I just noticed open in one of my other browsers that Norman Fenton, Martin Neil, and Scott McLachlan published about Simpson's paradox in the UK data. Good statisticians usually make this kind of correction in population data. It's not too uncommon.
The source data should NOT have been merged on age axis "due to low incidence". It should have been the time axis. Due to cautionary principle I'm forced to I suspect they are hiding et harm done to the younger ages in the massive group 10-59. Due to higher mortality at older ages, even a small benefit to ages 50-59 can hide most of large harm done in the younger age groups.
In any case, there is data from Sweden where the 14-day deaths post 2nd dose is higher (3939) than the average 14 day death rate for whole nation (~3000).
I applaud both you and Alex but there are still mistakes here. If, like the ONS, you are using the population by vaccination status during the same week as the death then you are overstaing unvaccinated deaths and understaing vaccinated deaths.
Mathew sorry but this is a little bit above my level intelligence wise. So please forgive a stupid question. Are you saying that the vaccinated are not dying at twice the rate? In fact the difference is small? Therefore the vaccine causes many injuries but not many deaths?
I have calculated efficacy, NNT and NNH using Pfizer, Moderna and JNJ's own data submission to the FDA and adjusting the endpoint for "All Cause Morbidity and Mortality". Here it is:
All Cause Severe Morbidity Endpoint
Moderna Control Moderna Control
Risk Risk
Randomized 15210 15210 VE= -322.59%
Day of follow up 56 56 ARR= -20.00%
#Severe Covid Cases 0 30 RRI= 322.59%
#Unsolicited Severe Adverse Reactions 234 202 NNH= 5
#Solicited grade 3 AE, shot1 848 361 one harmed every 5 vaccinations
Bertram's discovery of the importance of the 3-week delay between infection and death probably affects the curves in the early months, but when vaccination rates drop towards zero the later results will be accurate.
If only we could have known that the most sued companies in existence, paying out many billions in settlements for health product malfeasance, could have anything to do with exaggerating their products benefits, while hiding it's flaws....
Great article, but wouldn't it be even more interesting to focus on the category "Deaths 21 days or more after first dose" since there seems to be a VERY intriguing all-cause mortality signal here?
And Simpson's paradox should be less of a concern in this case.
If you could calculate the area between the actual curve and the expected curve (or the sum of differences for each week) to compare overall above expected vaccinated to unvaccinated death rates, that could reveal a difference. However, most important is to also look at the all cause deaths of the partially vaccinated, which England may be the only country that keeps and reveals such records. I don't know what the difference in expected mortality should be but, perhaps, exactly the same as for the vaccinated, however, you estimated those numbers, given the age groups taking up the vaccines each week.
Thank you for putting this together. If I am understanding correctly, all-cause mortality for unvaccinated and vaccinated both remain slightly elevated at about the same level in the U.K. This would seem to dispute both the theory that these vaccines are efficacious and have saved lives AND that they are harmful and have caused unnecessary deaths in the (relatively) short-term. (Unless somehow they are capable of killing those who haven't received them, which would fulfill the darkest wishes of a number of resentful jabbed.) Still damning data for the vaccines, especially for those who are trying to mandate. We are not the ones who have to prove anything.
Thank you very much, I arrived at the similar conclusions regarding the Berenson analysis. Furthermore, I kept working on the dataset in question. It seems to miss a large number of residents. The total for 10 and up is about 39 million (check if I am right). I figure another 7-8 million under 10. That's 46-47. England's population is about 56 million. What happened to the rest?
Sorry the formatting got destroyed. Suffice it to say Pfizer 1 harmed for every 241 vaccinations administered, Moderna 1 harmed every 5 vaccinations administered JNJ 1 harmed every 74 vaccinations administered. Mind you this is THEIR OWN DATA. My experience is that these data have been altered to hide the severity of the adverse events. It is not possible that the FDA is allowing these products to be used.
Michael, take a look at table 2 of the raw data set. It has age standardized all cause mortality rates for non, partial and full vaxxed over the full population. The numbers for "partial" categories are off the charts! Using the rollout data you can back out the age ranges actively shooting as well.
Btw I tried to do that painstaking work of reconstructing the age stratified rollout data from the plots and only got through a part of it before giving up. Would you be willing to make that file public?
I just noticed open in one of my other browsers that Norman Fenton, Martin Neil, and Scott McLachlan published about Simpson's paradox in the UK data. Good statisticians usually make this kind of correction in population data. It's not too uncommon.
https://www.normanfenton.com/post/paradoxes-in-the-reporting-of-covid19-vaccine-effectiveness
The source data should NOT have been merged on age axis "due to low incidence". It should have been the time axis. Due to cautionary principle I'm forced to I suspect they are hiding et harm done to the younger ages in the massive group 10-59. Due to higher mortality at older ages, even a small benefit to ages 50-59 can hide most of large harm done in the younger age groups.
In any case, there is data from Sweden where the 14-day deaths post 2nd dose is higher (3939) than the average 14 day death rate for whole nation (~3000).
(Thanks TB)
data laundering ... that's it!
I applaud both you and Alex but there are still mistakes here. If, like the ONS, you are using the population by vaccination status during the same week as the death then you are overstaing unvaccinated deaths and understaing vaccinated deaths.
Mathew sorry but this is a little bit above my level intelligence wise. So please forgive a stupid question. Are you saying that the vaccinated are not dying at twice the rate? In fact the difference is small? Therefore the vaccine causes many injuries but not many deaths?
