Knack.be posted an interesting fact check1Fact check op Knack.be in response to an article by Herman Steigstra2Article We have it!. He contradicts claims from various studies that COVID vaccinations would significantly reduce mortality and would therefore have saved countless lives. The article was also picked up by ninefornews and nieuwrechts.nl
The central claim under attack is:
COVID vaccinations have (in various periods) reduced mortality by up to 44% and therefore saved many lives.
According to Herman Steigstra, the observed mortality rates cannot be reconciled with this.
Korte samenvatting van het artikel "We hebben hem!"
Through observational data (CBS and RIVM figures + timing), a striking connection with vaccination is substantiated, which undermines the dominant narrative of “life-saving vaccines”.
After vaccination campaigns, a consistent pattern of excess mortality is visible, which is now strong enough to make a concrete estimate:
- About 1 in 1500 vaccinations would lead to an additional death (≈ 70 per 100,000).
- This pattern would repeat itself with multiple vaccination waves and in different countries.
- Excess mortality peaks in the months after vaccination, while this is precisely where the (alleged) effectiveness should be measurable.
- At the same time it is emphasized that:
- Only part of the excess mortality was explained by COVID itself
- There is no clear evidence in the data that vaccines reduce mortality (correlation is not causation)
The conclusion that can be drawn from the article:
Previous studies claiming a mortality reduction due to Covid vaccinations do not hold up in the light of perceived reality. There is even one strong indication (no definitive evidence) that vaccinations themselves contribute to excess mortality.
Herman rekent kortweg door wat de cijfermatige consequentie zou moeten zijn van die claims - en dat komt niet uit, eerder het tegendeel. Dat mag een factcheck are called: a statement is tested against observed data.
The problem is not difficult to understand: if vaccinated people (e.g. 80% of the population) die 40% less from all causes, then this should be clearly visible in the mortality figures. However, only excess mortality has been observed.
(Previous articles also showed that if the remaining excess mortality is attributed exclusively to unvaccinated people, this also means that unvaccinated people die more often in fatal car accidents. And that always after others have been vaccinated. Ed.)
From drug check
Hoe factcheckt knack.be vervolgens die doorrekening? Het begint met een 'guilt by association' citaat: "smeerlappen met bloed aan hun handen", alleen is dat uit een willekeurige Facebook-post geplukt - ze moeten toch wat te framen hebben. Maar het staat toch maar gerefereerd en geciteerd in de factcheck van deze statistische analyse.
Na de inleiding volgen er enkele belletjes met deskundigen die bevestigen dat de aangevallen claim inderdaad in die studies staat of onnavolgbare afketsers naar voren brengen als "je moet naar de doodsoorzaken kijken" waarbij de factchecker zich niet afvraagt waarom dat zou moeten. Dat hoeft namelijk helemaal niet, dat hangt van je conclusies af. Ook vindt de factchecker zelf studies waarin de claim wordt gedaan.
Herman zelf wordt door een expert neergezet als iemand zonder relevante statistische of epidemiologische scholing. De doorrekening van Herman wordt daarom uiteindelijk beoordeeld als "onwaar".
Rik Scholten schreef een brief naar Knack.be waarin hij de kritiek van ondeskungheid betwist. Hij somt Herman's achtergrond op:
- MSc Physics (Utrecht University)
- Long career in medical statistics and data analysis
- Former head of statistics at an organization for quality assurance in medical laboratories
Based on this, he asks whether it is correct to portray him as someone without relevant statistical or epidemiological training.
The answer from XXXXXX XXXXXXXX, the fact-checker in question, states:
From: XXXXXX XXXXXXXX
Date: wo 8 apr 2026, 11:38
Subject: Re: Artikel in Knack
To: Rik Scholten
Dear Mr. Scholten,
Thank you for your comment. I don't mean to say that Mr. Steigstra has no merits. I also mention him in my article as a Dutch statistician. However, we do not mention his name because anonymization falls within our fact-checking methodology. As a journalist, I have no opinion about his abilities and the statement you put in my mouth is only a quote from the biostatistician Geert Molenberghs.
In addition, this article not only quotes experts, but also a study by researchers from the University of Utrecht who worked in a manner analogous to Mr. Steigstra, but had more granular data available. They were therefore also able to effectively look at the mortality of all people in the Netherlands who had received a vaccination and thus observed that 44% fewer deaths occurred in that population in the first three weeks after vaccination. https://link.springer.com/content/pdf/10.1007/s10654-025-01334-6.pdf.
