Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf] • Hacker News Board

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#71

Post by logicchains » Sat, Oct 17, 2020, 3:45 PM UTC

gridlockd wrote: post #67

bearjaws wrote: post #49

Earlier quoted context omitted.

Only two states have continued preventing elective surgeries, one state (Texas) that was overwhelmed stopped but has since re-allowed them. Around 60,000 cancer patients will be under-treated due to corona virus. The current estimates are that 10,000 people will die due to missed treatment. Just for the "fun" of it, lets assume all of them died due to under-treatment, that is still 1/4th the death toll of the virus i…

In medicine, people like to talk about quality-of-life-adjusted years. Why? Because there's no point in making somebody live more years that are subjectively bad (e.g due to treatment side-effects), rather than less years that are subjectively good. When it comes to policy-making that affects us all, it's time to extend that to the collective level. We need to ask ourselves, how much suffering can younger generations…

>Because there's no point in making somebody live more years that are subjectively bad (e.g due to treatment side-effects), rather than less years that are subjectively good.

This also applies to locking down the elderly, as lockdowns cause more loneliness, which increases the risk of mental deterioration: https://medicalxpress.com/news/2020-07-highlights-loneliness....

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#72

Post by rndgermandude » Sat, Oct 17, 2020, 3:46 PM UTC

logicchains wrote: post #61

rndgermandude wrote: post #50

Earlier quoted context omitted.

Of course you haven't seen papers describing long-term effects, simply because the virus hasn't been around long enough for any long-term conclusions. Therefore there are no studies saying there are NO long term effects either. There are clues that long term effects/permanent effects even in mild cases may exist. E.g. some six scuba divers with mild symptoms/asymptomatic progression were checked afterwards (Innsbruck…

> Of course you haven't seen papers describing long-term effects, simply because the virus hasn't been around long enough for any long-term conclusions. So you'd support locking down based on something for which we don't have any evidence yet?

We have seen evidence that the virus kills. That alone warrants action. In the beginning, the general lockdowns were fine, simply because we didn't know enough yet to come up with better solutions, and frankly didn't really have the time to rectify this before taking any action. Now whether or not to lockdown and what to lockdown can be applied more fine-grained. Of course, everything is still in flux, and some decisions will turn out to be wrong later. The entire response is trial-by-error, for lack of alternatives

In regards to potential long term effects, those should be a concern as well when making decisions, yes. Not the only concern of course, but not something to be ignored either.

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#73

Post by nradov » Sat, Oct 17, 2020, 3:49 PM UTC

logicchains wrote: post #58

nradov wrote: post #54

Earlier quoted context omitted.

Florida went even further and removed all state level lockdown orders on September 25. So far it's too early to tell what effect that will have on the death rate. Case numbers are roughly flat. https://www.worldometers.info/coronavirus/usa/florida/ I'm not advocating for other states to follow Florida's approach. But it serves as a natural experiment. Depending on what happens to their daily death rate over the next…

>Depending on what happens to their daily death rate over the next couple weeks that will tell us a lot about the effectiveness (or lack thereof) of lockdowns. It's already been more than a couple weeks since September 25th; how many more "couple weeks" do we need until we can draw a conclusion?

For patients who are going to die, that typically happens about three weeks after initial infection. After the lockdown orders will lifted it still takes a while for businesses to reopen and for people to change their behavior. There are delays in state level data reporting.

So I would say that if we don't see a sustained rise in death rates in about three more weeks from today then that would confirm the "null hypothesis" of lockdowns being ineffective. Or if they have a major spike in deaths then that would indicate that lockdowns are effective.

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#74

Post by logicchains » Sat, Oct 17, 2020, 3:50 PM UTC

rndgermandude wrote: post #72

logicchains wrote: post #61

Earlier quoted context omitted.

> Of course you haven't seen papers describing long-term effects, simply because the virus hasn't been around long enough for any long-term conclusions. So you'd support locking down based on something for which we don't have any evidence yet?

We have seen evidence that the virus kills. That alone warrants action. In the beginning, the general lockdowns were fine, simply because we didn't know enough yet to come up with better solutions, and frankly didn't really have the time to rectify this before taking any action. Now whether or not to lockdown and what to lockdown can be applied more fine-grained. Of course, everything is still in flux, and some decis…

>That alone warrants action.

In only warrants action if we're sure the negative consequences of the action won't be greater than the negative consequences of the viral deaths.

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#75

Post by gridlockd » Sat, Oct 17, 2020, 3:53 PM UTC

segmondy wrote: post #52

thrav wrote: post #41

Earlier quoted context omitted.

This is more or less what Texas has decided to do. I’m in California, but my entire family is there, and many of my younger family members are going about their lives relatively normally. At least one of my cousins caught it and their whole young family (parents 30’s, kids Everybody hated on Texas because of the Lt. Governor’s over the top declaration of sacrificing grandparents for the economy, but there are definit…

You have no idea how it would affect you or your family members until it does. This person lost 8 family members. https://www.cnn.com/2020/10/14/us/phoenix-arizona-aguirre-fa...

"Then, despite all precautions against the virus, a family member got sick. Then another and another and another. Now, seven months after the shutdown, seven people in Ricardo Aguirre's extended family and his father, Jesús, 67, have died of Covid-19 complications."

