I’ve been trying to reverse engineer the Australian study for exactly the same reasons you mention (and some others).
Did you notice:
# remove unvaccinated as it is biased group
This suggests to me that the unvaccinated included in the analysis are somehow a rolling cohort of people that went on to become vaccinated, so the comparison in their analysis is probably not between the vaccinated and never the vaccinated which I had assumed. How you achieve that in a study where death is the end point beats me 🤷♂️
The second part of the study was an analysis of Aged Care residents. The total number of aged care residents over 65 in the database is about 1,250,000 but the number included in the Aged Care analysis is only about 175,000.
They seem to have excluded roughly 85% of the Aged Care database. So, who was excluded from the Aged Care analysis?
Most of the exclusions were non-permanent residents. The number excluded is HUGE, it’s just over 1,000,000 aged care residents. This is a remarkable number. It would suggest that the vast majority (roughly 80%) of Australia’s aged care residents are in fact “not Australians”.
The study also suggests the % vaccinated figures quoted by the Government are exaggerated it is very high at about 95% but not the almost 100% (or over 100% if you do the math) they claim.
Why do you think the total number of aged care residents should be as large as 1,250,000? That would be 5% of the population (25 million).
Why do you think the study claims 100% vaccination rate? 0.2 million (unvaccinated) of 3.9 million (in the study) is around 5%.
You are right, of course, that the fact of people changing categories (from unvaccinated to vaccinated, or from n doses to n+1 doses, or from 8-90 days to 90+ days without further vaccination) is a kind of selector (you can only change categories if you haven't died in your current category), and obscures such studies.
Thanks for the reply, I agree the change in categories obscures the result, though if properly done I think it can be account for in the survival analysis model, or you can see its effect in the residuals, or something else along those lines but I’m far from being an expert in this regard.
The total aged care population is shown in their supplementary material as 1,259,410. It’s their figure not mine. According to the United Nations; Australia has the third highest life expectancy in the world, Monaco and Japan have higher life expectancies than Australia, Australia's male life expectancy ranked second, and female sixth, in the world. So I’m not surprised we’ve got a lot of old people in aged care.
I’m not saying the study is claiming the vax rate is close to 100% the figure they calculate is roughly 95% (surprisingly it’s even slightly lower in the aged care population) which i think is probably about right. Other government departments like Health like to exaggerate the % vaxxed (particularly in the elderly) this study confirms it’s not as high as they advertise. I’m more inclined to believe the figures shown in the study than other sources even though I think their methods are a bit shit, their data is good. Other sources claim the rate its very close to 100% in the elderly but they don’t show the actual figures they just list it as >95% in their tables. However, they do give the raw data they use to calculate the tables and if you use the Heath Departments own figures the vaccination rate works out close to and often even exceeds 100% in the older age cohorts, which is nonsense, they underestimate the population and don’t subtract deaths or people leaving the country is my best guess at how they get such a high number.
Notice they don’t give the total number that died in aged care, they only give the total deaths for each period. I think the number of aged care deaths was too embarrassing to actually publish because they wanted to show how effective the vax was for aged care residents. It would be a bit of a stretch to say it was super effective at preventing deaths in aged care facilities, if almost all the deaths were actually in aged care facilities. The dead “foreigners” in aged care facilities would have been included in the whole population (65+) section of the study if they were included in the 10 August 2021 census which is highly likely as only 1,735 aged care residents weren’t matched in MADIP. The population included in the 65+ section of the study is N=3.9M which would roughly be made up of “aged care foreigners” (N=1.1M) + “aged care Aussies” (N=0.2M) + “non aged care residents” (N=2.6M). If you extrapolate the “Aussie aged care” death rate to the foreigners it’s not hard to imagine that the vast majority of deaths actually occurred in age care facilities. It’s probably even worse if you consider that according to the ABS people born overseas are much more likely to die from/with Covid. Then if you also consider the previously known problems with aged care in Australia it’s probably not something they’d want to draw attention to.
I actually think the whole study is mostly just an artifact of changing of categories, I believe they had a outcome in mind before they did the study and carefully chose the cohort sizes and timing to get the results they wanted to show. They say the timing was chosen to take into account the different variants, it’s close, but it looks to me that they chose the timing to maximize the effect of the rolling cohorts and the covid “waves”. There are too many moving parts in their analysis suggesting to me they either planned it that way or it’s just a bunch of non-sense if they haven’t accounted for it.
Thanks for having a look at it, I’m happy someone else has paid it some attention.