Since this is reproducible and peer-reviewed research with extensive data analysis, we come to the conclusion that the method used by Mr. Steigstra has some holes and therefore that the figure that mortality due to vaccination is 1 in 1500 is incorrect. Good scientists also make mistakes, so I don't want to make any value judgments about that either. Hence the explanation about confounding, as that is the most likely cause.
Sincerely,
XXXXXX XXXXXXXX
factchecker Knack
Het is opmerkelijk dat deze factchecker spreekt van 'only a quote from the biostatistician Geert Molenberghs' terwijl de rest van deze 'factcheck' grotendeels wordt onderbouwd met citaten.
What is more striking is that the quotes cite the studies that were disputed because they do not correspond to the observed figures. That is the repetition of the disputed position, but then used as an argument: a well-known fallacy called petitio principii3One circular reasoning (begging the question) is also called a circular argument or petitio principii. This is where someone gives their point of view as an argument, essentially saying the same thing twice.. The results of their own desk research (such as the Stanford study and that of Dutch researchers) also repeat the challenged conclusions.
Nergens wordt de transparante analyse van Herman ook maar geanalyseerd of nagerekend, laat staan dat er een methodologische of rekenfout wordt aangewezen. "Het zal wel confounding zijn", verder komt men niet of de professoren wensen liever niet teveel tijd te besteden aan een werkstudent die een stukje moet schrijven.
Herman's conclusies worden afgeserveerd zonder kritiek - en dat die conclusies afweken van bekende studies, dat wisten we al. Dat was namelijk de aanleiding van het schrijven van het artikel.
Kortom: reproduceerbaar en peer-reviewed onderzoek wordt door Knack als onbetwistbaar beschouwd. Daar kan geen logica of rekenwerk aan tornen...
What exactly is a fact checker?
The knack.be website contains a short bio of the fact checker.
"XXXXXX XXXXXXXX
Journalist and fact checker
XXXXXX XXXXXXXX(°1999) is zelfstandig journalist. Sinds november 2025 werkt ze als journalist en vaste factchecker voor Knack in het kader van onze samenwerking met Meta. Eerder was ze bij De Standaard aan de slag. Ze rondt momenteel ook de Master Handelsingenieur in de Beleidsinformatica af aan de Universiteit Antwerpen."4Website knack.be
A journalist/permanent fact checker “in the context of the collaboration with Meta” means in concrete terms that the editorial staff of Knack collaborates directly with Meta's “Third Party Fact‑Checking Program”. Knack is also listed there5Knack at Thrid Party Fact-Checking Program. I just asked AI. This usually means:
- Access to Meta's internal review tool, in which fact-checkers see flagged posts from users;
- Financial compensation by Meta for fact checks carried out (often per checked item);
- Feedbackmechanismen where Meta automatically lowers the visibility of certain posts or adds labels (“partially false”, “context missing”, etc.) based on Knack's rating.
Like one Knack‑factchecker works directly under that program, this implies:
- Institutional embedding of Meta-policy in Belgian journalism. Meta determines technologically what is seen and what not. The “independent” editors do assess content, but the consequences (visibility, distribution) lie with Meta.
- Opaque financial and contractual relationships. The public rarely knows how much media gets paid and under what conditions. The impression of independence is retained, but journalistic responsibility partly shifts to private Big Tech interests.
- Perception of neutrality is being eroded. When one party is both gatekeeper and financier, the dividing line between reporting, commercial interests and behavioral management disappears.
That someone like XXXXXX XXXXXXXX is explicitly presented as “permanent fact checker for Knack in the context of our collaboration with Meta” shows that this structure has been institutionalized; a structural integration of legacy media and private content control. Big Tech controls the distribution channels, national media provides the legitimacy. In practice, this is a private-public censorship infrastructure with a journalistic face.
Covid is too big to fail.
Footnotes
- 1Fact check op Knack.be
- 2Article We have it!
- 3One circular reasoning (begging the question) is also called a circular argument or petitio principii. This is where someone gives their point of view as an argument, essentially saying the same thing twice.
