Hacker News Reader

Comment by system2 | Hacker News Reader

If 0.37% is true, why are the hospitals in this shape?

Imagine all 65M people in the UK caught the virus then a 10,000 case fatality number would give a rate of 0.015%.

The rate itself has no time element so it tells you nothing about the number of people in hospital right now and hence isn't very illustrative of how hospitals can be swamped by this disease.

More interestingly (for me at least) is that the mortality rate is dependent on a patient's access to a ventilator. This means if you caught the virus early in the spread of the pandemic your odds of surviving are substantially higher than if you caught it right at the peak of infection.

This is impossible to reflect in a simple global mortality rate and so really paints a different picture of the virus - the death rate in South Korea vs other places really highlights this.

I guess to your point, hospitals are more in this way because the virus is so infectious, rather than it being a universally dangerous disease as may be seen from a higher mortality rate. Remember, a high number of people experience little to no symptoms from catching this.

This variability in symptoms doesn't seem to be as true for other lethal diseases like say malaria or dengue fever

> This means if you caught the virus early in the spread of the pandemic your odds of surviving are substantially higher than if you caught it right at the peak of infection.

I'd say your odds are best if you catch it some time after the peak. The hospital will be better prepared to prevent infections there and have more experience on how to treat patients best. There may even be some treatment that prevents the worst symptoms.

Also, if you (egoistically) try to get Covid-19 early, you may inadvertently infect others against their will.

Yep you're right, your odds are best towards the tail of the pandemic

In most cases, but the 1918 Spanish flu famously got more deadly after a mutation, if you got it early you were better off.

I'm still hoping to get it late assuming better treatment (or better yet vaccine), but I recognize there is no way to know if this is correct in this case.

The best case with mass immunity is that R drops below 0 and the virus burns itself out.

The worst case is that immunity primes the body for a massive and lethal immune system overreaction if the virus returns and/or mutates. This seems to have happened in 1918, which is why the second wave was more deadly.

Generally viruses seem to become less lethal over time, so this is rare and unlikely - and we'd be very unlucky indeed if it happened now.

But it isn't entirely impossible. The smart thing to do would be to have some high-quality monitoring in place to catch it if it starts happening, with high quality contingency planning for an international response.

Below 1, not 0. Below 0 would mean an infected person on average heals other infected.

Though my understanding is that the survival rate of a covid19 patient once he requires a ventilator is around 50%. If that's true, at worst, if we ran out of ventilators, the death rate would double. That's bad and we should avoid it but the doubling of a tiny number will still be a small number.

Which is why I am still confused by the ratio of 1:10 that the White House has shown between the death toll under lockdown vs no lockdown.

For the reason above, I would only expect the death rate to double. No lockdown means a lot more people get infected faster. But lockdown still requires to build herd immunity, we are only slowing it down. So the same number of people will get infected ultimately. So either they implied that the lockdown would be permanent until we get a vaccine (at best 2021 Q1), or it completely disregarded what happened after the initial few weeks of lockdown and then the numbers were at best grossly misleading.

Some 15% of people become “seriously” unwell, and require medical intervention such as oxygen. In a situation where the peak is hit unmitigated, these people will not get the medical care they need, and a lot of them will die.

The 15% looks implausible given the fatality rates mentioned here.

But I would expect a lot of the medical care to be scalable so unmitigated doesn't mean no medical care. We have been building field hospitals in the UK and US (though have barely used them so far), using hotels, etc. Providing oxygen should be fairly straightforward. I understood ventilators was the part that caused a serious scalability bottleneck.

Yes - part of the strategy of flattening the curve is to buy time to scale capacity. Dedicated covid19 hospitals also have the advantage of isolating covid19 patients from high risk non-covid19 patients (e.g immunocompromised cancer patients)

As the disease is so easily transmitted hospitals are partly buckling due to contamination within hospitals. A regular 50 year old probably has a good chance of surviving coronavirus. A 50 year old who is midway through chemo, probably less so...

All of this nuance is hard to express through topline death tolls and mortality rates.

The "We shouldn't have killed our economy for this" online comments miss some obvious truths - which is that the economy would be crippled by a high peak regardless.

At best people wouldn't go to work, at worst they'd struggle in and pass out - or maybe die - in front of their coworkers.

So the economy would have ground to a halt anyway as people stayed at home to protect themselves. But it would have done it in an uncontrolled way that didn't inspire confidence in a return to normality.