Also, thanks for posting the Austrian study for comparison, that’s how it should be done, even though I’m not a fan of survival analysis, it’s much more rigorous and transparent.
I fully agree with your paragraph on timing. It is frustrating to see that there are quite a few countries with complete data (Australia, New Zealand, Austria, Scotland, Singapore, ...), and we are only given carefully curated conclusions.
“Fascinating... so Australia is some kind of aged-care provider for the world (high standard of living, decent care, warm weather)?”
Seems like it based on the supplementary data. I was amazed by those figures. We do have reciprocal arrangements with a number of other countries with regards to free health care I don’t know if that also extends to aged care.
It might also just be a mistake, if they used criteria something like born overseas = “yes” & current visa status = ( “none” or “expired” ). Or if they used tax status = (NOT “resident” or “none”) as exclusion criteria it is likely legitimate permanent residents or even some Australians would have been excluded because permanent residents don’t need to have a current visa to stay in the country, as long as they don’t leave the country they only need to have a current visa to re-enter. If they’re in aged care they’re probably unlikely to be traveling so they probably don’t have a current visa as it would cost them about $300 every five years to renew it and isn’t required as long as they don’t leave the country. Technically, a permanent resident only gets one permanent residency visa the first time they enter the country which expires in 5 years, any new ones are called resident return visas (different classification number) this resident return visa is not needed unless they want to leave the country and return. Basically you can be a legitimate permanent resident and not have a current visa. If they based permanent residency on residential tax status then that could also be a source of error, if they are retired (highly likely) and they are on a low income (also likely) they don’t have submit a tax return so many would also not have a current tax record.
Maybe he's a buddy of one of the authors. Curiously, they constitute a team of 11: Alena Chalupka, Lukas Richter, Ali Chakeri, Ziad El-Khatib, Verena Theiler-Schwetz, Christian Trummer, Robert Krause, Peter Willeit, Bernhard Benka, John P. A. Ioannidis, Stefan Pilz
Spot on! 🎯
I’ve been trying to reverse engineer the Australian study for exactly the same reasons you mention (and some others).
Did you notice:
# remove unvaccinated as it is biased group
This suggests to me that the unvaccinated included in the analysis are somehow a rolling cohort of people that went on to become vaccinated, so the comparison in their analysis is probably not between the vaccinated and never the vaccinated which I had assumed. How you achieve that in a study where death is the end point beats me 🤷♂️
The second part of the study was an analysis of Aged Care residents. The total number of aged care residents over 65 in the database is about 1,250,000 but the number included in the Aged Care analysis is only about 175,000.
They seem to have excluded roughly 85% of the Aged Care database. So, who was excluded from the Aged Care analysis?
Most of the exclusions were non-permanent residents. The number excluded is HUGE, it’s just over 1,000,000 aged care residents. This is a remarkable number. It would suggest that the vast majority (roughly 80%) of Australia’s aged care residents are in fact “not Australians”.
The study also suggests the % vaccinated figures quoted by the Government are exaggerated it is very high at about 95% but not the almost 100% (or over 100% if you do the math) they claim.
Why do you think the total number of aged care residents should be as large as 1,250,000? That would be 5% of the population (25 million).
Why do you think the study claims 100% vaccination rate? 0.2 million (unvaccinated) of 3.9 million (in the study) is around 5%.
You are right, of course, that the fact of people changing categories (from unvaccinated to vaccinated, or from n doses to n+1 doses, or from 8-90 days to 90+ days without further vaccination) is a kind of selector (you can only change categories if you haven't died in your current category), and obscures such studies.
Thanks for the reply, I agree the change in categories obscures the result, though if properly done I think it can be account for in the survival analysis model, or you can see its effect in the residuals, or something else along those lines but I’m far from being an expert in this regard.
The total aged care population is shown in their supplementary material as 1,259,410. It’s their figure not mine. According to the United Nations; Australia has the third highest life expectancy in the world, Monaco and Japan have higher life expectancies than Australia, Australia's male life expectancy ranked second, and female sixth, in the world. So I’m not surprised we’ve got a lot of old people in aged care.
