Vaccintrials safety illusion

Looking back on the introduction of the proven "Safe and Effective" MRNA vaccines, the "effective" aspect has disappointed: vaccination figures under COVID oversen are not reassuring and are at most unintentionally released. 80% of the deaths to COVID-19 during the Deltagolf was vaccinated¹See steig.nl: The big secret of the Deltagolf. More than 90% of the participants in the Long-Covid government project have been vaccinated²See What disease is the post-Covid report?, also something that should not be possible after a trial that left no doubt about effectiveness. But it doesn't sound that you say: "Safe" because if it is not effective against Covid, what does it do? After all, it is not a saltwater solution and the increased death of all causes and many reports of side effects also arouse no confidence. That really shouldn't be the case after a trial that the signal "safe" had delivered. Pfizer apparently has a different definition of "safe".

The basis of "safe" is becoming smaller

In the public discussion, it is assumed that the Pfizer trial³The Pfizer report: Summary Basis for Regulatory Action – COMIRNATY With over 43,000 participants provided sufficient evidence for the designation safe & effective.

Everyone who has viewed critical vaccinations studies has probably thought: "If a sample of 43,000 participants offers sufficient safety to inject the entire world population, why have mega studies done in the past with hundreds of thousands or even millions of test subjects?" The statistical knowledge that such an research design required existed in the 1950s and 1960s. Yet those investigations were not a little bigger: they were 40 times as large! And for some time we can finish it with a few tens of thousands. Why?

The larger, the more precise. The smaller, the more statistical uncertainty, which increases the possible insecurity.

Although 43,448 people were enrolled, the primary safety analysis on which the “safe” claim is based involved fewer participants because not everyone completed a full follow-up. The safety analysis involved 37,706 participants, which again makes you wonder how those 5,700 dropouts were selected. In any case, that results in an upper limit of ±10,000 additional deaths per 12 million vaccinated within 2 months. More precise is not possible. But a median follow-up duration of two months is also reported and this means that an unknown part of the group was followed for less than two months⁴“Safety was evaluated in 37,706 participants who received at least one dose of vaccine or placebo; 18,860 received BNT162b2 and 18,846 received placebo. At the time of the data cutoff, a median of two months of safety data was available for 18,860 vaccine recipients and 18,846 placebo recipients.”
— Polac Fp, et al. No, 2020. This results in a higher upper limit that we cannot calculate precisely because we do not know the size of the final sample. This will therefore be smaller, and the smaller the sample, the greater the uncertainty. As a result, the statistical upper limit continues to rise and “safe” becomes less and less safe. I will indicate the size as <38,000.

The cut-off date (November 14, 2020) chosen by Pfizer determined how many participants had completed the two months. That therefore depends on the registration date of each participant and accurate registration thereof. Because individual data is not public, the data integrity is only guaranteed by the manufacturer, which must be sharp on possible "sloppiness." That it is difficult to report imperfections is clear from how reports were made from an employee to the Pfizer-Trial, Brook Jackson⁵ Interviews met Brook Jackson, die later in The BMJ (2021) the disorderly data collection and defective monitoring described)⁶Brook Jackson met video in The BMJ: https://www.bmj.com/content/375/bmj.n2635.

In the phase 3 security study of Pfizer (Sample <38,000), 21 people died in the Vaccing Group and 17 in the placebo group. 23% more death in the vaccinated group. Pfizer's conclusion was: no statistically significant difference in death, so the vaccine is "safe". For the argumentation theorists among us it is immediately as ready as a lump: absence of evidence is not proof of absence. Not only a statistical rule but a classic fallacy: the non sequuitur⁷Uitleg does not follow by chatgpt. But pharmaceutical companies don't think that way, so let's look in their own history to see whether that is a new insight. The study design (size) at Pfizer contrasts sharply with previous vaccine studies (move your mouse over the graph to zoom in):

If Pfizer can manage with approximately 44,000 participants (and < 38,000 in terms of safety), why were such numbers earlier than necessary? This could be related to a change in the meaning of the concept of 'safety'.

What do we think "safe" stands for?

If you help someone with a visual impairment with "it is safe, you can cross" then you don't say that if you think there is a small chance that he will be run over. You only say that when you are sure that he will safely reach the other side. As a citizen you are blindly sailing on pharmaceuticals. But they have other standards.

In clinical practice, “safe” usually means: there is no significant difference from placebo in this study population, in the limited period of the trial. So that's the fallacy. In everyday speech, “safe” means something completely different, namely: no real risk of death or serious damage. But again it is unclear what “none real risk ”. That is not"no Risk ”. In which numbers should we think? Let's assume that a risk of 1 extra death per 10,000 punctures would still be considered acceptable (we will come back on risk categories later). It is the second safest category according to the official EMA/WHO classification. (Often:> 1/10, sometimes: 1/1,000-1/1,000, rarely: rarely: rarely: rarely: rarely: rarely: rarely: rarely: rarely: reld the ownership: 1,000, reld the ownership: reld the ownership: reld the ownership: reld the ownership: reld the ownership: reld the owner: reld the ownership: rarely: rarely: 1,000: rarely: 1,000: 1,000 and 1,000! 1/10,000-1/100,000.)

How big should a trial be?

