Coronavirus Mortality Rate | Hacker News Reader

Coronavirus Mortality Rate

worldometers.info

103 pointsrgbrgb74 comments

Okay, I have been following this closely as I follow JS libraries and this is the headline: The virus seems to require intensive care to some subset of the infected people and deaths occur when that care is not given in timely manner. That means it's mortality rate increases dramatically when your patient care infrastructure gets saturated.

Eloquently put. Agree 100%. The logical conclusion then is to ramp up your patient care infrastructure proactively: identity and clear the bottlenecks.

This leads to situations a society has to be willing to accept without crying how much money is wasted - namely: You will have empty beds/hospital wings/hospitals most of the time. And people will sit around doing 'nothing' aka training for the day when they are needed.

I'm not convinced that the current political climate in either Europe or America is prepared to do that.

> And people will sit around doing 'nothing' aka training for the day when they are needed.

This is already an accurate description of most fire departments and militaries around the world. It shouldn’t be too hard to accept that a similar necessity of spare capacity occurs for healthcare.

You are woefully underestimating the size of healthcare compares to emergency services. Neither emergency services nor the military are slack. They are both scaled for sustained operations; peak utilization is made up by volunteers or allied forces/departments.

BLS says there are 334K at firefighters at $49k/yr.

There are 240k NPs at $113k/ye, 3M RNs at $71k/yr, 728K licensed nurses at $46k/yr, 300K lab techs at $52k/yr, and 1.5M NAs at $28k/yr. Go ahead and double that coat when you include MDs and Admin.

maybe the military, but a typical fire department is orders of magnitude smaller than current health workers force. At least where I live however, flu outbreaks are tipically worked by making the extra workers double their turns an paying them extra hours, as they have the effect of increasing pacients, but also making workers ill. A bit more of capacity would be appreciated, but tell people that you are going to increase taxes and you are going to loose the elections no matter what.

Exactly. It’s like server analytics. Averages are less useful than percentiles.

The bottleneck is the number of mechanical ventilators available.

Ideally we want someone on a ventilator to be in an ICU and under a doctor's supervision, but in a crisis you can put a ventilator anywhere. A patient with COVID-19, will likely spend about 2 weeks on a ventilator.

Unfortunately, there are only about 60-70,000 mechanical ventilators in the entire US (and maybe double that if you include older less automated ones), and about half of them are for neonatal wards. [source](https://www.ncbi.nlm.nih.gov/pubmed/21149215).

10% of COVID-19 infections require mechanical ventilation. [source](https://clinicaltrials.gov/ct2/show/record/NCT04244591).

This means that if the US encountered an outbreak on the scale of what is happening in China, then we would also quickly run out of ventilators, and the mortality rate would be 2-4x what it would be otherwise, rather than the 1-2% mortality rate everyone is expecting.

Assuming non-containment (starting to look more and more likely), then an unknown fraction of the population will get sick.

The Diamond ~~Princess~~ Petri Dish Experiment, shows that at least 218 out of 3700 people are susceptible ( > 5.9% ).

If we knew how many of these would eventually need respirators then we could estimate a lower bound on respirators for a 330M population.

Working the other way around: 33000 respirators / 330M = 1 / 10 0000

so if 5.9% gets infected and we can only give 0.01% a respirator, then at most 0.17% of covid-positive people can be given a respirator.

This assumes they all need their respirators simultaneously (pessimistic) but also assumes the Petri Disk experiment stays at 5.9% (optimistic)

EDIT: what kind of query did you use to find the number of respirators in the US? How can I find similar numbers for European countries? How hard is it do DIY a respirator?

If we look at the international statistics for dead and serious / critical people per confirmed case, and interpret that as the fraction needing a respirator we get:

22 / 523 = 4.2% of covid-positives seem to need respirators

So we expect 4.2% * 5.9% = 0.25% of the population to neeed a respirator, while we only have enough for 0.01%, so assuming perfect triage of who needs it and who doesnt, 24 out of 25 who actually need one won't get one.

It seems like we really are heading to a CFR of 4.2% * 24 / 25 such that there will be at least > 5.9% * 4.2% * 24 / 25 deaths, where the 5.9% can only rise depending on Diamond Princess evolution...

