@outta_so @Sjanderson86 @Neoavatara didn't mean to be short, I was on my phone. Basically this idea has been around for a while. IF there are lots of old cases that simply weren't identified as COVID-19, THEN it would drive the case fatality ratio down. Fauci talks about that here. nejm.org/doi/full/10.10…
@outta_so @Sjanderson86 @Neoavatara Numbers for influenza-like-illness were high this year; that wouldn't be inconsistent with lots of occult COVID-19. nejm.org/doi/full/10.10…
But people may also just be going to the doctor for respiratory bugs more often
But people may also just be going to the doctor for respiratory bugs more often
@outta_so @Sjanderson86 @Neoavatara Simulations based on R0=2.2 have suggested from the first that there should be lots more cases than being reported, 10x at least. systems.jhu.edu/research/publi…
@outta_so @Sjanderson86 @Neoavatara The number of confirmed cases in the US is growing quickly, but so is the number of tests. Right now, between 5-10 percent of tests are positive. With 100k new tests per day that would imply 5000 to 10,000 new cases confirmed if there were no new infections at all.
@outta_so @Sjanderson86 @Neoavatara I'm not ambitious enough to compute what the real growth rate in cases/day is right now, but the data is available on JHU's github repo here: github.com/CSSEGISandData…
@outta_so @Sjanderson86 @Neoavatara Wu et al estimate the read case fatality rate is closer to 1.4% than some of the wild estimates, but someone is sure to say "that's Chinese data" and discount it. nature.com/articles/s4159…
@outta_so @Sjanderson86 @Neoavatara they also put the R0 well below 2. All of those are risk of dying after developing symptoms and so is almost certainly an over-estimate for the population at large, since we know pretty certainly that there are asymptomatic or minorly symptomatic cases.
@outta_so @Sjanderson86 @Neoavatara That increases the denominator and thus lowers the fatality ratio. As an aside, note that paper says "the novel coronavirus SARS-CoV-2 was officially identified as the cause of the COVID-19 outbreak in Wuhan, China". Chinese *science* still puts the origin in Wuhan.
@outta_so @Sjanderson86 @Neoavatara The CFR in Italy, and to a lesser extent Spain, is a weird outlier: Italy, at around 10 percent, is many standard deviations from other European countries. Anyone who is telling you Italy is the best predictor needs a very strong argument — or they're trying to sell something.
@outta_so @Sjanderson86 @Neoavatara So far the best "petri dish" example we have is the Diamond Princess, in which sick people were closely confined under quarantine for some weeks. There only about 1 in 5 was infected, and the fatality ratio among the infected was about 1 percent.
@outta_so @Sjanderson86 @Neoavatara No one really understands why Italy and Spain are so high. Older population? Poor management in the hospitals? Misidentification of pneumonia deaths? No one knows. But the rate of NON-COVID pneumonia deaths seems unexpectedly low.
@outta_so @Sjanderson86 @Neoavatara One thing for certain is that measuring the tests *per capita* is innumerate claptrap. The US has a very large population; even doing the most tests, tests per capita will be low because the denominator is high.
@outta_so @Sjanderson86 @Neoavatara Or at least I prefer to think it's innumeracy and not malign attempts to mislead. In any case, since there is an active program to increase testing, number of new cases confirmed in the US should be taken with large grains of salt.
@outta_so @Sjanderson86 @Neoavatara We've seen a recent big increase in cases in Florida. We also started a big drive-through test program in the last few days. I doubt that's a coincidence.
@outta_so @Sjanderson86 @Neoavatara In any case, I'm tired and this has gone on long enough. Here are some points I think are important:
1) the gigantic "worst case" estimates all depend on a bunch of worst case assumptions and use very long time scales, which indeed leads to very big numbers.
1) the gigantic "worst case" estimates all depend on a bunch of worst case assumptions and use very long time scales, which indeed leads to very big numbers.
@outta_so @Sjanderson86 @Neoavatara Reworking the estimates with less dire assumptions does indeed lead to less dire predictions, as see newscientist.com/article/223857…
@outta_so @Sjanderson86 @Neoavatara 2) as I predicted some weeks ago, everyone with an agenda has found a reason to claim their agenda is critically important under the pandemic.
@outta_so @Sjanderson86 @Neoavatara 3) As I observed pjmedia.com/trending/who-b… some people stood to benefit big time from a panic, including short sellers. Coincidentally, the places that were really pushing the panic were skeevy finance websites and hedge trader. klfy.com/health/coronav…
@outta_so @Sjanderson86 @Neoavatara So my general advice is
* check all the math, even mine (although I am of course right)
* whenever you hear a COVID-19 story, ask yourself "who benefits".
End of thread
* check all the math, even mine (although I am of course right)
* whenever you hear a COVID-19 story, ask yourself "who benefits".
End of thread