A lot of recent studies and papers suggest that there is a negative correlation between serum vitamin D3 levels and hospitalization and mortality. Eg:
https://www.sciencedirect.com/science/article/pii/S0188440921001983
https://www.medrxiv.org/content/10.1101/2021.09.22.21263977v1
We are also told by Dr John Campbell that Fauci has admitted by email that he takes 6,000 IUs of Vitamin D3 daily.
How reliable are these studies? Are they committing any statistical sins?
I have calculated efficacy, NNT and NNH using Pfizer, Moderna and JNJ's own data submission to the FDA and adjusting the endpoint for "All Cause Morbidity and Mortality". Here it is:
All Cause Severe Morbidity Endpoint
Moderna Control Moderna Control
Risk Risk
Randomized 15210 15210 VE= -322.59%
Day of follow up 56 56 ARR= -20.00%
#Severe Covid Cases 0 30 RRI= 322.59%
#Unsolicited Severe Adverse Reactions 234 202 NNH= 5
#Solicited grade 3 AE, shot1 848 361 one harmed every 5 vaccinations
#Solicited grade 4 AE, shot1 5 6 NNT= -5
#Solicited grade 3 AE, shot2 2884 341 VE= -322.59%
#Solicited grade 4 AE, shot2 14 3 EER CER RR
#Total Severe Events 3985 943 0.261998685 0.061998685 4.225874867 -322.59%
Deaths 2 3 Death Rate= 0.0131%
Pfizer Control
Pfizer Control Risk Risk
VE= -52.38%
Randomized 21720 21728 ARR= -0.41%
Day of follow up 81 81 RRI= -52.38%
#Severe Covid Cases 1 9 NNH= 241
#Unsolicited Severe Adverse Reactions 240 139 one harmed every 241 vacinations
#Unsolicited Life Threatening Events 21 24 NNT= -241
#Total Severe Events 262 172 EER CER RR
Deaths 2 4 0.012062615 0.007916053 1.523816866
Death Rate= 0.0092%
Jansen Jansen Control Control
Randomized 19630 19691
Safety Subset 3356 3386
Day of follow up 28 28 VE= -80.32%
#Severe Covid Cases 21 78 ARR= -1.35%
#Solicited grade 3 Adverse Events RRI= -80.32%
Local extrapolated 135 23 35 6 NNH= 74
Systemic extrapolated 357 61 122 21 one harmed every 74 vacinations
#Unsolicited grade 3-4 Adverse Events 83 96
#Total Severe Events 595 331 RR control 0.01680971
Deaths 3 16 RR vax 0.030310749
Death Rate= 0.0153%
Excellent analysis, thanks!
Bertram's discovery of the importance of the 3-week delay between infection and death probably affects the curves in the early months, but when vaccination rates drop towards zero the later results will be accurate.
https://bartram.substack.com/p/the-importance-of-the-delay-between
https://bartram.substack.com/p/the-importance-of-the-delay-between-6e5
If only we could have known that the most sued companies in existence, paying out many billions in settlements for health product malfeasance, could have anything to do with exaggerating their products benefits, while hiding it's flaws....
Great article, but wouldn't it be even more interesting to focus on the category "Deaths 21 days or more after first dose" since there seems to be a VERY intriguing all-cause mortality signal here?
And Simpson's paradox should be less of a concern in this case.
If you could calculate the area between the actual curve and the expected curve (or the sum of differences for each week) to compare overall above expected vaccinated to unvaccinated death rates, that could reveal a difference. However, most important is to also look at the all cause deaths of the partially vaccinated, which England may be the only country that keeps and reveals such records. I don't know what the difference in expected mortality should be but, perhaps, exactly the same as for the vaccinated, however, you estimated those numbers, given the age groups taking up the vaccines each week.
Why is ONS grouping 10 year olds with sixty year olds in the first place?
Thank you for putting this together. If I am understanding correctly, all-cause mortality for unvaccinated and vaccinated both remain slightly elevated at about the same level in the U.K. This would seem to dispute both the theory that these vaccines are efficacious and have saved lives AND that they are harmful and have caused unnecessary deaths in the (relatively) short-term. (Unless somehow they are capable of killing those who haven't received them, which would fulfill the darkest wishes of a number of resentful jabbed.) Still damning data for the vaccines, especially for those who are trying to mandate. We are not the ones who have to prove anything.
Thank you very much, I arrived at the similar conclusions regarding the Berenson analysis. Furthermore, I kept working on the dataset in question. It seems to miss a large number of residents. The total for 10 and up is about 39 million (check if I am right). I figure another 7-8 million under 10. That's 46-47. England's population is about 56 million. What happened to the rest?
Sorry the formatting got destroyed. Suffice it to say Pfizer 1 harmed for every 241 vaccinations administered, Moderna 1 harmed every 5 vaccinations administered JNJ 1 harmed every 74 vaccinations administered. Mind you this is THEIR OWN DATA. My experience is that these data have been altered to hide the severity of the adverse events. It is not possible that the FDA is allowing these products to be used.
Michael, take a look at table 2 of the raw data set. It has age standardized all cause mortality rates for non, partial and full vaxxed over the full population. The numbers for "partial" categories are off the charts! Using the rollout data you can back out the age ranges actively shooting as well.
Btw I tried to do that painstaking work of reconstructing the age stratified rollout data from the plots and only got through a part of it before giving up. Would you be willing to make that file public?