COVID Deaths are an Undercount

[_Icarus]

And then we find out that 15% of all negative tests could be false negatives.

https://www.wpri.com/health/coronavirus ... sed-in-ri/

[_Icarus]

So this was published today.

https://www.washingtonpost.com/investig ... rc404=true

According to available data, the official COVID death toll is a severe under-count based on all-cause mortality. We're well above average, and the official COVID death toll only accounts for roughly two thirds of the difference. In the five states of New York, New Jersey, Washington, Michigan and Maryland there are 12,194 excessive deaths unaccounted for over a three week period between mid-March and April 4. That's just between four states and over 3 weeks, well before the peak. It is also interesting that only a handful of states are reporting deaths as COVID which are "highly probable." The CDC requested all states to do this in early April but only a few Blue states have complied. Unsurprisingly, Red-states like Alabama are counting 90% of its COVID deaths as people dying "with" it instead of from it. One can only assume the intent is to keep numbers artificially low to help their cause in November. Since there isn't enough testing, many deaths we're finding out about longer after when a family pays for an autopsy. For instance, the earliest COVID death in the USA is now known to have been weeks earlier than previously thought.

That "excess deaths" number is even higher than it looks. Because of social distancing and people staying at home more, the numbers of deaths that would normally occur due to some causes (e.g., car accidents) have actually dropped.

[_Some Schmo]

It's hilarious that the right accuses Democrats of wanting more deaths, when the GOP seems to be doing everything in their power to cause as many deaths as possible.

If it weren't for hypocrisy, the GOP would have nothing to do.

[_Icarus]

https://www.propublica.org/article/ther ... n-reported

There’s Been a Spike in People Dying at Home in Several Cities. That Suggests Coronavirus Deaths Are Higher Than Reported.

[_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

Recently, the state started backdating cases, graphing them based not on when they're reported, but when symptoms first began.

So, for example, if you get diagnosed with COVID-19 today but your symptoms started 10 days ago, the state will graph your case 10 days ago.

With that in mind, the state warns the data in the most recent 14-day window (in the orange zone on our graph) is subject to change.

That's because new cases could be reported in the days to come that would get backdated to 14-day period.

There are some who say we shouldn't even be showing the 14-day window since the data is likely incomplete.

We want to give you all the information that's out there, and since the Georgia Department of Public Health itself is reporting it, we're including it too with the caveat that the data in the orange 14-day window should be viewed as incomplete.

Everything in the "blue zone" outside the 14-day period is more concrete.

That data shows the state had a long period of case growth from mid-March to mid-April.

Then, things flattened out a bit, with a 7-day average of around 700 new cases a day.

Then, there was another bump up closer to an average of 750 new cases per day and then a slight drop from about April 20 to the start of the 14-day window.

We'll see how things go in the coming days as the orange zone data gets solidified.

That data will help us better understand if that dip is part of a trend that will continue or is just an outlier.

[_ajax18]

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

[_Analytics]

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

That is one of the advantages of the "excess deaths" approach. People had heart disease before Covid, and some of them died from it. The excess death count captures the number of people who died under the current situation who wouldn't have died otherwise, regardless of whether or not they have heart disease.

[_Res Ipsa]

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

Recently, the state started backdating cases, graphing them based not on when they're reported, but when symptoms first began.

So, for example, if you get diagnosed with COVID-19 today but your symptoms started 10 days ago, the state will graph your case 10 days ago.

With that in mind, the state warns the data in the most recent 14-day window (in the orange zone on our graph) is subject to change.

That's because new cases could be reported in the days to come that would get backdated to 14-day period.

There are some who say we shouldn't even be showing the 14-day window since the data is likely incomplete.

We want to give you all the information that's out there, and since the Georgia Department of Public Health itself is reporting it, we're including it too with the caveat that the data in the orange 14-day window should be viewed as incomplete.

Everything in the "blue zone" outside the 14-day period is more concrete.

That data shows the state had a long period of case growth from mid-March to mid-April.

Then, things flattened out a bit, with a 7-day average of around 700 new cases a day.

Then, there was another bump up closer to an average of 750 new cases per day and then a slight drop from about April 20 to the start of the 14-day window.

We'll see how things go in the coming days as the orange zone data gets solidified.

That data will help us better understand if that dip is part of a trend that will continue or is just an outlier.

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.

[Guest]

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

That is one of the advantages of the "excess deaths" approach. People had heart disease before Covid, and some of them died from it. The excess death count captures the number of people who died under the current situation who wouldn't have died otherwise, regardless of whether or not they have heart disease.

It's funny to me when people try to draw comparisons between the H1N1 epidemic and COVID-19. The final figures for H1N1 are based on the number of cases confirmed by tests alone. They use modeling with inputs like excess deaths to determine what's called the total disease burden. When the COVID-19 epidemic ends, the CDC will do the same analysis and, if the politicians will stay the hell out of the process, the total disease burden from COVID will be much higher than however we choose to count deaths.

ETA: Because that variant of H1N1 continues to circulate, the CDC continues to analyze the total disease burden. https://www.cdc.gov/flu/pandemic-resour ... -h1n1.html They do the same for seasonal flu every year. Most people who compare annual flu cases and deaths to COVID-19 don't understand that they're comparing an estimate based on models (flu) with positive tests plus some diagnoses based on symptoms. Here is the 2018-19 seasonal flu burden estimate. https://www.cdc.gov/flu/about/burden/2018-2019.html

Note that the CDC is using exactly the same terminology with COVID that it does with the flu: deaths "associated with" The CDC is, and should continue to use, the same methodology it uses for measuring the impact of disease -- not change it because the Trump administration doesn't like the numbers.

[_Dr Exiled]

I hope too that the politicians will stay out of this. I also hope that other entities don't try to skew the data one way or another in hopes of some benefit. But, too many follow the adage of not wanting to let a crisis go to waste. In any event, I think it is obvious that the covid-19 death count will be significant. People are dying from it and my circle of friends have had relatives pass away from it.