Currently 355 of 3700 Diamond Petri Dish confirmations, so replace 5.9% with 9.6%

so if containment fails it looks like 4.2% * 9.6% = ~ 0.4% of the population at large would die.

Agreed. It seems to me there are, in the main, two problems which include many other problems relating to the general problem.

Or just cut the budgets.

> I have been following this closely as I follow JS libraries

I feel like I'm missing something. What does this have to do with javascript libraries? Or does the JS stand for something else?

If I had to guess, they are just a JS-library-following nerd :) the extent of which they have been following this topic might resonate with more people here through this fun comparison.

It goes without saying, that this is meant without any offense whatsoever

I think he meant to type "I have been following this as closely as I follow JS libraries".

@tommit's explanation is correct but I see how my typo might have caused the confusion. I didn't notice it until @Nr7's comment.

Watching this closely based in Singapore, the true mortality rate is probably an order of magnitude lower

a) The stats coming from Hubei are adverse selected; they are from patients who already had severe enough illness to go to the hospital

b) A vast amount of mild cases are not diagnosed or tested, this is for reasons related to self reporting and minimal screening efforts globally

c) The virus spreads rapidly (R0 of 4 in some estimates) and has had 3 months to sweep through Wuhan ; there appears to be a lot of death in Wuhan but that may be because most of the 11 million population was exposed

Maybe I'm confirming a bias, but I believe its true profile is something like a more rapidly moving seasonal flu

They're also selected for people who made it to a hospital. The most frail types, or those who got sick after hospitals were filling up, could have died without ever getting to a hospital, and not even be counted yet.

Agreed but even if you think the real Wuhan death count is 10x the official (ie. ~10K deaths in Wuhan), there's still an order of magnitude lower mortality rate assuming there's been widespread exposure in the 3 months

We don’t know the actual death toll in Wuhan. The number we get is people who were diagnosed and then were determined to have died of the disease. What we’d really like to know is simply how many people have died of any cause, and then subtract a base rate. The total deaths figure is not being provided.

The number of deaths is not being accurately reported either, so both the numerator and denominator are inaccurate.

There are mixed reports coming out of China that claim 60% of bodies are collected from homes, not hospitals, and that only 20% of the dead are being counted as "confirmed" COVID-19.

Also, we do know reliably that about 10% of cases overall, require ICU & mechanical ventilation. I would not characterize that as "vast amount of mild cases"

The extreme steps taken by the Chinese government (which can be assumed to have the most accurate -classified - data), belies that belief. You don't shut down so much of your country over a somewhat more severe case of flu.

Two possibilities,

a) a faster spreading version of flu, which produces a greater overall quantity of severe cases, could be mistaken for a more virulent flu... which is what I believe most reports are doing now

b) the unprecedented lock down is because of the origins of the virus

Another possibility: the subsequent ARDS is catastrophic in smokers. In China, 41% of people smoke.

Good point. Is there a statistics about mortality in smokers vs non-smokers?

This is what I've seen so far:

https://www.preprints.org/manuscript/202002.0051/v1

There are also other hints that line up with this. The mortality rate in the recent 1000+ case study seems to be much higher in males, but the rate of smoking in China is also much higher in males, which could be a clue.

https://jech.bmj.com/content/71/2/154

It looks to me that when medical system is not overwhelmed then the outcome for patients does not look as grim as in Hubei. I think that we will have more clear picture of the situation in coming weeks the cases outside of Hubei are more recent.

> a) The stats coming from Hubei are adverse selected; they are from patients who already had severe enough illness to go to the hospital

I don't think this assumption holds true. In a panic, many people are trying to get diagnosed based on rightfully held fear. These are people who have the energy to travel to many hospitals and stand in lines for hours. In some places, people who are seemingly symptomatic are also encouraged/forced to be diagnosed despite not feeling the need to go to a hospital.

A lot of discussion seems to highlight the problem of unreported cases, but surely this is a problem for all diseases, not just novel ones.

Are there any epidemiologists here to answer how we account for unreported cases of e.g. flu in mortality rates? And if we don't, is it reasonable to ignore unreported cases of coronavirus as well, to compare like with like, or is there some factor I'm missing?

