Why we haven't paid attention to the new data from the UK yet

Last year I threw myself into the data from the CBS of the United Kingdom, the ONS (Office for National Statistics). The tables contain detailed figures. However, based on hard counts, they generated age-standardized (ASMR) data that could not be understood and could not be calculated back. In some groups, I came across more vaccinated people than there was population. Then I started working with another database (NIMS) myself and ignored ASMR. After all, you only need that if you are going to add up age groups. In the case of a strongly age-discriminatory disease such as Covid-19, this serves a secondary purpose: an overall average. But primarily it is about the age cohorts.

Excess mortality is now a global theme and real statisticians put a lot of time into it. Despite insisting on working with the ONS data again, I only waited a while. And indeed: it turns out to be dubious and suspicious data again. Various substacks and blogs have paid attention to it. The following is the translation of an article that is easy to follow and gets to the heart of one of my favorite substacks: boriquagato. At the end, another postscript. But then you must have read the review of el gato malo.

The new UK US data is out and it's worse than before

This was not a solution, it was a further detriment

el gato malo Feb 23

In May, the UK Statistical Office (ONS) stopped reporting the number of deaths from all causes by vaccination status. In response to a number of serious criticisms, they even went so far as to admit that their data were in fact not fit for purpose and of too low quality to assess the effectiveness of vaccines and their results.

One of the main complaints was relying on an old census to determine the number of unvaccinated people by taking that outdated census and subtracting the number of people pricked from it.

This causes a serious underestimation of the number of unvaccinated people. the population has increased since then [and the registry didn't cover the entire population anyway]. As a result, poor results (such as deaths) of this larger population are attributed to too small a group.

And that in turn means that the percentage of bad outcomes per person or per year becomes higher than it actually is. [10 deaths among a population of 5,000 is twice as bad as 10 deaths among 10,000 people, half of whom are unregistered].

Comparisons lose their validity because the basic risk is greatly inflated. [This is consistent with My findings in May 2022, ed.]

It is a huge and persistent problem with their methodology and a problem that they have proposed to solve by updating to the 2021 census. None other than Sarah Caul of ONS herself has described this update and the reasons for it. (Interestingly, one of them was to include information from a 4th booster and this data doesn't seem to appear anywhere in the new version).

In particular, I had a lot of questions about the number of unvaccinated, because if you used a better number (like UK HSA) the signals were massively reversed and it became clear that vaccines, based on age stratification, were linked to much higher mortality rates from all causes.

This seemed like a very suspicious moment to stop reporting.

My expectation was that this data would clear the negative signal for all-cause mortality due to vaccines. It would be too big to hide.

Well, I was half right. It was too big, but they seem to have gone to great lengths to hide it. The new set of ONS data (available here) is downright worse than before. They have not solved the problem of too low counts, but have added to it. This data seems highly manipulated and contains such improbable/impossible assumptions that I'm not sure what to call it other than deliberate misrepresentation.

This is not a clean sweep. This is a money laundering operation.

Let's see:

We start here in Table 1. I have selected all deaths from causes for December.

The main claim from this data is that vaccines are associated with fewer deaths from all causes. 944.9 deaths per 100k person-years compared to 1026.7 in the never vaccinated. But this claim is based on a truly Herculean data manipulation.

1. When we calculate the simple ratio of deaths per 100k person-years, we get 206 for non-gevaxt versus 1113.6 for ever gevaxt. Gevaxt is more than 5X as high. This does not mean that it is wrong to claim that gevaxt is better on an age-standardized basis. The non-vaxxed people are younger and thus have lower expected mortality rates. But it does mean that the age adjustment does all the work here and that the result is only as good as that assumption/model.
2. Even worse, if we calculate the assumption about the percentage of ever-gevaxt versus never gevaxt by comparing person-years, we arrive at their core assumption about the size of the non-gevaxed population: 15.2% of the total population. And that's insanely low. There is no data anywhere in the UK that remotely resembles this. And this assumption spoils every other aspect of this analysis to the point of extreme inversion.

According to the UK HSA (Health Security Agency), this number is absurdly low. There is not a single group of British society under the age of 50 that has been so heavily vaccinated. Some have tried to claim that the UKHSA data cannot be compared with this ONS data because it is a subset, but I find this objection unconvincing for several reasons:

1. this is a large subset that comprises the majority of the UK population
2. the data is of MUCH better quality, as it comes from counting actual medical records rather than inferring/assuming the number of unvaccinated people who pose the basic risk used for all comparisons.
3. The subgroup included in it disproportionately excludes the very young and more recent immigrants. It seems likely that this will underestimate the number of unvaccinated people relative to the entire population.
4. we need to check these ONS claims somehow and this seems to me to be the most authoritative. maybe it's not exact, but I suspect it's of much higher quality and less error-prone than the ONS estimate, which looks progressively worse. it is the one used to measure most cases of disease in the UK.

