COVID Deaths are an Undercount

[_Icarus]

Georgia is currently adopting this technique and I wonder how many other states are doing it too. The effect is that we get a rosier picture of the curve bending over the past 14 days when that isn't even based on all reported cases.

Here's what Georgia's COVID-19 case curve looks like as of May 11

Icarus, that's not a conspiracy. That's standard epidemiology. Epidemic curves are displayed by onset of disease because it gives a more accurate picture of the rate at which the disease spread through a population. It should always be displayed with a shaded area at the right hand end, to call attention to the days for which data is likely incomplete. When we talk about flattening the curve, this is the curve that we're trying to flatten.

Ummm, I didn't say it was a conspiracy.

Just pointing out that the recent "downward" curve - which states are using to justify a reopening -is skewed by the method explained above. As far as it being "standard," that doesn't jive with Georgia only recently adopting this method in late April.

[_Res Ipsa]

Oh, sorry Icarus, I thought you were saying that people in the Georgia state government were manipulating numbers to justify reopening.

Listing cases by reporting date skews the data because the results are affected by changes in the rate of testing. Listing by date of infection gives gives a more accurate picture of how the dose actually spread. It is the information states should be basing their opening decisions on. The problem is, of course, when states are are spoiling, the data won’t show it for a week or two, so it’sa good idea to watch new cases by date of report.

I was not aware that basic epidemiology was defined as whatever Georgia does. I was thinking it was defined by epidemiologists. https://nciph.sph.unc.edu/focus/vol1/is ... _issue.pdf Thanks for setting me straight.

[_Icarus]

Listing by date of infection gives gives a more accurate picture of how the dose actually spread.

But when dealing with asymptomatic people there is no telling when the infection began.

It is the information states should be basing their opening decisions on

Are you sure that's what all states are doing?

I was not aware that basic epidemiology was defined as whatever Georgia does. I was thinking it was defined by epidemiologists. https://nciph.sph.unc.edu/focus/vol1/is ... _issue.pdf Thanks for setting me straight.

Wasn't my intention, as you were nowhere mentioned in my post and the point is how states are reporting, not what epidemiological preference might be.

Having said that, Georgia isn't exactly South Dakota where epidemiologists are scarce and have little influence. After all, Georgia is home to the CDC and Emory. The point is Georgia has only been doing this was since late April, so why did it suddenly change course and only right after Trump announced a three phase plan to "Reopen America"?

And after reading this piece by the AJC, I may change my mind on the conspiracy thing.

‘It’s just cuckoo’: state’s latest data mishap causes critics to cry foul

[_DoubtingThomas]

And if you have heart disease and COVID, that counts as a COVID death right?

In April 2018 there were 233,000 deaths (all causes) in the US. There were about 50,000 covid deaths last month, about 1/5 deaths. The total number of deaths last month is probably around 270 to 300 thousand.
https://www.cdc.gov/nchs/nvss/vsrr/prov ... tables.htm

[Guest]

.

[_DoubtingThomas]

And if you have heart disease and COVID, that counts as a COVID death right?

The only thing we can be certain of is that covid is much deadlier than the seasonal flu.

[_DoubtingThomas]

I hope too that the politicians will stay out of this.

When do you think Las Vegas is going back to normal? I do plan to go so you can show me around.

[_Res Ipsa]

Icarus, at the beginning of an epidemic, an epidemic curve is not as helpful because of the information lag. But for finding the actual peak of infections, it’sa much more accurate tool. Washington has posted an epidemic curve on its health department page for weeks now. Sometimes an initial report is missing that information and health departments will get it through interviews. I believe asymptomatic or presentation cases are added by date of report. That’s generally explained in a footnote. For an example: http://www.bccdc.ca/resource-gallery/Do ... 141105.pdf

[_EAllusion]

Vox has a piece up on the Trump admin's emerging efforts to manipulate the COVID mortality stats lower when they already are likely quite conservative:

https://www.vox.com/2020/5/14/21257247/ ... eath-stats

On a related note, one of the reasons that democracies are better than authoritarian states is that democracies can't fake information regarding deaths of its citizens which, when combined with democratic accountability, leads to fewer famines and so forth. The free flow of information that exists in a liberal democracy prevents that. "Sure you can," you might say, "Look at what the US is trying to do."

*taps nose*

[_Some Schmo]

The right is really good at creating a unified message and then having their little foot soldiers spread that message.

Well, sure. That's what happens when nobody in your base has a will, backbone, or intellectual life of their own.

I swear, every Trump supporter who takes any effort to defend him has to check the Internet every morning to find out how to argue that day's Trump “F”-ups.