An Elegant Demonstration of how Efficacy can be conjured out of thin air

Hat tip to Norman Fenton

Photo by Julius Drost on Unsplash

The graph above is reminiscent of the early days of the Covid-19 vaccine campaign. What started out as a highly effective preventative therapy touted by our authorities as 100% effective was soon shown to wane in its ability to thwart infection and transmission of SARS-COV2.

By August 2021, the FDA authorized a monovalent booster to counter the diminishing benefit of the primary series. It wasn’t anyone’s fault. “Vaccines” can wane over time. Moreover, a new variant, Delta, exploded everywhere in the summer of 2021. The mutated strain now possessed a new ability to evade immune systems primed by the original vaccine.

The solution was to give everybody another injection of the same stuff. But why would the old vaccine work against a new strain? Asking this question was dangerous–it might lead to vaccine hesitancy among the public. Once the monovalent booster was distributed, the question became irrelevant. CDC and ONS data proved that boosted individuals enjoyed restored protection.

But did they?

Here’s the thing. The graph above isn’t actually a plot of the Covid mRNA vaccine effectiveness over the first eight months of 2021. It’s actually the calculated effectiveness of a placebo against any disease, whether it be Covid, cancer or the common cold using the CDC’s method of classifying outcomes.

Professor Norman Fenton, co-author of “Where are the Numbers” explains the deception in detail here:

The illusion of vaccine efficacy revisited

Mathew Crawford, author of “Rounding The Earth” summarizes and further clarifies Fenton’s elegant explanation here:

Adding to Fenton's Miscategorization Illusion, Part 1

Both articles are vitally important reading.

A Little Background: What exactly is Vaccine Efficacy?

Vaccine efficacy is a term that is used as widely as it is misinterpreted. In the fall of 2020 the FDA told us that the Pfizer formulation was 95% effective at preventing symptomatic Covid-19 infection. Everyone knows that the higher the efficacy of something, the better it is at doing whatever it is supposed to do.

The mad rush to get jabbed at the end of 2020 hinted that the public had little understanding of what they were lining up for. 

I have spoken with dozens of health care professionals about the idea of efficacy. Some think that 95% efficacy meant that if you got vaccinated there was a 95% chance that you would never get Covid. Others thought that it meant that if you didn’t get vaccinated there would be a 95% chance that you would get Covid.

If folks in the medical field held these notions we can only guess what the lay public thought. Is it any wonder that anyone who eschewed the jab was generally considered foolish? Moreover, our health authorities claimed that getting vaccinated would prevent transmission even though this could not be verified from the Phase III trials. “Antivaxxers” were not only stupid, they were a threat to their own friends and families.

Efficacy is a function of relative risk. Because 162 out of about 20 thousand placebo recipients contracted symptomatic Covid in the brief period of observation in the Pfizer trial we calculate that the risk of getting Covid was 162/20,000 = 0.0081, or about 0.8%. 

The risk of getting Covid was 8/20,000 = 0.0004%, or 0.04%, if you got the jab because only 8 out of 20,000 experimental vaccine recipients got Covid in the trial. This is a substantial difference. The relative risk of getting Covid was 0.8%/0.04% = 20 times less if you were jabbed. 

Vaccine Efficacy is defined as:

VE = 1 - (Risk of getting Covid if vaxxed)/(Risk of getting Covid if unvaxxed)

Hence, Efficacy of the vaccine was 0.95 or 95%.

The point here is that even though the VE is impressive, the chances of contracting Covid was very low whether or not you were vaccinated in the trial.

The most important metric of a preventative measure is not its efficacy, it is the number needed to treat (NNT) to prevent a single case. This calculation is also straightforward. Vaccinating 20 thousand people prevented 162 - 8 = 154 cases of symptomatic Covid-19.

This means 20,000/154 = 130 people need to get vaccinated to prevent a case of Covid-19 during the period of observation (the average period of observation was six weeks). 

130 is the NNT for the Pfizer jab.

Is this a big number? It depends. It depends on the risk of getting the vaccine. The Pfizer trial demonstrated that six out of every thousand vaccine recipients suffered a Serious Adverse Event.

As long as the therapy carries a non-zero risk, it is the NNT that is relevant, not the VE. Nonetheless, our public health authorities touted the impressive VE, never the NNT. I practice at three different facilities. I have asked nearly every person I work with if they know what the NNT was for the vaccine they chose to get. Not one knew the answer. The reason is obvious. They read their newsfeeds, not the published trial results.

