The bumps in cases around April 2021 and January 2022 occurred around the same time in all age groups even though younger age groups got vaccinated later than older age groups: https://i.ibb.co/7tVGHY0Q/cdc-cases-per-age-group.png.
You can see all-cause mortality by vaccination status from the Medicare datasets that were given to Kirsch by his source at HHS, and you can see COVID deaths by vaccination status from the CDC dataset titled "Rates of COVID-19 Cases or Deaths by Age Group and Updated (Bivalent) Booster Status". However based on both sources, vacinated people had a much smaller increase in deaths than unvaccinated people during the first Omicron wave in January 2022: https://sars2.net/i/rootclaim-delta-3.png, https://sars2.net/i/rootclaim-cfr-2.png.
Age-stratified data exposes a major weakness of the Rancourtian approach of correlating spikes in deaths with spikes in vaccine rollout. In Peru there was a big wave of deaths around April 2021, so Rancourt blamed it on the rollout of the primary course doses. But the deaths occurred around the same time in all age groups even though younger age groups were vaccinated much later than older age groups: https://sars2.net/rootclaim.html#Rancourts_paper_about_excess_mortality_in_125_countries_3u. And similarly in the Czech Republic there was a spike in all-cause deaths around December 2021 which appears to coincide with the rollout of the first booster dose if you look at all ages aggregated together. But the deaths peaked around the same time in all age groups even though the number of booster doses administered peaked about a month before deaths in ages 80+ but about a month after deaths in ages 40-59 (see my previous link).
I agree that age-stratified analyses is the way to go and a next step, but combining the 3 variables for the same age strata and over the same time period is difficult, if not impossible with existing data (that I could find at least). Also, there are arguments for not restricting the vaccine administrations by age, for example, possible shedding effects as well as vaccine-induced fatal accidents (i.e. a driver suffers a vaccine-induced stroke or heart attack while operating the vehicle and kills other people of different ages) could be deaths in other age groups attributed to vaccinations in older ages. I like the graph in your second paragraph. State to state differences such as urbanicity/population density might help explain why the bump is missing in some states. In regard to the graph in your third paragraph, it wasn't clear to me what the y-axis represents in the grey vs. non-shaded region, and whether the plots adjusts/corrects for percent vaccinated across time? Isn't the blue line higher (and red/green lower) in the beginning because less people were vaccinated at that time? I haven't (yet) read Rancourt's work in detail, but based on the graphs it looks like he doesn't adjust/account for COVID waves in his calculations. Is this correct?
I made the plot in the third paragraph based on Medicare data which was missing population denominators, so in order to plot the lines for unvaccinated and vaccinated people on a similar scale, I plotted deaths as percentage of deaths during the period shaded in gray, which was a period I picked because it had low COVID mortality. That way the plot also ends up being adjusted for the healthy vaccinee effect where unvaccinated people have higher all-cause mortality than vaccinated people even during periods with a low number of deaths.
Yeah that's true the number of vaccinated people goes up over time. There's also some people who didn't have a vaccination record at Medicare until they got the first booster dose, which partially explains why the drop after the Omicron peak is much bigger than the rise up to the Omicron peak in unvaccinated people, which is because the unvaccinated population size included people who were in reality vaccinated but who had a vaccination record missing until the first booster dose. It's not ideal data, but it's one of the few available datasets for all-cause mortality by vaccination status in the United States. If you want to look at it yourself, I uploaded the spreadsheet I downloaded from Kirsch's S3 server here: https://sars2.net/f/COVID_vax_jan_feb_mar_2021_plus_ALL_deaths_in_medicare_per_day.xlsx.
The bumps in cases around April 2021 and January 2022 occurred around the same time in all age groups even though younger age groups got vaccinated later than older age groups: https://i.ibb.co/7tVGHY0Q/cdc-cases-per-age-group.png.
The bump in cases around April 2021 is also missing from many states, even though the rollout of the primary course doses peaked around March or April in all states: https://i.ibb.co/DftmT42L/us-states-cases-deaths-vaccines.png.
You can see all-cause mortality by vaccination status from the Medicare datasets that were given to Kirsch by his source at HHS, and you can see COVID deaths by vaccination status from the CDC dataset titled "Rates of COVID-19 Cases or Deaths by Age Group and Updated (Bivalent) Booster Status". However based on both sources, vacinated people had a much smaller increase in deaths than unvaccinated people during the first Omicron wave in January 2022: https://sars2.net/i/rootclaim-delta-3.png, https://sars2.net/i/rootclaim-cfr-2.png.
Age-stratified data exposes a major weakness of the Rancourtian approach of correlating spikes in deaths with spikes in vaccine rollout. In Peru there was a big wave of deaths around April 2021, so Rancourt blamed it on the rollout of the primary course doses. But the deaths occurred around the same time in all age groups even though younger age groups were vaccinated much later than older age groups: https://sars2.net/rootclaim.html#Rancourts_paper_about_excess_mortality_in_125_countries_3u. And similarly in the Czech Republic there was a spike in all-cause deaths around December 2021 which appears to coincide with the rollout of the first booster dose if you look at all ages aggregated together. But the deaths peaked around the same time in all age groups even though the number of booster doses administered peaked about a month before deaths in ages 80+ but about a month after deaths in ages 40-59 (see my previous link).
I agree that age-stratified analyses is the way to go and a next step, but combining the 3 variables for the same age strata and over the same time period is difficult, if not impossible with existing data (that I could find at least). Also, there are arguments for not restricting the vaccine administrations by age, for example, possible shedding effects as well as vaccine-induced fatal accidents (i.e. a driver suffers a vaccine-induced stroke or heart attack while operating the vehicle and kills other people of different ages) could be deaths in other age groups attributed to vaccinations in older ages. I like the graph in your second paragraph. State to state differences such as urbanicity/population density might help explain why the bump is missing in some states. In regard to the graph in your third paragraph, it wasn't clear to me what the y-axis represents in the grey vs. non-shaded region, and whether the plots adjusts/corrects for percent vaccinated across time? Isn't the blue line higher (and red/green lower) in the beginning because less people were vaccinated at that time? I haven't (yet) read Rancourt's work in detail, but based on the graphs it looks like he doesn't adjust/account for COVID waves in his calculations. Is this correct?
I made the plot in the third paragraph based on Medicare data which was missing population denominators, so in order to plot the lines for unvaccinated and vaccinated people on a similar scale, I plotted deaths as percentage of deaths during the period shaded in gray, which was a period I picked because it had low COVID mortality. That way the plot also ends up being adjusted for the healthy vaccinee effect where unvaccinated people have higher all-cause mortality than vaccinated people even during periods with a low number of deaths.
Yeah that's true the number of vaccinated people goes up over time. There's also some people who didn't have a vaccination record at Medicare until they got the first booster dose, which partially explains why the drop after the Omicron peak is much bigger than the rise up to the Omicron peak in unvaccinated people, which is because the unvaccinated population size included people who were in reality vaccinated but who had a vaccination record missing until the first booster dose. It's not ideal data, but it's one of the few available datasets for all-cause mortality by vaccination status in the United States. If you want to look at it yourself, I uploaded the spreadsheet I downloaded from Kirsch's S3 server here: https://sars2.net/f/COVID_vax_jan_feb_mar_2021_plus_ALL_deaths_in_medicare_per_day.xlsx.