re: Daily COVID Updated as of 11/2/20 8:00 PM

quote:

---

I reverse engineered a possible mortality rate at .05%. 18% positive * 330M = 62M infected, 28.5K deaths on 62M infected = .05% mortality.

---

Explain?

Numbers were higher today for confirmed cases of Covid19, but when you look at the number of tests done it makes complete sense. There was a large increase in tests completed. Thus, with a large increase in more testing, and still lower than most days in the last 7ish days, that is a huge sign that things are really slowing down.

As testing increases there will be more and more people testing positive hat have very mild symptoms or asymptomatic.

Tracking death numbers nowadways isnt worth it for two reasons:

1: Lot's of people with mild symptoms don't bother getting tested. Look at the study with prgenant females in NYC. Most tested positive to have COVID19 with no symptoms.

2: Hospitals are adding deaths not from COVID19 as COVID19 deaths.

Now for new cases, the number will rise if there is a continued increase in testing but will continue to decreae slowly over time if there isn't a big uptick in testing.

I predict we see a lot more testing and more confirmations that millions are asymptomatic and the severe cases are less than 1.5 percent of the popoulation of the United States.

quote:

---

I expect the IFR will end up between 0.3-0.7%

---

I would agree based on a focus on the NYC data.

The 5 NYC counties have a positive test rate of about 50% (118,302 positive of 237,371 tested). Extrapolated out to the population of those counties (8,400,000), that would be 4,200,000 infected. Those 5 counties have 8,215 deaths so far, which gives 8215/4200000 ~= 0.2%

Of course, that positive test rate is subject to the sampling bias that you mentioned, so we should take 0.2% IFR as an absolute lower bound. I see a lot of people arguing that since a lot of infections are asymptomatic, the infection rate could be a lot higher. However, it seems absurd to think that people without symptoms are somehow more likely to be infected than those with symptoms.

Cosmos, we have administered 3.2M tests so far. Of those tests the cumulative % positive is 19.51%. Today's Positive % was 18.7%.

What I did was apply today's 18.7% to the US population of 331M people which yields 62M infected people.

The current number of deaths as of today is 28,529 which is .05% of the assumed infected population.

quote:

---

Cosmos, we have administered 3.2M tests so far. Of those tests the cumulative % positive is 19.51%. Today's Positive % was 18.7%.

What I did was apply today's 18.7% to the US population of 331M people which yields 62M infected people.

The current number of deaths as of today is 28,529 which is .05% of the assumed infected population.

---

The main reason I think this estimation is quite far off is again because of the NYC data. There are so far 8,215 deaths in just the 5 NYC counties. We can assume that 100% of the total population of 8.4 million has been infected, and that still would give an IFR of 0.1%.

And they will likely add 3k more deaths soon.


Edit: Actually NYC has 8,455 deaths so far. I was looking at location of fatality instead of residence of individual. Not that it makes much difference.

LINK to Twitter

Tweet has link to the paper. (and full size photo)

quote:

---

Cool paper builds a Random Forest model from ~4,000 Covid-19 positive cases to find risk of being hospitalized & entering critical care.

---

Put simply, if you catch this, don't be old or fat.

quote:

---

I think that would significantly overstate infected numbers because you have a sampling bias towards those who are symptomatic, which is not representative of the US as a whole. I expect the IFR will end up between 0.3-0.7%, still significantly lower than the current CFR; however, that would result in a current infected population of 4-9.5MM (though this estimate isn’t perfect, not least due to testing constraints and deaths being a lagging indicator).

---

I am interested to understand the basis for the observation I am seeing here and elsewhere; namely, that there are a significant number of infected but asymptomatic people in the community.

Is there any scientifically credible basis for this view? Initial estimates were that 80% of cases were mild in the sense that the people affected did not require admission to hospital, but early in the piece I was reading that cases of individuals being truly asymptomatic for the duration of the virus were much lower - e.g. below 5-10%.

I am unfamiliar with the reference above to a study of pregnant females who tested positive for COVID19.

It is however interesting to note that NYC has recorded four times the average/projected number of deaths of people in their own homes for the past month. In most instances those people will not have been tested for COVID19, but of course it leads to the inescapable conclusion that the vast majority of deaths comprising this four-fold increase are due to COVID19.

Missouri is over the first hump.

New cases has fallen off a cliff.
250 new cases on April 6th.
Less than 50 new cases yesterday.

Total CV hospitalizations fell by 17 overnight. Total of 1024 hospitalized, about half confirmed. The other half are awaiting test results.

The IHME model continues to push our peak usage into the future. They are projecting we’re on the upward curve of utilization and mortality still. They have upped our projected vent usage from 240 to 393 at peak. It continues to mystify.

quote:

---

Put simply, if you catch this, don't be old or fat.

---

In my state add to that female (55%) or black (70%).