see the usual confounding: higher CFR/severity percentage. Why are we observing the opposite?
9/
At least two explanations:
Confounding somehow weaker when 1) comparing later time of 1st dose with early time, and 2) comparing early time of 2nd dose with 1st dose
And all is good.
At least two explanations:
Confounding somehow weaker when 1) comparing later time of 1st dose with early time, and 2) comparing early time of 2nd dose with 1st dose
And all is good.
10/
Or, not so good.
There is no reason why strong confounding would be evident for fully vaccinated (vs. 3 other groups) but will be weak or absent when comparisons are made between early 2nd dose, late 1st dose, and early 1st dose.
Or, not so good.
There is no reason why strong confounding would be evident for fully vaccinated (vs. 3 other groups) but will be weak or absent when comparisons are made between early 2nd dose, late 1st dose, and early 1st dose.
11/
We are not observing consistent benefits over time, but rather evidence of EARLY HARM
Early period after 1st dose might be a riskier period to be infected!
(1st dose increases risk early on, if infected)
Like a surgery that increases risk early on and decreases risk later
We are not observing consistent benefits over time, but rather evidence of EARLY HARM
Early period after 1st dose might be a riskier period to be infected!
(1st dose increases risk early on, if infected)
Like a surgery that increases risk early on and decreases risk later
12/
Who really knows what happens to the immune system when it is simultaneously exposed to the virus and to the 1st dose of the vaccine?
After all, part of the pathology is due to exaggerated immune response to the virus.
Who really knows what happens to the immune system when it is simultaneously exposed to the virus and to the 1st dose of the vaccine?
After all, part of the pathology is due to exaggerated immune response to the virus.
https://twitter.com/prof_shahar/status/1359612825568645120
• • •
Missing some Tweet in this thread? You can try to
force a refresh
