Good analysis. It's worth also looking at the excess mortality stats for the UK. They are showing very high excess deaths for this time of year. Interestingly not just COVID either.
Very interesting and disturbing observation. Just one remark: the population of the UK is smaller than that of the US by a factor of five. This doesn't change your conclusion, however, and the situation in the UK may indeed portray the near-future in the US.
I was a few days ahead of you on this; I run the PHE numbers as soon as they are published.
What I can't understand is why the case efficacy is collapsing like this. Compare that to the Connecticut and Massachusetts numbers you referenced in an earlier post, my comment on that post with the Tennessee data, the data I'm analyzing from the raw numbers put out by my large employer - and all show a current case efficacy of roughly 70%. Yet the PHE data has huge negative efficacy in every age cohort, age 30 and above?
Why?
Is it because of the heavy reliance on the Astra-Zeneca vaccine in the UK? Is it a case of Simpson's Paradox with the US data since we only have a composite efficacy and don't have the data broken down by age group? Something else entirely?
Igor, can you help with providing further details please? I shared your article with my community and someone has come back with the following comment:
"I am merely comenting on the erroneous/fraudulent use of statistics that Igor is employing to bolster his case. Specifically he defines Efficacy as 1 - vaxxed case rate/ unvaxxed case rate. The correct formula would be Efficacy = vaxxed case rate times percentage of people vaxxed / unvaxxed case rate times percentage of people unvaxxed. If you then measure efficacy using the correct formula you will find that, for weeks 32-35, about 1.5 % of vaxxed people tested were diagnosed with Covid while the percentage for unvaxxed people was about 3.6% For week 41 (the second url provided by Igor is invalid, but you can easily access more recent cases) the percentages are 1.7% and 4.2% respectively - slightly higher, maybe because of the greater prevalence of the delta strain, but still more than double for the unvaxxed than the vaxxed.
Is there a standard way to calculate efficacy? We are based in Australia so I wonder if there is a difference between UK and AU? Anyway, any further information would be helpful. Thanks Igor!
When the latest Vaccine Surveillance Report came out a few days ago, I looked at the efficacy numbers. They all got worse AGAIN.. Every week I think, "Maybe this is the week things bottom out," and every week I am wrong.
It would be interesting to compare all cause mortality in the vaxxed vs unvaxxed groups. Then the rapidly diminishing protection from covid death afforded by vaccines can be shown to be more than offset by deaths from other causes, exacerbated by weakened immune response. When / if vaccine protection from death goes to zero or lower, I think we may see a perfect storm this winter in the UK and other Northern European countries, as other viruses like boring old flu get their teeth into a large target group with crushed immunity.
First UK jabs were in early December, mainly to doctors and nurses.
The jabbed in the UK have gone from hiding in their basements to mixing with one another as if the 'plague' never happened. Many still wear masks in shops voluntarily or, as in the case of Scotland and Wales, are forced to. They act fearful in a shop but them practice no social distancing or any of the other guff once they step back outside. The jabbed act as if they are invulnerable.
PHE has been rebranded as the Health Security Agency. Nothing Orwellian in that.
it is very dangerous how statistics can be manipulated. It’s important to provide context. different statistic that will likely become more important as we get more cases and have a highly vaccinated population.
Example: when over 50% of the entire population is double dosed, The vaccinated population is greater than the unvaccinated. Often with minorities (unvaccinated) they over represent in statistics.
A simple scenario
100 people 80% vaccinated - 80 vax, 20 not.
the virus is bad and 30 people get infected - 66% (20) of them are vaccinated and just 33% (10) are unvaccinated
Now.. this looks bad, it seems like the vaccine just doesnt work.
But look at the populations - those 10 unvaccinated are 50% of their population but just 25% of the vaccinated.