- 4Website knack.be
- 5Knack at Thrid Party Fact-Checking Program
Knack, with money from USAID, was fortunately dried up by Trump...
I left Herman's article as it was.
You call it facts.
But the excess mortality figures are not facts, but interpretations of mortality figures based on suspected mortality. You call that standard mortality.
And that involves a lot of assumptions. Slight changes in those assumptions result in different 'overstepping' and therefore also different 'puncture damage'. And that increases exponentially, or turns into no damage at all.
I think that is a number that you can compare with thumb sucking.
Anyway, our fact checkers don't understand that. And you indeed get worthless articles that cannot be called a fact check but consist of fallacies.
I don't see the word 'facts' above...?
I can't imagine calling it that. It is a calculation based on easily explained principles. I would like to hear a more plausible calculation. Of course, there are countless nuances that can be added, all of which yield something different, especially now that the intervening period is becoming longer and longer.
If only they would go into something substantive, such as the calculation method (which approximates all CBS forecasts well), then we might get somewhere. It may indeed be better. Gladly even. But all you hear is “no, that's not right”.
There are many alternatives. We are only interested in a tool that is in line with how the government has always done things. Or how the actuaries did it before they added the 'excess mortality terms'.
“With which all CBS forecasts are closely approximated.”
I have already explained that the mortality forecast from CBS had a huge artefact among the 80+ and women. It is not entirely coincidental that women are overrepresented in the 80+ age group.
By shying away from approaching CBS, you intrinsically admit that you have made the same basic mistake.
And I have already shown this to you several times. I don't understand why you include this as a 'plus' for your calculation.
Sometimes I wonder whether your desire to prove that vaccination causes collateral damage blinds you to such mistakes. It seems that the end justifies the means. And don't get me wrong, I'm not saying you do this consciously. I understand that you are driven.
But that CBS expectation was really very bad. That helped them in the beginning with the exaggerated corona 'excess' mortality, but eventually it caught on and they abandoned it.
Bonne, you say this. Slight changes in those assumptions result in different 'overstepping' and therefore also different 'puncture damage'. And that increases exponentially, or turns into no damage at all.
I think that is a number that you can compare with thumb sucking.
This is a bit of a strange reaction, isn't it? I know you don't always agree with Herman and that's fine of course. But there is https://sterftemonitor.nl/ made a baseline as objective as possible, something that they have not done at the RIVM, then you have to see that we are dealing with a very significant excess mortality of 50,000 people after 2021. The RIVM baseline tries to wipe this away, you make your own calculations, but you have to do your very best to wipe out those 50,000 people, don't you? I find this a really incomprehensible response from you.
Well, quite honestly? I really don't think the green and black line is an objective base line.
https://i.postimg.cc/9X608Cd9/Exponentieel-afnemende-sterfte.png
My red line makes a lot more sense….
That's what you get when you use models. They are always open to criticism.
Therefore, it is better to just take the raw mortality/100,000/year group. See the article and my response there. You can read enough misery from that in 2021 - 2024. And that is simply objective, not contestable. Facts!
The excess mortality from 2024 will certainly not disappear in all groups. Let's focus on that. But it is much less dramatic than suggested by Herman (and Anton) with their “aggressively declining standard mortality” (the green and black lines).
Aren't you under the wrong post? The green and black line do not appear in this post, that is stated in this.
I answered that too.
It seems that what you find does not correspond to arithmetic results.
And 'logical' in this case is a subjective concept. I don't think it makes any sense at all that a baseline of absolute mortality remains horizontal or goes up, while the group size is clearly shrinking.
You actually expect that baseline to go down.
You could also do it with just mortality probabilities, which might be a bit easier for you to understand. Then you have to divide the observed deaths of recent years by the population sizes of the relevant years. This will not change the differences between expected and observed.
That's right, that detailed discussion belongs in that other article.
I think Bonne has a point here. That's why I refer to that other article here.
I am not writing/suggesting anywhere that mortality should increase over time.
I just find it illogical that mortality has continued to decline linearly since 2010.
And I think it is conservatively logical that the decline in mortality is slowing down. And I wanted to express that with that red line (in the wrong graph, because absolute numbers). You can only meaningfully draw that red line like this in a mortality/100,000 graph with small cohorts.