In order to avoid (as much as possible) the problems of the opaque ONS age correction and to get a more detailed picture, I will switch to the age-stratified data. Unfortunately, almost all of this data is presented (perhaps on purpose) in a Bayesian muddle where it's broken down by vaccination status as if that's some sort of independent variable, when it's not. Everyone who has had 3 doses has ever had 2 and 1 and has gone through that risk strata. This creates a "left-wing truncation", where only the healthiest (and least culled) reach dose 3.

[Again, this is exactly the method I used last year and which caused me to be seen by statisticians who themselves could not explain why it was wrong. It just had to be in ASMR, so you see how dogmas play tricks. Occupational blindness.] Ed.

The only real exception is Table 5, which contains very little data, but is still instructive to look at, since we can see how many % of deaths per age strata occurred in the gevaxten vs. the unvarnished. The problem with this data is that they give deaths by status, but not a count of people or person years. This seems like a strange omission, given that it occurs in many other parts of the data and they clearly do have this data available (and the data is part of the kind of "hide the ball" games inherent in this data). But we can still calculate the percentage of deaths in each group. If we then use the UKHSA data to estimate the group size, we can get some risk ratios.

These look rather bad for the vaccines. Mortality from all causes is higher in all groups. (I had to merge 80-9 and 90+ into one group because UKHSA reports vaccination that way and there is no granularity above 80)

This is nothing like the data that ONS is trying to pass on. It shows a greatly increased risk of death among vaccinated people if we use more plausible population statistics. Is this 100% exact? No. But it's good enough to give us an intuition that everything is not OK and what trick is being shown here.

ONS provides the data on personal years based on age stratification in Table 2 and also adjusts them to age despite this stratification. Unfortunately, they no longer provide "ever vaccinated" vs. "never vaccinated" and split it into the bad Bayesian cohorts that are now familiar. Unlike May, they no longer provide enough data to even assess the 21-day period after each jab.

This seems to be a serious omission, as that period in May showed a greatly increased risk for many.

If we value (and ignore this problem) their stratified and Bayes consideration-based ASMR (age mortality rate per 100,000 inhabitants) data:

You can literally see how the risk is pushed backwards and out of categories. 1 dose is extremely negative because it has nowhere to go. 2 doses have a strong negative signal, but suddenly a third makes it effective? (by the way, only in 40-79)

This seems biologically implausible. If a vaccine can't teach you to resist a disease, then more vaccines probably can't. The idea that you have to keep antibodies high by constantly stimulating has never been based on facts. It seems to imply that you can not remember the reaction and even repeat it 6 months after vaccination. None of that ever passed the smell test. More likely, these are just statistical games.

Based on what Alberta accidentally admitted, you can see how bad that can be, but if you stop boosting and redefine "vaccinated", the signal may come out. This whole idea is like "extend and pretend" using math games.

Even if this signal were real, it also implies that anyone who makes it into the "booster" category has probably already taken a whole host of higher risks, and thus is a cleared cohort from which the weakest have been removed. This in itself would make the claims about the operation of d3 problematic.

But there is another reason to set aside this data and the claims: The basic risk to non-targeted persons seems enormously exaggerated and this was not only not remedied in the new version, it was even accentuated.

See this comparison between the previous population sizes and the new ones:

The use of the "new count" somehow ensured that every age group except 80+ contained even fewer non-gevaxed persons.

The margins were significant and if we zoom in on that effect we see this:

In the middle groups, aged between 40 and 79, which is at the heart of the "reported effectiveness for boosters", the size of the cohort of non-gevaxed individuals decreased by an average of 10.4% between May and December 2022.

Anyone who thinks that 1 in 10 people who made it without a jab until May 2022 has decided to get the jab, raise your hand.

I don't know a single adult in my entire social network who has made that choice. Somebody?

This seems incredibly implausible and is literally nowhere to be found in other UK data or in any anecdotal record I have heard.

These figures are fanciful and these data are a tool for misattributing all-cause deaths to the unflated by further underestimating them.

The effect of this is profound.

When using the UKHSA data, the risk ratios are roughly doubled and are all above 1 (a risk ratio of 1 = no effect).

Those aged 50-59 who, according to the description, were 37% less likely to die from all causes, are now 25% MORE likely to die. 80+ jumps from a 4% benefit to an 88% increased risk of death.

But how accurate is this? That's hard to say, but probably good enough for this kind of crude analysis. The UKHSA data looks much better and is probably heading toward too low a number of non-gevaxs, so I think they give us a pretty good clue, especially when it's clear that US is playing silly games with the denominators and moving them to unlikely levels to make trends they don't like disappear.