That is correct, though I'm not an epidemiologist. The number is estimated using statistical models, such as SEIR and others. http://www.public.asu.edu/~hnesse/classes/seir.html

Not an expert here but isn’t there an unknown in the ratio of reported cases too? I doubt it’s safe to assume the percentage of flu cases reported is anywhere close to that of publicized viruses.

The more I read about this, the more I wonder about the reliability of mortality rates generally. Or at least, for diseases where diagnosis depends on DNA-based testing.

I mean, it seems likely that pretty much everyone who dies will be tested. Except for victims of auto accidents or whatever.

But it also seems likely that many with minimal symptoms won't be tested. Perhaps because hospitals are so busy that people can't be bothered. Or maybe because people are afraid of hospitals, because they're full of sick people. Or simply because there's inadequate testing capability.

So isn't it likely that mortality will always be overestimated for such specific diseases? As opposed to, for example, auto accidents or falls.

From what I've read it's highly unlikely the dead will be tested. Just cremate them and move on.

At least, that's what it seems like with all the tales of death certificates with cause of death as "Unknown viral pneumonia"

> From what I've read it's highly unlikely the dead will be tested.

Not only that, but are the Chinese authorities going to accurately report that numbers?

At this point, I'm not sure to see any reason for them to lie about it except for the force of habit.

Another reason would be that on a day-to-day basis it's easier to say "yeah we had 100 deaths yesterday" than it is to say "we've been lying, there were 1,000 deaths yesterday and 10,000 so far with 500,000 infected in Wuhan by our estimates". Which would probably also cause a wee bit of panic, too.

If that's actually the case, of course.

That’s why they usually use a statistical model to estimate how many people are infected with a virus rather than only using the number of confirmed cases.

> I mean, it seems likely that pretty much everyone who dies will be tested. Except for victims of auto accidents or whatever.

Availability of test kits is the bottleneck for accurate statistics, and they aren't being used on the dead because there are many more living people who are symptomatic and need them.

Really? I mean, there are so many fewer dead than potentially infected who may just have a cold.

But if that's true, it just confuses things more.

If there's a thousand people dying of pneumonia every day in Wuhan and you have capacity to test 2,000 people, do you spend half of that capacity on the dead? Or do you try and test the "maybe" cases who might or might not be infected and shouldn't be put in a ward with whole bunch of infected people?

> it seems likely that pretty much everyone who dies will be tested.

Anecdotal reports out of China estimate that less than half of the dead are taken from hospitals.

The ongoing nightmare of the cruise ship in japan will provide a worst case unbiased example not tainted by the reliability of Chinese reporting. I just hope that most of them will escape what looks to me like my personal nightmare.

Agreed. The quarantined cruise ship is, sadly, an excellent inadvertent experiment for us.

It won't help much with the R0 calculation, but it will give us a good idea for mortality, though it might take 3 weeks to get that number.

Does this website not look a little suspicious ?

It is hard to find an actual name of a person behind this site. When I go to faq -> "Who .." there is another company called "Dadax" listed, but no link. odd

I am asking because people have sent me other suspicious sites over the last few weeks (wuflu.live).

I am not denying the virus, just questioning things on the internet.

I am loath to use my real name on absolutely anything I put online. ...and frankly, I do not understand people that do.

There are crazy people on the internet, and many of them love to hold a grudge.

The initial death rates were shocking. The reproduction R0 number seemed high. This was a very contagious virus. The world went into a panic. And rumors, conspiracy theories, and bad science didn’t help (the German and Indian papers).

But what was really behind the numbers. Was it old people that were dying? Was it middle aged and young people? Was this some kind of stealthy virus that creeps up on you, and then just kills you?

It’s possible that it was a lack of medical facilities and equipment, that contributed to the crisis, and thus, contributed to the high initial death rates.

For example, when people started getting sick. The first batch went to the ICU, and took up available beds and medical facilities. These people had the highest chance of surviving. But then, the next wave came in and created the beginning of the crisis, and swamped the system. Then, the third wave came in and totally overwhelmed the system, that the hospitals couldn’t help them at all. They were turned away, and left to fend for themselves. These people started dying in their homes.

The system suffered such a massive denial of service attack, that it was impossible to help and provide service to everyone. Thus this created the shockingly high mortality rate during the initial days of the crisis.

If we are lucky, then this might remain true. So in other parts of the world with more medical facilities, knowing this might save people, and give them a higher chance of surviving this virus.

