The day before yesterday, Maurice de Hond published a review of a CDC publication stating that vaccination offers better protection against Covid-19 than going through the disease. Because the NRC also writes about it today, I am posting that article here again for anyone who missed it and may see the NRC piece.
Intussen schuift de bodem onder onze voeten vandaan, zo hard gaat het toch wel met de afnemende effectiviteit van de vaccins. Omdat er op één paard is gewed als "enige uitweg", staan we in februari zo ongeveer met lege handen. Boostershots zullen tijdelijk enig soelaas kunnen bieden onder ouderen. Ziekenhuizen melden nu al een sterke toename van gevaccineerden. De rel rond Maastricht UMC ("Vorige week waren gemiddeld acht op de tien patiënten op een corona-afdeling (niet IC) gevaccineerd.") could not be satisfactorily debunked, no matter how hard the NRC tried.
In Flemish media it may even be written that it is in 70% of hospital admissions vaccinated About. This has also been announced in parliament. There is every reason to believe that the situation in the Netherlands is not much different and that it will otherwise go in that direction in the foreseeable future. Can't we already bring in that ship from Brautigam, including staff?
A study from Qatar, published the day before yesterday in Nature, liegt er ook niet om: daar hebben ze de "real word effectivity" kunnen observeren. Pfizer was na een half jaar nog 51,9% effectief, in UK en Israël was dat tussen 39% en 66% en dat was begin september; we zijn inmiddels maanden verder.
A Swedish study (pre-print, i.e. not yet peer-reviewed) shows that vaccine effectiveness against hospitalization drops below 50% after six months. That is the soil percentage that a vaccine must meet according to FDA guidelines. So hurry up with those boosters to boost that percentage again, otherwise there will be legal problems.
Vaccinated people are increasingly likely to become infected, presumably also because they think they are safe.
It is common ground that natural immunity is not subject to such a lapse, in terms of months. In any case, it wears off with age, like many of our other functions.
In short: Vaccines protect very well for the first 90 days (that's how long Pfizer's trials lasted). Then it collapses. Research that is partly based on the situation just after the vaccination campaigns does not give a realistic picture if the time aspect is not included. Now that is not even the main objection to the study discussed below.
Below is the republication of our co-production, published on November 2 on maurice.nl.
Data manipulation around the value of natural immunity
Article Written byMaurice de Hond in collaboration with Anton Theunissen ofvirusvaria.nl
What gives better protection?
The immune system recognizes not just one part of a recurrent virus but the entire pattern, despite small changes. Going through an infection therefore offers more complete protection than a vaccine.
There are many and also large-scale studies available on how this works with Covid-19. Inthis large-scale Israeli investigationone achieves 5 times better protection through natural immunity than through vaccination.At the bottom of this American article you will find a list of 96 articles on that subject.Natural immunity is rated better than or at least equal to vaccination.
A few days ago, the American CDC (an agency like the RIVM) came up with a study that seems to turn everything that is known about this subject upside down. On twitter, many prominent people reacted extremely positively, such as these two:
If articles appear with conclusions that are at odds with what has been assumed so far, it is important to first look at what kind of study it was and what the basic figures are that lead to that conclusion. So this striking claim certainly qualifies for that.
What does the study really say?
This is that study from CDC.Reading it once is not enough to understand the design and conduct of the study, it seems so unlikely.
It concerns an analysis of patients in hospitals across nine US states during the months of January-September 2021. In total, it concerned 7300 patients, who met a series of admission requirements of the study. They were divided into two different groups:
- 1020 who had tested positive for Covid-19 in the period between 90 and 180 days before hospitalization and were in hospital with symptoms that could indicate Covid-19.
- 6328 who had been double-vaccinated in the period between 90 and 180 days before hospitalizations, who were not known to have been infected before and were hospitalized with symptoms that could indicate Covid-19.
So just to be clear: hospital patients with Covid-like symptoms were divided into two groups: one group had had Covid about 4 months ago, the other group was vaccinated then.
All patients were PCR tested during their hospitalization. 9% of the unvaccinated tested positive and 5% of the vaccinated. And that would then lead to the conclusion that the second group had a 5 times smaller chance ofXXXXThen the second group.
WhereXXXXis not a typo. Because if you see the CDC summary in the tweet and the cries of triumph from those who retweet the results, you would think that the study has shown that those who have been vaccinated are 5 times less likely to get infected than those who have not been vaccinated, but have had an infection. But read along carefully:
It really says:
"patients in hospitals with symptoms THAT RESEMBLE COVID-19, who have not been vaccinated but have had an infection, are 5 times more likely to get a positive Covid-19 test than those who have been vaccinated."
