TennTradition
Defended.
- Joined
- Aug 14, 2006
- Messages
- 16,919
- Likes
- 822
I can somewhat reproduce these results (but still not quite) if I use the CDC's ILI data, but stop at April 18 instead of April 25th.
However, the CDC warns that their modeling that produces "total deaths" numbers in the ILI report is currently suspect because of how CV is affecting the data.
They instead are tracking these differently in the CLI (covid-like illness report). A very interesting plot that they include in this is the "Excess Death" plot. You can go view it here for different regions:
Excess Deaths Associated with COVID-19
The key issue is that it takes roughly 8 weeks for CDC to get death certificates. They trickle in initially and usually the have a good count within a few weeks but they can tail in within 8 weeks. So, all current data is estimated from models that have to be tuned to new data, so there is error. They explicitly say, even on the CLI report, that "Provisional counts are weighted to account for potential underreporting in the most recent weeks. However, data for the most recent week(s) are still likely to be incomplete."
If you look at actual provisional counts, for example (without this extra weighting), from here:
Provisional Death Counts for Coronavirus Disease (COVID-19):
, then can see provisional death data from 2/1/20 through 4/25/20. New York is broken down in the state and city (the state excludes the city). Adding those together, you get 16,345 CV deaths vs. 63,395 total deaths. Pulling CDC Wonder data, you would normally expect 39,651 deaths in NY over these three months (2014-2018 average). Also, this data isn't fully populated yet for 2020 due to the delays, so they'll keep coming up. Also, we are missing the last week of April entirely in this data. So it is clear that total deaths are way above average over this period, and almost all of this is in April - which was just a horrible month. Also, this tells us that the plot referenced above cannot be correct.
For example, if you just look at the last four weeks of data (weeks ending April 4 through April 25) - the provisional death data (this doesn't include all the death certificates yet) is 33,027 deaths in NY state, including NY city. If you look at CDC Wonder data, there would typically be around 13,000 deaths in April. Of those 33,027 death certificates obtained by CDC for those weeks, 15,079 listed COVID.
The data is all there in the links above. I would used the CIL data. I would definitely not take COVID tracking data, combine it with ILI data, and subtract out the difference to calculate "other deaths", which is what this plot did.
