Deaths in weekly-breakthrough.csv vs data-by-day.csv
HarryLeroy opened this issue · comments
I’ve noticed that the total number of vaxed and unvaxed deaths in weekly-breakthrough.csv frequently exceed the total number of deaths in data-by-day.csv. This doesn’t make a lot of sense—the population in the breakthrough database does not even include partially vaccinated or fully vaccinated <2 weeks ago. Both data-by-day and weekly-breakthrough appear to be based on event date, so the numbers should mostly line up. But sometimes they don’t, most notably on deaths.
By way of example, the total number of deaths for the week of 12/25/2021 are listed as 320 in weekly-breakthrough, but data-by-day is at 174 deaths (confirmed deaths from the full week preceding the week ending date, i.e., 12/19/21 – 12/25/21). I don’t know if the breakthrough database includes probable deaths (I assume it does not), but there are only 15 probable deaths during that time, so that’s not the source of the discrepancy.
What’s the reason for the difference?
Thanks,
HL
Harry, I noticed the same thing. I have never been able to reconcile the # of deaths by day with the deaths by week. I've just been assuming the delta was partially vaccinated deaths.
They need to just list all the deaths by vax status (Unvax/Partial/Fully) and also give the crude rates instead of the mangled age-adjusted rates. Or just give us the pop denominators so we don't have to reverse engineer it all.
No answer on this, huh?
Hi @HarryLeroy, thanks for noticing this. This comes down to a small difference in how dates are assigned for these different data. For the daily counts, deaths are aggregated by date of death (per documentation).
For the vaccine breakthrough data (rates by vaccination status), deaths are aggregated by date of diagnosis (documentation). This aligns with how the CDC reports case and deaths by vaccination status.
Hi @HarryLeroy, thanks for noticing this. This comes down to a small difference in how dates are assigned for these different data. For the daily counts, deaths are aggregated by date of death (per documentation).
For the vaccine breakthrough data (rates by vaccination status), deaths are aggregated by date of diagnosis (documentation). This aligns with how the CDC reports case and deaths by vaccination status.
Very valuable information, thank you.