If the virus was going to infect at the rate it did for the unvaccinated, there would have been 50 cases (or more as vaccinated people have the power to break chains)
At some point, a higher percentage of cases will be among the vaccinated. but the thing with vaccines is how many cases its able to prevent. It is impossible to measure this, we can only report on infection rate changes
NZ statisitcs for rate (we do not have a large dataset yet)
57% of population is vaccinated
43% is not (includes in eligible)
as of 18 october, 2005 cases meant there was
4.62 cases for ever 100,000 of the vaccinated population
88.2 cases for every 100,0000 of the unvaccinated population
0.14 of every 100,000 hospitilisasions for vaccinated pop
8.22 of every 100,000 hospitilisations for unvaccinated
In the UK the DEATH rate per 100,000 (reasonably vaccinated, no restrictions, 50,000 cases per day)
Thanks Igor. From mid-June to Sept i downloaded / extracted / imported the periodic UK Technical Report to perform time series analysis.
My analysis mirrors your own, I initially conducted the analysis to share with family & friends. I am grateful you've made the time to post your findings. I know it takes time to (1) conduct the analysis and (2) to publish it.
In addition to what you've shared, I noticed that there was a substantially higher risk (rate) to infection between shots 1 and 2. Have you considered consolidating the 1 and 2 jabs as a single group vs. unvaccinated? My analysis is over a month old, and at the time, there was virtually no difference in the rate of death in these two groups; however, as your article suggests, the rate of deaths among the vaccinated will likely increase.
Thanks for that. Since we know efficacy drops by time since vaxxed is it possible to get those numbers by time since vaxxed? That is, if we look at cases/hospitalizations/deaths specifically after 35 weeks since vaxxed, the numbers should be even worse, with the numbers worsening for longer times since vaxxed. Indeed the claimed usefulness of the vax against hospitalization and death might evaporate for these long times since being vaccinated.
Good analysis. It's worth also looking at the excess mortality stats for the UK. They are showing very high excess deaths for this time of year. Interestingly not just COVID either.
Very interesting and disturbing observation. Just one remark: the population of the UK is smaller than that of the US by a factor of five. This doesn't change your conclusion, however, and the situation in the UK may indeed portray the near-future in the US.
I'd like to add a few more links: table 5: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1018547/Technical_Briefing_23_21_09_16.pdf
table 15 https://publichealthscotland.scot/media/9030/21-09-08-covid19-publication_report.pdf
I was a few days ahead of you on this; I run the PHE numbers as soon as they are published.
What I can't understand is why the case efficacy is collapsing like this. Compare that to the Connecticut and Massachusetts numbers you referenced in an earlier post, my comment on that post with the Tennessee data, the data I'm analyzing from the raw numbers put out by my large employer - and all show a current case efficacy of roughly 70%. Yet the PHE data has huge negative efficacy in every age cohort, age 30 and above?
Why?
Is it because of the heavy reliance on the Astra-Zeneca vaccine in the UK? Is it a case of Simpson's Paradox with the US data since we only have a composite efficacy and don't have the data broken down by age group? Something else entirely?
Any ideas?
Igor, can you help with providing further details please? I shared your article with my community and someone has come back with the following comment:
"I am merely comenting on the erroneous/fraudulent use of statistics that Igor is employing to bolster his case. Specifically he defines Efficacy as 1 - vaxxed case rate/ unvaxxed case rate. The correct formula would be Efficacy = vaxxed case rate times percentage of people vaxxed / unvaxxed case rate times percentage of people unvaxxed. If you then measure efficacy using the correct formula you will find that, for weeks 32-35, about 1.5 % of vaxxed people tested were diagnosed with Covid while the percentage for unvaxxed people was about 3.6% For week 41 (the second url provided by Igor is invalid, but you can easily access more recent cases) the percentages are 1.7% and 4.2% respectively - slightly higher, maybe because of the greater prevalence of the delta strain, but still more than double for the unvaxxed than the vaxxed.
Is there a standard way to calculate efficacy? We are based in Australia so I wonder if there is a difference between UK and AU? Anyway, any further information would be helpful. Thanks Igor!
When the latest Vaccine Surveillance Report came out a few days ago, I looked at the efficacy numbers. They all got worse AGAIN.. Every week I think, "Maybe this is the week things bottom out," and every week I am wrong.
