Quote:
Originally Posted by TLO
I believe the big scary number came from the Imperial College in London. I believe those were the numbers shown assuming no social distancing was put into place. (Up you 2.2 million American deaths)
That model then backed everything way, way down not too long after. For example they predicted 500l deaths in the UK, then backed it down to 50k.
Not long after that, the White House showed a model that predicted 100k - 240k Americans dying even with social distancing measures in place.
We then learned that the model the White House was using was very similar to the IHME model we all know and love. I distinctly remember Dr. Birx referencing this.
Dr. Fauci then explained that models are good, but we can't base all of our decisions on them. (So many different variables that go into models)
|
As more information comes in our ability to refine predictions will obviously improve. I've mentioned this twice in the last day, but I think it bears repeating. Regardless of what you initially believed about the models, we are getting enough data in from New York and the infectiousness of the virus to make some baseline assumptions about the death toll that are fairly well grounded.
What we know:
*New York State has a population of 20,000,000
*New York has a death toll of 17,671 at this point
*Tests of suspected COVID patients in New York were 38% as of April 1
*Deaths usually lag about two weeks behind infections
What we are assuming based upon some clinical research:
The low end of the R0 is 2.2, the high end is over 8. One study put the R0 at 5.7
We can plug and chug to help us understand fatality rates, but this is still back of the napkin stuff:
Given that 38% of suspected COVID patients were positive in April (and those are the ones most likely to test positive) and the test has a false negative rate of 30%, at most 47% of people in New York (and I'm counting the entire state, which will greatly elevate the potential number of infected) would have been infected at that time.
That gives us a pool of 9.4 million infections in New York. Although this is highly, highly unlikely, it would give us a lower bound of a fatality rate. As of now it would be 0.19.
Now, if we assume that the R0 is 2.2, then we would need 55% of the population to be infected to reach herd immunity. If the R0 is 5.7, then it's 82%.
Thus, with no mitigation strategies, and assuming that New York had a population that was actually 47% positive (essentially impossible) with no excess deaths, the total death rate from COVID with no mitigation would be:
330,000,000*0.55*0.0019= 344,850
With an R0 of 5.7 it is: 330,000,000*0.82*0.0019=514,140
And that's assuming that hospitals wouldn't be overrun.
If mitigation strategies reduced the R0 (known as Re) of the virus to 1.5, then 1/3 of the population would need to contract the virus for herd immunity. Thus, mitigation strategies, even if they only lowered the infectiousness of the virus by 50% on on the low end, would save this many lives:
(344,850)-330,000,000*0.33*.0019= 137,940
So, given what we know now, even if mitigation strategies were only 50% effective and the virus was at the low end of infectiousness, distancing, shutdowns, mask wearing in public after reopening would save, at minimum, 137,940 lives.
I can definitely see how the model came to an estimate of 1.1-2.2 million deaths without mitigation, because if New York ends up with 40,000 dead, then, by definition, the CFR couldn't be less than 0.2%, even if every single person in the state was infected, which is an impossibility.