Covid-19 Data, Models, References - NO DISCUSSION

[ImBillT]

I’m not anywhere near assuming that 968 deaths in the US today is a peak, but I expected A LOT more today considering yesterday’s numbers. Hopefully that’s a sign of treatment success or social distancing working. Since social distancing can’t go on forever, and even if it could, it’s probably just delaying the inevitable, I really hope it’s a sign of viable treatment.

Edit: clarified my first sentence.

 

[huronmtns]

I’m not anywhere near assuming that 968 is a peak, but I expected A LOT more today considering yesterday’s numbers. Hopefully that’s a sign of treatment success or social distancing working. Since social distancing can’t go on forever, and even if it could, it’s probably just delaying the inevitable, I really hope it’s a sign of viable treatment.

Peak of what?

 

[ImBillT]

Peak of what?

Deaths/day in the US. Two days close to each other could easily be interpreted as some sort of leveling out. I think we’re way to early in our path for that, but it was a lot less than a 20%-40% increase over yesterday, so maybe it means something positive happened.

 

[Dave N]

They are using a different treatment, it’s called the kalashnikov cure. 7.62 mm injection..

Must have gotten that treatment from North Korea...

 

[VikingsGuy]

Italy’s Coronavirus Death Toll Is Far Higher Than Reported

Italy is undercounting thousands of deaths caused by the virus in the areas worst hit by the pandemic, a WSJ analysis shows, indicating the human toll may end up being much greater than official data indicate.

www.wsj.com

 

[huronmtns]

Deaths/day in the US. Two days close to each other could easily be interpreted as some sort of leveling out. I think we’re way to early in our path for that, but it was a lot less than a 20%-40% increase over yesterday, so maybe it means something positive happened.

I'm guessing this is a wishful thinking post, since there is some contradiction. I'm with you hoping this is the beginning of positive if that is what your intent was.

 

[VikingsGuy]

Article about variations in covid models

http://strib.mn/2UBhwOK

 

[LWC55]

Coronavirus: How to understand the death toll

The rising number of coronavirus deaths is distressing. But what are the figures actually telling us?

www.bbc.com

 

[MTGomer]

They are using a different treatment, it’s called the kalashnikov cure. 7.62 mm injection..

While Americans worry about residual traces of lead in gut piles affecting condors, look what Russia has been able to accomplish with much heavier doses to humans. We are a weak, low energy society.

 

[SAJ-99]

If the test isn’t practically perfect then it’s going to be letting sick people out in the community which completely defeats the purpose, and at what cost? In another thread someone said that an accurate flu test cost about $800. What’s another $200B? Are you only going to let someone go to work if they’ve had it and recovered? If you haven’t caught it yet, and someone with a false negative is walking around out there, then you still have a high chance of catching it...but you tested negative....are we gonna retest everyone every day? What good is it to know that I tested negative last week. I could still have it and spread it today.

Already happening. The rapid test for these things isn't very accurate. For the flu, let's call it 60%, and I suspect it is about the same for CV-19 but I can't find the data anywhere. One HC worker case I heard - shows mild symptoms test negative, two days later still shows symptoms but test negative again, and then two days after that finally tests positive. The way that testing system works, most errors are false negatives. It is hard to get a false positive. Testing is the key to get an idea of how prevalent it is, i.e. where you are in on the case curve. This is why we can model the flu, albeit with a huge range, because we have years of "runs"/seasons. In this case we are changing with the curve. The other test is different and has a different purpose. To Vikingsguys post, if you could know herd immunity, you could start to loosen restrictions.

 

[ImBillT]

Already happening. The rapid test for these things isn't very accurate. For the flu, let's call it 60%, and I suspect it is about the same for CV-19 but I can't find the data anywhere. One HC worker case I heard - shows mild symptoms test negative, two days later still shows symptoms but test negative again, and then two days after that finally tests positive. The way that testing system works, most errors are false negatives. It is hard to get a false positive. Testing is the key to get an idea of how prevalent it is, i.e. where you are in on the case curve. This is why we can model the flu, albeit with a huge range, because we have years of "runs"/seasons. In this case we are changing with the curve. The other test is different and has a different purpose. To Vikingsguys post, if you could know herd immunity, you could start to loosen restrictions.