We should be able to readily see that the better a vaccine is, the higher its VE and the lower its NNT. Although the NNT of the Covid mRNA vaccines was never publicized, our government and agencies of public health were strongly implying what this number was.

“If you get these vaccines, you will be protected. If you are unvaccinated you will surely perish. This is now a pandemic of the unvaccinated.”

They were telling us that the NNT = 1.

Another wrinkle

Now that we understand what VE is and how it is calculated we can examine what Fenton is showing us.

In order to calculate the risk of getting Covid if vaccinated the Pfizer investigators chose to only count infections that occurred 14 days after getting the second jab. This was reasonable, they implied, because that is when the vaccine was most effective. Our FDA regulators did not protest, and the public readily accepted this approach. If it was okay with the FDA, why should anyone have a problem with it?

Although the published results in the NEJM did not mention how many people in the vaccine group got Covid prior to the 14 day “grace period”, Pfizer let the FDA know in this briefing document:

We see that folks in both groups got Covid from the time the first injection (vaccine or placebo) was administered. VE, in this case, is still significant (82%) but less than the 95% we were promised. Notice that even though the VE is lower (bad), the NNT is also smaller (good):

NNT = 20,000/ (275-50) = 89

I am bringing these numbers to your attention for three reasons:

1. According to the information that we know the FDA had at the time, the vaccine was shown to work better after two doses compared to one. Dose dependency is one of the Bradford Hill Criteria for determining causation.

2. The rate of infections in people between dose 1 and dose 2 was still less than in the unvaccinated (39 vs 82). According to this data, Pfizer and the FDA were not hiding any period of negative efficacy.

3. These two reasons allowed the FDA to justify their approach to assessing vaccine efficacy by only looking at infections two weeks after the second dose.

The 14 day window that was granted to the vaccine before tallying cases (i.e. vaccine failures) may seem like a minor concession that seems justifiable in lieu of a major public threat. The reality is that it would have large implications around the reported Vaccine Efficacy once we take a hard look at how the CDC started measuring how well the vaccine was actually performing in the country once it was authorized.

Vaccine effectiveness (the efficacy as measured once it has been deployed upon a population) is calculated periodically as both the vaccinated and the unvaccinated get sick. The problem is that vaccine failures (cases in the vaccinated) were only counted if a person was outside the 14 day window while cases in the unvaccinated population were counted immediately.

Simply put, as the vaccinated population grew from week to week, the cases of Covid-19 in that group were not being attributed to that group of people for another two weeks, yet the CDC was counting all people in receipt of the second jab as “vaccinated” no matter how recently they were inoculated.

The lag in case counts among the vaccinated artificially reduces the risk of infection in that group while the risk of infection if unvaccinated is not. VE, being the ratio of risk between the two groups, will necessarily be exaggerated.

This is why Fenton’s demonstration is so powerful. You don’t have to mine huge dumps of public data to see the problem. He demonstrates this effect by simply using a hypothetical vaccine that does not reduce the rate of infection and calculating VE as uptake of the vaccine increased across a hypothetical population. We would expect that VE should be zero if the intervention has no protective benefit. But the way the ONS adds things up, VE of this “placebo” will be 100% at the jump and only approach zero (the true VE) as the vaccine uptake diminishes.

Here’s a succinct explanation in his own words:

If the CDC methodology was devious, the one used by the ONS (Office of National Statistics) in the UK was outrageous.

Fenton cites the deceptive method ONS uses to create the illusion of efficacy. Basically, Covid-19 outcomes that occur within 21 days of the first dose of the primary series were not only excluded from the pool of vaccinated individuals, they were attributed to the unvaccinated population!

The CDC isn’t being as audacious. They provide us with their definitions:

“Vaccinated case with a primary series: SARS-CoV-2 RNA or antigen detected in a respiratory specimen collected ≥14 days after verifiably completing the primary series of an FDA-authorized or approved COVID-19 vaccine.”

“Unvaccinated case: SARS-CoV-2 RNA or antigen detected in a respiratory specimen from a person who has not been verified to have received any COVID-19 vaccine doses before the specimen collection date.”

Cases in the vaccinated are only counted if they occur 14 days after the second dose. What is the CDC doing with cases that happen in people who are in the process of getting “protected”? They create a new category, the “partially vaccinated”:

“Partially vaccinated case: SARS-CoV-2 RNA or antigen detected in a respiratory specimen collected from a person who received at least one FDA-authorized or approved vaccine dose but did not complete a primary series ≥14 days before collection of a respiratory specimen with SARS-CoV-2 RNA or antigen detected.”