If the US is correct and the UKHSA is wrong, then the group outside the UKHSA must have been vaccinated at incredibly high rates compared to what is widely considered to be the "representative group" of the UK in terms of health. It's not even clear to me that this is mathematically possible, but I don't have the data to do the analysis. (if anyone knows where I can get it, please let me know).

Igor looked at this and saw the same thing.

His conclusion that the unflated individuals are under-counted by about 50% (meaning the group is roughly 2X as large as claimed) is very similar to mine.

Its uniformity is really striking.

... and so we end up again at "there's just no way to trust a vaccination claim made with this ONS data."

All indications are that this is a mass manipulation that was suppressed after May to find a way to further manipulate the data before releasing it, because the old manipulation was insufficient to hide an increasingly bad signal.

The vaccination rates are highly implausible and seem to greatly underestimate the unvaccinated in order to double the basic risk attributed to them. amazingly, this is still not enough to make 1 or 2 jabs seem effective and if we applied the reduced basic risk to that, it would get even worse.

This was not a clarification of the data or the introduction of better practices, but a last-ditch attempt to manipulate the data and thus hide a debacle.

We can argue about the UKHSA data and its comparability and how accurate the assumptions are, but it's real data from medical records versus a calculated figure from the ONS and nothing else seems to match the ONS data, not even the OWID, which is notorious for over-counting vaccinations.

To believe that this is effective for the reinforced, you must accept the following principles:

1. Vaccines that don't work as a double-dose "full course" suddenly become effective as a booster.
2. The problem with Bayesian manipulation is small and the duration of single and double doses and especially the clearly dangerous periods in the 21 days after the jab is small enough not to reverse the risk ratio. (it's not, you can see it here)
3. that the US doesn't play games with age standardization (could, but not proven)
4. that the ONS correctly calculates the percentage of unvaccinated people, despite the fact that their data does not match that of others, especially those of UKHSA, who use a much more reliable method of counting versus modeling.
5. that the UKHSA database here is not a reasonably good proxy that is conservative rather than exaggerated in counting the unvaccinated, but somehow massively underestimates the pricked individuals and overestimates the unvaccinated by ~2X.
6. and that either from the end of May to the end of December 2022, ~10% of jab refusers in the main age groups decided to get vaccinated, or that the outdated count used by the ONS actually counted the number of people in the UK too high.

Sorry guys, but this is a bridge too far to accept and is no longer plausible.

Nothing about this analysis, or the claims that underlie it, seems to be true or even logical. None of it is consistent with other data.

I think they're playing the same games as CDC and putting Bayesian data crime on top of it.

I had really high hopes for this dataset as a way to get the legendary "all cause mortality by vaxx status" data, but I have to admit that this is just junk, unfit for purpose, and probably manipulated to go against it. gatopals™ martin neil and norman fenton seem to have been right about ONS all along. see footnotes 6-11 HERE).

These are bad data, possibly on purpose, they do not pass any snooping test, became more, no less smelly, and even rudimentary adjustments to bring them into line with higher-quality data greatly reverse the supposed signal for the effectiveness of the booster, in itself an incredible inversion of double-dosed results.

I just don't see any point in working with it like that.

Another reason to regulators and government agencies can no longer be trusted and open data requirements.

"That doesn't happen to us..."

Really?

If the CBS of the United Kingdom plays such a peek-a-boo with mortality and vaccination data, it strengthens my idea that we cannot expect pure information from our own CBS. The shady fuss so far has only further confirmed that idea. See also the occasionally distraught tweets of Ruben van Gaalen. That that man hasn't resigned in the meantime...

CBS spokesperson @rubenivangaalen praises the ONS[ report analyzed above, again emphasizing the data obfuscation with ASMR:

And no, this valuable ONS analysis is not the final answer to all questions: thorough research is still needed. Even better, people's health is also taken into account. And they also look at very specific risk groups. Many investigations are still ongoing now and in the future.

— 📊 Ruben Van Gaalen (@rubenivangaalen) February 23, 2023

CBS spokesperson @rubenivangaalen puts the severity of excess mortality among 40-50 people into perspective: 12 unexplained deaths per week, particularly between 40-50 years, he classifies as "numerically a small group compared to the elderly.”
(He mentions 650 deaths in 2022, which is 12.5 per week. Mainly 40-50 year olds. Unexplained, no research. That's a national disaster.)
It may have escaped Ruben's attention that in the elderly, even everyone over 50, everyone dies. All. This is very different for people under 50. By the time they die, they almost never fall into the 50-min category.
That's not a statement for a statistician... What are they doing to those people anyway, is the grave they dug for themselves not yet deep enough?