And then a mortality that is below that of 2018/2019 can in principle no longer be characterized as excess mortality. I think that makes a lot of sense. Even Maarten Keulemans cannot deny that. And the point is that there will be a huge excess mortality from 2021 to 2023. In your graphs last year, it was still there for a number of groups in 2024. And I am very curious about what image the 2025 figures create in those pictures….
If someone makes graphs of 5 previous cohort per 100k and a baseline until 2019, I would like to see that. We once did that with English data, but after 2023 they will somehow no longer publish those reports. And they are not the only ones, strangely... I am curious about other graph methodology or baseline but brushing away the excess mortality, especially among young people. (Where there is relatively the most excess mortality) seems impossible to me to eliminate.
Give me your email address and I will give you my dashboard based on 5 year cohorts. And as an extra you get a division into winter and summer comparison. Because annual mortality with divorce on January 1 is also distorted.
Do you also have them every 1 year, including the population size?
No, I personally think that cohorts that are too small are too dependent on standard fluctuations. Especially among people under 50, one year with high or low mortality at the end or beginning of your chosen trend can generate considerable deviation in a prediction. This is much more evened out with 5-year cohorts.
On the other hand, I have added the Belgians, Danes, Swedes and Portuguese.
So divided into seasons, male/female/total, smallest cohorts in 5-year groups, but also larger cohorts.
Well, here you run into a fundamental statistical problem.
1. Within a cohort, mortality is greater as people get older.
2. you cannot solve the problem of statistical noise by making your cohorts intrinsically heterogeneous. Then you're comparing apples and oranges again.
3. So you will have to accept that if the numbers are small, you simply have a lot of “noise”. You can only determine a significant increase or decrease with a T-test or something similar. And that implies that if you cannot read the excess mortality crystal clearly from the graph, it is just noise. You do not strengthen your argument regarding the presence of excess mortality by aggregating. Artificially smoothing data can lead to seeing patterns that are not there in the raw, unaggregated data. And that undermines your authority with regard to other statements, which may be strong and correct.
Jan does not make a trend or prediction; he takes 2019 mortality as a baseline (or perhaps 2018). Striking things will come out of this, such as with those 40-year-old men. I fell for that too when I first saw it.
I don't remember that artifact of 80+ women. However, you did not want to include 2019 in the reference. Well, it seems to me that this is less decisive with an exponential trend line based on 2010-2019 than with a linear trend line based on 2015-2019, which we started with at the time. I still have to work on your dashboard, I haven't had time yet. I'll be able to see that.
Jillis, I don't know if I can find the energy to explain this again.
In terms of objectivity… I have written an extensive critique, mainly on data point 2019, which was a “flu-free” winter. This data point mainly had a major influence on the mortality of 80+. Since this group creates the greatest absolute mortality, a small miscalculation will have major consequences.
Data point 2019 was the last of the trend line. Is it realistic to assume that the following years will all be 'some kind of flu-free'?
The result is then a naturally low mortality expectation, and therefore 'size' excess mortality.
We are now 15 to 7 years further compared to the chosen trend years. Is this still realistic?
In addition, Herman states that there is one death for every 1,500 injections.
Two injections were given in 2021. This then results in a mortality of 1/750 people.
Have you calculated this to the subgroup 50-65, for example? Where avg. 8 people die per week.
With 3500k people and a vaccination rate of 70%, this results in 3250 extra deaths in a few weeks. If you spread this over 20 weeks, this results in an additional mortality of 4.5 per 100k. About a 60% increase.
However, we saw an increase of about 10%. A factor of 6 too high. And if you spread them out over 10 weeks... all model choices and assumptions.
At the very least, such a rough estimate should be spread across subgroups. Same as with the virus. Vulnerable people die earlier from the virus compared to healthy young people. This also applies to vaccination.
And I can still expose many flaws in such a calculation.
In my opinion, this does not help at all in the vaccine damage debate. This will only get you laughs.
Because one thing is clear. This vaccination had quite a bit of collateral damage. We all agree on that. But those exaggerated calculations make no sense.
That's my point too. Totally agree.
But: Herman did apply a triangulation: he discovered that peaks in mortality occurred shortly after the vaccination campaigns. And that is a calculation via a completely different route. If I understand correctly, that 1/1500 is based on that. Not on the generic excess mortality per year.