We assume a hypothetical maximum risk: the additional chance of dying within two months (the duration of the trial!) after vaccination may not exceed 0.01% (1 in 10,000). For a population of 12 million vaccinated people (NL), this means a maximum of 1,200 deaths in two months - which you can also question, but we agreed to accept this as being 'safe'. It should also be weighed against the severity of the disease. What is also missing is the age aspect. All ages are tarred with the same brush, which raises ethical questions and can lead to statistical distortions.

To rule out that the risk is greater than 0.01%, more than 1 million subjects are needed. (see heading Fieldwork, below). If you want to achieve a likelihood ratio of 100%, it will be many, many millions. So that no longer makes sense as a 'trial', which is "rolling out and seeing what happens".
However, Pfizer worked with < 38,000 test subjects. This means that you cannot statistically notice a small difference such as 1:10,000. In fact, with such small numbers you can only rule out that no more than approximately 10,000 to 12,000 deaths would occur in the first two months among 12 million vaccinated people.

This means that the Pfizer safety trial indicates that the total mortality in the population can remain the same by half Because of the vaccinations, because the figures come down to that. Everything in between is therefore declared acceptable (because "safe").

Moreover, it is important that this calculation only applies to the duration of the trial: two months. We now know that side effects and deaths sometimes only become manifesto after 3–4 months (± 100 days) and possibly later. But also the suspected height of that risk of death would not have been observed in this underpowered study⁸Observational estimates appear to indicate a risk of 0.02% for the baseline series. That would be 2400 deaths for two months. That falls well within the insensitivity range of this design..

The ethical dimension

Medications are given to sick people. A small extra risk can be acceptable there, because the patient runs a high risk without treatment.
Vaccines, on the other hand, are given to healthy people, to children. That makes the balance fundamentally different. Can you let a healthy (younger) individual risk of death to prevent deaths elsewhere (in the elderly)? That is not a purely scientific, but above all an ethical and possibly culture -related question. And that is why perhaps a political one for a referendum.

Age dependence

Like many other disorders, COVID mortality is strong age-related. The percentage aboutDeath per age group is similar to all age groups. The trials did not have the power to prove safety and effectiveness per age group separately. Results for an "average test subject" were simply generalized to all ages. That might be justified if the goal had been group immunity. Furthermore, it is not necessary that the entire argument of group immunity did not go up for a means that did not reduce infections and was administered at the wrong time (what a panic football anyway, if you put it back in no time)

Semantic and framing

The words themselves have suddenly evolved:

• "Vaccine" stood for an innocent first encounter with a pathogen, so that the body builds up immunity without any risk. With mRNA and vector vaccines, it has become a medicinal intervention because immunity is so temporary that the drug should preferably be administered a few weeks before infection - an impossible timing, as the 'boosters' have proven. The distribution through the body is also very different. Traditionally, the further away from the injection site, the weaker. mRNA vaccines are micro-submarines that can travel anywhere in the body and only release their active cargo upon arrival in a cell.
• “Trial” Once stood for long -term, large -scale field studies (Polio, BCG, Hepatitis E, Cholera). The Covid-19 trials were relatively small and short. Looking at the required power and the investigated investigations, they were mainly for the stage, shortcuts to the approval. Netzelling data were conceived as safe. The absence of evidence for extra mortality - the predictable outcome with this little samples - was framed as proof of the opposite. That invites you to view other vaccinoth studies because, as we can see at the top 10 graph, almost all studies are number-up under the Pfizer trial discussed here.

Conclusion

• Pfizer’s Trial did not proved that the vaccine safe Was in the everyday meaning of that word. It only proved that in a limited time there was no significant difference in death with placebo. But absence of evidence is not proof of absence.
• It was known in advance that this trial would not even remotely demonstrate that the safety of the vaccine would be at stake. The popularity of this range of sample sizes is understandable, from a pharmaceutical perspective. A lot has to happen if this is to produce a hard (significant) safety signal.
• To demonstrate real safety, much more extensive trials are needed - and that is sometimes the case. The claim "Safe and Effective" has become the result of framing and repetition, not of hard science.
• The figures are a hard limit: less than 12,000 in two months after the injection, the trial is dimensioned on that. That is the statement "safe" too. Less than 12,000 deaths in two months is "safe". Fortunately, there are probably no more than 2,000 per 2 months.

References

• 1
  See steig.nl: The big secret of the Deltagolf
• 2
  See What disease is the post-Covid report?
• 3
  The Pfizer report: Summary Basis for Regulatory Action – COMIRNATY
• 4
  “Safety was evaluated in 37,706 participants who received at least one dose of vaccine or placebo; 18,860 received BNT162b2 and 18,846 received placebo. At the time of the data cutoff, a median of two months of safety data was available for 18,860 vaccine recipients and 18,846 placebo recipients.”
  — Polac Fp, et al. No, 2020
• 5
  Interviews met Brook Jackson
• 6
  Brook Jackson met video in The BMJ: https://www.bmj.com/content/375/bmj.n2635
• 7
  Uitleg does not follow by chatgpt
• 8
  Observational estimates appear to indicate a risk of 0.02% for the baseline series. That would be 2400 deaths for two months. That falls well within the insensitivity range of this design.