Two observations: First, this happened despite all precautions. Secondly, extended families can be hundreds of people, including many elderly. Lastly, in a large-enough population you will inevitably find clusters that are more affected by some illness than the average person. That's why you need to observe solid data, not a media spectacle, to come to rational conclusions.

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#76

Post by nradov » Sat, Oct 17, 2020, 3:55 PM UTC

lima wrote: post #59

logicchains wrote: post #35

Earlier quoted context omitted.

For example? I haven't seen a single paper showing such long-term consequences in healthy people are significantly more likely than from other viral infections. We'd especially expect to be seeing a large number of such cases in Sweden, should it be common, given they had no lockdown and a large number of infections.

It doesn't even have to be worse than post-viral syndrome caused by other severe viral infections for it to be a problem. There's little immunity to COVID-19 and it's highly contagious. This is like saying "COVID isn't worse than the flu" - even if it was true, having another flu-sized disease burden would already be bad, so it's a bit of a weird goalpost. That being said, there's a lot of literature describing post-…

That depends on what you mean by "little immunity". Many people appear to have some immunity based on prior exposure to similar coronaviruses. That won't prevent infection but tends to reduce disease severity.

https://www.jci.org/articles/view/143380

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#77

Post by dragonwriter » Sat, Oct 17, 2020, 3:57 PM UTC

logicchains wrote: post #58

nradov wrote: post #54

Earlier quoted context omitted.

Florida went even further and removed all state level lockdown orders on September 25. So far it's too early to tell what effect that will have on the death rate. Case numbers are roughly flat. https://www.worldometers.info/coronavirus/usa/florida/ I'm not advocating for other states to follow Florida's approach. But it serves as a natural experiment. Depending on what happens to their daily death rate over the next…

>Depending on what happens to their daily death rate over the next couple weeks that will tell us a lot about the effectiveness (or lack thereof) of lockdowns. It's already been more than a couple weeks since September 25th; how many more "couple weeks" do we need until we can draw a conclusion?

Well, you'd need a couple weeks before you expect to seeing measurable impact on cases/hospitalizations and another couple weeks before it starts really showing up in deaths. The daily data is noisy to start with so it's hard to eyeball trends, but it looks like cases have started to trend up recently and deaths are still mostly flat, which is about what you'd expect at this point in time if the lockdown that was lifted was having a significant effect on limiting COVID impacts.

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#78

Post by gridlockd » Sat, Oct 17, 2020, 4:03 PM UTC

airstrike wrote: post #22

White_Wolf wrote: post #11

Earlier quoted context omitted.

(1) Is generic and not entirely true. Long term effects are indeed present but only in those genetically predisposed. Doesn't apply to the majority. (2) Actually we didn't. My wife works in a small city hospital. They only have 4 beds with respirators and almost never had more than 2-3 people at a time there. The city is being locked down the second time because we have 5 infected in 20k. We also had in the region of…

> (2) Actually we didn't. My wife works in a small city hospital. "It didn't happen to me so it must not be happening anywhere else"

It's about as valid an argument as appealing to the pictures of "overwhelmed hospitals". Yes, we've all seen those pictures. We've also seen pictures of empty emergency care units that were created in anticipation of a wave of illnesses that didn't materialize.

In general, we can say that in the overwhelming majority of cases, hospitals did not get overwhelmed with COVID patients like some of the models predicted. That may change as we move into the second winter season, but it's not a foregone conclusion.

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#79

Post by gridlockd » Sat, Oct 17, 2020, 4:08 PM UTC

rich_sasha wrote: post #65

rossdavidh wrote: post #2

For comparison, from what I've read the typical annual influenza fatality rate is around 0.1% (don't know what the under-70 rate would be). So, this puts covid-19 solidly in the uncanny valley of viral mortality rates; not low enough to be "just like a flu", not high enough to justify shutting down the world. No wonder opinions on it are so divided.

I guess not so many people get flu? I don't regularly get vaccinated, and I only had 1 flu-like disease in the last decade. Covid is way more virulent. If 10x more people get it, it doesn't need to be more deadly in terms of IFR to be a concern.

> If 10x more people get it, it doesn't need to be more deadly in terms of IFR to be a concern.

Sure, but what sort of concern? A "shut down everything, COVID-cases are rising!" sort of concern?

Re: Infection fatality rate of Covid-19 inferred from seroprevalence data [pdf]

#80

Post by lima » Sat, Oct 17, 2020, 4:21 PM UTC

logicchains wrote: post #66

lima wrote: post #59

Earlier quoted context omitted.

It doesn't even have to be worse than post-viral syndrome caused by other severe viral infections for it to be a problem. There's little immunity to COVID-19 and it's highly contagious. This is like saying "COVID isn't worse than the flu" - even if it was true, having another flu-sized disease burden would already be bad, so it's a bit of a weird goalpost. That being said, there's a lot of literature describing post-…

>It doesn't even have to be worse than post-viral syndrome caused by other severe viral infections for it to be a problem. There's little immunity to COVID-19 and it's highly contagious. But it does make it highly inconsistent to adopt policies so extremely destructive to economic and mental health when the same wasn't done for other viruses with similar long-term effects.

We have a chance at eradicating COVID-19, unlike other viruses that have been endemic for decades.