Fine if you don't think that baseline until 2019 is realistic, but 2020 had, apart from the flu, idiotic measures and enormous problems with primary care that was skipped and hospitals with poor protocols. So apart from all this, this bar has the least excess mortality of the past 5 to 6 years.
Suppose that without all that idiocy, the excess mortality in 2020 would have been a few thousand fewer people, because no one was really saved by this misery. And if you then create a new baseline, you will still see enormous excess mortality and a break in the trend. I dare say that is a fact. Meanwhile, I don't know what will help the discussion, people are short of arguments so I have no illusion that nuance will make any difference in the denial. And then you just have to mention the matter and wake people up. Even though they are e. A few thousand deaths too many. Everyone just needs to know that things are not going well, but almost no one knows this. Or they don't care, that's also possible.
“People” are not blind to the arguments. The expected mortality model is being challenged. And that can also be combated. Because it is “just a model”, with a fairly strong decline in mortality since 2019. And that is an exaggeration.
That is why you can only be convinced if, even with a conservative model, there is evident excess mortality. And that model is: per age cohort of 1 year, there may not be a significant increase in mortality compared to the series of years from 2010 to 2019. And then possibly. still a small decrease compared to recent/lowest years. Remember: 2019 had no flu, so it was already quite low.
And then you see that there was a huge amount of excess mortality from 2021 to 2023. In almost all ages.
And then you also see that this excess mortality had been largely resolved by 2024. But certainly not for every category. And a special feature was the baby boom of 1945. They were quite unhealthy and therefore died more often. You could read that very clearly from those graphs. I have not yet seen the correct figures/graphs for 2025.
I agree with you. I don't want to give hard numbers on (excess) mortality, because then you end up in discussions about models, instead of about tipping points.
That is why it is so nice to divide it into summer and winter mortality.
Winter wants to fluctuate (viruses), summer actually does not.
And then you see that the summer mortality of 2021 is a tipping point.
The summer of vaccinations, QR codes, dancing with Jansen, and the rise of the Delta variant.
In my opinion, this provides more 'evidence'. Statistically, summer mortality is much more stable, and it is easier to see deviations. And it also coincided nicely with a large part of the vaccination wave.
https://x.com/i/status/2041910400904482971
Thanks, that's right! With a slightly less negative baseline you can still demonstrate excess mortality, and I say so. That summer mortality speaks volumes. Hopefully you know that people like M. Keulemans, Van Galen, Andreas Vos, M. Bonten and associates only claim that mortality now only occurs during flu seasons. That is simply a very wrong statement, especially if you base it on a RIVM baseline. But still, thanks for the answer gentlemen!
Good idea. But does that still work now? Isn't it a problem that the vaccinations are now given after the summer or at the very end of the summer? So then you don't see that effect in those stable summer periods, right?
The level has also been increased in the summer – at least, depending on the baseline of course. The one from the RIVM is not too bad: little to nothing to see. The excess mortality has been put on the map precisely because of the summer mortality. That was in '21. Only in '23 did things seem to go well, but the other summers had excess mortality. Also '24 and '25. Pre-2020 was calculated in the same way (Normal mortality) and they look normal.
Hi Jan, the big problem with these vaccinations is that they affect many different places in the body. Acute mortality is likely to be multi-system. Too much production of spike protein, this can also be expressed in different ways. Brain haemorrhage, myocardial infarction, vasculitis, blood clots.
That is already a long list of misery, but the problem is not yet solved. See, for example, the heart muscle inflammation that people suffer from for the rest of their lives, with a very increased risk of dying from it. Blood clots, same story. Spike protein passed through the l.n.d. ends up everywhere in the body and causes inflammation. Consequence autoimmune disease (LC), brain problems, intestinal problems, cardiovascular diseases, nerve problems. In short, not a simple problem.
That's right. Someone wrote that about half of the damage occurs within a few weeks. And the rest may only be done years later... So the “Bonne Method” detects that one half better if it is administered after the injection after the summer.
yes, in that sense for sure! Quite interesting to take a closer look at this. Different age groups, different cohorts. Create a 10-year Baseline per 100k and then calculate. I saw your PDF and I see that you have already created a lot! Now women are generally even more sensitive to these problems. Perhaps you can highlight a number of graphs and see how this summer-winter difference works out.
See also this example. https://virusvaria.nl/2024-eenderde-meer-40-tot-50-jarige-vrouwen-gestorven/