Stay strong, Wuhan!

Wuhan, Jaiyou!

For "realtime" numbers there is a chinese web-page which shows the mortality rate is of the order 2% and the new infections in Hubei are decreasing. (Google translate is your friend.)

https://ncov.dxy.cn/ncovh5/view/pneumonia

I'm not an expert by any means but I suspect that the Mortality rate is lower than what they have said.

We've been told that the majority of people will get very mild symptoms if they have it. So the actual number of people that have been infected is way higher since many will never even think of going to a doctor.

So the 2-3% case fatality rate is higher than the actual number. If the number of deaths is correct or close to correct.

The scary part is that the virus is so easily spread between humans so the number of deaths will skyrocket when compared to other epidemics. Let's hope they bring it under control soon and there's a vaccine in the near future.

The problem here will not be the virus, but the healthcare system capacity (I.e. the hospitals capacity).

I.e. If you managed to get care, the mortality rate is X, but without care, it is 10X or even more.

So to judge the current mortality rate, you have to know how many of the infected received care.

I also assume that hospitals are like the phone system. I.e. they do not expect mass load at the same time.

That would be a very valuable number. What is the actual number of people that need to be hospitalized for this virus vs other epidemics?

Based on what I read (mainly a very good analysis of first patient in the US), I think 100%, as everyone would need oxygen, as the virus infect the lungs (even with no preconditions).

And this is on a healthy, young patient.

https://www.nejm.org/doi/full/10.1056/NEJMoa2001191

No, not everyone who contracts this virus ends up with pneumonia and needing a respirator.

But it’s still a sufficient proportion to serve as a denial-of-service attack on hospitals, if the virus spreads throughout the community.

The scary part is "acute cardiac injury" being a frequent complication. The virus causes heart attacks!

https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

"Complications included acute respiratory distress syndrome (12 [29%]), RNAaemia (six [15%]), acute cardiac injury (five [12%]) and secondary infection (four [10%]). 13 (32%) patients were admitted to an ICU and six (15%) died."

Believe it or not, this can happen infrequently with "ordinary" viruses: viral myocarditis (inflammation and infection of the heart muscle) is an under-appreciated killer of the young, with a mortality rate of between 25-50% [1] and is a well-known complication of, for example, influenza [2]. In fact, it's believed that myocarditis was involved in as much as 50% of Spanish flu fatalities in the early part of the 20th century. The difference is that if you're young and healthy, you're likely to clear infections well before they get to the point of systemically infecting each organ.

In the population of those who are acutely unwell with this novel virus, you have to remember that there is a massive selection bias for those hospital doctors to see the sickest people. Myocarditis is unfortunately one of those things that is therefore probably to some degree expected. The real question is "what is the probability that you end up in this terrible situation given you are infected with coronavirus" -- and in order to do that, you need a good estimate of how many people have been infected (but not died or seriously ill!) with the novel coronavirus. This is very hard to do in Wuhan, and estimates vary by orders of magnitude.

[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3370379/ [2] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3533457/

(from the article) 80% of people who died were over 60 years old and 75% had underlying health conditions including cardiovascular problems, so the fact that the virus caused fatal cardiovascular events in some of those cases is not unexpected.

About 80% of Americans over 60 have cardiovascular problems, compared to a third of elderly Chinese people. That stat indicates that fatality rates could be higher in the US (and other western countries, although I don't have those numbers immediately) than China.

https://www.heart.org/idc/groups/heart-public/@wcm/@sop/@smd...

https://www.jwatch.org/na48562/2019/03/05/china-cardiovascul...

These are from earlier small scale studies, I believe this study has much larger cohort (1000+ patients)

https://www.medrxiv.org/content/10.1101/2020.02.10.20021675v...

The TL;DR;

Initial estimates of mortality rate are all over the place due to uncertainty in the data sets.

That said, the best estimate at the moment is around 2-3% and is subject to change as more data comes in.

