Discover more from Where are the numbers? by Norman Fenton and Martin Neil
How to create the illusion your vaccine is 90% effective
....even when those vaccinated get infected.
Spoiler: A major study claimed the covid vaccines are over 90% effective. But when you look at the details of the study you find that a whopping 37.2% of all vaccinated participants who were tested within 14 days of the first dose were confirmed as covid cases. None of these ‘cases’ were counted in the efficacy calculation. Also, out of the subset of 1,482 participants with confirmed symptomatic covid, that were part of the study, not a single one died, despite 812 of these being unvaccinated.
Deception is now endemic, and this article describes the ‘how to guide’ on manufacturing high efficacy illusions.
Here we outline an easy to follow five step fool proof method to ensure a vaccine will be accepted as highly effective and look at a specific research paper to illustrate how this is done.
Step 1: Suppress legitimate criticism
We have reported the statistical tricks and biases commonly used in studies of vaccine efficacy many times. These tricks inevitably lead to exaggerated claims (“95% safe and effective”) that are echoed throughout the mainstream media and are routinely used by censorship units like Full Fact as ‘proof’ to close down debate. We have also shown how, even when the flaws and biases committed by these studies are meticulously laid out in letters to journal editors, they refuse to even publish these letters, let alone retract the studies. A good example is our recent experience with Lancet and its published study that supposedly ‘confirmed’ the 95% efficacy of the Pfizer vaccine. The same happened with the New England Journal of Medicine where another study reached the same 95% efficacy conclusion - they refused to publish this response by Mark Reeder). These widely publicised studies essentially ‘sealed the deal’ for the efficacy of the Pfizer vaccination in the eyes of the public.
Where are the numbers? by Norman Fenton and Martin Neil is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Step 2: Publish in a (bought and sold?) ‘reputable’ journal
Nothing has changed. Since then, there have been many even more fundamentally flawed studies that make similarly exaggerated claims. However, the study of 109,865 healthcare workers in the 20 weeks from December 28, 2020, through May 19, 2021 (reported in 'Effectiveness of mRNA Covid-19 Vaccine among U.S. Health Care Personnel' - https://www.nejm.org/doi/10.1056/NEJMoa2106599) is one of the most startling examples yet. The following screenshotd are all from that paper.
The study looked at both Pfizer’s BNT162b2 and Moderna’s mRNA-1273 and concluded the following:
In fact, for ‘any’ of the vaccines the effectiveness of ‘complete vaccination’ was stated to be 90.4% (with Moderna coming in at a whopping 96.3% and Pfizer at 88.8%). This information is all provided in their Table 3:
But before we look at the revealing information provided in the table, it is important to understand what kind of study this was. Most people assume that if you want to evaluate the efficacy of a vaccine in an observational study you simply compare the outcomes (e.g. infections, hospitalisations, deaths) of those in a large cohort of vaccinated people with those in a large cohort of unvaccinated people and then adjust for any known confounders. But that is not what was done here. This was a so-called “test-negative case–control study”. What they did was look at a set of people who tested positive for covid at some time during the 19 weeks of the study - they selected 1,472 out of the 8,365 testing positive (so this would contain a mixture of vaccinated and unvaccinated). These are the ‘case participants’ in the table above. Then they try to find a set of ‘matching’ people who only ever tested negative during the 19 weeks. Again, this would contain a mixture of vaccinated and unvaccinated. These are the ‘control participants’ in the table above.
This method is similar to this study described in a recent post and (as explained there) we feel there are problems with this method. But any inherent problems with the method are not relevant to the major flaws we now consider.
Step 3: Ignore infected cases that are vaccinated
The main trick to getting to the high efficacy figures is to completely ignore vaccinated people who get infected within the first 14 days after their first dose - they are not even counted as ‘partially vaccinated’ since that only applies to people post 14 days of their first dose and pre 7 days of a second dose.
So, imagine the most extreme case in which every vaccinated person gets covid within the first two weeks of their first dose. Then, assuming (as is likely) that none get infected a second time within the 19 weeks, according to the study definition no vaccinated people ever got covid over the whole period of the study.
