The P&R Plague Thread (Covid-19)

[deedsker]

45 minutes ago, RedDenver said:

Note: this is just my own estimates based on very simple assumptions.

 

Someone mentioned that the flu caused 80,000 deaths two years ago, which got me thinking about how long until covid-19 hits those kinds of numbers given it's exponential increase. The current doubling time for deaths in the US is 3 days and there were 553 deaths as of yesterday. So the doubling time is:

80000=553*2^x => x = log2(80000/553) = 7.2 => 21.6 days to reach 80,000 deaths in the US

Fun math time:

 

WHO says it takes on average 5 days to develop symptoms.

Quote

This is because it takes between 2 and 10 days before people who are infected become sick and develop a fever.

 

That means we can reasonably assume that it takes at least 7 days for someone to become infected before dying. As of March 23rd, we had 553 deaths. 7 days before that, on March 16, we had 4663 case confirmed in the US. That means that the fatality rate in the US is something like 553/4663=> 11.85%

 

Now I know you say, we don't have testing, so it is impossible to know how many cases there are. If we say that even half of the cases are mild, then a week ago the US already had about 10,000 cases.

 

Or is you build from the back and say that this is just a product of bad testing, if 1.5% of cases end up in a fatality it takes approximately 67 cases to get a fatality. This means that on March 16th, there easily could have been about 37,000 cases. This was all before restrictions were put into place anywhere and even with full lock down now could be exposing their households to the virus.

 

That being said, we either aren't catching almost anyone who is sick or the death rate is much higher in the US. Both prospects are very concerning, either way.

[RedDenver]

4 minutes ago, deedsker said:

Fun math time:

 

WHO says it takes on average 5 days to develop symptoms.

 

That means we can reasonably assume that it takes at least 7 days for someone to become infected before dying. As of March 23rd, we had 553 deaths. 7 days before that, on March 16, we had 4663 case confirmed in the US. That means that the fatality rate in the US is something like 553/4663=> 11.85%

 

Now I know you say, we don't have testing, so it is impossible to know how many cases there are. If we say that even half of the cases are mild, then a week ago the US already had about 10,000 cases.

 

Or is you build from the back and say that this is just a product of bad testing, if 1.5% of cases end up in a fatality it takes approximately 67 cases to get a fatality. This means that on March 16th, there easily could have been about 37,000 cases. This was all before restrictions were put into place anywhere and even with full lock down now could be exposing their households to the virus.

 

That being said, we either aren't catching almost anyone who is sick or the death rate is much higher in the US. Both prospects are very concerning, either way.

FYI, data from Wuhan and Italy estimates the time between exposure and death is about 13 days. I linked an article way back in this thread that shows those numbers, but a quick search didn't find it.

 

17 minutes ago, DevoHusker said:

that's in a worse case scenario, and still .02%

 

not saying acceptable or preferred, just pointing it out 

My estimate isn't best, worst, or average case nor is it informed by any knowledge of pandemics. It's just a very simple extrapolation based on the current data. But I'm not basing those numbers on the mortality rate (or any other percentage rate) - it's just the rate at which people are dying from covid-19 in the US.

[Nebfanatic]

20 minutes ago, DevoHusker said:

 

that's in a worse case scenario, and still .02%

 

not saying acceptable or preferred, just pointing it out 

That isn't a worst case scenario and that is 3 weeks from now. In a year it will be much higher. Also you don't calculate fatality rate based on the entire population 

[ColoradoHusk]

10 minutes ago, jaws said:

 

Some colleges will do anything to pay the bills. 

Jerry Falwell and Liberty University isn't too far from The Righteous Gemstones.

[TGHusker]

45 minutes ago, schriznoeder said:

And so it begins... 

 

 

The Dem Gov of Virginia better shut this down quick.  Another blind kool aid drinking trump cult follower putting people in danger. 