Okay, that seems fair. They are dividing up people by their stage of vaccination and aren’t burdening the unvaccinated with infections that arise in people who are partially vaccinated like the ONS is doing. So where’s the problem?

To reiterate, it has to do with how they count the number of people vaccinated:

Above is a screen shot from the CDC website which states:

“Weekly age-specific incidence rates by vaccination status were calculated as the number of cases or deaths divided by the number of people vaccinated with a primary with a primary series, overall or with/without a booster dose…”

(Authors Note/Addendum: The citation above no longer includes the definition of Incidence rate estimates as of September 13, 2023. A CDC response to my request for clarification is pending.)

Read that carefully. Notice how they count the number of vaccinated people: “...the number of people vaccinated with a primary series, overall…”

There is no mention of people who are within or outside 14 days of their second dose. This means that for any given time period, the number of cases thrown into the vaccinated column will be in those who are 14 days out from the second dose, but the calculation of incidence rate will include those who are less than two weeks from their second dose.

The bottom line is this: the more rapid the uptake of the vaccine in a population, the greater the illusory effect will be because the “partially vaccinated” will drive the infection rate among the vaccinated lower.

So how fast did the US population get vaccinated? I estimated this from Our World In Data:

This plot shows us the number of Covid-19 vaccines that were administered from the time of authorization in December 2020 through August 2021 when boosters were authorized. Using this profile and a rough estimate that at the end of August 2021 60% of the public was fully vaccinated with the primary series, I generated the plot given at the beginning of this article. This is what the bi-weekly reported VE would be for vaccine that has zero protective benefit:

The point isn’t that the efficacy goes to zero. The efficacy of a placebo is zero to begin with. The point is that it took 9 months to get there.

For those of you who enjoy trudging through numbers it becomes obvious that the quicker the uptake of the therapy (be it a vaccine or placebo), the greater the initial exaggeration of efficacy. As uptake approaches zero the true efficacy of the intervention emerges.

Notice that the VE of the placebo is effectively 0 at the end of July 2021, right at the time that the FDA authorized the use of the monovalent booster. Once the public began accepting the booster as their duty to protect the population, comparisons between the fully vaccinated and boosted began to be reported.

How did the CDC calculate VE in the boosted? Take a wild guess…

“Vaccinated case with a monovalent booster: SARS-CoV-2 RNA or antigen detected in a respiratory specimen collected in a person verified to have received a primary series of an FDA-authorized/approved vaccine and ≥14 days after receipt of at least one additional dose of any monovalent FDA-authorized/approved COVID-19 vaccine on or after August 13, 2021.”

Once again, they slip in a 14 day grace period and once again proof of booster effectiveness could be conjured from thin air. Randomized, placebo control trials of monovalent boosters were never conducted. Why do them if they seemed to work from the jump?

Conclusion

To be clear, this is not offered as proof that the primary series or boosters were no more effective than a placebo. All we can conclude is that their reported effectiveness has been exaggerated and that this exaggeration will at least partially account for their immediate effect, especially if the public gets in line quickly.

During my recent encounter with the movers and shakers in the vaccine industry and regulatory agencies I was struck by how “concerned” they were that uptake of the bivalent booster was so poor even though they acknowledged that proving effectiveness would be very difficult in a population that has either been vaccinated or exposed to the virus already. If the signal was so small, I asked, why shove it down our throats? Perhaps this is the reason. The slower the uptake, the harder it is to prove that they are helping.

Pfizer mRNA Covid-19 vaccine trials were gamed. The investigators were unblinded. We don’t need to believe whistleblower Brook Jackson to prove this. The evidence is mathematical, and it is damning. It’s only a matter of looking hard enough to find it–a job for the FDA, a job that they chose not to do.

Here’s a great summary of the details by Josh Guetzgow, PhD:

Comirnaty or Comirnaughty?

If the responsible parties at Pfizer are ever cross examined in a courtroom with a judge, a jury and expert witnesses that are not on their payroll they still have a way out. The evidence appears in a memorandum to the FDA’s advisory committee. Pfizer let the regulators know, and the regulators looked in the other direction.

This doesn’t seem to register with some because they insist that these “vaccines” were shown to work, at least for a few months before their effectiveness “waned”. Why nitpick if lives were saved? But were they?

Observational data used to demonstrate vaccine efficacy is fraught with confounders. Fenton et al assiduously dissected data from the Office of National Statistics over a year ago showing that even official datasets have been corrupted. They proved that the ONS was misclassifying deaths by vaccine status. Even after requests for correction, the UK data remains too unreliable to draw any conclusions about whether these things actually do more good than harm.