How this factor of 5 was arrived at, from the established figures, is very unclear. There is also great criticism of this in several tweets from experts.(See this article, among others).
It is certain that more than 90% of those patients, as we can also see from the original figures, were not in hospital with or because of Covid-19. And really the only question the study has answered is: "how likely are vaccinated people who have been admitted with Covid-19-like symptoms to really have Covid-19 and what is that like in previously infected people". A completely irrelevant question, because you will first have to focus on the reasons why those people were hospitalized and how that differs between the two groups. These results say nothing, really nothing, about the degree of protection against Covid-19 hospitalizations from the vaccination or having had the infection.
Because these are the numbers: "Of the 413 hospital patients with a positive PCR test, 324 were fully vaccinated and 89 had previously had an infection."
And that can only say something if you know how many people there are in total in the population who were double vaccinated and how many people who had previously had an infection. Only then can it be calculated whether with a past infection that results in more or fewer hospital admissions than after being double vaccinated. Suppose this had been about the same in the entire population, then these numbers 324 and 89 indicate that the chance of getting Covid-19 is four times higher in vaccinated people than among those who had previously had an infection.
It is a classic error, which can be described by the following example:
"Suppose it is announced by insurance companies that red cars are twice as often involved in collisions than blue cars. Without knowing how many red and blue cars there are, you can't say anything at all. Because if there are four times as many red as blue cars on the road, then relatively speaking, red cars are less involved in collisions.
If you really want to know what the effect of the blue or red color is on collisions, you will also have to make sure in your research design that the drivers of the cars are made comparable. Because it could be that red cars have younger drivers on average and that blue cars drive more kilometers per year because they are used for business purposes. And that also affects the chance of causing a collision. And as long as you don't have this, you can't determine the effect of the color of the car on the chance of a collision in this way."
Now in that American study a certain calculation has taken place to make an estimate of those national figures in a roundabout way, but that is a very error-prone approach because quite a few questionable assumptions have to be made.which is also described here.
How nonsensical this study approach is can be explained more easily in the following way:
- Apparently, there was a reason that these patients were admitted to the hospital. In more than 90% of the cases, this was not related to Covid-19. And the cause of the admission can differ considerably between these two groups. And that can have major consequences for the final percentage of positive PCR tests.
- The group of previously infected people was, on average, much younger than the group of vaccinated people. So the chance that the second group would have ended up in hospital because of something that had not been related to Covid-19 is significantly greater than in that younger group. And that in turn will have an effect on the proportion with a positive PCR test in both groups.
- Those who have had an infection before may have had a positive PCR test, even after a longer period of time. So that may have boosted the figure.
- Among those who were double vaccinated will also have been people who have had the infection before, without knowing it.
- Those who had previously experienced Covid-19 may still have symptoms that had been the result of Covid-19 and who therefore ended up in the hospital, without being infected again.
What was also striking and what was not addressed: the report showed that in the period June-September, 5213 patients were in hospital with Covid-19-like symptoms who had been double vaccinated and only 189 who had been previously infected. In total, 325 of these patients in the hospital turned out to have a positive PCR test (6%), 94% did not. Of those 325, 19 had been previously infected and 306 were double vaccinated.
The vaccination rate in the US is about 60%. So there are one and a half times as many vaccinated people as unvaccinated people (40%). If they wanted to show that the vaccinations help, you would expect that factor vaccinated people to be below 1.5 in Covid patients. But that is not the approach: they want a comparison between immunized people who have not had a vaccination and immunized people who have not had an infection. The fact that one group in society is/was much larger than the other and also differs in terms of age, origin, social class, etc. is not included in the results.
To whatever extent each of these factors played a role in this study, it is a study design on the basis of which you cannot draw conclusions about the protection of natural immunity versus vaccination. But to be honest: if you read carefully what CDC wrote there with their statement of 5 times as much, it was also not about the greater or smaller chance of ending up in hospital after infection or vaccination. But it was only about the -apparently small- chance among those two groups – if you had ended up in the hospital with symptoms that resembled Covid-19 – that you really had Covid-19.
It is significant how enthusiastically many people respond to that study as if it were stated that vaccination protects better than natural immunity. Whereby apparently one does not wonder how it is possible that this study comes to such different conclusions than any other study on that subject so far.
But then again, that is not unique to the US. Data is either completely omitted or interpreted in such a way that it always supports the conclusion that was already envisioned in advance.
Recommended reading:
And, also about vaccinations and with references to peer-reviewed studies:
- Article by Guido Versteegon virusvaria.nl
- Article by Martijn de Jonge on virusvaria.nl
