Interesting article. What's your take?

boots

Sat, Apr 25 2020, 6:48 pm

The fatality rate quoted is most certainly an underestimation as they only included cases were either known or probable cause of death was Covid19. They did not take into account the fact that the underlying death rate has nearly quadrupled in nyc since the start of the epidemic. That means that if you look at all the recent deaths attributed to heart attack, stroke, etc, there have been 4x as many of those and that increase is almost certainly due to Covid19 but is not included in the official count.

amother

Babyblue

Sun, Apr 26 2020, 1:03 am

His math is moronically absurd. 11 deaths per 100,000 people in population means that the fatality rate is 0.01%? The fatality rate is fatalities/cases, not fatalities/population. (The current case rate is 1648 per 100k according to NYC DOH, and fatality rate 13.59 per 100k, so a .82% fatality rate among confirmed cases. There is also a 10% hospitalization rate (166.31/16.48), which is staggeringly high, and would project mass fatalities assuming health care systems are knocked out. Even assuming that 90% of cases among this group were not diagnosed (a pretty generous assumption, albeit definitely plausible), that would come out to a 1% hospitalization rate (unclear how the current hospitalizations will resolve, but it's unfortunately likely that some will result in morbidity, thereby increasing fatality rate), and a currently .08% fatality rate, with a large number of unresolved cases leaving the real rate an unknown.

I'm not going to go through the rest of the article, but this should provide some context as to the intelligence quotient and mathematical aptitude of the author.

ora_43

Sun, Apr 26 2020, 1:44 am

There are two huge issues with the article (and the zillion or so similar articles):

The Stanford study. The researchers themselves say in the same study that people (think they) are basing these articles on that if the tests they used are even slightly less accurate than reported, their results could be significantly off.

And guess what? Other attempts to use those same tests have found that they are not as accurate as reported.

The Stanford study is interesting and important but it is NOT the final word on the death rate, and it would be extremely stupid and careless to treat it as if it is. It's a single study, using (potentially less accurate than the company claims) antibody tests, that contradicts many other studies on the death rate (a Berkeley study done around the same time, for example, found that the minimum death rate is over 0.5%).

Apples to apples.

Let's say my neighbor and I both decide to count the number of cats living around our buildings.

I decide to figure out how many cats I can see, and then multiply that by 5, because cats are good at hiding so I figure there are probably 4 cats in hiding for each one I see. Based on that logic, I decide my building has 20 cats.

My neighbor decides to put out a bowl of cat food, take a sample of fur from any animal that eats the food, and test it for cat DNA. Based on her method, she finds 2 cats.

Now - how certain are you that my building has 10x more cats than my neighbor's? Not certain at all, right?

Back to covid19 - people are comparing the death rates for covid19 and the death rates for flu, but calculating the two death rates in completely different ways.

When you compare confirmed deaths of covid19 to estimated deaths of flu, your results are going to be wildly off. Comparisons aren't completely lacking value (they can be good for trying to get a ballpark sense of how relatively bad illnesses are) but they have to be taken with massive heapings of salt.