Which is the economic reason for flattening the peak. Serious damage was always inevitable, but there's a huge difference between a barely controlled but orderly temporary shutdown with mitigation measures, and a shutdown caused by mass panic and paranoia.

In fact even laggard countries have been forced to take some measures, because when their lives are at risk people and their managers act on their own initiative no matter what the official line is.

I suspect you are assuming a large proportion of people who get the virus have severe symptoms. The evidence of the contrary seems to emerge.

  Old people, those in the middle of a chemotherapy and many other people at risk aren't going to work and hence won't die in front of their coworkers.

We've been on lockdown because approx 1% of the US population has tested positive. That means 99% are not. How can we come out of lockdown when we are effectively still where we started. We're not even close to herd immunity.

Not sure I agree with your numbers but I agree with your point, we are very likely quite far from herd immunity.

Which means another 18 months in lockdown if we want to stay well within the current ICU capacity until a vaccine is mass produced. Which means an economic downturn not only worse than 2008 (it will be worse already assuming we leave lockdown in a few weeks), but many times worse.

Or we can swallow this thing. It will mean more deaths. Trump says 10 times more. What I am saying is that I don’t understand how he gets there. If it is “only” two times more, then we need to have an adult conversation on how many lives are worth saving vs our society going through something akin to the great depression.

> If it is “only” two times more, then we need to have an adult conversation on how many lives are worth saving vs our society going through something akin to the great depression.

Why is it that only two choices ever exist when this argument is presented? Choose either death by virus or death by status quo economy. It seems like a more pleasant alternative would be reshaping our economic systems to be more resilient against such a crisis which will almost certainly occur again in the future. Maybe the current ownership and distribution models and how they relate to globalism aren't compatible with something that is now simply a fact of life.

The ownership and distribution models aren't the problem. The problem is that while everyone is locked up at home, no one consumes, we don't go to restaurants, to malls, we don't do tourism. And in some case we don't produce either, construction is halted, etc. There is almost no economic activity going on.

No economic model works when everyone is effectively on house arrest.

That sounds like an argument based on faith.

Edit:

that we can not "be reshaping our economic systems to be more resilient against such a crisis" seems like arguing from faith. There is no natural law that dictates that everything has to go to haywire just because things are on a pause, or that the current systems, particularly in the US, constitutes the optimal way to handle this.

> That sounds like an argument based on faith.

Faith? That when everyone has to stay home, no significant economic activity can happen? You can call it faith...

The actual CDC plan is here:

https://fm.cnbc.com/applications/cnbc.com/resources/editoria...

It's not just "full lockdown vs no lockdown".

Right. So why were we in full lockdown then? I can't imagine 1% is that significant.

> But lockdown still requires to build herd immunity, we are only slowing it down. So the same number of people will get infected ultimately.

Yes. I've been saying that, too.

So, in blunt terms, currently we are trying to keep people alive until we can get therapeutics and/or vaccines that confer immunity so that we can get to herd immunity with fewer infections and deaths.

The actual CDC recommendations are here:

https://fm.cnbc.com/applications/cnbc.com/resources/editoria...

The survival rate of a COVID-19 patient once he requires a ventilator is closer to 20%. So running out of ventilators wouldn't double the death rate.

https://www.sfgate.com/news/medical/article/Some-doctors-mov...

In addition, this virus tends to prey on the weak. You don't want to have a pre-existing condition around Covid-19. If you look at the top causes of deaths in the USA in recent years then match well with C19's "favorite" targets. This instance, high blood pressure, diabetes, lung disease (e.g., COPD aka smoking), etc.

The last stat I heard was 40% of USA adults have a pre-existing condition. That's a significant amount of potential hospitalizations.

A study of patients in New York showed that obesity was the top chronic pre-existing condition risk factor.

https://www.medrxiv.org/content/10.1101/2020.04.08.20057794v...

Right. And obesity is, in general, preventable. That is, C19 is a symptom. The underlying "disease" is how generally unhealthy the average citizen is.

Access to ventilators has only a minor impact on mortality rate. About 80% of COVID-19 patients placed on ventilators die. (This is significantly worse than other forms of ARDS.)

https://www.sfgate.com/news/medical/article/Some-doctors-mov...

From what I read the mortality rate of patients on ventilators is near 80% in NY. Are they really beneficial?