I’m not saying the study is claiming the vax rate is close to 100% the figure they calculate is roughly 95% (surprisingly it’s even slightly lower in the aged care population) which i think is probably about right. Other government departments like Health like to exaggerate the % vaxxed (particularly in the elderly) this study confirms it’s not as high as they advertise. I’m more inclined to believe the figures shown in the study than other sources even though I think their methods are a bit shit, their data is good. Other sources claim the rate its very close to 100% in the elderly but they don’t show the actual figures they just list it as >95% in their tables. However, they do give the raw data they use to calculate the tables and if you use the Heath Departments own figures the vaccination rate works out close to and often even exceeds 100% in the older age cohorts, which is nonsense, they underestimate the population and don’t subtract deaths or people leaving the country is my best guess at how they get such a high number.
Notice they don’t give the total number that died in aged care, they only give the total deaths for each period. I think the number of aged care deaths was too embarrassing to actually publish because they wanted to show how effective the vax was for aged care residents. It would be a bit of a stretch to say it was super effective at preventing deaths in aged care facilities, if almost all the deaths were actually in aged care facilities. The dead “foreigners” in aged care facilities would have been included in the whole population (65+) section of the study if they were included in the 10 August 2021 census which is highly likely as only 1,735 aged care residents weren’t matched in MADIP. The population included in the 65+ section of the study is N=3.9M which would roughly be made up of “aged care foreigners” (N=1.1M) + “aged care Aussies” (N=0.2M) + “non aged care residents” (N=2.6M). If you extrapolate the “Aussie aged care” death rate to the foreigners it’s not hard to imagine that the vast majority of deaths actually occurred in age care facilities. It’s probably even worse if you consider that according to the ABS people born overseas are much more likely to die from/with Covid. Then if you also consider the previously known problems with aged care in Australia it’s probably not something they’d want to draw attention to.
https://www.royalcommission.gov.au/aged-care/final-report
I actually think the whole study is mostly just an artifact of changing of categories, I believe they had a outcome in mind before they did the study and carefully chose the cohort sizes and timing to get the results they wanted to show. They say the timing was chosen to take into account the different variants, it’s close, but it looks to me that they chose the timing to maximize the effect of the rolling cohorts and the covid “waves”. There are too many moving parts in their analysis suggesting to me they either planned it that way or it’s just a bunch of non-sense if they haven’t accounted for it.
Thanks for having a look at it, I’m happy someone else has paid it some attention.
Also, thanks for posting the Austrian study for comparison, that’s how it should be done, even though I’m not a fan of survival analysis, it’s much more rigorous and transparent.
Fascinating... so Australia is some kind of aged-care provider for the world (high standard of living, decent care, warm weather)?
Vaccination rates are indeed meaningless. I wrote about this, considering the example of Germany, more than a year ago:
https://cm27874.substack.com/p/measurement-uncertainty
I fully agree with your paragraph on timing. It is frustrating to see that there are quite a few countries with complete data (Australia, New Zealand, Austria, Scotland, Singapore, ...), and we are only given carefully curated conclusions.
“Fascinating... so Australia is some kind of aged-care provider for the world (high standard of living, decent care, warm weather)?”
Seems like it based on the supplementary data. I was amazed by those figures. We do have reciprocal arrangements with a number of other countries with regards to free health care I don’t know if that also extends to aged care.
It might also just be a mistake, if they used criteria something like born overseas = “yes” & current visa status = ( “none” or “expired” ). Or if they used tax status = (NOT “resident” or “none”) as exclusion criteria it is likely legitimate permanent residents or even some Australians would have been excluded because permanent residents don’t need to have a current visa to stay in the country, as long as they don’t leave the country they only need to have a current visa to re-enter. If they’re in aged care they’re probably unlikely to be traveling so they probably don’t have a current visa as it would cost them about $300 every five years to renew it and isn’t required as long as they don’t leave the country. Technically, a permanent resident only gets one permanent residency visa the first time they enter the country which expires in 5 years, any new ones are called resident return visas (different classification number) this resident return visa is not needed unless they want to leave the country and return. Basically you can be a legitimate permanent resident and not have a current visa. If they based permanent residency on residential tax status then that could also be a source of error, if they are retired (highly likely) and they are on a low income (also likely) they don’t have submit a tax return so many would also not have a current tax record.
See my analysis of the Kristine Macartney paper in the Lancet.
https://geoffpain.substack.com/p/5-or-more-shots-and-you-are-out-say
John Ioannidis works for Austria?
Maybe he's a buddy of one of the authors. Curiously, they constitute a team of 11: Alena Chalupka, Lukas Richter, Ali Chakeri, Ziad El-Khatib, Verena Theiler-Schwetz, Christian Trummer, Robert Krause, Peter Willeit, Bernhard Benka, John P. A. Ioannidis, Stefan Pilz