If you get that many false negatives it does nothing useful.

You can’t get good herd immunity with the restrictions. If you want herd immunity, you’re going to have to let young healthy people get back in public. They just can’t go home to elderly or sick family members that are at much higher risk.

It would be nice to build that herd immunity with the MILD form. That means letting those with mild symptoms out, and keeping those with severe symptoms in. If someone is completely asymptomatic, but tests positive, separating them from society may actually do more harm than good.

 

[ImBillT]

Coronavirus: How to understand the death toll

The rising number of coronavirus deaths is distressing. But what are the figures actually telling us?

www.bbc.com

View attachment 134405

Now the question is, how much is additive and how much is compensatory. If it’s completely additive, then you stack the curves. If it’s completely compensator, then you overlap them as they are pictured. It is certainly some combination of the two.

 

[SAJ-99]

If you get that many false negatives it does nothing useful.

Don't test, just guess? I see a few problems with that approach. I agree that it doesn't solve the problem, but more knowledge is better than less. After all, we run these tests for the flu every year but the millions. Rarely changes treatment, but knowing is better than guessing. Given we are completely inundated by information every second of every day, at the very least knowledge is useful to sort through the constant BS from the administration and various "media".

 

[VikingsGuy]

Don't test, just guess? I see a few problems with that approach. I agree that it doesn't solve the problem, but more knowledge is better than less. After all, we run these tests for the flu every year but the millions. Rarely changes treatment, but knowing is better than guessing. Given we are completely inundated by information every second of every day, at the very least knowledge is useful to sort through the constant BS from the administration and various "media".

Yup - in large sets folks would be surprised at what can be gleaned from "noisy data"

 

[ImBillT]

Don't test, just guess? I see a few problems with that approach. I agree that it doesn't solve the problem, but more knowledge is better than less. After all, we run these tests for the flu every year but the millions. Rarely changes treatment, but knowing is better than guessing. Given we are completely inundated by information every second of every day, at the very least knowledge is useful to sort through the constant BS from the administration and various "media".

If the test is as bad as you claim, it’s just a guess...a guess that costs money. So yes, just guess. Also, what are solving with the test? If it’s that bad, it’s not giving you “more knowledge”.

Think about it, five guys in a work truck get tested. Two test positive, one gets a false negative, the next day he infects the other two guys, plus the two new guys. What on earth do you think you accomplished with the test?

 

[ImBillT]

Yup - in large sets folks would be surprised at what can be gleaned from "noisy data"

Yes you can get information from noisy data, but you can’t use it to decide who goes to work and who stays home.

 

[SAJ-99]

If the test is as bad as you claim, it’s just a guess...a guess that costs money. So yes, just guess. Also, what are solving with the test? If it’s that bad, it’s not giving you “more knowledge”.

...says the guy who keeps looking at the Cases and Deaths curves.

 

[VikingsGuy]

Yes you can get information from noisy data, but you can’t use it to decide who goes to work and who stays home.

Lot's of reasons to test.

 

[ImBillT]

...says the guy who keeps looking at the Cases and Deaths curves.

You saying some of those people ain’t actually dead?


You could literally randomly sample 1% of the population of a county or city or state each week and get an EXCELLENT idea of whether or not cases were increasing or decreasing, even with a really bad test. That alone will tell you about herd immunity. Declaring an individual safe or unsafe on the other hand requires perfection AND it requires EVERY individual, AND it requires retesting all the time if that person hasn’t already been infected. There are valuable uses for the tests, inaccurate tests, but deciding who stays and who goes isn’t one, and testing EVERYONE is completely unnecessary.

 

[ImBillT]

Lot's of reasons to test.

I don’t disagree that there are some reasons, and don’t think we should stop, BUT I don’t believe that it’s as huge as it’s being touted to be. If it isn’t nearly perfect, it won’t STOP the spread. On the other hand you can sample very few people and get an idea of whether the rate of new cases is increasing or decreasing, and even without a post infection test, you can asses herd immunity simply by noting that very few new infections are occurring. You can also asses herd immunity with a post infection test, but there’s no need to test everyone. If X percent of people in your representative sample show immunity, then that’s close to the actually percent of people with immunity. Good enough.

It doesn’t work on the individual level unless the test is perfect. Not for treatment, and not for going back to work.