According to a couple things I've read, the US will not follow the UK pattern because we have enough percentage who have forgone vaccination and have much better immunity through covid recovery. You can thank us later. https://www.geertvandenbossche.org/post/the-keys-to-unlock-the-golden-gate-of-herd-immunity-towards-sars-cov-2
It would be interesting to compare all cause mortality in the vaxxed vs unvaxxed groups. Then the rapidly diminishing protection from covid death afforded by vaccines can be shown to be more than offset by deaths from other causes, exacerbated by weakened immune response. When / if vaccine protection from death goes to zero or lower, I think we may see a perfect storm this winter in the UK and other Northern European countries, as other viruses like boring old flu get their teeth into a large target group with crushed immunity.
First UK jabs were in early December, mainly to doctors and nurses.
The jabbed in the UK have gone from hiding in their basements to mixing with one another as if the 'plague' never happened. Many still wear masks in shops voluntarily or, as in the case of Scotland and Wales, are forced to. They act fearful in a shop but them practice no social distancing or any of the other guff once they step back outside. The jabbed act as if they are invulnerable.
PHE has been rebranded as the Health Security Agency. Nothing Orwellian in that.
looks like the promised "superhuman immunity" is not going to be forthcoming. They'll just have to settle for the non-existent variety? https://www.mirror.co.uk/news/uk-news/super-immunity-double-jabbed-who-24938837
it is very dangerous how statistics can be manipulated. It’s important to provide context. different statistic that will likely become more important as we get more cases and have a highly vaccinated population.
Example: when over 50% of the entire population is double dosed, The vaccinated population is greater than the unvaccinated. Often with minorities (unvaccinated) they over represent in statistics.
A simple scenario
100 people 80% vaccinated - 80 vax, 20 not.
the virus is bad and 30 people get infected - 66% (20) of them are vaccinated and just 33% (10) are unvaccinated
Now.. this looks bad, it seems like the vaccine just doesnt work.
But look at the populations - those 10 unvaccinated are 50% of their population but just 25% of the vaccinated.
If the virus was going to infect at the rate it did for the unvaccinated, there would have been 50 cases (or more as vaccinated people have the power to break chains)
At some point, a higher percentage of cases will be among the vaccinated. but the thing with vaccines is how many cases its able to prevent. It is impossible to measure this, we can only report on infection rate changes
NZ statisitcs for rate (we do not have a large dataset yet)
57% of population is vaccinated
43% is not (includes in eligible)
as of 18 october, 2005 cases meant there was
4.62 cases for ever 100,000 of the vaccinated population
88.2 cases for every 100,0000 of the unvaccinated population
0.14 of every 100,000 hospitilisasions for vaccinated pop
8.22 of every 100,000 hospitilisations for unvaccinated
In the UK the DEATH rate per 100,000 (reasonably vaccinated, no restrictions, 50,000 cases per day)
1.9 for vaccinated
8.4 for unvaccinated
Thanks Igor. From mid-June to Sept i downloaded / extracted / imported the periodic UK Technical Report to perform time series analysis.
My analysis mirrors your own, I initially conducted the analysis to share with family & friends. I am grateful you've made the time to post your findings. I know it takes time to (1) conduct the analysis and (2) to publish it.
In addition to what you've shared, I noticed that there was a substantially higher risk (rate) to infection between shots 1 and 2. Have you considered consolidating the 1 and 2 jabs as a single group vs. unvaccinated? My analysis is over a month old, and at the time, there was virtually no difference in the rate of death in these two groups; however, as your article suggests, the rate of deaths among the vaccinated will likely increase.
Thanks for that. Since we know efficacy drops by time since vaxxed is it possible to get those numbers by time since vaxxed? That is, if we look at cases/hospitalizations/deaths specifically after 35 weeks since vaxxed, the numbers should be even worse, with the numbers worsening for longer times since vaxxed. Indeed the claimed usefulness of the vax against hospitalization and death might evaporate for these long times since being vaccinated.
Robert Clark