Observational studies of COVID vaccine efficacy are riddled with bias/ Not counting cases 14 days after dose 2 is a problem
A look at Peter Doshi's paper
I finally had the chance to read this paper by Peter Doshi, a University of Maryland professor and researcher. It’s quite clever and teaches principles of epidemiology. I am going to summarize a portion of the paper, and in future installments, I will flesh out his other points.
Point #1: Many observational studies of vaccine effectiveness exclude cases that occurred within 14 days of dose 2. In other words, anyone who gets COVID 36 days after the first shot (doses are 21 days apart for Pfizer) doesn’t count against the vaccine. This is because— the argument goes— it takes time for the vax to kick in, so you can’t hold those early days against it.
Of course that’s a silly argument. A medical product owns anything that happens after you start taking it. Of course, if vaccines take a million years to kick in, they would be useless. If they take a full year, and everyone gets covid in that year, they would be useless. Omitting these cases is irresponsible, and yet it continues.
But Doshi’s point is that this can make an inactive product— something totally useless— look like it works. He provides this thought experiment.
In the experiment, he says, what if we compare the control arm of the Pfizer study against an imaginary vaccine arm. And for the thought experiment assume the vaccine is useless. As the table above shows, both groups have identical numbers of covid cases— just what you would expect from a useless vaccine. A straight forward analysis shows no benefit (second to last row)
But in the ‘fictional vaccine observational study’ cases are excluded for 36 days. When this is done the useless vaccine, looks like it reduces infections by 48%!!
Doshi makes a very good point in his paper that the solution is to subtract the 36 day infection rate from the observational control arm. Sadly most investigations don't do that.
This is one of several biases Doshi discusses, and it plagues the vaccine literature. More next time. Subscribe.
Fenton, Gato, and others - some of whom you probably blocked - have been all over this from the start. And it's not just "some" studies that do the 14 day exclusion. It's ALMOST LITERALLY ALL OF THEM:
https://www.researchgate.net/publication/378831039_The_extent_and_impact_of_vaccine_status_miscategorisation_on_covid-19_vaccine_efficacy_studies
https://wherearethenumbers.substack.com/p/vaccine-efficacy-cheap-trick-by-exclusion
And the handful of studies that don't do that TEND TO SHOW NEGATIVE EFFICTIVENESS IN THE FIRST 2 WEEEKS. And that is even though an acute healthy vaccinee effect is helping to conceal it.
And as you know Vinay, the healthy user bias is another beast and that has been shown 7 or 8 times. You've only discussed one or two of those occasions. It's not just the boosters. The dramatically lower non-covid mortality among the vaccinated has been going on right since the original 2-dose series.
By the way, Fenton has shown that the misclassification bias can make a placebo or even a negatively effective vaccine look exactly like a waning efficacy curve. THAT'S A PROBLEM.
So to recap, before even accounting for the biases, there is a signal of negative efficacy in the first 2 weeks, and an outright statistically significantly negative VE against cases after 5+ months of waning.
So once you account for the misclassification biases AND the healthy user biases, the data is entirely compatible (i.e. can't even exclude) with the hypothesis that vaccines have led to a net increase in covid deaths and all-cause deaths.
Vinay, you are partially awake, and a lot of people are waiting for you to get the rest of the picture. I second the sentiment that once in a while you need to actually read the comments and learn from them and respond.
Edit: Another commenter reminded me that a negative VE in the first two weeks would further cause higher natural immunity in the vaccine group. Vaccines can do no wrong it seems.
Edit: See my other comments explaining why trials do not exclude a possible acute negative VE.
Please correct me if I am not understanding correctly. So Doshi has shown that the observational studies' findings for Covid vaccine efficacy are seriously flawed, is that right? And this is supported by the fact that most people who got those vaccines ended up getting Covid anyway, even with multiple and updated boosters, correct?
It looks quite straightforward.
So why aren't the science community/medical community/government health officials all saying some version of, "hey! We thought we knew what we were doing, but it turns out we missed something pretty big! We're going to fix that, and we'll make sure it doesn't happen again going forward."
There were quite a few people who lost their jobs temporarily (some even permanently) for refusing the vaccine, even after the CDC admitted that it didn't prevent transmission.
And I know at least 3 people who got the vaccine under duress -- because the alternative was to lose their jobs -- who had new-onset autoimmune issues within a week or two of getting vaccinated.
Leaving government out of it for now, if the science/medical community does not make this right, and very soon, I think they will never regain public trust. Surely they know this?