[FrantzHardySwag]

1 minute ago, Nebfanatic said:

That isn't a worst case scenario and that is 3 weeks from now. In a year it will be much higher. Also you don't calculate fatality rate based on the entire population 

I’ve seen a number of weird fatality rate calculations on this site. CFR = Fatalities/Cases. IFR = fatalities/infected. Infected Fatality Rate is impossible to calculate right now because we have no clue how wide spread it is. CFR is flawed as well (arguments can be made both ways), but sits between 1%-2% for the US. 

[Nebfanatic]

Just now, FrantzHardySwag said:

I’ve seen a number of weird fatality rate calculations on this site. CFR = Fatalities/Cases. IFR = fatalities/infected. Infected Fatality Rate is impossible to calculate right now because we have no clue how wide spread it is. CFR is flawed as well (arguments can be made both ways), but sits between 1%-2% for the US. 

It's all going to be flawed until we get past this. Realistically you should only use closed cases in determining CFR because we don't know the outcome of active cases yet, but that too is flawed. 

[RedDenver]

3 minutes ago, FrantzHardySwag said:

I’ve seen a number of weird fatality rate calculations on this site. CFR = Fatalities/Cases. IFR = fatalities/infected. Infected Fatality Rate is impossible to calculate right now because we have no clue how wide spread it is. CFR is flawed as well (arguments can be made both ways), but sits between 1%-2% for the US. 

 

Just now, Nebfanatic said:

It's all going to be flawed until we get past this. Realistically you should only use closed cases in determining CFR because we don't know the outcome of active cases yet, but that too is flawed. 

This is why I didn't use any mortality rate values in my estimate. I think we can be reasonably certain the death count is accurate though.

[FrantzHardySwag]

Just now, Nebfanatic said:

It's all going to be flawed until we get past this. Realistically you should only use closed cases in determining CFR because we don't know the outcome of active cases yet, but that too is flawed. 

Yep, when you’re only testing high risk patients with underlying conditions over a certain age, that’s a selection bias. But at the same time many cases now, could take a turn for the worst in the coming week. Nate Silver said it best, looking at the data now is like looking at funhouse mirrors. 

[TGHusker]

There are still a lot of American's trying to get back to the country.  I didn't realize it was still such a large #.

 

https://thehill.com/policy/international/489143-state-department-under-fire-as-americans-remain-stranded-abroad

Quote

 

The State Department is facing the unprecedented challenge of trying to bring home thousands of Americans stranded around the world, as countries shutter their borders to stop the rapid spread of the coronavirus.

The U.S. is tracking some 13,500 citizens abroad who need help getting home, according to a senior State Department official.

But despite nearly a dozen successful evacuations of tens of thousands of U.S. citizens from conflict zones and humanitarian disasters over the years, their current efforts are coming under fire, as other countries have already evacuated hundreds, if not thousands, of their citizens in the same situations.

 

 

[Nebfanatic]

1 minute ago, RedDenver said:

 

This is why I didn't use any mortality rate values in my estimate. I think we can be reasonably certain the death count is accurate though.

Right and the rate at which deaths are occurring is also an accurate value so it is easier to base statistical models off of these values.

[deedsker]

29 minutes ago, RedDenver said:

FYI, data from Wuhan and Italy estimates the time between exposure and death is about 13 days. I linked an article way back in this thread that shows those numbers, but a quick search didn't find it.

 

That is why I said at least. The numbers only get scarier the less conservative you get with data out of China.

[knapplc]

I think you folks running numbers are doing interesting work, but I think we're all going to be surprised how this shakes out when we get better numbers months from now.

[knapplc]

When your grandpa who can't set up the Blu-Ray player you got him for Christmas is in charge of a pandemic response:

 

 

[FrantzHardySwag]

5 minutes ago, knapplc said:

I think you folks running numbers are doing interesting work, but I think we're all going to be surprised how this shakes out when we get better numbers months from now.

Best numbers will be a antibody test, some have beat the virus and don’t even know it, they are carrying immunity. In a related story New York is testing the use of plasma from recovered to treat it, that’s a big step. Science will beat this, we just need to buy science time.