Hacker News new | ask | show | jobs
by usaar333 2276 days ago
I actually don't think deep analyses comparing covid to flu are outright bs, even if I'm skeptical covid could have such a low death rate. Reading a lot, I've generally come to the conclusion that IFR of covid-19 is under an order of magnitude higher than flu, even if likely considerably higher (my guess above is 7x, could be as low as 5x). Regardless, the pandemics speed is a huge problem which can overwhelm hospitals.

The only data broad enough to predict true cases is Diamond Princess or perhaps Iceland where you've had enormous testing. Even on Diamond Princess, you have 10/712 infections resulting in death which is worse than flu (for this population), but not 10x as bad. Iceland has 2 deaths and 25 hospitalized against 500+ cases 1.5w ago, suggestive of a sub 1% CFR.

Additionally, flu CFR is reduced by targetted vaccination of the most vulnerable people (demographics of who get infected are more likely to survive than general population).

2 comments

> Even on Diamond Princess, you have 10/712 infections resulting in death which is worse than flu (for this population)

You still have to compare that to the baseline of the demographic. I.e. once you’re over eighty, you have almost a 1% probability to die within the next month.

> I.e. once you’re over eighty, you have almost a 1% probability to die within the next month.

True absolutely, but not true if you condition on me being fit enough for a cruise today.

> Additionally, flu CFR is reduced by targetted vaccination of the most vulnerable people

Yes, and COVID-19 would be a lot less bad if we had a broadly effective (even if as imperfect as the season at flu vaccine tends to be) generally available vaccine that would naturally be more likely to be taken by the most vulnerable, but alas we don't, which reinforces the point that COVID-19 is, in the real world, significantly worse than the flu, even if you could construct a counterfactual scenario where that would be less true despite the diseases each retaining their intrinsic traits.

Right and the "flu iir = covid iir" hypothesis would still lead to the real world pandemic being far worse (that's pretty obvious to anyone). You still have a disease with no vaccine that transmits extremely fast.

What would change though if that were true is that the calculus could shift away from locking down (SIP) to limited social distancing measures (to reduce r0) and a complete isolation of your most vulnerable populations. Basically, go for some sort of herd immunity if you can keep the herd's IIR below 0.01% or so.