Hacker News Reader

Comment by TYPE_FASTER | Hacker News Reader

59k deaths out of 54 million infections is a 0.1% fatality rate. While the average fatality rate for COVID19 using global statistics is around 4%, let's use South Korea as an example because they have been testing much more frequently than other countries. The South Korea fatality rate is about 1%, so we'd see 540k fatalities in the US for similar infection rates.

Also, initial R0 transmission statistics for COVID19 look to be in the 2.5 range, where the R0 for seasonal flu is around 1.3.

So COVID19 is 2x as transmissible, and 10x as deadly. I don't think the conversation is around dismissing the seriousness of the flu, and more around emphasizing the risk of COVID19.

View on HN

> using global statistics

Let's keep it real, the flu infection rate given by the CDC is hindsight knowledge. The real time "global statistics" for COVID19 are with vastly restricted testing. There's not really a valid comparison to be made at this point.

Exactly everyone who is quoting numbers for flu deaths needs to realise these are estimates based on the statistics of how deaths vary year to year, i. e. looking at how many more people died in a flu year compared to other years (simplified). Most people actually don't die of the flu itself but secondary infections (the number of deaths linked directlybto the flu are orders of magnitude lower). Corona directly kills people.

The other thing to remember is the lag in the calculations, the people dying now are the ones who were admitted to hospital a week ago.

"Corona directly kills people"

I'm not so sure about that. Covid may grease the skids, but the vast majority of people dying have health complications already. Kids aren't dying. Teens aren't dying. 20s and 30s aren't dying in numbers. If Covid were directly killing, we'd see kids dying.

That does not follow. Yes the majority of people have preexisting health conditions, but they die directly of the pneumonia that is a result of COVID-19. For the flu, people typically die of secondary infections, e.g. bacterial pneumonia. Someone younger without pre-existing conditions is just much stronger to fight off the infection.

Regarding children, this is actually one of the many questions that is not fully understood, why do children not get so severe symptoms (initial thoughts were they might not even get infected, but this has been disproven by recent studies).

If you want to know fatality rates you need a controlled population. South Korea isn't that, its level of testing isn't even much higher than Switzerlands at this point.

There are only two known populations in which everyone was forced to be tested and their movement was limited, which thus makes up a controlled study: the Diamond Princess and the town of Vo in Italy. In both cases the fatality rate is much less than 1%. In fact in Vo it was zero.

https://www.theguardian.com/commentisfree/2020/mar/20/eradic...

https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-a...

In prior pandemics like this one the final fatality rate ended up being much lower than the initial reports: like an order of magnitude lower. So we would expect to see the same this time.

0.1% or less is well within the bounds of possibility, as the essay by Ioannadis explains.

The Diamond Princess’s population is also small, extremely poorly represents the general population, and happened in a setting unrepresentative of the rest of the world. Passenger age is just about the only factor that would drive the mortality and infection rates up, everything else lowers it. As a model the Diamond Princess should probably be treated more as the floor rather than anything representing the average.

everything else lowers it

What are these everything else’s? By far the best predictor of patient mortality has been age. What else is it about the Diamond Princess population that would come even close, even in sum, to the predictive power of their age?

Quality of care after the fact is the obvious one. The patients of the cruise ship weren’t being treated by a flooded hospital system. Quarantine on the ship is less than perfect, but it’s still better than the efforts made by many governments which was to do nothing until things got out of hand.

> In fact in Vo it was zero.

Isn't it a bit disingenuous to use a city with 89 cases as being statistically significant?

Vo has 3000 inhabitants, it's not a city, that's why they could completely isolate it and test every single inhabitant. It gives a 'random' sample, unlike other testing regimens.

The population size isn't that big indeed, but it's still multiple times larger than the cruise ship.

>The population size isn't that big indeed, but it's still multiple times larger than the cruise ship.

Which cruise ship? Diamond Princess had ~3600 occupants (passengers and crew).

> Vo has 3000 inhabitants,

How can you consider the fatality rate over an uninfected population? A year ago 8 billion peoples were not infected and 0 died! I have no idea why we worry. /s

What's matter for the fatality rate is the 89 persons infected, which isn't statistically significant.

Reply on news.ycombinator.com