Yes. By comparison:

--- Mortality rate ---

1918 Spanish Flu: 10-20%

2020 Coronavirus: 2-3% est.

1968 Hong Kong Flu: 0.5%

--------

The realization that China censors heavily makes both accurate, timely and complete reporting of facts much more difficult. Intermediaries are used to exfil data to Western journalists, but this is haphazard, slow and incomplete. A journalist who has contacts on the ground described the situation, including that a few bodies on the streets and in crashed vehicles (people who were too sick who failed to make it to a hospital) in central Wuhan were not being collected: https://www.democracynow.org/2020/2/7/laurie_garrett_china_c... We may never know the exact numbers because it's likely face-saving deception will be used for geopolitical and local political concerns.

In the US, Spanish Flu mortality was ~2%

It was far worse elsewhere; figures vary widely.

https://en.wikipedia.org/wiki/Spanish_flu#Around_the_globe

I was looking up the current flu season in the US here: https://www.cdc.gov/flu/weekly/index.htm

It says the mortality rate is 7.1%. Does that make the current flu in the US much worse than the Coronavirus?

I would look at those numbers again. It seems to be the percentage of all deaths.

Based on National Center for Health Statistics (NCHS) mortality surveillance data available on February 6, 2020, 7.1% of the deaths occurring during the week ending January 25, 2020 (week 4) were due to P&I. This percentage is below the epidemic threshold of 7.2% for week 4.

If you like, every year the flu is a low-level "pandemic" of a relative sort. It goes almost unnoticed like antibiotic/antimycotic-resistant pathogens that are rapidly becoming untreatable due to a combination of factors including over-prescription, meat agriculture, human over-population and low-cost travel.

Oddly enough, I used to work at a biomedical informatics department where one of the projects used commercial data from retail stores to predict and identify bioweapon exposure and pandemics from people buying OTC remedies and prescriptions.

That seems pretty bad, no?

I don’t know how to interpret this that’s why I’m asking.

It seems like the media is captivated by what seems like a relatively small number of deaths in China compared to the 12,000+ here at home.

A Chinese supplier of mine actually pointed this out when he told me to “stay safe” and I was like what the hell is he talking about, but then I googled it and now I’m like wtf.

In late January, Chinese media widely reported "news" about an American flu "epidemic" to distract from nCoV. However, the US CDC reports pneumonia and influenza deaths as a percent of overall mortality on a weekly basis, and right now the numbers are shy of what they consider an epidemic. The rate of lab confirmed cases is tracking 2012-2013, and mortality looks about the same as last year.

The current nCoV outbreak is more worrying because the percentage of patients needing intensive care is very high, and because it looks very easy to spread.

That's not the mortality rate but the ratio against overall deaths. it says: "The percentage of deaths attributed to pneumonia and influenza is 7.1%, below the epidemic threshold of 7.2%."

You beat me to it. (Sorry to the GP for seeming to dogpile!)

Ha, right. The completely standard and obvious threshold of 7.2% to declare an epidemic...

Seems like the number is chosen so that you can say an epidemic is anything worse than the flu?

I was just quoting, but also wondered about that seemingly random threshold. Looks like a crude measure

The mortality rate, as calculated by trivial formula, has stayed close to 2% for a couple of weeks now.

It's interesting that MERS was about twice as deadly as the Spanish Flu. I guess because it happened in a certain part of the world, the Western media just decided it wasn't worthy of wall-to-wall Snowpocalypse yellow journalism.

MERS had apparently a low infectivity, making it a low level threat to the rest of the world.

It's difficult to say that because it was still transmitted to people more easily than Ebola. And the data and research show MERS isn't going away; bats maybe its natural reservoir. Pretending it doesn't exist or not doing anything about is playing Russian roulette with public health. If the US Army Corps of Engineers can eliminate malaria from the southern US, maybe a vaccine can be developed for the natural host and domesticated animals to prevent the risk to humans.

Threat from a disease is a combination of tendency to spread, circumstance of initial outbreak, and deadliness. The latter alone isn't what makes it a matter of urgent global concern on a simple pragmatic level.

That's obvious. It's more contagious than Ebola but less so than Norovirus. You're discounting that MERS hasn't gone away... it's sitting in reservoir(s) of wild and/or domesticated animals right now, likely on two continents thanks to travel and clusters of human cases in South Korea as well. If that doesn't worry you, then I guess anticipation isn't your strong suit.

The point is: hantavirus was a rare disease but was hyped up in North America, MERS is much worse and has the potential to be a minor Ebola again and again.

Reply on news.ycombinator.com