But what does that 80% mean?

We know that New York's hospitals are overrun. We also know that in countries where the hospitals aren't at capacity the mortality rate is much lower. Look at Wuhan vs the rest of China. Or Singapore/HK/Taiwan/South Korea vs Italy, Paris, New York.

So if ventilator access is low and doctors are triaging ventilators to go to the most urgent cases, is the data skewed towards severe cases and late access to ventilation support?

This anecdotal account[1] seems to suggest ventilate as early as possible to maximise the chances of saving a patient. So again, a global stat could be masking a time element and context to the data

https://www.bloomberg.com/news/articles/2020-03-24/covid-19-...

These were among the very first cases in Germany where capacity was abundant, hence the low fatality rate.

Also, this is a “up to now” value, more cases might turn out fatal.

“Only” 7 deaths is furthermore not a good data basis for calculating this number.

Moreover, the report is not even public nor peer reviewed. So I can’t even cite it here. However you find criticism of the study by Prof. Streeck online.

At least, here is a link to a preliminary summary of the study, in German language. Please note how shallow the text is in its current state.

https://www.land.nrw/sites/default/files/asset/document/zwis...

The problem is the high number of infections. The virus is having a run in a population that has no immunity, unlike a flu. So it is spreading fast and in large numbers. A tiny hospitalisation rate multiplied by a very large number of people infected may be more than the hospital capacity.

Which is why both of these statement can be true at the same time: this thing is no more dangerous than the flu to an individual if you get it, but it can kill many more people than the flu because the population doesn't have any immunity and many more people will get it.

The majority of people that end up in hospitals actually survive the virus. Their symptoms are treated, not the causes, unfortunately. But you have to take these people in as they have (among other things) respiratory complications, hence the great demand for ventilators incidentally.

Infection Fatality Rate is certainly coupled to the number of people that develop severe symptoms, but it's not this number that you need to look at as to why people end up in hospitals.

Because it's spreading so very fast, and nobody has any pre-existing immunity to this new disease. The flu kills 60,000 Americans every year and so far COVID has killed about 10,000 -- over roughly the same flu season time period. The difference is that the flu is spread out over time and is fighting some pre-existing immunity and flu shots. It also has a lot to do with excess capacity. If there suddenly appeared a new disease exactly as fatal as the flu, that would kill another 60,000 Americans per year and 650,000 people worldwide. There's not really capacity for that.

So far Covid has killed 34,000 not 10,000 in the States.

Thanks, I grabbed the wrong column from my data. Either way that’s half the flu total over the same season. My point is that it isn’t unprecedented load, it’s just all at once — and additive.

If you're going to compare the deaths you need to count them in the same way, and you haven't done so here.

Well it’s fair to say that neither has the government, if anyone dies in possession of COVID they’re automatically a COVID death, pretty much even if they get hit by a truck. That’s not true of the flu is it?

It seems unlikely that someone would even be tested after being hit by a truck, or that a doctor would choose to list it as a cause of death.

No, because people die every day and no tests are run if they had the virus.

And indeed they can be counted as COVID deaths if they had respiratory illnesses because the CDC direction is to count any suspicion of COVID or the contribution thereof as a COVID death. [1]

[1] https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-New-I...

You're incorrect to say that if anyone dies with Covid-19 that they're counted as a covid-19 death.

It's likely that covid-19 deaths are undercounted. Here's one example: https://ltccovid.org/2020/04/12/mortality-associated-with-co...

For a while the US was only counting deaths that happened in hospitals.

This argument -- that we're somehow over-counting covid-19 death -- doesn't hold because we can compare excess mortality this year against the previous 5 years.

https://www.economist.com/graphic-detail/2020/04/16/tracking...

Spain: https://twitter.com/robertwiblin/status/1247500657386557443

For the UK: https://www.cebm.net/covid-19/tracking-mortality-over-time/

> That’s not true of the flu is it?

Yes, it is. The figures that count covid-19 vs flu are available, and they use the same method. They look at death certifications and then statisticians code those. There's lag in those figures which is why worldometers don 't use that data.

> You're incorrect to say that if anyone dies with Covid-19 that they're counted as a covid-19 death.

I'm not. In some countries they absolutely do count anyone who dies while in the possession of COVID as a COVID death, for instance Italy. " Italy’s death rate might also be higher because of how fatalities are recorded. In Italy, all those who die in hospitals with Coronavirus are included in the death counts."

