You wrote: "I think you forgot to analyze the non-Covid period After the full Covid wave was over. There you will find that the vaccinated are dying at a higher rate than the rate they were dying before."
But that's because in mid-2021 there were still many recently vaccinated people who were impacted by the temporal HVE.
You wrote: "I think you forgot to analyze the non-Covid period After the full Covid wave was over. There you will find that the vaccinated are dying at a higher rate than the rate they were dying before."
But that's because in mid-2021 there were still many recently vaccinated people who were impacted by the temporal HVE.
My ASMR for people in the Czech dataset was about 969 in May 2020. In comparison the ASMR of vaccinated people was about 606 in May 2021 but about 900 in May 2022, so both were lower than the 2020 baseline, but 2021 was about 37% lower than the baseline because vaccinated people were heavily impacted by the temporal HVE in May 2021:
year vax asmr # ASMR per 100,000 person-years in May of each year
2020 FALSE 969
2021 TRUE 606
2021 FALSE 2081
2022 TRUE 900
2022 FALSE 1584
If you would've used the second half of 2022 as your low-COVID period then it wouldn't have been confounded as much by the temporal HVE. My code above shows the ratio between unvaccinated and vaccinated ASMR was about 3.4 in 2021 but only about 1.8 in 2022.
You wrote: "I think you forgot to analyze the non-Covid period After the full Covid wave was over. There you will find that the vaccinated are dying at a higher rate than the rate they were dying before."
But that's because in mid-2021 there were still many recently vaccinated people who were impacted by the temporal HVE.
My ASMR for people in the Czech dataset was about 969 in May 2020. In comparison the ASMR of vaccinated people was about 606 in May 2021 but about 900 in May 2022, so both were lower than the 2020 baseline, but 2021 was about 37% lower than the baseline because vaccinated people were heavily impacted by the temporal HVE in May 2021:
> cul=\(x,y)y[cut(x,c(y,Inf),,T,F)]
> ages=c(0,3:9*5,50:89,18:20*5)
> t=fread("https://sars2.net/f/czbucketsdaily.csv.gz")
> a=t[month(date)==5,.(pop=sum(alive),dead=sum(dead)),.(vax=dose>0,year=year(date),age=cul(age,ages))]
> a=merge(a,fread("https://sars2.net/f/czcensus2021pop.csv")[,sum(pop),.(age=cul(age,ages))][,.(age,std=V1/sum(V1))])
> a[,.(asmr=round(sum(dead/pop*std*365e5))),.(year,vax)]|>print(r=F)
year vax asmr # ASMR per 100,000 person-years in May of each year
2020 FALSE 969
2021 TRUE 606
2021 FALSE 2081
2022 TRUE 900
2022 FALSE 1584
If you would've used the second half of 2022 as your low-COVID period then it wouldn't have been confounded as much by the temporal HVE. My code above shows the ratio between unvaccinated and vaccinated ASMR was about 3.4 in 2021 but only about 1.8 in 2022.