And so a source THAT I've depended on maybe grossly biased and WRONG!!!
Date-of-death versus date-of-report
At issue is the way in which the websites record deaths. They do so largely by pulling from state death totals that are presented
by "date-of-report" rather than "date-of-death."
The distinction, at least during a pandemic, is a critical one. Many local health authorities have been overwhelmed at times with high numbers of coronavirus deaths over short time periods, which — coupled with standard administrative delays — can mean that COVID-19 deaths take days,
weeks or even longer to reach official state logs and from there data tracking sites.
Yet the three major COVID tracking sites all tabulate deaths by date of report, without any option to see COVID-19 fatalities listed by when they actually occurred.
This has led to all three sites graphing recent weeks of fatalities as if they were increasing — but more timely data from the CDC suggests
that week-over-week numbers of provisional deaths have actually been falling.
At the COVID Tracking Project, for instance, the data show the three-week period
with weeks ending on Oct. 10, Oct. 17 and Oct. 24 had deaths increasing 4,420, 4,396, and 5,262, respectively.
Worldometers listed those weeks as 4,755, 4,603, and 5,282.
Johns Hopkins, meanwhile, pegged the deaths at 4,600, 4,700 and 5,300.
Yet provisional CDC data paints a starkly different picture:
COVID-19 fatalities actually appear to be continuing a steady downward trend they've been on for months.
That data show that weekly COVID deaths in the U.S. appear to have peaked in early August and have been declining by hundreds
of deaths per week every week since. Of the most recent three-week period that appears to show a rise in deaths per the popular websites, the CDC posts numbers that are both starkly lower and trending in the opposite direction from those listed by the other sites:
from 3,635 at Oct. 10 to 2,841 the following week and finishing at 1,119 at Oct. 24.
Major COVID-19 data websites record deaths largely by pulling from state death totals that are presented by "date-of-report" rather than "date-of-death."
justthenews.com