“On re-evaluation by the National Institute of Health, only 12 per cent of death certificates have shown a direct causality from coronavirus, while 88% patients who have died have at least one pre-morbidity – many had two or three.” [1]

In New York they're not just counting speculative COVID deaths of anyone with respiratory illness even if they've never tested positive [2].

"A subtler issue is what to do when the patient has other serious medical conditions. If the person suffered from chronic lung disease, then became infected with the virus and died of pneumonia, the immediate or primary cause would be pneumonia as a result of COVID-19. The lung disease would be listed as a contributing condition, said Sally S. Aiken, president of the National Association of Medical Examiners." [3]

The CDC has guidance on this but it's fair to say its interpretation will vary from place to place. "COVID-19 should be reported on the death certificate for all decedents where the disease caused or is assumed to have caused or contributed to death" -- that's pretty broad. [4]

My understanding is a few folks who committed suicide and a few who died in car accidents due to head trauma were coded as COVID, but I can't find my source on that and I assume it's pretty limited. I wasn't being literal in my earlier comment, however, and nor should this paragraph be considered as changing or doubling-down on that. I assume such cases, if they do in fact exist would be few and far between.

Hope that helps.

I also don't think it's fair to pin 100% of excess deaths entirely on COVID. We shall see, however.

[1] https://www.cebm.net/covid-19/global-covid-19-case-fatality-...

[2] https://www.cnbc.com/2020/04/15/coronavirus-new-york-city-st...

[3] https://www.inquirer.com/health/coronavirus/coronavirus-covi...

[4] https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-New-I...

From what I've read the CDC has told med facilities to code any death that is even suspected to be C19 as C19. No test? Doesn't matter. Patient had other serious conditions? Doesn't matter.

With broader testing limited and deaths from being inflated the IFR should be suspect.

You are in fact correct, and you are welcome to use this CDC guidance as a reference in the future, should it come up: https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-New-I...

That number only includes deaths in hospital of confirmed cases, a lot of people are dying at home particularly of heart attacks induced by oxygen starvation. The data from New York looking at excess deaths compared to the same period last year indicates that the actual number of Covid-19 related deaths is probably around double that.

How do you know? Those deaths were not necessarily caused by the virus, in fact some of those dead never had a positive covid test.

There‘s bound to be a high number of Covid-19 deaths that will only be obvious by looking at excess mortalities. At least preliminary data from the Netherlands and Spain indicates that.

Yes, and it will take weeks before we have decent numbers, so it's dangerous to draw any conclusions just yet.

The numbers aren't looking pretty, though: https://www.euromomo.eu/index.html

The 2% number is 2% of people who have bad symptoms die. 0.37% is 0.37% of all people who get it, including those who have mild/no symptoms die. It simply means a much larger percentage of people get it, not that any less people are dying. I’d say there’s a good chance that many of us have gotten it, and spread it, without knowing it.

> The 2% number is 2% of people who have bad symptoms die.

I don't believe that's true. In countries such as Spain, right now there are about 80k confirmed recoveries and 19k deaths among a universe of 182k confirmed cases.

We don't know. There are probably a lot of cases in Spain that are flying under the radar because people without symptoms don't get tested often.

OP's claim was that "2% of people who have bad symptoms die."

From Spain's numbers, around 10% of those experiencing symptoms that lead them to get tested by Spain's health service, which due to scarcity is reserved for serious cases and people on the front lines, have died since the onset of the epidemic.

A majority of Spanish fatalities were in nursing homes and other restricted facilities. If the virus penetrates these places, average mortality spikes.

Korean testing is revealing a fatality rate of about 2% over all cases, with about 8% requiring hospitalization. You are off by an order of magnitude.

Korean testing is largely confined to the symptomatic. They are still suffering from the same issues with missing the iceberg that everyone else is.

One possible explanation is that hospitals were already operating close to capacity. So even adding a small number (percentage-wise) of the population very fast, will cause disasters.

They're currently mostly empty so it matches.

.0037 * 83,700,000 = 309,690 additional people

It's still going to be a giant number, but I recommend including the fraction estimated to catch the disease, in this case 0.7 because diseases with an R0 of 2-2.5 typically infect 70% of the population.

Because the infection rate is so high.

Reply on news.ycombinator.com