If only one person in the the unvaccinated comparative cohort had got covid, over the same period, the vaccine efficacy (defined as one minus the proportion of vaccinated infected divided by the proportion of unvaccinated infected times 100) will be reported as 100%.
Now, while it is certainly not the case that all the vaccinated here got covid within the first 14 days it seems that quite a lot of them did. If you look at the first two data rows of the table you can see a total 353 out of 948 (37.2%) who were tested within 13 days of their first dose were positive (the percentage is even higher, 40%, for those tested within the first 10 days). Does that sound like a vaccine that is effective in anything other than giving people covid?
Step 4: Don’t look for covid if you don’t want to find it
But the paper states “During the study period of December 28, 2020, through May 19, 2021, a total of 109,865 health care personnel were tested across the participating sites; of these persons, 8365 (7.6%) tested positive for SARS-CoV-2.” So, over the whole 19 week period 7.6% of that population got covid. That means (crudely) on average in any 2-week period the infection rate was about 0.8%. Yet the infection rate for people being tested within 2 weeks of their first jab was 37.2%!
Unfortunately the study report provides no information whatsoever on how frequently people were tested. In theory if only people with covid symptoms got tested then having a high proportion of those being confirmed positive would not be surprising (if the test was accurate). But, as these were all healthcare workers, it’s likely they were tested very frequently and we know that during the first two weeks after vaccination it was fairly routine to get tested. If people were tested every two weeks then we could reasonably conclude the vaccinated were getting infected – within two weeks of their first jab – at a rate that was almost 50 times greater than the general rate for this population.
So if you dont look for covid, by not testing for it, or by ignoring the test results you wont find it.
Step 5: Ignore outcomes that make your vaccine look ineffective
And here's another nugget from the study: of the total of 1482 participants with a positive test and at least one Covid-19–like symptom there was a grand total of zero deaths: an infection fatality rate (IFR) of 0%. And 812 of those were unvaccinated. Bear in mind that this when covid was supposed to have been rampaging globally and causing widespread death. And of course that nugget somehow never got mentioned in the abstract, mains results, conclusions, or discussion. It only appeared in the detailed results section (along with the fact that only 2% were hospitalized):
Conclusion: Why all observational studies (and many RCTs) of vaccine effectiveness and safety exaggerate vaccine effectiveness claims
All studies on vaccine effectiveness suffer from one or more (often most or even all) of the following systematic biases:
Misclassificaton: Participants who got a PCR positive test within 14 days of first vaccination (resp. second, third etc) are either classified as unvaccinated (resp. 1-dosed, 2-dosed etc) or simply not counted. Yet in many studies there are especially high infection rates for the vaccinated during this period. As explained here this would result in high efficacy rates even for a placebo vaccine.
Delayed reporting If reporting of covid cases is delayed (e.g. by a week or two) during vaccine rollout then this results in exactly the same illusion/exaggeration of efficacy as misclassification.
Illegitimate comparisons: Efficacy assessed by comparing the never vaccinated only with the ‘fully vaccinated’ (based on whatever the definition of fully vaccinated was at time of study, e.g. at least two weeks after second dose, third dose etc).
Different testing protocols between vaccinated and unvaccinated: People who were unvaccinated were generally required to get tested far more frequently than the vaccinated even if they were asymptomatic. Hence, covid positives were more likely to be found in the unvaccinated irrespective of the true difference in susceptibility between groups.
Survivor/selection bias: People who were symptomatic or PCR positive when called for vaccination were recommended to wait until they were PCR negative before being vaccinated; this means that all such people had natural immunity when they did get vaccinated and hence were less likely to subsequently get covid.
Study took place during period of naturally falling infection rate: Timing of study coincided with period when infection rates were decreasing: This would be certain to create a statistical illusion of efficacy.
Vaguely defined outcomes: By not being explicit about the outcomes and end dates for the study, many studies can simply choose which outcome makes the ‘best case’ for the vaccine. So, in the above example, only in the detailed results do we find there were no deaths (in either the unvaccinated or unvaccinated) and almost no hospitalizations); hence the impact of the vaccine on these key outcomes were conveniently ignored.
Where are the numbers? by Norman Fenton and Martin Neil is a reader-supported publication. To receive new posts and support our work, consider becoming a free or paid subscriber.